gmm/0000755000176200001440000000000015053352522011034 5ustar liggesusersgmm/MD50000644000176200001440000000620415053352522011346 0ustar liggesusers77d360e49cce21b345b43bdbe3f5c541 *DESCRIPTION 8dbf0a1bf9e2ef0907845214e724577f *NAMESPACE 6372ea458199fe4e15b03dadd9f7289e *NEWS 34060d64cf14b2bc3ed236bf761fe875 *R/FinRes.R 8d15734187f2b41f1bc52c42213564ec *R/Methods.gel.R dec5ef37c3702d9603e2ed4401246664 *R/Methods.gmm.R 609683327c9bdc00a30420873939e865 *R/Methods.sysGmm.R f7fec7a4d5ec4e545ec61b3f3675e7d1 *R/PlotMethods.R 989f6a73d773a4113d55c6e80f04f845 *R/ategel.R 6edab88cccacdc9a87dcb228a934d699 *R/bandwidth.R 00a45548f548138d30c5f7c660ffdbbf *R/charStable.R d504335bd50c16265c02d128b3445601 *R/gel.R 0a3584ef441c1d563ada2540b2589001 *R/getModel.R fb8fac9af83a75d64a8364c75680537c *R/gmm.R e50bb0d4f9c06ef3bcaf29a6ff9bcf62 *R/gmmTests.R d04bf27581e1bc1f1a305a83b2d4209a *R/momentEstim.R f175f4ff505f6edea1b2e97a2a45d72e *R/specTest.R 5c43d9f22c0c1c9f028319f57004d8a4 *R/sysGmm.R 1bf18d9c216b8d95033f0bd2fe2906b2 *build/partial.rdb e72fe4d2599b2c67a08f3bdc438f8363 *build/vignette.rds d01888b9a5f72d995bfed87ede69194c *data/Finance.rda a5b7324d7f0bfcabe5a91b60a5f2709a *data/Growth.rda 1b3001ceb0de2c2bfce1720ca8d16a22 *data/nsw.rda 6dddfac8f53994832cce85697b6b731a *data/wage.rda 70ec1dc7a37c75468cd81bc9112cbfcd *inst/CITATION 8e3ba12a04239f738f6d4f2ae5a782d7 *inst/doc/gmm_with_R.R 3e9fe3a0e1f9f739f49c6fcde5ce32a1 *inst/doc/gmm_with_R.pdf 379d4dc7881ea419410aa5abe8308745 *inst/doc/gmm_with_R.rnw 827934d3cf90f600c67a991bd5f6ce8f *man/ATEgel.Rd 5e37a93a689634ee38cd7fdd177c2827 *man/FinRes.Rd a91a58e7472a68ca02c9cea14055a9c7 *man/Finance.Rd f3672fdf3a8e1575afeebd6cf1269744 *man/KTest.Rd 83dc94b87a175ae526b811f60321897a *man/bread.Rd 14c5b9c9d8f776721210e48bbfd1005e *man/bwWilhelm.Rd 14fe0fdc342559be4e3f371835ef8656 *man/charStable.Rd af701418d7b6f2bb0360e0d12311d3cf *man/coef.Rd 0f7e154c11fe4b909e75524831c431d7 *man/confint.Rd c4cd9a9027c011355eca2a34a2569140 *man/estfun.Rd e2149d386f6102467f98b95b9e40a634 *man/fitted.Rd 9ab973a6d0f9157e45a3345ecd2aa474 *man/formula.Rd 3dbd92a8872d58560cfdc500f53f0580 *man/gel.Rd 0d586d2638432c60dcf1c0f3ffd0b02e *man/getDat.Rd 47577cf99ff1539239775621e439d88e *man/getImpProb.Rd 34154c9016bac2e63a0386e4024b4e46 *man/getLamb.Rd a3e02ce173d3b2f6fb5a73f67788eca2 *man/getModel.Rd 9eff861ac07a92778edccabcbc1cf858 *man/gmm.Rd c4df6739d8fb7cf62cddbf2df1c21f42 *man/growth.Rd 512a3c4267f6146e373a1bdd6bc3a9e5 *man/marginal.Rd 19c918bb116d31ce090dfe3109d5e76e *man/momentEstim.Rd 579a6cb97e9ff6bb4000e1ffe24b2236 *man/nsw.Rd 9041105be7d299cb1059ccb3a38b9a6a *man/plot.Rd ce22b21d434e85b31242ce58e0c976ec *man/print.Rd 05c3e41ccab1554d745de9c899f638c5 *man/residuals.Rd eb57b92e2aebae7773c84b0c6e42bb12 *man/smoothG.Rd 52588d1b32001f7a264674c50dc1ec1a *man/specTest.Rd bbeb210028428733ce829e5d9620e94b *man/summary.Rd ef34708f39759290653bc2d8b5dbddf5 *man/sysGmm.Rd 84958cc9840247ffbfdc2e34a3c4a8f6 *man/tsls.Rd f9861b90cefceafa0fd7f65157ae05c9 *man/vcov.Rd 63c2c7e4c4083a16361d4c3949a73807 *man/wage.Rd 3996e7c16bfb96fad295ee425815cb4d *src/Makevars 27461960ddf6e06cf7e55e51cfeea7f5 *src/gmm.h 16f0236c9e73143592e6d9ddd5ce08f7 *src/lambda_met.f c1c162760356490e4bf0119e0682ea0e *src/src.c e1cbec0189aa6e4ce46551f005822730 *vignettes/empir.bib 379d4dc7881ea419410aa5abe8308745 *vignettes/gmm_with_R.rnw gmm/R/0000755000176200001440000000000015025245305011234 5ustar liggesusersgmm/R/FinRes.R0000644000176200001440000001776314247643115012571 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ FinRes <- function(z, object, ...) { # object is computed by the getModel method # UseMethod("FinRes") } FinRes.baseGmm.res <- function(z, object, ...) { P <- object x <- z$dat n <- ifelse(is.null(nrow(z$gt)),length(z$gt),nrow(z$gt)) if (!is.null(attr(x,"eqConst")) & P$allArg$eqConstFullVcov) { eqConst <- attr(x,"eqConst")$eqConst coef <- rep(0,length(eqConst[,1])+length(z$coefficients)) ncoef <- rep("",length(eqConst[,1])+length(z$coefficients)) coef[-eqConst[,1]] <- z$coefficients ncoef[-eqConst[,1]] <- names(z$coefficients) coef[eqConst[,1]] <- eqConst[,2] ncoef[eqConst[,1]] <- rownames(eqConst) names(coef) <- ncoef z$coefficients <- coef if (!is.null(z$initTheta)) { initTheta <- rep(0,length(z$coefficients)) initTheta[eqConst[,1]] <- eqConst[,2] initTheta[-eqConst[,1]] <- z$initTheta z$initTheta <- initTheta } z$k <- z$k+nrow(eqConst) z$k2 <- z$k2+nrow(eqConst) attr(x, "eqConst") <- NULL z$specMod <- paste(z$specMod, "** Note: Covariance matrix computed for all coefficients based on restricted values \n Tests non-valid**\n\n") } z$G <- z$gradv(z$coefficients, x) G <- z$G if (P$vcov == "TrueFixed") v <- .weightFct(z$coefficient, x, "fixed") else v <- .weightFct(z$coefficient, x, P$vcov) z$v <- v if (P$vcov == "TrueFixed") { z$vcov=try(solve(crossprod(G, P$weightsMatrix) %*% G)/n, silent = TRUE) if(any(class(z$vcov) == "try-error")) { z$vcov <- matrix(Inf,length(z$coef),length(z$coef)) warning("The covariance matrix of the coefficients is singular") } } else if ( (is.null(P$weightsMatrix)) & (P$wmatrix != "ident") ) { if (dim(G)[1] == dim(G)[2]) { T1 <- try(solve(G), silent=TRUE) z$vcov <- try(T1%*%v%*%t(T1)/n, silent=TRUE) } else { z$vcov <- try(solve(crossprod(G, solve(v, G)))/n, silent = TRUE) } if(any(class(z$vcov) == "try-error")) { z$vcov <- matrix(Inf,length(z$coef),length(z$coef)) warning("The covariance matrix of the coefficients is singular") } } else { if (is.null(P$weightsMatrix)) w <- diag(ncol(z$gt)) else w <- P$weightsMatrix if (dim(G)[1] == dim(G)[2]){ T1 <- try(solve(G), silent=TRUE) } else { T1 <- try(solve(t(G)%*%w%*%G,t(G)%*%w), silent = TRUE) } if(any(class(T1) == "try-error")) { z$vcov <- matrix(Inf,length(z$coef),length(z$coef)) warning("The covariance matrix of the coefficients is singular") } else { z$vcov <- T1%*%v%*%t(T1)/n } } dimnames(z$vcov) <- list(names(z$coefficients), names(z$coefficients)) z$call <- P$call if(is.null(P$weightsMatrix)) { if(P$wmatrix == "ident") { z$w <- diag(ncol(z$gt)) } else { z$w <- try(solve(v), silent = TRUE) if(any(class(z$w) == "try-error")) warning("The covariance matrix of the moment function is singular") } } else { z$w <- P$weightsMatrix } z$weightsMatrix <- P$weightsMatrix z$infVcov <- P$vcov z$infWmatrix <- P$wmatrix z$allArg <- P$allArg if (P$wmatrix=="ident") z$met <- "One step GMM with W = identity" else z$met <- P$type z$kernel <- P$kernel z$coefficients <- c(z$coefficients) class(z) <- "gmm" return(z) } FinRes.sysGmm.res <- function(z, object, ...) { P <- object x <- z$dat z$G <- z$gradv(z$coefficients, x) n <- z$n G <- z$G v <- .weightFct_Sys(z$coefficient, x, P$vcov) nk <- z$k z$v <- v if (P$vcov == "TrueFixed") { z$vcov=try(solve(crossprod(G, P$weightsMatrix) %*% G)/n, silent = TRUE) if(any(class(z$vcov) == "try-error")) { z$vcov <- matrix(Inf,nk,nk) warning("The covariance matrix of the coefficients is singular") } } else if ( (is.null(P$weightsMatrix)) & (P$wmatrix != "ident") ) { z$vcov <- try(solve(crossprod(G, solve(v, G)))/n, silent = TRUE) if(any(class(z$vcov) == "try-error")) { z$vcov <- matrix(Inf,nk,nk) warning("The covariance matrix of the coefficients is singular") } } else { if (is.null(P$weightsMatrix)) w <- .weightFct_Sys(z$coefficients, x, "ident") else w <- P$weightsMatrix if (dim(G)[1] == dim(G)[2]){ T1 <- try(solve(G), silent=TRUE) } else { T1 <- try(solve(t(G)%*%w%*%G,t(G)%*%w), silent = TRUE) } if(any(class(T1) == "try-error")) { z$vcov <- matrix(Inf, nk, nk) warning("The covariance matrix of the coefficients is singular") } else { z$vcov <- T1%*%v%*%t(T1)/n } } if (attr(x, "sysInfo")$commonCoef) dimnames(z$vcov) <- list(P$namesCoef[[1]], P$namesCoef[[1]]) else dimnames(z$vcov) <- list(do.call(c, P$namesCoef), do.call(c, P$namesCoef)) z$call <- P$call if(is.null(P$weightsMatrix)) { if(P$wmatrix == "ident") { z$w <- .weightFct_Sys(z$coefficient, x, "ident") } else { z$w <- try(solve(v), silent = TRUE) if(any(class(z$w) == "try-error")) warning("The covariance matrix of the moment function is singular") } } else { z$w <- P$weightsMatrix } for (i in 1:length(z$coefficients)) names(z$coefficients[[i]]) <- P$namesCoef[[i]] z$weightsMatrix <- P$weightsMatrix z$infVcov <- P$vcov z$infWmatrix <- P$wmatrix z$allArg <- P$allArg z$met <- paste("System of Equations: ", P$type, sep="") z$kernel <- P$kernel z$coefficients <- c(z$coefficients) class(z) <- c("sysGmm", "gmm") return(z) } gmm/R/charStable.R0000644000176200001440000000401214247643115013432 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ charStable <- function(theta, tau, pm = 0) { # pm is the type parametrization as described by Nolan(2009) # It takes the value 0 or 1 # const can fixe parameters. It is NULL for no constraint or # a matrix in which case the constraint is theta[const[,1]]=const[,2] a <- theta[1] b <- theta[2] g <- theta[3] d <- theta[4] if(pm == 0) { if(a == 1) { if(g == 0) { the_car <- exp(complex(imaginary = d*tau)) } else { re_p <- -g * abs(tau) im_p <- d * tau im_p[tau!=0] <- im_p[tau != 0] + re_p[tau != 0]*2/pi*b*sign(tau[tau != 0])*log(g*abs(tau[tau != 0])) the_car <- exp(complex(real = re_p, imaginary = im_p)) } } else { if(g == 0) { the_car <- exp(complex(imaginary = d*tau)) } else { phi <- tan(pi*a/2) re_p <- -g^a*abs(tau)^a im_p <- d*tau*1i im_p[tau != 0] <- im_p[tau != 0] + re_p*( b*phi*sign(tau[tau != 0])*(abs(g*tau[tau != 0])^(1-a) - 1) ) the_car <- exp(complex(real = re_p, imaginary = im_p)) } } } if(pm == 1) { if(a == 1) { re_p <- -g*abs(tau) im_p <- d*tau im_p[tau!=0] <- im_p[tau != 0]+re_p*(b*2/pi*sign(tau[tau != 0])*log(abs(tau[tau!=0]))) the_car <- exp(complex(real = re_p, imaginary = im_p)) } else { phi <- tan(pi*a/2) re_p <- -g^a*abs(tau)^a im_p <- re_p*(-phi*b*sign(tau)) + d*tau the_car <- exp(complex(real = re_p, imaginary = im_p)) } } return(the_car) } gmm/R/PlotMethods.R0000644000176200001440000000611314247643115013630 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ plot.gmm <- function (x, which = c(1L:3), main = list("Residuals vs Fitted values", "Normal Q-Q", "Response variable and fitted values"), panel = if(add.smooth) panel.smooth else points, ask = prod(par("mfcol")) < length(which) && dev.interactive(), ..., add.smooth = getOption("add.smooth")) { if (!inherits(x, "gmm")) stop("use only with \"gmm\" objects") if (!inherits(x, "gel")) { if(!is.numeric(which) || any(which < 1) || any(which > 3)) stop("'which' must be in 1L:3") show <- rep(FALSE, 3) } else { if(!is.numeric(which) || any(which < 1) || any(which > 4)) stop("'which' must be in 1L:4") show <- rep(FALSE, 4) } show[which] <- TRUE r <- residuals(x) if(ncol(r)>1) stop("plot.gmm is not yet implemented for system of equations") yh <- fitted(x) n <- length(r) if (ask) { oask <- devAskNewPage(TRUE) on.exit(devAskNewPage(oask)) } ##---------- Do the individual plots : ---------- if (show[1L]) { ylim <- range(r, na.rm=TRUE) ylim <- extendrange(r= ylim, f = 0.08) plot(yh, r, xlab = "Fitted", ylab = "Residuals", main = main[1L], ylim = ylim, type = "n", ...) panel(yh, r) abline(h = 0, lty = 3, col = "gray") } if (show[2L]) { ## Normal rs <- (r-mean(r))/sd(r) ylim <- range(rs, na.rm=TRUE) ylim[2L] <- ylim[2L] + diff(ylim) * 0.075 qq <- qqnorm(rs, main = main[2L], ylab = "stand. residuals", ylim = ylim, ...) qqline(rs, lty = 3, col = "gray50") } if (show[3L]) { y <- as.matrix(model.response(x$model, "numeric")) ylim <- range(yh, na.rm=TRUE) ylim <- extendrange(r= ylim, f = 0.08) plot(y, main = main[3L], ylim = ylim, ...) lines(yh,col=2) } if (inherits(x, "gel")) { if (show[4L]) { pt <- x$pt plot(pt, type='l',main = main[4L],ylab="Implied Prob.", ...) emp_pt <- rep(1/length(pt),length(pt)) lines(emp_pt,col=2) legend("topleft",c("Imp. Prob.","Empirical (1/T)"),col=1:2,lty=c(1,1)) } } invisible() } plot.gel <- function (x, which = c(1L:4), main = list("Residuals vs Fitted values", "Normal Q-Q", "Response variable and fitted values","Implied probabilities"), panel = if(add.smooth) panel.smooth else points, ask = prod(par("mfcol")) < length(which) && dev.interactive(), ..., add.smooth = getOption("add.smooth")) { if (!inherits(x, "gel")) stop("use only with \"gel\" objects") class(x) <- c("gmm","gel") plot(x,which=which,main=main,panel=panel, ask=ask, ..., add.smooth=add.smooth) invisible() } gmm/R/gel.R0000644000176200001440000005103514247643115012140 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ .rho <- function(x, lamb, derive = 0, type = c("EL", "ET", "CUE", "HD", "ETEL", "ETHD", "RCUE"), k = 1) { type <- match.arg(type) if (type == "RCUE") type <- "CUE" gml <- x%*%c(lamb)*k if (derive == 0) { if (type == "EL") { if (any(gml>=1)) stop("Computation of Lambda fails because NAs produced by log(1-gt*l)") rhomat <- log(1 - gml) } if (type == "ET") rhomat <- -exp(gml) if (type == "CUE") rhomat <- -gml -0.5*gml^2 if (type == "HD") rhomat <- -2/(1-gml/2) if (type == "ETEL") { w <- -exp(gml) w <- w/sum(w) rhomat <- -log(w*NROW(gml)) } if (type == "ETHD") { w <- -exp(gml) w <- w/sum(w) rhomat <- (sqrt(w)-1/sqrt(NROW(gml)))^2 } } if (derive==1) { if (type == "EL") rhomat <- -1/(1 - gml) if (type == "ET") rhomat <- -exp(gml) if (type == "CUE") rhomat <- -1 - gml if (type == "HD") rhomat <- -1/((1-gml/2)^2) if (any(type == c("ETEL","ETHD"))) rhomat <- NULL } if (derive==2) { if (type == "EL") rhomat <- -1/(1 - gml)^2 if (type == "ET") rhomat <- -exp(gml) if (type == "CUE") rhomat <- -rep(1,nrow(x)) if (type == "HD") rhomat <- -1/((1-gml/2)^3) if (any(type == c("ETEL","ETHD"))) rhomat <- NULL } return(c(rhomat)) } .getCgelLam <- function(gt, l0, type = c('EL', 'ET', 'CUE', "HD"), method = c("nlminb", "optim", "constrOptim"), control=list(), k = 1, alpha = 0.01) { type <- match.arg(type) method <- match.arg(method) fct <- function(l, X) { r1 <- colMeans(.rho(gt,l,derive=1,type=type,k=k)*X) crossprod(r1) + alpha*crossprod(l) } Dfct <- function(l, X) { r2 <- .rho(X,l,derive=2,type=type,k=k) r1 <- .rho(X,l,derive=1,type=type,k=k) H <- t(X*r2)%*%X/nrow(X) 2*H%*%colMeans(r1*X) + 2*alpha*l } if (method == "nlminb") res <- nlminb(l0, fct, Dfct, X = gt, control = control) if (method == "optim") res <- optim(l0, fct, Dfct, X = gt, method="BFGS", control = control) if (method == "constrOptim") { ci <- -rep(1,nrow(gt)) res <- constrOptim(rep(0,ncol(gt)),fct,Dfct,-gt,ci,control=control,X=gt) } if (method == "optim") { conv <- list(convergence = res$convergence, counts = res$counts, message = res$message) } else { conv <- list(convergence = res$convergence, counts = res$evaluations, message = res$message) } return(list(lambda = res$par, convergence = conv, obj = mean(.rho(gt,res$par, derive=0,type=type,k=k)) - .rho(1,0, derive=0,type=type,k=k))) } .Wu <- function(gt, tol_lam = 1e-8, maxiterlam = 50, K=1) { u <- as.matrix(gt) n=length(nrow(u)) M=rep(0,ncol(u)) dif=1 tol=tol_lam k=0 while(dif>tol & k<=maxiterlam) { D1=t(u)%*%((1/(1+u%*%M))*rep(1,n)) DD=-t(u)%*%(c((1/(1+u%*%M)^2))*u) D2=solve(DD,D1,tol=1e-40) dif=max(abs(D2)) rule=1 while(rule>0) { rule=0 if(min(1+t(M-D2)%*%t(u))<=0) rule=rule+1 if(rule>0) D2=D2/2 } M=M-D2 k=k+1 } if(k>=maxiterlam) { M=rep(0,ncol(u)) conv = list(convergence=1) } else { conv = list(convergence=0) } return(list(lambda = c(-M), convergence = conv, obj = mean(.rho(gt,-M,derive=0,type="EL",k=K)))) } .Wu2 <- function(gt, tol_lam = 1e-8, maxiter = 50, K = 1) { gt <- as.matrix(gt) res <- .Fortran(F_wu, as.double(gt), as.double(tol_lam), as.integer(maxiter), as.integer(nrow(gt)), as.integer(ncol(gt)), as.double(K), conv=integer(1), obj=double(1), lambda=double(ncol(gt))) list(lambda=res$lambda, convergence=list(convergence=res$conv), obj = res$obj) } .CUE_lam <- function(gt, K=1) { q <- qr(gt) n <- nrow(gt) l0 <- -qr.coef(q, rep(1,n)) conv <- list(convergence=0) list(lambda = l0, convergence = conv, obj = mean(.rho(gt,l0,derive=0,type="CUE",k=K))) } .CUE_lamPos <- function(gt, K=1, control=list()) { getpt <- function(gt,lam) { gl <- c(gt%*%lam) pt <- 1 + gl pt/sum(pt) } maxit <- ifelse("maxit" %in% names(control), control$maxit, 50) res <-.CUE_lam(gt, K) n <- nrow(gt) i <- 1 pt <- getpt(gt, res$lambda) w <- pt<0 while (TRUE) { gt2 <- gt[!w,] n1 <- nrow(gt2) if (n1 == n) break res <- try(.CUE_lam(gt2), silent=TRUE) if (i > maxit) return(list(lambda=rep(0,ncol(gt)), obj=0, pt=rep(1/n,n), convergence=list(convergence=1))) if (any(class(res) == "try-error")) return(list(lambda=rep(0,ncol(gt)), obj=0, pt=rep(1/n,n), convergence=list(convergence=2))) pt[!w] <- getpt(gt2, res$lambda) pt[w] <- 0 if (all(pt>=0)) break i <- i+1 w[!w] <- pt[!w]<0 } n0 <- n-n1 res$obj <- res$obj*n1/n + n0/(2*n) res } .CUE_lamPos2 <- function(gt, K=1, control=list()) { gt <- as.matrix(gt) n <- nrow(gt) q <- ncol(gt) maxit <- ifelse("maxit" %in% names(control), control$maxit, 50) res <- try(.Fortran(F_lamcuep, as.double(gt), as.integer(n), as.integer(q), as.double(K), as.integer(maxit),conv=integer(1), lam=double(q),pt=double(n), obj=double(1) ), silent=TRUE) if (any(class(res) == "try-error")) return(list(lambda=rep(0,q), obj=0, pt=rep(1/n,n), convergence=list(convergence=3))) list(lambda=res$lam, obj=res$obj, pt=res$pt, convergence=list(convergence=res$conv)) } getLamb <- function(gt, l0, type = c('EL', 'ET', 'CUE', "ETEL", "HD", "ETHD", "RCUE"), tol_lam = 1e-7, maxiterlam = 100, tol_obj = 1e-7, k = 1, method = c("nlminb", "optim", "iter", "Wu"), control=list()) { method <- match.arg(method) type <- match.arg(type) gt <- as.matrix(gt) if (method == "Wu" & type != "EL") stop("Wu (2005) method to compute Lambda is for EL only") if (method == "Wu") return(.Wu2(gt, tol_lam, maxiterlam, k)) if (type == "CUE") return(.CUE_lam(gt, k)) if (type == "RCUE") return(.CUE_lamPos2(gt, k, control)) if (method == "iter") { if ((type == "ETEL") | (type == "ETHD")) type = "ET" for (i in 1:maxiterlam) { r1 <- .rho(gt,l0,derive=1,type=type,k=k) r2 <- .rho(gt,l0,derive=2,type=type,k=k) F <- -colMeans(r1*gt) J <- crossprod(r2*gt,gt) if (sum(abs(F))= 2) { w <- c(w[rt:2], w) w <- w / sum(w) w <- kernel(w[rt:length(w)]) } else { w <- kernel(1) } } else { if (class(w)[1] != "tskernel") stop("Provided weights must be a numeric vector or an object of class 'tskernel'") } if (length(w$coef)>1) x <- kernapply(x, w) sx <- list("smoothx" = x, "kern_weights" = w, bw = bw) return(sx) } gel <- function(g, x, tet0 = NULL, gradv = NULL, smooth = FALSE, type = c("EL", "ET", "CUE", "ETEL", "HD", "ETHD","RCUE"), kernel = c("Truncated", "Bartlett"), bw = bwAndrews, approx = c("AR(1)", "ARMA(1,1)"), prewhite = 1, ar.method = "ols", tol_weights = 1e-7, tol_lam = 1e-9, tol_obj = 1e-9, tol_mom = 1e-9, maxiterlam = 100, constraint = FALSE, optfct = c("optim", "optimize", "nlminb"), optlam = c("nlminb", "optim", "iter", "Wu"), data, Lambdacontrol = list(), model = TRUE, X = FALSE, Y = FALSE, TypeGel = "baseGel", alpha = NULL, eqConst = NULL, eqConstFullVcov = FALSE, onlyCoefficients=FALSE, ...) { type <- match.arg(type) optfct <- match.arg(optfct) optlam <- match.arg(optlam) weights <- weightsAndrews approx <- match.arg(approx) kernel <- match.arg(kernel) if (!is.null(eqConst)) TypeGel <- "constGel" if(missing(data)) data<-NULL all_args <- list(g = g, x = x, tet0 = tet0, gradv = gradv, smooth = smooth, type = type, kernel = kernel, bw = bw, approx = approx, prewhite = prewhite, ar.method = ar.method, tol_weights = tol_weights, tol_lam = tol_lam, tol_obj = tol_obj, tol_mom = tol_mom, maxiterlam = maxiterlam, constraint = constraint, optfct = optfct, weights = weights, optlam = optlam, model = model, X = X, Y = Y, TypeGel = TypeGel, call = match.call(), Lambdacontrol = Lambdacontrol, alpha = alpha, data = data, eqConst = eqConst, eqConstFullVcov = eqConstFullVcov, onlyCoefficients=onlyCoefficients) class(all_args)<-TypeGel Model_info<-getModel(all_args) z <- momentEstim(Model_info, ...) if (!onlyCoefficients) class(z) <- "gel" return(z) } evalGel <- function(g, x, tet0, gradv = NULL, smooth = FALSE, type = c("EL", "ET", "CUE", "ETEL", "HD", "ETHD","RCUE"), kernel = c("Truncated", "Bartlett"), bw = bwAndrews, approx = c("AR(1)", "ARMA(1,1)"), prewhite = 1, ar.method = "ols", tol_weights = 1e-7, tol_lam = 1e-9, tol_obj = 1e-9, tol_mom = 1e-9, maxiterlam = 100, optlam = c("nlminb", "optim", "iter", "Wu"), data, Lambdacontrol = list(), model = TRUE, X = FALSE, Y = FALSE, alpha = NULL, ...) { type <- match.arg(type) optlam <- match.arg(optlam) weights <- weightsAndrews approx <- match.arg(approx) kernel <- match.arg(kernel) TypeGel <- "baseGel" if(missing(data)) data<-NULL all_args <- list(g = g, x = x, tet0 = tet0, gradv = gradv, smooth = smooth, type = type, kernel = kernel, bw = bw, approx = approx, prewhite = prewhite, ar.method = ar.method, tol_weights = tol_weights, tol_lam = tol_lam, tol_obj = tol_obj, tol_mom = tol_mom, maxiterlam = maxiterlam, weights = weights, optlam = optlam, model = model, X = X, Y = Y, call = match.call(), Lambdacontrol = Lambdacontrol, alpha = alpha, data = data, optfct="optim") class(all_args)<-TypeGel Model_info<-getModel(all_args) class(Model_info) <- "baseGel.eval" z <- momentEstim(Model_info, ...) class(z) <- "gel" return(z) } .thetf <- function(tet, P, output=c("obj","all"), l0Env) { output <- match.arg(output) gt <- P$g(tet, P$x) l0 <- get("l0",envir=l0Env) if (((P$type=="ETEL")|(P$type=="ETHD"))&(!is.null(P$CGEL))) { P$CGEL <- NULL warning("CGEL not implemented for ETEL or for ETHD") } if (is.null(P$CGEL)) { if (P$optlam != "optim" & P$type == "EL") { lamb <- try(getLamb(gt, l0, type = P$type, tol_lam = P$tol_lam, maxiterlam = P$maxiterlam, tol_obj = P$tol_obj, k = P$k1/P$k2, control = P$Lambdacontrol, method = P$optlam), silent = TRUE) if(any(class(lamb) == "try-error")) lamb <- getLamb(gt, l0, type = P$type, tol_lam = P$tol_lam, maxiterlam = P$maxiterlam, tol_obj = P$tol_obj, k = P$k1/P$k2, control = P$Lambdacontrol, method = "optim") } else { lamb <- getLamb(gt, l0, type = P$type, tol_lam = P$tol_lam, maxiterlam = P$maxiterlam, tol_obj = P$tol_obj, k = P$k1/P$k2, control = P$Lambdacontrol, method = P$optlam) } } else { lamb <- try(.getCgelLam(gt, l0, type = P$type, method = "nlminb", control=P$Lambdacontrol, k = P$k1/P$k2, alpha = P$CGEL),silent=TRUE) if (any(class(lamb) == "try-error")) lamb <- try(.getCgelLam(gt, l0, type = P$type, method = "constrOptim", control=P$Lambdacontrol, k = P$k1/P$k2, alpha = P$CGEL),silent=TRUE) } if (P$type == "ETEL") obj <- mean(.rho(gt, lamb$lambda, type = P$type, derive = 0, k = P$k1/P$k2) - .rho(1, 0, type = P$type, derive = 0, k = P$k1/P$k2)) else if (P$type == "ETHD") obj <- sum(.rho(gt, lamb$lambda, type = P$type, derive = 0, k = P$k1/P$k2) - .rho(1, 0, type = P$type, derive = 0, k = P$k1/P$k2)) else obj <- lamb$obj assign("l0",lamb$lambda,envir=l0Env) if(output == "obj") return(obj) else return(list(obj = obj, lambda = lamb, gt = gt)) } .getImpProb <- function(gt, lambda, type, k1, k2) { if ((type == "ETEL")|(type=="ETHD")) type <- "ET" n <- NROW(gt) pt <- -.rho(gt, lambda, type = type, derive = 1, k = k1/k2)/n # Making sure pt>0 if (type=="CUE") { eps <- -length(pt)*min(min(pt),0) pt <- (pt+eps/length(pt))/(1+eps) } if (type=="RCUE") pt[pt<0] <- 0 ################### conv_moment <- colSums(pt*gt) conv_pt <- sum(as.numeric(pt)) pt <- pt/sum(pt) attr(pt, "conv_moment") <- conv_moment attr(pt, "conv_pt") <- conv_pt pt } .vcovGel <- function(gt, G, k1, k2, bw, pt=NULL,tol=1e-16) { q <- NCOL(gt) n <- NROW(gt) if (is.null(pt)) pt <- 1/n G <- G/k1 gt <- gt*sqrt(pt*bw/k2) qrGt <- qr(gt) piv <- sort.int(qrGt$pivot, index.return=TRUE)$ix R <- qr.R(qrGt)[,piv] X <- forwardsolve(t(R), G) Y <- forwardsolve(t(R), diag(q)) res <- lm.fit(X,Y) u <- res$residuals Sigma <- chol2inv(res$qr$qr)/n diag(Sigma)[diag(Sigma)<0] <- tol if (q==ncol(G)) { SigmaLam <- matrix(0, q, q) } else { SigmaLam <- backsolve(R, u)/n*bw^2 diag(SigmaLam)[diag(SigmaLam)<0] <- tol } khat <- crossprod(R) list(vcov_par=Sigma, vcov_lambda=SigmaLam,khat=khat) } gmm/R/specTest.R0000644000176200001440000000362514247643115013165 0ustar liggesusersspecTest <- function(x, ...) { UseMethod("specTest") } specTest.gmm <- function(x, ...) { if (x$infWmatrix == "ident") { gb <- colMeans(x$gt) j <- try(crossprod(gb,solve(x$v,gb))*x$n, silent=TRUE) } else if ( (x$infVcov!="TrueFixed") & !is.null(x$weightsMatrix) ) { gb <- colMeans(x$gt) j <- try(crossprod(gb,solve(x$v,gb))*x$n, silent=TRUE) } else j <- x$objective*x$n if (any(class(j)=="try-error")) { j <- noquote(cbind("Failed", "Failed")) } else { j <- noquote(cbind(j, ifelse(x$df>0,pchisq(j,x$df, lower.tail = FALSE), "*******"))) } J_test <- noquote(paste("J-Test: degrees of freedom is ",x$df,sep="")) dimnames(j) <- list("Test E(g)=0: ", c("J-test", "P-value")) ans<-list(ntest=J_test, test = j) class(ans) <- "specTest" ans } print.specTest <- function(x, digits=5, ...) { cat("\n","## ",x$ntest," ##","\n\n") print.default(format(x$test, digits=digits), print.gap = 2, quote = FALSE) cat("\n") invisible(x) } specTest.gel <- function(x, ...) { n <- nrow(x$gt) khat <- x$khat gbar <- colMeans(x$gt) LR_test <- 2*x$objective*n*x$k2/(x$bwVal*x$k1^2) if (x$type == "ETHD") LR_test <- LR_test*2 LM_test <- n*crossprod(x$lambda, crossprod(khat, x$lambda))/(x$bwVal^2) J_test <- n*crossprod(gbar, solve(khat, gbar))/(x$k1^2) test <- c(LR_test, LM_test, J_test) df <- x$df ntest <- noquote(paste("Over-identifying restrictions tests: degrees of freedom is ", df, sep = "")) vptest <- pchisq(test,df,lower.tail = FALSE) if (df == 0) vptest[] <- "###" test <- cbind(test,vptest) dimnames(test) <- list(c("LR test", "LM test", "J test"), c("statistics", "p-value")) ans <- list(test = test, ntest = ntest) class(ans) <- "specTest" ans } gmm/R/gmmTests.R0000644000176200001440000001354414247643115013177 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ # This function compute what is needed for the K statistics of KleiBergen (2005) ##################################################################################### .BigCov <- function(obj,theta0) { insertRC <- function(A,w,v) { NewA <- matrix(ncol=ncol(A)+length(w),nrow=nrow(A)+length(w)) NewA[-w,-w] <- A NewA[w,] <- v NewA[,w] <- v NewA } dg <- function(obj) { dat <- obj$dat x <- dat$x[,(dat$ny+1):(dat$ny+dat$k)] k <- ncol(x) h <- dat$x[,(dat$ny+dat$k+1):ncol(dat$x)] qt <- array(dim=c(dim(obj$gt),k)) for (i in 1:k) qt[,,i] <- -x[,i]*h qt} if (attr(obj$dat,"ModelType") == "nonlinear") { Myenv <- new.env() assign("obj", obj, envir=Myenv) assign("theta", theta0, envir=Myenv) gFunct <- obj$g assign("g",gFunct,envir=Myenv) res <- numericDeriv(quote(g(theta,obj$dat)),"theta",Myenv) qT <- attr(res,"gradient") } else { qT <- dg(obj)} qTmat <- apply(qT,3,colSums) qT <- matrix(qT,nrow=dim(qT)[1]) gt <- obj$g(theta0,obj$dat) fT <- colSums(gt) n <- nrow(gt) q <- ncol(gt) All <- cbind(gt,qT) All <- sweep(All,2,colMeans(All),FUN="-") f <- function(x) all(abs(x)<1e-7) w <- which(apply(All,2,f)) if (length(w) != 0) All <- All[,-w] if (dim(All)[2] >= dim(All)[1]) stop("Too many moment conditions. Cannot estimate V") if (attr(obj$dat, "weight")$vcov == "iid") { V <- crossprod(All)/nrow(All) } else { class(All) <- "gmmFct" argSand <- attr(obj$dat, "weight")$WSpec$sandwich argSand$x <- All argSand$sandwich <- FALSE V <- do.call(kernHAC,argSand) } if (length(w) != 0) V <- insertRC(V,w,0) Vff <- V[1:q,1:q] Vthetaf <- V[(q+1):nrow(V),1:q] list(Vff=Vff,Vthetaf=Vthetaf,qT=qTmat,fT=fT,n=n,q=q) } KTest <- function(obj, theta0=NULL, alphaK = 0.04, alphaJ = 0.01) { if (class(obj)[1] != "gmm") stop("KTest is only for gmm type objects") if (!is.null(attr(obj$dat,"eqConst"))) { if (!is.null(theta0)) warning("setting a value for theta0 has no effect when the gmm is already constrained") resTet <- attr(obj$dat,"eqConst")$eqConst tet <- obj$coefficients theta0 <- vector(length=length(tet)+nrow(resTet)) theta0[resTet[,1]] <- resTet[,2] theta0[-resTet[,1]] <- tet testName <- paste(rownames(resTet), " = ", resTet[,2], collapse="\n") if (attr(obj$dat,"ModelType") == "linear") { obj$dat$k <- length(theta0) } dfK <- nrow(resTet) which <- resTet[,1] attr(obj$dat, "eqConst") <- NULL } else { if (is.null(theta0)) stop("You must either estimate a restricted model first or set theta0 under H0") if (length(theta0) != length(obj$coef)) stop("theta0 is only for tests on the whole vector theta when obj is an unrestricted GMM") dfK <- length(theta0) testName <- paste(names(obj$coef), " = ", theta0, collapse="\n") which <- 1:length(theta0) } V <- .BigCov(obj, theta0) Vff <- V$Vff Vtf <- V$Vthetaf qT <- V$qT fT <- V$fT dfJ <- V$q-length(theta0) # the following is vec(D) D <- c(qT)-Vtf%*%solve(Vff,fT) D <- matrix(D,ncol=length(theta0)) meat <- t(D)%*%solve(Vff,D) bread <- t(fT)%*%solve(Vff,D) K <- bread%*%solve(meat,t(bread))/V$n pv <- 1-pchisq(K,dfK) S <- t(fT)%*%solve(Vff,fT)/V$n J <- S-K dfS <- dfK + dfJ pvJ <- 1-pchisq(J,dfJ) type <- c("K statistics","J statistics", "S statistics") test <- c(K,J,S) pvS <- 1-pchisq(S,dfS) test <- cbind(test,c(pv,pvJ,pvS),c(dfK,dfJ,dfS)) dist <- paste("Chi_sq with ", c(dfK,dfJ,dfS), " degrees of freedom", sep="") if(dfJ>0) ans <- list(test=test,dist=dist,type=type,testName=testName) else ans <- list(test=matrix(test[1,],nrow=1),dist=dist[1],type=type[1],testName=testName) if (pvJ=length(obj$coefficients)) stop("Too many constraints") if (is.character(which)) { if (any(!(which %in% names(obj$coefficients)))) stop("Wrong coefficient names in eqConst") if (attr(obj$dat,"ModelType") == "linear") which <- match(which,names(obj$coefficients)) } if (!is.null(argCall$t0)) { argCall$t0 <- obj$coefficients argCall$t0[which] <- value } if (attr(obj$dat,"ModelType") == "nonlinear") { eqConst <- which } else { eqConst <- cbind(which,value) } argCall$eqConst <- eqConst res <- do.call(gmm,argCall) res$call <- match.call() return(res) } gmm/R/Methods.gel.R0000644000176200001440000002552714247643115013551 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ .invTest <- function(object, which, level = 0.95, fact = 3, type=c("LR", "LM", "J"), corr=NULL) { type <- match.arg(type) argCall <- object$allArg if (length(which) > 1) stop("tests are inverted only for one parameter") if (is.character(which)) { which <- which(which == names(object$coefficients)) if (length(which) == 0) stop("Wrong parameter names") } wtest <- which(type==c("LR", "LM", "J")) if (object$df > 0) { test0 <- specTest(object)$test[wtest,] test0 <- test0[1] } else { test0 <- 0 } if (length(object$coefficients) == 1) { argCall$optfct <- NULL argCall$constraint <- NULL argCall$eqConst <- NULL argCall$eqConstFullVcov <- NULL fctCall <- "evalGel" } else { fctCall <- "gel" argCall$eqConst <- which } if (length(object$coefficients) == 2) { sdcoef <- sqrt(diag(vcov(object))[-which]) coef <- object$coefficients[-which] upper <- coef + fact*sdcoef lower <- coef - fact*sdcoef argCall$method <- "Brent" argCall$lower <- lower argCall$upper <- upper } sdcoef <- sqrt(diag(vcov(object))[which]) coef <- object$coefficients[which] int1 <- c(coef, coef + fact*sdcoef) int2 <- c(coef - fact*sdcoef, coef) fct <- function(coef, which, wtest, level, test0, corr=NULL) { argCall$tet0 <- object$coefficients argCall$tet0[which] <- coef obj <- do.call(get(fctCall), argCall) test <- as.numeric(specTest(obj)$test[wtest,1]) - test0 if (is.null(corr)) level - pchisq(test, 1) else level - pchisq(test/corr, 1) } res1 <- uniroot(fct, int1, which = which, wtest=wtest, level=level, test0=test0, corr=corr) res2 <- uniroot(fct, int2, which = which, wtest=wtest, level=level, test0=test0, corr=corr) sort(c(res1$root, res2$root)) } confint.gel <- function(object, parm, level = 0.95, lambda = FALSE, type = c("Wald", "invLR", "invLM", "invJ"), fact = 3, corr = NULL, ...) { type <- match.arg(type) z <- object n <- nrow(z$gt) if (type == "Wald") { ntest <- "Direct Wald type confidence interval" se_par <- sqrt(diag(z$vcov_par)) par <- z$coefficients tval <- par/se_par se_parl <- sqrt(diag(z$vcov_lambda)) lamb <- z$lambda zs <- qnorm((1 - level)/2, lower.tail=FALSE) ch <- zs*se_par if(!lambda) { ans <- cbind(par-ch, par+ch) dimnames(ans) <- list(names(par), c((1 - level)/2, 0.5+level/2)) } if(lambda) { if (length(z$coefficients) == length(z$lambda)) { cat("\nNo confidence intervals for lambda when the model is just identified.\n") return(NULL) } else { chl <- zs*se_parl ans <- cbind(lamb - chl, lamb + chl) dimnames(ans) <- list(names(lamb), c((1 - level)/2, 0.5 + level/2)) } } if(!missing(parm)) ans <- ans[parm,] } else { if(missing(parm)) parm <- names(object$coefficients) type <- strsplit(type, "v")[[1]][2] ntest <- paste("Confidence interval based on the inversion of the ", type, " test", sep="") ans <- lapply(parm, function(w) .invTest(object, w, level = level, fact = fact, type=type, corr=corr)) ans <- do.call(rbind, ans) if (!is.character(parm)) parm <- names(object$coefficients)[parm] dimnames(ans) <- list(parm, c((1 - level)/2, 0.5+level/2)) } ans <- list(test=ans,ntest=ntest) class(ans) <- "confint" ans } print.confint <- function(x, digits = 5, ...) { cat("\n", x$ntest, sep="") cat("\n#######################################\n") print.default(format(x$test, digits = digits), print.gap = 2, quote = FALSE) invisible(x) } coef.gel <- function(object, lambda = FALSE, ...) { if(!lambda) object$coefficients else object$lambda } vcov.gel <- function(object, lambda = FALSE, ...) { if(!lambda) object$vcov_par else object$vcov_lambda } print.gel <- function(x, digits = 5, ...) { if (is.null(x$CGEL)) cat("Type de GEL: ", x$typeDesc, "\n") else cat("CGEL of type: ", x$typeDesc, " (alpha = ", x$CGEL, ")\n") if (!is.null(attr(x$dat,"smooth"))) { cat("Kernel: ", attr(x$dat,"smooth")$kernel," (bw=", attr(x$dat,"smooth")$bw,")\n\n") } else cat("\n") cat("Coefficients:\n") print.default(format(coef(x), digits = digits), print.gap = 2, quote = FALSE) cat("\n") cat("Lambdas:\n") print.default(format(coef(x, lambda = TRUE), digits = digits), print.gap = 2, quote = FALSE) cat("\n") cat("Convergence code for the coefficients: ", x$conv_par,"\n") cat("Convergence code for Lambda: ", x$conv_lambda$convergence,"\n") cat(x$specMod) invisible(x) } print.summary.gel <- function(x, digits = 5, ...) { cat("\nCall:\n") cat(paste(deparse(x$call), sep = "\n", collapse = "\n"), "\n\n", sep = "") if (is.null(x$CGEL)) cat("Type of GEL: ", x$typeDesc, "\n") else cat("CGEL of type: ", x$typeDesc, " (alpha = ", x$CGEL, ")\n") if (!is.null(x$smooth)) { cat("Kernel: ", x$smooth$kernel," (bw=", x$smooth$bw,")\n\n") }else { cat("\n") } cat("Coefficients:\n") print.default(format(x$coefficients, digits = digits), print.gap = 2, quote = FALSE) if (length(x$coefficients)|t|)")) dimnames(ans$lambda) <- list(names(z$lambda), c("Estimate", "Std. Error", "t value", "Pr(>|t|)")) ans$stest=specTest(z) if (z$type == "EL") ans$badrho <- z$badrho if (!is.null(z$weights)) { ans$weights <- z$weights } ans$conv_par <- z$conv_par ans$conv_pt <- z$conv_pt ans$conv_moment <- cbind(z$conv_moment) ans$conv_lambda <- z$conv_lambda ans$CGEL <- z$CGEL ans$typeDesc <- z$typeDesc ans$specMod <- z$specMod if (!is.null(attr(object$dat,"smooth"))) ans$smooth <- attr(object$dat,"smooth") names(ans$conv_pt) <- "Sum_of_pt" dimnames(ans$conv_moment) <- list(names(z$gt), "Sample_moment_with_pt") class(ans) <- "summary.gel" ans } residuals.gel <- function(object, ...) { if(is.null(object$model)) stop("The residuals method is valid only for g=formula") object$residuals } fitted.gel <- function(object, ...) { if(is.null(object$model)) stop("The residuals method is valid only for g=formula") object$fitted.value } formula.gel <- function(x, ...) { if(is.null(x$terms)) stop("The gel object was not created by a formula") else formula(x$terms) } estfun.gel <- function(x, ...) { stop("estfun is not yet available for gel objects") } bread.gel <- function(x, ...) { stop("Bread is not yet available for gel objects") } getImpProb <- function(object, ...) UseMethod("getImpProb") getImpProb.gel <- function(object, posProb=TRUE, normalize=TRUE, checkConv=FALSE, ...) { if (!normalize || (object$type == "CUE" && !posProb)) { n <- NROW(object$gt) pt <- -.rho(object$gt, object$lambda, type = object$type, derive = 1, k = object$k1/object$k2)/n if (object$type == "CUE" && posProb) { eps <- -length(pt)*min(min(pt),0) pt <- (pt+eps/length(pt))/(1+eps) } if (normalize) pt <- pt/sum(pt) } else { pt <- object$pt } if (checkConv) attr(pt, "convergence") <- list(pt=sum(pt), ptgt=colSums(pt*as.matrix(object$gt))) pt } gmm/R/momentEstim.R0000644000176200001440000013240614612777354013705 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ momentEstim <- function(object, ...) { UseMethod("momentEstim") } momentEstim.sysGmm.twoStep.formula <- function(object, ...) { dat <- object$x y <- lapply(1:length(dat), function(i) dat[[i]]$x[,1]) y <- do.call(c, y) z <- lapply(1:length(dat), function(i) dat[[i]]$x[,(2+dat[[i]]$k):ncol(dat[[i]]$x), drop=FALSE]) z <- .diagMatrix(z) x <- lapply(1:length(dat), function(i) dat[[i]]$x[,2:(dat[[i]]$k+1), drop=FALSE]) if (attr(dat, "sysInfo")$commonCoef) { x <- do.call(rbind, x) } else if (!is.null(attr(dat, "sysInfo")$crossEquConst)) { k <- attr(dat, "k")[[1]] x <- .diagMatrix(x, (1:k)[-attr(dat, "sysInfo")$crossEquConst]) } else { x <- .diagMatrix(x) } names(y) <- rownames(x) <- rownames(z) <- 1:length(y) df <- ncol(z) - ncol(x) k <- ncol(x) q <- ncol(z) n <- nrow(dat[[1]]$x) df.residuals <- n - k data <- list(y=y, x=x, z=z) dat2 <- getDat(y~x-1, ~z-1, data=data) attr(dat2, "ModelType") <- "linear" if (!is.null(object$weightsMatrix)) { w <- object$weightsMatrix attr(w, "inv") <- FALSE tet0 <- NULL fsRes <- .fsResOnly(dat2) } else if (object$wmatrix == "ident") { w <- diag(q) attr(w, "inv") <- FALSE tet0 <- NULL fsRes <- .fsResOnly(dat2) } else { res0 <- .tetlin(dat2, 1, "2sls") tet0 <- .getThetaList(res0$par, dat) fsRes <- res0$fsRes w <- .weightFct_Sys(tet=tet0, dat=dat, type=object$vcov) } res <- .tetlin(dat2, w) par <- .getThetaList(res$par, dat) names(par) <- names(dat) z = list(coefficients = par, objective = res$value, dat=dat, k=k, q=q, df=df, df.residual=df.residual, n=n) z$gt <- object$g(z$coefficients, dat) z$initTheta <- tet0 tmp <- lapply(1:length(dat), function(i) .residuals(z$coefficients[[i]], dat[[i]])) z$fitted.values <- lapply(1:length(dat), function(i) tmp[[i]]$yhat) z$residuals <- lapply(1:length(dat), function(i) tmp[[i]]$residuals) z$terms <- lapply(1:length(dat), function(i) dat[[i]]$mt) if(object$model) z$model <- lapply(1:length(dat), function(i) dat[[i]]$mf) if(object$X) z$x <- lapply(1:length(dat), function(i) as.matrix(dat[[i]]$x[,(dat[[i]]$ny+1):(dat[[i]]$ny+dat[[i]]$k)])) if(object$Y) z$y <- lapply(1:length(dat), function(i) as.matrix(dat[[i]]$x[,1:dat[[i]]$ny])) z$gradv <- object$gradv z$g <- object$g z$WSpec <- object$WSpec z$w0 <- w colnames(z$gt) <- do.call(c, object$namesgt) z$fsRes <- fsRes class(z) <- "sysGmm.res" z$specMod <- object$specMod return(z) } momentEstim.baseGmm.eval <- function(object, ...) { P <- object x <- P$x n <- attr(x, "n") q <- attr(x, "q") k <- attr(x, "k") k2 <- k df <- q - k if (is.null(object$tetw)) object$tetw <- object$t0 if (is.null(attr(x, "weight")$w)) w <- .weightFct(object$tetw, x, P$vcov) else w <- .weightFct(object$tetw, x, "fixed") fct <- .obj1(object$t0, x, w) if (attr(x,"ModelType") == "linear") { z = list(coefficients = object$t0, objective = fct, dat = x, k = k, k2 = k2, n = n, q = q, df = df, df.residual = (n-k)) b <- z$coefficients tmp <- .residuals(b, x) z$fitted.values <- tmp$yhat z$residuals <- tmp$rediduals z$terms <- x$mt if(P$model) z$model <- x$mf if(P$X) z$x <- as.matrix(x$x[,(x$ny+1):(x$ny+x$k)]) if(P$Y) z$y <- as.matrix(x$x[,1:x$ny]) } else { attr z = list(coefficients = object$t0, objective = fct, k=k, k2=k2, n=n, q=q, df=df, initTheta = object$t0, dat=x) } z$gt <- .momentFct(z$coefficients, x) if (is.null(colnames(z$gt))) colnames(z$gt) <- paste("gt",1:ncol(z$gt),sep="") else colnames(z$gt) <- P$namesgt z$gradv <- P$gradv z$iid <- P$iid z$g <- P$g z$WSpec <- P$WSpec z$w0 <- w names(z$coefficients) <- P$namesCoef class(z) <- paste(P$TypeGmm,".res",sep="") z$specMod <- P$specMod return(z) } momentEstim.baseGmm.twoStep <- function(object, ...) { P <- object x <- P$x n <- attr(x, "n") q <- attr(x, "q") k <- attr(x, "k") k2 <- k df <- q - k w = .weightFct(NULL, x, "ident") chkOptim <- any(P$optfct == c("optim", "constrOptim")) if (chkOptim) { if (P$gradvf) { gr2 <- function(thet, x, w) { gt <- .momentFct(thet, x) Gbar <- .DmomentFct(thet, x) gbar <- as.vector(colMeans(gt)) INV <- attr(w, "inv") if (INV) obj <- crossprod(Gbar, solve(w, gbar)) else obj <- crossprod(Gbar,w)%*%gbar return(obj*2) } } else { gr2 <- NULL } } if (P$optfct == "optim") { argDots <- list(...) allArgOptim <- list(par = P$t0, fn = .obj1, gr = gr2, x = x, w = w) allArgOptim <- c(allArgOptim,argDots) res <- do.call(optim,allArgOptim) } if (P$optfct == "constrOptim") { if (!any(c("ui","ci") %in% names(list(...)))) stop("You must specify ui and ci when optfct is set to constrOptim") argDots <- list(...) ui <- argDots$ui ci <- argDots$ci argDots$ui <- NULL argDots$ci <- NULL allArgOptim <- list(theta = P$t0, f = .obj1, grad = gr2, ui = ui, ci = ci, x = x, w = w) allArgOptim <- c(allArgOptim,argDots) res <- do.call(constrOptim,allArgOptim) } if (P$optfct == "nlminb") { res <- nlminb(P$t0, .obj1, x = x, w = w, ...) res$value <- res$objective } if (P$optfct == "optimize") { res <- optimize(.obj1, P$t0, x = x, w = w, ...) res$par <- res$minimum res$value <- res$objective } if (q == k2 | P$wmatrix == "ident") { z = list(coefficients = res$par, objective = res$value, k=k, k2=k2, n=n, q=q, df=df) if (chkOptim) z$algoInfo <- list(convergence = res$convergence, counts = res$counts, message = res$message) else if(P$optfct == "nlminb") z$algoInfo <- list(convergence = res$convergence, counts = res$evaluations, message = res$message) } else { w <- .weightFct(res$par, P$x, P$vcov) if (chkOptim) { allArgOptim$w <- w res2 <- do.call(get(object$optfct),allArgOptim) } if (P$optfct == "nlminb") { res2 <- nlminb(P$t0, .obj1, x = x, w = w, ...) res2$value <- res2$objective } if (P$optfct == "optimize") { res2 <- optimize(.obj1, P$t0, x = x, w = w, ...) res2$par <- res2$minimum res2$value <- res2$objective } z = list(coefficients = res2$par, objective = res2$value, k=k, k2=k2, n=n, q=q, df=df, initTheta = res$par) if (chkOptim) { z$algoInfo <- list(convergence = res2$convergence, counts = res2$counts, message = res2$message) z$InitialAlgoInfo <- list(convergence = res$convergence, counts = res$counts, message = res$message) } else if(P$optfct == "nlminb") { z$algoInfo <- list(convergence = res2$convergence, counts = res2$evaluations, message = res2$message) z$InitialAlgoInfo <- list(convergence = res$convergence, counts = res$evaluations, message = res$message) } } z$dat <- P$x z$gt <- P$g(z$coefficients, P$x) z$gradv <- P$gradv z$iid <- P$iid z$g <- P$g z$WSpec <- P$WSpec names(z$coefficients) <- P$namesCoef if (is.null(colnames(z$gt))) colnames(z$gt) <- paste("gt",1:ncol(z$gt),sep="") class(z) <- paste(P$TypeGmm,".res",sep="") z$specMod <- P$specMod z$w0 <- w return(z) } momentEstim.tsls.twoStep.formula <- function(object, ...) { P <- object g <- P$g dat <- P$x if (dat$ny > 1) stop("tsls is for one dimentional dependent variable") n <- attr(dat, "n") q <- attr(dat, "q") k <- attr(dat, "k") k2 <- k*dat$ny df <- q-k*dat$ny w = .weightFct(NULL, dat, "ident") if (q == k2) { res2 <- .tetlin(dat, w) z = list(coefficients = res2$par, objective = res2$value, dat = dat, k = k, k2 = k2, n = n, q = q, df = df, df.residual = (n-k)) } else { res2 <- .tetlin(dat, w, type="2sls") z = list(coefficients = res2$par, objective = res2$value, dat=dat, k=k, k2=k2, n=n, q=q, df=df, df.residual = (n-k)) } z$gt <- g(z$coefficients, dat) tmp <- .residuals(z$coefficients, dat) z$fitted.values <- tmp$yhat z$residuals <- tmp$residuals z$terms <- dat$mt if(P$model) z$model <- dat$mf if(P$X) z$x <- as.matrix(dat$x[,(dat$ny+1):(dat$ny+dat$k)]) if(P$Y) z$y <- as.matrix(dat$x[,1:dat$ny]) z$gradv <- P$gradv z$iid <- P$iid z$g <- P$g z$WSpec <- P$WSpec z$w0 <- w names(z$coefficients) <- P$namesCoef colnames(z$gt) <- P$namesgt z$fsRes <- res2$fsRes class(z) <- "baseGmm.res" z$specMod <- P$specMod return(z) } momentEstim.baseGmm.twoStep.formula <- function(object, ...) { P <- object g <- P$g dat <- P$x n <- attr(dat, "n") q <- attr(dat, "q") k <- attr(dat, "k") k2 <- k*dat$ny df <- q-k*dat$ny w = .weightFct(NULL, dat, "ident") if (q == k2 | P$wmatrix == "ident") { res2 <- .tetlin(dat, w) z = list(coefficients = res2$par, objective = res2$value, dat = dat, k = k, k2 = k2, n = n, q = q, df = df, df.residual = (n-k)) } else { res1 <- .tetlin(dat, w, type="2sls") initTheta <- res1$par w <- .weightFct(res1$par, dat, P$vcov) res2 <- .tetlin(dat, w) res2$firstStageReg <- res1$firstStageReg res2$fsRes <- res1$fsRes z = list(coefficients = res2$par, objective = res2$value, dat=dat, k=k, k2=k2, n=n, q=q, df=df, initTheta = initTheta, df.residual = (n-k)) } z$gt <- g(z$coefficients, dat) b <- z$coefficients tmp <- .residuals(b, dat) z$fitted.values <- tmp$yhat z$residuals <- tmp$residuals z$terms <- dat$mt if(P$model) z$model <- dat$mf if(P$X) z$x <- as.matrix(dat$x[,(dat$ny+1):(dat$ny+dat$k)]) if(P$Y) z$y <- as.matrix(dat$x[,1:dat$ny]) z$gradv <- P$gradv z$iid <- P$iid z$g <- P$g z$WSpec <- P$WSpec z$w0 <- w names(z$coefficients) <- P$namesCoef colnames(z$gt) <- P$namesgt if (P$vcov == "iid" & P$wmatrix != "ident") z$fsRes <- res2$fsRes class(z) <- paste(P$TypeGmm,".res",sep="") z$specMod <- P$specMod return(z) } momentEstim.baseGmm.iterative.formula <- function(object, ...) { P <- object g <- P$g dat <- P$x n <- attr(dat, "n") q <- attr(dat, "q") k <- attr(dat, "k") k2 <- k*dat$ny df <- q-k*dat$ny w = .weightFct(NULL, dat, "ident") if (q == k2 | P$wmatrix == "ident") { res <- .tetlin(dat, w) z = list(coefficients = res$par, objective = res$value, dat = dat, k = k, k2 = k2, n = n, q = q, df = df, df.residual = (n-k)) } else { res <- .tetlin(dat, w, type="2sls") fsRes <- res$fsRes initTheta <- res$par ch <- 100000 j <- 1 while(ch > P$crit) { tet <- res$par w <- .weightFct(tet, dat, P$vcov) res <- .tetlin(dat, w) ch <- crossprod(abs(tet- res$par)/tet)^.5 if (j>P$itermax) { cat("No convergence after ", P$itermax, " iterations") ch <- P$crit } if(P$traceIter) cat("Iter :",j,": value=",res$value,", Coef = ", res$par,"\n") j <- j+1 } z = list(coefficients = res$par, objective = res$value, dat=dat, k=k, k2=k2, n=n, q=q, df=df, initTheta=initTheta, df.residual = (n-k)) } z$gt <- g(z$coefficients, dat) tmp <- .residuals(z$coefficients, dat) z$fitted.values <- tmp$yhat z$residuals <- tmp$residuals z$terms <- dat$mt if(P$model) z$model <- dat$mf if(P$X) z$x <- as.matrix(dat$x[,(dat$ny+1):(dat$ny+dat$k)]) if(P$Y) z$y <- as.matrix(dat$x[,1:dat$ny]) z$terms <- dat$mt if(P$model) z$model <- dat$mf z$gradv <- P$gradv z$iid <- P$iid z$g <- P$g z$WSpec <- P$WSpec z$w0 <- w names(z$coefficients) <- P$namesCoef colnames(z$gt) <- P$namesgt if (P$vcov == "iid" & P$wmatrix != "ident") z$fsRes <- fsRes class(z) <- paste(P$TypeGmm,".res",sep="") z$specMod <- P$specMod return(z) } momentEstim.baseGmm.iterative <- function(object, ...) { P <- object x <- P$x n <- attr(x, "n") q <- attr(x, "q") k <- attr(x, "k") k2 <- k df <- q - k w = .weightFct(NULL, x, "ident") chkOptim <- any(P$optfct == c("optim", "constrOptim")) if (chkOptim) { if (P$gradvf) { gr2 <- function(thet, x, w) { gt <- .momentFct(thet, x) Gbar <- .DmomentFct(thet, x) gbar <- as.vector(colMeans(gt)) INV <- attr(w, "inv") if (INV) obj <- crossprod(Gbar, solve(w, gbar)) else obj <- crossprod(Gbar,w)%*%gbar return(obj*2) } } else { gr2 <- NULL } } if (P$optfct == "optim") { argDots <- list(...) allArgOptim <- list(par = P$t0, fn = .obj1, gr = gr2, x = x, w = w) allArgOptim <- c(allArgOptim,argDots) res <- do.call(optim,allArgOptim) } if (P$optfct == "constrOptim") { if (!any(c("ui","ci") %in% names(list(...)))) stop("You must specify ui and ci when optfct is set to constrOptim") argDots <- list(...) ui <- argDots$ui ci <- argDots$ci argDots$ui <- NULL argDots$ci <- NULL allArgOptim <- list(theta = P$t0, f = .obj1, grad = gr2, ui = ui, ci = ci, x = x, w = w) allArgOptim <- c(allArgOptim,argDots) res <- do.call(constrOptim,allArgOptim) } if (P$optfct == "nlminb") { res <- nlminb(P$t0, .obj1, x = x, w = w, ...) res$value <- res$objective } if (P$optfct == "optimize") { res <- optimize(.obj1, P$t0, x = x, w = w, ...) res$par <- res$minimum res$value <- res$objective } if (q == k2 | P$wmatrix == "ident") { z <- list(coefficients = res$par, objective = res$value, k=k, k2=k2, n=n, q=q, df=df) if (chkOptim) z$algoInfo <- list(convergence = res$convergence, counts = res$counts, message = res$message) else if(P$optfct == "nlminb") z$algoInfo <- list(convergence = res$convergence, counts = res$evaluations, message = res$message) } else { initTheta = res$par z <- list() if (chkOptim) z$initialAlgoInfo <- list(convergence = res$convergence, counts = res$counts, message = res$message) else if(P$optfct == "nlminb") z$initialAlgoInfo <- list(convergence = res$convergence, counts = res$evaluations, message = res$message) ch <- 100000 j <- 1 while(ch > P$crit) { tet <- res$par w <- .weightFct(tet, P$x, P$vcov) if (chkOptim) { allArgOptim$w <- w allArgOptim[[1]] <- tet res <- do.call(get(P$optfct),allArgOptim) } if (P$optfct == "nlminb") { res <- nlminb(tet, .obj1, x = P$x, w = w, ...) res$value <- res$objective } if (P$optfct == "optimize") { res <- optimize(.obj1, P$t0, x = P$x, w = w, ...) res$par <- res$minimum res$value <- res$objective } ch <- crossprod(tet-res$par)^.5/(1+crossprod(tet)^.5) if (j>P$itermax) { cat("No convergence after ", P$itermax, " iterations") ch <- P$crit } if(P$traceIter) cat("Iter :",j,": value=",res$value,", Coef = ", res$par,"\n") j <- j+1 } z2 = list(coefficients = res$par, objective = res$value,k=k, k2=k2, n=n, q=q, df=df, initTheta=initTheta) z <- c(z, z2) if (chkOptim) z$algoInfo <- list(convergence = res$convergence, counts = res$counts, message = res$message) else if(P$optfct == "nlminb") z$algoInfo <- list(convergence = res$convergence, counts = res$evaluations, message = res$message) } z$dat <- P$x z$gt <- P$g(z$coefficients, P$x) z$gradv <- P$gradv z$iid <- P$iid z$g <- P$g z$WSpec <- P$WSpec z$w0 <- w names(z$coefficients) <- P$namesCoef if (is.null(colnames(z$gt))) colnames(z$gt) <- paste("gt",1:ncol(z$gt),sep="") z$specMod <- P$specMod class(z) <- paste(P$TypeGmm,".res",sep="") return(z) } momentEstim.baseGmm.cue.formula <- function(object, ...) { P <- object g <- P$g dat <- P$x n <- attr(dat, "n") q <- attr(dat, "q") k <- attr(dat, "k") k2 <- k*dat$ny df <- q-k*dat$ny w <- .weightFct(NULL, dat, "ident") if (q == k2 | P$wmatrix == "ident") { res <- .tetlin(dat, w) z = list(coefficients = res$par, objective = res$value, dat = dat, k = k, k2 = k2, n = n, q = q, df = df, df.residual = (n-k)) P$weightMessage <- "No CUE needed because the model is just identified" } else { if (is.null(P$t0)) { P$t0 <- .tetlin(dat, w, type="2sls")$par initTheta <- P$t0 } else { initTheta <- P$t0 } if (P$vcov == "HAC") { w <- .weightFct(P$t0, dat, P$vcov) attr(dat, "weight")$WSpec$sandwich$bw <- attr(w,"Spec")$bw P$weightMessage <- "Weights for kernel estimate of the covariance are fixed and based on the first step estimate of Theta" } if (P$optfct == "optim") res2 <- optim(P$t0,.objCue, x = dat, type = P$vcov, ...) if (P$optfct == "constrOptim") { if (!any(c("ui","ci") %in% names(list(...)))) stop("You must specify ui and ci when optfct is set to constrOptim") argDots <- list(...) ui <- argDots$ui ci <- argDots$ci argDots$ui <- NULL argDots$ci <- NULL allArgOptim <- list(theta = P$t0, f = .objCue, grad = NULL, ui = ui, ci = ci, x = dat, type = P$vcov) allArgOptim <- c(allArgOptim,argDots) res2 <- do.call(constrOptim,allArgOptim) } if (P$optfct == "nlminb") { res2 <- nlminb(P$t0,.objCue, x = dat, type = P$vcov, ...) res2$value <- res2$objective } if (P$optfct == "optimize") { res2 <- optimize(.objCue,P$t0, x = dat, type = P$vcov, ...) res2$par <- res2$minimum res2$value <- res2$objective } z = list(coefficients = res2$par, objective = res2$value, dat = dat, k = k, k2 = k2, n = n, q = q, df = df, initTheta=initTheta, df.residual = (n-k)) if (any(P$optfct == c("optim", "constrOptim"))) z$algoInfo <- list(convergence = res2$convergence, counts = res2$counts, message = res2$message) else if(P$optfct == "nlminb") z$algoInfo <- list(convergence = res2$convergence, counts = res2$evaluations, message = res2$message) } z$gt <- g(z$coefficients, dat) tmp <- .residuals(z$coefficients, dat) z$fitted.values <- tmp$yhat z$residuals <- tmp$residuals z$terms <- dat$mt if(P$model) z$model <- dat$mf if(P$X) z$x <- as.matrix(dat$x[,(dat$ny+1):(dat$ny+dat$k)]) if(P$Y) z$y <- as.matrix(dat$x[,1:dat$ny]) z$dat <- dat z$terms <- dat$mt if(P$model) z$model <- dat$mf z$gradv <- P$gradv z$iid <- P$iid z$g <- P$g z$specMod <- P$specMod z$cue <- list(weights=P$fixedKernW,message=P$weightMessage) z$WSpec <- P$WSpec z$w0 <- .weightFct(z$coefficients, dat, P$vcov) names(z$coefficients) <- P$namesCoef colnames(z$gt) <- P$namesgt class(z) <- paste(P$TypeGmm,".res",sep="") return(z) } momentEstim.baseGmm.cue <- function(object, ...) { P <- object x <- P$x n <- attr(x, "n") q <- attr(x, "q") k <- attr(x, "k") k2 <- k df <- q - k if (q == k2 | P$wmatrix == "ident") { w = .weightFct(NULL, x, "ident") res <- gmm(P$allArg$g,P$allArg$x,P$t0,wmatrix="ident",optfct=P$optfct, ...) z <- list(coefficients = res$coef, objective = res$objective, algoInfo = res$algoInfo, k=k, k2=k2, n=n, q=q, df=df, initTheta=P$t0) P$weightMessage <- "No CUE needed because the model if just identified or you set wmatrix=identity" } else { w <- .weightFct(P$t0, x, P$vcov) initTheta <- P$t0 if (P$vcov == "HAC") { res <- try(gmm(P$allArg$g,P$allArg$x,P$t0,wmatrix="ident", optfct=P$optfct, ...)) if(any(class(res)=="try-error")) stop("Cannot get a first step estimate to compute the weights for the Kernel estimate of the covariance matrix; try different starting values") w <- .weightFct(res$coefficients, x, P$vcov) attr(x, "weight")$WSpec$sandwich$bw <- attr(w,"Spec")$bw P$weightMessage <- "Weights for kernel estimate of the covariance are fixed and based on the first step estimate of Theta" } else { res <- list() } if (P$optfct == "optim") { res2 <- optim(P$t0, .objCue, x = x, type = P$vcov, ...) } if (P$optfct == "constrOptim") { if (!any(c("ui","ci") %in% names(list(...)))) stop("You must specify ui and ci when optfct is set to constrOptim") argDots <- list(...) ui <- argDots$ui ci <- argDots$ci argDots$ui <- NULL argDots$ci <- NULL allArgOptim <- list(theta = P$t0, f = .objCue, grad = NULL, ui = ui, ci = ci, x = x, type = P$vcov) allArgOptim <- c(allArgOptim,argDots) res2 <- do.call(constrOptim,allArgOptim) } if (P$optfct == "nlminb") { res2 <- nlminb(P$t0, .objCue, x = x, type = P$vcov, ...) res2$value <- res2$objective } if (P$optfct == "optimize") { res2 <- optimize(.objCue,P$t0, x = x, type = P$vcov, ...) res2$par <- res2$minimum res2$value <- res2$objective } z = list(coefficients=res2$par,objective=res2$value, k=k, k2=k2, n=n, q=q, df=df, initTheta=initTheta) if (any(P$optfct == c("optim", "constrOptim"))) { z$algoInfo <- list(convergence = res2$convergence, counts = res2$counts, message = res2$message) z$InitialAlgoInfo <- list(convergence = res$algoInfo$convergence, counts = res$algoInfo$counts, message = res$algoInfo$message) } else if (P$optfct == "nlminb") { z$algoInfo <- list(convergence = res2$convergence, counts = res2$evaluations, message = res2$message) z$InitialAlgoInfo <- list(convergence = res$algoInfo$convergence, counts = res$algoInfo$evaluations, message = res$algoInfo$message) } } z$dat <- x z$gradv <- P$gradv z$gt <- P$g(z$coefficients, x) z$w0 <- .weightFct(z$coefficients, x, P$vcov) z$iid <- P$iid z$g <- P$g z$cue <- list(weights=P$fixedKernW,message=P$weightMessage) names(z$coefficients) <- P$namesCoef if (is.null(colnames(z$gt))) colnames(z$gt) <- paste("gt",1:ncol(z$gt),sep="") z$WSpec <- P$WSpec z$specMod <- P$specMod class(z) <- paste(P$TypeGmm, ".res", sep = "") return(z) } momentEstim.baseGel.modFormula <- function(object, ...) { P <- object g <- P$g q <- attr(P$x, "q") n <- attr(P$x, "n") k <- attr(P$x, "k") df <- q-k*P$x$ny l0Env <- new.env() assign("l0",rep(0,q),envir=l0Env) if (!P$constraint) { if (P$optfct == "optim") res <- optim(P$tet0, .thetf, P = P, l0Env = l0Env, ...) if (P$optfct == "nlminb") res <- nlminb(P$tet0, .thetf, P = P, l0Env = l0Env, ...) if (P$optfct == "optimize") { res <- optimize(.thetf, P$tet0, P = P, l0Env = l0Env, ...) res$par <- res$minimum res$convergence <- "There is no convergence code for optimize" } } if(P$constraint) res <- constrOptim(P$tet0, .thetf, grad = NULL, P = P, l0Env = l0Env, ...) All <- .thetf(res$par, P, "all",l0Env = l0Env) gt <- All$gt rlamb <- All$lambda z <- list(coefficients = res$par, lambda = rlamb$lambda, conv_lambda = rlamb$conv, conv_par = res$convergence, dat=P$x) z$type <- P$type z$gt <- gt pt <- .getImpProb(z$gt, z$lambda, P$type, P$k1, P$k2) z$pt <- c(pt) z$conv_moment <- attr(pt, "conv_moment") z$conv_pt <- attr(pt, "conv_pt") z$objective <- All$obj z$call <- P$call z$k1 <- P$k1 z$k2 <- P$k2 z$CGEL <- P$CGEL z$typeDesc <- P$typeDesc z$specMod <- P$specMod z$df <- df names(z$coefficients) <- object$namesCoef if (P$onlyCoefficients) return(z[c("coefficients","lambda","conv_lambda","conv_par","objective")]) if (!is.null(object$namesgt)) { colnames(z$gt) <- object$namesgt } else { colnames(z$gt) <- paste("g",1:ncol(z$gt), sep="") } names(z$lambda) <- paste("Lam(", colnames(z$gt), ")", sep="") if (!is.null(attr(P$x,"eqConst")) & P$allArg$eqConstFullVcov) { eqConst <- attr(P$x,"eqConst")$eqConst coef <- rep(0,length(eqConst[,1])+length(z$coefficients)) ncoef <- rep("",length(eqConst[,1])+length(z$coefficients)) coef[-eqConst[,1]] <- z$coefficients ncoef[-eqConst[,1]] <- names(z$coefficients) coef[eqConst[,1]] <- eqConst[,2] ncoef[eqConst[,1]] <- rownames(eqConst) names(coef) <- ncoef z$coefficients <- coef attr(P$x, "k") <- attr(P$x, "k") + nrow(eqConst) z$df <- z$df - nrow(eqConst) attr(P$x,"eqConst") <- NULL z$specMod <- paste(z$specMod, "** Note: Covariance matrix computed for all coefficients based on restricted values \n Tests non-valid**\n\n") } if(P$gradvf) G <- P$gradv(z$coefficients, P$x) else G <- P$gradv(z$coefficients, P$x, z$pt) allVcov <- try(.vcovGel(gt, G, P$k1, P$k2, P$bwVal, z$pt), silent=TRUE) if (any(class(allVcov) == "try-error")) { z$vcov_par <- matrix(NA, length(z$coefficients), length(z$coefficients)) z$vcov_lambda <- matrix(NA, length(z$lambda), length(z$lambda)) z$khat <- NULL warning("Cannot compute the covariance matrices") } else { z <- c(z, allVcov) } z$weights <- P$w z$bwVal <- P$bwVal names(z$bwVal) <- "Bandwidth" dimnames(z$vcov_par) <- list(names(z$coefficients), names(z$coefficients)) dimnames(z$vcov_lambda) <- list(names(z$lambda), names(z$lambda)) tmp <- .residuals(z$coefficients, P$x) z$fitted.values <- tmp$yhat z$residuals <- tmp$residuals z$dat <- P$x z$terms <- P$x$mt if(P$model) z$model <- P$x$mf if(P$X) z$x <- as.matrix(P$x$x[,(P$x$ny+1):(P$x$ny+P$x$k)]) if(P$Y) z$y <- as.matrix(P$x$x[,1:P$x$ny]) class(z) <- paste(P$TypeGel, ".res", sep = "") z$allArg <- P$allArg return(z) } momentEstim.baseGel.mod <- function(object, ...) { P <- object x <- P$x q <- attr(x, "q") n <- attr(x, "n") df <- q - attr(x, "k") l0Env <- new.env() assign("l0",rep(0,q),envir=l0Env) if (!P$constraint) { if (P$optfct == "optim") res <- optim(P$tet0, .thetf, P = P, l0Env = l0Env, ...) if (P$optfct == "nlminb") res <- nlminb(P$tet0, .thetf, P = P, l0Env = l0Env, ...) if (P$optfct == "optimize") { res <- optimize(.thetf, P$tet0, P = P, l0Env = l0Env, ...) res$par <- res$minimum res$convergence <- "There is no convergence code for optimize" } } if(P$constraint) res <- constrOptim(P$tet0, .thetf, grad = NULL, P = P,l0Env = l0Env, ...) All <- .thetf(res$par, P, "all",l0Env = l0Env) gt <- All$gt rlamb <- All$lambda z <- list(coefficients = res$par, lambda = rlamb$lambda, conv_lambda = rlamb$conv, conv_par = res$convergence, dat=x) z$type <- P$type z$gt <- gt pt <- .getImpProb(z$gt, z$lambda, P$type, P$k1, P$k2) z$pt <- c(pt) z$conv_moment <- attr(pt, "conv_moment") z$conv_pt <- attr(pt, "conv_pt") z$objective <- All$obj names(z$coefficients) <- P$namesCoef z$specMod <- P$specMod z$df <- df if (!is.null(object$namesgt)) { colnames(z$gt) <- object$namesgt } else { colnames(z$gt) <- paste("g",1:ncol(z$gt), sep="") } names(z$lambda) <- paste("Lam(", colnames(z$gt), ")", sep="") if (!is.null(attr(x,"eqConst")) & P$allArg$eqConstFullVcov) { eqConst <- attr(x,"eqConst")$eqConst coef <- rep(0,length(eqConst[,1])+length(z$coefficients)) ncoef <- rep("",length(eqConst[,1])+length(z$coefficients)) coef[-eqConst[,1]] <- z$coefficients ncoef[-eqConst[,1]] <- names(z$coefficients) coef[eqConst[,1]] <- eqConst[,2] ncoef[eqConst[,1]] <- rownames(eqConst) names(coef) <- ncoef z$coefficients <- coef attr(x, "k") <- attr(x, "k") + nrow(eqConst) z$df <- z$df - nrow(eqConst) attr(x,"eqConst") <- NULL z$specMod <- paste(z$specMod, "** Note: Covariance matrix computed for all coefficients based on restricted values \n Tests non-valid**\n\n") } if (P$onlyCoefficients) return(z[c("coefficients", "lambda", "conv_lambda", "conv_par", "objective")]) if(P$gradvf) G <- P$gradv(z$coefficients, x) else G <- P$gradv(z$coefficients, x, z$pt) z$G <- G allVcov <- try(.vcovGel(gt, G, P$k1, P$k2, P$bwVal, z$pt), silent=TRUE) if (any(class(allVcov) == "try-error")) { z$vcov_par <- matrix(NA, length(z$coefficients), length(z$coefficients)) z$vcov_lambda <- matrix(NA, length(z$lambda), length(z$lambda)) z$khat <- NULL warning("Cannot compute the covariance matrices") } else { z <- c(z, allVcov) } z$weights <- P$w z$bwVal <- P$bwVal names(z$bwVal) <- "Bandwidth" dimnames(z$vcov_par) <- list(names(z$coefficients), names(z$coefficients)) dimnames(z$vcov_lambda) <- list(names(z$lambda), names(z$lambda)) if(P$X) z$x <- x z$call <- P$call z$k1 <- P$k1 z$k2 <- P$k2 z$CGEL <- P$CGEL z$typeDesc <- P$typeDesc z$allArg <- P$allArg class(z) <- paste(P$TypeGel, ".res", sep = "") return(z) } momentEstim.fixedW.formula <- function(object, ...) { P <- object g <- P$g dat <- P$x n <- attr(dat, "n") q <- attr(dat, "q") k <- attr(dat, "k") k2 <- k*dat$ny df <- q-k*dat$ny w = .weightFct(NULL, dat, "fixed") if(!all(dim(w) == c(q,q))) stop("The matrix of weights must be qxq") eigenW <- svd(w)$d if(!is.double(eigenW)) warning("The matrix of weights is not strictly positive definite") if(is.double(eigenW)) { if(any(eigenW<=0)) warning("The matrix of weights is not strictly positive definite") } res2 <- .tetlin(dat, w) z = list(coefficients = res2$par, objective = res2$value, dat=dat, k=k, k2=k2, n=n, q=q, df=df, df.residual = (n-k)) z$gt <- g(z$coefficients, dat) tmp <- .residuals(z$coefficients, dat) z$fitted.values <- tmp$yhat z$residuals <- tmp$rediduals z$terms <- dat$mt if(P$model) z$model <- dat$mf if(P$X) z$x <- as.matrix(dat$x[,(dat$ny+1):(dat$ny+dat$k)]) if(P$Y) z$y <- as.matrix(dat$x[,1:dat$ny]) z$dat <- dat z$terms <- dat$mt if(P$model) z$model <- dat$mf z$gradv <- P$gradv z$iid <- P$iid z$g <- P$g z$WSpec <- P$WSpec names(z$coefficients) <- P$namesCoef colnames(z$gt) <- P$namesgt z$specMod <- P$specMod class(z) <- paste(P$TypeGmm,".res",sep="") return(z) } momentEstim.fixedW <- function(object, ...) { P <- object x <- P$x n <- attr(x, "n") q <- attr(x, "q") k <- attr(x, "k") k2 <- k df <- q - k w = .weightFct(NULL, x, "fixed") if(!all(dim(w) == c(q,q))) stop("The matrix of weights must be qxq") eigenW <- svd(w)$d if(!is.double(eigenW)) warning("The matrix of weights is not strictly positive definite") if(is.double(eigenW)) { if(any(eigenW<=0)) warning("The matrix of weights is not strictly positive definite") } chkOptim <- any(P$optfct == c("optim", "constrOptim")) if (chkOptim) { if (P$gradvf) { gr2 <- function(thet, x, w) { gt <- .momentFct(thet, x) Gbar <- .DmomentFct(thet, x) gbar <- as.vector(colMeans(gt)) INV <- attr(w, "inv") if (INV) obj <- crossprod(Gbar, solve(w, gbar)) else obj <- crossprod(Gbar,w)%*%gbar return(obj*2) } } else { gr2 <- NULL } } if (P$optfct == "optim") { argDots <- list(...) allArgOptim <- list(par = P$t0, fn = .obj1, gr = gr2, x = x, w = w) argDots$gr <- NULL allArgOptim <- c(allArgOptim,argDots) res2 <- do.call(optim,allArgOptim) } if (P$optfct == "constrOptim") { if (!any(c("ui","ci") %in% names(list(...)))) stop("You must specify ui and ci when optfct is set to constrOptim") argDots <- list(...) ui <- argDots$ui ci <- argDots$ci argDots$ui <- NULL argDots$ci <- NULL allArgOptim <- list(theta = P$t0, f = .obj1, grad = gr2, ui = ui, ci = ci, x = x, w = w) allArgOptim <- c(allArgOptim,argDots) res2 <- do.call(constrOptim,allArgOptim) } if (P$optfct == "nlminb") { res2 <- nlminb(P$t0, .obj1, x = P$x, w = w, ...) res2$value <- res2$objective } if (P$optfct == "optimize") { res2 <- optimize(.obj1, P$t0, x = P$x, w = w, ...) res2$par <- res2$minimum res2$value <- res2$objective } z = list(coefficients = res2$par, objective = res2$value, k=k, k2=k2, n=n, q=q, df=df) if (chkOptim) z$algoInfo <- list(convergence = res2$convergence, counts = res2$counts, message = res2$message) else if(P$optfct == "nlminb") z$algoInfo <- list(convergence = res2$convergence, counts = res2$evaluations, message = res2$message) z$dat <- P$x z$gt <- P$g(z$coefficients, P$x) z$gradv <- P$gradv z$iid <- P$iid z$g <- P$g z$WSpec <- P$WSpec names(z$coefficients) <- P$namesCoef if (is.null(colnames(z$gt))) colnames(z$gt) <- paste("gt",1:ncol(z$gt),sep="") z$specMod <- P$specMod class(z) <- paste(P$TypeGmm,".res",sep="") return(z) } momentEstim.baseGel.eval <- function(object, ...) { P <- object q <- attr(P$x, "q") n <- attr(P$x, "n") l0Env <- new.env() assign("l0",rep(0,q),envir=l0Env) All <- .thetf(P$tet0, P, "all",l0Env = l0Env) gt <- All$gt rlamb <- All$lambda z <- list(coefficients = P$tet0, lambda = rlamb$lambda, conv_lambda = rlamb$conv, conv_par = NULL, dat=P$x) z$type <- P$type z$gt <- gt pt <- .getImpProb(z$gt, z$lambda, P$type, P$k1, P$k2) z$pt <- c(pt) z$conv_moment <- attr(pt, "conv_moment") z$conv_pt <- attr(pt, "conv_pt") z$objective <- All$obj z$call <- P$call z$k1 <- P$k1 z$k2 <- P$k2 z$CGEL <- P$CGEL z$typeDesc <- paste(P$typeDesc, " (Eval only, tests non-valid) ", sep="") z$specMod <- P$specMod names(z$coefficients) <- P$namesCoef z$df <- length(z$lambda) - length(z$coefficients) if (!is.null(object$namesgt)) { colnames(z$gt) <- object$namesgt } else { colnames(z$gt) <- paste("g",1:ncol(z$gt), sep="") } names(z$lambda) <- paste("Lam(", colnames(z$gt), ")", sep="") if(P$gradvf) G <- P$gradv(z$coefficients, P$x) else G <- P$gradv(z$coefficients, P$x, z$pt) z$G <- G allVcov <- .vcovGel(gt, G, P$k1, P$k2, P$bwVal, z$pt) z <- c(z, allVcov) z$weights <- P$w z$bwVal <- P$bwVal names(z$bwVal) <- "Bandwidth" dimnames(z$vcov_par) <- list(names(z$coefficients), names(z$coefficients)) dimnames(z$vcov_lambda) <- list(names(z$lambda), names(z$lambda)) if (attr(P$x,"ModelType") == "linear") { tmp <- .residuals(z$coefficients, P$x) z$fitted.values <- tmp$yhat z$residuals <- tmp$residuals z$terms <- P$x$mt if(P$model) z$model <- P$x$mf if(P$X) z$x <- as.matrix(P$x$x[,(P$x$ny+1):(P$x$ny+P$x$k)]) if(P$Y) z$y <- as.matrix(P$x$x[,1:P$x$ny]) } return(z) } gmm/R/getModel.R0000644000176200001440000006505614342733440013136 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ getModel <- function(object, ...) { UseMethod("getModel") } getModel.tsls <- function(object, ...) { obj <- getModel.baseGmm(object, ...) return(obj) } getModel.sysGmm <- function(object, ...) { if (object$commonCoef & !is.null(object$crossEquConst)) { object$commonCoef <- FALSE warning("When crossEquConst is not NULL, commonCoef=TRUE is ignore and set to FALSE") } object$allArg <- c(object, ...) object$formula <- list(g=object$g, h=object$h) if (!is.list(object$g)) stop("g must be list of formulas") if (length(object$g) == 1) stop("For single equation GMM, use the function gmm()") if (!all(sapply(1:length(object$g), function(i) is(object$g[[i]], "formula")))) stop("g must be a list of formulas") if (!is.list(object$h)) { if(!is(object$h, "formula")) stop("h is either a list of formulas or a formula") else object$h <- list(object$h) } else { if (!all(sapply(1:length(object$h), function(i) is(object$h[[i]], "formula")))) stop("h is either a list of formulas or a formula") } if (length(object$h) == 1) { object$h <- rep(object$h, length(object$g)) typeDesc <- "System Gmm with common instruments" typeInst <- "Common" } else { if (length(object$h) != length(object$g)) stop("The number of formulas in h should be the same as the number of formulas in g, \nunless the instruments are the same in each equation, \nin which case the number of equations in h should be 1") typeDesc <- "System Gmm" typeInst <- "nonCommon" } if (object$commonCoef) typeDesc <- paste(typeDesc, " (Common Coefficients)") dat <- lapply(1:length(object$g), function(i) try(getDat(object$g[[i]], object$h[[i]], data = object$data, error=!object$commonCoef), silent=TRUE)) chk <- sapply(1:length(dat), function(i) any(class(dat[[i]]) == "try-error")) chk <- which(chk) mess <- vector() for (i in chk) { mess <- paste(mess, "\nSystem of equations:", i, "\n###############\n", sep="") mess <- paste(mess, attr(dat[[i]], "condition")[[1]]) } if (length(chk)>0) stop(mess) if (is.null(names(object$g))) names(dat) <- paste("System_", 1:length(dat), sep="") else names(dat) <- names(object$g) object$gradv <- .DmomentFct_Sys object$formula <- list(g=object$g, h=object$h) if (!all(sapply(1:length(dat), function(i) dat[[i]]$ny == 1))) stop("The number of dependent variable must be one in every equation") clname <- "sysGmm.twoStep.formula" object$x <- dat namex <- lapply(1:length(dat), function(i) colnames(dat[[i]]$x[,2:(1+dat[[i]]$k), drop=FALSE])) nameh <- lapply(1:length(dat), function(i) colnames(dat[[i]]$x[,(2+dat[[i]]$k):(1+dat[[i]]$k+dat[[i]]$nh), drop=FALSE])) namey <- lapply(1:length(dat), function(i) colnames(dat[[i]]$x[,1, drop=FALSE])) object$namesCoef <- namex namesgt <- lapply(1:length(dat), function(i) paste(namey[[i]], "_", nameh[[i]], sep="")) object$namesgt <- namesgt object$namesy <- namey attr(object$x,"ModelType") <- "linear" #for (i in 1:length(object$x)) # attr(object$x[[i]], c("linear")) <- attr(object$x, "modelType") attr(object$x, "k") <- lapply(1:length(dat), function(i) length(object$namesCoef[[i]])) attr(object$x, "q") <- lapply(1:length(dat), function(i) length(object$namesgt[[i]])) attr(object$x, "n") <- lapply(1:length(dat), function(i) nrow(object$x[[i]]$x)) object$TypeGmm <- class(object) attr(object$x, "weight") <- list(w=object$weightsMatrix, centeredVcov=object$centeredVcov) attr(object$x, "weight")$WSpec <- list() attr(object$x, "weight")$WSpec$sandwich <- list(kernel = object$kernel, bw = object$bw, prewhite = object$prewhite, ar.method = object$ar.method, approx = object$approx, tol = object$tol) attr(object$x, "weight")$vcov <- object$vcov k <- lapply(1:length(dat), function(i) dat[[i]]$k) q <- lapply(1:length(dat), function(i) dat[[i]]$nh) df <- lapply(1:length(dat), function(i) dat[[i]]$nh-dat[[i]]$k) k2 <- do.call(c,k) if (object$commonCoef | !is.null(object$crossEquConst)) { if (!all(k2==k2[1])) stop("In a common coefficient model the number of regressors is the same in each equation") if (object$commonCoef) totK <- k2[1] else totK <- length(dat)*(k2[1]-length(object$crossEquConst)) + length(object$crossEquConst) if (sum(do.call(c,q))=length(object$t0)) stop("Too many constraints; use evalGmm() if all coefficients are fixed") if (is.character(object$eqConst)) { if (is.null(names(object$t0))) stop("t0 must be a named vector if you want eqConst to be names") if (any(!(object$eqConst %in% names(object$t0)))) stop("Wrong coefficient names in eqConst") object$eqConst <- sort(match(object$eqConst,names(object$t0))) } restTet <- object$t0[object$eqConst] obj$t0 <- object$t0[-object$eqConst] object$eqConst <- cbind(object$eqConst,restTet) } else { if (is.null(dim(object$eqConst))) stop("When t0 is not provided, eqConst must be a 2xq matrix") } attr(obj$x, "eqConst") <- list(eqConst = object$eqConst) rownames(attr(obj$x, "eqConst")$eqConst) <- obj$namesCoef[object$eqConst[,1]] object$eqConst <- attr(obj$x, "eqConst")$eqConst if(is(object$g, "formula")) { if (obj$x$ny>1) stop("Constrained GMM not implemented yet for system of equations") if (obj$x$k<=0) stop("Nothing to estimate") } obj$eqConst <- object$eqConst attr(obj$x, "k") <- attr(obj$x, "k")-nrow(object$eqConst) obj$namesCoef <- obj$namesCoef[-object$eqConst[,1]] obj$type <- paste(obj$type,"(with equality constraints)",sep=" ") mess <- paste(rownames(object$eqConst), " = " , object$eqConst[,2], "\n",collapse="") mess <- paste("#### Equality constraints ####\n",mess,"##############################\n\n",sep="") obj$specMod <- mess return(obj) } getModel.baseGmm <- function(object, ...) { object$allArg <- c(object, list(...)) if(is(object$g, "formula")) { object$gradv <- .DmomentFct object$gradvf <- FALSE dat <- getDat(object$g, object$x, data = object$data) if(is.null(object$weightsMatrix)) { clname <- paste(class(object), ".", object$type, ".formula", sep = "") } else { clname <- "fixedW.formula" object$type <- "One step GMM with fixed W" } object$x <- dat object$gform<-object$g namex <- colnames(dat$x[,(dat$ny+1):(dat$ny+dat$k), drop=FALSE]) nameh <- colnames(dat$x[,(dat$ny+dat$k+1):(dat$ny+dat$k+dat$nh), drop=FALSE]) if (dat$ny > 1) { namey <- colnames(dat$x[,1:dat$ny, drop=FALSE]) object$namesCoef <- paste(rep(namey, dat$k), "_", rep(namex, rep(dat$ny, dat$k)), sep = "") object$namesgt <- paste(rep(namey, dat$nh), "_", rep(nameh, rep(dat$ny, dat$nh)), sep = "") } else { object$namesCoef <- namex object$namesgt <- nameh } attr(object$x,"ModelType") <- "linear" attr(object$x, "k") <- object$x$k attr(object$x, "q") <- object$x$ny*object$x$nh attr(object$x, "n") <- NROW(object$x$x) attr(object$x, "namesgt") <- object$namesgt } else { attr(object$x,"ModelType") <- "nonlinear" attr(object$x, "momentfct") <- object$g if (object$optfct == "optimize") attr(object$x, "k") <- 1 else attr(object$x, "k") <- length(object$t0) attr(object$x, "q") <- NCOL(gt <- object$g(object$t0, object$x)) attr(object$x, "n") <- NROW(gt) if (object$optfct == "optimize") { object$namesCoef <- "Theta1" } else { if(is.null(names(object$t0))) object$namesCoef <- paste("Theta[" ,1:attr(object$x, "k"), "]", sep = "") else object$namesCoef <- names(object$t0) } if(is.null(object$weightsMatrix)) { clname <- paste(class(object), "." ,object$type, sep = "") } else { clname <- "fixedW" object$type <- "One step GMM with fixed W" attr(object$x, "weight")$w <- object$weightsMatrix } if (!is.function(object$gradv)) { object$gradvf <- FALSE } else { attr(object$x, "gradv") <- object$gradv object$gradvf <- TRUE } object$gradv <- .DmomentFct } object$TypeGmm <- class(object) attr(object$x, "weight") <- list(w=object$weightsMatrix, centeredVcov=object$centeredVcov) attr(object$x, "weight")$WSpec <- list() attr(object$x, "weight")$WSpec$sandwich <- list(kernel = object$kernel, bw = object$bw, prewhite = object$prewhite, ar.method = object$ar.method, approx = object$approx, tol = object$tol) attr(object$x, "weight")$vcov <- object$vcov attr(object$x, "mustar") <- object$mustar object$g <- .momentFct class(object) <- clname return(object) } getModel.constGel <- function(object, ...) { class(object) <- "baseGel" obj <- getModel(object) if (!is.null(dim(object$eqConst))) stop("eqConst must be a vector which indicates which parameters to fix") if (length(object$eqConst)>=length(object$tet0)) stop("Too many constraints; use evalGel() if all coefficients are fixed") if (is.character(object$eqConst)) { if (is.null(names(object$tet0))) stop("tet0 must be a named vector if you want eqConst to be names") if (any(!(object$eqConst %in% names(object$tet0)))) stop("Wrong coefficient names in eqConst") object$eqConst <- sort(match(object$eqConst,names(object$tet0))) } restTet <- object$tet0[object$eqConst] obj$tet0 <- object$tet0[-object$eqConst] object$eqConst <- cbind(object$eqConst,restTet) attr(obj$x, "eqConst") <- list(eqConst = object$eqConst) rownames(attr(obj$x, "eqConst")$eqConst) <- obj$namesCoef[object$eqConst[,1]] object$eqConst <- attr(obj$x, "eqConst")$eqConst if(is(object$g, "formula")) { if (obj$x$ny>1) stop("Constrained GMM not implemented yet for system of equations") } obj$eqConst <- object$eqConst attr(obj$x, "k") <- attr(obj$x, "k")-nrow(object$eqConst) obj$namesCoef <- obj$namesCoef[-object$eqConst[,1]] obj$typeDesc <- paste(obj$typeDesc,"(with equality constraints)",sep=" ") mess <- paste(rownames(object$eqConst), " = " , object$eqConst[,2], "\n",collapse="") mess <- paste("#### Equality constraints ####\n",mess,"##############################\n\n",sep="") obj$specMod <- mess return(obj) } getModel.baseGel <- function(object, ...) { object$allArg <- c(object, list(...)) object$allArg$weights <- NULL object$allArg$call <- NULL if(is(object$g, "formula")) { dat <- getDat(object$g, object$x, data = object$data) k <- dat$k if (is.null(object$tet0)) { if (!is.null(object$eqConst)) stop("You have to provide tet0 with equality constrains") if (object$optfct == "optimize") stop("For optimize, you must provide the 2x1 vector tet0") res0 <- gmm(object$g, object$x, data=object$data) object$tet0 <- res0$coefficients if (object$smooth) gt <- res0$gt } else { if (object$optfct == "optimize") { if (k != 1) stop("optimize() is for univariate optimization") if (length(object$tet0) != 2) stop("For optimize(), tet0 must be a 2x1 vector") } else { if (k != length(object$tet0)) stop("The number of starting values does not correspond to the number of regressors") } if (object$smooth) gt <- gmm(object$g, object$x, data=object$data)$gt } clname <- paste(class(object), ".modFormula", sep = "") object$gradv <- .DmomentFct object$gradvf <- FALSE object$x <- dat object$gform<-object$g namex <- colnames(dat$x[,(dat$ny+1):(dat$ny+dat$k), drop=FALSE]) nameh <- colnames(dat$x[,(dat$ny+dat$k+1):(dat$ny+dat$k+dat$nh), drop=FALSE]) if (dat$ny > 1) { namey <- colnames(dat$x[,1:dat$ny]) namesCoef <- paste(rep(namey, dat$k), "_", rep(namex, rep(dat$ny, dat$k)), sep = "") object$namesgt <- paste(rep(namey, dat$nh), "_", rep(nameh, rep(dat$ny, dat$nh)), sep = "") } else { namesCoef <- namex object$namesgt <- nameh } if (is.null(names(object$tet0))) object$namesCoef <- namesCoef else object$namesCoef <- names(object$tet0) attr(object$x,"ModelType") <- "linear" attr(object$x, "k") <- k attr(object$x, "q") <- object$x$ny*object$x$nh attr(object$x, "n") <- NROW(object$x$x) } else { if (is.null(object$tet0)) stop("You must provide the starting values tet0 for nonlinear moments") if(any(object$optfct == c("optim", "nlminb"))) k <- length(object$tet0) else k <- 1 attr(object$x,"ModelType") <- "nonlinear" attr(object$x, "momentfct") <- object$g attr(object$x, "k") <- k attr(object$x, "q") <- NCOL(gt <- object$g(object$tet0, object$x)) attr(object$x, "n") <- NROW(gt) if(is.null(names(object$tet0))) object$namesCoef <- paste("Theta[" ,1:attr(object$x, "k"), "]", sep = "") else object$namesCoef <- names(object$tet0) if (!is.function(object$gradv) | object$smooth) { object$gradvf <- FALSE } else { attr(object$x, "gradv") <- object$gradv object$gradvf <- TRUE } object$gradv <- .DmomentFct if (object$smooth) gt <- gmm(object$g, object$x, object$tet0, wmatrix = "ident", ...)$gt clname <- paste(class(object), ".mod", sep = "") } if (object$smooth) { if (is.function(object$gradv)) warning("The provided gradv is not used when smooth=TRUE", call. = FALSE) if(object$kernel == "Truncated") { object$wkernel <- "Bartlett" object$k1 <- 2 object$k2 <- 2 } if(object$kernel == "Bartlett") { object$wkernel <- "Parzen" object$k1 <- 1 object$k2 <- 2/3 } gt <- scale(gt, scale=FALSE) class(gt) <- "gmmFct" if (is.function(object$bw)) object$bwVal <- object$bw(gt, kernel = object$wkernel, prewhite = object$prewhite, ar.method = object$ar.method, approx = object$approx) else object$bwVal <- object$bw object$w <- smoothG(gt, bw = object$bwVal, kernel = object$kernel, tol = object$tol_weights)$kern_weights attr(object$x,"smooth") <- list(bw=object$bwVal, w=object$w, kernel=object$kernel) } else { object$k1 <- 1 object$k2 <- 1 object$w <- kernel(1) object$bwVal <- 1 } object$g <- .momentFct object$CGEL <- object$alpha object$typeDesc <- object$type class(object) <- clname return(object) } getModel.ateGel <- function(object, ...) { object$allArg <- c(object, list(...)) object$allArg$weights <- NULL object$allArg$call <- NULL if(is(object$g, "formula")) { dat <- getDat(object$g, object$x, data = object$data) if (!is.null(object$w)) if (is(object$w, "formula")) { dat$w <- model.matrix(object$w, object$data)[,-1,drop=FALSE] } else { stop("w must be a formula") } if (dat$ny>1 | dat$ny==0) stop("You need one and only one dependent variable") k <- dat$k if (k>2 & object$momType=="ATT") stop("Cannot compute ATT with multiple treatments") if (attr(dat$mt, "intercept")!=1) stop("An intercept is needed to compute the treatment effect") if (!all(dat$x[,3:(k+1)] %in% c(0,1))) stop("The treatment indicators can only take values 0 or 1") if (colnames(dat$x)[k+2] == "(Intercept)") { dat$x <- dat$x[,-(k+2)] dat$nh <- dat$nh-1 } if (!is.null(object$popMom)) { if (length(object$popMom)!=dat$nh) stop("The dim. of the population moments does not match the dim. of the vector of covariates") } if (is.null(object$tet0)) { if (is.null(dat$w)) { tet0 <- lm(dat$x[,1]~dat$x[,3:(k+1)])$coef } else { tet0 <- lm(dat$x[,1]~dat$x[,3:(k+1)]+dat$w)$coef } tet0 <- c(tet0, colMeans(dat$x[,3:(k+1),drop=FALSE])) names(tet0) <- NULL object$tet0 <- tet0 } else { ntet0 <- 2*k-1 + ifelse(is.null(dat$w), 0, ncol(dat$w)) if (length(object$tet0) != ntet0) stop("tet0 should have a length equal to 2x(number of treatments)+1+number of w's if any") } if (object$family != "linear") { if (any(!(dat$x[,1]%in%c(0,1)))) stop("For logit or probit, Y can only take 0s and 1s") family <- binomial(link=object$family) if (object$family == "logit") family$mu.eta2 <- function(x, family) family$mu.eta(x)*(1-2*family$linkinv(x)) else family$mu.eta2 <- function(x, family) -x*family$mu.eta(x) } else { family <- NULL } q <- dat$nh + 2*k+1 if (object$momType != "bal" | !is.null(object$popMom)) q <- q+dat$nh if (!is.null(dat$w)) { q <- q+ncol(dat$w) namew <- colnames(dat$w) } else { namew <- NULL } object$gradv <- .DmomentFctATE object$x <- dat object$gradvf <- FALSE object$gform<-object$g namex <- colnames(dat$x[,2:(k+1)]) nameh <- colnames(dat$x[,(k+2):ncol(dat$x), drop=FALSE]) namesCoef <- c(namex, namew, paste("TreatProb", 1:(k-1), sep="")) namesgt <- paste(rep(paste("Treat", 1:(k-1), sep=""), rep(dat$nh, k-1)), "_", rep(nameh, k-1), sep="") object$namesgt <- c(namesCoef,namesgt) if (object$momType != "bal" | !is.null(object$popMom)) object$namesgt <- c(object$namesgt, paste("bal_", nameh, sep="")) if (is.null(names(object$tet0))) object$namesCoef <- namesCoef else object$namesCoef <- names(object$tet0) attr(object$x,"ModelType") <- "linear" attr(object$x, "k") <- k attr(object$x, "q") <- q attr(object$x, "n") <- nrow(dat$x) attr(object$x, "momType") <- object$momType attr(object$x, "popMom") <- object$popMom attr(object$x, "family") <- family } else { stop("Not implemented yet for nonlinear regression") } if (object$momType == "ATT") metname <- "ATT" else metname <- "ATE" if (!is.null(object$popMom)) { metname2 <- " with balancing based on population moments" } else { if (object$momType == "balSample") metname2 <- " with balancing based on the sample moments" else metname2 <- " with unrestricted balancing" } metname3 <- paste("\nMethod: ", object$type, sep="") if (!is.null(family)) metname3 <- paste(metname3, ", Family: Binomial with ", family$link, " link", sep="") clname <- "baseGel.mod" object$k1 <- 1 object$k2 <- 1 object$w <- kernel(1) object$bwVal <- 1 object$g <- .momentFctATE object$CGEL <- object$alpha object$typeDesc <- paste(metname, metname2, metname3, sep="") class(object) <- clname return(object) } gmm/R/ategel.R0000644000176200001440000003442514247643115012636 0ustar liggesusersATEgel <- function(g, balm, w=NULL, y=NULL, treat=NULL, tet0=NULL, momType=c("bal","balSample","ATT"), popMom = NULL, family=c("linear","logit", "probit"), type = c("EL", "ET", "CUE", "ETEL", "HD", "ETHD","RCUE"), tol_lam = 1e-9, tol_obj = 1e-9, tol_mom = 1e-9, maxiterlam = 100, optfct = c("optim", "nlminb"), optlam = c("nlminb", "optim", "iter", "Wu"), data=NULL, Lambdacontrol = list(), model = TRUE, X = FALSE, Y = FALSE, ...) { type <- match.arg(type) optfct <- match.arg(optfct) optlam <- match.arg(optlam) momType <- match.arg(momType) family <- match.arg(family) TypeGel <- "ateGel" all_args <- list(g = g, x = balm, w=w, y=y, treat=treat, tet0 = tet0, type = type, tol_lam = tol_lam, tol_obj = tol_obj, tol_mom = tol_mom, maxiterlam = maxiterlam, optfct = optfct, optlam = optlam, model = model, X = X, Y = Y, TypeGel = TypeGel, call = match.call(), Lambdacontrol = Lambdacontrol, data = data, constraint=FALSE, kernel = "Truncated", bw = bwAndrews, approx = "AR(1)", prewhite = 1, ar.method = "ols", tol_weights = 1e-7, alpha = NULL, eqConst = NULL, eqConstFullVcov = FALSE, momType=momType, popMom=popMom, family=family, onlyCoefficients=FALSE) class(all_args)<-TypeGel Model_info<-getModel(all_args) z <- momentEstim(Model_info, ...) class(z) <- c("ategel", "gel") res <- try(.vcovate(z), silent=TRUE) if (any(class(res)=="try-error")) { warning("Could not compute the robust-to misspecification standard errors") z$robVcov <- FALSE } else { z$vcov_par <- res$vcov_par z$vcov_lambda <- res$vcov_lambda z$robVcov <- TRUE } return(z) } .momentFctATE <- function(tet, dat) { x <- dat$x k <- dat$k if (is.null(dat$w)) { Z <- x[,2:(k+1),drop=FALSE] } else { Z <- cbind(x[,2:(k+1)], dat$w) } tetz <- tet[1:ncol(Z)] tetb <- tail(tet, k-1) ZT <- c(Z%*%tetz) if (is.null(attr(dat, "family"))) e <- x[,1] - ZT else e <- x[,1] - attr(dat, "family")$linkinv(ZT) gt1 <- e * Z gt2 <- sweep(x[,3:(k+1),drop=FALSE],2,tetb,"-") gt3 <- lapply(1:(k-1), function(i) gt2[,i]*x[,-(1:(k+1))]) gt3 <- do.call(cbind, gt3) gt <- cbind(gt1,gt2,gt3) if (is.null(attr(dat, "popMom"))) { if (attr(dat, "momType") == "balSample") { momB <- scale(x[,-(1:(k+1)),drop=FALSE], scale=FALSE) gt <- cbind(gt, momB) } if (attr(dat, "momType") == "ATT") { momB <- sweep(x[,-(1:3), drop=FALSE], 2, colMeans(x[x[,3]==1,-(1:3), drop=FALSE]), FUN="-") gt <- cbind(gt, momB) } } else { momB <- sweep(x[,-(1:(k+1)), drop=FALSE], 2, attr(dat, "popMom"), "-") gt <- cbind(gt, momB) } return(as.matrix(gt)) } .DmomentFctATE <- function(tet, dat, pt=NULL) { if (is.null(pt)) pt <- rep(1/nrow(dat$x), nrow(dat$x)) x <- dat$x k <- dat$k q <- dat$nh*(k-1)+2*k-1 if (is.null(dat$w)) { Z <- x[,2:(k+1),drop=FALSE] } else { Z <- cbind(x[,2:(k+1)], dat$w) q <- q+ncol(dat$w) } l <- ncol(Z) ntet <- length(tet) ZT <- c(Z%*%tet[1:l]) G <- matrix(0, q, ntet) if (is.null(attr(dat, "family"))) { tau <- rep(1, nrow(x)) } else { tau <- attr(dat, "family")$mu.eta(ZT) } G11 <- lapply(1:l, function(i) -colSums(pt*Z[,i]*tau*Z)) G[1:l, 1:l] <- do.call(rbind, G11) G[(l+1):ntet, (l+1):ntet] <- -sum(pt)*diag(k-1) uK <- colSums(pt*x[,-(1:(k+1)),drop=FALSE]) G[(l+k):q, (l+1):ntet] <- -kronecker(diag(k-1), uK) if (attr(dat, "momType") != "bal" | !is.null(attr(dat, "popMom"))) G <- rbind(G, matrix(0, dat$nh, ntet)) return(G) } .DmomentFctATE2 <- function(tet, dat, pt=NULL) { G <- .DmomentFctATE(tet, dat, pt) #k <- attr(dat, "k") k <- dat$k q <- nrow(G)-dat$nh if (is.null(pt)) pt <- rep(1/nrow(dat$x), nrow(dat$x)) if (attr(dat, "momType") != "bal" & is.null(attr(dat, "popMom"))) G <- cbind(G, rbind(matrix(0,q, dat$nh), -sum(pt)*diag(dat$nh))) return(G) } .psiGam <- function(object) { n <- nrow(object$dat$x) nh <- object$dat$nh lam <- object$lambda q <- length(lam) k <- attr(object$dat, "k") theta <- object$coefficients gt <- object$gt rho1 <- .rho(x=gt, lamb=lam, derive=1, type=object$type) rho2 <- .rho(x=gt, lamb=lam, derive=2, type=object$type) if (is.null(object$dat$w)) { Z <- object$dat$x[,2:(k+1)] } else { Z <- cbind(object$dat$x[,2:(k+1)], object$dat$w) } l <- ncol(Z) ZT <- c(Z%*%theta[1:l]) X <- object$dat$x[,-(1:(k+1)), drop=FALSE] family <- attr(object$dat, "family") momType <- attr(object$dat, "momType") popMom <- attr(object$dat, "popMom") if (is.null(family)) { tau1 <- rep(1, n) } else { tau1 <- family$mu.eta(ZT) tau2 <- family$mu.eta2(ZT, family) } lG1 <- sapply(1:l, function(i) -(tau1*Z[,i]*Z)%*%lam[1:l]) q2 <- nh*(k-1)+l+k-1 lamM <- matrix(lam[(l+k):q2], ncol=(k-1)) lG2 <- sapply(1:(k-1), function(i) -lam[l+i]-X%*%lamM[,i]) lG <- cbind(lG1, lG2) G <- .DmomentFctATE2(theta, object$dat, rho1) G22 <- crossprod(rho2*gt, gt)/n if (momType == "bal" | !is.null(popMom)) { Psi <- cbind(rho1*lG, rho1*gt) G11 <- crossprod(rho2*lG, lG)/n G12 <- t(G)/n + crossprod(rho2*lG, gt)/n if (!is.null(family)) { G12tmp <- lapply(1:l, function(i) colSums(-rho1*tau2*Z[,i]*c(Z%*%lam[1:l])*Z)) G12.2 <- matrix(0, nrow(G12), ncol(G12)) G12.2[1:l,1:l] <- do.call(rbind, G12tmp) G12 <- G12 + G12.2/n } Gamma <- rbind(cbind(G11, G12), cbind(t(G12), G22)) addPar <- 0 } else { lG <- cbind(lG, matrix(-tail(lam, nh), n, nh, byrow=TRUE)) G11 <- crossprod(rho2*lG, lG)/n G12 <- t(G)/n + crossprod(rho2*lG, gt)/n if (!is.null(family)) { G12tmp <- lapply(1:l, function(i) colSums(-rho1*tau2*Z[,i]*c(Z%*%lam[1:l])*Z)) G12.2 <- matrix(0, nrow(G12), ncol(G12)) G12.2[1:l,1:l] <- do.call(rbind, G12tmp) G12 <- G12 + G12.2/n } if (momType == "balSample") Xi <- rep(1,n) else Xi <- Z[,(2:k)] nj <- sum(Xi) lam2 <- -sum(rho1)*tail(lam,nh)/nj theta4 <- colSums(Xi*X)/nj G13 <- rbind(matrix(0, l+k-1, nh), -nj/n*diag(nh)) G23 <- matrix(0,q, nh) G33 <- matrix(0, nh, nh) Psi <- cbind(rho1*lG, rho1*gt, Xi*sweep(X, 2, theta4, "-")) Psi[,(l+k):(l+k+nh-1)] <- Psi[,(l+k):(l+k+nh-1)]-Xi%*%t(lam2) Gamma <- rbind(cbind(G11, G12, G13), cbind(t(G12), G22, G23), cbind(t(G13), t(G23), G33)) addPar <- nh } list(Psi=Psi, Gamma=Gamma, k=length(theta), q=q, addPar=addPar, n=n) } .vcovate <- function (object) { res <- .psiGam(object) k <- res$k q <- res$q addPar <- res$addPar qrPsi <- qr(res$Psi/sqrt(res$n)) piv <- sort.int(qrPsi$pivot, index.return=TRUE)$ix R <- qr.R(qrPsi)[,piv] T1 <- solve(res$Gamma, t(R)) V <- T1%*%t(T1)/res$n allV <- list() allV$vcov_par <- V[1:k, 1:k] allV$vcov_lambda <- V[(k+addPar+1):(k+addPar+q), (k+addPar+1):(k+addPar+q)] if (addPar > 0) { allV$vcov_Allpar <- V[1:(k+addPar), 1:(k+addPar)] allV$vcov_Alllambda <- V[-(1:(k+addPar)), -(1:(k+addPar))] } allV } vcov.ategel <- function(object, lambda = FALSE, robToMiss=TRUE, ...) { if (robToMiss) { return(vcov.gel(object, lambda)) } else { object$lambda <- rep(0, length(object$lambda)) res <- .vcovate(object) object$vcov_par <- res$vcov_par object$vcov_lambda <- res$vcov_lambda return(vcov.gel(object, lambda)) } } summary.ategel <- function(object, robToMiss=TRUE, ...) { if (robToMiss) { ans <- summary.gel(object) ans$typeDesc = paste(ans$typeDesc, "\n(S.E. are robust to misspecification)", sep="") } else { object$vcov_par <- vcov(object, robToMiss=FALSE) object$vcov_lambda <- vcov(object, TRUE, robToMiss=FALSE) ans <- summary.gel(object) ans$typeDesc = paste(ans$typeDesc, "\n(S.E. are not robust to misspecification", sep="") } ans } confint.ategel <- function (object, parm, level = 0.95, lambda = FALSE, type = c("Wald", "invLR", "invLM", "invJ"), fact = 3, corr = NULL, robToMiss=TRUE, ...) { type <- match.arg(type) if (type=="Wald") { if (!robToMiss) { object$vcov_par <- vcov(object, robToMiss=FALSE) object$vcov_lambda <- vcov(object, TRUE, robToMiss=FALSE) } return(confint.gel(object, parm, level, lambda, type, fact, corr, ...)) } object$allArg$g <- .momentFctATE object$allArg$y <- NULL object$allArg$w <- NULL object$allArg$treat <- NULL object$allArg$popMom <- NULL object$allArg$momType <- NULL object$allArg$family <- NULL object$allArg$x <- object$dat return(confint.gel(object, parm, level, lambda, type, fact, corr, ...)) } marginal <- function(object, ...) UseMethod("marginal") marginal.ategel <- function(object, ...) { family <- attr(object$dat, "family") if (is.null(family)) return(summary(object)$coef) k <- attr(object$dat, "k") p0 <- family$linkinv(object$coef[1]) p1 <- family$linkinv(object$coef[1]+object$coef[2:k]) p01 <- family$mu.eta(object$coef[1]) p11 <- family$mu.eta(object$coef[1]+object$coef[2:k]) A <- cbind(p11-p01, p11) V <- vcov(object, ...)[1:k,1:k] sd0 <- p01*sqrt(V[1,1]) sdd <- sapply(1:(k-1), function(i) sqrt(c(t(A[i,])%*%V[c(1,i+1), c(1, i+1)]%*%A[i,]))) coef <- cbind(c(p0,p1-p0), c(sd0, sdd)) coef <- cbind(coef, coef[,1]/coef[,2]) coef <- cbind(coef, 2*pnorm(-abs(coef[,3]))) colnames(coef) <- c("Estimate", "Std. Error", "t value", "Pr(>|t|)") rownames(coef) <- c("Control", paste("Treat", 1:(k-1) , " versus Control", sep="")) coef } checkConv <- function(obj, tolConv=1e-4,verbose=TRUE, ...) { if (!any(class(obj)=="ategel")) stop("The function is for ategel objects produced by ATEgel()") momType <- obj$allArg$momType popMom <- obj$allArg$popMom conv <- c(Lambda=obj$conv_lambda$convergence==0, Coef= obj$conv_par == 0) dat <- obj$dat$x nZ <- attr(obj$dat, "k")-1 z <- dat[,3:(2+nZ),drop=FALSE] x <- dat[,-(1:(2+nZ)),drop=FALSE] pt <- getImpProb(obj, ...) pt1 <- lapply(1:nZ, function(i) pt[z[,i]==1]/sum(pt[z[,i]==1])) pt0 <- pt[rowSums(z)==0]/sum(pt[rowSums(z)==0]) m0 <- colSums(x[rowSums(z)==0,,drop=FALSE]*pt0) m1 <- sapply(1:nZ, function(i) colSums(x[z[,i]==1,,drop=FALSE]*pt1[[i]])) mAll <- cbind(m0, m1) n0 <- paste(paste(colnames(z),collapse="=", sep=""),"=0",sep="") colnames(mAll) <- c(n0, paste(colnames(z),"=1",sep="")) if (!is.null(popMom)) { m <- popMom } else { m <- switch(momType, bal=m0, balSample=colMeans(x), ATT=c(m1)) } chk <- all(abs(mAll-m) 0) { var.fit <- ar(umat, order.max = prewhite, demean = FALSE, aic = FALSE, method = ar.method) if(inherits(var.fit, "try-error")) stop(sprintf("VAR(%i) prewhitening of estimating functions failed", prewhite)) umat <- as.matrix(na.omit(var.fit$resid)) n <- n - prewhite } # define kernel constants kernConst <- switch(kernel, "Bartlett" = list(q = 1, g_q = 1, mu1 = 1, mu2 = 2/3), "Parzen" = list(q = 2, g_q = 6, mu1 = 0.75, mu2 = 0.539286), "Tukey-Hanning" = list(q = 2, g_q = pi^2/4, mu1 = 1, mu2 = 0.75), "Quadratic Spectral" = list(q = 2, g_q = 1.421223, mu1 = 1.25003, mu2 = 0.999985)) # fit approximating model to moments if(approx == "AR(1)") { fitAR1 <- function(x) { rval <- ar(x, order.max = 1, aic = FALSE, method = "ols") rval <- c(rval$ar, sqrt(rval$var.pred)) names(rval) <- c("rho", "sigma") return(rval) } ar.coef <- apply(umat, 2, fitAR1) Omega0 <- ar.coef["sigma", ]^2 / (1-ar.coef["rho", ])^2 Omega_q <- if(kernConst$q == 1) { diag(2 * ar.coef["sigma", ]^2 * ar.coef["rho", ] / ((1 - ar.coef["rho", ])^3 * (1 + ar.coef["rho", ]))) } else { diag(2 * ar.coef["sigma", ]^2 * ar.coef["rho", ] / (1 - ar.coef["rho", ])^4) } } else { fitARMA11 <- function(x) { rval <- arima(x, order = c(1, 0, 1), include.mean = FALSE) rval <- c(rval$coef, sqrt(rval$sigma2)) names(rval) <- c("rho", "psi", "sigma") return(rval) } arma.coef <- apply(umat, 2, fitARMA11) Omega0 <- (1 + ar.coef["psi", ])^2 * ar.coef["sigma", ]^2 / (1 - ar.coef["rho", ])^2 Omega_q <- if(kernConst$q == 1) { diag(2 * (1 + ar.coef["psi", ] * ar.coef["rho", ]) * (ar.coef["psi", ] + ar.coef["rho", ]) * ar.coef["sigma", ]^2 / ((1 - ar.coef["rho", ])^3 * (1 + ar.coef["rho", ]))) } else { diag(2 * (1 + ar.coef["psi", ] * ar.coef["rho", ]) * (ar.coef["psi", ] + ar.coef["rho", ]) * ar.coef["sigma", ]^2 / (1 - ar.coef["rho", ])^4) } } # compute remaining bandwidth components W <- diag(weights) Omega0inv <- diag(1/Omega0) Sigma0 <- solve( crossprod(Omega0inv%*%G, G) ) H0 <- Sigma0 %*% t(G) %*% Omega0inv P0 <- Omega0inv - Omega0inv %*% G %*% H0 # compute optimal bandwidth nu2 <- (2 * kernConst$mu1 + kernConst$mu2) * (k - p) * sum(diag(Sigma0 %*% W)) nu3 <- kernConst$g_q^2 * sum(diag(t(Omega_q) %*% t(H0) %*% W %*% H0 %*% Omega_q %*% P0)) c0 <- if(nu2 * nu3 > 0) 2 * kernConst$q else -1 rval <- (c0 * nu3/nu2 * n)^(1/(1 + 2 * kernConst$q)) return(rval) } gmm/R/Methods.gmm.R0000644000176200001440000002171614247643115013556 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ summary.gmm <- function(object, ...) { z <- object se <- sqrt(diag(z$vcov)) par <- z$coefficients tval <- par/se ans <- list(met=z$met,kernel=z$kernel,algo=z$algo,call=z$call) names(ans$met) <- "GMM method" names(ans$kernel) <- "kernel for cov matrix" ans$coefficients <- cbind(par,se, tval, 2 * pnorm(abs(tval), lower.tail = FALSE)) dimnames(ans$coefficients) <- list(names(z$coefficients), c("Estimate", "Std. Error", "t value", "Pr(>|t|)")) ans$stest <- specTest(z) ans$algoInfo <- z$algoInfo if(z$met=="cue") ans$cue <- object$cue if (!is.null(object$initTheta)) { ans$initTheta <- object$initTheta names(ans$initTheta) <- names(z$coefficients) } ans$specMod <- object$specMod ans$bw <- attr(object$w0,"Spec")$bw ans$weights <- attr(object$w0,"Spec")$weights if(object$infVcov == "iid") ans$kernel <- NULL class(ans) <- "summary.gmm" ans } summary.tsls <- function(object, vcov = NULL, ...) { if (!is.null(vcov)) object$vcov=vcov ans <- summary.gmm(object) ans$met <- paste(ans$met, "(Meat type = ", attr(object$vcov, "vcovType"), ")",sep="") k <- object$dat$k if (!is.null(object$fsRes)) { fstat <- vector() fsRes <- object$fsRes if (class(fsRes)[1] == "listof") { nendo <- length(fsRes) } else { nendo <- 1 } if (nendo == 1) { fstat[1] <- fsRes$fstatistic[1] df1 <- fsRes$fstatistic[2] df2 <- fsRes$fstatistic[3] } else { fstat[1] <- fsRes[[1]]$fstatistic[1] df1 <- fsRes[[1]]$fstatistic[2] df2 <- fsRes[[1]]$fstatistic[3] } if (nendo > 1){ for (i in 2:nendo) fstat[i] <- fsRes[[i]]$fstatistic[1] } pvfstat <- 1-pf(fstat,df1, df2) names(fstat) <- attr(fsRes,"Endo") ans$fstatistic <- list(fstat = fstat, pvfstat = pvfstat, df1 = df1, df2 = df2) } ans$specMod <- object$specMod class(ans) <- "summary.tsls" return(ans) } print.summary.tsls <- function(x, digits = 5, ...) { print.summary.gmm(x,digits) if (!is.null(x$fstatistic)) { cat("\n First stage F-statistics: \n") if(names(x$fstatistic$fstat)[1]=="(Intercept)") start=2 else start=1 for (i in start:length(x$fstatistic$fstat)) cat(names(x$fstatistic$fstat)[i], ": F(",x$fstatistic$df1,", ",x$fstatistic$df2,") = ",x$fstatistic$fstat[i], " (P-Vavue = ",x$fstatistic$pvfstat[i],")\n") } else { cat("\n No first stage F-statistics (just identified model)\n") } } print.summary.gmm <- function(x, digits = 5, ...) { cat("\nCall:\n") cat(paste(deparse(x$call), sep="\n", collapse = "\n"), "\n\n", sep="") cat("\nMethod: ", x$met,"\n") if (x$met=="cue") { if (!is.null(x$cue$message)) { cat(" (",x$cue$message,")\n\n") } } else { cat("\n") } if( !is.null(x$kernel)) { cat("Kernel: ", x$kernel) if (!is.null(x$bw)) cat("(with bw = ", round(x$bw,5),")\n\n") else cat("\n\n") } cat("Coefficients:\n") print.default(format(x$coefficients, digits=digits), print.gap = 2, quote = FALSE) cat("\n") cat(x$stest$ntest,"\n") print.default(format(x$stest$test, digits=digits), print.gap = 2, quote = FALSE) cat("\n") if(!is.null(x$initTheta)) { cat("Initial values of the coefficients\n") print(x$initTheta) cat("\n") } cat(x$specMod) if(!is.null(x$algoInfo)) { cat("#############\n") cat("Information related to the numerical optimization\n") } if(!is.null(x$algoInfo$convergence)) cat("Convergence code = ", x$algoInfo$convergence,"\n") if(!is.null(x$algoInfo$counts)) { cat("Function eval. = ",x$algoInfo$counts[1],"\n") cat("Gradian eval. = ",x$algoInfo$counts[2],"\n") } if(!is.null(x$algoInfo$message)) cat("Message: ",x$algoInfo$message,"\n") invisible(x) } formula.gmm <- function(x, ...) { if(is.null(x$terms)) stop("The gmm object was not created by a formula") else formula(x$terms) } confint.gmm <- function(object, parm, level=0.95, ...) { ntest <- "Wald type confidence interval" z <- object se <- sqrt(diag(z$vcov)) par <- z$coefficients zs <- qnorm((1-level)/2,lower.tail=FALSE) ch <- zs*se ans <- cbind(par-ch,par+ch) dimnames(ans) <- list(names(par),c((1-level)/2,0.5+level/2)) if(!missing(parm)) ans <- ans[parm,] ans <- list(test=ans, ntest=ntest) class(ans) <- "confint" ans } residuals.gmm <- function(object,...) { if(is.null(object$model)) stop("The residuals method is valid only for g=formula") object$residuals } fitted.gmm <- function(object,...) { if(is.null(object$model)) stop("The residuals method is valid only for g=formula") object$fitted.value } print.gmm <- function(x, digits=5, ...) { cat("Method\n", x$met,"\n\n") cat("Objective function value: ",x$objective,"\n\n") print.default(format(coef(x), digits=digits), print.gap = 2, quote = FALSE) cat("\n") if(!is.null(x$algoInfo$convergence)) cat("Convergence code = ", x$algoInfo$convergence,"\n") cat(x$specMod) invisible(x) } coef.gmm <- function(object,...) object$coefficients vcov.gmm <- function(object,...) object$vcov estfun.gmmFct <- function(x, y = NULL, theta = NULL, ...) { if (is(x, "function")) { gmat <- x(theta, y) return(gmat) } else return(x) } estfun.tsls <- function(x, ...) { model.matrix(x)*c(residuals(x)) } model.matrix.tsls <- function(object, ...) { dat <- object$dat ny <- dat$ny nh <- dat$nh k <- dat$k x <- dat$x n <- nrow(x) hm <- as.matrix(x[,(ny+k+1):(ny+k+nh)]) xm <- as.matrix(x[,(ny+1):(ny+k)]) xhat <- lm(xm~hm-1)$fitted assign <- 1:ncol(xhat) if (attr(object$terms,"intercept")==1) assign <- assign-1 attr(xhat,"assign") <- assign xhat } vcov.tsls <- function(object, type=c("Classical","HC0","HC1","HAC"), hacProp = list(), ...) { type <- match.arg(type) if (type == "Classical") { sig <- sum(c(residuals(object))^2)/(nrow(object$dat$x)-object$dat$k) ny <- object$dat$ny nh <- object$dat$nh k <- object$dat$k n <- nrow(object$dat$x) hm <- as.matrix(object$dat$x[,(ny+k+1):(ny+k+nh)]) Omega <- crossprod(hm)*sig/nrow(object$dat$x) vcovType <- "Classical" V <- solve(crossprod(object$G,solve(Omega,object$G)))/nrow(object$dat$x) } else if (strtrim(type,2) == "HC") { meat <- meatHC(object, type) bread <- bread(object) vcovType <- paste("HCCM: ", type, sep="") V <- crossprod(bread, meat%*%bread)/nrow(object$dat$x) } else { object$centeredVcov <- TRUE gt <- model.matrix(object)*c(residuals(object)) gt <- lm(gt~1) arg <- c(list(x=gt,sandwich=FALSE),hacProp) meat <- do.call(kernHAC, arg) KType <- ifelse(is.null(hacProp$kernel), formals(kernHAC)$kernel[[2]], hacProp$kernel) vcovType <- paste("HAC: ", KType, sep="") bread <- bread(object) V <- crossprod(bread, meat%*%bread)/nrow(object$dat$x) } attr(V, "vcovType") <- vcovType return(V) } estfun.gmm <- function(x, ...) { foc <- x$gt %*% x$w %*% x$G return(foc) } bread.gmm <- function(x, ...) { GWG <- crossprod(x$G, x$w %*% x$G) b <- try(solve(GWG), silent = TRUE) if (any(class(b) == "try-error")) stop("The bread matrix is singular") return(b) } bread.tsls <- function(x, ...) { dat <- x$dat ny <- dat$ny nh <- dat$nh k <- dat$k x <- dat$x n <- nrow(x) hm <- as.matrix(x[,(ny+k+1):(ny+k+nh)]) xm <- as.matrix(x[,(ny+1):(ny+k)]) xhat <- lm(xm~hm-1)$fitted solve(crossprod(xhat)/n) } gmm/R/gmm.R0000644000176200001440000005763114612776367012175 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ gmm <- function(g,x,t0=NULL,gradv=NULL, type=c("twoStep","cue","iterative"), wmatrix = c("optimal","ident"), vcov=c("HAC","MDS","iid","TrueFixed"), kernel=c("Quadratic Spectral","Truncated", "Bartlett", "Parzen", "Tukey-Hanning"),crit=10e-7,bw = bwAndrews, prewhite = 1, ar.method = "ols", approx="AR(1)",tol = 1e-7, itermax=100,optfct=c("optim","optimize","nlminb", "constrOptim"), model=TRUE, X=FALSE, Y=FALSE, TypeGmm = "baseGmm", centeredVcov = TRUE, weightsMatrix = NULL, traceIter = FALSE, data, eqConst = NULL, eqConstFullVcov = FALSE, mustar = NULL, onlyCoefficients=FALSE, ...) { type <- match.arg(type) kernel <- match.arg(kernel) vcov <- match.arg(vcov) wmatrix <- match.arg(wmatrix) optfct <- match.arg(optfct) if (!is.null(eqConst)) TypeGmm <- "constGmm" if(vcov=="TrueFixed" & is.null(weightsMatrix)) stop("TrueFixed vcov only for fixed weighting matrix") if(!is.null(weightsMatrix)) wmatrix <- "optimal" if(missing(data)) data<-NULL all_args<-list(data = data, g = g, x = x, t0 = t0, gradv = gradv, type = type, wmatrix = wmatrix, vcov = vcov, kernel = kernel, crit = crit, bw = bw, prewhite = prewhite, ar.method = ar.method, approx = approx, weightsMatrix = weightsMatrix, centeredVcov = centeredVcov, tol = tol, itermax = itermax, optfct = optfct, model = model, X = X, Y = Y, call = match.call(), traceIter = traceIter, eqConst = eqConst, eqConstFullVcov = eqConstFullVcov, mustar = mustar) class(all_args)<-TypeGmm Model_info<-getModel(all_args, ...) z <- momentEstim(Model_info, ...) if (onlyCoefficients) return(z[c("coefficients","objective")]) z <- FinRes(z, Model_info) z } evalGmm <- function(g, x, t0, tetw=NULL, gradv=NULL, wmatrix = c("optimal","ident"), vcov=c("HAC","iid","TrueFixed"), kernel=c("Quadratic Spectral","Truncated", "Bartlett", "Parzen", "Tukey-Hanning"),crit=10e-7,bw = bwAndrews, prewhite = FALSE, ar.method = "ols", approx="AR(1)",tol = 1e-7, model=TRUE, X=FALSE, Y=FALSE, centeredVcov = TRUE, weightsMatrix = NULL, data, mustar = NULL) { TypeGmm = "baseGmm" type <- "eval" kernel <- match.arg(kernel) vcov <- match.arg(vcov) wmatrix <- match.arg(wmatrix) if (is.null(tetw) & is.null(weightsMatrix)) stop("If the weighting matrix is not provided, you need to provide the vector of parameters tetw") if(vcov=="TrueFixed" & is.null(weightsMatrix)) stop("TrueFixed vcov only for fixed weighting matrix") if(!is.null(weightsMatrix)) wmatrix <- "optimal" if(missing(data)) data<-NULL all_args<-list(data = data, g = g, x = x, t0 = t0, tetw = tetw, gradv = gradv, type = type, wmatrix = wmatrix, vcov = vcov, kernel = kernel, crit = crit, bw = bw, prewhite = prewhite, ar.method = ar.method, approx = approx, weightsMatrix = weightsMatrix, centeredVcov = centeredVcov, tol = tol, itermax = 100, model = model, X = X, Y = Y, call = match.call(), traceIter = NULL, optfct="optim", eqConst = NULL, eqConstFullVcov = FALSE, mustar = mustar) class(all_args)<-TypeGmm Model_info<-getModel(all_args) class(Model_info) <- "baseGmm.eval" z <- momentEstim(Model_info) z <- FinRes(z, Model_info) z } tsls <- function(g,x,data) { if(class(g)[1] != "formula") stop("2SLS is for linear models expressed as formula only") ans <- gmm(g,x,data=data,vcov="iid", TypeGmm="tsls") ans$met <- "Two Stage Least Squares" ans$call <- match.call() class(ans) <- c("tsls","gmm") ans$vcov <- vcov(ans, type="Classical") return(ans) } .myKernHAC <- function(gmat, obj) { gmat <- as.matrix(gmat) if(obj$centeredVcov) gmat <- scale(gmat, scale=FALSE) class(gmat) <- "gmmFct" AllArg <- obj$WSpec$sandwich AllArg$x <- gmat if (is.function(AllArg$bw)) { if (identical(AllArg$bw, bwWilhelm)) { G <- .DmomentFct(obj$tet, obj$dat) obj <- list(gt=gmat, G=G) class(obj) <- "gmm" } else { obj <- gmat } AllArg$bw <- AllArg$bw(obj, order.by = AllArg$order.by, kernel = AllArg$kernel, prewhite = AllArg$prewhite, ar.method = AllArg$ar.method, approx=AllArg$approx) } weights <- weightsAndrews(x=gmat, bw=AllArg$bw, kernel=AllArg$kernel, prewhite=AllArg$prewhite, tol=AllArg$tol) w <- vcovHAC(x=gmat, order.by=AllArg$order.by, weights=weights, prewhite=AllArg$prewhite, sandwich=FALSE, ar.method=AllArg$ar.method, adjust=FALSE) attr(w,"Spec") <- list(weights = weights, bw = AllArg$bw, kernel = AllArg$kernel) w } getDat <- function (formula, h, data, error=TRUE) { cl <- match.call() mf <- match.call(expand.dots = FALSE) m <- match(c("formula", "data"), names(mf), 0L) mf <- mf[c(1L, m)] mf$drop.unused.levels <- TRUE mf$na.action <- "na.pass" mf[[1L]] <- as.name("model.frame") mf <- eval(mf, parent.frame()) mt <- attr(mf, "terms") y <- as.matrix(model.response(mf, "numeric")) namey <- as.character(formula)[2] if (ncol(y)>1) namey <- paste(namey, ".", 1:ncol(y), sep="") xt <- as.matrix(model.matrix(mt, mf, NULL)) n <- NROW(y) if (inherits(h,'formula')) { tmp <- as.character(formula) termsh <- terms(h) h <- paste(tmp[2], "~", as.character(h)[2], sep="") h <- as.formula(h) mfh <- match.call(expand.dots = FALSE) mh <- match(c("h", "data"), names(mfh), 0L) mfh <- mfh[c(1L, mh)] mfh$formula <- h mfh$h <- NULL mfh$drop.unused.levels <- TRUE mfh$na.action <- "na.pass" mfh[[1L]] <- as.name("model.frame") mfh <- eval(mfh, parent.frame()) mth <- attr(mfh, "terms") h <- as.matrix(model.matrix(mth, mfh, NULL)) } else { h <- as.matrix(h) chkInt <- sapply(1:ncol(h), function(i) all(h[,i]/mean(h[,i]) == 1)) if (sum(chkInt) > 1) stop("Too many intercept in the matrix h") if (any(chkInt)) { h <- h[,!chkInt, drop=FALSE] h <- cbind(1,h) intercept_h <- TRUE } else { if (attr(mt,"intercept")==1) { h <- cbind(1, h) intercept_h <- TRUE } else { intercept_h <- FALSE } } if(is.null(colnames(h))) { if (intercept_h) colnames(h) <- c("(Intercept)",paste("h",1:(ncol(h)-1),sep="")) else colnames(h) <- paste("h",1:ncol(h),sep="") } else { if (intercept_h) colnames(h)[1] <- "(Intercept)" coln_h <- colnames(h) coln_h <- gsub(" ", "", coln_h) chk <- which(coln_h == "") if (length(chk) >0) coln_h[chk] <- paste("h", 1:length(chk), sep="") if (any(duplicated(coln_h))) stop("colnames of the matrix h must be unique") colnames(h) <- coln_h } if (!intercept_h) { hFormula <- paste(colnames(h), collapse="+") hFormula <- as.formula(paste("~", hFormula, "-1", sep="")) } else { hFormula <- paste(colnames(h)[-1], collapse="+") hFormula <- as.formula(paste("~", hFormula, sep="")) } termsh <- terms(hFormula) } ny <- ncol(y) k <- ncol(xt) nh <- ncol(h) if (nh 0) { endo <- xm[, -includeExo, drop = FALSE] endoName <- colnames(endo) if (ncol(endo) != 0) { if (attr(dat$termsh, "intercept") == 1) restsls <- lm(endo~hm[,-1]) else restsls <- lm(endo~hm-1) fsls <- xm fsls[, -includeExo] <- restsls$fitted } else { fsls <- xm restsls <- NULL } } else { if (attr(dat$termsh, "intercept") == 1) restsls <- lm(xm~hm[,-1]) else restsls <- lm(xm~hm-1) endoName <- colnames(xm) } if (!is.null(restsls)) { chk <- .chkPerfectFit(restsls) fsRes <- suppressWarnings(summary(restsls))[!chk] attr(fsRes, "Endo") <- endoName[!chk] } return(fsRes) } .tetlin <- function(dat, w, type=NULL) { x <- dat$x g <- .momentFct gradv <- .DmomentFct ny <- dat$ny nh <- dat$nh k <- dat$k n <- nrow(x) ym <- as.matrix(x[,1:ny]) xm <- as.matrix(x[,(ny+1):(ny+k)]) hm <- as.matrix(x[,(ny+k+1):(ny+k+nh)]) if (!is.null(attr(dat, "eqConst"))) { resTet <- attr(dat,"eqConst")$eqConst y2 <- xm[, resTet[,1],drop=FALSE]%*%resTet[,2] ym <- ym-c(y2) xm <- xm[,-resTet[,1],drop=FALSE] k <- ncol(xm) } includeExo <- which(colnames(xm)%in%colnames(hm)) inv <- attr(w, "inv") mustar <- attr(dat, "mustar") if (!is.null(type)) { if(type=="2sls") { if (length(includeExo) > 0) { endo <- xm[, -includeExo, drop = FALSE] endoName <- colnames(endo) if (ncol(endo) != 0) { if (attr(dat$termsh, "intercept") == 1) restsls <- lm(endo~hm[,-1]) else restsls <- lm(endo~hm-1) fsls <- xm fsls[, -includeExo] <- restsls$fitted } else { fsls <- xm restsls <- NULL } } else { if (attr(dat$termsh, "intercept") == 1) restsls <- lm(xm~hm[,-1]) else restsls <- lm(xm~hm-1) fsls <- restsls$fitted endoName <- colnames(xm) } par <- lm.fit(as.matrix(fsls), ym)$coefficients if (ny == 1) { e2sls <- ym-xm%*%par v2sls <- lm.fit(as.matrix(hm), e2sls)$fitted value <- sum(v2sls^2)/sum(e2sls^2) } else { par <- c(t(par)) g2sls <- g(par, dat) w <- crossprod(g2sls)/n gb <- matrix(colMeans(g2sls), ncol = 1) value <- crossprod(gb, solve(w, gb)) } } } else { if (ny>1) { if (inv) { whx <- solve(w, (crossprod(hm, xm) %x% diag(ny))) wvecyh <- solve(w, matrix(crossprod(ym, hm), ncol = 1)) } else { whx <- w%*% (crossprod(hm, xm) %x% diag(ny)) wvecyh <- w%*%matrix(crossprod(ym, hm), ncol = 1) } dg <- gradv(NULL, dat) xx <- crossprod(dg, whx) par <- solve(xx, crossprod(dg, wvecyh)) } else { if (nh>k) { if(inv) xzwz <- crossprod(xm,hm)%*%solve(w,t(hm)) else xzwz <- crossprod(xm,hm)%*%w%*%t(hm) par <- solve(xzwz%*%xm,xzwz%*%ym) } else { par <- solve(crossprod(hm,xm),crossprod(hm,ym)) } } gb <- matrix(colSums(g(par, dat))/n, ncol = 1) if(inv) value <- crossprod(gb, solve(w, gb)) else value <- crossprod(gb, w%*%gb) } res <- list(par = par, value = value) if (!is.null(mustar)) { if (!is.null(type)) { w <- crossprod(hm)/NROW(hm) attr(w, "inv") <- TRUE } res <- .mustarLin(par, xm, hm, w, dat, mustar) } if (!is.null(type)) { if (type == "2sls") res$firstStageReg <- restsls if (!is.null(restsls)) { chk <- .chkPerfectFit(restsls) res$fsRes <- suppressWarnings(summary(restsls))[!chk] attr(res$fsRes, "Endo") <- endoName[!chk] } } return(res) } .mustarLin <- function(par, xm, hm, w, dat, mustar) { if (ncol(xm) == ncol(hm)) { par <- par-solve(crossprod(hm,xm),mustar) } else { hmxm <- crossprod(hm,xm) if (attr(w, "inv")) T1 <- solve(w, hmxm) else T1 <- w%*%hmxm par <- par-solve(crossprod(hmxm, T1), crossprod(T1, mustar)) } gb <- colSums(.momentFct(par, dat))/NROW(xm) if(attr(w, "inv")) value <- crossprod(gb, solve(w, gb)) else value <- crossprod(gb, w%*%gb) list(value=value, par=par) } .obj1 <- function(thet, x, w) { gt <- .momentFct(thet, x) gbar <- as.vector(colMeans(gt)) INV <- attr(w, "inv") if (INV) obj <- crossprod(gbar, solve(w, gbar)) else obj <- crossprod(gbar,w)%*%gbar return(obj) } .Gf <- function(thet, x, pt = NULL) { myenv <- new.env() assign('x', x, envir = myenv) assign('thet', thet, envir = myenv) barg <- function(x, thet) { gt <- .momentFct(thet, x) if (is.null(pt)) gbar <- as.vector(colMeans(gt)) else gbar <- as.vector(colSums(c(pt)*gt)) return(gbar) } G <- attr(numericDeriv(quote(barg(x, thet)), "thet", myenv), "gradient") return(G) } .objCue <- function(thet, x, type=c("HAC", "MDS", "iid", "ident", "fct", "fixed")) { type <- match.arg(type) gt <- .momentFct(thet, x) gbar <- as.vector(colMeans(gt)) w <- .weightFct(thet, x, type) if (attr(w, "inv")) obj <- crossprod(gbar,solve(w,gbar)) else obj <- crossprod(gbar,w%*%gbar) return(obj) } .momentFct <- function(tet, dat) { if (!is.null(attr(dat, "eqConst"))) { resTet <- attr(dat,"eqConst")$eqConst tet2 <- vector(length=length(tet)+nrow(resTet)) tet2[resTet[,1]] <- resTet[,2] tet2[-resTet[,1]] <- tet } else { tet2 <- tet } if (attr(dat, "ModelType") == "linear") { x <- dat$x ny <- dat$ny nh <- dat$nh k <- dat$k tet2 <- matrix(tet2, ncol = k) e <- x[,1:ny] - x[,(ny+1):(ny+k)] %*% t(tet2) gt <- e * x[, ny+k+1] if(nh > 1) for (i in 2:nh) gt <- cbind(gt, e*x[, (ny+k+i)]) } else { gt <- attr(dat,"momentfct")(tet2, dat) } if (!is.null(attr(dat, "smooth"))) { bw <- attr(dat, "smooth")$bw w <- attr(dat, "smooth")$w gt <- smoothG(gt, bw = bw, weights = w)$smoothx } gt <- as.matrix(gt) if (!is.null(attr(dat, "mustar"))) { if (length(attr(dat, "mustar")) != ncol(gt)) stop("dim of mustar must match the number of moment conditions") gt <- sweep(gt, 2, attr(dat, "mustar"), "-") } return(gt) } .DmomentFct <- function(tet, dat, pt=NULL) { if (!is.null(attr(dat, "eqConst"))) { resTet <- attr(dat,"eqConst")$eqConst tet2 <- vector(length=length(tet)+nrow(resTet)) tet2[resTet[,1]] <- resTet[,2] tet2[-resTet[,1]] <- tet } else { tet2 <- tet } if ((attr(dat,"ModelType") == "linear") & (is.null(attr(dat, "smooth")))) { x <- dat$x ny <- dat$ny nh <- dat$nh k <- dat$k dgb <- -(t(x[,(ny+k+1):(ny+k+nh)]) %*% x[,(ny+1):(ny+k)]) %x% diag(rep(1,ny))/nrow(x) if (!is.null(attr(dat, "eqConst"))) dgb <- dgb[,-attr(dat,"eqConst")$eqConst[,1], drop=FALSE] } else { if (is.null(attr(dat,"gradv"))) { dgb <- .Gf(tet, dat, pt) } else { dgb <- attr(dat,"gradv")(tet2, dat) if (!is.null(attr(dat, "eqConst"))) dgb <- dgb[,-attr(dat,"eqConst")$eqConst[,1], drop=FALSE] } } return(as.matrix(dgb)) } .weightFct <- function(tet, dat, type=c("HAC", "MDS", "iid", "ident", "fct", "fixed")) { type <- match.arg(type) if (type == "fixed") { w <- attr(dat, "weight")$w attr(w, "inv") <- FALSE } else if (type == "ident") { w <- diag(attr(dat, "q")) attr(w, "inv") <- FALSE } else { gt <- .momentFct(tet,dat) if (!is.null(attr(dat, "namesgt"))) colnames(gt) <- attr(dat, "namesgt") if(attr(dat, "weight")$centeredVcov) gt <- scale(gt, scale=FALSE) n <- NROW(gt) } if (type == "HAC") { obj <- attr(dat, "weight") obj$centeredVcov <- FALSE obj$tet <- tet obj$dat <- dat w <- .myKernHAC(gt, obj) attr(w, "inv") <- TRUE } if (type == "MDS") { w <- crossprod(gt)/n attr(w, "inv") <- TRUE } if (type == "iid") { if (attr(dat, "ModelType") == "linear") { if (dat$ny == 1) { e <- .residuals(tet, dat)$residuals sig <- mean(scale(e,scale=FALSE)^2) z <- dat$x[,(1+dat$ny+dat$k):ncol(dat$x)] w <- sig*crossprod(z)/length(e) } else { w <- crossprod(gt)/n } } else { w <- crossprod(gt)/n } attr(w, "inv") <- TRUE } if (type == "fct") { w <- attr(dat, "weight")$fct(gt, attr(dat, "weight")$fctArg) attr(w, "inv") <- TRUE } return(w) } .residuals <- function(tet, dat) { if (!is.null(attr(dat, "eqConst"))) { resTet <- attr(dat,"eqConst")$eqConst tet2 <- vector(length=length(tet)+nrow(resTet)) tet2[resTet[,1]] <- resTet[,2] tet2[-resTet[,1]] <- tet } else { tet2 <- tet } tet2 <- t(matrix(tet2, nrow = dat$ny)) y <- as.matrix(dat$x[,1:dat$ny]) x <- as.matrix(dat$x[,(dat$ny+1):(dat$ny+dat$k)]) yhat <- x%*%tet2 e <- y-yhat return(list(residuals=e, yhat=yhat)) } gmm/R/Methods.sysGmm.R0000644000176200001440000001102314612777306014251 0ustar liggesusers# This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # A copy of the GNU General Public License is available at # http://www.r-project.org/Licenses/ summary.sysGmm <- function(object, ...) { z <- object se <- sqrt(diag(z$vcov)) k <- attr(z$dat, "k") if (attr(z$dat, "sysInfo")$commonCoef) { seList <- rep(list(se), length(z$dat)) } else { seList <- list() for (i in 1:length(z$dat)) { seList[[i]] <- se[1:k[[i]]] se <- se[-(1:k[[i]])] } } par <- z$coefficients tval <- lapply(1:length(z$dat), function(i) par[[i]]/seList[[i]]) ans <- list(met=z$met,kernel=z$kernel,algo=z$algo,call=z$call) names(ans$met) <- "GMM method" names(ans$kernel) <- "kernel for cov matrix" ans$coefficients <- lapply(1:length(z$dat), function(i) cbind(par[[i]],seList[[i]], tval[[i]], 2 * pnorm(abs(tval[[i]]), lower.tail = FALSE))) ans$stest <- specTest(z) ans$algoInfo <- z$algoInfo ans$initTheta <- object$initTheta for (i in 1:length(z$dat)) { dimnames(ans$coefficients[[i]]) <- list(names(z$coefficients[[i]]), c("Estimate", "Std. Error", "t value", "Pr(>|t|)")) if (!is.null(ans$initTheta)) names(ans$initTheta[[i]]) <- names(z$coefficients[[i]]) } ans$specMod <- object$specMod ans$bw <- attr(object$w0,"Spec")$bw ans$weights <- attr(object$w0,"Spec")$weights ans$Sysnames <- names(z$dat) ans$met <- attr(object$dat, "sysInfo")$typeDesc if(object$infVcov != "HAC") ans$kernel <- NULL class(ans) <- "summary.sysGmm" ans } print.summary.sysGmm <- function(x, digits = 5, ...) { cat("\nCall:\n") cat(paste(deparse(x$call), sep="\n", collapse = "\n"), "\n\n", sep="") cat("Method\n", x$met,"\n\n") cat("\n") if( !is.null(x$kernel)) { cat("Kernel: ", x$kernel) if (!is.null(x$bw)) cat("(with bw = ", round(x$bw,5),")\n\n") else cat("\n\n") } cat("Coefficients:\n") m <- length(x$coefficients) for (i in 1:m) { cat(x$Sysnames[i], "\n") cat("#########\n") #print.default(format(x$coefficients[[i]], digits=digits), # print.gap = 2, quote = FALSE) printCoefmat(x$coefficients[[i]], digits=digits, ...) cat("\n") } cat(x$stest$ntest,"\n") print.default(format(x$stest$test, digits=digits), print.gap = 2, quote = FALSE) cat("\n") if(!is.null(x$initTheta)) { cat("Initial values of the coefficients\n") for (i in 1:m) { cat(x$Sysnames[i], "\n") print(x$initTheta[[i]]) } cat("\n") } cat(x$specMod) if(!is.null(x$algoInfo)) { cat("#############\n") cat("Information related to the numerical optimization\n") } if(!is.null(x$algoInfo$convergence)) cat("Convergence code = ", x$algoInfo$convergence,"\n") if(!is.null(x$algoInfo$counts)) { cat("Function eval. = ",x$algoInfo$counts[1],"\n") cat("Gradian eval. = ",x$algoInfo$counts[2],"\n") } if(!is.null(x$algoInfo$message)) cat("Message: ",x$algoInfo$message,"\n") invisible(x) } print.sysGmm <- function(x, digits=5, ...) { cat("Method\n", attr(x$dat, "sysInfo")$typeDesc,"\n\n") cat("Objective function value: ",x$objective,"\n\n") for (i in 1:length(x$coefficients)) { cat(names(x$dat)[[i]], ": \n") print.default(format(coef(x)[[i]], digits=digits), print.gap = 2, quote = FALSE) } cat("\n") if(!is.null(x$algoInfo$convergence)) cat("Convergence code = ", x$algoInfo$convergence,"\n") cat(x$specMod) invisible(x) } gmm/NEWS0000644000176200001440000004026214275335277011554 0ustar liggesusersChanges in version 1.6-2 o Fixed a bug with iterative GMM from the previous version o Added the option onlyCoefficients=FALSE to gmm() and gel(). If set to TRUE, no covariance matrix is computed. Only the coefficients objective and the lambdas (for gel only) are returned. o In previous versions, the HAC weighting matrix for linear models was computed without specifying that the first column of the instrument matrix was a column of ones (with its name being "(Intercept)"). It makes a difference when computing the bandwidth with bwNeweyWest. The new version, takes it into account. It is not possible to obtain the same sandwich covariance matrix with gmm and OLS. o In previous versions, the weighting matrix and, as a result, the standard errors were based on the small sample adjusted vcovHAC for weakly dependent processes. The adjustment is not justified for GMM. The option adjust=FALSE is therefore added to vcovHAC in .myKernHAC() o Added a method getImpProb, which extract the implied probabilities. It allows for negative probabilities as it often happens with CUE o lambda for CUE is solved analytically (it is minus the projection of gt on a vector of ones). The option RCue for optlam can be used to restrict the implied probabilities to be non-negative. Changes in version 1.6-1 o Added the function bwWilhelm() to compute the optimal bandwidth of Wilhelm(2015). o The argument $vcov of tsls object is the Classical covariance matrix. vcov() must be used to obtain a different covariance o Changed the default value of prewhite to 1 in gmm() because it is recommended for HAC estimation o gmm with optfct="optimize" was generating an error message. It has been fixed o A new set of functions to estimate the average treatment effect by GEL have been added. See ?ATEgel o Some other minor bugs that were reported by users were fixed. Changes in version 1.6-0 o tsls is now a real 2-stage least square. Before it was a 2-step optimal GMM with HCCM weighting matrix. o Fixed a typo in FinRes.R file. It was preventing to compute the proper vcov matrix for a very special case (fixed weights) o Added Helinger distance and Exponentially tilted Helinger estimator to gel() o Fixed the LR test of ETEL in gel() o Cleaned a lot of the codes. In particular, a single moment function and its gradiant function is used now. There is also a common weighting matrix generator function. The main goal is to make it more flexible and give the possibility of creating weighting matrix based on other assumption such as clustering. o For GEL with smooth=TRUE, the weights were not fixed which was creating problems of convergence. o Add the option of fixing the value of some coefficient in gel as it as already the case for gmmm. o It is no possible to provide gmm with the gradiant function when constrOptim is used. o There is now an evalGmm() function to create a gmm object with a given fixed vector. Not estimation is done, but the bandwidth and weights for the HAC estimator are generated. It is possible to give a different vector os parameters for the moment function and the weighting matrix. o There is also an evalGel(). You fix the parameters and only the lambda are computed. o Fixed NAMESPACE to avoid the notes given by CRAN. o Modified getDat to allow for ~1 as instruments. Useful for inference on the mean using empirical likelihood with gel(x~1,~1). o The Lambdas and specTest from gel are printed even if the model is just identified. It is consitent with gmm which prints the j-test for just identified models. The reason if to allow evalGel to generate the Lamdbas for testing purpose. o confint.gxx() has been modified. It now create an object and a print method has been created as well. o confint.gel includes the possibility of building a confidence interval using the inversion of one of the three specification tests. It is therefore possible de construct an Empirical Likelihood confidence interval as described in Owen 2001. See manual. o On the process to create a sysGmm for system og linear equations. The function sysGmm() will do the job. Not yet working. o getDat is modified so that the name of the dependent variable is the one included in the formula. o A set of functions have been added to estimate system of linear equations (this is a beta version not tested yet). o A set of labor data as been added to test the sysGmm function o A panel of Macro data has been added. Changes in version 1.5-2 o Added constrOptim to nonlinear gmm() with fixed W Changes in version 1.5-1 o Added importFrom element o Added element df.residual to gmm objects to allow F-type linearHypothesis o Removed the unnecessary ls estimations in the first stage 2SLS Changes in version 1.5-0 o Thanks to Eric Zivot. Many improvements are based on his testings of the package. o Fixed many bugs for the case in which vcov is set to iid o Added options to tsls(). There are options for different sandwich matrix when summary is called Changes in version 1.4-6 o Added the possibility of providing data.frame() with formula to gel() Changes in version 1.4-5 o Replaced the deprecated function real(), is.real() and as.real() by double() Changes in version 1.4-4 o Fixed an instability problem when the model is just identified and wmatrix is set to "ident" Changes in version 1.4-3 o Added the option "constrOptim" to optfct for nonlinear models o Added the option "eqConstFullVcov" to gmm() to allow the computation the covariance matrix of all coefficients including the constrained ones (see ?gmm for details and examples). o Removed the dependency on fBasics and replaced it with stabledist Changes in version 1.4-2 o Bugs fixed when running gel() with smooth=T and g() begina nonlinear function o Now, print.gel() and print.summary.gel() repports the bandwidth when smooth=T Changes in version 1.4-0 o The method for GMM-CUE has been modified. Before, the weights for the kernel estimation of the weighting matrix were flexible inside the optimizer which was making the algorithm long and unstable. It is now fixed using either the starting values provided by the user (for linear cases) or the first step GMM. o You can now trace the convergence of the iterative GMM with the option traceIter=T o The weights for the kernel are also fixed for iter-GMM. It is faster and I don't think it should change at each iteration. I am open to comments on that. o A function tsls() has been added to facilitate the estimation by 2SLS. The user does not have to worry about selecting the right gmm options o summary and print Methods have been added for tsls objects. The summary prints the f-statistics of the first stage LS o SmoothG for GEL uses now the kernapply() function which is more efficient. o Something the GMM converge to weird values which is sometimes caused by bad first step estimates used to compute the weighting matrix. Summary() prints the initial values to have more infortmation when convergence fails. o There was a bug in specTest() for GEL. The degrees of freedom for the J-test were wrong. It is fixed. o For GEL type CUE the implied probabilities are computed according to Antoine, Bonnal, and Renault (2007) which solves the problem of negative probabilities. o The data file Finance has been resaved which implies that the package depends now on R version ­2.10.0 and higher o The function rho() had been changed to .rho() because it is not useful outside the gel() function o The function getLamb() has been modified. It is now more efficient. The default for the gel() option optlam is now nlminb. The gradient and the hessian is provided which makes it much faster to solve for lambda. The argument of the function also changed. Instead of providing the function g() and the vector of coefficients, we provide the matrix gt. I may be useful to call the function sometimes to solve for lambda and it is easier that way. o The choice "iter" from the option optlam is kept for cases in which the optimizer fails to solve for lambda. It is often the case when type="ETEL" is chosen which tends to produce NA's. o For GEL of type "EL" and optfct="optim", the algorith constrOptim() is launched to make sure lambda'gt is always greater than 1 and avoid NA's when computing log(1-lambda'gt). The algorithm checks first with optim(), and swith to constrOptim() is optim() fails o Sometimes, problems happen in GMM estimation because of the bad first step estimates used to compute the weighting matrix. The first step estimates are usually computed using the identity matrix. The vector is now printed for better control. o Cleaned the codes. The data are in object$dat and we can get the moment matrix by calling gt <- object$g(object$coef,object$dat) for linear and non-linear models, where object is of class gmm. o We can now impose equality constraints of the type theta_i=c_i with the option eqConst=. See help(gmm) which includes examples. o The K statistics for weakly identified models of Kleiberben (2005) has been added. See ?KTest. o (gmm only) The bandwidth for the HAC estimation is set by the first step estimate. Its value, along with the weights, is saved in attr(obj$w0,"Spec"). If the user is not satisfied with this automatic bw selection, bw can be set to any fixed number when calling gmm(). See ?gmm for more details and examples. o The function gmmWithConst() reestimate an unrestricted model by imposing linear constraints, using the same specification. It also use the same bandwidth for the estimation of the HAC matrix o A regularized version of GEL based on Chausse(2011) is now implemented. The Lagrange multiplier can be regularized by setting the option "alpha=" to a number. The regularization can stabilize the GEL estimator. A preliminary paper will be available on my web site that analyze the properties of such modification. A method for selecting alpha will be implemented in future version. Changes in version 1.3-8 o A bug was found in the computation of linear GMM. The weighting matrix was not use properly. It is now fixed. Changes in version 1.3-7 o Until now there was no way to know whether optim or nlminb has converged in case of nonlinear GMM. The convergence code is now printed by print.gmm and the message and function evaluation by print.summary.gmm o For cases in which the weighting matrix is provided, you have the choice of assuming that it is just a fixed matrix not necessary equal to the inverse of the covariance matrix of the moment conditions (the default) or setting vcov="TrueFixed" if it is the inverse of the estimated covariance matrix. o If gradv is provided, it is not only used to compute the covariance matrix. It is also used to compute the gradiant of the objective function when the method is set to "CG" or "BFGS" in optim. For CUE, it is only used for the first step estimate because the gradiant can hardly be obtained analytically in that case. Changes in version 1.3-6 o Bug fixed when estimating model with timeSeries objects o When NA were present in the data and the instruments, gmm() was not working when omitting them created different dimension. If by chance it created the same dimensions, the data were not correctly aligned. Now it is aligned and the observations associated with the NA's are removed. When it happens, a warning message is displayed. Changes in version 1.3-5 o Bug fixed when estimating linear model with Z=X (OLS with robust matrix) Changes in version 1.3-4 o It is now possible to enter the instruments in x either as a matrix (like before) or as a formula. See details and examples in gmm help o A data argument is added to the gmm function. Therefore, it is no longer required that the variables in data.frames be attached before using gmm(). You just need to add the option data=your_data.frame. o 2SLS is now implemented in a more efficient way for linear models. Just add the option vcov="iid". An example as been included for that case Changes in version 1.3-3 o It is a very small modification to avoid errors in the installation. The function linear.hypothesis of the car package is now deprecated. They have be changed to linearHypothesis. Changes in version 1.3-2 o The functions HAC, weightsAndrews2, bwAndrews2 and all the others to compute the HAC matrix have been removed. The sandwich package is now used to compute these matrices. o The option centeredVcov has beed added to the gmm function. It allows to compute the covariance matrix using a centered moment function. o The option weightsMatrix has beed added to the gmm function. It allows to fixe the matrix of weights. o The methods bread() and estfun() are now available for gmm objects. It allows to compute a sandwich covariance matrix which is required if the weighting matrix is not the optimal one. o The results are obtained even if the covariance matrix cannot be computed o The vignette has been updated. Changes in version 1.3-0 o The method "getModel", "momentEstim" is now used also for the gel procedure. o The GEL procedure as been modified. For the case in which the option smooth=TRUE is used, the appropriate scaling parameters have been added following Smith(2004) in the estimation, the covariance matrix of the Lagrange multiplier and the tests of over-identifying restrictions. In this case, only the kernels Truncated and Bartlett are available. For the former, the optimal bandwidth of Andrews(91) for Bartlett is used and for the latter, it is the bandwidth proposed for the Parzen. See Smith(2004) for more details. Also, the option vcov as been removed since it is irrelevant for GEL. By setting smooth to FALSE, it is implied that the data are iid. Changes in version 1.2-0 o The new method specTest computes the specification tests from objects created by "gmm" or "gel" o The structure of the package has been modified in order to make it more flexible. The changes include: o The new method "getModel" prepares the estimation. It creates what is required by the method "momentEstim". o The method "momentEstim" estimates the model defined by "getModel" o The method "FinRes" finalizes the estimation when it is needed. For now, the method is used only by "gmm" to compute the final resutls. o Some bugs are fixed o The Vignette has been modified a little and the errors in the manual corrected. Changes in version 1.1-0 o The name of the vector of parameters is now "coefficients" instead of "par" o The following methods are now available (see lm): fitted, residuals, vcov, coef, confint o There are now print methods for gmm, gel and summary.gel and summary.gmm o These modifications allows to use linear.hypothesis from the car package so the function lintest is no longer needed and as been removed o The following are now available from gmm and gel objects when g is a formula: residuals, fitted.values, model.frame, terms, model, the response and model matrix o Because the presence of the confint method, the option "interval" as been removed from the summary methods o A new plot method is available for both gmm and gel objects. It is a beta version. Comments and suggestions are welcome. o If there is only one instrument it can be provided as a vector. It does not need to be a matrix anymore. o It is now possible to select nlminb as optimizer. It allows to put restriction on the parameter space. o The package no longer depends on mvtnorm. It is a suggested package as it is only required for examples. o The packages car and fBasics (and therefore MASS, timeDate and timeSeries) are now suggested for examples in the vignette. o The new function charStable has been added. It computes the characteristic function of a stable distribution. An example is shown in the vignette o There was a bug when trying to estimate a model by ETEL with gel and numerical computation of lambda. It is fixed. Thanks to Márcio Laurini. o The vignette as beed rewritten. Changes in Version 1.0-7 o Modified some functions to remove dependencies on tseries and sandwich packages o Convert de Finance Data to data frame format Changes in Version 1.0-6 o Some bugs fixed. Thanks to Justinas Brazys Changes in Version 1.0-4 o documentation enhancements o Added finance data for applied examples Changes in Version 1.0-3 o Some bugs fixed Changes in Version 1.0-2 o Added a new example for better understanding. gmm/vignettes/0000755000176200001440000000000015044423073013044 5ustar liggesusersgmm/vignettes/gmm_with_R.rnw0000644000176200001440000017532214271746755015722 0ustar liggesusers\documentclass[11pt,letterpaper]{article} \usepackage{amsthm} \usepackage[hmargin=2cm,vmargin=2.5cm]{geometry} \newtheorem{theorem}{Theorem} \newtheorem{col}{Corollary} \newtheorem{lem}{Lemma} \usepackage[utf8]{inputenc} \newtheorem{ass}{Assumption} \usepackage{amsmath} \usepackage{verbatim} \usepackage[round]{natbib} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{graphicx} \usepackage{hyperref} \hypersetup{ colorlinks, citecolor=black, filecolor=black, linkcolor=black, urlcolor=black } \bibliographystyle{plainnat} \newcommand{\E}{\mathrm{E}} \newcommand{\diag}{\mathrm{diag}} \newcommand{\Prob}{\mathrm{Pr}} \newcommand{\Var}{\mathrm{Var}} \let\proglang=\textsf \newcommand{\pkg}[1]{{\fontseries{m}\fontseries{b}\selectfont #1}} \newcommand{\Vect}{\mathrm{Vec}} \newcommand{\Cov}{\mathrm{Cov}} \newcommand{\conP}{\overset{p}{\to}} \newcommand{\conD}{\overset{d}{\to}} \newcommand\Real{ \mathbb{R} } \newcommand\Complex{ \mathbb{C} } \newcommand\Natural{ \mathbb{N} } \newcommand\rv{{\cal R}} \newcommand\Q{\mathbb{Q}} \newcommand\PR{{\cal R}} \newcommand\T{{\cal T}} \newcommand\Hi{{\cal H}} \newcommand\La{{\cal L}} \newcommand\plim{plim} \renewcommand{\epsilon}{\varepsilon} \begin{document} \author{Pierre Chauss\'e} \title{Computing Generalized Method of Moments and Generalized Empirical Likelihood with \proglang{R}} \maketitle \abstract{This paper shows how to estimate models by the generalized method of moments and the generalized empirical likelihood using the \proglang{R} package \textbf{gmm}. A brief discussion is offered on the theoretical aspects of both methods and the functionality of the package is presented through several examples in economics and finance. It is a modified version of \cite{chausse10} published in the Journal of Statistical Software. It has been adapted to the version 1.4-0. \textbf{Notice that the maintenance of the package is converging to zero. The new \pkg{momentfit} package, available on CRAN, will soon replace the \pkg{gmm} package.}} %\VignetteIndexEntry{Computing Generalized Empirical Likelihood and Generalized Method of Moments with R} %\VignetteDepends{gmm,mvtnorm,stabledist, car, MASS, timeDate, timeSeries} %\VignetteKeywords{generalized empirical likelihood, generalized method of moments, empirical likelihood, continuous updated estimator, exponential tilting, exponentially tilted empirical likelihood} %\VignettePackage{gmm} %\VignetteEngine{knitr::knitr} <>= library(knitr) opts_chunk$set(size='footnotesize', fig.height=5, out.width='70%') @ \section{Introduction} The generalized method of moments (GMM) has become an important estimation procedure in many areas of applied economics and finance since \cite{hansen82} introduced the two step GMM (2SGMM). It can be seen as a generalization of many other estimation methods like least squares (LS), instrumental variables (IV) or maximum likelihood (ML). As a result, it is less likely to be misspecified. The properties of the estimators of LS depend on the exogeneity of the regressors and the circularity of the residuals, while those of ML depend on the choice of the likelihood function. GMM is much more flexible since it only requires some assumptions about moment conditions. In macroeconomics, for example, it allows to estimate a structural model equation by equation. In finance, most data such as stock returns are characterized by heavy-tailed and skewed distributions. Because it does not impose any restriction on the distribution of the data, GMM represents a good alternative in this area as well. As a result of its popularity, most statistical packages like \proglang{Matlab}, \proglang{Gauss} or \proglang{Stata} offer tool boxes to use the GMM procedure. It is now possible to easily use this method in \proglang{R} with the new \pkg{gmm} package. Although GMM has good potential theoretically, several applied studies have shown that the properties of the 2SGMM may in some cases be poor in small samples. In particular, the estimators may be strongly biased for certain choices of moment conditions. In response to this result, \cite{hansen-heaton-yaron96} proposed two other ways to compute GMM: the iterative GMM (ITGMM) and the continuous updated GMM (CUE)\footnote{See also \cite{hall05} for a detailed presentation of most recent developments regarding GMM.}. Furthermore, another family of estimation procedures inspired by \cite{owen01}, which also depends only on moment conditions, was introduced by \cite{smith97}. It is the generalized empirical likelihood (GEL). So far, this method has not reached the popularity of GMM and it was not included in any statistical package until \pkg{gmm} was developed for \proglang{R} which also includes a GEL procedure. Asymptotic properties of GMM and generalized empirical likelihood (GEL) are now well established in the econometric literature. \cite{newey-smith04} and \cite{anatolyev05} have compared their second order asymptotic properties. In particular, they show that the second order bias of the empirical likelihood (EL) estimator, which is a special case of GEL, is smaller than the bias of the estimators from the three GMM methods. Furthermore, as opposed to GMM, the bias does not increase with the number of moment conditions. Since the efficiency improves when the number of conditions goes up, this is a valuable property. However, these are only asymptotic results which do not necessarily hold in small sample as shown by \cite{guggenberger08}. In order to analyze small sample properties, we have to rely on Monte Carlo simulations. However, Monte Carlo studies on methods such as GMM or GEL depend on complicated algorithms which are often home made. Because of that, results from such studies are not easy to reproduce. The solution should be to use a common tool which can be tested and improved upon by the users. Because it is open source, \proglang{R} offers a perfect platform for such tool. The \pkg{gmm} package allows to estimate models using the three GMM methods, the empirical likelihood and the exponential tilting, which belong to the family of GEL methods, and the exponentially tilted empirical likelihood which was proposed by \cite{schennach07}, Also it offers several options to estimate the covariance matrix of the moment conditions. Users can also choose between \textit{optim}, if no restrictions are required on the coefficients of the model to be estimated, and either \textit{nlminb} or \textit{constrOptim} for constrained optimizations. The results are presented in such a way that \proglang{R} users who are familiar with \textit{lm} objects, find it natural. In fact, the same methods are available for \textit{gmm} and \textit{gel} objects produced by the estimation procedures. The paper is organized as follows. Section 2 presents the theoretical aspects of the GMM method along with several examples in economics and finance. Through these examples, the functionality of the \pkg{gmm} packages is presented in details. Section 3 presents the GEL method with some of the examples used in section 2. Section 4 concludes and Section 5 gives the computational details of the package. \section{Generalized method of moments} This section presents an overview of the GMM method. It is intended to help the users understand the options that the \pkg{gmm} package offers. For those who are not familiar with the method and require more details, see \cite{hansen82} and \cite{hansen-heaton-yaron96} for the method itself, \cite{newey-west94} and \cite{andrews91} for the choice of the covariance matrix or \cite{hamilton94}. We want to estimate a vector of parameters $\theta_0 \in \Real^p$ from a model based on the following $q\times 1$ vector of unconditional moment conditions: \begin{equation}\label{mcond} E[g(\theta_0,x_i)] = 0 , \end{equation} where $x_i$ is a vector of cross-sectional data, time series or both. In order for GMM to produce consistent estimates from the above conditions, $\theta_0$ has to be the unique solution to $E[g(\theta,x_i)]=0$ and be an element of a compact space. Some boundary assumptions on higher moments of $g(\theta,x_i)$ are also required. However, it does not impose any condition on the distribution of $x_i$, except for the degree of dependence of the observations when it is a vector of time series. Several estimation methods such as least squares (LS), maximum likelihood (ML) or instrumental variables (IV) can also be seen as being based on such moment conditions, which make them special cases of GMM. For example, the following linear model: \[ Y = X\beta + u , \] where $Y$ and $X$ are respectively $n\times 1$ and $n\times k$ matrices, can be estimated by LS. The estimate $\hat{\beta}$ is obtained by solving $\min_\beta \|u\|^2$ and is therefore the solution to the following first order condition: \[ \frac{1}{n}X'u(\beta) = 0 , \] which is the estimate of the moment condition $E(X_iu_i(\beta))=0$. The same model can be estimated by ML in which case the moment condition becomes: \[ E\left[\frac{d l_i(\beta)}{d\beta}\right]=0 , \] where $l_i(\beta)$ is the density of $u_i$. In presence of endogeneity of the explanatory variable $X$, which implies that $E(X_iu_i)\neq 0$, the IV method is often used. It solves the endogeneity problem by substituting $X$ by a matrix of instruments $H$, which is required to be correlated with $X$ and uncorrelated with $u$. These properties allow the model to be estimated by the conditional moment condition $E(u_i|H_i)=0$ or its implied unconditional moment condition $E(u_iH_i)=0$. In general we say that $u_i$ is orthogonal to an information set $I_i$ or that $E(u_i|I_i)=0$ in which case $H_i$ is a vector containing functions of any element of $I_i$. The model can therefore be estimated by solving \[ \frac{1}{T}H'u(\beta)=0 . \] When there is no assumption on the covariance matrix of $u$, the IV corresponds to GMM. If $E(X_iu_i)=0$ holds, generalized LS with no assumption on the covariance matrix of $u$ other than boundary ones is also a GMM method. For the ML procedure to be viewed as GMM, the assumption on the distribution of $u$ must be satisfied. If it is not, but $E(dl_i(\theta_0)/d\theta)=0$ holds, as it is the case for linear models with non normal error terms, the pseudo-ML which uses a robust covariance matrix can be seen as being a GMM method. Because GMM depends only on moment conditions, it is a reliable estimation procedure for many models in economics and finance. For example, general equilibrium models suffer from endogeneity problems because these are misspecified and they represent only a fragment of the economy. GMM with the right moment conditions is therefore more appropriate than ML. In finance, there is no satisfying parametric distribution which reproduces the properties of stock returns. The family of stable distributions is a good candidate but only the densities of the normal, Cauchy and Lévy distributions, which belong to this family, have a closed form expression. The distribution-free feature of GMM is therefore appealing in that case. Although GMM estimators are easily consistent, efficiency and bias depend on the choice of moment conditions. Bad instruments implies bad information and therefore low efficiency. The effects on finite sample properties are even more severe and are well documented in the literature on weak instruments. \cite{newey-smith04} show that the bias increases with the number of instruments but efficiency decreases. Therefore, users need to be careful when selecting the instruments. \cite{carrasco07} gives a good review of recent developments on how to choose instruments in her introduction. In general, the moment conditions $E(g(\theta_0,x_i))=0$ is a vector of nonlinear functions of $\theta_0$ and the number of conditions is not limited by the dimension of $\theta_0$. Since efficiency increases with the number of instruments $q$ is often greater than $p$, which implies that there is no solution to \[ \bar{g}(\theta) \equiv \frac{1}{n}\sum_{i=1}^n g(\theta,x_i)=0. \] The best we can do is to make it as close as possible to zero by minimizing the quadratic function $\bar{g}(\theta)'W\bar{g}(\theta)$, where $W$ is a positive definite and symmetric $q\times q$ matrix of weights. The optimal matrix $W$ which produces efficient estimators is defined as: \begin{equation}\label{optw} W^* = \left\{\lim_{n\rightarrow \infty} Var(\sqrt{n} \bar{g}(\theta_0)) \equiv \Omega(\theta_0) \right\}^{-1}. \end{equation} This optimal matrix can be estimated by an heteroskedasticity and auto-correlation consistent (HAC) matrix like the one proposed by \cite{newey-west87a}. The general form is: \begin{equation}\label{optw_hat} \hat{\Omega} = \sum_{s=-(n-1)}^{n-1} k_h(s) \hat{\Gamma}_s(\theta^*), \end{equation} where $k_h(s)$ is a kernel, $h$ is the bandwidth which can be chosen using the procedures proposed by \cite{newey-west87a} and \cite{andrews91}, \[ \hat{\Gamma}_s(\theta^*) = \frac{1}{n}\sum_i g(\theta^*,x_i)g(\theta^*,x_{i+s})' \] and $\theta^*$ is a convergent estimate of $\theta_0$. There are many choices for the HAC matrix. They depend on the kernel and bandwidth selection. Although the choice does not affect the asymptotic properties of GMM, very little is known about the impacts in finite samples. The GMM estimator $\hat{\theta}$ is therefore defined as: \begin{equation}\label{gmm} \hat{\theta} = \arg\min_{\theta} \bar{g}(\theta)'\hat{\Omega}(\theta^*)^{-1}\bar{g}(\theta) \end{equation} The original version of GMM proposed by \cite{hansen82} is called two-step GMM (2SGMM). It computes $\theta^*$ by minimizing $\bar{g}(\theta)'\bar{g}(\theta)$. The algorithm is therefore: \begin{itemize} \item[1-] Compute $\theta^* = \arg\min_\theta \bar{g}(\theta)'\bar{g}(\theta)$ \item[2-] Compute the HAC matrix $\hat{\Omega}(\theta^*)$ \item[3-] Compute the 2SGMM $\hat{\theta} = \arg\min_\theta \bar{g}(\theta)'\big[\hat{\Omega}(\theta^*)\big]^{-1}\bar{g}(\theta)$ \end{itemize} In order to improve the properties of 2SGMM, \cite{hansen-heaton-yaron96} suggest two other methods. The first one is the iterative version of 2SGMM (ITGMM) and can be computed as follows: \begin{itemize} \item[1-] Compute $\theta^{(0)} = \arg\min_\theta \bar{g}(\theta)'\bar{g}(\theta)$ \item[2-] Compute the HAC matrix $\hat{\Omega}(\theta^{(0)})$ \item[3-] Compute the $\theta^{(1)} = \arg\min_\theta \bar{g}(\theta)'\big[\hat{\Omega}(\theta^{(0)})\big]^{-1}\bar{g}(\theta)$ \item[4-] If $\| \theta^{(0)}-\theta^{(1)}\|< tol$ stops, else $\theta^{(0)}=\theta^{(1)}$ and go to 2- \item[5-] Define the ITGMM estimator $\hat{\theta}$ as $\theta^{(1)}$ \end{itemize} where $tol$ can be set as small as we want to increase the precision. In the other method, no preliminary estimate is used to obtain the HAC matrix. The latter is treated as a function of $\theta$ and is allowed to change when the optimization algorithm computes the numerical derivatives. It is therefore continuously updated as we move toward the minimum. For that, it is called the continuous updated estimator (CUE). This method is highly nonlinear. It is therefore crucial to choose a starting value that is not too far from the minimum. A good choice is the estimate from 2SGMM which is known to be root-n convergent. The algorithm is: \begin{itemize} \item[1-] Compute $\theta^*$ using 2SGMM \item[2-] Compute the CUE estimator defined as \[ \hat{\theta} = \arg\min_\theta \bar{g}(\theta)'\big[\hat{\Omega}(\theta)\big]^{-1}\bar{g}(\theta) \] using $\theta^*$ as starting value. \end{itemize} According to \cite{newey-smith04} and \cite{anatolyev05}, 2SGMM and ITGMM are second order asymptotically equivalent. On the other hand, they show that the second order asymptotic bias of CUE is smaller. The difference in the bias comes from the randomness of $\theta^*$ in $\Omega(\theta^*)$. Iterating only makes $\theta^*$ more efficient. These are second order asymptotic properties. They are informative but may not apply in finite samples. In most cases, we have to rely on numerical simulations to analyze the properties in small samples. Given some regularity conditions, the GMM estimator converges as $n$ goes to infinity to the following distribution: \[ \sqrt{n}(\hat{\theta}-\theta_0) \stackrel{L}{\rightarrow} N(0,V), \] where \[ V = E\left(\frac{\partial g(\theta_0,x_i)}{\partial\theta}\right)'\Omega(\theta_0)^{-1}E\left(\frac{\partial g(\theta_0,x_i)}{\partial\theta}\right) \] Inference can therefore be performed on $\hat{\theta}$ using the assumption that it is approximately distributed as $N(\theta_0,\hat{V}/n)$. If $q>p$, we can perform a J-test to verify if the moment conditions hold. The null hypothesis and the statistics are respectively $H0:E[g(\theta,x_i)]=0$ and: \[ n\bar{g}(\hat{\theta})'[\hat{\Omega}(\theta^*)]^{-1}\bar{g}(\hat{\theta}) \stackrel{L}{\rightarrow} \chi^2_{q-p}. \] \section{GMM with R} The \pkg{gmm} package can be loaded the usual way. <<>>= library(gmm) @ The main function is \textit{gmm()} which creates an object of class \textit{gmm}. Many options are available but in many cases they can be set to their default values. They are explained in details below through examples. The main arguments are \textit{g} and \textit{x}. For a linear model, \textit{g} is a formula like \textit{y~z1+z2} and \textit{x} the matrix of instruments. In the nonlinear case, they are respectively the function $g(\theta,x_i)$ and its argument. The available methods are \textit{coef}, \textit{vcov}, \textit{summary}, \textit{residuals}, \textit{fitted.values}, \textit{plot}, \textit{confint}. The model and data in a \textit{data.frame} format can be extracted by the generic function \textit{model.frame}. \subsection{Estimating the parameters of a normal distribution} This example\footnote{Thanks to Dieter Rozenich for his suggestion.}, is not something we want to do in practice, but its simplicity allows us to understand how to implement the \textit{gmm()} procedure by providing the gradient of $g(\theta,x_i)$. It is also a good example of the weakness of GMM when the moment conditions are not sufficiently informative. In fact, the ML estimators of the mean and the variance of a normal distribution are more efficient because the likelihood carries more information than few moment conditions. For the two parameters of a normal distribution $(\mu,\sigma)$ we have the following vector of moment conditions: \[ E[g(\theta,x_i)] \equiv E\left[ \begin{array}{c} \mu - x_{i} \\ \sigma^2 - (x_{i}-\mu)^2 \\ x_{i}^{3} - \mu (\mu^2+3\sigma^{2}) \end{array} \right] = 0 , \] where the first two can be directly obtained by the definition of $(\mu,\sigma)$ and the last comes from the third derivative of the moment generating function evaluated at 0. We first need to create a function $g(\theta,x)$ which returns an $n\times 3$ matrix: <<>>= g1 <- function(tet,x) { m1 <- (tet[1]-x) m2 <- (tet[2]^2 - (x - tet[1])^2) m3 <- x^3-tet[1]*(tet[1]^2+3*tet[2]^2) f <- cbind(m1,m2,m3) return(f) } @ The following is the gradient of $\bar{g}(\theta)$: \[ G\equiv \frac{\partial \bar{g}(\theta)}{\partial \theta} = \left( \begin{array}{cc} 1 & 0\\ 2(\bar{x}-\mu) & 2\sigma\\ -3(\mu^{2}+\sigma^{2}) & -6\mu\sigma \end{array} \right). \] If provided, it will be used to compute the covariance matrix of $\hat{\theta}$. It can be created as follows: <<>>= Dg <- function(tet,x) { G <- matrix(c( 1, 2*(-tet[1]+mean(x)), -3*tet[1]^2-3*tet[2]^2,0, 2*tet[2],-6*tet[1]*tet[2]), nrow=3,ncol=2) return(G) } @ First we generate normally distributed random numbers: <<>>= set.seed(123) n <- 200 x1 <- rnorm(n, mean = 4, sd = 2) @ We then run \textit{gmm} using the starting values $(\mu_0,\sigma^2_0)=(0,0)$ <<>>= print(res <- gmm(g1,x1,c(mu = 0, sig = 0), grad = Dg)) @ The \textit{summary} method prints more results from the estimation: <<>>= summary(res) @ The section "Initial values of the coefficients" shows the first step estimates used to either compute the weighting matrix in the 2-step GMM or the fixed bandwidth in CUE or iterative GMM. The J-test of over-identifying restrictions can also be extracted by using the method \textit{specTest}: <<>>= specTest(res) @ A small simulation using the following function shows that ML produces estimators with smaller mean squared errors than GMM based on the above moment conditions. However, it is not GMM but the moment conditions that are not efficient, because ML is GMM with the likelihood derivatives as moment conditions. <<>>= sim_ex <- function(n,iter) { tet1 <- matrix(0,iter,2) tet2 <- tet1 for(i in 1:iter) { x1 <- rnorm(n, mean = 4, sd = 2) tet1[i,1] <- mean(x1) tet1[i,2] <- sqrt(var(x1)*(n-1)/n) tet2[i,] <- gmm(g1,x1,c(0,0),grad=Dg)$coefficients } bias <- cbind(rowMeans(t(tet1)-c(4,2)),rowMeans(t(tet2)-c(4,2))) dimnames(bias)<-list(c("mu","sigma"),c("ML","GMM")) Var <- cbind(diag(var(tet1)),diag(var(tet2))) dimnames(Var)<-list(c("mu","sigma"),c("ML","GMM")) MSE <- cbind(rowMeans((t(tet1)-c(4,2))^2),rowMeans((t(tet2)-c(4,2))^2)) dimnames(MSE)<-list(c("mu","sigma"),c("ML","GMM")) return(list(bias=bias,Variance=Var,MSE=MSE)) } @ The following results can be reproduced with $n=50$, $iter=2000$ and by setting \textit{set.seed(345)}: \begin{center} \begin{tabular}{|c|c|c|c||c|c|c|} \hline &\multicolumn{3}{c||}{$\mu$}&\multicolumn{3}{c}{$\sigma$} \\ \hline & Bias & Variance & MSE & Bias & Variance &MSE \\ \hline GMM& 0.0020 &0.0929 & 0.0928& -0,0838 & 0.0481& 0.0551\\ ML& 0.0021 &0.0823 & 0.0822 & -0.0349 & 0.0411& 0.0423 \\ \hline \end{tabular} \end{center} \subsection{Estimating the parameters of a stable distribution} The previous example showed that ML should be used when the true distribution is known. However, when the density does not have a closed form expression, we have to consider other alternatives. \cite{garcia-renault-veredas06} propose to use indirect inference and perform a numerical study to compare it with several other methods. One of them is GMM for a continuum of moment conditions and was suggested by \cite{carrasco-florens02}. It uses the fact that the characteristic function $E(e^{ix_i\tau})$, where $i$ is the imaginary number and $\tau\in\Real$, has a closed form expression (for more details on stable distribution, see \cite{nolan09}). The \pkg{gmm} package does not yet deal with continuum of moment conditions but we can choose a certain grid $\{\tau_1,..., \tau_q\}$ over a given interval and estimate the parameters using the following moment conditions: \[ E\left[e^{ix_i\tau_l}-\Psi(\theta;\tau_l) \right]=0~~\mbox{for}~~l=1,...,q~, \] where $\Psi(\theta;\tau_l)$ is the characteristic function. There is more than one way to define a stable distribution and it depends on the choice of parametrization. We will follow the notation of \cite{nolan09} and consider stable distributions $S(\alpha,\beta,\gamma,\delta;1)$, where $\alpha\in (0,2]$ is the characteristic exponent and $\beta\in[-1,1]$, $\gamma>0$ and $\delta\in\Real$ are respectively the skewness, the scale and the location parameters. The last argument defines which parametrization we use. The \pkg{stabledist} package of \cite{stabledist} offers a function to generate random variables from stable distributions and uses the same notation. This parametrization implies that: \[ \Psi(\theta;\tau_l) = \left\{ \begin{array}{lcr} \exp{(-\gamma^\alpha|\tau_l|^\alpha[1-i\beta(\tan{\frac{\pi\alpha}{2}})(\mathrm{sign}(\tau_l))] + i\delta \tau_l)} &\mbox{for}&\alpha\neq 1\\ \exp{(-\gamma|\tau_l|[1+i\beta\frac{2}{\pi}(\mathrm{sign}(\tau_l))\log{|\tau_l|}] + i\delta \tau_l)} &\mbox{for}&\alpha= 1\\ \end{array} \right. , \] The function \textit{charStable} included in the package computes the characteristic function and can be used to construct $g(\theta,x_i)$. To avoid dealing with complex numbers, it returns the imaginary and real parts in separate columns because both should have zero expectation. The function is: <<>>= g2 <- function(theta,x) { tau <- seq(1,5,length.out=10) pm <- 1 x <- matrix(c(x),ncol=1) x_comp <- x%*%matrix(tau,nrow=1) x_comp <- matrix(complex(ima=x_comp),ncol=length(tau)) emp_car <- exp(x_comp) the_car <- charStable(theta,tau,pm) gt <- t(t(emp_car) - the_car) gt <- cbind(Im(gt),Re(gt)) return(gt) } @ The parameters of a simulated random vector can be estimated as follows (by default, $\gamma$ and $\delta$ are set to $1$ and $0$ respectively in \textit{rstable}). For the example, the starting values are the ones of a normal distribution with mean 0 and variance equals to \textit{var(x)}: <<>>= library(stabledist) set.seed(345) x2 <- rstable(500,1.5,.5,pm=1) t0 <- c(alpha = 2, beta = 0, gamma = sd(x2)/sqrt(2), delta = 0) print(res <- gmm(g2,x2,t0)) @ The result is not very close to the true parameters. But we can see why by looking at the J-test that is provided by the \textit{summary} method: <<>>= summary(res) @ The null hypothesis that the moment conditions are satisfied is rejected. For nonlinear models, a significant J-test may indicate that we have not reached the global minimum. Furthermore, the standard deviation of the coefficient of $\delta$ indicates that the covariance matrix is nearly singular. Notice also that the convergence code is equal to 1, which indicates that the algorithm did not converge. We could try different starting values, increase the number of iterations in the control option of \textit{optim} or use \textit{nlminb} which allows to put restrictions on the parameter space. The former would work but the latter will allow us to see how to select another optimizer. The option \textit{optfct} can be modified to use this algorithm instead of \textit{optim}. In that case, we can specify the upper and lower bounds of $\theta$. <<>>= res2 <- gmm(g2,x2,t0,optfct="nlminb",lower=c(0,-1,0,-Inf),upper=c(2,1,Inf,Inf)) summary(res2) @ Although the above modification solved the convergence problem, there is another issue that we need to address. The first step estimate used to compute the weighting matrix is almost identical to the starting values. There is therefore a convergence problem in the first step. In fact, choosing the initial $\alpha$ to be on the boundary was not a wise choice. Also, it seems that an initial value of $\beta$ equals to zero makes the objective function harder to minimize. Having a gobal minimum for the first step estimate is important if we care about efficiency. A wrong estimate will cause the weighting matrix not being a consistent estimate of the optimal matrix. The information about convergence is included in the argument 'initialAlgoInfo' of the gmm object. We conclude this example by estimating the parameters for a vector of stock returns from the data set \textit{Finance} that comes with the \pkg{gmm} package. <>= data(Finance) x3 <- Finance[1:1500,"WMK"] t0<-c(alpha = 1.8, beta = 0.1, gamma = sd(x3)/sqrt(2),delta = 0) res3 <- gmm(g2,x3,t0,optfct="nlminb") summary(res3) @ For this sub-sample, the hypothesis that the return follows a stable distribution is rejected. The normality assumption can be analyzed by testing $H_0:\alpha=2,\beta=0$ using \textit{linearHypothesis} from the \pkg{car} package: <<>>= library(car) linearHypothesis(res3,cbind(diag(2),c(0,0),c(0,0)),c(2,0)) @ It is clearly rejected. The result is even stronger if the whole sample is used. \subsection{A linear model with iid moment conditions} We want to estimate a linear model with an endogeneity problem. It is the model used by \cite{carrasco07} to compare several methods which deal with the many instruments problem. We want to estimate $\delta$ from: \[ y_i = \delta W_i + \epsilon_i \] with $\delta=0.1$ and \[ W_i = e^{-x_i^2} + u_i , \] where $(\epsilon_i,u_i) \sim ~iidN(0,\Sigma)$ with \[ \Sigma = \left( \begin{array}{cc} 1&0.5\\ 0.5&1 \end{array} \right) \] Any function of $x_i$ can be used as an instrument because it is orthogonal to $\epsilon_i$ and correlated with $W_i$. There is therefore an infinite number of possible instruments. For this example, $(x_i,x_i^2,x_i^3)$ will be the selected instruments and the sample size is set to $n=400$: <<>>= library(mvtnorm) set.seed(112233) sig <- matrix(c(1,.5,.5,1),2,2) n <- 400 e <- rmvnorm(n,sigma=sig) x4 <- rnorm(n) w <- exp(-x4^2) + e[,1] y <- 0.1*w + e[,2] @ where \textit{rmvnorm} is a multivariate normal distribution random generator which is included in the package \pkg{mvtnorm} (\cite{mvtnorm}). For a linear model, the $g$ argument is a formula that specifies the right- and left-hand sides as for \textit{lm} and $x$ is the matrix of instruments: <<>>= h <- cbind(x4, x4^2, x4^3) g3 <- y~w @ By default, an intercept is added to the formula and a vector of ones to the matrix of instruments. It implies the following moment conditions: \[ E\left(\begin{array}{c} (y_i-\alpha-\delta W_i)\\ (y_i-\alpha-\delta W_i)x_i\\ (y_i-\alpha-\delta W_i)x_i^2\\ (y_i-\alpha-\delta W_i)x_i^3 \end{array} \right)=0 \] In order the remove the intercept, -1 has to be added to the formula. In that case there is no column of ones added to the matrix of instruments. To keep the condition that the expected value of the error terms is zero, the column of ones needs to be included manually. We know that the moment conditions of this example are iid. Therefore, we can add the option \textit{vcov="iid"}. This option tells \textit{gmm} to estimate the covariance matrix of $\sqrt{n}\bar{g}(\theta^*)$ as follows: \[ \hat{\Omega}(\theta^*) = \frac{1}{n}\sum_{i=1}^n g(\theta^*,x_i)g(\theta^*,x_i)' \] However, it is recommended not to set this option to ``iid" in practice with real data because one of the reasons we want to use GMM is to avoid such restrictions. Finally, it is not necessary to provide the gradient when the model is linear since it is already included in \textit{gmm}. The first results are: <<>>= summary(res <- gmm(g3,x=h)) @ By default, the 2SGMM is computed. Other methods can be chosen by modifying the option ``type". The second possibility is ITGMM: <<>>= res2 <- gmm(g3,x=h,type='iterative',crit=1e-8,itermax=200) coef(res2) @ The procedure iterates until the difference between the estimates of two successive iterations reaches a certain tolerance level, defined by the option \textit{crit} (default is $10^{-7}$), or if the number of iterations reaches \textit{itermax} (default is 100). In the latter case, a message is printed to indicate that the procedure did not converge. The third method is CUE. As you can see, the estimates from ITGMM is used as starting values. However, the starting values are required only when \textit{g} is a function. When \textit{g} is a formula, the default starting values are the ones obtained by setting the matrix of weights equal to the identity matrix. <<>>= res3 <- gmm(g3,x=h,res2$coef,type='cue') coef(res3) @ It is possible to produce confidence intervals by using the method \textit{confint}: <<>>= confint(res3,level=.90) @ Whether \textit{optim} or \textit{nlminb} is used to compute the solution, it is possible to modify their default options by adding \textit{control=list()}. For example, you can keep track of the convergence with \textit{control=list(trace=TRUE)} or increase the number of iterations with \textit{control=list(maxit=1000)}. You can also choose the \textit{BFGS} algorithm with \textit{method="BFGS"} (see \textit{help(optim)} for more details). The methods \textit{fitted} and \textit{residuals} are also available for linear models. We can compare the fitted values of \textit{lm} with the ones from \textit{gmm} to see why this model cannot be estimated by LS. \begin{center} <<>>= plot(w,y,main="LS vs GMM estimation") lines(w,fitted(res),col=2) lines(w,fitted(lm(y~w)),col=3,lty=2) lines(w,.1*w,col=4,lty=3) legend("topleft",c("Data","Fitted GMM","Fitted LS","True line"),pch=c(1,NA,NA,NA),col=1:3,lty=c(NA,1,2,3)) @ \end{center} The LS seems to fit the model better. But the graphics hides the endogeneity problem. LS overestimates the relationship between $y$ and $w$ because it does not take into account the fact that some of the correlation is caused by the fact that $y_i$ and $w_i$ are positively correlated with the error term $\epsilon_i$. Finally, the \textit{plot} method produces some graphics to analyze the properties of the residuals. It can only be applied to \textit{gmm} objects when \textit{g} is a formula because when \textit{g} is a function, residuals are not defined. \subsection{Estimating the AR coefficients of an ARMA process} \label{ar} The estimation of auto-regressive coefficients of ARMA(p,q) processes is better performed by ML or nonlinear LS. But in Monte Carlo experiments, it is often estimated by GMM to study its properties. It gives a good example of linear models with endogeneity problems in which the moment conditions are serially correlated and possibly conditionally heteroskedastic. As opposed to the previous example, the choice of the HAC matrix becomes an important issue. We want to estimate the AR coefficients of the following process: \[ X_t = 1.4 X_{t-1} - 0.6X_{t-2} + u_t \] where $u_t = 0.6\epsilon_{t-1} -0.3 \epsilon_{t-2} + \epsilon_t$ and $\epsilon_t\sim iidN(0,1)$. This model can be estimated by GMM using any $X_{t-s}$ for $s>2$, because they are uncorrelated with $u_t$ and correlated with $X_{t-1}$ and $X_{t-2}$. However, as $s$ increases the quality of the instruments decreases since the stationarity of the process implies that the auto-correlation goes to zero. For this example, the selected instruments are $(X_{t-3},X_{t-4},X_{t-5},X_{t-6})$ and the sample size equals 400. The ARMA(2,2) process is generated by the function \textit{arima.sim}: <<>>= t <- 400 set.seed(345) x5 <- arima.sim(n=t,list(ar=c(1.4,-0.6),ma=c(0.6,-0.3))) x5t<-cbind(x5) for (i in 1:6) x5t<-cbind(x5t,lag(x5,-i)) x5t<-na.omit(x5t) g4<-x5t[,1]~x5t[,2]+x5t[,3] res<-gmm(g4,x5t[,4:7]) summary(res) @ The optimal matrix, when moment conditions are based on time series, is an HAC matrix which is defined by equation (\ref{optw_hat}). Several estimators of this matrix have been proposed in the literature. Given some regularity conditions, they are asymptotically equivalent. However, their impacts on the finite sample properties of GMM estimators may differ. The \pkg{gmm} package uses the \pkg{sandwich} package to compute these estimators which are well explained by \cite{zeileis06} and \cite{zeileis04}. We will therefore briefly summarize the available options. The option \textit{kernel} allows to choose between five kernels: Truncated, Bartlett, Parzen, Tukey-Hanning and Quadratic spectral\footnote{The first three have been proposed by \cite{white84}, \cite{newey-west87a} and \cite{gallant87} respectively and the last two, applied to HAC estimation, by \cite{andrews91}. But the latter gives a good review of all five.}. By default, the Quadratic Spectral kernel is used as it was shown to be optimal by \cite{andrews91} with respect to some mean squared error criterion. In most statistical packages, the Bartlett kernel is used for its simplicity. It makes the estimation of large models less computationally intensive. It may also make the \textit{gmm} algorithm more stable numerically when dealing with highly nonlinear models, especially with CUE. We can compare the results with different choices of kernel: <<>>= res2 <- gmm(g4,x=x5t[,4:7],kernel="Truncated") coef(res2) res3 <- gmm(g4,x=x5t[,4:7],kernel="Bartlett") coef(res3) res4 <- gmm(g4,x=x5t[,4:7],kernel="Parzen") coef(res4) res5<- gmm(g4,x=x5t[,4:7],kernel="Tukey-Hanning") coef(res5) @ The similarity of the results is not surprising since the matrix of weights should only affect the efficiency of the estimator. We can compare the estimated standard deviations using the method \textit{vcov}: <<>>= diag(vcov(res2))^.5 diag(vcov(res3))^.5 diag(vcov(res4))^.5 diag(vcov(res5))^.5 @ which shows, for this example, that the Bartlett kernel generates the estimates with the smallest variances. However, it does not mean it is better. We have to run simulations and compute the true variance if we want to compare them. In fact, we do not know which one produces the most accurate estimate of the variance. The second options is for the bandwidth selection. By default it is the automatic selection proposed by \cite{andrews91}. It is also possible to choose the automatic selection of \cite{newey-west94} by adding \textit{bw=bwNeweyWest} (without quotes because \textit{bwNeweyWest} is a function). A prewhitened kernel estimator can also be computed using the option \textit{prewhite=p}, where $p$ is the order of the vector auto-regressive (VAR) used to compute it. By default, it is set to \textit{FALSE}. \cite{andrews-monahan92} show that a prewhitened kernel estimator improves the properties of hypothesis tests on parameters. Finally, the \textit{plot} method can be applied to \textit{gmm} objects to do a Q-Q plot of the residuals: \begin{center} <<>>= plot(res,which=2) @ \end{center} or to plot the observations with the fitted values: \begin{center} <<>>= plot(res,which=3) @ \end{center} \subsection{Estimating a system of equations: CAPM} We want to test one of the implications of the capital asset pricing model (CAPM). This example comes from \cite{campbell-lo-mackinlay96}. It shows how to apply the \pkg{gmm} package to estimate a system of equations. The theory of CAPM implies that $\mu_i-R_f = \beta_i (\mu_m-R_f)$ $\forall i$, where $\mu_i$ is the expected value of stock i's return, $R_f$ is the risk free rate and $\mu_m$ is the expected value of the market porfolio's return. The theory can be tested by running the following regression: \[ (R_t-R_f) = \alpha + \beta (R_{mt}-R_f) + \epsilon_t, \] where $R_t$ is a $N\times 1$ vector of observed returns on stocks, $R_{mt}$ if the observed return of a proxy for the market portfolio, $R_f$ is the interest rate on short term government bonds and $\epsilon_t$ is a vector of error terms with covariance matrix $\Sigma_t$. When estimated by ML or LS, $\Sigma$ is assumed to be fixed. However, GMM allows $\epsilon_t$ to be heteroskedastic and serially correlated. One implication of the CAPM is that the vector $\alpha$ should be zero. It can be tested by estimating the model with $(R_{mt}-R_f)$ as instruments, and by testing the null hypothesis $H_0:~\alpha=0$. The data, which are included in the package, are the daily returns of twenty selected stocks from January 1993 to February 2009, the risk-free rate and the three factors of Fama and French\footnote{The symbols of the stocks taken from \url{http://ca.finance.yahoo.com/} are ("WMK", "UIS", "ORB", "MAT", "ABAX","T", "EMR", "JCS", "VOXX", "ZOOM", "ROG", "GGG", "PC", "GCO", "EBF", "F", "FNM", "NHP", "AA", "TDW"). The four other series can be found on K. R. French's web site: \url{http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data\_library.html}}. The following test is performed using the returns of 5 stocks and a sample size of 500\footnote{The choice of sample size is arbitrary. The purpose is to show how to estimate a system of equations not to test the CAPM. Besides, the $\beta$'s seem to vary over time. It is therefore a good practice to estimate the model using short periods.}. <<>>= data(Finance) r <- Finance[1:500,1:5] rm <- Finance[1:500,"rm"] rf <- Finance[1:500,"rf"] z <- as.matrix(r-rf) zm <- as.matrix(rm-rf) res <- gmm(z~zm,x=zm) coef(res) R <- cbind(diag(5),matrix(0,5,5)) c <- rep(0,5) linearHypothesis(res,R,c,test = "Chisq") @ where the asymptotic chi-square is used since the default distribution requires a normality assumption. The same test could have been performed using the names of the coefficients: <>= test <- paste(names(coef(res)[1:5])," = 0",sep="") linearHypothesis(res,test) @ Another way to test the CAPM is to estimate the restricted model ($\alpha=0$), which is over-identified, and to perform a J-test. Adding $-1$ to the formula removes the intercept. In that case, a column of ones has to be added to the matrix of instruments: <<>>= res2<-gmm(z~zm-1,cbind(1,zm)) specTest(res2) @ which confirms the non-rejection of the theory. \subsection{Testing the CAPM using the stochastic discount factor representation} In some cases the theory is directly based on moment conditions. When it is the case, testing the validity of these conditions becomes a way of testing the theory. \cite{jagannathan-skoulakis02} present several GMM applications in finance and one of them is the stochastic discount factor (SDF) representation of the CAPM. The general theory implies that $E(m_tR_{it})=1$ for all $i$, where $m_t$ is the SDF and $R_{it}$ the gross return ($1 + r_{it}$). It can be shown that if the CAPM holds, $m_t=\theta_0 + \theta_0 R_{mt}$ which implies the following moment conditions: \[ E\Big[R_{it}(\theta_0-\theta_1R_{mt})-1 \Big] = 0~~\mbox{for}~~ i=1,...,N \] which can be tested as follows: <<>>= g5 <- function(tet, x) { gmat <- (tet[1] + tet[2] * (1 + c(x[, 1]))) * (1 + x[, 2:6]) - 1 return(gmat) } res_sdf <- gmm(g5, x = as.matrix(cbind(rm, r)), c(0, 0)) specTest(res_sdf) @ which is consistent with the two previous tests. \subsection{Estimating continuous time processes by discrete time approximation} This last example also comes from \cite{jagannathan-skoulakis02}. We want to estimate the coefficients of the following continuous time process which is often used in finance for interest rates: \[ dr_t = (\alpha + \beta r_t)dt + \sigma r_t^\gamma dW_t, \] where $W_t$ is a standard Brownian motion. Special cases of this process are the Brownian motion with drift ($\beta=0$ and $\gamma = 0$), the Ornstein-Uhlenbeck process ($\gamma=0$) and the Cox-Ingersoll-Ross or square root process ( $\gamma = 1/2$). It can be estimated using the following discrete time approximation: \[ r_{t+1}-r_t = \alpha + \beta r_t + \epsilon_{t+1} \] with \[ E_t \epsilon_{t+1}=0,~~\mbox{and}~~E_t(\epsilon_{t+1}^2) = \sigma^2r_t^{2\gamma} \] Notice that ML cannot be used to estimate this model because the distribution depends on $\gamma$. In particular, it is normal for $\gamma=0$ and gamma for $\gamma=1/2$. It can be estimated by GMM using the following moment conditions: \[ E[g(\theta,x_t)] \equiv E\left(\begin{array}{c} \epsilon_{t+1}\\ \epsilon_{t+1}r_t\\ \epsilon_{t+1}^2-\sigma^2r_t^{2\gamma}\\ (\epsilon_{t+1}^2-\sigma^2r_t^{2\gamma})r_t \end{array} \right) = 0 \] The related \textit{g} function, with $\theta=\{\alpha, \beta, \sigma^2, \gamma\}$ is: <<>>= g6 <- function(theta, x) { t <- length(x) et1 <- diff(x) - theta[1] - theta[2] * x[-t] ht <- et1^2 - theta[3] * x[-t]^(2 * theta[4]) g <- cbind(et1, et1 * x[-t], ht, ht * x[-t]) return(g) } @ In order to estimate the model, the vector of interest rates needs to be properly scaled to avoid numerical problems. The transformed series is the annualized interest rates expressed in percentage. Also, the starting values are obtained using LS and some options for \textit{optim} need to be modified. <<>>= rf <- Finance[,"rf"] rf <- ((1 + rf/100)^(365) - 1) * 100 dr <- diff(rf) res_0 <- lm(dr ~ rf[-length(rf)]) tet0 <- c(res_0$coef, var(residuals(res_0)), 0) names(tet0) <- c("alpha", "beta", "sigma^2", "gamma") res_rf <- gmm(g6, rf, tet0, control = list(maxit = 1000, reltol = 1e-10)) coef(res_rf) @ \subsection{Comments on models with panel data} The \pkg{gmm} package is not directly built to easily deal with panel data. However, it is flexible enough to make it possible in most cases. To see that, let us consider the following model (see \cite{wooldridge02} for more details): \[ y_{it} = x_{it} \beta + a_i + \epsilon_{it}\mbox{ for } i=1,...,N\mbox{ and }t=1,...,T, \] where $x_{it}$ is $1\times k$, $\beta$ is $k\times 1$, $\epsilon_{it}$ is an error term and $a_i$ is an unobserved component which is specific to individual $i$. If $a_i$ is correlated with $x_{it}$, it can be removed by subtracting the average of the equation over time, which gives: \[ (y_{it}-\bar{y}_i) = (x_{it}-\bar{x}_i) \beta + (\epsilon_{it}-\bar{\epsilon}_i) \mbox{ for } i=1,...,N\mbox{ and } t=1,...,T, \] which can be estimated by \textit{gmm}. For example, if there are 3 individuals the following corresponds to the GMM fixed effects estimation: <>= y <- rbind(y1-mean(y1),y2-mean(y2),y3-mean(y3)) x <- rbind(x1-mean(x1),x2-mean(x2),x3-mean(x3)) res <- gmm(y~x,h) @ However, if $a_i$ is not correlated with $x_{it}$, the equation represents a random effects model. In that case, it is more efficient not to remove $a_i$ from the equation because of the information it carries about the individuals. The error terms are then combined in a single one, $\eta_{it}=(a_i + \epsilon_{it})$ to produce the linear model: \[ y_{it} = x_{it} \beta + \eta_{it} \] This model cannot be efficiently estimated by OLS because the presence of the common factor $a_i$ at each period implies that $\eta_{it}$ is serially correlated. However, GMM is well suited to deal with such specifications. The following will therefore produce a GMM random effects estimation: <>= y <- rbind(y1,y2,y3) x <- rbind(x1,x2,x3) res <- gmm(y~x,h) @ The package \pkg{plm} of \cite{plm} offers several functions to manipulate panel data. It could therefore be combined with \pkg{gmm} when estimating such models. It also offers a way to estimate them with its own GMM algorithm for panel data. \subsection{GMM and the sandwich package} In the \pkg{gmm} package, the estimation of the optimal weighting matrices are obtained using the \pkg{sandwich} package of \cite{zeileis06}. For example, the weighting matrix of the two-step GMM defined as: \[ W = \left[\lim_{n\rightarrow \infty} Var{(\sqrt{n}\bar{g})}\right]^{-1} \] is estimated as follows: <>= gt <- g(t0, x) V <- kernHAC(lm(gt~1),sandwich = FALSE) W <- solve(V) @ where $t0$ is any consistent estimate. As long as the optimal matrix is used, the covariance matrix of the coefficients can be estimated as follows: \[ (\hat{G}'W\hat{G})^{-1}/n \equiv (\hat{G}'\hat{V}^{-1}\hat{G})^{-1}/n, \] where $\hat{G}=d\bar{g}(\hat{\theta})/d\theta$ and $\hat{V}$ is obtained using \textit{kernHAC()}. It is not a sandwich covariance matrix and is computed using the \textit{vcov()} method included in \pkg{gmm}. However, if any other weighting matrix is used, say $W$, the estimated covariance matrix of the coefficients must then be estimated as follows: \[ (\hat{G}'W\hat{G})^{-1}\hat{G}'W\hat{V}W\hat{G}(\hat{G}'W\hat{G})^{-1}/n. \] A \textit{bread()} and \textit{estfun()} methods are available for \textit{gmm} objects which allows to compute the above matrix using the \pkg{sandwich} package. The \textit{bread()} method computes $(\hat{G}'W\hat{G})^{-1}$ while the \textit{estfun()} method returns a $T\times q$ matrix with the $t^{th}$ row equals to $g(\hat{\theta}, x_t) W \hat{G}$. The \textit{meatHAC()} method applied to the latter produces the right meat. Let us consider the example of section (\ref{ar}). Suppose we want to use the identity matrix to eliminate one source of bias, at the cost of lower efficiency. In that case, a consistent estimate of the covariance matrix is \[ (\hat{G}'\hat{G})^{-1}\hat{G}'\hat{V}\hat{G}(\hat{G}'\hat{G})^{-1}/n, \] which can be computed as: <<>>= print(res<-gmm(g4,x5t[,4:7],wmatrix="ident")) diag(vcovHAC(res))^.5 @ which is more robust than using \textit{vcov()}: <<>>= diag(vcov(res))^.5 @ Notice that it is possible to fixe $W$. Therefore, the above results can also be obtained as: <<>>= print(res<-gmm(g4,x5t[,4:7], weightsMatrix = diag(5))) @ In this case, the choice of the type of GMM is irrelevant since the weighting matrix is fixed. \section{Generalized empirical likelihood} The GEL is a new family of estimation methods which, as GMM, is based on moment conditions. It follows \cite{owen01} who developed the idea of empirical likelihood estimation which was meant to improve the confidence regions of estimators. We present here a brief discussion on the method without going into too much details. For a complete review, see \cite{smith97}, \cite{newey-smith04} or \cite{anatolyev05}. The estimation is based on \[ E(g(\theta_0,x_i))=0, \] which can be estimated in general by \[ \tilde{g}(\theta) = \sum_{i=1}^n p_i g(\theta,x_i) =0, \] where $p_i$ is called the implied probability associated with the observation $x_i$. For the GEL method, it is assumed that $q>p$ because otherwise it would correspond to GMM. Therefore, as it is the case for GMM, there is no solution to $\bar{g}(\theta)=0$. However, there is a solution to $\tilde{g}(\theta)=0$ for some choice of the probabilities $p_i$ such that $\sum_i p_i=1$. In fact, there is an infinite number of solutions since there are $(n+q)$ unknowns and only $q+1$ equations. GEL selects among them the one for which the distance between the vector of probabilities $p$ and the empirical density $1/n$ is minimized. The empirical likelihood of \cite{owen01} is a special case in which the distance is the likelihood ratio. The other methods that belong to the GEL family of estimators use different metrics. If the moment conditions hold, the implied probabilities carry a lot of information about the stochastic properties of $x_i$. For GEL, the estimations of the expected value of the Jacobian and the covariance matrix of the moment conditions, which are required to estimate $\theta$, are based on $p_i$ while in GMM they are estimated using $1/n$. \cite{newey-smith04} show that this difference explains partially why the second order properties of GEL are better. Another difference between GEL and GMM is how they deal with the fact that $g(\theta,x_i)$ can be a conditionally heteroskedastic and weakly dependent process. GEL does not require to compute explicitly the HAC matrix of the moment conditions. However, if it does not take it into account, its estimators may not only be inefficient but may also fail to be consistent. \cite{smith01} proposes to replace $g(\theta,x_i)$ by: \[ g^w(\theta,x_i) = \sum_{s=-m}^m w(s)g(\theta,x_{i-s}) \] where $w(s)$ are kernel based weights that sum to one (see also \cite{kitamura-stutzer97} and \cite{smith97}). The sample moment conditions become: \begin{equation}\label{gel_mcond} \tilde{g}(\theta) = \sum_{i=1}^n p_i g^w(\theta,x_i) =0 \end{equation} The estimator is defined as the solution to the following constrained minimization problem: \begin{eqnarray} \hat{\theta}_n &=& \arg\min_{\theta,p_i} \sum_{i=1}^n h_n(p_i) ,\\ && \mbox{subject to } \\ && \sum_{i=1}^n p_ig^w(\theta,x_i) = 0 ~~\mbox{and} \label{const}\\ && \sum_{i=1}^n p_i=1, \end{eqnarray} where $h_n(p_i)$ has to belong to the following Cressie-Read family of discrepancies: \[ h_n(p_i) = \frac{[\gamma(\gamma+1)]^{-1}[(np_i)^{\gamma+1}-1]}{n} . \] \cite{smith97} showed that the empirical likelihood method (EL) of \cite{owen01} ($\gamma=0$) and the exponential tilting of \cite{kitamura-stutzer97} ($\gamma=-1$) belong to the GEL family of estimators while \cite{newey-smith04} show that it is also the case for the continuous updated estimator of \cite{hansen-heaton-yaron96} ($\gamma=1$). What makes them part of the same GEL family of estimation methods is the existence of a dual problem which is defined as: \begin{equation}\label{gel_obj} \hat{\theta} = \arg\min_{\theta}\left[\max_{\lambda} P_n(\theta,\lambda) = \frac{1}{n}\sum_{i=1}^n\rho\left(\lambda'g^w(\theta,x_i)\right)\right] \end{equation} where $\lambda$ is the Lagrange multiplier associated with the constraint (\ref{const}) and $\rho(v)$ is a strictly concave function normalized so that $\rho'(0)=\rho''(0)=-1$. It can be shown that $\rho(v)=\ln{(1-v)}$ corresponds to EL , $\rho(v)=-\exp{(v)}$ to ET and to CUE if it is quadratic. The equivalence of the primal and dual problems can easily be verified by showing that they both share the same following first order conditions: \begin{equation}\label{lam} \sum_{i=1}^n p_i g^w(\theta,x_i) = 0, \end{equation} \begin{equation} \sum_{i=1}^n p_i \lambda'\left(\frac{\partial g^w(\theta,x_i)}{\partial \theta}\right) = 0 , \end{equation} with \begin{equation} p_i = \frac{1}{n}\rho'\left(\lambda'g^w(\theta,x_i)\right) . \end{equation} Equation (\ref{gel_obj}) represents a saddle point problem. The solution is obtained by solving simultaneously two optimization problems. We can solve for $\theta$ by minimizing $P_n(\theta,\lambda(\theta))$, where $\lambda(\theta)$ is the solution to $\arg\max_\lambda P_n(\theta,\lambda)$ for a given $\theta$. Therefore an optimization algorithm needs to be called inside the $P_n(\theta,\lambda)$ function. It makes the GEL very hard to implement numerically. For example, \cite{guggenberger08}, who analyzes the small sample properties of GEL, uses an iterative procedure based on the Newton method for $\lambda$ and a grid search for $\theta$ in order to confidently reach the absolute minimum. Using such iterative procedures for $\lambda$ makes the problem less computationally demanding and does not seem to affect the properties of the estimator of $\theta_0$. Indeed, \cite{guggenberger-hahn05} show that going beyond two iterations for $\lambda$ does not improve the second order asymptotic properties of the estimator of $\theta_0$. The function \textit{gel} offers two options. By default, $\lambda(\theta)$ is obtained by the following iterative method: \[ \lambda_l = \lambda_{l-1} - \left[\frac{1}{n}\sum_{i=1}^n \rho''(\lambda_{l-1}'g_t)g_tg_t'\right]^{-1}\left[\frac{1}{n} \sum_{i=1}^n \rho'(\lambda_{l-1}'g_i) g_i \right] \] starting with $\lambda=0$, which corresponds to its asymptotic value. The algorithm stops when $\|\lambda_l-\lambda_{l-1}\|$ reaches a certain tolerance level. The second option is to let \textit{optim} solve the problem. Then, as for \textit{gmm}, the minimization problem is solved either by \textit{optim}, \textit{nlminb} or \textit{constrOptim}. In order to test the over-identifying restrictions, \cite{smith04} proposes three tests which are all asymptotically distributed as a $\chi^2_{q-p}$. The first one is the J-test: \[ n \bar{g}^w(\hat{\theta})'[\hat{\Omega}(\hat{\theta})]^{-1}\bar{g}^w(\hat{\theta}), \] the second is a Lagrange multiplier test (LM): \[ LM = n\hat{\lambda}'\hat{\Omega}(\hat{\theta})\hat{\lambda} \] and the last one is a likelihood ratio test (LR): \[ LR = 2\sum_{i=1}^n\left[ \rho\left(\hat{\lambda}'g^w(\hat{\theta},x_i)\right) - \rho(0)\right] \] \section{GEL with R} \subsection{Estimating the parameters of a normal distribution} For this example, we can leave the option \textit{smooth} at its default value, which is \textit{FALSE}, because of the iid properties of $x$. A good starting value is very important for GEL. The best choice is the sample mean and the standard deviation. By default the option \textit{type} is set to \textit{EL}. The same methods that apply to \textit{gmm} objects, can also be applied to \textit{gel} objects. <<>>= tet0 <- c(mu = mean(x1), sig = sd(x1)) res_el <- gel(g1,x1,tet0) summary(res_el) @ Each Lagrange multiplier represents a shadow price of the constraint implied by moment condition. A binding constraint will produce a multiplier different from zero. Therefore, its value informs us on the validity of the moment condition. In the above results, the $\lambda$'s are significantly different from zero which would normally suggest that the moment conditions associated with them are violated. As a result, the LM test also rejects the null hypothesis since it is based on the $\lambda$'s. Notice that \textit{summary} reports two convergence codes, one for $\lambda$ et another for $\theta$. The ET and CUE estimates can be obtained as follows: <<>>= res_et <- gel(g1,x1,tet0,type="ET") coef(res_et) @ <<>>= res_cue <- gel(g1,x1,tet0,type="CUE") coef(res_cue) @ A fourth method is available which is called the exponentially tilted empirical likelihood (ETEL) and was proposed by \cite{schennach07}. However, it does not belong to the family of GEL estimators. It solves the problem of misspecified models. In such models there may not exist any pseudo value to which $\hat{\theta}$ converges as the sample size increases. ETEL uses the $\rho()$ of ET to solve for $\lambda$ and the $\rho()$ of EL to solve for $\theta$. \cite{schennach07} shows that ETEL shares the same asymptotic properties of EL without having to impose restrictions on the domain of $\rho(v)$ when solving for $\lambda$. <<>>= res_etel <- gel(g1,x1,c(mu=1,sig=1),type="ETEL") coef(res_etel) @ The type ETEL is experimental for now. Although it is supposed to be more stable because no restrictions are required to solve for $\lambda$, once we substitute $\lambda(\theta)$ in the EL objective function to estimate $\theta$, we still need to restrict $\lambda'g_t$ to avoid having NA's. The solution used in gel() is to obtain $\lambda(\theta)$ with \textit{constrOptim} with the restriction $\lambda'gt > 1$ even if it is not required by ET ($\rho(v)=-\exp{(v)}$). It is however sensitive to starting values. That's the reason why we used different ones above. \subsection{Estimating the AR coefficients of an ARMA process} Because the moment conditions are weakly dependent, we need to set the option \textit{smooth=TRUE}. Before going to the estimation procedure, we need to understand the relationship between the smoothing kernel and the HAC estimator. The reason why we need to smooth the moment function is that GEL estimates the covariance matrix of $\bar{g}(\theta,x_t)$, as if we had iid observations, using the expression $(1/T)\sum_{t=1}^T(g_tg_t')$. We can show that substituting $g_t$ by $g^w_t$ in this expression results in an HAC estimator. However, the relationship between the smoothing kernel and the kernel that appears in the HAC estimator is not obvious. For example, we can show that if the smoothing kernel is Truncated, then the kernel in the HAC estimator is the Bartlett. Let us consider the truncated kernel with a bandwidth of 2. This implies that $w(s)=1/5$ for $|s|\leq 2$ and 0 otherwise. Then, the expression for the covariance matrix becomes: \begin{eqnarray*} \frac{1}{T}\sum_{t=1}^T g^w_t (g^w_t)' &=& \frac{1}{T}\sum_{t=1}^T \left( \sum_{s=-2}^2 \frac{1}{5} g_{t+s} \right) \left( \sum_{l=-2}^2 \frac{1}{5} g_{t+l}' \right),\\ &=& \frac{1}{25}\sum_{s=-2}^2\sum_{l=-2}^2\left( \frac{1}{T}\sum_{t=1}^T g_{t+s}g_{t+l}' \right),\\ &=& \frac{1}{25}\sum_{s=-2}^2\sum_{l=-2}^2\hat{\Gamma}_{s-l} ,\\ &=& \frac{1}{25}\sum_{s=-4}^4 (5-|s|)\hat{\Gamma}_{s},\\ &=& \sum_{s=-4}^4 \left(\frac{1}{5}-\frac{|s|}{25}\right)\hat{\Gamma}_{s},\\ &=& \sum_{s=-T+1}^{T-1} k_5(s) \hat{\Gamma}_{s}, \end{eqnarray*} where $k_5(s)$ is the Bartlett kernel with a bandwidth of 5 defined as \[ K_5(s) = \left\{ \begin{array}{ccc} 1/5 - |s|/25 & \mbox{if}&|s|\leq 5\\ 0 &\mbox{otherwise}& \end{array}\right. . \] See \cite{smith01} for more details. The model will therefore be estimated using the kernel Truncated. The GMM estimate with the identity matrix is selected as starting value. <>= tet0 <- gmm(g4,x=x5t[,4:7],wmatrix="ident")$coef res <- gel(g4,x=x5t[,4:7],tet0,smooth=TRUE,kernel="Truncated") summary(res) @ The \textit{specTest} method applied to a \textit{gel} object computes the three tests proposed by \cite{smith04}: <<>>= specTest(res) @ The \textit{plot} method produces one more graphics when applied to a \textit{gel} object. It shows the implied probabilities along with the empirical density $(1/T)$. It allows to see which observations have more influence: \begin{center} <<>>= plot(res,which=4) @ \end{center} We can also select \textit{optfct="nlminb"} or \textit{constraint=TRUE} in order to impose restrictions on the coefficients. The former sets lower and upper bounds for the coefficients, while the latter imposes linear constraints using the algorithm \textit{constrOptim}. In this example we want the sum of the AR coefficients to be less than one. \textit{constrOptim} imposes the constraint $ui\theta-ci \geq 0$. Therefore, we need to set: <>= ui=cbind(0,-1,-1) ci <- -1 @ and rerun the estimation as <>= res <- gel(g4,x=dat5[,4:7],tet0,smooth=TRUE,kernel="Truncated", constraint=TRUE, ui=ui,ci=ci) @ The result, which is not shown, is identical. They are also many option to compute the $\lambda$'s. From version 1.4-0, the default algorithm is \textit{nlminb} because the gradient and Hessian matrix are well defined analytically which speed up convergence. The other choices are \textit{optim} or "iter" which uses a Newton method to solve the first order condition. If the option optlam is set to "optim" and the type is EL, \textit{contrOptim} is selected automatically to restrict $\lambda'g_t$ to be less than 1. It is also possible to change the default values in the control list of the optimizer with the option LambdaControl (see \textit{?nlminb} or \textit{?optim}). Here are some examples: <>= res <- gel(g4,x=dat5[,4:7],tet0,smooth=TRUE, optlam="optim") res <- gel(g4,x=dat5[,4:7],tet0,smooth=TRUE, optlam="optim", LambdaControl=list(trace=TRUE, parscale=rep(.1,5))) @ \subsection{Comments} The GEL method is very unstable numerically. This fact has been reported many times in the recent literature. The method has been included in the \pkg{gmm} package because recent theoretical evidence suggests that it may produce better estimators than GMM. Because \proglang{R} is an open source statistical package, it offers a good platform to experiment with numerical properties of estimators. \section{Conclusion} The \pkg{gmm} package offers complete and flexible algorithms to estimate models by GMM and GEL. Several options are available which allow to choose among several GMM and GEL methods and many different HAC matrix estimators. In order to estimate the vector of parameters, users can select their preferred optimization algorithm depending on whether inequality constraints are required. For the vector of Lagrange multiplier of GEL, it can be computed by an iterative procedure based on the Newton method which increases the speed of convergence and reduce the instability of the estimation procedure. It could then easily be used by those who are interested in studying the numerical properties of both methods. The package also offers an interface which is comparable to the least squares method \textit{lm}. Linear model are estimated using formula and methods such as \textit{summary}, \textit{vcov}, \textit{coef}, \textit{confint}, \textit{plot}. \textit{residuals} or \textit{fitted} are available for the objects of class \textit{gmm} and \textit{gel}. \proglang{R} users will therefore have little difficulty in using the package. \section{Computational Details} The package \pkg{gmm} is written entirely in \proglang{R} and S3-classes with methods are used. It can be found on the comprehensive \proglang{R} archive network (CRAN,\url{http://CRAN.R-project.org/}). It is also hosted on R-Forge (\url{http://r-forge.r-project.org/projects/gmm}). It is shipped with a NAMESPACE. The version used to produce this paper is 1.4-0. It depends on the \pkg{sandwich} package of \cite{zeileis06}, which is used to compute de HAC matrices. The packages \pkg{car} (\cite{car}), \pkg{mvtnorm} (\cite{mvtnorm}), \pkg{stabledist} (\cite{stabledist}), \pkg{MASS} (\cite{MASS}), \pkg{timeDate} (\cite{timeDate}) and \pkg{timeSeries} (\cite{timeSeries}) are suggested in order to reproduce the examples. \section*{Acknowledgments} I am grateful to the three anonymous referees of the Journal of Statistical Software for great comments on the paper and the package. I also want to thank Achim Zeileis for his suggestions regarding the way the \pkg{sandwich} package can be used within \textit{gmm}. \bibliography{empir} \end{document} gmm/vignettes/empir.bib0000644000176200001440000004743414247643114014656 0ustar liggesusers @article{jagannathan-skoulakis02, AUTHOR={Jagannathan, R. and Skoulakis, G.}, TITLE={Generalized Method of Moments: Applications in Finance}, JOURNAL={Journal of Business and Economic Statistics}, VOLUME={20}, PAGES={470-481}, YEAR={2002}, Number={4} } @article{garcia-renault-veredas06, AUTHOR={Garcia, R. and Renault, E. and Veredas, D.}, TITLE={Estimation of Stable Distribution by Indirect Inference}, JOURNAL={Working Paper: UCL and CORE}, VOLUME={}, PAGES={}, YEAR={2006}, Number={} } @book{nolan09, author = {J. P. Nolan}, title = {Univariate Stable Distributions - Models for Heavy Tailed Data}, year = {2020}, publisher = {Springer Series in Operations Research and Financial Engineering} } @Manual{timeDate, title = {timeDate: Rmetrics - Chronological and Calendarical Objects}, author = {Diethelm Wuertz and Yohan Chalabi with contributions from Martin Maechler and Joe W. Byers and others}, year = {2009}, note = {R package version 2100.86}, url = {http://CRAN.R-project.org/package=timeDate}, } @Manual{timeSeries, title = {timeSeries: Rmetrics - Financial Time Series Objects}, author = {Diethelm Wuertz and Yohan Chalabi}, year = {2009}, note = {R package version 2100.84}, url = {http://CRAN.R-project.org/package=timeSeries}, } @Book{MASS, title = {Modern Applied Statistics with S}, author = {W. N. Venables and B. D. Ripley}, publisher = {Springer}, edition = {Fourth}, address = {New York}, year = {2002}, note = {ISBN 0-387-95457-0}, url = {http://www.stats.ox.ac.uk/pub/MASS4} } @book{hall05, author = {A. R. Hall}, title = {Generalized Method of Moments (Advanced Texts in Econometrics)}, year = {2005}, publisher = {Oxford University Press}, address = {}, } @Book{wooldridge02, author = {Wooldridge, J. M.}, title = {Econometric Analysis of Cross Section and Panel Data}, publisher = {Cambridge, MA: MIT Press}, year = {2002}, } @Book{cochrane01, author = {Cochrane, J. H.}, title = {Asset Pricing}, publisher = {Princeton University Press}, year = {2001}, } @article{chausse09, AUTHOR={Chauss\'e, P.}, TITLE={Computing Generalized Empirical Likelihood and Generalized Method of Moments with R}, JOURNAL={Working Paper, University of Waterloo}, VOLUME={}, PAGES={}, YEAR={2011}, Number={} } @article{smith01, AUTHOR={Smith, R. J.}, TITLE={GEL Criteria for Moment Condition Models}, JOURNAL={Working Paper, University of Bristol}, VOLUME={}, PAGES={}, YEAR={2001}, Number={} } @article{chausse08, AUTHOR={Chauss\'e, P.}, TITLE={Generalized Emprical Likelihood for a Continuum of Moment Conditons}, JOURNAL={Working Paper, University of Waterloo}, VOLUME={}, PAGES={}, YEAR={2011}, Number={} } @Book{campbell-lo-mackinlay96, author = {Campbell, J. Y. and Lo, A. W. and Mackinlay, A. C.}, title = {The Econometrics of Financial Markets}, publisher = {Princeton University Press}, year = {1996}, } @article{newey-west94, AUTHOR={Newey, W. K. and West, K. D.}, TITLE={Automatic Lag Selection in Covariance Matrix Estimation}, JOURNAL={Review of Economic Studies}, VOLUME={61}, PAGES={631-653}, YEAR={1994}, Number={} } @article{andrews-monahan92, AUTHOR={Andrews, W. K. and Monahan, J. C.}, TITLE={An Improved Heteroskedasticity and Autocorrelation Consistent Covariance Matrix Estimator}, JOURNAL={Econometrica}, VOLUME={60}, PAGES={953-966}, YEAR={1992}, Number={4} } @Book{white84, author = {White, H.}, title = {Asymptotic Theory for Econometricians}, publisher = {Academic Press}, year = {1984}, } @Book{gallant87, author = {Gallant, A. R.}, title = {Nonlinear Statistical Models}, publisher = {Wiley}, year = {1987}, } @article{zeileis06, AUTHOR={Zeileis, A.}, TITLE={Object-oriented Computation of Sandwich Estimator}, JOURNAL={Journal of Statistical Software}, VOLUME={16}, PAGES={1-16}, YEAR={2006}, Number={9}, url={http://www.jstatsoft.org/v16/i09/} } @article{chausse10, title = {Computing Generalized Method of Moments and Generalized Empirical Likelihood with R}, author = {Pierre Chauss{\'e}}, journal = {Journal of Statistical Software}, year = {2010}, volume = {34}, number = {11}, pages = {1-35}, url = {http://www.jstatsoft.org/v34/i11/} } @article{zeileis04, AUTHOR={Zeileis, A.}, TITLE={Econometric Computing with HC and HAC Covariance Matrix Estimators}, JOURNAL={Journal of Statistical Software}, VOLUME={11}, PAGES={1-17}, YEAR={2004}, Number={10}, url={http://www.jstatsoft.org/v11/i10/} } @Book{luenberger97, author = {Luenberger, D. G.}, title = {Optimization by Vector Space Methods}, publisher = {Wiley and Sons}, year = {1997}, } @article{smith04, AUTHOR={Smith, R. J.}, TITLE={GEL Criteria for Moment Condition Models}, JOURNAL={CeMMAP working papers, Institute for Fiscal Studies}, VOLUME={}, PAGES={}, YEAR={2004}, Number={} } @article{kitamura-tripathi-ahn04, AUTHOR={Kitamura, Y. and Tripathi, G. and Ahn, H.}, TITLE={Empirical Likelihood-Based Inference in Conditional Moment Restriction Models}, JOURNAL={Econometrica}, VOLUME={72}, PAGES={1667-1714}, YEAR={2004}, Number={} } @article{blaschke-neubauer-scherzer97, AUTHOR={Blaschke, B. and Neubauer, A. and Scherzer, O.}, TITLE={On the Convergence Rates for the Iteratively Regularized Gauss-Newton Method}, JOURNAL={IMA Journal of Numerical Analysis}, VOLUME={17}, PAGES={421-436}, YEAR={1997}, Number={} } @article{carrasco-florens-renault07, AUTHOR={Carrasco, M. and Florens, J. P. and Renault, E.}, TITLE={Linear Inverse Problems in Structural Econometrics Estimation Based on Spectral Decomposition and Regularization}, JOURNAL={Handbook of Econometrics}, VOLUME={6B}, PAGES={5633-5751}, YEAR={2007}, Number={} } @article{donald-imbens-newey03, AUTHOR={Donald, S. and Imbens, G. and Newey, W. K.}, TITLE={Empirical Likelihood Estimation and Consistent Tests with Conditional Moment Restrictions}, JOURNAL={Journal of Econometrics}, VOLUME={117}, PAGES={55-93}, YEAR={2003}, Number={} } @article{carrasco07, AUTHOR={Carrasco, M.}, TITLE={A Regularization Approach to the Many Instruments Problem}, JOURNAL={Forthcoming in the Journal of Econometrics}, VOLUME={}, PAGES={}, YEAR={2009}, Number={} } @article{parzen70, AUTHOR={Parzen, E.}, TITLE={Statistical inference on time series by RKHS methods}, JOURNAL={Canadian Mathematical Congress}, VOLUME={Edited by R. Pyke}, PAGES={1-37}, YEAR={1970}, Number={} } @article{qi-nian00, AUTHOR={Qi-Nian, J.}, TITLE={On the Iterative Regularized Gauss-Newton Method for Solving Nonlinear Ill-posed Problem}, JOURNAL={Mathematics of Computation}, VOLUME={69}, PAGES={1603-1623}, YEAR={2000}, Number={232} } @article{schennach07, AUTHOR={Schennach, S. M.}, TITLE={Point Estimation with Exponentially Tilted Empirical Likelihood}, JOURNAL={Econometrica}, VOLUME={35}, PAGES={634-672}, YEAR={2007}, Number={2} } @article{guggenberger08, AUTHOR={Guggenberger, P.}, TITLE={Finite Sample Evidence Suggesting a Heavy Tail Problem of the Generalized Empirical Likelihood Estimator}, JOURNAL={Econometric Reviews}, VOLUME={26}, PAGES={526-541}, YEAR={2008}, Number={} } @article{guggenberger-hahn05, AUTHOR={Guggenberger, P. and Hahn, J.}, TITLE={Finite Sample Properties of the Two-Step Empirical Likelihood Estimator}, JOURNAL={Econometric Reviews}, VOLUME={24}, PAGES={247-263}, YEAR={2005}, Number={3} } @article{groetsch93, AUTHOR={Groetsch, C.}, TITLE={Inverse Problems in Mathematical Sciences}, JOURNAL={Wiesbaden: Vieweg}, VOLUME={}, PAGES={}, YEAR={1993} } @article{carrasco-florents00, AUTHOR={Carrasco, M. and Florens, J. P.}, TITLE={Generalization of GMM to a Continuum of Moment Conditions}, JOURNAL={Econometric Theory}, VOLUME={16}, PAGES={655-673}, YEAR={2000} } @article{carrasco-florens00, AUTHOR={Carrasco, M. and Florens, J. P.}, TITLE={Generalization of GMM to a Continuum of Moment Conditions}, JOURNAL={Econometric Theory}, VOLUME={16}, PAGES={655-673}, YEAR={2000} } @article{carrasco-florens02, AUTHOR={Carrasco, M. and Florens, J. P.}, TITLE={Efficient GMM Estimation Using the Empirical Characteristic Function}, JOURNAL={Working Paper, Institut d'Économie Industrielle, Toulouse}, VOLUME={}, PAGES={}, YEAR={2002} } @article{anatolyev05, AUTHOR={Anatolyev, S.}, TITLE={GMM, GEL, Serial Correlation, and Asymptotic Bias}, JOURNAL={Econometrica}, VOLUME={73}, PAGES={983-1002}, YEAR={2005} } @article{hansen82, AUTHOR={Hansen, L. P.}, TITLE={Large Sample Properties of Generalized Method of Moments Estimators}, JOURNAL={Econometrica}, VOLUME={50}, PAGES={1029-1054}, YEAR={1982} } @article{hansen-heaton-yaron96, AUTHOR={Hansen, L. P. and Heaton, J. and Yaron, A.}, TITLE={Finit-Sample Properties of Some Alternative GMM Estimators}, JOURNAL={Journal of Business and Economic Statistics}, VOLUME={14}, PAGES={262-280}, YEAR={1996} } @article{pakes-pollard89, AUTHOR={Pakes, A. and Pollard, D.}, TITLE={Simulation and the Asymptotics of Optimization Estimators}, JOURNAL={Econometrica}, VOLUME={57}, PAGES={1027-1057}, YEAR={1989} } @article{imbens02, NUMBER={4}, AUTHOR={Imbens, G. W.}, TITLE={Generalized Method of Moments and Empirical Likelihood}, JOURNAL={Journal of Business and Economic Statistics}, VOLUME={20}, PAGES={493-506}, YEAR={2002} } @article{imbens97, AUTHOR={Imbens, G. W.}, TITLE={One-Step Estimators for Over-Identified Generalized Method of Moments Models}, JOURNAL={Review of Economics Studies}, VOLUME={64}, PAGES={359-383}, YEAR={1997} } @article{smith97, AUTHOR={Smith, R. J.}, TITLE={Alternative Semi-Parametric Likelihood Approaches to Generalized Method of Moments Estimation}, JOURNAL={The Economic Journal}, VOLUME={107}, PAGES={503-519}, YEAR={1997} } @article{newey-smith04, author = {Newey, W. K. and Smith, R. J.}, title = {Higher Order Properties of GMM and Generalized Empirical Likelihood Estimators}, JOURNAL={Econometrica}, VOLUME={72}, PAGES={219-255}, YEAR={2004} } @article{kitamura97, NUMBER={5}, AUTHOR={Kitamura, Y.}, TITLE={Empirical Likelihood Methods With Weakly Dependent Processes}, JOURNAL={The Annals of Statistics}, VOLUME={25}, PAGES={2084-2102}, YEAR={1997} } @article{kitamura-stutzer97, NUMBER={5}, AUTHOR={Kitamura, Y. and Stutzer, M.}, TITLE={An Information-Theoretic Alternative to Generalized Method of Moments Estimation}, JOURNAL={Econometrica}, VOLUME={65}, PAGES={861-874}, YEAR={1997} } @article{owen88, NUMBER={}, AUTHOR={Owen, A. B.}, TITLE={Empirical Likelihood Ratio Confidence Intervals for a Single Functional}, JOURNAL={Biometrika}, VOLUME={75}, PAGES={237-249}, YEAR={1988} } @article{singleton01, NUMBER={102}, AUTHOR={Singleton, K.}, TITLE={Estimation of Affine Pricing Models Using the Empirical Characteristic Function}, JOURNAL={Journal of Econometrics}, VOLUME={}, PAGES={111-141}, YEAR={2001} } @Book{owen01, author = {Owen, A. B.}, publisher = {Chapman and Hall}, title = {Empirical Likelihood}, year = {2001}, } @article{carrasco-chernov-florens-ghysels07, NUMBER={140}, AUTHOR = {Carrasco, M. and Chernov, M. and Florens, J. P. and Ghysels, E.}, TITLE = {Efficient of General Dynamic Models with Continuun of Moment Conditions}, JOURNAL = {Journal of Econometrics}, VOLUME={}, PAGES={529-573}, YEAR={2007} } @Unpublished{Bindelli05, author = {Bindelli, L.}, title = {Testing the New Keynesian Phillips Curve: a Frequency Domain Approach}, note = {Université de Lausanne. Dernière version: 2005}, OPTkey = {}, OPTmonth = {}, OPTyear = {}, YEAR={2005}, OPTannote = {} } @article{chamberlain87, NUMBER={}, AUTHOR={Chamberlain, G.}, TITLE={Asymptotic Efficiency in Estimation with Conditional Moment Restrictions}, JOURNAL={Journal of Econometrics}, VOLUME={34}, PAGES={304-334}, YEAR={1987} } @article{nelson90, NUMBER={}, AUTHOR={Nelson, D.}, TITLE={ARCH Models as Diffusion Approximations}, JOURNAL={Journal of Econometrics}, VOLUME={45}, PAGES={7-38}, YEAR={1990} } @article{hull-white87, NUMBER={}, AUTHOR={Hull, J. and White, A.}, TITLE={The Pricing of Options on Assets with Stochastic Volatilities}, JOURNAL={Journal of Finance}, VOLUME={42}, PAGES={281-300}, YEAR={1987} } @article{nowman97, NUMBER={}, AUTHOR={Nowman, K.}, TITLE={Gaussian estimation of single-factor continuous time models of the term structure of interest rate}, JOURNAL={Journal of Finance}, VOLUME={52}, PAGES={1695-1703}, YEAR={1997} } @article{yu-phillips01, NUMBER={}, AUTHOR={Yu, J. and Phillips, P. C. B.}, TITLE={Gaussian estimation of continuous time models of the short term structure of interest rate}, JOURNAL={Cowles Foundation Discussion Paper}, VOLUME={1309}, PAGES={}, YEAR={2001} } @article{bandi-phillips03, NUMBER={}, AUTHOR={Bandi, F. M. and Phillips, P. C. B.}, TITLE={Fully nonparametric estimation of scalar diffusion models}, JOURNAL={Econometrica}, VOLUME={71}, PAGES={241-283}, YEAR={2003} } @article{qin-lawless94, NUMBER={1}, AUTHOR={Qin, J. and Lawless, J.}, TITLE={Empirical Likelihood and General Estimating Equation}, JOURNAL={The Annals of Statistics}, VOLUME={22}, PAGES={300-325}, YEAR={1994} } @Manual{gmm, title = {gmm: Generalized Method of Moments and Generalized Empirical Likelihood}, author = {Pierre Chausse}, year = {2009}, note = {R package version 1.4-0}, } @Manual{fBasics, title = {fBasics: Rmetrics - Markets and Basic Statistics}, author = {Diethelm Wuertz}, year = {2009}, note = {R package version 2100.78}, url = {http://CRAN.R-project.org/package=fBasics}, } @Manual{stabledist, title = {stabledist: Stable Distribution Functions}, author = {Diethelm Wuertz and Martin Maechler and Rmetrics core team members.}, year = {2012}, note = {R package version 0.6-4}, url = {http://CRAN.R-project.org/package=stabledist}, } @Manual{mvtnorm, title = {mvtnorm: Multivariate Normal and t Distributions}, author = {Alan Genz and Frank Bretz and Tetsuhisa Miwa and Xuefei Mi and Friedrich Leisch and Fabian Scheipl and Torsten Hothorn}, year = {2009}, note = {R package version 0.9-8}, url = {http://CRAN.R-project.org/package=mvtnorm}, } @Article{plm, title = {Panel Data Econometrics in {R}: The {plm} Package}, author = {Yves Croissant and Giovanni Millo}, journal = {Journal of Statistical Software}, year = {2008}, volume = {27}, number = {2}, url = {http://www.jstatsoft.org/v27/i02/}, } @Manual{car, title = {car: Companion to Applied Regression}, author = {John Fox}, year = {2009}, note = {R package version 1.2-16}, url = {http://CRAN.R-project.org/package=car}, } @article{diciccio-hall-romano91, NUMBER={2}, AUTHOR={DiCiccio, T. and Hall, P. and Romano, J.}, TITLE={Empirical Likelihood is Bartlett-Correctable}, JOURNAL={The Annals of Statistics}, VOLUME={19}, PAGES={1053-1061}, YEAR={1991} } @article{ramlau-teschke05, NUMBER={}, AUTHOR={Ramlau, R. and Teschke, G.}, TITLE={Tikhonov Replacement Functionals for Iteratively Solving Nonlinear Operator Equations}, JOURNAL={Inverse Problems}, VOLUME={21}, PAGES={1571-1592}, YEAR={2005} } @article{brown-newey02, AUTHOR={Brown, B. and Newey, W. K.}, TITLE={Generalized Method of Moments, Efficiency Bootsrapping, and Improved Inference}, JOURNAL={Journal of Business and Economic Statistics}, VOLUME={20}, PAGES={507-517}, YEAR={2002} } @article{newey-west87a, AUTHOR={Newey, W. K. and West, K. D.}, TITLE={A Simple, Positive Semi-definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix}, JOURNAL={Econometrica}, VOLUME={55}, PAGES={703-708}, YEAR={1987} } @article{andrews91, AUTHOR={Andrews, D. W. K.}, TITLE={Heteroskedasticity and Autocorrelation Consistent Covariance Matrix Estimation}, JOURNAL={Econometrica}, VOLUME={59}, PAGES={817-858}, YEAR={1991} } @article{chesher-smith97, AUTHOR={Chesher, A. and Smith, R. J.}, TITLE={Likelihood Ratio Specification Tests}, JOURNAL={Econometrica}, VOLUME={65}, PAGES={627-646}, YEAR={1997} } @article{horowitz02, AUTHOR={Horowitz, J.}, TITLE={The Bootstrap in Econometrics}, JOURNAL={Handbook of Econometrics}, VOLUME={Vol. 5 eds. J. Heckman and E. Leamer, Amsterdam: North Holland}, YEAR={2002} } @article{imbens-spady-johnson98, AUTHOR={Imbens, G. W. and Spady, R. H. and Johnson, P.}, TITLE={Information Theoretic Approaches to Inference in Moment Condition Models}, JOURNAL={Econometrica}, VOLUME={66}, PAGES={333-357}, YEAR={1998} } @article{hullwhite87, NUMBER={}, AUTHOR={Hull, J. and White, A.}, TITLE={The Pricing of Options on Assets with Stochastic Volatilities}, JOURNAL={Journal of Finance}, VOLUME={42}, PAGES={281-300}, YEAR={1987} } @article{yu-phillips01, NUMBER={}, AUTHOR={Yu, J. and Phillips, P. C. B.}, TITLE={Gaussian estimation of continuous time models of the short term structure of interest rate}, JOURNAL={Cowles Foundation Discussion Paper}, VOLUME={1309}, PAGES={}, YEAR={2001} } @article{li-he05, NUMBER={4}, AUTHOR={Li, K. and He, Y.}, TITLE={Taylor Expansion Algorithme for the Branching Solution of the Navier-Stokes Equations}, JOURNAL={International Journal Numerical Analysis and Modeling}, VOLUME={2}, PAGES={459-478}, YEAR={2005} } @article{ramm04, NUMBER={N2}, AUTHOR={Ramm, A. G.}, TITLE={Dynamical systems method for solving operator equations}, JOURNAL={Comm. Nonlinear Sci. and Numer. Simul.}, VOLUME={9}, PAGES={383-402}, YEAR={2004a} } @article{airapetyan-ramm00, NUMBER={}, AUTHOR={Airayetpan, R. G. and Ramm, A. G.}, TITLE={Dynamical systems and Discrete Methods for Solving Nonlinear Ill-posed Problem}, JOURNAL={Applied Mathematics Review}, VOLUME={1}, PAGES={491-536}, YEAR={2000} } @article{ramm04-2, NUMBER={}, AUTHOR={Ramm, A. G.}, TITLE={Dynamical systems method for solving nonlinear operator equations}, JOURNAL={Int. J. Appl. Math. Sci.}, VOLUME={1}, PAGES={97-110}, YEAR={2004b} } @article{jin00, NUMBER={232}, AUTHOR={Jin, Q. N.}, TITLE={On the Iteratively Regularized Gauss-Newton Method for solving nonlinear Ill-Posed Problems}, JOURNAL={Mathematics of Computation}, VOLUME={69}, PAGES={1603-1623}, YEAR={2000} } @article{watson93, NUMBER={6}, AUTHOR={Watson, M. W.}, TITLE={Measures of Fit for Calibrated Models}, JOURNAL={The Journal of Political Economy}, VOLUME={101}, PAGES={1011-1041}, YEAR={1993} } @article{diebold-ohanian-berkowitz98, NUMBER={}, AUTHOR={Diebold, F. X. and Ohanian, L. E. and Berkowitz, J.}, TITLE={Dynamic Equilibrium Economies: A Framework for Comparing Models and Data}, JOURNAL={Review of Economic Studies}, VOLUME={65}, PAGES={433-451}, YEAR={1998} } @article{Berkowitz01, NUMBER={}, AUTHOR={Berkowitz, J.}, TITLE={Generalized Spectral Estimation of the Consumption based Pricing Model}, JOURNAL={Journal of Econometrics}, VOLUME={104}, PAGES={269-288}, YEAR={2001} } @article{christiano-vigfusson03, NUMBER={4}, AUTHOR={Christiano, J. L. and Vigfusson, R. J.}, TITLE={Maximum Likelihood in the Frequency Domain: The Importance of Time-to-Plan}, JOURNAL={Journal of Monetary Economics}, VOLUME={50}, PAGES={}, YEAR={2003} } @article{seidman-vogel89, NUMBER={}, AUTHOR={Seidman, T. I. and Vogel, C. R.}, TITLE={Well Posedness and Convergence of some Regularisation Methods for Non-linear Ill-posed Problems}, JOURNAL={Inverse Problems}, VOLUME={5}, PAGES={227-238}, YEAR={1989} } @Book{hamilton94, author = {Hamilton, J. D.}, publisher = {Princeton University Press}, title = {Time Series Analysis}, year = {1994}, } gmm/data/0000755000176200001440000000000014344367335011757 5ustar liggesusersgmm/data/wage.rda0000644000176200001440000004011314247643115013364 0ustar liggesusers‹í} x\W}ï±$K–dI#iöEfF»4²–RÞÜı $…$ds'qg±IÌ ¬}”R ¼”< ”¥}Pºñ€òúàZ–BYJhR$%‹ã$N€$tîü¿sÿstg$%VßCßw}çž{Îßιç^ŸtÜé›zNï1Æ´™ö¶6ÓÖ^ûÙÑVûgé0ÝþÅó.ºÀ˜öáÚïhíè«u=lž¼¿uOìǃs-é|*þÖ‚Ÿ§B—¿ý[ÝßZËqµð«×Æ¿ßÊã×ãoÝ2ÇJú¬äpá¬ÿrt¬Î¥û©†³¸¿ýûÍø[+›ýM=ޤL~êû#å‹O•O?Q¼îøµäc5°ÿ‹‘O5¿ëšü>R0/ÜV>»<·ß‘”ñZÊh¥qêɰ™µÀñxk #AËSígaOVÞ6!׫Áïö5&ö‘®Ö¦+‹µ¢É„œŸèßSeÃk‰w­`¯$G¬µÝ>^ÿ\k|kícÍêûåð‡É¡Ùx·ÿJÆ´â7 ~Xß0žšõYΛ]‡ý5£¥Þ0Z[[ }.­ðº}ÃÆ7»¿+…¿ŸÍ~/gO­hoF+û[nLܰöåèZI{³{+ñ‹°¶ÕønؘVv½œ¿·²›V}ZѺ˜­p­Ö7]¸a¸ZÑØÊ†ZÙ€îß v+:–£­üåxkÅ_3ØËµ‡á}¼>ÛŠ§•Úg+º[É)¬KKmÍp6ûÝÊ7›mEWØ9 ßr2lFÃJ`6ÃvÝJþÍà¶²ßfö°X. ­ÚÃøZÎVWãOÍî7£- ÿJ`‡Ámu½_kv/¬Ïr¿[é×í·yµ¢9{Y®¯>X߯ÔWcCËÙ×ZŒ “ÃZѱZVƒsµ0WÂÃjú¬DnW†«•·yc׊¦#MO38nûr׫½·R¼kÍÿ¯›=Iý‡õi%÷å`.×g-dr¤d~$i3«„¹r9Ò²ÒGš–#E÷r0×jì‘Öç¡óHÒ±Z¸GR†kI+—·•ây2xZŽn³Ìý'ëhU߇ÉͨëÇËûZèö7íÐrtåjœ~aã~“åµRÚ[õk%»•Žk&וü-§×V}~{¬Nÿ®,îÃdÞê¯ÕýVvö; Öjù\ -OT–OæJÿÜþÍä¸R^õ½f0šõ_ îÕÚâJÛŸ ¿X‰žÃxmƘlÜs+kÉïãááÉÖËSiÍðš³û÷xà‡ÁÃÑŠwìjéj6¾¼f8®—ãw50]8­ì»M+á=lÜZ×÷GÂ64ܵì÷DivÇ7ƒ·Z­p­µ4£óÉ<Âp.GÇ‘¤s¥2ëgT»{&çåî7ëëþ^-Íþ¯~šÑºZÚ–“c3¸nÛJû,‡·•¬[ÑÒJwGÒ6×âã5ºj%§f²Z _ºo38GBfa´6k;º])/O+å»ÙØÕò°R:Ý¿Ç+#Æj`?‘±Íú/OÃtû®ÖWƒ9­„Æ0.®0¼¬ï7ÔŽ~kG'Ú6¨¶Z_ÁïÕO_oÀY·õªqÝw¨v ¢m}Âàõ«ëåŽ^ÀäugHŸî¶  }ÏÇßåôëW¿#êqððØàœ7ªkÍkiNWiÎCn–ÏChïÄõ†\ZO}Mð¹°Ù¶Áiëu~s\®»”lIC/Î=!2£ÝèþZž½8úÁG˜L]~ûŸl¨³ £[áÒ4k]ô«{|f©Mopú†éYë‡ôPæ=N?úŸk£ëq ª~­øpñÒö]Ú5~ÆꆲíCÛzÓh«}ŽîÈ—ÆO›à¡õµÑ§å¨ãÛ»Õom³½!°ücÀ¡•c;œÎu¯‚«m Í¿QÁreÝÌçèßÚ:Mx¼kf÷×ëÀï31ø:L#z®ýiÛím—z¤½¸öïêSËÖõ ¶¹qUÇŸfò@›û;ÍRÜnÓ¶°Á‘›Î­òÎzuÝ­úÒ¦HWŸâƒô¶;2Ž„ðç^fG ™»rèQ¼õ𥶦ùf|¡éœLZhëšà§\ «[Ña§Cõ]o–Òïò­ùÒôjšØfT7Þh¹ô›¥±r¹#¬Æpñ»qµ7dìhë¹§í.,Ò¶#ŽœØÞ£î (yñÜguË8 ã-ÏnÎè1A<Ñm®ÍÒú=&ˆSšÇõ€éŸÛqÐ'i'F¡êØÍ1šÒAºÃîñ~—:t<`]¢yÑü»1U·i;tmDë¹±^ߟ];ÔŽËjÇÎÚñÜÚqfíØW;®¨çÖŽ×׎ kÇ+Ñ÷šÚqqíØ[;žƒßÔŽWÔŽ«0æm€q-pìÇøëjÇEµcp¾ðϪ—Ö+pïåÀ±c_Žû瀞½¸~#úï©WÖŽ—ïEÀ»0wá÷ é$ôõá^ x>ͯ½€yðìÆýàóÜß;ÁûÅáNÈðrÀx%Ú_„6ŸçSAÃ% Ý¿ÿbÀ¿Ç% ýbנϹŠßW€ß—áÞ+ÁÏuÀ½g¿ÿéø}`? p(¹ø2=x¯=»M`/þ¸W£}?ôq5ð^:öƒ¶½8NÅ8ïKÀë™ÅñFôº²¼Ðú¾ 2¾}ÏÜ‹g§’ç.À¸<ú}ÏÞ«•Üö£mÆ}»ÐæóîÛÔ¥øý"ÈíjàÚƒóùh;ã.Åõ•€ÿèæègx£­²?õãÃ{x8€¾+˜¾í¾pvAÞ»!£ ¡;¿ß€ùjÀ})äþRÐ}~_º.¾Ó Ó+Ñ~>Æùôì?»@ÿ à{îï„| o'褽Ÿú/ÍW˜À§váž^ ºÈß…Ó§ùÐÙk ×Kq} Æ=g¢íh¢¯Ó¯}úN†^wš î¼ãÏìsqMŸxŒw»Ñg7Æ_9ø6v x¸ô2î1¾^z¯»0æeàWÛ»Ž=/0b¯‚ÎNν€x#M/5ÞÏÇø³!§ý ýptr`ìŽý÷vèêôv­ ìj§ |q7`0–]çïëM o¿Ï‰&ðÚìyû:Ð{"à^ ¹ìƒÞ|šOEŸÝÐíÙ€MÝ=4\=îü30žpöá>ãÊ>Àx=d͸räèËãÕ&ðñÓq\9]~ßç@ûAßYê7sÀððàÛÜ—Ï9 {/à¾ØyÊ—ã+¡“K1Ƨ酠Ûÿ}®ÒÑIë·ùvíû¬ï¯Œ;0öZøðE ÕçëdÀ$Mû1–ñûdô» 2¡ ïì30îdœ}º|;~.drdôjøæ±¦1·¼4ûqéi€Á¼I^N¯ûÐ~äsè8²Ú ~<ðy Žã!sMà+Aö׿ó0Ö·ÕWþe8ö×~ȃvËœMØ Y¿Ì6º¸w˜ n06]ޱWÖ… o'è>¿_P;Þœ[!«ó!‡«ƒºÛ šÏú>k¿Ó€÷r¥ÿÞËMP¿ˆ6æg_&:wú~àÛÚq ‰|^ÎF;k ‹p'Ž=¸w è£ š_„c®}XÇš ¦í}—¨6^_h?dnÙ>NSòd è_Ÿ9ú0NQc|xÛðûLè‰1îEÀì“€ïlŒß ºYß׎©Ú1_;òµ#S;’µcSí(ÕŽYsµc÷po¢vkG®vжˆ¶M€Ç¸ àÇõÿ÷4àŽ†OKxfÑgúO÷çíEÀXĘî^‡?.Š1þý2îû¸žŽ±s8J8gÐ }G0®„ßE\o/SJ~ãê·/‡,`- _2˜ÂýiàÛ ØyŒÉBÖ³ÀU.Ê–²žÅïd3ŽƒøójÌ4à>Íác¸ŠèçËt‘÷EÀ-ïf%_^û8g€cp˜À–²ãëhð}Øiœ)ãàŒšÀö&kýóªß,Ú§k8hS3hOcLQQcóà1ÚæT{0GMà #×d_6 çp.â˜Qã¨sê×—mEébGÖ64®t‘>òX†|+¸¦ÿø0» °GAC ãq.b<ýu4NƒöhI¡ßä’3=N ½Œ¶QÀ›AÛ¤â1 Úüñ›Ñ—2Sô•M`7Óh×÷¦ eu/|ôÇ8xÌ€ò5z&€#ãÈHÓ¬õ_Àyp¦pŸ¶8¦dIšs +‰ßô뼂;vúYc§ ¯¬ ìÿ·I}|Jéd |U+gšS¿³ §#¦dͱG¿´5êŽö>m‚˜CÙSYœ+8§M»˜Sh‡”sÁ±-X#Ðcʱʇ9o_Ëþ¸’#}¬ ¾Ç•™[¨ë‚‚ËØS”Ýd ¯©GÚý>==0éSŒG gAµSÎsŠþ ŒŸ†žÆÐ–Q²Œ‚ö9äHæÉYØÑ” üaÆ9Ì‡ÇØÁcÚ"àgq¦=Ó~˜c˜oÓˆ›þËVê”ø§MP'0_ú<ÆpoBõÍ› ŸæÕ1Là—ô§<®ióÚ׊Рí|Úq1t÷ Û ø‰uG½ÔAxôŸyÐÏXÀ3ëö7½N˜ 7Ò¾C©ê‚vO¿cލ¦®wGœ” ò1ãKÙñžõ½?oòç3[pޡκ͟ëmÃoöñíªï6lßêÀã±]gŸãÔï­¸GXúÞvu‡Âuœ»Ýi#,Íë6…Ÿ‰ó8‡+'Ò¥áiZSth¹éc‡ãX³”ÖãTÛ6uOóºMõ%OÇ90·;¸·8øÉßv\ouîoU÷v(¼;L£¾·+|[­´•- e°Ý9¶™¥2 ³SW„­u«å§mWÓ¿Mõ×ò¡^U}·7§åÀ>Úž\ÝkÐv¨m¿ì³EáÕ~¶Íé¯ýˆ²pùÚf–Ú–‰ë§l×ø´­j¿$-ÛL£¾h Zÿ[ÔòMÜ”«¦_Û€4ÞÓ>¸Å¹·Å,Õ iÙnmHÛÖß6g¼¶kíSÚ´½k/í†4ï0¶§iÚ¢~ßç>å Ûõ±UÝsõáòìò©u¬yría«jÓþ­ã¬ÿ\»qå¥cÒÖñÚÿuüÐþêê~»‚­ýÉEîµÖ‹¶;w¼Æ³-FX~ØbÂ}ctœ¢ mwàj·ªþ®Luìв ßk~´=í¡ó‹æÍå]۾εnWî:Oq¼ëÓZÿÚt¼Ör;4n~ÕñEãÝ‚SÛ¹ãuí°Íéú‡ÎyŒº–rc¡Ž?ÛÌR:uü×|h8[œ³ŽÑ:¸öêú¢Ö™ Oç¯]Þu,Ñø\ÿ§ì¶:×Ûœ>n~Ô1]ã":þî0KíéXgŒŽ•Ô«S—®Ö«ëW[œ~¬ïýù?ñç4CFæ%#s¿=n‚5’(¿O m9üÂÙ¿—Tí<2¸çÏ?Rj Ÿ vÌëY œÓøQx£&X3ña¢}HÑÍ6Ò5¨àgýS€C<‹° ?¦`kÙ¤Ô9ÃJqg<ù£L8Ä9ªð’渂Íõ‡A…7ßIÀD߈’-aP6Ô)ºÖüÇ!¯˜ l(¦èÉ&õ;h›Ä1¤p$| uÖ´’gRI+ÚrÎØ¬sŸt)’fÚ¯)ߤ ì›ôÐfã\ÊsDÉGÛ’Öý0ƦÔoÚJJ#ŸIèÐ¥™ºÑðI?}(e[2Á³ÒA_f[Tõ§6â§|Cði¿ -Ó4Ÿä:4ݰmŒ>E[£ß ÂIý )ÄGýÒ©SÚBRõI*\i%§¤#G-Ú-i‰™¥¶:l}Œã‡Ô8ÚyBõ¥<é§Ô×§(mW”Ç j#ê<ªà©ñÄ3¨à’?Úâ iôÂÊ:xyŸ¼¹>ÈûìKßbp}ŒqBÛmAÃN„Œ‰ª6æ*í› …WÇ:}OÇP^»1‹qjPÁpódÆñÑü=yÆU_ÚV\%­®¾hËÔ'㊶Kmûlvî鸥y¡²Ÿ–;aù±0‡Cç¿hÈXúã5åA]鸧Û)ó¬#G7WoReÝÃkm{Z.Ìk„üMfM`‹Z~ì›U×ìCÛa¬T÷S&ˆƒ æ ƒ'k‚xǸ–U0鬴_ ›ÆX˜tÎnÞÒ>©m±š2ὤ©ë#Æí¨Kû`Ž Ú¦µ3_0Þ躔÷÷´?Åò¤ceªé×pY#Ò.˜[“ªko:ÒÞ4:'ŽŽÍ„ëæÚ1í…~Àúžû¸®_0Á3ÿ¢ žÃŒ™ fñûŒ«{9ÀUme<«7Á³†¢ŸWý8>çà$,Ž1Ás7Þ+™à¹RNÁci%-eðAøyìG(›`/G^¬Q…¯h‚çO|[TãGM#_„UräWPýÉŸÕVâ ]ÄSgòZVí„K8”u[Pyc;ùQã —4Œ™@ÿy¡}œr¤<?tL§,uþ¢<Š,ÊWÛ8eI»ÑuiÖö¦åÃøåÒM›ÖñKç$­Ë’£÷àòž†AºtD1¦ix®_躇t¸0™stü¢\ÝÚaÔÁC}3†1ÎS¾¾¶Õ¼’=ûSÆyÄ ÚiÕñž>BÙº9]ç*½Tó©uG>)³¼’7áSæä—4e¢ã;óœWrâXâΪ6CIò kNÖ÷þ~µˆi|Ù?ü¹Â€ ÞÏî7Á;âþÑcÂß玨sD]ëyÊFÀêÃuÄß/‰˜à]lŽÕ4DÌÒ÷ᇜþC¦‘Ž ¦{ 8cCx ½Ó‹q=fé;õ åÚoùêU0{~{L8<\œ.~Í»þÝ ^êÛ«Æ’VM»¶—A|ó(âÔ·K_¿YJC·ÃO˜Ì]{ˆ„´iPþÍôîÂ8¸‡BhwmB÷gû€i|Þ¢õß ¿ÆÓ£dK˜ú7í£ ÚOxOÓ ý<¢~7“)Çô™¥xµ-ë¶GN=Møwí¶×öíÊUû–¦_ËŽ²re¬c a—î«ýp(N3Ÿkå}&\Ïš-G-_ÒAy8}†BÆQÞa4Ó—#N[¯s_ûC|­Žé}´3mZ®„ÓïàѲáXòíÆž纙M¹vG?Ð×=JÚÞûø:лÑF½ôšÆ|¦ãeÂx»Ñ4ÆgÆQÒ¡yÔ±YÇTÚˆö ×ÈO3;Õ:q}Ük®.Ýþ¼Çœ1KeèÆ׿µ Â|S®=Ó&]{ÒðuÝÁX©uêÚþ=èÀvypiÔq›2tõ12n@õu㉶—?æcÚ¶–»®Óh³®œ#ê^Äi§}6³íS磕gæê²Ï,ÍGîáÊDë®Ç4Öaq¾Û,å¡•¹´>kÀˆ ì©Y,rãÑ¢KÓ«c¼›)£þ{!2²õ½¶sÊ#¬¾×y»Ç6±ýNÚŽÃê{í7Õ‡òf?]÷h=·ªï³{¸õ½;7ˆ8ý(#]ßkŸwë{-³°šÁ­ïÃì¸U=áæð'Rßšåë{S?̧ ÏÍ-aõ½Ë[¤I߈Yc"Múég_ÍpQW}&°AâX®¾g®pùr]ó¸<õ‡´…Õ÷Cfiþ!Þ~Óhÿaõ½¦ÁÓ¸¼i_pÆë>n­¤uìÖ÷ºnlv,Wß3þO· ç§Y}Oþµ]»ùÍ­ïÝÍå³ßÁÑÌWÃô¬i £AÓ»ÚúÞµ¯ˆiôöÕõ½ [i¿Û4Ú¶®Kô†¶Ïˆ §MËo¹ú¾×´¦+̦©‡•Ô÷”5ãeíÖ÷n¼Z®¾§k›wå½\}¯sõjê{m‡ÚN›éÄw”‡Ö¥k3aõ½[gê<ïŽwëò0¿Ð‡®ï9Žõ}ØÚ kW·¾§ž[å æH ›ùÍ-®ÞÜCǃ°ºPÇÏ0yô:íº^jVß»c›Ù ïiYn+ûÐ>1Í××"ÎY×¶n½1KiÔó öÓþèÖó¼Öö°šú>lNõ\}ïæÇfõ½¦w¹úž2jUßÃFÛWo~Ûž}ŸÊÙà¯úƒ¡Ëßõ–o­~ïÄ[ŸµûÅ9¯ÐùÑ{ß÷Ws¼®Þ¿ý“Ñ—ßöœêÒß3»žÿû>þHõÁû?½õƒ_û¢…sÃY“þ¿·žUýI­wmˆ×žý_Ëwzƒ¯ó¼€ý¼±·ß÷ñ_þt¯wc­ÓKÿõ±ê#ïxÉç/yÕ/ª·Èøêwn«#öòŸôÉ{Võ ú?äwô¯½³xóBúÞà?ûßðÚjWÿóïTïú›gžÙu‚—¾ØÏ8Õ‡þØÿ{Ûîµ |{:åѧ²týr÷‹}Î,œ;¯ü“zž}UõÑŸ^öæ-ÿý5®\½u/êHÖXóÒc_ûÂ놮ðò›ër«Þåƒýà׫w¿ééÇÿãGzír^¢Ÿû…~·Ýë©Iñ²7W½kÇçÎ9ºúJÊÑz½(õâk³ÿEÕÇêjí¯ÞXàÅjØj-Õ‡jÜÞÿémÕŸ ýÕ‡„¿êmßÒqO±.`‹¿÷Sù/ý?ýQõŸ¿×æÈ÷ç'?ôÍ›â¿ç­÷ÕúóXz¿ öè¿ñÚk»æ+¾íýC5 3m«wøR¼äÕäÓö¹.Ñcä¼^ðƒÿêõwÖ÷2>×¥ê7®^{÷ÛöØñëëæÆê½ë÷}vùÓêM>Ù—¿ÛÚÿ€{WÝm^Ðá›Å½ï÷º¯­#¬ÞôƒÂ·nßy,ýÓâ½£¦Œ“úoø`ݽåW³¶& /ûH=§î(Õ¾Y°½?ø^ñÛ¡[ê±ðo’8àuùÖÔ‘ò"¿S“âµ÷x}r]}@äáê¡zãí;yÓÓŸgåöSÈðªáW°ïêמ]w(ë7ÃÿÞhíÀ}cƒ_{Å ±ß‡E®TïùŠðëÈÍKKü¨Þ;ûy­Óó¿s¨ú³ºùÎz8Æxæ±}@ì»zcÝN©Þ==*ö_=ü…úoz»EðX¸ëê—ϵף‡ë„VƸì{Ý5_]"×Göþ¼èƒÆ«‚އÅîíøŸÜWwˆê/qÿvèñ†º;«¡ð™"Çù~ØëFÆ^Ú¯o=/ýž—€ÿ ×Ã÷nË߆F¹ÚöÈ?Ô°o¶ôýXüÝk÷Ñ$ßQ=ØèŸ}9ñ€ö=˜¨ãõ†î‘üA>®—8æ ÿ]Ý-œë%>zëÉâ3ýö°»îÆ8hñö¸-€ûȃC°¯ëáaÎÍK6¿‘Οo®ÿ-G}²?óÏM¾û|é½6΢Ÿ¥¯MìÌ‹Á>6¢Ÿƒ×ˉ‡s£ð3Žº~ö]'þ«ÈÑÒ±ŽñSÚi¯ö>ò—ÍCÖî$®Uo…±>yùlp¾H,ÞGgïùÌ ö†QoðÓõ@èmÄu¯#ÊqÈëD=ƒºÆæêé´ÃÒk-×CãÆ Û9QÞ±%ÞPï·C°ãê÷kÁ¥æé^a¶!/úhRß°þÈIü¯þ´î^ ^äF>»|î&3^‘OVï®»ßÕ^òäAð“ïù¨>vaý¯ú É_^¤®¾ ¯½Ñ]{±rê÷Ãö7oòú™Ï™¼®7ÔW½ÕÉW¨Ï,ü_‚ï¨ÏÖƒÿP½Cø³÷áÏÞèó~é¿„ž|œõ]½r»±±¨ÞóÁq#Ÿ9ɶßüÁ†º¤zØw¿Ÿ³×Ù.u†7,vàu×ÃÜþfòòRù£z»Ôö>ìÚ['ýl¼Nägó'í‹vG½öI]dá ŸS¼6ß'~âµýk£ÜYïQ®Á¾Ö¡>ý¡èÓöÿ†ä5k‡ý’7ª7 ÿ^j¿Ôɦ.XkÏmN]z›ÔGÕ/ûeÆñÿäu!ÎB|ú©ø}õ0øtâ¢/ë@’ùyrqðºúñ†‘¸t‹ä;Ï ~P¾<#þ/yÍ+|ÑWäǼð!>{¿?@þ”sõÈç1øùÅ«÷³nôË„cÞäÅ^V/¼X½\·u1óeæwùê½òÃKø¼ öÿ0äËùàP= ì­¿¯>†ó}â§^âsR'"žØüô]ðÑ+~æÔ]wÊÙâý1êlÖ‡ôÌg\»·yzö 9ñ_ΣîÄ|«q½ ùìPÍk%TõW2/¬þu+ê(¯uÞ£N½:ÊKK½ìE†ãÃýˆŸ8ô­ÆxÁ¼\§„Ž{ù´÷G¡ÇÈç!Ôç9賿±î·xW«·ÉüÃK¿¿ž—ô{´q~Øê¾CN=y£Ø ÏÿŽ^¬Ý‹¼ß«_]ôó^É|ŒKÞmkBôT½É×–wÚþ¨i·Vnôs#uЇ¿Ð?(ë^YìÒÂ;?û•Ä[ÇBþ;$p<{N‘™zCÔ'æ!ß¿°ðɼ×+=½î8Õ{êi§µ÷ûáG”לyæúÆz±ú(æ‡QG÷¼µnð^{c=nÇJ¼õ²¨YD¿Qwð ^à>ê¯ðXP '-ë6–w½Ä­i7‡õ`ë-Ö›·@÷n©2¬w«?sø¶ëI¸fI9^^¿ÛñßCœyPì©z§3O¿¾1?îr¶®uÖq–ô3Ž^c˜WÞ.önýèÖÇœ¸Q=»¿³q=©ú ìáWÈ“ågˆ=7©Çª75ÎKmû#RÇT(ñÌÆsæq®—} ôw Îu =6Ôþa;oívæ?ª/›mñ>€8‹ù๵ÃÎ{±ns×Câ7ÕŸ`Þ ×CõÖO6Ì–ðý€Ô·œ7ŠxÛѤ>8„õ»„Kïá¯ú}YGóFA×ïPçY»¥Üxý]©k,¼É·Õ Á<ªqÓÖqÎ|ºúózXxqõng]‰ñ0&ëX¶ëWÔûzÔíôß<æíwK¾°ã”õÕêÃz¬þôuË<Àúí£Î<¶땎_{#2ÿ«Þâè‰ëL?¹Øxº¯­‘_‹‡v|·SÇß ú“yYõ{õiUÆå£úüä ·³±^·v9 zÚ%¾¨üP_·÷J[¥~/ÿ]=.yåAÉ7ÓsõüïÍ"NO_æ]#âßÞÖÓêéöt/ÿ'’Ÿ ïÿˆ îçO ÈË‹Ýïb?ÞægÕ×›¼ÌO§fü Å]ÞÆ?S׋7ã—%…o{‹ÇÖíÛ›xq —ÿG™‡,ôÖç‘^ù¬€õ±Ñ£dݾòÁº<½ù§I>œÄzd>/q+ºe,#óÃÒ¤} óºÒ¿ˆúc™ÿÿFÖŸ^(v:±Kêá‰wÊúâØGêöâå>&ó°2òp¥MúÍôÕã=­Wø­ð5M>Þ,ù½ôv©[FÁÏä;ù»2/ .ŽI_œ’õçIŒÛtQÝ^¼Ê[ê~á1ïŸÄ¼}âÙ²^2<œÿˆ<Ɇœ¾*umñÞú¼Ê+^Z÷W¯Œ~qêõT—š¬‡ÌÐþ&dýkëõ…J}4‡u£‚Ô/^=œ½ÂË_%ñxúßD?“ßÁzÚóçÁž`w“ï•ùåPÄúq ü§°Þ»X©tÞâ;ïÂ[Ä®çß#v”“ùWþ¾\§ž+ö0yV$x¥3Å>¢%YWàøâ ü²¬+{e<¿˜»EÖ… ÿ[æM›Þ'~37€çbï¼…Ôó±7;(J<òJ¯–ykAÖ¼ò"Ç"æaã²îéCI©'½ À;,ú¨œ%òœß-zzÇÛôqÑoé÷Pçc}3¿^âÂâ3e~·é'¼…¿®×GÞô^ßSWJÞÄºÃØ³êäE¤NÌËÂWø¿ÂWø çè6™¿•eÁÞËH<÷*°“™{DXß(¿Rì·2.üd±·¸Wä5Ö#z,ÿƒÔ—Ä©[Ån§zOõD~1vl}Ç+ÊüÛŸä«ø&Yç(m”烋/þ4‹8šú/¢‡âã âÜæùÿßô ±›ôOd}‹ö»ð ÌÛóGK4÷,‰Ë³¸ž~¿ÌsiƒŽ_¯w½ÉõõÄâï{Úô‡"¯M²¾é;@ì´øOS²Þâå!×±Sä9k ñ0{‚Ì·óß’q³.ñjª[ì{þ]A¼(KÞò;ÿ”¬ëLþ)ÄãøÛâoúÍã9káGÂOjñUæ¿^ëLãˆÏ…ÿ#rÎÿ¥Ô“˜ÿÃ7Rè¯ì½oú¬ÄïÙ7 ýã/ó ù®·°Wì,O}#ßM‹ÞЍG¦_ï|EøqóÆQÉëÞ"òZôfeÕKÁo{/¶‹ÝN£^MÃϧƒagq¸ôMÁÓ‰¼¾ ó:oæ÷&þV䜓uUoðæ!ïøœÝZ¨Þâ"—1Äâæ4øL"ÿ$Q€wëYÚ‡äO›‡*ÿ½ü“äƒYôG\’:Øs—нT.þòXoŸ‡œ`wÔqó¿x»œDü. o$YOÀo/»ŸC<®|Aì´2-öXü ä€úìè‰ÇcÈ·å©W6ý™àAý”üoçÒ–8µxÀD{Näiëè âØ8좄õoTøÞ|Nì`uQ!‹ù)ì` õÜ0âë嬧OïŠ%~³é‰ËÓÓ§§F$îO#Þ¤Q'µCä“Cžžiè§ò\9¡¾äÉs*oö ‰ç‹Ñë<ž»Tê>¼ ÔKsxî;~¶ÈoöŸÄŸKÏ”xQ@ܨl…]}Kü¡Œ8VÙ.|'è·'H|,Cþs¨;çËb·…¡>C½;ø?‚¼_Æ<¹"ëBÞQgJþ]<ˆúnQâRå9xž6*ö:z‹ŸF=ŽÝ$r¨¤^ùF‘×ø.ÀžÊÈ/óˆGÁooC>y»øÃêÚè»/ Ž-£®žÁ:bó”‰þ<ê° äÕIÌ'ÊÇâÅ¢ßJRêÇÒ3Ä__c¨Kg±NWõ#/†zgþw¤~§Ÿß\w`oá/Q·Ëº±7…:´‚8]‘ç}^Vž§{#?D^¿KòåøkêvjŸÛMÁï&ouÃì~©Gc$öQA½0ºgv¿À¼ ¿.¢ÞZÄó¡qÔáYøKÉÿžÃ|xóÑb'• %Þ. ï "þÎÀß*“æPMc>œÿ[ðµEô=•‘z$ÿ]É3³ðóü‡e~;{8êÄ~RÈW Ä÷±÷H=4öc‘Û$žÃŒmþgN:ó¨cJˆ{ãß».Â>ó˜Ìâ¹TóŒIă1àË€¯±oÈ/ ¾ŒbŸVõë‘Eè{òsb_Ô1GË:£Å]f©-¬k2.LÆ®ý¸÷ lNóNØm¦Q—ºïzN—«{ÂÑ|„õ3&ܾ][_ßþr2ѶFy´©{ä©-dl›:´P>¬g ïµ›pÐ86(:´n\=…Ùj‡+Ìþi´Mâérú¸|‡ÑKÛ5 ·¦»•Ü\?ÐÇzÓh#ú·ÖygÈØf>f+®ý.Ço˜ 1~º0MÈØvg¼ëSa84íSßwëñFÉßëÊžc]¹˜]iûÑñÛÅÏs»s½^ývùÐøŒs_gß0ßÐcI+ÿ\¿é ¡i3è£óiMnœsaºñRë'L¯”Q‡¢Ñµ‘fzÕ²!:ÆuµfÛFýÓãX§¢7¬¿k§Ú›éÀ˜F~ݧᶇàhSôé{Íü×¥cƒs­ý­Ý4Öl×2£ÍÎXÝáÒÐ,f…ù6ypkfûhºÜzÀ…ïÊ—çîœau}TûÖ§ökWþaºÐ²sÿ´ŸºúvuÎúÞ¿Ž€Ÿ.øb?úELð?iíVçŒÓu};xéB{ûLàßþy£‚³Ñ9»í~?¾CÌïó¬ÇÁw’ù $Êo£ æ]¸æ™øùÆÓ:ÀmÃ=â&ÔýuÀGþXoh]’ʓߌ#~òAÜôï4Aì3&x·ºÝï|w™à½õa%‹^À"žú»~Ê9ûžõ€¹^íR4‘òË·†úÑNÜ„5hÛ€æ§Ó4Öz Oð3–i[é7Á÷™º\û|ü}Š'ê²OÑ4`‚˜Bäcðw(ZyÈ:mÌ ÚùD›Úh{ոܸ4Z‡”NÙ—0i'LP3“¾_O¿íƒþ|¸1Ð: øüÞéç·{ú•~Üæ·Å?” ý”²f, möÆ ’;õСä¨çÊF~÷€~Aù¯WcIoŸ âÏzÕ>ˆ3u»Á6aL£÷:za,бlÈÁ͸˾¾Þ¢Àуkú ¿e¡k^S–Œ[”1cmŠq“öH[àÜöÃo¯p<í’úꞥŸuÓæ´é˜O˜@ß„©¿qÂïÔè|JÒ–©Sú#ퟲ0AL¡-ùý‡´)-?ÆÊŒö©ý†qƒúgÌ$ßmªÎÅÌúDZ41>0v|‡ƒ6ÆXQ2¦.X70¦r~C>;MàS] ¶/f–ÎÝ#ŠfÆLÚã%c ã[·êO]ò;(a¹‘6¢uDü~[» l„þ1AL£Ü¨+·È¿­BÖ5ç÷šÒçQäM§hzmßV¡=±Ÿ».Iý%/æ*Æ_©o]Ké<ÙçàaާLXs°Ñ߆jW‡®áº,Æý-ælÆUú”ö_]#nTýè;=&È Ã¦±æÒ1¡ú lµLè3ü~#}•¼÷ª1´KèCœÛSöÃÀI]êïiûdý©yѾ2ZéߤG׈ú›7Œ“ŒK´--;Æ~^“½æ+ûî¹ÿ=†çGq<·ŠÝ%Ïßxn;Œç^i¼žÀóŠ8žK§ð¼2þ°¯K߃ÃsŤìKµp3Ø—Ä~Š$öoDñœ8ý©Ëåù{ûÆÈéŠcDÏØß¶ã=×Úùœˆû’x_'ç_qìGIbÿPò y.“÷€¼8ö£Ç±? ‰çcq<'c_w ïo&þPöEñ<‰ðcxï°ä1²"Íý¸Àoå=Xý`@ÏóãØ¯D¸ ~‡ç ú#<Ú—µÊ‹ïOb_TB¾Ÿã%°¿8} 1<ç¶ð‰úH^ügò¼<ŽçË|^Ç>-ŽOÁþSØ/ÍqIØáFŸ'pHOû“£Ø·@¹Ç°Ÿ#Ž÷I’x$çmQ<'Oâyœ…ñl']Iì[ŠcŸí1Iz°¿ÁÊÏ[ñžõÄcçô{ŽKÞ ¹a¿í•x“Økýˆö‹ý’ Ú'éÀ¾Œ$žGoϱ ‡ãRØ÷‘¼®lûÅý4öÐŽRØš@\°òc¼Áósê'.¡½aøå0öIÒÞ8žzN‚Þ8ö/kïØçƒ'—¸¿û—l\_1ðO?OcŸé³|@Þ ìûIÈû¨ÆÆìû‰!ŽoôzÙgû¶ìC¦\­?s_8ìrñÒÚ ø´v¿ˆñŒ÷³lÜŒ¿IÈÛæìgI`¯™è÷qìCHa_ŽÕ?Þ+´ñûE(¯´ý`Ÿmû…ãxŠöÇ~ÊÓÊßˉÉ÷0l>J€Ú僟Ƿ§°Ïú£}3o&ñþíÛúí’yü[yÓA?ýˆòeÄû<–?äíöa$ñ‰•?ø‹Ó>Oƒ!·(ö‘îaèÛÆ ì&æÊ߯aÈ5‰}Pö³%ñžB‚ß™¡œàŸ1ì!_¬ (‡$ö7Z¾9qŽûØcð[Úóó áØ8K>áÿ ÄgúõǾYê‰üÐãØ‡Ã}N.òXï=¯(öó±ó‰µ#Ú+öwÚ<À}†ðw¾‡Æ¼OûçýôG=FxVnˆO¬óâw4ê'J¿b;ÞÇb\ˆaÕü‡í¬’¨_–èß“H2þÑ¿h?x¯# ù'`O1¼§@¹Û8Ž8bó àYÿ†§P_±^ ?¤±ýØnëc¼/B»¢=$˜Oåûpö=0›‡ñ]æCÖmIìû¢Ü¬½2nóê$+oÚòYïÓ.;tD±ïÖ­Ï™Ÿ†±/•qdßß¡ÿ,ÁÏø@»Aý•D‰aŸRûÇ’€Oºlœ£Þ௠Ô'6Ž¡gÝÃû$Ä;$ßM²p§¬=Ál^¦ýPOô ìSf}‘@>§=EQ7ZúOkÌô§$ì;êä‡!ÄÉ(ÞcNà½é$¾—aõżŽ~6¯ fœ ¼X×$¨WÔ‹Iê ñ,*߯ðÒW‹S ÛÆOÄ$ãü?†}¡¶>D\²u$剸•Äû\Ü·Êöü›ü¤°œù}ãè¬CY‘æwê…u–+_òcã<¿—„ù“­7ãxÿxm] »dÞ‹B_¬#l}ƒùSóÎïYÇXù#ßÇPw&°Ï2Šk·~¦»N{Jž­k9Ÿ!ˆ¿Œ3|8þU¬G0n9öI=ÙøÃy䛆߲>&=6.`ÄÆMð‘ÁþnŽsóë Æ%®#ÙõÚ%óð[0OÃþÚ$¾Ïl×3¸þÁ8Ⱥ€ë¨‡mä:Î]NÇù.æ«)¼×ÄzžgÎçY2?¥ðž°­ç±ïÝæ5Ø‘?x_Œõfûƒ9IÒX7a¾EøqÔ™Ö.™i—˜§%àgv>‡÷-Yqžhç=ˆ“ööÍõB;Ç:W óhk?¬‡·XŸÓO¹áÊß® Τ°OßÖ\gÜR°_Ž2^ažCúìú ë,Ö€3¼»ƒ_S_¶ÞA~Â{"Œ«Ì×Ãx?ƒõ3ã‡Í#ÔëÈùŠð¬^Qw¦PGÙx…zug õÏœ¿QÞŒ;x¿ØêïÑ3þÙuÄ?»>ˆ¼Jü¬ í:×5Xpþpb#~Úëê•|[ýr^Gy1>ÁŽÜº*ÍúƒïA.vÞ{àz9ùa½Àuææi޳óÄ ÈÉ®Ûs$ :ÓÌÇðŸæçŒOÈ›q?ƒùcõd~™A\ÉÂNrŒ«¨ß²\g`¾C=Âz'Ë®Éàû*#x7ƒylîõ¢‡,ìëº9¼W™‚½’Îô>Êõ2Øgë•iÄŸ,êá4ì‰|§‘ÈWq?C>1¯Í>ÞÃ¥¼² åÅsÞÏáûˆ¤7‰õ5ê)ü‹õ‡üßò üYÚê° à¤ð}_Ö?ÔkþžÃ|$‡yÓô“ø,æƒÄ• ì„ô¦ñýò•<¾šiÒ|•Æ|3ÿJã}ÐQ¬Ogà'Yø%¾3âeaï\¶rý„¼Ç–Á:$í‘ë9òsîLÉgYøYëÐiÚ=ùcþ„\9¯ h)ı,êAê…ôf@Ou ñdñžÜÞïÉb½!y¿ÇÂvÊ—õZï£Ò.xÎÈ÷ý ê‘ Ÿaž›ay'éÈâýºQÄ¥4ýy?‹þ¤‹ñ#|™ÅwFÇåýÝä×4â}q/‹:ŸöCÿ³qþ:‚ï!d!ÿ,ÞOMa>”A}B}Ù8Åõ~êxi·¬ÿìsÐG¹X:é·xnB|OA>´«ŸQ¿h·~Éï¬@.”Sv›F–ãû䌻xO6ƒ¸Îú…gÊöÀù”]ß…½[;d\g݈÷BS\'E]Gygð\ÓÆ>‡"|Æ È+ƒºÏ^c>kí—zÀwZÒ#ólò³þ¦Xý .È¡®à÷f3çJ݃ÿÇÂú]|d)<ßËp~H}"¿o ï¦Q—Øç ˆg¤“v˜DžÅ|Îæ\s~Ï|hã?¾[bíú²ñŸñ y0u÷ æ%<ÿK¢.Œ£>Éa~I{ÌÀ~­~0Ê /2Þ¥å;¤^tY}€ ü7ƒ:vÿ?ÂúUvKyf°>•Á| ‹:Áæ_ð?‚ùZë¶.ÁóÊÑÆYÆmè‰þÅ<œtçWSX¯åwT켉úúvcaãâ$×{Ò´3ú;ê^Žãw“²ô|ç, {!=Yøå†?Øü þRÈ—Yú3ñÂŽÓà‹qƒrK¡~g¼H£n$~¶g`/”cû4²Œ·ð·êQÊ-‹ùVï±Úºëe\÷ás¸ ìÉÖ;˜1ßgŸõ–}¿óÖ êâ,žÓ1?q†úJa^•†²ÎM n±Þ¶vϺþ‡ï6°N ]à —^ù(Ëuè#‹º0ƒuûÔ/9Ö= ò·ùù}q<ù]ur ï÷Z½ žc½lípHWšñ‡úba|dÝÆq€c÷ëÀÞݺ•óÍ4í úM#®$YÇC/üŽÇÑ/“N>·qñ!Çü‡u:~§‹uõ’C~°qŽüðÿ7‚,ߌƒ€G;aÞ´öŠxÃ:*ͺöɸgýñ8‹u_æ-P—ؼJ?'ÿ|ƒ¸ÈyŽÍÌ¿¬ûX¯ÒnPOdPGÚçyÈ÷Ì?¶ÔKëVi®sc=ÙÆ)Ì“Y×$wl|åscÇŽ³œ¿"O&¡ÏÿŸ2Ðkë<±ó&ä?[Ï0îã{œ‡Ùúq‚óš4âŠNÝŸÁø$æ9òI½ÒnXÏ îr~’„år¬[?`v^…ù­Å¸I>“˜±Î±|0_Án3|>zÜÆw^‰3ç[YÐÁ8J<Œc|Äæ)ä'öË nf}–Æw¯r°ãú³.±ûPX¹áùvuˆ§±fžãÿKy ç¤/ú uéͰ@þ¶òÆ|'‹<‘A¼ç:l ü$¹Ÿgÿð]kGÎ|ÓÖqŒËàÏÎcÉã9ã×ß­c]Ïæ ÖU°7ÂËA.#È?¬ÓX7Zù¢Ïâ>é¶q‘uâVë’YÖã¬8ïDH¿Ïr]‰paÿ´“,äÇø3B¿F}㼆óDÖ£Ì+Øeã/âÑ’¼Êy<¾¯•Áº„µÚ?òØúÙxCùCž9®/0ðy®Óx>B?·õ#æ‘Ô¯?°¾Çú¯Ô/â%õÊú‚ñ9Ëz„uµ³ngõɺ’ßá| uBóØ,ä@»‚?°^ÊÈÿ‡Ì¿¡Ëãûcß)뻈û®ýd0O³õ7ì&…8oý—q‰õ ê›oÇ­^°¾ÈœÏržÁõ`Ä;/ev}ç üÏÆ]ØQ ßÃIÁ.ìs6ÀãzSÏ1˜¹Áîw!ݬ ð|öž@=’vâvóú‘]Ÿ‚?0Žd8ŸqžÌx‘Ã~Ž4×íÏYpukòO NÍp?âlûàl}„ï0¥PO'1_ƒÓ!ïjtÈþÿËÏÛ{Á~lꢱý¤NæýÚÏgTØþü“Na{í§mï8ùù'ƒßþo5bëqøÙ¶ã…l<þ´Ó8òŒ­ÇœÄæc¶o%ôÚO cii(éØzú 8®Óÿmïtž²õ„SO"  reï¶=Ïb~ÞiµVGç_vÞ~ ƒ=»Ï{Éyóî«ÉÉéÞ½ïŠó”_ý%éæW¿zÔÉÁÿ>ûO®³­ï|gmm/data/Finance.rda0000644000176200001440000143633414247643115014023 0ustar liggesusersý7zXZi"Þ6!ÏXÌáfýïþ])TW"änRÊŸãXa“ÆqÅjnçj-&é±}nQmaqFQBÚA‘ËRÏJ³çè$âWSÈhiÝ›@ÿ§ãÙ¡à,üËD–\5Æ®±6±°~öí’ø ˜¾pNgõ&îòuó&{Düð·ZÔ[’ÃÂñèâ<·®t^¼$èM à«Da"bï$Y‹ ™s€¯ùJËÛ@â_‘þD=—¼žiú9Ûj:ÂF+èîPãgà× N ™i VxêëD†Q žrTÕ×®›™PçGìEw)zUé1ÂÇÓ¦8¶vpíÒ,Ì©X`‚A†d›¯TÒ8Õ# q« »¶&Ø—gЦsUÒÄ%€“~Pà“Á^üÓü_í£›V\å$ój–”ÞZõ.±gáWŽÆ!®4•eñ“±94`Ÿö㣦ŸÃ%f¥ &uÒÀ äôÊêþêtõ—æjD‹fÚîÆ|Ùÿk9gž­üÝÃJ•';€½ãÌE6ƒ¸únÏí×þ\„l½òÍ!bé,oZ£o2Ö Ò”jªØQxòZ¥ÎX²£¤;3N ƒäŸ &ç[?Ó%}vÆTôëmaÁý‚J³ø“[YP&–G”È!ûtrLþû ó ° _ÌÜ| £êñ\ÓÙ¬Al¿BhÈ/µ#btžsƒ£6=¦(꤭Ù}X%›M;¬ðÛ¿tZ:×Ü·óŸ(òo›?)|Ò(¤»b"ür—›+èžÂ‰O»á"R$œ!I|§Å>Z¦Ê;k¹ý9ÌäS¦úøí)—PLO=~F*·ˆ›çÌJd:‹e9_d´dÇÏ›ÊVHÄIÉýÁÕO? ûͯ YêF5ériX:º³oú1u—Ô-üÏ\ø€š>rÄ3{ʹ-–,ÔÎÖ9œ1’¢;MݨL’GuÂJÿ.Ý÷Ù‘$ó£ñ»ÇF>y¤lºî“Xò"Å¿àh¶ÀPC!ž–IÅùè‚ìP õøðdyÝ æÎŧqo¼+Âα(ƒLiq„Ýi©}/ï\Nï‚$3³m‡jòž=µÏ#\—Áï`"g<™÷% sw~DN04ûœuâØ‡AgÁê lçp¿Ã]@® dØe°Nz¯ Bؽÿ½®GÍ¡[é›tÜGžŸÍ5c}…²tÁ7™äÅH­U¤/û<&ëÝÃþ*»ZtŒíRÍ¿()“(Ôk‘/jœž›"v“žV“”„ð”'ÔPkÙ+… U*ä²zÞûñ³ñA>¥‡`€c3zOì‘D‚–æßËß's?uKvGÖ]„+Ë}¾÷áiêýU¬¬ Á~T΀¨=Ç#G´šyÐòã5aúEøz;ª8‚MSŸ9înÔqÀ°I’.œ:†óXKŸQ)qãqC—?ôŽœkœ :P©—Û3>´ý[k ­•±Z¹z[†+·×êxgþë—›l‹pòàŽ[ˆˆgOMfZÏï÷ªþ¤Èí[/™âV9âïan|DÕµEïMþ+% ýƒóz¶ì*˸µ‡+D- BþÞ”-{Të¼r&iî¾±·/~L>×x_@€B n]÷娧º¹8É\+)ï·Y4kÄd«ùÄ;ßj üíÚÕˆûxÔÿ ¢DæZ‡B¼ën×™'r‚=ÄXE€_*e~㨠-FfÃwÊÎ<‡ã©2›µ–õýäêø8¨¼¹õÒ¤ |äÉS*ù:Aødolzš”°ÒQö:1—LÑ; ’bqî¥Ò»î(›~n<ÔÄ?3Ù½ÿCÀóhÄŽ§hj“äÜ6|è}š½¥6 ¥q‘²7,{ž?¨‘Ôÿeцl+!ºFóã(nq™N)ZǰËïÄ—ºymÝQã㻫Փ³íFújÎûÚô8«i}fö$ŠµÅžf2Ò8Gø04$}þ¾Ö¤Ûƒþx·×7U¯åŸä%Ýél ‘ JÆ™½O´ï‡í:?Uó¯©é«ÖR°¾çOîõ>Wzn™L›ÞD牕‚jê·ÃC„~©ú’Þ ŽÆR^ªèdqسÑI©gŒ9–Ûƒ>p¥À… t`v2Àoyn =}eW),í›l–;0ЗQÿ7âúë«Ô¸cü¸¹e”ƒu¹¤ÆCøõ“TÜî"Gy󌎠îøÈB\q‚7ÍÎb™;Tì)œºDfÏ㑱@«mÃzt»f¸6jG±IÌÖ%©²ßfÀ‡¨œ”νßcÜzìrÆK6|ÜCÌh˜RYSc!ù–Úݺ!Ñ·Ë¥5§PéÀÖ‚B±5#幎ÍO˜>²”˜ëÄÙý“ Å2vE>Qô°~³ƒ›+Ø`ÆÁ·,‰³¼(¹Ìñr‹ó6Á),c7÷ˆ=Þ7Ò‹óéR@B¡ˆ›¿«_ !eT÷­ä»|"dïº-ì¶8í”Lõ­5¾m*ÀS£²ÑÑó~o sÄQj ‚ØÓMgboyÜ”À~!^-ämXAÒo[`^ÊÑD,ÜêÙ®ù°HàS¥šlùìÉU ™çÏ»ªs'ý ßÂ0¥íbµŒÎ÷ø6CÜï'‹óÓÔ"îË%ð¥E“ßYûø_4 iBcÌČɑoŸ‡ÝsWQè9@àæQÜPTD;‹´%¨¦FVmr(819›“~®§™Òä©às÷iÄL1ý›|€½¸}êݹVÆýfìåÅÞïÉÉ@Í#çÚqØ´ÕÆ’Àí0¿MÌ6šuvÁÑm£¾bÈ;ƒ–Íñ_æå°ŠP†©!åôtçÁç\™Ír'XwP8EÚrv3ÁŒ†Ýʹ‘0 µÐ—|E\Ѽ¹@cPpî·÷ÁUèw–8gÿÈ´Ý1`÷žœˆ ØÄõ¯›X¨-Iµt+²ïÙØô.¤wn@.nÿd(uÑùîšhjÈŠWCÝLkzE[戫®L¡Kü«Êwx¦·fG|æàC¹##ôÔpÛˆtÄåc,êd{:Íž3 “™ˆ‰ÖÁÍÑ›U¯ –úµ;ú]Ó”ð/÷)Õó(ü¹sÔ,j\š$\lÒ@»DW-Ù ð(ö~òq¨ÍÓ U‘3Æ÷ÝØž ¶u¸2e\¥05t1Ñ7ùóײtÞÙ¼0°Î‹·ò¸ð‘ @æ‚ÈÏ{Iø•Gž•Âk§*|å‘þ¯@„¨¼4§ßÚôöÙpÕ o¿LªÃ3DK•Zú*Üÿà¬'|¸Ur1éFñó>ð&¢–UP{ø¾{¾2›bU+@âªÔÒr¼/´­ Ph@ûWƒèûuñdÙDzvd$Õs Ž„üÇmïò_7îkI]îð–%NåXH~b(u«vÀafIº¡ê»ß88JÇa+'‹L½öO9Ó‡n†1w¼V‹Ú_ ¯‚q>n·ZæMvZ>Û:}@•€Å|ËC·K¯)ï.â-û•ÆoÀsô¿éÄ¿¢!Ö6ÈL¢iJ4¬ö|_PTA/ì,ó•ßI°"¤*é0e`Õc}¾î‘o<þõQP!:ìPŠ5c¿d¶¬#Á>•{ú° ΆsuDÙª¾£ª¼jÃxñ™=]%ùÐ%¾¸üAûˆ¦û¥©ûµAÀyaÎ*/Ÿ½~JÍh­Oõ·¶N/ê’¡±'|”δJy¨Å­øzŸ+&.Ãê0å(ˆnë¿ç P0VÀ%²An qŒ”w´aõ >3§1²°“ß'½ò×ÙNU$cJ„]l¿Vø[ZÕtªK3ù‚~¯‰ì-›u·ñ)¤ám;§Æú§?bÙ<òR2ìM$ÝdÉŒ$›½çõ 7¤ç› Þû3L{ÊéŒÊòVeö“7,ÿdee¸±Ü4/ !n¤Ú{ÄÈ+(ÞX,ˆ]õA-´&êµ)Ów«!Ì¥¸¯‡P”ÒHb¿F]V“­æÎ1#åˆ?=RÈøÆ»?Ïú¾ð8¿ÈÎ6’:Èýè«÷]æ°5Z[³I\V­ˆ±1h)Ïê \³ Z˜m÷GùÊÔF,qD%¡•¤('y¢ ôK³¼ÌVŒÈ\zZž|S7TÀÐfêÄ™§ ÑQEu=¹:W‡Tq'˜ñw³Xf‚Ìé ´Ÿ©húÒKYÕšÛ|€ž×…Þ_;ÞUÝ¢6ÛìÚª©ícáˆ8,~•³§P( ’úÜõÎ@ŽÉ}©"ÐUp€ß¦Š(t*äµK§&‡•jrqË8cfÿÈ2fسøèŠm!ö¼ìÖD[í¿ˆŸ!“ÄüN„Qõ=ço‹Sõœæƒ¾²Ž~‚L²3Õr»ÝÈÔÑÅqT‡5*ÉõØJ³Áƒ±F;–ÅG(­‚l+%ÏÎc$6 Æœ×ϰrËHñ-¶å±KÚÛ°b‹àÍsÉòjB ã.ýÆ9ëì[á^Ëê @O~Í’±œH•éBùo‰k‰}iog­[IM{9§umµõ¥u³ÇEÍ]éóÜ«ã[ýér€1wô‘ÏX¤L§ð£ ÎÎj,ÖÌ׬¥Š·p`œ,M¿º¡;ï^!díRŸÄÍib „t\^k—¸TYüðá$*AñêÊ<Ð1ÂO XÚä9†Ò”;…‡…¨Ëµb¯ï<ǧ{h¶Ñ/›w. -AÛu%!yÑG#˜#7á+41»™6m:ÏÖ²wuÈfx§Z{ãvx™$j‚6>¯ÙÃj1`þǤDÀ¥;ôˆ&à²Ñ¢~’êß® Îô&òÔ½Afn9ˆ»$B+gl âCrÚã»Î¶Î jUÙ?ƃbÒ3äIÛé°ÏË6ÊÀZùnÒÌV5B®h´*duEJýmäü{ú€ŸÒ5éH$ÄyCjMsEá<—üÌ„Û+œÄÛT†A€{<“v´ ß[žQ.'äe3µy鼨 ¬ƒsÀ#U×ÖªSÆLK…GTZ§—\.´„=±X¹h ÆF0s×e•[2…ê)Ó_ö FfDœß&Œ‘.,^\¶ NäøæÉäo.“.­Å´oQš}¾ÉŸ›o7ªmÊqRß×ÈE•Žhoô#Å–úh;ÃÿS|¦?>¶„NÔ•xFnN슳5ä "ñNݤeï D1-ÉùS@L²÷) Îh=`#`‡€ÒL½n¶¢@¡Ëø3µ·RwoÅ2Û %½ª9aý6ê%–BSJjx˜š'-²N(' ¶ÛS›þ±EÍ#ý!Ì[ çÄ¡ó”éCLÄrø#óéwå oìU`A–¦;¾·mHIJí<ð/†±áR{2½¯ä^]W:qàŸI ÜÕ3ˆ¨rWóG”x”?‡ývXËóþÃØjžñÛ)ÂmFËX7•^…žË)ÙHIîn {9rúœ˜~U½¹c3ƒ’ébzYÈK ÜÊ,}{Nglù˜±>NB,:sæ‡eΘƒø¦ | þï,«2»ˆõ®´òqé~P ŒÓDŸÜ@rzCö[óš6ÿƒº ÔÝø{ÅéWÁÔ^’í)«68úÑVÊ·1"ó¶¿2,Ò´VþXb&¥v{Jw§ŒwÁÄH”˜8ß ÆÑ_Ë6ƒ¢£x(i¨Ï‚K@ÄæÚÿiuŠÁseRoô<¿æó*ácÐ c§ÍUe÷g¢M½zyÞ(»ƒ¬'[‡EÕ•ˆc^îÖ´Ÿ@„´ëaY^f%iõ|}ãc„ tØØ€ú€û*“Ïpssò»l;ÒmK¸XDþýÿç²ÛÓÒlÌäUÓçö3ÎÏÓß'4‰ª‰Ð\.­€`6¯ã›²’úX_pˆ@ˆ-Ãë¾"eM3?uµt±mTËÎÃD‚ñ%„é„Ë ]á}m朙‚t¡Žç„9ÛêÓcänþJEAo| ŽÕb¥+d¦µG¾ÈÔä¹h9Õéðúö¥Ç”…ßàŒ•ˆb¿Ú"ëî?á’›·?ãÛ0 ´óø«±OÚzØ †G[û5êš´‰2u­Žœ}ÙûåjãŒ>éG޲;«ƒÚWù†qãÃÅÎqûñ$è]d³“O…ÂÜGžâ—þ,z[¡£bI Q³±5`ª_³Û_|ðØùôäÊÎÑ•«»/Áy}™õBÇÏ!ä¥ëÿÈPžÎ>@Ô2¹A×Çé@µ¦ïnk-*v¿y™úzF4¥š¦Í¤ûwXoѦ˜ˆŸ…—mg’MÙhœórý™ÍÙt`û˜híÁžpÛÝñ¨JmÔ:)MÒ¹ë©WT—ðE1&AW²­që׈ßÞð(î·Rx½°bú†)¼áYæq?®àþ©Xýå©^¹®d `‹Œ1+TßIÄËg2ð5éL‰Õ%êfÄS¦¶M7À. év´³RFW4aãjãëìB–^ry xW»}¢ÿ]-åÇ™ñS©HªlÃçÙëF&È)È}À¤7£Ä™f , ¹†ëõLû®“­­™á_)BžŽ©×0öcNhÕjæ"–O¦GYl§a¸ÀÂ÷SýËÏj>v£ä•ÜTjÐgvpç¤Ì˜z’Û4„“›¤6¢üe€ž<»ÍŽŽñ Úøî™jš!îÍ }š¦°gIL‚MÕ9E%e¿” oMEJñMHý\/D<‹k¡HÐ2Lûx•àçEÍ€JBÏœ±éÊíÿ¨U/4°€*¬f:rÏ‹z´#FôÁüžfeÐw–¬|Ækº‰«Nù\sjÜÒù¤o¿© >F]Œ–™jý‰v?ŸÁV p´¼èc›²ßLöP{ «ˆg}Ö˜¼"§™ ˆI晘»ªGº ¨ôN“Â%â×åM÷Ÿxe8½¶ËÝljGÜ;ñaQ8T „dˆ—Êñ0† ` †Òö,{£í…Àa×Õ˜½ÚÓ³»Œ'0PŠŸ3OTU–|ý¤¡FFÄÉ,‡B.¬raU?Ç&ü“a²1Eå;¾¦»Gº–ï6Í¥ÖM`ßô»¹fsQóˆgÓ§¤7c§}¾£l@) .-^F¨Ó¥&ˆû}F-ÄHïÜŒœÀg§Ÿ ^¥‰ˆhó³jeü·b)ù»è¿ÇÐbE›諎êÇÐS'ÚšK–¢ß+|ÅÃ#½Kùêd¢ütŠk0&ŠÑ Ž>[xþR–×õšP _O±Ï°¬yA9J_±™¤G±{bÌàjä­œécõø ÷ WdÍ!îKG[„bŸ­2nˆPXx Áb k_S¦Ø8+Øæ|)Ô©J]ôLÖ»©†‡—c ‡:¼…~¤efEhéu~\¨SüÒXÑi©hlèÊ)gÐ4B¶á|i³Y~ú2*R_ì*Øo– Xâò™E3(aX˜)7ùïü1Ùù{±ÅŽDèx:ŒkëC]Òo>h íîßY‚—¡¶šBMÚ9•ÚÌ—SÆ)Õ!xÔÒ®è^UÄ¿!ÀávVñLx)×!1œ~¿$áá».íɬY€0^ïì,»qiÙ~C8µ¿¶fz:ࡎ+I¤·ï!^—!vh72úŽ1bò´d•9(¶¾Šö?áåßv‰d0a¨¹ã„ÏÌ/æ”-Sìó¡9rÚº@/Y’ʈ‚IJ@1çŒiBShw¨Ð\ *Ù5¹íiX@$7³dŽÊáÿ¥$•œ_ñ*ûèØ&,ÜOPGr{ÔA„ìUO´·Z­­_É<3¾dWœ–gÉ«‘FÚ¤‡y/ÙZ<—MÑs°?ù?éx!R&.¾‡Uäº7·GËkÚ…rA˜Þ~¢Zó"Ü„fþF–EõÜ(ÿêN6àÕlLxcµ,-âïÅs…»!üZ'cühÊ-âÁÉnèÖ4á¶ÜWæ5JO¹Ø§ñô¦€—Õ`0Ü<š1ºÌ&›…-4“SûËØv@J¤6ÙRc—éLÓ-3‹üo4·‚"ëVñ—ÜB‘ØäŠ &»à.•[£,xÕr‰©U=RÕ1cI\S ‚IQJý8ÜlÅJ)PF·Çŵ!§Æ™.¯vƒÆjðZÕ<ËXÒUQY+ÇŠU¾ÅÅÙ^‚î-!ýÇ“ëÞý(œ¹™Ô„¿e Fp{å«> ƒìíÜŤsŸøÜ^á¤~¾ŠÐº)—¶#4—`ßñCÁ««X3ß¹B-N¯0^šÆÐ«Ø!Ëzä§' jOJ@HouÇÎ(žMSd¸¥Îº%[¾$+7.nF û–åÒvò åhH¡s ídÈÞºÎuÿñF*£ÉzÙdŠ“„£ì;ÍÔu{)o>¯Qåò4 ˜ÅáïÒA¹ m¯:B XËQÍao¬Û»$ªL²È« ÎÄî”qÏË\ Ïn˜­ÇÁ…&ɉ/HÇXªDJ¼öz…â܉¹miÌx4À̘îV6ãEÍ›CÁ²V­¾×Ö‚€Ý:Üb?‘!8VÖŒNBÆË ù©/yà€ å?²3U?«sÏýkš¦»Ì’æ1‘˜¯cföV¨š0 óV ¯4ñÌòg>¯üýRÔ‚ÜÚ9ÃŽégniQ…¨mNá§O( ¨;*/<]Ê-¿¶Œ·ßòãÞœƒÊcñ:³R² ­žt©‡ñˆŽ ²‚ð2°z,ê †áÚ¶ÕV³XK1ô‰qŸnäm//$50³;ölÛÖÇ* àÁ¬³º0?ê®éfËR°@zLi0 -5Ãj·ŒFw2K3ëá0 ƒ¢_âÖmµ8ßS$U`Þ¥Û´\F¯Çl¦)Ë‘(-¦ÖµÈ¥t}Rb°›!ÅZÔ¬å‡øÞiW6ˆï«…*nÅäøe=”Å;÷[0w§’ð(—8õúoÆášœíÄî´Õz^òO§ò=@´ã§ëÞÄ­EÛ-ð>JÏ:JMÃôÂ|Ö:„$²F ÅÊÅö> ‰ÿæd"NLtɇ,ûF ²W Ƚfö¤h?É«™F Ö¦Ú)˜(àzÖ÷L8¤®ì'^šæ0»¨ë‡üë’8fÕð^xO¶ÚÎú¢ÛVD²£ƒân…ƒÓ±"ìç–$‡-Ûl÷ÆßpÏ÷ݤ –{J¯€W¡š].VgÖ{`ÀbÇ-‚?ð/5Êö²ÄA§~ÇÔØ¼ùH§ËYðü—F{‹÷*4©z㩼“y–=¿xøc±lCºœ@ ÒaºNGlaÝ>•¤—ÍryK;‡N˜ ÐѽdzæQûøæ]°ÐÁø~ ªe°xâ½kgøßïxt©`[\Ù|ê{¯©>Í*óôµ¹˜ÅJ½ûâÜÃÄÍ?Bœ·¥·+yŒæéZ°óBÛ ‡°žï ?]-Äï)pC@È §ñã°Ó@µ¥w³¦¿n;ÔSßév\ÊÁ¨ÔTCžãÞ0®×R®·‰¶°bHmæ`éX¿ýåq#ê(–H®sðh±¢€.RÓ«#íi@#lø¤€Ôò·)I„>”s¢‚ÝrK'e•)±&d"@É&ÍUÂg¡e2ò%/?”%Ø¡gC½º2æj}óè ZäG|ƒ-„À2_vë`%®¬pMt,›}q_[æÿ5CéVù'›uø#2Ÿ,z…þì€)ÕM¹™.W¨GT%Q—×B ½tyP!¢íå Ò )!XF:Z—wî«hé 7²õ üç3ÐHõ‹Æ7*F:À Ü],0©!àÆ2°Qúc‘ôA˜¥¼—»©5n QoÂG]–»%r9¥Š~£qS-¢½ `ÛXIžÁÌ΄Òo§ÎÙZ@1»‚ý<;ʯý¨ÛæNìqgX$VËtž„·ì'“ ¤D Ç"Œ[7Nkß°Zó˜z›D~ŽIQ£¯øQCËÀ®ÿ`5œÅ–õ04õŠ2Ѱ”Ô™·žw¿Á£ziD&-Ž{¾QK†5޶ӦS ~¢ ªd²l ¼’•—•¢Ð#²&ÍÒüЍ¶ñ ÄnEz°¼°µ+Û·;TI(¡5¢â™s0¸¡é[IùÛØ¶Ed$k§Q편k<ž9}íÖðÏ.ý¶ªxÜO„¾$Ï”ƒÇܾ„–¸â¿ -û°5(¿™l4¦Ÿ¯ø¿\ ˜kè®V3ŸÔK" Äc)Ð}O§¾8ØÚÅ}*[—S•tüÆèÈ¿xoÞ¨›½o©¯]Á~Ñëþd¶½ã?xTÐfWÒSµ±© ].Ua>wT7<ĹÃÞÏQ˜Ð4EÿÒºê1ÖéÁºLTh—ãJbJ[ 02v?ú€@-ßÈñÓˆ(‘R¿FÙ­žǦ^Áÿ”0á$§|€{ ê7áF{óR¡/Ãì´I¥§N‚¬M}_܉@“I„ãë_ßíd¼)MYFÿƒ{p>ß ÁB;£w6ßW„uÍ\ær9WküŠ\•5[X$Zäé-Á°ùÒ}­æ„=Ô¨q{ô9â¶tÙž#ëÉ÷˜Ë÷m:¼eËæ+GÁ@Fs5ʘ;œI™cµ὎º½šÎÕë^œ¬‰oYN¯xüFŒ³hþÊŒézôé= ËáÙ±UKӆЂ,ß»S«J ªZŒ°Z˜W£ôÒžŸSÝÆ« ¨yiˆQg=Ëg@Ýç°Þ ‡ŒÑîÝ D¼TéùØáh5¬¦íÚ0fEôiqç'޳;½u÷V黲f¾Ñj96h$HÔ«oZÆÁHøGgR&XÞI×›ƒ‡Ó z"Ö±¦<è-fåKd¿ Å¥žÃ©üý…QE)p%ÁcC™èìí‚ÉXu…´ó,ePËaPõÕÐ,c˜´Å…ÓЖX {‚œx÷5<ÐÕ/ƒÓm=éu©ÂLÖ08&V··${‚Öi’]Âê©C¸¿”å" | Iw¨¶«¼AÌ^ æðÐc6øÖℽ`),o†`ØÎüZœ¤Ðµ—e§ ®í•÷­lŒšgk謺Æ…’Oc½ñ]  Àz }é–‘¬[Êã4úºVÑiixõ3!¸Ì[|yë*[âÁö £”ï\”LšFheò¹ÑEXµ°©ô#„Ë®k¬”ë;"­Ý¯YZZ–r£Û8ÑRÓ2\œ¯áP¯x¼³q\!·+­‰? ‚‚w»t‰$›ÊëñÌ1²m|,‚ƒ?õ£·p¡ÜêÏ ÿ¶‹ÅrsN uÌ䌈*Cér„{ȱ-4ÖÐVËbÇî «Õê%¡ÿuPƒ²Ñ_Rr«Ú/„GVæ¤iå²6€yk0Ñ©c;P÷1t2UCo¤ê[ªybJ[Š7z®’u^Î`.œ5«÷SØñ†^täüdK è]ºÙcçOÿÀt²Æ…L>ŒódGB4ýSƒ=…N=­¹²žì2ÃÀÛ|¬Í‰ÙÂðïUÅÖ„‡ì¤f)çSŽ©â >á²ü÷ßI× Õ ájýŠ‘¤všfÀU>NRs™zfŒmó{c¼LbUDàÈ­ÕòlÑ!×_æVœGààÙW½î•ˆ¥«ÒbyปˆaWS_-.=Ä¥nòÓÇœáò€}Ý·û‚æfTTŒ} _QÀ}Eœp)÷˜·æÜxú*ÞÉTðWöÕ©ý!À4x5 Œµ÷ŒItzr¶U†ucÒÆSSû[žg:ásôX´9]ù ¡îçÉ™$y™|Tîšô莜ä•öwÞ—ßÿö…@,@fÆ„´ˆÿ®º—{vâ.Õº^¨ŒÄ/úd{2oAF·t—JðÃV¡æ=R¾ˆR’´Iñû~7¯|ƒº£2`Š­+›xãÊ—ÚÎ:æþ]@2 í¤./|;Iž¼ -$¬ô:ˆ°µ»Æ~nÍQŒFÿœþ‹3ÿâ¬XœýâKjI8‡&½CŒâ¦sðØÖ¦šdÿ$«èÌIû*”1<,³'¹ñ8pia›õ1Kþ_™„ªøFª^g«r)f­­AϽø¤×ŽVç0QÇV“Çÿü AÅ´Jµ±3 C7¿iÜk?ˆØsPÔšÂQ{T伓è‘~Û©M˜´Õ3Þj·i‡ø@{£B°•´ z‡@³ßãó÷àÅj é½ ˆ?£ÅjÞ swƒp ÁÿSÎÁw É%Y3É-Vm+CRXÌTªâ’üu_oÛ7I^Fb½‡Á-cäÍ †åŸpV¬j6Þƒ`Rô±œð%ûÉÀñb=1cÁÛ¼¢'Ü ¼ý˜ £4ºÁVê4\.`C‘×aƒvi Ž_—›ãô[ cŸâÉ%ïc‚Ð`ßuu4k)2Äþ4ûºï‡ç¥UQ,„6óf K}uò`œke‰^=ÊÅé–žLaˆ1â4YôžëAÕIëY‘䓤¬i1§'2usî”ÜÒ«ór%Ätjzå±êÜN\ ffôd§—NR`n«•×É:2vX>Ê]Cž6ª]FíÂ׿/©š»sĺÄB 3Ðd…ɽ6wXXÓ”O›Ãb3OSÒ"IÞëãä×èú\ïpIëº>×X“·i¯P‹I­ƒjùÇ9¢P×äæ‹ÜÛEGyCÆ*¼t¬Øÿ½bÜrËI…¤Ê %–›Ëú„v­¸Ì×x”²(ÿÏÄ^ºƒ‘üôôŽaß"Ÿ¡T¤Cïä áUdRÀŸ—½²Ú36/ mtõë;#0pÚë·{¸ë˜z.%‹Ö¬¾Ù×Ù/93´X³Yr‡¼ðJ)`v²Í¥•<ü4ümØ”~¦m>'“‰fq=±„éµÖ|¢º·]¡”ܽTĺ‰n‰±“¿ôòârÊê‚‚¦‘<_=¬˜ÅÇ Ï@‚Sâñï­‚üyÞ–â„]åw¼­4o?»ïomÞ,ÀOÂå+u91S1-¡€Ý™§ã§ñ`2óêóX¶±vºž¹´ ÚÀ›®®<Ÿ~,Ha67DáRþ Y¯Ïü³.»û=›™´ò'Ë¿ª=‘Ǧ„°jÝ®L>„ë{XnŸØèÍV7èðçõTÌlšŽ[}™$P­Ç×Ñ•þQ!Ýû½xØ ¢˜@äýÀÎv +6sWÖ†f­Ú-&{¦­ QX¸Ü¶ öUйŠ."“@Îÿ¸"\.9‘¶`4)w¡ÝmdB¥bö”4ß DˆšèrzZ0Ĉ«L#òÆôà–Që§ü‰sÿ™hwr‚ã¨û>šHlÀžò›´Mø_mßGg¡qlЩç,BÁé?ÈŠ Œ;L§çßÓShS†¥ƒ‰›RŒÍ¼‹LXþÖPôrÏéÌÜQÔf´iÇÍtmòRkuj×TƒBÕUÁb•5º¼‹“j6¶3´ù%‹ Ö§ß„ÉêÚõš¾ó…ìæÌ«y^ðÄT…uFÉÅ 9Å8z…ÿÇo=Yð0†]8ß9Nž#´ë)š—¤¢Ú˜¤Íñwš»÷ÆÕuª4ŸÖ‹ž&…›×;œ‰¶v´@›_¥‰ëL?öª¡¬ùP­Ï(*(*}÷Èk|aÑö²yó‡-š÷äÄ `âmÕ`_ãT xe¶®©Ò/®vn½¼"þõ)Ñ!gfDRÓÅYE†bÑ92äË«éO²Úy2¹Ù#rSjÚè 9[Ö~ˆõè˜Ðuœ„a¹ª€7Ä7c¬Éuãßfàú461T«£ï©–ÉùÿÂéÞŠlgØ+¾‘öî<±O»ÛëjÉÜrµŒC¨èNhÉ|¯u>H½3шÌ6Üä˜5}¶ãI;*ÒK‘¶È«ß–]TÍ …t‡ÜpmÁ—YWºk¿¼ ÅJH¸ ÓWH]—™f>å$'›«kõ¾)T|2 åe}^ý qZÐa*Ž{>¢™é‡t•ïÃívò'ÿçïM±qÔºËFÅF=;Œß·» ó³W=„î„nFå3(,‚+ m`bî)ÂêŒÇš7¿ìµt¯ô>ÍÚ¾gá¡â¿V *°U!Ÿ±ðÖyF‹LýÔDÈs§ì<ÙÂ*t¶-šq…öéP÷÷ 9 Œè+,´+P}m ¼Mز-}¸thYÓœbyº1õY;Ø™ààg{âHu¨˜~x’ãñŒÕ D‹u³6W)ÜEN:#\k%ÆK²ÓÒó5§ê;F7%ßñ‹2Š [Í_f77ŒùáÛ· úžôÂ4Z¬qàõ¡¯(K«]¸s­ù\ ÂFíG!‹»Ü«¯—P®]â ºÃb¡·¶úÊð—^è)Œý!55|‹R;PTóMt:¾ cò ¬•Ý…r*«%†w^¤ÕËì;Ü~[.8Â5¶\ ÅÝXeʦ6ù®1ÿ¸1û½V¶„ZC|‘6¨SÚë áÉI¦ËêÒCúfÛ_ þI£ k?Á/¬ ßuš\qw+À¢Oæu—Nþ¨q”1ýôŒê¤ë †-wZæ.‘×U¯VÂü£[éëQHw¼1A–ê·5'l(Ì»:‘@šMèŠyób§wNtÉÊæYFݨt(×캱.NÖî_—,öPNÁxéÿÈ=Á—c’Bo)q¶Œ1’·èáÿZ®þ3šMh:jrŒ*¨3xXtc·¡_óÑSÂ4:ZRS´¢ì¾÷y%×ç û–¾ì%R8 {ÿrÂN0eÕ‡,Ô3n›Så©3®þ4•qM¼$-§]J;—q7ÚIL½ä™JUQ§$Za?wJøEªbÁR¹8ÑMË-yÐ1Òêà韯,gŠêDߩ췆_6¬Ð0u\¸Šó '…úã°sÕÞ©ËÉWß#¡’<}¸ï^+kÔ‚.—e ÓÖa½¡›ìv׊/1ÑEçV2/ Zf J¾}0ÙkuÊÊðØäy8»È£(|ºÂŠÉSì!ÛÅ~Ý4·‚Å]¥v~MÕ¨$‹ýï©«£R¼¬!Q8]w?Ý~w—Z8‚ºr«÷^§‹Uß«!yWÛ§kAÈç¼i’¥åcÇ6v¨0ç€ïj›ªæÆq{ôŒ2»¡Ð%·FØÎv±€†ÓH°_½èɮ֯XNÖÌÒb¼×ªFý0{‹­ˆ¡²¿ Ž+™ÙÈ6QÅp!9õÉrG¡™ºðº…±â­žw€ñùi{Š#Æ!o³ºåT*RDØ5úR^vͰsV0Š,–ªá‚žºŸ‚ͼ÷”ªH׸Çsñw7‹îª«­Æê,¼ÔÄYžY}vêTzøµÕæÉ†Ú{dÿ…û)4Ä;:xWòÅ“ÞJatŒ-ëY:¸³‡,ŠÊY'N0÷îšëZÅ0±ï»¶†ã¸?>ö"µt¥ʺöؽB™ŒÍh°åFÑ|/šhŒ†ÃÐ\Îå­Ün½6&)q8ºŠÕõšøíHU5J’tðér”m¿LðÀHÀp"g(ö0ÕRô Qع~ݬÿáÙ{n;[FòÁ4ÃyO¡AL‰ð/©±>_¦ø~FÁïÎMw~wFC "&{I°¤Ž}@uÊrÎ^AB¦Ñœ‰_¶ôœõ$­÷㚀q.o;„t•*÷E £¹¾Íâþ\?×L–C ÈQþÆÝŸòær8— (Ããyôš)X-€#öÏNKSÊö˜<Ñðãh¸gÒ„=9³”±Ç¡àÔh”ÕÀÉUÄM?EŠTâ ŸQ2¾J!ŠåÙ•¢w)!ÃTJ}æH¶ÓÑÖÀ~½‰{âùÞ†¤v¿âh¶;Ч O”ÚEg®ªO(uᅧLWAœBß!Sâ–Bq-ÎÓ‘ÈIÒvwÄ©\áíl«ÍRyA æ¹w“ñz¢g±­ÙiiZð»º÷œo± É“כïŸé;ØÝfŸ…GTû¹çúä_P%1]òÁ,T£üR”mwrŒŠX@v¦Ðºá¹*s³Ô(,t²ÖXÔ6 GþEVãzSYØ6š-K(Þ†ÀýsF#~¥]L°¹Í ψ|¬;y±•â&ÞÊæCú3 º2³‘¡´%o"ûZ¦3“ d––ñÓ&Q¿$¯$o.ÂÝÆ0$û{/Dn!·Z‹bŒ®\xW°üXŽ×ºì”‚ýQZßô¬Ó·»/ç‡Ý gÁÿmÿÈšÞК_y-ÔÒUE›b…?q»ë[y¡¥éëo@SuÄë Ý•õ¼T:‹Û Ñà¡‚â§ë‰îXê;»²ÖÜÐÖÜ[“Ë jâ#éyØ©Öeãv2Öùn4"2ÆÈ„Tb¨G¶\f• G0´Ó ŒTÖƒTXR!؈BÆzõƬ‡Ë 6ŒüÿŠ[Èu_rÉpµûn',܇¼!mW.­–ñ4É…Ñ9˜æéó×þ#Ðg‡ÜÌ‚øˆ}ûÔ÷ ívÓ0Ò*ô†–±†Ú!˜ ª£ÏdŽ€hï•®&™^ Ê ¿7sŠà!„ǃXcÀhÖ¿Ž´<‘^:4øŽ[öG^}ÕX¬òðÉW0}a þvºÈl‰%¯4%'ÿ}F¤Þáød 5†é{býºÇ}ŒÜÚW¿íçå sCþÞäÙ>n9!Yž ê,ž:åk9þÓ”ä½3èô2?)«ñ5¹ óÄZcãïÌÄcŸ€ZžGK8„ÛÑ_ˆ'ƒvdDN$¤ûE>ãžàÓDoxÍuz˜5™z¸ÈqžÿC‡HFq#X>ákï£øÔz×,·Nº«ï*®Püõ×ä¢ís冲9<ö³¶âõx£ß ö›Þb 2½â]:à±Û×YOsÚ^^,¡G§Ê*Õ!Öß•ž0ÞAKl¿½>{ÇÃÕ]Ê/på¤ÒÎeí;Ië••o†‡6@Üìóôg aÃqK3¢_tÜÔ¦ÜÍ~;ûÿ5ßÈ´0¥Çfí¤‚È~ênZ¼ÐÈS¬—®æwWôð6œØkõ;€GS;¢³®EŦ¤ãµœÑ‰^N¬<€œX bîM.ZÑYHÆt‚ÆT ÷U9ä²~ tîF‹n+[é¸Åßj¨V,2ÏÓÀŸJ ß UÚØÕ„¥µ5ï©þÊDñÏêîÞ(ÃZŒkt˜Â{Êh;+Ï¿îî/ö¡õ'â(åp†@·ô~Îüã̳!(”šéo‹›Vvǩ߃´_^^nµ.`]‚JÈ`'h+ê `à6“ “qø4špG«Èìç]bŒ¢ƒ Þ_¬, CýYÝÂÈ‹¸x>Ó¾#0lLè ]Ë ¨<æ·êˆ¢¸”\F¨‹€” æw¨–•æù ÖX¢þñ¹¾A‡0—_“ÈtKzæ'”SHZç@®ÃÕ1÷KîÄ›Š2B Ö¡S´6K¹°ÀdV0rttg}©ª^ Ç*ØÅ$•`Øõuä^eÎÔ'3_¸â˜–ͧ,s}¡ Pð„FŒ¸áj [ÉK!l—Û°¬”»ôò•Ë× ¨«NCŽÉVä-³ô²©áöeì+;]¸ÛZ@,ô`¸çá}6Þ ¼øp®wÅ b¸,Ô*n×*€Îž<1íÑ¥žWÿ@»†ëŒgA ¿©p#à @>Àæ™])L4z†im°qðt>¹º¸`Õ ßâ=tôë:v.mõçîÚÛTÆ¥1€YnºEØý{?iýCÿj„ȧðÌQlH’B(„¸ÖÑç-“F úòWC;qÄà0¶ßÜ—ysQÕŸ®Ì¶É«\Ç"„™™‡0½úþ*N(,$r{ëP Umì;Wy·ŒÓp³¢-¯")Aseò2)Bü„ÅŠòEÆtÊ(wJ2Ö«q¥ëŽS4`“¥$’­e¡°ÎP“ÂßúlG¦-\;–3 Óƒµ£å`©‘âµÝ…AÝO.ó߯%HQaÉGÂ1dÊq»áÂwCËÀ3¹þK'‘in<Û­Ç[̋DZQ “ óŒÁi‘;C‚bS7rõ"_”[p¯]ÍZð`y¿¯‰·CÀÇ©¥*UJy&U¾è·oH]ÕaöZVG#7Î;@£Áû°ÔȼL“÷rÞ~“aiB½7Y¸GLF!%ž ôt’Ü;â“h­äª XýnŽîDPæþñ |A›n!M4tdÄ2ˉ…1§—i~5†m?¸˜2öŠ»šB¹ô­óiTÕ€"I÷N/í<§†pKiìB±Å6ùlAÀáeü\%7ëbµÓìí-‡fyŸ3Ézó­œ ,!ª0ÂY{ñ^Âl0§³½ÚÇd|:³ 9¡wUãƒñ£WÑ™¸Îrx€“â.ˆFy<—©Z!žA€š°Ÿ]m”wÍ×(¡¹PËå>&ãá») z‘’Κ -¯#å²"ÍÅaoF!¢w DŽªÙ²™¶v ëf¨a˃CêçüÉîxöÆé¦¹TéÑvqhÑ?MLŒ­_ïem»ûöMøu4UëY“_o£ S¿0Ø’’Ëéì­D¿_îzãv#¥*3¯A˜…mS+Á‘ÍýŠìºwš/*Ôç]kÌÕ ¥/¸ñeZÌaÓ‹ Fñ'7¥ÚÌ}n“b«’“7^šv½ö–·ÆÇ… ù¤8l¢#øõ¸Iœ”`NŒ>6"î¶DäX¬XB›3žÇé‚eªðtcÔš´¶x­(LÓ˜¥À=Y¹6t»~+‹ÓzÍ…ÔÅam ݸYdÆ_Ö4ø¾¼&›‰Á„>íc H³YLj¥ Â S½‘xp6îÕÍã£Z|º}L‰ÿ™lÆ^öcVD!“`Ÿ ÆFÄ­[õHܨgêVÖ@y]'¼ÓfKËÖ‘Ͻ¶êsÆßa°Ñ~ǼÓ ãÉsºeOÑY1ú¡Y¿úû^-ÍPØN/¬ã½Ïe/{85Geñtl¶T3`>BÊž‘Ü^.L綉¢u©C (aŽÐ•ˆø³í"€­¸¸"í7Žü‡¡oÁ§•ïbôüôx Mðòc)ý'òi'´FSQ«aSHÊiŽ`ýÈáÇQú<)fæJu¯¼šÖÍiF;±Œ90çںƈâ”#©ne;-*h û`L:1ÛRô2ó`4ãýëÔÃ4멉É,¢.Qð­‹/€~½'eQ°U´AøùdPÔb(rÁF26sO=Š4Qh\É×$0ý cÉ.Èè•k>ØñÛe °‘¥HfFß5ÂÍEîÞnD½Ä¼ÄÝ6 ÅÙ¡’˜Ó:ñé87Ž@¨`„W»ÍÉŸ‡} ­gþ(paꇻ@069¢iB)‚ò ŸLxÌÅ„*V÷”Tÿ®{ª¯}Þˆ„Ýq"Q©;ä?·õ¤/AD¼ £16‚£F ©<ûH¡YfCÛÜ ›„XÔ»921a°_\¦ð¦ÖÊ ÅÕø§‰Ãs¼þ<O^×ËPÄ%ÚÏ*G°} h%¢Ú÷±Ùð$J.–ÓHydœ\vÞÃeêd£tÈrT&ôgÙ$Ì×s^ihJË*Pí~XùÌõ“ðL¤Q©ÂDˆ1_âÒ(ã£w-ßãòz¹îGÙcfîŒI \s^ Ö+Qƒq³"ÓQ?x@v·P¼¦ë|³'kÉ\ôeyÀÝÆÚ&B._û¡a¦ÕÃcsÄÕÞÊ­}>r}XªÑ¬¾ÄT.[Ç„@e‡‡Wû놥öðªYýœ¨Ò#P˜Fø¶þDãL‘u?Y²GN<¢úF Oñʪë%Ô Ê^ÞÚè‘%€˜ðŸZyâcOÚ.¦%c\o–YÇt#Ä<¸Ï8 YÚïãCvø-‡[‘ÚÍÊs .%²öê5ÿGFÅ—ñà¶Ä³q饘 ªú¥´±é åòd;ÅùŸéÇžÀo¹",Gôoe«àòŒ® ÑBtŸ³Aâ£7¹a.zb’Ñk@c ýi˜x†=ÏS¿Q·‘Ô£„›KÊÑÈKJÀú,R„-GæÎEÀ#=q2f_~nŸFSyÔWïi#¨a×&åbðM;Áj…ôÄmዹQ¤å8kúŠJìa°›¼VñÔ4&½ŽtI3[š§§Â él_ÊñÈÀOÁBVÏ»êœCŸHxú-à`ˆZ¯_2´ó»œaÏÌBD|–ïŸ.I©±’®2ÁBß‘ÂîþòCãCxÐqëž¾ýÇ ©ÈR²S<@AÄÒ”!SÝN '·›äÇâ©T÷ëoŠï× 1x¹«¬dî ¢C\êF~°ûªøsY³­ýû…P™›u’((ùŽ%ü«Êc#¡ €º1¾ˆ“ê`l*’¯ëÉ1ŒAxbw=h¹›J9“‰6÷ÂãÑR'°Í*†hrT#y-ÑÄP7MU k#åÊ)d XŽ3ÿüÜ Éìÿ"rÖid·RO“ëC±ŒÉ×fË ÿNxáöž#y=µ…£ˆ×1àø’(,ª9(ÔÇ(YË8âN ŠÜ½–g©ªÒé÷¨ƒjK1BÛMw{(EVI‘GznXgŽ{w'Zþ2Yí¬œn´ÂƒQÜsݰŽöÒ~6pî8/8Ù ²ɰ*·«´Sšîá\šG¤®µi®3ó¿Èñ§‚\ù|ÝêPnºÇ©ûóß\0qÚÒÜ;Н}º¼¹Ü×häQ³/ ¼2çò¿§snÃ’ëmﳫÞëÚBVÑÉ- õQ YÇàέ‡Š’l ˜´¸w3U8Üw䥸#Laןé§V’åô]dÞÝ¡'ytr‚2!iXe{Ö°’— ̘r$5‰YDNÿ–c{>_^D'ƒd™6 uPAÌ^„‡†vˆz±ìÎÚá|1/áãXg5Þî<`q-.ö ªö,륩$B{F¡x­Í6ŠÝ)Ç«k¯ x¸{|]Ç'Äþ±ºŒV˜îJǦS¹aüÜh4÷Ö¾ÍA ç AïÝÍèªL ”–$³"[Ÿ¹þè2Kà`—wÄ;;^H¡ùe».8TŒjR¨ÎÛFsö-V<.ž ‡`Ž2†/&Æ›eÂ1c: 4›?}Þ|”„<í²CG5qwh#†z×B¸(=là>’§ýù·¼¹”íâç@ ª­ÓE¯Wêþ$ÄÍîDœéñŠaÁOŠœÞ±  ÍÄw®¹}ßA¹+k¦<òmH”Hûj%ÊçÞ‹†14Kûì$ZãæöNч¼…ºB…´ Õ ï9h¸Žd¶Fº½2Pö Âi_•a#&ž«Ñ¯q¦öÄz5E‹ì®òÆíŠ _»:®³ù{é솷’)°*T¦¥BC¤M¬:KuhÙßÝçqñ)ÎJ—â(šÚáPéÅîÏRNå’žÂþ–¬JD BŸž5ê~­Â€§œEV›´2\¾ÄdI/N€ýêC?\Ãܳòh”=cNP-msw·à£,Vj.Á|¦)VZ¦#FUpRï• gUïñºqÇ–ùíìŽÎ'§Gè!;âQíQfhÖã§=ò¹n}»÷hUkü¾oÝ&ÿTÕízÏeZÝ`!sÙ¸R/?Z´§¥2…ýyâXäÏ‹?_tCôëÞ¸h¡à yePãF’ˇÅü‚È©® ïá½ÓnꯗúHÿ @é'˜åj˜…7\ê´ó²`Z0š‡Mqˆœ­'³ßhÄÉé²ìjº¼j°›—}؉ÆõþÇ}bm-ˆñzUâ‚äµtrcUOÙ–q’þž/õÃÀ‘Ä“’úº&Õ-[Â?çM¨X„dˆÍºvT®©xþ–>•K9‰.Ëe\›(iŒü;ð0 +÷áoó„Ÿ Ôª{ñB-Œ£ ±Ȭâá¯öo ÷°h Ûþ(²HwÕ•BàÞר}sYû‰Vv®ÞÖ95ô?BáÑ-Ñ<{¤‰)³Ðí@ªk^pSo€’áâ@ Nú˜@®Ñ¿…Nt=?·6Sÿ§´&=i¾Y£ô€6ö¾Šæù ,Ȇ«ã‘õ?Ô/zyԠ굫žYyÛ1è›e*)4Çh𙻜O›B­‚_]£vÀmS™£3 â,À ôÆPc¥ ó=Ôâk`E^Õ¾ <¾òNìH¥ q Ò‘&I‚wµuºc´Vùªãà”œ Cû³ÿ@ô©bfpåKÁ^æIÂX  ›//´2Ô5j¦¤"}Ë)"õÑäSîn6m1NBÚí¬àjvvö;Ë‘êÿ¥2 ¸{‰Þ=eÑñÏ•9´¬€Ë)=|üŸ^Áƒ ‚Ê1†ú^ü*àÛ¥--EV R`„µ©aåeà76BÛ¾×(c',µ+¥½1ºÿ¤˜ô©uPÒjA_´[-/ $@åB¤_¹™‘Kø£ZðÁETúû)à45¦çôÞȦšY‡ ”½î‘¤Ã!*ð}³ Ûú-´R îs™àN½Ûi*’L½Mjaq¶l‹j³0ÕW-Ä¢‚‰Ÿê*[5ÚßÈ%óÒ¹Éb¹$ ç~E&œñµøÎeÅíË­ Hv™ 3„K×ÄÐÕMc›ÕÜ5ûŽO;9h˜\}+@95[çÂWr„³ÊDî½gýìe†LÉΧRd¿5¡È5•6d-)cfØÚ¢¹.ÒÇíÄ+—¶2OóÆH*%/6Âß3’%¶ÖU0«“€·dñ÷F ^÷y÷IôÌ¢ ©rxêÒS œTð¥KÎé ƒäË2”$h1™Cþ/öæ˜êi-r2ËgžqÅnRUí²ü†µ`óûh9³]¨4åÉ\)ÕÉÆNMXv‰ ~î$jo¹ÒuËÇÉÅ¿L[êóÉ›gƒ]5l(o~Ö¨ÎÍ!ýcprëÖNéÙzw*¡¹¸–w‘v‹¾ÝJ´zš¸üèÓöYç¤ Væ¥sÞñ ôUF3-bkZõ6€·¡À‘ïöº —qR‰éÜI–9tÒ]¢#€†A;àzXÎ%lòïÒÔô¼Bèä××r9chiVó„¶SýÄ.×=d R ÕÀæÃ(¤LtÕ©x¶€É*=fq¾œ7“—ü­í#KšƒseÞêvIl«í˜g K&ñê~‚‚ÌkêF¥&©ÙqFóÈ(’ÿ¹KgÀˆÐ!7˜ÍÎÿ@ptƯöËZË-#,S߆3v<|0îklïhZ/Ö™Œ9dÀ_(û°ê]ªÛ:%¦`uíªéY+Æi•$ys½?ïç”ü›fÃǼ¦q–»gn“)GB²*Ã_Žù^  × m<[ïœâ|TË w FÄl‡*dF*~¸ ]M舳»F€ì½Cõë›TtLŒÒvzœv#jÒŠï³Òf¸±<”Ýå ×zÔx¾ta76œß”‘má¨1¨êjÁ^ídjN~¿¡|ݬ±sÚjŽõáMÊv×*B½£ÈÒ{Jg}Ö¹ûÓ Ó|¹ý±hþ̫ɥ^Ú¢—Vã·Ô@u´޾ ´ƒ¡«½Çl3XÛuE¢„¿¤vÑæW^Úz°sV†Ò7°9ƒA1ÄŒÞ>IäKúãÒŒ*8€î)è\èþ\Xä)z• V»†Ÿ1f:\ ×hl× Â:Ê’V'+j¼ZyuìÌ`ya›ªÆþ‘×ÄÎh-ß-sqLÖ϶²Pï¤(>ºyPþ^ØÄííCŸð½»újKùŠÞBRq@xÂØíºQ¡—½Ë¸g °“"¿0ˆF"šTg‘O0­icpB†¿r!lë¶ ~¸÷áeÒbgÏ0Ñ1ƒ²NáÕu“Â»ãæ½·Wƒ˜îû0 XïsYæ*sC[ñkc¯y0rå_UÿÌo"þ6ú½Õ‚~ èåë¿ÈÕ¡4£¢. üçÝJ×ä ¨ì”±æ91Ú´]7Láw¶óÏåÓ†&EV#Y+d5¤Ç­¦ÈX²sñ>{òihеWd <ÄþôäèJ Ò´ åûÁ1)* hÿ¡v¶÷ËÜì˜:zMiÚ¬p dŠÉ?Š;Þ[û·°{Ø´ìIAÏ›Äh¼—Ïm˜òôå«pY‰›‚qón'Ô•Û×iµãÀ „ %O°qÉì„øŠöM Ž»#Ï’ÈæØäò.á¥8b³tv¸Ö„é@O?rm¯pú9¡U³9“”7NÖºÕšÔ– ˆ&4ëcpO&EGkç® ÎšfTÒQÜ©¤êåvbñ`pÎŽ:Ñ+éÞúi ±¸Õ¼‰æëߣа:ÅíÁY¹}<åÓ+ž!¢c³ÇGÝV7ã¼ßùWëq˜3µHŒ³>yó½^šÊ’8Öv2¦‰¥n*ííº ¡]#TN…ÎöîóX•ât!x™/Ú‡ÌI´3@@¦'îCŽ›S*Зǵ4òô(ò“\ê¥p~Á(z‹.:ëh‰yDò§¸á…GW“1×á€5il蹃íÐÊjh3/fí÷¼Äc°"ÛׄDkë‹wÅP&'*­ßRƒ.Ñ–î’Pb(ï ?À;ZOŸ )!dÃñóÀÄ ^ù§ôJŠd†!Ó Íü ¬VÕÝSáNÀE‘˜Jù5ƒ¼%v+4B–Fã´ r­èŠN@ð®8:þhýóÃAƒó@.Tõ\\lâ)«c$­/lNÐv£4›=à)¶ÇÄ|¶‘‰ŽIòDCéC0!‹¸SGmbT!q~pÂOµS[ØÖh²$];Erp~ôÍYþYHfè`N]J¢+“Þ?o>¤©‡3æš›¦ÜËÿ|¥ÉSþ8þ]n•ˆ›­!ÍÐ úª$ñ;^S·þÂøƒ¯22®¥Û®¬ÉÖÿõ~fèA‘—^õµd’,( HMgzà)zʾ_Rqð}Ÿ/?«åýÇMû$ûèѤ‚+©m2-XE-5ejÂHîlMÓ$ñËiSõœì0þöl‹`®vØ%J™Û#?œ*h"\(¢ãF’ê¼|É9´ã<Ö îŒŒÌ”w]g ›m= ÆMŸ'‰ÙÐu[¥x$îÐm‹¯JwØožÇ•ËÇ´†¸Æ^Çäóe75¶cõ¬î…£¢ª³D<Ø$ ySæ6ؘ+U÷Á?Çæ‹ÈeW±-ÇãYÞà¯HREMé2ÝàCeã¾ß'‘™Í¯Œ³†¥Ÿy üá็Ý,ÉMq$)´š×S–Ò‡­;_È«&´Ë{-úOÌx´ªýá˜dѱ>z«‚‘ C¥ßWêW|cðU+ïœg‡MÙñ~ù]F^%Ó¼ë(…›&ñ˜…¥›žA.ù¨Ñœu·¨_À † ¾ÔKw«i›7hláµb¡»Öë ,šJÑÁa mg5­s a÷ÎóþC©ÛõE§Ëf©&Ò¾tñô'7ìÎpÛ FÂêý>3MQk~XNí9R—¦³åabLŽ][/Ѳ{ƒZ~è»®fÚóÄ™ åÒ-‹yuÊ-ØD·¼%謼º§ ~`÷@Pƒ†i»6ÿ~ • ó@$¸-µr︭>ÿGô ÚõÕ“ ‚o 0~æµXOÉÈ6ô&“?)&ÏÜQ[ݵ–rBxe1øP_ÖCÒ¨-æ®1hgÜM¯à´?ºÌG›Í/ñ-ÿ¨E¥ Ñ‘fô–ØÑ¿“1ÜhBe=IjFú*A Þ¦$ðl­˜÷fųe£ŒuÊo÷u- o2Õ9_&Ãt«íâÔ¸Q)ÊyéÓz7ESn£¾3ân Ããê*`í.Ò—ÊßÌ•]ƒªÐÀùîž$C:·±£æ>ÜÐ#‡°Â5Uç£0ô£ç¯ß­ùëQ~÷d+!Ä/Ê•VÜ1vå°ŠürRiQBFQø/o&‰£]MVÉšÕ¶YFsuñˈ”:D<½<ñøB¬Í(>©ß,ÔD‹K§¢]`-GBgëæ †|A’OÈLc'Ul¡©3/é×p­Î„#“^6ߎì MÛú„(>ä´m  tûµÿXumøõ±¿µ @ü)_3‘ÇDä° 3yw[4°^¨UgÅe‚Þ …³K&–8†Úö$XÿM€ÙJŽ8n1@ÒFN¾ŽA¸Mð'!¹=· ªý£DAšD}¥Qà&ÜÒ0û¥† ~ýU–"½Ú¥¦‹Ÿ”pkéÚ96ÔÞ³$áæGùYH*úÃâÂBÛ¯) HÍý׈‡>ÁDÑ|BS'¾°e¥T{ é`$Іˆ‡u;Y)k6¾Â»Û-’ʺcût@/ýül‘ÝÜ¢‰œšD_A0ÿMÛõÊ(×zõ<áLq*“ûýb–Ž£©£µ^²)úªÈ5i¨ã姪þ¬€Ï@¦šÐÐôPˆ®a¹öŠ{Ç·†œ—à&nA¥" LÔ^d7p,b#çA;£Ê-E ‡÷ZF á]^ ú7z ÐÙx99ý1æ•ø™ïWßcêç¢â›çshï/1?>½.?Œ õHÞ,Ëuš>ƒÜIvÍ…ôg"js¨ÍNÐÙÛ >–zÁ™´%Šéå:½ù…Õ‡#T|;NWAY3nÊÉßt)‰ü;E­ç=gðN•šƒ0]s竤y›¨iCø¦NÂé8#?d‡©½úˆ‡õ5?Íù9ÀÄ ò"„x1f†’:kãûKCl®$ü͆'•Ýsm+!Ž}¹*Qëì…3ln¨ ÓÐ,J²(º™:+ÈÐ!3Ò˜¥;Õg¦{7  4ë§rj±r¿¢WxMCQ«Ñ%ôtÙ™·b-Je";–Ì2Þ™j•bÀŒÆ"ÉO>aÆ’B5z‘çþ*IšÅm$çeÕR*wl(»`ñH~Vßaï4Nô¤ßà„î%Àf½†N8¶qÉ]â-h «„2Ù¾.QGG£ÇÒUI .D ¼ „3|Û5ÓÊì4i5 /.W«Æ5•áJ(?Hof£€Zpöâíò‚MSž½®ñ<ý›ËæËHϸ)/óÉC? ]ìñĂ۰Ÿ-gø¯¹|CN¡š.õw’¬®äˆ†éÌD$7æÜjÊKt¬Öâñc@aÈæûJS´Â¯ÆqŽd¾\fHȶ´¼JªCû¤óðÙ>ž†ÙÄTl-ƒ~–:“Ķ[ð:~Üâ®)AaõJÍ.›¯.&±ÀwL™ ûì=æû þÙ$%CXp½î!û}æÑ‰ŸOÑÓþ§’P„fœÔxºô“j:¸±î~›½ñ ¼oÐØ¢†20dZ„pè·°N¤C™—Ôµ¸‹õ“ÎOÔ޲”!fµ•çöäe°Ï)ˆ!œKáòÏ|:LnçÑöÄHÊcŸ¨‘+ÜØà1³%žÅ޼r†=°Žèÿª?S½^y~Nýg <*‹¹“Ÿ£c³òÉ ÜÅÕùS8ÐOeÏ“ƒ¬sŒAY"¡è?ÕË)AÀužŒ%׆ˆÜåNƒÐ~º—§ÈÏÔL$ì?  ]®~a&[.;ŒC{ËÙòÃtåøEÑDcOû°EU?;]–¤=êâ@ƒ_¢vÝ%úd—þªÔTƒþº>bddŽ('Ørb½c…9Ùmï2å3V#òuã|;@|:9’””r›ÛOmxÍùÊ€ÀÂ{ô^ܰ—€4\¨qâ}Çm›¤2Á&¿¼AñnÍÃñýéEj‘İIìÐíP}'uÀè@@"0µ8 KnMnI°;k&m!Ï®t%m‚ÇM£Q{ñŠm…ËâYh-‰æâÎ^Ä9 ¾Œ™>¯âp1š©†3‘Τ¸² #F¥h<{ð"ö¨)ÒpJxuF1ÕT}ÂåN}píöýèzë2&Ö/¸ö:L9çžhE®‹·(Þ‡Kwv¯»eÉ*«þeü:]â´¶£d²“…ˆE|ƒŒ[ùš*N^Üé=  $uPß¹Ö=  ÊPÓ¸ô2JšÅ>BùAx8j3È.'ŒYY2Å^šò9™¦dëM¶ŸÓ>9ytcD1½C¼ÛK§¶˜Ä]ÉdÅâ"rfp s’i\|¯·'û³A'g„í…rª€îœý È€V¡ß³‡³ ÀŒ›1y” ¦GFØä/YΤqG‹TˆÊìs?7zöÙ7ê]ë4ˆ!Õ›ˆ[YÄC#«y î‘L"Ð!·«ÒËoXØÝV’%™­sV³ 4‘cú>;’&D§¿?âÂô“˜Žº”D òQ8^üüÄ=u´¾Ì‘†L‰: wû£ÅUé1UïùÖÿùa·yÌkoDŒL ÃM–û-ÓN=ò¡A²Œ¯}1 â4¸¶ÕîŸRWßSœj0Ì騣 ¡ÊäÌ"ÿ!L;%å©tfp ju*ùK`nÛ»uÿé3þ8Ž$«,ž1ù]ÉÃÞ‚ƒ_!6N²óêe®• ¢ ÎªóQíHHð<…ok‰NJ ¹Aþ¿ðu³tûTe^ú‚¼q”XO®N)a1²Ù1@ùN”AåïVíFr<+WŠm*š«õ‰2p›÷Í&)ÕÃ^9òªŒj¡Ç%s°k,˜d4ꇫ¼å¥z?H;jšóý6µY.R\žúÉàM§Ø§ËÖªˆ·&üÝù !ÊÛ¿e j% .•®Öéžö¬,½w&>©àR.*šö!•#è«âå3¨'ÆD*`¸+NÎK•€ój[è‹ëþ'íyeaXTT@jõÖ$áR 0ÈÜ¿°\µý†ð¨Ð@¬ z1ô S)ù» ¶ ¬ä«õ ÃâŠ=?*ü)¹8Ĕ޸õ,þíþè›=/ÊþclÎ8îp‘Rü}²Á9\=ÐÒ›yNì©qZa\¶Žhﮆý e;²†ÉOºf¡Nh F¡^¹½¹Ì£V½NuÉ‘@ýÌëO–‘o+¥ÅÏHýç!Þ± §×Ð倿O¢0XdšVT=ÿî‚Vê2„äýÓõ_÷<òèæilX¤Y.—@\å]£]HCa\[7^¨íÊéVV²6gÄ„Ahç_î°$ —KÎ+ †ª áá\Î,FZA>|Úå%¾â€£Ë~÷%˜†§µ­x“¬a\¹ÚêÃpµ5¨Ò‘Êã5°‡¹°¤Ì2ç˜5TÄê;x±¸qL¥,,™•ºƒ÷‰œÄ©œiK¢¹³/d¯ç§AÞNM»ºoC¹)\ÿRóÅ0{£ïT@'éØz½ë0I>Þ¨®0¸¤~M²l `Õb¥yR’?®ÇE¬zmM¢¢oÝZŠ1Fú^…•ÊÇ|NÍHÏBÚÌ”œ"ó!O)Ô­Qòä”—Ì,þNÍãŒ×fì?k>f¾Z:_#2\©8Õį9qx -§(ùª±ÂªÐœi6OÉoIšŽ2&TG pMQ-§—8Ž1Éçðö^̧ßÿÑm¶[?o´ÄTÃծͲ‘õK¸åìýk† ÞAi¯ï!K!s²qøPQq6ÍUbx“>nc¿ÝÛ0 Tt9HÙ)ITq`‹ìû9&w@1h(õk ‚dÓiwÖŸ¸˜O­ÄQÀ.ª­SWhŽålÛ€@}A’B²Ù×àæe¼o™²ùi€÷Uò¡³Ê}€^F9ð©=¥ÕfZ…\5$ŒÈQ¢ÝÒÇÞ÷²F”É„´‹ç,‚Õ}p›9³UÍË6ûéðôeÕe@›<ÚËV,µ­-HƒA}ñŒëÛr‰l=‡CûvI_ ¯Ä”mÌL—;ذ2'˜â ¿µ7°ªe£dƒjÅ{ ÈÃ…¬5<\m}û€_McíÙ žʸίPÏá6Êìâ[övÎÛ~ `š¤›@\ñJ(sÂvÒÍ.tø —çäÄ©l[äZ36ŸJ„]þÃ'yQ]_ æ¤ÏÙ—ÆÊ­HÒΨèÔ ä=RdŠÛ,’¤šqm³¤}KO@Æ|lsCñÿ_žvZwyk“s^ÜäQÚ‹b¤" ¯¥o\k§B[ií“‘Áß¿4 ÷ýÉäŠÑ]¬ÿ¹gçfú“ÿcòJXçäë…¸osE|šŸ †½z…Жv8ÆSžÜ—¨µ0‘ÍÅ×z.QŠpþó2#HÃ_2W!ž"íeZš }ûeŠ”odØFXî#,vm›”Q „}ž!áö@õŒP36*ï©ÜÌÌz+eAÛ>ú¥Ô…t݃ªÜ¦qŒÄ/BÐÆLVjn²Ó|YgI… y´ul=Tåì%¬æQAXûÔÐô^º¬›ç§yJúÎ&qIýµp4¢¼Ô×,¨êJÚMZÒÄâ9(»;DuRÅÝû‘œ«C[à’•†j“4j^ñÒ™¡©Àˆw«£× “·æU=˜ÈÒ ¬Û×°|F„3ÐÁ•ǧá4V®Zá¥îÕ}¥ L~•°0·¥ß‰Í.HiûígÔѬ]ÅĬØq¹½¦MÇìE"}‘÷é Ø3Ù‰gä`s(;'¢ö‡E{µRÚÒÝ¡™gföò=§€ÂIÌU\`b”Üýþ>7qE¥EµµAænyj8/70Ⱦµ¼|åÄÄÑãÕ gž¿¥š/™&—?J`éFbT®¶ÝÝI­‡ªÍGçRà¦Íò®R Oí0M_ŸPökkÏ›¤ÿÙ=æB*@-.Êèy,†4Ʋ™õQÂØäoCóÜ þÒ¨Ÿb„¸(pƒÞžƒ¸(;pF¬b÷‰E—¹Fš‹TgEb× `E2{}ñHòµ”SMzC>eºýÐE7­ìyJãZ'O’á1Qú3‡b¡uýÞøÄDT$A{ì× ëˆ8Q}£S†ÔŠ»á†ÝÖèÙÛ­Ý{Gœ(ŸÀxKj¤f’~g ÉYeÌÁ]èLÁî¹é‹L.—¥ús®oFÛ*%6ý€Pä+èfʲ ¬¨ŸºP‚òè7ë™0~üYÿmïÜ÷£:©7ªI£dOx^ŸÍ\ÖÀhÓç{k™ã…{i­Ä¶¢Þ„¦‰q¦Áø­Úž›Ó¬Ò†ªÍ|ø´½ª¨#-ÚÈ›ÍÉø´†›€Z7¾C%Õïižpà•I„F‹cn¢È—cºô2Ä + ƒ/²è¯Ìh*ç{Å-ÀÔT¯sàbØÓW1¶°òòR›a þ\áâwu‹D¨Ë²Æ5OZàTûû v: ¸¨–5ÍHܨ3*5׋anorÀ“ûgô¢”®Ôà Éæv((¹ø#^P%è dij‡õ°ËÛ<–ÎJ¼¼cþuN" £ƒKד3PBm×(‰ØáÇŠùxðe¹b©x2éÒŠ·óVòƒ­€D_rpæmÂ0ÕJYº®fM?Na~'Ê`Ȉ'C¯[#fÃàPÒÁmS“›—9ewæùYÈ s˜ÄálŠnèa¢g¸oÚôÄ~}tüŒCÇßT$c½½YgìÐÛÁçóàÏV¥/åÐv¬ÂÞèI•:ñç¥f±CÙ{ ·u68èªÆ^Ò³¡ò :3”ÐQ„¨ƒ­ˆæq{[+Íî0fÒŒœüÿí:F@T··Û· ¼ÏâÔ‚­Š‰%Sÿ D/–Ëd5°-„Ð2Ÿ©¤í%è’K[.éÎ÷Å.]Ö[œùÄWŠ&µ!Y¬I¢«räA=‰ÏïkGäúè bØJV>áäh x™¾;»UC½³Åyù’Eôsx‡“ïqý?('‡|Šžø¬Ò%^­?Ú” `-ê„Y*¿öOGÆìãDÇ#¨åt›Æ4T?³ÖBuãÏ‘±ø´ÄI ÂV•o¥ð5 èD§’N]„ðÞ“F§kJT(V¨lJ›$ƒG“=zv•º—ÙOO;¸DŽbJ‚²Z²áyî‘ ^¾vdÞq4a<Úf¡œ ìΚäŽ}`þ,wE÷ú­,‡¡”BÒó ËNeeó¨tq=€¿BÕi”+³k ­df׿±×)þ7~QD +˜‰ sSìÐGÚz*€Ãú)S+€_¥ÓG,´ FF‘DZµË—ä_-"Àw5pyY¾Ooi[ˉN §ÖêÙFßO ÏÖ„i:”ìÛ”îº7{±òQ›„c¬P²ÇkÉ÷›w‘ŸÊéðçër D=2EžÓùSâÐv}Ãp=´qøµ%—_Qª $j!>•¸ö„·¸äªÎt)Ù–œwV%þ¯ü®Í'«œé˜Ò™–ëæç­N›9G˜Eu,“ÒÜ]j“R‘îˆòIÖc~¦óPÙÈJãȽ¾¹Š¥8~Bt›Ð>‚ÁX@¶3$ŽU§-%Ôâ9U>Æÿ¿¾»'÷ÞN¹O¢ÔãfâUª<1&$oùÇ# ’?B‹½bÜ›òÌ¡ÌCW¼$ ÆÂddОü`×NõÓÝÙ\”üÞоÓ, ì•t]÷ÇÐ>"}‹b†¤äÃ%©†ÁÔD¢Ã` “‚4R®,Sµ¦'ë™Tb¶ÓÚ[3•¿öC3«ŒBRŸlq“öüí$&x6¾£y®®ö‚i¸Nµpª Ú—9Ð@A¶|ÉæSUE¢y¼Ñàsž4HTÉ3y¬™f¹­â–$Ìœ'ü<ÜCù(î„5µÏÄ!ÿEE÷’«äÞÔ ÏŠÆIL*ˆVnf5hÒ7 ¡ 4zÃ1nˆ±§ºQ!ïê‘ øÔCP¹àqÛ°ÁS€/.¶Hí ÷ì[i,ÝÕèaÏlÓ&ÞåbÃýÚÐÁ ßp9M¬=ù±¢ÿ‹ïYeö’cðÙ"´dþ®¦5«•áÂݹ~HO–ÃU`o7ƒ­02¼¶¢yàj{"À½¶§„ðA€´¬tÅkûÁý[És­W®ÚÌwÇ9ݦ´nsr•Æi¦^—]»¢%-c!ŒªoÿøÂÊ“Á=m°zBƒ p§å»$ † ié¶+ÀήãÀß™Žd{é-¦^ó´4ŒÜõÓŽ"Äqeôè»_t}Æšû­é÷IW#' uJ᛺ 2ýѵz&ë(º·;nÏ“Ðe6zòGÌ8‹@òkÏ`}ô÷d.ê²V!%®tʲåO+¾ëa3 ‘7yÃP…ÓÎF¿’Å–çö£TÇ\ñ~dƒeJךØjPikY<Ó»~­î”k?‘rT<l>Oó‡0RƒÜ•åo>ŒlqS ß¿Íêpð×Y¨{zòþÝSÚžËü¸W³õ›R=¾šûÁ&xy‰Páÿ«ˆgš´Ù,léqƒœ<}!‚÷ŠŠrÕRO:"­ÂárÚF×¶ §îÂq"ο¬—õÈ–s¦—†×H‡ˆ[©›m˜{|vº[Ý0ãÊO©Äûµ1þ2‹|ž ÷LüPÛ {Ù«²²ªÓ’Iç%Ä_~µ¶¬!Р‡Þ282òqÄÕAWΗü„`w‘vàâUÁD¥m0ÆchSð˜™ú˜¿k.¸ØTY+ýÈ ˆ¯s˜øAa+éù‚([脾½gÁ !—£ÂPœ1EžW°¯=4±n_ŠÞ5ņgÖ­SëXÒÄVŒdf‚»R Œ)BtAšŸ€™›Ü‚ë |3 Ú§X2L£ÞÊ|T>”Ã…Î’ºE~K½Vr•\E™ ý #®Ð£,¶G~ÿz¹Ù ì®M“#ýJ~ð¥Á—@c,Y:^m •a”õŒ'€]vÙ‰Öƒ!dÉl㗩᫦¢OjM³¢™%a±Ü¶ûg°+¾¡iö¸EE `hyw ª¶„*µÞDÕÆÖÛov|í¸ï®ôÀ2ÇæÂÙ.•ÝeÊ…©—‚D[¸ä:ÜA+òÐt“ =ãƒCáhŒŸ‘æv[LÿL¶œí´ñS0¡ù?ÒòŠK§­ó³4Tñ{«]ËÍPÞÍùÜ$DÀSöMoþC€5h¥nùLQJ÷[”@ôF+¬~Nwì4ÝîCw+!-ˆÀÌ÷i˜Îa;‚;z3 % wdÓµpÚvu9×íãpºzþÁb¶Æ)ûš‡nÊÊ™¸w&DƒxüçgÏ-U×Q(üÏ]qú»èíȯéáÜëù¾LÝ"Q„³ÝôLÇNþÙöácû°fq+1ÑHÚƒ3‰,»Z{¬¬«31Òë‚@ËÓùµŠ{&­äliH–áâÒn€°‚ä¦ïÖÐø(@-úø*W²ÂÀˆ.}‰áìÐÞ]?ÃjWÆKä[ªÔÄ„v>'LKWPÀ&}:Pd7çÎÈÄS/ÞÊ{}!~²@)3XÕ&+ØSU4G­Ã¶õÈ´!˜=¨H]êi3‚¿ 0¬kË”á?ª}©³ÙIxnûzƒx^ÍÈù·•ß‘#´u7¬+ŠŠF5aEÞ¥|šÚ­ÃðB~3®ÐΉžë ´oã’¿UʶšgÿÉ_žã×ÒÌ—[öÙ¾o²%ʰ5Ç£kD"i…3›Ùn+–âRº fÛ„Áùü—+ØÕéYîåtë9ç³Ô ÂÔ»d ¨p—ÀPÐ~»— dx’–¸™ÛžÏ$¬òh£¢M»‹iªšîƒ»¡I òRPLÓÅKe8Óp¹X=á2mé[ŸÇâÒñ} Ò„°œpò{ ¦™Åk"Mœ4n§JÛ5äÃV‘:bÒ7s;à r²=z§L,0H¦òcȽC6,I{´¬€œÀ€øÝ–¾~ðÍ‚kO™—–<¸ÕÓ.¨Ù aÈáôµid}¼£Ó´·”¾9$ ÚšÜçC5¼ÈÒ Éûìv?lTê×r&ç$™’ýÙ·I÷üãmhóÄæ×®*«¶'ŒÄrvA~7gB´‡ˆÔB7\µ‰ôi˜¥Ï£ï¬‚ PØå‡ ññ7ÆÐ.­¿ÅÞ&Ä,¯†þ&ϺªË}FbåÆs“£Qò’Ò}§êÒ—3‰(Å¥ÌTC¶n»`dRVŠa';=°èýô}x8±\x¾™þ-²˜¥a€Ð,äí$ <Žz™N]£™½à†)N ¼$ ½û”7†”mg]Çq¡*þ$Í…˜.*¦ÿßóŸöØÊ PéûªèCú-]+¼fÙ3ÿE¡¿w($ç^»+ŒWNd&±}ƒÊð-³~¯øL8fÊLmš^¡|*[”¯ù&«‡ûÕVjã ØQ.«¡@þIÐÒBŽ,Õ¯?¨ã™®7k#”…ب’XÊþOjï_›êË zP7“-iÆ7¬N¯š[j✹²ùn¨ÌCv[«KE gŠFß,×Ïõ:ÑõMS‰ù{è1F«Iq9c2|W~÷õÇKþ°'U¡Û!+¸³oÒhÊ–RŽÜS‚¸»ito ¯xMÒî¥{±~i`øÿ±Ã2kçe‡5}Ÿ½]<8ëBp”žJn»xߦ£îݹã*Wœ\±Qk`ý)»à$¬B¿€,*`@ÀM®¼xÐG$¶ôÀFn>&F„î“•k1)©Qf”ÊñV;£,ʲËýƒÊ^ã$äP¹,xËÔöÖ,åéfÿ›c] Î×3J$kšõ ƒ“£.€ —|˜‰¦¦ =‰7 2ÔÉR+Â(HÅ"Œsü‰éí ^alø Éü" 2¡àóLܬIŠ_ÿpÉÔ oÊ̆{çVÌÝ=%7{Ö¾5åEÌ%-ê7è×ÍŸ ?=<{ã?ð¨_î•«“ ™Ïm}¿Ô2†'Îs]ýò0då®' Õøðú-£¹ ~‰–ZbãE3újeÎýÙñ{,*O|º`Yh^Ä—¾ E ·RÇ ÎÀûAÅ.€ÏLI‘¬ÓBã7I_ü7L?èÔ ²ÁÁÒ½­˜3³T ŒlÖÚ²¹íÉ`.ÌÉ>®ú$ÞAƒßŸ¿†oD¿ÕÃt÷à{¹ÅÀÞ™tAÙ…Ì%¦…€XóÌFus·ÇZŸtÍ¢xHV=ìÿïpä½-¨!÷SKOó|æë¼mn;ºW„t\¢г ¦7ÜàÍvñVJ`ݨh5äܘX~‡+wØØ@ã%|g’›b%¥È*i¿j¶Ÿ¥(¢lÏÖ‰çˉíém™¼+®û”vÄv"¬;QÉ„fl ˜¤§.bžätuyÿšEÒtuÌOµæd¬ Ó1Jœ7#Ö”bWòëHÆZ_ g½óÄn0ìºT¿Ou».Ì·¥ðWÒG‡6õ!›ƒ>0àj°"¼ßå:³Hj]ÅãU©EM÷’±hhXß̉|R»{4¦ˆòåÈ8#ÍÚMèÖÇ- @_/rÙwœ”þ\ä‹Å@-Ô!‡=îéH³92”~³Õ×v*›º CÝó HÙ&‹aÏ×I±pjUÚÐÄ?!Œ÷ðª/*?ØpNö×M„¾¡á.N)þ¸†éé´€Ÿçœ¹"keyÓαÍ|Oy‰1F ›’eñê.,—c: û‚3“o¯^‹Ñ=§ÿh‡÷´‡µ§Q¼+Šc—’i.Ì)ñuý„£¨+žzð’Ûè£ïíu7ó`3ú‰³s0!â¾JÒÏ ƒòm!²ÖJ¥îi?½8QO\¤!âÞÝõŠs: :GÎpíøät»ä$èB{ÍESJçSjÒ¿±ôl„o'’èÌ =>2ºsi)wí’6tØ`õä9ílMØ‚Qùw”êNŒ Û¡z­ï½pËgC½+S¨Dø–éùÆí‡Wê-ñŒÂÜs*(^N²vÚìjÜàÏi&„Q8È(V⎤Ÿ±Iá$¨€ŽÅ^_Â.”ù\Nm)ôÞÚ0Pì*ZùÂæ Ww¥ªÑ\àq2Y0_Áq#~`Fe³–¢­/½¯z3ÀÆá¹a1M Ó/¡~ggˆPjè Ø’“ߢ¨`Bì º|Yá%øoE²‹É-2‘Š6N!>C°ã&üï¸ì“‡^vcز™êb!Ä­”‹šÏ—9u*,õA¡ ›ÞÙfrX›ÜOЛ 3[©±d/f'ñÖÌYfG•Ìm$lìÿ÷X{ÓÎV°‡IÞ€æ‹.8¯Ôm΂Ù©,BvºÿßwÊ!÷{¹ï&0fA¤€©¨¯ž¸mä#‘öÁÕUƒ¼¼ŠJÙÛô:hY» \J/åë°¤†-Lþñ 4 ® ÏðZùêÿGôôSÛ/=Rk·„ê(³=PtÎöWÔcgeÓ¦ÿ€£ó(±_4 ¹‚¶ém"´¾‚{~yú‡ë•y¸B˜¡ÒÔÏ4Å mTõè:ò3œÛšºNKYÕ x·J"j#KOïNÁÞ}S¸ž\j^Cð™°Âý÷ã±cŠ™ ùˆ¼¾vk_)Ž!!b|j3«ö»EI„‹˜hƒ  ›ÒýÞ³¾ômATQ좈yôb0ɧ¸¬;G訨ÉWQ^î:UŽ­ã~?¢ì€¦å¨[ÕLà սS·z‹ã9¯A^ùy t‰aʺºi&%%ÝD‘“nm&VCF Ï„ ú  È’6«núë¥Eb5Õ.«à×Üý,{i8GZ}±‰Wµ¬ßŸQ:M*’µvå.«u=ú°©]dsº§ê~zôï ˜Ùܪÿp-PÄÎzÑÖ³roE9ùõ¢žµžôd—Pþ5öê Èɶë–üåw;å¢ÄÖ}*‡’a–R×k‹EþãKŽ•¸d‰.RºàU¬(îA»ßL"ǺY¸WKÚÆÆöÕYÙý7¾ŽOŒ;>aÅ…Ø'4B·†Ÿ(M•>bˆ 1Qlì£y‡O†2×ÎònF@|0—¥³tcƒÚlz^ß01ÖJ•Ó'M9Z¦úiwûÇI` @égž:C¤ø€„Ýù»úœö­¨:*NJ®[$Áè€Ó·Çz¡†é‹òÒÃîÃø”é0ðäÉè^7`¥Â£^±Wy¹¢)Û åž^æ|(ç“oï·¸¹ù,”Á.ª“‘Ý0<œú­ £$›?9$‘+Ⱥꃳ;uLC‘º©ÝwO;Aé¹aPNÿÇû•ÝÅI.·é¶…bÆY·óÙG-KÏW<+j=]ð ­ »‘¸s0Ö­ÿkZk˜×Ì—Ì”"ûЧó+Ë>Šrf…°RÍ ±Ä«ڣ•{2«™1ýhcÒ1<`)¤ËÓ]3 «Ìî§€÷¸ÊÙ¿_tÿô‚©ý[L‚ê­‡J½y´%Ò…aäÐÓ‚Š?ø¡´µ^Ò¿åqI¹KêÖÒG½¢gÚ¬±Ë+Üû{#†­±yl¥X •xæ…ŸÔt"+áôi ® `\ ~µ#Û …d¯›Z}7`x§Â˜º,O$z¯²ÀØê¾_ã]šÅ:Ö¿m„‡–€¯vÏȇ6¬=͸hŒæ¦wùœà¬“öÚÇ ¬n…¢a2ŽçšÂròÖÅCZgí2OžñpÖí0¥­Ô=(•°Ó‡‚^sáC&½Äs¯çËWŒ ÅëQà¿KÇã“aø,?Ý3#õ–T‡êz©ŽÔ ô©Ë•‘UË”z'ÚˆÖ~oùêf,Ÿ‹'G‚6s“Tƒpe±\"âKC9>ÌÚˆhŒËÓœ¦›aÚ•F—Ø>Õ–^3HóæZÅÜèáe`ÿî8„{óØç7¥¥ û ¼Ü§^ÿTÖï°‚ã0ááÔœ¬!Dì-ðu@N]ÇÏ}Ìy iu™åtsÄÒ‘ÙüȲhúÞ'ÿMFfF‡ô/ „Éûµ¾Wƒá/D¾ãP1‰ôÄrmø/Ásÿĉ! îv;£¬Fš\¸nDtÍvüžT#fs†I1®¤¨‘Á&z£©XyÑí_YI‘b™ì?’…uTçQ§£¤ÊYûî Ìr(Ö>¡„ÁJ™¢ÅŠÇ͆Ûàä½Jº0:§ø¸-< nô»Ít[÷õØH%gü[®Züâó¶\<¦£óc_v " ©ï7–þ\©ÍÅjëÛÒEÖÕ1ﲃýÓãšqнZÁ57Ü£ 348¼Œ pì ôu$¾]±gDó_TwÇ›õIΆ@ºbÁ•ì=ûZAÙ7ÖyÁ+q‡–ˆˆ„]·Èþ¯î™›qÔ½<¹öÁºa]ûªïãŇŀAÚ˜°…$¢2_ –w5}¼.ˆ Nð·›®³£ÆPâ&A(Ÿû4SXC8Øààæˆ•Ó½Ý/]Ô/œ‚ê úC6ã¦8wÑjÜ ’§£4•x¡ä%Ï,C η øLZ{}(òKÉÕÊy ´î<(Lçþ5PðaþäщM²\Ð'×¹òV©µ.¦—û”9~€;êÝžõ³²ùL€ÿü-þÄÌÎÄKYC‘Fjx¯"ˆ8ýzH(ƒö¤ÇP\sí.ˆ“ «Jg$’ß«{“øP°Añž¼Kpœåµ+’t²l=ÖÞ[„f»$v{ý o™ãçdzˈ«é°-0_Wbâ'9˜º=îÝÿ¡‚ÔQÛ‘¦Yg›Ó`ÅV³šF‰ó•~Ú6׋Gk£|3è4xÎS^·B9@ä¾¶ïKÉ)ëKä0;(³crS°3EÀ¦V0^vÜýOmÈC¸Š1%fŒÑåÛhïýîæ¬ÍHq·8ëÅ\Ý~,œ<(ëÖ»à57*M“TÓaA\Ãv1‘)eV™Ý¥rO[IöýÃY³í9÷ÞòA¤2çUm¾ùÜìæõ¬ˆœÛV²%pV £X9Cî¸VŠ‹ïýqŸÀÔWÅkü¸ò4ù‰ uW-GìÓbºÖ;ž·Ž»iˆM%e‚Å[>ÜCÚœçR#”i6=³uM‡(q#æe"oåè^¼Ïƒ~3¼H+KôÝ—W¦ˆ Ô߾NyzÛ«÷ÉÖ ¥ý¨Ìº¦P—m¥xÈQìâíèb›ª'¼üLcR«}uuÆAËü<³´+ÿ­ãIPÁ“µ§Ÿ ´1í¸”ú&T¨u´ÛY~ vá\ûFü`¹a³ 4—æO#}=Ë4Ã0§ú™áa½õ/ÞM|úô›·%1ÞLs{ ꑃ–D ðØ_à(†‰¢òf T†V7»L•X:J#? +³,DŒT5KeÏ`Jîð—öâÍq4Tâfæ›= pºOJ‰«&‹¦W4Ýö¶-ãÇ01ªÜôÇÃŒòÐp–h.eùlÔj‡oƒvœÇ!šîI¢åíÓçˆ~TGѾ}ëø þ '}Gû¹hD¨ii"ÕÜ„A×ú-gÉ`s¥PdÅ /|M2ªŒm•Ì!µFüÜlL Œã„£Ûþ‰Þ‘”¨!¯²—]–2þ-†SÃÝPdY¬ Ú€Bã¾O®¦² sí~€ xáår˱y\iÌ5ƒ“Þ>‰ÓÍš8Ø»ùZO1íäkžO¤×àa=FIœ„¶ø"šô Нpɾ6«C…j›YÞ/¹)ÖéÒ_/jAwÊÔN£”‘¬3ÿŽ$’r^xþl’øÉôørb¨³oDê “«#ýq¤óþBéÒ)­ÿÛÐ5»ƒT~x÷Õð˜dgVÇbŽh,Þp›Ù,HUÄS> ¨®]®ŸŠã´Q€ -ai€Ç5È&ê+»|@×ïÙyê*Uô«r®?Œû%ƒÛ:ŠSrïáëel#Ÿ1š2½z~ÛW´@-¬×u£C†¹fN\Ðééš.ëïå.2WEÞ§}%ÇŸã:U ZŸˆ„?z¡,¤âlŸ06Ä'B¡æ½õ¬zâ¢óà†q`°01·¸ fDÉú~®dPW…Ê¢hdõ€ X䜲­ÐHËHU?ˆù#½´ãf°š?™&§ >[áÕøÖW¦1(!Þ`„`.3£…5 `oÁlwß/sÍÕ¬§i"þJ NŽàOð:­èp¢ŠŸ¦`–õ‰sn¼ŒŠz÷f×—*ãTj?Æéu„zYégµJXw|"Eа®wò¥Ñ W[ðMè@u"n0e'å—ü& ‘¦A»qxO æãU”"*9™‚FܯfZ†5©°[Ÿ“Äì=c-JQÞÚß×Xo®H'¬œXÄ7y0LYžv•ñвï<}Š(=%ÜgÃV5 Æ:¼ú SCÄ»uO½:çØ¢‹È`áô6OE-¹YK@³ëÞ5­ò„fúp‘7Ðé_ÚE’œÐœ`ÿ€Œan¨·—éZ„ò_32Ãy]líI¶'²˜¡ó°g‘¯˜":Ãí€o ˜–/P5}œ»?~4½éAYàÔ½X/îS6L¯ÿŽI/Eœ+­¢f.¬6Û+ça¶¸,£ ¿á~’õybB|…5‰áœÐå¸ÝæÖ Ae€P¨ülÓ]Á?ùp ìž_’$E¸‘|»”ê5;Fòå†ÍàÂ?~ M謷#ŠP{ªW[0s1&Œþ Ô¢^p@اë_ù&q¨æ‰MìgfOˆ€©Ñ:aÃmk©q¹öÅÈ(zÒ¯©ìnõm“:él³üXwó4 úÎoy^ÑæÏ®r&)ûD ÑSÂEQ—Á"¤²©ÀØ òÀ*T-pfÃ¥ •ʹHNXqr~ ñ=ÁíÈUƺy“nwZ7;Æüs‘£ã:çÙ{ÅRfÉ©#‚¨~„l±ó¤’Yú/Ô"§5¹Ê+UR£¯A•¦èCѲhj¾‹—“ßRçüÎßf<6•މ=y³Iñr—ìåQ/h>ݤhBŠþer»…3/=¶¿ÕˆÓNç1®´åE/ ªå®ÏêB ñN&ªk YÜÆÌàÙ8r6Ù²8£àÉYå%¶áþù># ½OÕð‚æýEâ™9Riiª¼"©¤ÕÆvJ^׌@Ú‘#ˆV¿Å5ë•ZC7¿0;ñ‚Žù¾¼"?·]‚ÓLWܬTÉÁÉ›,µ9_HÒÐÙ]c‹ÔpšþƘªgð …™ jhqy]³‘S½ÀðWkÇâ§ºÞ ‹£Q:n: ObCÅ=¢c5dœÇk/Ûþ©fÜ=Ä^²ÎCützRœkFé©JWM:EÌŸî]vûâge)E›¯8Jyç}¥4¦Nò|pQª››™ºúfHÆ ý^%þ}(t>™öÁuV#gpvK¿ÞÄèàXAwY|è£ûѵÛÈè8=(N vóG³Ç• {–Jß#Ò± ñÌê²õ°\`€ªÏ¸—òB¢Ù~'çÓÇÅ1,cU!ë+¾lݧÏü—íhr‡Ü‘M`FI`) ïU)n$iÃ>XÑ F["s|kwÙõ1 PZ½~èJr&ã­âÒšÔ6ƒ­©ü|×ä´Ä†Ä*’ܵëêhÎvÓw×@šxOÇ&0++*Qg|*ùcþîBŠwÁFÎÊg;Ðþ&öFo†HÒã÷Åyâ€{¬ÍXbZ/l÷á`ØÇÊ1÷ÛÆi “±^hÂK¡7ÍlóU mó¸¿ ’e”¥ÊzÇgË®ZÅl 3‡Ý<3ý—~CY!¬’ü”IœÇÞÓò"”·¯BçÜÓÄwCdÎ]ÚA­Žý*ÇrønÏ’g5³úÅKÁ¯6Übí¥÷4 r¾G/½Ü¥þ!ö™c)›û„›áž*ÎtïðR¾*og)³ž2çI»'Ά_Ï»·õ†vw™äøU܉]:!É ,ó6½G2Õ“ÈÎ N ÙÀÉp´€¸ÿVuˆôTêîÔQcð씢G™¯Qª«¬>ý:¨ùr¶»P¿‰äƒ¬,¶BG“|õŸ-nÞÕàg¥Ÿ§{¹àcÝæ»C5F„@-Êœ¢êžîÞ°0yí·z q²:>ém,² Û2ä7]Á‡ÿžÏ¯xa ›ò~ å·›c0‹û6»šÇ3Âí=\-z1±¡– ]±£²™âÅKÜfdÑ®¶…9¬%k‚_÷ Å“FžAóZ§ô‰|òON„–Vù .á‡Ïl‡Le¹–:žoÝ3ƒ$©ðÏ€¾.[dÜ^L3K®‘ ´Œó­3Ñ' X›Àk]2ȇ×K-º¢y†‘”Ôl C¯¦7ê“‚#·#I6‰qg)µ~m§9igýV þ)û-)±€§ÿ3JÏšÿãÄ4ÄB‡-Üï„­Ô²#— šYbîÖöUóþŒø [tGä­‹ÈGÐ&x ¥Z`ÄÕ7rÓA¥]–&æ¡€<_·¾RµMÄÉ"Å B/ìz¤"=ß>ÄÀ™wâCq3]ºÔC ¶ap!lí&(ÂÌMÇÝÔëGøÚÖ„T»üV-ÏGßx@bï-F¬25hã3u±U©ûÞ rdÎu“#Ú:h ¯o[Ã@xÞ/OžÔ†f4ÈÙÛ>99Ò–OiÚÀòÀ¿ˆŒÞàYGyÓÈ VÑA"ì^¾vÉsbûøÓ›Wk„yW×"±å·iËî-¿’—½3(Òzª´FŬêù}ðž«ëN™QŠÊøJ e„ªÑe»‰ bfâT”³ÚXÊhê™5ÛÜ÷½ú^XRŒÊe·D£¸–2#Ѳ˜ãœy‰¯_Fb\TJÑu¡Ý€>í1Žtô“˜ÍmL>¨¤»cšÚ[|.(õQß †ÒæC¼¨a¼Ÿ•i¾ \ 5WÜ ìŽ·§3²…¯UmÛýØr£ètŸki£Œ€½ M½•ª·Zݧ™áòh •q)ðÚ#WŠðæÚǾMÒâáêw}~Iq$ü’nJ»”Ø¥{Žšé5î{ç9ðo`E]A¼}³æ;#kÂÆ+9 †ž½O3"]Ë%ŒÃv[E¤EÚ.U„ŸŸÒöNU[f©[<çrfy9¥Ÿ/ÏUÔ\ðÖ ]Fά†bæV‚i&Ɉ·? 9…ÿqôåí5x»Ÿ"p+ø&•Xá†$wiCå øðšeØ~2ù@ŸYaµM¾O6«aüs{g¼Ed×Ζõ!înRTê0¶ø@Q×öÑ©¢ëm# .4ÝøZ[Êæí >²›šQåJžvqõ5¿ëHºŒ#®Ž^–âW¼ŸIõäA>GI¤eÏéæìÄ7š9Fø’§ï‚’Ïï~(õFº¼Ë¨à ö3 ˋڲË©pÔáz!_‰ÆÔµ?'û´ù§)¡'$Üz:B‹ Î(ÕhA ­<<X‡ [ÉK½¹w¶pDÆH¤9½ßöW¤JC5ºK”XÿÝvmÉ.¬IåiFBÅY‹ ,§¸¯ 0ùx[IÞCÚ1zø¿;òÝæõÅìh¿ ÷çg‰'©æÕçC¶yç²Æ<}jK˜Y¼2/séé‰r#=ëæ)ðA<a-QžVbz¯ø†|døáËr¯Úw½î¬+ZÇ•jf‡¦!e¶¸&[F°Ç#%ó}·¤?*‡¬Ëü?t-ÒTqÍ=`Aäã N%lÕs BâíêFdÞ·¬§«”€\çŠüðV¸c%ØkúBßg˜Ž“Êü+ÝNkügkòÝHT=N“/D“Ae¦.È aEr·åÖå#P‰CÓU>µ_ÕÚ¸z=b°¹ECObyÃÛJMr”ÎξvÅÂ÷ `²|¼¹’M)¾ë`‰ímõ¶ÐF¡¬l7 }L”«<»Òv\Pï Çæ˜c_;S›UÿíQÖÔ :}”ÖüEêm¥*«ÍÇq#‹±Vü`…•êeðϾTzF„#ºæ¹Y¸à°¹Ì€ÖÍëÁvâÝK‚ñ ]wdí^ãùtnwUcQ7@Ç håʆ†ÓËÿ¹~'@ä  È™…9d3d0n‚?߈Wð „2_.uš2çí&kÀ§¥ÎÏy‘̦`ªÖ§spÛ•ÂXW_tÂ*ËÈÃ⤊NÀEóþMÆ]Å5M€\WéÇÌJ£Ñ¡ˆÆý‚”‰ ¶6†7 íVAÜÇÅŸ0˜¤a1‚ÔSEr»5“ô=ÈÅ=(½f)ÓN€òÓ­¨²TS¼¬Òe7ke8 ŒLŽ\Þëß'¤*q¾FhžÆ.A ñ×’÷ʽڞ«ýo×¾cÕþ®Õ?«±Û=CCF´ ÁJžÅöBæeI-øôÀŠR€§v ¦U[À¦lHMýWMÉ}ä‚c&ö‹ Ýf7îNÚi""‰êe‘#¬Œ"–­Ô&›¢®8¢¦­–=ë®ß_4·Ìa?Q©è¥xÓêWˆƒØä„Pæsx¾†Ä$ ¢x_0tˆ˜ÖÈ­E\XQkœ®H@2,«/:Ê]ÈäžÞ…7ß-öyŽŒùü~öê Ûú@TÞ×döâ.÷¿2j1ÔS.ÒËcÔ&UÚG ¢§§ߟnaQ…ý¼Âd°†±=Ó/k,Ä|©œ¾çŒ"×4²¿2JQ¥Š™º§´¤:¿ïN¶ÝéY°`M7JEAÒoO€üë!&oýŒLDÞÎÈÏ‚ÃBx&¶/Âi}&׿’„ÔÙk’qv„ØÛýC•;¦çµÈ’ã­þ‹°,ê¸ÉOÒ“9~k%ëßôXŠ)›œ©CúÅË‘dšŽ¼Ž°Å6Åí¦Tσ%Ì#„â!9z;ÝÝ!…ußí(@VÏeµÛìXA‘lwR-4…ø˜# „ ¬Õ¤‰è±§B­ø%W?Ì©Œ¬@dµL)ÃŒ?÷& ?“ÕRX+ÄŠ*›P% ÌXŠC6] äèÒ½£îÂïxôFÏšéú^ê#YmS*ÞÀód°j&kK…'C½|¸¹¯,@¢ºk·Ò]Ä'ءǒÙŸ1ʺ ¢¥lQÐ Þ„Vž‹»MCšÉ{0kxKÛú_U:6rzæÔå'ìRÍüɉï!û`ø;’æ«Í@à ,·}éú™a:¾\9ðbñ¢žtÔŸ¡Þ´lmvÏÿy½Ud@ÊKD^|ôÅLüIOÿK÷é )Î×€­\›#¸éÐç‚uï·Çz%‰ÁEïRDãÚëë§Ågá©tì=³IlÆ–àüàçn_ص6%äð¨ˆä'8›¹õ-Š$jÇD¦UZÂæº,]xïHkD°N}(ø—9cúé,žM†£ÞfaG;²ìÖ$°5»ßæ©9&´U…¶ ÛHÑ5 É•A1£óÅÞÑî& õltæ"@B¬ø;R0b"ó •Óª^^`¥#Ð齕ì^bê½<Ò=T}»¬§;BCὡˆ…¼õ–øëâ~W nJ?`AXŒrGÏ1Iö%úiËÛWž[â í±‡Íœ€ce[Lý¯uN nöÂsœÎèá0ª¯ 4Ö’dÁYÞõIpþÎÍqž`\œÒÀRsë§?â6º/˜ABµoiWVÒÒ’è›hдÖ4ÉÀ!?÷9 Ý­,£ƒá\Ý!8œÑ­£û(fâ¶Ýf¯_ߖ˵,LvÿZ g‹y±´u´cu…žFŽ[³e—ÖŒ'XV Ií µ²]$´nÿä(fCÇÕ2Aƒ¥JÖï2ã)Œl§x AÍ^µÉö‡7Ðú<´6ùñë8#ΆF ì9 |ÖblKYeAxÙ`1†{?'¬kêI_F“PŽ «˽ß'Ÿô‚ª¤JÂqóyȸt„ É9{†Þ[ÉR*mjæþvvqßç\ì¶Á9û…rw¤á’&s_Àúå6¢#ÊNöþ–1ZKa•ŸåJw:FÖ—å~j%xOŸhXŒ7hŒ{%rOPר¼UHgRçX¥º|Ïšy0ùÅOñƒg„×=|qÏ”y7{p¦âû¨>;Êt’ ¸Cy÷VÈK0ê¶óQÕa'i÷jAo¾;‡>"ò,d Ú†K1'ÕÖÁ?®èøÚf`XE!q'©#ÔuR6 “’‘<,M‚’…N×~|¸¯{äÞn¦ÏÒøŸ¢r0¯qaöìñPpØU¤Ÿ*Vã YP³ç{2Wõ:™Û®úéßrffëëÄŽ]Ì´»ò ØÛ¡ÂRxQYA€òåŒc,•ׄóÍ艈Á)´Qv•sOtJÁ ÔM¦ÚjÀmvIêTY9•Ö‘K.§ò°ÕV(Ârfâ=†EU/NÝ”iÊÖêÛÍù³"œ7§oi­Â¬ƒœHïÁ?ŸºF¬EuøÓÚ¡_Ç™ËGvGc§IZ¹Mëä'•¢‰Û üJUGåõmT®¼pÐ✅]öYa’–ó©tmæ©´ºÝói#YÉÈò/ÂÒrI;¶Y”Ð.tŠ®M÷VìÃYÊiÝ©©ˆã{UcÌû\%cñVÊZÐ 5ökŠ>ž¤æd}ño)|IÏ×ÑçØû±±ý€Ï¦Çã^³Â1 ˜§OwûúG´Ïü5d‘ö˜;•Kõ€Ü5ÆâCŸ!yÜ;3dƒ¿¶¦z‰L4`Z#ˆ™“GR¸tºü&ù¨8°.ÿSi˜ê‚Œ3š­Ê8OˆE`+6;¥ øwM¿m|Ð’b±ÀZP·×±>uzõëZžÌG¥‘/‚,ÁMnB— ù‘Ì9¹œZäríT™i79Éxa ²ÚdgF Ún!ÕªáWM¤{µîI‚ˆ‹`"]‚†R»ßk`H”9¬AšzKæXÿ—hqÅ­NƒbhN‘%ýøóDÌ⬊²`ð‘s ·òX¾ ô?±Æt63(æÉtåyùH¨¹SCÍÒ+u€eØpìZ'»_VÎáÖ¶¦¶ÎjŒM ’¤Ø§n~ì$Ú'qª™`½f*>È\ᥤŠ!ÈIC[a6_W'†øK™Ó\ÜU4§|s‹k8¢¢“µF‰¢cBú¿¸ÒLž‘9£àÍ'%¶^ ã§b%«íÇâ`º)½ĺZÂ2 •…ÉhâIAÝ®û ë<v\——OT½dµ 8 ¦§$Ý®ï¯Ëº†ðœTœäˆdT™¡ÞûÍó+‡µjÒ¢¹z'Æ[ƒŒzyÛÜmÅ‘º÷¡~‚‰zAðSÇi¯w3–V-Y,±†öíØxð3ÐãVŒ¼W_V?–\¹µ¹ÙVŒt €Ê8¥_nnk¼zòÎ-b(“„CµEë—Ò/Éç²¥Õ!ÿ;u=+uµbeA5óMëÇ´¹¯ñ‘éÙ¢xº·‰ùjlÉƒŽ‘aOá™›–Õ¦úAêe8â®2MBr‚fY °«¿)Wÿ)r“À†yŠé_‚@ÐTDŠâêúL®]f½—Ã0o:Ú‹ÕìƒX.;åò"7õažkò‡qL­²×Á3*;P?Ñ#‡*AšÉ¯ÁÚæýÉ{Š£¡‚‹ë<žUöš‰yÖҙ˧ 80‡BDVî~À\=š›î‚ ‘–o7ûŒªòt ôt ©p×?_ùªL@BK:»R(ÖU­§øÔ´å»fR¼Y¢­"³/˜¢Úc`d~ë‚2DÒã¨a Æ8s……!чe€øÿÊ~au³!uÌtÀzú$Ö=ûö„Œ ²ý›DUóé78„^‚bWª:ާÀhñµ¼âTp— ާxmA㌿ü:y w§H  Øž ¯~;W²ŽÍ5g+äd$Mˆ7÷ÂNEç~54KÜy‹øÂYÛ4àAÙ¼BÓa¾Â«™íø1éN?:ÉÖÜ…RÍÒäEÔŸÏc¾n"‚ŠnúQSOÙ&ÏëÖÐ'f+æ¢ãi ¬lCeÙ #Y] Mßåq§¥¡ºß—=¾â<´íṯeÛ\èÔ®ì¥Þqxã‡Ä ü‘†P>Ý™`™ä‚½¦eX1q¤û£«žOYë÷2Œ‰û*+g"ÍéŸèèæ®óU“9Ï¥Wj¤d¡9DñבúÍ.;),%N¨ ˜„§Ÿ|ZÆ\ÚzGÈÞÊ1³+=hÌrPš-TÎ8ï8²¨Ë{£wƒßCàð·½²üêQºEöÎT£ð"-¾ÈßË–ÜG ðR£½šˆø€uÞеæ<ÃAd1ÿꀊ³ü ^£ºF…Q-Ø7s¦dÔÖ9üMNx=«ÒràÃ3žGÀ¿·Åêìö’xe±l¹é¸*8”_‹r±sJŠªåßÃaLäQš+ày¼¬¿yø¨˜]xjÝd1ŠÝ‘eÇ¥Â&e±׆Í7´‡íÁõ™:ybWŠqZ1Lïó–oàÏ1^*7¹}м„u•c˜ ïw§ÏøMT쥼‡5ÝÃK`á‘<‰úæKÁ”I§&6Ç/å_à!ù€I04¿L»ÇÊrôÈçû ðíÄq¶ QAàdT÷D(‡þP”åä%EÝsf¬ïÜ7óÓ½N’´1°±d†¿gi¦áì´ t­xмÛÝ-=%¡(Lœ€ñÖt¬¨‰™sç´?c½½{ve;DÚ˜^VïI}½^1…OëøÖ›‚i®ŠÞ¸€™d=¡³+XK=Ž\y+ÒÅAM/¦¯&\þî¯Ñú mÔ.ŠÝGèÍN„×c}Î]zl7¿læ,º^²•Õ'Ï Ç»ZÇ>éÕO¸¿á4¿_L°ý½Š(DöÝö+¯€Ñ©‰ ¬N5­¨Ïœ1ÃÑœŒá™àéı¥„Läò´Sƒì!E·ŒÑQÛ„·»Ø£vøg/˜´Éþ’4×^Ú§ ã$)'…“†ƒ/·Ñþ´/° uã»A>Íb€ö¬²v™#ÕÅÚ;ú$C Gg©«žžÜGt-Öí065 £ÎŠ£o±¦\7Ñ$r . m~3>'ê›%л3>+¿á/}=G àÔ ‡åЙ a1­O¾ðGʲ†¡[—Èú=ݵæ^Ú̇“Ê~¥Gç8‚¿Ìj&yòPKhûDÉÈfvìB~ cÝpÚ)ì½®`Ë&J› e§ïé±xµ¤@Ò…—‡ì·GéÂ.Ö6ˆ›-àUm !1çQ÷håòÐA‹uŠ6¯-RêÜÖŸ‹Bo&$ðm­ˆAM¦d:1+V¼Ìn¼ e?|Š4exyØd×L{Ãßä4yڒϨ“!nMªAxh½S˜/&3¡¤ôóP˜¹³Új¿a¢-ÛZ0Ñ-\øRfçr«•Zîñͨ›^øÝ¶Š,¼ž¡…© ç=)'œô ƪ Ç[X¼œúÖ5 ÕAE|vÿà» ^tgÉÚ>¹zÕ?MD6œ¬ñEK@ΦF|‘~þ’•Y¿Ú{²ÌÝx»åKÈ- }%‘9Јâ·rªO9Å¿v(a‘2<Î?ü]ý­VNHØûúõX_O©:¿}òí>Ç’ý²G}n|Í–ü|åŠíCáØ äƒ(㯠zI--®Ê/ÍÈ ~ ™|Wˆl"L9ì³J€ŠŠ5ÔQV™,'¬‡.HžÀI ¨ßÁQ«°Àßwed¥ðK–î°Œ÷„u2C”0nMç‡^v=êÁ ô6P@4™¹ÄÒãÙÆcNÞKxйl™+#îýV<£Ê òšlè5ì½SÉy^QáDë\õ…¹æ º^·üHmäܽ·ˆç¸×Ó÷¼Ü…5îÚüOo«|I9T,p|U§?ÐXÝÿpDmÿõ(êþ™uûö‘ØmmZVG#8s⪙Ö,µd;$®4~×ûînTÕ<I†ue‡êæÊâ˜ýôëIž(Qk¿ƒúÙm¹s‚såÁ×%ÕïïÍM„е\ußÖa–#Ûežؾ؃i‚îÌKã+bÍ ¥DÃ9s¯dã®{™i^[ò'_ƒ/$š3Ñ–µ¾O–»šp5n.ùù‘´JpJÊ”(×. jŽP•ê `;…JF>ñµn˜˜Â|G¯ÀÚRŸ¯7ƒßܸá´Q&} ²yFåÀÀp¬}5 × ŽJݾ?æînÂxÎÑÕó‘|auþç¬ý&¾W%Rÿi¤QÌ0³# ¥¥Ñ‘CNY&6Tp©•6KÌÎ?·'zîëÕzäy.ª=Hcå! ÜävÏ+¾ýudÔÎ÷#VHåZGßs¸Š+cÑc5~Ü*–ͺñõ…HŒ’ïË*9_JóÛϹ×ð€lî$ßµÖÙ½*Š•ºtýº†žµ¡Ñá²ù¤ÞÆH`ccCY¿u©X¼”fÁj„ÀçþÂ[ÐK€rŸ%•¼L³Ã2’«2ž•NW™ P~«3rÖ4k‘ìan:š+UüãCEƒØÀ2–òÀNØvPYÒ“®}îö ÜÈydñ\‹‡ª³_LBeÅ*z¥c3a&[¸aºLi˜‚ç© îÞÆØùtV¦FÍ!ƒ\?ŸNëþõÇM¥æÂ{ cÌc±d`E¥ôåM¡¸·TM6HTµd8giVïró%‚4‘/ôÜY¨‘§âlÙZ…†è„: ­W‚âJ¡ç§‰D‰RöhÀá<óPÁD&õUºwÔ—«:\T8>×D …´N©Q<+¸òú:Åš5I:‘¨ðÂê®þöä'rø3\í•¢D^³Ý‘ÓÁ -Ÿ¯-W}܈$Ü—7§Æ´ƒÝ×ÝB®x—Ñ úÆ#á;¥À:ÿÀß‚·oVóÃÉíð?sͦpToP³Œí½t‹Ù”gIÇáQëç¦Í‹9?2fqØ£ÂQ¿¬ì+çþFzÒàI¬Eý $(µÆKNMéÈí£·EèÄ`a(s<—¢O]Sð(Ò5[Zs \kAÏôŒ?`ç5U³ƒîŸ´/ž©Ä?Qÿ+”Âi )!m¹ë°"ä¦?LzBGÿNª’ŒƒÇ«S¢Ÿ´4bs&séžf^Æ Üºý}7m+Àsf>Iiü+ÌsËV‡u±Ý™%—Ç—ÄÅßäZ3i°ÛãÍQ@·(M=ztØÌ#àÂÓ}¤yç€FÏ{ILMWKæ¯æ¸ÀYÄ€‹übñôÃéÐT™¡ù=r?Šíj|ªØµ@f¾àʘ|ÔÚ…ôF'}ê’I“:Ðã/7aörîF&”÷€ÙÓYµ`4”V ²ïÚ›ê,}Y¾ú+ÓÿÆ£7[óì™´ÜŸ€M“š%w˳5¾.3|MäÍQ(§Ë{¨u¶ôµÙ| †ë î‰ýÙ+#K+@†˜G6™üZÙAd «óZâÜXÊó+»¤mš¤m)hߑưÅÓ ªŸ ï¤Ìqs·7Jn üöEªÿ>ªˆ÷|h¼ÔTyUŒSSнð[Ä»r“V=¯d³ý}êø”²q€UU}#IK¨ä8î"9S²g|ZW– âË‹;U½µ9Á;t«N± ûžßÓ`[Û(¦ÓáörnE²&æÖ$Ëò(vDlú`5èÜÐ(T Ì·:Ã÷€$-kGÏ®… %•`Ê¥ßÐ2àÎBÏÖ­‡sÖƒïŸ×Ë2Õ·@j°ÚzÌaG'ñ¿øj](`:Bz´wfE_R‹|å»ñ²ÖC Ò'Ï!•“ïìTP ý†]Ì¡Œ‰Û<ÅúD©þqù=Uó"¼„unjh%¹w¤kˆþ¡®üYO¸bþ©á’–ÂyFÒãòD$l\|Bn‡&‡ÐZÌ6ªïV‡Sþ(q‘ªH¾pÓý¹SH²z–¶ÈbízÀu®–#TLòŃ÷hÆz¨ í£ÎòÝÌ©šÒÝ-ô*ÜùM–b}Á2ËÙ@ôï›ñœ«.L?WƒF¶ÆgÖ-À®×úÚ¦¶(ŒVÛ’øï¤žÀ¬Õjúo­£æDï{yàu³°ß¨z!Û숹æ\îÐß%œ†°>¶;€öñ‹¥…d²ý‘Æ{ É«' â}qè óé‘õcvèçyVy"}rÌÝ)Dìö¯(ôc؆ѱn p„.²M‡W#iÅi­|¯A%AÜ/œ¦¥m­'2æ[Gm­‹‡Eþ@ü&kðºècŸdÜââ[žþðñ‹¯"È7 šm#Öã™X3 Q¿ã&Ù:_”=æ¸Ë›¬ê2—dñrB댗´.­ Ýz¯ÿm³„¨2÷Â¥.çìt35tîràúÃq¥uÜXÝ‚þ;uÿ"7˜F­©Ý.dÿ¦¿Ó6ø©ˆy!¾è¼çì¡Ô¨SÔ KŽ\Æ) þUÐr©q®w/Å$Œ‚edƒ)í¼ ×Þã8òCcq ½Õûøí“z±!]?”«V-æß{ÈHg¼ˆî‹ËòV´=Bã´Z ÿÜ|4—O~F{R&ª›"¼Ç˜üéÊ(Òÿ)ò ÌîPý–/Üb'ѹÖ0œ_vÅ$ÕU {™Ÿ ˜«'²ÚMÅ*×4öŠ«² É$êK…ŠSH4â'èè²2¾T‹ðƒý✺­©pˆÉ3¶PÕüwº(&«ÁÔFó9ÐÔl½ÿ…vG¥~´b#vEÏÚ,ö®EV~!Aé—DHϘHÇ­£ùòd5µúBÞéÕÝ[ÖÁ×L‹“WÕ¡²MV›]#'‹`zíù C¶ U'ýŸŸ%Ãïï^£ŠKþ‚¿@@ÂX{r¸¶—¸µˆùë¥é‹“ñoMÝ*Û…d!òiøÝb5fçäG’Ëý53¹SX\FŸW­–“XÝ—£GDÐÿ coÍ Mºã$ÇåG:Ф•Z{r:‘Þ ÄþabWûhoQêBéŽ7““;E²ã*JšY—ÄFˆçrIšJþÃúÔÅÕ›š‘>|ÔøÅ?–‘Eô—/ËábÛŒnŒˆAX•Š”¾¸{Ç»ù­¤>äÒºVevì}ìÑŒqOf—­ÂRÔÄ>"¾I-|‘µ<˜!Ð:Èò 1êg.­™4¶òì m‘Áhq%ÜÊ+ÖD”¨èº¬#6Ràê2.«êu‘º]ÔuHz} ï5$S±¡&5J‰¥»%ÀpdÓŒîNÄûŸî9I54ÌS*4ÿôÔÙ]Bõì6Gj«ý›ùÁØè br?í}€ ‰îÅ0út–Þ¯m=(Í{^5r@ ÑgÉ&c„´¹½NÔRg6òŒ!®hà ³ù˜RÚÚ¥¸/ñ°vÏ÷:¼Nß=nNltÍ‹y7E—eeÔ£E£üa•5=¥²ßåqžõ9eZÿ”#˜½œ<»Èu&ô­~è[i霚š¨1ªöÒúóoY˜1„Ù–°œ`ÌDjºŒ©WdX"±òFéÊPÙ¥çΟ(^pS¡ZCd- ¨Ñ)eÕ:+¬šZ˜ôéBåWå:$¥‰Ï¼¢ilòfºŒvU2 AkD"´ 1yNû!7œ©šPhÍ Ê#3ðŠ¡Š1­k}ª´r}\s¨wùLí0 k°Ícú¥d0Åb\l¾wU,(h‡ OyµˆÒŒx+…“ÑÄÿÑ„‡Ôö©ä7AÖÖ‘:ê¨ ÝÙjÒŽá{ˆª<9 r˜mŠ5Ç8˜6­Î’ޝ®q´8ßáë_Ë샚Ã^¾1h Jî A¥ tW¦¢™æzK þ5HÄ¥²BTo %aå“é­ƒgz˜,‘ì{IDñ‘ÒñâÝ%{Á².y‰%4%0 ñ0 ³›Ð—Ž«ÿ Ð(–It¶7™ñåýB™YŽ…«äPàÅwënÕ$UL Ôîbn]}* ƒ©T>žÔÍûľHì-¦è;·dn!†ãë¹¥™˜©âpJø|qÀõåòƒp¤€Ã5´àÛ 2Gâ߆_¤éÜêìHžp†aJ¬âxtY¢‘¦ —9Ųg~>ø·Œã*¡w:ʦk¾$7?CF½§ÛÆ$ÐýeX/ïûz›;驃Yä°×%þûëÂÖ³~^ûjrà5¦:ß•*m×µ© }û Èç«•[$´,Y¶°´Ö# *-³Æ“³Æ›Z­Т¨ŒG ){!=)&À¸Owñ²ˆ–žÅ]MDßxD>$2)\ùÙr`,œŠ‡ƒIóâAä‰Å…ÖX²+v{V«`ã\m<_…u$ÔhoMC¶×¨t­ïû+¼f/AqD†a]—ɬñÐÉWÚLŽÙ„ôAfï$ G¶±ªÇûÕÃŽ>MŽMÊè{š× cê¡!xd‡!ì»`ÞU¤sÛL®b]‹Øâs×)ÇÉH5°8PþRøµŸ¨"¡¢§0ø_+!wšem^æ}’EfîËȰ瞓…ÞiÊÕ‡S¦‚úh’æT,Æedåh´%”.ËŠ–YÇlú鈸͠rÞÙU…ó g„˜¡[ùÀH‘žS¤»ôS…ÙKRü)üˆÓSÕOQ\ÀUÖÜ@%òHØ- ^^Ý©à—ªéñÖŸ/tÌ…ÕW=žáÚ,±Ç©pƒQÔ\¿dIUïþºì E§y_#åð?§íïm÷èÔéÌdŠD ÛL©Ø-R±Á6û!ÓaNêMo¨‰Ê‚gÊ(oöËÍ“ÍRw!ë•¡MšyÔ%»cdÞª‘*ª["ð‹Åç·2 q[½Šc’‰B¡žQè føŽ·)qÿ<+g_ñѴàçf2œø^Ë}¡T.Èyí=´È§^K/-㩾‘|0Êú'Û¹: OáLƒYÇA¡+i"òô¿A{H®˜WÚWô(”ȧøY}ǧŒ¹OnÓÅm¬â/Cz}†Öû‰Z5]¶Z)Eâ}8‡:‰ÖC°.8e2!OŠ´fÅg¤Î7 ŒtÅ.NÙv$éU°·Ÿn€-¸A¶-J¬2¯¡ ô`Ù¯4¶~ÉÕìÙ4äzŸà€ƒ¼ã•D ³„–7 Œ¹ îR›¦ê°ï{3øË߈ãÑcÊáÞèÝGwÅÅD)gäÈÙZÀé·L»í$ØËŒÍ?¯R¼hÉ¢üæC7htA Œ9ïo”¡ž”˜túÐqkƒ&1(öqOÄ@Ò‡ĸ<¹ p2ˆ°C#Og„ˆæs[‘ß0†Ýo9èX¸• hvËFp¢ïÓl7zƒæ1E¦gEEl¤(C4 eX£àšä²<ÿ”̼Ä3¼‘DDuxßòÁ¸òY¤u2ÒnŽnèòðÌ92Üž¡œâž.—ÎÀC Dr“PÑœ ì¿ã® Ö}^ü){BÈõ¸{EÓJ¿EvW˜ûÊ”Úe{½2/)Ì9ªd‡”&õì0ñÅX*“øœ{ÄwiùotÌR/X˜ì¤âÌ~‰‡h¢wr¦>‘t€øY›º½o¥öîꢡïQ®NÏ{†á†¢Cw}¶&úògÌ9úUMÅ¥‰€l„ÕÓo_â:v+ÿ‘éN'’òðé¹íš,>}8ª8†Éý’¦ŸÄò¯\]´™õ×S„û bá¦ñÐw:uôg„ɲ]²ì»àH² Tjõw ÏÚ3@°Ù©b!0öK;øS¾I¬¾qd;ê ‚Q“e wÑäè1%Bn6'b`ë¬òQœDš~:1-ÌÓµi ®@°*e²<eÁgg¡÷…órÍZMeFéZüœõ¶®)ÒÐO, éßSì8ÌE‹ïµÞݪ_wybh£hîÍ]8œûýÑ0qRLy½6åøgÄÿm/’UOm&Œt“ƈÿÎ}é$°Q½à6 Ö‰¹}“5,Ë]óoðž02³„rÚ3)–ÆQ÷5]Àÿ†¨ Ø ZN ‘à Î: â;ߦb )9Ïb—?OàAñ{”}3˜íRÑ÷è0½·ÄÝ¥Š›)Ýü:}õu´‰w¿O­ òñV›‡XGæÑÕ*$Vìºt:á–@ÌêA= M›¡)h% fÿl 6Ì 2¤g’¤XqÒüÔt;ÄÝg•"ü¦-²ôƒ‰wÖ6ÔmÔ(›"1E¿ÝƒðQelüAfLÕõy?ä*n1߉ÿù—ë o–¿~“]Á÷êÀ ŠHŽðbÓ˜'G'‚:¡˜sRdÝ‚ó+ IQ®‚uÞ}ì×M[*Ìx<ІLëtÿF߃'÷j—…'ìloÇ* ˆÍÇcÐ]rC‡/cß‚š¡±[o…ÁŽy¯ö‘žš†NÆ%Ò•«‡ôÁ„D}UÚy‡9˜@z“fß%ûÑ;ôâ"}dn .Ÿ^¯NŠÑ®2ç¸ç¯M+›äL΂?›íNÇù·3‹”+:SYÝ‚ƒÂW¿±Í«™¾‘¾…ãÕÏ·Þƒ®WÂhÚfnÈ[DmL}$ùø -?*VB*óÓ[àAVfÞ»‡¨:0Š æ¯4cº‡x !Ù‚†×Dúw=JÏc3s’ÉhÉþÑœ‚Ò{€(³É‰0'sÅ—_ìMŽÐ´‚–³æ‹p; œXõ®ØBñô9ýž{¯½ý·n«*x¥ãÌEœ¢%¾rÏYÕlj•Üþ2 f!ù£||ÓpXÀ÷pFF8”ìÉÎ6>Û/ØÉç}µ«¡ÉTx ÷ÆÚ&D£2ð©ÊXŽ2N¸¿-V`ÑY ÃüøOW÷|l¼ûëe¬Á߆0L±ÍE' Þôzv0ÏîÇõ£î6 õIÐlèŒÐôF"¯£'Æ>—s¸Oo™¦×Gÿ —¤§âO°µ §~0šOØß|O³Óä’q«_礋‡Ïd¨ÀR§6Ü%iå]`Ë1Ûv^¾8æ{1k6qWmŠƒxó)°Ü¿*v°zzL”ÒQà5Ö=£*ÈÞùžðg—³`±\Å^m_RŽÁ¯¤ÿd{Ç5iæa›{ Å0}åTé{eÛTýXšuQÑc£º'~òñgeeúD]})ô½–~Ícï6ÓûÍÖ¸WOÙ}á'lµØø+ Ùh`͇4'”oʇǘ%01c³nüÅýE‚õÛË=ú³†›‘(<±¨D)_*ËðòñK·í2[ÈSךxÖ­!o«f ‘¥é•O<@aÍt*¸vûÃaÖwŸÞœsc[¢©XhAQr*˜$87µÜE °0?5$ð† šii 1Õûª«"ìFð}f7ü²E—vx笛¦r†¨@(/õ]L‡Á‘ Ù ¹ ÑõŒlRP*M©í"Ö亚y¼Q耇œÎÖMåléHô} FKð&µ¸yÚª¸Fµ!•åÓ ¿´ZY ÚóØNäüY_Ø— ˆÉñ˜õèV«ôôÆÁWæ™àȆ+Æ#8ž-Ê!þ58azgçC¥þ¨Ä´6nó¼\ø›yÜù×¹iÈ~ëør^¤KÜ¡­æ)Ã.¢Cª‹\Ò×Pv-5A5Ñ}BÙáÒ¸?×sž™R‘íy–¦U+àlô'—"dXAý)s‚z°žéîs6ÕVJº“$· !YÃ,Må¤P2I˜;f¸|²n‹z‚ …°Íi>x$”î„1 ÖïxJh€J4/ŽX"‹«l+oGxy¢ÓÅbVH$…ñcœH/bH Ä6`$3¨ë Ml?5üŽÎb÷(¹^IlºŠg¯Sû ÿ™ôk d¼¯Á¨|÷óì¶Uɺ˜‹½x뗅ßàÆykHB&¯*ÃañŽïlœù‹»àÃÕâ”Óµa .ØMR¯döÆ=Ã]ÕJ¥v×ëB^¿½¸>°õ—ýx‰ÅD|꺰ƒºi(S—’>‡(òU€sᔽì¦x6\¼ qI`[À`í îâù„Y§\ˆ +ÛÒÒG¾Šƒ€™ðúšç$Ù`ÆY¤9êd!„>¸ÃêLJGí|l‹Þò1@‹ýüøÚÒºrm{TÇgflY¥“ßÈ"zZ)œB¦w.G:ÇX±ËŠͺNª¬€rû0àF%ƘŠ]]¦?Ò]úÎ_mr¡ý›ŒufNî ½{\ ,ïäU±oÑI˜)wät#p¦=Ë·ßÅC&PûEý¥€¨~£mP<ö )ιDp€„E,#]©ÇbWQe¥t«pA¯–¸åËi©”ϨÔí³‘ÃôÔë°Kw¹Ý†7¹.â™5]Oüއ¥´ügC Oô:½ ‹ú~—ßbÚaÙ2;ÀÏ(ÈñØ¡KÞ?ÀüÎ*Cês°ÁÎÏ|¢5Zûp-שYCËö°¦˜épW†Uà®t+.›Ê *{˜„À´É­m ˶+7"n?=êYSÃN¹01ºyy¼B ¹'1ºAûІCÌv¯j÷tõÂ}ø`b†Dò½[ *›lÍ!ÿôGCßÌ»6 @Õ –õV’ÎÝZíHa ý²Jcs!àEVéÀK¢í’k¥åÀqBñÜùÕFp‡é(øÉ—2ç)£†ž>ð»§* òÈ2>o§V† IÏÛ]ˆ¦ïͶ£àŒa}­ò!ÖnW>ÍÉ…)• O'GEÿM¥™ãðXë¬Ú œš#´e²ŸvÛw²ˆzh<_fBWøè*Ô@éºÁÓI‡õ27×Áù@{$!à[,Ÿ¸|\)µ÷Vvs:¿Ì8ÉôÄÑåltq ‘RùSn%š·­iÚÅlÚ‘l'8›t¥¹ïÛiÙw¤WŠ›}ÊVÄBL\'Íž©Wm5’X ˜ƒüEr€ÑåÞˆâ5šEÏ gë$ë˜6ÉÛ«iœ$L¨òWô´@s<¢Gì*Øz±pEVû¯º‘ÅB¶ìXè?úgÜ£i0曯À…cʹӇLüå.ÅCwEÍj„~Ðx­ öiHýä§'©Ðå…TvyzöØbÎùL µBß¶i‰ç&UN¦O+—m°ïþÇ`2žÚMm‘àŒ–”p‹p´ø°âbÜŒX¼‘²óÒ4áòç_QïÑã„ê¥aÁQ uª®¥¿i—L›j• ^ü.Î÷?\Ã@ÖžâÙÙÈYin0Boõ—Öœåwn¥òú‘UÀ¸.bÚ\ê¼Ô¨æ‡iÑêÉß¼—ÛÝAàmŽ Ã M=Û cãÔí'.ÎáWE˜©™(œ{ÝAŒ„4°F*½ÿ–º¢üøRGÅŸ1æ3Ä´ÅKa¬Oúá£& 7æ¬}¸,~ÃŽ±!Za Ñb'-¿™¯®`ø½~ŸA0¿êÂŒM­ŸÛÿ±žþ ,_ø…îšuAåoðD¹ã°üáÕ&3eDcÚ7Í_P»Ô7ŠÔ¨ØðÅ¿‹Ç vúe¯U35‚RÄš_Dåy¡„âŒõc`CŽKä@¬G‰Š^bVS*;Qžf±÷ÍY“xõNÔ]•Iy‹+}ÁZÛ±ö\¬Ùh*AêáEqé ž&0*Cqþt÷“ÚÏû úÖ®?Ä‘‹¤l¨˜áBeÁQóÊ9ö›ë{—"ÆÉ.’…Í$ÈÚH°ÿ_²RE‡/¥Ð\íz åL[ÇTº”P²vÃ×12@)öjÚÃþ™ñÞŒH°Øoj#vRN…!æU‰à4|¤”DµÞRÜ$ÉÇYÈxj\H – Üþ¤ÿŸ5C…‹Ä+ÎÿQÁ¤¾Ö, v"MjüôNãQ´B4²ú—4JàkBef·Ã^PÒÊ#ÞA€BTõ÷MϯfqP!2rP`mOŸ7ó¢·ÉHa1ÎèúêúÝ÷r]+Sáå°oŸ¼É.2¦ïn"îj‰YQ· t¬O³ÙC3qº°’$Sò¤ éÙg6¨ˆ4ãö9$üèІҫop_¼apÅ‘§Õûó’K¡Ü¤h©Ãh5‘YäìÇá*Q…\צjPËŒËpO–Ó•Øýܵ¯ TCóˆ¿<Ôé¢õ¥]Õ$, 6“¸å@kŸžŽbXz5‰æq{§ m39Ùë.•ó= ZÎÁ+sáÈûýþ΀bV‹Å€úÅi’VkMáÅ~ä;Ak.Ú’võ~§­ÊÂÊý*¹ö .…ÒŠ—«­[?^!‹biîÌ[„³zÝþŽjKäÖûrÒ PÀå¹ÍìÆ@=3x92 õ}Ó›'¡•ÈìŘ6ƒ–ñß(±ZÍq[j¤KÏp\K]í– ç.óËg°1àÚÙ¼ècBi—xýIéýZýu¯äwá2ŽØ°b4$ÃeLýÜCÎܳܩRä%Ìn æ!<ôuN÷Ö\§b]Ãâ2–9¯èêJâ¥ä”Pa€È Ï.ºÇš€ÛùÜŸ6ÕÄJó r°Áúô+ª6'H'[7ÓµØCÎùЙš¿ý³Cµ·YÓÆXÔ\€*ý½Sv‰N&ƒ*z`G˜†H\=­‹ŽÅòhy"‡UqPݺ!-Ûñ-Ó×wg6/\£™¦ZFå²jÊÚ¸6ˆgЖU½àCðŠ87÷3! òç§i†G‹bõÕfv£Øš'£²tOšË#س\Õ[ŒC¯ŒÄEöW=jc,r=Ë÷öÜ—¼J £í³ÔG „ÿ* \ø‰`ªÞì—/x $À H«z–*f±T`súZHk¹œ ·N†Éþ—H¥šsÍb³ž¨‚ÜÕöâ(¨–Eà÷/y¼ó6ºÎ›bächÿ „ÙM×$=`±çûøÂ±5/+t‚Ó @•+SÄm™%YÛYí½]%ˆ¡ôFäRÃG+‡¼fqà“OmZ¯U ý5ùhfwrQÙ/fK 7"©4‰‡Pè]mš§[ÞOÿŒ®Ð<Ån˜'ŸÕöDrhßWºà³$/©ÌcìÌ[«\nCïIdìW†ŒS÷·$¿´éJâP£Ð‚§LMi¨dbó© ŠœÕùHžþ]a$Bú7[¨ï¨jÇL4n7ü»ü[Ùt+J³rf%½h]¨Òì\e¢‘/Ü5Ÿó,a>˜ ~À*¢=µ¬$lbaPM–%•øh`Gúb/ˆ wú]+™ÕؼÚn/˜–¼Z.­`orá£.ÛäåH¤Ž[`ô¸ÄÖéÿ‹;ˆßç#ÎæC4žÅýœhÑñH¬²†1ŽKůyÃe¤Í"¬{M·3Y[L38гË%ƒ \ÛU´û+ I9•µJBeZRj»`…ù¥®V­í|µyˆ4—Ò¦õèoÔë *¶*•Ì¥¤<ü굤VLŸ»ao¨žC &ÚOŒq‹åÕsÖ!*'1Zfª°Ê5_aù¾ù6F׎©ØÎŠþ Q‹­ã‚ˆxÉ» ×­J=!=ŸcWÊ aY«ÒÂ߉/ÑÕ%írEÚf//B&›U|™looÒ|ÒÅG÷F!!%–´A Þ¿¶œ±5Á¬Á&¢FT§K)¬F§DÆ‘ËÞÙÂ麾ܧO4ˆ"‹:_Ý(òý°™òh^´îŸÔ1ýÂø·Œ±büà«*º‰3Z0H»‚âµr1´úX¢ÒÓsÅë;ÀÔ­•8ü¤^Wt9p$cqR9¯½Kf²5ŒÎj!lO&QšŠö霗4¯iIÂW¦% (ºÀ÷ ªéþmy!5:’Þ2^ƒ¹_õ€ œ¢„IFe1—H¨¶ÛdBx?MyŸ%Û%£¦ˆ¿[ ûÿ2ÓxßË*‡9:qµÔ}¬±y†ßÒÙ™)ÖLßeùEÂÐ*__äeï«ÄE$}›fì®Àý–¾3éªãöm?ÿüá„ \‹EäCüxˆ.Emñ†ïÕ1m:j„˜)"ÛPN–^ @fƒ e¤WÄ98«žµè†£)=jRn5Xd¦j¶døùŽ£d<öŒå =:ï»MðK9ΙDAŽB5e™vñÛ!_Uë¤ ?í±a­“¥<&·8Ùÿ¼”3l]?%’ŠÔ£äæ]v¶ún¤ c”F>Æ 5é•“Ýd y6øõF³bt,A“Öà¿‘Y w|û±šTet*øå`ÜNJùùts¬¬>#»®`þà4½•æK’kö‡(©¿„Š]àÚA»2X¿Ã ¯[lFÎØðæpÃg•®?î/’öxÄf¶þ¹wp‹YmÐ@:´¨+td¦ƒûv_C'¬"~õ@Ðc†]%ëI,ÞÚ_0²¼ä…ŠVœ£‰mPØyqSAòµù ‡Tâ'+/\K¹%Ù¦¼|Z(Ø¡x¿hÊáì™V£91­€qE§/^Ý»äù{JeÕ‰?qQ47ËõÿÂoÆ’±óˆ—7Û$Ê`Ú¼™ïØPÆÖº‚©6ÂSÓ„ãTb/ÜáR‰s>ºK}ì<*`ÁŸÌ—¨Ï•ƒàˆIý´Ž‡ÖI¾³}'ù^+;oH¸ÕÁkx÷Âg nY4ä †´è˜ƒğ†Íd}1ÈkÀ+’@º`_‚à‘œúl‡’•ÀÆÅh“NXHP¿Mˆ’±Âq.QÞ믎,†0°ûE§‡5ô©‰œBO5p ¢a\Uk™Çœ\¸â´Q…Ü÷´ÞSàÁøzcd¼]°¶ ª™m°ÃÚ0ÅùÑ™æ!æá{™·˜`'_k2ô„p9Çœ%Ðj£ÅmÛ#³ »SíÌâérmðÖ¾®e €åÏlš";°_±k‰W§# ]ÁLÚ—·ˆÍÍ•ìþ¤¦ÊF´¤ÜÅ&q˳s–īҎh¼í ™ dÉåÞ¶b³šáè–ïVÜ#ü¢W¥"À3¹1´O7Ç‘*·²ÙñÅ+°r“ac~Ùl[·T»]š%~ð«|ÈÖ„ÏI ›½Ý‚ÀŒ‚ƇwùÄ¿ôÉÆ-E`O¿óRÎV€…'þj`/x9øµ[È\Ýþ+”Ê×[1R÷TÂë¬Ò½ CÅåüÀ:Ï4Õ?—J€HëÊmçŧ%Íñ]eÀJP¦ŽÝå›äøuŽe)Èuo¼7¯ª©ßrë6f±&%øâÝ×¥Â×´Ák±lkCÞQZ¢n½JÿkêܧŸ’/Ài¸gCÌýK²^©n£Etÿ‹¦}wÛ0„€ŠÍ”óQVý®ù:¥ˆÿV;ÝλU|&¶çÚÆŸ¢Vø_ ª¤“¥EqK}lú¨Ç×HʬBήK@¼ÿ|vˆqq¥¿k¤Â›3ÔéOâ8 j.©V«)ÞÛHÌeL@ÓÒÅØÖËñx¯e„1FÕàsÂÒÚ³]S?ø•*Áìy,ÜÀíþi–tzž@D tCö¶È½qM.ÎÓµsÜWq˽Ð4ð²lDŠH-ÿ› Åó«Ê¡¸'T’ݤ¬Þ<<æß%PœIx×BµAl 7Vj¥E²`Ôâçз¾Zeéà%.æ’îu¦¾¼_´—O~œ" ¢høŒ4‘Ü…'_ u~œ<¹#…kB3ïr¸&4ôtroÍ19ÿ³w˜8lmºÌ}òËÁ Ö6üX™¹z#Ãò/Êy– -æJš_IrDȇÎêˆÉ¼(R™´Ú ©L¦ù=úÓÁŒns¦­é`"R}V`=`©b²B=«ÈÔ\¿FGžŒh¨ðT&DoèJ›ú5´ ôÀ¿SÄöÚi#¸ñŽ$ ¸ÚŸm_—àéÈÚ=ÿ7[¹ƒ¢W¦ð+>–NÚWÆ£Ìß(IÞª¸ÙF ë«ø‚ªá¡ŸÇxþˆ[x‰‚ƒgÜN³*ÖË@°ÀåSv2ûxgûC+#wmVú>|šè­à„4¶‹ŒÁЄyÄ™7:Œ7)Ö-¢@W1"ìç9âyÏ NN3Ù§ïÊÖ½AÑtŸžyþ'¼ž-•ɽ«cø›·›“@:—¦{`,Ùª&hïTñw/þì½»LëÌRk› Pù§„õ¯{¬8¬Ø8ÈâëA8²„³­¨¸z¼{~ÈT2 Ø^¡rš?Qº d î¡}LÃÝ[TŠÎ'¸·úV«ëXeŒÁ4íÒÏvÈŸI4ïâ!×§¥pØÒIzÄHuÒÜŠš·Vúµ´N-©¾E|àJ y·Ã@ÊßbÏVf;Ðvò¿Ð7û%t6J³ œG„«hž×È×Äy>σ-¶øà‡SUQpjî+^3a,#Øú¼1Úò9…À;$P‡ÓâþCIÙ…ô:¯ÿ)›á'ç0e–8Ýeü»ï:ÓŠ˜9N oÎR‡ôżGܯ,óS¤¿²«Ÿ7Vº*X^ÏÀµ8Ø<º^/@5¶6Fp ‡m‚’qÏ@¢c\²H?Lÿœq³-XBXîpBÅ|ru߬º×°Äh ´ˆÎægCî±BáqÃãçž/€Í>\»?m.’öKÚÀ‡°RÇnt¯oŠü”ˆ{·mãE ‡àÊð:Ãòé±D˜šVS§)<}¯ž‚ûY²Ë(Åj‹(,ºgÚñïÌÏùÌWå县?azð»ü¡7´°O @Ñ»uW0Ö38š{ØÄ|ûŽø±hh£Ûâ« ¦öæÏª;?ãÞ¡§7`™9ݰ f ·Öîu@u‚úw‚Ԛܹý4 |An .Ø%”_ øtºèôKY+Ô°£Ù°#…².ò‹îq—#|FåÞ)¯^½·KÓb}Œ Ï_ßO^—).Bô{ÎòF9,v?1’pfp}€? ‡YntÅucC­B´»÷Û–ž7å;¼xmdÆ µ;º­”õ¿ò)n–…µ>=9Þ/ š™åEÍ…›''=`¹ÕîÑNq梯¯Ö jj{ÄÀæõ5ù~4—m{шì 8xÁú?=O0±i‡EѼ9#Ù4I}SÒô8¶¡} Ë3‰ ycÈÂ)fñÜoTÛ:ý‹}i᳡t*õæÜ`z¼Ù†€šÙ7Ø·—IÉÎ5PbZ¡Ÿ™ïî0–ðkÃÏ;Þ’Jô~ü{zò¥²^¿S_q_¤(]m—FOºo\ó` Ö~ð%ÔŒA%ŒÓ ©$a^{$Œ …˜¹Ú©ØÓÝn h$ÿ³Âçñ$ïF´ª/f5šo^uš¸‰ÒïüÊ,¶Ã‹¾£†uú _D‡eóÚ⎇¾§ô8(ñ~‘%ëX•Qþ_|ñŠœøI­Q=°5‘m«1º®p€¦ÚDÖúÝ:´+Àü®æsª E>¤KézÑ «]ÍÏZ‚ ðИŠiæx» z´—Äùðd‰±óöXáÖ†òEITƒý½ØûéaÇH@‡cfXŽ¢ Æ‹êJ2ú¯?f=ZσoÄx&â: ±h=‚è¥&«êŒ†¼‰sÓú)g€Ñœ>âß~¤W–h13çÜÈä×GL† Àï‘·LÞGKtÐúoÀøH@hwV˜ºŸÿ!ª»œ4… äë9Œ£¦üE&ùRÚ)í»D#r®{¤l%ÿþG\±)ït¤{!'If?r{“£3߉ª]#ñ©„ÙÞeyOEVÏ ¼†y‡g1·É5F—ýâWŪL­Þï(Ü¿{]+`eVLy¼!íÞ€•kЖX©dB¢—éªUjÛ™|ßÒ‰f‡Äïþ3KÈ/Ll1QÂâ†Ð_ MÑç¥T¶” Ðc ¹›Kt]Bi8û¢ÈSÞz‡Aü¶å‚Ô(£P’ˆRòzÁ=è>]Ù*¬Øeá¤aAA®¼(XjO*z—ûà ¸W‹ôHß>‚eJ„Ûåîx¦×šŒ‘¨Þr„ýµ,ÆE “l«ô…x&ÇB´Ÿ7,jÄÑ}œC/%÷Ž>5–‹8õDfB†ú}pûØk«qƒ ¬T0½'Xè°ãLÞêrPŠ®®˜ûaö&ÿ÷ºXÇB5Í2ª5YHœ#Ë[2Âr1m%É7½_(õ/áFªl >qZ>¼Cå8¶½^G°¯aDQšà{³Û¤uü•uò%aèr¼žzÝ$*l†­H ¥Þ^i%×^ø*I<]þ H§²ø¤ÞÝBNÈÇÿP“d`æhÎ?*Wܪž½–&ÏAqÃ*{KBF°ûš2”­ó†*]2ÈúÛ$qUÅÖIPVqVªI4øä0«¹ùÅ ’•ä \G¿ŠÎÿÔÊM„ÁÉf 4ö7º¸Äy–ñ Í~" zBåò\G83bÌP«WÔX5L ãq!öŸ{,‹Œa\Ý#ê忦¥T>÷Ž6 eŸ]¢TíÁMÚÊyS"÷8™f7>A©Ö],¶0bb”»kÒnLîœ;/ Zè´ ŠX±Æc-Ê=| WÛo̾ô?ùÚ¹Á…’(‡£5Eë¾ÏY›j1>¤³íÀ­4¥Wõ‰ê½çMö¸yý=‚QvX¢\ð¿£=Õ|ŒôÀ‰»Ò}ËÝÄ:‘ɨ媔üïÒ·)0¯_©Á˜BTµ°AøŠ—ÆÊäb‹4¨)"ÙË<‹"ÎJZÅGîúz>UÒ±g>…K&T‰„²s’*M´‡-ïæx=T$ŽT`!"i¾ׂ­b‹žsSȽ¨æÓ£Å{u.§¸h»¦Sja†&„x§R=–GìlÑ\`ÝOÌð±¿üBõDLP‰ójqÄ`›‹³Y¯ Õæø­-.Y‡Û¯Õeϯ‹± b“ú°Côr$¤¿;’ÅcVQo™ÄÊ&ð“É{Èi—Á„úæX2E:LieÓÔ3]e3'*w ¯‰,ûºJº,ÞWyøÜ*];5E+I¥ŠJ³dáð Å ¸B_åœï2îa(; ‘I:néœ(”Ù°­5±‹rÜæ^1D_°½¡²C€ºçAo²ÞƒÙuTkù4§U¾Ë¯üäXëʹԦøc[«§.B¯v¦®tÃ`!|"¸Ée%PÎ&IO¼æ€¯þfsIq1.®‹J(zîI¹*ê×rP5[V6×%Ü-q: Q^²*įœÖ,¬ì&eÂH†Ü¸*nÛÀ™†ÚÓ±Ãc뮫`h=ÉL0†I>XT¡ù=ÍÕR8—çÃæÎ\ð0½Zü Fh_l3¿<·D¥3ÀÖ‘2NU^«ɆUQ}›¢½÷×kð§,¿üåaPܱÖÃÈ7¨åíz”"†ØlˆÿÉMUÌvéÜUÂ1%÷ªnЊìYì7àÕ¦´m“e”b6cØÄñRƒÔôâTK'ªÈâÀ6=] o `4Zkèy€ûddl ŸËH³ŠZ<¥ŒÂ@ Y°/Á§à¸1‚nï¯'%”x1õî´v´(Âeý祠¥E6j@ø¨i–O_¾‡Èeà ÷ժפrëh.jêÈ!-„‹—º¸%ûä=,Ïâ"‚§Fo°MüÔu €¢«zB‘<á#lÛñˆˆX²ìêþŒƒŒÙÞ}ÍáOÂ5a¸¼Adòé‚c¸÷Ù¸ºÜØìÏ<†Ä? ”/¿òš¹[ ÕCªj‰mpq$ÑÑ?«67(¼#ùµ ¦¹óX©›G¨V ¥ö#Å‘ŠàÎÿ.OFøO2ÕA)µ¬ã°»(c;/„`=5ñØ“°ìCa)æ]…ñŽ@ܰÏÂZû>ø¼ÚÓ]¯¬á¶D Ò6У=ÖœØ\¾Ò3±Å-bJ¦l"éO$;€õ0§ùÚ;Up9`†™±ˆñõ뼨ÿ/®V )ñ“TÕ¨•ˤîWƺŒ[eyˆ×È ®H)̘KóíÞ1¦.w#õ+G¶€™™/oÚÿxAŽ……/»â× åu€Ø,$Q†±ÑÊyKu‘q¨e Á\<6˜< º4ðÞ'£¦ ü ´Ól²Æ–ð,é#¶X…' ÌgÈãÝ`~c‰f€-›Ùçz0 –I?É*-+7bUåé†Ì'Sï¼”ïARï«x\‚Û`ˆKnš:¼™¡tC·ÀüO@SêçGßN©­†+ÄHZ•~PÛÀ³íªy{§†¨ž~ŽWbhB> jR™¢ÆÏp^A©†ôC¸Éå]" X ¹íe i•×Ô&×OtŸ6¼nBm“l*WÊQ(V›Ú€CëWƒŸU'sõ”Ú‚Ó ('‰ìZBöl¿•8æFkƒËä4Ρô4XBýl½µíyô søÛUòåä0O=Ž…NüíUh…˜uÔucM ârÐk6ÜA  þW`öƒï`ö}‹Úþ8м!Ét 5i]Mþ4)7I!ŠøÛF·2®ŒÙ™ÞãžJof#®wµ VKMµ‹ÄWø€Þê˜Ãj!À[ºP6?*â6Û¬1jOÝ>ÂdÈTÑÚõ†ù‡mi-°©e! ÒܯêˆÄá?æÙq˜‘¥×æŽf¡®æeñ¹l¼-Á7¨¤ó/G]!ª£ùÅË´7›AïyÆ»eßZ¢=“4ï†æi ,—]‰+¹}àíá‰^] vѽvàÙûÚùJa…Ȫž@m¹¶#}èIщÉðÄÑz¥e.`ï$•òÌWæS9a9î >+ÿÕ0æÇÝ{ûF´7_×ÓÜ¢)ý=H>€¼Wçhˆ¸"2~U  Ónå(D nEîÖ–c[n˶_dæó¹:ˆ½I±/¿@Á”Y?¡IFë~á=}®°‡šÿ¾]ÖáóîWðQnÙGÙ´ $KúŽl R*uW½ŒØ+{Ëïä2îaä¼ÞU3àrö#É€’á íd"VF`É:x×õ]~^d`V ¨†D?ö/aADA³AÎÙëÕ”ý.MIÊåõ"tÈ@†å™Ìñ0åª%–µ÷î™l§MÒþ$ÕoÐø~…1ÚôX+‚,ê|¶s¦à„F›t©èóˆ?àJ“¢ŸÇ1ïAn_dÅüçYO†hæär¥‘Œ… áO\Hok¿3°µ¡Lèþak0ÔE³e΄™[Q>É`¹ÉK{–— Jš¨¾,uÙ¹oo£$ï×#Ü\ —,e›„JXÌ¥ýc4¥z¨7G•ÃÕdo·£=rfK±‹üd\ÚrñSîæ˜XšÌ¨t¶û)!¨›©!bWŽŒµ‰ æBw) o2@4•N2€nÏUÀiËreãøz£zxŒ©cµhOרHºÀ¹ð™|J±PlyÇXéÆcÓÿÚÆÓj”&”N™xÊX)q×N<@0ªãê"À™pAuš\¦ƒç ¢2¶ŠÈ¯²—k–ÈöKÍóÒÖÙ–ôgiÛ„šôÅ+?Üð6•i8úïAŠ¿9®;AÂzòk3 íZ~º¼¤É¶ƒežéA ß Î‚Þ}ò|!g5âÐf{åÝ2cÔÚ MÇæÁ¡F§ª†H=Ng‚åT°ü¤›%»¾º/î’Eº•~Ý¡Äé׬…MÔÖ)t[9°÷X´a¦û¡¬é8_E ;Ù“Ô’¸ sg¨üÌ-²)’†+‘ô™âsb®†Ã‡hŽÈÒÞ Û@óÍpΪZj'Šñ]é5#@:Ñ}ÏœšÅéô,}JC“•)ÅuRW{)½&ô¥NZ;JÙb:Ô¶%Ý$G:ò·ôëðÀјÚ;Þ0³¿y+ѲÛ×$Ø¡=Ú qÁ\®²@ÏØÔÌÜ_ªFÁj%°ÔŸ­1ð«$#ˆãŠ7<¾¤¾ïîÇ!ï¥ü¾Ò|3ViFÞ¤,XSŒ68ë½8å®n—G‰D¯+zž¾RÓ^X„¤ùPÙ‹Zˆ° ÚnK ™+×ߤÜ0«g¬ÌÜ~ ‘Þ%ÞÚ„ñj t솔DG»Ò—It#ã3-xå?<;•í¦¢’„BiLr… Ÿz=9^:ê¿“«“lí1W$é€1xI´›Íbò£ûbqÍŸ`Ê_(û‰O?+¹uK´mTO鑌ôóáãŠ(¯ÔÚâõ#Ž6Sñƒ=@…Xô(/v½ªÜÊÒ°äÊìÃ^¸Ù]‰YXz|!è8zèÅŒ@çYAØ­®¥‘ઑÞÅŠõ»™ÑXCwÅ6ÒX³Ø™¹înó£Ü„È{­ðóæ]¾Ò¤WJ6ñZ¢Ro ¸ŒÛ$¿?pœÕw$¤H 6ïÐS ñ‰5¼ªÇçî0@­Ã%VÒu” P¤×ZЀ³Äe.O˜iõQ…àAÈâÕ 6ß_š6ÄZ¶¦ž_5ŸÃéD¾£ÄÒ¯«fÞñò¡,ýÉU±µbL·V…/Æñ’|òœóÁÝ¥dº3ÀV`¤dáWÆoÁÔ»¶vÌøŠà Ýé'AÓlC‘ûlLPÕTR :‡þ; ¶/Ù¤JÁCÍb?1y³JmQµ ¥¾R‰÷*H3]ºÄ#7œÏß\ ؼ‹œ`,qI¬1HaòÙx‰Ó¶cÕR/‡ý;Åû¯1NOZŒqÇT6(3“ßÜhPülÛŽ¶øš*¥gwÙ¶c~\‰|N? @#ÇèÂH©}èܵ÷›µžð·ÁŸ¯©®þYëÅNóô=íhƒ4´šûBÅ%´¥7íémzöq 5WdìÓªmòluÃE·´³ÿ×*¢F’œÞëú…Ö8“Ó\~ÊE$vPƒGb*±ŒÔÉ<ÑìsäЕÄúˆº¤ËãÖØï ñÙ·;šyõŸLüÎl“®¢¤D=¥Qªš¤­ýØ_È@Õ&†Âª  !„G¢Þ¡‹*Y8@YX°Ù;jÕ ¡®­ŠÉÐŽ–|¢¥8Ë…Zþo¶+Ñ®ué²ê¼²YÉë?¿ÂÏÖÈËJˆŽ„½Fýe´tíJœÿ(~qÙϾ*5‰ÚG"B1I/ÑJ³ªrâø¿ ‘…u¥b,¬²è{c)$LÓ‘/f:_xÑóíz­Oã¨" ^S°:[åhÆM†WHâmM9×öbàN+‹t8!¤ºM±°i¾ìáp9ãMý:ëM¬TˆRg ƒÕ!`+#á=-ªn%é:Þ¦=¾ÖíGÓnÜ~O»‰•ÿžª_¼K‡Âpé®ÐS5*ñë …*Á‹˜¨†«…ý%uOç™ã9ùBA3Ž2ø)hCÖ 6šÿZ4P`˜./7ÆÔã!€ï·…ÏñAÍmÒ¶½MÍ„ w É S1'ù=¡JÐKÙ¹QÜt8dÐ6þi‹ç~Ö5’–rfÙ³—Ì> ÎW†ÓDA†|ÖsbýÑ—ÇV˜{ÌèâQdš4{§¬üyJ*á(±4 %D;Õ¸˜oRbº8'Ò6¸—Þšè;-°£Mºß=yž%Ìú(Ì-ú Í~WêÅ•Ì<±þ!5ˆ©B$‡Ÿ•Phΰ\ÿ gŒxÖš#eSƒyr"eàŠhvL½‰†Õž&­Ì U%q¯mãkE@3ŸiÕ%QøO_.ç%Ä—30å· $i;†a±OÏv ÌìŠÒáDÐß’ó=òÑ€¢ 6sÀÐá¹§‹ø‚éè·«¤2\VV9䇔K­Ð¤³¼.=‹È€8ü7Œ 7’t-»”^(Üîìú«iÑ<±·³Ç”S×—æaõjüˆá¦L0!(yˆYtSáÙE©PæÉµJilg‡ùôÄny@;Š8MB /ÿ‡; ™c]ûý;¸buà_?aò¡9[rq$ã>>ZÛV_ õuvÛÿÉvŠ `íc)‡Û¦£´È²é/}”©¢ýZ_@—Ó·_Y†d” ’‡ÿo$÷•s%Hº!u ¿–0ÖùæÛþ¥ÙkeVm`çy dO* ®A€ãJ£(‘oÂ16å€ ‰´”œÌg™¥a•Q•’+$é5Z®_ 1ê„{rw"¨:p‹žöZ¡Ç±Æ“#ìeùT-‰óSÊw÷Lƒü\#Éof!YÖ¦/0Ýf‹•GUŸf¹´³Vl¯sC›ß`GpŸ÷.öÐÖs[|­Áî6›uºü¶¾=E=æR‰ lLˆP>õ‰.À£‰Íâü½}š@ó¹<œŒM8/þä‚¿O ®;Ôú(Ò7L³âöLOXm¡æ:¿—dZpE©@:‚–•L Gë­[ëiÇ•.çZñgå¸/§ª v¾s–C}¦þŸ¹¡h›}ºíà©Ä[ý@·³Õ³6Ÿ—¼`8L¿²É ëç“eÄB/þdzòQ¥ðDèÄË刃È=Ôp»×¢šÑ 0Îíö X%)G0vJr@Z(¨ú!o7ÿ (Ø /¤Ž;òq¡Î>ÝèFSì Ú°°8i¾,3„“¸]¶èãßÔY£Mi5r>ðÁèË©–°Y¹Vž“œókà·¾mgÅ0[À»ü7lŸK`™ ½3¡2È¥îc*'«žRÛ³c™h–1[˜rב@‰¸3¤†Æ—F ÀÍ¢ûÚ‹Ì\¿ÊÂ2dOUÉïqhxÿÊ\Ðs,p$NX ¯_è£[×&ÇÏiÿ Sfçíážñ{ÉRufêæ1±NtRÜ4b®¥è`8çg.,~YEh>  \K$hc7ÙXRs!f.Üå÷;<¹3ØäIˆõ¤Z0=+›‚ãY×°4†+JJïï0c»Uö\®•¾œáÑÔ#³Ÿì-ƒˆð"‰)sÕ᪻¶S¬U¨ÕMÊ Šó­K,¸m¹i¬Vsè}6î;>W[è[ëÎæäfÈSqkU×±ÔâeØQWašá€µ¤PX$ÚQ{]·³l %uš*RÆù*(±vª¶¢aƒ®‡Fóv-©¬êö5lI\T®Ï<.~y sÈ/üu…‘¾r(KËÏtКnU ~ÖÍm«—ôÀ¹ÃäÐ…r˜o4&öá_v‹èuÜ\xiªøžTož2¼Kg=—BÄô¼Ëæ[s_›tt¸ø“I'õ¸„0\Ïd9øÆyŠÚWü ÉO˜º6ì—R¼õûþëàs5#x$ÔžM¾y0îÞ&½¸V18û EéŒþ^n(©å€RüÝzeúéÞûwQ[ â÷¯*-Ÿ>ø7³ „)ÍÖkŸï ÿDRƒr•*žÝqÙf\/ý ½à«/Þ‘_«vä»Ü½SMœ%2(ó;)«]yDf>mü]¨¬¡Iˆ€!™–$# ‡Ñt‡…ùçZwx¶…Àh;ê# œfù1Qúi¤=2¡ü¦5!;Mû gh?#±¼Æk:‹öíAžH< 2?4)óÜ¥%}2íÜ•ídqÝ‚QàÕvµ“¶ƒnTGÓp`¸ˆœAÃi¹;`)‚Õåç5OJãR«úCFcÖI‰$4ÀËãwûѲ°Rów›9²¹j‡ï[6b¦!ôœ± ÄËIôcSxbí‹|î“X ~´%y®äá¶C§wÓ¸î|A•i&æT¤èú(ñ‰›y;ݶ#€5«p|J ³³„ú«C˜ÿËÄÊ& ŸÁ`æv9ØSóÒ «¦ )oG¯#>RˆÅت—¹ÏÈ~¤G<²¨q@IÜ,p¯t±«ól0iú~˜ðÑJÅàÀµ¬Æà 7~–/…£ >âÄ[ñMäZéî/œ„ßl¸ä†Ù܈߮dIn*Íè+¦ƒÌ>d€AfM(KKhü¼$‡ÆØUûDìpC5=T-3ææ‰ü\ò†dWržUdk›Gîκ¯˜ù.¾(7úà½TÜm[Lj—å vRû…ØâIΔ¿¹nCm>¸©%UÍ?|ò°ÝVè ÖÔ±ùuÂ|Ù õq`a¼ADSywHàóÆ'™1ÅÿÌÏGãoõGz.¾ Þ\p—­¿Ø0ši€ëÕ‰]šúq‚v‰K@iN¿Xý—®®_ž§×Òq»aå6Ï1ý¨ŠÚ|A鯒‡Ñè+å “Åµ÷ {â!(§+ôŽñ^¸2?ˆÌ%ÝÛ3#À©>ó@ˆ­öu.ñ—i$ ‘1É‘ûµôydÄiªI£ó%ú<†Ó¢¡¯Ž%¸pÀs(À8:¬ÜsÓÑü72 m]O¤7õ ÆÓ½^¾ÅwÔ€7®1±à=¾sÝFî9'mn¿×g_Ú½ìFCšÃ ÿ€îÉŒÇÙá23pÖ3¡‰ýV;…ù{·²hÝaà·âbmöniO°ípXýT¹§Ü>ËÛóÑü÷æY?869$OÔºn…¦1_Sï8¹êüç錫¢"ó¹&°Ûõ_]ÏB(9¢­s63i¾Ø,¹ «>5#Ù’¿Ý_D™xÛiž%-÷B(: YE=Ä“¹ht®\]—)/]W²“ümÿÚF³ÙÂ}sCœæÕg÷8<ƒ^ ;­Åˆ~•LÌ6:&Ì—pBBaF9läòmü¥3AáJä§tm“Gtâ˜U­*À“‰CM,¦—™6µçÿÅ‚¤½t䣭aæâ]·é|$»3[ô‰9ÇßHm­Õ#<:é=yHu-O;»“¸iè»d9÷—ù .ÓF4*±Â¼WÊŒ´5Ãñª¨?{ã÷ü ¥`àvC„ŸJ‡Çk½^~v«0 .IóD_æk×ñ¾‹é7TSÀShCóŠY ªî¨ºãXºf²2´«’C£ëýëéRx?ÑYàÊ)‰ƒëUe\—»Ç„òÙæP¸[ût ]#ÉñKTÄï–Ñ¡›ôÉû’¹rñ‚OC†ø!Í Dt%leÔƒ5Ù KçÄÉ®€_Zå–• ŒF¬¦Øñ ¬ƒ¹™Ì¸n9gÊ”“Aú±5-CÄ_µ‹9EaÌïŸxF.ýÈþ¦à…çnö¥d]§ÃìMͱÊÁîéô’ÝäÆ\,–%RÀúŸáú‹\ÁÖE\¸áAîöÞ†=+Š÷¥T¢¨§’°gåŠßïé߫䡘QY½ÒèÏ× c™Ä•w¼eè_ʼnâ D™RÔÔ ø/ŒÕª€µDs œä>”a‘õN_!?‚ECb‚òn¤¯:©XôÂm»’ôCt]Ý€·6~Ük'ñµ.¢/a6:Š&:Á¸Dè=Í6ó¾YðÒS‹£'nÚÍkÂfáU=º#YRÓœV÷Y4ùTX7Š>Ä\á¿¥ÇÊP¤Öá¾õòžÜgíFÛ¤¹«K‘ïøŸþçvÄY”ßóÙð´ªÖ†[î`ï}ùž™p™Ð:±. l¾¬ØZ%”hÓó,d¬BeÁæ¬.äNä£v{¬êñ«¤c£üªßÕ‡4COŒ‰ø‘ ™[–᱕w®ç|<='°Â;7µ%ŠhåEôŒø‡Œ~Kj™L¬© û°sa8#ha!¡o³éë`¹= áµÕ+eñ4Pf+ýhzb}J…ôQ§\Ò²íc¯n€"r6ÛÆÃ~]ÀK&{ÌQZþ6jš°|‰‰PNbŽô¾§£µ?¾îòáü·»‰U!ÜãApv¢¼¢ó§tذÃRF»îz?CÀçº‰š¦‘¢DÓ”•950R@2ÕÙv–o5•ùzr§²ÍÙÔ¿=O¤èg!*¡ÿïHÉüIp‡ïÄ9lA£=ãâ/Dý“·8ŒÂ¯®`•p«“õ„˜¼g=g¿2÷¤fª¨bJ(âý?ð#­Œ ÉÎzÊX)%u)…šÛSsýR!ƒ÷ˆ5Ý¡à!/QÝ!ت]®:J«èûÀ't` ãz¼Dßé×0$6Ejy“ë*Ò©¾¦Y*ï¸)-R­Å§»E%_R–ÚKó@;±9~#ÍAH¤ 0ÎÇ=¯Ë¯)°4ºòA ¤—ÅË‚po^ÉJµZFíÊÝ{=  ÷~=º›D~ózÉèè%^Ö²}©UX©öfÇlLW:§|=£6?âó¨%`÷ væôòÿ~J´ãÃg²¹2ìÀæˆÁzJ¢Õ±òÝÂM5‚³¨ÖýE”ä`ñÚ§ <ô‡#x @Œù¬Ô0@vï‡.þõE±ØhWuÒäŠp\¯åA{™sû#/Ù‡¾>ß»ç‹=“Üè;±þªÁ+J] ›êJñ6­UnÀäü– –Ü9M%`ËŸôÛr‚öÌðØö©¤¡kŠ…Vá»ØîÓÎ 4˾5sÂ`Yj6Ö™x‚b½KÈFdxB3älê@Î|ùÖ`wOïF»7¼«ž 3g#zwñxAÓ÷·C§2䋞Âunx¨êÊpAÆgè•S·{&;¶~Ø®.£§Q‘%ªŒfºzÄî™z®PVò¼•Y,ˆ–iE‘ámù1kñŸ©]Év`xV° ¢zZJ3;rE¶óس25¹± [ò6‡Jþïf“UÃ},Ó{¡««ÇÊ0Ï5È{ÍÆÇú`{ÝrÌçA$öfôÛ*ɽ§‰°éhsuqNMD¦S_V*C¸µâÆPß œiÀ)Ö“yèÒÙ+ 0±ÇT»è ˜©ödbZÊŠ³ì)ùùwn… ±¬GïaòB@òëCb2$Zb 6¢›k,«÷˜êë£Yøe°‰Ùµ:‹Ôø’Q®gêËd£©øfð¾€'€½#Š–äÐüµ«s) ' …¤Fí@×îv’ãÏŽ6¦òÆ13h~ tf%4PÕ%~iãÊ!ŸÁ†Y¤ìY|îœÐmOâ¢QƒàqŽZ-›!š<÷)õt¢,)rœ¿;ÿq1ÁF•**Œm5©‡<¢:äÐÁ*ê±2¦>ˆ§0¢J·Ê˫ʿÿ q× ´[2ú$1BÓ‚S^rÉ‹†Åä+”j.§0EÁ¼Fêf&œqß^¡Iܯ;1^?¯¼qD¤)V*V꿦¡÷eˆÎþ\ϘVé‡a^¢ÂƒW¡Cx)KJ¨fä)£îq‘í-¸ÌŒ<Œq òä¹öŸÌ&h“(ÆLû™ëh†PÛØá(* ·yCTõ^&ÑEÊÂ-0¨„3¥²±Ž»;u¨6ÈÁmhN1(¯Àþ4é~i¢½ø¥ÊÛ‰ÑÖmØ‘*aèKìû]ËP8ø,›/1³¦‰Ct~ƒk#6  >%“¾Š«»ÜÚK÷`¼G4!ÁQ÷Ž?Gx¬ #rzžfJ±kd姇õMo‡ w. _îA”!‡JÇ“9–º‘À 쿚ž´Æ‹{$ô`Ø™ ËŒöØfóÉOd”ŒñŽÈ1ÌFƤpx‰pì(K´¬£¤…°Ì‰–;Ô¹ìl¢õ9}¦*‘Z¦'pˆÜÉè±{lã±_ñsw"å´ÏPïnŸ“Î*&÷ƒí›Š„-Ïœ.Uÿçäå©ÓéZZ5ôüUÖ ãGð":I%ÍP?Ýg޵þõ&QèÎÆ»ÚågŸ®é- ;õ,™×â)Á§~…{ßÞm¬UXÑp¾F©ïX•u†pÐ Š…¢ÚÁÅVJß`—““jC.2GNüç©ú;Ã"±ƒ•œˆýןÇq@¨þ($îòá|Š·86„ Ë¬ W®¥Fn%ùgªš5j§ܯrvõ0 xð¼µp á¡5rôDü.Ñu. +0–8•'J1Éÿâì=o·‚'¶k4lSæqÙòæ9Ô+`à­61«®ÝÇêþeœ7Áæáæ @™Áñ«¯ä¹éXkFU4rGT"›•¨· :Æg]‡þãÀ\l|@X‚ˆÚ÷ó=æKžµW° BrÍ(µ[K»·,ãñãk…MâSéÚhoš´P×òO6gÈ¢HS¿ ó#µ„f¶ÔñlC*ÇÂS*mznEYÈAœ¹óÃZvFÆï­†ý=ÕÂŒ8Z´ˆá%䕚‘wŽŠÚ¾À¿3¿—¹8gåHìVŒF…Ð(·d1{ÚÛ"wøt¾Im­Ûånèè±Pp§µÚžÊ/m=½%Í÷Lø‘gØÒó`ò3\Œîß_ç_Þ–väó/Ž­(™µCùçß Â²¦Ùè`¿™'<ÁÞïí:^׺ˆ'&x7ö:×Gû Ç—Wý²Þ”uÞ£xB±<Ã:1KØùPvô8“×±¶å}ßéš&f9c ,×9Vì‘$× T|Cå;˜ö2÷í¶„-1gÇõÅýàêf®XÒЧ†€YÆ‹•Û¨„’«K,pd‹FL7¨ª3ì…ä$–4R@‘²%¤êæM!xk̾-^–=ߨðƒú 1î×4&ˆš¥d•Ã0;Ô­€zt.•ÍÆ ynHÚ†3óéÚJy³IÎàç~„^™Ç]Ó>l%õ„dGJÞ¦ $%ÿ榡ü!ý™ws🸵—$ƒ^€¨ê›£Ùÿä<Ì Gi ¿lW “¯nÑšPØ×5“ÑA—¢Ù󸶑¸Ö3­^ªæ ´I¬zþL:b««Xç+º±qqúftÃ&¤¸¸@Á󚥱n-‚D%ÀBþS¥pGÔ÷Öõéxt«w´&jh†&&$ÜAJªŸÎ-UP|œ¡»úî“GJq™,Šð-çS·«ïï6}†Ô_'m‹§vŽüÂ徘KúfÖ_-c ÃõóÒu¥¬ËÇDÿMÒ1"ÝáßÑ»üDÛã¨BWRúG,×c×ɲìÚ*"ôS­þ)‹}&$â¤/rÇ\óÔêœ`ÄQŠ£ëÜÈL±9Ϩ…tœM?^üoŒÂ˜È²ð†69:”ÈÕÑšàm˜ÃÎ|uÍì_L.>Ö? ’nF7ßg¤øÐólOÏq©Ëçø¨ÒmsHÏ‚žÚÜãjò@–sOvÓ?{ês›øJý’CA]ŽõîyÝb½ ¸•!2ŸìуsD—Ç®ˆÌôú@;;ÜNÏ&8ÌKj £rV îsG|v }|³«o¿>ƒ|ËriŽóÌQ&ÒÞyE 2 î|Kç—ŸŸ|µuc”ÎðžÖ¶T1Ø­qº@WóG!ÍýàmfU4Ÿr/oC/* –Y˜0¿fónŠïk¿ÏéøµÖMŸ¶,…Yv µ-a:wô¡)œ™T¾ê-v‰Âqv‚-mÑðƒÍ“ކð}ý+%M‰¨úcËToêgŒê;ÌùK˜;9›‘Å1¼”K¥snW›×]¾*½U¡¦Gè2–éR”&õ`§È¾)ßpÒœö*$¿Èé*mëOmìWá›v¬mÊ`jžª¼ì•lþ­¢ê¯˜ùÔjŽ˜|o•á´Ïô£¿>'ëµ\mÕáõ"ðÉéí?)ÝXbŸ<7Pè>hbvø¼ .¢,FƒSAîV}†“p) óáF/©wWªÀ¹…ù&Gtò3 %‡3ÓÕ;‹­M«-©Š´£žÙhyíÝý#hSAXšÔ®âvºô™ŽWIíÏrSó6z_lc“U€Bcz_ *`¤és%¥:¤°fp–A¸åñªuF÷Þ£r?Ú™KЕAòÓ-ý¶þâpOyDóX²è˜HoMcwú.ëßÙ:–.ŠŠÚù[+ø5k¾Q§¹Öø^ìWÔÈ#¬åk… ù ³`‹Jkiïæ›qç:ZìÔ®h4‘¹É Ïkylä3„<0Táœ(#<Â?¿O‡|þŸM€ÃX6CoÔX§;)â“6*ÇʾٌA^|sö  ·‹È ¡¦ð…©^!¤8wàReOu¼¦t&Cq†„èIÎa²:¹Ÿ²j¡åMì}W3{?Müç•,Q™;tU…«É`[Ú@ˈ½Êûå3fñ® gÍWz¦,œ c8åJ‡&ôÚÍxgì?‚xÉ×Y',HU\u¥HJ ;³rÄæ‘ݘãݺfÊ¢ÞX­ÿÖb!Þ]õ;l ½o]I+w\¯u[OY¯c$àƒ°”[ˆƒdÔµð­½R&²µÃ?‹8&”üáU >îéØ+ÑM=Hcб8\”*ݳ¨9¸î3¯y à åqª3Ÿ1sjÙ—'·MêyeMIhÄ'Xš)sŽUÓUM/Ã'ø öRf¶°¿ºPþþ$EC®ö-1ÄGôÕ1¾4SZ­¼î÷èÝhçgTö;›¨LAx'HømÒ5I î‰YU(¼™Y™W’Áõ2’èìß…qËuaDß½:Ú-¾%¿«Xj[ãÇ.Åz`#ž¼aMÓAê×ú îx-aÔ¦Õ;\jÏ›•ó¾‚ÅCJ×ÌeÈú¾ÆÅüœöq&ÑÊ„cçhÒ»h‘3°.7ŸÃc†‰ªv"l#šbžt{% ÁKKÅȤK+V»à5t.Äu ²XíR°áJX¸ú{[ûž†–µåÌP™™¥Ûó}c´ šžû­§ (žýH:ºY‘­’ð/Ìx3P’qPÐ/±ï˜1ïî´0=X¸¹Ãm!ö 0‘Ò€òç„tüûº¼C½/Åн„_zΖ&€93豊 ©Z?kn½o—ø‘Ûû™@Ú¢¿¹¿‘ˆímlàpºlÇÙÂhuœZsÍ™³(fkJ˜y~8j•¢FÊgž¶ÒÆRÔ³eF?gã"ÛB¥{Þ“ú¶Dƒ:¿ÒÕÏùaí˜<¡E @.íô»5~ u§×B²qD7Ù ²ãw–gâf$P< ŽÈèkT‹–'¥ú&¬l¤æÞZ/óùl ¡Íãàÿñ×Ë×^tuŽS¦qÖÕ+ö¼x”Üná ¹H³M&òH@¬ùÊ8BÎ@L+ðo}Éë!°«ÝÁküT©…"Ÿ«.Bf1®–¿Ø×&˜Íb络­‘øÓ*Iö²û5Ž—ý¾GZNºq/?º­îâà:;ÛZ9ب-¨r¶QPÛ¼×Ëym ¢ÀÖï3>;¥«\PÚß?r§¦ê®Q·5•ö!O{¼_gÆì¬Šu‚¡Åb5÷bÛšXïSº}ª¹"ÎE9²¢`œOØ>IvJD~GÝSµ4WYº TÂëøÔ4¤‹C\eê˜çcæ2áŒ]̦rÄ.S ø±+²(ô+ÂÍ‡š¶¬…¬_ótÚ€NQQ?C„ùü·A¦ÊWzÍÝG\sž& )OTåÈœf¢¼–;oÈíÛ®!¸ÆÆç~§ŸÉQ*6àÇß/•™·mxà8–ÝU{OŒ7|:±Ä3ÑpËÄŒV…qÏgÖ @FÖŒŒèú×YCàl]m|Zþ¼}Ü"›‰zƒS2º®!ßÌ´Q·ÚåóF HÉ¥gqÎ_.LiòÝæSCÔ)ïãIaTt«01 >ìθuæHÓô4Û”•<Ñ‹ ÿEبœ»é<šµlÒ{SÀF ÝÓûŒbËB⥊iW™×nÝ“ÐÞ;ÿÉíDj‰›’ñˆáºÉ7R íBàT”¨¦ÒLåÛw—|ç&ë¾¾EaÏë º¸‰Àèpª3’…ÛP¼™­ŒxX†mèŽù0 —ï,s,¿´ræMíbà€¹5D'lMÞ+Ï Dÿ,˜ã´æ)4÷·\ê÷¾uTf¯_E;F¨Or<È•¡ÈX¬æ>UPFV_{k寠D°>%+úP_ç·ü¯¤`¢Û¤(¹8¶¦Jcp“;°Kp S™Tdb"µ"ïE+ïSž“ %ÃS«@N ™\:Ò6Û‚~‡qw½Ù©3ï©’$ÇÜ2ÈŽEÚÏ5Kò¾{è]Jù€7ÍmŒS&wÓ. ¹4úSèœØ¦°57_E1(¦¦-àyLCÜxýøûNÙQ ÀI{nÅÂVääinΠ•WÉ©ôþÓY^6œ;Ýó0 \åý…X¶<íR€Í[N‰q]ð©.=Àfë0ÅøŒXiaçøẹ̀—éÇI¸^_ªx_¡k ®]T—^’Ó3G'åÿ¤ªóIŸ¥wÜZµ)T¾‹¾¡\7{8’aÈV!¨ S·a½¡MvÐ*Ìò„5~à87ñ9×5Í­`…ÞÒ‹—v…,À a~£Å "X[|¾)¡#iêÏàÝss—þ”tfxŠ5Xâ*‰¸NôoÙ3}ð[AZ¬—<5Ïj3ס5ºö9Ö¢Fœ¥÷ÈÜÑ®ÿß ¿á08ÓÊhöN ‡‹u3)¤æ‹Põ-ìÊÁ°ªÆ# &íÌúX?2„ž$Ì{Gp}[Ukía½<Ÿ}3óª em÷Ý%RŒÑfl‹Øøçe!yÛÉ#yUØ1¦¦a½>.1.Z´úÿ%—µ°ô—¶zg/msD–ZƵõØ dð$ï ç¿)=ßï7Âaó¶Ö°Úw_uô%\H­Ÿ‰]k>îšÞtVºÂFoPœe¢CÝ€ï§v¶xΑÜv®Ê<%sß„ÜÒJRÝͲޗ÷•R–7ß„HzØÑ¶Ø’¾Ø3Ü«JÁMP¸’ƒ}!3 y×äít¨ …„«.ûŒóròg{œ²nԱݧRÌ <]#K7à›c0÷ló+zõæÁ×¢õ~Œ•/ÇI¢li¨Zš6”n¨7¼ ÊšjZuÁËá?ѹ¡"FÃzòÞQ{'°â@°'4üê-"›ˆ´30ãl^©{'ŠŒ³1h—@zÑpÇèHÈÃä¥Ü̇I’ÂÿÝjfáœèªUÖ²â(4žôÕÈÐéŽRAl®HšÈAZB¶¬ªzñ¬ù2ðü×òf39õ:b¼¯+~Ý¡CEèFˆ9€BLJY]33XJ‹6ºlVYOY¹é¾™çÜ ŽÌ¿-®U…òÄF¹5)bH¹dÉžyåâ•Ãè»ÇZ Õ5yó˜Îò}‰êq+ z ë‚ Dl»‹ñÛ‹ ^΢˜¼7AtÛKõÀ¿Á7Î81½|ê߀ÛÐöiã´ÁäS`¥/_™uó2é{·rgb Øö-DÖ˜®ÂÉœýxÂKq ãrjEª‹çv0XWe§´R\…Ÿª{(ØÅkEi?ÙOH ÍÞwé6‡r2GйnV³oTÈ?:͸ªd30Ûés•· ÂÓjg˜îQ"oÉR4ýç§âU¸Þ=qm”}}.Þ~Šøg¦nµ]bämæS–f<ƒ(8´—úë2…S¥œß¼ƒù–&}BêAÔ^¬z–Ò¾#ô€ÂzÀžNáYÑγo/J˜þ5ÌVDá½}•þÛ‚ašC8¹‡äûæ =›«…wÝuü°°pKâ§Xò ÍhÆéŽf÷&–cM  Îtö{€ÓÎʇL/”ÎÍJ †öÙçód‘RNS*›íQQ³ZkÊøïšÂÍä©–áŽÔ˜ü|*Ïoø­&7ÇK–0F“¶b¼6txÊûòÓ3 þã="g,¹ŒÓ-mx;D¾ý{/C›¾¬¦•ÊÉ›ª•g~$"”IÁ\£ÉRâljãKC:¨2$i*ÌÙ¤q“3ùGPî Ö‚éb‡#’“+u>´÷É]€˜„š³´çiÛ¯¤•ˆE¨h1ÑVåîëyÄвÌ*Wn´%Ð¥~G:Í (±À÷Þ3lùbÌ}ýgk)˜qÓEt@ÏvÍE¼‹Þ þ»Ê_Ë ·µVVlV^Ð÷?¿¬A‡è^Š! NŒK#昪A½àõ$S4`Ì“Õ?<àz])L„ôŽèÁ ¼ŸR7´¯åh§#ÀF¯Ž,cNT#ÿ=±MpóæÉ%¥. fïFe¢ññãþŸøUN‘Áÿl…„”å|Hù¯$[%‡¬ÑeÉ]TÚ\¶vvþcõ¾Ôð¹h™ûç©i«;8Ùü ËQ-([•Ùž¶Ë6’lí´Lƒró-"Egy£ÊÂt¬!õ@mˆz';œh+\жÆ$>>»ÕÕݯ2w;Åø›™5‡ë”.´ ›pÛð‚Vaü—Á%ø++M®.] ÅÐV¢©a'À(SÖ4téË£¢OT’‘T @GúÃÿfhxb»úök—ý^o­çÎÛXÉJcK Vl@ÞÁpRU;º^ƒ ‘î**ËžÜä>BíÁy9¥_¶P‚,^Ã{!÷±óטץûlPL<¶†'y•ýÙu¼h‚™GFâ$®og&zÙèÃñÍp­ÚC¦ôp–‚ÕÝ=”¬·Î ó¹ÊÞüŽ(ϧv)Ùg’ÄǬ ºBM1’*ÝÄžâ¥}CK‚ä÷þûG§Þ>îªX8µ¸­°Ðù)›5)êí°¨äy¯³F3[n:6½ö³F`ÇO Á ö¸ÃÈI-GêÈ‘ùpýí^×m…ùiwvÍiíÆeÔR ùS3ûËÖ „¦­J·ÄE¶{l$ÇF'& ÄøQ«ô7ê¤jã$XÅ&³qñÒÒÜu6Æ÷rý â‹{›«4;¦”&áØû–¡Ž¯ÿk›·—™ÒÔ¥üWÙàQðSô²ÔïÂ*Ûï ›•àP¸Ë©@}ë£Dñ’ÅÜFRÀHKL±V¸X7!»ö[gE›¤«ë¯ŽõO7 !Y–© vöÁðn—dû/¿y²ÏL¼òú®v.8,µ£éM™CG²ñ½¬Ç·ëñþ?Ëã ŠÔá©Ò`Nûß–q`Á'™ù œä|F‰DYt}¡/ J|€Ó°ªÅ¿Ñ ¼þ³Ô99•ã3G=Óâ…þü q£ÃFSÚ1ü< ˜ˆeò ÷lê6¤Úšªc†’#Áü†ñ¾ŒMîá¿»2•ê’Ñ«±{Œ Wª\qô€,‰à—Ée„€]€(F;Î_ȹRhljí_ÃGEŨä†w¼½ ÇïÅ"p„“a~£?Eü\Fø¯eo_ûõr=~ÈÛ’-Aäë>LÓo±¿Üó°ûm£1­2XÇ$Æþ÷Ó\µÞšYÑç[17­0왨RÊ€»io¤)Q¨_ŒEúÚïe!SC½S%Äq>üí(Ç¢£‘h>+Ô=µGð¹p0üÌÑFÝf Øœ‚×±²Dî¹þÎâf…h§•–´åâ«sÿÍä…Öœ¼×rŠb8ûOan+(ô)Ç(ðF!#'/,ä_0ƶ5¨„9n Ùïߎmuóÿ>c”ghë¤d&ï·Ø'•Uè/Þ [îòÂmÐÃ<ô&úÌ풃߼PoK2äæ’þîÕ°èhq qþôSSý·aNAYë3=ô}Éyoïo3ÊùòýÅÅyK’ýq`z…HdüD7‡3ï6%‚ïËÖ8<ëíd`ø¤ Uª9ÛεeÉÞ>‚©L#Õ+hÙßÍ ²LÀšð<°øß‡zª:Miz¨™ãm¸·cèË÷óæqš$S_œÚÀ,Dæž 7´âÆ_!\á…ÙyÛ^#ØåÓ,gÌ·GôÇD^ôÂá¼·q™Š5p@(Ïb°¿ëcØ s¶ñ+¼B;åoùVÖ¢vùØgsõ[[4Æ)è ÕÕûÿi›ž3=Ùœm[m5ˆâ2›ÐùÂ[¤yhB‰€<à)}@ÛÍž;U¬Bšêu¤zàqS4pPõü­Ø0Ã.“bDèõÒÓ8§5Y²vœÏ[0yÈ;#N´ÃWma½ŒèÌ¿¨1Sز­)ï`­S–²b$9û‡É«Ãu;I¹ni\Æ ëb?°—|ÛÜ>NÐ`¼(¥ÔFÑì…gS&ÊD…/ãºú$âG9Ÿ•"ïS´ê¸rÎáw·×öp hB«„ßïÕq‘ï—«ÊEFVm˜—ögˆ0dº`ÀN k§°MV†êwÈÅõçÉ£G“«ˆÕq£‚Ö2$S9…'À`§:º£YòÊãð@T÷zÁNðøGÔûˆ8¹qŸ|f\$Ó£üB²SÔ²UËø^¬–¾P2küá-·£.©Ìé§F€ã \-ŸÞæggà Eû^tgwŠ|ÞÑ‘… ϘôÔæ˜È;4gZ©Žj™2Dðôñn¯xD,BNð(¢ofâŒ5v“¢é±˜¾Å§¹Ð¸ù]¤·¶k£Þ;Ô‹ê«KaÕ¼r¦ko×S/_(tñ}xL‘Aö’enG¾6Y¶ €É` U‘ºvën/%ÐÔ·Í<˜xóˆUš|â×Ú9"KoæKë4j_N6n^a*Ñ›ÞÒþIŒòD…ÎÄ¡¹Æ3£§$o2\©¢´¹âY­K­Æ;ïôG:FúÞãu™Ê<¯‚snÝ÷ªð¡%ŠJÿë û‡£KeÎáR.ƒsDô¶ B}¿mJ<ñ‰a%7$ÛÚHtì=HO€ö#‘ï9Ðûä,ì 6ÁàboÊßèC¾_Á¤¾r2CÛOQEk&ÈóŠ»'U7“YZk%šöˆ[fÄvEßî¬iŒÚ—Ô]À½#¿hoIÿêù䪖]:;&±Ñ¥vQ‹ÜêBŽZ_™žAf´/ìá²|(›Émî(¹Íb×Y;\ V”ÉèÜ?”;DRRÆÞ6é#Kñ¢<ßbah ¹ z±o]¡âÙvà;}Õ›å…ÃOÒ6ÔGÌ›˜ÑMlq«7ÊÊø6½J£…ÜèÆñR‚Ý®(C8œ5rBðj³KìL-b1ïòž²†@¢ cE+œbjd«[<Ò›õ^Uh³}ß"U‡dÞú m,µ‡‚zÐî"ËPÙû"íwÁ,Ý›}¼ö`çDÛ –«¶ú›\9(†¬8bu¸­×DŒÐöï `Ä©¿¡FYZÚ%<Ì“Š6iÃì›°úÂmP55Ý3ëˆQ²Ì]Ï3毹?åìü®"J÷ížžÓ6ÛÊ6ƒbÛKsÓ d¹P¿2ùÝ7—ÌÝíà¦9 šxûi°º øÜL~Y$â ðPˆ¶ŠbµÖ6+h£ã½ªÝXòÉO8rmkÀÃ.¨cн‡D+V±Á®¡·Œ2è¯ô©i ™‘®D þÛZ··ª1*¥æ!6}kÚS‘§ÍB‡AuÂòŸ¯.ë9ý¡Œì+T³ò¬¶Õ1@™æ2¸„¬ì1æF±b×8áVåù;e­q+ Qé‰rFå¸â䎱ÑÒ'ì¡ à¬éó+ÌôcÛx)Ì!äÑ ’Çýš·g‹üý_žç ßÿ%dLƒ¼G?P/”%Mïîð&(Î Ó¿Û£Y³%ì%gužèDc7ø¤½BÒèj…|Ñ\8Ýs;¨{™ŠÖÿï~Ü+›Ž¾>+géÕˆg†  ÿ¢3΃4C¶ñõT»v`RÉEÂfùÝ”Ð7?ðŒRêÛÊ•"q.ÚÒÕŠlh^JÊJË*¡!¥=·4€4¬Mï²½!ŸkK¥Îký¹îTI‹ Ö"öB'¢_pé@çrµŠnOxßDåëVRÉ®šîî»%ò æü))0êœáÛB¿÷ƒ&&ß ¹ŠÊ×ÐzF"òáïìGÓQ¥ðW¤ŽGö©’Û—Ï–êêLOÑ…÷KæÜ;Ç„a žÈÆ’sõ'#1©<êTì&ˆ ’ ?j¡f/IîÄ@Í»ý{>H·1°Ù–-­ÉéÅõÕJ=«©}M,ÞxT.µ;A ÜdÁøAwı…€†2·.¯dyoMrÖ颓©„³By<ßX˜ÚÊ‘•‘´·¸·èl°2£Ó,J¾ Š°’ý!È0vãL£€’žŸO¾ix‹rg«x#+?Ù5‘-Ad‹Û7§^q‰ ¹$ö¿€cTjñöÙWÜ[T!ÏpCÁÈœ?O€Ú'ÍßΚGI¼×<ú‰Âò²ÁZpˆÒ|ðWAˆ†ªÁ"ý³Z©ªHï¢LC§îQnZ|=·!´Î3!|3Ï÷ÕNQ’1Äý½ úú1ê{zúó.Yîñ¸?î½T>Ó9¶MܾÜM¹ èÐ@þн¿”yÞ­Ï>ߊ)½*mV]—¬7ïIX›šïäê’] c8hð&Ò0ç¤=^5ê•ÓÞ¨hû6Lò±ÙÝe…<Çâ[ZKÈgN6ÄñŒÇÆ×)ʹ~G†{%|gpúµÇÍÃ*Ê—¾8u“ˆvà`h„t¶Óékî¾<àÏ<—“åã‘cÅ.4{Qà䌧áoR‹U?¡×l¶¯&s6èKIÄ„ÚSU ‚ÞKGÇDn¿ %_­ƒƒ,޶5 i¹ý±¼Âf¿H¿ñ1Ò¨ÈHîgÉÊÎõ©!÷ÊNß„Ê&t<óa®íªi>+5¯²ßR ܆¬…·½§$Άÿ#¡ã[#â~·)Ĥ1^õh*].í*cé,6œ}Ât:d£êÍ$”S*cá^~'¬Á:ÀúÉ`Q(@Ô¢„ß:Q¶žªòï0½ŽÏô.<û¦ÖršåÐ'i#xñ]749ΡŠÄ¢˜²è^ÞŽ9G€6l<ÜkIØ.²Šy&æ:1/8,ZÅyf>cøl iÈi]K7sÁT1@Nb°¼+{nxàþ-âþ@šu…k=ážÀúD²©õ%0ÂÅØÝÕ¼à”5'ã ³‰¶û“lX=ï-Z &®.ÑS&ØìßËÛ^b‘x”„jÞ/#^‰¯‰f"ûz#­#‡7hpÖ°øóËŒñO•(¸îWÍYÊx…‡j5Œ´dTýëºÐ~miôViü>rX³èaðLÆ"oœëü¼Cá2 f² l†íß^hÔàà xÜëõ |{\|—ñÑk#–¹z6á_uþȰ™Ø‘AÑ£Ûq$ÿܼ=½ñîMæÏj x]ÓKq+Î-nÊË"£ÄÜšt è+(¤?0 Vó’{†.;ïþ°¨Æ«d+Q¤œRÿö 2~÷Ñ7D¢Rè ½®‰º™€#@10 zÉ’Ç»-ˆ¢@*©·  |¬)J 0F[¹5¡¨^C¢“ F‹©‡R:ò}`úÁw,Ücº¥Ï‰\>òA=»ÄLÔ~ùw©ÐÎ*›-ÀTÆ¡S5-ãaæÚ÷ËU]³H³ª5^”~kO¿Ù7{SÒ/ðmª›pP ¦í½éˆð&é-ázqtB)Á»ý:“K0uü¯˜í­L¡£ƒsãƒ`iöÐó?€ÙŽ_iššS‘«¿’-æªU¹—>?7ùâhñÝŸè¸ßóê\‘>ż¯ºrΦ.ÄÎmE¬õÈ¥bJ–2ˆÉ賓xçóUh‹º‹~7RŒVqÆ-rzÇüá— çösºëœ-Øs?Ñ8‡zÛË­œÆÚ«–O`ŠðÔ*ÊkÁÝy ÇSsÉ߯ñ>¼â®º$Aγ6jÏ#eÎ3!¸àˆÿ®˜70cx¶åóœO*­W¯iU‘œ7×ïÚâhÐê@68>ò­'–TþFöÇ`³NTþàÖzÒé‘[¢  l|pön9ćWG»}‹îuò<É?uÆ&É>”‚|l9Õ)Á|o‚A};õÏÀ 7á%ùºä~!+è¿®‡ÞB>Ýô{`ºË„ú=7€AVôŽBnø]z Ú9_•–½Å›T³¨±vÙ#,ÜÓò ®`j ÔÌöÉAPLÐ=È´ÁÊx¯~—4h §Ó¢lÏ“*8R ä¾+[^9¤J Òzëüp•1‚?ŽÿÎò ›†Š,¾Pÿ±\ôø\U0»c÷¸08%žz|ñ·WÊ‹Žy±GÆ ¦ÿÓâa w-ÅÁûñŽa¨QTè멪AkÙS¼yy(!:êÙà㊓[õ˜{Üx—ŠN¾&—» liºÒŸ&#Ùž„™ŠõCܧ½ FE'v¶£õÎ…*Æ„nQþFJoÂÊûn Ôœàü´çěݯqór4 tN|%¼ÅpSAB¢€m4ÛÄ›€EkfÛUÇôõmž@èk~eÛÓzM& Ÿ Äøû“€ãôδ§Ž&Wòræâ:%äkøcÏ8kŒBwý !ó7šŽ”Wöwq¥Oœ‚ÂE*3ØJ“ùðÍ ; ¯’ãÈYÅÚgœ?Ó~ÚëŠGɆ†2àˆ÷µ†E@UÕßqù’8ÉJ@üŸ³UNiÏÔv23a ˆš+¬ÑÚ°ôwwßÛ=_ÛÒܦ’E)>É.‡r]ÿ‹Ï‹súS馇ðÈŸ69-é!U!­L…wqRÂv*[$Èâ Ý‚pˆ¬§|#Jr.Ž›ªMöp.8HF_D"2îÆîü>c*‚œŠ¼q“ªžÜΘø´mW>™ ‘úð„m n7á|*Ìh¯0Ôú¾2\‰ë‡ Gð*=;Ÿ|”>¦ƒÄœCË–â°w02¥Ÿõr7ÑCZ+ö NÎ[…m¾[þàCähh½ÂÆÔQÒaØ^˜ 7[}[0òȇ…»†ZÙB³p¡÷¥—æ>.ްï–`ßjÙ×32@L¹üôµ`½ ™î:ðþ3Î-¯Í¥? Uø¶©AXOîz9û— Ó"#fÿÑ4¤J-7nAàTfS.fðJlõaªê¦™RaR"-e³Ka$£ ÄI„:.Ý aÂé þ: P@&Ü;³ÉtuT-ßXräÃu?îâ5|±jÿ”.VO,N~ ,Pê‚ ¹á­J>úíÏHÜg@žÁÚ%r¥x‹}i[»µ3ÎUã–sÇ·_‚4Ãÿ± Ï¢„Ú·•+ëR[вֻ^Ü&|Ñ@ûQ¿0z¾²ƒVÉ S,ÅÇKn›+u¦ÕBwj¢Í“9í¸œú E 3û•ܺ_êJ!™G^ñƒ^ÚîwŸ>l®'ì…÷/‡¥àhJ-‚æ9C©¾Œi,Úã0;Y­í»‚ƒyQƒ¯«Š¿äÁãËbY3+–ð"e@u§ÇrŸ1¤EoÿÃi ä9Áy;änAPÜ=R§R1G¬ÍT…{P”kn G°µ 0 ÁFiù¤Z óhiÁÿ¶Ö¥rj-¦ßpB;ˆÒÀåXEÒ×ZVø·#ù‘ù²+_”ñ¡F¢•;kU¬ zç?55Hq³ôQþ–´#M"&€HD5hÂù'ÄÏP^ŸMìñ#¥/š8e¤¬ÎcaìŒF{M7ü{,ïJ¢‡QCi¥<Ú=õv­ÄÈŒ¡›ú·Š;GN4ŽW.ö+ ±CfR ßcÉecäݪ®¼òŒr®ïÛã½Ju«sˆßG–"Á‰yÇÜøÕvx¬Þ;±j΢Ŕ4µ{¨É&NDÕb7K1nžoX¶é¥ßÔ{)½b¡vŸPËßx¾šaÖÄPLáJs䜠3¨ùr/f&>FÙº¼9K7¤ÁÓÅ’‡&ÕQœXƒ¬ô†'mFî| oïšDF5ŒŒl[ì!Þ‡ó°ðqJ/ëq´˜×ˆõÎCð¨J:Ȉ?O Ô/±¢”Ë·éÆ• æ„%Þk!ìôw#À»å F§ÎÈÅÞÖ]]ÓˆÙàDE+r°vºq¢øÕI}ƒOôFT³EÍ|Úë&«4‚³G€®¢žÊšƒÊÄz]Ági•.¨±é¿Ae¯¼;/ÿ›ŒøÕ›A5‘í“4¶Í¦æÚø%:æGK{¬¸†fFû†ÃÃ.5ÍèY~]ôåàöGæß‡çӌո¸Æá¤Þ¹*e $·§”\$}õ$YÇÕš^m‡wÑE›®Gæ•L o{qì2j²@àh9J°eGíÉq˜L°G¨0f‰-ÕžAob޲ŠTÏë¾­:B³|ªI¤ Ty3æ9÷@¹àòLõ>”vwê§FmmDðjjŠá•ðk«‹T¡K¨§Z„¦Äs†Dï,~é;U‹6ge€%ÓåcdȇØßùÐÇDhc®Ðp$GGBkKâÜœ¥´'óîe‹&„õïÙÃÄÊ€¯Xù°ïü¹dñL' „º ¿í«¦ØèŽÐ²n} üüPbòV}õZ1%1Ì7X„ü ñ”ç¤ °M>[‰~Åh N7…û´”‹,9N^.$œU}¬Î Ksy©GPæß‹vÞqLõž ,ÌŽ¾…âé<µìlú«z~Rïs °Ïén|ÆÏÚÅå)DsI.“ŒJ:bá¹Z9¾œµÎåî»QîTn ¤r;Î&¡Ïõ7ïQ=|ãÑüµ ï™ >T±Á–"eH›;¡âî¹K™Áñ^ñÐ_+øÝ“–×¹@êg Ï…ŸÚ vØ;ûVÿ§²-ÚÛ­ ðÓ0ð¯NHµ„V¸a2¿ŠJñWy3ŒK™ ´>ì[r-9P¥Þ„Ú̘“TmôÇèZþŽ„PP4wÎdôµJéÞåîW ~ØŒ"!ïç%{¦ÐV0jz/ s8ak« FKç2þø€ö5¦ˆÙÉ õÁm· -þ—Ä ´»U³ï½èM#ϯ¶ìwSøBÚ–å\ï A-ä“ gžT»ÓòíY¾¡ Gmq¥It£(t4¿ {Ü‘ŠÝ@Ô¦Ç\[NvœÁÚ•öóÿÝaÐèAÚHÖ ÓÂsF" ÏŸƒõ£m I8†P3¯õÉùÀ"Û@`*ŠñÔj4!¼è²°XÅšpöìB»}kOƒÕâ0`뱬Á„}@lRF¶ªî©.JÖX}JþÉ*»¾h¤–òxÃшàæp©€À¦ÿÿ®ˆ›ù¶­!­«±µþdk[[ëÕ\¥G¦7œ4 ûœ|‰‰"*iô³x­¨dzù¥¾ð[¦ï?þ‡âÒEГ°4/bË~ðCÞ¤oÀ}Í—]í¥[:h€7?R1þ 6Žõe8•!YÅ*Ø($Öhè@ݯðc²ýë6Eî¨=/tìpÍìET2–ôsµ5Ø€"üaSsGƒˆyì›,8Ø\ôâM­?*ù¼îwÝâÁµ·úøáê\½m@Azž#ãxZ@ï+s RˆvÔðXÓtº×çˆÃCÐ@†Ãlsès€„{ÈŠ¤j[ËÊdwŠÝÞ½qÈe¶€§Àÿ*™óLSkjên³•¥Ð…d?`‡]c·áa.Øp4Rá5x$8\å©Ý–ZŽBd%^Š%ÞKÿé_ž>R¿MÄÑfI¡(óeÆqà@`Å;ôXý%'¡re `I!M ¡/úGf;ˆÿƒÄˆJÂ)ûŽ|îKY[vL–Ç”(<ßʵ‚'k2nv•63†±V7sUpZ‹l¬ë|sXU|sÔÚª™(™q¬G<$f}»ˆ]ûB´v•df fæ–@BI©›ø¶@MTè ñØwÞHá†Bê5Ó÷)6yÝi=Ý€6>Ñm‚SžŠ‚“6]er»×;øiåþ<̉m„T4æÌSùP"ïÒ| (NØ ‘‘؜NJ[a.q¿:3sHBhˆC!¬#Õº`©íÄ´(Pc2NA&Öpn[A0Ë3}³‹~ ™¿L9›qYzpSú‘€ÂEŒ¿Óå*ëÚ`ä/cz7ãJ­~”å†GC/SÈÖ{Ö¾1#º†è5Ä+Œ×lyÔ~¨OÈÍŠ˜´è¡7ùt Ñ]ßøÂØ¼O? cìñç£íÆÊ # ­…”jÕË#o¸‰º·èâ×[Þψæí3e+Nƒ…9‹`\FÓ?Ž·ŠNç3gLÖÚ©ùœ-S³äV…?ðÍ?ø \YÉþ[‘»¤ c?#m]~2Íý@¨™¡­úÓZNÕ(“EÝaè‡[“ö¿¤¹à€í;U:zƒÝl¦UR¨ è4ô•NK7túÓ –íúåÃ.«„Ý·"í!4Ûôt"Qˆ«¯Í›>uæ¦eçͺÉKsØöõë,¾„ñ¸4·©Ašø£ ‚…GEuÈáÁ˧ÿ&gÔ¾:$Ecvíì¶I£ 6Ïé¶ð–ýsàµFëO[ Éòµb{V‰ÅøRÑL~#q[ žÅ‡*ü¬/‹T9âgÒRõÛk ck ±ƒVºÉx7Täh«Ï?ÌhÒ%~ÊŸg1Ó0Ze¨¨ª2­¡Z©¼½OåT<§Å” ¯G5¸…K¬öÎ>h:Û0â O¡ÂBtzˆÔéí{:ùê\ž{h]£ã50Ävá;j0̦g?ÓxÃT抶©Ð,ŽøVVwœîã‚w|>˜ÄT[n‘͸Â,Ë Â^ÂežÍ]çLa§uŸÎü×ú¥ÉIêÖýð…4âî˜ ârýt”OT mÀjê/V¦0‡ šÑ§Ú ­VÒÙ[ÓÔA`DNn¾T"üŸãx<â}¡"N¥Ë–^…=8 @¥5¦¯by^•±rø6±3…£ýª0°TÊJÕ¶+~B›Wæ¾ý€¸ f¸¶dKWmIR«QW\Ýßñc߸~sš(àí(­bÑýë‘NG ~øS䉱$Æ8åR•U«XûCƒZt± ! |>ËùÌß-~qÜÃcJ⤥›mhv”!´rĬòäfúüÌkÒn ° ºÛTýÊø¾?·þo–‘‡W-ªDc ]Ò¥èË_ -ZÂÒª d¿’§•ï'"ܼá(†Èºl¥û Qw.æìI~ϧE«¾©ƒ*uÕÙ<^éfnv¼ûv>Nú)6]“«û6 @mX·—ûÆ'[ AøÉß®åP[o U•o s‹ rw¶ŽLždppÂjê©–“7N/€WBé§âX6U&Šñ{ôH¯1e(S¬Š@XôofäÐ=üõO²Ÿ¼[ðã©Tßù!W óôÇÑ_Á)8W{«þœuüÕs¬S£—Z1ééX˜ñÓK4®©Ç>Ÿ³PáX$Cà…Q¤÷âJ4CœÏÏÉ/ŒŠ8o3pÌ¢øznÌûa®$ñ¨sÂÅÇX ƒ¥ð%wÛ3÷ê’Šµ·o@YfD_xy³†JTYÈRG&g¥"ýf1#»@m͹TâKhņ¦å/GÉT€hUÁ&ý%~W;ð£ óéu»yà‡!ô·\ØD2NîK§“Ôĉ`¸¶óz·c¿‡'­4˜Ñ&1”ìêjü_¦nÕÉ£¥Þ8ˆTΟ+)‘¨¢ ‚²&Ò*fÇMζAf¦UXÌ!‡vXô¦"•üÙö»z| úº£Ff¦ãti3¡`oêu`acþ÷d@ám&±"ÀnTÒE½%PkE$ÚÚ´rýÌå‡”Ãø‡ëÅR`õJÚŠ/;Ôl .lçÎO9|u‘Öf^±A uëÖÁ.%±.ot€}!’›fš‹HÔí­Ž¢=B2öîüò:>¡ìDÿ…8j± ö5̦vbbÉÞÕÉM»k^JöfclCÜX2?¯½øIV ¿~©¨_Q'š¢*—ù´RÉP Æ73)ÄWÆÙVÐ=5 º¼ÿfŒŒèXµƒwm7ò$íÙ¼D¦Ø(+¯O­ ¶Î~eüŒ™¨Och h Tˆ}Á3µÇŠ?¯!©š—G§®ð8 ˜ÜðäÚ ŽHÕ*éóÂ'<@*¥#Û‰^kOAÉÓò)ÃtÎXº¬s—âK]´%ð#ÂØ¬°/6b™›Û»ŸÃHHñòI<1Á ,A©7%—Ô¯B~-ÙhS·¡ÎMã©@OÕ ?¢êçT€ýæ`õú³„y Ì%óªx‚åà³<=+ÐŽÍ?kS³#}èeä¢n –z;¼-Û¨i*à%ë¨NøÆõumS˜ãšrꪺÀ› M>Äb ±M¥\Ò¹Šwh%iê2\ÉÁ»tX=Í =jÞ’{–;2M¹>ËnÛ¹åéIž:.e"UÔjx»Gu(Mã› be¼ƒOþ7­‡;£Ü’~—æÉ®±Xþ$ë» 1)[1¾. ÁQx}F•Q>5ÙbÑa6«|2Õ9kÞ!®S8>†€Ä$/+p¢p@a&£ëñ…dÎF5t"ŠÔì3y]²>PËGâ=3‰¡ Ô©ÑÑÜ›wL—Ýœû(ô4-÷¢\Ü(½°òèd³·/W`8YUÝŸKËR;ÜW‚ú™©Cꉙ|¾Äùµ¬ Äþ$³·¤O~h'\ñS;@qã•«¡ýÝUötŸÆBF̦×k8iv‰7øç­ã=Ó“Åmhð0`˜þ%^“‹üNÜ$’gªÛÓ½)…ßâmŽ/`ck9°ÒéðÇN8fX›±3åC®VÉ䫀esž –ô ]ô[ðè³Ön”-ãì भøF&ðp®Êž¤–tU¬Ú¶,^V ÇÍÚ7:²n3XŠ4<âñàVœKÕ‘{Uµ³hó 8@XÔ·%–‹c0wHbçzˆXάB×EVT•vʧÜQú¼ö«Av¶nÚ?üMþ_±õy24ƒå©Ù‡ÔÚUÆ@36Ï®A|hö7Ž(MrSNRí³öÔ·Ø1{ÀCè߯‰#’Ž‚· ³™à^m½`pÕö<Ðrù½Æä›,¶ÍŒãj`6Úc÷•* ì pãÂï)|O{D=Pµwî³Ï4íËkò>9P½Õc솾ve&û}ìt™}5Ó[ì¿:Ðæ… zeIkÕRÕ>÷Vš»Ã7âÅ©ÒÑh5?*«w¨¨gq]§ï ¡×ÆÄg´j¬„Ç׉ –露F’]Þ¸ÝÈ~‹ãiø_ÈGæÅgÔÁW¶¤1…ºH–öÔàÑý>F¸>þxEìê²d ¨Î¶MuÄQ.”áw¤ølýZL„t˜.¸‘™CN¬>¸e"²– ¡T"4¿im˜Õ8Š ‰Ôè9EA ¬í\ìJºLe?ƒIÏmz¦ @£Ñ…AÏÏkv3=ÃzÙ•.Q®½§s{PC—“¿¬§f"÷j (» DS¦NøÇ¿ýD¸ÿ‘rª¼“ªw®c—ý©RûäÿË­=šLö]~ORó|ÜÃr…ÏÐ=ëùÔAgcÞlÈÏð~ ©\$²lq9Í†àŒ¹ŸwBhƬ|Í ’®8ìYlF¹-hò;›ˆÊë=ãek¥L€Î\1Ò{½ˆDøEÁâ&i‹óû€ò…WñíàZö‡oÀ‰n¤»—”é 2)UÅÒËÏ5žõ|ñwÒ³S!èd׌æ\þ’s§ïl‰Ä~Ä>oÈ—~¾ó˱R–¡ä­Ã}“EZ@QáÝW«uråKf°y!“—YÏG1¨AϲFú ÓÜ!õ ß•‘½é<°ÀV-#q‰î+÷êiÈU‹Ï—J%q2¬Üu { ÉaÞ뜪¦e§#‘š¦Á…Ÿ“D•"ôXÛnf—5+ÈjÛœ¤wÆà.;SK´ŸéGº1Þ*ÆÓ‚3¿þÁLR·Ð s!To†GkÛ»cؾ´„ž¥‘lz?R ññ&3Õ#¼*ö®Ñ¸¥¹)U¬Äú‹VDo±f §õwtÀ`/‡ÁêÈ[9’èch7æ"è¶T<ÚPý"B#êzrïÕu—$fSe(\(ðYê”—c×íõ­©‰ŒÐ­oê°`­„?gý™Ø@]R÷E3R2lž×"E2‘¦MÖ‡3Mb¬üꂌÀ.+ŒÁSZÉÆÙg1¹ÅˆbBs¨W¾9 yÐñ¢_„ »G£/&ñ4²-¦J 5ð©'ÿÜ< R¸Ú5o% L ,¶°®Ä´™OÍYV—²ÍàPŠÐ¦+-òSsM^Á/7 ®/q~§ìSs‚Ol”dµ?-é4X:Ÿ2Zm‹UÒC!hÅ—eÆÍ!žŠ¿eÓAÒÀ&®E½w‘4ÆüPÐrÉZàgdEÏ"µmÄ…-> üà~ìEËÓ|pï+é4Ü%Êy¬3û«îoéÁÃ×ã uæyR™û›+3é1_ow¡òšÒ°‰”ÂöMÅb®ÈÊî£7Â<ì{6ó_%CŒ@Y$õ®ï,Ùý“«ôz-~(|«¹ÍP«ùØw!>o±Ÿ\áªÉZ;CZýMOæ!@ku ØÎÛìŒ5ÙâÙrj=&~ý\[ôcJÞõnÈNBy'n¾P=}HëÞÇì&ñë]Ú}zÙL;ñSzk[Ûý¯ãͳ0©”}é½eýà“‹äÑs\Ó—»™á¥ÚŠrýôp³};ÄK,I·œ±ÆwVØÁ4sâ¯ålåÛmX:ƒÁÁïo4 ˆpµï&Á&€ð,"¦€E@›£2ð¨±Ð z‘,D›Û¿•t›€}LUÛéŠ[s0‚Ë&vnt÷³ÿ¸ñÜ(ÞUf$èiÌöö3yâe=4¦7ÿF%ñG5bv鱋é4¦6ºüI–G!½mŽø¾Y(þ•6Ä–Q [\Cc,¯.~Sg­Å_”þúì^sÏZ)ãR4•Oé¯y&c`ŠÞ³TÜÉ|Í Žítg/jŒÅ¥RÔ›á=±ÝQšÑ9 gï\§ÐX3úƒ:»ÂÖšw§÷߃—,^¬Nü·šî…ÊjµHªÆúOçtÄ÷Tk£ryÑ 2'ôXQâÙ"‘µó9Ú€nô7öOýŽÝê%ñ?[:´±&4Ñʯű•­hœHb%Þ•–û¸3愈@¾äÀØp–ÔDŠ0/7Ï¿p©udM*ÊTŒn  Š©±K¼ñÿ“ýé“.]tÇ‹»­L™èÕÓ)¯!çØ˜R*ßT†pf¡ é™o<2Sõ*et›–ÜþýÓÛ⛑-âŽý¤™ty±ÜY+Z€СERŽ œZ™cG"Å.Å¥ÆÅ¥ØÛËÃZmÀ/:VZ¥§6Ј»Ì“_DZ>„²ŒÑǨœ5?উèR=¬¶Ð-/`ðs.‡ÂÇ·…FS9újÚ*x4Û$â³ ç¶“vÓ•"רbTzbv<ÛLÔ^ Ÿ-QŠM‰Œ©@óºÅ6æv`˜<Òâ*/0µå'ò¹×V³Á›)•ƒQöJ8Ks¸¬½þÂÀfï Á~R·`,%æ0xRÅþôe¯ „Cu%kv˵,¤º›v· Å!EòùpýÙ Å–.¸iO¡BfÑG„œ'ìùêK"`q8<*bXÀïa Þ7ç†öˆÈøsø²ËûR_¿ ?1«I¿éö“×<–¥¼Çt„ƒ hˆï%HðÖUÀ³4FÀm¾C’‹¼+¹š,ÞÊäiPì/}§5Þ§vðsP ëeU«±hÅ0ŒŠÇ‚L½¢±P½÷CôÂi©Î ¦HîëÍÖÔE½Ê¸ãye÷Œ«Ytè6}Ú_ùÙ©å±ôIlˆ÷KüŸÊ›ê¼Ÿßó¤“oe¬´Lõ¤D6.jY5ÌI.ëŠs)¾A±¢Ÿ޹ð‘ÑÀø»‚%­x¢Ëݘ\å«K¥¹ÖÄ1Ê"Ÿè3ëÿ³<‹nÚ«/ÿ›le¸3.Ê}yþàºá…Êömó>\i纊h·œe·¹Uû¾ Rà²3q)P¾Ý ‘HH°ÛLÎ4ŒÓ…@1.ŒÎx•»ª¹ô²?½{I7ôõ’®öo¯®@&îbŒ*˜þ0ÞðN¯§Êé¢ÌƒÉ:¾%àz›<\±ÚI:`sºZ1¤éoÞœ¯;%øyÚž>$‘ ÷ q`»1ƒW*-#-è$ÉéÊí,Q…›wG 0µâ3ÆU„íñC¼W ý3 $7 +(Žý‚‹3u3• BÞôæž8ó–'#¢YÏŽ`ŸîÛ†dw¼× )ŽŠÖ ˆP•>·ì oïÀª‹Î ?]ýÂö…>òFŸ¾E6p¬¨R‡TÕbHr–`‰2ûÈrµúùåÍc °iZ iáÛsÀߌ§vH“1"™S³`çgð XÎè ‰€¥Ö¤:yP óŠÜ˜ Bao•½Á#U&]×y3f:5ÌZ^óMiÚèÆêܨ±}H_ˆ¾¶ÚÊÓVfË”wŽ °£›Î(á¾0Ú;ÍY^@äü›àN´øèi @[íI¿¶t×.ÎLí“”‹Œ‹«é»ŒÎí£=«º`ô¶Vá‹iH]£@·TþÜ/üæc"ù·‘´b ¢¾ní‹zìà ÈLª*ü™Ž«R†…bŸÍ\Sugu䟫B‰0Š{´ÕDq{AeÜñªy ê8ªs&&-”À8‚R”çKÅ#´gՅ¦QDô2Ôù3óa½¶j¯¾®Ñ«BÙ 'ÿœA†çÏœœ/±©L€Skí#x[©‰ô+jð+Ñvlˆgº—ô‘ýß»tˆd»»´u´/]ÈOHÍååwÄf¹„=Ý eöTãD6ÆN~=hýÿm»«ŸO>ƒsꓓߎœ,Õ)VŠþÀ œ†rŽ)3ð$f.æ.i^µ Ò×{él×K8;~Ä&ίß:—£µÝK;R`nhr"Ù?fúÙjòø™=äsûâaèa©ŒMÚ;Ï?ÁÂm|l[¡kNpD¥=½³|»+¡ŽØâe§¹q $Œ‘­<¾@kÉÊM³AÌÒò3t ¢Ó]‹üÛXIÑ+ã¬ð°Úá=Ñ+U|¾Ì­Ä%Õ×°y‘q—X'É"³Ú«¤¼(jÝYò5ÐZ‡ºD™«¼ìÈ|\º©á‡‘@x›ûI©ù1O÷xÙ 5¿7Xgà=.sè…XÔ­eò0cÉdmSŸ“âû±Ê~Ç&“&Œ/Þ ‹‰8œ\ܦ¥ï-£ä^ñ6Ž)˜=äÓ=_ìËç2oÓ=~-¡¯ÊÂS? ¡Šs ‘8®?Z«7%>­Íh8P†¨[ãp— Í¡|qê•ÏÉ £1?ŠI ÌUý Å5Ì^Ÿå€ Õ»nRnOóä««1òÌ×lïàsxÙòPZ’Š£8*CA9ëÏä4AÅ'Wä@ëò˜¤Ÿ¹±m$!ŠQÑ&é=°ñÀCêí±x@;$Z¾ÌíyØ7,}Š©K f‚<Œªuöáøåñ¼h|ÎD©sxTÚí|&ÌOª±Øm ë<ØœIY aäY@¨Yà^û¿Ã­dŽ'¢Æ0qÝJg2S}ïÆ,]m<;æ<üLÙ‘ÞPp8.\q¸÷ KçBžù¸=B›ýÙ\\B6/(›xr}ºøqøÞœ/Ø‹’7 3=d€7ÒÜ;w•íÕÂ2Ñ^„y¢¬Æš‚:æ…|Œ:iÉ5­¼Ë±ö\*þ@l¬Ýƒçѧ¨^A¸‘0b€v+ˆ+J’PÂ;Ã' rH‹§E>­YhsWƨxwuD<ω{ÏâݯÚ©Î{„$Ù½SZtn«Ú—Èd–é}UÚ¼šÕÚB‰ý#³šD÷væÚ­Ùo¬GçÜ©†þ‹BçP™ûÊYªˆÇÜ$ÏìM¬“×´ª÷âU‡bxâé~‹ ”ËÊ”a«£Æx­ã§æ†*IdL­ïx¤M#½®©s*W},‚Í[™øQ^N9ô:ë3÷/<›ÚQ´Áy:®)ߺr–;ïeðUñ@Òœ¤¯â6,ˆ3`öîºåÜKÇRÞЧWUHxSPæKY(*^=eí(aÓ^<–W˜c‰ª½TPÆp£ôdt[-Ÿ×i@ý‹¦hÓ{QÍÈæÖÐép6k]ˆ¬f.âÖz™Z/ñDæ.ŽÅ`º§1@GÛ?°zÁ'%Ÿ«TÓlYwíAêZéSþw êkF²€s‘vÛªÐïãÄôûtM¸ EÖÁO­|ËΊ±¢iD­ü©%Šoø`*gèø3B@‡*ѶèíBÇüÿTl²ö›^lA4õÔd<@±‹`?³&Ù?„}+môPþ÷\À¦Ãñž™e`­Ý„÷tsŽzrAT“Û‹j>ÆYŒ’’Vh!/Hœ +…Á§³…°çûÔÁFgtqc8ÚÝY?•#ᆂ@«ªøm\ÿÍ1‘ã¨EGP=”ásQÆ=ï0ù²acd&‡ È!iÔªóA‰CmJz¶ü í¿B¦ZTÆYõÛmJì¸7œh­0[ƒÈŸgÔæfA9Ãfqhâ§ŸG~,%ÉY¾xboÂÞñg rº`D•ø-$–˜€àÐ1–™>¹+¸?%÷‹TÑò$hqi €·T¯É‰0² õ±>4#RÇ_£Ø’Ù'§{óЧobÉi;긛Ë€“½¥§^çóÈÁ²&¾Ž€×#ÞãˆËÇmÇÍr™'„óÞi Ém]SÀ|Ew§-±Ë|+ª£–-ç“Cž7BÊ.H¿OÇ¡›^§²ù噊>¬-`þ#Òàsã< ö ÀK»‹ÅÄóÄIéáÞiØ}àÛSN»IšÕ±Ã\9“̇.µ aE…ÙWnŤÁ¥¡qÄÍ(^¾yå #Õ ùPƒ.Êí:FÜïj EûÒÛx‚C/ïµ}\Çâä›ûƒY:*+ÝM~œr˜ø†},—yÓ?Ø*^n C¯1Ò¡²²ÍÎ9-pƒ2·–MPˆý-äý¤ßx­:xþ²82jo¼7À=´þa5ÇiPÛ.kÖkÏHÁºkb®Qp»ø»†qojdçvÄ7^<޽š‰5©R&¬{“N]Ǹ@³Æ’Áê -!¬‰ŽùŠØÌ¹PøÝïr¥:­ rhè¬d Þºs°æ9 wKu˾Ñ,\4xÛD .ÊŒëÓÇxeÃèlIÁ#ÚOɯÐklļEDýÅ· Ø[´CŒê¾å ÅÛknÓ}ÓÖj^sÍEè@¬‘Vªdº¬O¦3Ì6aô*fq+[×e©N™T•ÑpÂ[Ùè1·—Ä¥Îë%ôYƒól¥©(Õ´[¤Ý÷ûÃà25@.Œovƒ;PÅiß|Åg:8Äàw'§ù¥ ÊR)™ÚÂÃRl w¢ÒåÔ`íÚ @l…0p’Ñm8ÚyX7N }wNhÿÿ]tºË2o˜¢ÆcYÉ3°.0+D>ó×maމCáA÷aK¼a$þ´C¤¥§r… ±ãÑ:²rDZÑ@ »-X t¸Ùiÿ@Üð×}UGU,¼Ë¦ª!<šV B[µ“פϹiŽôuerwúìmÔº ÍÿŠ Æ×r³Óÿ5»’<ÕÛþçh¿Ã#p»5±µÞ ¹ÛÜ8| x55UJ‰¶œaI%ƒ&Û)ör€@ ^ïP í¼‚Ÿ5ˆN¯ÈéæôMñZ1ðV.r<ƒ%ué ð,§—d`#mþ‚×6V©ùÒ,¼NE¸`ÅM¤n]O ˜GÑB'¦£F&U÷ÔvŠÕâ>Pch°ÉöB'ù‰Œ,ÜoRrùü@ÜZµÍJ JC“Lœ#lâÒÌ»³±+Þíß­Ÿ,íà¶}ÉÖm’}+›ƒÃˆ Ñ@¬¶ KŸæ+u mëPºùw¨u[[¸mÏâêÆä¯&•[.›Fà§}k4dCõ xá”i(õC5ˆ#sÊÞÊjijvGé/‚8ŽëMiRÓÃÐÀ˜'Æ âvíª4¹y]ú†éÆÍ¢·•Cö>Ô]_¼"?bà:C§>îýéõ¶\ƒL?öœàð|ÿBl¼X5ÏŒSÒ_g«›áØ“—ƒçô„TDUØLŒnÖI]§Mä6¸À#ñÙ½{b©ý£¢îž/ò®8S8 ^/ø„ÆrÃØäF³ô#ß U5“©àPo8Xä¼í‡ãÁjí¹8D à¤ç'+r+C¸På}r²m…¨ôÞƒ/[’ÜÄ«WPÌûÅ0Ö4!æ±èTùø—…ÞÔð%f®zµs-ézðqC€VR¹‘ÓË;6εhe¤&I€íhp2…3ì?#c„ïݦCb~ÚjW|‚í\v˜ù±‹…ijö-„ç°Ívñy­’aYý;4Ô!(ÊÚ8!"×§ÌL<çØgµµ\Lùh¯°'‹èe!Ñx’ÖTaU:¹òGP*»¸+xI™[ Ây¿°,ù„Q“˜Sï6:04´()m˜ÊH·ÝzÛ¤5<öY\äý •Kp º*út¾Í[Fãì›ñžåì'K¸W&Æ%´àeÍp.Å} Sò^vm¨<…¨iÚAX  ó¢K¡VàÑšÕïháB|;ÓM3;IÈ@\¢=•ðS[纇ˆÈ‹ªPs÷§k§˜"ð±F‰ÛÝb©œ àÓžƒ?nTwqü³Ù=Ãþ8.[¨˜¹s b¶k301‹öU°=Œ‰áP?È÷G*È[rÉW2°ŽêDˆ ?{kœúR{’ "þój½>KÁ…° N3tX àG¡‹…%.Ñ/ 6í”ìÛ9~ó(zÑ•tž;"þ>C¦a–­á_Má0X —Ž#ù¡üØò ŽúÀ¶Xà”,PW¤øNk`2w=òs#âÚŽ ÎS-™7 =ûg«©WŽª¾¨´+wª êoÍB(b=³m¶?akæ øHÜÇw ZWDr Õ¾ h IC㬵3H6` !+ž3ÐÖà q¸—Š ó±4¼jÓc¯„0só0ˆö¨ F• Á5­¤ä¼£¼œ49™Öƒ»gwiñø$öb€_ΫlÄeË÷M¯Yk‹›*¸°È•k27UŒ8µkpP!?£”€_ =àD}Tkm&¼ÆöT¥@ÅbÌßšüN¤ãÜìåx^ˆzyHªw)¶&Û¬¨Ê"ñ±°8©$Iu‰åGVÝ+”öòÓ&ry‡E+Ð.ðÈ ïÞq!õÝDŶ“aGÏhkÁœtÔTY6ŽRÂøºT ©ûI|Ç:®ŒSKÖS›9À¥)¥dè.F@@@Aúº¯Š?b‡³%Ɔ-ÚÀîÔ,ñÊÀ̘ëÈ£é¤ð*¥¸b„@–׿•à`‹XUòÏTÐÝiDÄ´Ø:ˆ 6Œ-¿Ĩ«~γdÊè*.B6ÙHûÉ|çÝà”œì> ¸„ñsΨ€¤KùEË%Á®°H4¨ùmä%Ë=Æ–Ó 4ÏÔé±ëДo‹5@ÕžŠŒ£¾/è4†R6”ëDHfò}³êѦoß,«¥:Gí»L¸€xÃïx ê–qåþò}‚àizEŒˆ;>î (øØÏKO#³ñ[§™Gmc 1§ì€²CJd(hL h¾™iô5í…NŸð¼a¶yHî„ Tî-Ý7ep_Š™žY“§~aX&/`§‚Ó8Ž,á ܯ1Ñ—;Ê4¯jjàȻ鹱F"[ºmõƒKtÞÃA PIŠmúåþ{=„V’CÈT?(1« Âl^Qÿåˆ)e½d:]»ÊßÎùڢטµMµcð'0OPå>švßð&L”š¤œÇ´­$šCŠÆÎUÊøò10lO;&­™Õ©vð4jzØÅ»K¬óc¾á£Ùuókc%Ë^‘3æ™Ã†*3i»Ú3’©ã›ê1'ÝE) Þþ'fÑ–"ÂB㺷DÆH_ü^$úÌx+Þù˜–ÎU>H#ÍæÆöÚb«(OÂ,"¨ŒôÁ cªš”ÇBf€žaÝaµ‹˜ £5.h¾¸œÍl&CXÿ—a@^¡Dêß8ÃîôåÌûl.ÝÁ85 (c½j­XœFJ ÂĽ ÝÿgGí86åÞtÙ„º3¼#Ü ì¯v»j z ’¦íõRÇvEHΚ±w˜˜ÿÕÏž4 .Ü3p8›QÍ\œZ‹ O­aÂw1Ã&éÒKÒkILH€5 é€c_+rÙh?¨ÃmâÌ‘ý—l½¢²‡çèšTÓ`WsfEÿýd¸ŽN†:Üê°*þ÷s]Æ£Ûxv²ÿÿ·I®t)ë4%ÉÅÄU™ö£B5 =I!¯lK —/É Ô¥u•öjúìaMc”â÷Ö5H)]§¶!ûÃùU¢Cí)!'ý ™Ãz@Â_ÆO›uu¹yší¨ÝdÄ$Üè“{îXʤþuѹ„B—Ѐ&ãTÚ­X`’Éÿ­¿{sÏ|/*T©»2<—γ± ¬˜åó†Ëˆôf8½—öp1š[|Kjq*›5àP…³ÏvçÆoºÛ4)uÿ¬L5@Ìæ+"1¢_mÖ¥û”6¡f^Âþ4òi‰éy 9‚µ„~è—…!g_fö0 ‰Á­Š Í—ÌÂë³[äZ»þ·Ðä}[üž øø"5dÙ¹ãQ€ ŒV°…Í\«>Œð°ÜæòN4©“…ÙÝä`Lv0,É8ª”² á ¼ÔŽ Êä(çR™×ü¬kKÖÅ硌ˆòò¼¨å¦Ú?³¸Å `¬víy ¾-œ\­îý£'TÑ@…§tk‰ïŽ+ôËÔ9 Çû¤dÌ·ÐvÔ…j¿¡ð¬Uaõo „˜Ü 7è%wâ>Ü™ÿ0*ß}KŽ,ùIpIpÊâæ²Ô!§…¤'ëbðZŽGɨµÝ–{ƒ?xàíÀ«B³ˆÿä?c=-´FJš R’,Ó©bŒ1,?tôã"•Í3(%÷Æi¸ýÀ‘%©¢B˜³€ýz_-5;6î让-Âÿiýpx•”ºÔE¦O ¶çô÷ ’Èb€A›˜P³Æ÷KuÕÒ599)ÉS¡Sfj_¥j±ÈݦÜÇù´»d³Üza [ÛVÔJP‰ffŽ0^‘¸}#ØFü–/oF–ÁHÔ"¯H‚î:žõyr’½ÏVZÍX˜œ)O–ÆBýó¬E|~ÃEh*˜o¾¥GŒ+âξ° ˜¹Æ>ÿ·l c—õZŸ·¯Œ“P5Ñšmͤ²ƒ !zjª¨eqê£á­{ìÚ³¦:oN˜youg5Ñý¾äŸõNTuꊰÖcúQ!k¿3ŒÑ^=S»ÛHÔC\|×úè8Í÷a²ëØõ*¦·0v±OVG##ÓyÍ?\ŽW‚{`ÍF¾L" “$sX“—!:^_ä ©‹àÛÑü!׎™õ€ß<ãÅœ,æ_…Iuù¬ëᬣÉU¢„©xÏì»}ÌîÝ4µ:AûãS>¸tQqJªhEdÝ}Ï÷ŠNÑŠ\ecr‡M˜â÷™ÚæÔ¼[qkÎGÿ1Eó«VGIZعÊ)õ6ô_å+>®¥D£-ÖŒÌò÷›£ø)ô/ð¤<¼üÏ\ð§² ˆÔ£GÞ5N§½b«ÿ…DògoŽ}RÉ  ¡®Üu`DZ ý\ÇvÂtKK¯h¸œØúVôhÖÔOç `«ã1{xÁ;F½})|{¬qe(G_=3ž$€R‡°Æȩšl+põ XáPá&埈(ÃM Æ´Fð k›³r¿®uAW ĘažÂ`ƒƒŸÔ½n©qsžoÃVH…\õòX²yÅK†> ÕùÏÂÊ ÉQ‹ÀQ’3”´Å –0î®ï§ª. °÷–†ùò²‡Ô¤˜õ#J~ûãmt†Éü0`…^t0(ŒÍ03/ã€|V’Ø—Ö£òÃ,¡‡Ï;l³Þ¹Ï1ÚRÓΫO©$&õ…è “~»ÂBÑAvÎшjf–Ô±†o¸Ü8mätæèüÒB†Óƒø³yEO}GpGR*_Ì®4¥pÛâyêÁ…<ðÑLnyÛ²fJÛRáØîÏÞ]´Ÿ“®ý¯ºëŽQ oüÌv„ÒÒ|‰X™öÐÚeqÆ Aë/Ÿ©|É eañçÖHîðÝP‡P?ÆÝ„˸ÃY–áÖìèáP¤"è]©‘HþzýaÂ"ãGOo„!MÁ‚¶ ÇœyD›|°9£])#ѨÓõ{/bygNòˆžÕv¼*ÏN ¤DL¬2c?å™Ó„aÄb1U vì\µóS±žx§]&-у@%òŸŸÊNÎyVSòÙ|Xøö@NYòqZø½ï˜P©—¢³ºªñŠÇ€LÒ²ˆÓfy£¦ëXßúspèÆÊ ¬è#Ž¡r´=]oJ…ISý¤ª!åq©¬Jqã‚ó\¢'[TTMîä±Õ;g»/«ÿM˜ÙµêûÝŠMÒÿ(€ÀÊ»ú¶âE&%#è»°©0‡à¢zíÃÉ»èÒ쟴=!¬ÍÁ¡åÁã¶Û)9ÜÂá3€|Ô÷o ‚óö!k·±ÓùêÿÚÈb‹N²FmôÓ >ëúÓ/µJØ|õÐY3f+Ìç²>â+F_äªÊ´Ø üÃEÐBŒë¹ÞžÙÉW|Æ8ЂøF›þ·MÆ™©þ1{§ý¤Ú:Ý?ž‡,š‰µØûžUÍNÚfnþR\ÉJ†qupü™üo{×åŠËt¾ÌbLW½#Zªc—ÖØ ¹Õ‡B¬Äø4_,‚è¼ßË9õ Â_ʆP¹9¢ cÝažVɳ8Åè1žI KÐû_pø¨e\€mþD•íT…Œ…Ü_'Îø€ÞÍæâ±ÅZF~ï¸pÈëXr…Ú?’ms"úú—¥ §±µ¡ÅXp‚6þ·ÎHl{ê(ã/_Í©3Yw&?ܤ½¶†Èg÷òD"±½å€,¶Ÿ¼ciϸFˆïŽ1BÑòŸ~%r—¹$ vèsÜälí•/ ¨À³²E‰0àêmÜGèìÂSWbôû´M ºøÝ£:'ôõ×äCñˆîý¥3Û4‚[·{©I@ñ]ßlcíkïzë6N¯¥6“7Š0§OªŽÌî³/°0¹øŸVž0ÈN2£‹•½Y ߨíVĆjw»«y=v úñ¤Ò[¢ûðn¥†ÑDTÍGî3£Gƒš$øLÃGzú’C@GþPto0‚EÅ °Zý•÷娦>õOy ‹£Þ¡÷R™Ö©e7a±@“®>îˆÏ1ˆ èî 0D^eö ô¿Âu?[ßí¹ƒ&sÚ/¿>쫺áŠR+þÐÙÊùºÅAüÓAýGGí»í®ËŽíCÓ‘ÇúÓ2LÖýôzˆ°b)Óµ<5ÒY¦Óê’gú±hÀ]ŽƒK„dú¼‡ p+¿r‰"W¾ô$SCGKÔM³Uö¤ÿÃåägRN—Æ©'?qAÕvŠsPî*dûÛMy*amU{¡hÅ]1X­×ƒatÒADe~È×5D¿0ÝC©{ƛ֜¯ÞÙNŠQÖºVªHöNV²5?ܨ< îjúÈ}ûtê-zº…Æië¯E Ü:ÉÃÖK7“8%ÖjI§öÀ©·O‰Ç_5jç‚ï%Ïó/¬ì•T{­§Ÿf¤Œ“‡Ó‡+qtï(æ]o‰ 1kMA_K'@´e•!5%áëfˆ‰âõAO‰ô¸Ö»Êh7ˆqì©¥&ORÄ¥ÒÙ³œ7Z¾×.†Å„“dšw—ì+Ûì’.«ú}Ör7 ‘ `>t\R/è` ‡6!û:Úþ f ”= 5vXq~i |ý™›´àÀEù±d¸Mü­ûÕ“ éÝöv\¾õaŠöã8àÞXo¯WQ‰?‡°ùëS–|ÇPè+ðÙ¶=u—ØV™ ¶˜%UGÜ’™+üNB˜¼ùf¶3Þ‘Ä20Fƒ¬Ä§ßI°r Iõv}Ô=éI:þÄ8™_=ˆ§¹p=Šr8„ª¼®Žà‰“w—æ lé­©MNÒ‚¸I—Äá­©$_ÖE÷ÄJó„®ÏtÓCµ„\ü{äñh*ÉÚ›¼!+÷s{ÍÒ°ˆ¼­Ï ¼NɵãÅ0¿˜âlVñôé5×_Ñ4n(´ËüýŽÒG“û“È`寜6ÝŸªoìêâÈŒ s˜%°;†jéÞ´Ÿ5°yViÅØŸòªJu¨ôc¡ÔQi“ ™­Ïü3?Ã×ãϵÓóŽú½¥y´+Ëÿ¶JÞÝ:9-è6´9‹ù,ÊãjU„þ„ù×b„¹^`éŸh°[O09 ò'ÛY!eq2•(®y­c`™·ÎƒÏKHEþš²D¿GžÿæÃVÞ*ôüYð¾ßbÄqyõað ¢W÷eïV¥­ø²¤×…m•>?_Ò‘9ŽdãÀ>V¨'N»k逓€å‡›µÊ™Çw•ï™’«„S /zHÌ~nùpA>ê–9(¤®Ó˜°[̱ `ÎEqž8og{Gnÿ4¥¯‹Æ1Ôö\º$*F»€—*Pßëò¯A‡RdoUúïÑç€J'‰·\u¶—å(½º™¿ m 18 7Œú_6EŒF'D¸C5[’ŠÛ›õn¯âNë’±Nz¢ßÿÞE¦äÁL‘õýØk¦d×¾yBã—òº‹Á ¬Ó{Ê-7Xʾ¦­öSÔÈTÌÎ>=¦k_¨™kÑ:ë¹³­—G}]»“lÚR6ÁìvÓ¡§…ñ*]ÎôZ£e 5ßBõ’ùÞ@¡|ìÓB¶Õ!šbÿ ZóÆS²‘àÑ÷Më…Eù˰‚­Üâéªr êNNG¯¾@ÕD]'ÅNjl™z¯Ô×ödoæB`×p‹êR(,MûõìB`qåKETB½1É–R)Ò£vÆm…ª}+ë ³Ò¬+ªº ä ñÀ°à•©},Ók²»µÛ÷r¦þ=p€Sý‡ƒè~HÂQ0¸êçÂ([ë=§äYp¬ªvÀëV×Ô~bÀ—Z©½7z¾¡ÞpçêT0ºiëÿ0Ý5H#žÆÉM@<ú¤æ½=TöfNWí)P¤ ÈÏ僎¯t[“¨÷hj ðqÓŒ—¦S'|ÆþtÿÉÂíU’CêUý°­Ò¶k„üg“íœXÍý+$Q¼’®êEŠLF¤8 "ö7®ÜÜoö÷ìÌ«Î~ =7T1†²[ú9IéM¹I;j%8ÐD-èÀuCZõÄÚW¦Q„œzvhÕKnâI57IÞ´+š¨¿=í¸0òcxÉf¢ÜûL.Ô¢¢¾°2„*øJgamªŠ-6½;—ô~xT޴Ѽh:¤>Î=£ŒòÀ¾m¹ðGÒÚU )áÕåä"¦¤ÀKýŽØè†¯›ñݙ೟ðI"•‘G&Íâh‡ôh¢a˜fÔŸ¤¨9$ñÍVž;ï5/¾ñÉÕ…>ͳø³né›PUÏb÷Ád)ñOg¿õÌRÖºç"\ ž_juóÀ , šÕÆ»Gæe¯ãµÇzU„Ìaz졚$Òþhk‡dûa¸âÇ)OklÚbËùzhÝ. vb¥ªúêQG ¡Û„†¦¨ov¹«å§Ï_dû ÿê+zÆrC )VÜ6 dg£Œÿ»Ü$6“ lÑÚ÷«ŠMÍQßqˆÎ™ÿ-t—U݃õàãHÈì£q×Ö@’FÌÐøº›ìž¸Z#Þ2í5‰Rp¼èKdm΂¿@jYô¯¡ezC7ßÇ íþ1 ¨¦)–I¸®ù½£ëÕ$C­³š %Ú´l`WöOÕù]9ŒÆ–$ˆ¼òQ µ ÿPÆá­­Ä€K¢mç oEï¸,r}ÇÑWIšé h~Eþ4'%Æñ©0O‹øþÑq©ÜKG ¡GÕêœvµ*"ÃØÉ„ŒÀÚGï'E H P¼ Á.Œ¶fåÿ\¸»ÆNgJ°¡À …ývbô¿åHD"p Â$ÿhï—„iJKL}~zd˜C˜ò·ÀɵÊ:SJµ„G°Š\T+Ê_–&|úSnMûQ±5“¡Ž]'ݤyMl@᎓ÆYú‰”Ⱦù"c‡l£ àŒ¢»‘ª¡Jeù31ÍoÍ{:£8OkŨø|üu˺ï úþÕH«í9Íò>3.L™xî꯬cZ•`! þJkÍš-áÛm¿ñŒ÷g™‚)[‚á§_­°òngKw?Í©„ :X&º—-E¯cQÛÀa¹ª1Ôh£.Ü`Òn´Ö…SµÌTi(êDÒq¬†{¶‚yAß88‰=¿SìáO¡­§íªÃŽÇ 6¡¤)édûa}è4”}ÝS|¾•³yåK²·ÀV{5pt‰=ü‹KSÚü»þO¶Ì÷ÔU„øíÈL•¾9a÷€·`Ž­Qt–ÓóA×¶%Sø§s_åéÅÂÇ7ðj?%ãÿ¶òð(-ÓÕ ÛMe‚ˆ¡”·÷Õ,äë¤ÈÑI£3+:Ýý";´¤æÚÁrÖ¼J¬*LM¡Ì>¨¡dårG&4ã¡Â:¸ 1–¶!‰Ý=-O6Fþ¼j‡þ ÿ{}Ð26O›õÍ•×Ý‚˜ •x[™å–ï¤÷ðŠ›©ð)5öWx”`LZõùf5õîÀ¼î*ÿ}~óvÛíY ^x»ÉF¬RÕ¾mIôþ¡Ôp®¹£}¨ûÚ —­²LȟôQÊ$ùósÙð½éì €P$•ÛàqÍ.n@Y\ÛšrEî¶>èaÜòtŒð®­ò¡~)b€ÁˆB5S’–@Ö_2ôˆœY,ÐXVå×!°A¿ø}r‚FШ‾ˆ]ÌŠAjÙa€ô)Ü¢­À!/|eÉ/\·†-æ)Àý¾‰íd÷2ú›@fšâe~x*"a⊴Åp“,âH\Vn¶ÍZýŽ‚šÅÒúÝh~|6c¶ÌÖ8¼×^4SUf5­žÐÙrÛÉ6߉ h—FXçmæ¦ ­1X F&ñ@þ©{¡W‚cåâLÉŠm:zœ›:–¨&+)6œä­ôÉA aòÇ00•_ÔùÜÄŽYì9˜ŠÕJVfÕ0­¡õ- ƒ¬¦u~!´"}2Æ_/z4àAŽšÉç1“;^I=yJ“®;H…#m³·eôVL .È:¼ÈÙiùM æ§¡â‡^~T#Ôá 8ãÑaõy»x y*Óñ¤m¶ÌUÁhf«œí…êÓ½d²”û Žóå ÍΞÙƒö­¤v(üt« ç~0ê•ÆÈþï!?ðÿØ 픤aW\œÖÉ?Ú9b7À&—¸ÉÉg8ÇÓ%9£Ê0µþ=¶„EëÖ¶SµoÇÝ΀á" ŽY„Úž­žNèE !ÒœÄÆÛ³å siÝv_ OŸè èÕ²ÄGæ7\Cy²+"" Œ“‹Ë;ò ^ãFAqަAÒElûBz#Q2@!ó¸ßåÜ/­ñN$UÁ¥u h;ŒGKÍõW|ùòÚ]°Ü¥ÏO\βM ›t0dCqˆ»~ÝÝŽt²`Yiå{qNÖ)EÍšû†Îi%ß"«, \&“ÚPÄZ‚ž;±“‡ Ö¬ ûÙ|ácû1‡£[ÿ(÷ߢנê©áöò@d‘nl·3 L½c³$üïo+ÿ{pi½¶zZº!ŒŠ“Sû+Ó#Ýsè„€L… ÕTåPÖÑŸc“/%Úp¹–¢´Ý§ôìš‘az|²öþ}`ŽmÕ}jT8iú6z²*ü¬Ã4Ý6ßnå=BÝaV«;¦=)¿ð°Oâi4Ìã?ðÜ}¬ì G¿JŒ3,†-šãóÇÈpÆÞ YùH›‚î•qÄ4Q¯ A§MNV­$è9f]߃Â˪zήP>Ðõš.ÀÀUK€»§Î‘{¯?û»z¥­ÅvÊ7o²ã+Õ:iØP:mÙWi„t/;wõ ð*;: N`eƒ/L è³% Lžì”â·E¼ rJ?â»Ç².6%½ Bw¼V¥ÎG¥¦'þ£Á%‡gRÊ9ËŸx=0.+¬×sÎ £m† ('ëÆ ±¤Â¿pÝQ}–H užÃ?ׄ?è~h¾¥˜ V†…¡—©Šö_1Ü%z’O2}¶=Û5|zâÇ/:M›Ñ¢\>ýt¹Ô4và&žl®ÁnBV+%9%“¸w޳fÿA‰²ú$ehbªð˜ó¸E ¡Í¦ƒ%øÚ™sêj·6^ ä1(7Wº†0þú·FPV©õÑoÐ7cƒ±€2Ìïä›òS&*29µ+}°¬­‡®ÕYaÜ©—ŠOO¹"Ä– ˆ)y7¼Ã þí⹯ÑÜÆû\û¸ÇÞºNEÜ>®?èÔ¼+cÅÕ•F b /šÌ›¯V×Í+”п†Äwå,9{ØÓüDÒk yˆßI~Ç"PµŸp¡­ÞSŠR6L7 †ßBî øÁÃÓå}ôÇ-ó:ÏãúÒmR±Î€ë'NW. œ!¦c² 2BGC¡%Zªpä”Ô–eyÊëèšÄÍäJ×…g[!òô9Kµ]"B»¢ß)žVõIèqìg™…*·ÂCçÐÆMÎ:7à2Á­´™Û,è}qZ¬èÚ¡j}o)N˜<ø«4„CY‚„«³w_.5oÿ*Š“#zŸªó›Ë<_æéY‘p<[ìPÌNñÊ̲4HfñEþ‡—Rí÷2ƺY‹¡•¬–“±` àÀŠÆ€ µ;r²?Äùw_(©:jØÏÆátàJz¨Ðg 8âÒø¸J æX9ˆ²^‰ÿºvGÞ:­4ŠLAº βû=ué%*¯ ¦È&¥<¹Ô¡Ú ê)”àï@“¢þ{9!ä{Ь3ñz é®Ä( ]á ›ì: 6†¿³]…=˜Ì'IÛûUå~·[êdõß9~žS;¸8ÓüS»c*ÄÙÃ"Lþ ã*ç\+€G=…"½þ·*Ð} Ã*Âf÷¹2‰§òÃ?‡òD¾A0Yéñ³nœâJ¬ÚÌÉ’}*Žõ2è^´ /`âPžÿ.½ › ûRæ*bZ|_ÙÑD9ž‚i—7*z¿ºô^D¦“6{ÍuÃd;ºFÄö~«¥™Ÿi¥“k¼:‚ËqŠÍ–w cÐib¬vÌ4šùóôͨëxƒÝ€èÞZ$¿£äìHÌ”C–ümœ&ðpZsí*¨©*¿LQßf ͿϫwK6iü~Bó9’y‚xIïÖ¦2 +ÉèïŸ#mŒ•)Ð|Œ€Ö«œ;PB4[žHtzгÆÍN¯ ~ {/sÍ›W“ò¯"ÖÌ9,‹hFZ&ˆ};ªØ a»¨—¹ŠC6÷çæ9öÿeÌ€7 ÷@Žø=}1Èòoù­Ø±½,‘ü”¢|@öØ<¤ßʳÛåïˆ1‰'ª¹5 ^"¼¹o,PBË_¶¤¡†ámÒŒgòjÝ…"zínnøŸr)Õ3ùöKì·—j¦FÔ!i€±­x)zJRO~eúÌ"ê$OæÓŸ/£7`#%©žó÷• ÊN–ŠÅÔ©Ó€æ7{ý7Õ÷³ŽÆ¾~"ÄÇ­¤Ê-Á"ä8C)”Òö€P»”.Ì€ˆÌ5Óy°“QC ö2Îv¹|×:UL½ÁÍã à\ 3” Çoÿá­ïM}á_ÛªáBÅÑnþ"Ÿg3WXéEJ²º³Í n-*¦fQ¾q}Õ®ðÇns‡Jb¾,•g?ôæ"׿—WqèoCÅgVc­‹{.øÏÖÛçU‰„H wüó¶Èâ›õåUƒx—ÀÌŠíËUÆP{`¿¹‘^>»­­5¼”VÐß”È#3-tô3òÄ‹ñ¾Òøuï-U‡¶åwÔÍeR«¾II.‹Ø¤A\M=zÊ£C~cžîæã–]Aò ÑýùGÞ«®aÔ•ëõ‚Ÿ_>Õä*~!)ó¼$· éǦ©bìŽB={Ë2ï;šv~õgÎŒ'Ib“ç‹P2ð“ÛBʪR]µÜƒ–+²àJÔ"ê²s¢âö¨¸tíí;ŽWå&%€µ ]V/äº0g´eÀQ\ÕK'ºÁ Ö^ê4 $ ’|ßïBOW…ÿ:ÅhôȽ“N™òb¦ ,ç§¼ýÅÂO²õŽpŒkm/˜QÐ>kÝ®HOmà#Ê%Q¹"Ê«¼*ñÃezds“µŸ>ª©3c%íò’ZNJOQ[ýù]jµ_ï„L3ÝÍÍÇÿË僇0x»¡láØ3Ęí­µþ9I0 A!ãQ ]K$BÝÜZŒ0ßÈÑæ¬=3°nžªúÎgÐÍÓ›‹ÊFö“øú(Œ/šÔ•55ŸQ¬°Á6ϼîü&>¥à|†@b&Óri’—3 é™È£ºns¦YÓå’.üù;¡zh/ƒù£Ãë¤ùRf/7ÿÍìTÖ÷‹EæþÃKyêÔ-T_˜Œ1«¼¯„óa¬w”dñd( 5y¸´§¨ä£=¯ŸsSôŽ'Õ '»‡£ ¬sé7~DÅF¸Óo\P ª›Þã>{oÌw¤¹+™M/ilÜÀh¸+ÓA‘¸œ܈xHmbòµ3W‰ žàʆµ“ãFàHÞLíC/ŽcX ÝñyÐM“ Š[Õ,7®eYöŠÜ:zj“`੆Øo‰# X@H Uþ”õBÖx—Þ©?æ§»tm®áQ›L,,ŸØX®è¾2ˆ^šWgÐþ†ÿÄäƒ'(’7ÀLMlR?®«´¹Öñe8ÃÛ§fRjÐ}ôÃ#ê1I{¤úànŒe¡ ®h™ö¬ñ.Êû×½â_ð™\¤u ³‚Ò «ß-6Žt´S*-ߺbÊNø†< þ'×öÝzûpÝ-…ÉÆ¦3㙕àáÕ˜SfjLÅÈEùl h_U-Ý)ÞÊXZ¯Ðo™ÔSDéý°m7?ƒæ–¼]4VK<>Â!Añ‰í§ÚæHò¾|,CPìU!´@o¿óñµI„ÂoÎU}—ü²í4ª?ìmõ¢|}Qu¡&r·*6ÂzòqÈôJùZÏÐ;ÆàòQ”HÅ.¹ßQ²Â=o§ƒQò³93ŒÍ\†.!F¹öcèzÞ¦Yõ>î!¯tæS=M_ÄnDJŒcJ§Ã¤A™%Òö¥ÚHäÜôÿÁ— (bÕ¹%¾[õ8²m놠Îñl‡WfÑp1„$+¿B )K¼³cö)H”îCÉ`v9î™y‚EYW¤_ >¥䙆’ɸ\ Ò>YÇ ¯/Ñ9ÓíN5VñªWv$J¿æ<´ #‡ïblCM[ü§Ûý8Œ:ò=Eî̆œX2<È7wª ã{°_V˜F5pú½Öç@Õ· cëo‹¢0ž¶¢"I2Ø:»fK,™A0b¶ši[ pT3!ùÈ'•¶(ÄÙâôyW` ñýÉJCKÂéÏ»/xÍQÌ%ÎOHý,ö=G8Ëë§#ð à6tïªÄ›Â^‰¨êЂ»¶ßóJQªBpT5ónš´‡}ð7d_³YƒìW“ûò<Ó´®‘ r5¦Þ@´ÿs'* ù¦ùf(¯qôW–©Mz3”ãO¥ ­Üøµ0È€E.«Þ* I.éÐ^#Š4Ór)8ºï¶…³GÕÛwÅ-wkS%¤ÞwMã«kúm³(Žš¯ln:ùÙ§ØmÕì¤B­¯ýø8¡Ay$aŮָ×'5d áT¨ýQ,³"£ &è¶ÞÚŒ×í?É'¼”4sƒ¢ÌÖÊ¡ 0?óß9â¿L›2«®hvÀúßG[U"­±R£—dÎËÒµÛº¤íƒ\l)KšÊ}rÛÁ6…;2‹ˆÄý ×׿W–b†¦SU RÇÒÌÓ Ú¸guú »ÈÒhý0 è¿ßP‘9Ö´°¹3Úl."i‹D»Á½Sµ´ÏQ¹FÓUZŒV0KìàYaê@VH¤= Åì “F .¨ÍÕ".#@~qÿŸ §Œ‘Ô±xè°”ôîÿ=¶)™ é¡ó$ÊCÒ ­RyAyÒ·€´±˜2]Ë÷N=œ.¸íäºO²ÚBϯ2ç·ÌÁyE–y¥Ób]åmž¯uáwuBbÕŒ tµ¸§©è™0UÊ•õºs’äïP¥ ÿV¢ržúÓÞ,.s8ÿÒÁjkôü«{Ú€'Ù SvêPj©w>‘µ¦È Õ× 7Èx}J! © × Œ|ÁüZö±PY_触Ü4è*ëçFÃcÀ¹RYƸ'UÃhRêô*+“ë“u}Ç!nkU/w²éô eÜ*WåP·gÑ7ùG ;Xß|Pˆ÷w¥5 ¸¦¡me½â7 /#Ùz+MG{:ÜìÙâÏC‘ŠšVò©¹bY muánÞ=½Ã6ð9éàQ3æ†Å$ÛÜR'ý™Í>Á-z!ÖÜ7µÊç¶¹sdÓýÌ”²\M’÷ªR9øû檠_H w&h÷þ—í9ã…›¬·oÝ/òå«›$›{vÁåâùHiÛ³‰>|ê™õóã'ʘ´¤°§ì«Á%y»Uåî™°(^5ÞÑÍ'2ÏMéîè÷oŽÍÏ,{´‡?ùKˆˆ;£¡ÌW(zcA¬`î´¤ 6uáWÙyˆ0nßÓ𞊠6S3ÇrŸ ¹A§Ù–XæÖÎFìÚª·ÿÀöxÎ`*û}C÷‘L¥¹ú®}˜+)_ÉÃPÁž'¡?yîx~¶²cÌ òY¹¿»˜îHV39b'kYz8©ý Û­ƒ+òÜùËLc„ p¤sVàEÍ9pFEgf¦þU€ÔôX9ˆlóc(°}>’ìù‰Ö©’ºñÐ9‚Ëû¯R|QrLœ/Êj\â!¢‹oÄRÛ)³"|i‡'Â7íŠ,â:<ãÌØ)xÍ•é‘>’¦£ˆšðG®CJ.ê{8eZ“„hΡqS§ØÀ}°ñ¨°ö9-ëÁ)0a $Fqù’£ípä§§pë5o‚¢®90"›Ü†¡Ê}RËçÖÞåy[w,L‹ö<Șߦ¼xœ˜Àú!ÍØ`!ÈI‘@§ÂX÷˜_´'™ß„*TéMMì:Ðs>Ï=Oûá¿`ä!_Šá‰[ÅŪY¸L,ÃÌá|õ;&ŸÐÕ£Ì/,ƒ ¶l–K }Éâõ7‚RJ\æ­ùyø@\Ú~m|ÖϪô궸ֿ–~ý=ýºOÙªKÅw@Î.r rŸv“”¥¸Óí'8ï°wáÜŒE,ÊñÅÑ›aÚ+ãYÂZ&]Álï%Diž~<Ó°hbÓ#JåL…1Ôà/$pûðóVÿŠE74ÙU“…•÷ÊÒVãôlê¯ÂE%ç¤z1Ü7Ñ_þÖ~” ±ØJ6„¨‹#sZ•m5“§üÈRiâ¿CU¦ `N”ðb~ï÷ßÏ,gù{‹BÞ¢ê“.iRÈ $0•á™Ã–õSuH‘9Îç¡€“ðæËÏx¾¤o‚€[@¡7=HG4Ñ9Ñ++oIec…09Íf7¶ñþ(oVá»fH|jȽ3%15ðd•¯rß¹ºuÑq8ÇášUš,¢¡¯ó>ãƒ\>äÓù”®Ý¾l‘¢Gb1N·?"©=[üغ¯™¤žé°«Á OC‰ôª#È.+e«¦ÏõÁÁ‰±Mr«Šu~÷ÌNO෶ãû,G**£nØà* ×o¡’áTÿ•1*²z ‚ ¤›ì×ô—x; äC÷ú/µ»üÃJ :‡™œS x”æ° »ù2@©ÖÉ ÅÅ[â®Eɇ¯ûAŒÃüÉÀ•: ‡h l;ª’Äjó©ŒìÙ7üyJÛQ¾×OÖ!¿øÄðbÌËéΪŽËýXÛýl¹:–ñ&Ín`ýê)nxb4û“ÈÕ›WIɬ‡%^%aú@ˆÞì-Ä!è/ñäÝ£")£í]îéJÕ_oˆêLPy×Ú>ñúJ1_UÖÈUŽ ž§6êá£ÎÄn6ašè∫;¡d¥Ò(oPR{ZÕuÝSö÷Ž%"ÐW :“œ©kšÁ¨Ðûo E â;'ÓçCG¢½ ¨!’o‹(PÀÄŽeOøÉDp»mmð¦èæ7 úl-þÑõâ]ZkÉA%P[ªY#¬^RµäsvBã¯\Á‰A¢‘ŒH˜Ã?m _+ò'À¨& £AŠ—Ò¢ÚÛÛA݌ˉ|ï'²˜y‰¦Ê‹EÝ© G 7G3“/Õ·>¡ß€î§ ¦sŸJ—½ñ®Þiζ~q|Ÿ‚çFh2Þ)vþT$*œœDfw†­»ÖjªÖg%uyèwº5Ô,˜:¦V™À«*àq¯V)S»½69iÀIªJ™1z4cè-½Òz¶k;âªà‚c (ŸUÓ‡:(Ä}ÍæßLe6Ôñ,#ÿiTí­;·Fÿ΂ý6„F‡†,ñ ŽøñTÂþ5’MžT#ªîtô2K¡b‡í™¤ÌS¤öG‡šI ¢|'g øè³Lìâ–(þ0ÛIÚ¾TfêBš^Æq &‰x=,ARã®ýKjNæºÍŠ”ÐsPˆð ˆ±*Í™õ(ØÍ²iräÁSÚRªQŽ@BþÌSDØÐܾlK)7b ¹ûÐä&˜å‹½CyÿåYò×q1Šä+¼>'.hÛøœ¬×A\!Ça·48Êu̾QáoßüÑG¸ñi¥¡(P†º|‡MÈ€”CV6}¡Û3ßùÅäº ½!^¬U”é–TºcW ç續 ¥ëXMÒ7ŸØc@ ±šŽBÝ‹ËeçàK².?dš›å”¥|&ƒ’â†$cwHáe’Cg•=#в»íþËÕ>c%êÁ¯r¥tM§v‹D†kúX̤ðbzÈèCyµv0ž¥[t¶„Ã~ !¬Ø*؈K“7[¥`–Öþ”„‹©ÍÌ‚t6ìíÍ~åRÿ#ú!­é–^!Ÿ+ûîÈu6‡Ö5ÍSˆ™AF{Ïz\ xÜŽF½5ùúÜöià‡Â¿3ôL*XwuÌNð “PËó/¶:SäÚ T@i[Ìv4«jUÈ Kr"ÄM¼ð³ì±>Œ~åÙˆl‚u0vý B¼)ôB©8[kB­ž™1Àˆ~wC¦9Áß'Ô|É ª€ÓÐaíô¦¤WUÎx¬Æ–ŸxüT+=}žîƒƒÿj3¶×1¼ºèÀ[$¦½|º¤‹0Ü+{tñ–Ÿ›šëê^\@&éøh<¿›¿d䔦Ð_ÏÊŽÁÖLñ›¶WuÇQ`Êðc ¾*Ü*OPĤñ’Ö0iùGá»/K‹e硚ae¼ßšäË«#çðNeË4˜{-D¥lõ~ƒ5gÃÃfÑrÍX‰^»¥™K8euK9Èø&l"Î` Êq”Y­ü±°aÕÝÕUmâ^‘iÜ/à1åuU”™B¼ßè†N ½m4Ê3fGé _4ˆÖª[jþ®TdÄÈH&€\>*¹m¤tÐê$ ˆ@|"÷Ó´¡ù«ø"0`í„\’‘Üî&Ñ)?ˆTeZDh¤½[³2?¥p~œÞ ¯¤ï–¦CwbgðÉÉ™CÁ‚OýJÃIÇl(­ŸÈV 8ˆ›Ðl“ß[|¼ÀÄÌó?gÜŠ˜Ÿf i_˜€SðL3ЄîS§eôÌOBM‰îãÖZ@ Žç¥'Ú‚t$ÿˆi`äà¼^‚ì5<ñ–nÇ[}ö¾‰SŒD ,Pqv®²^€¤[* Ÿ’ÔΈð©G`V-bbõ_+¡­KRD»Ìð^ˆöº¦Dí®ò¯ñ'œëÍ s±ûîunƒ´àÑ‘rë)†¦ Xៅ‰uTã´gSÈ 6gZן ®`¤ª®4Ü«@Ò ÑÓo•ò6‡ÅCv¸;©£_bó}Yþ¯½®Êý±šÀw¼ô‹ETÝó·XÙÒr,VºÔÁƒ):T˜ù_3®õJéÙ ÜÓ[âÐ08Ö²6¨£€îS0¦Ê}Œ:«ÈÉÎs ̯¡Ej3k¸%È9{DýC E~ò^'”æI67@p)6Yãè~gíʪþùšÈÉr¬ûÁ"Ô}ݵÉ_€—t±ô̾¯ø)f#=£NAõŽ666é ¢¶¡Æ[ôq½ ¢±b$=BmÊÄ'v}ÿФZ”s÷7dÃ^äçË•bEÅË¡¶âÝ‹0Šæ3sDKÉùXÌÁweƒxß§ã·´xÂÒ»ß-Zh¿ÊäÊ}ĉFUfl€ 1 WtÄ)úƒ™äáRv$Ĩí0Z½IÂ¥A#2&;zI#âÇѳV·`¢ZkÞbœ6ÆïÉÀiþ&#;w“9¶ö…WõÞ¶ß·ÿ•lv‚éßÎí Hxö·›Ú:`…Õò]¿pÁD‡,4ÈÍÊ[µà·¹%ss¢Á.` ÷•öÎyìn¹ÈÒC3–o¢ +‡&³,íÙ¥.•æ8DÛ'“c@tö÷XÉ €T Î¬xÔ X4´^Ãë§oRî¬jáº'>‚ÕìRÐ÷APù¿ ¼EvUoƒî ±‰`øÃûR4QTq–N²K¸U8-|ËLwM­KgH’£7Õê*õ£Î™)²‹Þu'aeÒrQÔ@ ÝLFiºÊþ fæ¦Û·q¯ ­Ðêp³¥ V°{ü=Øxˆ\êO4ìƒÅW$R:3‘gùÇÈWp>N®®)ýUÃ8 {ÍWž„çùÆLÔÏ'D ²Ç’—âsé»Cë„¥oˆ0{O@V„+Ý{®©ó”D*iCj¨Ê^[ÐKYÙùן´N®GÊ.[–ïw?,65tÅHj]U±C2SpÕjRÎ ý-ÂÈ9wµÜÈJ0º8#íøö­,ÁÖPÊÐ{YEñ\úp«Ð»mM‰]ðŒ“±]…´égi¿øE7£[6®Ë;žÓëýVÔ¹/Û“«ëݦŒ~Ž+7'ÿú  QQ¨n±í™Ã¤ÎÑ”Ì`±mc>5–\œô<ÌgjT•N<`z-ÚFs} œìHf=îµ{È )ÞÆE¿‹­±¥D) C¢toxÈïè[²7%E%ºRv8˜ÓMTLªo¥ì4úb+³ÓkŒ=Í8õ^¾D¥˜ïÆRùRF98bn§‘ºL±þ÷ßCVi.›Ë|52=@‚÷[’ŒÞýìvsÞ:¬zTè%ËPua¨>¯øïˆ†:‹‚tcãôM«fj‚$ ÚæPk\’aJf•DÞw2©·q•÷d´c¤@¼]~ÙçzF9‘d<î6÷Õc4ñ¡ú÷Íen•ñ¶8•ôà‚¼±ÒC§³Ó\c™Õ"æ5Të%U]ž…>u0Ëtî<8å›+üžDƒÄ«°)%Ÿ=â¨jè‚pïþYsꨬøbz>¿,áâeÎro_©@??¸8ï‡yMÞ^=õ¢t–=žFËc5…€ §mø÷Š?ìã;]j¶wçêHfýŠ<keé”ê80ÿ¨ö/œµKÒEÞ’ FÓ´¢Ê#º-_Ú†î–+PvKµI1N‹ëF´Ýh¬¯‹T{zÀ=g°Ã×áî¥ÖG*?*@õÉBòCUìÌá>šQóËÙõH%Ⰼ̑»$®“²–BÁó ¶§þql™hÒ¿•§Ù­ûÐbü|JJ‰oj“#Þ‹„oë@Š5k-±˜BálÎUùYñ‡ã{Äò…0†ƒ@ÀÍ*€Ýâ­Á@Ý”®D¡úw™ñ ø8¾ºèjieÄgÙø.$é±a5Yl#4ÚuL›Lj s":Û‡Ïg¿Ó·L,§º(_œ•ZñË (jÚ`P@ùþE*Uq9‹ØoµÜtí¬:;·OžîA,Ý×Ð oë™TÐt&G’ªJWaÐG€ÙœïSçà{s>è ü@}ç#Ú¢`Î I[Ž¿æõZ¥å¦Ì{ýéŠcÈ^lyTr*$2‹‘¸\c¶–´´Y™JÖZŽrËÿ Z%üÃëô… Y$¬¢ïŒa?DFøVŒ/Sý%µY ÷ƒ5ï Ç£;PˆITVuø‚ûÛøŽD+¥‡ôx¨P.¸c{µƒIêíè9Dvµ›¿»i§Ú/ÐÀæÀµÕ*”Ý3‘#›²ÈŠŒ´p÷ ªa‰j95Nu±¼‹â‡˜2µ]‰¯·ø[êFSÉ÷ð‘qà[7JF©uŒíÑV9Wõ8÷ª^%ðK…×§ QãêõW-ͤ¿Ø‡™îžhßT½M‡KE*#MÄqo—ðQýÊšå¼Î¶˜Ä:/yœŠÆEà; ik‹ç©A±¶µ¼ Yö¸Ð)¤C0;yóQƒDúÌnÊ}DæŽÅTÛ­V&hÁbyX% fjKàz¥»Ã¼3ð²ºâ98nÓ=‚dOæÛçZÒ+XØ”€VPôõ0R%·vÕ¯SWÿ—›M¼øc„XɪÚåu£"Eì”ß*i\#†‘ù%¿+y~ƒx8=:2_òù‹æô0‹'ú…¾vñxÝàZ:î …HXâdÄð §8÷±IjF¦'wÿcîè«ý¨@Lwa·6=EG<ûC{j–Q(k¢ñöwºôÚ|,"@ð÷tKvR åy~G‚븛ÿ*Òà–ÑÀUÐ ³íëãüå°¢Û"?1DZw°­šºléB~ ˆäдä䆶ÅÕ¬oöáFs¹á¼F|à–µVá—‘èM€®VTã«AqÆ{0§ãI¬jû/ÉÝ;Ÿ±¾Ü*Üè¨Ì–v'5òô±zw­¯ÙH‡ëb¢yo ÖK i>sí5)À€J+M8ß™zcÓjù‡ˆ·wAY|âÔX¡Õn6Äít—ýåÔf²ZÀUŒô¥5î9R=©åÉå‰$B´0/þßÓ"î…E^ºà'©uI€põùy8Eƒ®Ùò‘©Æôh¨ÆdÎB¥÷&“@ýl¤þ@R0£/Ý9?¬P·m f3[¼ÀjtdzÙLÑéUëð¦`ôu‰ QIwjkÑé\úô@¬„´5ÕN"xªX¤4ýJáÔ‘Â;jÞ•)ËÉÃÝ)^Ї§øÚhöp>ÔÇEäz–·5ßþ^ 4Ѽ2‚0©‘v/í+¦‚ŸÆM³Ó,yÂÛ\©ŠhÓ.mõ¡2¼Æ(µ™R±«?)ø‘lØß)>œï*Þ ¡'§"±'ë¸Mú)'ÃqÎÆo´¡ãg †ûåäQ‡v¼R)ƒGa“|øt âZ/Òqq€h»¨$Ÿz¸£{ŸÏœÎWŸmÿx•÷ø|örë :P%‹%Ç´1„¾‘ÌÌb-èWk|L ¸}svÑf#ðÀbXl\ ÇöƒÖ6ÍXÞéµÄ‰fûÖz¹íY§ö ÔùÔõLzž‡%_$ŽŒ“)^©Œ&±„•P¼Æ¢Ù˜3¸n‚äwÀH{Ù·é«c÷<‚~W¨ò[tpÝœ6)¤Âu8L}Ümÿš®ÞV#JÄbG†•§S[ Ðì¸^¹Ekÿ¾ól®RC+c1K8ÚX£¤zcÌG;Q”x™¸š›)†Í*ƒÛÌïÅ86 ¸º4u~'ÿ#v—›fØ)cj9¬Ï|-ÑÕ4ªÏï(«Ž#ÿ˜oÙ°¥Q* j&)Õ+öL´Æ“àdap¤^”½p%âFE-©ÿ‡_Æb¸EaÏæêy›õ^¡)[w0á¹ÎÓ2÷p³\³XÌéòœ™Œ„ÕÉÁÕÒN.n+×ã¦ä0*ÂTQ“âù[g pPÊHR‚Îv¹Jµ˜”·xÀ#©rªk%~wa´Q#Ø•Å×¶ÇË’ŒQ)Å”Î:4½ü‘¯DA®×$Ùg-•ZkyäVÚ,ŒlÊÛ@9º¥‘d’ÊÿM6¢­gøŽBëêÿéØF–í£{ºó;ÅÃò.(£ñ>^„‹ÐòÒ±8 &¿t§µL¨'Ç€­{&/ÄOßuk´¸ÂÛ-Ùv!˜–&ä é$D[5ûM£.Ыø ÉÁüÛ6²Ÿ èAÇ›:zÆ›î<"­äÙƒP£"ª1Ì@rp9àRÙUVêXB➊(8Rþ¿ ­Õ8|‹>ck2ê¾–ý×P¶€|„!À±ñJ:O™ÑJUIVŽ%䥬èÿ5T±8‚)ö;|[Wj´ýƒ›†)åwAP§_ufIuAiì»æ[9ð¬æFñÝó¢Ïxú’ƒÅôHU²îz€¢ÚòŽžåRcRi!#˜ÈŽëÚÿ™¦•}ux•¤Ù«Ó#Š2 yÈe˜ºv¡~ˆü.^ß4†{ÂK AØØ9-åÿÑo[‰Szê1Š€GlKu Ÿ„O¡\Bpžã dJOøóvì|­ÝÖŽÖÈÇ“uG¾\ºcÙp–)ÀFz{ùÔšj‘Š~ˆÌö÷É6_èÍ:£nüa9ôaÄËØFì½™^ϪzKÓ0XÎ)i_:ÁÜ!‰/‰;äôoâ$î‰å‡·Šxd×Ó‡Lø»Þƒòƒ°Šå›Û ºßëaƒ5Œ ¾H7Õ&¥…˵BSwó¬±Å+´¥ÝÜ6°IOLɤýŸõÍN  ͇‡‡”¡E¯+Ø–ÕÛž]ÏÙ Z®]šK­©å`Ø,ÖóïUöŸq8ú÷ô'æÿKI=>¯ÏŒ¿Ç;Ðpç®ÈÔ@(+á:ô*p‡ãÌÌn/:”‹ÛâÔ©{+Êð±ÉzÁ”È!>—,ʵÅ>Pœ©S»mÂ÷gÎDÑnÞP7oˆB-¿WÊå{åöf 5g˸ò3ïÆ´¬ÇC™ ÊÆâ‹o9Ñ÷Çêq<`‡$ap¤áÌ{Q‹ç3d® Á.µ¨ž’¡³ÍÙGXᑸká^ˆÈóÅi³òx¾'·»aå=Š8f7BAËÔèöéadîr ßIõp­ƒÀ:<ÀÂÌ#ÿEÛ-eì¤RÜú›Bsí±„1ø1 îÆ‘aWt O*ÓÀæ|òdú&Þ‡z`#Nc¤OÍÍ'–…‡’‚ÅZšžï€¬3É9–»1ÒÚ¼±Áo)¿…»¦_ø?Nב{ãòòžñ “»M3 @ ä©B>Ñ‹úš½Ri¯T«Úä»M½3ÁÆûFã€a óßAë`ŸÝ·P‘•xn(‚‰ç»¸V«PD–‘lîL–ŸuÅi¡ª¬O(,ß¹Q1,¯…³µÂ æo$ƦKPFÉŒ€œµlvEçT•oJ,@%Gü¸ëuµýôÏ€aîûM«$——°ÿ^ÓTÔŒšyýÛA(_ù_â<´š“Ó±Xœ ó_õJ¿ 5R?=BWážÒSy3 õo‡GìˆkºÓ!'è³Ö'³çà{XÁ/»¸ï™™›¿*âí:w m7És½X™(dþK0 Ë$>æÿAËã Ƀ´€ÃôFöëÐ]˜l½óÄ/ñÆ4ýWÚº.ìi¥OÿÃÒä(-¶ÿˆÍ;—;'i ï]õ£9çõjÄ‚'!ÿµº…rCؤOþ*\3Y†HjBˆÑèZm)Kõ„“f3SŸu¤·8~$vl.R,ü&pa%Ó‚EHÆéÏJÚPÞÚ~a¡<¶Ÿà‹<Ô&á̘aÛ¬ö}ûaÛþ0ç•ÖuIñòÃŽY+Š'wä-lµ•¶žàdakÎv:ÔÎ!w¶šO6~*éi•˜Â³$÷ažº“›Ñ^WA—à›Œˆæøkxm`Á ƒiøqþJ ±öëß² 5ôƒlø³&¸Ý¢6a S!X£«2@ù’&Žú±rÐt{½Î>–‹¨Œò¶ÅoŠ#˜>í ¡Éæ¤ê5œ\Cºü/¯•€UÐ#„{öÀj,ÝjN«W¶´þ6Rö|}[–‰|áÃej£Pô¤Ea;uèâë·:³l)O°{,^P:S\æÅÂ\!׋۟¼ øR°Ès½4·²[2d&í€*âŠÊoÙwŸý¥¨ËږЩIhò…Y]Õ¢=½ÅšÌC­{a‘6å6‘ª—e¼qñO’¶p>HÊ6¹àbi´ïK¹9Î’S 0ÿýHd–›Üæ¢ïªmnè ææ-Kº™³ã¿ÃY‚òÆÊ…¨*ñ}éLø/6‘Ò|…ÿB½‚6¢˜$ú‰Ñ~ô·zP0¹j{jóW-ØY1Íóˆ' ~Úà l‡sZ%Ò{5õæ–¯›ÐK“OrÞ;Ãá}}xÎË$¾ÀrÎÒÉãÙöú›ÞÕxqÔRÅ‚ÃÉrºãBÝs{ñÄãÝûhVsüïïC”xÆ'Û]§*ÞœžoÔù ¼Ðyõ¬:’õ†9zÖÛò˜@ñç7¹¡×§‘­Ëz©ä¤šsV4}óÎ~ûg§bÚ81jðÐÌ¡BøŠ{WçJfë°U?bPp–³Ø9ý$NÛÊôãÒ8@5‘ßÖ–ÓÇ€ÆülœsÅ ŸyMÿ šØ¬Ù„Ùeýøð˜ªá|ŠÕbL¹k…ÃmÛKa,Á*ש̼‡Oƒ¸¦áx¼~Ïc¿‡TiÖ?Ò€32Aœuþ3M #ô°­žÚ"ÑVl ~yKÈU `®¦ö»¯H 1žùÚa÷Ò‚lrçÌ•,b,Õ!.eH¹>ÔR¾Òƒ¦$§¾–2ƒ`êûȶ6T=€¡VÁ[ÊIf?CÅ´úçBïTvàÀžÇþWx`‚2­Ý㇫“æPYñ’btŽ×“\úÙˆCy5ŒþÖYX)^TLF_+5Vè/ã]xËêw"æÌ\ÚLM ¡÷ä!ü*‡ƒ{qvVª¸ö ¨n{ ?ŒÇªžb‡Å(Dñ˜dˆï˜„ktºvÛò²%!*BÈyA]ª…ËñÿßðË@-à ½+x—mšÈbÁ±0¾ú“(Ýð}þ¥Eúh}*tGi7mqVC ³0ñ(Ç(#X7‡¤ì€é|S¥ ùÈ›Æ>l'™ ÿ•íía.Å!%°1TfÞ¬bg¾ËM^þL+ +lÓS½ç*wO–Ã8þ)àW5¿ÕSs-ä¿xƒ“^ÑÂwÍU¬uPmÊtAV«°_»z£¹· íÔ9²AõŽê–‚Ìù6þN+ìn4a*Tìq"Æ$ÆZa \³à=Ül2Û^ìÉ2¯’üÍ™M‘KÂ|NÕ6™tþõQ¬ù®ÖŒ¼<ùjÙõª\¦N*â?™u€‘T×iÕMNNÇ“OÍV;¿;nÍï•ç»Dv/8ª<ß|ïÌ49#˜B`O' L“NñÔþÝô@)s¸ÖÓOhø.‡uóEKw›Ðeu³ò» ’‹$§:&ê<Øôe؈“©üɘñ ³ÛHO;‹Åáw:Y.óŒƒk”íÆÔ!RC|RrýX =v˜¶À´Ò ¡2ú~Mâ¸ô 6ĺ,;¾ÅBÿ¹‹¤@½XÛ×wvCâRÄ{8{÷;šž‰üú©Ž·‘wÆD8†ÕRhý‚[±{!ÄlÒàÕ÷Um~‘>Z£g³ä [7€ÔD¥ow ý(¾Ïcm‹m¤ÂÂLØùd*õEâ(Ò÷å£Óo4´v§R ÚŽª—'à7þϵ&€•aÝ胜#F$?=.—ŽÿýÎ?2> *±Xe ò‡ÈSüôÉb£k°ï§o}­ð’þ@ããSxNÝÞÚzGPéω>‘v½Ô¥óÌ%JH˜2Þ—l×¶Õu¾5–fò-ùÝ»N‡§)=Y´Ü–MÛ†°pï6h…Ÿ~N. `—¶¶4ýk8|Z±Þé»ÄIÉE£†¢òÛÚñ ³)Öu´èÝàT ÷º€à÷¢(‹ç` Q\ä9ßÝJžÇC¢Ñ='ñ;ÿœè®®—iG_o1NŽŽHzÁ¶!Œ¶œÐþsíÓ<òDzíEp ñä$žF¦;Á~$¨ ö~7zð.êßéê‡>_Á]I];þcà’=Ôš üÙD°!vOz[–r°>š5åê]eÃIKv+6¿=§Öˆ«[Púo]ì^‰½¿òNúÕv}q<Ëh}V»³fP†.)vÇýD®i©).jêÎ}ä8G¢ô"†-"pߌN/´Òª;žrÜ/*‡tå(:‰©q÷Uz1k#\¸âÎZD«†J÷PùÉõÚy“{|;SÌÕrý¿žq's“ÎS

3ˆ /h‚º:¢Õö|Ù†IÔ+¿šCÌ؈´y»O“Žr5+*ÎWzçàLÊ^žßV3W1Lh£‘@u¾ß6†ÏO‘!Ž*.(´‘I^„Eòå‘Äݬb’±¡ð€²¿Cl…!ë“;f!þM!ÎjúHeœ‘öüRðTTcKy6ìv É€‹<Ô¼¼«†Á4ÎâÃÔn4Òé᫽T#} ©²ÇÉ eR*Dù3G°Ô]CÄz…ª€Z¢LèÌ;O«¥÷v×@åbÂåâex4;^IB{Ýô ´ûÞ˜Òé\ÉjßÄÂÖ0,AŠv#bØqƒ¸“œü97õ›Õž1X;Xм;yãÙñ—ËØ ”‹G©vÚºµš%þ ö§–kðÛ/#;õÍóÐ*&–ïüÙx>ŒØÖÏØ¯UL+‹Yä‚>hk|Ä]Þ)Lyé#EÌ+«Ô¶ÙŠ ¡ñ?jvÏÔ”1fºõb$Þªº¥Ê.À¥5Vþ ìuøæNŸó/þŸÔõ-¤P)>ŠYÚDËo¬Á‘Ü\ _P·„t¦‰¼¬–ý¿1Ùêÿ †Ñ@¶šà7Õ=’¹{!òÁk”n•ТA_ïv@*€°ËnM£Â[*‰¯’ ~âAüÔ\-3òY¾ë—ÜÏ65ge—ÝNºg²[§wôΛöʽLŸV'áx.O4¦´è‰œ7h³—¼þ¨¼àTÕ5ŒzÊ™7ÌÓë^"¨›Jûžëþœ.Z(d?h3M9¢¿é•JÛ×@H¨Ã{8ˇ͛½|î)br_èJ¨^^át ø¹Òl/x¾™Á·e²”Š{uqUëiFÍo¸ø’zš9ªå“±t q|§.û½ж_{úu^ô„DGÊŒÓØŒžo^Úâ¼ ÜíXäD 1$Ÿ]L·‡ªµßö¬ù“|{â?@ÿm løó쮩ÀÇöl[Œý˱˜Å­ xÏŠûª´N ïüüž•„f˜è„Ðgëí¤F•\ß9±u•ß§fa­ßk`R›CãÚŽ R†§!”K‹­FŸeY•…÷·°Z±·ª‘jΞ)’)LÕoÝè‘YU\” ÷êžIšmÆËÙŒ·qªiƒUŒoºù‹`u`M%H¶¬ôdІ5Qψ é-&]™7ƒ(g¤Li5pX'oå?2BÝ‘ê;ØŠ]ŒQdÝ |—/×uÚ©ü‹¬¸I¡^ÕV‘Ⴙ]Àä#_\•{L8 §•2}Æ]ÓIÖhTˆ²á,™rФ¿ï¶góâ½Ð ¼À%EØc/»Ò|è ¦,>OÛy+vj;üÄÈY’‘K/îyçj˜ZˆfŸE@²*õòÄĈî®ènW¸ð2ï$;n:2‚j í,ñÜÉí>úû¸ëÜß÷AþÞÃV×Üm4 Õ¾™K{ }}jqÙ™?¤R&ÜV… Hªà< ±Š±ª„ZçhóUú­öí(R¬Ë&n¤s{¯tâ ©À‡˜&Vk¤‰¯ÒÕ±7WF^ÍyóÏìœâwúÊS2)SÝvXg ƒ %$Ë+õö.(Êku¯î¬_$¦¾ãâ3Ï© ›{o¨dsŸémL½¿–»}éKäØ|]b‡ þÕû D.-r`¿àI& ÃˆÝ7FYu}orsJ?ý•}dP%–äü“dÌÿ³?U5ó~Yø˜G®4ý£, Ð &‚Ø`T4l±Á¯³jq½¶5ÜÁ—mSeúcÛÃÁÖE å>ζ‘ßÓêj#xÂmÀúyÿõˆZ“›I”ŽL†Œ•u *Œ”þ;õïƒJž<ÂßëÚ*|°bV–˜Æüx'´s“u”¸¨À¿ ×ɽ´B‰ÿ×r$o.ÅJ¡ï}æõÝÉ=Šoìe^f?TÕχ·È¥ nCõ&÷M †S¸-Hᇈ~ûû£Œ«À**¤ÅveåÔ#gϵZÀëûFæj*©8hÚYú“ùÙ”ÌKx‹nšÕ^B6XŸ„£%å0Nú݃J,1† YP¢hì*5í/–O8ÛúÁþQßF>ŒLý"ôFê¿â,í1eÆ#ZÓ}‹Šguè:'ŠFö±?Ø2éר£#b©ÄCÚJàµþs¥ÂìÏe¼Õ6-£±3ƒÞ¦LÓ„:%ŽO~*óâV@˜“7™×xÞ@b*à £Ôx3ÄÖoU¢òãLl¤ æí8É& ¿ì›Xôê·Ñ•óèyžÙ'ã üC€û^F–A-'†)+ÔÔ¡‘s‘Åiºß(_„d· QÁyUPzP@Ì–$¨ T—÷xÔ½U|æw~±°8Ťu{$SúEçˆxP%6EŒP@Òt°(«‘6î›õä_뽸[T½©îáu£Ÿ”ŸÐQË.Ób•MßÔE[L¸ÛaÜÃш²7u9î«õàÕ‹.Ñà‘ÜÊdCÊzé¥-¶,”m]ǼÜÓá·¹á¢jB(¼5XÉTïù¿v`!<+¯7éPfC?–BIûhSú‚éµo.i=ÃdÊVIYŠ7*X<_#Ÿþ$ÿrE V8e¶hà•ºÜ I`^´8o,f'Xo¢z´8ïv+Ò}-¡¶ÿ^øÎçY¦x%-dÊÙÛ£A<ûTï}¹k¯ö¼åË;Õ…ï^¢ñé…¢r#ú?R˜ÕËbˆ\ê®´&š]Åã¤ú-œ|\“©R’…Ê£1¦qÅY·ZŠe¾Fnƒl4kî3לœŠÄBÒœ'+R˜úà@ÐZÆàšW–"çA6 @ñÉ¡x^¡ž« g½˜ú­P'ªGµBÎ+±hï®94÷æ›ßÅ ~q©|ôAð‡>‰øÌteŽfR;f“g?æw˜¯ï»—Eä §yñ‚e²âKÜíÕ׸~tœ1D•yÉú¼ÝÓÖ¡ÈÅSÏn?¯`?TPtܶm˜BYɵ¤ ‹¬è4ïã0¥ÑPÄOþ{ c;©5(–0V¸3£õùøü9~µ?‡7Z1¹”Ú÷¡Ìy\g;árzìœK‹ÿZ÷ÉÉMA:t­H¥³·(¿(IDäÉÚª‡Nô£¾)ózóË‹öŸÿLϜŮõþXP‘5Á±¬Ú¶‹´¸éŠ4G=fæLoq2É¢¼X"ùz.Èm0‡÷}Î*-‹”°œAHÝ 1’ÇňU—PìB ã¾EIÚ"ÿ7ý³KêhÍ]JÎgÉù2¥¡C d¸OßîÏ ;š…ûx‰¨¨œgð9Ë{þ$³ µ²† ¤hÐÅ ã³VIÛ—dœ !´Püºnq¹9àY1ÃØ4>×IµùÊÏ;«`¨‘¨<Â*ÎÙ{ߺ"«Ì#HÓé^4öÓÁ±B”—Ä¥Q„ºé(xýË =ôSmÛ9\ï R]jKÊ}˜Ž ;áΨËfqGìõ.”Mf«ï” œdÚüÿ]ýÞרß¶¿’„€.­ŠeƒøI“¿V0°N'J¬Á ‘/ç&ÙÒ¼™®"?à…Tj#û©NÁH–ålmðšÏx®³ÚVÔ¾ÕÄtâüS(|sA…Ðö­#ŠjöØ/ÏÝóðë’Ø×n¥˜g›Þ•wøvc†ëü¦ÂZ¼œ²Bí¸>Ýýr*ô_£§½ê8 Ë})®I`UÝŠT6[$ðÅ,C÷ möºËOa„PŸ\,ÀciÓ­Ôáø•ÙSöQ¢ÿ€ŽwþlM±È@X\ú»HÌw=(œÝ Ùˆç†0§nܰ çí’ÏCw a9¥Õq^˜imj½KcÓ!EÍ…ØÝy¨ ^Ó”œÏu!³Ì&ç“”¶ÌôÙ€Ü_µ5W¨¬—”âÈp/¸¶Ùoh/|– sú“y¼Í´ /d¦9â¯GU>:-“z*úLÝûÃ9üø±‡ÝÿôU‰”šä8<ž´ÑwÀ{¦ W{B‘¤ë¸¼g"‘ W*]”‡KëKËÆTr|gžÚhx„PÞÙºþ7—q_"&š'@ºkޏVª,‚™ÄÀ݃^ ›ÃœøÙäÖU›pBErBÓÿö쒚γ£€üWÙ°ú$ê÷–Ù™ Ÿ†¹ŸUjš…2Q˜õ‚ñ‰åM²&'úPàhµÖpB‰JƦkÎ0¨¶JœKD¥òßÿ*ÕXÎãq¸€>h¨º!íH%FÐ;¥Ä5àž÷âÓ_Alø. ./ÄÎ( þuø3½s2ÌA0÷-ÀÚlß½r·uÅKë¬ç”÷÷köó-£þÿ)c“÷y׌¶$“r¶VëÚ6ÄòÚáñß-p®„R” ®ËܲÝTzç†Vñí¢‰ìoÿ—à{ ÷·Úq4`äbþLED°ÍÉÃ.hþ»¨"_@Úu‹Ïoû˜ÿì)c²13jY§ºÓ¯ë’#´™ž¼2²Ø@uVë,Ç«ˆbÞ'fĪÞ$­hhr\aòmU·²j&8àX¶"æäáÓ—còæÍÝSå@öNËTÜÈ™Ñðp Ú(#™â¿âZšÐZø’âÞ¤•JtwAoÝÂ95 §O¯8CØìGçR+¸;c â&’:ÊÙãD·˜õ XÇ5‡6—§R¶4ÙY­ …·3!ó Ø©ø^ÏœC¤n¥Ò°(Ɖfç¶oÁC¤xÒ¹Þ ¶òi ÑZòp´× ïVJš^ÓŠÍã´,3Qtª|Ï)”¿ ÕŒ=kMBª(L_f‹âB“lzŸ…GƪK²ûî#xéºA±{W6`½6ÎÁ”5ä„»Nu^ËYæDRƒ›«êv­ºˆôäÀø¢ìš7Y®mÚ½R2Â!Ÿ K¯6ÑI¾ú ¨w’?%=æªP‡i‚‡Õ õNL ¹ \Nà{ö}ª¦~b?@;)ÃŽˆ©q®ÛOûˆTM‘+Í86¼BÖ7ÞU>€Ÿ[ü…³L K‘§ 1GEÖôVb à›|{ÆËÁX<¨‚®ONB@»ûgÚgŽÍ»©0‡EoÚÅêûu&?èä>DçU?;ƒišé¹Q&¸0Z»àègžã¬çл˟ÊAi°.Ì'xf">x ž9 ³jT#è»· <ÖÖÛ+`‡gñ‡Qe¢¢·×*ª6g+d'd¸X„rHÈ­€èòäB¥ZyYseäÞŒ)ÑFîOîËÒc·%5†õÞD'‡¹ÓÕ,º ºzâåá]íËÖj‚)Êv›¹Ä”ÖºD1ŽÈЀõ3褉~œfÞEÞ¡‚gY6ì)ªéÙÙ 9c¤Ó€åU–¶•ã{íytåtÖ¦Wð‚éá¢d€NpÁõ ’:U6Î2V„ϳ ¢½Îÿþ4Œy6Üž®ê§àzêd¨stVˆÐ V½–Ò‡±_ÙÐÇ÷ý "õ`Ž­,ðÞ°Kö4¯;0v2$:7,—\∆ÝFAédŠãWYÒoŒé 1b6*·O’ÙÍèqéNà‘9É‚REÎlâYÂú@ñTÀ½6¾s'ñ'lB¢üºŒjË#òPµÓɱ­š2-8r¬eþ`Õ„¿è’É´½´ËäÒõJÄœK¹¼²ý»Í^YHCÛ;´ÎX&¤Gè’ï…;öpÝ’ýÜ*¬˜Ž«?®L,Fl¡–%K%Bô-‘n7ßMÔ`º¶K—›[U`»ä íy°š™VŽ4׉I}´"òMºcwˆÑÉÖ‡4 ,?˜/Ô–×M{ü'(õ»uMuk_÷3vhÄv™mI—`qå™üo‡ÊzÔEÞ1”Ö´šÿžgÎKËÓψŒ 2JpÈç8~4ìGÃŽéù‰ñutm°®K"æ#V*S f¶eE[Énÿ'‚@Mqe5ôE`oZy¿ÂÎVßûÜÛZå©[`òý§/ج´DnÞŠ<±Åð¹¥èý—¯l^j²7ç<är|ŸÂ'9xï¯/y¨Žî3l®ì[OÅTΨ™D"=èäzÍ£‰‡  Þ´tcœæ¶çnA‘Q!ÿaîÒH²¸Ë|¶ö0r*‚»ÿ ߪ®í ÅM Þñ ÷1ŒX•þ¦(ÅeñsY‰aÒ61€¹ß  Þ`Û”E¡oÍWgíÞo8håîbO(Ï”*EGã÷®Ø¦4³iÓÙ*h q\žÍíkùêjÈÃô‘Hw‹Õ²¡,¼’¼ÐwvÒ–ˆA߬ÚNÈú’ö`£’a¤ˆëæ‰z÷5I»õ£lnìZ¥-'b¾©¨;ÊPIA%gx{Ð"Ò°|Ok’f·Tê>c¸9› Û±wK©¿^„+ĵw ûEÿ—W$3²Õ^CÆ%ffWb_;q»×¿é¨,°f½2{˜§äõ­‹æÊ¹ÔÙ8à&Ñ>#@æmÛŒšIËöpÌêtÔtg¨¦0ìÉa»K›¹#ÂJ™Xɪ2º9Z×-»ã‡®{ƒ X Eˆ×A:1Cíh·K;> Fûw€ÕèwÎÿ†ê¿©Îº]Þ9×Lëè‰OQÕé÷êcÑÕ(˜®ÉP…SÖw‹å˜ñfeØí¢2$¢þWvŠë¡„)ß+Ø t.E®]1bOÒó"’9Rž+3¾/MFl~è5˜á_!ª¢(êlÝXÙûlÛ—¸ÀÕ®zµ±ÄÎÑ>=*øõ£"ßáŠéÛ¶¢ýá.rÉ@+ËØ»†Y;±à·înzHL§Á\10v¼¸Ï¿!AÁ2äŠíq5[ Û¡üµ½›S²™C‰ö‘‡öÓ±1hÝumuKnÚásô8V#çŽc=‚1LÌT&Çáv0…t°t+– c{y]<Ž~‘.hMáJMÐDó¥Éú{YC÷ب™FtlnFÜ\å£K§€|,´ÝŸã%¶k„¨]¬y<oóÆ¡Ðìmïnò x¿L”fCM$ñ~?ßÎ@!v©Ëz4’¶×luañþÊæq©œLó©e÷·SMi–á{‡O½]qÞŸ…öç4T#ZBÿy°hŒÖ¯¨Èd™”»Eûv‰NѤ5L¾öFB2þ x 8þ*ºÞT)¢5´Þuk —â 󇿋P.—^ŒI* ͦ¨&0>X²'÷åLÊùÐ !ÜÓÇõy©ù3^?ünfÇw‹Rý}8÷ÚªáÒ …§ç-fÉ\o^Ò¨§Ü´µkzøÔâ‚¬ÂøEƒ²t#ø"¸ahè‹N»îÖ’‹»'åÅ"ú;íë^ñ]M¿_Îê>_u-^zªÈjöÏ}3ßqØQ‹q­¡·¡([˜ÉÃÈ‚<±q¡Ûß š#Ñ&€mVÐ5ã¸äkÎØ½ô¶bª((ñ’ÝW7C>ò+ah-43 È¥JýšDÖ54Ô ,f”«I ÀR‰—e-te8ã$úâj D÷8"˜¿¢p!“íȨOcP #fÈ¥@QaZ–èWªµ˜æ0š{„ŒœGÖ™ /ãæ6=íÇ ‡£e”{uó n#øïáwµ¡¸Õè[ä|ÊÛ—{™s¦mPaZh]l0–šm/JZö¨‹…É9"NeOOðË"gïö´$qWªW Š’}®FÀ‘™°e²ˆ=é$ëh"¦£µ‚–pÜõ¯Õ™L(œ`K£Éç8æšU½©(c^y‡° Üü“¾PÄ8iZwkáŒD§Š`8[°%.®[Jsú\Xä“ U¸ü¿]¬ÚFÐ< po5„V=¦=ÝÜ0ØÖ?dµŽ5F'NþÇ¿Œìq1=8ñœÕªv‹!Q :žN©$Aú:µÅgzDîŒØœb²Ä sËe®Žr‚íÀ/žôÓ¾>Gdv#g=% Ø­€â½®{ÁuùÞÌú<€k}Ä/Û§d-Ä-vâFY÷'—[ES¡s(q5¤yJqT›øe$á”yÉg¨hž\¿»\ÁÊßuAà"]îá';LïfР&{`™ˆó‚/>o×çu¾v{ù+Rè¢b’?ì xÜÙîè‘3ÃÏáØÅ Ê-ŠX-æFF·L®ãܾPÎÜŸ_×`”å ÁKÁÝÐCºîhjÔ‘½ôžRBåÄi¶OãÞ[uÍ`ÅåîUÀt¤ j€SîyÛCÎË5²¶èüQR•%L­‡¶µ¹çŸÌiݳ´îpº¯]_HK¹ê¹¿]$µc?Á"ÞÔõ Ü SEu”šÛËÏøÙ‹@ž?Åƾ.íÛñPÏOc'Zr”2Pîë/ÑÔx"‡í î,®Ã¾j¸ÁŒú“]Êfª+:ü6rˆ½Í‚(ï„ÚTAƒJå)q²?„Øxyɹ‚üU” õ“¢BêøBl1Æ_E1L’Á†–8ñ[ºƒ!øzý´ï% ´;]^Oy=B›3žåîRñ½"$#,^µ½ry^$©Á B†{Õóü¼? }º´1Ä Æ®zµ$”4w†Ý÷„?ÙÃÆp9Rí~Š»FÊÚ-·,"øÆW>YÑ„QÆUœG¡Ü`œ†x‚ ½ŸycøþÓü»,»ö´TÉ¥Ç6±;ž’„mïlPeò4‚ÅõȬ’ŒÄ_ ¸‹J‹‚=8XFßyœ5o湌à|JÀÈʵµbM­.Í|&Ž.žVÛŒî]yf¦õ€*¤ü¬PÑwòJŲ̈l…ƒDÊišIˆBÎÿ-ìXvźI!÷–©Ýgăïc‚ÃϱìuÍuZY>Œë­j†œvÁû–ò± ÔK„aÎ ’?J}޽§ U$p‰¨ã+C4 ˜¬&Z½Äa/–|¥Æ¶/ÁÑó¿‹Á~²›¨+«Ô¼›£’FÉ—€KŸ Ô4¬ãFlÅt9Z3•ó¡~8ýÛAÛÆ¹\Œ`ŽàšÝ&ûfÃ}9ç ›y¿Õ-(÷ƒ³ØL‹&8mu¸JËØÞýG†w¥ïã÷%•‚Eùê><»Ñêë¹ÅÎyŽ™ŸéêʆFý4l+Þ‚|ëžnÈòu“òÄ+íõo–&ögÁZæ“§´,†#FÔáJ×=—±£XÞW’‰ìÕQ©A[Á¨Mqÿ޾qO7Ã1À.Ëã“òtè_}Ùëx!\åßâ¿|%6L·2ã9/0dÆUΈby:_’jo!î:–¿°ÉZF¼¨säjÉŸ‡H[sö—+¸EïӢ;Â")¶Ö4û“8 “v%¯=^æX|3 ×IQF%Ò ô·M_ÿ]’›€Ç™æËêä #Á}F^¡´ÝåMÌ™ª ˜ªže:˜M¡3jW¯Ž+“Èwºk(°L¥NNÆÑÓÁä)Ø" Š"¯<ÄDi¶ôGÁŸõЇ÷(T_èO8Ù¿ˆõ‰ªwÖIïYÒ«f".§Aôa’äš—£Ny!kãõ NaöÁ R?úÖ/ý¶Œˆåü˜!Fª³x²“aÛÌ”#¸mïUlú¶˜ ¾îL`Y¸'ÙGH³X‚£f³ ú¬dBûdÅ'4«¡m%\ð_UYu vüX;S­ Ǩéî¦^€ðGs&þ8ß)ïst!êÓõõ åöä ²éy/ Ä mÆ v­îQ…õ¦ò.ßO3îry”.™Š]‚såY/”âð×…³‡–Æÿö®ýXÄybmŦ8öZê^H’µbúQÜ2Õ÷ØB®ãôœ¬gë8YÚŒž£æz/»,Æ£.6æ‚?V.ÆdÿÖlò*Ž’X„B±Á^Nê×Vƒ·ÔÞ{þ7,ÔOœÕNT!s,®Ç¥KýZ4jÄÁ¥êçJ&‹Cµ¢àÚ¶À>¹PqN[WÔö™:6ô WÖs±ó§h'[Z½gÚ!i¶ø6¤œ[(í°8AC£±÷5y? ì/—ƒ<°®C3³ou$'Z½—ÿÂdÒ}ýɽRFÐ]Ús êÅE©™lÆ,Uí­¸hÝ0"ž¦“Ám¢IÃ^­áI@#¦®ÀÞ5Ûm è…?=?ÐÑI-öíÝ>´ÃÈÕœfvÆ›C“¾Ú7%“"ÃÄCnU¡Z3&sG˜0Ƥ«—/qó³¦J*wšÞëN™ÄÒGU+Æîò¹Ø³)Ëñ¾hrøËÔ‡1š ï"›oß©-pÌ ¥¿%.kÒ™¸%BÏÖ|[vLVÃÕ¡/3vm´rN’$a螪â;AÛm½"åA9wš¹Í ’¦¶úypý …Ñ@Çi’èGƹ`ÿ&Ú¥0”y,õ«»Û-þÊÀqƒï“"¹úeTnèÿCÓö{„”pÎÆ&æ×Ýļ}ûî›,(I~ˆ-¯”™ÍŠ ];4PùÂ%úE€ef‚)2ŸÅȈ"ê*c‰Ó¼Þ{ͱ=³­•H¢a_„$T“*ÞÂxU\Ìzâ/{}˜¬sïœþ­ÒœkyÌ8` ¦ߪ 4Â=¬èAÉÓT^fßó’©"œËCЈ, ¸êIl×® >«2ÛÔá˜&ÇŠÓZ:#½¢-ª“8Î̧ܦ~rNÓxTèVvà·¶R¸ÑÆÀ]`KTòàp®æÃQŽ5žö•Œ…%ðöâ°é+t?/‘œ«¶°‰ìÊ—Qvw aœ¡hãrYB‰¹¡æëtG Z.¥ÞèIÅ𾃅 ä‘àú[ R #¦ííÈ.!`¢ö¾LÊæS÷&âm¤š=â‰áÀZw¡m8|-b½td³¡[;±§s)­Èð Õ}䈰ìæÎ){`Ah-,žˆ£ ’Õz5LÕt Â.f¯T4ApKd/\¹=܃9å¸ä¹<×Õ¡Z’ »;Õ8rÆ*Ì⪆µóÜš™@`Q&Ø÷Š‘ÁÚÛE<-=B=w> Ɔ¶H¶o׈]±J§Ô‚ü±Fá2è, ç‡ÑŒH¼œÂ9#øÆQd@ßENr®ÐL’ácG¹´qƒ}4î¢K9}Ï™êƒ^Ɉ‚ö¾®œö~RÈ0•ž9ï|dì`¸kâ÷·œc‰÷ÿ1g}ÃvÀ}ôÈ'ÀJ¿£¤× šü¬3BXµí æè­èûÕ¥kÄ)q™Ì hvMzªJ)î믔 ?Úkš¶ß6ä/ D|̨¯ßè64ÞߌqÚ™P°¸ÜU.-ÑB|:qO´þ D\pGÁŽM{„9té'R6ÊòÚü,Õù;Wê¦oZG^_òd˜½ê÷ÄÁv€¨~58ŸjLéÓ6‘¤Ý¤¡c Ç·@ iM%†Q{;¾œ) ð5qØ¡tÞœ©ð‡ë ¶qwÚŽŽ¹zŸÒªwôµÁâ[°T0âç¨R«ÃÌÛ*JÜ„‚2…Pº=/‘'©ŽB€³™ µ76¯‚*‚R©ôIùÀ=êN­i~ 7¸•m)oì òur &Gv£k‡Ÿ½ÒAAÑ£<ýkέcø¸©(þª`ø£ã‰Ü—\`gýSùADèù8¡”±Î› ÝDÇôÇ"S´¯éÌK§¬Ædei˜ä¼ëùȲS#w⸡TáuP¯S@Õݽ$â3p¿õ¿±´ê»¤IŽT&êZyžÂm¤_2X´Ø ~g½iùTq€I‰ß¯I@P ÄÐÜŒ´ks ÐÜ2$ÌÞ¹»Ð-_ÔRàëø´¯õ‹?ìV‚QûŒ],«I±oLÕ‘ü×gùYCK»IsÃ×.Ú‰@AÙ¤Ív—ilMÓÊM`$\ÏèiÇao{þXšÌrçšµn¿ÙÐ;Ô¶àÛ¬©0ŽÇ“ç$ùh$e[¹ ëÖ¯‘‰#dT£_÷(ñ8GiÏó ªFÛùAžU ‚zóGd¹¼¦Æ™õa„]‘6ÇÕh~ÌnŸBR €Ë]њȩ–I¹Ì@Ü ÒÙí­–.XÉÇAL¼L³Þ‡b@ÛÁîSáCÌ–;ìȨdP«è²ža½ ³èyKM3aøvAÜ{þ*Z‘’4µe2Œá­ñG¯àí±jó—)5Õµ]>k4zmA£çÙ{"%ýœÆÔ‚˜—5âRui”%} ¤ˆ‘¯Š1ç¦øVX´™1dø§üòAN9×um€¨Xús}Qêv×+«t[ªUÚ4šJþ¨¤8Ú .EÆ9 šÊxò,¤Ý!q1gô®³Õ·¿®C¯üt(žCwBµ`þJÁÀŸ9ÅØÇ@‡˜kïÄÖã–ŒÌM¬¯X г‰T ðßMÉrÝVØçË‚0ô±CüÒ徨K5=Æ€E3ª8†oÉŸ H+NHެ6ÝIW&²ˆW0ªõ^ZùÚ£F_'¨ õ~ƒw_‡b5`J ·£_9+/¯i:ß,äðòáÆŠS“5+'MÓsc^a óÕw³×)狎à2‰‰;´ÿ¡…¡,Ärü\|皪K¡Pºˆ œÓþàXâW…‹/"L½è úŠŠÂ¤Ø¨ÛX4Æ¢gBK'Í(}ìJMçŸS8ªHWár3ÞˆªÐܘS°9ÊG‡›³]+Èù…€}¥÷ÏQ¸\nÜÌp.ÆÎ¦Ñ…?n˜öv]Pô -ëû:éInWkí™è Ί“5©yºPàÆ3·1øQÑj|ff¶ŸÇÿ4`a8fOv¾ix¦3&;>X/wÚZë1m|†Ÿí˜õo‹2ÇŽvÑh•T:ÁÊ’gBXÁ [ü7YTõnG×{àr5Ƥºx€ß©1§öÅ1»5ä´ßX Í„É;°ÉlðgB˜((uذ1ê‘üK&Þ•‘ý×:Âò9·~×ú0M8ÞiHï¢k·Mϋ ТӃÙ"Ì6£»DË$Ñ=æÄnh‰Þ”žr5nÔ-@›IÆšº3QHIqÏh¡NhfQIØ#T|8£¦2F'/åKÑØèvAÀ<ã(ïÓ ‚iåÚÛ7ÄÕq/ySËæò 2øJÂeÁÕwªeï—û«"2´"º¾ÁÐ}ß`.Bó#f`þ_„OJѰ·mRžB¨+Æ#ÎÂ;óNðl«k> ‡ž 0ÉS/4ú7˜f‡â®·¥×Ÿ±Å7C‰Ð]Š~ì“°¯EiЬ\u³u7¯J†ì‚,eu>&7ŠLJÔû†¢ßx?%‡þ ñ(½QÏï„£Œq&!Q 5‹±‚œŒ–Y`þY–!‡å>䧬ƒX(œRųd&“{¿©hKº6~[õ6ÙF-'üIô[i¡í#!UØámuúx¬•¢aÇߨÐ@Q¸¯”y¤½šˆiíùW#u§ â°0®ëûÍòÔÛöðEh| °ÿ} ¨ Q¡¯¶Ø:¯#BwØ&0n¡Rm^õÒß _Tr‚¥f”ßÃD¿Ï›VŠ·î…D³¸½éÞ"çVÎ8†0i²VõŠ Ndg%®ÛøŸpUÂÛ.1µC–6‚ãç´îdÃ: ÂŽ 2&æmt¤_ëjÚØ ~XëP©yÿº¬hÎÄ7ÒMZ¡# KçµpÓ‘‹­ew£é)t©èÅn¹¯ê¡óv"<옒øßn-²íâáŒøôÕ h\2HŸåW9cc ÐÐJÅzqT@Açéo™CzpZËÁjmøóB˜ü_<®Ñ難WÁ½iW̹ •0æ*º@¤5§™œ@€Â%D[û³ïIªD¾ÐÓºQ5µ7yV5‚ùÖôÕ¥¢bé_‘oMpjlG gDª&äãÇÝ;úÌ懤bü¿ 5By4ÈÔ[ ‹±#ÆÐ-–§8*¸$eXg¼™ ¼Æ¥ÿƒtg_5Ûq'ÉH1hÎû…¢R"58[S+fY\Ñ-Ôõ(p z• i §(ª1ô¢Ô‘†4³ÎkíÆÌmì¡u!×åpÅUtð©3 µ@1·ÎvBJå­-Û­w®NõfÅelЏ”œ¦Ï¿ÁŽQñs0I“·è:-ˆèðÿ¼t¿•·™ÌÕ¸p€Yǰ÷>1_•wÓíi¦ù‰›Ä “Hc&å´ºH‹ë¿ •bί¯ÙS&¢ó²,©¯ÝZŒ°Î5Ê:F—fí¾˜z‹A\ WëèC?–‡Ò;WÓª3Ì8xJ3g㽇ïÁZüñ]#o;Q5îeÎxraHÖøòÕJÕt-¶b(¡/ sŃ[IVŸdu܆‘ÅDUw‘<Û- þP|¨Wy¢ýZ¾õ–ïÀvúM·®¼+¸zµÓ} ÌÖÏ0žr½oÀã:8ÈȽ€êyÇžÝ`6KºîMR9EÇÛ×£—=ƒÇ„dÚ ü÷.9ùÀÍW/öJaƒ(G-Ÿ·¹•RŸBjÝ\ªÏ.û¥<œůy(;wÙÃòP£ÄaUßLéHVBâýsHÅ‚wïê~èaïÜ1¾"ŵ Îј˜jÖy/𭳂:‘| ‰+Y]ñÅsV;°¯Ø{k|ƽölТý{/±µþ¬·í?àݦAztWª8„ð޲(:=ʸ·Áo@zý0n°1£‘6É€ù¾p82äàÙEqO%@}­ŠvŒ¾UBuT¤‘8TdcFÄ| ²DXI÷Zãͨ$L&œ&JÞ;šÖƒý.ÂÓTÄh‹­ÚÛÕ¤K‘àŽÂ jG§¾„Ó%FWrlž6c}‹Hðò-¾“ùÔ?çùÜ]yw÷i­Ò§óý:Ϭîg¸æ¦µÅf6¾ÞæÌtÄ*9SdDg‚çüåŒ9Éï­—Þfë°ž#TXòµäó‡áÅV¢¥YúUÅÜT¨À±:ÑbE–Ë:¨.S €‚è´w6ò<’Wz5Fuûù×m Ý&$¥µ”à Õ¬©”¸þ"+O_æE7ŒÖŠÖ›ó*ïªìû²¨©Ñgèõ´Tûb™Oª`ÜÉ•Õ`ÛÑ߉ Âñ>`>tÈØÌ ÚHPú|4¨P/0ŸO2[ ¹íFa’j{BmÖ&<ÖÐ1Z‰Þ¸ ìÊ[CóƒQÿë÷;©Ï0OÖÿØÙ /pðb Í"×2YCÌôô(à»í‹­R#[,þô‡›y1ßBèbøý®&Ï7ý8ÄëÑ6kÙi3Þ¼?r—iá~@ØC@$N¡Áæf:·éª©~DW '‡>˜„MÉEt ±Æß qW@hì°øã¿4ö·M3Fv ¬/2?9ô<ËâGŽAÃç›BèZÒÑbwýO Ægè¡Äßy JÉ‹°À{7àWN:»µã±Öª5ïLê 46zÀ%ÅÕ|œ‰M7ê.„Húcò¯4.g¯g•QåÞ÷‚òu°^u޼L¿¸¶¤Ÿ˜gLj2d%~v´Épdz?޵2Rɪïè5E6‹7l!‰ß>„8¤A½áz½z¢ÒëÔF0˜„k¢êЛx¢* ê¸-wÁÅíG¥òþóÜï0šä?,4Kd8Ëq Ë·q§yqåXPõüµ3g·mýÆhª«qXwÛu‘û®D¦¦>ä ׋^ÝÀ¯6‘ÕKT!&o¿¶³5‘ûÈ–…¦õÕ>ŸVÒËå ‡¿Â„MÍÐf! ƒàæ nýïx€Ì’À-h¶­•ÄÕ‹ˆïK@ZmHlÕñ ¡d±Î-­@úïî„¥Úx’™ó æ•H^NÄêÓ˜)u oÇ*áhô Éq±xìncUÓôP#E¿5øÇ.…kúâ(ýÍÐOtè;ýAà sÈŒZy9}x–ÍJ&›c½Ç6ãêŒÏý­æJáðÞ¾Ù‚µ†á¦W†}:ÔÃq×;&ÕNΤ:FV¡P§V%íÕ9Ølkx¥JÇštΞGì”j»a‰K]ñ«Vþ±—;«4m®eY¨sТnJµÐ©îž‡+ùÛøŠ®H0_{p]jZ%\RvŠ»97²adKðDJÀà]¶ ×Ð'õ’ŒÌ«ߺ?L+˜Ê¤Hj{U´r좎Z41¤%Žeײ¦ï•Dç%Ù$A@‡zÜ®ª›Üæ0oü;¬kT׋O¯»K^†V¶·Wð¿òݦ’-~­pÃ.öŒ3üVGwA—@`~‡§A†î2H a²»’(z]t¸ÞÊÅ$k up‘-á™É÷BÒµŽŠÎÐ*ó/zlI ·âK»€wƒö†r%ãFÎ|RÛAy$N]‡@ƒ”¶'çÆï~Ñc7ôHL›º“ú>KO÷&‰¦½ñÄyÕ¸Ûw9gòaž;Ê“ïcMŸÃàaå¯ÿf±ßx.CÍãøfOˆo©üãÐËYx êDƒWºÏ­mR:”›™ Gí·Ã>Σƒ>ä)×^ö0fk.í²±tÑ…%cÒˆü~ÎN ´c:¨=Ù7êwgÂèïG";•/"@Ä"¼mðضÊGâIËå‡Þí éióÄQã«øòSŠóI0iŽæcÒÉv+ù«*è ß?×È‹èÓr9èQ2‚6ó?+”µ¢PÉŸ®‹i¶…ÿg§áD׿]Ÿ{„ºå[ÝÀ/hÿS4µuÖE¢WSÕ‡éSýQ_< ŠÓŠËöH.Aç MËœ2-O˜;Þ¦<„N¬…0§‘Td=ðy#éÄ<ЏQu¦[ C.žôÁä}®E¸Í+Ûýì¬Ö,zøýôMÕ¯²€*€¨wvým¡¹±äèž ¿Ýš‚mÍ;ëk죮·Íúåg'1´ö¥—ïnWd1?ÛNàÄÅâòá\eÌk5H»\<²µ ì ÁU^vîVûx0"zGA+')¤)4@¤ b áÞ‡5FmÓD$jY ŽÅℬàœ(š‰Ý^È“ÉXôÂn¬¤•DåýlüÈÇ%¤‹D÷y·^ðýU•Ü%dÊ”EMcj0aÿÀµ0ƹ}«g‡ºzÂÚPÂóxð´©¼¸±ÉfµÑy‡ÎÌøˆë\'þ]Õ±ˆ–1?ÕÓ§êÙ¥«lÂ+øNóþkŠ'¼‘,>6Ú`R¦“öaO8€G“¶$¯,C{g—/¦#"ו›d•¨uñ…ØXYws÷´Ã|û¥?ì)ñŠo—s­,”ämf]î<ª¸T ®Bdñý F† N!ŠÎåÆŸIÕšÄ9gäRÑćBý§µûˆÔo’1èM "qÑÀD_q¡„—÷4éì¥PÉÏp:ü>²yøQdï#S£Ôþw‰åª¼uŽó°+â۰\¼FoN¹+9/cŽ y{¡Hy¬«b³1–c‘É!éS2øwGåÅPV„ÉŸ` ѽ3Š”Û¸¦?©ßrŒŠNÔÜŸmÙ¬—µ¯û¨¢þV™5ÛçÈë!Ö^š)ž§ùmìpðkäG^};Ü‘ÛÈq¥—ÿ}¹nVâ“ß³[9{”§Þ¨•®ã_—¿.²øV3 ظY¯wH ¼yc^. !ËódôÃëA ާ4hä?Ä8ù™½hãC†Ì•€/µE£}ðf0u ˆ«Rz·/oVzFdo00d'›„tùáTséCÁ]ÆM?ÉŠ&Èšêöé+r¬ß€õˆ\V‰ µïŠH›h7Åê¯Mùa'í­U’‰!a"F…Èé—õXáI†Ü¡#l¦ Ò™ÓuMå¸ã÷Œ×—‹èÕ·ànÞŸ•pþ.öa;t“ú äg?àEœ[µ¯Æ”ŒÊ‰v‡pAªÆevýa¼,2Þ”ÿ¹ò¾^–"$+ãdB‘œ„ÂÝ~·ÈZb ûjTp>q‚›n»KHb¶ÜsÖ/Ms.›7„«)W¬«×ê:Û8fiZ}À¶ÊÞ>Ù~4™'6Wš/…¥–¢V;Lû„‚¬Ï¡ƒÃ*Þzë8ªoãUå‰ÓŠü4L+lê;L2E{Égëé~ç+5€„Þö°aNÿÙ%Âù…€óVu„6VkBmà ébñJqaòŸB½á‹»'ƒ¼¶?—{ò'7N±ç!¨Ü^˜ä25~ÈÐÄÏÑV@êáO£½ ¯#Z-¿[Á¸*Ä´6C!uÉPpß顨ôÚ,/­æÙÞñð4i§b(m_1Ù[‡°d˜jŽÖSUëŽì“Êó»rj‹ŒZ\¾›¹.r,KNÜýÆl&ã¹ +yûkÓq,»zú­%Y¹l…OάxëÆŒnSŸæ‡>§øŸ è¨x1²é„öXS]l+(»Y¥0ÕˆÁ¯Hiù“Ò¯Õh%.~)ù—Í;Wy!Ìc¬O@üP.¯ÚÅâªf ­™Š/ËÛÚ`J͹¨|g]ùØ"^çþ=k:¤ ŒÁìჺ®¥É8$Óà¢iá"ýŽB` Šj’ï!tŸßE6¬¬œö(œ~ÔüÅ}ŽF›·Æý²y¹Uš*»¨)­îm¹§ýgYyº@Õd”CRWF ûÏ8;Ð{õ‚H¼x'ZÈ £sö[ÿݯíû36`aBYD7,“–oÕèÎÜË'eÉp —댿޹Í#X{#x‚_áº{&ˆçÀj–*7…ÜÄçÉCaì9ùbÒÁɘˆ ÊqY×å'*»!íp^ðîŠ:”3Ò‰&×s7¾³]ö=ý®Z :» âiõ±˜¶.€A¬ì ([ü|@Ï^ÔÍ+s«ßœÐ5aJòÏȪ„iÕAózSºˆ²DYÚLY›ÚfxtªW7kàyBÐ+÷aÚÃt·^÷$«Lîhì¬ó>Õ)™3¢ÊÓòºÁ¦R¥¿›³ÀrÔµx„E’f߯¬½¸ðw•<Èç}’íÝö Øxy‰ÅÑE6+Õ{‰¤Þu èÀ$5ŒÐ‰µ\IÉ$¦Qû•|@™“áûöÛô6%b :aØ!y‹§,IÑlX¼½Èã Rº\v—ǾLÕfàßt0~'?'i,f4),>Ë2­ß¹A"× ~é3­NàT¶wÅocß’½7×:ïÍÈXéJεé”]¯œh~ÃŒf<Ó=â UÿØÆ]•ÊÛµ ¸ÐzU;ÝÉ$U‹Ž »Ö‰…Jº¢ˆB¨2Ì'{¾æÿz<©\½çª¨Xéª.X”W5£§×°QÒ!Ë–r* º˜ª^:Åé¾]ôY<¬á F[\ËQjH‰;³fJÑÄúzLuÕc2 {ˆñàŸu}¼üªÊ¼Qz 0ç.°[Lk´+´oÚ W`FGpÛϽ†Pø²š´¬ié×&»R6–8Ö@éõ¨PQü0£ÅF5v ùc¶«3¹óß RZ{`¸ýÆaºI<¹2¸{uĪ$: Ç‚º«N‹P/·ûIO5Ûæ: œÏ&J7ß¼ÿ"x_A〰L·¸q(¾Ú€ðRÏ{ GÒ +tÓ.ˆH'}£a š¼b4 dåÍ/–^ÆL:o”Âp÷k馯JeÑÿ†!SºÖ~··q2Ø98' öO°?Aÿ°£Âƒyu{¼œ>Ò(¹jxÖü/ðÇöö·ËáÄŽ»2ëKÝ3Õß9Ù|<«ÐŠÑÂ[!n>H…s)ã÷DǧéË"0 Uç6Âú]ux™Þ@b2M›x|ÞYZPè¿"“½EΛ֜Eó S4LèÆ×%\ç¬ä\ig¯ä&—DIKJH„Ê&yþw9îèikk[º}Ž…ÒðTNÓ™ÊÄUV»nIM©Ê7ÕÕLã0 4%R=5lOO¦S&¯ š%mæ@PŸ_ÏpBºvHmZ óÍ®/·/ÛÙ± º› ¶$0A‚ÛUõ‰›×? ÚU_¿> G|(ª`úFþ€$Šs‚X0#èµf·{}†iüu,ÊÎ1òJcíWræÝž©â>óžs«mP^ÎNh¡¾¾|›[5aê{¾ô9:ë‹êþ9×ÄØ=zÅïÖ;öê77o¿8ÃH.5jW@sË¿ö§PD5ˆZ§^p·/  Êç\Úã)½ýÂõW rÕ´¼~Ý>[uaZ¿þÉáý$…Mîoä•nhgÃn«õ\˜¿=êúmVp*²šˆ3d¯Qè8‹ˆídd¤åŸ‘ÿM Ökã¶È/°9 ¼;qÍVst Æ×ž â$Ç•·B®^… V!ê5ç…0öO¾1­ hé©AbÒ€8fÂäST[â»1HÑß]ŽöĶƒW,&’2ùhµüFîJšRPŸŸG)ÚEöÂG Z-¸ÒÀ…÷¾WØM®µ4kl¨ÒFé¤j›)ðÕ.î®N1÷;Œ@òx”WXAÓÏ@· h„Ôgð­ƒ@¡&¬Ç RÉ^Qe—s5%A<¡Ÿ, (n>¥çzI/±WãÓáü/h‘@Ró£,6e7¤¡¢”ô-+ݹ´¢Eêël ÝÃ34qëjÑeë“PGÒ¥lyW>¥ƒwmÖJù2ˆ`µ¸‰Êñ(qèr6ÅÄãy…8U83 = ¿Lª¼…Ljþø.¡Nà­Jï­¿îÑ %_¸¾ä³ÍÎÇ^¤x¹ÒÐ<õ»voroÂoYf¦)èè>¨©fgNÃjo˜±Æ àmŠa^þoTЄªE32djšgçÖ½#g5p³nt.Z•T À=J‰¡{Ol~ù9w"ºìžÖ,\“óy—8°¨óÏá·² Ãô¾»êˆ{šP;˜˜Ù‘4x$q—/#Lt9Zè¶$ãâòÊÓEµ«¿7Z?¯úiv#±÷Õ±hþ»¤~9ÔÊËù¤ƒ¯Ý¬:«,tÎí*%â^ç€ÉïÊ£ŸÛ•¢Yäs–¿Eu{RW4Ô,H †ø#ƒ©²8—Gcëî“„ )âmáÜSäépÐÓüÞ<•YBAP8ˆ‰8yƒØS+1Æh§"…ÿ¦¢ÍDþæÛðïÖ¦´) çŽÅÃ'‡*¹¹›,4GÚxцÍã x$Ù)¸6›Ç7Ã3¿F¢²^q:D˰ï+FÇï7ƒÈLLâ:ªa_#EŽ”A1’°™y«øÎZæ‚ØoÀRŠ»»¢›ÐÐòÀ?„ dšÏ$‰b<P'ñ•ž»«Ö ÚjôEàÂ0}û£ZrI7}ö$7mDÙ$’U×™y”üÍö¤G§/ŽÏK_IÁ"Û=Áª‰QÖþZÚ%A~lÊú¦ŠU•æ…µ`45Z°±Šy›µÍÆÕŠÉË&e(-Tü*Ò­rjlzGå'8’˜÷0MV÷©Ôh&Ú ÂYn†ýÿþA§¡:ëœm]jò¨ Û”öŒ`A;:p´nåJî:ü$ŠÞƒ¼:¥"½_9 ¹(Ûf®ª³r1.f™…¢Hœ"öÙ|ÐpM §ÃŒù Ôuþ0$¤ù %þ£.­ûT`…æÈ„…ðþtLP»o¡÷¨Î2OÏý-§žŸ¼Á†Î_gغš7«µp'<Š=|$CbC#þ—ñÿˆBg·øëk1êóFs§ïíÏeâù,á×6ÙXPFÂ3 3`¶êãE¸:½›üל½²‹–Œ(3»µq.pÄT5˜Ä’ã”üL¹da:·ú ɝС¯`ºù5SbÎíô”¬FÃC$u/z€éúÎV¼.ÌÉt‡:Â&‚ý q5·PGz[I|²ž°‚Á{zèC_xé±¶xQµQ}ú®—Šúþß(ðM#ñ½ôdZåmêg±VnEM´…ÿ§ðœ"Xµ‚;H;ô²`ñýÊ2o*÷®$Ãl«ÙÞa)uB3N>R¨ÅñE@¿¨»o"ÔÊthÛ1u“Q´¥jHNÿªåÂRšñg¸™ÀY!û3Æ$9’¬ö „‰2ß^‹Ã6}½/ø:ëCzqÆQÏ šþ5p(¥ZßB”ÛÂæ¢|~0þŸK×~Nçþ‘“ÊJª™Nö‹*Qw©…Y¿´›üøÁ úÌ{¬S2ºæé-š&‹\ëÉ ibwq.å“.FÜÖ”¶IcPü@Åqþu8%êϤ†|!¤âlfðˆ,¾5žž³•'DCþ´kØ’Ôä†rßCB'ð˜ÛéÕ”é'Í1%îv5jÕÓ#fôHHù 8ï•!Y6&!AëÍÚÓz èüŠ„.Ç´>vµ®YT˜4ó Ǽ/¿:5¡6ò7VrÈLO©dýáÔ°!EJ->ñ•‰Ò!eh¡¤¬¤|#„sê‘$nƇ»£lLRÃþÏÍíÂKbÆH¼Xä.c‡àH&jÕ š¿Íqp þ¤ü ÕV¼³–=Ê=Áå@»&Üo∋4ÈÅ@‡#3´?Â0Úª ÿ&œ ›‚}Œ›¤AÁ¦D^iͰÅìwås¤eON%FÊ…5cÇ:¨¤(±±ŠðÙµØ][Rg“8›î_ê;<™«/âeˆèg/»R(n. \¿äa:Ö,ÖA0‚ýWÿcøguÈøŽS"r½ÈbU L³“6È32/ø¹ ¿rxŽ?M*ÉŸôÒH "ÉÑÕ¿–Žo2Š(ø‘ Çå"ÛˆÑ/{J¥l²>#NÜu²xSª!ßÅ 邃ë¨â–—­²µ1Nájíoòf°GÙ:Þ±SåÌûn˜±Õx *ÂÍ`£ªÇìEF.¬í›Þ/’é‡B.0ÝÉ®¢Ç¯K–Hˆ [\é),ÕêsX0 6¤}f‰¾E§‘÷Ù¶‡úr!|—¹B‹±šÉ¾ŽÈM®+ƒDÙŽOÓ³8|8᮵u ÊnBºÆî*%Ê~c“GA±Kµ§Ð¨,ËåÙùdƒ«‡1¤5ý©Ã‡U[ÿàÜZ…â¨îÇħGB²¿Ó¹òk²±ê"¼0̤Œ½w¡Ý‰GcîXä?Þ&eÍFe ìà,Ón ‚»Å_3õMÆ?§ª˜B2øo(vB‘JQ±x"˸$H—8Ó8F^{Úw<6dwLØÝž¤€—àiL`4` AU 0g0—†(gϵy8ý-‚^·}ªho3 W^-l¤³û*­—ž‚†à©¤ÄõÊ«‹úº}áüÛ*ˆó> y/D¬uéz Øû4ž¡?Œ/©x97¿£WÔ£’hê¨GŸú ÷¡Øß5É=$ua2-»¦µNl5(Ç6’KΈ–GL x«P8G=Ë1+và\à q`ÏÂyçè|$ð³±ó&üsõãÕcWb=¿Z=êâ Œ¥‡I¾Î¯i»éŠÀã$k¤v¼Ê±Á7Ãkº LBÝ&Ž'ÝG“Uô­`ï²5ôÐ)y"9®£ª_á€ämÐFQ7ýöjïfñ±0Kã*ùŽÙgÇpè&Ú°£h$# {ç2›¿èßv†¨ç°LOøÔh°ÇÚ×i9Ž :ç5ƒá~[6䊔ðbé"“›îy«'»ñÈ[_iž^¼ßŸà9hwÖ˜©mÓ8¶_…_dËîð5  ƒ¾:6,¿&ÃÉÛnx'‰Ø<¬“gÑD—#å[»®µ³b5 @¾ï‰ö có˜l2&g)çrC¿bIʼà¤,èm?Ãôĉâ¼ÿš!›ú!¯”W.Q?ßõ¡¤ž—2Ú‡'ͱŠ^ “q‘cô¦©‰EšÙf’Äç6Ά"Ç¿&É2plÐçÀy”kôFâì B—Øý\¿÷yË‹¿±)O¾0T.zòW)8Ù‘Š„ùÄüõâû\> ÕÙ²¥!ऽ²/2Êãù©M"ÊêU6DÌã¤hX•YÙÍ „Öè/\c6ò7ìyÝcl!©Mªë|ˆÀrݯ ®>bN`©¹ßO:u©>ypIŒî?¡§äÛÉ´ «Í ýsV0qÖ˜´E{–£}"{žãÝWq¯.QÑY¿i †æ(–l%ªë@ßäy›µ@{œiΙôÈ÷çéðpáa=¬Ê¨®+1­ 8i†³‡Î².¾Û2¾cš5Ìó®o oACææø´tW˜?‘;·q¶O6~4‹öèÃJ· ›ã¾R…Ü›Ðw(ª J€jõªaAÌY ]sˆŸ±–æñWTV¾õîب–H‰†¡šzèÔG£â™«Ëe×;âË\‹EëàÄVÍ_M|¡•è ž‰ðÇž&¿Ð ÑÏ —Á5n;û˜'I¸}ÉÊçŒiMŽªIøê¢T!¨£ãÊŸwHc]·+#t% ^çÛ*ŒkžÂ_ù{Á•ðH¤ìÉÁÒûíu©Êôä6P”-|ãùŸÎT'[ ÕŽöÂ1ÄüålÅL¡˜z¬BGXKXåö¢eoxùüÓ‹µ§0À>/t]Eðoõ®Ìêu7)*¬ž)Dü:4¡p8Ö[3'·æ‘;¬ž]D˜˜-j¡ýϰOÕë?–ÑTž­˜ÞÅrÏ!WŒ|¨êæô6\›T¸@!ÁíH×=í_ýš@áiñ» ÞQílb¯žâgKÊI³µ~à˜€Ë©?u°EŽ{õzõο· î)YÞ¦Ù$Á²é|¶wwÝgj褞¶87« (ŠêÊCø“5’¨;ðljÇ®KvBÕ0ÿÔnoî)Ü4úãl„ôBéJyàè÷æ¤Q> 'É›«t-º½O´¨¢„ÝÂõ~× ÞWìÛOžT\ð“O†¥½†_Œº]îy×3µsS“âWÑ_×\òÍj¨ã,á6^ŠAÿœ!ÑÕÌã ¤ÀãðÂØÑÛÈDz¦ÎDT±±3%<ø×¬Á| ç ¯6WCL–8e<Ï•o¡UÝ4wZÖæå*5wëï3-ìÕzFÏa~IxË{06žªƒ7­Râý %({¤ Áßü6¥Ët›DõO©7çY÷½,yK ½øaÁ7I›d&8BÕ¾»—·¨Jõˆ™Žâßaà:+š,ì› ÿHò¦Ë"Lèä¦=‚lˆÙäæ¥á—‚ºÁwFQñLÌȳñáõSÛý²Å_DJšÝ\•6g)2®c&ínWWo4 ÔkêÞÖŸÉû‚aÒ OJí«ý]* ³›K=ìôqéÏå»YŸ1–Ô¯3æm£Ñxó׬Øt\\ÇÈ‚Sr&zïo7=4U'²5RL½z^“mt¸ºß#‹Cæ]54¥BPê6X«üJæÃ ¥LÍ­S5Í8)¨©ý­[-tEóOB‘™¸aU"Ú¯,"Á“ã'ò¯rW…¹ÚùŒW ª„(Ð †oØ õ€H ,‚¶åŒ”2,V*Z}¦¾‰Š³ÒˆA­ír±bZu¥sÞXXÓ¯fä ýçD[8!;íÃ~ñ/‚n7ûù.¼*ó§‘”Oï®7ÞBÓÝü¥‰t]ª5÷\x$­"”±Î' ó °{ìI ©L£"(|½.âe1Ó'`’z áþæÈ69ùÑ>gùfI<œgcšÀ©J@ìO ¸~Ɔ¿‚x3o\³77ùÅzÈØËÆ¯™“à[Ü%?ö7fI{Ó.#}'Ëcõá‘ço·5vÇN4\V¶È9i׊B}š"AZQ×7µ±¢«Q¯7í#`?E*“åFeŘk&s37m*¾ßESý©ªD¼-O\Ȭ+ùöU@üÌíz„ã„íŒ0§M¶¾ }q3ª·1à)o;‰ZG?^Ù˜‡7w°7v©+Geˆ®×Ϥ_T¾éÉPæX’ð¥EñâÒ† aR3¹{oÓò xUê–R‚ µäÂkbK5Ì/®³—¢ÀS)¤ë”‰4P ±®äϤw|³h®û©Ä¹7Ô.ݰ-ÓôâûF~† ¢±¡Ùê£ø{  ÆÉ ¡\xîÝ›RµÊ³¨‹5ÿ^ þóô©Š`6”…?iИ#½¸@ :¯J²b¨¿ÀÔ3‚·,ôÕ·MãèàÝÊx«àP•¬ƒXêâ<J«þå š!g¼’y¬ÙÓ&Å\ W­Çê1JòŽ,Ârùà|Â\µä[´–K[‚HÜ PgâŠ5:IOë%ž±1†× *ØNC#‚£ß‰_P‰ÓZp¥—ƒ?C¨o g€ú+«Úá í½þ;SÆûÒbØË©jkXËÚÇóe«ïZªì´yíDztbHû¾g‰Иgc÷X–¦C²=ùˆŽ#ÝÍК {.Ò÷.ò¼"–=Ý]ðKv`»i‡ñu6}|ôS_©²¹õÉ«éð+©®“|…£- '¿¯ÊÐó@ÓùŒÙCü5, ÕP £N ÷r5ú‹Üb ¾”nÙ`Ù4µHÕ†aÛ†¥c+÷H×ìë*S¨e¦Lã3í 9§ïÛÏØ@”b¼Ö#ÖL`!A€60_EKµcYÞšERv½,ƒE/0²©ÕÐ ëV­!Œ$CQ)k±ÇÐÞ 3¨lÍ:ŒX?2üËྶ`¾ÊÇGýÎ0Š`Ñ ºŒ(tC¶še糆"½ŸÈ~HR…N(/ÑgI˜ a|Ü]ý7º)Ú—*´DzÉÖê «³$@‡8’y •ÔvÏú£‘‚Å™T‡®.+ §îWð¸F°bÒÍjöîû‰§«~2<Ïd€“»éñ4Áµ¸¨ÛÉ G<Ëx«øQ\/®Á+´ îÅ=œðßß~YEäÞRUcÇäMîU» $:Ô)Wš!½ g{Å¡LƒlÿÓ4Ò\ôwuCØà~‚—y'æpì^ÂË¿#ÑÐis`~:U´^9À_¹Hæ½Î¶$ÏoH,XF¬DJ±8*ª-a<û,ê³¾[êÁoòLq:„~(oíãÄ…aÿO˜ÙÌã<Ö±ês–ÅÑë”(óå+XkÒä²duŒWƒ+øðD…Ëq^¾ØÚÞÚ@¡†µjα=WÜX9Ôi )(‘ZS0á ÇáÜ!ø÷ó–öëíUÞXvÆfÃzš‡âÕÄÃ>ÌP£22`C¢ƒäЬŒh×ßà+Mzx¸ÏH¹„óÀ¢øô|ÚµÉûЂž‚‡χƒ‡”¼Ó+©âþÑ@¿çïÿ´ÒkD<8÷ÌÃh/ÈË)+w¤»íÕ‡®¾;hi¯é$Ê+oöáp¥¬¶]Õ¸+Ü -ÜÖ/ôã6V{ÜÈ=etng1ž¶4VX¶³²³çn¹J®Ä%p——?þ°¾Ò»á:b3‚pªøÑôŒÙ˜–kä ]r â£boÁR‹ð¨ô ¸w<ÒivBGól4ošÕ‡Ù…¦ÙQ ¡€Dóqéî“ÊIA¾k™Üu{ÌÈiÅÂô¥Âäì¬gësò˜|ù%»Ì« lv‰–ˆ»ÂN¢Ò˜3þaeǺ¤¦Ìþ³/I›Õª‰.Vñ‰iŠqïA;Ò›õ;J¢B†ùPY“þiç#÷'ÝÐí#PhÉ»mHHÖ˜Û»{®o—ŠÀ ì€R”³½ÛQJnê™<8Õò±íÖ‡À’YÊÆ¦—²ît»[js }W ŒŸèLR½ˆÆ´‘ŽoA±ò¬ ;GË/"{±Kãî•–6,¸2D*ƒ­m…~ÛÆ½}Ÿ'O6^àf{GÁáƒáÙ¢ “Í#zØýcç&wùªÆ‡jTFZU£Ó˜äXêúÄLÌÆ˜flž§-!óéIFX°ü÷çÈàà”|eÄXàÉš÷m²ktßÊi&©{—6ÏÝo:¡í#ÒRüç ’bÉ<~¦j9ü7ªê5^À×ÐÄ ¥äYÊ ðcÊ>Oôl²GápÕ:hÉ`jSþ35gR(kØy!uç4Žxœ`¤ŸÔT’Û·¾Ë¼e;ñªMBdŽÙB¼‡.# O#’®;N£Ò1vªèuŽÐäìÃEC·g6Cúœÿp²˜bÍ70×? Ñ„çõSxÙ8KQ âÛÆzJ˜”Úé¦sÕt`dðK—O5ùz¹ÉxZ™ZZÂG¾åÌweI®¥@,‡xA–oëåö÷Wz®zö3(S^öÈaÄ Z <"€®¥µì€ž öøI»emyfWr¤Ó" Qí{ °Ê:£Y‰,ºPÔƒô˜‰j.cE ˆudcžÚèÃAý¸Ç¦»ÎqWÜÿ9f¥÷š(A­KkÀ»ÉG2Cð˜E³ã–ÅSßßrŠnWs|\˜ëJî´]µ¨Ø ÅÖ§è,~#½*Îäwçk#À õúØsð;ÛF¬?’Æ ¬8Ãÿí®‹Ëé’ ™Ö˜Ž¤Š{Òþîpûìâ-á‰Ón{Q­žmtÀŠ}>‰ÂˆKÁ¢=qç …·(þõ9ÿQX¨5 ß 4?YHyFMY/@†Ê`ŸÌn1á¾²qäÐØß0ÿ,ßù㜳òß?ËJäñ¶«ÒÞ‹é²%0<"ê3Ûè[ÞäPm¬ým£ØEÊ³ÏÆÜ˾ø…ÒIX³ÔÑÑÁõI™k3ÊQ(âÃ3Ä­wAœÀVù ×컳‡ †ËŽÏ»ïÅ…{n,wc*×̶<¹*4æc¦À€ ß vèJ鲡ÿ^ÙUñ£“IÃc‘¥´Xøµ³)yÞæV˜dz§³‰O¿X¼öl.D'ÁZ;ÔÑŒãðXsÒà_/Š‘¤Uc‡?Ks#ȸR9äÔ»j©Æ—Æ`X'®ÒZM¿éú›þ2ÕúZi4ýÙþpºùÍEÞ&äž ÇÌþ`ï¡ç›G‘óh£ ƒ±ù­µfd·,Xó y5Fò.í±+« ¸u=Iëãé#¾O .4žä£ø×yO”fìŒ Ñ¹‚{¡$ à 4½‡`à'ÆG-ÙwЃ¥e=êLt>ëë·;ø¼½ bôÁ‡h„½çg»$Áw*>ö›À«'{0L0Azãû"ÂeUüúfaFýìBÐoìAl¡A=º…ã tF/ÓÅàƒ_}ŒkàpEÃT÷m…YE |Àò·Ï(½rw1 þrÜf¾!§ô ;¥,ŽT^Æ"ìvÙþ˵+°zGHÜîèÉrÌ=—>ödDxÔ¿øð2vVâ ¥Šß5G¶÷•½—]Ÿ}ËXÊ„Á&Hþªn¼m?d˜Î$Œ>ËV,’ÈÉýYýÅbgBžü°©ÏC6—%Q›…X‚xåýbGÅS¬Ov<‚øÈý´·‹™|ÄÚa¾zPà&2MÊ>¡ÍTËÔ|>õwh'ö kôêÏOð9µ º48r´ðéÜùNÙãåà Ev*˜ÎêhªVWÑ´;$z“‘‹ëVsZTx¾Œ±òñJ{«^’qÔpæ±;"¶ ‘’g»ocáá!ã'ŒóJûYÜsœ•uTBcÚø¤ãçÌÑäu++¹oK#ä3-ëvÁv&Ù™ŽÆÀm²WŽÀgÒ˜j5gùïìÒù¢íƒ-Œ#Šq°¶ZS¦TÖ9l‰ÔJBgx^ÚsÄZ&\,³ûŸ0<÷)^} @ZÞHtŸõ·³ÑË»r*CçHñçúw¬ËÀ[ÓHÕÍPâŒt[¤A¢y½‰çê0ÄŠ+|1Åáb(ÁßžôÜ,n[-K^|ÛÅ«• /33‘Çi®è¿ôscÞ O–þmÓO!ÍdÇg_få¥5 ú·Îï8ÔÑñ3EË×ù!ÇRgg2Þ0´ÝáV¼•0«@¹C Úd+ÁÇè¹\g÷UþîÀù(Ñÿ*¸=ø`åûiNÊÚ?×bÓÅOR8óNö¹ÙX½ˆ«3µÈfþüÇShÒ›œå—„rØd"×Ú6@XËCˆ& ž:'œËÊ,†XÃüÀÎÞsû]þ5¢{¡X4 ”ßê| Ú®ê¥çKÈeœÊZy+÷Ô VËËò콿ïÍþD®6\^±Ïpë½®!áX‹æLã"²æNwâË ‡˜GHö±Â…»Þ-ÎÞ‹øm)n do¤Ùv:‚1©(×võ"°õ¡CØo >ŠôÑILtcÏ=öDl¦ct°”ʵ§` °dóA+VÞ˜ªI€8Ý¥˜ïaþÍ÷¼¤­Þÿ¡¨ó;{ËíQÞN<#Á P[¼V8å¦Ó«AÝУáëyÝH: ÍED½¬Æ\Hó (ž›ïÜ3–Œ²Q[|ÿåVgü^»Çþê‘B´¤atwÄu{f L–C•2~¯2s((whU«ÛÒe7ﲘ-­­ÅM§¥*ZqëVn}hõääôÉf*Ùò¼ðøƒ:©{öØ›jMÓO‘ ¹aÛ)§;F W.Xn-*»G™=¢ÚØ¢'ùCÀð¨]»˜Ý,M‰º,KH£àÛþWߨ±õ¿É…ª°1€¢à¤)”²Ö#Å R¶Ib4†âÛ=BµV>Ýz'³ì}LáÀAmŸ™J†à„¡i€ÿSªø•¨‹±=B­€ü—·?÷ï+v!* w4 cNÄê^¥¦¦pJªX±Í¹õ½Ô=_b Q14~ÆðÒº_ÚI€{&òQÑ”e vV{}5îAÙIÏ6´Ú^éÆx£á£'>+D$œE„%ÞkðüÖA+q@k‚QÃ5x™À׃ Ih+äm‘Úä\Òrâò•e%€jþt(oþ0k=9ÚûDAïxD<åÁmDÞØa¸]üÊl¾Ó¸ïKcbà€@ùZj¤ð¤Å„@ÿ EÅgI[6óÙ']à,#²à8Lr7†>5†ä»Þ¨¾dU:%?«z£Q¥‡yÎö¼Oÿ·¸ F:f é&[¶5Ó´ˆ¼Y³'7¬‚bvb¬‡Žþ¨âîí !½-çÕD!ÕÛíš -U 6$àŒ^ç5HݰBð¼”l3Y«Âðƒ€~ï}ïüÍfܾ ·T’U¥Üé4=ße™Aç×Q‚/¬Ý}3‹hîšü} QûÜ,M±?IGûšíÅë½Ìâ{AЫôe‚”3ßwâ[ø ‡½:šÁL±5ü:Ä[®À¸™Öàõ~ ÙÛYñ¿|l5†;íVçpõýµ‡]!VW"äHŒz®}d¶bRìÉT€“›“ÊÞ]}=ÕhÕ+ɽ#Åiý€cº†å›—=JèÏÞ—D@÷°³¨NèþêG•ÕýL·¼§ÂüMr,=ÝÑ蘥ñ¼ Uõ߬Ò‡3 ¨ IÚˆ…•[ðx¯ú+æùžPø[¨Du°ÐçbÄ.í·+%¶!ˆÄîÍB^냒ÿb}‚Ý€-Y I_uo×ÌXîÞ[¥I¿¿‰ÒÌËÝö%nºµÛ­ÜäùHSD§†Y£HnKìjvi±–ú`¨ž[%Ëúñ6²-N¸gÆáYWÏf×FF—|;µA¨Þíç¿n¬B4BiØØW¦Þ¬UcÈÅ‹$àå¡’š ß´›3¦_}[NŽë‘ëÅ#Çz¡Æùƒt󬦵Å]pœ¸j{ôÜŽ!ÓàY­Åzõ+“WˆíR&âÜF^|¬K×x mdøl ^úZ^y“r|™¶¤yu@F¯ï`?ÅúMóv8´þNx/èšÍxb6ÑS€ç"h¶°Áî?EýC.Ðkú¾’Á¶'èbÔ|#,ð‹bJA¸'{×FIo¬‰Áóï§È€ XÊK–Í_üaÃnUÈØ·SÓ8Ô"î†ë%Õ«9ºeè>ÏT’Š6©PËQœ®*cVæ¾;ÂÚœZÝ“ï»\+ù>C’@>}ã“ÈH×2j•–vÛ¬¶´þ«‰ O;¸àWôª%Øc%5Wg¨ŽŸèøVÕ¨÷0úPìðdP.òÖWw`.§.ÌþJù Ðì&‚‚KÌÐVêù¦Põ øœ¹MlSéí@èÓfœ˜À\›2N5Šf–s åÎv jjð‚E³>$ý¡Ó-„á"†_+Stȶ|Í ƒ ¾J#⃶Ž{ XÑ]@"à (¥»8I ä†7f•'ŸJ0¡"f]ÒMFŒ Ÿe:²ì`qõY‘nʦÃyWþ²_ƒçÆ©‡®Oz=Z}—ˆD-íçe *·lX+4„+ k_A Ð :õèý> ʉÏt«¤'âÀÙcÃÃ- Ÿ9–"Yµ¦aJ†cd.KN¦÷“Âüñ‰f®g-h^Z ¼×¿MƒoiÝÈgèåé’í÷¥,hŽÿ¦þ_›t½Ú—ÀA\Ë;®XÞ¶ì ¥Õ¤F=A%UVM\^Ô–làŠ›0¶n9¶<Àfháh–´ý†ÝõÕA#}(jÉ$ÐϰÅÇ‚™þjåæˆð㜓6ô wU×ë=‡=E*@»þ Il}¬ye#=ñ•HË©ž¦€‡øv8Ëþ­úÉ(ÿ½b¯\¢ZÉ"ㄯWŒ<¿>¢çÅ4p0üÆYU2/‘æÀ™ÜÁ"æÃç=…!F‘‡‘Ñø9Æ%sƒã¡çÒg]0Nþ¶Í¸0l7¿ëì”Åk(Ðgé Й @j ¢'e¤:¹Ê?Ö9º¯[=‹¸¤…ZNžVȶSy#oÝ>Ÿž9..(ó´z®iBÒ÷w‰ärPË©h  vOO&yŠiúm6£×jq –eÛWQÃDï ¶|- CUx5šžD)xr¦1Ïž>¨¢¡é-š5WɲN)”±ßÚ_—ƒáªZÒ8õÕNRÚêiÁ"sgXê )ÎÕ7iÛS¾ÅrPÂÀ+oiW.ó'‚&æߊ֓©÷iÒ(î K¥-# ¼ˆ~<‹*Sdx3©I5Þ/·|7¬d3r,P?"ßlU?“öf0I6 ‘KD}ѩڜÎöhm5+GŽUÿY§`«³¾ªë͹y;*õ æh¤²a–æ“ÈÚ´S=Æ—ÉèĬ9®b2«ÕYM\ˆM;!àÁɕ䨯*U¤MäíÒs3^@MpBeú?·3@ÞE€CÅö!Ù|ƒØîÆ[ËxcÃÈdÕ*A½ŽvÀŸ¼²ÚpÞ©cÛګˬXõy/’û ¦ ²œZÅ×ík¬Òp)ê“Ýn—5¶;, 4'u¦‚« `Ûµß$ÛÈŸEtþ?$;ã=%LwË÷ÊæÉ6á d#jŽ{63Z3‚‹M¹nsG¥)dkGŠÊòTÁA/¾*y´«©ëò•— ëA x·5íhkGÁW¤çÀܪLnêµ} ¹êÖºh[X>nwKðÏÇð±´q†´D“ªAvŸPÜ-UlBßÉ]_Œ`ï{Î8VÝúá<9¯ùŠÑò@³’ÄCqÙŠ[`†¼ê MÑ)k†?}EM÷)¿Ši&ëÉç£hí}d³ªm¼zpJ,ÿúøòEq•mµñŒ²µ>ÌÈk%Ç­X‚Õ ™%+éPo`~èù©ªªÄ6Ñ:NPd'Ž=DwfÆæÔjUk©tûò6Ý6¬ð¼’_*MЈÐë:ë 1¹‘úHêfÜ€T@ÚÃ&I_]{¢Ð¶95B KÈWEùÅ höæTz÷â±Õ8W: ¶ÒB9-‹Ïu}×1ö8‚”ç¨AKé¿[•vݲØNiÜGáÀ½BÕ¿6PáðYðeCt‡5^GZ¨^Œú1ýF¶(üÆí‡Áµäã.Í@·Ÿ6î˜j÷8Øa?,JµâQÅ+ÍvÄßâPÒ Ðf%~"ªSÀ…ùÖ<`Žå§E<aœ ‹0ãÚç…g “Bè<ùCÂé£c^m!­P›¯T÷=ùíÜi&”Ö5ñЍtpO¥7‰jùIX\'~™ýþ†ýƒ€VœîRA˜ á&¼cÙðˆ+ê ¿uUG¶6–½GØDëñ>Gmóíþ½ &QëøC ‰=ì–=õ·.^D)çÐ3º6Ã<(ãÖT§¹Ë ˆ»ïs«ÿ¥B5£¥k!š^â®;îN‡+Y@¨¨;[C†ãàt¹ {FÂðŠš^amá•V…ïþ #Õ‰ÕÚ죩U-Í½ÔÆr‘ PŒæÍ(D/9ÆzB ɘº=ÍŠdZãß󀫗]ÝÌáÅè>^MÉ6½Á(¬vÆkzÚ^Vgšˆ?£x¦ÌN*ÐÂb±Ê¢®"µà˜ÑÝDIv€U]Oú)Œ7ßñDQGšýÑ5}ݺçÎ뤙ªý%UD€jp²åhD€ò€–e–™Å¾ø½±mm8í2›qÖ\>íãzê÷c7ÉÊ£ûð¸ç‡’›Æ®Ð:R5Iäç?SBÌ–#£LìvÆ8 ; 2áîÿÒ(Lî+1P«Cí éÁé)õõ÷ Ö2“‚ï pÈùC0Ns››Sïø•Ÿ)£à-)í㨭y*7ðF^ãß.¼P(5UÛ««OiÐîÏ¡Û âì‰ÎÚa.LmJ_Ɉ'Wc|ìšÑ€Í^Éyø>ˆyÄÞä}‹îCZY]˜«%Kù3°{“¦åÅY & =QaI\Þå³J®ñgýk[ÒŠýÉn.‚Ý÷2ÍŸÜÛp3ØÜ¹« g!¹¼J qÔïAÒ&4x¬ž¾nçÄgPÃzá5ð©ÁÀkžñ§N†ãʉôO,Lb½Yi<µåN^z¹ªšÑ¨TÍ•kVä*ºÞ@ZŒYE¸ª¹âv®ôŒ³p m“Ÿßþj›8úmŽÜê‘‹w»}ò¼.ì%óa‰æ=¼=>ò¾¹$T®ò;ÞZ (¤ÍÜ­eòx`1¦¾bô{[º CŽMÓˆ qòá‚ 䃠¥§Óíu7þŸ›wÊë+0¼L£—¯zŽoeúüƒÇ–“½ñ69 Ô™Û3b FA Pj—HÅ|} LïÄV9ë„]'1U ƒ—­ 8ru!Ï=¥)žƒl ¥U7ƒmRñWƒž+nýÀ)àÒº ŠZe>´U ¡Y¾ªܱuM¬…ÌßSr•*»¸×Rçr!š?<ÙÅ”Žâá…Û' æ\p£o¤Ê+¼˜Â{֩ȈJ9±è{°K>£LE‚çðÃQ·dëÁPsBm`dð×®šé+VŽpW0õyF·rß@Ѭ†­€X³ªQF“!SgcAŠTæQj&¤Æòmç±i­¥ qÑÖ/s9¡ÕÄëîÁBCŸØ`gn mýpÉ(ÿåªÙFÌN%;Ü€«Ù.ëjÞÔæˆhé*Ûê|(~ÄïECß¾1ϹlW.»pÑq i[A&{ßïºaõÔÕ¾’ªs/ì5 {­î‘¯5´)…ü/WK q¹  ðP™|áªGÕN½æÌ3œTN°¢+RÏqw^OÈÞ€Šo0¾ KJÇN~}]àDY·äÌ«*³Çs¥Mú /¾íæõ–™ÃÁ“â2*Ë'¤¨[SžJ\¹Nvà?9Žóõ)ÐÄ‹ûm2Ñå 46Œœ³x³Bd“b}â &Ñ5ù´Eò|t”B2õ.ÀMØQâ |/E›!¶ÂÊŽd5Ð[+äÜÒÑöt:&µ`oðÏ5ļhô¥a–ªÛQmªS£©ñÇè/y…FÅ'ø¹X†äÚ·¥49B¨ýD†øi’«‹sf0- RpúZ ä o Œ€ì9~¸s\gè(ÂvÞ®ñÖ¹U5iqo®[À´K§ó™[Ó± hûŒk6ÎÇäê“…‹‰\ «ã¸ü‚åâFº(ïlåJé_ëï',ºçZí+‘1–f4ŠÐû&@­Ú’_™àøYñ–Q=£ßÒó'3gº½8ç:’ $ŸW™ÈÔ(®aù‘æºX¿ãŠ/±VÒ„²‘Žl*q1W¥ÿÝÆb>”ï=çœ%èñZæ´Ûe±ÈDÅàmNÙ—‡ø8¾u“XJu²a9öÒ{ QôO†àÌŸDпË+AhÜ)&ÜGÿ†õÒ¼ÖÅzÉ5èPÐÀÖeߦ›æawöw)™ HAZ…ô¬-ñêg]y×v\ÖÞ[´ÎþK§µ2¦¤fA$ M.v7vÛ šà› ´,û¥é ýw¥|§À!›ù,>å7)ò˵ÄÉÜã‘´êz/ƒ]øœ;Õ8Ãɉ¬þrž®âGGAõ Ìû[d¯˜r6‚Î1è-ëbmªwµäiË{>bYOl[–ŽÜhÆAŸ à±?!ÈýÖ#Ã0-'LߌLÁM|çЪ$|$vQÆã,c .DZÑ\åÄ?TÁd¥¨ßÈ¿eŽGØRj-ø¾¶¤8I+ËMOŸµìÜt Ö`6`u}wшòCüü^¿ò·£ ¨fg[ˆç®òW ?sËyÅß,‡6x­]›,h¡„ãR pB’Öls[沞Úþu •;vtUvÛ\ÿŸÆženØÿ¦FjV {KHÿ¤|z¨AÓ’–—å§”®ï,dybŽ=òÏÌ]›íüÅæšîé«J⪗ò r¦jÖekµugŽrQAØ-6Óˆ ~©žÓGé °«ëØj”4Ç÷ .˜UèåÂt}¨>ûžTIÅ'[ݺ²ä¢&¾ÖjU¸‘Á‹MûHW™;DüwèÕOò>àTD±$Ü“ VÌ©c;Wy©[}ìô­ðx‹³‹hj~´IͺڥMJuž4x¡Ø ‹ÞubˆÄŽ}™Ò1èßžÓµNÞ¢‰" ¯;”èA²cÇ$ÅœuBrT9Aá²SQFWÁv¨:2 ÜJù;]àsÑJzè§øA å;LÎ.gfOÐ[æ¸X¦òC;4Cá’ øaUX«žïCˆÎ?°ÔjïT¡äUiG„ªýhzVÙ ”Yd²øˆ½F«#LÆ—^L‘cÁ@ Öy­_åjÚ]´ïêÅcŒ:ak~=7 7ŠP°om%#ú³˜^D~ìs·1ðR½_¡ «óf+á§ á(÷uÚO šR`Q›#0G+Ú1ªº5£²lC€ä(×±€´¦”I )U\ŠTr••»Ù ­4Ú/ú.¬fq¢{xÀR!€/È ¯¦eFÎ@ñZ;áAsúƒn‚ukÑdm.¿öj?ê#d´ÅDOì7`³CŒØÁCó„é2Ó[½ß ïé©£v·w|-¦¤Œ(™†­Czà;S _ª¾6æšk®‰ãšDT¾Á=õu),…¦žÙ;VÖª3zKôö»–ߣº¾ŒWuŽNŒÀ:ÿ!*-6ºªs¼»T¶HLˆˆµ&)^Ô@AÍŸÍô^ç–ѯE‡;r'à ž…@y!µ­ôUqþPïñÒÞÆv®‡±4v.¬±ÆEφ#ý] ÜÏ`Û M~b†KP†HëƒÌ %%ðŒÒ;çÊÛOZÂõ~HÇg) 8ÖñìÖõF3窛yb?¯Ö;Nçã`Àå±¹hÀ-ʽ¿úº¤;bXîÇce5¶»‚!Z/†¡´ÐYÛa;uJ¥7 ÖÄjlËC¨ƒ{Åñ £®ZÆô)gGoóÜ Ù!BŠ&¶û£Z›"™JKáºöIºéƒ™AJ [¼šµJ+¢ìëvbù´ºlŠ£.ûö]jG¥"ãùB<¸übö6ÙnhŸÿñÒÙæâ\G Cõ×'•Ñųå! ä6˜#£m C;oSbîõ¤ž7ÚÅ¡·uЀ+sY|œÈ?H.”¿’”ÍwûxcË]Ãé)Ânéš«Žæi®ZOC|g•Jƒ»AåùM°†“L‹üŒØâ,´o/ü12ˆJÊóªŠ{AŽ”„ÕÔ~5·ˆˆ235,h,häù^›Ê‰ËÚ ¡gbô &…U×w­Õ~žek“Þ[!“s?°ÓÚ` =Pá8•›'B‡«uägk’Ó•FMW¶üm÷jÊ7ANZ\}[ []IÛ<ùøGXÖ`P¼ýR¦åB«1çÑ/AÖ…¯öÀ\ŸõÛb†›)ãëåɳ~è“­MÊ®þ®·ÈÔÝ>›,öÔª51ótd•*è‰á^ôЉ»ue;Z窡ê®I‚ õû›á@ž=YÛg)à>˜‘m•Jî}±F§˜²©Ì1µÄHß¿ïß‹¤ÕÑÙ–Àì·’iÀOJFÝ®íþˆI*ªÈÒRÌå‡L°7IV Α(q«‹j”d é€a°DÂÞ¾-‰5Uäõ8*ÆíhïüåMîYK|‰Šeæ¨rüÏêrpÒåYºñN:´2À(›á|¥ DdÇåZ'l`çÆwdæ¯•š—œ§Þ«- ‘΃‚þÆùöRÁ È0\<)àM;î’НöñÓÌxÕÀÔ\vô•̳_=ceÎS‚¯©€PwKêÝu‰É”„ú{ª¿m}(µ²"mÖvOoÐmsR–Á”hsÛ@\šñÔTïyLG¸ÑR ýÁ×! å$«ÉõÚ—›4Peù£ƒ¸ÜyÎûCH*ÜB(jŵj•º§‘uÓÆ ³k\¦þÂE^Y™ŒgÖÐÊÆ dè 1dÙÿ’2|ßÖ£*F××8zuN"4]ô(ý/úÌwlŠÁü‡DDÛƒ`ùs‚CV¬™Kš±º^7Œöˆf®,,¦k_ ö ¹“üêaå9'âLäks‰ã%Lû´˜ª RþâÎÎ-Wî‘b®(3íc-BBR$æv¦œæwgRI÷gì´„/ÞÖfßo pB˜³#Q_1€¹Õƒh@Sä»}?Q;íÜ<µ+I¦òi¨&"ÂÅkÅÔ50ñÆî,/­vƒ?·„#膘Óò™IŒ`íN ×÷ ­oÂ:¤AxK Q £#S<åÎÙ(¡í¯Á¼w`ï%ÀœK[ÀþžÔYyr®sj¹/ä«Â–'U.Öž’xÝZ<Í6wbG§%"­)ZÓ ÀÀ‡º<_ØÝP;s2ö 8Êæ(YÄ^Úh^gcXæ–K­mNÛ¢š+2Qòˆq‰Í¼'bÌNʤÈ"¾Ö=ÉÿÆC9ÓÅ·û®ä,2Õ[F΢ÔÄ)^Áu,Q·Üe1§ûg,Õ®ÏÐV„ù(&EOþKë$Ÿ2ÚeXMµWÒÉ£ (*¾ßW‘}íçÍÊdžn‘«ܸZoX|“8üטVÐí+Ô¢¿I¾ó|DûèûŠ-ãIï_Ä g Žžå›“&`ƒªIR6Ø!UÏ1u(óÞgZ™ð·€Å"çFÞ“Ø"×òï”Æ´}õ>$ÄQù€,Z¦ °é+k¸v{úþ’Áã,±w¸}/|—ëéCžòõ!VÿÊ8D’±5i>ð"•6óò…½Ú¥…èÌÂõ0L!0ß8wA\Z@Ï§ì¡ J£DFô¦{ó!±‹¦š©3yØXk‰- Ú˼>hðò¶"xŪ”0R­åî-ŽîPPÖ‹`k䨚·TëpÐZPxxeÐTlžÞž%Oß«<úšØx†˹ãjÛ­‘·ÎDìeÓàÖÞ6 ½$è ¢üXP&fâ±5,“’YݳŽ(bÍ ’g¯K–ŸDF±;íÊoæ½×ÐÕa [Ð ë¹ß’*°[¸Œr—e‘RÊFÁülŒ vwÞ"ܤz©¶cñÀêÍ¥„!ö»Ü†Âº!¸h–¬Îð+* 6SU†ÆPÂ_Â¥²æ„Ç\un1 %!:aÓ®óW(ë= Kùaþ'’vG¤JüÇJO¤1ÊôŸö¯‘@+•‘YäV¼k¨óµÿâIÙâ·–8ÐÁÎ úÖ(¬ÈŸ÷ñ¨Æ/ücð£Þú4ŒîAÖÌ¥BzC}9 ø™•Ž{c ŽkòögÀ-À(~ÚÑè+sF/ŽCT=e×/éó=ãÃM³Û”:~¼w^¤ùúC e–¬î\­ ³bÃYÇ[uü¿êN«ªWÉÔYþ}GdX{7zí}&‘^xJ°¡0ð_°ßýÆ…X0œ¥Èd¬«Ô¤#LÙấŸÖnh¦ƒû”ÑÑΪ;Uß<Áõr0–øÊ¾\L mÅçÐ ›\ñ]oøÅA´]WŠIèíG¿ÓÐ `IÈ׬–< mI…ëÇ8T‘Ú¬m?ÕÑ(îaÇŸphEž0g·K"òBkÓ,¬ \TÚ³šô%‚L“ò¿bÕ9‚OL£äQ¸f†‡Ag·ÇÍ?w¸ª÷ †Ú‘@ÖÆáÈ!˜Û‘å|à²Í^I2ÈëQa 2õøµ3^?Ú/'ÄD™(HQˆÎvËXhq)ä—Ù,ý´S¶•–Z*H¯/D—”Ÿ ü:ê¼ „ò{ M¯Ýz9ɯÌGÂt&ù^k©‰n6Uò J¥ƒ‰:nuÊT¯€àaŠRžð2ÑÖ?´s/Ð)¥ˆÚÇ€$ªŸ¦žµ hÇjX3¡5¨7ókÑ3†´l? 8TwGZxÓ;¡T¹¤ÛòîÁýûÛ.ÐR¹«¬œf­.ë V¼ n‘ÑïBhº£Š68DåÉá¹KÝÃD] wSÌý­v¼»£Œ¯àÅÇj$Oy0à5„ãö!23›téúw÷–‡¼ŒËÆté©]I§2—ë:C ¥ˆ14¾~Þ°è R\6T†É&˜™2qyN6X|¶Z Å©í}FÉ}æVˆ?E ÛŠ Ÿ- gñBkÈ+&¸Û¹ÃýÓ|pš .¦‹„ÌZ>‚¿l¨z‰YÐ"­9¬>Ã6, Ö]d‘!xBòè|ˆO÷§Ï6(ìü˜n…æEçS¹Q Pzú]+ B´°—òºRµÜþX™ÚŒ[âöË Žk2s £ú:øòšÚåî¥ë¶¡ÚßêOMÞUcí¾XtʇuI•>ÞŸ•ASª@gO=Ûÿ—‘4¢8©)&Hu³B@xõ=owesÐ cǶÓH,„"Ç@zØëj;Öí㡘ïgöÖ©–BòÊ!šÓZÂ'È03´!1Q߯èÆaŽ%Ê^ ¢6×QhZ éažS/J( ´/Yx9h KbNG4ŒU‘Vçü´ a¯SUC»ËWgß ßlGø×‘ض¼iAwœ½²m«sƒÔ×µ9J™åÑ€ÖMnÁOž2SPÏ©W8ç"µnè&”·¸bò“«fÍ ´bÔ>£òÓ¼=Ø8 XgfÕ?) :Ä>ïÀFÛ;æAþÇß½Å!Ìû´­ÀF¤ðuöEåáù\åÜÍc³êï,ÐŽÞqæA¯þË™öü%9T¥¡WQÝF4îû†³nW8 ˆ¹+ ¡YÀCiÇ©æÒÌ®ƒqeŽM$v"+‰ÕÌJ –Œ)èÁØß 0òÜ.¶ä›Y°Èï½lËv¿¾“¥IÅYß÷AjUÚ€UE(Ò¬»u'Ý„Æy½¿êôXéä¡÷ú»åá.¼t¦þo~åã¶?´°øÆ‚NÊÅœç¥F:û+›¤v|šæ;§8ïÀ"2½ßšâ_j*uÐÆ(r:°_2àÙƒR…ŠY˜—RO&øA›ˆ Z‘¥NVn3[ÇÌÂÂÁOL‡#ØŽÚ ’¦wúµä›ÐN$º:VÛ''fÓ¶+(žÂÚßN3°Ær™H³û Õ’Ð\OÇ›ßt“”¼BãdGâ¼ùm´*ãIf+Ð,Ï 3ŠoyUV`Çì†ßdÎteÐrûÙiå„´&2E;GBl=mkfÜjd0Qû©3„ˆÝ€JÒfWn‚„ˆÅ¶"¾c2+Û «£B!› y\$ZóÄR‘j¡CÚzrYdØJHè(É$°¯ n0 {F(7†ÉkæýËÔ´·Dó36SÏÌî±b-Äcõa¸—Ò²©…O DÏdkq:|ãdË5ü©íßÏN“„;~ÃÅÔ ˜Î/.„§\{(*Öÿp} û)dÙ¦@U…;²t¦Ë€UWÛ6ž%o‚w§+}krï6³·íKíPxµºxËýþ‡·•1\­¹®´ö$Ú黥éèöñ©ê ‚K;´Kç5_‚""i*z×Cvd>{¶wÒZ¤t¦›þ|TœÁ/•"¤*óöɳiG­³V³ÈÊ6kæ½q/•¿µ8ôf_Þ\_"1èprÅàÖT™àاV–ËJWYŒÒÖM)Ø39Å›Ãá~lÀ¸BN%€¬¡Ox&dÔúùÎ î›”þšªc7å±Áj¶;ÝØ¹òX_¦>«¨­”3º–<5§]ÖצœuÇìÙ:311ÂþµÍTÞpÚå5eýq÷že´ ³"{«®^ØxbûîF•äãœÎ‹j¥6ï -üñPÀ2Kc÷|w5aQÖ.‡ƒldå‹ú•vë*ñ+å¡%›’\ò¬÷ƒ1÷‰-–«î’ £¨AP+8S|Üyè >Ýšh~UHq"éÒIATeC„¦?ó™1j–Êìö((Áí% ŽìStÆÍo×bƒ æÂ #„Œ\hž •3Í™[(žÓr¨:%c0î8-_i}£›¶N€þ,zyåÙyˆ,¨n•`Ó‡µ)Ù0øõ[Hôp"Ò©Jf}y>½¬ZõOPô˜yZñ­q†˜5³Ur¤†ÿ©µ°@¼»#>u})¨ÚñÁ²¸ÁÂMGÛl—“!ÂnIÓÐL÷Õcd1md–ì: 9l×F'Ͱ}oQŸÚE:2_ýûú%‘Nuò{¸å {gÏu|u¬z禡qñ…Þ£ƒË-`8Œ šny„ޢD³›BÃ8ùÿ…óˆ{¾ ­BS.N#o²Gè¥s6Ÿ^ü°覾)›‹£n\2+)ZùÚîCF”=¢ý0&€'‰ØòÙ†#q¢º@Á™{©çMœÅ4îÿÈŒ8/ﺹm³ü™I_Ê´ÕÞc;7æ E æÖ‚E›ÅZŒ¹:ÈÀćÖGÑx‰z½´‹Òð­õ^것ÑCÅ¡ÛIÒ¾0xÏÉ$“`ÎØÚæ—ÉÎïJÈýƒ‚;ÃaÝ¡ß}½ÔFÉ^)ÅG›L¼óäþÐ>ÚT#b:%Æ r#KÃåVü¨yº–V:ë¥(ý=pNfl-iHåϼÀût££Í•-Á[¥ ç î×þ˜å1[-h­Åt&Ø0iüsœ¬“ñ}ž¾¡”#i¯”(eÉyZ±l‹˜p™ Kw+F&2ƒX¼Ý¶Z¨3ÊWF—9PÖM"‚ÝgyU‘°tƒ¬×ÄR#ôÚ€ˆ¤shMßܻьóL0n¶w É×J Ý?ú:Ôþ:þG%a—$žâEø"q=•Éè¸÷öÏøZÅÓ9ÎcmA¼›¬m¾¸N_L¿žúëXvt/lyBñΤ¡©²wy‘Ü.‰M›Ì>˜´° †È³³wm›H@5M—SÃÔÐ'ÎU±”ð«Üº£™v_ÞIÎA4¬ ”Ñfðl‘®!-ŸT.šêØ–Œ‰GYíÂi&’eÙ¡(‰"6ÕMÖæŒ+ŸªΊ)·êîÅœ½/+Á£„ÞiÅBsêæzMcËíu2ìnÇI¹ ˜)› íÔÃC\ß…# pôÓq¨kþ– Ãq Fç£!?FnʸŸŒþO Òn¬Ý?˜uû¯º‚‰†à¯ôä<…í§¿Fƒûð-雜ž JÀü ÛÙÕXè›´6öiIÙ€þŽ™¾lÕܨjÁ·n‡C•¨¦?N™zwY͇Ùh€*!,üy°‘1qÙÞò„i TÏ;˜f5š95LËW›à”½¥ZsT6™–Ë’*RÛ*‡knÏ}2ÄЉ’qRòÃîG(šZzj¡v q@æPÌ „ãß5Éäb%ÇÁËC9}}YÌçCtC¯JËóÏ«ÖUg¢Ó£|0Ĩ;›ÅM¼hÎ%€˜½„íÖg?°p¸ÆŸBœßÙ} R¼ ÁIUífÕ©ÙÓ´u·` ”ušºyÊWÓP'ÌÐŽ9ýæ*|–2Ï‘˜¥Š·f§P•qØáQ¿àUŸ¬_Yåµ-‰>w׿dI†øYO•d´ @mÝOx,9þ¨”ýå¨È¤¦g¢ã‡­~zèi#½¯[ %Rã™W3Ù$5.ïðŸÄ÷›Õq&à¯+⪫”ˆÇ5éi«•h°ÉDoÿE½MæÃü•oe¾Çïìš}Û!äÑ)´L’û E“Y—Ç¡¤hûˆæÝ«cÝ}-˜hM¸ª5Î$Œ_Lq¢e»¸]\h&4+½to•¨¾'Û2vÍ÷a7¬1‹¨2†):o™¸ÎËAÝpMtŽóè'ņpíŽ3Ø.í›jüqwÜ ]OÊô²Õ;ÏÓ•‡(£âÏcÃÅa÷~ >žÎÆ¢ÍóöÁÇ·ÄSœ @Zô}²rqDf,Y¶(̸;õ%3Ÿ÷ŠwS<Ýrpì°Äf4Hîh€£õJ#ìqÏ­àͬUëï+èâo³ ß(©4‘õÚ¾-šÞ /È849Û½‰k ã¾'¤c¶¢­ºõphhn8ŽqW°Üÿâ,¢«xmÕ#!ó9†1+Lǽ)ÿoÝÃÑ@‹V«qk„^úJ×ÒùmõQôЧŸ/@þjLp“HbMÇ_„ÓñAøOø:ØôÁØR˜‚äÖWÙ5|ßhŠ•ú4•ÕNÙ½¥U ô”g€ÛgaHÕ ô¨= u¾ÔÎ’áp„ò©FŠÿéž×ŒçJ/C÷ iýú(µŒPtÑ0üpd»+Ú;]˜Zi7b±M›ÿ³î‰¸h8óÎÚÉÝÏð"¤€°ˆÑE= `êÇo6ˆÕ±'@ýâÒh~äqq)[€¨œÇnÍ=Ðh Ó¥ñSkfÿ1¥x|(·ÿµµz(°õLÄØµeà7!Øeçn釸‹ú},¢Ÿ¿hˆ©ôér¾ÐP•ibä¡vL&óêæ¤Á¡iï´:Nf4  ±ûÄÈÝËÁ½eŒÛÝ ?!ãªåýE^ qÀØi:˜C¸ÃIjé çÅ8œY!sp~$]G?±?]ºäѺ Ÿ‘‰œáø”­UDÓTwÇ)ª;è@Æ"¾ËHnr½žÉ{`—_¥3¦™Ë»Úü£67äœÑÝÚE_¥Âô™4žªÙ‰|…½“¢Š|]ô`­¡;+§Ìù©Á;ÖÒKeMÐB\'D°«Ÿõ!p¨2ØH‹Ö½:¤$ÈHMìVý^H`ŸV¹õõw¨H§¿5z?wz*ŠX¬¶N¼yº­­ìðˆr6r0âoûi}ÙÖ€øÊ±×¹Eø1h:ï&”fOJPÖÂ3ª­<¶æÍìN¦ˆùª´À©¿ fWFºDèÓ<–Ï^蜗Î>Óƒrˆ+¸rq}X,hÐ:pÜñ‡í‡S‰,º¹ç$¬š²Ô”u3ÁKå–¯£¡0íü|/(9FǹæÚLYÂQq î+^‹Új½!ÕfÔœíÞqÉY•,5gÓ$»:|¾'÷w¿8ksV)okfDΗ¤±©¤÷¡˜áIA¾ž.r®Ç‹xEhјÙÄ_µÖG¶õ-k¾RZ V‚%ÔB&sq:Z×¾¸s5Rî>L+U§÷(?Α¹ÌPr_w6n0K(ʪ ÕòB’d5Æž×Ìn%p¦O/³.¼Š?Ï^.ݘ œo'ÇÙRë~NX1¤bãþþ;£þ;ôb´Ç9a ȰÕÙ®ô‚¤ØÅ½—‰ÚwøûnMI.ö¿UVù-XüÁ²F [~},ÃtºÜ#ésHeÐ(Vô; 52•íma( ¬`¡(tÄ(l;‹êã†@*n¥÷réSaÙ +«ùÓ,Ïq‡¦»°¢=™‡Kj&NkjĘÔ.9„‡g3Ü’öå|›fs ¶þD´tEAÓ»î}3¹¨R§¥ ^R/¬j5Nÿ”§cmë§•iÃ]ÑY™£ÇwaQ 6K¥5¸%S/Ë~m’áÛ>$Í2¯CÅ9¹ž×õ N¡¤º³XÊ®5e´ È6F†Xèdš9v‚äc zÇ9Ö«à‡“{„Üd©ÀÅ0¼õÐ&gð÷Ã%f1z GT¡½Ìm2h0ŸÓa¤4íû@Ñý¾„ø„i'R^猤8»кûñ:.˜€ýÖ6~p•KW«Â¿Ô ¬`@¨Ù`8ãQžù2­Øf ÜÈß„UËMµÀ4%D#6ÎÉœ9¾:ÈÎɉ€<ðB;¤NÝÀ«YÓé<©ŠÅWXãXa•y_éS'W&Äz}uÞë'L¸%¡„Ö-†R¯>ZÑ¥„gÙ´d¬ü/dŠË|%wx[ßÕv¾Wj3‹9‘VÖ9a~M‡ýgY4!ãˆméÌ­C$¥&L2ÇAÛ…M%º ­Qðƒ_†ÔÒ٤и|•é[¼ì;’\áž_qä.ôùsĉŠNy¨QTà{‹>m<<8Æ'ÝâÔÉ6£ï!õñ?€$Iá½qg¡Gû¯Ò•Ã8´'aÇ5}¾úùÞvN3&’È`§‡#YO/>0"`¶r ßˆ2hWÚ²¨$¶e©”T¢¤Ø‡˜Ñï`Ù¶Ç@»hÛÏúã¡éaÄù;§<zß k{(+2t½G[QÚ÷¦Æ°Fˆ§´iþ$Ù4‹;ü±¶MÙ—ÐP¼4ŠTi1þÒïel3ϱþS˜ŠH Ô Ñ•@Óƒzô xb5Æ­4g`bÝ(Úƒqn¹+ê^v}æî†A‡‰ÎôÖÑJɰ¼M—*ôñ•þÜ Êg„’ïNÞ]ë5Ã݉€—÷ ÄDúˆ±gŒ9 ¨Ñ ¾µ5ËĤ=3t+BRÎJ>ùZdë˜2Ì$kàÎê™9Ö«wÚ¸ìÏ€/cà)íN]?Y{ð¥tæv¯ƒ™X¶ÕõÀlÃvR™Þ*ÄDï1ž\Œ‚­œº¥…žEf.ÛTxN óèÇç௟k;;^õ„RŽ~\¦t&ím tn›§Aê^ëóÒkÒÏ¡©¹JÜHU …1A¶¢Ž§IÊnñÖT9½.R«ÛIo«³kX›yÇÒ[!Ð…r,ÌòEÅì'CF0¸2…ä.ˉ¯vWe»Rƒz£áàÙ䨯ÆSrv%;ù?^.ªw#+JB: Vý•óÄdª·Ÿ²˜7ð5áw°;Šú—ÙÁ<;–ú ;¨ÆÚ‚`Uù4P¿É—'¥Ý´ ——½\Z¶$í nFŠÅ2X}äËCa “l_Òáïú…õÊ¢rÞ.s˜BüëSïÑÌ€?¤Aß0 )€„úd:]‰ÄV/Hú>0Ž[æïq?lÀ˨‡‚°_‰KÈVPµ@ÇVØnVå×Z%y€L@ÜR°…§ ž¸¸ŽÄ%ˆmoî ãüÚqsì¤/f‘õþ5!¤3ª¼ª[®q¹§Yl:ÛO“NDð£ òÎi€ØðPJªÅ§—f<ͨ6à)ôJ 1,3öigM(ð4ä*!6Öëy"Ve¢‰W+‡ðA‘žb„`EªU=èFöŸk×~²¢¾œ^ú¢šd‡“qŸ÷Áá©ÔE á™}ÈEû&x2XÛ#\ì–vnå# Œzñ¹çw¤!™Â$ª9O?§çü§ØÖþû‡çÃYÎÓX.G½ŒÚUðÇJb#¹û Ó–Tì1ÿ|âè±Õdwáç'ÒÆÌŒö$iñaZ“·2g[;¢:ö53ý9ó”bs>º_ß7ƒ!€Å†½6Îðáñûµµ\vg caý2—5]û¡Þr³ÁåÄk†xGÞöEï«‚“!è}ª]uW}±óÞ÷”Cnul½Ò Åæ@Þ’û/à[,´óßß’ö!gøÈ<Œ–‡Gûõ–§è~VÚ*UBÇöå^±];}6 SG…'ÐSj¨Â^}¬ƾ¬UO·¹t7âßþ!ºZcÝÂ+Gífû”™ûOŽJúÒœŸú¸¼~ÃN‡ÀákÜ£\à¨*×àP•Ê´…KÁ«ö™”÷ÜÜ“,2³Wè ×b‘qWz(=VÔ»ÿIä‡9šÉw«,èÛj•SÙMݯ4OHúúáï»ð¡ˆŸ>G•gÚ·—dÚËîöGrXðþq´–fE9–þE„tü|<ãW'׃gë Ž3O«( 2=uia<åìëÈìõiòvA„îˆÇ/.?ĨŒ¢W…쥽.„­MM+ãŽN[k†t2”Î$RHÙkޱUAsp&au–¡x†÷áŸçÊïGµÞy”·bœ1qpû@eÀ^±ÞjÂ#ÙÏ}¼‡fü›µánH_f  ðB’.ÉX+^@ÈFÆõ­O·<$н>žÚgjÇE¥Â½}œ däÃiP(:)A‘®JŠ´Ê5Yš h‹äw°[±Ç¨È[®²øæiíd ‹R¸õÍa_úH•Ì¥h—ÆeÎØˆÝ}Vȸn¥Š&ÍV@d¼R WYÖÌL.r™º¤=ÔÜPYEÉú¯1z"¬‘¹úÉHTñ”OÐ}+ÐÉ ìR&h§a»–wâvs1Ú\^BeÑçm×t¹ì…•J¤ol Ð¥QV@…Èÿ«»Z¢'XÂÀE;fòè,sìê+òËææ`êK8N5ÀÄ÷ÆÃÇ<21ôÑn÷°ß{39å¥Ïaôâ­®>$•mK¼è±rÏôŸ«õ_ߦ+<Å=Õ¤/vߢzÁݮ㑱p^x¨Ø/µ¤¦YEqb,ë„eJ®›Ãƒmƾ¹ºICyÊ’É‚ë] þYhïò5׿Á5›ûtï‰Ô‰«'éÛ€ÓÀÿnDÍd†àfB!ÛŸªØtc((e"t˜GdîààípšÔSzÝB.J{ú?`‹ö#©Ñá=€J#Ú/{îâ0Ñ\æqPèP K(ÁÅÓ@´£“ @ÏÝä­ôšè¨¨Î{«„ÖwdL‚LÒdTÅìL*Òh­#u\ÔŽ²ú¤´ÆË*~LG&òžÃÉ ÃVŒ•KiF°”+Ò±@Ã]^m3Z°ñᜊj‰% òst£ u¼?ð8påS¬3‰H«éÐìÑ^e~U¶Ï'9@ßž”±x‹ÝGaˆª“†‚ø$ûÄ“)±M9˜$ F øõŠì4÷ËáKZ¿ =qœh–›NÕö Ì¢'‚Ä6´xcâÏŒ8Ü4 õQ/y÷z£jv+Ø¥?õ/Ìpô6ÿÀÞ)‚ƒÁÈåôÚ á‹PæZ^àm:¹¤Ã˜az @¡í4ˆQ‡‚~rÅh$™ lÊqþÚûžˆ5cr9ù_­î?õ¼Ø lþ‘ésf¢7Á?% I“– rÞ…ÐRRàá_§,;ƒü øåoW²ý ä¢ Á,ªv{Åp]Ram§€Ó¯Ã;ТùV>%ª­·-ÐÊîœPƒª’¦„ÖÁ8﬛²×’„ÆVÜö OÂJz½£/œåG'?VüàH’”&—Äß8T´fè •x$ñ&xïy0ç}\Ùð›Dxߦæfän/E'"&¿ßäè»Zø–Zc:Ž+Ñò7­=€4¨gÙÔÒ/…`o[6jôHsâtŽKwّʵjê»[‰Þ\°–Ø0üy)¯."Qí_RpzW‡f%{ Tá#Ê2ñ>T9 3»€EÊâÞ%mSéå‹fÕtÓZf? ªÇ¸2ÑfGKœ}½DAêf"󯢑ø¾eí@àÈËO¶è`ÔÃ’‚bH­.£²M\BKX*/Þ‚?à’ºÛ­ú‡xÎþfvZþ鯖¸¿ÐÊÓýäµ Õ'Ëìô§vdO»ICí¢Ûk bÑ팋µX—þzóX€Ã4¿õø7|¢¬š±LØ>’=ö[)„÷}Ú‚˜22‡CaÐ_µYƒ0òdF••í4šk/{W£3j¬Ù檡ùƒ“Ë·œ_¨QâØ«±1"aÔJ!ñ¾íwXçÌÆˆ,ƒæ¤á€‹ÁÛ`ÑEÛ…²÷Ä”ÜDÆ|µ3€? ¾û‘ßÁtÖ[ô‘ %ˆ×:©?Šcï§ù«'+±ë&ªF¿KŠ´¥Qΰ¼½„)êØªqøwÌñ8w©Þ"­§ì£J?ºfŠn¨žåJþJ®—ÜÙ"÷“9áX_ÒYwÚ2²t[Š{tU­Nà"ûñDŽòN|7…È‘iÀDÓY;ö£[ÂØÕÁfÝD£µkV¡Ö³Ž¿¡ò‡bv9.5å„úª‹tô‘™M=‚;ÜÉ¡ñÉá3–xŽvöÜ&*8Û%ß©IqkïÅÕÐ(Q®üé­sú“þUD è›R/ò†~sWR^“»ÎµóÏK¼Õ þV,`x/‚¢G «çÇ‚‡Y猬xT—±iVžÞ¼b9–ö&ÕqY°Kß1ê3Þûv /÷ba”û˜lE8  $©]M¦¹!ˆS¨léîþ+e ™8—Žât³ºM7÷‘4š{wJ³&~‚ƒ†éÌoµ’8¢*·‘ö\ïB5*ñ”˜fëÁöŸ‘[ôb8Ô¾[(~ÿÏ¡5ÚÝî:©¼h†¥F¥PZÍë«fáu]::éw]´Ëè¬ FºãìXä°%j)¸°4q(ÿãk‰ƒàw>ŠL²ß© 6ä4‡ëö+{w(ë&ÎÅ:ªkAÿ5ëU§Õ¶ñà2½n*ÒÜè𑟊êü&"õÿ˜@՛Ѕö%ˆròüZ#ê¤þ ¶]ˆü¬<Ö ð¸ /<&j³„Œ° QnóDèbzbjNýí|·Dõ2A]Sq2 Ñ#QYøVÁò嵘R…¿a‰2*¤Á 7Äï?ê¡¢uE²¬¾wAý A£ÿçáþiù¦F J¥Z°OÀ/ïŠXAt*wù",S‰`ï‘;øH›ôU§¡¡j†«ì^«ë§yJkcVÔIOeÿ!Û,—£¿ùbJ*Ó·jÿ°„îé»`VN…9GÄPÑ#eí˻Õ¯½H9O­sìÍIB„K]Rìi‚c“NÅg?²÷4ºùÂÀÀžñÓõø:€€¶~¨š¢£:œ1™‹3ÂÊùn²ÉŽEÀ+–}ÐuZ²kÚÊZj(™¨Sós„jK{–{b¸£mË) XùPÓ±!ÎÎZõÏl¨›«ª$¤"þE4Q—qsצ’H”¿KÉ•úhÇ5.«qpß ´:uÄÎ@"¯J¦üå5p[ÈÙRÛšúVa5Jc§{¨›;—#“¢_«»¾zWŠÿNÑy„rÜe,¾¨·¬¡í `ÞP ÷gÞCôÊÖÏÐé}¼ &l‰ügùï7¿Ç¾õN¯ Ó9ùß4*˜»aÞ¦-•®{ MÁîc¢¡îô;*/xtFp ¬\o%<ŽÙû?˜QZ¡ Õó¡×ÖŒÅ`"è‹í!$‡„IÎd>@3Ö¶;°îŒ{|ñºoÜìµ { 0KªH8èñ–òØt¾Ød(‡°2¤‚Œv~¹zËéÈÜz­Ô­Û6tÈËío–Sa¡’»1"å«rÑJ„ázð”¾òæ¬T4ûµ™Õüu˜ /rtч½Êœ¦ý ß¹íQ™к̒'¼÷ÓŸrãg÷<Ú(Š'ç áö¹Jy“bÏvd!„5‰mPKlK?Fpó¿ÖŒðÙ‰ÿܵgp>,ÉP`Ÿ­äcÑW$—‹ÂD<Ó~é ª(öcrçèNŠ2º–š|cý 9³ùÊŒ=ëx".¶áTÚìW§Š#{ULR¨ xSªÏõëgÍR&)Sæj0=+Ù¦=V’4êŸxùÍ]ÉÄ51ÑÓªg:å¼U_.6¤Š‰8].Vd\’RÀd¶7¸ œ]ùŒ­@lá°±S4·D†ª¤ó³+š° èÎ È÷Ý}Œ­õ¸_Ê¡jì€Û½$‘é‹t%ñzþ·’3¥Úê¹yi´pp)§vzub·\o.Ÿçõöà%KHVx¦²rc a÷¸¸Â|‡ ú(Mh´¶£ÈJÔCÀ*F)A²Ì7XrŸÈË"tèÐ i³Û²îh¸Éà*ù[Q‚Mä”>pß<ˆH'p“WÅ|qRÁ<·a؇Ák8J:Ð$h»S@g›À³/Í9/–Šb±ÃùiÒ²ÃÛoœè 3±A‘-•³|ù;d¼7yE+A|‘Ç¥‚º“‚Õ-ËL;Õ¶' ób!ŸÚèÁœ„p¹a9Q1¨J<ïâç¨åS«ŠD¢ï§5™d×aÅIøÃ3®3?ø¥eï—(öÈjG$_"æŠÚ ke¥×DðZѺµ«€Þ¹;¦‡(—토ººÇ÷ŒÇ]ŸˆÁŸóš—ÓÀõ]òÒ%Ù‘{Š±ÚøIl×È ¢$]‰V´VžŒH†°Í/¯0¹`ð,µ\J‚"UZã`©ÂR1€† _‘ÃÉ[‚¢WÈCo¶³]j] ,×€©iaw‘µ­×ѧg[j‰ʯð8&NŒúØ#0aHí¦öp„@iÊ+4­F»#Ø”…AA^‹UéäÒYs8J«xèÔ¨È4„JA>‘¯BQ.HÇþ`‹¾qEóóÞd³H¡z ÛHì Qò1gb·\H@î·¥-Ë9¢Ì]Jt€%^”%>ÿ8cTXÖ-{[C…¯(]EŒRì€nLù°uª§L˜§Ûª{àHwé1ð{yúÞs‹·eÒqvЬZˆÊú T¢z‹‚æJ>Î} Uc?áìÃÃw ‰(ï-nrz ¼‡ìÂì«ÝZÜ´ŽÜþ‡3±Bp¤ªÿ&Éq:rI€Âä~g é朰˜™&ë;¯3‘†;’l—Ä "ö2Œó†” ´N&ˆªŽtÑEIÛ«Îÿƒ¯¢ U2 û:HŠÿl_29w]êËÕeÿ LQ1õ`@%Z†v¬'(C~G¾ä†UçÌ}e–ÝPʦ¸Õ"¢½°W!Ó¼7ÃêÇ«Iæw(^©ø?À~Æß'6b®)ÀÀ"¿ö¬°ß3æfé¼-#þu>šÚ§«sûÝrví)g;˜5œ'ƒR -g5¤€—4w}®¨ªW0 —`‘R¼/ÚÕjʦï|þÐùܶÖbŽtô`'㙲– y)°ø DêW(µ@>CÀƒ3*‰ÿ(NÓx¶]š+1ý6&ò^ž·…Å>íÖû‡‰)ÜÇOñVüeõùä*„Å < ÍóÛnàs}ª€@Ðq‹4f-`ˆOOÙ£¬¦Ø­½æ¼×·úŽª;äÍ—Î=9 Æj,}´š(¨Cñݧ¥ƒ-.#ÚU‹ª\0ü†ß XhöîT?ÔTp÷¼ñ¬\¯ˆrBü\§vbžýChp' Zn°êUÅáß?:ïçAñR»>kO¢K«å¤}]´Ð…ñƒI4¤ O_ñ¹S^ yF‰ú"äăRHz3 6ÑA¨Ã–2#"ézNéRoyQ*F7þœ%Íd­Rã°„ô§Šü]ÞÇÀ!†(éí]ûk(NîÜõOJC±·®1+“ñ¸½³1ðæ¿Ý–é3Èêå¤ÈŸâ0£·{þœ—ý’Áê13Y3©Ø'Ö>f¤Ù k¾j‰VåŠC=ÃôÒÐ%ab+ñCm]$73Ç<Ј ©Óçu:Øl¸*ô²äPž¥†Ktú/0ÐËûŒí¿€gKƒg‚ËÛCúßp£ ‚§ÃÈ9ü0Нa(6êRãn¥Ÿ3üî;op¯:þÕr¥uNpCÄ­TËÏŽÅQm#£î\F`Ëûôó5aJZ•Å㇀S¾öbo“¿6ËPd$´G³¾{>“/ ü4×&ûx>ÃßÐ}9Z6€Œ§Úò3u=k)Êç’ó³M‰ÿ~õ#ïd‹I=q{¤UªßUˆÖG¥d‰Qˆ[å¤Dsš P‹X;3Ûk]Ø>¶_Û¶V͈¯± çꈄÅ»ûZ˜€ÔǵÄתdhK YÓ:KˆMÌý]H´c¤lî“öCäêFfËWsÁCÌBHK?.W˜ñ[‰öÑfÑÔ¹É1€ü±9!l€Ÿ²#¶%ê³9rá4©WT‚#>èFbsÀŸ`òþ`†°S#¨/´¸ÀQEf?ß³ôäY5›F”ٸƟŽi$§×ð†ân7Ì®Ûöò7AèJ'¥§{Voû×°ôÃÜ—òå°ö¼yU!©È g¹ ™üªD«`žçîå,¦úºtÐ*Ñw­g^<ç̱Ýa4K~ü«ëµâL'›:Jâ7ýéŸNµäöN! /äÑ”,èÑ1’šßò2c%QÛLªÜÉ!³çµàpÅà†JÄfMüMyn†‘Xï3zá©øY_Ü,ÆøE‘סֿK?3×ww…!})ÈBg^q[¨¤3¸?qïÃ?²øïQñ$­\‘­¯!H—Â,7ç!¼m†òl'tx5es% Õ[Ùûv[ ü#!ôû{$sX¹´÷ŠÌtÚˆÜ ]'B4Õßñ«Ù€€TûD+O„bØ1[ž¶X+týò“õ·¯Ÿ7™¸°ûsqlYþóx‡\t †JýûÍE-ìé‰%1Û[;,þêù™™`.;îÇÑÍ¡£n‰räjÇB»½g¯±˜àq r!µáYz©BJ›s»-ÈŸ!êk†°„vÐdì\EÆ~áÆØG1¥ÛAõz94s²fì¿Öéh>´8èPßOÆõ‡žˆ ߥ#‚K1“0¯xŽ97†‰¾Z'ÑPY ›2 󷇶-#.žä;Zªâ/H¿üìÇ7î‘Ó– °±¤§~Ï5a\¾Mk7  EY£‡‘.mxÅíŸtæ´µÐÿ”^¦•*çýûƒ%IÕ3¼ÍàdÈH£c§b߬þ©0ÏغÇ;ᱸ{ïGûZ“å%¾¤ý~]CdS$&`ÀcíÍŽçŠì}íY}žºL…”ÜD¬ü¸ªqàE’âØ¸­Þ(‡¢oßï}~W(î (qÖ0Œš zä£l½2_Ì_f.eU;!W7<çì=Y¨Èr©›PDs@d9›ý-ß ðbæY0þgFÙ—toœe«ú0óÍ[×x’u:! âŽzŽ0v&Ñqƒ—CæA÷ð‡„Ñý?­qh`€>Ø’J(ŽYÓSÁ{KyÝBºäû@uhkT—2_‹õ–ôç“áúZ7®Ê¿£#À~Ãyå4+æ[D=#¾«uÜ|C `º.•Ú0œ"EìÞ™±ÛJíÛù\aÀ9ôÔÚ¿©‰y#ÑÚC!ê&ŽÉW¯âK&¬ƒð‰±ß”¥Š­˜yÓ>{a¶½F¯e$^E“›ÔKËr¢¤ÀJœø…Ä%$ùé¤ÅRü÷£}l¨é€PÄ™ë %šÖ•ù;ØÈz¶ƒÐë±^ßm&5¶¿?ùÄ`ÇF*ãñ…USH bM TªË‰(2‹“9”岯s;G"P÷\š?$‘Ýþädú6n@ PÆMÐÁÒоÀÛš°‰fA Îӂ뻬Š\Gº?œ}¿æ“#løé)6xk¡#‚¾ñEñ›KC?Ù–åÃR5©“LJ]B‰’æ%çÑ[øß–jéâ³¼¢*„d3ßïüd¹-/1©>QÜØÀKt‚9w¬\¬4vàô«$Þùk`'gÀòÞ¾kUj×5ÐLvÎ R±Ì{ä&†ÿðKbl(‡]SIc,ËÙqó6èµ¾gPl|„ùÑêc±Ó#ŠÉ)ä&ᕵAX_äÝ~ž!`ÿ±Føßàt¤½€­– ‰vW{ÏNZpõPÜŸe|õø‡Z=Í[œZjD†Óv p·Yqòþ+OðøC¦­—»p‹oOÀÿl}!}]ÿ£ Ñ” œ‘Dëqáš½8†O2X½KüDA÷ÿƒX÷ ‡ð]m& h÷£2 öôÙå@Lø}ÎJÓ"p®\:ÎÈÉ•øKâ~âS:÷{Ú9ZP?MîK+óð/(ë`·`’Ï~¾P¤Î§”›‰>/Ž êQ ³ø<«L[œ!_¼RÂÊdõ¥€w»N ‚ôVµØÌÝÔqýbB+Ù=²ÜÄlà‚Ñœi"¡o ÓÖž*õ„ú_m ¤qIþ¸Ñp¿ aaÐ’¤~‘zÿ†ê¸|½A!i°L²‡ÑÌNRk44(êõ€û@h<? ]~ sThò¼Ï<Ýø¿+Ž|± äIžâHo±n1æ_ý“h<5a¹†'=$uŽu\Š‹èðÐ6]g?@gÂ.5kS» Î\_Îë~¯3ZìþSf­'óÓÞEÐûG/%|ë`ͽۢ¯¡‡>Bã7-‡¶:ûg®ï8"uù¯ª%±W~ë'ad>gm4Ì­m‹÷v­Dñm¥¹D›o;n)îÌ¥h«hÞ€ •]“£‰u”êa¦Èݤá@_Ôº¾HˆÓa5 Þ“ˆüÔâz†‘¡M€C•üÜH'a¹ÝH0jÅ© *ã,{‡,½ª9κqð“=Q)ÚyÃp ™ýGѤƒ­¾ÃX™Mî ±â3¥& ßL¨)4@>FÇ%Âøî/T¢¬*ŒëQû ¥0Þ"ð…6;Ù©g{¿ˆ@ƒ¦¯ÈO7•—!øºIXy  62f¾ÖG¯ŸF->^:]¹fìëx]âÖ M¾¿J~ÉP.¹$S½(’ä…¤`#3ˆóDOé÷c=pÝL “Ò¬€on*Uî×yG›¡°;¶$î\R¢­²<ëÀD9xnqÆqÓÎΖ“"N{r”œ×áT!LºUÔ¢v3óZå(ý±{ýeÔ6/€†äú-DLΩ¾”ˆÙL/%«-¯ꛛ´¿Žå¸ô]-hD}Ü.APç=ËSÕT|]ýÑ;?ØàëD*[ž‰¾]j$Â5µ?^-@¬?v‰À:1¸y¬¼@¡í­T9·Ûš¹ÓrsaÇüü\á¶±w/-dFÑ ËqP0s¶%|_ 6ÂO†¤ZÒ~g汦&¦*¡½±N§ï*ó!×ý‹V´ tàÁ_à$ÊBWçÊ`åŸ^ðz¥õ1$ö&à©D²„¤ûq[êh(6Wj:8Ãq•<¯h­Îÿ²§Óõ»ÝÄ*0Äð)ëÖÉö(ØÔ/(Àì0öó¦u®Ãk ÍŽË“?Íà›ÿNSh–+¢3¥„já~U@¨é8«G0À$qeom •ä nÒVˆñŠQH•ìé. tæ [üÙ,:Z÷ªd:@mŒnôõß§A÷ušhÄkO­™1,‘ð(íïÔÖÎ÷ûÊ>¤Qì½cE‰Þ ĵÂic¿þ©Å1¦Œß\3/rÀÇ¢¿´…ܽ‘ÑW/›[9µ¢Š[º\aÒ{HqK¬;˧bÔ¯½ÙZé‰vŸHýËçÿÞÊ:”‘„ªÿPÂ̺»]¤1ÕÅyJÕÍ›ÞóÛá‡þ~èÂbºW#´rLäY ð”¹Tî Kþ}RÜ7A¤ÖSü¬}GÝ­iƧÎ4e†Éýå>¾À$(ØU‚§Nþ·óÊì—˯à«ã#‘ŽâlÅ=.‡84B¿˜|s+‚Hâ}÷*/pUb­6ò££eÖëÄïÖyëÔA;M¡¢š15Ùv *vþ0ÏezJ󟄈|hUéöúÕâò@Âö„òÍŠËâ=•{•”\[]Žèi ý+©JH—66Ãb©¿}ÝIÊýó1• Y¢Lþä`DIBMØ.¶Ÿr&\WÀ&÷xŽzaýOf87 è”EuÕ±è¼Ä™“ÊË`nÄ}¸þ Q‰:¨?+×RÃuÐþÜv×ü¦0Üd`»×šñöEûkR!LR¦ºø¸‘´ î]¤×¼Í£¬ÃÅ ÄШãã>«¬@)øî¸ûa¸M†_T‰"Ò“®”ƒàëµàf Xöfº½»½×²¶|“B#Ý5!°ŸéòÅshqÐ)“`‹Îï|ã{©‘øŒèS†º}‹É+yÿÓ ñnOû }ÐÙØhb%vG¿¨âîºýÊ÷öU;_µâ—l·ƒß9Ãíe8X‘+O¸JnU®î¿Èæ>é´¬75É’OM‚~g_¨™@ÕiË£‰Až?°¡ÿ"èS±™\ÿŸáž(6µªLDvåµ³P-:nè” IŒfiÖŽ„ó;˨7õȵ Ð&õà’y“ÇÐzÀ˜_Ð-Å•ðxƒCßÄÖjxθ¾ïßá=W?:»ÌæíMÔrGøÀ¼wP÷Ø4Íù Ldî({c çiõˆ¦(ÆÓAà¬SÎQŸóÁÑ5rò†  J{‘¥ ³¡ÖàXœ‡ÍQdPÝ¢Pqö÷¸C–R¹…ÂÕiÅ%÷„/ݹQ±ÔÿÁFww}lÍ”Õqd~VâÃv’ÇAå6&jDv€'1H°eI¦ø¦#iƒ¹‚aÔ›XgÈZT«ýa”ŽôA™Œ]š-mX…¼5Ùñy0Í8EnŽ4B³KBiþhŸ¤Ï‘x׋Kjé…9`xå5òî㣉ù9Ö*VìÝúVÃaB¸‹÷Z’ º'CXã‚2ï9-uó*_oh ëBEV¨ï"&o¿òÒ«Ô¼÷4Ö!ë!8O³¯f m´iKF .2+ÒA–eXu){i2/œŒ),¼‡æ˜ié¯}þŒÈºâávþ–gH*߯Zõå@g–óÀÕ)Rй1¯Â¥õãëØ‘ì”/C M J,ãI*ù^Jlü îˆó§3ëÀ-ÓH¤{‹ƒ8èxJôzÐf2£œ÷æ*xüOÜìo®Ö—Ò¡’ÒÇ:1çÎf¨b¹©|d8oÂ\‰ù Âú»xÞžÅDI9¸JÅŒ•Ì L§^çÇ“Šü×C¬˜múFÁS AÝ,Šw˜š¿QÔn§ªg>Þ¨'ÈšüHqðW‹ÁÚ+œúy‡‰xðÙHõñÈSùÜáÅÂïs&Äx›TgÚ»m@H®sòì£gãIú‰îÚˆ‹n£KÁº:gFAe¦€më4yPZ®kÑÍ®†úàèÀ7Dn³ê|Þ8=Ý)C‘•´ø1æŠñc©ÆG’Ë7¨ýü{û£Dî9D#Y ÊAž Væ33äy;3Ì™FDZFTz˜-Èhpʃ  æoNëFx1™wÏÊ“&%|‚2´:O[Ô )ƒq}-ÊÜ©º_¡{ïÑ&ä)_Zu2ða²§Í¬óS+Xi¨ðÛ­*6íÚV)š%ž¿ÝÙx^ L/ÁrìXMô!Örš6µ¢l·IOÍèAÚ÷Ť¹…ìàDáà¿:ÓÌùbX§þûh…„ƒÅµŒ®Fta©ðßßz&—}"PˆjéhO äƒ:pÔ`¿F=rçÌ7ƒ¾$Tû=4º0þ?B!LûêøDrŒd1QÔYJfQ#š~ƒoRž=uAú¸3Aרý½)×qŠÙ{ Š€¼ÁÄÂoÙcN+‰ <;ݱ8ü*EÔú«eáWJotÿÒ‰ÊüpRÿ²|;€³w胋\žóëPqC£jÆw3ŸË’çÒ;+A ”0ãóöò»¤ðµR0µ¶T[F”hŠy\å]²À}Å%ù,E,^s⨕&½Ûï}±½±¾yrÞ·Œ¾ !K¦ua8/W Aì—"ƒñ «Û{" `eÕÇú¥”Ú©:笱hqAÓµ½“, ü«^OJXûµ¨Õ…Á"1–QäÒ_#.€š:¥:Óó‚徬/—Z3Ko’@cý 9àyìëüö}%/ ¬¸·’ýûŒµ2Wx’Æ'©`Xrò Ï:GŒQ¢bÌOrŸ+A[\ñ˜?Ò:»hïþ+#‚|¹ŠôX‰?öwϱŒ©Wô¦Áɕ̓óæ‚ú™ƒm¾ç`ÝNΰEajÙ’éÊTq8 ºò ÍâbC“/Ô ‰6pÏG,\{Q¥ô¯˜ÀÅ ÔWu}{r“¢‚}ÉNÄ#~ælwøy›æúÿé•õHh²Ynð[±’rnùDžà¹C!ó–~×aé÷hb*ööP…rŒ8Ÿ,`™–¬LðbEþ1±|†IµðØ.HƱYR‹£[èƒãløð$¢“ç=q¾µÞAB8Ýøm>§Ýgº«÷%ç ,!S…q6}˜?Ë Z†Ùß¾mCê+@`Þ¼ «W,Uj FôßÂFiøGÖº§Ûñù‘EU¯E¡Ò§ E=ñ~Ê⇋“ÌgHó@iñÖ̈džüÔI”ŽŽy-±Œ»%MÈ9ò´*f ›ÿI‘šjf›EÙ‡ñf¢ÇáÓG6+Æÿ,7a+Æ1»Èk÷Ûý«ÎŒ›H{ÙáÓ1'£O»}ª­nþ:P\y/ƒ~ªB^ 7›4 sE@Õö¸ÑH®xÖ2Þ¿ëÙÙÉy¬£”YŸÅ‰nmr›‚¶’,©‹¯±Xºà#ÔÝô 9ùŠšfÒ6,tùh`šäÿÖ§ÙX~_"aŸðèWo6%³É¯lÊï–u18APt- V·=t ¤ô‘Šp3ÝKp `)0‰ ñÒ¾¸ÄM˜ çr)#m>éº)¤ßÊ\.c'ÙT­¦@êÁ$ÝßÇ´%o¨gØ—ˆc9Êp¬^{Þí~µS#¶zLsžiEüFf–úe6EùÎÚÔcb>Ÿ€l!#º™sílXù /½¢ƒ„¯¿»X~)Ö^DdfÔŠñ²iú;±0¡Œ «Öä¡NOIÇá…@¶Gi+aع dº*º˜•$A„%gËö;¸û?a&€l‰¯Dܾё¯s ºÇ°Øpð®>È®KbÅôU¯|{ÔµYµí—¥«»•VóüàÈw4#Ù£tTéü I“ðsÿ5ÛöW‡´–&²“õ âg‰ê,òræ§¶§ÎV™oám+ÈrP .bQˆ…X±Hå`4+1(ìpì;©§~yŸ¤¿pL‚1YvÑûȘÿ¦Ö“b'vK$è¿ÌfÙ#ÃSÈ(ï’ìÝr³Æ ¯—Î~ .x€ëµ[§ÈíO!&çlµ/C>b‡8*Áb‹‡Àñqî¹»^bªÄ$œLÝ­È1Ü ‰;–ÎwoX—® hT ù¾‹òÐO$;Ö}ômi)6I £ÜÌ6›³]‡¸ÔÐ…n€&çA¦l¢ãÁkðûX¼»Iîô ¦"¹ý´fº0P5¯—£²ÓÙ˜B†èÅiª°ßË^òæÚÁÖÀo¥”WÒe}~ðch`}&\„]2Œwñu5Ûéúî…®ŸÚ|yVÒƒÿ¹ÑÁ³>òwDP¤¤Ö2™qVüðÊÖ¬º @¦š‚ÜÎ…eÖú틘]£$éó/V3ÖãÒ¨¦­§ž>ë˃n")*«9~…p• à³^-’<®%±bÂy N*:Ðõ÷–®Ä<>žÌ'º)/ÉmžÓë_[-îÔ‡öéê@Ù³øL$l¼jf°iy&òbl‰ |mn‡à¼U¨Ú°{ ÙEÈ™úñáËRW¡Ío(gC¡À¥M—ôz¿‘·Åµ €4¦ç­Ø†òØ4¤¡"ÿuíg9Iœ¼eë¯d§‡ô·jÕ¨Ú€X!xçSfºÏƒ>ëw¡'¦&C\ ¿•H ЈÝiRBåËËìŒk„¿ ÿ^xü]œÓÍ4Øì–ÉÛFf_f¤K‡ÐgÆéÊŒ—¦ïè€gà¤8NÜ]zÒ…ôT¢Òf@ý_‘°„cðÔw¼ÚÊÆë APí80¿?!ØïÊÖÂ"‡uL»3‰»^@d .‡r§÷ÀÝFa—Š]wÈ€;EáâŠQ®5|Úý+ BÝaï²& 3çȲƒô)Ô’Bu¥Œ”æ'³ã£ðU¸^H&jìp—µéTsû¯ÙS>H•c¤•Ü1eÈ2Y·îEε€?fßÃ6ؾ-ãä±-Ê7±ì¬¾ ½_?/;à^W‡úx¨ã‡ üŸ¼Tý—;½%ÚŠFÍs9@™IW $Ç”:]>­3K,*Á8˜Ú(õÇè+#(Má˜ÜÃÁôRŠ¥Jýe3L…¸o'“¶lužHØœbúK7òótAP•Õ\õµKj{MñÖ‘;29©,ÿ¸‘Ÿ&\ÂL\D²ÁÔ³1gx©$yáp±çˆÞ@HéߨJ¡DÞíÐ¥ª¹J‘ŠDÈWíkHs)ôåIØ^gšìNk¬¼à %IÊ)ˆd½3H\U9 !Š¢‡áQ÷þï0b1J¸ÔÂø£Ý“1¿]}¾pÏgúû4Ÿ#H˜º öïÛª—Iÿމ›Þó¾†áäCÅ´ÒM¨c”nš»Ã†`P=,ÝB–Éßò¿¨†Ú¸Ð‡˜4m@ܪS³£ç´O{7zì.øw{È„=/¬m± DtãêG$=«ÍÚ² ÓÞ|ÑC%€ 1öy‘?å[s2-'5O\ñYuÖNv_Bî­T¬¼üyÒé«WàáÒúþ È¿¯[? Ë#[Áâ/ù@¹7Z9ýÚ?•(Žv—–òÍd¯–Ïb).#öØ=ÀéäÿQ^)­‹»Kpê¹¢!…ÍðÇÌöáŒT³r¨ZT^áî,µÙâºjâ½×Ûå2ÿ"(É5IòÄe¶À §àLI™¬¾‘:Ûss\ñŽU;Ú¯ƒ¬”\v‹é "AÓ­ |5ÊnØow°Ýá)àž}2è“jOµº“Îf8i1À?N䥕ßZàáA…Ê(áäåkµoŒíéUÁÝôð]¡ÈîZQJm†½};IA¨v´˜ùу¦A,‰â ¤õ”®™Ÿ±¬mä‘ÿ-• YXEßV[•j~(7É`õA‚_Ú›$îÕ}·ò­è‹_'h <~,u7%t@ñŽò•æXÕ6h¼/Ÿe825›M_"!ãj04Ï”ùà1EÚ?,FäȰŸ©ºÎ‰²ÚPs…{8¦]9÷Î>¡o“à¬ÌžTz:Ìò«©OÌ5k…zΞC}OcZÍÐ_r‰.I)9V="V²å}Nm'ô¦Enlcœ…òEGÒY6 ÖÏðonZcUÛÏä俤¯{Í*ûúGW]ÿŽ"e5Ãâ²å.é1Ž·³M ejècÒ^‘c•Ÿ @f·î|Ú : Z³òn‡z÷)‹^ÿQ-s±Q»}ô…ZéK  05²ê¨½MSFng›Ø@$ZÖ©-xÌ•ìß‚I|nà}Ñ,}Ü䯇¼œa~Ûœ8ÇËV´3`¥Zy¯y¦NNXs4×$ð£Ä[Ñζ¯*“)¶¶‹´Ý¸¬çfÃÒ(ñ O넌»å®€˜Tݼ¡öÒÕÎ!¤˜ÊîGqþ˜iø2P‘¨ÓÚ_­ÜÉ(ÏÆ7Ñë0X+»F¹  À6ËíÇ<îaGãÝ8²s¹©³Ï4´ú‡ãôßiДùóú¼j3ÅMž„la+ ÊnbÆÜ<Ô@z?·¨¹yb«îÚ\= Ñ=,º°¶‹ÿŒEdG³é&ÜìœÓ ãšÀ »ø†YºÖ3ÿ8ÚqØuS”‚"ÒäjU"°rbÛåN÷æo™Šá\˜–l{€åLfˆ™†, ªbw ÁüFsÉi€?ä÷¿"ëò(ÀÆ÷Õ΄”›À¿ðEêÌúïÛR0Ì{­žÿþv¼*#îO‚/§Ã>¯Ÿ„Ð+ˆ—W8ÌÆåx’0 j¯Q]Þf¨¤ˆÁEz —M”A” Ô¶(ö–©µÓ®‚ìL T§šŠì¥<ÖçŽ(ÚOZZ2"å=1‚ø? 4ßõƒR+_p)B»Uå|lVo? }ŠÑh€¥XØ¡º:’Ru2ë MÏÛ>ÙŠ°0 )VíÂô}¼nrŒÁ¸Û$ËfúÌ9íïbW _$ûœÉE„|¬†²zÁZ» 9®DõþÃûþ?õ¾Ãrò&cÀûøw‡HÛ8‡ƒw¸¨ZÌc||# îb—²Žñûÿ tõ©…03Ý^?T™âÁS`¹¾lbà\ŒªtÊšfÉ[‡UŽËºëê•TåÁ&¹Œ–›ö¹û…N‹´ö€y¾W¤¨[Ú(bMš’4JÌ‚òM0æi0U=Ũã+õŒÇLMÃÍ}Ý5Òçñ¸÷™G¦"àð-âþµa(ø‡Z•³´p¨³Ïÿ€\Èyœßnp$1´ä&ÂÞØþïZyLHêŽ~õ]˜A»-\ö…dM¨LÙªwèó 6ÃÔ‘¤ÜégåpÒç¬h[ú œCCOµ­ÛÕ ã’EÏb€•ÐsÆ€}ã×UaœGÿ±_hrL¿íØ'Ú_(V¬åÔ›KEì§ÁœÙ·M¼[жVH#"ëaúÉ›ÁDÂEL³Þ_Ô*uÉ /³râdÕ§¾Òtæ”øL´(nìt‰oè^£ØbøÃÖ äTK™ä·qÞb?®éN¹ïr—Ί÷1g;µè¾ÅrÀÀ¹Ÿzì@·ü¾£äÐÜþ ¤ ÕpC5À3— yᎨÇO2¿‘ TWZÛèó Šõñ´‡r8Ht1ET`äB~µvá‡o5p²µ®¡"Ìô^1!T b¡qçtyA‡h#³äâ`Kg½¼žE åÌ ]ÃÀrC2îb£öѤBI›Ëœ§K>^õQÿñxàKæðb4¬®Ð1Í´;Q†äÔº1†;†+Sh:mS¨”è~¸-KÇÿéx^ú|Üe“ì–»yœçkŸ¯.DÚŒõS ÖæÖ }T~„÷¶9÷MΓ¦Z«µVBÐÖQO$›ü{{è¶àä݈ŽÊIÁ(F:ï†'•BèJcÈHŒR};´âË*ÏÃaØ•uçs h—ÖVu†x:½Åhôg€ÿòzTÜ•§ äÑÓEHpÓ]§ïÔ\`ø[q<×å1ªË¿ð´ÞÖzÕçø…ãieoOnj8EþœÄáû›µ]¯R¶™´þÏ%3¼ÊÁ3LÏnÉ’á4M·9<úbzLò,3´ç7¸^€E $;7R·A $K8>ÛÌó`Û)êÍÅIç­ü*ÆÿúXà“‘ÃXQ4mM À›hÉCËQù¼eD  R„Þˆ ÇåИõòÏïƒûo \õê¼~{–'ðš2–íAjJÌä5)wû¬pžLrØAQØ”ð/?zÐ!wDƹ 4‡®rO-7\=0.@çæ¼ÚÓ{ŽŒ™ó›ãš°O ÞåïYB¾ÍôØ]ÞñXÿ wHü XÊ W,M?q÷1n4å ÿºòeƒ±åÎ\ÿ"Ç`¶Ë±ÆNjļ¹T;úàCZ› w 2ø”Þ¿Gmä”w7”s6LŸ/ª°÷á,,D¢f$ÏD¸Aá1ùŸ¤˜kF‘(Nž|¢x3cÐä…vµlLŽ¥·ÃûÙ îÃ*äŠâ¸:©þÌæˆ3eä„ë b[”T–ÎYø‚«H»Oj”.f™¹l(‚G>YTRãzy¶»°p ±Ú©8¶Ø4M&@e´ŠVìxxÖt}5–ó|<›{ÔWô‚avâæ¼çT¦Ã\ aÚç 'W|œ2>*ÌœpA÷Ê.…ÀÏÀúÌk&»LŒa·Ö¥±ç"¾­9€þZ‚g†‚€çˆI¹¢7ÒéÅPýû¦W)÷ Vk;\ ¸oòòt¶»Q®vÇ…0Ê!Ô×ÉDÒYñ:ÍÉ|ê”°ûí×&"¨× §fÌ@AMÌ>ŽßÉ­Áxü’£2І@%Ã÷©w‹mŒ;Î/0ìj ºh#Põi‚cÁM üK&¢/XÖ±X¦õ³¾×Ã‰Ž³¡wýŒu’ž-éÂH©D.UFjF%eÔ)ë÷À,ö2Óe”VŒèmÕjòÁ0®±Ü3§hÉ)\îBÿFxjSi,D΂—8 ‹IÜ»3ZT)#Ù¶r­pï1”S¹¬“E ø¦‰3‰9‘Ò,´–•ë†P´ÊªÿdJ"¯w1ô —gRí…¼P5´’’'vܼR ·L®Ï—¼Ø`]Eщ„˜ßƒhŸx¶êÈ¥Äê~ãvwÌŒmt2šô™8ìrÞväöO ¨K3gSUýk^‚ºõÀùø?/µ[ï‘÷hÿkYeÈ\õ¢€*þ‹ø±ÁçT7}À«º@¯ÌZÄ=IC0©Ýh5JyäÃE‚èf pǘ)öì®å‹:‹yzPè!%¯É‹Òž¹Þ Ú"íü{me^HE½ {ÅW½.…sh2ÕjÓT[‚¨Ë%×u.ÒúÈÀÿ©žU[ûvê¦!`ÑÁt¡/ J¶ ·çˆ:9QØ8­’€ÕæpÛ|vNj|,wéw凖J„Zý¼†ÜŠbÖýlKB•L¬–*Û l3áâ•_ÁnT‹ÿw;Z“Øt’óÕ"š=óµur—.†uÈïQÚ\_g¬y»áÔ¬ø®fGÊÀžºvn¾Ôà?*¦#µ&o…„Lú»»š{!âq-ÛE6—¢ÓÑç¬ÿŸÏGCÙXD@ŒA/xBe”WVûÒàñ[qf±HLûÙ¦i?U‡É\~”K¡‚7$2s3ûÒL5àN=¾6>U^pXÈëmgµ_£¢ïoÞîuµ,Zn^Ò.ðR­ŒzŠ N=u.Á´gyÛB•½ï=vÿ@_ªìX°€KZ±üë•ÄǼí“ï …DU³v+)U#_¿BÒ#…L3D6å7Õs›Ø§{½˜Êë_­SîšØíª}ÃlŽõ,Þ‹7}±Â3dèƒíùŒæ%ûƒö×ú”šIþäõ¯(–óªÒ7N$ŠÅî3Ìä(³BÏ„[å([3Pú_”|¢%“‚~0÷Jeê~Óöä+ƬX£é’É =­[† ¬åÆ zê Î9£á»u½ˆi›œG¶ŽŽõ‹’¼µ/£uÞZ¹Ùû~ù›‘€¿ª*èôT.ývkß è»s{ÝçöMDG,Ð3 WÝ­¬Åm£‚oÌPɧsñó5/1qín‰Ú3Tù³,­IÔçùŒÏîZKÿ–b'³)(,÷˜JÒ, º[pï™ÁˆU摲g|¯Tƒù2&¹¸ÃW aÈïãåõȀ脇Ð󤱻ÿòè`•ÑŸ2s÷¼¡¥=öõ ®ÕŽFæ1NUŲ6¦ ¸„ë‰ð2±ñâË‚½Pÿ#|@ ý(Ùð3}8óï©ÎÞÁìÐñ""ËþàÀ« B;¶Óª¾Õ@*ý3Û¶½Ü;Áýp‹U7©PP9ºå[ÆŸ KÕCäô|Ä0ÖH¸_y‹þÚ¿ÐS›n9V¡ˆ€SVqÃNÚ]I,òå¾2ý* ûH©UöŒà­8”Yßo±Â”Ãp0§Ðïõ[:ã®-§eï¢W[µÿKÉD„Lßã?Ð1Gó}+~ '—_%é$€0~PSƒ$ æ r,í+!ïí)ŠP¢á û#Ù`a(W'D‘ ŸÂ;ØÝß‹DÝLšÒ°ö å¢ãnÕˆx ½‘ `,¹BcAÜøZc„Ô|Y.z4$·ýçòžN¶U8 ¬O¦}5[éÃ\LɉKòKgßi­ïýŒê)ºQƨGáòº0ïÔË“^¾ÿË[˜xóNùF±o[çÀ÷(HJ¢Ý€Å‘9*­‹Ä£¶ÚýÆta²Z··‘Ø3*®íP.‚žjæÚœ­¸™ ^ò’0µŽm& 9f¨²ŠÍç÷n6øƒP/>î)Ρ¿ØàUk>'ÙFƒ&D³Ã Ù]ŠäÎN¯¥ÃkO(=9ºÅ¬‡÷ë߉착ž Èã‡D‘ö=?[äPçÏêÔor7åIÁ5(|”Àé`µGú L¾=”L¶¶¤»˜¸ÓVÜrs„j¼Ó$¸@_Údü~Ác†@Âì®`%*—»å„lnÿHRñFÞ׎ÚJÙÍ>D_0Éɧ͊Kæ×ùÜ58¼‰ÛY?Á4ˆoî¥ÌîøBÓf¼Öùº_ûVqמÅù7aðìþþþ ü[úEe“ð€ã†âÑn•Ƥ" ÂÙf¸x.íj!ÝåŽnmÒgpö¯­ÂªW+ói½¸¥í³¬·@®Í©G›Ï¥ÑÀ‘8’‰ò¥Þ¡19Y‹½ “úÃ9»Ôo׬Q¦~O¸ cPà=ŠpÁP‚Êê8JEŠò© º7–NVzŠ£Òõ;©N›õå†÷§ ±Àލd­W‡¾ã ’«Ì#ι±Ù™”úÖ-±iÖ`"óî`_ö¡Oºã¼€‹ùòLwô'¥JV §8]Ñ8‡üÔpi4ýߨ&'kX  h^Õa©ý"ØÛâ‚q:zTSã?ûE4Ɉx±3§Ž³!k €´ üì|öªzA²£©Ožm{†ÅîäbªIùI4n›®jXž 3R“nU.‹‚nb „”–¹­KJ:ë@¤ð#:p+Õµæu?’’ê…ïÅM»/Üå“Ô*½/üµ›ÌÍu—$ª ÄÚÇ[Üwò±šô5~’V»d/šŸHù‘ TšÃS†ŒçÕ¨RØ–Æv‹ÆR{'KãU!È5þÂ$¨‰b˜‰8¹æšàcÈ€ÛØLwšay¡‹æÊË2Û<4‚ 'é`=߬® …̈«¼ë.l/¼;‘§6'RIa垤¥“k…W"Ê W†QɢͷöX¸ÊÇ)rËï©–ôÏæ’›ÕNÚ„ºÝ>ú~g¹­¨ˆvE%„º3[Þ{0ò—ÉNÌ¡’ðo*ØLÉë¹Úò¾7÷LŒNF½)àp¨JmSÂâÖ)X[“TÙèÊyìçuø®Ì4ªdgÙà\ø¦ª§uHØÏ Tr!š—…Ù?o¡¹U3—ü¹+fd%³˜ëŒÐ r±Ô7Ññ¦†Þ?p ÑbÐË—Ô0Žôý¬~Vž©Ì1“CêÄ^Â2Šqb.ϸÛËZÀFÚnè ˜”ì»`1¢œá\$µ8fó«¿û—=aæfZ T`ðÇ£s‘?Ø x¸£EÈÌqIÞUý×òÀc†\Øëú×׋üXIŸK؆c"~wRƒÞ CE0ƒ°™8¹}‡í=ÌcÔ›‰©”x óh¬<xÖmÙù­S0¼ÊŒ`&*Ôc$ÀÑБT¡1ƒ_o3d 2’‰þ¤Û‰VuƒÊÍëýMLtj3îÍH¾NÐ5S3ã F |²ò¦ÁPÿööÖEàŸ^“/MÚßþ'kT\¨WÎÛЀ²À´ó“Ãr敦û)·c‡1]´ëjÍÍ9,£ŸT½47kÕ(NsBIZiA5öå´®ÀjEޤQr%£E}ÐKt¹ZçÚ®ò û÷¿%8+²Ð77X¨¿Z2`—nÅ%‚Y8âÑ òlË4øÕë¶ogÌ87ißz_žÝ÷d±‚•öÅ<¹žrä[îí^>ƒIbÎj¨OÁY]pÔ©éÕ¦”B?J²Ê ¹¬žv@úV–­­‡6{LYU†O Ã_¿¯E5­ÕþÕžGi^X‡õäÊ&@ZÛÖà¸àüáÞ‰Ò4P 쵂CŸºª,Αʭ¤¯B8î¤g¤–*@1OŠO(ƒaÎ]*Š[ÈŠ¢óàé@ Y">¬ƒÓTŒÅ=½ [ÙZÛƒ9¶©ÝJ%?ÔI½°gà0Ýeê­ÙYN¸Ä;4ÓÒOùÊÀß’Ž~$¤4bÒf¯Áz°Tí3~E©EJe2ê'Ç…†‰V™Ïwðaæ1r},(?1p+OP‰€€1hˆ]Ÿ¦9}Òù4€º üÛUEú…M‘ËXû>ŽXPFæ;Ôj¦j>ì¹áû‰¦•¾4Û«vL$¿šàUYiYºÈíaŠº¿¸€Eb¦:;WTš€÷˜{úÅH•9ÏÎwÒ_³Le~J:ÞHp½'dÞuö{} 83ƒ“£W—Þל&~MD¥ƒU¨ÿÏEÉ.GoúíÙTÇ8PŠíí;ÖV3öF*ùµê—6òßcô—Î2è ¡íü#Ûsz³—Iš'ùœu|ÅE'Žàør±­ÇŠç…k ŒÀxë`‡RpÓ¹¢âNp?·ª© š½rTŸª©b'Ý[Ö’=˜\_Õ‘­¤hz›oº¦¤å7½*¬Uά zí”߀³®,ƒ†},mYÆYÛFX.º66„ƒx;c ®-}ÈšïùÃDÇ\‡\ÇØI˜RÅ‹ø8¾tè£2ecô¬C+Ú¤CVÕ$CÚ –ØV KjT1!Ú±“Š[”ÒÁbOBÃ"{ùeïþúåí¼CˆBÓ“bF’&®nŸš×¡i„ó;qè èu—;·ûzzö„ZvÓ{'ùšj‰JzÞA8ˆ¾4k˜‰˜+¡&YäÝ‚¦ÛË]r¾K¯–{ uf¥™¼^ÏüR’ \ãxzƒuHÓ×u(žÇ]ÎØ‚C£·wQâ‹×B6…, CÅqÆ'2¦hxU¶GH(®¯Šýüg íR›Øks“À“a±š(VY½65?ÇÊØqÈ>AÍ›‹@™ßɾYF©'æ4l‚Õ&áDxðxmÔ‡ë\휫žbò ÍCÂ?XÚq^ií@ 2%ÏTÔ%Óy¹ZñX÷å=™:¤ÏÅQ£}–£`ðvæ©·Qž-%» ÝñŒë=`¼ˆúˉ<¯šhÙ¶ ÃkÒÁ€ ‹>_ÅÝâC (¼¸Öˆßc–ÙeA9e¡ÉTQP|IàØðM”SxŸˆÁ8¬ÒühNZ#w 'ürÆ wìt ÃÓcFäÞýóízú×'bó#]¶‘0»ÓJY[­ñ¾Àzb`Ò¼EûöÞ?ð}%WÙoñ­ÏìGzT)î·21åÿ&‘˜-äñÀý]ô »¡u[¹¿H,ÛS!âwB,sLxØvšgÞ6Ú’ââöE%¥ÈDÌ»Óì¹~ Lè$ÎK¯'8·!ººÈ7&‚O&¶)Ÿ³~ÕuäQà¥[$›7þ©ÅC£X‘% è·\ÝËu4_#¼ÈßxI +ƒÜØ&G{E¼Gü¸)Š0ó¦tðÀTÛ6[žŽÅô@ÄŸ %8ðYÖ™EU´¨…pâ-<º¨'šõ˜Lú㟳‹tÚš¦ÖŽ€·f5< —qËÁnC‘ȹNÆ7.Ÿâ_ ZJ­3¿Ê4yžÙ¡˜õoxmÝÜmëèÕ`¸AaÉê–Â~ÿ‹|þ‰u¾·_ƒ`ý­­„ñ5?R÷×ê Æv<¡ï}·qDõ–õ¼yÕì>ɺMù(«h€ x ª8G l5D,–foíè>h¢®±Ú‾»v‚kCŒ(œß/Ä{)fðTGãœ÷㱫Ììébˬ­é¿Ü§aZoûÚŠ ×TÒÞ`×v ©Ü½VŽXÜRê ¸þ£n" kñ±RxýÀÆœïÙCÃkÐÕÆ„¦õaÆ›7Cá„æaµÈµ’ÝDøª1xc‰§srI)Äýž É’ª²¦«àß ñ½i !”Ï߸š°Þe{û³ 7ëÖcœ^Ëjú©è" ‡¨.zŠ¡« æ³'ºÃ¶në~„>`Zç>›Ž½úäF¥'¼øA ^.³Zó©ô þ:p’&(ƒ–•ŽèŸma@s ¯ùfƳ3çÀ-²ÒÕžÒgÁ goŸ’»ñˆ’è”Ù$õ gíÊ.$á¡¶±­`¡GuPÀ!űñZ —GUlÖy¸”óö¢­¾bK”&:‹|K÷Aeà™2ßögâ2e—Ðr<Ô9;Y°²îøáÿ[¾ÆØ?bèÂDxs3ð 6N•¹ â„ý³¶³¾ãQ-xϨIs…±$ü%ž€Å‘—WñÔ>œ—Ôî¼nµ2ÛÇ܇ϓ¾=eØM¦›x ³(ζQÍà³g+ Õ¤ûV‚uÇõéK žš7ÅlØ_n·MI®³ú€y³w4¼xâÈ”ø×œ¬Mì~5ndÙ¥áRuÿ$©Î£ÙÚ‹cx\®3ÁØ5Õ¯Š_–xüÀÃýõîÍš…2Å”FýYò0¥mÚtïý|rþᘿU°ãFòç¸ÐtíýÓ®ì« öàJ;úõ÷Úl웺 ñ,þsÆ›Éjˆnî ÿ ´íwš¢3: ‰:õ¨Žê‘h[A¯—r¤ÑÙw¶«É¼»©¹ä²üs1ôÈ+a` VúГ¢J´†wWß‚9ûо¬F¥ù 9›QÜÃ^4ߣMû“O Æ»A×u¨/¢­êKˤ›ê‰®júõõ)žVù+JÕ¡Š X'˜±çû:«N†äîp©ÞÁ›]HmÙˆ¢ŽýjtüU³ñ½U…X¡ NãTaè¹û«‚Ì>‚V[«ÁÀõ c[5{Ó í2`)Ði×Ù>ˆàÚ(@þö´Ï¢IŸhñSŸ´ønÈ׆¥î'V[Ï•vè~>(úŒ5µå-ÐÄûäßemo,?™¹Ë™ng%vÀ¬gpkÐyÇu/íJZNߟY˜ŒÃÞ¶Ï­?£Dÿ4¢$03æË™2µüÀŸ<ËžMf Þ TxU¸äIfÖŒÂ(®¼lMÝòú{:íµÍ|ý¨6Ⱦsµ”u#´¤¿®•üz„¤ƒãÝq9 ™sÃÃ$’ž=RI¦Þ ÊËe&®À_39݉‘†ê|ÝJ~Š”šc$aL™2ºB«Õej,Ué U/VîÒ¹8l§TßÙ–ÓëÍ[=|1q8­›-Ç_s /@èjÂóàÓý\@€ErÀË`Œ“ÇÛDSðÿÉ•…OnÞÄÆzYOI"`þû¢îÖ@Þ¸‡ÿ™w,á„U2?µèå/@ä—ß[Shla¬»V6|n¸%‹ª!Ì/±-qeÚÁHµ†¯‡U,ÖY”Þ -V[{³-.ô£ŠkŸŠ‹«ÜÔî¸P 4±ª² ßX‚Šlªqf„LÑoƒÁŒÖks¸!W@±|Æ™¨D„»Edóÿ0Nj†Î):©¿¯•õ‡ÈsVÇbœõª;rêskèÌêÝä*¡Æ Z&Â(Ý’¹Ùf‚P®&gó¾M9 h‰W„ƒéa¿Q*eW¨y¶à*Úö< YŠŽ/)·ûÊËÃ?ƒÝÞ;.à`)="DWª›»È·e¬«“N‰}(ræmŵ\Êý ²1ƒ\-~¨$e#¬ú”n Ud©,Ú_a§R2ÇšôÖèÃvi–dÊ‚8*HcBÎv?m$ËMª‚ñ—éó>‹CXXÃI#~…Zûú”ÝlGô—‘î"5)-‘qìØTdÔ`„·ÌgC­~l*—+Ïý†^‰Ò_ÐWÒ ²§pŒ\t£ò‰2~^™ ™Î¯šƒ°BÉ„S]¦Ð_*×^À¢¡ßƒbǃ s¦“6qK²&W¤Öö$•ÇGR#…Í×£ôp÷w™*÷FðmÁqµ3m½zC\ÄTSâ\Ã{¼H±%x—Üš ÉK!´¶$—,U|ßq§3ñ³o”øfb2Èð6†QÃáL Énèè“ %’ÇT"ê½Ræ¹ ½ …¯_Á$ûAß«jô-G÷Ó’Ér²>È4S·{•8Sî¢Û¤BLû©nÀfõjTå¨z¢zwŽ4g¼ðÿ•@¸Nt ÷¤·´¸ì–ïŒ÷&mGTð†ž ]C+zïXûþ[‰«§eT/xhû9µ›"ÙßÕu[ @ÂŽSkS)ßE‰ã ö<ÜÇ{„")ýôFöÖOHù„ñc5ß AÈʵZê;Êð&kt~ ëÛP¼ÖV.A©3ù ^iªצ¡Ò»ÄÎû†ÒÏ{¡øA›Lw󌙞Ä!t`Ä*Å«›Ÿ¥e‚ÊÏ.†Ñ¤as1uìÅDûlÓ«'z¼{åÁçÀTê¿b*ñv hdO Ö S•Z\Mmñ÷aXÓÔò±}N|Ÿ»bö4Vóþmœdª4`¿ ªÈçá!–„ƒŸê„âñÙñ¦¸ü4bb U¼QÐéÊ áûCõ®H whëùÙÒÜ®–,¤³®¾UïyŸÏÿ‹Œÿ±öâ, ÖÃÏïHì  @Î-`·Ûniʬ|-póâ¼Ó—`º§Ò¢¢x æÖ>!Ö•àðeà"Ðî ›ögD‹_Ñ$$Ð'ÊÆZ.ÿ«ÞJEgð×­óF–1Ö Ò˜òI&oñ‚wõ¬JRNÀ¸zo+3óN®eÞ3t‡Ê¸f‚­ªÏ 1‚¼‡Çø•çЄîlÉ%2=ÃHXkå5†.XÿäaóìH8µ&äk\>©¹ ¯ Ãˆ’ ÅOZÓÊ–ÐÅBa¦5€s§Ø*¹¹í´†6dPãYÜŽ îÜw(ÌÎ>E¿þöñ0û‡­>}ÅòR³Ô£FÄɺ¯rårAu”%µ|#b–ßL þð€©ÜÀõoƒoö©"KU@tÁ—3 f²M´$'N£bŸŸ?l7"Sá:¸G&ähÔ&Ž(ês#u廯'}NeÖX ÅVòŠÜX Âw°$TQ¶zÉ:кWRLBñ do­~L÷[Ÿ<>²-èåIÁÐ? Îˆ'ÊU€ž \¼ùqR6E)–^R‡âHu®ÒkŽÖvI_<ÜÛtw­Ô,+dsíÊÌ•FžÂ‚®“p½©-„bI¬†ãºyM®hõ`õN®ßJF€%4SyÂh €fëþ-Åe^±¿‚Йvq½»ºù(F+a^¦¤É¡üt*Ô:"ƒÑ è·ô‰ùùÒäþýSRi¦WËjÖ¼©3"¸šW¶”®‰R´\|>Z( ÷p o™—\ ¾Í®£?T]®Ïs&þ¶=µ,ŸC27üŸ7>AšÙ·.‡1Š­ïždt›x'ÞË—ÊÜÞóóO- ‹nÖæ•ÉžÎøx‹ÀàÉEfB ÄɜʾÓC+„ rF³¡ŸÉQÉÖ81 WB•už»RUïm彺¸ ±`VÁ…ÛK´á[ÌœzFç¦ìo;™•@vš +‹0oa–Rò“hn4f"ú¹S‚X{ò¶I›U™j¾'ÁÙÇ4¼­¤¯í ràEÔ¼;ŒqY˜ï½z<þæÄô½+é3~!ÍäÊúW˜®Få;.•CÄr…0_×:µ3g-ê_k­g3¥p!¼]Lžh’Ø;¼j¥Ò?|+ÂÐñÙï ¡6›#C¾’Õ¸grº¦Y¥2®eÙ «¸CIOwÙ*-ÁÇWªc“Ó–ÂRä(©£kK„†—úÚaŠkNƯ Ð6¶ÃÖÚ£!ýYý¹Wº$Õ¥>õ”b±býÔüà0qÓa†ÆäA™ÈÁçøH ÁóâDf€öÜ—ºb´s¸pî¥mÎDº«üâHÁi ±U`B ‹ÿM·Ï-ûs%<ŽèÀv­nN¡sêSÇôèG€¿_èJørȽ?±Ó³S…©ÔŸVHGÈå«0J~ˆ$ü_CÎ{ÃÑàÆR½ P¬Æ?ÑŸ-deÒ®_tè¤ òeãÐBÇkSëv…>ûv"¶ô?ãò …Òåù :äC‚[ wÍÙ³Y0 ^×P0¥›¼²¨öt@(u ù )S=±€÷ñ>tïCØ™Åhcì¤A-#{€üo N£0~äÝKúRTZ}^ãäÆèæo£¼öûšÍ¥ÁO¦·i ìó=ƒÔ•óM©nwö¼RË­O\ñK,­1zqèÁ㿊Þ·@¸¡ÐÔ˜çô>zþ `Döer<¾r芵†×C]/ÿ•ù1dÓ¸ ]h¸Ë÷‚4ïG8MbA) æ©XÆÈ ù‡ûÎ)G 6~¶q ™Ý<ÆÈWÙN’o➥ê=ÃÈO†N!w£¸›F!P ÖwΙ¼Ûyœ›†žp.«2u@Þó8´<ãT…2ÍJoƒQ«’èO‘nÿŠùö;‡j,6xY¹•è3µ»Ù‚]‰Âòû Àˆ]…¶¯å9æ1ûÎ{D¤ÄYû€¨ëÃõ`å©Áb óëývE„ç<rÆÏ.X`ÃÓÈüo®vÛä)ÎÚ»£¯OßËMü'÷LŽÕ³0êÃNø*íЇ-3šÀô(¢:¶`Wàë{v5xÇß[Y?Bî35û!L ¿6Æ??«Ñç$ 3iÁ :T²Q¹Œ ÆÅš5"ȯ‡”ë]›w–€ìÓwÁÝoî »Qbt•t§r@â@jÚ=/äÎ8Jûþkµ:)Ú"]\EÑÁ½Ë|ÂÚ³1C¿s4«Ÿu_á¡EWðȬ? Ýx§€(Hg=5CÑ»"²}NîoÐÐu­ ì\I rG!5¬ñ澡?ZB[0W DUk«xÁž>ËÍéãO¬Ž¬÷}ið`aí¯¹´2‹e-ƒïŽÏ‡ŽœPßúñ·Ó;‹&±ü´ÁޤZnN™F+§róÖÝ$\e`Ïö¯ù&Î.€ª0i8ŠÅ"ž(¡—P<ïL·Œûªà,,ouóH¾Ú º*:<vìr‡42†˜rŠûôa{>$™štÎó,(ŒK#UžÒòÜ?W‡âJ“Œ×Á>€K½jQOó%ß3º Œá÷Ë6ÙõŒf’¶íí8ÌŠpiúm‰»6rg©õ[gÁ&ý|ƒAÿ.™6›Õž=®ÖéÅ:²`.úƾXGR#€¢0–méâ´üÐãUD™ƒ›Z#” Ò[k$á%æ¦T¿ûæû#"A™›*ï¬Ü§‚W‡“í—w…Çá³ßJ–`Bµ=ÅñGœ‚¤Ïý¬¦s_x   ½UDÓæœ’hkDëå°­ P5é­ÓG¬žÌVE¼¿º'œUô5t§ Î ¤Gð¿÷C1ÒEci¡ø„®ý‚J$Ï c¹¢æbÚ®>of˜ï°7 1BœOÄ„Ø!#}-â~_U"ÿpf bìn5(MÙûö_ˆœï=@úd Þ¾hÛŸû†âªa78ußTw ëÊ5á¯ä&oµ‘lŽ(ç¹òÜ©2œûÿ¨Gè†9*KsUŽ6ý)½D5ÒbA¼¶µÖK„1‡½7Â`C™k^…¹r3Íšõä†ãÜHU°uô™ÂuÁdHÆÂZ§Ã>Lcx[të…‹«¼+ñ/z)"Ù ‚!âž =béªØˆ8k„£Ç ÿÐ;Ï‹è7,€”@¯Ê d‹§M§©Ú¥]kŒ<o~Úˆx/ÅŒxMÍ®¯"”CÝ}w¿¯}h–º €WZxðä)]'èp³‚~¾ó­eOfÚ¥y: _n™8D y-'C7sjGI·Õ6ÕÚšª ðA“¸¤ÙÏ>a±Ð-Jæ¾þ¤›ÍÅÖ~Û¶‚@¯ºö}鈌n8û³i¹ÙK+ÐÑFcl.Sþ¹×Æ $n‰Q6¦®—w,‰OÑKô…VÉ ò¦ Ã_à 34|Ïb¡k¯ª¿ÓÐCQŠcQ/^)© ÿØ}‰A%tÚfˆÃ⤒fz÷;_å6­vÏ1ÓfÛ¾©s;»gTAêEu¤«ì{õ7ŠØ+ôèb£í'ñéúYª!ãUcEù¿_wÀ-^áÄL¶ØûØ`o×wor±éF,>§œn‹Á…â¾SúÍ|Ñz¢[]7ñÉd$ë¸èÑŠ5J'Ž”Â£^ €œ°bEp¡¼ÀmÔÌ©™kóÛFèâýŸùµðíé&Ÿô:bpª™ï3ÂÙÝdñ!­}Yë# —Ió“&ç&`](ȦD ¼JëÚ[l ‡–CjXaÙÃ𥴱shüÅúã*å]ûÖ´µcšvû{}¡+¸™0'¨·SkÃâU<Ú=Y™³@iÆ-$l#™„<38P÷¸?Ä>Rû¼I÷Б=s@1?Dº3‡®Ð fM2½ óro·u¿£~êÅ«^–ÆF ¼˜ì™Gâ áE¯á7r¹Ã v7.ß)È|¢´A±BšÐD`”ÈV³joT–90ÙrQ²þ:«ãyt)lúQv©lÆì“z”rŸ†§ð^@LŒà[S÷Ò¥d©ÏŸÐäû„â>+~Éù–¤o«v£½c-xÍ\òNðÁh·@ÀXÌÓ(<’Csþ£{ ‘E °9ÝvŪœùÂdq—¿)¿uðÓɺØJ™:Û–SÚ^v8Ö±©:Ó¥¤Tp‹wUVpôź ðâ-¹ lrü[d˜9ZyEøÍ›×bo±¹M]ã¥âÒ.V\óbœ¾Ð„ºÄ3¿ ·îQR°!n³zìì‰Þs™ÃoF¶v½J¥Ùþ0'£DËm¥E ²q/É™©à]EOuºÙ VªtkgàÆN0Éû·}¡¶úáäžt+lGL gÛô¸Ó÷ç ñ{Ñwè+Åó76êH6cSÆ1DÃ>šê9b™×åªB£_?ô[›hqkÏ}zOïpÁTm„Ä,ÖínìªÄŽèÅ>°‚ª»·Ä‰~X—Èg ©7r髟;v„ÍÝžô¯çø9m†Apy:j#P®OóÎW°þå¶’ü €Ÿ…Ö"´0"Üžü·MN0§§f‡ëx9®ò[w“¿)Ë.ìã!ÀËò½ÁÝh£½ʲáq­à€ˆ¡€Î¾Ù¨|V°ª¿.á†(UHç1+ºÈÈm ÁØIo³Ú¶ÃÐÒqëhZš=Fn$Ô|îéÌO-/öš‚Ø—”2nͦY¬’oOÛÈQ²·ª„?Òͪ…ʶú¹ší5gNv/ÊjdYr:cË&²j¢8ù¹çmU¡¿>ÐÑÃgíÁqÙñ˜¯ AYS3½_ŠçgNÔ¹å'µé}Yÿ7Ã:FU"S÷i©Â¥½ü# ­Á‡Ú¿ ÿÀçjŸ69‹€åw°× å6ß/¯/KyÒ³’íÏg™vÜB;'ó»=õ)݂ή4DõÅ\æ[oÂ×õ_s•¯=…³³AëJö™bÈš"t°óÝÅÐÒÝ9„{ƒÛ‚RŸÛ+œªqRÉrb<éoÉ:Ÿ _”¹½¾’d †Ÿ§ Bæ%À—„ÿ{ßpÚËóòÓÚ­°§ˆ€#@Øò“N¯€±øÂ-톨…¸}„ãÌ'¿sŠÁ‚6ÛÇ©Â~JÊ:pgôMú38N ‘2=í†ð£Ç1¿2»'aÆÀ°ü;wíŽJª±÷2Ö¬3rÇ=8.Ë1&uOYJ9PôôÙê9ÅOäÓvYI絊bêvN%?õ£’ÂcMBt¶ÆpZžøÍÀGœÆâý¶ÜžßnHv”¯Nã€PyMÊÛÏÂ-–U±ŒÔ j¾;ñÐGíCäNfù*Ržü÷s™$¿ÿ"0c)Y•oéчuoZ ¤Ú¦9á™Hb¹zè%Ä„U‹ÈÄ||«p°¶ð¬È»úߟŬ2;¿>(;u€9Bn|Ðà\Øj׈"±³ìÞÆ}maˆPyf´4Î^Æ_X£‹ø²³ UÌØš¸G!£j¢+À¼]»àiÌÝt“$¼«ÐÂfÓÝQåá&S«18È_ó˜*&?•7)|ÉŽIqÉ笄†h¯Ž4,cžH.«žb~? -j6 ¨²¬Áÿ@f‹Z¨=1n©éZ9f£ùØ‹ƒ1l‘a–‡˜«¹8–¨ï9éS%ŒsDÎâ²ïdžØ^cEw¥yò›ןªzJ6B9Û…¤'×ä°Yµµä³8#¶[Eü ¬ÍÞ¬)RRjÿVk¶Ö'ƒ1¾6C!rŒ;Ò׿G$—Óx§jçWžä‘\J“%*:7ƒÔÙØwÔ™.±í ‹ˆcZåAð¤I«¹¨íúLÄ.-3æ ÏÁfëy­e‚Å€xð_'—øe4g#E êÖQü÷ËfÕ$uÈ! u Ü*¡ÄéÕê â®p]‘‰ÅHŠS®YV¶…•åÝ+îïã¸oVïBɧ ôÞ$*^Üw)ðºÿX|ÅðKw#¡Æ>ÜWׯ…Z!­y~ªh‘E>˜e´KEÊ›àJþ­ì ¢p1šrDw›\ Ó¸-¾ª¦±Á/<èzf@œÀ3Þ> Á òD¯ N¬J¡¤/V­Ú&'^áqlÌüf[øê%óèîê’ƒ™„Á(ÓžS[Θä3Ý”íí@½W³¤‘ЗÉ!îãéªd+OÀD •ŠZØ¥ª€w.xe"šñY¨•y¹`ZÊf¼j«%ÎõŸ@B:Ï Gbp[VLsŠp¡;H‚í ð¦Åœ4TF3êùüììf""ùÍmóÂË{ÎÛ`À°-ûëUIêQˆiã‡vz(z÷ÏU4ð<‡™ëS3d^ORi|Š" ãˆl'ƒÈB.¹÷𿢝~®sÆà“vO(\ʲؚ[8ïieÙ£W8èÕ÷|ÛfÝt…Ýø¯½óªqŒ7ïËk ‘zÒê|ügçÚúùAaÀˆ…=æ±äSübÒ¡D.f2Ê60Wmdù-Ö¶vgÒÕ›Eß5ЊºLùÏ-Æ4h §pÐ9˜kvuÕoɵž!iúÇ›º[­,9¦HÀhø^½Ö{ø6ì›”‚BôÞw‰Ú‚~tâð3_w†¹”¦¢ì½#N˜äˆàÍe?«R»€Jë¿ZŒî Ç‘m7, Ãy:¨èò47‹áÞ—‘Z°¸1ÙÊÇ£Úêu¼ä]'~ð-ÂÒ\O9Ã9u5$Ämº!Z¿Å!~I‰ ¤) R]$úÊüKLRÔ)ͽJ/#Lm]g•§»Å½t0þãe¿½(g•Oé<€ÿÓ1f„¡ ùÀ°Ñ”„ÑúY$Ç:rM»ËDÛÏkÓíþ¶GH’@u©—21pÅŒYËÎP…Yù._û2 ×xfmjâœX]£ Å´bëŒÖcõä£ÊË>ॶNŠÈ½)žÙFÑ÷cêy\ÛáÁ*Û1&k‰™ZrÖb–Òq?6'°«£.ýà ¦mÅü•54¤6Ï:Ê„´ÕVM…Õ>+Ò¨Î0ÿ±^3ÌR÷ÒZ”âjí<·š£úȨðItØwhæ9ê|y|qý;÷“¡ŠkªW fàËM÷‹·gº"ë{Ëæ²ô›×àÊ®¾ŒgB£XˆaÊ¥4Ÿ‘¼=m¬HžE¢§î ôä%¾Äˆ€+†S<=â†^¶GR¨†Ø—s u¸4íZøAÈ­EuTµ*³ àÐM =k»¹><êx#tÈ~ •[€„“]Z‡1kìbnçSØm022:ë/ 1^`ÆçQ.‡§L]dŸÝ­Eòý¥–MEØ‚K¥ýR„9Œd1š©—½VˆvÂX—EøßoÖM†0Ü5Ñ2(hýá[[_î7B¹é1‡¯±õVì 퉃եï¾`ödø‹ž¿¼–sÖãÖŽprúÿÇ ƒV¾N£´ñ,+/g™„’!#9§<ö®x_0£¾ºsë#æfÃwøGŠ8bâ©LÏ_ D.4»Ž¸èF'ðj¡›ï0¸ÌÄIº$®ĈÍÑ 5Ý1â£!Šþ¬-­I.Bå£vîßÁ÷'ÈZÿ]ÂŒ¤ˆZwr&¸’3á¹–tJð‚|DA{4—:róÛ·ñ|’Ú„!l‘úcø›;ç”;u3Ã$Ðâ…¨uÞÇ´ ë7›~âM1ý£Öñ­æ614ߨ ô"º˜yLÌ Vµøl²%¡Ð݉¥Ú”/íã…9-ôc(°g¤mH:øOjS¹«ü5ºþ&>e"4oÇQ Pý$y|¾#8®!‡qüâSò>žµ–¨ÍB <»eþÃHjá‹«ó¢¹Ài öàv?]M£wÕ±ÿ©v”†”Ë ¬‡Ò£IØA‡oL';xÆÊt`©Gʦ–ˆt]±;íˆý¢ý¶AšGj+Ȭ¾Â$˜?¡éMè€eÝA‚Êe>´ÈC³.”2<æ/D*ˆ<€;žßŽd²ë苼èlÖbwÁÓöqõæСÝîªÀÛÑX±;N#é³$rÕB­¯dxC"ôàÅù¯ u54ÌT¸¨}8Å‹¥µºÂ5ÔqÉú |™ÎaX_™×a?ù+&‚á#c1öC®‹JyŠœ:K9|„Ërë›멶(†Î‡^ÙYFm72µ)jß0§­Øí‡#›‹J z6¦@ÞCDéåme§êaþ]Vzùè˜ÈP7vÕ€UVž6ë oÅ›ö?1šêK`ú§0tÔãåþÍZ#¦Á}" Åä¼ æ:%Kà”eìHѬr"“”+›Â5oõï;›†Ô|µE~ÍŒÒ,j dy$Ø(öý¡©ï¤(¾ˆLã5•.ýÒ²ffIŒG±t/Íía†aŞΩ’sónò×”Ÿñr$YYéH°F”,^!ãtÂéòÙžnR]åøóÕÁfWóÏÃÆVvOÁ4œ¦@¼hD+Á=ðôxžXµKÔÉ…¤»¡ºÎ›Ïq.†Ì³L:êŠ$Ø‹­aL¢›Ù†È2Wos½¾6l˜e#´Uí²€dÃX…òp£Ým4™ÏŒ})^˜\£aÖÿž–€Z i’‹ÕP‚G9`šF€c;YQõºÚÊyAô¥s¶7 ðf goPÁ©¦;"âPSˆS)>ÝSËò­g[÷õý7¦–®&à"6; ±:| l.‹€g“î2#=uÔá! Ì%Ú&:¿GÂÜÈÉœûcd@”·—‘²¹%îŠwÙŸUôù–åoŸ¯ÑmÏmH‘S 3`YBýç̲ô*ˆÆ’G:]ЈûÔã‚,*Åäoªmi sÎ)äFl꤅®¸-8ÎðÀ*B‚aqܬ ,dÆpÀy52ž°$•˾|‡Ý6ÿÝ9cîuˆ>Ä.÷«D‚ˆvI=Ýòê÷35Šä±áG齬Ê'nÇaf4«ù/šÕåPlº8ã¤2ÌÁeÍÉ© HzS@‰kÊt†ê_RŸ6Ș'ƒüŽEåJRmdXº @àC~ ‹¼8¬nøŸ­¨õÊüÁˆÆ 6ÔÛ_–Túí!þaÞD?È–4„0GJ»’`~ö"· .ceíäÖå”ú_ím7ø$!Ú„ù*9žC£¶ÈÎ(©†yyÛËIf¨øå™¤¼’zCMœK þñ'ÇÕ¥†W¹±Ö€Þšzq;êÚðäê¿uÝ$ûä/®ÕçûI™:Y­g(ç¾Óyù<]GL‰ÕÛƒ¡˜!•ŒªüÖy÷d»…UõÆr)k˜ßö–ýǯÃd)½&â69V&©ºôBþ|)©RÆfÛFŒ‰Ýaè;ؽ¦&B„TéæÇ`¸9’ƒ‹?YrÖ¹"D0àx%‰¯Ž“ä5”_ˆ¤†àWŠÅ.S–5Sð/CV'œ¿Lþõ óÍ9¬‰£ï¹Â@_“Iîà–gd‰J“ZMûÅ߀Nªî>‰ñÎÌ{ ªJÈ/Hµ‘m/—½Öó³ñJ¤@ðÅÔ©t¡ÏÕù\ÞƒQ˥ƕ¸§äá«,GOKͽ¡¶õj¹îh(Þ6ùB<ýGxmù`õˆa4cKà kı-T²žß‡ÒéLa‡yÒ0å³i¦›UX‡¨“)U\ðn±¥²Ön’f«y÷aÛh·Lºl'¢ûÉlÀêÚvæóW>cqZ¬ØÂWúëo5¤<†ºkopðûÍ ½J¾§çWCÆË1‡ªž°á$»×—®ý¶%pö-@þî,»®HÌkq»AÃPµ‘Û—~7cíýãHíÄÈ.‡[^IÁEíù²mÅ N¦Žà] x©ôØ:ÇŽ!j¶©oséH Íñ1:-£]CMS`Õ\»q¼VÙ’„õ=YÀöK10⣒N‘A½%À$ qʯY–X˜vb’µþΜ¸ýÈȳ%¿ÇäF˜`Gë2–T*9fÀö ù6DIK':øZck+R½zažcš<©´ÿžùö«Gùl´fÔÆ¥&"cê¬dAx˜“ÜÆ~~l’~ Y)Íù¤Ë=,ÎGH1VµÌãf«~jnkÛ‰â0s6 ¾6Êv¾6¤Iƒâ:¶FæÑ6tXÄÅpêÁضJ¤×Yîö3ëøÁUÉÞµ1~{ÒËPrÛVLƒH üžZÔŋ޳6±OœŽ·oo™hƒ˜ë?n†XÕ¬Jw.Zþ–áykµ_÷3ßmÿ[Gß#>——8Ÿ‚D–pùÀÏŠ×#” ›ÜÏ+/²¤!ðâ…I„ñ¬ ½Ôæão¯%'N•Eþó¡Ó«жìhÞ߬4VÍÖºÁ¢éÔÿªáGÏvtÉ<ÛP nFÈ5bõ¨Ê\’›Ñ°C|¼”Òõ¬Ü¶¿„òN±Ü G~)œ°#Yä&Ed¢Hƒ0·& äÝí@ÐV¾ôþŽœzË&]ž²Ícêlà;-lJîðÃk´›Ê1r€ã Å‹-lŸãË}mƵزƒ²U²˜Cs˜¯s«¯y¾^ë×õl7/ØìÖ f-™ö:/ÓÙUfS‚ÙÆf€JüüùQò;ÛæëÌ[“à0¯‚è×påìVÐÚ¥¨7Å1¶Ü„î û…ºaO·Â °}¶Ì¨€[ÃòûAý8hsз•2ÍôÉ–¥á^bªUÎM<ë¼TFy‘xi%HØÆ*¨Ùµ¬©š©*T5·fÆ;y2S™¦¿Ê^ÑÒԇ׋óùuåîÔ‘Bh¦ývüm‹©0È)ãcœ´‚Jyû>B«úc’ýY©H=Di^´)Ä9kWý¯j÷çH¯5|­J ÂÙi{©ø?÷‘n±">Q®¸íLýñ+“ù¤×y‹âÈAm8Oóë½Ê]ÚSPo_G¶;ù³_™þdA@ÉŒ¸`†å°ûÂÌ‚ŽÌìeJIÙ…6Z¼G,•* Žˆ †gÝtø“ã]zÍ?sU¥uò”*c~O&kWB aIÍsS“~_{1êînC¹¬$Õµ›ÃWVâ¤6®o"ê[gq|ÿ40RZÎã†jYÐïƒT£Á&NM9N‡Ý Ñ‹WÈU°z¨ë¯Q¸ÿ%±GÙÛuõŽV<µ"Pžù—JäËÇ9±dáü)Ã`ðz'χIHìÜEŒñÙªg9ßš¸ýk0G¡t.ÂÒ&½Ë±fÖA$½Ò¬.´¬%C ªä:Ý›Îë²-òBÍWd –Vúçh®…Wöž€wì \¾Õ,¥jà–¥ü¿{?¿•¯oHºcE»% ³øF†âlt…ïI‰Ó•¢¾ª†G™¨æ£Wö^%3äF¥"œ¹4Ò×¾;‘­„ã@Ò{]#žësoVuŽclÑÉm9•÷J,I©B|ž ¶4  |ˆq5tlïEÎÜ”V^ßï‡yü^>ðÇ`^ªÔonÜINsƒþ°ùIÿÞû£`Œ:9±4êÉòeÀK톣ùL°©6߇Ӄ’cÑ€/ÖØvÃ)ÿA«?UØ­yj…`êJ4†fý©BêÌái®ƒ˜f]­7‡Æyš§ºûàÐ(Ý•ŠöÂOð¥àÖUœíf)„*A0uï-ˆûX-b‹®º£s1˜ù¾ þ ,QiàÐjþ£Ø)ù„µÏµ-Päòv´Qê€)Óf`R¾À1UJþ@´š̰ äB=UKmNýtÿ¸ ùr¶Ú&DÆ‚³ßÇÏnID$]Ã}E3i”S¨§Š¿šým¬wÑÖíÅÏÎús°u2¤@¿ô­dCæiÊÚWóüâøOýŠ½Ò².ËG.Goßv'Ì̼oýI/L.žŠ]½d*È•ë ºþæ5=Ë|]5Ö½vcÇÇ0„VFn4—ÒâtL?R· 7~ÜòbXFÕ 4þŸ‰Ç`’]—ÛIÆ{OÀ¬ m¦ÁËàY€õ=7[qÔd–$?ØK*Éܧ‡¾”0 Ñƒ,8Í(¢o*¾²ÛÃ*6ËÚ!—0— b¨ÅxäÊb¤Õ`±ƒK2ú4;tƒPQß ¿$Œ(h«ü¿ ¡Á>¡ìxñg? *£gôƒðõö|5,÷Ÿ$ËÖ¸`\ã߆#¹M²¥Ëܶw·BAÂHâìÀg».­9¨ÐÚ®à–òÀ{Wp˜ÎSÖÎ]8ùã/¤Á¯Ptøj°5ÜòžÝ±TI¤½d”Û¾†%È[ÌkÃ2 A} àiwü>$äv¢Z–‡$ò%Þåáï ócv'éY£;oh€q ¸^JøUµù`RÊFçƒPü '*áõ dWC´6†$ß@\å ÁšÞèI¹Þ”«–¬ PF@Šyw·oźªV‹#âB¥3cæÔi*†Îþ y+Ó6 éB#XŽJ/¦»áˆ·> uÓâ8î¨aØ”0à ß;7V­°\ï„>†/pϤ¾ƒ âʰu± š­¶[•LT’ˆ´è>NÎ)ý8ÇÝgpQžúÙó¼Çz«½Y¬³ïÔû ü¸d³M°8~þSuë™ñ)8~o½µx 5dïAŠ*¿•ÍäÙµ(õ£±‚>ù6åk¿¡nÝ~É´­ß Èr«7æo#ýŒm$·¦èøsšƒCvu{b Ij¸RB•}âá5®¢bŸÓiN™Š5ïŠ2ý7×ÌW"ï"å’O‡ÝФ¶g¥Â-z‰c/:tÔU¥G2¼r6Æœé}À›X}9øÛ2Ž‘ÚcæJŽq¤I){¼Dšö€ˆ-+|ûB–‰•ÂDa¦szÕ$g(àq8ºpú·ıNö÷À.øs¶ßLHcM âW±)K‡ÅR]ôòT]š\Ç"m¾¥øÈÓÀãøÈÝM×_`•U´äðJó¼µSp” ܬ[ÞâùåS–K¼ÍpÙO§ú°8%½C;j«LŸ Ì¦¡jJ1Õ¬)jv¢‘Ðt·o ³¿’E4ñXs»e,µˆÊÀ6•Ç—èû¹mmP{çdXÔâmv–ƒq÷>×&àê/šóå1r/XПÉÉæM“{)šm Àx“jâÊŽî]ºý*]DÑÌØ,ûç-éè-ö„«æ|¼áNr@ñˆk(ž>{ +ÖÃ{s©K˜ìO-ø7ôò{W4ÄO‘è|ƒ%=Aug²ŒWˆ}lœ_²ùTç§2Þ„Æúì:PçdùÐRæGSnWÓÚ—qCÚÄÿî}YB±Ô:ˆ%‘;ûjñ¹¸:#Ø–o[BÒ{1z¾k!ñ8Îû<í5†Uz( –X£Ú$ŒÔÅFtÞ­n`ÑëE«g^ìЦŒš¡tO…ö’šóáÇæâñ ›c!®Ãà5cÝÛ ÷F€J6Ñ/m»Øfl–~ Ñ…–X Ì©Ï ÑÀ[¬Iî†Î\Fp?& ‚Eñ4e$—σQdeWˆ%{¦üoóœçåH‹RŒ2R›-~bpIÖî˜.'¦Nt>–¼w%®êÉÅ'|k£$o.P• ãïA°¾Ó7/>.‡|ßù‹Ã÷æGqœûàzÓU(#¥D8‰>Û•¶¶Ìy¢:2 #cò‘þ?¶L «P@¬³ 1¢AÉÖaz˜nÈÅêx:Ïe ÷¨¾„½¶ÍÃÆs5«å±õªE&¥ÙEL…ÇsoŠ¢¢š­*É›n-9¤Uèu®\,0_ôq;dÐãìã0ý9C‚¿Ôh•‰ŽœãD9çì]—ŽGðL'4rµeXAé@Yg,4[‹D|c¼ëgñÉڅshĴ܃}ÉíÎÍj™âqvOª_¥,Õ^7f¨ûb ¢t]jAÍd•ŧøH×WL™ØGH§»¬ qW4—(‡÷r¢'žW{ø³½dmª@r˸·1?ŒëÌ/Ù¬ AÞ6§ö¢“B°R Aðµ:§ J€65ÔÃÙèœKM6Rxéur2Ñš£E‰Q¡ÓV™2ßL?HAç"rS©Sp=[Wÿ12˜ÜN×Oªö|ü(ˆ„D©¡f Ð<³:Å ½ #‡¦Ì Ø!míó,ÌöÖ…”1a­“ÀÄ(bÐ,£h÷Û4„÷b‡:YiRFÃÎoÊ´Cì¿H4B¢ƒ,ùw°2¯Z;¡,c Á@ÙRŽÂÁéh×§I‰"Bk!7þlb†Sq@•ÐÿDü´}t,ù(V K~tF«­êŠíPcžÝÔ:_¡¾š„‰’òV²îv_%|·—ŸPbà5SɤOíH|QÐôC_¡¯—4 ‹gdŠÀT¡]ã”ÿŒoÄþ¦ÅnÖà áÕ:ñrqM„)=}Fƒxv˜R¸”Bí 3}râ0ÃFµÛuCøë9mrlCP&ö}>ƒ~²ÎyØ€òÎ"»ÅÔÞÕ ‹Ã¼t$žB\jùÙLFˆØgN*|ìÛýMg•”_uöGŠ`D“"eP­ 6•…¿tòåwp±¬+H¤‹ÚSB/;ôÙ~ö¤¡®E€¼G£ž¿HÔõƒÄ (Ëd'g× û“z(r (Îm0×N Lç%3ÐAlo}Úw¶_¬yö"26¦eûóý J(N–Î Ù‘ÜA‹lH†b"˜t¥ÞuŸ „˜¤ñÄb`Ü¢zhp½„®ùéƒ&;¢m17‚&T;5­îˆû;„"¡à¿[Aòˆ„¦àÝ'üû¦aÿžœ&F©9¥HØÑ—ƒsqÆ»T»­À6 „¸×¥9åÿ‹g/ý^A9YLžÑ׃ô»çV d»êÌ\~[y‰Ç÷r oó{,°'{Ì[@¢]Å4 «³ -:gÿô¶êöóSÀšlNä¯xµÎë­«6`S+ÏU„‰ßm½ƒS„Øwiþ^c*+v„ß›í*Íþ”»Œ>OX*„ÏÃ,“zµº´OÃÛ³3A d™ñËÕZ_p AЦ^K#’¼–•/k2P*ñGM¦Óü/Á€ è߀‘Õ§+í’u¼i)µ§K2t ÷ù"'lNÔ©,ì¶Ô¶Wã+οÓ=n[y³s¾b‘Ò·Uæå)Éó‡]|Q´çÊxx¢;ðu ©3iÜé@ÚUP&#¹ öwS¯]UÙ 9®Ëx_«Œ,â¿§‘ù©åãØÎCñÍáß»ò¸×ø±Á‰x ÝÔz~ÐU©)¾„“h…zCE!7á ‚Ý_Ÿ_~¿ì{4†@Å%ºðmË ëâOÔIRÐ;fÊ ®uS–èoæ3ŠW ¶}Ÿ@ñ«Duc„±,»³d„œ®=:še 9àE›×6­äR •Z] ŠÎìÐÀƒ²b.%"ì(ˆ²û7s0xjöÊC¬ŽmíªÓ¸ƒC?$Hµ, ˜C@lŒ}®Â Ëý“q}ìù­@ÒÉÇi[ûÒ #S¤‡:H‚:|‚žû ¯kº1t¤‡#0Ä¿²õ)°‘lq y”é;ñÿFã'eíjñQˆÃšpB£›…¦]‡&SE_šdSdž4o«„ëJ£ák°¢)úy#¢xÐ÷^‘5tYö²Ò´}¶×7lïkÄ‘øÞLdqO!ÙX]§æ6×fàÌg°1ýO °™ïÝlÜÿ’ezvÈ‚MŽñ…*J‰õL q$é·§þ–ÊÁ„n‘¤s0ÒýÆqŠ]°ÝZAÙ\ͱÈõ&—ú·?AìG×åjvl¡ òuFùârŠÂB­ë=×™’À%L}V3h¼óW^Ù`åʃ^NˆkÞ‘Q1ú˜ÁZ%Šðøw¼Íè€Jâ r#/u*+ H˜®ûÍGïÿº¹ùz.I7#‹$`-X¡ÎK;8ä´7I®¥ÆzAÁ!‘½AÍûfÕ’|ëUýªñ¢ÐtT·í~­k¸.üÁŸm~p,•§°éŠ6ýwègÓè7³ªQ‘•7ÚœøVí|Owo¹|'.øé‰?Ëã­?û`ŸËO†×D"¼3Jj<ÓËé+µª•{ÁƼ¸9Q ÕÿÓ$v­éþ{ÐQy„Æš‡–ŧT¥wkÕaÕÖQËÖ E“Ož =qi{"¢@¶Œ/¬9‰ükááH¨j_Õ^¾¹ásMo‡n½X6ih>§J©3ïsàcÃs;S À†"Ú§ ¥²¨¹ÇÇ«nÅ?œ­7“f(zHAk´=3‘~gc¦rëOß9È8z`ÛüÊà:Aš7îú#PÛý=ÓrÇ …ÂÂ[fx}’’^ 0"}@nÔĬqA û*Jû-™nlÂ4†‘¦²ÌØùÚ}U®šxè0Àìò‘ LÀÛû9`ÔÕ-ɘìT¨ªC3®ø•³Uºšú›²Q "™ ®:ðï´Ìü?›C`Ž•â$zw(;bfÞ“ê|ù³*-;‚ |H”LÛiCnÙ·ß}ˆxásê¹h ´.•ÑvDÞMe^éÛ> zÌ6Ä$>âC¤ÌY ˜Ä\©ÃU[Pû´Þ6Sn—O3ш›ÆCvÕÅÆ91q˜jo™Î0õWŸ'Áò0îæ¼âø^HðKIVñ“N©£¦%íx¡ ÈT¶%ÄÚ°":«Mv›¢íWuÞéx:¤ÖÀê°<;ªØ ( ̪¬ÄC?nG÷,#Yöf5«–xß÷7üJA©œ t®_¬®±MB„ý_…æÞ*û<¯U’—èi<àsðZYĨü –³Ô³šìÕ?0x¢l·¤‡œSØŸéU¬‘y{A³Ÿ$åÄ^ºçY8ó; ð ~Ñô¶éùÆ)¬ øû‰ÍuÇ›¯ù˨ðéUðj`>s‰ y Éâ€Ç&  J0À‰UÞ†ZD»$®(OJÒŽ:qî&ûSã4ªó¿‘¦CLŠéSÏj Çdü5žûÿ k;ü5ªí†Ÿƒù:3 nØU}þR%…À¬´¤c²·¾x’‘"X&³ØHž ä½bÐ*í½Ú ì"S'siðWßI2xÌììòrDz!O=`Šš×lvă±WsûC æ Дեú1«k<ÙZFà×+íE6,Ç:Ó1{õ¯Ûnœ,UÓ‰+Cc©¨ðNWœóåb£5ÊK\Þ®Â.Ù¡:9Y`D”Õ¹Žd°$ZXØ9¨ìóõ4ìWÇÐ6³è§9Ÿ&Eð<%Èـ̗ª¹ Ù†Œa’µÆÝ7lp C$V ?ödå³C‚–ªú<õ$r‰'WµÞ»@eëÀ©-ãcZÏ”lã|Và·Æ'ö‰“tI¦]Å(fU#ä¦ 5-Y‡@ƒ–;jôRrŸ44KÒŒbÿK°8hh®¤q—¾´˜»çÜõ_èrõ©Ý Ý³¼úUT³å„zŒ—!EøÇo«Ô ÷é5µš%tîÙC1 DãùvHŽK‰†X¤JT(žVðM2³ÛX™ÏÛÔz’¯'màÉðCÏr®ƒÔç¨+BÞ2ö/çÔy¡’|Sšòð1‰Ç UNöõu)Þ†ð¸8‡¡ŠýYúsÌiL2„ˆÕd:@ÓH–ûԓ̱ÌyiÍÃ|Õrƒ¶$ŸýÿÍœ~ù¯¾Šuâ:«,ÿªÑdKÛ괳/ë¦w—@©0ˆÕŽªD"Mîlé½û ›¾äI~ëš$—+%Xän ;õ*oç@¸ï£#)_ )–Ä—&†©÷.XÓY¾r ²F³Ú_ wjvÛEà=e±§ ]Åìe'à^o(ÿº§µŽ·º¾'ºÀ0qË(ãA2±ŠÙäÐDßo*øémHb “ÃS‚BjªçÉwß+^£Ž kBæ7¾ô„º§'È6ÙÙ}TGBÇ5Çß'p{©*-¤Ä/K õÇ·rU[[Õjl|1ïû¬°sÝàa%•¨tY™j|x‚?Å–Hkâ*Í<1ÃÌ¢{ÓE0)°iZÎHƒRÈIàe Ö˜áÉq÷ÆvÛsƒžážÛÁ'w½ö Õo¢ñ‰×”¦ç9Øw-nÒ?"¼³ùËøšï 5ieÍ䮲"#­=˜:QhúÌ„X¤›ÞÈFÕMNwm^ üä%´Ilüý‚¯vÀüÀò|Æ“v…(jþÐKbKZï¡PlŒNúdËÝd³ÿaôúÕuþÇ^:œÆâÑ0ôrBNØôjzs 4ã‘ ÖÌ|…¯oŸ=: ‹Á#´û¶EHK h~ãºÆ®ê)m²²6Þ'ÞÁĉôŬ\5bómüºµ¨õ€„{wîGÃÿäú·ÿì¶?“ß¹® oúBÈùñ#Lw á¤øÛÇKö¤ µkn½ºXûe{&Vð¨ýÖGvŒ2›g/ŽÚ5ö”+ù¬óô ç:_ëäálkñÿ¢Î„i9v¥ö*ïEÉÖ‚º^”óRwÒÇ´Þ­tJˆmÔ‚üó¶Óm•ÄñºXÆ”p|BÕÄŸêµ'}CLx!Õ«¸à´¦q_ °*-*À@ÒöÃ|ÂÄWäò–Fæ Y¢+SCSËx­šâÔ þ&£¾"YeÙûŠW4Ìñ$”Ã[d ƒ¤yxC§ÉqFó8eåžÌ|9GÜÓµ¹>Ÿ<èXŠîö&tZU ÈG¶zè\õôSѲJ¾#@ÿú‡TõJº^ç–JÄÆU©¡žòŸH M8_Kdܼ÷UQ½4MÑ`ƒ'Êæé¼®0¢q¨öq3^÷ÙTŽôø§örÆoW* (•¼ì=,õ½‚ªDF$Mlò¬oQPÄxàÒšÞ™”MŠæƒ­¶á„_¬2ÁŽ*8û'£ -• A$u)QIðøä‚gÄÆ•ÀGµº_=á³f#¢›ß·b¸ëÊq9=]ãõ@À.-„˵u¨W ÿív7Û•}èTÈwʪžéıÿ“ŽÙåY/OkûÞ0H±¶‘óÚô(â¤fEÌÞ‘p‹C$Ko…h1Bmöÿ¶¹«t×q®Uyúÿƒ€ÆC¦ðäôÄ ›ÐHxŒØôõåÈÄ1Ôw•[Â3+9Å3Ò Àl#¦ú´‰,Es}é§È"¹Á­ ¸Y7 àìpöVl|Á‡–8žjñTÒ*A©Hl8©e*:§B¦qÍÈ*U)×l,¡ßU'G¬•à ç+O"uùܾQñp°æ£h €"Öjf3¦ÔÞ %{çp“jߦ‚OàÇM>”šš³»dŸ£o²2=g/zXQu~|ð–ѧz}€Álð‚¸ÃÈ6¹Ÿ¹rp…ô›*Þ"þȧ1chµ5ô`wBøˆ˜eê¼—¹±œ|%^®ù™éKoRD<µ»ÎúIQ¬ÔàœöÎ3oo•÷ó99Õò{ÿÙlyÀ/&1@ àN‰›Û¶Y ˜ÇTÞr…Û¥9‰m2èCÚx8ày*¥V¿¬,mfS2enˆvëM ÍZ5¿‹¬èÑ rÝ÷U«ÔëR_ ù9|ò²“JË~JK~©ôïþ!@Ví’¦a Å™¢jiC Q<ª àËÓ’¿ÚÎ×A¹Ž¬öÁ¨Q_Š™, &ÒÞ“jé÷d€wÑ¢œM®^,pß\û’™ç‹6 SÉE »R#aR륙8®­Ñ.Õ×íÙ„i{é'$G-`xcå"3É462 CŽõÐ[?>~âQÎw¦¡p˜ÝÆÈâ`èg)(j F%f{el¡°¬fè#RÒ\­›Æ A¢èžŠhuèdÍ|/ìö #DJ¬\q»é>@5¿õ7NWyäâY,÷‚hÅáÒ•ïÈÇ+jBhÀ~"+9>&Ó_ò“)WŠš½±0ãCÅY˘aýȪÂrvꔸof锃þæâûd3râw®Ÿ;eÉÀµ èè·c§3ø×Nû×eëd¥ŽU lJƒü²fàž% êP)Üò|t ÆJ‰6 !˜Ôq¸[†T¼ð„[:ª¿ê~ͽñ3 R¢â¶…¢“mª¢¡µ«»”`¤Pyüºƒ™Šhö˜Iò@«i©–>˱í f?°yŒ…ê¬ùoȱ錞¤°*”W_ÐR+Þ\Ì)µìÔ‚ŠCšN®F½<—ëϧ]mbzM;lþÚ÷w/ju7 x¹%$ÿp È`¤$–Ï{Ç º±®´8ÂO´:_ïñSk„®ßˆ¡P»’Ž÷-É|¨?‘¿NXÝÏÖjÒë5Š0KÙ ¬Ã‹XôŸÿ™!ˆÕYž+ˆw¡ê¹˜äß<Æ«sáDç}廌áw²Œ‘>#DýÕ†%—e¥,Ò.Ã!M.(ûKˆr“—a(*g\Aa‹¡êùk Š ÷. wr<²¸þFç™{©¢Ñ™œ,›¬dÍŽ+]³ÈÃtg¡½aßâÚÊ*ÀÙÌE§Ï£Á°ðEî$û SEóhow€`ú9;Iõ{.ðŸI)ŠÙ­–Äv8òEÌóp¹Š®Åíàª{¢Pÿ©HKšÁx®t),OæêSÓ‡—áØW2 –È.Ò•ùÓ†Xˆ}‡µÏœnvöšå#â"” ØìœËÓSêNfŒBÉ‹ñs ÌgŠö¾òw,™Ôíe·Þ£¯¡ú?YÊá!DÂÎô¿¡ô‹Îp÷ÊÃÕUp¥Ènø’¹TšÚ½¿K]°@šõ“L (кØ/Pzб~¶[U䪭ƒ_NEÁåµËéÊ—?[vŠ30á‹æä÷$+. ÷n%Âû¢ÖLøàE÷“äÆp‡{N+§…ØŽ@96й×AŸs¼é 뤉öpÓ@ÖLÒTN®ØÑ˜kã&i{s 6cÎà“kÍ p¡„‚7 ÓN•ËA xx‡Ð”ÕzW³*[Æv[ #/ë/>•‘ˆu¨˜º¢\ïjD÷ÌYX"µmnÍú$Å×\Jì®¶ôìëâÒPæ‘Ó¹ùq%ù,ÆÁ_A\Þ§nž>‘_ ÙÈ $CøpÑ{(Õ¬ñØžµ}.’E Eàwxž= ÕÔEÚ ø0/MEi¬Òh>#¹õ¶\¢©ò¤»„÷Fƒì Ð-xòÇ¢„çÅ7Ž%ä‹} guß2ÄÀ—Sï ZçáaØ Â/ê ®ÚÍ膲ßÉüþA´\s€­ȰÐ"+ßt‚TT³È¯Nw}—ìô ìѾ8cšìÄ+™Åmfª8楔â37f@þúsKV³ÛLAÕífB€¢Y¨£lu,ù„v¸Ø!àÿ @Ç(_X7Ž"æ™b›XÓæîɢÞ%Ù³"ãëÞq»»­ãTc,Ù„%-”ujYA‰Å 'º)1Г|i\4º©=Ù½Šü¼xT¸åáù@Ri½elagºÿ“`º‚R¹¾EλH4XÓ×RÒ7y€cùq(3ho»eCDÎÔI¸[ü‰Œ-‰)kŽº¤*Ž@™¤ÏÅWôß ùñëe*œBKÒ¶á°_F@LDa¹î<ÆMïÓ–ŠâK{¨ÿ1ï*—6Píj-ö˜ÄÊÚç7}†Îüܤïá¼eôk߸3/òRõpµ{÷ãÛ&<7HÜ>¨üö"pû¤T&~€]"¼ígÞ‡¾¸‚¤š@+Ep–“no³=蕬…oýÔÃÐDÿ椎mõÙãòf?Pë¯Æ¬ªÌ‹|e€î…T‹€j7·ˆÏU¨0r¶Ë·WÞçôwö û!ý`ñ>›Þ–,€f¸¹˜ €ÁÝŠO¦}C>t¾/p,Öö7IÞ¢˜qHÑ¥T Öº{‚ô9’óS)€øû¤®É"&{þ\‡x1ZK,áÖ´ò ¦§óY›ûʆ¢'ÄOsHUáómd“¹9-´}×ïfÑ1º41=Üò>~-Žæoý¦ºMXh_kï” ç¿~`·}œ38æd$ökNS•)±½ÁYýÇFqž™Ë…O^5D©þþŸž?ô+‚]Ç€Ò%× îdæÔõIEš‚¢ Ò… f€=sF……·EUµÀœ(|¦Ü‘¨œ ‡òåýçʳ‡Åa$ó´´ýŒm`‘Ò7­ÀÈÐÑjÒü¾.^ ˜Ä6ì³§=m3ºwÖVkRR¡ŠóõTæÃî›Ñ.âT«®ÔÒcVêG¿7¿Æ·æà#ÌMâ©ðiªÉ^¢à«Ä {4™gÜ= ÂÞ’ŒØÜ `(ò%ݪ4=°pO•‰+Wú_×DGž³Ðµ™-'— ¹[è Ñ9pÀF@¡¡n“óÑ[–eøkfŽáÕýXaz£ !ÂÂ^!,5íñÒæÕs·Â,„#Ö›Òó¹ÄÊÔ^6 {½î•(ï·vÛ¬ðhø‹œ¯Za‘óÉÑÏ’Ú [žÎÖIªè)ÊsÍÂ…º.µ]&ˆÉHäq£&ÿpʼnç+e ãš»¢ÜÉ_À:|µ èœú Õ“æ*õŒTúâÞÇÛY5½£pY²-K4f'ÙX›Ý®â7a^ÒFì¤8Á²íw¿ù)d.õiyˆœ^€ŸÞ‚äò€¿êEÙã$̹/¦¸Ñ ò+¸ž0\½Ÿý$qAfÛ[½‘(ŸÇ}9è^ߪÖ¤÷ÖðâÇõ`²— öµå£6öƒÂ€Š€ø|Éq¯·÷±dšª´ŒWhèœÃVœôÕͬ?üSÓ2±¶„äÅ lC·§Q"§—Ë"K(fñƒQJ»0ÕL½9É´ú ¬Ôðé¥7µÖÑÛ}ð˹¸äO퟉ZM+ô’—ž>>Úqf_l¤VµÁüæÒ«<0Ó㡱h¢ú.¢M‚Ol逅)1? +%fƒœ°qþcq×ù“é/¸ßÊÈ.ãÀHæ»àÌŒWÅKcéäü·H¤õ|ÓûW¸#‹ÕoÅ·1ª˜Ö†¸–»¨åù³ƒÂiký©mÍøÌvìâSQpÑ]Z"Q[Jp÷…éˆ8\¹+:–ÜoHU4ÝŸ×qYîðDïº*WŽ [CB{²Tlš¥™ñÕ­ò6SãOÕGf„ îi¡õ¹hü>Q9eûÝˈ_y÷54®¯väK×g'®ÎéŸ_ Ÿ«'æùc„±¾…]þóPguážZØ_¬«#†U+¡—~þðz<žÐªµÞÀ´-4H3kQc&±sD§¨32j@ÿÈ™{àIÉÄh†ükSßÀ£F1Îò`Ñ6KÝn>;|ÉïœïÐÏLÑÀ{™MÆ øR1k‚ªRA¸(}6þR©7Þó2‚ò%gK±ý[câ ÄÞµ]”.;M28’ ‚"¯ÂE„h®¬bØ8ß_ŠîNÓTÅŨ˜P«­œ8l´uS:Ü¢«ÎfcIˆü{FÆO?œ&Ë>b÷ç·o^Æ.¶‡û¸å^ªøòí;?|ør5>íÃS”cî)ú‚жçÄGºFñÖP‹©‰AwRÌœ¢åî}+øR^¤GtöKÛdÍw¾±ÞÜ{q¸­v™ó)û" êýîú‚Éça«aqL×{—wìwö¦1Ó¾M°nƒAé²ð!–êÅÃOøÍyÀ*]µÆôúÇ£Š²Îåé§}m˜~îõ›}üþ²NŸu’EÓ:0pmøb4ð «v18“ÉÀf …ã s›eg 4- åØâcR¦;l`æ4æ×u#&²>Wï•?ú•/aà\߬{­qoæÉ`šÏ5Nà=è¤q×@ÜÔˆŠð¾ò³ÿ´Zrõçzæ 'µ…fôЂžG5ΘvåH¾ŠÆéA;Ÿßò´]9 >JJÒIºý\ ¬²# i"¶6ãúÊVÎÑš£JHp+@~Å¢þJ™¾öÛušQÒNª|a˜ \nY?w•µh?Úá ȺÁ}Ðj}ª_ĦṠ…œw£kï‹‚~4-ÊKð®Ö–<è£Qu¢ADˆæ±Òíb(‡ÙÙ³"üàÊ^€ióXWªÃ 9ÏW-[r Fÿ‚X©AÍѾûwo<‹ct$ïŽÎœw¬¨_VÛ½b¾ÔG„ÔÊ ¸QwEë&!î”éSìf$J`Ãì; »NL€=2“>¸ãþu±­úcK™g±]Ô xÔ¨v.“Œ¡³x,G/r¤jcÕ_ŒY²Tå¶m‹ ÛXh$ iµ¬U‰J"> aSøs†ßÑ^Çâ"X¯¸Ù:/™„Ú"K QÓäaôvñóøB˜\:Ų‰æ,Mº‰.Áã³<Ü û‡w©:y°=§‘¦WûâóÄÔrXÞb=H±rܵ6äűêƒ4RV»„ÿ¡³«1ô'ìiyb€['íÛgR]”¿šˆOp_¬­çKx/®p«\~oo¡·¬Z_¡Â|zàžfYx­8ŠAaL»š8°Ès¾P6'Å˳¤Iêœ×+ˆ¹Ã0ëºsiÙâ Ç ФDk§HÂ@¹¸y3_>ý@?’¡L1®'DFÛ–“{Wp«[²#·Ú12@y°6Ì&oN$²•P˜ú¶²…°Pálu­3wõ%6ïÁñÜŠÔÛ€½Ëù´(úqlžëeŸ\RaXæªyw´knÖqûǧ&)¦©eC+ܯ„Ë ÐäÛÃÎMe”3V/\¸ì«¹'› äwüÄPtìà@}°"†—òŽGˆ'U$S]¬ôaP ꚦŒydá¡6ºÊ¾“c½JuÄV-A*qc§ˆ#¨¿lõGYâïÔ9dMê ³+™¿fâ^„´áé ÜyãT¬³WíŠ?•Lêû¼´DÊFœß¡ONl=胂‘ßÍ4!ZÜ’ŸÈÎ ÚõtbºÈ=’G{rÀôµ…Qz´©Þv`³dƒÎ,á>v.q@î!“qMìäºÞ}pÏ•AöZ5Þ÷äÌ-•I¬2§[ÿîýYèi"Kº§ô—R'Pט—QKL” ¹u—Ê&B^ð}/tléôÙ± ñ¿Ö‡‘¼9DúÓq¶X Ïé¹yU¢ð),Z-= nÉÙ¼\ÒšPa]¬=ˆ]ö´n¾½¿¤7¼¶~hÉlKRD²Z©^¡mHåî~B:ž´Hªa Î,EíÛÞ­Ê„*¾ø¨ Â5Óòì+­6§dpNºh/»$Tíßñ=ؾì`Ä÷çº ~ÿ(ØQz€Íºë.©2ÙBµYìóXÅ爙9æøØÿÇæ fUòJ¹|†ÎÏë–Sóþˆ­¤³Î‘ke­¿vïÈu²åRW“Oø“~l^¨t=Š6KÅC sîR5Oþ¹„‰epÀǬPÚ\úKæšÎ&9\dƒ›:*6êe¬‚QÃatDTU<¬‡bŒ•2Лߛ:6?S"[17!GÉ™%ïw‹¥·Ò¡ôÓD“´e? À 1³AÚIu=„çî™ÿ4÷÷ìs÷7èËÝÕBËøóÚ5àr¯X‹#€a¿ÆÎ³þÆ­taós* m\v6éxsJ)ãγÔ>ûü;Lp:Òö—ˆ'EÊ ä1¯³‰/qÍÍîtžà€ù½nÀE¼±Q«Šý ( ñ§/p!gmr›ò?`»˜&7rqÿÂUqǪ–`J±Ø~M‚Ä=›aC¬ò°DÓÁdEÖ<µm%êÞ®ÃÆ~hÌÐ ²Ô˜ïÜÊ\}˜Íð¨îÞ3§n\º›ƒêlÑ­×l?òœÅsƒäÈr;a_(‰QaÙpÿ¨!â!qËL/° åg69oÁq—xàßÛÌ:Æø’×Zv¿Ë«˜%"¢• ¶ÛßgnIŸ'¶˜së‹4דä}†qùª:"`ç᪬œv± cÖ2f5§­ÚÞÕ }”Kã¯T,uÙp…þnƒÕÚ{~r¼bt˜ à~Ýì$îI ¼×ï~®¶ÁÑ@iè?ï!ÀôÉz*45mÇ’Ã.mœŒƒdðÔ„Äåâ$!þ ùba"ï0Ôº|Õ_&4ÂïAÿÆ¡‰$$nm]ž§iMÔtyUI2ý:4½œ_×úZ÷L¿ >Ÿ8_¹{{yšÿžü@vèv܇H²„‘fOcÖ›àæˆéêcµ¬‰¿î4ýqªa¢ª>&0c¢¶GìŸ)˜}ú†•”N­†S#ÿâ- }x¢©+Ë‹³ï4u.ÙþCà’"ÁˆYŠÔV$Û&¯‡4þQfÞöUå\ Øå…É%qÿü²öLñí¶²! }'&(*¬yaHÎOÉòná*!Î ¤œT5æÃãzò×D&†¦0ˆ‚¼ß—“ÎOjv6€²4Þh¤ÔnY­G‘0auoý&ô>.Œu!Ì&æ¸÷Íf%s(°Ãyë3·`o;}7“ð ¿ÒæY s‚ëA>)o-Z,ñÉâ ”êk[¼ªÌ_?”iñÆ2žºæ)=É[¸uýx«Ô³ð×Úz©˜£"[¹E-~`s`ÌÍ q\¹blívY™îìYP&üÁ–qoß°Š®*<µÀ£g‰Ú–ƒ²fÔåÆÂ×aœ¶cgf1Ñ,9U6@ì|owêÿè­GGâˆ=É‚ )žõ¤KPÌ_®›¡s½‚BÆY¤k^‚²£ý¯;=ý÷7¼¡À úÓÙî JqÇ[K‹;›7žÈëÀ´™AU=t£­[|9°=£w?î€ds@`‰-Js»·»•X<å¡Ä±8‰¤r#òN’ÅÇŒokâØ¾'±1.¶žÚ`ÕòÓk0.1‡ã6,94±Ï­^*öîý¿ÿJ™.X=½gœf±Ü>æV6—u ѪìÍû ,Â!†MÑ@AZn3Bä¯f¢k¡†²îôŒ~×Ïû"¼ÂúN ›K¦“M1§gç’kÚ/ïÿô3sÙl…¶ÓšBô+³°z~xðƒ-c!ôEK¸yþ–ʵy71½•ñ" _û§³†«f•°ÞÁIˆ<Ý£fõß{bKü¼^rßcŸëÁ﯇“ñ€O\‡(ñëËò³´OØsDÇ]ñ'ñ˜"ÉãÑ ¿gíô]Aëb&1@ ᦥÙSÿ·ô¢pó“ج˜<“OÅ¡€°ë/díÜd6:bõ½:ŒaäXàÝžø‚Ú“£ò„þ ZnKð[“òòÑrÑÓŠt*Z´ãªÂOm“BøÏÒ›³’Q^®WWO¹+9ÑzÌ ™®þØ¥ðÙL§ ­S¢+ØF·šœU¦ž3c+Þ’ˆ†Â9Ý2t%ÈóCù¶Áôó<ÜSÞ§Y°’ÐLÄ6äë’ÉJ^Úõ!‘^og»LSÌ"Ÿ¼ ˆ È‘ß÷<2öê;—~ªL(H] 7þ`âaõ2Q,gö4¦4‚ž…•hÈUît‘®‚«ÞªËüÑ3_'ˆj"Õ?ÂáIî¿*ËX0‘ytfÆàRI÷,m/÷ÄÜDQ¿“øêlP.ù(ùx‚µóÊâj$@!^>H¤RS#¦€†%¢õ)b~ÒÙº%ðkæËv-ÞM}r@U ½ZŶÐ|üÇJu"q0mmD _·ì H㤠©óÁ浃¦)OÀ/Ùÿ)ÆEh‡ÂÙîèÄ¡›â—bF$ŽÿL5¡…-9ñe²Š¼æuwáWE©>_5Žxòš@É‚›R$úÙŠ³BÞ¾ µ)—¿Qت¹C=‡]U1Z7¨N[ñZ)$)¤AÐ àK²^T·‘ëÏ0ù;Ç %Nx¯­]™hÇàÙzü“ø¿¡œ^?H ¨0©8éj9i‹²þ‰ÙCdÅ¡éêG8<Ì$,³b:ð“k†5%MrÏï1‚\©ž(V9þ¢ê~r–ý3…Ž=~ðŽ÷­9ϲ S¤:ùôƒ\¶‹q’$¯øG”a˜Œ¼ú‘bùe¶t@€ÕMÎáŠÆ’õî#¬\s¤qkÍ ô£`à£ÒêÚâèæ«ƒV4SÅyÓ‚Ó8|)ƒå‡ù­œnÚç?zbzE7¥L}[·¡ñ^Å´ëð Wkv/4Ýôc9Óï(–)T:<²ä]&(ÙýXIoeS}òw¦´!ŒÔXL7OÄòÊÝöžRe^g&S i è’;·ÿÔÎ D 0î™8Ç)òѽn§€ÿ—Eâ –ê°©‘9…aS ¨?ƒ¡ÛXûÖU:_ÿT¼p_z¥öœü@«¡âbÅÖÞ8•"U¢¾eÓ|~†HI4ù&„Kóák‘v6Á¢ƒ¿sÐÚmÚЈH7f¹=ˆH«Ó+öçñ² ÚøöüËn‚í×H{ ïKƒ5}b©hÍWöõœæÇ[‚çG¶ýrÓ˜¾´òXrÇ/îÝlNŽÿ”-ÉÄtÄtSݼ6±Ïi ã¥P'Rs¼æÜãVÉå8æ)_ME«—ñ ¦+ï^vËÅÑzYš>¾º¶)4žƒÆ¡£ßÉrçD’ž¿¦iû]'N›rà@ÙD¼à˜yÄDzf¥Z ¼i Ë Ùoy'Î4c6ìòòíÚø5çû¹ô/<¹`L&ѺݤpMœ(­Cy)žƒ‚&E7Å3’Ñ÷J©lvèÀf5”d­0ÍY^‘Éï-Aµ—®Ÿrø/ý£€_G-¸…€Ô©–LÔyxëJØš8¶ÞŸm÷è¯ëkþö úÔ~€Ë X-·5ñÍu2š ª¸ ¡#BÏÕt„|¯'uø9¬+J€FëÌ[°af¹èyug›îα14hr7¬ ^Õ<¡© YðIW'Ë€!mϲ‰æµ…Äâ%GËyj7¶Ç”ôHBŸ%׌-Ù:F<6ý“w Öëp"ƒÿq"%¥ôEçâ.½&õ Ò03)më"0ä!8\P<£%»|Ù1ØÒjr–Ó(ƒåQ|ev§uÕ;-–2JÃâÆ#º¨†ÇZ¸Þ)ôSütê¶¢ȃcšnùg ØÃ§:|ü{âwE¨Z[(gT¬;2óJYô´n)¡°Tädš¦Û’Ì û¤LUôKAü·©’s-+ôJ=ÞTÑx9Ü˳°\.ûRŽúOÅÈ›“¢fäü(´¡6+±"ë"øžy‹Á´šÅ&|ã¤'ã+ž­þCAÞ2=Ÿª›µâFO;Z ßqW<_¶xYk€}ûi¶®¢Wã|EÅh…ñvqVR ðË3iÑIIâ  &Àn·¼,‚§F:µAjø‹¹óõ½]ž2×YDƒ°‘i¬&ìÓwgŒŠRŽ*Ýn' ×-s`»"ÇTœñ´^X¶¤›;E8c>•mâwô†;L…4¾debC§z4MÿÄ©ÕÜIbrdSaé.ÜT=ó¼Æm§ÞñieÚš¿}•@ïF;, #v)æDÕ¥ÒaF²S NFâTÀ³!¡^:=_®lžs³M`T„Š"à+ê{bZã]˜ Œù‘¾\÷ÞQ‡ê³esL¶•[§[ºVúÂh¢óÉýA:G¹C)¾N‡Ð&Ó÷¿¨GÖ¯Õ=( ÖA×J³y—ÂÌ Cf7€ƒ(Pb4ýåÞTbtCèKâ5¯Ó”CÙc·ßS‰&¥Î_õÚ{y–:€H…±Ê””tÃ%ã1+XUµ™F€öÂéÕ<º§&ðP{Ó›ÛÁ ÒW 9I|rŒéêM“Í“ÙCLz,5ÙXºæÇ/ÓÍ‚ÌM9θ¯Ô«Àñ*ôr^ãE†?öõ?­•QÁˆÆÊF[R¸t>›îÓ…ÀK݆ʉ!G2Þi8õ1œÚiZÃl‘f¸ÿ7qNj«®½D‰–بFB3%ÃúUqýQ6×<ÛEÈr–‹6rC¼׫þƒ«Š` Û¯ÎN\ô[ÔsÄÙU³ã^ïÌ8ešÉf”×ÇÚr>!e0¥0ÖÉ-Z ÏIAÔK’„i˜bƒD'íC.SÉšãÖ§é÷)·JsŸ=3mlN%Ôw³&?ìy î+p­{%ùÔܰĜ=«7K©XœCÌb#„Ç/ó+’1¨•€ ñ×9óá†äÐË®ºD|aîe¼Ñ•ÄHCÅ·Ô]Ÿù}ô|\¾ž–‹ yšŠc#E×'ý¹eâ»ï }™ã2ïÉÓ«êʃÛ‘õE—0¦fÓmQºPtKü>núk£:šGÿ“JY-J,ÐYÎíÙwø %z‘‹W¶ƒ©p@çl¸Ý² V¦(·_‹z,¥ì>AäNAŠW$'ãMD)YHø—VÊ0b*Žþ=ø·Ò$¤·Gõü’êëa¿j.õs0ßÖ_âÌýHú ¹[;)–°w‡J|ñ“X»ñhÕxÆÂ¤Ÿè{X[8¸æÌ²„ÜëOgd-"²È kÎX°ë,%ë}÷NPè`†ª¶8U°CB¯l?£’&Æ¿4‹_u°qž\*ø.=¦´!¥è½˜BÑ+Gk"Öî—Ý“ÅoŸçŸªKïJ iÆË /ú¸k» ©‚Ïi¼FMZÆÂþ’úø¿¥×0?ìTŒÇª=O¶+5DŽ)Iÿ9.gðà}øFdð9;² Áû$Þ†G¼†a@üÇ;,7{ƒ½³°#q”Nk˜¹þf‚[ [-ÀÔ¨jðš`ê«`HÑêT¿GÊL$ WOÎ:´î Œ,ŒL{SæþVkû.jTô*¸úºÐ~½bˆ_»«O4k¨|XbõE@$=ƒoˆþ.ri·§=u›[où¯6+U®Õ—…µå¿kÂôÛOlâV"£èÁRÊ€Ÿ#†ôCÞ?È—B°ê@Lð°ÆÛ›v}ÕkºŠ®ä`ê[E¡ ¶%L\®ñ̹Ft³†Ø†l:‰ ûEq!¾å qA»Ü¢ÖÛôæÛWåçãuIx’Â`Ì/Pû†V Žúý<ŽéOÚ?ˆ%|Ah±ItÕ–[V@þ|…BÒ/ü…}9€yŠ7§†Òµ1âȸeo¿ ëˆ1sCN§>[-ÕÐe“”Aø[ø}9`²½V”˜™|GOdƒ„fjÛû¹ŸVñ!LuPÊ€(³NK[ÿ è3>õ ¥‡ÞÞâJv>Çjù÷÷á‚‘›ÔÊ)øõ¦ž6 #ý]‘ë6Øï- âš§J{”Ú‡s«†©l8P åýŽWñãü¾òÀ¨Ý²=Ç»\ ß_›•§“¾ëAª™òØÑ\ô0CqéÁ4xÊ[o¢a¦Fj^ÄÛE9.dZÜÁÁAó|.¶Ñ mž‡o)¾âÁªBÒŒŸ#cD¬ÝÞgØø0ÔÂŒèê†m`»jQ½˜íMP®‡’ŠOÓNY|””$шï5ëR¢ž†qîXù*Ú)­#´e#Z©Ú«Íï²Áå¼W…xj±ß±º*2eU»¯¦ãp˜Cx®³aȳA°Z4(BÍr¡xg¼2*–JÚŸêu‘íë âŒ'ósì‘0ìTU¼¯'–Ä“§+i,í ô3?“TÅî1ÍZ7s¸±™|KZ©í<öÓñû™Fñ¡Àb­†ò¿[ÅÐ){Fïè´†;ƒ”C#'¸‹Š—ZkU¿'5?Zý£F°1µqæbçÛ到ÈRäöQÎ/ ÁÅ;Ña;ÓÜ!úD†–^ÂTñãXW(¬Þ+[GÏO)™c÷ rËÓµk—øÍ¾p‡ ™"—W_÷³ž¿Ã¨É:KÑãíµ‹!2òàEx?ÒcjÆÇÉ% a½2¾ü¶üóýÇVÍo)ß)L¾ä‰ÁäÖ*ŸT„È„Üç9[w£L.„8·±¤ð wxu°Ž\ân'hŒXh‹„•Dm‘'· …ºckèãböû jvª¦Œî³aÃ#Õâ[?˜Þóa̯R›†·õ¦æ¬?÷ù¿±¾ÎDên Pþ~Etí›ØÍ±y| oñ  Ðû±r Ë]ÇI»à ¯¿`¶X;ZG õ›Sh®fsËüm}AÐVl…b¯?ø/äq€ùŠ~WkxXv†öä}u1̈Ï_Œ˜Sud¼"™=Å#ªwµ>Á6­[ïäÕjc£¥+ (T–[‰Tïgª¦ˆÈȼhùå$^äu¾J­/T¾H›ûêxÎRjx<Ñï1—)}ÅÃë§+m<²Â©â;XÆüºä ‰5,$Ô` Š{XæÆùÆ¿ÁÝ"[–rÊ¿Ðpí wú¤‹ P©C‰Òþ£’¨rÇŽ Bõº¥¦¥5[]DÖÇ0ÏÝü\O.êË6ϼ{0ÊNUà‚"^˜ÐÚÑ ÏKzØ'%hÐe:J‡÷¹P²NÇmZ¾igm…á? ¨X¢æN=â/ØÎ£ŸRñ:ö&J0%€+v+ž¡záÜìÿìùaµcXDX7<¶óÜú¤¯š2ˆþ6œNе{ )TwÔ©¸¢‡)—ÞÂJ#†@X9·¤ì¨äÉÖ×)-î‡?Åø“ˆ27‚y©Þ12^ÐìN;X[á®%³Ó®)àg8W½ôB´‘XP‰ûå —;oz˜y!ÆOûןæœ%ÿAˆdêZü0ºFãÊ:€´Ó•c£žïvCP|ëOwj¶iÖ>T÷4&‚ºdv›žÄþª”ÞOÅC -¢û‘&¿Ð›;'‹ò†EÍâ`àý˜ÛVÔ¯Š¯š*ª½(wë$ý¶f›2㵬ú êÔ2ÎF7Ž™¶2$ïó†Á” ‘âùu êc¬/h_:57†ÆK cAÊZA©kà…°ˆ¥qÖ-©0„¤^²ˆ¨uøAiW¹y÷1sWL?…@ 4…â:B€ÿúÛæ!/á'¡¨]®‡z &ceM*m#KZŒº=´c-{Ϥݒ`®ª‹ˆÉa_.{šaYòÕëe ó¼ýÕJõÙ} ¨sëNá r͘ÜtW"¯ÄúWõ9>¬¦¦èZ(o©e÷=‡¡¾¤Øò !²(äŠßÑÈkñ™ª|Õî¿ ¾hN/]j}ããSjè{i¿ÙÊ1#;Ds³“·šËuÆ’³Ã’Pö7ÞºïÕg ì ÿüAŸw”E6i»´dò¥0nìÃß3á¯x+“VâiØi ’Ý,áÀÇD.¼üÅOª 3ÚàFÅÉbbÛüÄK=x¾t6Ö½uê3ÿ½riµ s,@z"u)PÈ7“vcêÌÞSÆ3,ÚW´GWt‘_ µN/Y±Ã/F¸b»È4ñüIœ§/Í7ú0˜…m‡D—º(‹ ¾=ÁÀ;ɳ†q˜RøÊÞ¯øh5E1üô’´žÑÕŸ¦ôÙH¤õKäÍo'1ŽÌ|Kn,TÊR ©¹ø¹žáÇrØij µX©0„ß•=÷¼Ë¯ú$‘â¸nuþäû%¤¥+¼×öý’ˆAè;ŽcµÉZÐÜßJÅwØÃ5~Bšù@—UÍtìJ ôßòv­ãå¨/ÍG„ÏËrs©ô&¦žp …(yÌö™"˜{ê¨Ò’ Ôít R}©cÜ;÷é§æ{ÇÌÚ)÷‰ª¢X+jW:_Ðìׇi$9ØÆ rµ£è=—¬Uù¦ ¿›´á1ÝçÉÍ|hÑ–»ÔŒñ”ËŒzvÉ:¦úÇV'Ãü¯&Ê“ÞD·JÚµ•~êE,`’(wáÊ·XÍ0Ð6€?]qÛ2F3%Ò0°ˆëµr2º5˜ Øùç\m!GŸí8;YõÉ6ña %9w_$1µ‚xpbzá&™„/%er.µàQ·‰œç4£N´só`CÚ†£ ’Æÿ2^ï+¹ç"s1^òtëBWž¸þµ ™…§ø kmö½Ês'†U>UeèÈ"ï•×"*ØvžTߨÝY¾«?Á`”C÷Kþ…3¡WÐÔ¸j“¿úPvxP1ß^Œ£øð€§“¡çÛ1è¸Ñdé–¢é‚ ÿ­OSè/bJVvŒÀ™¨Y‡áÓ+ÚvQÉ+ûüHhtJîãšÏR@A—šÿpïUòÊÃà¹#¿½ÞRÜ\-q}?0b-‚^ŸF'p÷:h¾)º—'žbYØBçëKôèDܸ¦"z‰Øä½•ІPwŠ ß¯ÒªX³4¥Y¿q&&ºnv]œoË·rB"e7æDGžZ/ë†óBcxeÈ£IX[’Å(€ 2§DŒ'Û1³‰¤IÙ*®ç†5øZ5ïeë.]„&¢þÙ§Åp2D׌áºh÷Ýãhƒq™é VºgÐ¥#ªdÇçOæÇ¯¾RAÔ ÔeE&·ÕRF"2¡Ú%µù«®Ê‹u&ŸCÓF+5Žéþ«"í§GiÄïªÜuø¦)†1íÌ @ýb·ã'‚†šMÍHÄ7_d5û'vÄXGÚdtkÆS›x8T'jC Ç»C²ÎR)ë·Ã½ßć‰mTþzIíQ‡€£Kì±`rNº½*mǤӹ_¬"Ëê-ÞŽ}š‰úPÛî¹ÑýS¬”A®+7AK“Br¢$ Þ¨Ò|V %Ä‘',Êžó m­U$&Áá w§oœÙsQO¡J7À•Gò1Š3­ïq?yߥîµvpìºÓ(„F%tar˜Ú4·Ã óP^Ïέ§1ýí8ÅÙ@‡?x|Õ0FÜÖLÎ.]7ÑIކ_µ}H”Ÿ³U˜ÁLöÑ*`  ×ë@­,ÚÀ'ÿµël 7Ï Õ©Çã˜Tß°¦çLw¥fCÍîÍ¥ŒÒ§øEK­4OcP/Â9|ˆ²HoÂc"z7ŸXæÚØÐÎÅÓËÚ‰‘®tÝ2Lƒ¿½Ü£ü 5â&¾ßângû ­_®5, ‚ý  ¯ ljîe—„Æ„LTµÁx¼á#•z·{¶ìLž %F·{òML30$ÉÐ ¿ aiAÏQ8r/š&8dQ%ൺqmÁ’³ù=”Eì2ÎWxÖfí¸~&¦Ëñ}N2Ÿ·ÙÄÔkE OÝ"Y#!Ì«t^+ʘ)M<^®W€X•ø1-ùo¶·}€«÷m©ÍЉL8V[ú“†Œ½jÍœ ”}fý>¼‡ç:Žs+aÍ»Åò°HLäjäUuG×f ÏÀ¯šlHo¾§¼Þr oBþ)zZ/ø%4ë½—ˆ×J¨û`¯C-ą̊Ǝsb ÷ÌÕšßaREQºíaš³=¾Ç¦ž$aƒÖê¡H3ïé,W«/zðÒ«öâ*&ü9¯Rþ·ÎjÖÖg ¶/c GïîëuC“RÅlÒñŒ¾+« C&¨Ä%›PÈØ/Nº=Žñ\5$N0>Û¯~bBÓ¡® i$;Bêc&ÓÑÄ ÏLˆ£íü-Ÿ[okÇéCpP´v‚õyï%åö#â„^Q©º2þEìÖw‰¨ü×· ó¹ÿºdø± ã.oé’ð-Yr…TÏ¢/È6„ÍòWf~ctEƒÍOÚ¯½†å¡7=Y`X©¦ÀBSàíŠϾ€Ô”ƒ&’kÄ÷ɉ(îšX¤øšØß/Âú.Ï„°Œ×Ú§Äÿ¿'w P a—ñZ­‚²Ã ·ë d3SÖé£0¿P¥éJQÓŒ?»Û÷ r-Áë.1à•ô(Vñö•‰¥I/ÓØc¬4×:3ŽÈsbFı!™¯ï²aü=3~x;Æäší¾QëY²C’öL9¢« èb Ø…$e¬  è¦‰§ö¡JdÓ8ãÃhöR<›±‹\`äz—;ßxÞ–¼Â×–ñýD±;pÓÉŠŠqŽ™wÜ¡+tݳ«·ÛŒÓ$ÂÏ_!˜Ü¥²S“ÍIXN‹°§ÜjXG6bê,Ê^-!³Q«œ¥#p9ž¤+8ó_Ò&M¾n¨Îˆpà$8ÀSú„Dz¶Mă#”8„P‚"Ä™S š_ÌEJÙìVdØÆäþ}nr0Ãc\Ó0"´çõáÿVƲL@ˆêØ»µ©ºKÜ|Í÷@,p õpjˆÑDZt¥Xjk!Ï—Y w³¿tãÔT!íÀºó:ìLĵ’°¼SÑ€úƒSyÍ9¾ç~¬J=€ô‚ Y%˜#`Ñݵì€Vîš„èCó8p &;Ïo(+z„ª^΃¶?$f6YŸæËäÅT©Þ/Ü UaþÉü2‘ìë Lt Ôß“MÏÄÅ ¹6Úߋn#XPéf³NÂl›/Ùϵw&ÿÐ Iƒ|M¹–ä°Á¯œð%q˜¦¤Ñ°X   dÁn[3´†²iJN-ÝO' î¹3ª¹úZ6}gVß`[»sZ?Q‡år ä$›Ø­Á‚|+`Îkzã1ƒp'Ìn_õ\7ÃpM%$ÿkU¸Iéž•ª¾é={Úü¦û’^] ¶|ÁV?‘cPõIʪ+=é*67§|äöù:QHÉò ìî§Œ ÚBÑ.CÖÕEŒ4xþ Y<¢»Š¸Œé´¦³§ÕŸ./›1) gs±´ÛŠoؾKÃvVŸ)ÃãFÕó£mróÔÁ^ò$Ú¤÷¯ÅÍ€‡«ÊÜ»k6U”Ä# )ÜŽ²`Œ DCÿÏè3žËTì¹úøÉ c¾ZoÙcræ½yT²MpÉ*^yZáúqÔ˜=‡ÆÛìù’ýœ 5$õ9y ¢@•<{l*ürk‰d£šÂTB‰¯¡ì ¨¼b€(å݇¼ß0gàz»–w› ¸±³±ÐW¶…¿ÛCˆÂ'2¥Eç1EU Écj)ꦵ™Êƒ' #Ôì ùFϧ ¡6ÔÛ¾ÖÛƒ“·2)˶ °A~å,´þ²—ˆŸ¬wðß‚:ËõP80.§¶¦µU-BÄ3÷€I¬ p¶î J—~ABèâ6ØaÏþX.ÎÿѶ¥ÞgÂØgjÍòºK|ÕÍ4öŽëÛÌœ(éMäDÞÞfNïjvWdêzÅ¿ƒ;À€šèv!sÚÊÈ¡ó!0ùÈ(lRiéË<ïûšMQ¼IˆO׌×"$l=!¾1UÕÜÕ“kÒ/–NáïŸ@Ç%Ìö2WËACÞ02¦¿ù:òÄÝl1>n¸È LjÌ×ʧÀyI˜YìyÞñù›ìåi êPn¹w´Zº¯ªf*zã)ÂNæWŠ|ªä €©ëÊnÌ×áƒä)˜¾h”Õ-ùa'²r¼”ã…®«D#hýäl…ýVrCЗuÏA ճг ÉÐÜp߯¡†oeA #.N•/¾ÙÐÎÍæq—„Šѹ%×I2¬%‘{ã2ÃcX®¹H%EÉlŠvì`Œ(’#Ð"?ú´a¹#o9~ÇÁãý¨wtÕ‡©Ü ›Ïºýk{rØÔíì2È›çÇüñœ6ßÚ‰à=šýÛ šÒ0$U’À¬kº®yþ2‘ÄæÊûïcSḬ̀‘JQ™SÅÚï+!ùÕ¶q€Î}út`4yZ|íшêæô´©ê=ÊC@âE·•Ëø‚W×° æ:(2¼™(loÑ®³štþ —ú›)l9@’ E|D7ƒSUîSç”ù’û_F#›£Vûq žý‡hÃxæBNQ¶®Kä­4>I®½‘»±WŸjÞ°9þ ‘9ÕŒEð‰Po"õá=9¹ëÌ~dɯo-Hå|É¢Ê5p䜤‘A®`$œ²=èët¼ªK.[šrùð~Í"!Œúçw|Þè’dl Ds¡r¯;aBi]]º<˜Œ 'š®-â߬tÆW×ù[y¡½¾”y±¼Û¬µ`…—¦N©H»‚ùŸ0*¤&¢‘Í#¸ÉÇ÷ äâvQ*”ÃAòû…®Ÿ8g NicgÚ¥>Â9uÒ«\2·ùòe¹!£ós숿qƒ»Õ€ÊZï-3‚’{Îg…þþ̓ê Ï¢&Þãá½Ø(Ÿ*×”=ˆ=‹MsfC«J OYÎÆ],ÂÎgB[Z¹ d Æ¿lm ¸™Œ›—ÈM&¬ÿýó¾;Va§¦Ã`ߊ]&F:&0Ùz©ì}+VãzSÌh4P*N0qç ¶,ŸMjÒrT9<ô@Ðy‡Tlyâ~È"ûØ M/1Õ;SÌ»2ãÍ?j‘©¼Z]¤ž±'‡ðϼÅ7ÊÊÉ}ì½ÏIŽ»A ?Ç1¬iå°·EuVIåæ‹÷xÞ«ðÁBgæ^¿lEC… žoje+j®}†­gÉDÅÔ›iÊy"ã‡K¦¨OãÆadU"lµ$Å&²Ë1 °ËÔ4NEM‰A±÷ÅDŽðZnö䬻OÙÏÍ4‘¦ eºÂsEú<‚43Šü ²Q¼Õ(}ëyc¾}<Ò–S†ÿMT£ާ£‚Lû1‚Ä·£Âfò%û6¯°ï¯iËo1çHÄ€£Š~L[úhÌ|‡f~€ÍbH$¯TÝëkÄ6å_‰Hå¾h Ób^¨ºØT/}2ëc 9vHçRcÂ×厌“5î1†G€Ði@…œÚsùù"rDGÒJXéeÂlûW¨+»F/¬ ¢ÙÑIÍ8LÇÐûà}túwOßh ¨ò⊱PMÍÿ¯z>5¶ê‰\Xêûÿ âMª}hè s0“õF®ª°©-%lU!3ÿ6@œÔ„ʆSb>{AÃkL¿Ò»úëxÚÀBÔx‡ÛÅ‘dÎk¸Ÿ ŠŽC(ÜdwfȈÖh óµ“+LðëQ©þ‡4éæaÁ3¢NE'ÓÏ”äÅ)ÕŠ‘(?„ìÚ4zÏÖÙ6Ë3â¶Ó=Ûż¹]-nô ¹v±WÍ)ÙÜ›nç´cvïnƒ®†Ö MÄË¿¼Ý š%O"Ý÷}Ü©`8ÄO7mb|oöUWp>§ª7Â~xFàÑúAíÕ¼N€àHs«ç’½Ý‚Á*¬‘*›fʰV´Ù8mF|Rú\oÿD“,ÀœzwEd.,ïO¥Â[Òký…®f~â 4nЀÑ,¶Øë=ò[P#±æÛ43•X€†´¦‹X11÷ñ[ÙLá3_¡w÷ß‚øÝrÏòî“å˜è$€^µÄÝÐ…æOjȤ:Jíõž»ˆQ¸'µì2+“ŽWËåÜñç4ó/è:÷ö¾Ð^>1LÐO+Óãü…úÏ^r(±j£‘Mô{*HñéžΆ¾^˜T´•8ž!r—ÞŒr%ÊØr0ÀÇc¦ì|‡Ï#Ûˆh;UY5 iù×ûõ³ ¤\ͽPaBoª$Án–-£y;­D[¬êÒ–dhJ솋ºåÊ5ïN¥}æ°\ò±C¿FÔ‚©ëý)'•mÛŸ¦È7B’¯1u1“ݲ ePeñ„‡ ³”[1^UÏé ô¢;-ÃvÌ1ª*\îÞuë “º(&O¡Œý”é5”VÍý#æk¼.Q†©^IýÐ:,ý„:÷Ð |3À@mêåaðY|û#°`£ʃÄÑ7²²‰U߸%I„‚þ%I’Z[Žó–üzånUòð†’;.Í$ß¿êû2$lz ÅÀT“ ‹uo½ò?JÍŒöMöª£ÝWé(ŸóRü{(Ñ0~ì¦-òßmWtâCP)ù¸¡×̶ykSFÒ$¦ñ"ÏY*ñ@Üö„G¸Á«ÙŒÑT™Ù°ÒäuúQ Ä]‹Ýêð–Ê Á–ÆÙèáüÞÁ÷*^wH6%j•û$Ô;hìœæjDÆG'M=¡C…5is¢ÇWg‡ŠW¤iåýT$šâ¬;V¢±0o² ÎÇBÇæši’_î$ùò솫\\s Ç_<ÌÅ5jJ%޼«—´c]ª±Œ%@°H%÷Á¶† ‚ÙÌ—AOl$G˜låß2˜aòÛpfLxF›-º¡1‡x™*ÏÕf>j–^Í>Ò”ï½Ç) °jã2rvBVOð¨ú×n•£­ÜÔÆ÷Fˆ åi æ€ø¥+è9ôµ#èó|ÜY¦ ¢²Èï$‘òT¡Y¸%©šbÁV¹áRä#½S¥®CE} ©ÞEâÔüøkìˆÂ˜2 ,À#µÙ?º%sm# c>H(ÓÂÐqBÆ .ŒÅc©kíY™‡/l¥ bŠsL—½KHKhƒßvË-]™<ͨ¹ÿË‹p.enÊi’*ãZí _T„º_‡ðÁr’ÞÀHw)j-£Ÿw ÷>¼¹üòÕ{tkÉ Ê#¨'Õ¶?'õꓵÐͺþäØX aÍaÒ¯PûÐÐoöï†yªÿ‚XÇ:Êõ¾ö¿á¢;’C÷C j-'uÙmÑÜ|]6ëêO½KÏg(yt­¼óm/ä+¯‡ÝÖr€i Qu4åè“H?`iOÀô#SËÄ:/Iè,ÐæÎ¨¥#’x¢Á[¤€ÌÂ15Æû}ÁßgbÂc»]M‡’Ö€}Iúœ^ ¶É"¦Åé§½ñH„Ä`l®Uõ(”aúGÓ®>Ø9C~¥0ÿOí¹¢É¥UÌ&}É'“9™¯ò˜’W€Dpö‹“i•€…‘{™Ši~õT:®H)fÎUVë†À¥1¦i@Õgb½¼ðk8p/™Ñ^©!s9P¼ºuJ^·P'eÊ×Á¬xÛ'/ä‹x]*gÎeƒà,×O'Rb®êZÙyõP£! g(Ž|ˆ5KïäüZzZ×>­ï~Û·G`,ªuÅc¸õ’@×v–ùƳòÝѸ® iù°Âx~]λ£BøÒÔ䲯‡e„sƒÇó–°í2¼ï¡©)pé’¸õÊÝEá9„†9ˆßrÕ”·ÉèeÔDî\¦g]¤çÈÑ(- Dè#–M$+ù$ûº6$lz—?ÓHáÏ‚¯cîå¿(Â-ðì`‘­ÒÁd-/ì£É„|}: ˆÅϸĭ0V7“öÆ~ cKsÊ#¤çuC{ßA½$”Ù^7žœ‡Ö‡/ÃÛK„‘i&±‚d[Œ¡.õÉ:¯«~#µ¿½ºñë.¾›*«Qêô¶=q¤È6êšé6NDs´ë€»¬³°¨Ò M¦öÂä× ¡¬Àë[;›’œYy­u¢ïó,±àI€ÿ­Ø–ÃD"£2šeCƒ:‰¿–\?2SˆúÐg¸€ãšÁéÇt´$À‹÷âë}ÓÕ‰,w½ëÝh']goÓpæ¥R±!ºª»ëÇ<ÒFZÉ:iÚæº§$çwn¶ÿømá(ÈÂ’§UÙqT=]PºÈ ð‘[é$4×–|¢ìÐ×£ÓVxÛÇÊ×¢d°ºz-mûîƒw‚EÈurÃe°±0ûé–mäG¬ä`Jö¼YÆà×dV'.—Ióc¹%"^ùªÆ-&ȵo£…í^ÙËVÒÒÓC£ dƒ ‡´¾üi1ðFàþO-Âïah¨-`–Ó>J‡ü°9"€ì!˜†«H#”wËG#gm½¸ƒ›§Cv¤%ÕB²ô2C ÓãË¢má$²C–,Ïp(¶g+4¿ŠL «+IO:ðP¼Ã{qY{¨#¹)?êèúe廿•©Ò]Rå¸ï'÷ÁØÃ€¿ŠÆÕO•˜‘ÿ•»e–Ÿœ¾8öÈ—«ÊÂl†p/¬R×Uó47Eã,Y}{“·jÓC#ºþ³ßV EƒéY‚YL©­ÏÑ(a•bŽyÞ;¹Öù‹ý©kh9¿ ƒAgÇàTÿð†·”#¿g²+篊’ráv³W¥*C:ëZÝÔj|AÒýt=&ÀWÈÎQ­¿Î) ËmÓd“%Ï"|c,|`Ät´·2œÕ)›ÛèF<ò!/rîj‡içÙÑ“‘)ì×QàîT™UÚ±¢7ž4ΑF•ѤÆLÉ¡Å) ©¦žÞÐpD‹&ß89ÁPÛ üñ0W稠ç×°¥3ƒ#œånBØÈðõ:¡éØe¶zž ×ôAéÇB¡ü;:¬SŠèèR„ž×B‘€QIH_ÇóE¼òŸ2½5Cbs/ÑŒn”Ê¿Nbè $­ ÿJ¾R7wEFã¥_¦¢m–×Í[¶¦æ–Ñ‚wߘ¾Q:ú'g.ϳ˜·¥°¦‚˜§3[†ñšþBΟ8¦L0C.@Æ–Pމ®£~Øzž8^¬N†múd-,S$y«ÜXn­Ø:‚Šiz¬¾O¤y5KXmöá®æÒæwXHSÖ Ø– —QIÖ o›0Ü+‹cL׸¹3áþ,Çp¾ÄádƒR®à¿ñ^€OIºœ0T úm¾ˆcß@‡íÐÜaðŠ…ƒb¦géÙñ&ìKó b›­KÔ\2=…oT2{šR!à×±tQøDÃÌØí¹›šÿì­'4´Øþ,×CŠ®e@¾Į̈«Bµ}R›[ª6?ÚûLõ¿®ÖKRß¡±›I¹ªÿŸ'S—B¼%œ~þ|Ù1b¾ÉÁE$>Á‹_û¤É.Á‡ÂW×;z½à²»·NSÄ àŽø9×·r,Ô|7Ò7eK,Z¬cÔ(\1)Tv°Üiõ¤ßôÖ]+{ ™A8/–ƒ7 ÏP8L¦†öá‹‘œ×ï„òn‚Ù!5?0"$ ûSðžÛ0åù3ž°ç´Þ\R~#~@8SŠÍv=Ë?r¥ #ÄêØ ³mQ@×»¬N^fé>PÚílhhzŒ¶ÎJ\iÀý`²ÙŸ¹SÓ¾JŸ×7ò_×ÏÔîÔ¢¢¾ÀÒZ&¾´ÈÎv´­L ¼mPìÿú¦û^åFV½1ø›ë?.¼w\>ÊG¯ÕÕ§1• ©J§(…úúBúàÍ ¤C”_¹ ½ ìÒ¸"v´X#¾…Û s1Ä2VØ6©²J͟ؤæ.:? 7m$ÔZDjaíæ½Ä2T\©>m°Ã âf}¸øâ£ªgi6”J­“exˆg;áƒJ×jmól/—«.+ %?ÙŸÇA¹ëB݇„¸ró‹]^( JÏ}k’`öTÒÄèu5:¹xŒn”ùFUv*!¨ÿ7W¾ôªØ…Övlèí™1Ï<KÊÀ¤Eô.¤Ei¿Á-2«d—1_¦«;¿–!b;È¢;_)¡§0fÚ–LÜ}Û\¯&/ûå³[3š ªU‰\¬ÐææS¥ÃBv)säáí ÿ ‡.”¦¥Ôi{>ÔQMÁ`±»ilë¶t_c«U!›\|)z¥ƒë…Ÿ¸“%õ:Çâ(úâ-);òñDV]òUy)LP³ •ЧeÏÑ!ù@îçl¦÷9s¥¨³k^ÜÓץǨDÄ=u¥Pî‘Xv4;¶à$‰s³˜¶XIÝÕ;DšÐûˆ9Vƒïáí$înuL«^Sñì4}‘Þ°<^Á’¥¡IRˆ8G#q${Š%WNvu¥¥“U%í·%ÕÈ*±EÄ”¤½(ÞwФÕh–ÁXéqžuÙÇx—¦ß0!òxêGEì{ŸóžôŸr}~%†¡b’o+ú3zsÖ pùÎ3j6ñ¢õú-(µLéÅUC¯t,|ô\Þ¬HÝfÞX6ÓBYeä%Â|Omæù˜IËÍh‚%ÁBÛ­‡¨•Š?ɬq M²(,V<• ràØxyá–¶ä0߇¸rCþ®È[÷Ñ\¸•Y5Ü×&šêY›˜ŒÑëÅÅ,.gªã7ñì³ÑÓ,l7©Õ¸©<=“¤ŽÇvÑ`5°¹É0Úç67¥d\»{D' %ˆ|5·ŸwÚ€Dºlñéôj¢4ƒ$- ¸M… ðQ¦b7”âKcWëZ0ÙÙðý¥;ÿª°*ìNH˜Dül¡BÌ!……|„”K˜j«¨­Š!K_‹ð¶Ý©þt¥Û1ç,.zÀÊ,Öú5<© Õ‚»«"‡øxfÓóÀ&ÒÉ ¢ò.Œ,°„Å?Ñ TxnË:žtãlÂIÊ‘RÊo2‘¯<ïòi Që|›H°ÄÜÄÝ•$ŇÞ,ÿ)ÜÓFïg¸ûªBøèâz˜À‹–ƒ]ü‹BÕlõüÀ~PZmQÉçt¶{¸Å/œ¨ïë#:À‘D è;Þ‹åë~åÜŸYÆ€ù/”FGtð,nKÜÙ0?Íÿ£ BN+³²)ÕŸlJÚñåD”?B““•ôÈÈ3ŽaTGH¨0Š3¦ylùZ®ëņvå¨Öü9 ߈«¿a™ÓÅ3TÉ<¥—\U–ª`REj2Ö3cÑgó=bÑŸûΑ® `íÙÞ\ZÜ2ƒLÉvØ«C‹îäådCñZEsäù Ú‰[ [VÜŸÆ÷´þ.h7ý>Hß÷eÓÞÌžTg+ùÔT#¯‘ò·šaû¿ùmI›¦–”óﳫ³L=ßHLL@&Z ÏÀ:·“òÃwƒN9±áPæ>ñj  t"z{ƒˆJå‹Çž, G™´Ô{É’\Øý» D:bõ ã€}5ŠÄÉ#ˆ‡šÄz¦RÛñyãÔ•ðSêiNÐ=±««Gñ/Ê®ƒ7w¼TjÐÍ‹{θi8|cRð‚hE¢xé9/µ› a|VÞà[!·©æ§KpfsÕ×=ù´®@¨‹`Q|CJ»ç>#ˆå4æ‹xUÈ o7XfŠÓLÙ?ëj:¤‘ʺ׉¾÷—NGïh]¶Èc·nä@lþÚkγqxnÓ "\›/ð¸’Ä|Á<Zà¬_•¸”ÄÖGAiâk,kG-†VºHSuÈUX# oÌ©#ë@’IæP+¾4ùëHE/ Æ>k“µ%gT;ÁÃ3vŸbí½UÇž¢âøMÄ( :–”ÅDÒ „/ÅŠ»ÚS‘6MÀÌG{‹ÝÑð<ª­é•JÝã@äç|§—ŸÈ8–®ö:ødŠì[ùç¬ïªNj‰{^}¬g»^­£‚Iûµáô¨2Òu¡ –þvŽVBŒ–4 šc9(<7ÂÄ'¥çÇÅmŸ Š5‡–ó&´Ì’Ɔ1&ª/¡®@ÕËÅ)¬µZso”ãÈÒ†¥É&|ó„%k_2ñTQ-¼zÔ4¸n'!Ñ/øÌ²¼ îelZ _PÌïŒm²`öýÞ±þ†E~ nÉhF¥ìÂ{ÃÉúalsÄ|õÑú„‰@Ò€T & ¯«„÷-É]&7T¾_toF⎅ˆë>¸ÞØ2(€&¾,C⑞·‡šxD»^Í3Š~Û~äõÉ~„k ýÚŠíçÎ@ Ëâ»T•jH>¤÷&€ë⫬ŒÙÐe Ñ‘!* 5Æ/›4?PÅ´ îeEÅ£Tâ•Eè…\_ê„♪ì eÊ·²UÐàÿšÍ}7{¸tÑOné±tR,ÄŸ»S’]‚;,˜G‡ŠIlûã¤veðpF»Èž“##Sj²ÄÑ®ü-ïëæÐ¸÷ÖÈÎ^Þ3OL`ÐN”¢âøÜ3¢Ü,ZMËÞKúú7ðähÚH@¹ø}¾í}Ékm”-øúòa—ÇÔ/è‚™T) "…åoY‚Žæ—VUioÛ™Ô{z~ üñzoè¯èu#Ѹ×ûïÃT­o3.›Õû’@—øß™ØáÒ®¨]ÀãD‰rµ0ŸŸÆïÖ¬Ÿ‡•g¡à„Ð<ŠºVg>>SnxÄ„Oý¾bÌô JËÐy©¹% J*jWG˜á!)MQŽüÓ%:ÝH—KW‘ã­ÊŽR‹èÙÇs¨ô½qãßäT)̹JœüŒÃ€¢Ì‚Љ¼o»oyðo™·öT˜õÊE†+ØoŠÂŠæN­Ubï¬B‚¦ëïÄ’KÈ@c¥ñάaÁ[Mè Úåý—&Gu©ç žÆë‹¦KÍYòŸ¢c`—ÀK¶•‰æk(í]h:K¬¦ŸúëŽVf3-Ã)¨¥Ù—Ï\³@IŸŠôò:lƒ$ Ü!×蟾ù-J[i›@Hùz†sCóu`h‘4¤Ps^ôšhµlæ+'˜žèfd¯~ÈaçwZtž”Ï»\~\›à Ã, ¹ë(’tfŽí^E±íçdŒù Wú4 Ñy'9$É,ˉ'emEaP6óNÏ:5Fø¾¤Â:ù¦™žj¢NÃI˜= Ió¡ƒÄ»”bE̽ø…mú?v”I$_f>¢îG9蟑;hPbŠ_HN,[#–DNs½-–Æ7Sú á ¤1³­¡µï{ž'œ–0šZäºäæ9ö)åjM4¹ö@4͈ŽÒãºÕÊqoPíǪ¼€¥Ä¸©{î]ÃÓáûþ¯fÕ·Ôúê=gjxú]”¬ç=«CÍk{­Õzn“ëö# õ2“Ú8‚¾s×Ó¹‘õPÞ!Jí¾Æº9“Vï+ÿnDCè’ï+a::»èé?: ìoÿÖÿ·7¹XùµñhÕ„xóqøëUc¤µgâ•Fy6›ÔŽìœ YÖŽ-ÎBÐAaùý†€·t+ Ëq¯ÕΗTöf-¿ø8eìÀöã•3³dê\&±" !“ŽT˜}: ÿ± Ê~¡•¢j6«}¡†³‰ÑÄ_H;-0Q0øœƒ‰Õµš4ñmOš›XeÍ?GK¬þN2ÕkЕÉj)‘ 'f÷ǘéGœþ5ø‚š&²jÄ’=úÿÌK]²ãÙÚd‘8^˜rîܬœñàznžÁ“N1p ²}Ö8A NÊÀ~YÂýò¬~ôMŒ|YÏÃ׳³ø–SPWQh–XÍéíËÄ¢zý÷¤~äšwĆ+Ž·ÎcËx§£š/õÐzŽ9Éß‹1=ÛÝÃÄò–àª\w®<ö…ê.ž;”K sPGîª¶ë æ~QIÔb¶Ý¶òdòG!¨DÚ­gçœn°me:éâ“Å©“,Åh«?öeýntŸI[èfÞÝxuN¬@ֱǪfÐí)ª>j¥çÕ1uÿà<¤TÉÚ{ÿ`UUÌ&Ô’âð•ë -ñ=(ÓyžÄ ½\•|Œl"g}”I(€™÷gå·ÛËAÚ@QC¥›Gªo}.IøP‚«´ÿ+ÿbÃÌBbÔs<Ö)òÇÜX\C§œRMÔ£åöXÏüG6¦átŸ³YÀù[ 뜵v aIúÔ•æ3É’r£mï„·5yèìQJq]ŸÏ§aEƒ©\"Ùæt_ÊŽ©€¢EädÐ<ЍˆøUæäÝ7/Ü>¦ñD¶‘HN¿îçî¦ÝdÀÚÛ«0Ô ²p}wz<ÏÚàc³¸OfâquJH^N­aôq léÙÄÀL¢bôî€á fá™Áá2Uç’ž Ð’Éým% ¯bØ`D†œ5FîÆ9‰#úRšªŒ ¤jóì@Ó£éîBñr¸ÀÔz‡ý*QóÜú›UÙTM]hÓéÿó)i…³ €°Á½ÎýÕGýò9g}JYž²h‡)nWºAd¹¿r’`ŒËÈŲølÑuÔÎn$QLÎÑlÞ‘Á‡;Í¡Æ=ËÑ÷ Œ™rxÓ7¨NÉ)ñÉ)ôõO䇳ޖMêüÀ ¡ÒÓùã0Ó‰ä§jdRÛ½t×owo=MoF@ÞÂO HØÃ‡¨X›"W…Hâß¹‘,Òl9vBÎÁïâá(ÐÜp6£*wÊé²"«é—(Jq´—ä ùé\¶Sßçø9ƾ â4éEB:€˜n¾}ÈnIcd²UŽ|o·ÏajlkdcƒrÎÞñKE¤€xúÿ ­4³¨½ÐUÍFÏÌêõ óÀåL Õ™Ë ŠÓ¶ª¯F¢AE=G6jÔT0{2qsßÏn™ü!ö詳ªq»J¦ ²ÀœöŽºÎ¾áI«IL|’ ‚Ÿ-ó´M‹2qjžC –û€Ôð6âºo¢à÷ÔpoóAò"·ê-ò~²(Õšô}T fg0„CúT¿Ða¹w@ڣ퓬ŽvPÃV>5±YÔjô¨lóýœCéb¬û™ÝO­ÐÔ'.’ˆ±å!tÀ—1¾ÏjsLIÏ4 ª«eáš¼¡Å9E£ÿOkzÏ4Sn,VL­œ²"í>{Ó†JÚ3¶µ’0u%Moa”ž‡JÛûQ «²GaÀ¹h¢Zø kO#»N&ò©%¾»2x†|ëK†RnD‘jXƒ»!qÔ$“KÖ…ê¾qùxaG×-=WÜ9èÄ' |÷I•TYåùþc´0µÔïòjÝ+\Òx*Ë{r°‹¾ýuaRf¡Qþ®Ä/“ŒzΊ©“ÊÿXÜ#Yà®k‘ªž¢n™úàÐHïÑ õ$µÛZÿ)­ænH?s›vÜ¡ŠÊýÜCrÄsU=h¦w†{Æ%±âG>š%•¦ªÈ´ÀoG_ç` fÄÕ;I!‚G-*5c6¥³¯M×p´¯aÃpJìs}.ŠAyp&W¢;Y(òùGOû£:_¡ä|"_!H×GÊç±Z•.¼^Ál¢})/-4:C6è™,ÿ¶ß­W¥ú/¨+aÌTé†a9šÿa˜Æ >ƒP(^m_âúøs²`3ðÙ8üòÛaý Õœ…X,ESÁhZ)ää´&Ø<;Í&OÆ%Y¿³Æ†å¾Z•Ùz6÷ÀÃÇ>‚ó`úp…S¢ß\ñ¬ÿb×$Ohr ]˜Vü®p6ÛiöÌ1ÿÉÌœ!ÆÇ<¤ãxAVA_ù ß’3/íeÔøûž ÐO*,|Ó®nUßU)»ÓÇ÷䫟Sœ½ ):\(Xüe:BÙ}žljñ+LµN)¸õó³ …!êq‚¼‡qu÷N Ei‡˜½W \>B4!Eu0Ö(¯C$÷s¡¨ ÞÓÃw¥}Õ÷A.œºÞy•Ÿ¸ûÒÞ¤¢»Na½ ÑœÊÜÊaáÏä2tâ={<€†^^Ñÿ€oQü$D)®‘4»*ÉÙF×@B¤`4Ôg¬™ö¢‹—« ™õ;ù+,¥ø'»L7Á &87kádÿ´)(í õ~ó—Wy2Ù n !È:YK®p¬XªE\!–óÄ*£%iúÍí­pOu{铜c9 §Á‹Ô±”FåÇÌuõ<¼ú·úh¡8c\¬¢Göûíª#¶• ®lü3¥0{[„¶¤},VYY:]y~XõN±Co¿}7 ÔÁv?GsQ9)"'Ñ\ÇŽÀ'sµdYïLõ^ðYõ^‹§ „ÛHŒzøCÃ÷““~¥Æy ¡òùâØ‡7œ@F.™bM™%¹ß›gD•:z‘è°Æ€öí"Û°Vì@¿;_yèE÷À - ;&ã2¹ÞÀñ“g³™& -ˆƒ®ÉøÕöZ>lÛµk¼ÀgwÀüÝ9õ¦üE¯þÅÑEè¡¶žÁfXxä×Taƒº 3)Wé¦ggˆ?WyñÂn &%f¢¡×y‰±ð1šaC镟•à¶j.I ÒqØòkÙêuCómn‰rkÍCLÚl1øµï4ÁlÑc$ âî«{$à YJ¥N"ñ¹¯ÃI˜6ðÕ­åht0§ë±( G`“LÜ'O ך10 ; M%Fb»%")Ý+%zæ¯;ð MY$ÜÞÈܪQpnfj’KÙ3›|O“èþÆ…õ^ÑÆ 0ãpˆ=ÏÑ» <]iwÌï)™Mªb¯^¢×?Ój|&A›Ax9öó·k9èþùڸ÷´ os¼¸ÙX¶µ„‹æ¦ïOAx~Y¤ÜhK›xV}tì“źr '«$ûôñ6%¬ÊºõIÁËô.nHÀö`O~EüEpÊUè`n­¢çø@ð= œÃyœ¥ôl¤üÞÒ ó=""Õ“iï6\òá´Læ²320»žV3‚®À#æÞ%+ëgâ¹H1êÎ8*´I›{²“zDéf±²WÇ‹ê+£è™…uÖáÍ¡?–]Þ– cý:Ä%ºþDu\LÜ[Ëä¨Ã¢_;q•UðЦ­W¥ñ…ý}Ýê‡nG{íµ*âyóÞƒxq鱸øŠóø54B¿’5ƒåµ*¨˜Ñ1Ú ¶Ø¢kó‹ï ¡HÀÃpþÀ§kn s|/6Ç ÎÄÆŠ÷?ìàÙbPÓ°…?-Ô&|Ná=Ë͸º™3vºùÄÓb«úð"§jܳ.>+Ia͵9»g©èÐXêmµ‹©¼RÐfEû‡ˆt¾ªÐm€ŸÏƒüg¿<@TЩ&~g²̃Æå¥µc£c!›Â_O}“[lù R±h~nV0¬›ƒH_1yÀcºÜD·&榙d±©éü€1†XO’œ¢ãfÁÎFlxëk!’Æ£—9-ýƒw7|kêþJ„ aNEÓr]døä‚ÂünÞI„ëÐ;ÍÊ](“œnÁ9˜1qdïyfšfäŸ|Ïž®ŽWœ‘ ñ2 ¾joƒ}M3\ NÁ Àéäàw¤Ü ÞÔú ’É5¸˜öÁ!£©–ÓÓ˜µ–ÔÓ{c¢WkTÍÖ¾”T¿3€×è `TiõÈ}¢£ò&»?£pV¤ÖÞͱ°ï·VLÈöbÄ-˜sƒéKÑóØ´"å<2~Ú. fŠÅ™GïAÂÂÂcgš¤cCËE…„ŸÝQÞ)±Gø·ªNº«bLq›&ÊA1[rý«˜ àÿ]¢`Z/îÊ’”ȘJ‚F“EêR¿ðÍ€ôž^7¦ÍU+Ei"ZÃp÷uš€KñŠ×=aŸ(1Ú bFïcÄœ¥ ƒF_W"d`.ËRPö-JNüÁiÈì‚8Íô3ú°EäоÔll5uü€w›æ œýœ¦d¥¡ŒÍ½ó>€cV¥gï*âNl‰uQxôž·4’1ã+tò¯ý4fBøbYúh÷)ͺܲ†Ç¥µ €4èüºOðR‰v=_X%S«&*õ\Wvë8ÙÕțߓ³¥î&„Q4"kaØ÷£zçú»õÕæ±*¼ªñ¨|Wx”l  v—ßNçp˜•ñŸÇ^yE’ïðúæMRì²OŽu¨–ämŠœYÈõ£·‹]Õ¯#ƒòŠq¥$Çé¦ïï”·VKâß»Œ·£Ê€hE܈‘·^Ü+…®a'wº <X^ `¶rÍþ0‚Ô+ ßTðã¡Ø' ÿÝd™‡‡£wî’GÄå¼aÍ3­¸ç¿k¬O”M¬ÍIxn;þ8ân1χ|èù½îYcF(ìø$5lÂÊ$PãîÏYõë–š$´ 蛆!€ë(Ú÷³ó@àVçÔ©6˜ÉÖӶצsžÚ%˜pàæs ŸöÚ:¼eXk;R C=AÇ›Ö}OõSÆú–Û2hqŸ xõ"p ~¹5þæ…Æ“YãOµÙèrš wÚü§«ÄôÑçAŒa¯-=ÀÛL†4×ò›ÄNÿÓm³ÆO–çx4ÔbMù=(õõá:õôMÏYmÿfZ‰æjC]v+lšìd Ÿ¼5¬ÿŠ.ƒEZ§óü0väËi‚¬KØt °Vwƒ˜¢V’’®)S'æë¦‚z@¦¦‘ï¸1#>Ý‹تÇ›NréP¸Áˆ¶ùs0ùÔá²1œ OÐ +ÁÌi+ì°(°ßuŠw(iØ=ó’µPþÀY ¿æÛ¢—SuÍd²§%§sßÔ¥$àʉÍŸ®+[ãÌDô›÷˜¢eMvéÆ)m«”£ÿë¼­îN!)´3EdP~šÿv}ôÁ;È›7+ÿæ³J¡5­ZÄ2–ˆÄÍS|‹¦¡á ?±>±y°O«f" ’wyÂtÞŽ÷|ÌúËöZ3Þ9i'êå¨ÍÇéŸõ<åÊ&ú{sY“¤RèíkÊ“¼¯3Öûqk$-d ú3ñüÐI9þI'Áít°ÝP£ÙîãúšGDÃv’—ô êp"` ÆÉÏÍÄ juvFç !ÉÞÀð¨#V.î¶fi‚Eýö=ªµ¸Ú ©ÿ€H2b‘œ5œh(G eY[uJÚrzØ ¼£ÙL*<4FŠÑ“xéû*ªý‚¡jñÛ%@Uƒ™Œò#u«96F§a‚³»sõB;†uï€Üžé µ%ùì¯Õt#©<¥ñ`Ñn·ùuu="²:Íî n=4péaÝ’VÕ×Ù”l %>·“CÂWUûB x´ÄØ„웇¡Ôé­`;ØPu<{Üþ’Š}7Ý= ÏupBvÀV¶ß,´^÷RRÊöáûK[Ô;F1«ŽÌ×m<ßv@uQ”{= 0ye½ 2ÂÔå…kÒsÎ`cÇ+σ qÜeØE,„+­i?®Á °’Ê5çT™)ùüA£Î%æŠÊàý¯Ð v‘WZüÆ(ø÷CÂV\>s RjÚý¶f}÷/¾ “ÙTÌ$)䙘ž¨FÑ ïVPc͹T6àóû¢p®)QAc§«Í vÎÚÒÓä¸OöÒ½Âͬ¦1‰©ë÷î˜ÐR#f´áž¸Š‚Ëj¶J\Ÿ¡©p–q²z7.¸),5Á¤´_ªAÔÿvbIÁíšœ'‘XÆÌ£T­-—F„/ÖÂ\¦|žÅ¬ Š1pÈ…Ö*4ÅÖV(.äZ‹5Á;i0’~QmÊÁãjɦÂw«y‘ÁÔ§2㛥û¦KO %."Nð ;€/ÇQ$¶#ÿ¸W8\°IÂ…ràƒÓáÁrqå3öø?È[BD$@ÝÞFõÉUÐ!ï!¹÷Ô™õ<:ÿ>þ,dIËõª5,ÜúÎÝðãÄܩŹϮ|Ù›0˜æJÈž@HúŠ!9/¾‡»Ý+& ž‰|e‚j²š:%†V÷²˜´'ßÕ_eòZ·è‰—–<N°ÕA'ø7Žbøø ŽéN·À&m·éyšß¨¼$;vþg¡€S4¥¨9 ’xÓI¬my™”Ò†Æöþ!Ú*ùgõ»ÓØ!Â<0±V{âz¨@‘G6«ªè/ °EOüáÙ'Nkt ¯?q÷]v/½_õÿoPv³lp±¦Ó&[úzèËé^w¡% >[6'¢o-M½tš<%˜nMSÌ9Þ2·*äÄ/FOø§Ç½b<ò¯,¶ü-…?ßû†ù Añ·À¨ö_ËüôKÿ•µIáÿò“6Ó%u‡E {mâ}+ÒnúhT75ñ±“ÆQÁv†‘—Žmìæ7²ðÍñðz+‡ˆ(3ö=½Cø¤‹ÒþDÆÑ-ræHÅÖb½øIJªì~ég¢ù@Ó[aLg}be"‚4%Ⲅ \OÁÔ#?ÒBø%‰å~‘ .Ò±ã\èœ@1ÞàC‚ËÐõ",­1å†C¶Ðº“"ö×ï0«áã–ÞLó™ÂR`†…*ÐÀ‹¦S!Yéó¾“ÐQM©7,*|îLóÈu=œ‚ Ÿ–&dΫÆÈÕ/|®öæ5PåÀ‡ä Õþø¸ÑíL€Œ0fìMONjw”sEÊL.úZ‹1PíÓÖÊ Ôxl†0…±Üm&1nÜO?ZÆ-#.¨8áZ}`­-Ü&Y¤wÅO&@Ei-þTP“Þ~Ë‹@Q^󖉨Úñõû¿Ë7XÁL+aœv ´Ž^£‘w^…îܵÐ"Ïà«C¯…"” ÚkÎm†á…~ŒÈšd_çjã@s–Ã. ÜDÅWhýšÂ@ÀA¡k&WáD&7̹'Õæ»Êw4²›ã(ÄA1:D0ó„^¦F¡“}<É8ޱSº„,çŒëvEâWˆŒ¦yE'(J±V¢¹‚°ºi›ÝÄ^eÏèYò¾c‘¾´§êpqrÿ$þs®B»$è%–ƒñ-êé–ség¼Ò“R¤ÒBÔíÂù"ÆÐoš7‘ºêp¢£×ôϹÐò „Á$)Ð/«ˆcOþ²<NÌÆÝD”¿EÄ5$5¦»HÜ'S•Æ'”¼'ÑÑVö­ãÜBS®ƒqšo‡ÍSÃÎï‡ë€+3ãø˜‡Mê.Ž~§Ž´À/ìY~–^ôT¹âfEò£tÒçà‚ùõŠ-]Y˜Ú-:pÊT¼¸lP”•Ie‚P]î\¯ÖÃhaòç@â³y7·,’Uψº’ ÂÚ“£eÓÐ/c‰áÈØÉ®=c=2ŽùæO½Ü¼èÅí½Ì—îâ5[0´˜Ð&†NU€ÇXšgz0ì‹ J[ãOOH0æžñ‰‹0µÜ¹ñÕGÈDË ²ÛBÙ™{Òƒ þF÷µ ¼Ò(,J«s*†ú.ì\…q]ßÏÂÌ`Zû…†Yq-†éQl¶‰!¶w 9Š/â°±¾Œ Ò]}Sfžõè=.ó7B»¿‹bÓËç Þ^¿Â*˜7ÞåÄLËæ%2Z›;%;[ca1‚Ä×p¼i°|¬l&ÎeËïŠÁò©^ך¦ÓäÓ7¢5¶—¸ø4–ž9I«Iw±{©bÙX¯ä×# È|¸•d6Ú³PÓÔ|šž¤dKb,jÿý HômY„®á>Adòp÷ÛÝÕ¿¡‘56à¦ðš‚õÓ‹,K¸£¦RÔ˜í16q·HÎÈ´¤kÙßÊB¼kŽ›Ûâ1Y7«×Ë@,ZÅþ9@QÁ…]R‚·lÝ{‘«ÜfÏÀk/Ç“~rTÜM±ôàQC2þWºr 5íÀ4Ü|:A ÂGÈøû5½ü¬ª˜º ˆ4l‰¨†×ºÛ° ê;ÜÿâsÆì_šíˆ·ÿNŸéÄG¹§\Cg…±4øÏRí‚æâZGhdà [x/ã©XŽq¼Æv¢o0×ϲôA¹¸¿õÕb} à}¡•¥Ç ÕÛs-þpä)rô9CkÍKW«ú†r# Âè·Õ\XÈ&“.«ï’?Jä-_…®(ªÀ|,ºxsôß1’&éûŠ£Rø-ù˜¯„ç\cH²–ÑüHèÖŽGÑ›ðG1͘aª*Ü—0`˜%ƒûIßh{‡51&jÞóÍß,•ã“Ã<üŒ½‹þÀyóv0+È€†ùôïƒìm(WŒ|ÅLóUq‘ Ò³UÁ>+ÌIv©R³±÷èPâQfÂÕ€ªŽmD"™à]d4¾)’KÐmób׬îüµ‘Û9ÏV¯tTª/@2`QâÈIÎ$’Àè´!‹òlr¡e íth¬Õÿ¯1Ù4$o§S£QRk™Œ+…¤;œï ç<•Ûd°•¯øõ dË“Ù"¹ÇåÄØ$!êB Å‡ûŽ]íîmä¥Ó7zÞy¦8Äã²ä+íŠFsä &}ò6WÆl Øo„-‚I/7‰6šD4@Ñé˜‡ÔØªñA•¼hY Šox]£.\ðŒNê³Êéèn´:=ÙÈ1Rß«÷ì%ùz‚L\n1&‘GA³»ÄÌ÷F^ägý¥WyûuW×½ØÌVajñ}ËûõM]K*&ì}-¶˜þù"YÉ૞/­ýQv£@MôáX63µ\ùÇU±%5)·&2~|ÌpÅ-MVf+Æ£Š«DÐCÀÜÀH³³•”`}²¨÷™¿x©vW”õÊk¿Ü/êÒZÞ[¹õɦaÕHWø7 ¿V½¿oY/S`<7ÇF´^iK6S€Û)’Æê¢Ê+­ÑÖ‚=xO .÷å;ÿaÂ2'Ì/¼´CýhˆqeA¦ÀõhÄ•q›Í®f ¼ÅYU%¼¢©öJ5bíEx3 sb§„Z4 R÷·Ô%¸á—@“ŒƒÍ©®Å’Š@0ê™”òÞŒF>Óh;Àd@›…]I×óÅoýtazðØ•ïÜá-cÆú«JÂĤˆžW/‡‡‘÷9X[" `ãŸ{eòИ,RÚ¥AGÝ`A"­Õ #j÷v*·yÖ'ÿ'ïÛ‘Á¬÷ŠŽâ·ÚHßföo¦Ãs6"¸éݦD^ž]ânÀ¢¤|€·bmёŽšÂÔØnîµîÞƒÀbî,ô¸ÂÄŒëü¡¹ƒ;N™ú´\ Òƒ8‰K+Ac•v8Õ`òõ«ä‘ŒfÜËÃX/9Â÷¦5ñ“Óô€%ÓË—xš‚^¾Ílét!ñ¡ ×Õï±ïZÕ•f¼J¶¢%ùûÆèÃù„Þ—…~$—¦±Ë@¦É€…ÿÉ?×^CÔ”ï¥t¶·Ò´÷Ì]ïþ¾’Nƒôfå?&~"¢‡ám“ó³ïæ¼9%§g×4^cJ ¿Ë,€´ˆð·ðòo,h·Ù$áEf³Là‚ã3/Ì`—òZõ2ŸíŒ1u+ü¨:=CK$ŠVªîíNšõç.² ”ÐßNcÇÓdŠd±sü%Óó”£ Ï6…££]=r©ÁØmNØV§: EŸúͮ˜¢=ÿ'šè™¼1j)Ôeª›µŽÄàp‹A¯8@–” ñ…m]8È’–6‰YLÚ+KVY€*|©Âëè"ÜÁ¿Ñ• o%†ô]”˜ùoQ\—ps÷Ú?e¿Í‡£CέçO¶¤¬Ý¥†¤µæþd¾u¤Gaµ~4€5|d”ý\ 6¯}PÛC½¹h]Û;üëúÍ|AøgkdjáÊ´¾âDÔïÝx`ŒïÛè‘S ÿžÀFaÔá b¯<-º¢»g´3*Roªo¤¡³·öÍÐߟx=õHÛvÂ+³' ¢e[) …~˜Jp˜W’;Ôê½Ë#–¨‡×S=7¨@Ò²ŸáIñ!Sÿ{™À‡-¼ Ý/Ð]ÏñÿÞl¼u‡ruû%ï˜zí(˜®;Y_ݧÀ±ÍLCžj¿¸©µ&£?4TzȺ讨j©ƒçºÄ2#*-æƒ4‚\bJþÓ2þÙÈ‹¸µ„[|qè0EKðÔ•*ÆÆ;Iƒm¬t´)J·=]œ¹, ‰Þ“5§ißSH¨XºÝÒòɈãðgù²³_•‡˜ëbcOwH“Ea†uŸ£øŠJÃøÊWÀëÖˈ&¨`<(í=»|qòHp~ >ËGr£jž§2¶¸•[hÛM3”xŽ"RÉ1Å ÙÿS²™ð"ŸÇÖÅÖŸÑ×d3„2ɉ’ïèÐÒÌm¦^÷o$ÃoR2„iœ¹Õ|—œ$ä86#C2]Î\Ì(…)ÆGñ'û‚Bcw’Wk]âGžô{i`Þlj•[ª¶;iùÉDØFy%o3ÍŒ(WÀŒ¸Yº?‰IOÇÝ®a;„Ç%V Â®~îÚ®"ŸÑ-ý|C*-0ÊCAg¦g(©M à8¤3vXós†é Äy}ó%<„î ªô±àk5(è´Î˜Íö²ÅÜ¢ýì%:ÌuM‡ÕÝïƒJ<ßM`¾ZKÛeø¥‚²p"»63=§&÷@%¯8{‚¸”‚Õ¶[ožž\œºð'BÕ ú&M82–ÜùK´<Œ\ÑŽøò¡×³pÉDþÌ/bè bZdp98Pãk“¶’¼*¼š4Û+ã’¬ Î~8‡µ¢§L0*9µ:Ðáfº<®¶Óãk€Êýκ„¡æ8_‘ع¯/AAžn\x·GÙ5ÀÕã}xîÖâD°©,Œ¼Éš.ì>¼?ÍcÜcë:§p3³ ÉiùX¤ÂH–L ŸT]‘»^¬r@M›Î m8ÎÍÊ'iàe&5 #7ªÍ6ÈÈæ£_µG0»ŸJÎ<Ëq‹ˆr Îrû1#Üb\ƒ‡IZÙqô.Èôê.LÆB³aÔÂÚ‹C•é€q] ÚîvqÇàs9/b€““7ÑÌuJ Ì–Du•ðµT§‹3€»¤XL à ˜Œ3ÑÈa¹o"+ìÏç-;Ê)Ü,±‰)û󷤸°Œì"Ýc’RÕ Î|¢™ÊÊDÅïŠ\{…£¤: šâÞyó¦À8§Oçù.¯¿¢‚á)S&åÖÿ‚óã]½7ÙÃÆøIŽÁE)ÌÁƒKÈažFKï“3ä‰Nƒõ}ƒRÒª©}øq9½)k[' { ˆ š} `êN:N3gˆ §XÀâ`…¼x Ø«¤^!o¥–ã²²ðý}Ä ‰G€–—ì+Ø>¢’¸ö0Hÿ{šš]ˆkjŽ®üK¿aŒY­"‘û¬Ñ¬Ã¶}„xÖ—*þà ¬s¨Æí€EÔF˜~ ˜yžDK˜ ˜Î(ÁV‰ïÒÀ¦‘BÊŸÔ ÷ h‰(Ü Ý‚KþVWFDIqˆí<ÖB¸ÆøóG AB‚–)GÏœËXM¥’Š#‘u‡ ʹõ Û¼£:©yB5ññäFŸm)B¥Eû1úöwrý.ú>LŒ÷šÒ{˜úâLÉF‘_í V386Ës‘ÿ“Ð.fðÒžÿò8¦?j^"Í x¢ñ®ª/Rqid#Pl¯ i‡ìÕ«½¬õš§·´3òo@г!3ßJVw°ÓwGÌÛxß„ʵ¬²‹rèæ1E/$ 9|tc9&á.|†Ý×ø£mVÌ“Xyü§”ñßYÿ裰§' ¹_ª»²©DÃ6nË:IÁ(µºÂrÔŸ J‹D@½—*3Éb`g€÷‰ `öY:Õ2h¡c‚)&4®>Ê‘ŠýåÁ‡õ0ÛÜOJKþqq,%¥‚È“>÷@ÝBÜ8aõ`áÁÍrŽ,µÅÚ£p“ rߨæˆ ø å9aÔ5±»e¿)á•Êÿ Ù1ës—kØø_ùGçKãVÓJëô ×ÏN*]­|Oñý@#ÈÝŠªX2öj Ohm¡AçöÅjRZî}—ÓßðR‡¼jcIdØNœŒLñ0Ä…ÄnœÝ8Ø#j'\öI.,ß¿þÈæE¬ ¤O’~+R™q:üa¶j8L%;G& å¾»ó¨þ]ÉkÀ ÙA‘S¢\å\$Ñ$'™µ†F²%‚ø:í'@ýmP'aÛç¶åC™ykãä¨Î9H|ƒ]ØB°!=ÒZzBÅU"\lV¹g¤&¼âßàùõ“ÅFÚ ZòÛ™Ã-ÙÛ“J;R÷וhŽN„ª(¹¦BÃýáº('[ÆRôfW_hzŒ};Í’¦T]Æ4\â“Ì(¤Ïò„j°¼œžWRØô¿;ÁBˆ-ו©XD#f;Ýî¼I#×*³6Ò(õŸü ?©›–ú­Þ⌠ Îl9ôR`¶ÊRÍ4Í9Å'(L' sÉß“ó_‡ä)¸ÄÜîôÎ@iÿq4s:ßÿ¼Â1‡K¨Ó ¥JÄ.8JÖc» ÉÎóýæTÞ_q¨ $­O]²Å¹2%ÿFédÆqP•² 3‰žÙÙ/Ù±¨›cþ‹€|Qíq¯8©1$}Ôò,pÞOfEãI~°æ.\ã °Y S:O7ˆ´ÏƒýðSŒ9XŠÆd]“E+»+¹gµÒËP­ åßG½ä É ´[¬øTº‰] sÓð5Cû?—¢V0øñ¦<îÃÁ«_£Uæ£pKµ©XÂà9£„óÜ‹Bò` kò¥·E:~iDÃÓ‘•¼^dxÛY-QÐ+¡ ­•D.Èà°å€+µ\r“$^ájL‘1]ÇÊ5µÕ#/âr…dÒ£Š)ü|˱ݰÌ]7×oÌ)<)ÉÍÀfk³@€æ]ŽXpõe_ú,8þŸ÷͆i6iâ1ѧ^bàÜb¿vcï ÿ‡¨¶É;g ©¦‚ÞË—Zˆ*óŒZš”¨wÞ‚I<œ×EÂÍšÓ"L¬úwêÀ«å«ˆÖ¥+ìWtܽ‡€{?s‡ŽQCb9Ø|¿8 Ã?(s›há6íë®bXËŒ¢40f[š°Õ¦Y¬|å×½¾ž±d‡ ŒèøMWD3äej¾’ŒÛ%Y9½÷=X‡¯ˆšK^6mž+':gž‘ïϪE«ïŽ“è¶1‘ º­·ù°Ãº¢[·p£œfŸ37,šm–±“Ï4– ¤Þïª/Ñ]dF¶ª=[%ô™3šè")¡M¡nåHc{0‘IW²‹”· ­›lå¥cÀÛ‘¼×;$”™ÊöqŠ ]þûÔõh¨"Ùi™½ªe†$×­Þã§ÏÇ^þðO’Jj§¿ÿH³Óã úh&¡R2”kï{ðáõ´»x:X<íè1ßã8 9ea€®)|&+Æ Û•{¨$ÆõŠ­X;}{Ù$›cµ¾ññW®zÐ&gÓÐ y¼d³.*Qð¸þo|†ŽÅ_J-ÃC-z€\HF5ž¼Ón5%±y‚ϲ >bæ=í"H­}^Ë x¹+âÎkß:ž^,À‹~:dåÎfbõ.Âj;¡ y·Úk¤óÅËOµÖ¿K’¯Gi¬>R1‚‰È]ío¿¿`õîp P”(ôp·ä—ÜÔŸØ%Æ)´ç:¾ðõ³Q)t èífá÷¥ô õg¬ò—xÝ3C¢ÙsLö%(h¶{`À'5¡+Z+7匷°nºB‰G<'2vÂM¸Fbœiô»{îÿ£3w–ð Ñ¢Üѱ*î?e„D/ùž ¥ñC ŒR¿b.àô­Ö—£÷LŒ”{vz ïÄ ~;èz½åh‰y™Õna‰p#0–&ïyóF:&±Á8ƒÖóªì÷†á6é¾XÞÃnêZ‚¦{yX›º·R~ÑÔ’Ù2š C<ݶsÿbƒeÈÜzÅ|/}Bup°*:Pã1²o£÷IömL…ß%J,Q§Ü~=7Ϻ‡Ÿæ¦¡æ¹§¤«Z*6®Ñ”3'pæ†>¡N©>Êü€ …‚{‹,€D‰…ÉÖ¿Ù†·êƒ)á2’œ‡,°1÷ñôö2hÂ\ž æ|ëÛ¹ç%Á¨¾LúÎo­»æ¢þƼ껲F,*z x‹Fô¨j® tŠèb ê.”"*ЍÖó³›4¾“á#ðµ0Ì«âŠ7C| š›\sÑÈZn0²ªRÕJ¦må/®ÆL™#þi , _îÎÏË °"Á¤æ¯ÂtüV$}çÛQ;_ê\ÛNëÚ?’¨0j¯ƒjhÃi߆Yˆwän­ÑŸjÆ÷+(ÌVÑC”$Çt…LˆÆ6ó†lF± Ãoü²É¶eLé0ÕIu~æ(o®`¤¬“wtw4W+¥"à½"oeuù«×ãMÌŽPêÏÚnÇú£fÀ½ÀŸ$ˆ8Ø479œ<˜}·(L,ÛÌzQ@TK Œ)…˜`×ADc‚‹Ú_„j›§oaÍ‘UÖ˜-t›…Ø«f=[˜ õ}ïþùnüïÑZ< ùÈÛŸ>Fë–W~ë£dÌÙziM<§Ôÿîíÿ¦LÐ$)Ü€f–ðÈG×¢”9ÂõG\"¨Ìž€wð$qW»_ÈØ—~ Ø<øq^Æîtã‚+?ÝEæÂN²à‡˜Y õdÿ“(ˆ]1,½ ¨b˜æÀ¼½5a5,¹ý]¾Å/û ÓW„z]˜-Ü´ ^_þË“í[h:´µý"&c¥ï¥,Pa;Øpœ÷šê\ÓéìϔÀæý“O„šÞ^Vû `»á,é½¹IŽXÁ;çK•"?Ø' oÙiãgsdNµAª£Úž²ö¾~Ãò‡¨\E¬œ:b®¼—ˆ ŽÜ¨9=;´Ÿf’‘V]¥ÐYKr„¶St–Æ+Ó k*öhšQs/qÅ=ºàðw3<3]$–µûë‘+Õ€¼¥œ“̽˜=MÊPƒiˆ0=/;Ò™Î+íuTd,{’…Ê…´‹ÐÜ>°s̬¥‡ì6°’,Ù~¹œ-ÁÌÈ 4F¢Výk˜!»×åo=â&Æ}ç‚¶.öÛ9ŸðàÃîä§¥/X.Ežp§X^A·ULá‰h\£bÉì øAòQ å‹ë$w¡rìgèr§nž9D¨Ò;/µ-Å™HÖçoðµ¤bjVäg %MaÔÛP®W5““l(#aˆàvÁ?üp˜va«‰‹€îâ´àòϨéÄ# ,¡Ÿ­Yxúa]ù¾³'QkEg™Òµ- ñ$|Š0¡0Òçä=´”ŸÙq¨n*þÖ# ‹Î.ë6ù‚CÍ…>ôÈ'Ú×4¤>_°eCÖŠ´‹âͤ<ä hN๿[.}˜Y9rXz­r¨®uÞœx€Ã…À–ÿ;u¿2¢ ¡´è‹QrÎX>Þd¸-Ô¯Ò+‹¢ñ W££9ˆ™†ïœCyw½Ó;ØŠ÷/HÔ׸ &˜Õ®ÁlÐm/®”¶²©, Øo]ׯ ãÛH®èÏnYç¨(ÛçI_NV Å´ue@1#„Vu·Ž\­€“ 8UÙà³\§\ß e?¹n éFùI`(åz¾I¦~p ÈIœƒò “[-wQE^LÓ»KÏ¥ˆý¿QÔ ;óÝF?þ×n@Qµ!hu;èÆ~;1/EAdÀ˜Çp^zwLÍDã[ZŒßQ`M¾„×&ÿOØ ÃÈûðq¬œ üCiuu[I¼ýAjÕMá¯Ü¼Fÿvž}$A+œ06‰ÜÑ è(tª†çÇ›NÏñt€KnEõlð¸ƒD$ðÜŒCÚ½Ú欄¦G\*ŽÓcˆ{R ¯¬odVÏ©€×¶M,Œ&ZU|ÝÒÙ,;”•â>qšzÿ2Öüšõ䳦Q&œ…ó4ü*þ*2*5îs`D«õ`ÙwC ôaßÿ©:6~y[$&s.!zU”ýG’ádâF¥<îÞ y§Š%Ã@¾¡z/ò– È–Kê/iT!A›ý&½_·°Ã˜ò'ÿ IŽ4‰ì#šâmM0[…ó¿åyâÛTmoN™K€§TJDýñã]=a›]ki{YªFUàá“Ö(b†ß¬>ƒWàŒ¯ŠX9â„W©1"Þ÷ Ï01¡Í»ÏaŽÜë#êʼnÔafAÇIsLAÿu"CTøTó(75{ƒ€ÐEü·UUÒ¶ÉZES–~à,°Ý‹jÈ)WÏ@O‹²ÏS|ÌãIkún´¾›a—ôµŽ˜{Q˜™2€#Åï"”l”Úmð±jû*<¯7óx,’µžG<Õòå9[Ø»ÇÙ†N\¦“"\Kãï€-â- (Òêþ;b ŽdÑù‘5pÍÏjû<-ö¯.Ú€=Qiáð+aë¹ÝN¥Ú…±%›ŠýȜƖÍóôÐý—xäf(=C} o‹¸ÿ%÷AäBtñ{¸r*ê§Ø-ÖpÇÿ¡6š. %{š P Ý¥|P5Ïdzގƒ€¥¸½ýðâ`a©e]Nøm‰>åT|àžæØ-ÁÂ0f°ìzKzÆ—tœšTf^Aüý øHÓ+]vlé³b;è-ªåR»VÓí¢£"îêxÒSÚ÷û>>:³ErgëÆ~ÍôÓ·‰¨«ÇPB½Àûºë>G]‰›rÝk2jóÁº­“×–| ·4Rqs‡DpªwDpÉ0½@éW6û¡äI¢š¼*2ñ‹ª’챤O$>Ÿåìk8vëáTÔ#fëÒlÔt:ôâcòzN°›Ô/¾ 8žzš.ùvîŽrÉ~þï ÄqOhÃŽ|ï:;šÔÈGH'¸/H¤¾Láæ3œ´¼ØJº,’ƒ™a´ Wgq<ꉧòR;€õãûfµ ¤EÃÒ¨Må£,opZs[øOs±šÑ›(RÔ)2Œ  ±¥Ñû^ùˆ-Ïv÷-d(6²˜aj<¢l7ž\sˆÊEBHlTÄŽ8ÞÓ!˜L£ø(¡Bå}ªºœ¹A¯ºQßK‚7¥¹²JáƒùZRИªžŒˆösÏk ³ÁÛÉ ~Ò Áè¢8öœi›z}`^¦EÛ™" éâßÔ:Ëအp/½EXý‹DZ®6+?’lx»_•_ mÖ*iî`£¨ò-JIà´óß>dÞ³—ùai×ÒÜö–É/3»ŠcÎɬh7,¾I`LÅ~Í·c¢$_d]<÷EéÈ­f¥bÒΆƒ`Åé,"ãÇfrLA-òú\²|Mb¨E½ŸEk¶làÑ´DÊÌ^ ³Hµj—¾Öô~R¼Ye5À-2bMw2WÖ½_³´gÖœR‚¿–m|oV€î»| FÖ¥qKð 3=4¾øz°‚¸‚4Ãtpõü=˜ó”È;T ÛÃ{.>õŘÂmK~ž‹áèæ* Š?XÂd1Ÿ€?‹}\ ²ÖëãL«ÈAoùžpîŠJîkw€tÑËÁþ³ Œp¨¡÷/¾lÝ·m÷M5>5§øyÉ×À¦ð›TUšóÀðJö0ÎÕ Ø¨YÚ´uSæ/8*ð~lðzÐlÈ„ †Kh¾×û®ef/AèÈ BÆS.qõi`½Bähï³w9a€æÂ Æ|lK&«uÓ=xŒK,u8"E‚ëwàzÞ§[k¬uT_/XĈŒ(l9ç{¨ÿ¸KÀS^&o‹;ÀÃù|{«æ[Ì«ƒý4 ½øý1iÔ¶!~)š•RJ½–u“L+ßYV_]o)  MN®+;c*v2&kx†% ×¶Œº•¦ Rs ŒZÑ¿¶G?—2džó€j .L¨ƒõ¸,†R?IŸG¦ßzÚ ºÆËŒ]I3ÛKp sc<)ñÂ%a‰Ê>ü*äú†1ÑU4¼?òeJßjàÆ†PÇçkñUÉ*X7•J((…K“¡5ŽÞê€è²I!–ì²ðî66X„<½Üœgׯo `£nG9¶™/á‰!?¶m©1ΣÀ÷ÕÙdÒfΈ¢në± Ú™PN¨ú¯6«4x¤¿R²¦¥Jࣵ1yBU¬a9ñ=Óú®ä¬ÐïšÓ‚çt¾ˆDéóÖûèµð†]z{Ö¸+(µ”­ÓJáŠ*¦Ïð ã 'Oüø<'œ7,Ò²½G ¶S¹[õvø1•§ P¾C”o‚a»÷ŠÆ;™âQt’¶A_ê·`ÿ,´|Qà»l(>.©Ìü œ w&ìì0A&/x¸]vGvº¿ÇÈV ;Æ9aý>=”í±]hÓS§ º>yÊ|Ø™Ñ!Œ×¯¨¤ü[Ù⢞pop‰U'äÅ‘ÄÌéÕpÌ“F«©g*&‚sûeg(£¨Ú±|s¦"~HÁÑ¥q!ƒñ3°{•§M)ž º¢úDHY6?f=ªÓ?dø¶B|µ(º*ÛÛB1…—ÑV‘ .ç¢J6ÄMž>ýëÖ©¾˜Ý´¹ºOø¿&ßÈR NûO„|øÈ„ìŠË¡¶afú˜½‚6 A²…§Ô1G…Ÿ¡ló›>f`žp–ŸÎ¼s›$NÝ“f¢h`-¹˜êîD¢H:GP/’Küd‚fë¡7ºl+z;×ÖÁÎÇžG¹oOý¾bûälxºfЕ;ȱ;Y>ijÏå³ú8€ciø ÷“ÿ‘ ©;ø±¼!•ó¸%;¾Tm¥ëÁ¥Ç Ü÷#°¼wÒºZŒ|B²^}Áã±”>úìvÏÎcã(£&2ø<(“v¦ˆ9àÞDd™<%VƒæT©qd?yö˜<é€É¥Åþz"”Ì”zfbÿµ0¦sE½Ÿôz¶ ¬3W|ï9OsÊ|¾ÑŽaYøKÛr]2UYsèÆ%\µQu›-`’d[[ý½vúG:í¦#m 5h¿¹Ëdï¡‚ ö·;á¾¾@· 2õ7öT¬1ž†¤}›ß·G­þóäÝ¿íïPœâTA“d½”:™â$yó9|‹oØ~ ]øeòÆØ;zÞ¦ÿ2ÃÂdà·Aå½LïNr#àlò;æÉÔ„Ë«Z:ÎXÌÃ…€¨ýf0±ï7gúáo2ùDÈêv¨³!ÓP] ´±å!€k%8aËUf߆aª:§ÞaéŸÒÏÄ•=<Ì×8Ѫûÿ“Ýùê‰0N^]3x^›,Q”‚°W¹½f×¶¶Ú`Áì ü:'ª½Ôg<„A;G, e-€Ÿ}uômñ5%¬ÌIPy"çØþ%žæX&œf\7)›lú]Šu­¤<\¡øý¦â¶ƒBN=ž Ö‡ÇФ¡‡©ð`ƒ/2©fŸ@_ å àdÅÓ|›Ç»…O{»Ã6ÈÈçÞèÊÉdâ³°ë¦Sõ‘/ÍõœeÇ¡Ï0~¤›•ÐñÈp2£Ì%c ’ÝŽ‚[Dæß.®ýiÈMîD)QXß·“sWG?rè£sí2¦45¥8¥}ß@eË“¡.Ï\ìÌ ð…}‹1 !¶ý>z º¶¿éî_¡LuNÑ4jûÓÞ"„°)WåKË2€íYÍx#$x{š^36©yÇ ¥’s ])`aÔ"”¨ý‘‡m_-3¨twŸ D(ß™šõ&P¹BDÓ6>(ÍðáëÜdà¸Å?8¾P¥‡Å&ñàˆên÷ð=ÆRƒ×ÝñËÆ2Ý‘J@²h\ÎæñÃÝ:qt@æ¡`íLàâ½½<ÎË¥Õ¶¸ ¦ƒëŽc&±ŸÃU=go•w4Ý ^_ÆZu+U¤pîð-•JûƒRÜ_cË9J­BåmÓ§®c0‚g€-€W»ÈWˆš9ž8‰6´¯Uçp|ñrT|rôK?"¾Å¾º­ãY­Ë“Ç äÑÈoæ×VP¯Bp‘ÀçZÉÛ•Ý$0Æ>9 Ì>/¼u×#~¯ß4ïˆÂBÚFvÙ.²œ°BA-pî{•N¯¤Z¼‘TÆi¡{h°]3ø Ba‡úÙ[qöÔ©º•ίþ‚%ïçОfDÿáó MWkŽÝT£"Ëå!z=pXðž+ì?9•¤9@$#¯B>"ðÔO>„~´NÃŽÍ>ù/ëM—îžÊz¾L!˜Ok}[?EÛõà¼y¾p£¾¸T˜B³Mr&‡C°§’ôÌN·@µRU>…!?7ÝÁb5ŪûTÌ€þKí”ÆÖ8´Ü{¨¬å¢cy}ÔDµBé%^a½¦—3[q…vA` ÷סb»¡Yå5Òßg½MÚÚ€Èßo,$Ø>ÍãO’äÚeÈíKù (ˆÌ˜knä˜öØŽbM`V¸Ò3+îâ°Ç*ÐD÷HH^)N\} æJxVÁQó—È,ÖŸLeËGigÎDÑxV,™ó¬‡Ñlf—¾Œõmb|Fù½¸Jù|go\ t|þ@8+“é0”ûæ—Uù(ô#cõnVËÌ$f,c€W¥ÀÄÈ Ô=¡ ¶›ìòç"ƒòa°û*T$ óå…þu,à‡ >˜5¬/‘õkš6ROÖjÎû¡MbD+÷oN5{NÖ×PÉ`}-‰”[QŠ5 ý#œÀ\|)•/HÍÕ³–C< ³ÖvÑ% æò¼Æ;ÇàÀJÉÜ# “±£)ÇÅG!‘,¾#Âgu¥›ÔƒŸP µ2…‘‹#!Â@GᵩâQ°gÃ'ºÇû“¹z4(D„9G}¯Htà¶B–GºÀÙà('-á&¢™ñ´Sw ~ØÙ—¶xš³-Ç|×#òí¿ èìóùEƒ8qËEBþ:l•ýu?¢²I ïœû¿ccø6kü ¹üú®ÞÒf]î»LºŠ¥Wt‘‹M"^ÿþ§Ù—6ÄA8ºwõ§¤¨*„vmˆâ} *i©ÙËIô H LÒ€vˆ›VµX*;}y"鉊¾DÑ?+·{à°O¢ÍÂáïf¨Ê˜_>¥Tuø,G·?6pžÝ-) 'ÅÕÀažÏ$ë;‘Ì­Åܘç®áK8eÅû{ßüŒU'WÑÙ¬Íj'”7»sÊzF—ÎÇ‹yK¿d ïy àßLŒÿ^žÒˆ¬¬ÖAg„_eÁ%¼‰ÁAؘ!‚wURæm þ%×›íUeFj¨±ö[‡àfT—cë“aؼ g*³‘2òç¤,xz +9gnoäV ž‹ú™/Ös vp…¨8ÒßÔq3øÅK%fnþÑÍÖÈûz:¶»pYöÑ?™¢Ò%^'ùøDBßï列¹Î“ÿ÷f RÊÄÊE07—–î·Zø ƒB4»QTü¯çÏä¸Y©u+bìZþPPÆ39¥BzpËðjÒ«ýÊ@ÐÔó[² ŸH‡±U¯¿Tm‚]d¾U?ÅMx…ÍãÇê§Èd-¥ñ:§b[¦ªrÈL¾æ—ƒ…ëÈeüd v¸‚ë£E^aeäð<Œ¡ƒÖîév]ï(ÅL¹{=3Df*O²Š^ÒðçõmæÄñš÷Ô .a©€³Gëìu•¾wØ"~² ãCd÷‹™{ü€‘`è|¹o†Íü:/ åHûçüV߯¢)ÀÓ1êÊb† `z¼`wzÔBßY^ñt>­Š5!”]o0t¥üqÇPŒ^Ròâûu×;I²ýºFûè¨Í-j±ââ>HSú†ËKªSžójN“ï‘ 4fäV}‘ØâëÔ²õ@–^ëÜL-ˆ‚Ÿ†¨~“½¯WZÊY·¿Ø:†‚=ÇbPq€©ß5FñLŽQçÑwÏÎBuÞ0!3 +xë‘®gO9·ª ¼½ZBV¼ZÊ/ï#›*þbîc¢ŠfE¬F¡5¢0õµD =îðt^À†ç5€Q „„gã«B¿º„¿A¬Æߢå Í©•­ÿWí|ÓôK‘aÂ&ür¯–5[épëÄﺲ¼)Fm)hŠþ]Po¶Å$:/Æi‚$µzà0¼ëµ¡ %Ri0ÈѹDÛÕ]£-A¬2,UHAªq$uA–$2À¶ù0kpÚ«J>K9.Îè£Tz9<ÉlëÄoÈê’µOâÌm5€+f & ][ÃäÊ·s§¥Ðórõ¿Ÿ«ºN4¢z]è.’•‹Z!¾f)¶¤Ð´ŠÒëÒFKâåW+±Ž§Óê1 v:y‡èÁ­O ö°jÁ~ºå¶Ø¿´·‹ˆ{ìŸGbœuåõ„l ´ÆÞÎ7&}ÕÚÐÄ5âDû®H¥¼Æ¢#B/,ë¿äþ½ßËÖ&‡ˆ)csÎ[Ÿ4 Á$¢Ct¿år×”åq |0$†ËjÌÄyº3iæAPé@ÀXÝD$yµ£Šô¢ z,û”ÝS'‡´[4ƒƒ“þ7Ù k¸CÿÓ¿ÿ‚eÇsNŒ,0 XŒjŽÅ>ï[øjÅЧP†‘'¯AK›b%ôòäì,[úÊ„:Q™sÙÿó+Ã(/Ã÷VËûJ¶Ò rïTRu×à¸5>‘X¬-ôc¶í¤zΦš€Ó÷`U·Ê¤ü0Îîä9ò`IÒàZ  Nq*§éz{ÊëI)M°;&†½¿ªsÔôƒdQqüÌÛ=[Ó¸I™n·…F<\j~cüG{p¨Ï 3w÷±¹?Qý”ÂÿË ìKJzfhªÀôdH[p3wÆý Är龚îÓýj9 ”¼8òIßÃÓ/¹‚>J~ž=sº5•I§Xó8{~œ—*Àx¹b%N¨‡Bogmÿ¡œ›/.=/ßÔSp*3¥¨ZŒØ•/‚³%£ìÌZÅP­káYù ,mÄ1]$åNa'cŽ*þXÇÄÇSó‘Z¸ZËb¡PF«ë!f>%¥ušÄ7ó)Ô»l:ÜÀ\ËÞ¿T|¡œ°ŽóÓƒ¡‚Å…ÎyDèÔAºÄ?H/6@$|ÖaËÊsSº‘EY€Ž'uÿ¶Î×kmë ¿,§ím|ƒ×¤ÐE@¡5⡤)Y ¯É0E„ò$ðòqVÔ6;]<›äV4LápÙéÃë'{8Lf¤º*ÌŒyÞþzè(¶×ìËh ­—Ï*‡àœVÞ 4Ò«²·*„T±ÒŠ 7g”+Ylc Þ Ãý^Ø" L„³h= Ä!¥›©Ç|—Øâ:9*Ã1A®0'5Kî³ î#eöªc”Ìmêø3c›¾oµÐæ]J ÎrdC¶[¦#û×»ˆûxNg úÜXÛE‘jD³â’¤¸ØÅ^ ™~ê7pýůÿ³ýœ@8ÀsFâ«T½ä‘u`æñ9ïâÇ8©:ëy;ƒZ5 YÍ2©§ÉTï‹È[ò4‰@ «¿8ýc?Ök¶ß9QcP5;¸z$b^dF@¼›ÛØ2zcJ2hGÅ$J««hn#˜°Þ¶JµZ¢÷®‹Ž‡”²Æð¦Ÿ怓~¡kô ¶^¥½P&PlMGrV¾ÿí*J…|f’a ‚q¼§Nâ?Žú÷›Ú€ƒn{ˆ`U:`¦¸¿’Ã=«Kw€ÉþyƒÏ…HaéßO“ÚO¡¨£ïI)J™†ðs¢µÅŸF7t ñ¼4 z{V<úB•¨u*I7xpZ¨c ‚¨à:y”{ÿ\B¦#¾GûùŒo\û§ %¿Dšf¸û_øçÃ7žê3ö×ñ 8)|w¨’ƒuÔÊĿF¹îu´èîŸgNá{Œ¾cíù apò/7&]*‚h­ŽŒT«Íßf>>hÞ2G\å–7ýa€'`å\úÃ`zA‘ò]©68YÑ×SõÅøl"AÈÀVqŠhòX°Ýß”SÜèG¾ê¢˜1Ÿ±8þ­1ÅÙ 4ðÔŸÔÒ÷©ë† â¡*,”ûÿ·ØøŠ*+cÌ%aÂÜ&ªßJ^e*;¤³µ žo…Á­ßƒ@´¦Ïc·wb~ØV'wBúa%6œl>.a ¦žA#-Lý_üçub(¢ÌhvzáTU Gõ®EB¹÷%ÄyM‡i¸~7…J‹g`“àw 0¾í™¼]¬ ¢L×!²¶U”ëÀ³Á ´¼Úc¼s8ï'o#<”••„a¢8½2ضub¶cÐË~si×òGj7é;‰¡5"Y "Y†q¥8<öœÈsÊbæcKu€Én ýIX±ì·yL‘ûŠêwÚo~"7å 'wTj|Ã÷>þ›MYtùVÏ5lh48꯰ì:œ-| ´’=›ìÇÿëÿêÕqÌ8ÉUù;£0uÐÄw–ÕÛ·4vt¿‡ZôÐX2¾ïÈ>ð–™jX/ó×yîÌàà9°”»¬cÛ¯x´ecßIÊÀÁPKçÙ Ü(JŸ¡ãcxuÕ¶÷}ºöŸÆú>&ø*‰Ê¯掉(àMá TÒŠÈØÞì%»âêbÚÜìÑ6–©Üš¯Ä[ƒMø¬S#LJ@ åè²Çæ P} èÚ²LYùÌÊ€ ÷oct}»q9ÿv€Â½My’}mpVÁøÈeræb1fû—§,C=!\»A±ø Gåmaiö`òΈ¼°ø°d$Sm)®þ4‹‡¸4 –²Ü½<5¼qêþ¾ñðX hÕu$VWÖÜÄ‚?Ɔg#ͺšTÉmêPÝ¡X~þ²sò"vôó¼I9C^Ùðp“nÍnFyÁÓJ+n6ñ á ð¢ß æd¹,2HÅ9ñaðËÿ\  MH¦-Õÿ'é8%«ìëø_9Õß<•ÁxýÉCt&ö$š?ÜúUÜ:è5JD´5¬¢‚þ’o,ÏŽpâqý§–kÍ‹‚™ºû9¦-iÑ ðî–Í¢¤ëÿyIW àÿ÷cd9ãÁ®ª£íÏÌb8@ÛXZpÔzÛRÆyý´#+²lIaŸoC–rOÛ~¢ß’•-CáC»VèNGÈ&kqÜ ‹ÖAHx<í"ÞO“jpÈø³µ'eMÛën":ˆ[À(ä0ýì‡b/½îøKÌîF,!Þg8Þ&”Ô-'_s-sxÕ[×¼ÖI­¿Û ×fîk³”SqMØT"ÀAÙWÝlhè-ý@p´Uœ ð}9ØE‘j õÓ£`u#- ¬l”ÐR;9F/TQâÓrtÖÊ«ÓWãD`Ü…wñ—'ùk\Êì[eHù°y6ó¿NFµò«—o'ʶ‡Õc–¤ëÙ3ÅOpí2´öáùâêòõž·ì¯&ßÜì ÒÆYoš’Å®)2OÚXJ°G+ôm̃ÁÊ@?)ŠUR’ßKü «yÍTãý»š&)#Jp“^GH¿8*I‰apäû²sª&+sT¸´4çïåöæ€Qà•_­øuSîaaîÅÑ|þ®$²@ïFåÑ÷ÇÝjÇÖTð€ÿ’–ÄÚ%Bæ³C¾¾çüd™:™:î{%SmVCÆS™ñŽ÷5[áË:YðÏürg0’v¿ê¿y®äÍ— ê>sQ°ÃÍÛۮΔº1Ï’v_uͽTí« mô ]xºû¹ÇÝv”c!á\Ò9„~K¡“k=c¬Ï¼Ùï^Çê€kæY’t`D`ÊNN8ˆp&wþ×å87o>»X²˜w{w]´ K5ò< G=t&Tù@JÄãlÀ¡ü-[/§Ç®óqp!$Ú‰7ø·œ£àÉ‚±˜Rƒo¤[z_ð⯊ª¸ ‚&5⎿£»IĜހ(ŽÕ(èÿ!%lX—)àð|ÉF™¨Ÿµ0)¯£õx ÇçlPŸî÷„pžf‹s볩£ÔÀ‰aª_©ƒC·jÚNôÊÃ1*ŽV¬)s‰Ç›YƒPÞæØ¡,觤ÖNC?ÐxK%.ëeyžuͰô»²‚ˆÞX¸Í€R-w'ƒdü¢˜±]À…wŠ&ӱĔ˄‡Å'Е/¡c>ïÇ.g¡·zU‹a?ñ:So\¸;Î2‡9õ¤˘«·Ø¬JO_ ‚5ŒklŠ«p póªuöãÁí ¾ÂÛ&âüÉÓ9|Íœ4€ªÐDÏïƒÆEŸäI«Vþþ97Bð¡Äf÷»M$€¥z5°ÌQSŸª+I”JDtÐvI`Rnâ.ÃLžx,‰ØÓ¨¯v¿ÿ«]o»uå-'â¹Õ~ÉeÝž G±Æaö]t¯~U¬!ûã¼&Å ×L¥•¶Éy^°$Ãf÷I,Џ@sÂô~’f‚’Ép£ü I{…Ažpoâ\ ¶J:^ªŠÿ¨Ûö„ÃbçÃ&vy$P/#sÿè^+©Ø ÎÕNÂé<$³’b#ÎQmDÛ¥Ð>gØmìÑ[O*—1s¡aèU³‘ä Œ»}M|;ã©óoá5`õJde>Ñšø§w9rgŠûZ\ÎÜ5•ö2Ø7P²TWîÒBp§Ô=-CMÚ÷±…NFÏ(3EûiµFûد™„ßäšž8AÑð’‘Eߨõ³kû¬Š2S NÏÜñ"¯MoÛEgúmÍ<%Žõ×2½ynÞÏûq“ø²øÅiÝuùZÅ nXfM¨ðGQ“­¼«øk1ºº];Q‘CåÇ®ñ° 5[€åí>ðAhIOª•f¶öÞ¢A¾†— ”[vG–Dø"imyéë:›v´ðt#H‹&¥nqö`Ò#È>pã~yXü s²áLÄ­wm† ósŠvÓ¹Óˆ2°Ïmê,ù5&8Á{ÖðÁ:<:²4® ƒk¤jôãJÑ€À¡)5PÖ´lžºŽÐ” >ÎT^|yHÕ믡/H3­ñ &ì¸ü!¦™ã¡ú&‚…ÌqæRäà8’ o–¥V›Â˰Ì…\\?¶ù/ÌhG*d6”,Tö”å:1âH*v „Ë7“èÚ8v¯f]H¶Æ‚ð ëÈÄ×k”Ý_¾D% f×DTù`ãž„ ºƒÒŲìé¦@†;£®áî<šAcó–·²çê¡}wÎ+"ÒΗtN²‡¤ãSL`_Ù÷yzXïi¤ÈÄ,óAg, ¸R 3Õ—H:^ù?ËŠÁ²&$ýÑ]6J„.¼,ÄÙDçU‚$p”óßÙW‘Ù‚CÓ¡tyýeÖy±RkiƒöuéDWË_Jr~"= Òèî¹Ø›´vP=/ZêlG(Çö­ìDæ¼â 1—Ýr"Ñ}Gò—ºÍ8¯ÐX‚¼Úæž ùq~¤/ºŒ|Ñ®T’ËÅãÌÕñéØ÷Þšò¥S›"¼|Æl¥ QaUò ‡'ñîg*’ß/äàø[c¬dò U{ÁæÑ×2qߌöÃ.åwÖ„G__m ÷•Äo&]NFY…ÿm FýÁá&ËŽ˜¦ÕE룔“ã¤wGÏ+h›© PÌ´× ¾¡È“[°ºpCâF/%ÔˆO¡³ëû*dá€z\›’}rf¬°Ë˥̶ÿýdT’?ê¦åHœ½ÀÈìᑤ´Oóã`À•lAaÖ·¼ý?%Åý1'J®DÍÍj|šè(;|êGˆeˆÉ…(»Y$ÂÄÞ”ݱycŒQ×›H®±ÇCì8góø“Üó½q–/b b†§SÈö;q>F…`â¹ ÖŒÞYà%.·i{÷ê¨xMz“}£Ïg_ŒlèHé“•ÑJyâc¤Ò¯Ô8*YI‰ô¾ žy ¯ædÜd•sn,»ßÐjI*ÄMXöö5î³î;öànžÝxè9JâÍ÷¥ê"Ø@TsöÀg¹bfBÌ´òj ÷  P®j]Fæû¶ã›Àø Ä7 ã"¤G"7d ž@ÕìbL±i‹Ôøé³7rWéõp9þ-/Ywxa/=ùø¶ÐÒXo-mìØé¸•AÙ]1yÆTÊ;:ž'-8ÅŒù©u>ͦH!%«zꦱg¿%á£G˜yXvË@‹îשj"Ã÷·t#oZjµB zõ®çd¼¦Ô¼1ëÎñË|‚sДLcßg|A*œ’ Ë‚T“`Ã_ó6®fÁAÉ0hÖ(1Oš„$ã{1Þàx¾ð+¢(Ò‘¤‹dW ±â0™†ÃÛ1ôB©?³’,¬Æ%u8r¢VO¾Mo.ZfÌQÜ£é1*#^Oº·®F]æÊc0$ëènh³¤ºiV†sÀм²!^ÇDí),:`ïîÞá­¥tY•ˆ _ ÜêlVÖ0‡·rüç…† ìÎ(øÏ‹BXî çµÈ¶›Qg·É´qûž'Z{ÅßɹÔÞå;—$>”C—ªêkjñ~’àI,˜Qð„ŽgE×wG¬ðc´@{$añP> ù÷,7²K£Œg2ݧ‰tÑ….•eèA®¤}Þ–‡Ú•ußG ö§^4j&Ëà/™ÃzjÕêµÖ†ªÄ²zp-ªþ Ÿ ¹ë4ž64ÊÏC¿pÐ ß5–Ç iz†}¥IàÜÊùÅj§*m‚ýPšçÏjn_mˆäN¢¶ÈÌ)Æ’é†; Wºk¬w Z{‚þ(¥ï=lo²ïá6ï]ìÀ”_`C÷ri#`ö#0S@{…+€±Bi_ÍãåùES¾[ëó¶y2´Ÿ4ˆ÷e"Z3J©A©×Wàû¹^ ½ bMËrænøp9yòFÎoð¬nŒxþçƒÜ れ= Äf T"¯  ¿佯¤Ž€ƒYP4ŠfÄîZì þ÷<Ž›ýÿ_™d‰ÛÜ?Ÿd¸êÌú=/ûeÄuFJÝed…xMy‘i¢éæ~ À91ÌÚ„‡CŸÀ⠮âôö×Ú‘®C¿Q4Ê5Ô§Äk•Ñ¢Ôðvæ›"êÉn(&؈ÖnDºÈ0Þ¹Ú1Õš´¬†ó¿?.r¿•O^“X±Òõàa{>xt¥®FáTülGËHp5Ý$µ#-’žUV-,­œ‡98çÍ=NêËÞºM$§ú BK[”ú¯¯±³DÉ|‹QVg~ÐEhU áì\ÆD£ ¾™Û¦|Ž·exš~ó÷9˜‡Ê+ëØèµHÝ·ïJ€@…MV^LQR¾‚OAN+yJ¡1&æñ©|ªŒoA̳¹û/gÕ‰â‹ÕôìÜÿ*ÇÎÚÜýåûRK®cpT˜ô½/Ål» DzU1(¤›AGìržiËÓjòãh„[æ3j‰t&pGð5êõ9Œd¦¤W±·«O g¶¨jB¾ÍFhL“˜sU¹„LÏõ´¼+wd_ó轡r—zuOô†ÿ–Y’>ùTOƒØÀ€Ôí« £'ô¾ËƒÞ“ÉIÐ'N¹šì¾@M‹;šì†˜ø!}à}Í­šÍøÂ 2'«.VŸÈ˧kHɳb½rTl3 wfÞ»@ìzØ7—Qq§°¥nðo~*3>çÿ+)×è‡ù8_Üê}z@B…oÖË®‹P,ke ʶ¿^sáZŒ[¹' ˜W´¾ -NÅœY’R2ó+Zª¬£Å7X´ÈªQ6&–›‰^Æ$:¦7ú‡! çÜ\¯ê0§( †#C;Ã!_`0Ý…¼ˆõ,q§ÄV­g HV&_4h¨(fFFþÿ"·ÎûáÂÐi.E–fw“j€ë™ÉD©åÓ(Ó&q9™éZUn$C5¹p–¯»ùÔç‰ê%G$GÏŽ8¬Ü9ÓÅ!n™›”¤Ãø HEüŠ7ìëÄ9D]9u‡i<*®%7ÉÚ(ÊÏ®˜†ÞuI M!oÞͧü* ‘ a¶B­ì€D´: g.€¬`Úxœ[ÚŽSßÔºÊ âÔ0ÃÆÈé’"CÆÕÄ£û’*$(•F > ã³Üž‚Í 9dÚÏü¶ØÂD\(‰OHΙuù°}ÓYz¶Ç…7ÚºªÑ©cSe¥­·(ùö«K×JØc%´™¹4_qóéñ½‘hl£`naœÑ©² x!B…üéWŸ¥·jЊàDìÍYHŽÑ¢¤A0þJ-5ê¼ñ0° ÛåòSHû™|<­UüßÎ-]pî¯2 oŒ%¢1d%E­Ï\°ø»”û¤&Ëñx[R¾+ðµsÑ’ ö"j ¼Ô9¹·¤W~KMÄ ¢×k`¦GBÞWÑŽ œû”±ØÌ„»ä›Ù‘¯Çq%•,‘†–‹¸`t<$—M¬I¥ÕdÿÆ÷°±áèÖê?¬ÿ(Èì"gÛèá¦ÁÍ`v´ Áq™œ‚kZ¢ÛÝD Užo=aû…,ÀÀ^¬ƒÃSezc¤¿mÐC³®ÕkÁ'N}»î4.I$‡¢çŠ"–s„ºÀ0¯¶‹.*¼rÒÕYÒU€÷$ Zu˜P6.È®s¿DÉ! J÷ál¦z·;;´Ü' Ö³P/“4X—‚ˆ9Cï'aÎAU6‚Öä’['~®±T±œ"ˆ„£´ý²Ë…›¢LºQEìÙx£}N¶˜}œN;ÅæÓpˆ¼n×)¡‹ú»ÏˆðZâ†Ñ¡$%\Þ$*ÅæØ<ÀÿüRV`+|€Ò–ÃXŸ›,ÀÀ&’²< µúéÛtà"Ö=¥!mk5â¨,²æ·¿¹Ë-8x±l”9áUëZ ½p㦠å "´ÙTkº Úµ|±R„äÐp°$²Fô‘YµªÓâcV…—ÐõLS‡¾ªX)bn…ߥÉóô›WTgÖTéšÍÀÎÍ´0Âò$#óbš’ ²‚uÜ^ÍNBé Œnµ£ÍcÕQ!ºzyŸ€ÙkPô”\""/ª´›kNjÇŽÓ‡ÚþüßxŠ6z9 ÎïÔL’ÞE~@׺ ïSÃ27,lxêM8Üåi|ël•OkKµÑ“i¿öÛ’(/Â`eÓÐr?Ar±eT~þv‹¼¸qÌ fÂ,3¦ð ûì^ñ† hÓê‚YÀ.‚Û¯.áÉ-ÚTwvñBŸ.[ JdÆh‚­gFœ{ltÌ•.Kâ%#Û\4ç8^†•4d#¿ 7ˆÙÖQqÁëÉg_ ÚÕ¿Ì< éoI"( /ma:/ÑÓ!\‡JJòŠoO Ž?èAÀ…—”/œŒ/›€&ÃJ¹^ŠÔ©ƒ0uXP¥‘’•¹fï û)‚·m4-“/øx§Ãˆ®ºUAïB$ʉ¥áÈg  WجDÀ~ ¹ÔÛ”8‡[/ß3ZO©èø}sC~ù6Uåc^[ô×¼3 þ¾,:(d¢™RŽ˜]@ñ)A×Iš`gÅÛ€Aç"²¸ª7V°¨8—(§Ò9ö½.ÕN~H`z&€'6÷àa,îSç-O(“­ŒÊ¤OÓÕ [ñBz¥)Ovþ2`“©³øÀ8=ÍiœV5½dS´ò|o¢T;:ö®üfZáóÏ1~1ë`å|nŠÚ,¼*$(oD8z ÀÿŒ€hÿ#« `Dpß›qNôd$çÛ0¦¤Q–©±òmyÀ*m‚ìµÂ8°I<È8„AÓ=½r|¹+ó±uyÇIÖ*IOóö^kãI#GJV…°dÆQ­]¶½±Y™³3Xö>ê)‹ø3WÀOMˆ`Íq°·mœ™’˜îÖ`ÛuÁš…ša¶aˆe5·Öe¯ûõú1°ºÝ;6ûk饵ÂL‹Amdä´~ïE/ïÙþxŠH8) %P£©$£±â0‰uˆ›C<àCc¥œ\@½Šo\ñðüTí µÜv±¡n‘`ÈJª½q25érñã‚ħ)H@ûZu2Úêð^Ñ›]À‰ô› ¸p-в[BÉ6!×´ÜÊœá¤\ñ]í·vJ¤«à 0ú§pKÚ<(ˆá‘dðxóhÃS×V0ŒS¹¡ÄäŽC„ÔÊ翯¢Hè a†™Tÿ¹7p¸Âí\& ]¥‘VÔ¨þ‹„Ùþ«cíäÍÓd^xû ŒÓ† ù©/$h”z‚çÚ»•&¯†| í-Çð‰›9‰öÊ(–JëPu×À5SNþ°Ì…︴–Jî”?PxC*š÷á–‡òÙ€‘•]øb¨¸~H»¹å•ì™C’m¦°6ë@Oú2­3àý@hÑd½IR¢„œkM·Ms— R9£ÇI¹ópnp¸ %&Ql‚`B¯4®w`+] 8‘®LÜŠN‘)•>I™› ĺt?8P¼¥ê£¡HNU¼ÜË}M\98ÄÄ™gq/šÏ¢r([Ù†FeýŽIúS yìõüø|£nEÅÕjÆå§Ð*©ƒ>¥T©A8eˆâªÂHóœÔo’Ì?²‘þÓ.èáœt¶ÅP%ä›U¥šÜ5½*³XíWõÉÛ¶"ÃôLf`Ëìéa‚̹#DŒÉHœ¤UœnÜÊÊJd7¶ : E,ÒLØÌö`Ah1©pé9`5ΤŽF¥á"ùô¿þ—ÜÏ×j¦“Is;qU' [N@Ñzhçéö"縚¦&»“d:7;Þ—Å ’[PÔ_šVå]ÝC†R}÷~%kª'ç JJ½ŒSþ²æ[ùxí±#’ ÚÂ]+ïHbbþ˜ƒy_Ktï‹»r«È©¹ž†“…’Fi 9Ù8˜¿eGÜšê éA‚~Ž›~ ªz?Õ&ŽÔNé’ZBb ´æøaEÖ3·Ø%nH² ‚$Ñüc¥ˆÝügœã+òÏG¡Ú¥’ `Èžš–%13!ºŽ"3›X3Ð2w\j±lôfoœM0`q·žyA­(ü‚Ú n¾YÛ@WÏ•€‚Wv채 >Ÿ4ªë±d›±=>‘'…Fý½Qæøwn—Xµjq.wšÐ*Ð"¬%˜xKy,˧p¹~Örj¹êÇiòÕ8˜Ó¹j Ç)˜øëEµ]~£ë,gŽ=°³•Û_Š*ç7¼%Ùø„ ŽFVúŠà¤BÊKÌjÄß@^÷ ¿¡Jë¦ãî÷úúKD>f7òÛÕ€[Åêæ‹ñƉû|-ðÊcH{¡˜;•{÷r¼´˜×mó €P N£ }cÕ„¡ózW§Âq¥˜úëDÀ(w!úa¿ —=¯ƲPýzmoO%¥!Ý™Ž6äŠq‘¦'ms¶¿V»Ç´' ÿW™Õ[˜ò‚<Ùø¹®~!Ñݽ,+=ÕŒ™*2æÙ %rÀU`8?_.çÅÓa÷€Ôì³Ü˜ 2«@y*è,6¬­¤5Éš“ñÈŠþ54@­Ë a4+qÌ5沓¾pƒŠƒÉ¿6¦ —ÏêúÊ$2fô}§CM*÷.(P)ØM}´RomoÝoGMôiOž¨ Âeæ޵gÙÏ¢¾Ì½V/ÔJ¬‚ðÀ3i’ˆfÙ}×õZ ´± œ†'ùÿ›ûíÒ}þÀ€4ÑY•Ì-Ÿ™Î7@Ë’ÓHæo² «Î,7à‚#U¿—±úöQ XnEU·™÷˜¬9É+¬TžÏ¸5(ÌäÊpó%_¿k8—-VN!cwÜFò=ñÉâ(d€K'©Â”YGúÐÚ BD<ù!ÏUye¬"SÀ·ßg¥è«ÓƒsuÏ#î»]1”] _[«Jh¸'ع…L-†¡Å½'г&üMpÞE5Eç æ/ÅfÌ\Ϊš‰ZèÍt¬áþE,±ºy«ÿ¥¦>xæ"°ºò]‚KÄ”|ô¬¤JÿçÈŸôÁ°L-»JÊæŠþšR1q÷AwBóC2gÞ9@ñ¬¼Òù¤ |‰Î•uf pµW$bj¬':[lˆ¹²¯&vËÃþqÎ72ñcs÷owÎ…2`fŠcþF\ ʽˆa/•Ž™ãXœ"i ¸éÔy¦ÈjÅ»eˆ`7³º%jÏ ß/.íaÅûÙq–Ó_¼ ‰¤¨k‚Y{Õ¤ŠÝlï¿¯Š€÷k„Nö Í]Ìö… ºtÍz¹\ŒlÆ`¤e¹E9é1³f–D´%;nÚÁÿ"gBÝñ=NÛñU£I›OÏø^w±j³6†„¼ƒ0Ô&¿Fà/Kuˆ‘B]us©†_–ŒbˆÌ'°{…NŒõÃQ@ëIët¥î5Ü–™Ï‡êâpÈ÷cµ›z¦Ã—ú;Œ.YäH x!Í~½%»ÓBY‡ùà ]ž¶DAåÆ$FW_™Ô„¡±R˜¦†Z‹üЏ>ÍôCŽ Qa >ÌÈo!Y¬3dk(ÿÕJª‘Õ]›hÑWpú&Cv€K 3þ ‹ Ÿ¦3<•-_ÉØag Ãpy£&&Ä6U—ä(kj“$(ð¯Òú÷ZäÉ üìÚpýúˆ¤(Ù&~!%¤î>)N3}«Ì~Dìm›w5öЮáÄò!æ0>ŒŒ– %)w¶ —)…´qϓȈ+í3Ø ŽµiEõÑ\ÕõjŠw§kùÛñÓ]ó èOÑø˜.¬ %{ 1hÂ/éÌÙ–˜­×QèKy†ßœP¤ì:Pýâ—‰’•+ш¯¸q. Ûeje6“ĉQíd’DÍA•b¹IŽÞ©Ú›5r8þ- vþÕ^¹|k(nPÀ# ¹ï;ß›T”׌Zäº1¢ •fr ûCD½üàPUŠ”òã³»Ó_FÓ-Á}e ³âr5¾* ôzð_I_{i±Èh¾„ú­¶ÿm›SŸV¤C©ˆHÄ"K{Ä¿mNtþS®j+†¦ÖñÁP,=ìöN=ô¢@´·6â:Ì6ŠpaŸÚçO™Öyr|¹Ž‘NX—0“ˆ~utôÒŸÁ÷_P‚b¤ AY'¦8Ü–÷¨HÚ @n°çÝ;2T‚ù—'ÆQÔ ù<î i£é~WÇùó/JA"ç¾íª09žªõx±h”ªºFÝÓõ„¥dsäÛgaõ+Öa5‡IÄs(ô¿²¶í[}q€y^f;bjÓ=ÃÐ[pš¶$ˆÊdä 'vDVÕ昸óX‰Ý¦šiƒ"LÎ$>iõ”¡(djYKWи\`³Ãæ¼É{¨l‹á›îº¬Süõp¯>ζÿÈ›Mq`wïdJ¹ /SqËy8hæµOA4zG6&‡¿QzºÚ!~,ã ‰°!†írêŽe |ñWÐh<¡ÐŽy_£öRÄ/ª8¶§Óøžø5· ˜ôå‚”áÙA3¯ôv(HK/Ò¶Ôm*o ®tÿ’Ä kÛjtÂ5ƒü™ÈĆ)ìóGCñh»ÌÇ­&°4ûꤚµY¨ÊA@£ |è%¤] }mÿÞpíG’Lò4Ù—Rëë2²±¥WØø Rt©reguÀÒÉDŽ]âó¬˜{U¿õØåT#¥¸cX«.£g;Xœ½ºßÞ7Ê@~wuÐäÕÁ.ÑѤA›ÀÐÔ7&zÎÛyf‹\ ÎÑ}ÞJs]õO؆ÓØ~%ÿË5Á‹Â ţȧÄù¤•²+ŠZ—`ùðt¨›þm“hÕy0þËì}ÝíT÷…FRqFqÈ|ÑŒÆ'¨ä'ˆ×µÄbÅÍâ|5„8²¹›È ú&f©‹Ú§¦ž áE‹þWQ¶Œä”­a-™¼ìt²h>1¥–» _¸# ?ÁœÝŽ·=1ÓL,d»?;šàõÞ`ªŠ\E©ù¹JçtÓ}>¤ & ‘f=¹³Ž_>*Ó¨¿ã“ŠGºZç¥õ¹î)©Õâ‰ø÷;ÝõRM-4•…4’µVGÔP5ÜÀ·?—-Ç|"„Ç›WOGÝå2¾xç4#äËL‘}v Òƒ-I¶QÔ‰qbí|¸»”Ä&åöú„Æ€½F~ öŒœ€´f”%¡w&P¥YäíuÅé$‚Òt3íw{øVÖÝ dz­¿YÕ{h~nÊ·?Ä;ÇØÐsàSaà&i‚bs5§<àcY‘|ÚEw%nnä”Èc ÖYµ˜Žº³éî³,v7_Ümƒª–´–½ê¿þwF.ûRr°0]«4ÅŒ–}\Sä?²+!*ŸìBA)Am¢`Ž\UhÈÏF!³‰¹mˆ&·V:½²Õ{âI9ìGvÛ9Ë ²-W[ýn“Ï"k'ciK¤lÍJÀÅxAà1ªBáeLSK³¥j¯"pz+0O'¯X!‹K ̽ÀÜv¸vúgÑY®Œ(˜K®ážAP=5KiIP ‡·9ï#EW ðAiÕ×Nk)ë³| 1 ’ÞP'•Úª • ~âì¤ÒJ–r¬½œÀŸþÌ–…Âú‰Ë[Ÿbâ„nÐ] 6IÖ´¢âP*¼€±Ö“€0£ão"1,̺ûΞˆÊß>ãÌÖ8XOó}]"ÆÄB‡®F>RWñ'j©WP Ã22‚>¼v!MFeìV$óö»‘Ð¥ö”480`}õCò]±!‚L룃£v¥!>WïA¯L¥8brQ2õ<îÓ fpïÿš×̺Àö”J<*íþ+>‰? º$h‚8Dñ1€Öƒ!ìbŒ®ÀJœX˜žz¼W³t éÕG••Dεj!ŠLòxÅ4CÚÚ=8ƒäû]”ò9¯GCŸ#s|g?\UüdˆÂÇe`£tã¨&Дۗ͹Xìß×zØ_%ðÍòØnl*Â7íâëÙ‹WöÉU€öçë‹ÿ¨¬èÇ,Ú¥™ ]D½§ÄŽx5ñdæ˜!7㳆³˜L®ÎͼP×Vä5MP!œW™ê 9gX+Gݦq3pŒœJñ½3iîì¸0•lŽAxÈ~ ˜_E[¤]2¨Ãà>ÌèwF­©[­‘JÛ„<—9ÝZ-ðþåœÁA¸¨nÈt×ѰrµOìÁT”á~ïÅ£UwòÒˆõòJÉÚ¶Jˆ)?Jõ+”R¢ƒ“=Yªl?®3'_9yÇ{þëAaÁ22Þ„G*0%—ÒÞ—¯”kžËõ)0PüMü…[ý曆ÿðBÔ~"°tõ§PRL.)öЏÀWáGP”¼—‰‡ ìΠ‘S°…¿¸!sæ^¹ö`|ÁÕi‹¢Úí› }9¨ëcµüùx:â ¢ðT±iæÎ{ˆ–ž t‰k÷ŽwK÷ìs}èÄöàNsŠvI‹ã jŸ7‚‡&å]ƒ|ô64SÖ¬–E‘UÉðÚ)ka»6!…¥n=bZÓ蘂,¡z«|ñçŠBX™é×dø¸û¼¾Œ¢’ƒT¤ Ž],® ¾X7 üŠ ó>å‘#’¬ÑÀIù)þÞÓ·è÷èVi¨”Ý#øÇjó-öîÖØªò,¬£Ã‹ž¼Hû 1R»á&‡1ÔZ×N5q-~ä#ÀÍÌñW*Á%&0a>†ÌR0M!ðuPdÍ_NðOÁЉqÉ·°„®X¡92ÂeqSH{$AM'4q|«I¤rÐn¢nFÇé0=\k--Å º’Ÿ,`Ú÷BÓfÖÂǬ²1¸¼›ÿãó&¦Î ¯ôŒ+¿^¸Ø”ýf»cØ¡©5j<ê`Îæá™^±»§L õ*ú®Iþºsè)8ÿUa|BÕU§v)¤©„L‘ú „r¢"_cÔ«‹éKÈès¾ÀØÃdØ=úXà?)½´&ó~©ï­òôM}0rÖ‘®¦ ÙÜ4윳amËŒŸ¡ƒ}@)p§»dÆaóÂò!®žì ¼æM@åN¹l6>ÝQ;këóÛÿŒºN§iíÀˡ ^áà+•£0uا3&>ëo¢5{ˆI‚7Èî0òQ ¢ø2äcV/Ño@°%ñO©òD>6†xŸ8»%wms‘ˆG!³ÝlÞá“ñèûLIF*Rs•ñ:z16éûþâ½o4øú•¡¹Âr’çŬÉzšmÉ:ço;÷x¡Ó~ù/±Ññè4׎æ l“7ìó8™û Œ'­ Z©3+ž–˜7“p*7©øñÞ(2o¸ˆˆwÅÔhÙ¦^äNðêR,z›e›]©ÚäeK¬´) 6‰Û¥åô¬qwS¾Ûh¦Cõ(îÅkU)|.㮸…yg0ƒ2ËJÂËnå¥úÕ”*Õœ?ËÉ9¸+Ö6¦öšÚ4¿½01·°&¨¹Þ¹Nô¥˜y¯É'Y¿ØA*`.bŒ„tŒÂX1 ÖqLß »­*bóov•K(߈=-&ÍLyÂ?Õ2åÍŸÜ–?¹8DÉún~4ô"ÊAé‘T©¥@©<šB¡G\ß—§™•t%—†ïW•¨ß?¿@¿Cðñzo…› o¾pIR>D§½=‹¯ÆG¸‚lˆiE4ÄGTïlL Ÿ |8Yk–)ŒôîÙÀðRãÕå>Y‰3øj’.ò¡ˆŽôéÄ慨_9Wʤ‰¹ôè;¤s~â‹èsaG‘½‰þ-$”FZ:y¡#›¹“€Íc§ M…žóï ‰•r†¿±-½1*ë]NÆ•ç¤ý6×û+¤æ »ÝDóþŒ'X»¹Ä¡ˆ3~–K2Èz[êæ&l‚wÛ©«nô‚¶ƒ£ï·êl‰»ó~CÈ3‹«wÑ Ukmä5Ö:lNÁG½¯®ñýù=¹ìäÂBB_GKVè ™qU)Ϧ“œ§¸ùxz·@X]/š¤NoC^±¾ÝƒÙAhlvJ¦8zÜò­(ÐK²À‡Cc+Kȳu%†M7æ•–éŽå™„‰Ë™»jÀ™Ò]ëákÞ}ÖÑwùçQë}‹8Ü,•4Pqqî±GÈtÊòÏ¿Íè,Šàfg_é@͹ [@ö-$l¹»ÿ »öS‡e³VNxq:Ÿ‰>Iëú¯ZX;OhK*“K”(µ§sVñˆ2hi$Áø«*Ès.kì“ÖKRú¸J¤xs¼á)ÂÃfÿ[f×;>»ð÷}ã¯Eú½?þÊ.}ŒùÉ’¨ÓO¹|Ùç›Y··Žì×5̘B@aG¿ùõxê ©?c¾®V¿á/XÒ.§B¥Án£É/œÇ)…ƒHj6›9̦p`œ SôEŽ…e’£!Zã›-;6n¿þí]‹)6! ðeþV8A1ñb°Âjò©¡ógK­6‰¦ó0Ϻ°°-´òçhYpJ–c=¦Â¥š #ç ù´C—C¹ãs¼*«bu=c®K`žGb¶5" ßÀêW•?‡5=t ·G;eó<Ì~gü"²-mØv¤w,æýs¾²õ›’ÎS±îbAW«¡Å%Ö .»>w)á.ï‡s|!]óƒ¨3 õ ¿_‹ùC]ÛÀNTõôŒ±©ã;äM¡%…ŠÜBÞ¡­cC5^‰îÆ­TšÂ÷ms QW‚Bô~Y&„i¶YvÏ¿ˆ øñwQJ¤fÂWÿòO­•YÙNZ(1,”lÖ pqkËØOó&.d½ x êo”ë”<²ÈZÎ…VÛZ5ûÑR»&t ‘̓ŠÍ%wžZ­p<ƒ{2AëJ1ãªÅ­‰8³žÐlw’kÂÝÅÕ)1–i…kÍ#yPû¬—ÝMJbŒÅígú‹ƒæXzøD{èìºë¦üÉèug"~q!j#½½¦}sëº`;!|`ÉÞÕßÑ¢mW ÈÉ Lßš·¢LÁ„S_Ñá0ÉÄL!uˆÎÿ‚9O3ÖŒ}M•ºk©dŒ‚íp­ â—ŸhÜ5Ù yµ¢ìݶB“4l)¬?*pUŒžÑYXŸ&NÓV¼é„2¡vWÉVeÿ‘k͵ÿ;eÞÁ;ßd’‰ž"ß³ ]q£àfía(h3t‚Ÿ+ñdë]Æ3·¹ëó‚îÒýœ¥Z=Ïš&¢(Ìxy‚156‚rÅLÚš5!¿kûI¶I7ÞÌúäè°ª-ØTé ÊÞŽ>±¤‹i±´ì´þÈËö$ï­ŠÇZöùXd@ç†@ߌqò°fzþÙ»{“‰XV§…€úš¸þf´Ù_Šp Å/ÁÜÎlÛ ŽÆøx‹!‘}’Ä‹ú7ºÐ8× 9fâ³$½Íb¿!'Sq~ þ%þ ªÙsrè·Á±\pÝÛîK/È7ë+j̦Š+à ƒ7¤Úlý)¬Çûðn÷6¶ý€è|%ÃbE+—Êj´–ñøÙ¾ŽtÓ½TaÁä6éãL›b«]hGF—Š£3ö-Sý—¿;z bQc<‰ßhX©Ñ…óB¥kÙ6 xk©XÁßãl2í©Ÿ2ò©>PK³®õ˜žϹDqPÿ5/mYv!M(:Ðî8ˆ€³H]?€mÿ-”žÈø7yy‚zæ 5d(7t~Ÿêgé#\L“rJ‘hüMA¼ãLU7X$†py×k@¦1ê( ’ªÆÐ÷† EŸèËïãsx˜@1I¦'$Ýðcïw²”¿½sUX‚Ö³0O ~Lì3Ÿo AV:Ò!h ‰…Tà.}Û ‘“5(ò.¨í€E‡G=ýí 3qþj8ˆ'îÕnØS7X¨—PnVuºÕˆ—çÕ0Š!øœÉ) wÜ[œSJ3ª dMbÍS[{Å%| ß6í "¶p<ÔØ ¸—?¬è „ž¡%z4Ict”2ˆLcG ra‡—µú ¢Q‡ ¥öŸG®†Ž¸ 扖dúGMíG™tó:\lµpÃPº)…߾ެJ `»ˆÕWfvü‚5ìÝ‹*ïúýø}Ï:´è='T5ÛºBÊiæÚ P*^j<ú“!F¤wXÊaº‰‰Aõ'±šçI~Éi«è¯¢v#þ?WGù’ƒ’ÿÒ|&,lJ’ê«Å9û4¯Ç€¥·xÍ6MÎEVšû]!¾Òjt>WÆÚ»B&þ6úlÜ;¬AtR:’¹V™k«V ¿‘ޱ–Z–¤›H°-÷ìVè·m¾NKíN_àÝ÷ìÑäÇR3X©K'Ç[ú&œ>ø¹¼vƒdú•1):4þ{! ôþÔÒ„/ûSm‘ë²C«Dt'µ'¬«/t’Ðò’#DR 5ý¯“—3û¢ôu–"Š¥éO‚GRz9šcç Ûáèõ¯ÞGºþ©øQ£@¢£€¡ÃíWUqs™H‹|nÔiìòß2Â)gÏê¼¶…“S¡zLàt1‘‡ð -?&päåûCï‹æ9®¼í{¤©×׿ʉ$PÅך8R­¢N%(ïœ!Èý£å›À±£„Š6¦q„>8EÍ$Öƒñp;ƒk˜þð{IÜûhe ¼ÞÔ4AFÇï éèë âÿ‡†ÜjMð5¥[(¼‡¯<¯CR]` 5Óì«K §ŽW¿Ì&-ЮpÅuò»‹xž'Þ˜z Ëþ­º:Æ‘óŽÈ»m\<æóe=u¸xÜe†u „BÛžÎoð…£ü™…£_‚ï}Ö¶`’f¡9•ô<ì–®Îq]±-+¿² гwÓ3çÐPÅËœá…nO3‚ý«L™¥%û“åG·(¼’uÞé7ÎK3±” uqY‘ˆVI'„?У϶»ÐÑêßÁoÂÂ0½•Ö‰zM~wËoŸ § ‚Ž¿_ûêö;›°Wš½tŒläŽ\X¨³ûÖ|²ïÒëh¦ ½¯Ty·… -›S½•ù",¿@DÃmPQ61kC dÅIí‹·ñ‹ÜÁ` ÿµC¢Ôš@ÇL¤µ^\“KQ§y|¢ÎP¾¬º Íê´×*4KUžüMêÁCHˆÔåˆ*Û`^Y2ÚbŽ'2n0.mƒ¢ìCï~ª¾ÇGÂÅ£è1} †9æÝ½Ê *éQpyrÅÝzgì{L0Œ%W‹Ê,Î ƒÐG;•›C!•ã|iêŸéàÏ©xp×Â,ËÑÅâ<Í[™üdx0¯¾@½ÛqwÖ@à¨~2)¹ë ƒNŠV¤±€ЩúÓ%ãÄÂ.˜x,®š„tØš÷ºóœªÕãÔò}“j¨¤wê’n¼Œ7™VŒ Æ´®'dÈ!¬–|Ï…’\fmÊ”¬ê¨™fèÎ"~ÒÍQœxÁ¼“ÌkS Dy#û½³œ`瘊˜>RÆ'lŒDý²õR L\Qdª/ÛÄ{3>çÁ$1›ó0ÔøÈ¾Ø#Cö9£Va,j˜­ä©Kkùtm›Øô“ìÔ‡Sí§a4FdóM H€ {Õòa$ þû‹úº<˜/œJÀͽ2¤IDиYãd½O"ã;–6Ú –,¤µ¼=©ýN(äÆÛ7Vúíùè ¾•ª˜‡ªZa™§G}it%êÞø’d«ÕÈúVÚ¥{!ù!…Gkz$Äž‡ý×6·†ÂßÊâ>Ê«éw²˜Þã床ˇ4>qÝv1 ëQãaþ¥¸¦"+/XJ!ÀÎ)©r)‘”t†{m7õ6X_—b5.ÐÄÇé:¥w²ëyîPçÈ8©ãgÉ!VO™"JKÔ5ŸôOu?²æzÑfo'ѽŽdaebž/x=H£Y¼ ÿùåá NT[í|uÐÍ÷QÆ?3SAÝ÷©Ìb×Ku]VFå%Ÿé4À•¢‡i{&³ð|ܶJÔÚaÅõwn‡?Ð/ ¢V³lÕXQÏgW+¨à¼òJÎÞ>7?8ÉSå\þÕüÈó⇌øK2Áª£…ÙUèoŸlÞdy{”‘ú¡o-;R¥ôuÊßÜ*•w¾çXh´È©ÅzK½:ÍïǸÇ:范,ür\é¼)¹Q̵ô.êYzP•Œ>P<eQ„æÆ‹!?–ï(ÿD2ÑÔ¸4¿ZÉGÊ®IišOͧÄ$ÕÆ¿ø1¸ÙO³ Ž63epƒÎ] -áR¡ýi¿#T+€›iô’¾ÿ§$Ka#Ç–à|ú @ÓC^ WÌq :+”Æ¥§Å$òÎ| `Q챦ãøeélÁ^6 }§P&ŠÛwtQkæ21À¼PÈÑÓTß;Rà¾*µøkòDaLyx˜ L*¦XÁþÛy1¨eàwBSω^œ™7úðF¶¥”ÚpÎÑ"ÌÛÄ’\͈»ŒäþD<ÿ.ZSÖþ*a’Ss­ÚU°M¿ˆ ¹AB‹€§,j@N[ØrSl =;ÃT|_`&½›ÙŠ”ç/\Y™¥u\FW]¬Õ5ÝY±ä(vJ/¼$¬Îc:òÓ\Â6=A®ƒ2Ð*ã0Ãö¶ŒÚÅ^¯Á÷Ö§ W!3Ï#•Nà`L/4*\¨¶ùhx€‹Æù$òŸqÈ v1µeîÃ~¼³2°Ñ‚¶ö½~ Ô}#·Ø ë)ÒHÎ>oí.ÖY†#Œ¬E™¹Ÿ¤›f4®Ÿ£È¤ÅVªI¡ôúGŒšǹþÜ yx9,ùnf~».ЧðN ]È øº<¢æ¶»š,ÀŠúü6 _Ü7£™%Âà ÛI_UíÀðœ•QH[ŠØy’ÕÇ>X-‚æÖoÅäÉã¼4maàò­‡½ê)%øš§4¨s™8^3e1ÊŠ[2nVÁ[@ù1¶ÍañjðïÁÔЫ[Ï ý­1Vʤ6IxFwï ;Ó½†ñ=Ú£…³®Êú¾bèwº(ÞËÏlv 4sî·AâRïZÛ'w£`¸¯„“æ_róaœæV·²D– Ö趸ÞΚì’FË8.On!2`+Ë ÀBù^÷”›Wëæ|¬.ÈIÊùAè{ô C1LjLª˜‹„BiRë?H#>{1Åk/jf„«˜¶ë* õrç'~—ÿ¼¥B¢SC22þzéŸÁîiç½aH(EÙó~>æï‚!{öŘæL(š“8Ì!/åûÌ| ÑP|`1£¹~K¾èEA’ŽSÑÌ]^x_í+êßAƒÊHØj^€ø™>*#wïfgé]ÁP¶IÓ¯¥ãî¡­óÐl([t;ЬȺ8PLMwæß{î9Šñ¡U¼ð:?€¥¯Úmåǘ.ìÓ{ùctôÔ`Æ"ÜyÕo*¾Bó„úgþ PZª´,u5Ùk™¸..YT§ÅëWT=ü!êH/Ait×¢EK°ç³3í¬)úònÀÆ*ÀAõaÓb—#³<]Ó{ËT>õ.Hóµp(Hsôçgƒ»Æ *ç%.ö -©À7œ–ìyÕÞa:Ïã³`#wÏ-î ߦз¶"é=x°¢½/¸Æ³©K›õ¨ Ä¥yR©ŽOR¹oEËñ£n 1ó¶ßä Í˜^÷+–µR×á³ý€g‡Kn&ò°ÕÃM80uäÚF4€dX…+Œá8Éƒ«.æa'åqrúDÿDSÙÜðsñP²æ3šÆìæ­{ÜŽ¸ 3Ì“ºQ¿é»éé’Tœ ›TyŠË­÷KrŠ1Ûµ‚ô2 @õr–etìwÐ_¦„fR܃Ûo…:Ðq¾ó áìcã‘DÕÀ²d&Ú#5~ØöŸ–´x½‡âÿ¤œŸÂÿhº.¸†_©If¨˜tž54+®ÄLå…Q·Ê'žø& €.)Ú/ š•Ãü–ãæ:Úe _N%ý‘˜wpùj;A¨ƒ(ÒTkNë(“¹< ¡‹š”¨ôŸË>ðr†q÷qâÞh‘M±.A0ûØ*¹å8û¾ˆév"\ƒÕFNKI·¬À" 0æ;ÿ«ÙÔø}Æî5‘®öžÖøìãÑ•ÓØ§t[ûÑÞ»ÔÆNÊ›’¯\¤¾UaŸÊ¸0÷5‰9ÔÀ £ê¦ý "ãÇ_µ,:Š”p2aÈž¸gÈð|ù[È=E…ß&; Ì=pHüŸn!)O£ ¿ŠD“…ýUWTþîЋzÍhg‹¬TÓ ˆË˜qÔiÄ/ÌšyìŸD¼Ñ”Öû ¸AF“«C’ùežÎ\G›] 3r¶ÐÌÒ*åê½j O_ôäαЫñ=çÞdšR?ø=ÃPïùpðûêðÚÁnïl(TH`“Hž¬ÑïñO˜WÖ[ˆv<•]UUrO#”©‘€ÿeXí`'RMò=Þk¯É"ûÝ¡züÚ¾=>„Mõ‡íAc+ë㾦Y4+×ÙNæoÄÔý•Kp×ò‚PüêŒè‚½—6ê»Eš®i3ef gD1¦ÐHk˜XÆ—j•yЪ›(Þ Ø«ä÷L¬òA¸ÝOU«ßTödhDu±-g‹‘Æ0íÎÏ¢þýñÛ¸ÿŠÒ_îÞgm/É•ˆoÜqÌß6 ñž~bý?ó/ù±ÞúÔxi"Y_íÖva¬i[˜0íEo°fì'bÌÐ’¤+5 íÙò£§ïx÷BqŽø Ä:ˆ±ªþíI"ugI´°²ò¬Àp`IçÉ #Áµx˜aeÙâi#œ´Ÿ=ÅÛ€¡é}R©Jù“"àR½fš’ŽØ‹ü?×¹uˆ ö÷»¡pm ©zІ“F£¥±åŠíGç°öFP‹˜qј<ÎÀƒ£u¡K‘¯ÿ(éµ\ñÂähâÀyà€åµ^:֪뾎•ü%Kδ SJÙmDéÒ „C Ä“æL›ìèšÌã§YQ¿`÷»>žÄpºÁ}\Ê9uÔ …Ê8ðý\YnuRQápKŠ6÷¾mKAf—É £×s5rô+‰çaPÄ€¡³®Äˆqá½ä§î?lTò…¤±w›bÕܬÇxþZ¶ëºÌšF`HYÖQŽÚÞ$7ÂVõ8iH÷‹ûE+¨¡AVqeëú’š¬K·¦µkù‚Žæ¦ÈkŸU<ˆl…ƒVµéã™À´Ax u¨h¼Â7B{9Y¢@þ%£E¯ÊS”R*zäB¬Q®5™€jŠ—î­Ž­í¾âɦÅYÜŒt6aÿÈBdSËËÎC¢> мAÏÞ]©½ÚÁ7õÆÄĈEÚǹx]7‹x:x™“ºc(mG Fþ?Âgi঒ì‚áqáÈùç½Ýlãq‡ƒàs¤ òÎÅE¨!èÉF8ðT£ŠZhá_·±ƒ'ª‹x"W\„îê¨øákví"0Œ§.˜OZZNŽO¾ y$C&¼©´55QÝõÓ-NËZúU}-â ¹”ˆPòÿÜF+[#Þ}*¾§‰l;ù›Á\=ºŸÅ«6Ü"OæœI¶Þ¯Z»\ó`‘*p3f4ѬtÎk_xI¼¼ÆªD»lÙûn Ƚã«à¡ÊŸÆsåDžÿFNb¢W²Õ)´%0½•uI˜÷ 2í´ÿ|¤ å ,Q1˜·[y™×csÖn”RÜ]ni¢hÙÍž<Â-—fv—…PXqbðZ½ƒt¬ýr+Æ{붇¸†5¿Z•Ãü¾†™ª0ÁO<ȯNö+«y'dÔ <í}kC|é wo]Å®Hc¥‡¦ö±ïz¢$ê"fŠ}à2š@)—Ì|ŸÉYî,ý»ìr)œÇÕ‚<¹AÂÞqQ÷ÛXãÍZëO1pï"¶§>®¼Ð >tÌŸ¿¶îí=»ÑvoÚ©ÇE™[pàÞD!+ö'é‚Òǵ— S ŠoÜuVûD°ÓýXÙAûß.…4þTú™ñkñ¶§2¨.HÊÌL-úLƒ¬¥¤G‘è¢Uè Wô[6µvð·ÞäH’IíSÖz¡óìâRËÈWKÄ0Ð-_ŽjAïF íýp¯·G³U¼PŸÇQs6³só9ÎÕºÇã¶nCê¥;W»2pªÒH»ÞЬ aKÍÅõqý±™E±icwëḛ́ûl Â5òùð^Ž`=&é&µ ëvF§O/ àù©Á$²ŽASqTƒŠ2ÁkCÍÈ,D#Yä›/PN¶kóý“È:­_Ö¬}Ò7­K“ù›ÐO¨Ø±ã–z£IA{ž¼©rÞÜÝvè¼5Kå Ýèþ»FÖ!6æO™HÔý^F£¡â"Æ&âÖ3©«¬ü×·¾a¦’r¡.éû$çn‘çá"v°XNKÛ6eܵAãnÎ\~õ°{ ðЩ¬jÏ“Ô9Ç’Õ2"F]Å ªzrFû8Ñq$Äü˜èW ïÅƸ'äõüÇ7•›ØQa¨uêÜsä!³<Ü6W7ŽHÐ>Í ò w“Ü„’«B-¸áßTÜ‹óÔiŒ'±c>â•A§ibŽ-šÿä6ód¢ÓºÙZ…d³B" â /ˆ±Tž]¸bh–¡ýüIÓùeßj¢+É>,÷ÞZÐi÷©ÆXaê å¿;UqQ‘ušV8)àZÊTÓý÷ÄŸ8Â%RwJ•]Z²ÊÎ~¥5κ qà#(òËH“¾ïNˈ¢ø¥S"¨ ¤R')4çV¯SâÎ_ÚÁW:fùü ²a¦Á¥w›5†3Æû‹M0iVÎ PEñ±3€b 'ïÔªæ5±c“ï縨³:ìžpãbíj‡ˆü%û.<û£„»§­VLªgë-%)ç]šŸMcê ÖYø‡NÅÑÀ êd=4q¿*¯B÷e?w¦¿Ÿû4aT®>¹fiOo!†™Â‹£åÁ8¤vC+€`T·"ï’Ø„f‚#1ñh[[‘2A„îÈ’Û[)*ÇV߯ce7o&ÿv"Ê ®‡ß§#ÿ¿××|®w[¦n[ 9 ùUiN¶¬HáÎÌæY“À–Àóåãc6Á¨”›ê«%ÂÁõòWiŸå\Ä-— #ИBXâÚžmqY|ÅÀ;™dˆÇ¿&¦àL…‚]Ÿ_B©ýÊÙt¤,ÚƒŒ` NCqÝè ßí²äk»Â8‹K^0êL‰ô”Qg*?Šo¢†U¦››¥ËÊ¥b€x‹E˜÷$QWõ`•ðþÇÀu|s~6£E²vØœ–'¼œ%³)¶ŠòD ØbE´|XhûØú§G¤{ð”Û°d™—vаØá¦· ›ÃÏâ"óäûþydwfćëÑñN°»åîZÿÖnËÈçÁD¸ NÖÁ^O&gã•õV6ð*¥ø\5Õi¥mÔ @»(‡€RƒÈLî”…ûÏÏ9¼à´ ÁÛ_@wV$Ú[‘¸_í©­±É Q°uwFÑÐÔ§¨º£Ï8ó­Âv\¯;4Ôb†^v&R›;²v¨K¦×yr4³ÕÄ‹ ý¾žãúzTC–?ïö&Í·]ÚkgbIsvW´oœæ Y«K=õ¬r0d]žö.`’”|%—ZÖGî~™Iù$¹¿)+èÔêG†³Y¥´›h©ÄÀ~ nϸm5P)8P7)Pm~cBùöà—»ILD ÷PM}"HžÖ¥§¨zÊX€ŠË»àã,L¾U­´<¸ÊU¢yIun,³ój03r>ð9y\u.$‹8>ÁŒ ./T)Ê3Òƒž%`—&úfÊ›ÝÜ»·ÊvÓ4r ‡e@B¬þÕmX@¼¡%ÑõÚAËålø*3¾i³ŸÌT«H°È“¯þxzç;ÁÒ®CTLÝ\ÛQRPâ|™m­c”&­² ³ å/azMÚµØúÆRw‰’óÜz‘›ú5äFº{S/†Ðš*V yI¯“ÎTºëôpÙ FÁKZ €Î´™P98uŸ ¨XÐE’a®X³ª{‰ýOb|3bß…î£Q,wÄUÍC è¡Ðó h|kï?0žºßmR7Ì=>œè_qŠfàcV¸5¹/Ð6Ø‹ìŒCü›«™Ä[¨‰J‡l]9â…ØG¦€¦e=PÈ={~ðžà¼(J*…”!XÂ3NùL&BMï·ï¯ùÏRÙIîóeðÌ€ÑdÇÙêŽüÑõ…]3y¬ø|ÜçôÙ.ø3V8}g÷û³Õ‰ìÓ £L±;¶ Ïzê)I“¼jA2ryi2#„•qùRQ³ÃX0‡Jü4©~%±9›â7ìØQ3à$Ü€þcš¥{ †ÿUÝn&ÁFdåý‹@ÿœfºŒ¹ÝÆpƒPÉÒ‘÷ðü=iki”;ŒZ‰-{þÄ÷éöSâ¶™¹ˆ E!•„Ʋ¥@B|£%Bqã³ö:WÄàLñ#jA¯9~átS ²®[ò\°`G·. ÿá½T ‘TP]9¼N \Ÿ± ç… fSCŸ õ+Ää0Œk]…‚›¿Á[•&ÇFË]6¼úpÒ9Hé0íšð™2Oui" ­ñ5òfhÐE¼bhuzëÄ@÷g‹!ºŒØãî„2ç<Š!å×éDœ)˜6]²Öx·`(G¼=¦ý°Q1|N;¾à¢¾ñS~±¼~¥u%µl®ò¬ëù}éƒPF:J¸ mþ¿±u9Ÿ¨OB?»@qîHvºÕÄó‹²½Æ?¿L¯°ÝÃ;ÜK¿¿o:²Ä\ÓË{PT1˹ÙLAü¶«Ö1¬;ê“ $tj]m0z"çè¶Ây9«A1Cüä-žºq<Äï×'ûïbwÚ§uùÊV¥ÿñ×ïÑl,¤G+´èAÆ×À#6 Ý $ÿZ»ö--üŠ6ŸQHÚ² ­©n°Š69J6â×îà Æ0å¥.kس˜ÒÎå£ËŠj—ÜôÒÐ=Ö Ö%0_œYl7àW„›.BsÍ4;8§ýv·®ê´ÅQCü¡%'ØÝ® òx¾‹\ —Ih’ð¿>”ÂwÎ]<×&«Ü­ˆ¥ó³µãÔE›ÅîŠ^ú×8ñ¤û[®Ò-[66ý÷B„XöGªL‹¹¼¢nALywä ŽÖú¡óÄçv«eËW‡ìbáæn³J^ÀvnyêÍ2%ÞÅaómAä¿Ì:,ƒ¼hp¼£UTÔ/¤Ðsd39ãeD{ËØ€”˜Èndƒ×Øœi”`<Ûü`åŠÀ މgÐÉ· Üxþ5é4] eÿì¡¢G¼joI[3Näq[—¾{.ÔÜEÑ_È3ƒÒ‰ªëÓÿp Q†A\Že ãÇNúÁÅ­ÂTŽ#fsHä~¯ËúÑðôîvÃn°¨ï¯Á*GNïÉ.mM3,žV¸j-™~Øìâôg€i÷Kf_Î¥¶j"˜Jl¢‰ÀSþÚjòn?;í’ÑAˆdØàs*´«­•¨xlXœ•¬Õ3‡þ’×iIøÏ6 íšKôHÕš2Õ\%£ì˜"üÇ&ÈWR¼¶y S×HW¯IÈêMAx*gè­aZp¤•ð|þWóM´¸_1¬@€h}zÄ+Y«ÏpO]´ßÓ˜E*ø ^äi-o2÷ïµ—:¡à€ÕÙ@"Þ‘ú==r¯°½Sƒ”Qän‚á‰ÒG÷nýèŸm3~”aK‰ §¾—Ø0«B2ùË3õÊ—ä8fwÊÂbKQ­ Å?Æ‹®¯Únôsü4N)»aC‹ÚìF­¨„YR4ÑžO)©®igÿ÷ÚÑ‹‚ì:®Ïz¶¦îggñèfÆ„r&´—PÉêÚ$ƒ¼ŠËm†áÁ£kV¥8k«kË'7/¸Ü”³4-­gÆÖF®6™„5A2yYøÁ`@>³±R_7f?øÛÁ‰rn+ˆ­\À³ÝÕq¬AI‚-Õ]“Ðht,àleO2oí5kCR;*ÍËCØ×If#Ž…0[Ô_»½´„3ODÏíåúƒöQ%é‰VŠ6ÝDÑfTY§)Ý¢ÏAñ§vAÔ,}]1w«¿þšuGö3£z –x ]7uKº_út”›q#[~Qü+p†¼™¥gõçIQd ðE(÷~<ß󴨲¨Í ªéWâ,#¼µÏßþ“¥íçáLÔúàÆö·¹T·“–„‹"_‚{úMÆ¥3Æ$l|s²ÿSœøV¬ÿM°mx+*_³–Š;„“’…ͼ쟧¸„¯ØÎëi:«¥ ¢¥JSJÚüï¢É6‚g ÜôJ2»ÒUi:Ãщ&>[´az¨¶<9¥Ò22qb9>lsé&^ºi`ጮ yǶ„`ì£.ôí»2;¥Ö}r)ãöìÄÏÉe8š ªÇ­´hFá§Óy`J-îHÐÄÚ(’v*@±Ÿð`ù‡5{§¾ú{_@´$²¦å-ED‹ ®HÌöÚç:g+]OM¹3xŒ¥ìvU x.§ Öf2ùYvÌ1]=dgøIxÍ Òd}(naTíìÉ01ÎQÓ$ç¸ôz;z¥k]†ª2¹ü²z:íyâ|ìwð.ÛLG(™Mö/(%ðOÌ ¬É¤G7ž¤=`Ê6@õ§$ý·õ•ªîÉá¯c­þÇR «Žþ¿ùþÚ|ó5¤˜³Ã0rÉ¡Ð{ⱌm/¹` œ”áq/^’ê^T°åÌ…ó‰…J`¼¯ìÆ‹\¼u2Xù!U…h¶Ü’ïÁ;t‰šM &,ý -Ž‹ØL#F®ë•4u‚ç¼så²ý»û0.t9Ë/ÈÕ…B¥Åøò‰=Z2KzÙ¢ø®ü[.k e‚îð"ê-ÈéÜ”C#pÓaÑÙH륣3Ïœg'qÎ J0ÜÔLý›§_´Ã¸„'Îqš?´¾*­´Î]‚GñºÝ9^?é*ùždQ‚÷©x¡ê‚œÚªËYÑ´T™C®ƒ"äð˜¯oû‰æÊT±!¦þY`µ¤ûA’¢ÀùÙšø‰¿9ÄâŠ>¯”ÄÉÀ]@Œ2Ã/†5ªý]ÕNÆ.“ &‹†›]“Dvå}Ihaª³êäÒµ€@r^˜Ú2›õÔ#µÍÄÁJ™Å;wÜÄçž‘sƒN(ù¾?FÁEbªIãšúëA¶³íaºÐíáå€g],[¿Âzl—ûYÿwc$Ðf§ã„Ò‰“2ºüýky¹^:!N1¡”Øs°]}š!R‚Z'“6_ÌcŽŠ‘ºÚ¬XkYŠÃŒ(cØeÕp5 Ÿ¥ªú4½ÄA˜ŽÝKË©©M©µèÔ.O“ð¾t’É> ñZeÆŒÆô7*昴ҀR˜j Ÿ…°¸˜¾íºÉ°ZnËÖdxGã{,8JS BOkc'³à pË$7#3uS³åîãu5)BÞ‰ï!F\sk¾.¸ÈDÌvaÀèû’F®K¦{*;ŒX ˆ%2I¢-Ûµ6UGÐ$‘,Àˆ©óé»3јPØÎÁ•–s;^o]}z8ŸºÉ”³ ñ.}J;£¢ÛòJc¦$à i¼BAúƒY}oß4$$œ©5¶s[Rù¾z®§ôâ¿(aåØ¿Ëúpë(šÎ§Ð6ÃÏ0Y‘3Qh‚ISŒz泺ÅÕ»YhÝÝ%;Je¾í·B'$°™…R§)ÿ÷dæŠ#ÏÃ9®5Acëß6£á>êçKÙBØ/ãP™à×Ô;øqQÙ‚ö\µ*ç@p‚ȬhËõZôµ}#w^qáýßû¢Ò?h€ÒHï2õ|tƒ›4°¿7Oºþ{S ~£a@Â,ù—¬ÑŸg/P[×™?SîšF‚ºý¬Ó¯o;Åýï¼ B$à]/pH3å½+ÂàE3#j„–°¥‘nAƒ˜] ÏÏØüˆoŒé<€ü1IOS¦3µb :Ð^mÀÙž~¹ÿzS™¾@šã2©6¯œ{Îæ þ†y>ËŽ¦íŽ%מ֟ú¹†Bí{Å<”¯=0¬6<–Â9ÖJ»Yï†ì]:LjM‘tˆÐ嬢l ‹w™j½Eû²՜ˣÀئP.A%×!Ü#s~Yà…ö·6¨¹ ùìïW| ›ÝšlU¥êqs°f‹;ø¥J—ûBÂ@@®rô¹WK§öŒ¼ËúŒ¯†U•›ÛÇQÁO´·ÚR“øRÖ‘ò˜µqª ›ü!.šÈD%ÙXòÏÀ7݂҈ ð‘ÀŒouÛ­KzoÚ¶L$›;A+N¥Šó›JKÀÞÒ¾™°™Ò2ö4f*‡.|81(ä dÀ(O§…׈þ˜é@ í!È¥Äí¥i‹ð»YWJy¤*~ÊeeòÐþÉ1ë]¦„“¥DŒßú«”G Þ N òŽê²Þ‚B4ÂÆ°fhBUB¼g˜AÄ tÓš8 4ˆ^ïÍÛotª*:–=6OnÔ½÷_>í}=0§dÓie9—"‡é†ûеUš•…_U=ÞÝ]2,}OÑ"54&QHÙ!Óµ.®€V¤mr}¿'@ÖËÎy]Rܨ¼²ÑŸãËøüPÎÌù:áþý²Ñ¹ívÓ&ôb  á9{WÙjõ74_Ô¸¥)Ýï™ÿkºÍFdÇAÄyÔ¼ÈK.UÀ T¼Ç8Kmÿ±Ý& Sb4Û.¤gç’_¼õ¢8u}3Å­(Â)UÝG–ù3|pÍÅ:ÿ““»¡ ÚnP\i 79%Ѝ•%ÜÙ+Ÿ:æîÄ àœ?NUšORÝZ„æN bÅô `çì 20:*1x!ØWžØ°¾FmÿÎæ;­ŒBÆÎá]į•RqÅTdlõ´7÷ÂŽÝØÿ³˜;×Ç"ºvܴB$Ü=‡{¿y¨ \ÕÆKt3ŽgDÀ!·ì2yÒÝ–9Å(Ä{ÏN öI›¨ñ­ØŽEÀ,‡™ÊÇG¶Š.½l¶u˜Ù º4Ï2°/wé§NЉÅ:§Æc §<î/Å7êðrYQ×RÙ¦7Úÿ½^˜4븎S°Av2¾¯÷£Ô*ûWʇÃ?0DÞ_•’\%áÜ.‘Ï×ÿYh¹ì®5ù·Gø}1ñQ¹øvð0ùðóÄvud½#õÎ3âá¡ÏfyQEh´ççsŸo1`¤Hßûqcaöf¼9)ŽCvW¯¶Ç áøP3øB¿§R6Ê/°r©˜+rÌñ³ZËî!±ØF’Ý•ÄÚ@ëxäƒ#°ƒ!ñt"^¯Í¶ï£ÞYê{: ^  ’³*‰¸F—Ï_äÆz¬ 4ù°ÒÇa©@ŠÃv∃gÜÐP1SY¤l¶ÒÐÙVätä.È€^wNÞ_ûà™Ù¶`˜dãeá+1C ypflVÙ{¬ävt¦Í¶Ñ²}ˆð‚ÛÃFŽ£h éPnpgŒÅ‘òfK æþ ‚§.ÇÝ8mk÷a__Rt»8%%nçAA_±Lfƒ+ßêqAÓTL­\Xž ÔïàŸîüÒ¿ä,™)’ÅÓE×Z΋YÅ›‘x×½#nGt–¶Ò*3ßûŒt„Ôë·1Æ\ÞDíÐ^¢ºm8P]˜WM¸æ6gq™šÃzO>¬uÕâ)Ì],­|Eù¬dH ]Rü¬ÏŸÇw(}— "SDyþô"Óäj(Í”€¹…A¸n帗$>Ë1ãÉ­4X8š™ešÄ©êÀ—²é•„¬°:‰Ú»yr/0\¼\v×§JLŸÄW§_*µ˜ùCiœ‘:¯XJIr¶Ó2•W²‡Ìm1™ã‚§÷ÒBY"¥þ˜§w.×qVGºïP«92ÇÁ`›ÃV¦LΊ?p #'Ýu7ÔnXÁ‚^nøM¶nG©0¢-ke'}Ý&`d6Œ\!±LԠ೬;.».qÌ;Õ·©qDËõ~p×K&Õÿß¡ý#†L}(ÑTI‹äc"»áLGoqcOÁÇ(;›Dn yT/‚¤;)a§eíCÃ8šÞ—-H”pz>w!·ÁALv» å&\WYyËðÔnÉŠ¥ÉÏf-ˆá¿Z'‡ëó è‹Ô§hÐÈæcò·ïZòÏϰ¢·S _ðSTG!Nö$*03üë© "dVO=†&ǃbƒ³î1p¤Æçd^Lfö¾h‡Fó-œ…À lÀâµz³ý[qÅ 3,œ×ÿEùô”ó˜ËEëA<¥tvwÒžvjóV_CeØÕ;/õ—ˆþ4àD–Â8,&ö§4úO2»xÆXK.—ÔC®VÚNaKÏí•J mÌX[2OélcƒHf´>1í°½ÚQtYq„vWûij ´¤œ¸¼$£éæÇ$~º@1ñ³AH%¦P LíÕLÙ*H¾¸‘e7”ÀòöÓ*mÕ*é’ca!&êIK©Q!rËx˜W8z¸ …ùµöÂR@¸ßìŠwÞ\ÊVLÕJ¸˜Š,ã,R0gÛºæŠLqìä+û±Nr½wwQ Çù„l×5^°¨† ’™d²AÑOÒÚG—aVË:š´&!ùmÏJ¡hU’è°¤')á²ã»ÇãDQì ¦@!u„¥oìA& Ñ#q#‚ðç ·è›ªæíâïZXM’"¶ÛœÈTKÁ¸3ƒdÚpsä‹ÿ@`¶µNî ªYwb¿etPâU7Õiª i»m£ L1†ƒ~Èïå©é‡ÍÇöÚvmRö;Éï¨cì•=ÂCòÆLx#3.ÿ’Tm·±jh.à0׈$Eu½Vm˜Vñ1%´xWä”7ݨc()˜Ö^_ª÷ó“Zá<¿OpÙ3OŸÅÍnVœ5Û–$m5$÷ÃYÎÌ~n‘2™ã¡z”>:HE´¾·â·Ô"˜0åqy¥øÝ\¶XÀí¦ :2Ñú¥˜êú[¤äÓcn9亙¢æ÷Û›¨ÓÕÆ`šï©´Žé¼ oR¹Œø2ôo*ù³›ÇDIÖWLé¹n7Öj<¬ÐìûÿXã’—“Àæ¯Û}û=4[.:šl?Ík•Jö7zõ*´íO<¥mHª-q±(zï©3ðí0nó­ãl“´ TbH/¬Âj%y]¿Â+(þ´çZDíwT$žðŠ€$u‡Žƒ“´Úiçs¤w`p?¦fɵDg n°°Üuw4Z«XܼÛöØïò¼Ê"U Ò\Žvz‡Ôý+gz»¦ï6›ˆß{Ü4+Æ£\82ùÄûi¢aßôkÑ—zÌÿ!LfKØà^hä 5ê×2p€vÚ­#3L#Ø­7€¹‰–"­T”žÜ¯nWô9QÇHeÊzNq-å"¢8p}Ût6]o?ëÚÖ dÀô4$ÖP—ߣÒ,΢ü׉1Òú3ø|¯ˆÞ!ÙœŠ2´ãÞAçqqpF¯Ô™Û¿JSµ»'É)¡ÏØDhœ¯6ÏèÑ÷òL¦óg}Ë?}Â%j&¾eú£WÅcE#ˇœËk ûw·/%yºø!Ù±‘}.&'þÞ{®Å·$„hm`1¨zú=ÙØ…$´è–D&qÈ\BT!=WZ-ß’*àŒ_àË sOÞ”ÝBÚ'§5IæM|Ä•¬‹¦Þ>“Õ¢¤am‚ áÆÈòLhQN8Cã9aáæ”˜ÖÐ,p7ãVÝ=Gg¶Ì žóǬ~Ô%%Påx¼ ᜋ–ÙëÿŽÕÏSQ•3ë`’ùÿ›ÌX¶*ñÝŠ&WyÐÚ¶—@åÃ\Ùxž»Z•Ò*UªÝ†¯ëÒ.+ܚѕèn~w%úwƒƒÇƒi™´%¾!)Ço*âPɽ9±þ¢€èW1©#Í—Ã=üã ´æ05´ ¯œ/]LzÄc„ ì‚tïEÿ[ʦç>óªÂXfG¸µg㊺O›7ýí@öÒŸi­8Õº¬3òRà  œ€çˆÞ®è»kªxÒóXßôvFà4ØäµOÁ¯®Ï»aNÁq„©oÌî—±nœðåIbB)rÀ=U·XGâ4šÍ¶õ6×@v¢4>aNbà'rä¾9ÐÕ¨ZxëØÓÌ>IH„dµÇ]ùoc”ŽÒ`l³úLõ›ê'[Ó˜¥¼àìugOügÿyÝ71uN f]¨#Bž2?ªÑ$òÞâ;ÇIOÆÂg^txKïn@Æ‚!YÏñÆ à?ú–yÆTT« ý Ã%‚ø@Æ'1]Ö^n ëÅÃÏ{ŽëüòLr(—2„|5–\@I#×ó?°Ëa)Ú3q!gPžÅtÌa®Ì´!6¹C€H†vüN"¿Yç}ÞL(Ý¡z=%ha\à'ÂJÜ5ÿ5Àû‘XèVF¿„ª–CNqð³µÁd 푘OÇ{ 9k¢ÔÁ•äâi˜È ¸ìk€ˆðÓüþl8“ê<¶}ÿ¡M~ø&êÉLÁùI/s~(f?! %D­* þ¯FêÓùùgÉ„vcˆc&׸ZP“eD5ðy–‘Uš„]·ØÁ`g*IŒãå í5Α†y(&¦°ZÉ~•άӛâ-S¨huøÏ¤¦Lš¸ŠšanÍÇ*ƒ›^ÿ?´MsB+ϧpµú³½ß îiŸÌïöº$˜£3{ Ãh‹T}H“NKQBd®®Éø¾nÌ5þÍìŸüZa6ÓzT!jRûÈè×Ú쩪M”V@§ã~ô3-"ß'íÓ3KüϨŒÈ ]z9!éÞö(!þˆd.ªÐˆÑõ—d~]7V=tä÷)1®6G3 è ƒÇ/mÿjûn¤ïLÕïÜ.}IÇ"Õ×%ú¹n·Ä¼sr¤¶¯Á¼¢OÖ‡_µ^mû)ý¶³'~«-þ؃™CÝB£®ç>\ô°ÿÈ»¾ J~‰,Í”–®RÂçgg_ :šŽÎ5ŸA »rÇ;Þ¦–ÅT!sÈ…pÎÐ$ª±œEãIæálÚþ†àoX÷¾Á×Û«8yãI>jÏ€»ƒMXVaô2û®“<¶x óä’cñµã§Ö¦Còv7ÿ™ã42êÕ€ E*ÐPZÕ^_q—#NÊh¤>ìVö¥Rj«ˆrhF„asÖe#|Íø~µW&?%{Žsß–¡gËä:;‡\u4ó±¨=8ûˆ§h{þ¥ês9W²#l3aH‘ç{`HºY ÒµÎàJ T;ïÓQLV ÇìÓ /sÛ£Õ‘Y-ü,kJƒÏ Ÿ ú 4éÔyx?'_3Ž YrÃ⑌aɈ ƒæQ‰Pë™$’¢‹Dw‡WŽ;J\³-ÌAš}…¤£^t`{Æ:EÍ+UDôÌÛQÕ.¿+¬ŒÍí>~Ö0 n¶gÀT®t?‚‚šn1”xŠ*Šô(ˆÎ î Q×½åD†€'—pÐHq1ˆŠÑ‹¡€f¹Õ;©f…+l.(RîÒÊ(&ˆ,íø¨Ñãç1µ˜Œ)~ËÁH.Zèû ¥6Î"ïÄס•²`:ý œ$û6Á#Ïkw4' ;ì|àŒ“ˆtx«YQ»Õž–0a÷ €Ë¸9}ã­Þ’êRScì Ç’Z½KH¾yìÅ“Ž ëñòÿ”´Fš[ä§f}Ʋ"ƒdÆáµâ áÛ„+DÝTp-g+7÷Ws4„µYre÷`baâëÖïK<|>é„P…ŠÊ7«Ëoçï,¯}IÉcl~×E^@ºq‡ÀvðùÂïŒhqN;¥Ñ[”ºMôpHBZ¬2³tG‹R9Q@d­é•k÷´ªËÈz ñx‘ãÕP%Ö.o&XÛrÕfº—4Î Š_®Eyò®¬ g:ÿíA>#vkq}:dNPi?Šc â˜I–”ˆ¬µôlufÓžEÓÿÒw)ïê­èV}€Ý«Meù¯~“Ä|ü0!Ä+ÇQQÆ2‚]|Gñ}FN4‘عE´5ŸÎHeçy•æbÃM=ÁËÉ2lQ𺪯uRµ®(!+ûªh{“àv¥ §è:ùŒqG>N8§ìQ·‚ù0à¼HT¢ g}BYa›ÞX2Wè²@ré9fòßâRæÎ5™î܃Q«5­•7}ÿi¿ZL0%í ÷ —«ô`-ºUŠ]{fqò6޼Šˆ™x{9Ÿmª'Ù€£=ªš$"AðsìEo ~Ï—n…Af>´ù³|ZPÆØ}âN)Šc3ÛEÜÅ/ŠwøáËHF„×bß8ë°è‹ôëhä¢.Qª!`®ÕI{¾„c&-²p¿”…Í«%eG5Ù&øuZT: ã{þ‚æá™~Eþq@à]= ÒþÚCÐI{€‹qM×8›;éW#õúÏ“ÁÁõÙ]¿áçùÂ-ù+ÐKio±È†»³P¥µ ŒbZPaï¯oäL.?¼Ss*žnð|–¿ZÙ¤9\@7ÂÛë¶ÇM "¨4HÚR\ä3x+ù%“ôå0 r×ç›âÔú£¤­4ο|[Õwª?i4Ô“²á|%Aµ^îo’ÍHŒ 0•ªS­æsì .ú>-¹IÚÕ„ò‡…žvælÝ’ãÔ»á&¼ä(uêi³žíìhrR{;¾š-²`sX{`ªC‹êƉ/v‹ÜB|)®DŒq¢À÷4 ü¢þÀR ñ’bñ¨aômÇ/ûÂãù÷ô<:K­”;ùþbEÉæ‡j÷‹J¸/ùÿ:b“‚Òeºx©ÐWƒ Ø™ÜèþÜ¥V©;}*†Xå‘Ó±¨˜n0B:8¹º•XfL¥z¶5‹!£%¨ï†0ýÌiUGs†i7hEv+âÐ&:ÅØ„)¸”g<özÚ,lâ ’”Ö+¶|#£Ô¾€ßõ)ªçº¾í>XÝídølJq14ž/ "2Êhì&0/’9•ôÉ‚×(k„Ðá9Å/§!±›­ÈµÊ.xðúì“VЧ¤\_J…áäRÒ &q¨’Œõö«îå"@„Uâe«g¦ökË=ÑFjñþÆj@¯8Ã÷¬·àK ‡µxHx~î®ePž=òV§1=Ê`±*Ö~1E}!Ñt2¯’ñ°ÃJ‘-´ëOÀŽ'5ÕjätÆÉãzíɆô¶5©Bˆçât¡^ ³GZ(V‚”Hf‚@q ¯@öm '«à1‡Qè‘Ó}׊ )ŒËù;˜MÁÄ×9ÒÅK—ƒY©¦Wû/$!üvÇ<˜2áÔ @nè+#rùÃ×êõ /‘g¾T›Ãó^G]äz:-¿E.ÉØüŸ«q?u,³ë§’n! ±P™ËÍõbVÖÇhÚÔ¾ŽtÇ©{GG ëN¾ ³ÒNt€‹Ðx3Z×9ÂËʵï겄}³gæŠk:PkÿÔÁÚK#†WJmž°ô¨»½6¦ì2m¥FßG{6`´¢Þ©RÒ.­¼Ó©I›V“éé"׋r‹/Œ?å÷»ieîwLª^êŸÑ¿¬ ˜ï41_j0VE9›,WÂ"xgKºçZIž„¤£Y¡+Ÿ@)#³4ÆþúÉ9YQ ‡ú4–’…~ŒÅ…üÁ¨£D9jŸZš†Éä_¼Î!a¤ð‡búÀ êK\ÙjòzM­ÊOb9•舼µ¾¬‘8¹ö`§f{9HCÚŒÂ23eÂ]ùó½ÿŽ`1%gDÍBV PWÅT ËDï/BËÆòhΔgÖb3‰^ˆUð]×,Øò-GRùB :7´ ÊÐu*CËrC3ºøo¼H„f›g„R®t˜Zßµ$w|Êßw|Y¡CN@ÊúÏhÊèqÙûiĹ˜f'ƒ£·H­dÝ“K1d(éÚ–5rB#zê2§±Œ\:Zfü}Ò¦°ì]ÕúÍÑ™W÷wÛ»Þƒ´ÄÀ5ÿ‰æH†@§‡¯iæ7ÏV¥õÏQÝ#"­§PÖ±R³[í<Èa9Õö­ïÀ88íÔåÕ¹‹YA ±£>ëAþÎ=×~Sòà¿¶g¨«ÄÅ×6uÑÎÙ8ÜψifÔO#ñöP`e×”T¤‚ Bìè>ÙGc¯•žC¯|ÖJ2mF^ùÿ˜BVÅjCŠÍ?r sÏþÿ¬ ÀŽt¯JËåŒbƒè”MÍ”oý%4‰os„ó3¦ýµÜš/0ß'ƒbŸ[Õsr£Í˜Z ANLðzÐ-Ð0,86ZX.Q ül‹’$&çx¼¥ù±jK}ÚÜ&•ˆóA¹ >™ÜˆñMMÁ‚j׎•ø…LÍ 9‡úXzθèÚÄx|ËÜ ýÜWX™ Ÿo™£3çœÀlÎÂÄ; Îö/DŠ ˜;†PȪ‡¼î X…]”F‘ŠZkOKË5|¥:÷$ë¾ô§‚qßž…ºŠRØ©c¶*72c¤…–U ¯¶Ë~wEiì3a¿¢Ù41Œ?¨¤€^îIY§>³^³êC K0Ý Ûø‰*.ôàa\Åûãù¯l"汋²­)ˆ¹—Ñ‹>÷ž÷à“Ìð¬Ií}4Œ„2 æ¶×ª€NR÷N 9&¦Ô®Ú†¸Û¨}:…â?Ð1?t­Fí£rdÔ·ŒrÂ4雋X…|[–'‡‡kÂÕÇÍ`ňbÜ £ —©À4ßê%m Ø}ç!Kÿ³æk\]_#Ÿán’£¼ VS0'ÙÿUÛª';ÔC&ÌžõSWD,ýûrœiÐZ[ã.aa»È.O5‹ù¡Ã{ËYéh¤`ªH¨)ñ:A1Å9½•ûw Ãj…À’¥Œ úÊ\¬ÇÔÈ#Ãntˆç›%þg巛У‡Á——á|á;µ…ôE’ÿ®$ôÅ1#~ì ó¼†òUè,vó‰ƒîe•ÄTÝCa–²‘ZÚöÃNžÀ…°ÎS±Å ØWÎiKŽÀ¨ …iʃ`pŽSYd iŸn?=éÉ >Uáüéß*_´žS¼#ðŠówk…fQEj>Ú.žL~´;ž5Ïô·÷+*£szK߈ç•I\­ìóW0ñû·Ç‰Ø#º9 Y"úµ§ßYš¶õèú`X¿ÈÈο×4ã¿—qÂú „e¾ªx–r$Ì \U‰ŒJcWA ZÕ5€Ë>ëûÁÊ,Œ_š€†Ë Se+®;&Z\`¾© Eꋼº~/ž”بYÏ’¬ –8î´}¤{_¬Çº¡ÐÀvjG÷›v±¦4™¨º0jõÏL>T‚[9­×ÊqQ@G#°Yudƒ£Æ%¤í<~È2º»QA­à«JÒøiü!Bñ‡®Úþ|bgfÌØdo.?ÔÙ¼5Æf41J~9<ϸ“¾Ë MyåD‹“?.Ó€oÑèÒ€ßÀ߯U—:á¸Á¤Öª þ«Lœ—Q?!Gæ\°h›Ê=èÔE#ºM‹(òFvÀ)ÀÿGµõv:?Ò¯ö#(ëì‚”_˜RÂP*Ùó'›74,xÎÉlS09/¶gα>fgðX6ä/pÛp¼OÓü4éw†\á’¥²‡«ä›œ¼¥Ñ¥è7:…@)Ý×€xtƒ>T$×!{[~@3_(ŒAeWÇx—Á†4¸?ë%s>òŽ;îó´йH&†ìRk ¹¾lPÙÊÄs¸¸w)¾ÏÃîfä–&I+>OR °´±ï®¶j€b›rbâš•Ú7tpw¼ô.ôÕŠó–A8Ç7Ü¢™ë¥mÞ ßóGµÐÈý«GGïBç%Y3!—):Zå§s§¸>QÔ8¿›„›•É0jk*èÐM ý¡›yUŠþ²—íæäm¶Öw¬‰0[Æå¦ _VÖP›R'4Ca,«×zÌcÛ/tÝ܉5žør³lÞH0„d”¤]˜Órî<,ÓU³˜‰ ƒ:ŒZmê|J}¸JÒÖZz‰BŸB¥ciŒ¾—´ƒ)wZ·ÛÓ&Ñ€¶ý³¬O©XE11[sÀݲœý™ÌÇóù•ÑnÎôÌÀþAÖ¦§$ŒÛÌ̇µ5W`¢ÅÛ*|±\PJk£åÓcf Ž­1~ï®&6¿wÜ ªÔó£0ìmÉ·U#Ìõ¢§XfŒ+­«Êe„çvÆ8RîÖOtZôVƒ¤°rØ‚ïá苪ˇZ–ñ¹%ﻆ“ð,ÖžÔ,ww ªþ%´Æ‰§ ú¶DñËÈ]N:ë6:¾ÁÆÈ>6G1Àl¾#"q- §“ ßeƒ55ÀÖСjòf·@¿Eƒ'è –æ¾!ÞŽ®Ä²í¯“Úvƒ¤Ž»Â÷ÜØëþ;÷s5ùˆ,Å-”,é¼ÈäYò¤®‹w,B2‚¹öQò;¡©!Š¢â¨\ÌàA0^À¬6ÛåN‘õ½ ñ‘j¹Áe —Èå« !|zI8Óê++ØÇOкO tn@¼´«æ`]}Ø;Ìt§èþjúûxfÉH‰’'òùÀSqL²bÐLH—;f6š.w‡¬;°K]ãÞ.ªíx©mm]˜êCh0ðúÅ+7Sÿ—Àµa÷˜šÇù$®8ô¾¼WúY’3ø×…rðÛèèâ+}×ÝþØvR…¿VøR¯‡6üHd§õ Öu+Ïí9Hiì_hõ Ì°BÇõ­ï8³–Æ^\-Ù×_û˜œ–ÚÔ§¶óOØíøèçóɹ]Êë¼À²‚.;)¨3eíxyR¶07êþ"ü¸p!z“UE0µ9³Þ@k`›/o¼Pp§“½# 5'Ÿœ÷Îp›! h}Á™äËU•#f+-ù°Rõáãv¼¹M»ö¸†Ví ÐöؼƭÓGf¶‹BN5lÛeˇ+=|ŒÔe©4N,XcóKAúkt‰Ý©í.Š-ƒ{0wýDÚjNqïÜaj£±>Õ‚P¨Óìû²ìÆ¥9ëI¿¸¥‰ ¿Á Ÿ ÞÜnsG™¤Üæh­›&>ïboòVÀ¼ªC+‡jº!ÜÙ.W!²aº½ û´¦ lÀNE"Çãž·Gça•[ßkH7Æm8 žZB®Z;ÎYžû”‚gšt™’œuÔw-6¦w “ð|ÛÛü<àšÄÆk]9Âúg.CΨª÷ÎÜH˜t´Ä­ ol@X ˆFª×4÷çWq(ðc>öà‚€Ó¾ÂLÒ“œª^–Ì3,å‡åIAôCÖÖ¿ÕØ\¼dyª—Ì8éXæžéæÛ«XÿyV´$j¥Z„o|ÖÞÌÙU%|úÔ|Oª–<)Öòr„H5o[pÐÚ™‹.º»%YîdùÙ 1ÄRcLhyª åvý⼦ÖX½ p6s› ä¶ã®Ò‡ÙÒÑÆ^³Gº°,6cßòUò’ŒW PE¿¹v¸DÅÖÎÝd̵E-õ*WßF»¤‘ångf†8¼úW ínà¯ņþ Î ³.6é¶EŖݲøa‚ ý6†0Ñס²>&súг&\Ö\ úbG,ÒÈÁLòRSsÛ.8aäSÀL—×Åvº£ŠPO%¹íµç> 7ÎÜóðâT,LòO|´åÖæ¢!@v2.Ïù}XôÜû€ÎdOØ„v,XðæÄØRàìXV‰ë38>Ï»ÔUTžŽMkõÆæÈÇ•;Ê#¹Å¥'¸­Î6‚ ¢ *Ž«ÜàÄÉbBí1 Ü5c¹žqÿ ©sÄ˜Ì ¸£ö€£W)ø½A‰(ª–=…‹UåËÖÂŒ¥€·¼Ia*€‚ ³÷5B÷ð¥#v*¼+~±'÷-XÊV£«uÆ1AéÙ¤>yß )MÓ.%éá-Ñ!… ^A1â\úáS*OaÏ­ŽQædm2¡kÚ±¸¾ì][3·*¨O»ª÷e”~Õ­íÇgx!wôdÃöHÐð?½l ÷ó´Qä¹³\¦Q!ZỸ=·7öÖC”F§§£íÒü½`aKQe§™ËD –2°?C6ø·ù1¿ âŒsbdøœN'‘·îïçGÁCÎAá¬Õ]‚t€ÿ§´Ûßú€/@*(ù=·lPÊ1ëÄû ‘ðVl;â›ÒÈ Ò÷ͬ£JöÌ%Ðb=¥Éc@–¾WÆt{]ˆ;ÆåÝüઅ©S¿[zD$gö` ¶§.ßrô ½íRhÈ#1oÄfˆzÌÑ×hQk•4¿ yxu±WLNE ¦„ìÿiUš‰H[ö¾¿~¯ßð®ÁáðE ÚŸ‡ø¼¹>àOÄÖq›whÙ Û¼Í …;Lï}ÛÌs—ƒŒó¾€ú}jûj;RÌ@Áœ~Qíò¥„¶Z––ú æ ÁŽpœP¶f:Ï·u¹ïÌ‘¯ôG'ụ̊ôþ3qºXô¹oú š ñª #f\«DÔ‰¥Ñx…9º}PLW™ÚøBCÉ W²¬.˜1‚©+ ŒÝ§á±mQä+ú"†ÁpéÑ·• —IV™ 6P%p ©ÎÃM•*ðV<À¤ò+ÿUÒ"—jiÎH;y¸Ìm¯´{ìößq %D=çg+ƒ¾ˆ¦V¥¿T§¶‘Û£«€}¹J¿KÏø¸c#úÅ}·±÷#¥ŽcïJs|¹ðõvÊ\ßYª@ˆWÄ}…HÑœG¯|Jš<@·¬v͉‚ƒœ­5G•ö–hÇÔzå ÷FŠ3·@áRºX(ëÊ®Pò`ª_èsv©Ûw‚›µÂf¥é‹ÆÆ¯Ö·f#½äÉÀg˜½(u)†*d¸'Ð%ËݼyV4çOKc:fº³`þâ¾þ¡EŸ¦Ø©œM¯º ˜^„¥Ï7€±³%S’Ú®t³9¡³ùxwþ*ÄæFå}Wö·u»H2 ë÷÷†àè‘ÄØ"ð{÷%høõ>ú]Ûë¶ýâ)e"¨p—Ù@!¨Õ´µ†)ë' F1Å¿­ˆ ƒ–gŒ1ð"žäo¤Hßo0¯æÒ&øRE,MY…QTõ !ù)ø¶’ã n+»Îè¯7œE{Öê§qÞ4Þ7$Ñ=4#ÐNÂs¦ÍJ3ª=GáC;y™ì:êŽì&ÃN‡Qj-yòoÍ­ô‘ Ž)eÜNZCܳ_BÖ çÇö‰˜îœßøg‡èÉ?GG-¸z„lË.Úf‚¿ª¦OÕ6#’òu‚ÐÝ\>n-]Þ]J‘#lËI(m5âmÒ} !VXÜP^ÐËçL+§vø7¶Ì<(¯yG©<^õ4µ;“,¿I±]Ⓤì\Ò¹šû^fßß.ûöÆ,Nnà@¯S™1O, øN7w× ·xÿò>§I®>ߌÈo]R{ùÀ"Cd"V&³@uÑ0Ý‚µŒ„€\¯+r?CÛ7 vëƒÑ¶ÿÝ5è™—­õ9iN¡e knbÝl&ñ–´ˆ½1Ø¿-¤§|éµ7‹VV/ ¡FÔÄü4‡]Bú…Òºð¦!=ÑÁ²9ªM³åîâ½±{)ѸQú3|¼$Ë Z½Nî€æù1RNøPŒjÄmã=sQØ ÚF¶(s7βŸÃ_1pëd*€)ï‚óurô²rói9ʀʜ©Œí$ã `ýÜ0 vVUÚž2û€(€Ä‚wùøm³ëV"”„5Ï_¯ÝM1Xi‡GÃ(/AïzÌË,ü~›2‡Ëï)_µ+ èç`½z×B5ìPVœLþ4 ^cõž ¸®$P øcôe–}gý0n.­©Kù>.|Wm†›” »B‘y|FÿÐSQXƒv¦PÊœ¦³…våúHçjKœÆÝ4$ìn¼ˆ‹wßAÏÓþƒÚ·ÚR5ŽÉçi–êµaÕOq§˜U‚ ›òÈš¼”BÕt§ï¡Nåî5šfý¨gr?Ýb:W _‚RzþÖˆlG”ÄAØ7@‡ ‹Ã(ÙÄÎU²%;865EîÏÝmnIØ•>#¨œÃ›9Ø‹Û:š­Ø €Ás<ýG¡~;ÂÐxN`™¸¨svùª1›]ÿie©ˆÜ›¶Q*oɬݗbj«ØJë74­R6Yµ‡Ûf—`ÐQ3Û7éÅ·£ì[’@ úþÎt'†¼7ßHòÛ¨IIë/жmà5”Vðqvùšÿl~LÖ°Îøa’S}”øŽý«ÅÓî»ù’AL¬וœtZKÚ E¾”iìáQìt¼í/l‹(ü-ú †"Lý•=†í B,=øÔÅÒ¾g†9à–°v('Ú‘}½vå#C|0õ&YwÖ…µr5KêâwvÌ]ä£2eQÿ¡Òá^RxŸYñ*òŸEš¿7•H/èåÿ(NoYr~ªYöÎ9Š>¶‡;bÄvÑ©U ¶r‘89µ¤×8£#&ìï[þóĤ<¨«þÝ2£Š5è©cTôBH|øù°`³IÑ4[£©íû…VðåwK¿ã«Ï-C:+ûª†>œª)äÊÀ¿ú¹½[”û§B` ŠÄåj{Ë$·—D„M•Á¬5¹¿u´YÆPÎ ¡o.f¥B‰ãÄ­>³‘•_Ã+8˜Tšb°¤ìÀÏ”¿%ÌYoÚê°ÑÔŠˆ‹X±œsOâõ•dê¤%pІñº-²£*]W%ÜÂnî<ú;› ú&ÕÆ`£Eåøæfɯ1^BׯéB9']ñœyU¹æÿçj½Ÿ§ª4U{’í@/ÝÖónñ}èh}<í_ãe|¯úw,1~L¬ªXVcÄÂíäS ÖßO†~Ožsôã’¥¶ÃÒâðÏÌ‘ø5Å cË5ßtUß# V³Èkº<ÝÅBwèEÕEöÕ^–WõžtºÊà ª¼pÒ}7;=èbß3 µo•xž´ìS˜‹%ííwdìØ¶´q¾"TÞ‹•E½Iu€¾‚¡i™¼ióø7òMLÓeÖ “Eoùà=è2ø¯ Á\¹fÉò…G­Wýv“Êæ^ †®<³Ë;iÁôp_t•¼ƒ?$/î0š»¶æÛâÉn)šÃ5Üùã‹ÓI˜³ÀðÉuuPoY¢é÷Œ qL»N‚ñ¢û¢'QfD‰ð˜ Bų»Ë¾ ç ã¼|ÄØ  ÉMB©²ìñÂ^m^­ú) ÖË]@%‡¥ ¼ÐžYýÅ |ÄëdGøXœìz\J¬=m_È©gä<§ƒØ‹×Ÿ±Gv»æ–vˊÚÀ‡¹ÍSx{ùFy,"ò9Sx~¸ ö^e6äG俤_;Ôȱ025c tÔdRç§Gר5@KòÊCîÎXÕËÅgZòÃ<Ì1ÿ³´"Bksmu@•|Ò†5”¥°ÀÄ€xr¡‰Rm„ü®Opþ1‡hÍ8}V\ ¹ #³-íÄ+ù®5“n8ïSeêLÌD|˜¸ÚÙ;!À©©t‹_:{oænÇ_‚´Á7Q#C;y—f4ô"ØÜWrr?,ÞªøçT¶LöଟùOߤq» ¦mo¿Wç£T ˜ð-ÁY\v¡¨ò«,dìjˆä÷ªŸH^K³vÃviGËeÌæI-0hbŠ4&UŒšÂMmQÞ fÈ17Îoˆ‰këy¹=øð೪@§gV XR iZ`#ù樎„ ÿÇ+'L7þ¥0ðr”¦¼$¶€¸xfÁý†ÒÕfEnÿŒ©.÷Q‚N7¸á!v“zðHPäs:ÊGŸ¼}H/Ú´e¡U÷ÛçÁvØ3¯Ç焎 Ó~qç†!­'~KA¡ïn’‚eTÛÁÖp!ø‘õ´À¸n¤°A÷ñ+ˇÇzëüBnÝÏ9ÎR3òZ—V¯+`aJk½ï*e.=BGú!–EaÅøæh0ý  Ïñ\Ì Ÿ X¤0²©W;ÿ´.swgŠ£bRtr¾sAq¢Žv`ôpœØ¾“î¥ñ#Ü.·ìńР-a}wõ1Kõg× zÿ¤Vù/ÝE‘‘joªfsulÔyÃMÂåÜ‚®¿rOî¶qÂÿ—µ‚}0O†ý”"¼D‡H‡ü8éþøŸþ¸Èõå‘ ¾²ù<,®ÜW7šêLl …îATígý2ÅD[¡v×pÁi@–÷—º_• GË:]àh[V,mXX!×$À>ùØ”QÌb}õ6¬5bGØ DÀÍôö òÆ›|MpJýç28CôÓp€ê ³4¨Š“wn0`ªrGþKQž„únÝj(ü(ëÏ z‚ð’&ܪvPÊàÎ’N'ɦ$}Œ]€ÑpþJé|ð °Vä,5´±£»aw_ÈÆŒI7„ñc:ë~¶Që;§ €ÁÌ3ÐEÂoz¯¡}©v6¨X”v~¢}£ð,H‡aÔ\ÓÉoŸ³5eÞó°Aš®ÿ¼{Õ÷ÈÈó‘ÀͰj…OâUA*2`½Äeï¬Î‡TƃÈ0t߯°:{K Éo)ãËYy‡¡šµÌûø[‰ HhåÏA¹ÀÛ˘‡BÊ ×%·u¨*WŠù2«·µ>Ø·8É8]Î9Üá_AØÚA3¸ïsK¸@è©‚œÔÇ<Ë­aúˆÕ3kÜ?T¬; ÿ¶Uá?Ⱦ܂+çúµÝÈ®ö¹D!2£>˜7ËŒºrì }ïafm“æ—ªî¥/qŽ´B¹1‡»-ÛU›Þ_ x´˜4=3j‹’Û¿Õ¢™t+ Îyó‹ÁÖí=ˆïXlÒ±µƒjß­Ú«òD;çùÂétÁ,‰Óø¡¸úŸe :³Eý™¨æ¬Ì ^ÜNáÉDþ®Ýçøö<45 _QóæÕY70ä&§ qéB;¹&Ã.;—Ñ+¦5×m º Rû Q@XƒMOòãØj­ÇWtɧe¸¥´ÚŠ"þô÷ÃŽ¦ÆÉ)‡ñÊ »³æo¢•Ò2¹>A´bIµ'°g7æ”ñ(•y¿r›ÍÒ¤ù®kÓÀ÷›Oµ’åWU,J‘€° Ç^ÏÖ<©·TTßòÞãe(à˜¾ÖÝXÉÉ”¥‹‘£M°Á@¶=» E÷ðá}ÅÝkG@jûÁZ|×VDzÃáâ¤_+q rW=›«L†¨b ù[ÅUþú`c£íŠ›§Ämð¤O,ÿ9}}ýïƒ+âãA: hzÝ¢))•åt/" ô{u çEoþ)´aöà™Ÿ òÁæ•jàœG•Á5qæ–aAßpÀÞé\:õ÷œ…-£ ðr °P'/^qbèP”îX3¾3…AñC\¿-ob9êȃ¿luÆ÷ºhÏ»­‘ý²û^]%õïGC³–{8Cú¦Bé»AŠ´n¬ ý\¡í¶DhGè’Ñ쵇ԛ¢’ÚòrüME®• Ô…”8þHóP³‚0.CH7c©Nõ!"Ú®CÏUððïKõ ÚU+:FÁOS'Ëòâ-sÞóöÍWÓº6§¨ñ:ôlØ2ÅcÞhjiÛGݤû½ˆc…òдª5ªÀ3L£péZßÍ»ÞÀÀ\^»¸Õ…É+ûÁñRpÂÍH™Ý.ôólŒiMŒXåÝ#gx‚ù‹žáÍ$²ò!†(cPh,Þœàfz%£*ËÐôoŒª=ݘ§%ÏÙúõÅ-eÐÀ-}j±²ì]Ha(¿›Ípƒ·¾o5½Eù¡nO¬`éâý˜‰ ׿ ¡u.19þ$ß×î Kì¤&0 wAòK|qJ›çùÆÎ Ž’žÜÝ Fω¼+%Nž­ãÃR—Py0óõ°¡…ôüõj?Ýe•Öƒ/s  xk° ·)Ql*F NuëAë.=ÀŽ‚ÀÔi¼ ܛ准b Õ§ÙµŽb°7–´B;2×Ì]KÑßÏt¹c1˜zWÆ£[ ,<äø²MŒ¯$£šÞqy¶†Jqß$H)»eâºÍ}¾„G)°iHºŸOΔÈiÅ¥ß 9nJ‰ð«Äœ!qJé:¶7üñ»º…¢­²'Þ“H¸›^ÝmO8Œ»Èš¾Ì9J€ùÒ|JWb1¦Î°#¸;é¢=u^àà—ÏREo•?1ð›—"F5üTºâð/IM¥|„¶¶ˆù.È]KÇyÅÎùØ´³ƒ Ñ”¼œÖLç(¤8¿c±K_Üì™–¿.€î¨ÌϤoQrôúð‘aI„˸7þ¥èà]Ïûa³±RÆY/·ÄƒY†·äÓÃVúÙ‡°´]€¤J²²Bœ®>œr¾LqìˆÛ×÷›× =p¥»mmn­cØø©‰Ý›ªÛÛ^rqð&D@Év9™\8©¼á-ÊAtØ[#²ºyMItmNûŠ®†l&Dµðe/ß'[õ±£úçÖ0¤çåLßau⢸EØé×kšrñÇ»ȈÚAÚMñ£—`BÚÚÅ¥%Ž€ã}Ñ>ÕVɸšŽ ILè ÿì†pä›Ô‘Ñ¡‚òplˆì Í$ÞeÅ[ aó-¡ï&~*¿rŸ@ZÈŠÃ®Ö ^›Ÿ•ç˜ÈàÞ6ëÑ›#¢"†õšäWa›¾k?¸i‰4Ýbñ¦?é¥ÁU#ðµ3·Ë Ì}¸IßÞœØÕ·kÙÁ.Á´¿P#cx-âÑ3Éú%¨ë°ÒMíî`]?ãgÂõ‹»–0aÐqá(FŽ `ì¯mͰÀ)VÕKí¢~½)ãä.¼¥‘>'ØÒ²¨Ç ¤ £ ž:ª ‡ £¸6pÚ’c+•€üŸrZä86ë)*Û`íÑó”{_¿ômÍŽ#ùmÖ)%75…‘d”×òé°ª›'¡gåßýòk·“]6ÈÛXé’­mYžëäÊ󛘡œ ¸S‘ɹr´›Î²#¤  íßr,Ÿ‚µYf¦‡ö{Q.5q-ìRưÍy"ý¤²Ì;&|­þèü†µO󩮆J þVŠ2±Ñ¸ã|H‚lœ‡ŽßÓSÙ²‚¢M»˜û˜¼–1d¶!›s!öJþcKš@Ù•QÙ‹‰·+çèäÂO±h|Iв"=*D`ʇ-ª|a!>kãüø;Nƒ-Âoáõ2Kà6tÙ÷Ò¡é{,!ù ‰$•8ëÒ‚ä¹ÏÂSÖemuF~¡ºêáûdrì)ãH×:â×<ðo:e—‡÷Ä$¥¥9¦áUЍ–âÖœ_²e-Nt%ÊK …BZJ%öÊ‹×1¤OÙ€¥r~ÿ’ª].ÚMÄÑáfì/óC7ï¤9q~3Ž$‹ †«‡i môÊ¿Ø÷e¯±½õÍÀ×ÊX$2– ÆüR¡u0Áô`Ñw,æþìß³ivÜ“(uNmêNÛTi 4‘H›6¶lÞçPÌéWߌ^D9ûv{ؾÐH‘RÁ¡ Jݱ ¹Ó³ôßikäœ ol]«)ÈúSÿ[C[$DÌò’(§ntñ áÍ aA´‘%#ß—pX|kð-k$ÿéÉJO¾ÁÔFä<+LJì8ÿI]r¨N>Wœ©€Ú_Ò¯4Nf½;ø,¸}–JŒð¯5æERý3ˆ˜Ž)}PdÕã·Ù_HµÞžŸZí[Y̨ޜœ÷¢Ç!»èüúžÔ®#‹K34ÂêkùIW—ïÉ\K4„¿“ñ(%%X®¾»øØ¦MãrÝ€p° Šžî²R#WuâRHjÔ¦ ̾5 ƒ]2jßU‡XxOãÉõ–r Wz&Ù[w#\´º£Pn´KŽ|ï6H[ô_Ë—õƒ…1Åç®_÷¥YUÇÿížÖ$Ükã¯zx0A™¹³½Y'œh%ñ6¾1Ô„K9´ ÕIÎòo—ú>Z”Õ ©ð¤Â[/…iº$0¤ò3]vpqÖŽd©QGKR¯öû€ÌJ¼-ú+£ò²3_ɳË¢’/ó'tέ†ß++—D÷(jÍa⹄gþ4Ñ’ww* ¸¿"‡ÖijñÀ oD„ß¶êüllK^gCïM÷ð @¨4 øz$|{)yÒÿÔ™/žá{µÊÐu­ vlòú¾nô fØF*ÑG‰Wy‚î•¢¤~ZüžS }3Ó_(´ØE2ÍËþhé6òe2/Àë1\ECÈõ²`-èã߬ºï½¬—>•·]žÝ*31k·°ª9µÖæà53ìHpb%þ‘M«Tžyìä­ñ ÿ²…a(ÔE£:¼5*Do%ëvý,rÕ“w¤Û)6P<_½–EÍ®ô®`ëѽÉ>?0Y#“­Ÿ Íîöc;JçÀ_W|á6¥€èæ²vw<“¢q„$p”Û«›èï$0Î!z!£DâÐJ ™?VGŽ&k¡Þ0[îê DZáQ‚4[‘ñ²ÕCЏ[L—w‘tØ6¸À#ªöpAê1AÕ¯* Art!Y™-S*3^ÌOâyu:»‡­3GÔ;áîЫi)J …ÏÄs,ã isZ;’|˜7óã°(^4 V}7Ùge—‡SãxCû|OÑIÌ@6ï·!ìñÛ߃$·Ë[Êx'„r ÒÍÅ/ÄíÌó<¿»Ù'ôÒõ*ïÊÚúÕw¯¥ÌBõðfJòA­KG+ëXq-™[Åm½«bÌZjà5qÔðÝÙ†]óííÜy$8äa&ÕU÷ìÆ3«©PMZ[ýë0ÇÞ„÷"™4Àx¿ÝÃÓK‹ü°±–vÌÄä?)8€+6=àæøˆí‹E¼Õ~.åáï‘ÉOF1«-Ýn¨Ã›É™¨¥i{–ZG›˜%Ûãì‰ú2„AÈ‘„F†y¥Ò¼áëZ{²õÈÞGÛö¥±ÞÁw“hé7¤›Ê eǯm|*õ/ÛëÅ|À‚å®ÿÎMÜký¤òðt‘ñÄÕ¾oWbSVúÛÜRp½x“•w2=³fYÜu¶ÒBqÔùæ¦Y 广ŽnšØ;”H100|½Ò9á%Z’†ÔØÁW8Ó%lÚQºM.Þ* $?Ó¬‡cl:u¨û<î~z7à¿$ ÷ PÜÓ³³ nBŒ÷ ÎÀcÚvîÉA/ÅbñM>Uö=73>JìjX¯„„›hÌ5^2‡9Ì»î 2–Sq:ÂìÐÂNÙê÷'3 Oê¥ÙüÞ»KŸ̪!Šqdø\,×*qNÚ'ð8ƒÃ•…D3ZåÓ„[Qá¯ú„ü?AÏiv§8°èÉ;a‹DvçëâÉ˪ÒòŸèU«&4Fí&“‘÷„=†¤vžæ‘ø=êä=(ZçŒÍi4 'Æ7=Ù¤ÌeD*àÝ Þ@¬ÚAX©|ð$‰Yäþ›2ÂÔL²¤Ê©9f‰‚瘑_0·*x¼‡y_˜›¥Cï!ð‡½†ÈC­$éì­¶êÍH‰\¾«Ãm(ªÅ©Ba™~8I×Ò³èBD…ê?º&ºt0Ôí®#1k ·ÏÙ+„»¹ã.ý¼JØ 5®wVñï°ÆýÉ/ñ¯h -&ÆŒ+¾~¨ ÖŒ“ƒµ⣲›ö³ù3‹Z¦”h’Hx”ÇMó±ïv³1¦*ZLp¡DuVÅðüÇV”çcŒu{A‹on¥-Ì3"Š {õϺôÙM”n^Ê£šîÿ˜áSqžêí}aÑØôàˆªNæ c2^ß ÷°ÁÔ:: Ag-Ú£i`BK1Ãd¾$§—ZX2Šÿ&ÊÊ%î7LÒ&ߎꇳ¿½MÒ¥f,JÄóxZb2¿ç–•¦Å;\ôﹺ9J䨸‘ó2Q&p8›5 ©¿€}QZáõÓgÅÿâÚ]])íËuë` %ô³ohÝæËÈM9V³Ö¼"wó8<ʾõ#ú(ÄMÝ[ýÎÈÀð[÷;iôé_ò ñ~ÛÚØqàÚ¤™—¾Ü{URñŒš;^GHéJSõ™)!î6¨Dˆ ·Ó\ìòÞ죖Ìén6»{SÁÞ²|rÂÔë«âÝ=‚ëe(©äRw‚ôn[Ô ß Üy˜·}_cãÖbQ1ôU‹&<Ƈ(Æë±­ ×mvµÑ£ß£Ø*Ã:< 4ë»næmÌy¡KPçK âIN¡hnÙc"ùÐbKÈl½ƒw7¼Õ#+ÁÌk!ÖÁY6,KøÎµ[.ßn©›Ù¬·Ñ{ñaÏÇ^ã{¢§ŠÁ±šŽñ­YÇšéhb Iìþ`ù|ôÊ-€|¶#èa–E娕ŒÇ3#öÛ7ÿÄÿÎuîµZ‚¹ÆþXTË£j˜B Œ*N]»Qkú-M­NAñ͈_QÀÛôi£¶:ÎÓØw„ÛszÕ6݈^½cûÍña<¦â£ýäí·×8u¥ ©½ÿÀU˜ÅGS2‰Wÿðd2Yû@fÑO:´v¢­dÿ˜oo5´= æ»ãVÙi½Úé0bý*ý8)IÉçZÃâyL]ÏÚmUW5žgê!§æ°ñ4Âh{žÒ•j×¢£Mï’snƒñß„èJò¾ 6Žü¤lû:)h ”_™ŽØ d^.;ü‘‘µôífB°‰áK° •úV:åúò~Oxj50vŒÅûYñß&>”¥ñeŠˆ †®š˜BÔ^žçÃWä`š@‚©Œ¢Ì ž~†rÁ¿€Új_zKYð%öŽ·N¡à6Û×–[¢~ëÅòòõ£Z4œ‰Í3ˆ ú¿»%X†§‡^žÝµME¸'%(‚óÔ/N„TÝ»ZýÄx«fûS¬"» è;ìgÿrÌ ß5ÕwdbKv×±ÚØ¤B ¦Í¬¢ðÉ ÏûÃë¤|àZzP,¨³ò VëU†µk_ã³&V4Ÿ )U­ûò™ËH"èLŠ“V> F%òO9bcõ{ô 7—dô ©ì#?hîYíÓìD¡f™Ò0’6Böµ³kŒÿ4˜ L‚ÄÝîÍþ¶mo,™CŠÊ2(ÀšÎF}Vhß©Ud&£Vüªü*—{µS+Œwo2eC4«œÿœVúw'JßZÝ‘Š{H»?à]æ¼a”6u`®ˆ"ooaMèžREü£†ŸE~†I#ÖÔÃ6'®R5ëúÍìYå~–²q1ü¬hÜé"}Ý^¦VÒ†(N˜þ–Ú¿æ[¡[e6fY²~òÛF•Nà %ƒŒ×æn“¬]‡°ãzô';¥ðåû?ÐË"c9|K;ŵd[ü9¾ü7‰„)tL¿ŽÉâC™•§§i^ ÅðÉÀ©»r¾°÷šÓ…¯ˆ“…énï¨R>’Œo„|â—³öP1VÊnmR©_x3³©¿ŠSУrì¡A;¸ÔÐá4ˆÚ{=ÎH ÉWbøÀ`µÚKÞ¶?Õ?—B–’€‡¤Ä™YVV D2Iðiæ}H)ÛaUM‚@ôDª9­ 4X²”iç‡O£è´GÙp€ ~©°Ž4cñÙ÷Z›µb¹cDÎPÅ"zê)'gE6ÒdÙ݇êYÈòv†cÐ@m÷…Î(£‘|2Åô‚¥hŠƒÀ‰ä¡ú={›LÎD’§½@í‚Þ‰ ÚDȪd’­—bsG¥½0m.Ù¯±ù&étë |¯AÃƈ‘ KZzLÆ œR6`aÅ;Ç8Mœ¢æÆreRÛãÊaÏÑ cÔˆ¤[ÇG2eèžéh¤À?2wÅúeÓ WsØÊ†¢¦“E´tUÌÇî;œV”eŽ„².ôûž”sw…”L°ô€\’:Gfì]_µØ&’|º\ ïçã ÂE|wä%âÀÓé0ûÖDmñ]ÍùPŸÇžb¥Óì^­duSÓìiŽN½¦˜B›AÚWËP/SE»«iF“iRÌ*ëó˜Ð@Æe·?xÑq Q{ë?1ʹš¤? ÂS™ªã.±Ñ“œÇøI¹«¥EåÝ<|ªÎIuiícª5s@aKM"Þe[H0y_éwý¹Ü]>ý¦\³²€à8# Ãp†˜C@•ðLNcoÞ;¦°ÏßQgSÌŽ©!LAûÒgBÜß§!„¼ÿÕF׌lDHb®~z8{Uy¹äŒ |˜teTÐ$~Ý¡í"œå+âÈË”åü¶Ú­[å m†yœú5™CWgâˆ#z À*Óx¸°>J¿IÆ7Ó'«Ã¸L.Qj©©Çÿ%¡SöA,¯×ìð½Îa ÚAyÑðêÊ™p›ÊË¥±?‘®±‰Nÿ$‚0‰êÒW¸~6çžoj爋öšºX¿iËï2°:|,7ø¯šÐL·ú³  [„#Êç´HW|¹Ã;ÎÑ÷Œí—¬ÒÁlx<µ&øÑ­zQÚó•Z]þTŸOš~HJ®9´4UèÄg•¾ : R[“úG®uÍ9~:O\›ºrLOu°Åֱйã M4 ÈtªŒ—󃈆2yý¦94fCâv«Øôêûz­ìÿ~+œMþ Gš¯°ÅóŽágU0,2SEk,<”ûx•Úƒó|áÆ<b5–ûžqì&åá;[E§P9“iá½™ òãoÜËâºêÙ½ %'â9¬ŽDÄã&IWÖÕ¯ð…&ð8 Y´¸·.*$l™2Yê`s3²ì,ft>ÏæÕ“ÿ,.U3°B8¼¤½bsü6\eW´á2³ãC.Ä1l¾JÝ—ý;<^wWÏ蘲+'åøÜVIw$‚"¢Šá⨱0!ÄÓ(ÁUd>ð¶Â5äI¡vg| ’ ŸòE? aNÀIFH.{îXø<Í{h‹VÑ/úi WÕÜ.68€!s‡·¿N?û“¯xM‘+öæP<¿"Íbâi°û6%Í6R{"eû É O›,…=uùv¤õì§$.ÍÎÍÖ²9㾃0Ò¿Š„WoG[öŠ I%NZ¿ü,¥Ó-²Uüz+`:Z¾çbJ÷•ØÊÀÎYš9\÷öBHÿ(Û“!ñ|˻ʷøj fŠe=¶®ÇbÁ¯ís“Q‚œIácÌõ3º‡‹ç“ÂÃ"°!š}üj(¢;©0\ƒlWÞ;Ÿ©#›N¹*ß1…àÂüG^Ðìžo†‰ÒZ%Ë¥EÝ•›B4ØÞ¤ì¯9ð9—ÓÇ¢™3Ó3<Ñ%²>å”ñšl¼ZçÍ Ñ1Úu9¿ñ6NeÈ-'ùÕ†9Æ4ybp Ò^68]žKÉç79Óï4nž|©Z˜z'p¬Xf¹HkÁœxûiÂh|ý@ãíáˆe¶ýŸ1ÛÛK öYª@N„áÞ%`>*nûR8OãÈ3ÖOݵ ̱EØ”5dÝ…iŒk‘aý‰§ª³Y— Lêé ⦻³Ž„S7ÇŤIž)Ÿ)‡ jæ/Û4²™šMg_›†àF¹~±!1œ_?€`’²c­O+Ò ÉUfù]#{-eöê£ÝT`á6ÔÉw–ù¤“ yCÑ;ªÄG>¦“d ÏnHÍÞ©&¸×ÏÐ3inÖílÁŠ #” Ý#)iP™S=šz×6“º€¿õ$^×^ò”ßS7gñ¾i%í¨ðUF6q»ßNâA˜ÿô?ëÖ…@$©^\KÀ¦ðìmD"¡ªq`›Dλ7'†¾¹ÿïDž†P$Ñ«ÌRß]¡¾:ʉký}¸Ë­ÐšJ¾[ÙìseäA®‚ Q„R¤²°&1¤ù´œ„R j9.'©.I^QBÚ[ÞáTIM’RvaB4X«Þ•ç ®u½?êÚ9”ÅHº=kì„ö¯Ÿ4¥cå_Ð8çÝ%Ê•/YPÊ4³éDDæóÛmoL韞X8Ü“€0¿þ‘*€Â÷´Dúµw9[Çhº—&°áSOÖ4GÄÑ* Ejàöt€¥¸?ºÁ±¯ó8v­âœ ›Âv³ºd ¾» V¾ä…d”2"^” þ+«’Pg — 4À†—ðjúl‰©¯KÙ “â!Ý(œï¦MÈ{ZM¤)TŸÞò©xZè: øÿ!ûnž×–Pu1¸Íœ~ÿ¾Ö÷,oOJ-NFgU®œfÎ 2LVráݪëç˜Fê•v¶þY™Ép»ä¾ÞN‚}ÙÝ H»ásàn&r|Pu,À»œŒb˜Hð«LœTžy†È-ZŽVQãªKÎà92¤×›ð×öœ"~s\$ ³4WLØô êÀ[Ô5ò„µýÜñXï9âº-¦7ÃP:$UŸbO1t<ÈG-$éPÅz±n½tŠÅFJœÌÚ‰YOqtrb]|ذÁÙU¼ÝòÅz?Ån'l»“ HýJ¨þ6<=s_Ð|n:"‰xrrmí!7ÙŠŽ¿`§<i\94S†WÇ*Ñ‹ö$mÑWü=¦Ü+%úý™Â¨=‚i× åÔ62êi̲iq¡ndYK„ÞO¨µõ„Kãc:ÕéO j|ƒéöápT/<²zJ ܈W,EN D†«Å(µÀž²ÊØýØåØcïÝ6‘îΡí~”‰L‘„>+ù‘wOp«|V¶&¨&&y%FIT>µ‚B¸+¯é|òtÍbÅj]ãâÚ¨ÿ5bôqýÝêóEΙlWtŪý”§üôW ³Ý¸¦å>àÓKß{3µc 4YA"+Ÿ¦B´^Ò Vç)¥oÙšþ§9`?òÂEËrtc(¬×„ºoàãmå8lÿ?ÞÝ]Lä†fo¹ï²Qm¡ìþiçaóÂ#[ݺ}"`(ù§¼m§ŒMä„`Ä/íX˜ á“Lf¶:bç½@ü¯ó!þaBÃD‡B(0£ÑÂ;à=— Ï¡´ÎnK_NäÇ‹½õ’ÈFßHøOT{r…›n<2#Š dsô@J kßÔ¯dçÏÙ°ZÒ€\=U¦¥Ý%®ôS—¨·ß9^á³¢¥´¡b @”ù¶€ûéòMå¬-ñVÓ¥¡€Jm2‰2è ÖG4ªœõà÷K®C„\‰4¼v¦÷¡Q¹eŸâ•XÊ×Ó ¦êWïàÊ̓Â9Aˆ(^¦Ç4ÇS¿Öv¾Â4"Šã+´ÆCzM÷ñd ©ÏËÁS'IT³a!×pÇ›Qðp ê’,NX«ç=;)ò7Íàô^|ÿóåBô¯íN°D¢*UÑ)ÇfúO)ì§!Òþ%Gç. ÐÁt¬·†–KÇâ¹?i­£Z?¾]ï%Kµs–è˜j¢—ú­Ë¿i¹ ¯® Åuf'…i܈š~<¿Sô›vFùúÀ2Í’ÖB„.±»ÖÅPŒr/}ùÖÚ)¹¢âc¼2—=/9þÄÛ†:°c&3¶µÌüègòÂäáPý Œcõ=ÚÎkÛÍГ–g¶Í`·#òÔ«v.Œ¤½Jû¶Ò9ß7ÞÓ˜\¥Yñö>¾Ì–À™ô³»WÖ¬¥˜”)¥TC!“ús¤ê¾WþìÝçàÂÒ»l‹¯Ü}É¢+¥‡MÕÊ#B—¹KT‘—¥ÃÇDÖê›fv+éÁøüöh«‹R–ýzãDRäÐ̰Ӈcó-"-,;µf‹ÂRø+FT¨åY¡«|1ÔµFæE„âÊ2ÑîF·[Ž”ËŠr1 ›‚žüp˜1×wáßþôÈ¢“e}LÉ¥ªJ†}°8'„}Ÿu Ë1)4ç‚·¨BK‡§ÝA¿Å< ôxý•¶¥«ssñç˦€Äó n¿}v%Õ|²P,k0…F<2áûhvyïï'ðs‚ê³³Oš¹‰õ·%Of~¬éÿ!·B,ê_z“¦és²rfeð½•'’ÜÐPªÒ‡ÂÐC𑯤øcbƒöÙ™\ÊņôXëŠc¦åH¸]”ËN<%7€Ó\H{[«¦§ t‚„ž˜£O-j…Ï!à—·DS73°á–ûNÊJódaYÿøšzÌÆ¹ ¿z:Î1fºrÃe+éÖÑ R)”™1¼w;^™ Úo]£Mv0…Èf!e÷èüò’æÕ*hNE_àzí8D¾úUe2‡ýÌn;¹Æf{­p~fÅ…>)ª†+¸®§îÉí`´{ÓžzâœÿˆZŽq0®C¹ ½n_xçh^û¼°9D`þ´{KH£{e´H5›femFJPtP¥wu“/—Ñ*Pµ  3ÿ³ûÙiJ7Gw׬ðÆÊ!‘ò•x·3 ¤¼å>7Nb§#_+"éÍA_ð$cxüƒB:ƒµxëbñíùöí\±þ¼ ÐÈ'Öœ×72©ýŠ`&0`m(8XB♲|±¢DB|:ÏBBH©hÜÊÎblÈ¥QÛYͧo‰‹ ƒ]^JbåoÝ$m?ï[ Úz„Ü™Œjˆ‰²+¿¿[ù •°vë:ŽÜˆºvô‡šQ^BœýÑçt(CÑ«²Ö½Ô¶ÊúdB´gÅLßBžß6´Ñb¢ûpJ}òpÿ†{ÿ³`Kh²ã“¸f <þ›ÖŽÝ½E×32r*åbor$ïµÂô ¶•ìŸqòÃN¿p¦~Áð«wÂßcSŠ p 7C8]´>.‡:;:_~Ò$ ÓŸO˜âðQ,ñø7=ÐÃ6íÆ×§]úLw;®ê±=èNóÆçGøëõx‰y©·ÿGÖŸw×´…Eˆ…ÉAªÉèÒiuTëî¦k‘´¹Ø“Ò•«nº¢q‰-£è¤—ˆøÀÈHoç L¯ÝÚ¿HÖäD}UÆ@6½P`?£«ÐìËðY]¬ßH==¦=‚99r¹ºdmè4k“‡ã4­ÆßÜãÆï~'Ä6„ÉŬþ•®©±k7œ9‡Íd+ù¨Ñ*†,¡€k©ÜT=¢'ÃlƒhIÔ«Í^ÎÜO}”Æ3­íÇÿ‘ó+Δ»P©?ªe®•ra˜MiGW;êoãiÓæ_ (-„ø7óÚé/Éa›‡Ã&ä¯+—´GqIÁYè¦FÞðÞLÓµ¢W½Q˜·¸ŽßÚã|/»+ý‰†/ÉÐZ #÷K%Ø›EÚZ³`êÖ0’j`×§Ê.‹Ð´{±Py`=ÿ΢ÎÕ÷¨ggG8™2jPaIà@¡p¥Æâ ý¿XH!ç~i¿Þysf¡/Ó‚A'æ€H>Ý×ñOü´wM»¾Z'Wâ&‡øÿ”ïAW½8Ê©áá6rWÐìCFâyÁƒâ=éûiºïK_­fL~X¥¬Ñ9±õ,áõªRON'žH鿨ö"¬0‚¶Ï—èriÁ»Ëó¿„ rx㞢Î^Z LŒTßJÀ7üÿ¶­yp ¼Ÿ?žùëDí5¶'¤ášƒªù£€7J]¤÷˜ê¾±œgIM¤ö½vYr¾>rÖN‹þa/Äx‘ O ¦ðŽuQøëPpõ>?%~$Ñ-£!µv¨2ý÷˜7’ &¼…õì”~©½Â(8Ç­çT±ùAê 0±w „Dêžrò­L¨ç(Öצ"de.f|Xqñ »¡C‘’É› r¿À­ýAq5Ó-qLe\½Be!G9f9 ¡©b/+½eƒº)jPÝᦺã‚𪧻+<Ä8Êu{ óo£«¼ƒ¬ò3}LSr¥ñrF¡Jc[‰› ´ÙªÔÊ­(Ñ£RÚ"]z²Šûþ3W5;Ÿ@x¾&1ÞO|}Žj^S2f¼Lƒ¿wÆò[q«Ù ¸}ùø‚aÐtu´Æìð™2DÏABVóŸÂ?þCü¬õ˜tÁCÌV•œ ¼‹¶÷ŽÌ³oû÷»x‹ÿ¿D¼¿Îës¨ã.3ðC’G¥3‰œ8ïãt‹íÇ1ZŽŠ¥ ‹ƒ¸²ï©®]Ä}ƒ%3£v‰êWP‚¡~ìÕh×%» c2ƨ‡[Àrk2Ø_”݉åWA‡iX5ü쎚}g³dõw·ÍgÅD0#ß©ª ¦Š··àÅŽ©Åd’“KÕ×zÅ…Ê×°_Ö}Ù‘œ˜µ¸OSç·Â@æyÈ‘ôO˜=éO—OKÁ›êbptÝ· äã@¹í•šý¹¡ä’üø[#;riM U¿¤Vª´>’£ A•nXP1KÎÚº æ(NYÑe!‡¿¨¸¹X²Ø£²¿¿„,Ú“dc­Qø;³ïŠŸ©d§k•Y”×AdãÀâEûw /3…ü_`˜Õ¦Hc‚i¼³s],!^qý½ŒŽ Á(Ú9Ùžlú³'µ9ô5­’áÔèл'æhÙŠkãÑ¢ý®O ’ˆnéJ,¹©kzðƒð•úæ9ÜOh6¥zãkA‚Z™ÒÛìÐ<ÕŸˆ{€’¿wIÿZŒ˜ÅÞÜqs)àp,v)zÐdgŽt¨í•ä|Vª´a«4õëÎÿ×!ñ" ÔŽò<ŸŸÑÞØÎ4JáÃúéêÀ6ç5N¸\6¢&…›¯ /|Ì¿ÿK™ÕAÍq^9êÌéÍ_KÇ}p¥žçêoA%¨gu6K/霂ÂÿƒÝFüSÜ È·6 ¡-Q5wÜÍî7':¹* gÙ>i¤OYgVc²” /q &Rƒ¨'Žƒ€fÒÞ–zú0Á]P’Èñ¨-‰Ìo›\öõèüœVHoÐÈS¢æ;… ‘ ¿àq*¥3ªÝf»G+þGl+#3|Owoµ2rSgc)‘'ÊçÖ8—M€ƒ<cð|IÊDµ‡{Î6h‹]$kwØÛø9S„ºý]hà¹Ìn„ØW›fîë>ñÄíѾ?iž¿K&]éIË1¡ÎÊ.<ƒ[Sºbˆ¨Q™£°3Øw¢Cí¦È„”yyœ:ˆ7„Êž:\R{"¡ÛöÐÅ®¸–ÐGQa¡OÔb áÝ•k®)œ>ÃÆÅT9y¯"nüsÓÈ©ð4Ò+™g0ˆ*'Wùñ×øÄ{ Ì=«d'€94qøiíÁWÖãÑU#³CËÊCÄO}³c 6Rf䚺éä i+QD\Óyºíô_µ`ÒýÂïd {Èe¨¡>âæIA³&‘O r¬ C(8d_ÉÎ<êÛÃkÄC×Ê hâ7ÍÙÛq/à!¾†›¼B±Ç¨Y­Ë$†¶%• 3Eø¤®÷cjˆº>ÍÛºØmUtdç–Z‰®}Ð 'r±½ æé’/Wڄ盦ý“´\‘>xó]Å 5êÍ4Ðc]˜]å©`…¶+‘:ÆsSžû¿ÆwÀP¡t¥IÛBo·ó †>2äÚ°jçG/­Åuã"ì¨Dƒ—ð lH¹ÿ zƒ’r¬5’œRö/ò ¼µV„¼iˆW¡}%\£õlŸñsðF©rž–V¹ŸÜE\!‡­ÿ¨ Ç %jge”ŸÄm4î•'S—aý@·D({Tìûv¸ç§üP ]Î5ŽJ'] $Æ“ÒJ£uÃ98ª0¿ Žj§ì4±PW)C«§÷¸ãØ{*sø^v°/†kµáF¶^Ƨ I}¹ƒ®8Á<̯”鋎;AÚsrBæéjB’7U÷6ZƒDý™Õê"ðl¤g.‹$){þjB$½P¯k«¯UÝ#­ EÙž!'$[M¦òØ)E¶wbsp«SFB*¡dV¬j´juʃïÃ"þx—ÕìçÒ8à¡‹Ý !ö :LO"@`<ûBP¡ú‹Ü0ž4o@ŽúJAØSýi…¹]‹–›èøŠ‰Z6¢E0×H¶ ìÆ oªúhw=Ÿd‰p,Ûë‰Ï:þ0Þß¹iíO¦ÉK„j'Œ§<¢Þè#¶Àï…0£øÜÒyê#nÝwâÄMö§ 'Ü6Û-9M´4”‰a`ȨZØ|‹¸D„dª=”HáJÃ=XÚ <+ÏîÓ}¡}çd_^¹¼aScq,3þÏoœ9z+xRóöaæLÚrê¯QS¢Ú.:W•­¦åˆä&‚…úV C¨-¶1)UƒeozÄï„ ²®ÌÂy’µçH˜vרášy&ÚPó5PЏbPm1ã:Ù¿Â<ðpĬ›]ØýÒ0ŠÉL‚Ió¤è(B°™ Ô3µc}rSºÔ3"ð~º'KšÛ™ýÊó Ÿ"€º ²²ÉVˆŠr߆BuZ—£Éßn+Yi–K¯„¥]ïsìä9Ÿ®N¶vbÀ¦žKöö$v×Qó½÷ýs”Ÿ–Õ½þI[°÷)2Þ œEYJ×\ñð^€¸hN3ä›\±LKÛ%†ÞjÒô¨B¤ß§ÅQ#,夘åãF݃95´Âù™A•‚EaW€‰ákw̆èN)\‹b:ä§Í³Ú ò˜p<¦ÜûÃ-|«=ûA-O†³æØlÀ™'¿5{[!}þ3ç-ŒÕaŒ³¼&im³IJë…YXвM7#"Þ›63Š NÒ÷=°°ZR³Ì &Úˆ^x_ys$â·µÛ‰~Æ=ïˆöp‚ÿø´ž³m©»fRs¥Ôÿ—xƒ•\¦fÈAG}ÜTò*àª|Ùù$t™xîÖ½—úºÁr§ÇÔñß.=£;¹ªæÓ/…6ý6PTuÿ\š}<×újö·aÈÍ-¦ˆ|äªóyîTG¾5ãùCf$¤Jœ=ÞÊ«ñªå+iÇîtÏüš‚%?.ÒQ·ìç™lËË¢¼ñPÇy‡ÊÜäEQ&¸5ôH”c_h„(?1¹ÚŒ îÚD‘X\/§ ,†yü`´Óà–je©iZ¹ùTzQŸñ"@çe˜¡t?Ô¨BºbT”wZ¦v5à ðŠ7‹4•?6Qážé÷…~ asÖe®Å<[ÉzÙ–­q²tt—ù'?®£kºáÐ>ßehù |Ùu²QÈ>8s夡ÇÅ,w½Œ8«å.;¨á„m;ñˆ(IÕÝKem#ŸI'—†ÀçÝ!™4eÖgñÏÖY m¢ÌïÑTÖ{–/Q.Šɰ ¦,½rîµ¥M Ms‚ã©öÚ #A5¯e*YhõE‚K°:OYf1æŸÚHj&ìŠ÷?´iÙf‰ãcÜ—ôëÜζÉ9?çÿ ÅåX+5”¬*LIPƒ³Þ˜Üœp¶–éfmW(éu"ÇöÞwdæÕnÛãã¿`“ ¬p$„ÒÁ@#Î~5íªÊM¤Ÿ0 ›¢ðøtÅ/¨°h{;QÑ=”öÆUþ¿ BÞÜVõå+P£å^´‰òïa"lÖ9UQ­†i|ëÑïû#¦–c(‡6Fà ҺyQ ¥ÒÈÌäá·Å¨U³+Hr;<ºÅmO‘‚´d¡„¨Û8;Å—wYÅ…mC\¯>c›êª³›ì•é5»BÃúª5JN“›¾'†¬˜oÏ„¬ àUÝ]oÆšc“ñqô|¥b‚ž"ŽI0“¤KÄþîí) »,>=–ÜâäiNHÁx»Á5¼èâ9u} ÷£y$ÐpÿàÔòS-ËÚ¶À9¥!ßbä—´c SѺ-ýXé éÇ@W®O‘ô¸k… â€eVáùQ¸b?ð/4õšg>Ôã]6µuBï<Ì›¶¬ƒÔØ‚hoã`Á“‘Nu„NŠ1þ"…X·¥ºÐ@w„à…@$§Ÿ§‡‰êS|š¦ ÛG ï)gà2Ì6T/徘×pKÿôÖwd{C`Ï®2· Ëo‚Ç: fÕÒUíQ@gþ篊?]ìü0¤Š¤‘ †¬N3 ´¢‚!eQó &ÿhÓû/¿}v?L¼¹¾Ï"¬ë«¸ÿÏ*C ¾ª«¿Él‹ÌòœšzÀ(.ƒ‹”ÖSé7à £uO cMÀ‹D9M›1pF«¦llHKÓ~¯KÑwÒw8”ïÜ')r P&áˆ<¯¯çÂòÙÍtÞ%äÒsYÔ…j9•ÊXO§w$zœ™8ºö¡Ëà”©¾ "ñ,‡ —¢ÎáY:ÀêÜTŒ4w˜œ=Üig˜ÏœªAƒTéSüïq÷‹ ÿbsÚSú%ml…êÑ·üƒ¦9e3ä1!669:ø[Áf¤ãÝ0 u‡4ó«ÚSÿKÍœÓÊ è·nîzЬ2ôêÌâ5iiQçUg"3‹³Ô K4{¿u™¼Ä¹†9]÷!œ«e±Zùqþ™kõ^}“å†9“VcŸ8¶¡9øx]ok8î %º °+ìËôã¶1A¿$á<êz‘óŠèì+=©,-[Óé·P½Š‰îd­.¤ŽÞ+äh¥þ³ÌH´NÎDëEÙTGT©í48 Ã™T–XeãÄÙF.rHý§Ñ-uxü'ÙøníiØmö_ú¯"2ï½'8„çµäÍ,ßAŒiÆíƉ¿ )ÉÓ–Ãì¸äâ…ÿsrs¡P¹¼}Ö–Ý®¾’2~³Pi\Ñi~ [ÚGYmŒ=n°Iú¤Š^*Eœ£%Yø`ñþ0òlg1ŠŒ`„ùÅze1›7ˆùÃ…Š¦Ûl°!þÎvœ˜„C]1šålg%@½µ'Ma ¶mÚ­v€ÚOrªò“œƒÄ³·í¶zŠß7‰¢(—G?lË{åoTéQ¶GxpÃ:I¢GËëhÃ=•ðÈT(,ެù~[³nêDù`^Ue Ž.`–¡º‹:݆mÆÄH7ϰDóËî®yÌ%üd4ùm½rÕóÛ›6³hêÑVWv\'k¾âýñò0cO(° 9hFÙ†ê4§ÅnF ^šÈµYáÃ/×äÂYëOf4í»ŒßßGD§’ÅÈw>‰¾HšD™]ÉŒ…_ ÁNß"£`Ö'm5¿úįw‡"éë#>Þ)ŽuÈ^PäÇ]#¸« bIÆ@dZRSQ")·VkÞoCS‰ð±­ÂZûp')%Ö·_Ö˜ ]‚gþ\¡©‹Ú@²GΔ̵zCôgZ¡ò©î?q”?$Ÿ¨pÿeh” ¸¶a¦Š¤×$iy¦;v½¤ŽÔ„ÒwpôôÂJk¦¾ð.rE ‹.Oà3úyñTïóì$н>xÇ[4w둌ÑH׉”èyŽnm†vá)[$“döÊ&ÇÛ¸]L³1Ø9qMEìÀ–ÃÌ¢ø±8j­;ÿ¿$‚¡®–Ìœø§âq„€Iü]„)I`ìÜôè Õ(„Eo Ôt0Ìå)ö#Ýö0–ªÂ;)Jž1· Âð`÷wåÕ=[Qîé@tïXQa}—Uª'zœß{ˆhå1$Ò—Ç@9ö/äPHýŸ®ó}m;¾â| Ò+PôÇRå2Ù[WžYc-¿˜Àö˜œvÅtWÒ7ظúÝ‚ü²¬¡š;jÝ[Té©^þ“ëúD“šàݰÙ/Óä5¶Ò™äðø@Ô4 öØé­[gxæ]¤~rÀ ô’ëËuÍWðë¨Ql?M"úMV«Ð̦UÊŠÔh'±¸gfÆÊEá Z8£ßù2[ëÈUwhâb¨l.¾†ëðˆûÔ øžlõ@Ò…íÁF¡Fq‡”škÇ¿zÉÖ¬T¼àõé²›ÿëjdäðýå¨4sȶÂk|z­ËÑýù4^´.þ»© ·®_Ê@IR.ºšæŸf‚.¾à,%–`Á³lyëÅ#ÿ“„þ:Þ–r/ìÿi0Gà†×覄Êýv„lÖ™¯|Ö×ZÞwCcF|ü_¹J$í¿¹ <ýçŒ9‹¿:5ô˜üín°‹.Ì©âŒÆ Iet Ä£Äëû§e…ª@ò>¦DÕùvP©n刯ËB- ò¹ãÓÑ J¸ü×%ظžæŒ£Ï½+ÞÊRÆ"ˆ’¹¢¹Üí 0mŒ<’ÏßYÙ§`ûoa?;,š±Ë6µÞuäd¨ØíôïTC©EÓ ÷I¡µP“<"q¬‡~¥ž¦Ä§lYÕžhEOOn$7„þ3¹cªÞ`¶2Z¿ÖÎÂ^J":;l bmÇ´YÕgŸñÄc*³àŸ¾,˜Èz ‡æ«APÆX³ÅÍÄW¢2„º…ý*¤urryµ!ARo(9¡L8ñÂ* ¯ýæú¥NL€§žTL_ƒº@÷s±Wöò!ûn¢øT«ô@˜jµÓv‹ô7æ7Úý¨ÀŠ—§Ó coj~iÄ1î2§ö¦8ÇÆòûW­®Ñ}5Á LiøÈ‹Ýv~Œ¿¯Š™ËœÇÙHøÝñgh“Ûß´¬""’ûâ·ÚÜK'X8*#þ e0ï…PaÖý§%- Šå|ó ò×ÅŽÀië*ÏuÒ1Ü~ ].ƒ/éø“ eÍÐó8ôB³T¥"œhf|;·:ò·PÞ–Öh1ù Ñ+ïSuq),þ4­ðuß.ß:kÀ½LÖ«¯^³çæ{¼Ó Z—°u1×­îî`ënGoæ{ëVï‹zr¾«bwIzE¼ÝÌ‹Œ@ 'VJhˆÖZ"R÷«s‚_A#kzü~L¯s±Eq+”÷v@y{a®ûZŠø'hnf¸G†-Rg1D;gÏ¿"éž–Y-¯{D-®!«ËÜZ‡J†­NÀ¦CÞçí,@WƒÓ!*úŵ V·ú­v£ï“éMÈê¦^>ÚÚÿŠÚ`r÷ؘ9£è(óÈIšþ}0—r¦eÀzf?\¹ÙþÁ]F2º›ØÂžV£;<Û±Dò“gm•D;$W†` ~n•üßWCy¦-³“V ïºÒ÷/©a|øo–0™³ÃÁ´‹EýŠÏ»qùŤÛ Á&CNƯ¹‡VÍf*RÊAS²ZŽ÷Q"s"9–ÄÕ3Ê–Ú†aVAœxKýD{‰ A%6””/£KœŸ ;qát½sCŒW5Òe¼oy´a'ƒ íØ/ Šß…=öDb®’Ìã~/ãõ‹ù*xsßÏÏòˆNž @YØÊI"\ÒznßèHlkÖ:£“>Ô‰ìc\=–µ*Õà´ÛÀH3íu.ß++Ï„îç]Àp‹ºNÁ¾y){NÑôÈðMŸÃúd? fjžùÁà5^f×H >¡¬ ±– NxÜÝEÆq0T®f¼Õ)û´Wu*#§c´ä†% K>Ð3Uð*ÞXl=™ŸÕac:­Œ.âªPу ©…r’|'ŒhMÂTRŸÎÞÔ"ðNí£€³û¿C,ÆÓö‰>Ä܈éTA}ýZ›i¥ˆ+ìD @ËlBÝ¥ƒp3ºœdEÕs2$?ðRÓÇÛ°ÜõËÞÕÇ‚y»ÀõºÕçêÝøm>ÈÑ "<1Kü{¬¿Ìs²s— hPvûÒ„79×:­NÛ§ ©%b %ƒú¾Ö Q¡°ôÔ°I‘o2°O/Wf­ßꯅ.fû“g½Í’°‘‡9ïè_Ó´ô7îÎçPA2€ßÉCùHS&b.ÎiÚA°×ÑC:ë&äY!ÄèÁ_œPX×äy4ñYœã±úÃàû‘õŒ–bèi+÷3Wã _§¯+xOÃÛõ²Öú( ‚ ¸!…ë ¸½Y°F*F­„.2­ŒB…Y$ûM‚áú{-xGëø‚v*ál½KöMòUöêG žP“Y aœýT,ªÿ)GOð›•aB Òñ.ûÒÌÒ©1™æ5³F„‘yÚÆ€‹rMƒj.´97ÿÂñÌ.µ•¸ ¬r@:óÔ8©ÿHít`§¢Z-—¦¼^GÊ~¸Mù™ÀP%̾ŇÂ\Àn?+^Ïè¦â+¢ç9ðsÇkÉ‹¹[ô£Ê@›Tj‰» –Í‹÷—¡Ï©0I³Š§KFÊæ˜¨Q ŒÊ;ä$f=»¦Å5`ð¨„ÁA3no}Ùd÷ï¥ÃO·ý&éTe^ˆÙ¹×Ýñˆ3ÜÙT Á¡¢?ª*ö¬«»â“ÜïÖPÝ,ürµ}ºd :þWŠÂ»6^ÓP¾é9 A²”!ašnÌm­Óì™þ™4H'ý@7&fñ í~*}®Ây.ÿP/Î’y6!‹.É´7ÿ.>›À¬\Lgà×ÛýamtX—cèìýËzôÍ¢Œ»Ì1“K7uš˜-Óã[j*…Uwô-Ý–ŸÍ¹‚IÔÖTsEãïëôõä'¤Rå¤ÌF,>Σ­,ôx =üû¼ÁKQX<8 %ùË—ÊDkÙçtöê÷|mæž)Q™3]”Y1„næGÒ„D*›Üõ[[œñ¹pFzgdg-ÂôQÑ}l+ª'eGa%#6vÄì? 92³çJN÷퓊(<{¨qãíRë€h—­B‰û3“Ç®†ûc_Jçü Ô 0UJÇÒÌ)¾ËúïP&Ž :×&OÎo-CO±üéSA£Qõ”Šf¥m¯ûzötßàó SÒÌRW‹SæãþØœút „™rW*‡Îß+²´ðô×&;înaYx®:û\–ŠJݱF‡>þŠí±¬ôé5 0Uu˺á¼îÆ0ù-lÜ4a`ªúÕáÑ}ðb>çB°¹£—…|H¦³±|·¾ÉYv+y’²v>øzPþ !G-„Ó(r1úV¥ì»’š"ByoXù5ûR«ášÊ· W5oM%ï¾W9,Óc žŒé :VLÖ8}¬›/­öùPí\÷ ^oɰ=‡>]3þ…ôÊ1&žšf6—PíõÌWØ_èç×É[p”ܻ֔¾étôœÜNs¾n™W«“<‹.ÉÙ™±TTÙÌúy3ð(?ÄiÙ·ã{±6í¦NÑ}À]__7\!1¸–[HÞ79ÑØ•Ëc\*±Zw“9Ž-y“ï×킾+‡ƒ–O¨ÚɄܒ +vQ},ª-Îÿ³ ƒo#UkAŠ@¡!s~%m~_t‘„4«à•S5ð! ©£ÉYHa9³^=s“2è]=a[‡óÛÅÐ=†íVyÍÅÚ®|6ÿ\>Mzd°¦“6GA}jû)ÏËm'¼àúº’^ÊŸÙÙüŽ¥) Q/Ku}”ÁzReø]ýÕŠ¸§ “lú[r ÁbˆN]NßÍ2Ã3Ü¡ ÿÿ"—‡àºd©1ax¾É—l¥ñÕA0Û[H^®†¼¬tSJui·²£+d’îMâ×"J¬ÚÕd.4tT`èKæù¤ÉçmU5@[ ?­‹4_™}ÿe¸ùšÔàiZa„{ä‡Å=SYã·XçœNo„žlr¨·Þæjö±JÅ9%ú˜'cø ®|üüjÃóQ¨°,Ëîù2ztêš/œæìÞËg2ÁûÃ3¢z9£ É·%,F3Õóúp½®‰Ò,H-PHäluÝcÙžb6Ÿç;ã<lß¡jDýPßµIÑgÆ !©ßP@ÏÔ¢Áˆa£ÖQíYá‚Äðú<«/Ê+-ÿ,­õö²ÏÜ,eÈb‹_e>/BÉÿui- HõùÞk#}d"«bK«š6g¡x]n{Ú©ü•è°‘bŸFø±ƒ/¸/Ów†R U¼̰Ó ÁË7éB 9“Š•,v Äé#Ñ‘¿_ñÒÁ€¢b€‰­î:)?Kïy“m6ÒxgÝ‘2ä*ˆÅ¡ÅçõŽô³éë×_¬mÉÙ3ÙDZc›–îX#ö@,òÒªf¯âü€‚*©ŒÅ|Á³”¾žN>† F x$×”ÔG\ÿG¬6ÒŽ“çšt)2Ô=ŠúÄ =cžÚ†<1€®g5*Áî{1ÏüIûsk vƒêž…ð9ïöNÕÆYsÍà÷øñ3ÝTD*£Ü@¦ñ©r5µdK'  GRÔ§/!S 5ÅÍõªñ] é}´"_Ådú„÷ÄÆóLî¿…èÝ}·Ç¸N4žPxIWm1Y¯„¥×~Î6$R"Ò5“?[çIõ ’²‚>c BA hÀ$êÀ#O˜ï“^˜ EÒö)êŠiÊÓù5åË۪΅»»œ°÷õ5ìùÍEÏÕm¹¼Š›§áÉÊáÝÒôðÅ¢Ö+*åüÉÏÁçkªžØ½æ‚ΩJVÙÙš«š*<âiݹ#¢¢XhVJ±~JßsHÝ{i§I˜aŠãNÓ¤?NÃÌxpÀI5É1¯?Û»tÛëëÐRàů5f¯ïî$“¦·ð!§(ýüëéQ¸gܬ(àK8ó–hjø´û(O*0[FÁ'¶“wƒ¾³²ù+¥›>[/À±“™i‚S-wþ¬Ë#»ÄÅýIé•[ Ú0˜«D!˜#œ¼ü»Á«|ÀÙ‘v«”%7Z‚¾—gx¾!6g,š<×Hvý6Žu0‚A&v||nDÜÑ@®B­èPÏ,ˆµžf¹ö…3Šk»LFˆœ€ŸÆ<¾ä‡»|7iÏ pô¸Ö8HêâÓ‰î”î´ÿ¡’‹8òUEn× ?}3{Dö`aÉÁ&º?4³Š¼Øsµ±KHß¼¬‰fþwReÄùÓÞ¢¯qC¸0ÅISÜM\Kwyfj $šè]™jË‹æZçòQx󞉛žÃ"4 ”ZÑÉ›A çå"{ØÞ­ÊÔ“n¢á½»l€wAùÇßÿn4Ÿ…Í->¦ÅÄYVÂìCB¾âP,\q"ò¬9¨]!sˆÔÈ ÊÕ ƒy?Œ°{³…äüIðÛe5|)¼zK•jÒuKKØÍô #¥njË‘`’µvßn;]ÍÞ#W¥‚³ÿk-&/7:q€ÂáÁKp¢¸áÀ¦ ñáç÷Ù(ÂÂÓü$}ðlGÙ5ð±ûs=@[‰­l¿¡_Qw_ú®RÒ$ƒÓ7²Dj­ôè ;£T{”¾p|%âü¿¨ƒÉ«wxW PU>4¾Uö@¾øÚôâSÙ~ë⿘ÕGYhÏúèÀkšzÜû\k“ô` ÓM6rB?LwÛív"äÏ>’`BÖÔM;|50¾‘9†Šp‘gø›À'VNø€•ôZOË#ù5ºÌ´Ašÿ.õ|•úØöˆÒ”ö×¾%ïcé"ð3¾x¥='{0ðÉžì$çNíOLÖŒtOÁ=ÕÔoˇ¢›ʟzDÓFÍÁ‚…ÿc—јMú·)’kJS½µ|™VÛæÿ8¨;Ð D·è°襰;J§å@P€A¦.(hh=çû¸’Dt ;¯àÜãNj„ ÈönÍ/Û½Wx¤òâ=€|%/·¾q6¢Ù‡6þ3Ï’DM‘y!ìP»âð2©,ÔªÌ"MàÖÀ¿Rצçú³¾—ß§;Û2ÝóPÑcó€Üœ ÅêO§g¿ct¢"é5z}B‹KËŽf¿§Ÿ_»-¹fo*%¬+Q«y–xã µÈð- f%þ®›€*££2s6УƟ7Päÿ6ãÜÌ.h©Ëv@·%mŽ#:„è»–‘ ëìK#ç6‰'% °Ù3M q&ùùFILÓίBó{BÑúc§mZÅõêŸ×¦êÌ:zhˉ¶ŸÝuo4æ"‚{þx´}Ú|ŠåºƒI%*ÕMÙ›QJ‹ÛJõÏ©¬|»"wë]9°¿'wÓ^b4¨µÝ ”‚ýÍDÒ—$ª,ÆMö@ð\÷0’£³à–/+"·¿&'è^6|Fdw±¬,ï·ÿ¥UHÅ9“ò™?É|¶gywàEŸ“YpLü?_T\×î–±ì&ŽÛ}Z?u–U`'¦*_5ÕãÙ¨ÖÎ/ä®Ì¾×~oÔ÷}Ïc ê‹4Í$“ÓW ùãlõu|±'S¢›Î¯(ˆE@îTîG]Nþ©Ðb•x¢JD–|OÓ÷@ÆoÊË—¼ßëú ¼u® xœøÜcȘÚ+€‡=ëkï"áÆkŽº¼ŒM×t°¼¾MÆïI;',lKPx1.X‰¥Ë^ð_76òX#hûzŸ™È&w)„øÔÝ'¾¾Å³”Ø-ÊvK­uAà ?y³vNø·û.*K$Þ‹ivÉüκ›ñQèͪ-*Ä÷¡¸Ó]ìÝ”#C d²ËWߺ‘O;Ÿ}(>'ùÞiäAÉEX.ÑÖ…yR|ýðæ~¦ ŒÂ«=“ݽ†kÅå%qîÏùØRx¦GzËØ}׸úâIDS„söó ¸^9!ÉQ6M@ÊMa;ÇòÎPcSs9¤iŒ¯»ˆW ¼~ìf·_{ÁT°±x*Úrk?üh=pK9°}ò {Çqô/ i­²ÃdMºIº[Ц§éS]2RŒãPÄ‘^ûw”ï‘@€ýiñ !â2Ö Nêpßj!‚è~٣īj!oe¸ÚFã¨lñyâ4z²ȶýÄœ=7¹žòËÀþM5 èæÔ[¡ ðp‰6kI@Òöaô„N'ýšñîs‹¸ ÒUn/wã¤`M-ÞEÁU/¼Ôo%Ï1Fj„ êÓµt»ô!ì7oÁŒÎ:mÂ5gØÅºñ€òñva Ïψ¯AÛ•—÷ozn§™ Ø:Ú½¹€+[d¸4Œ¼uBT~7¾9Ð. Mm³}D¬;8÷7’›€A‹F[ödÑÀÊke]º j£éÏiÑó>Ë.n4ñÝ3Í`ä*‚1êoįÛK:íLâK®²Š^NZ6ÀÊ1Àƒ®é¸"N`ÎÌÔéßeƒ(¦8î µÁ€6ñ‰,ÇE$g>n3¨5Þ¾%ë%p Wå¢Gžïvûø ¢šKKÄZÏô9Õj¶Ûá?°Èòfˆ6Ç+fBd4çAg& çÞC£"$úåáTp•bÎÖØ ›G¤¯x &*½@™—ZUI%6ÑŽ’P‹Úhص˜¸ß¬4\!uEøoÂbÈx›PîúÃAˆöhpv4Ñp3XÞS]ž:ƒ‡Úcyž@F_þeÇ~;¤iEÃ7ëQT,í.D%Íl×,dïmÁ% ·Çé0Å¥ãE£µ:~¨ÚtÇ…Ÿþ±ùrÞޜƊé Ê @+NuçË׈üå#_-üYE¥¿JþÌýZ›[)oXYðüÒ)ºËì¸ÞIìeh1¹á²ÆÎòÅ$ºM¾–ÖV ¦ßî2N}nðßZ_Ä3$¼KÀaIí±ÈÂæ_›Ä¦³MS•ZlÆŒ)*ë2dxŠö«¨¬:%ùü 壟~íÁX¨Œç—¢|z½ žUpǰr똖oŠø¦EèQír“Ζç sÐ9¶Æéz¯‚i§¦õÇ,bc©ÇH@X‡­çI‡ƒˆ·+ˆ‰ÒG2¶O-Â: ÆmÑáMÓë‹ fr:ÈðRë„uê\ú]ª~²¸úà DƼµÙÎܸçÜ:šè»Œ<¸wÃÅXÿâ¨w<Èb5 ùòñaSmÿèõ>ŠÞÊquVRh¥ê[¶ÈƒZusŸªDå;eˆO?ex¬Mã„[¤ÍGý5pHÇF²Û¶£ôTÐþ45¥Ú!:Ï\Ó܆Ij¹0ÇH¹{¥~Û_ÿ(Þ‚r(,2œõÀœÄ@•èW)WÁ yÈw–\cÈ)tÂ_ãÔCÙ–µÑx–牗v­7Å–`2£È~&’¤ÀÓì_€jÊMðEØðÞD1ÞLd©ióƒѧ­®ïä,³–¨“=²G~…ÿ”ó 2ÓX•ʯýü¾B¡Mœ2þ# é&l‚ÒÅ­žáMUÚ6b„DÐçpg¶díNXÖEè}$f>HÃI únÈRŠÙï9~V'œæJèŸqØŠRD¨y§íZ/± 6RE]ó˜ —Î*J [¤b÷Él²"K ê`µL-=±.à›ŸVÊÉ98”ÝæCåMˆð$ÔËåv½ŒAÓ-¦É޶ ‡&Ôl£ !–f‘ý2¤Úû3‚xNŽ,bwšÆUÀ<²4äQÀ%û:7€‰ Ý=vI·º¹ÖØÊ d¸|Dµ®2ÀtŠÁÑZþÏ—…›õh2/`®UM’lE “¿ª9,ú ô+ò ÄæœÿRúœ~( ÙŸ Ür|¹‡D¡Þ š@7º?Ⱦÿ<]´_ð'îXwܘñ¸æÃ£Êœ™å™’N(¬ðØ·_Sr# 4¼žÔ!7a þiûÜ&Xch^–M¢-44OÁ_ç~…ã‚ËQÓçÐrëʸ Âé‚é-±›Ø¶«™ˆVƒì'¥–QÃI†FûÈéùío橦ØP”ß•­6í:˜Iˆ’wÄŸF(Ú[Ø\’Κ=ƒ6ö›wõ˜‰ì¦ˆý®$ÿ•Ž™-Úg dh%k—¥®ûB1‡»™¢Ô"M‘†6/ÀåØÑËh ¼¶ø«^jמ·4Qþ*˜Œæàwƒ>Ù¥Ðn¢ÂÎd$í0—ta×Xü|“k%IžÖBGaEùù½àéÈJPxMÆ}yÌÁäÕj€ ݤÒÁ©k¡S\„¯ŒÌ˜NྛÂ6K£‘6o´Æûu`¾ Í› ÉJSÓHä¡- xH¬xËøÜÄ YÈÇŒf™Îû½B¬ ~@ýP¢ –ÈÑæÕ7ÁÔÍrÊdSe‹áCç'èh“itX”\ªÖMmÏzå»aµˆe-aløô%pÁÓýÛEáª!ÙÌTÓ~0Àá†ç5À#(þhEóñ•ðû_ÜM´–š °ÚTK Ö3ÑÜþ§–|ê´JG:\L†)B·`bßD‘ErñäHþÎÇ_f_l®ðHì¶ÊYRT>Aù.Ñ–·×•Nùv£çDLQç_ v«ñºÛ ^vK‘ØÜÍ7ºíœñrú»f³`n`ÚÀüï9'ÿ5¬u8l©!x6Î)Û9äK&¦çý'ìRx9µNj´ÕQw«fãÜk(8ú#좋án˜^I£;ßå5— j³¡23r„Ì¡úfð\Tô©9"&`¨ú¡è WÖwäŽxò+áÊËú`?~dÉÕ[@æáÏ5–£ÏEfé=è¤X>ÐB…ÿý—ØäÖ6 hhª’R2nH8±¸i•!Cóƒ¥ð¨pG¶˜Qèuú?Ùn‹;ñ¼‘»‡%¸&.0Nh¾Ð í*&“ù§—àK—£–’µsBœjó‰LÿÀ9Iú¤«6MœÞE´Úýçöwƒl­sI;ªc¥w+7úlœ¤;ÿ0sBù8ËE?d¡l.Mìß"ÈŸ™%i9º`¨Ì #k2¤Ôš¹ \Ißwj§«rS¿/@Á?}îBî»RÌ;¤€ ŒÓ£‹ ‹.”e]ºnoÅ‚xw!¸‚ºþ(áª*Ò×SÍ”;±OðæÅÑ„yЭ$!3Â(+P'ùrõÜ%%ÞÞ #—`®Ĭf²Âä…wÛïZÈ0´HhþâN™²c“ð9œÄæ]÷Ç6ÁP½vér$øíÇ®’µ‚qµeB *¼OõE©å5)Ï´*ß³ëý- fîOéç?Õl×ÍòQîhªœ°|õ5Ùwâ™§("3¨Îå8›@ÒŸžˆ¿±'’³¸ïåw—ÜFŽRDÅk©#€.ÿÈRæ¥é èšÔ¿|. öK§Ð–­í„ÉÅXÝø',éA¡Ÿ¯Ÿ«í>w¾òmÕ©…ߣ2¿› ETj>›ù,"ý³”÷¾‹ ÑÅÒ„P×ðZwíiÇMÀ©‡ëÂÀëéÌÇìœ[Ñq,àRÅßkÊ•±à+3#4ÖNièh}ré+Ñ“‘- ÑñlÁ®êÅŠc:¾ºìápv­#rÝ–äè9~ðk4:XPBÝÔ^çc>¹&%Vy…&5]¢F´kü4Sdù¾¼ ËöÜÂoë›æºDÅ]ÏÄž%£×5O,¦@Í­§ÅÄ?@Lþ8éâ‰HÖ.3´­y8œÚñÆýZÖròü\8q!Åq”*ºŠ³$§ã»^ÏÕš§—ÃEÈWÇþ…‰'±®º×1Bèã¿‘£v;‹QL•£äP¡Èk|ˆúïöøð[5'ÒÛ³_3y†oCÀYXáøÒˆž,³“{dH¸:©ÝÑŒ—ØÆ)ÚùJFY"/[¬„:úTk iˆfìßDØsèï¾^CP™6z0<øƒ¾ˆVÎd¯Á½‚y°“Q˜Âá Ì}‘6†p 7(Ëœ‰—¾ó7ÌýWÚÛæÕöLíbëqÎÍÇ¡-ËØÉ^•Q‹€‰“Š® ­”µÇc‘Gœ©rÂos½ÐIDyÆÛ.Üb°N¿ù:â}Ø“=rè[½>giõôñ3 [!J^Æ÷-‹•çšJü·=ˆIfL^_gï=;eðÊUµcjÍ®æ dë£Êž\Æet„\Nº°ŒâMýðe¶«Ço›7EΦ۸m-BÍúäï!à±µ"1iÊ›eÍ'‹Bmî¼·7Îa±X@=+ãG-`[6LœÅc¼Àa ^ÜLµ¼I9{Qé0²¨w%¥ó‚¬ÀÄfr>÷˜ºÞæ‚滚üÔɼ”ûBÆþÞW¤F`¡î'öÐŒ™ŽÜ5ÁËÈAH©·f¹goü,ÝLÑò’¦äÉóòÈpé™Nï/±þíg´!/g¦HYʶØ.À;§-‡/¥E/H†*×iåfõ Ñ.RÒ¶=|uL:ÖÙ‰#Éã+ ¶]=‰äDP²pÛa +QÞ“Ù#;jÃÇüIÝ¢ £XAó-6Ë´VàúÚ€‰Ü¼ƒT ™±$ûZ$—ª_ÁKq“«V-ƒ©GŽä­PmxCž§Ú˜óã3Ñ@@Ÿg5² †tºr%þ•RÃXåçíp—ðƒÌAÇÖ´µÔ±ú& ¯ÓI8™ŒJ·E(p†·žAÒç¿def,â û™×Í›ÑÉcUÑȬ†yƒ›ið\Z¬G²wq/A¥²Ó.)R0è;¨æèàšÏåô#kÐú0PËeî$B3½~mAHî©&¡IŠ­¾,ûÇøßë_é\ŠTA ïÔC» ²ß{”!ÿN’:Æ:*M8´Õˆ2ò·€f*?ñöÛÀj9øX½ÊðGä¯^4Êï—Ä´0ô¬ì„ZÕË0&KHÑÆôÂq9)ÇmW%tvÉpã5]E‰j¬ƒq9ë!" ÄЂ7ðïöŠZ‰VVúÚCþŸ4¥Ð1™í¯ÜÚžÍV"µæ=œ+6UÍÊ_=߆=¦ëv£°`S‚Ÿš{ºº±CpÔ5y hœ- ~«°Ä^Mô³·w&Û=FÙ3"Ýr…Yµ³eOLk¢.IK§à¹ç¸wL§¾Òzàï Dª }Ìúðh{|PT¼k§½Ã³|i¡‡PŒërMéù"$š²ÀG‘L#a"® tˆX=$h’çUÈ”uù¥2Ý_äAv€ê±oÿ'þóf¾x¥væp÷-~ m±¶''<2¦½ñmTžûO ÊŒ6ÿg"i1·ñÒç­N\\k9¹{Çug@\,ÛGrAÃiîÚ‚ÇÂ0YÓÆº,-Þ¬±`êÏæ¿†šŽ™*^ä˨²:NT†™ƒøÅƒüÚk@)ˆhëuÆ×º7Ça‰ ;_h}q•ºÇT¬Ú¶¦ø’}Ù¬¦@å”V°ÆÙïC%z1?omàJ6#¸¸‘T›¼°ÙÄd.î·¯~þÐ0&õПh’ £6÷‰mç“=6t8¬3=íÎúYåEá5¾Püu >˜ôßѽÃRw÷nŽÚÑœý¨-¢ýTHO*Øëû›ã³ pcÒÄìàËK—%Ö6½ï*o4REÛ»¬ƒÿÊj».Æ ­óý8/Õ0žJIÞb¸W h2b"ñÇÄYt»Ž½#š­ýÕáÔU3úÂ}†o´š-ËŠ„†‘?và-ù c+ñ¬aÅxO%£‡² ›Ùj¹a¸äiË33t£-ÝbSë¨CËX%ªeOÜu°Ç<;ñªìK¤^@sÚøk’LE~µ1‹ñapådá^‘KqH¥s$f[P)«ˆ}íöwï¸U;aL˜¸àJïUŠ^H"Ó7SA&tÚ‡hC3U‚7ì 8>rï߇nÙã†Á¼È÷ÄYžVA—ýÓ0ãE4ÑF åýô ©»(:5‹¶« k"BócêDÎÌúˆŸôa<¿áÃ#œó3¤æÑØÏ&“†}ÈÄc&%Õ..…¢[CvÀcÖºt ­¾ŽÇOÕKæâ&mªK“: !±MË+ŒŠKoƒ•w©’c®þºÛ-ÕŽ‚îÞÏ©J­kñ°‘Óâ¸! kñ»¥Â|MuŽÝ}KC&sË7ò_½^”˜hjíx•ô<ž×Gæßò+°C¾¢¿óÅÊ./ñV(z»;j\ìAŽlí¯­øZ- ÷‹Ï Ë0*¶m§IÉtr•dP°dš««ÑX”$'Q†¤ ¢VÓ-’_û¸@“5üfÈŽ‚¡tX%di»/Ôi­ExL”"'À7åÖ§jk(~RvãFzKO2 ûWÒÐÜÏ—U#“í¸)KŠ?ÎÝ®°x<#§ãÄ|C°õåÓÊáþÆyµ*=Õ•-uZí€ÓÏnT9K(Œ°²ßq!³I ö• Éð'$§Ô<2sMú€r=Éÿ@çŒÜ謈¦wZ™T“»›$¢s<Ïø#5Áµwf‡l@;Õ;,™<Ê ÎS£Óâ‡ÒöxÒŽ…~­¥#"!`¬þüßïßSÕY:3½ÛTÄ3•S{ÜÈéJö˜¦šÕ˜laÏÚzÐd+Ç¢ôÍ—Ïž-öiÜmúEýx¶Mv•Q,ÒÙ•É4Vå—Ý-žß ÿ%ƒ<@Û³¹â¦¾PÍk;·ø¸Q\(ÊŒä KÔ€Y;ª©F”FßÙ¡åwíq¥«ÆÈ³Ç·©ÃïrCÝ–ÂeȬ–žŸ zF³ìØ >xý­×5J/v}:~ÖA;ª·OåR*uæ€ÌÉí¶ •Ì9éR±/‰ þÅü?K#åÇ~ îz Ø·Ïøï¾qg;mbEf2)°Äv¿GÂÁ‚ÚÙµËÑßMƒ‡ÄNåÛº‹=g!6J‡¯xûÕYh|ñ»: Vc±ãïè‘Ãf)ÓM”BŸS;®O.¾"ò€K6›«Dîú‰EC‹çD Ã`N¼ŽÌøÚ¦çÏT{ã½,RXSâ’_‡ë‘ºX·óQöº(>‡ôRóiÞÔ§Õ•ç'o–2–Yôÿ¹s[ãŠÌ×ëÚZIËŒÔà@%¦×u;§!ÙFÁJíµÐ‹~Øã.¬ [6Oç~NØÜáάÄK²ä»A¼éµ¡ÂA4fžÅn:Z"l²èŠÀßZݘ$­­f¦6¥yÝBzǧ#o}žpímw—eïÁáy™ S—‘L\õÌB0k“Á[£$ŠëPPÀyë • ¤¸y†–z°E|yØÌg(Á"/)¼æ|HÇoiqªÄÖéD.^eìK1Q˺EÒ]e¢'A“ Ò?®ºŸ¦¸¬×·×ä6®jÝb`åNé4œ](t¦ŽïÈ`'cCÜò½eÕ¥…ýbú”á¯æìô$t7© ¶ªÉ·ÔgK“l9Eq¬†*Ú¨'08å®GSŠ_2Ì &Ð|vßÚæ»žZMl±tþ«¨Báv±¢Cá“ÅfáÿÚ"^Ìõ²ökUÉ-Ù¬ƒÒ¦Y®÷¬—ßÈÒѵ=­îáîJ¿ÚØ[E›ìHƒSŒHœøV7gb`³…׈ #Ÿ:èiÞ¼ ¡GT/z\Ͳ~D¿è| 3‹%}÷ÿÅdØõCtIcž UÙô¯ÆL©%2]6mð ßËÀÚò÷sºO£ÝõP݇ÀŒK'aèžòÀÏQvªšþ4…{ôÓL²2FÓ·5xùõ6Ü|\™Ðq&Tò˜dv#Üøèáִöãkëê¾6Š]4§æ˜)ô`¸ß6®o*öÁ«tÖ •wè •mç TtÂþŸ3Ð6µuÛÏL:oJfp{%•¾bgù|ÆÎK|KZpÖc5‚ó”`ÉÅCñtåØT@ÁÏIÐlTxúI9®'‰³_°*nwí¦S'™>7ÇUQøó,AknG¶ç ÿíÊïìdö_¬F8œŒç õ—WŒŠW®„Ö•­‰ ã¥÷‚U)H—”®‹¯@3½¯kšéÜ@|ÔëW"r/ÙQ8Ò¡U·—fhµœ1¢“c8q†·¦ 8ïac°n©s¯%CÚÕõ[ U¿´l ioÂèùO±÷‹Pº—o·^sPƒÝSÂLÐW&.Û¼_$8ü1Mi0Tó|bºí¤ê·§¨Æ®¥ìP ÛùAÄ'ȵŽPlS**y>‘~‡Áki¬ ¸¾þT…ïý“®Zª){ÖYݹ´µ·Us/T“á¯õž’×É”Ä@Å{,ƒalH¡NÖü*ávõ’Ò0üSÂÄ1†—É/d¹ˆOX·WMg¢gÆ?—Áýo%a|²a ‚ìqïž\µ•áLÌY‘´ ²ƒù0 ¶Q!Z(μ)À‹+%J´Þ@ù2öËës΃õñuöñÚÞ@«DõäRåDX æÆß³´“ðžœ15ÖM5~¢UZ¬¹‡Më!äY~¾“žQlj~épÔÓnx`ÿöFQ‹^®1_Oq¨‹Žj)ÜBpÞ!!¹÷þrä0§7V ¥î'=5[Sß(^ ±7D6 nô:!£º”+¥@ÒC»[È»Ô &ô†û´©Xõÿ›é·û³W(|·ÏÜ4Ò<Àé­€Æå?¶W¯ML€žq€òC¥m%ªñ¦‰ÜóW¦(ù` .×»tH<†GCãîoC8ú £‚¹Òdߦ ‹0®O…)…&”à;ì¯ù=­òßßòwÜ(iDõ5“=Ï£9P¸£oAŸF´oúE­;©N2¦1¶33 ùÙ/ÅC„¾;zˆìÇZŸâ £˜Ãm?•›¨»*ŠE󱦥¬ÈæÔ3sKÙ]ñ±È ™Ú픘'PÕ zTÞþº‡*o؆³æƒ¶ÀkÉÓñmÖÁ룯J\ ¬–=/,CäøW²“M©T;c¥ê^M+8rn‚KŒPAh Kó3¨ËM¹B¶Fy¬c-úoÈî`±»úvE„°š9á©A[ Y˜á˜ òÂå²w-½/'üîô—:ÑÇC°Xhý?OÜG»xc²Žlõâ>ҜãB¤‹—½²m¦l²1CÝog¢bDSðú'âqÏHÆ©ïsµ4ÁüãÖÏ *ÚEB0~=\‚—¿}Av‡pÀ ™žÀög(‚k]&íýªC¿ê2\Ëce+š €ö|–ò› í¨]ç>nf,ñ=çÕ2ÝyNc‰2j–ê©€7"¨êº¯–„©ÐŠê"8¥÷éÀ㨢5©Ñ·Šæ±=Œ"Øâtýb}3wÆAbꥌ(Wôf&f2Œ/Y×·¦Ó Q„C£v : AèÓtVyEʬ\3´¿jõîGÂÀÏ]G1 à “Çâû·Šx{:Q`"ŽaU~ˆæét°´s°t0M’¯?‹õV-'øwËq©|:sIÄˆÌ Þz¥—¤J5žØ?“´åªB€‘‰T ²Úzš0Åõìz Îpz×¼^zõá ÝæB’WLKÒèÀ?Ȥ>s)h?¡°æ[4X„ÎÅ€û D ™iŒµÔÉ·ÑQxÉt~ܼÉ¿ÆÔq 5Ñsz=†ÿ—þÓuw7úA? $•®.¨c,gü$ ýOJÚ QCRñŠÀ Ç‚[¥ 8ª\J¿À—®+ìÙ¡»?¾vYöæÙNË­ûUw`o…V´6P¾Ð~1ð.ÝtBR†0’õ­/hUš"èüíKÎírà–¦³á㿆Â9fš]I{ËxzÒÞ£Äî¼D­Õ¿Ì9ýë"¥Tª¶^Ê9Ê6=C1˵¤”ºþ² ;’ ¡ë&²¦Sþíó y½Þ9€¾áQ„û&K7šJ §øB¿Eà{YÀßÕІTã]—y}T_ôåL8ú©Y4KŒ§…oÇÿ6†&Vn¬²Èê­PüPº@³Á>¢£m%š¤d•ÿ×íR"µÒÌ——’·’Ò…ÙôøIô›…®b£±/UÛŽ¨4–Œ¬ì"aÐðsŽeý¬Z”§(³ž ޹roòsD;K¼Æ01é?[!Tq§…íè[×]GÉ@¤c"©<‘оWñœJ#y~(/Y£$b­b;Õ‡À÷Ê 4hŸò‡æ! y6>7ã?­ÏÏ>Älo° Üûøqd¿ŽÇ¨š23˜0ײÜzñŽ‘œ¬a©¸"Hp—4%ß™šÌ—É(€†ÒÖ÷¦à½i”HÙÜ/C„üQ ;¯#¦xU›ç|”DYz!€HÃrR¤®ô”ª@s bæúÄjÕ ² uC¼å“ªÑ4N¥¢ü¿Þ±u–…ž¸{ÃÂúÚaå4¯[@ÚžUF3{D¢ÿnÚfG‹a°T­ÂV±LñusÜ‚@Ú 9²Ÿ„DCq³!3&‰Wg¹É(è9Ÿôç ± h±#dÝ¢Ø?PòA?`ÌÓï‹é·~ôRXM~WÚ ÷‡5£ãDeø_‰·/C´\>ŠÞxÑÔ§³* <äa†¸`uQg¯¶)Ó2úé¿=Fð/¦’­yí¸Ÿãå—QôÇIlk8 ùB¯ÁÆP[ œÐοw—I¿a¾0½WÅ(Züø[ó ¥SU™–io/³àò8Ä×:Íà#cѧƒŽ=êÔ_{×ÃnÄ»Ž``Šæ8Èg Omt{ŠÊ•Ê•««§‘Ã9#|dÙTß PÆREÏÞ;,[êB¡âøy ä±’ûœ}“¿°/1É|Ê ²ç‘ªÙú6v“kòÛBò¨ñÃÊÅ!T”*î¹äW‘8nй~WòÅ<ÞEhœÐìÏ·èSð’„BŠ- O„9?Ï‚ü?/œ—îI'$*•>‰XZ=­ãØaG£žÇPÒ2¤ón­F€Í±êi: ôÒ4 ‡Ž ã øtÁ=7逌Às€[Òn›à{3ˆqúô¥]èü:š‘NêÔ4àt•ÿ½:ÝÜ­:ÃáäCŽ?HÏEÖž»4”Ó6­CGë757ƒ;å¾oAšvÃÒ #EÉ®},à§ð]’($óºá¸!€þ*¡ IEòÒ+"4x×^uï2 “ŸvŽ>±rPýHÓð[z´a-Ø9ø{Íei漘6öø­P\©cé>k iæÎ‹% O¸ýž—”¬Å›C7Åf”} Ÿ ;oǘº*ˆüùÑïMß—;q¸“8…v+›¼[ïäºuÉÿñÜ"¯37y3µ™ï̘ìîÀ{Óí ËÀ~Gœ‚ü’&2ÅOKšžé†O›ò´&FiŽh'±î–FëÜWƒC©³ 0L\äŒÂðûaÚ*ôt7Äþî٘ –ã9Mî¸XÁ©¡%Ÿ-!ûž?‘±í•:ñé-@.÷JÖk_–AÃ/@º¶æ¿Ð¹³ãi«‡%2ú lШòýi¾úNï´À »Š"9†¢Èý§Eöò<*‡D`8hâöÆ>«2SRä /(.90RÏT̸>ù}ƒoB–†±:»…xeë(í¶ø_…ó°Mb™5&åMÌ$—ÙMV¿ÑóãÐÀbhOê”·¤í›*J`ËÖ§`ðL‚âQÁ½bóu»Ç®! ¡˜¢Æ3«Èßœ ¹Zp®5ŠAo¶ô­_¥;›[ü  ä¡k6z‰¦Ÿ0õïr¶qŒ![‹‰8m¬]’PG“oxÖð™ÅÕV¨×ªá¡ã,Œ•#ýb0)ÒˆÝIÑë¤àú&RÖ&‚k„¨òÌÙ ï11ÃÝ0Ž–é÷¿£~¤fî}3÷y¬Ï„¥ÅÇu\óã¨ó–C¶5¦¼CÉ;»Ò½uk}Ä6-íœÑ-ÀßÅýâ^`ƒå‚>TŽ‹ñ½£Ÿ“î2í?®°þqv¸ 6ï8!ÛâŠ^óAÒ6Õ …îä•heÁöµXeG÷Kíîõú¥mpKöãQð“mÇb÷hS4/SPV»äS8_sè[ŽÉ,–r,2Æd’eZ=äv® ®Öû[Ϭ'9f0ó[RÊmÞvŠ ¦~ÏÏò.Îë?Š ~j7w'ÅëÑš…˜?æàîeÊf´®8õR&UTG` _™1;áéêE*Ä RuŠâÿ"=K1 ¾²gb6r[ tÕ8m†u¶M0ÉíØCë·| ‡x½3c0Y9ÙŠòŽ`ÖÆ;ý‚’gTâ_#2±P,ÜÞT°éÄNXuåÁ3Vü»ñ‘‰Z…Òú+úÝwò!r0s¶j“iZçjÿT-/©bØ!ï8X’©Hß=nkñYCü6in^0¨”S¦¤ük$ûu„#^ú×oËeÊ‘OzŠuÐaŽB/­¯¿I¦¥Âb˜MJ3Ø7¢xw =§¸ cªÍ¹§ƒ>vw }ã' J›eý02Ÿ$D %%¡Â›™õpnïÒñý>0÷’·íl»®° “ð}¨î,nZ.c̺‚SþÏã{F$qý4A³jª¹Î¾}ÒÎþ^˜ýËí«I8cÉ”ˆ•º©KÓO]?½¡úž/ÒØÁ¤rbùú(—÷IEÜ?`²vìa×wÇ3ߘBª§›ëCë¥v(íýT%%¼L×"2O®m«,Ç,ØU¹=ðË^=G¬~sl“¥)œ.¡q‘1µ5ÐÒç°[01*Ñó†Dè™›Niåò÷qE|‰0/ŠGò_3úé×/æ “Ž=“£¦´Kƒ”Œß›¸>>øE.Xp¯f¡£qÈq‰b±Ø\šoé#´Î$ó©¼A¯.üÝë=D‹ÏG ³×J¨yQE™N€µÓ9ÿIÚÐ@Ùå}ëJ;ÜC–ßÙ”JP—æ xÿåßJyL ‚B6,š²EùAêz¬ë¿bCLwã¥ú%¦œrø†„†âXŸ- o¸ i}Æ5¤ ñœô³‰Š…Û˜ÊõÌC|]mÈ'ñÇŠZÿ:KN>,²öó½ÚCM®4J`ÇÛ[ÂÕsÅqÜÕQWµücH‹Û¯Î¶›G­ÞôAPV×LYƒx“dÑ3øMÀMHL¨É¬Â?1!Œ+7ëI½«ù-÷. ]†c¦7åÀ Ï >Lòƒ¿n7‚=6÷^©ít˜-˜›r•å?ÌîæÐâ‹«ÜX¸w¡ÞÝrR‡{k ‚B¨ùöW¾otl³Êìð>¾\`P$ÑU0³Ý„”õ½?. @Ðgßj8WáZ@.–£_Ä6ßš"ûáé`:°>UÐ)à&¯}0¿´Übë¬ ‡?°tÖîšWNžq!žî5:8Wï£h ë„í>™[5PhVÛ†³½8Õ0Ô¾ÈþÃGwžÎÜQ†LÙ÷‰ì튓Mñ|•SÈåát' Ø kùÇèCí‰ÍgÜþÒéVÁ¾Ñ,|`Í3Dù½qtZorFCÓ=8Á¯aÕÁ¿¶ =F½¬‰-öœú¶m·ùp¾*OW4(˜ÈÉÞöLÝç´‚¹—馴C:f%§\¶4¹rmÿŒ¦³À!˜ê¤u9°tP×yöhK‰Peu0EÖ¾cmR"äª:e1+Ÿ©Ð3ÄžÀ5ìK•„BSYp”©Îñ¡PŠ]—loÒaµ<~'¡ï\SèÕÇmã2ZÂ1=‘¸üNìüÜtã¡'—nOW …/ÆE5¸/Òï÷¸ƒ—½Õ0¹ •χX†‹·@ðÞTÈ\ 6dÏ]ªC´»ÊóšÞãîou ïV&SÛ<å(›¨d|õ\“ÐcÈ«¤&¸ä8ÝW^›Š&íH~÷~òRíDêwÀv>qþ•:­ cuÁ««ÿü®ˆ54£ör75Kóâ'vpN°´=† ›Ñ_;I »O¡q¹†“RÿîœòÉx¥ãcÆì–"xg.ñwÕÕÝ™¢€‹¨ï±ÊÆRH¦â•­vÅ›SfÞd/c´Mª8ßl‹ì[ù¹ª ¹mÞm3&ŵxý¦©ŒK†ÖºýìÈU ;šÌ?1gR Ô¥Ð<½ìkóÇú„†%­ÌS’÷B¥Û—…á¸/”s±¤Â·òæHŒ¶çü³Í‹ÀÖÚu/Àê:Ê™ ˆ„‚\€QF4#8¤ýÈÕ4¤ ywu ›àPúH<'(æ[©;o‰u .ä‹ ¤Ð_W2Ôái}_)üqŠˆUD˜%¢~MŒ'ßþuEpF†SóïaDY+_]–¶ÝTÔ˜W7õZj¨«ÁÇÛø”ŽâÕÜpÚ`ûÿ3SÜ_T#Ù­†ÊÞìÊÃ'óFóˆçʸ rPŽ$ øxÃþÑûò쎯ù˜Æ]ÖJn?”h¬ø‰r0üÕ—ÿ’ê —ã /ámã.'1 õŽ” `|`Õé=BCÉ%1çÉE‹^Þ…q­cõ¬!—èñý̳䚰Ûx›»Ðcë‰ó¦EÅÀQ¯‘Uaš?§c!¼'ŽË7‚çŸ]ªS¬ÿ¯?üt;ßó`†D:&PÀ¢7pN¸H;Éꪣ¤í„uýBž!û•à"‹‚Ãüˆûƒ6ä½»mcÆŸà”)í ¤Ü‹nߊ(îS²çpmšÒÈFú2 Dèv AF‡ ±UÍg{È•¥×í­Í¢KÏÞÎg›¼%'¹·MVFÿÇkÃÖ”°0mwLsk:Ûy‹Œ½á ›쉌ØpÐy‡p'Pëü§O–D®y0±×±sEóï¶À0q{j…ù¨OKm)VŽ[,‹QÁ^|Å;†¤Dœ¿!ÙS)ºšÊuÔ3ÎþÎÛņ»ãÝÙh·ÛCFoDiP×'#àD}€‰4²Mêq:!Œ¿Æ ¼_ñ#×( „Z𦱠"ATàlfvâtê ¦L_íXéE£ñKH H1Ø ZV¡|yœ––Ç#ŠË§}ú\xXIb ªüß‚µ6põä±Pè 9˜Ø¹†ÌB´õ›¸3ä£ä‚Øé~ ŠØ¦‰åÚ‚dÃO">†$™*ÜtüÒ_DT˜|^I–(P5 ÀʢÀÁq÷œƒlûÕ21Dæ³¾j9 ’'¼ð÷¥e5¡® õE·Ï¡á RÂZ§f/³5m7µÈJg‰e]Åë1è®\“î„¢–ó·]KØX®)Ͷ瘽‚³ ª¸öÍe#<÷‹]ñnàé(ŽÒÐ]SovÈs¤Žeà_f†=×*†áØfºãýšgÿ'D´ÎC›%B¡Ädq”˜Iø‹ù‰5 “è®N0™æÃ‹XÆ-ùœÉ€3YëáPÿô¢^%†¥ݽ§°Õm­ÆgðZuÌ(E¨»,°éDd«Ö5WòUqª‡¯dQ%ðe:ê–úgȯ¢^Þ°Â$`½høò/Qý©•Î7ô}u¯ð»vÂìL¤W—Ð)P¯g5$¿¾ zß΄EV°„GÇù‡[P͸½‘øN»ª—l2…1we#Õaea—&fyR›üZÃdìoNã)uXqÐTp˜|QÅ5JI-~ÅunäO”U9'¤´3PÅÖàèø6ÝÌiÀ;æƒ[añZ½é™G§dò¨ qêÖ—’/•± g€ˆzlV«×ëùi/‡;oä7Mw˜´a«¥çPãçGùÍßxÞÄlHþ_þ"*£N²5i ø?Jd—~ªß(­Ïë²k# £aÆ=Bªð1½³Ñ$I™îpøàlÍ}ãÔ£ævH$‡@dîÛ…‡dþþFå–+† ÷ ¿Î|ºM2)5Ä®:´T{Áµ: ƒU’¥ ¬À¿˜ŽÆÂµ\”«Ñ ¾Yá+@øA\W%h/%òB;Õ[ 34kùM–ã¼4°Ì6ù˜Î)'ΜӰp®ç¬¿iØR¤eÞÉ`n?G-…x‘/[?u˜—Æ.ÆtèÁØÝV®Á‘1pÍL@U«Þ8ç5À4ÛYH  ¡†y¥¯õd§ ØÂá¬}NS$º#± ±»£¦9¿Æ'¸jm×j¶á6X[·®(ñôÚ™”ïxzGr¯ÿ!ͺ¥ PB °J£ÑjL™ÍÈ Ê"˜‡Ç}ÀWYE±†  ~àÿpºïSúvx™0±<ô@αP"Ðp.##¾á¨wVh„ó… ¿½÷Ql€¤:Eáï­Du-Ï*ÒÇ™tí1‰ÙßÓ´}ÚæÅ||MZ_ÜIÒùÔTÊÚ&²`¡ñnÿ÷h'Ì… Ѐ§Ùdö«“!¨èåá¾xGЋ­‡f£‘QYU¯?I%† ÷T§}­`ùý-ZH¹ •­ÆkÈ ¯r¾+" g›‘…göKÈ8àŽ¿d¶*ÎùìSG·àñ;ÙÍ4¶‰Dº’À ïH³ªQéWz°¡Ñ9Ñü˜‰°pU„ó­Ç£ëb=©iòš˜štϺÛæ¶#— ¼7KóËœïÊXqT\*ž1ôÜ¥@‚Ð3wï`>·"xÆ.Ã×üp…ä,x“x:Šz’B*´ n  ‡×bDwpãިΠ¸ù‰âû‹ÜÓ{'w'à¥Ùñrp÷Ç“ÙÚÇÔ©4ƒ°°ó?Lju¤¾¸c?ïóPàˆ¶Žà ÞÔ››;llsyd²ÅUBÅxMd¹'Î\DO·'9î~’’ž'Yy«¿¦Ø5ö8þm?KÃÄ{Jlë_6ÏÛAºcŽ¥­ˆÂm8SA–¢€s(!s!@:ýJÄ ÐU&™Ï~Þ{ÕÈCˆßà (Lè)AùVÉÒ>ÿ¨ü&—ÕKÌ+H¬5ÿr릟LÚ™!\¬ŠY\›éÌÝV(d>÷]4ØcýÅG¿{@*éd¡Eà8Ûè3¢^ô3¸÷îo6Åš~šÍœS I[ÅU+s}Ì‚Z¸}a@¸$ÆÙbÄC x\W×}cqèþ¿ßÀ[ÑÉo2ÇÉ ã5žÚÉóªøÚÓT¬t»ã8D É îLpQTã™«ä¥ „攂ÁÔ¢<(€nš}ž×Èý)Wª5þšˆÆ!ÙCiõ/ó«/AàÈV¼‡ùåØŒx¿…†¼wÌò¸«ó$ ‹ku0Ç*BíE¥Ð_bó}iÞæÜÏk8[/‚iÎn1_î8#/||;;ó4´ý¡¯n²-ˆžK|÷¤”»ICØ ’÷ïïBÖwüDg%‰F’Ð7´žW5TÚÍ©åXENdóv%˸Nû o2îøÓÛ¡á SRAIõüè«Ox‰߇M=uÎè„ÙÓéòÓÌAÝù§ëâû”ª ñ¨á›HU+èºV•5Å,4ËëÈüÊñ#‹â¶U¼«XDíL:«B2.W®ˆ¶ƒq•ϰ?Á¯0d^ÁuÃ!Z_ÁA*Ê×Çy8ý玉ûàY‹–ý¦×ÇDhR”ÛC÷¥€3erÃ̼Fž@nª +‹¡€³<67ˆüi±JÙR͵²{ÿMUÄßRú—…Ð%uмwXš°ÈÔ9¨/Ý™k‚Ø ‚¶…èØ–!C?{·u<œÐñmÄPçt¿€OªN¸8ÈöÝ®²4··¡¦ùãõø>ƒã“—ÃôœßcX}^´®,ªùZ–ÐÌŒÆÿHåâœ×Õ¦ ½XÜÒ¼\óX#þh^™7êã'íç^•øATOøK[˜)£î¥p Dd0}—æø€­ù$ ƒæ¼i:¢Ä©…ЂbMŒwOeÄߌ>­w×CÞ8éCÀ:[ÈÌ_'‡1“=$¹ÔJ ç§K;…^MÕïïR'r„‹jz (#®ïn‚±.‰·íð¬ê%’um&¹‰îú"•2}˜2 ·!´‚Û:OL¯  Ê•BqúS×ð¼C¿-@4ú«]e[®÷(fðÎßcΫû×Ñùÿ†‡•Jå Ú}Á÷ù?ÎN´ã[DF»•Ò×Uênam•Ùe7îÆ1äüôàð6mã;”€c¶£ÎþWô¥—ÞÔ‡%p¬²¬öR_ÿEÍ7¨Ï_‘+Å``BnÃY«ÈʬÂa%ÎÏÁÊÁIpþ6ìò …¿IÑ/¼7iÂU¶ªnžv ÃÀwqÅ#8JF^“³ã3%”ÇKïLÏtp-Œ¿EÊ ~¡è°ØÚ-;µâõ±ÇƒΜoš9ôdûýåÎÜùaC$HîÒ[©éíG~ž{½eRÜê› K©ßh[ú¼*¦ÔjWXјf8Rý¤ÏäwîYÅÀ1 ÞÑrJ2dÖIƒÕÕ¾ÃWö濉щˆèø³$Øôâ·÷x…n9ÄöÝW ’ßù–XàÙÝês{íØÙsm4à@µ§œÀÉ”›¿«¬¯ r-1×{@º”ßTìÍ|I7û;T[Qø`§ÒÙ/âN±™Ið<çmtëÍ@Wu­8ºÙfŒÜP ‘Ï-iÍÅvÆÁ'?» .A4=4‘tÐØjÞȼ‰[AŠæ¿@ é1¾n7¹A‡gOšÅ˜½DD`ï°GbÉ·; aæÆsälò`1äÏ4Ó~™ú‰L€¼$‰žš)ÒK M×Fô&Âö¡¿—ÿV¬}¸û†}ئoÆ+¸b};¶ÄiªÐ ñÕ=Ⱦ$ž v>R½”>FD±—PmoÔ‹Ä Qä…DÒÀIµ;z8ã-›!{ÕÄœÐé vú¹X•×ê=JÊà߶ʤ+Ÿ.AuÄ3Uk†øÑ`Št˜e’AuÈY,g0Í›ºÜæ”HS"Ã`†t°"šVòº¿å‡{§¹w”²RØÈ5Š `3XUðº6áE"Ê·ÃÖ ½ÔÒí(®¹é”!F壄o\-C¯ Â:!Zƒ)DÁi'´qS£¬¹±3KÙWê[[3-]sìøÊ¯’Ëî6“|–,\/6N¦Â­æ`»Ù^ƒd{÷ì+‚F0ÅÝËØ6dq…d)mfY_ãéYrw xn >ÕÝq„ʦþR`oežm®úžBPõŸ×´º18 ͱ…ÖPzŠS©«Ä&UèÕ}‰™=„éèH¥wùþ³NXíÅCv^·Xof‘é¯ ]oùÅ)ñtD Ñák7×GÂÝÓù~ù'7ù‹Ë‹¹×ÄeQ˜”N^e"“6Qɨ¸®Ò“[&EoÎñ—o 1ë/»Ú-Z¦—œ~tCßÛ[ùКg XiÌVQ„B8Žá=؃žî&”é+ÐjÑ0¦Ç"×”Œ2¶”ò·™´åˆºÔ¡]‚§þü)µD͉¶éõÖ„«§q§z ×|op„ Êtc2/8¸W4· Õ”ó¤#9½Ï¼¶œ-i1®²Š)lB¹× -÷4WƒÛþvjÌtu¼‡Ÿg·»ÒAƒ@RuÁ}jï,^!¿; +ÕÜ$³û~Ϯӹ@’i7<Ù“üØÂ—Ëc}°Tªž–…•Èh„%Pæ BnlLnßÚ¸#ê°Ï)yƒožå<4>ŠÎÄ”JD …AAEe‡—ÉŸ+€ \œyç&øl&5ÍöjVß'a|ïtó'ýü`b ŽDÚ[Áêc-ô|¥á›N{#5}0©–â¡_jnBa nV¾–(W2‰¤íhcpÔf»*×›¿`T ¬ûï0ÙMœèÀìñáúöY@=€µCOïûÈ Ù¨p`ðå–Z1ÑØŸD>âß§LŸb‡,gª• Ö©ó-›hQ/Ênh4[]èÖÆ6mý[S%`Uÿ H =÷oÞÿÀŽåa,¨Ê­¶i”†“F 5y©á—¼ʃöa¡fƒ ‰Öo‡êqïã)6‚Ý¥Yž*fN ½àå² 1Ê(¤íY&©·gÞ1ã_”ÅÃ2Ÿ^¾î')3 Óïíñ†U?¿°õ@4Öˣɉ)u!½å„Ü÷þm½.¡Æ¥>›Æó´Ez0߀ò$ë=‘?TÇC^•oŒr¯]k¤žmÛŒNqœPcù›ås©|Êp\ñÖƒ9Š‚1 bAö>,é Ä'74¥"ãZ߃–—ÚêÃQu5Ÿ£ÐÁ‡~Xïˆ &KÂE­4ƒL¢s×rAž¯£@•æS;¬"Ó6²23®´kà X¤$ŠÊ‹ú ¼ ïžè“eÃ-diMg[64¢ì輄%q„…éNår‹0›@9`‘±©MþV M†w™&Zà«äÍO½xLñ^nMìdï—¾Æé_Mñt™ÉLåGvV™œ <^Ó=â¯.Im)ÝЧ:ä²ÅÏ{ÌÇ;c:N&»Î½vYçܦãQäŽ_Î-;>ègŽ7—ŽûIáZ·æÜøï¶ñiZœ" ‹vã¸U ]b3tp©Üwhßåà¢~¥‰ˆ0¹Vô±JAÚ9p€(fŽë Gè«P@{C߃&–Ëì&ûyøzån€*ÚÞ0ÇHûå2'2ì›õ7ŒÃæa¹Š¾„Wý=P±—EFOË™CGAñ 6sæd³NŽÉâ´**ì}À‘Gå„jÊQä-B9ßßóhnÿâA¹=’xy|ÄL}çïB°Ö!C~þ.Še#àóÁ%ÞpŽ™4Cn…(½ù1ůÊlå©!ÕȽ•†ÑÓ_ m®ym³¹ ˆ«#ŸIK×ÃAíÕ0ÐísmÂÃO í«È7*¬<}ÃC†©XN²Õ‹y ±#è¾–óy*`^%(‹UÅ$àL=Ÿ! Çý³¬ëÐuj-€ ¿U8ÔGÅU³Œh Y"ì–¦á¾Ì ôì½ ®W>é©K…Üj1¿`ÿJí0xÌYG`Xa¥ C?šQðgO†ûÖªRYDtû¥¬¿Ëv0KŒê…%9è­j ¤0÷ÓnÏÿÉ—Ú*¸ÁA+ø“óþ©â_?<˜I©W–F¼Mýɹ.“Ý&iºXÀÕ<“â!Ð,òÊ©×Ë>ÄîBOýF©+Þ`â}‚KÃD]në370‡²ú…«ûýc!!º$ׄ N1˜~¯úæÁdVZãÔ”ûô§Œ´î„–Iä< RÁDk2:Ì4 ÛoCêAÇRŽWÕrâEQ‡Kòz >ö¡ÉfÈ8¾5”m­b§Ë4ûÑ]j௠*æÉ a0UÁùô+a0F܆¢j|)Ò±¶ë)z"Ûãáv ÷My’‡á"ÞFûñ½7­hçï9¦ü*F™aÁ+ƒ'WNìéV]öíèÌû‹U•ç‚d6P_ôÄ/Gè–4†Lòš¹rùäCޙƱñ³¦ÁÕ&Ä8›ÎJ÷ðå ’÷à$ò•ÖBzæŽ •W”Ö)LJ‡ø…¥”J6„¿¤­Ø$Ó!ž^™f±{+Ì…75ãørƒV «‹Õ.Û ¿š˜éU3„£íØã@íÎ× þxgzAœ½,à6.@†F>Æ75_þ‹+²H[0 ŒRòJŸ­ËL"£M[¨>º£Œ<&¼{Ý*ð¶AvØOðylÏ—yÒë\t®`d»i¤abI<&¨tY®ºçwŽ£3«7X{J{w!¹à_¤‚š'ÝBÊu:óØàc+‰ÍEuÕ·2:dwî3×@§¯”é𬆠›ðpÐAÄÀïd ±ã™É«ž†GÉq’ÉîHÂõ’Š £,°eÉÐì2XËó±%¡å °‹)wv_¯Ât»¾­Âúl3¸Š7Hâï3Åá·»„µ›“‰+]r V—†cƒ}üé°¹¨ïß$l‹<ÚFe¨³øc÷5ˆ¦ë’ HW9Êêºç<»{NŠõÍÌÿ!`Dô ’¦9.7Zº ¦¦P%ôáÚ±<±äÆç¸ÅV£²ÕÉ{Ñ«‹Xà ¬|”oàû.Ž01ÑšS™NðúÜÙQó?rk¥ôµ ÍA=´–ôºr‘†7 &mB1ˆûIϹ¼Š$æ–»EWõbªûLÚ˜#G ¯Š…jÜù<|³1Ô äó–ŸxOù’°8=Ü4üˆ8Lª6WÌÛ4úéŽ(†fëèýg+”Ë}ô…Ýl+ÈÙ6AC¬‰áŒâJv¸N r7\ö8Bµx©)>ÑžçAÓ+hú_‹‰ËˆcíÏÑc=œè8\o —Æ)Ó°²— x=L vó'¸5(>­†—;jõòК ;F¸ZòWO•^­†¦Sxh —¸Ò凑̓¶õæ]A-S}!@ÏÊnÏàµn£²£H"Sèÿüj¥º.]; ¤2Òƒ·VnèÖ/qªœ¤uw]6 Y“]¾ù…{é=žKB/o·ÒσO¬¾=Û`Q`ü'y»Ùb“±‹8¦üHò?³EûdÔ®k$ ´RÊ þ ^ücx+Š íÈcTÿÕ‘•{¹òÇtP¼"•NSÚÌ ËÙ(2y|e}T<-5L\Ïù F/ &>þ¼àö\wë$¥Ú.´ó‘‘×Åí»°á&a)nI?9œÄº”¾œÑ¿Îl¤‡lð÷8zGhñÑËè&ýGpά8‘@´4£¤ê§²äWjm ܯñªÙd†¨X±•‡ „7á7X%YF LËÔ—$¶M‘¾_ßtµ“¹NÉ¢´ìX7XSÓ/­Æp/{²$vïQ‡ŽOÑÒïÅËð“:ÛÆþ˜ÌŸŽ.v)»]ê|è×.øa·y홊µyЄZ„ß)%T‘m²ê0h"×?¼·M,$Ša¬xœïíå*îe#^"ޝì¥q¢ÑB ÒlPS½ôIK{Àö9[øÉèÀ,JuLV–ÍøH?= gìlãxR¸séºA€Å‘Ú‰/ÇrágÖŠX–ˆ˜Š IЄÄËà|4^Å€þl=Ýc:GÊ!icG—©³ÒNH â┕ŸxßÑ3èÊ…ªÑw‰û‰ ýЃ3¤“|x”ÐZD{ ˆ´«[Æ]>øÝ,z&ÔÉ bD×3°ŒSQ%tvÿÁ°ðùÉš”òŽGÖ "ù-!åöVìàW±dUˆO’L€p%u2}6 BŒ³’X¼Ú¶_hœ¾ˆ”øËÁνf¼nGFÇu¤&W/åÕrìp§Ò”¬"¨í,'tߟ?%‡Wÿ­ÃE[­ú”‹”û™è½U€ªÍw™U«,ä$ð]y‚PçLŸTæe>°å§+¡]/s·Ë©æUبµé©,qê_g–ðªõZ¶o¯áœ_{!4ÓÔjLx/ÌUÖ aš=›#ï¥ÞÈnƒ>í¸bÎ ¾xCMd~ð!$kd**ãŒg:WQúzDôàã­v—RÝü‡lDç¤S-DqvÆ(2?áýÝØøb;6äB’‡- HET]ˆwèÏö8ʉ˜Y›_[À0ã6=²- O£Ùc!cXO¢¬,þVʰÒ;]X„‹— ,¦Qº.è)~p A-m'Vé ÒÒÃqß'WÄŠ¿¶/lÇc²‰€tynC;ÄOozôqhd9æÁU*ŸAkÏ•U¼Â‡SAx† 8òyµîÌóÒàfŒ’¡Ù„*Ø™!é<Ûcϳª¿¨ýq½(V/DÃ}³•ój2Ùz?­“靯AÅ=à Q 6r)¹âS ïZù7f ccÿ Y"©ðУäÖ@¨ìÔù]û3³¼§ø9Øåòaeï6¦:ˆ¯ÁÎÔ6ªþœ¸«È»¾Í WêÍ;+”ª0ã‘rÑ™ÚâŸÙûS%©å)Q$¸ÁDTƒŒÆÒxeZ D:;]žº3Ñ­Ô—eºá«2-d ¾ž/6r8›ï‹'áüMÓó‰še{IDÈ+ˆÈT_É'“Ì7¡•~^Ð*¼a¥["]ꮈ<3ϦÔJ‹ZÛ”šèDaäÜ„b¬@ë|Á !(Â8.÷¬|“˜‰rH­ª½ed¦ ,~ &Ý…Ø‘ö=ÉkfþS1ñ/aµ;5‘£}hE¨C/‡K¨T@!®±ÄK9Êü¾N?–Eý¥ÊÇ­/Ed äÁá§Ü¢{ÖãÂÄlÔØ«;ý§ ?jú½È–úr£¬^#¼ö‡á؇¯(Û%`OI¸"b²cËjUšÒ£UùAPoÆÐˆƒ]'…ɦN<îë$l28q¶d+jíÅIédE3§úd,1MäU…í¶…G®¸–l`IšWà÷‹„tbÿÖàAwÙÎZ*Å1ˆ…4F ï‚$ßAš|á±’'-Av÷¼ú»P䓞±8´²òÏÞòxq÷À&Dq†ü>I²ÁöµŸêÃà1c7Ó¹«“¢ebí²ÀÄFP?«k^€œIuþþ¼ån”+¸@ÄQi‘Påß°‘àÕ­ Ê/^|dj´¥ïœp©‘þ3?Õnªy‹CN4y/_eÁËŸYb0¬¼ü¶ÇÓc|ù£ò˜¦Gr<¯kRq.(íÁý©L“ºªÊDÄÄð¿?ÌÓYΑ½j>Þ{8§=DqÕn.¡^z^+*{³\…Èee^-Ÿœð@C‰®ÅŠÿ•ñiJ´¥-7oE8E¨¾âMˆê‘u¥-£¢nÀoË+cÎyö¿“ÿî‰v¶¸§˜ÍÅc·‹“<<0ÞÔ˜UÃ,çVþÆý”¦ý§m;:ôlÛ®U.n^Òç¡ÕèIdLÍö–’Äññhgwhë)ñØÿGjƒëUîøÓI@ÿ_(ž“ñ¬C{ìñ“²‚í/ü5Ëü•(TªÄL¥e÷SÍ[ï€Ê7Ÿ•Ù1 ¿±GÔ·6°ÂYE/—áÛ¹ˆÞÅ¢ˆðò•u$ådRG0L6îÛysA’Üíé·BƒƒÖ0änLV0úP¼G& âêœÕ\u‚[•¯}#I³P¯—w çE›}µGKuˆ¾7å~áç+çUº¹-ÈNgtu.ç™1ö²?ÖVövD”·"Oû”·M ÿoc´ÔéY{t›ØF™'ŸÕ7†Î'tù‡ÐªŠþåNkFŽGå܉‰‡°1 ¨¬‘®Ôt—1·g¤/b“ž§Ü øÙ¸¢ŽHÖbÈGä;šR£=£Ã×¾³@ÑJ€£~V= ²B蛉ّԈƒ2Ûˆ˜Ê¥Ð Ó¼©*|³„nÒ”Þ}tÕo-y0mÚËff“µÌª¯ð ÇAÍ«ëllÁMjšÀ¶ÜA6Oõ÷Î8¼–Z/AÐA@ŸøòmÔ„ÅT 6|} ñ­& õÿ½Ü¸êEVgèX‹˜ô>à÷„ú#ŽÙXl‰ãì;PÙ¢åÕL‹àÔ,¦²ÚÄ0’_,yÔ€yLºïF®5ñÊÊõ1V¨{îÐÏŸutžô¨B>Zˆ²‚µYÚiG І½ìK‰•aJQ¾‘ó¿ø¦7Æ8¾xnpÒhp÷½-Nˆ#®+kWª68§+…÷æ¦9¶dQÍ“ø÷Q±YIú3×|ù”èfjœXâÿuÛÄcó/øè5œA^ô\@Ū Ú¦vÝ Fš^°Îãþ”ß_QjHŠü‡ó1‚9yÀøk½³¬¹y؜ѴÀÔTµ?3Nðt@`™˜3ÓKN%#X.’[·!%Az­=æ×n( h ØÿøuQ£[ÍLVØÃðøîY‰œƒ±ëËc Æüø£³–¤à>@AöæìÝqA_ã¢(´0ûБQCVH·Ï裦~V÷|Öúü¤PÀ #OSYÂAAyãRkûj4öf—m­Ñ–´Q ?í ýן‰.ùämïo´=õk”0Ø`õ¨€®³ /¤*;tzH]CcmÓD5~ ŸVõ¥ºP //1zt¼ê89 ‰Áбd³ájìžÒ")´»J?ûÀæ«ÌŒô´—Œ¬ÛÚÊ®l31CÚÝ(*•ú„û‹ùÎE6œsÇüÚ|oÀ‘ÞVüøëä#Ÿ™´f~*n„Ëä¤Kúœê&hᙦ6ÅD7f²SJ–N·Ô¬\ìƒhh„®Ìöß—ã@Æ7N› Šºä!ž6ï9¢Ši€,K‰x_Ü{Ã^gS|0mªÞ,3ÐfõËÛc“8[е}@Á{˜¢ÿw!ú÷EŒ1ˆjê%Ô ´¶•Ïî¶«91jF[¤Š.–&ª6Ì)SÛ”–Â0r¿Q·Ÿ±ømQBœG2µÜ3pìhɼ€#çýÁQ}OÄŒ† Jà¯ÎˆÍg>Ê[3¼kŽ[[ã’¤åÁ$âdSSáéÊP«Åc¤£[{6Â"h3Ìrý–ó¯M6­bN9÷B´ö2¬ £ ÕõvÝhe£:\;޽ü}FSç¦iyÖÛÂww&ö¨6a{«b¶êƒªáhÝhi¤ªSAô\îû7žõëé"õl ãq}Ú¾º·DÑI›ŒF 0¾rªRÛ=÷ã»_K…\qNä*í@U#jnµìÙ2µ'R§“‚ =s^üóÉNXT‘¼éã¹Ê!áßk™|Iç׉mu.³gº 9e·ÝM$GÖ0Â…ëÝ@ñGîâȽ¡}²ê”æÇbàý*y"Ÿq^ߘË8<Ùþ]U“ÜR·?~0|ÕÕÐI‘›é¬ï}í‘ ¡¿T3Zó(Êu@Ð 4’SÑaµºJf—èØÖc:?¿ j QErq{¥ª>fˆCop»'wdƒ^ž$œW’Ai]]ˆS‰vÝñŽn¡·m©¹Le†u(ôÍ™¤´éwmUlÊÅÛL`†žÔ¨/Å’tçøÿpNuplC 7>1íWL’ëÕªÔŒ—êÙ>·XÀ#òyg†!²u•2Ü‹ÝðÏAÚÌn0¿ZaŒý…žgÏ ;`šÞà†xñjºâÆ·Åßõí¯æsùɽý¿|ù™Õ}×ômsVÌ?,ê¨æõ‚‹¥b©a±¶0ŽzJäáÅÿJƒG(9|ᜫkš.é–s ¹͉4Ö(‚ªâñqc|ॼH÷F ›åIžÌQpgÎùª#û<Œj«X#óè%*¨èø9!.ØÚ3iT§«äOº,^ŽÉ¦«õäÍÑ2}êûEiCú/%ï}ÿ°Í=”ä…,TÈï™—ÿ¥NàB?-€ååúÀ¾‘ C)Vå´…ðZí¡¨›å?¦VB1eÙÖæ¥ÑóÈ:÷ê›N7ì©ÛÄïªHÃê6!b¿@=T«©}\™öõ…“®¦ÌA'}@ØvBê›<`³šô¢ ×,úHÛb=Íõ]¦)FnáÔ»*­å‘à—ŒùÁ­d&‰5^­9‡Ûâg«&&”/htÿ#ha“”R6ÐNXZ½!BáTg5¹dr µù^_­(üÁ7h~Cpèy±Žv´’¤!í~(fc±‘s`j½|kÊÆ9ÅhP!`˺¬Ñ Y‘XzÂÔÆ”£ûMÕfz;OÒ 5“;÷ ;Ywhó¦ x –7ª©ïÇ8ïž[‚¢æà+UcJÀRÝõ|쥕_‘V5R$Ly_iî°XŠN_œÌ ~ ÜTÐL+óù{¡«>|Dä°~”/êpát·{ ÁHËJçÖpZŠ{j͸Ýï\ÚÍkœñ¾ÁC$ˆ—œC% ЈT<2À³P§ƒ÷©6‰ª¦ƒô’üHˆ×$i)³ÐiZANðï2ˆcž ÇêÑñŠíà÷ æqM’óµ¨/æ”#LoœËu ¿WN›O@~u ;[G‚…óùo(S$Lˆ[ê/i¾3LõB*ðÙ!<µº«*‹\ów@uЕ¤©æ«&Æžúˆ1¬mqX(Ù'ÔÝsòÊeä÷ ¢$„’a5Ëèü =3a­f%2YÈÇwG(Õ“Ô9"ÿN=^›»ž›±GA:ÐóÀÖ“u½gYDÂâÙó¦hHв )HzXfGÄ —êç Í4V}H­|-dfØ3g‰ö÷ùø§œ¸­«H|6ÁôZ?Gy°¡Þø„ž£eI%Åçq_J0 {Rñº?²_+`t'bóü“µI@­R ±r@±5AöúÔ fhü†ÛÁTßC]™Ô4ÿž]µÓz°¿©“¼7½;M9}C¼ê’€~…™SõÒP“øQ#©"Pî%Ñczç>}ˆ¹~8²K7ËÇíŸz5Ž$Æ©¾ž)}éE–7Ή•ÍIÅ{[’~â6…˜á…è¡Ëv©,´C Œ!§IÉÚ“/¼*Gÿ-æs]oÊxokØãïxqrã^ɾ#ê±ûvN–Ø×fâQ‚’ ˜Tº3Íû›Û[;ù†H/ÝNXë"ðî-8Õ2oÀª[©9fØ8c¿GŠ]S¯—ÿ™éx‘†>»«mêª9Âqâp‹æV”™SvtY«)6“," Ç Ò v’v©}ŒËÒ}Qš(‰m>TLð×ÒŸ¸ŒI~Àè¥\XÈJå*&›4;ݾ~k@ñŸâåu À§Z] ‹ËñÑlØ¡~*2/±>r…kʳª+ÎZ|»ì9q8“º¦a[N·)â±ÕÆ)±Ô‰Ãx–õ)Ê#¨E›/ã¿:»L•}ý.ÛëÈîÊ@š'L<ë‘îðÑnh¶s”+20²öó všöè¿éN„Âv‹;9s/fhówO,ùrÐò§xK\¦röÎ9ÞåEõž'É.™é’€‚Ž.  KR›ÒòH@g×Û»(Kõ~:VIoH°†t •àâm(û‰>N¸"Ò}ó¶ƒÌÑw©Æ[ŒA^² x/‚tª_t^®]Ý»EÎ;§¸wWl]f®F7j®b¡UHAxgvgã´r?ñ¤èdäF‚¬Mÿiôüì)dŒ—W@‚É4$ØÄÒ¸éög¦1dXcj$ùù4ny»kµ¿FÁC•ï œ™Ê•|±î>¡ŸHÂâþüm´Êç ‡L‘`L0åÄ3NOœ7pwˆFay¹P&¨¦ÿñWøqŸÓ„´*¯£¹]gˆcþËLh‚pz øÛ-îÑqw]˜!j‹¶î¶šÐ2UÛ=¶&õ·h‡‰™Ui[!$cdTOD³H[zYkHâOfüÓ¤3ÿˆŸ¬Ó!¡mðHþÞÖ%‰o¥H]§yÓl‹²¹­ciéD#:–À°$íU||Ý^„îþj§¥È>Í,Ø—œ&ä ÃÁ:ÁåP42¸ôÏeSŒÏàq£ø“¦» `'çÕ6(7ÇÄ÷U,Ø»Gè*Nʽ3œ]0†‚€åâa‡ô´Žš¾~n[Œ®W¡Ö7;­ŽµG@¡K-2ù¡B§ƒ½(ì9:ò žž>òm¶Ç_äÞ¬¨Ø5Þ[¬)Ù¤½¦è€7aD&ÒO†–hbøØ,ó‡BmÕÈoƒÇ‡ÐàM×:{¼Œ·¢ªÿQHà‘ûâEIÍFî}Ø òÀ©¥•â ¶æäZÌBÜû'J3­fˆF$«[Åw™ö mp¹Ûágã} í¬×`Ú³(é™1i»ö8µ+kSCº¼ßõoÅOªØF ¾)ô1u²û;Ãoˆ%¡³àï]~o3Ò½€Œ:Ußx§ŠÄ4µC¦ ª’:º[©¨“(Æòt©Rc˜¶Ö¯ÏBbÂô§0((K)¶ñ–éƒæJ©b’(¡¸%SáÃQØ=GåÍó„Qè®ÄÏϨn‚ñ§*³ª¼‰9£8üÈ1м»£§5ÜYò°Fã8M’x{OŠuþÏQÎù½ä_XQ—JV1éêï>÷ñÁWºlÃI“ö31™%b+.ïyvM>ž'´€’¬ªÛ‘rìy1ÞµE“2èb’r¼kÂïh^Ž •}Ý6 äg1²qQgl°Ó•V²Õ·¿ÑòôP¼0·¦ÑÙì@/36ŽYýš–6ÅŽãÓkíg°kU¼ÐE’íKØ’ì£i2ʾÀk/œº(›¿ˆƒbáà¡1o¢GÇ*Êg¯¬O@æR! úFòêšbŠÞŸ¼ë®¶®ðC ÿ[Fwˆâ Ò%ŽšàãÚýÀÖ°‡P¸”ây¯ß#0\k®!E„Ï¡2O“ÊIÛ¡²úzó\èw=CÇX»µŽBìVQTØ^‘8ÞjâwÓ ÝÆ{CÀ â2\ Ræ'Ió©-~åÂG~'ÒƒÔïÑ‚+jÇм8õ$%ÿð[]ŠùdÿŽYKÁ‰B%ö~Ä5È ªÿ±ùÝÉ4ÙìêÖz€òÒ͇…˜Âm!\~ýkè×X„:=º¢+€¦D&³„²‰òEõXzµX$S  y`#Ar¾Ôd‚Ø´ê}‚v:–7_ª^òÔÚÚpð¸ \ÇN³ì¾Q=SA+„ ‚»zº¿9þÈ´œ ÍcWõ6ÀFC]æv_Â:Îr~¶·Iµ¹:4Méu7bã2ù¦¶£²âM±Pw݇é®$¢Ëd¦¤Óèù¹tŒ&$¿ÎhE]»ü TÅà¿ÔŽc²ÆX®Bºg«oÃï=ÌmD"s›_uóÑ–GدäA1Ës»€ 7Q”ÚŸ‘áü =˜j0ƒ¹ì¯òyµûµ`ÍúÝvôgÁMPÁ¤EQ7MßwòžA²9­Œø„0è°ÿñpÎU/WÜ`ÇUž}hФî“ÍÐÑ£ÅAHØSúj¹NßðyZÉ·Q¼…ËyEt\JË Te›ÿ/=vÜýû¥{‚ ¤ ŸH¤“aP0eog{!¢²=fè:`ÀHì0t†š¡³Ò ZD»ü*H—t/ì¿Â®Ñ{bøM.™Ãݱñ‘—ü8£ULOªÅ 2’=Üz;Á1î@\¢"4¢¾I…•íýu_³ÚÆ?Õ½ãq.ë%O)ÝâäÂa÷n»¬!0GÎË•JH‰1:Üì`7ñâwÍ„®C;\ Ç奶ZõG†¦ 3oýŒ¿‚ê2tj¥ß>m´rîJsÀÓŒ4tŠ0 ÈÑÒ8—„•&¯YîM ÕûÛßT üqÁ7GÌ0…ùÞ3,¿˜#—6†•œº¶A1>¨/Ev]αžÕZW¾ó¯O/ÐÎÕ!þÏp„],^mÕ±QyÂ÷3‘×̾Öág€””ÚÜv]È?+ñ”„ô´Fí*›Ìʧim>ß)«£wД׎:¦X–êôú$5ëÓåxbµÉàëô€H‚=:rjÚW1TfÁdë09cÓ–­7ü^1ObsóBf û€ ƇžjîÞœ‡"|‰ÑÏ| ÆŽL~ 5¸„¼X}.—5ß¼o¼5l„…_¦&•9CäÝ•vÊ\ÖäAö\ÀQì" _Ü Í}Û5»J¦qƒ¥(¼LšÇàr)=òþ’39wEu“©—hËÙƒÚ)æ‰Ješú.fñŠóÿãê—kâ_­yc„—úÛ¼j©ŽŠÁ¾^¦,´IÒ‘•èËqŒ ‡œ[M¤¡3ò¢ÎÁ«!<:Gùp$æ½~˜ßú«Œ¥DM>õ ·Åõ ²x¬Lý¿)xµl+; eë VbíÒ3 šå ›¥o¬,ßg€ Ç£Yýü7¶Ï‹Ð`5ì$“÷‹ÙQÌ3h˜m¥n°_vPy@¬v¦Úi½ST™Œïy¤ñ½ê•BÐüf“Â%a©${§gäÁÈDîâj¬Î ô_iÆDmåíàuQè–f§¦ÅÊ…áñ}ä¿]¥(ocðíØ}$ÇîÅyù©ÏU‘ž¸ª@Ô5…±+êue†KÒÑŸdâˆàš":õ9¯v ùû4ò³%1±¡NÓ¢øu¥Ž‰]ßPÊKd®+¸güÂÙÅîóÚâ)ï¬CÚ‡ÎØc×€p”-6Â'I¹IOÙ )jG+yD«NpMdE–F¦Þ ø:Ž‚Q:ÇËw-LjÜk÷ƒ¿³Ÿqõ”éR– G~Y¬ðeR]òeµ-„r ¦0‰ öÎ…í 7&ïÒK šFy>˜À½ï&uà+³w3È?lÚPÉŸ¤7vǻٶ—b“ð{Ÿ*MÕØŽìøkÔÍ*}âí¦IÎ#ä’0©š¸eä‹$ê\ºI_ÖN¯4\Éwæ+¼Ù‰kØcÔ“jÛ÷ůÐz-ƒåò{`~ÕÈ@!mL†×ߺ„2&Š¡È"¹(IÐ5j«Â*çHêu1`@CÿûùäÔxÝyJööRC®Bj;^“¶S\ÞYÁ¤ÿÍü¾ä–4Ú¹„Ï÷úíݶ¶v±A±Žuñ!ØÅÍ øI‚¦õoÁ=çeí‡Ìµ\»âaÑŸ¬+s¸ø`œÌ…iCíõu€XœwóvÒsôê»<¢É/f^’ˆI½Kyüçcl«sãÇ×#§’ôýð´ç´Ì”ôQg9êA-»Ö7H¼ªÿ™ŽBmgò¤êÖ]r–ÆÝµáãØNy/T\½†Q Q$6Ž>mTM‘rt—jùÈàHc3ôˆàð~%Á8O–kh*\¶£Âú´aÞ´Bè#ñ4Gå”Íß‹Z)aqÙƒË\Mïê߆ÙjFAG`bv&XÂGÝtkLæ"N:F$I4ßQB­† `6~lbÍ€QÚêP7ä^hÕ.ID>-³wž•/k(ñ&aPDÕ1Óú7½Ž[±"œ ·EÖ_©½çÑúÔ3¥Àf,VcéϹªÄ‹ …\b_árˆN½ÇÏX§Z‹)³…¹UÝ ÂGl@¬‚ÉòΚηÆ~€2è®]å¸!MÜw¼¬vV@¯‘òYNJ{î¦Ý¡Çjлôï½¶Yðê«©@ë'N#‚Ð’·Xk1rIšGžƒ<Áa,Ú7¶®êÆy+#YŽ:Ûw?)ÂC%³ìë7fú!ãÆA­¢ÙóUߦéñþå€ Õ*뻼¦€™ϩ޶YÄE·H¥ÒŠsȉh1/óè/áéq )+í´¢×vlÜÜɧ«Ò’U<Ç9Q;G1°ZĪq}6ê3¨óÞíÀŒ™-FV¸òºšâ…ä)t,š¢”+JUËnnÖ¶Z!#*«’«çdRÌñØ…ú·îñÄ('órD.Ø:#ÿGàÍܘ+ß*7)'Èï øÈä¯QUš &$¢Oª¢ã`+‚*Ú•§•]hB-§=4Àþ'ˆ8Ù ÇË)¸CÙ,u»Ÿ+ °ÊêÙ]=ðÌ©±Ü¤pùhH€¶\áh:¯>HtS‘7Ï”·óþiàf¬K%æ°*‘Ùz¹lSÍÅh–…Øû>vßp WÑï3@1iˆ‚º=dì-ÂRÖT×ןkD®õ.£´S‰Ó":=¨ë¯,P¹#§¸eê®8RrWã­~eÇÄÒ–ª1Ý© tyáŒqø’4mŒ`¢†þa—ŒsYqc¨Wƒ”¿Ì®ï×a¦Žn²£l#KÖÇ)"€¶y9Ž>ð£Ú}nqm =G–ˆhÍ~•Ò"Mu;¿jXâøU¥Í¼ûy4—R£Aøô-7Ä-,}‰Dþ2;©oŠ„2ôit‘‡Ôƒ¬?ÂÌ0Õî›÷£RÀÃËÛ †5 3D¼€TI¡tnÉÊ< ƒ*"‚xÑ Ÿ‡Z×à}¸PùgíNxÆÚ§8´ò†F#&óƒR™&S=2ß¾_ï<Ï0÷hZß×½ñz}¥„áa˜×(ݶŠ¥êKôí!¨äT³b1TôWj”ò\&è„qØŒe?´¬ÙØ{È&<3Û®›ÐZøfÁñÀH`ù(‘,×€{”Éwñ+.¢&~¡5ŸîÊ}ÂmM\$%©ßʈ½rªÄêX$[ûÆr>úü8Þ•x‰PªnÎmz^çá/½¥*C˜ú'•Eõ‹J÷Vë±<À¸äj¥G âlZ(|¹Žób”ɱñEnù©%I¤ÛÌê!¥íéAÒæäî$îm55êàK¢ó°0Œë¯ÿ¾U2ç…RÂÉ_×DD戴 °–†7“F/pn†&øÐi×JŒ.¾1BGkpäfÕ)§Eq$ˆ^§VóÂ+§©#Œ ÞëÓÅ3d½AWiDˆÛb`TàŠCªx‰‹9 ¢ÊDhKÀdHïݼ‡uXë¨H,\€°*A‡W£ÌÜ˵b¨ÉÇQð:¯9–¸l|ÛMè¹`·3Þ’¦]ZFëpz÷Q_¼·%¬§TÒñé9Ñ<0zÿÿ[^ÂÝ_èïÄJ¯VÞa}dÈB?yç‘ùöú5÷}ˤXVJƒ+ÑöËÛNœ/ <PE~ÔÝïpã¥@´(‘ãÞýÊø0q=\òZÍ[”’´²ÍÅ"ÞP)³æö›¼!cφžð¿í‡ó7pzyj"3©=Ô„¨ru,‘%“!Ù•RZ)"^@c»ž•”ceQ¥xDŽÂ,PÍ8!§4§roƒ®0®éMe¤ZÄD :xh›{E7*…›„ÝêIÝÙÀ€&_<5üÊC~„O¥ïLÁ‹øjøý³Û¡CÉ¡µÁí²±Ä”{–Bã”Ð%ú»ãë¸] ¨rZ°é~n$t%gˆôø’ó½ˆËu¡rù‘âœÁ,ñ‹šU¢ö=š˜vú]³W»'ˆÍ¹9ÒT•ò<¢fG'ÌD*²cÎHå.ûÆÄݤä4…ÿ´_IêNâH°öIp÷ãÁJ ÛmeÏjðb>LˆTÉ4Ó<Ûå„Tö$þ$¨NЏ¸ªá…4–n’ã-bM‰Ìb³ÝÐ;MX¶å¨Íó¹…Þ¹KȾő];t¯˜‰a'œ¬|J~FZÕC¡+rA´ÜHx¯óMÒ@ÅZÕУm<1O>è8²ô9ŸÃ˜…\Z­ðüËô¿õËσþ^±´ÚRñëígÞÂ{ë\&©w¼=¹©œÞ¥ç²‹/¦ #Îϩќ`ø¥Èf4f8;Шj¿Õ\E3ïm±ð)­ñ\K† i¹E8»nxéuî•'õy…8·ˆ¦ scf d6§Íëó«JR´?â¾42æ n1¦È€Oõ\Zü@t€’·Éã8‹BÇñ9YZZºØbµ}ãÖÌo*û½ØF&<9®¯…°ž›¡Zü8Ãì˜ì¤ZE ÜY ‚/˜x–ÖªÛuª•æûàkíƒ4z»ÑjD^Y.? @¢à#žè'´¼­l{ÏI¸3’ôI¢ÊjXŸµ®! 4’"A»Â@ÉœÝËø¯µõ?¨Ž@꽞|3%½'-k²å@`ÿ’-\úS>ÖÁ5-¥}ú4ÿ‘ \MOðYNÃݹ\ÿp 9®@tFgÖÁT.fÔá&%rNº<+Âwݤ_î ¢‰“cOØ—oë §DƒáìxY—`PÑóÍ~ÂèìÝ~"€HªÏn˜½>ËU޾âOÒÚ+¸y!ÏVK)wôY ¹ÄË®¸xzW‰b³nïõ³8ÍN0úgcÌkþld,§\Ÿt‹ÕÏß@‹Û¦º­÷´Í ¤éç~¨bRôuÑö4ÃýäFuÝE.Ã×… /ÚFKë ?ã/ʤè<0ìPcÒ7Ì«S™,™@ÕÅRÓw ~L S7ê±`æ¸3†ñ^ŒIôTNîþ ÝŽk<{ò-S‘Ï™¾‚G·€GÙ§×䂸pó_yJ­õßnëå}f™ 3´ÎÑ«í©ç(@”¯6ÒUþér tÉ·Ýò´Sñ‰ü\n¸“TmðVb²{ÙqÐxßËKŒ /Ñ ä+oÁ‡l’»øècQWî¹IµynKÛ[0Mëïÿ{c;©‡\ßZdjNÙŒ@¸éîX°!M)u0Íg6ýP^ýN¤ ekIü3ÖXf=ÏrwØê¾ûUƒ;3òª0?«’}6NdwF‰·••1éme7 , îò;œ²ßª£´m‹Ð³y&Ùp?×ü½¯Eì ˆk»V¾!Á|óuØ[ãS•Œ·Øm™oã#dËuþ¢ŒÎ?2á™´ÔÑ•Wr¡‚uÒzuÁ0ïÎËRÊ_4h`,Õ©Á²Sô›f"Å&> €#$ÎsžfV,%š0";ík'>ø"¤:ðҥȯǸt)TI½¡Qð遲î²i£|hMÛ.á´T{vKÑ^^ £¿@¿Nçíñ©ø†÷0•;g*¼¿·,ATL.¹t¢) r\+Ù²ÙÄE“¡Fi—6ÉN¥\ÀcÅqÓ ¹•«>龇Ø4Ie[4­,š-:˜R¥¤>r£¶@/ÙòókÕg_óAe†´ ªFb:LZÎþTù<äºÅ;1T;¼¾TÀpá=R‘JØà’£Àn {¯³Ä›qgøoµ”qXRÂq/ʘîRþíœ(#ÚƱÉ;j "°j¹CEºûÏ,JÅrÏMx=ü|B›2ãå½÷î êëûF‰Z‰a¼¦¾Ç$HðhŽKÈ´„D"á|ó©#ª^Ÿ ìuîo732`/å;:›Ûó/ê½Æ¾ý¤(ø®ímvO•ÕC"4¶ °Ub{ ½˜Ð¥ã4‘y‹ó®$Í\RÈ*Qš[ykiEðÈ Ø‚¤nTË—Ð!v²¨P£Î>Z#£g?÷ý"¨}bqj–²tN|U&y¤zÂ!.Tܳrm²Æ¦v.r ¸äÁ¢G†JÂóùnz#Ì j£,Û*ÓŠº íKÛl *bl„æà€C¦m~;VN^±ÃZìs¦óK¢»„—k8qÇvXÓaa:B7:Ù’.‡€¬™¯Œži²^¢šü7`Э3SnÓ¶ú©Œ%84„à(À.ƒ%/t´*jä†hÈéQ3¦åR¸8pü/ÿuFƒžßϽ1ä 6‘dì„:ßùœ°i¥&ª×ç5wl#,¡6Åt7R> *_˜óÞ0t<»µ!ôiʳŒK<]ձ˔’$ô•và MÃŽo´Ï}*ã]½8¸M„œ†ì™qH\òæËìm‹ŠqJgÓgÚU[^.ïu½ä/òxµ,! ŸWCÓü‘Ý×Gó © S;$ϰYâ9UM¯è!£<ßer€ã¼XË5!Ý©GZ-äVµ‘ÖPŒBIÞ¥µd”õQøpaÊ—c˜E”á—º®Ã6Zlí^¨;•éõž´ôÂvíù5‰¯ÍÒÝ™i݈hVï¹KÊsºà}0Q„Mà ãû ‚2t50;Ý6zo‡\K·]±Bâ»j…ÆF.ª£ˆìÖÈ¥„ÑT†Š]½¥Çä^Çâ_"J]Õ*ìëL%äM ÝìÒš>Ø#ÿöZa²ÇmŸŒSMäcÇeìî{ž_Ÿ,À…èÆ;éoÜÛ“ãæúžmn;>u\£€Ïâ,ÙÆ˜×e”–-ã,ÝCà;„k•Ò,åCE2 õ sm7 \C°½Þi®˜þðçÞ{“÷iÒª±ì®n‹cWG¼¿ëÅôËÛª–TjkÅŒOЂZU?‹õfœïÚ—ŠãdQÌŸàrÒz4B«R`¶ܤýò19]5éò·Ül$l–+è׆PŠ&%#<º”¸Yÿ&ÄÍþâådˆ˜Yù_°Gц™*ŒÊjÂÒ€f p:›yÓûe$/&»cGCa”(¼„kØÒÐÀvgWŒV†a–îñÄQ^»EcŒ°/F'‰¦ö®“šžÇŠÑÓÒ•Ê!6üladXŽ'š ëÔðI¤ð¶Gø!‰µ‘þZÈ݆+‡'3²“mлb4·³0ßR¥Ú å³ø¥Ê@»t\Ëßè_Cru©J*G¢lÓ&½JÝ„s¯Nh7Á\]÷Xè`S›`»”>ÓÎßö²FÖÖOM©?ëÞí£AñžQiÓy ó7Ý˸({¯׋äZ¥wËWz¢‰J-„œ’• q‘WGŒŽ-¶[‰F¦ç+÷œ)ÓºÔ9f!¸U¾°•Gm¹Ç:èÝ·%I§ÄŒg„–à‘¶Ý—Å×M÷5¥’"lXCÅ6FÆÅ$¦¹’³ò„Ì,êNŒ×7G4ô½ºó©«&ÜW—ß\vë0]rwû£Xqóñø·á|˜0TØç×ü1|o.gaZñB€®·àyGÝSºO†4…‰9~Ôód™éth ›"ÞBÆUúai3(ºžˆ•‹ªK'3€T{z“rh´¦ù |``!’]wÇîŠq-ý;Ön–twý‰$7Œ‘kØÉY5[n‰yÜí…}¬ÜÝ0É-9T[|¦—žÏôî±Iãîý#Ü6ÇWW‡›Ed¶¬J•‰ wÚå¹¥\ǃ.'Q¢*¨oÂ_6êyèž“xÊh›éýR½R»xשå­ý—ècMûqÿjLútþ`› öTkÝõ„ŠÈú 8ª›Ü{ {' TMMòÎŒÔÿ°S(Œú稱øx;,þ6H¨ë”kðÌÊ^sÂöZ•k&gŽYÛÚæ'ÿ¼äŸ¤o„,O£›X²{²©EþýæBƉv»;ºÖ¼—&"î}9=݆¹wœŸÚP€dgŠ}hçBƒ€¸ *Œè……G}둞!7 Àt¾Xz®ú`ºé>ôñ˜>Îc>“ƒ-íŒ&î‘#eøèm„Qr¹2зzkˆ؆s31#Ä|.FÒ‡G§_3.D–~¹.Wõ%­5.Úþeiï Ÿ¯ŠâìQ¸¿ÙUS={òcz+#¼a6_äçÿVŒŽÕî£bB^ì›¶Þ‰¯ÅlE~nL'B·=yѨî6äê5R‡v}¶+\=¼óS¥¸êy5?±kU²*ÖóŒ¢î¤áIØ8ëŽçõ€¸Ômº{™èFg8POö6 نعw‰Põ^ady‰ÙJ’ú¥øEG¯&ç‡%/çm°Êv›š€PãžSÓCí5™Jéx™ q’ö'FBÌõ²ëÉß8ª_ Úæ9ñœÖIP‰s™ª¾oÊÀÍm=¸Ïþ¹™zAømöЕ¶! ²þHM ;U}ði0´@í!™+ÄgFî~i‘ì¿:[Ýân¼”’à!Ò*bô8–ß®Îà©(jË€³<¥{³ŒßWG¸í*Ë,iÿæ;0|VFùTÁ#”ã5Î#ꄌ[ô!Áh1=óÚºi¡I7Ê*@¹­Î' $¤(o4ˆè+ò€aú ¹ægdÑ"JPº •ÊñÃ"l².âj‹ùklµ»T¹¿½¿ Ôðõ”Ò|á${Ùìø$†Å!SõTjw „¨tÉDy[ôÿïÖCÈK |fÓ…_ˆ[|ÜØ²«Îß%«jåµ j7A"; i䬨Z`IÒ)Òu¸ëÜ ZRb¨ŠP½KдùøúÜzdëÛÐ¥œcãz®Ôè­°ã°r·˜º;äÍy0Öáî€UŸ]»¼!(”q“Yï1×T†rÖ¹Þ _8M’{úJÝQ9„¦§ö™ÂlW:Yï{¡{G1—+¨ (Ñȹ%¯hm MÀ÷+R¼î*!5~©ÞHg»pQ¢u§&Hn¯Ö¡÷W¹É0ÕÁµ÷ n R[r·Ú:š+÷‰Žù*jº¬6G¡Âþ¨(5ÇYÈ-*P•ùauÆðí7†7†¼K™Ä°Íçæ¾BzÜýíâÃÜBd»¾'œAOg'ø¹“ÅÛž®ÂÅz#LÜx·ÑÍ,ßîÁ¦ï£©úâ á%Zp[Æ$é+_j²g]CJsãa¯Ÿ p9ÔéGò™=w@¦âõIêxÀ=ž¨L~a±t囃æ16€†Š¶ŽˆzJüâÐû„àâíÚ­ÁÿíŸVÀ=‰ÁÖ”÷d§Ù.ÆÇ¤“§Z¥wÉUT{ÑäÅ?ªÑK šËËá.6ÇyØ êj$³ÆÆ_°&&atíÊCU¸{™Ìµ?u7M !`AB&Ðf±Ð R›`eDçÅ-x¤\çÐöZoZÑ£Hpÿ<m휻­ÎŒæÜ»œ7òö4B)¬ÄÀ|•`'wð´Q¯aüï¾ÍÀÁžhƒEkc°¡Ó3o•lß1h&]Ç*¾Ün.<Ÿ`®ÅÞ’~û$kÝÍ„¡¤aQ2¿ 7LÆKÒýËTK€êv׋ótožvtm¹;¹½ºrœšëYŠÄ½2§¯‰£?™*vNp}ÛÈÉU:´@ŸXwÂØšT©gåç— Ñ„¬K‹°Ÿ‚ªŽ©—HJf«ÐÍsÀÉ“H•’ Z! D$×ë,ÅðhÜ£¹n*l »Ùc±Ÿ­²Ñ÷R^î]|§Aÿ‡O¾Ã:FÌnqì2¹*Ø´Ýô“Uwª©þô tk—ž;dÚ©<¿’ý¨ï›¼½¸c»SŽ@ ªšÝb„&׿ÿoµµä?\€“‰¼{â ¶2ª7© bƒ—jsé‰FygþM¬XƒB;”‹P¦œÆ‡¼kĈ˜O<$ mÑ9ˆ›Êã-xf#ª™-®QÚ½C™5õª'°ºûB_ú¶mE\’Û Àå,¸Kúc‹º/Öz}[ ‡ÊíÆG wÝÖ¦ï_ÑS|ÐB_‘EAÆ>‘Ÿ„.c#Å<”noEâ$3èAZ$ŒÑ@‡U©<‚£Ì1¢‡_©±ðžP{\vÍnx&k‡£;ùa5OÑ¿~^åJgShWI›@S´6¯/V×·ßI^<ˆëó×±»¢bA±x0¼ËŒèꛦàeÁéØÔƒk œ†=—&ƒPì‹©,¨(0øÕ‘·µöÞ‰DCa"5¤ gѨËò0ó¡žê:!K‘Q äà¯_}ýv­÷‹æ¢hœ7nꘙºd~ð&6»H5½5âú¼K¨Â'*jK ÏEM*mÐõ¥Á˜l› å›/€R(ÜŸ{)ðçÙBX ’Îi€bÜÖ ß$®ú!ÊúÙ/u%·ë’Ø!9ËàðUPŠŸÂ éŠ}}ÈV¶ág [ôvv¼ÚZÓYdí{z9X1o‡8Å|<›§Hù»ªh@€qûV1W[~îVÝ[n̸R¼»Ö&ÓäLÙ÷Ó–‰‹ÞgÊD:™P3R¸mµêâÅa¶j’l_m§ÁÌ•P¾ýq=çðµxÌ¡ ‰ ¿Gà?ÆK5£¶§õ ÙÐ/"y O ÚS µ í„94‚š,ý_YâY[¢“æƒ[»V^¿É`æSƒáÕa<#ÝfxŠM={“4 †®Ì—Û&‹­å9Ð6êlüßx2Àß©¤Ã?¢ê5î,–l?/+5"{¥=ƾ)µ2W@•Ó@)zwìBBÉd¢FüF„¹fêÛ0©FgoœŸÈúÿÞÎUfÄ/p^{_ 9XwTÒñ² ÙÐÒ™9¶©ýNØÈd˜Ÿ"¨¦ûnö6éú ŒÊuókg-=]ˆŒúˆà«9°†6¼þÔAtÚØÊYgãDôG -уØB°Ùmdq¦Hà•Y’ípeØþàDªáºN÷Iç‰G9Š@¾aƒEºgÿ{WÐûŽkü£¦ÐÅ7'M×>Ç㟼1^£û¦J§“LY…Öàh{˜ŒÙ§¤û”¼W)3S@üœ|:¿&ß(Hog€ü/:Ì×ÛíÍ{¼ñ<^7‘Y+ÎY•k^_žj1Žqh°“ÊŒ¢GïÚø¾:¹^’ÙgãF¶ØÏïèr ÇÌ¥h41ß@÷‰Ñ*Ћ–fÕÔt¬fGØÈ"„’†î޲\ªë>†Zz#ë´Íwæ÷‡¨ZÊHCmÁÙ2.: DšÕÑ`gxƒPâA°¡—™ª’W8Ä«:h)z´nÌwÐÖíü(¬ÓüB6´2b¤¡ Õin# A³mÿ1âî² .H„.Áç?ŠÕÚ…µòcE‘Œ‡“óõCÕ>xÖ|QD^QÑ0^ì Ì'Z:8mÒX ÐjæÌï>Û×±žRÍ20ækQlO!+Çs2l#£…Èëv½.2ýøôÞªU@¨Èû»{\,‰ÜóíüÍMëXq¿ãØßÔÁ_ðŽp{à¾R¢a[¶\…¢` l9Ãç± f^o ý†þägªaÉ´ˆ"Þ-,NcÙßÉ´·Þœ.ärf.½‰þÂÍ–¾dt‡CT’k|ž*ƒíOXÝY²e÷ž–i»6Wõ#2+ƒ¯‡¬ð‡Ð>ñ¡Vpi{’èÝÈØë7êâMìÉ‚aœ®qÊc½tQÁŸX³[ÂE£ÙXX1;27´2SJºÙ”mó™àþ\$9V+N7ƒ Œ>»Šô{Kîß½*„¢£‰¸†ÍçF¸¿(£­ÓŽRiÒ¤¬›«oËörÌ5¦"¹è‚Ô@ÊãŒÓ‰¢œV·^›QÐ4š3b²ÈaÊÕbô‰÷¹ê”H~þÀÄ)åãÙÄà„ÅÎOGN«,ðs¥Hù»Zîп‰*ªÄ?(HÜôꪅòB\[»~›ûõœ¿%Õf‡c‘wcC‡|ò÷i?û!º·µÓÜø‚Y.Îö6˜‡šÂ3»D݃D¿G®¨‡=Ò웰˨>úGÛ|„êxñÿ” *ØÀpük#„CØv{Â!¶X‡ëž¾¢—šþFÍÑŒc¸Ži°ä4…fÒÐ Ý–V B,4í¡8¬,±ìF½J‚Ï0ÚC¥–½|rp/D­tÚ1”£Ä£ï =E˜°ªÀÆoöí˜*žK˜zš]†PÓìcð@Ç£oÄx>@#ÛÒÑa'bO¶Fíç"Ù›f‹Üî6¬Àㇾ(e•è÷¢X_¾`~Ò7aôµk`»_Ì»Ö e.ÕÚ»µj,½oiZ…'Køá`Gù¡O‹¹à9óR€ç`EÅ"ÇòÌUöÉó=dW·4‚¢Ä‘0¿5“PF/C#CW?7¾úWWN)íf¶±z­ÅKÈ!€{ã'tC‚ù€0dÜ€7qÕì…ok0¸5 #Fd7ŒÝÑ÷ÑåÜÇvÍpRNµ·{µ=]:A§‰ÖJhäìûâÏ„8PþõG=…•€,Þœ¤qU'à˜âYé~³@˜Õ_×·¸ ?Àvð*úQÞÇØ‘ÁÔìf†&™ÛLÑ£€³¯)€U¹ÝÀúÞÆâ  ¿É†QÝrk€jb ÚB”ül©t3`ÇYh,ê‹ÝÒòîú·ƒ¹˜KaÂö«—‘ˆ~ý6&|úŸãüÐʶdÃÏ ŒÝVGŸ¹$ãjÓQJÖ<0(ÊvjŸ>Òçyb]¼qé½—0î#R¥ÄžþYIkÎDCã q7hZ·ÚÞ&VRT]H‡ÝÊ!2·S|¬$|Táþ<ÛÝ¡1Yœy`¼®¤ÃØ›b õånøžTð´mPb¦ÞXòu¢š#ìM°JÕ{™¤_ÂúÊÚp¹D]mZØ9ã]-ò0?ÇSÁ΋f(o-ÀAè1${¦5æž½a:BhŽâ¼I]À8ÃjÛ½Úw>Sí 9~~hýe h ¾€ÔpD 9È ‚“A¡X¼V…òÌï²ì9’Ÿ¯FW®yMšÞ Wþ冢ƒ*I”–MæGðÀ*&ìùa¡ù¿^Íã½¥ÂlZKÙÞWr‘%,0ü^²ày-Hú+€,ø8QÜéȾÞL¸áR-ö‹KÈ^èCÅ5¬7G•,P™F”sŸ‡ -2L%Ïe0çûh}Ô÷#M׺ý…§‹‹ÂÖ¶?³%bZJ ßK#œ]|ÃÌNéë=•˜(pA ú[î0zFÜ”6*ܹûë‹|ü—w6#ÀÚCž$gôšÅöS´“ØŸ½‡Æð®qÄþGx™Æ ØŠž4g)t° ë^ˆ¦Ê|jŸë3¶ößckn¾w@(ò†p1,>iµÔRØbÀÏLº_1Å ªD¬Î¶äðñ ™Ô¹ä‰¿hݤ›µà0³}­VÊjRF`ƒÉ€Ÿ+y=7¤¡ x ò? qYÛ-­Lš¿(=¥S$+hóÅðâtJûx²oN*ïöqmuº õDT`o=ítˆaÃϵLS>黫ÜÑúl?àž‹ë9Ïã[¯Õ_vç¹\c™¬0ЯvŠ)ã#M*?át¿t"âg¥Þ)u~ïäoðjûv æªÖ*ñ’D^’A½0ý®ª=ÎmŽ9ƒ=|Eç#aÊ–“zÿPœ/‘Û`OŠªéX˜!/ã·¿ ñUjÁØÐ‰`Aö“Óz£R¸À"é&£3‚ë'>†7•t$~)ݬœE£·>bÿšŒX J›Y¹cK§X\\ɺÅîœRŸÙ¡˜&äh˜%äéåñt90È1¤€ÝÙ¹sꢰq% ûèZÐ"-‘!0²º`¬¸6*èô“…¿bòØÏ«>|z0Nc  ÍÖtHÆaSƒ‘Õ4€æÖ7TƒbÁh~–n(må’Þcv­Ûw¤å¢W|«§Ÿ¨Éë¾Ù2ô!+¼‰ÂÑ¢ù{úîä䇨7žƒöÕ=ênt«¹×•~žà{¾Oq‡ Ë-ߪ÷½"|8,:ìw^ R6”gý$\Š6„658"l­$V¶ã"½OÕ4²ûä3× >Nâ–C…S~Ǧ¦¿ßÅSIh«÷Û-¨tÿ"›a”'ÔÃ~L_zÜ h‹»S™`懛ÜnÚí3 >γ,[³!riý( f•ö¸³³¬âFWð[}jT7£êy21áèøo5l ìV•v†tÎh4èÉP‡=w•¤Z®Uå¼c!ì YX¼ì°Ë¿\Ë‹ÄbØŒ13ðN;Ÿà¸âyºkã\Þ¤e(Û‰ÅS•þBÏÇ:XÔ¸PUŒ?ËaʈAâ¼&¡Ð Šƒ':/(3Lžx=PmIßD ~äïµïö|P‚Lë”w{éâ¨EVË¥†ÂÉ#&‘ÈJ1OÎQ²?£t>:Õt‡‚ðºæÞçkúÄë‰Ð$ã&¼.÷!‰Ù¢ ûó…êÄΦóel5j~J#›•ˆ[&G¬_Rœæ Ýs¸F»¹fá;XÖ aÆç”@uI"Þøœèz÷YH1îŽ_¼ùƒJÚr•íDíĵàï¬ÈYŸ‘'O;å0‚¬ NœXúËR¢u¨ZÝ |)¢ªFG!;4EgÀü$>m¯:€Zû”­I!Û–)±iSÀyª4#nË#ÿ&“Þ´ÐmæÒ¡»Å OPæáøÑ=uA~¨H!2ô»`›â.LµÒ‰ÍuïÔSêð¢ÿ…fžÙ²¼m1¦Šu”j`»wßn×ç]bý›+Ve‘G«súP|Ë“Ø`«fÕyvl݃€‘ÂO¥giØßA­¿OûœêkõÕ:Cb8¨¶CGr¬‡Ÿ=T* I"èºX-ŸO°@wõæ=†Î tA‚JäÏ"ŒKQ'@ÔC4;pl!²x{¬Áñ5°/À‡ÜLç@ާ!é¥Ç`†–$@ÍÑ‘ûÑnÔÕÖ!U†DëPN+ÈšJ Ú¨«>ñìm®ÏïÄ»R†sÊ}¸MÛ?Û½·z*Fœ5ШõFCKÔ >j¨wt@Ÿ7Y §Ž,•?5·tQ :o‘'¢#W}û«X!a½·\–ó„4;°/•oõªØÜÿŽ‚A695MH š‹ñ*«ä¿N“}œ›v˜:$-ô’?ÇþÂÇÇ×IM5‚z§ëÕÿG§kckQÏ0k‘´ÅÐÛ¬žŠGÁþ.2k[—Pè@ÓÌÉÆà 3»mÓ‹¸€~€û84Y-¼ÍÙŽ½·¨k›—ºIU¹/U)ÌÚ ¦J.KÓááÃ1bÂùD3P: «±Ñjš&Òê~r+mšŠ™†«§äZ4å9þ®”xQnŽÁÞ‹·DÞjC”¥Â~`j8;xZÅ' ¦°eìm/*Õ:Â4¹Tî¬ÆòŠ«~…Aq2íЦ¾1Ô'k3êë«|mäÿ£ç%PmÒ)ÚÃWjPÂÃÝr2ÐåÅßõ° x†é¢<*£!­«ü¶]86ç[+¬°ÂݽZŸAº·Ô³VyifæOlA—ʉNÖá¡ß—ñS™ËEÏí1ÂÌ|„Ó§àèñW±Ù¯@°eѶb˜[ÆI^_–<™ µéhš”ØãÚ ¼ºÖ7/• ž.4ïòÉ8Ìî[0½*¹ ¤e:û2èÓ»UŠ¡xªÈ¾ìT¾ÍÌšX/4w#±9 Jë|øŒ¥5ŒlŸ/άÈi”¹U·PíBÆN6Ÿøðì%'¤s^‚ÊM q,l£þ¨}¯O808•L1? I¦mØPÄãó Ëø«1”Ô2ûÒ±>>ÍA¿7ñ.Câc1`|^0àûÓ*‹ý4©–£“é[£±ÚjDS›BC½7%ºÿÉZ[s°ÕÀýíéøü-oˆa,ZÿŽ”±«=Õ4æô) 2¶ï¥¦}Ö,(vÚÿ­„z:÷¦«TaÉ<™kúnÏ4ahKSBK}ƒÖÕ-i±‰u4·ÛÙ6¦¯‡=׸žPRÒ€{CÊmxÞÉ+ÆqÇSêÿš°J†®ÌÅZDê[Ü“hͤ¿ºaÂfPl`‚Ý{®°¸oZ„Õ~ì@$ŒˆÓ¸Ì÷¶kŸÀõøf°¬în*GÂeÞËs¨·ˆ1µaÖ_JEãÑw·É?ßüد­EÔÒ>öå.`y´±TS€ñÃÂ,ÝaIí‰W5™ˆ›Ý¶Ä£Å`L¸Š`)™’!_¥æ±,ˆM‡7%`à‘ðÑþ;Ÿo*ةѸúF¤‘©¿‰à(³“:ÒŸŠFªV,˜üHï%[ò”ë«y˜ª²ë´/ä)øÛIå£oœ›ÕóP Ào–î3î%ŸQ½饄,ì^7_þ ¡eû?LUU߬·¯‘Öú$o•’2,ÈôìÇ(·TŽwä0 Ìí4ÓÚîò”!¤pÛvãµ7©3¬vG¡¶3Fï¿7 y¨–º\ „mö7R?¼úC¢iÒCFüܾAGô}È*©¶¢ ›”£†ødq%K}-»FÊ&ð,ÉOjåô®Cm!rxsQœŽ^$ßç’”˜Ðæ[ý5 —Ô¿Ò`¯•ÕÕS§>ÁÂV›’ý¦b¾"ã—Fk,¨=Û36œ»ºè\£ÀOÜC$÷^e³bm¬z žiÕ¾¤÷#Mî*UŽû?ûÁ‚Sa#ÖGÄÆÚd¿í‹¿2AO¡”6sñÐDrÊØÜý0Þ[ˆN4…ùð¤Ïëˆé.ƒ¯‡f. Èòm-—^†Xsùxi¡!ÕþÜ7z‚ôʃ¿¦ÉHãîÙ†,7·Rƒ4î{‰õ”²ôÄÙ’1ˡȾ–ónº”Hß,K_BèLçùH~ ­oÁšEæ¿1ö9zàKÕÓЛñbeȘøÀKN‡Zäe‡™á;õR$ßúL‡]#ƒÜ;Õn®„M4»8MÉ7®á›B3õ$7S­âЬˆ½/‘0ÒÊïœàžýŠ ©eåp½e( ¡ –Ì«Yƒå—“Ò{éJ3.–X%DŸ”£(‡ŽB¥§Ö;}¢Í¡‚U .&Mç®w+'@bärYœÐLïí À ö‡Æqfqã ëƒA†4ÈQ éo;Ð:¹ 4ÆÈXÌÔ®æ'Š;,fµ¸>^²è1h Óª³c¦µ¥,Ûë:EVui• ï‹FqFN_„/ôÐÁ™¼0¿ÍR¾g‚ òØ5—O?c{ÐNjK)…~,°Ö’Vt®eqÊ‹ýgû]­¥zÌot—o¨‰‚ø™ôÌÃ|-ÕmKZbxoçñû71%h2¤3â?÷³0ëCAæRŽž¶õ°èåt¯WÓ TÛéHàë^„€Ó¼¸]`Çßž9ZqÔä¬ÕˆZUêkZ­4OÎo­dݨãZçšÛ+ñ2sMSÁZiîÒ$æaÃqþZ¡¾¶1Îfã?ac Êô4Þõš×÷›Ú·‚ýt-ù,èJK¸éoH¡‡ÅÜç ¬w¬¶!sÇÓŽ1½cæ—Àn |ˆ‡Be`N}„&Å”€3ÐÂ2ЧÛœi$¥›8Œ …‚t,H—›ÅÌ"Dèdk¿èå…BFšXÙ!é#oG4OUoFG•¾Û´ÿNe–OìÎÕ•ëÁxu’°-:êÝÜ_sàZW‘`¿qFd©Z˜/WMØWžsÃÓäÀh÷k –ôñ#­1ÿÝOyr§ ìñÀéjP2÷I°Íÿ±C TŒ¾a¾ìZÑ= ÓäS*§…ò”E“ŒJ'Î5ÄLiÊÖÛ+<•C#ÿçaï|åÇlW*RƒŒè½ Mî¡M¿“D,U÷S7ŒfªÒ9¼š² ßJ_¿‘ Y˜“údgØTohmÂ9wÔK ^`mï“*ö¶Œþ@µüÃcJ‚ÞBÄeÒ8Éñc$=¸akÅšsuR¿Üùª £,;D©çýuea÷¡^„¿jÙšP 9;ÿ-cìö”ùcJ©(Sè.ôw0ï±ZÁ™yÓEý˜QCÊKñ`l½Q¡³ MÃá ™…S¼³lúúµþ¹5žOøH?îÓïÜòÞO39o¹Îd>2Ç÷7ÒÈ c.üê"[’ O:ï®F*R÷ðø›®ŽzçáÞÍÉ)ù˜Ä$¶R±Ù0} _A²0‰?H'»¶$³ÓÒÔjÍýAÎD9Sè}Ø–6_80Q­A²® ewÓ¢™úʨ^IKs9z$PhÔ#Ñí­»x ‚‡ÒnºEßœjj¯ôÜ ÚÇ/öÔaͺKk‡À…Î}£oQâtë…¾RR‡Zaà•zV×ÝÂŒVÉh¥_;SôfW¸Š1üažJËÖrG ï˜ð[ʹmÛöŽVŽ+oE¥€«Â=Ïm´ D6ŽènÄÅ]rïŽ#ŒÞYª5ÅX*Ä•ufßÕîIÓì΄äÖÓ£ ìö´íÊîpö¶Ï·+–ô*C‹ÈÞ›ÞDÔ`¼­mÚÐ^8‹?ƒÄÞj6‡ð‡Ù¢›˜üU³ÔpЪ׆*¬ÚÀÞÛ ¾Ë‚ä»=£r\EÛýhíÍ™Úì'1B9°°… ÒDõ¢% ÌÙÌ8 ÓÀÇR7ô‘Æj$ uô£t\»¾ÉH¢* hˆè‡NSŠyª <À¢ÏS’ö€Ï<2B€ló {ŸoµØîÃÜëc^†&ÜøV¯¯RIó£lîpí$ÒQPƒ?±&Ÿk¡#.q—P¼å¥Õ k‡B—!ìÿ»Èq‘ùáÜ5¹ƒ:ºulç&²ª–2z­‹Qt²NTæ(²¢Ck- YWK äÝW‡fÖÿ¯k c‘g¦çÓ¡{»x²nþY» (‡§¿ß+xg܇3wÖŽ6€eŒÓßLë|²:)××â‘»Ö”ÇPɨ(X° ¦ÅØÖåøcŒ³Ý Û÷.RÀÛ3¡ˆÌyÛLŸÏ){qUºœ0²ñÅæyŽ>±³ÄyüÜþ&áA³«ä­Á¾¡ú©þÁ(µc†¯0?¼Ðx8uËO#梣ƒA$¬aÄÿËÝ(ªÚîɰ­8ý•à…ЪbüÂÅ|?Qû¼=¼f‘ àÇ'Gá0!)h‹©Í`ò…g_Z<Ü2Pþ\½ÒjŸ€³=øàI’îŽïù¯‚`>W3~á]ñ„Aë ù³]Ñ*ÔJÖo6Üé  Ñï½.’<…†Œ•F£ãx8´Væ³Á”ý Ýb(ÝêíIGÿPɪô»JµùÐn¼ÒÔ©2süË],X£tvÚ€Æ2)ª*ô›sãQ:” Àx3M½(%€m?4 |4„òE»4uZö%ƒÐ7-ÿ†C¯Ê}Î-õx9@&BtüÏi ë'爯Ígô¦M‰YçDx<ïÄ˳ù²ŽËV §Do½®^©Å[uìCë{ÍN;¯B1¨sÝ‚,Å+td(;ùÔÄ~Th—Kšüóß9LÇ/zѹ–Ïñ®¬‡hôꬽÎÖ®Ó÷)PiíŒoRzO”³¸É•SìQ`Œq: i¡ê ½–<á!yTÍP>,€~Š=H€¥¥?ÛE~*à}>‹¯d‚’Nr9<"Ì—ñ+ÕVf¢‚Ïæüì}óhÇ»8ŽÖjCƒ°Ê ì¼öì7÷n.ö¹°Æ‹9‰ên1µ`³À=§_¯Å‘ÄÊ%q‡C [´9Øn¬'i^´vϧ(ãæŸ ˜ÿ—Ð$Í¢Q)y9qÛa—–.Șà¼ÕÞa,v„Ú O±¹Hë`ž²iñ,z¢ƒ!«?IDïÀ4ëܵ>7ìK1·6„Ù;¥°Œß Ý]‘È£ •–ÇBnÁìû¬W;q‰`cìQîÌ-åýðõ5?GV;ïS©îٴΰȶµ’˜þn\^ÅÙ﵊ ) è‚e@Š_’ á£eV`ÄåÝ÷@ðüèà¦U)Ãl/Š7“ù“ÆÝYßxm_3ˆg‹ýrQ˜ëÈnòyPþü–yÝàAÐNoïî\Ø*eïSð•¸Sÿ3' ƒ©GŒµ‡«âŒIêÍû?®Î ¼ÙXÖ/®kõqŸÉñ¸ÈVV™€LÍëTî÷õ“e¦X™n Òù ±¬»gY÷Ñ”o6p„@‡ow”>ooµ_ ­ò ƒõ©] ý øEú õ‰Wž)ù®®…ÜŽ†(hàY†Ó&sVoÐûÝ÷ŽôÌÃú,ü=Ùè¹ µñÀÌëØ?>J6oo!ø –,±ûõxƒÃÍ)–X–p:ÇÅÛPçÊ•F„79~KŸÒ©ÓÅNÍLŽÐsô‹{œ/dü55fÕªÐ>}XN¦ûÚ0FW”KÄÓ¡N©`ØYð™! ”Üò»! sÇ?8€ál¨’Âç²Úò/7kæêd·%ôy~¦lÕ0Xœžcõs7VîärÛb4Ü–¿+j¶;9Êé*Ë2Ÿo¤Ô×÷U_ǹ«`ƒˆ†üËÖÜ\m9Õ;V”kWÜ@|åÙž\ÃÈnà‘@ƒ#QÖ®¬-|Ôż)¶lÎTbÓ@ë*¹›}Ï0z$/\Ê¿©æ¹!j…ª®Œ'½Z¶žÅÑÎ3NH½µOigJß|Ö§aÏÝLWÊV½$3”À5òjý¡íjé—é!-›!(„}JSÍò® ¥Á`æþS8ãÒĆYNŽR òÛ~ˆ¤Ò‘;/õmœ\ÚŽeŒZÖê0œN:hÅ6°ÊÆk5¦7Ð,i “’‚º&é:žÝ„*òïf ‹¥³¥l@øÙ”¿?P=¦ 1àúbÃz¨™UV‚ª¸ÓšåL}hÌ ä쟬+Éq·ð)_PDÄ^v|bVÇ3ÊÀ½½}Àqêú„—÷únV¹6À>Ñpw!„å7ü;EÞ°9‹;&¯®1MCd#ó'õ*kz´‰•߇Pà8Þß Í[ëÄÀsÔfßc#N؉ÛÖ×áK²Öò{ÜÔ¥ÿ?+µIâ‚{À¸6¬¤¯úAþ¥GÕH‘uVE™$¥êo&£k3G•ë cNÚ µ!} ŒìÑ?\5UEèÍʬçšòTyEg 0‚‹ŽnˆöÔb<$èoJüçÃÓßåméa‡™·‚ã ÖsàŽ± -@¸q-g&9üÆœ(y knc¹:!XªÃƒ Åe²¢ëXè®Ê}6wàúK•P-‰Ï£nXk€§BûóñcoáÞ4sû°~®Ó¶æÔã&bµ±šÞß!ÏÌôÿÈd—gt,ù<–…fÔŠ¼6óƆS1;C´,fÚ{Q4û ï^” Ê(äi)<5“¶Å +V£@Ç‚zƒwm‡l­0%fT2°J+òšà¸’×mö‰ÞBk 2šJòCMÔMôæxÔ '|Ø÷ñª/óÅÚõªQ¶ôüå5[q%–õéÛ¶l !GNâê^RC™-PÁ™À=lâ]å6öckHö­ çt=¾Õãóq¶³Ýv ÇÇܰðXT'U ¼†³^ó'6,ÝÖ²HîÄçÞóPk±—q&õ®º4;©™v+ÚínƒáÓ<.dÚCcúuw4®´O‡É+:G¯'9iÓûâ•û‰d¦žD~‰~»µ€÷Šùeî2ºêV•x)¥³ÔGÀãÚºÈå^:1ÛzÉÅÿ1¯›ÑºdUî’y°²±¯£x§äëäÐØÖLÔ¬æ~Ñ žv&§g|ÔìlÏK‚½Ç’׫UÛb%¡Â¤™fЪ±[~É8éºÕ”1Õ A)c&F9¨Vö¢çÎa8Rz’•oøY%5§NÓ¹ èæO³.[sôtuj œnr´ˆkäf?ØçœÄúÒZÇŠH€ÂƒQ鿸â·¼ÄëJБæ}„‚š # ¯êž d©Ewï \äŽÂ3ܦØÂ0çuŒ ×lG®ï®E{íYÇ«ɉ' è{“0“P×u¯´Eê¤lç\ÿµµuÉ©Íý\M¢Å‘Eu’§Çª–“ÐOìß»Ûå“r}ŠVU`Ü$á-dôzt .Pýþb‡ß.Ôþ¡HÎù{yó¨`;+M¨Í ßÖù¸ž"ì›æôy”¡`¾†5dG‚û;ŽÓnÃÓ’Á‹{ ûN³¼·X]ñ‚ÔrÄ +Pv.bÞÓCòì…˜zrý¨«a{ôº‰èµÊ(B4Uªæ!S±ùýí{w@^z‰¡†HU7@ÜAAªí„“›†ä $ˆTnœ ÍQ ~ ó6 ©R®ïyk×eAÖHk²³b˜·ŸÎ{2Xh|A-æ­vá»ç¾§¥§ä£o^×2²¢©ù[ÇïiÝ`ª¤t°E—İ‚} ‡é•‚ôÄ=¬|™hgüŒ²ïÉÒÞú\1­ûÒª—ªõ i‡c Kª#å?I{0ßlQLRÆ^Æý¥™£Û/Œû…&±]H¯àlúÀw³Š ‘,J,”e»5fd‘¶2Á/óâôbêvJL<öÀQçg–®TÕwl™mÁÆ*sñËh7¨lVB¦¿m!\O¤^w†DG'\Âä,ñOŽ% ¢‘gJ“Z+ë™ia½Õ‹­õ"Ïä¯c.íªd¯€1)sê]Ú ¼Jœ–¦µ:oÒ­ò¸™¼7!–þ·%uzÆìÛæuS¥a¯®4‰qÞÍÊ$­ÅlóÖ(=ðRSóíæ%ÞqFÌ |QQ}Iˆ»vw˜„?èO Eçî3¯ç°“f𡻟MàÃ"½+\4ICCÞŽ£4., ’õ9®þ.ü|ÏU0ðR :Ð8¢—çœX%ˆßQ+mŒ-T­˜ðŠ¡,²êú~eŒB>»kGa€¤Áb‹wŽ=JIZ×neÔÉ=³ŸYçZEeTÌçpÀ£ÃËt‡”mdÊ©n¯`2·¥~BºèëçGòYh $Õ@2v ÚúýåPTv¡Ùm¢µ7š+ÄrêMSD¦÷dÉ…,8ë-Ÿø­Î †:hóô”gž)„, аT0Ü˼ š@í FÝăŽþ[…‹:€%ÍÍs)ŸaÀòŠmŽËÑÓ'Á(Wõsx® ¨°–ñyÕÎÂSC ‚–a²~Œj28<@Œ;gSC½È8JƼ r&GÄ««`ø·ÞÐØ¸ ðpÊ8Ù@ÐjÎο(Œ~…‹{ÄüäÝïô.ÿ.bØx¯µãý^N‡oÃ{:Œì¥FLë Ðd"N‘ù¼ø I—V.­ðÚyæÜÕŠãÎc½¡kÉúáÈ6H‰ÖÂÎl¥ŠC|»;1Ü@S ½Æ9ží#ºLºì³5wä^b¿ßÿ&ðÔÐmùäég»ÜN3䉢i;$ &ܦyA¼Qâº*òìðô“²ÀE„üIj*ô®Îû=ã"õËU… ƒr †­*Æsåµ-Õ2Ý…wCœKè ÄZ2š-Æ=³ÑìR¨‘Œ !slÿ‚Ü´õ£Çdç®^î)×ÂäÆ"a~ù¤:¢þ¤ BV|îÁm¬è*±ßpƒ<™/ŽEÿkÝÄ©=œ4¿à²q¨ÌûSÃíký¦9U74a§üɉãepX•Åsx`Ÿ@릋ûÒÀeY¯ûfc*M?gA9UÙűJà›a¿$®iA»À¬ß¤R²æÇeè¯à»¦Dó½+ºá³ôïa“ãL¦õ4CYðsÌm;ñƵ&—&œK}bq­ëz4åVzˆïˆÁ•& (ï £ ­K0øÅwêô™n&Ád„E“)äb*þÕ‘Õã]|k¥lè¿£õ½9üƒ½1Íäõ߈ñ /׿nŠ4#9CY3à ]E¬ç}á;æI9‘ûêc?e§6WÖpg FÇ5&`-•~ŒkõÒæê ñâNÍzž£e͙̥¸5ÿáI…3QlkÅ ƒ°…!Da5¾°¹&{'Ùì®øO7ô­1 H»É ‚oáÉ/öñûá7l›§”‹ï ì([Ì’|Ô`”#XL$¹©ƒEÓõ?²GÞ`´”³A¶p/h‡Êõ˜§j`ƒø¹œ!Ç—µ©Wƒm¯”à«a‡Ô¡êµIž³_ï !K; €Ü ö¢±Åç'æDêÕsô|Æ5åü_Š·ëòÉ*P¼¬(F?v3ˆ„„²ÂfM4{Bœ^Õ dWØŽhÏ›Gí.8N×cˆrÑw•ãÍða4žâ-T¿ŒP§.ÏC4˜Oj¶ÂC ¼Ïê_(º&m6T»ûÔL¼ŒxØ àHL-!ÜWÆßß-§ˆLè2|HhD}¸P߉!r+ðl;G”mUwÌ•ò ]îϵÛZI¾ým cô{%Eæ‘Óïí¿#€²xMo†ð„ùo…:Þ„Lø cC‘¦4±®Bô‡û÷M'îUqõåœÉii>+«ÛŸï“Ñû9ð¶Ièv´Õ‚&wýÜOmdCâ@ž(nW–ŒùÊÑ#d%ÐüKÙW¤ð¡!•¼#³¹s—úsñnîY`?máýÖ›y‚™Oß§˜ÂQ•ÒŠVÇ%fŒÕ—âîH\7žŠÍhÜQ°·M¡±^åÖL›¸ˆ´!¨OV‹-ArJK¬ò4éU¦8ÅËî·¶BKxU-´ó­˜: Œ›ö;9Ï€NZv`D®†Æc/‡$œ»ŽKško”ƒ`ÀÊIŠº&£ïûµjM¬é/½=Í1ÓØ)óÍâ•ØkãQGYö˜WÇ’wÇqG,âÒ{ýýœ°KéM ³Wä3²¢Òa¾Ò|ŽVÀv•ºz;oˆ–‰GÀPûácù ^àjÜX:xî«y×›"T2Ö3öË™ßc{œHNÌ6îØìˆ€0oÔœjQdb››ê‹;ÜÏMa`‡‡€äÅ˱¢z¥êˆŽŸ·/'‚#Oçëe‘êq*U¸ôa:‹q/§ÔBÌ!ÂA€iÇî¬Fƒr¼Î3TÍóµ·Ý—G¤¾ùZnKÈq²@åœ̰eÝŠ=&§SX å )u´¸ž%ûgWxH0»zG©b„˾»:æ¯ÖL‹†ñccþac“dt—Ub¬LA`ŽXÅæa-O•Ì&ÿ7¸õMîãL"¤ãúN ÏšWÕÎÍ¿HgmcOX®`ÜI­*+–hžårÚ$0q1w1ë—ýº`ŽÊvò¦Åyhÿ XƒvÞ–!ºòµê[°:uåÅ xÂ Š©¸:#>\—Ú9ÑÆì”ì)ždý‡ìz¬åìà&²ér6 ¥M=~N$jo1ëךÄûé²ûU'¼eâ7~ë*Ï¥=Ùý*ÌDοÍ+J2-k¦¦~¸IRºR  W%lÞ#ž ’Ç¡®XÔ¡LV¿å÷Âd%›y;éóS¸EÇPÕŠšF¤Ç¿¿Ï?Ðj‹W¥ ˜àʸJць$j)s9B+yŸQ¦"óó:µ'¾//‹Ÿ²ŠŽ2hm˜iáÌ.€«ødΤÍÔæôýSÛo¶ä0K`åìÍ‚Ž1° zÖf¿é‚‡$¬‡Èû ½Ø%J#öõ·>wª,oº+±j‹§Bû‰–bëIy/!ÌÕÏ]šmS‡ R¥qðFÉãÀ±<–­úÏYœëóußÔ~#+Wd"¡¤•êgIrK¹žÅɯXŽ™×ÓôP(£#mÁU¯@H…Öi›‰¥bûŽž;ð¼ñsøÂ'#yŒ‡Èt`}çA?úìŒ%’ Pgö Îþ”0» 7AM$’¶U3 :{ÃJp«’Ÿç–-M²Âa%ˆ‹žæ£”sœ_ÄÇá•Èhó0÷d”“ÙZ7É;žãù6cÅ_˜‘/‡L2ž>Úâ%¯ å[B~ÞêÖ¬ÄÓTe‡ˆˆÉ@ëĆ©M¹oµ/Àu¦Jì)òA:cKïS·+´s-(w]2š’CçÑG|ä Nðìe"ŽŒäšý°³(Ó €C{ ƒ{Ó`“—ÿñ]4zÒÞ.‡ûÝ®žÚ"Í<%»1(•\fã×–¥ÊÃYÌ<ú¦lP0ÜPss'ØRòyßÞcè®ïm˜%)V¥ ®ò…{êM]ñ}š±-ÎþV ¢yKÄy}IèÜ{ÕfÚ‡çCV ­EZ£¢”§ %ˆ 0<¦Î7dÇŽ*l>,ó¸Jð™Ï4™…G¤ÈT{4T-Çê &aô1Ò8ƒá,~d@4UЋ]Ý=2T¢®ÉUG|ë,-Oª^ÝY‘>>ªÓ~ŒÑŸd{#Ü륾ÛlàD—J2a´o6Æ›¢9å5DkSJV=ÍsÅ{ –Þb h±7Ƨ–&7ˆóú òÖêMà¨\Ûg ¬›CÌÄV«×Uÿ¾ §l±Ñ_ ˜‹îè÷ò@@<Âb½y N>”´€31þbzŒgiII#2¶ÈD,«p"÷Z_‡13b0¡<új¬R@&Dó£MîÒ˜PÈÐZhDW…ô™q ‰”©hn ú’p r4Ô„+ëëá¸p»Á-I,8­$‰W.eùrÆwÍÊ 'L´J¡Ó«Ör·"‹´Ú¡aÜ;÷Q-°y”‡ŒÌבqu¬7hhS²Ô¯{ò¡vQúðæSM¡³°ðAœë;h€¡t”.b|¯}BNró”"™¹ð’9{3;Þ]9ae&€Ÿ«)ÈæÆÜàì {rgy/˜ Êü?ÀŒ ÞhLß £N÷vV4^‘ oâ}q*IRRè%÷ÝÖŒ½B_P·¸íFUާ”e´ò5£Ž¾Ÿ ’c%þJ‘Œ2ÿ^ìÔGÊ)æ€bÌ“iñ>`¹Áíær1ò¤ã}1 —j8sZ¼Ñšƒ?jV:nSƒƒ" ;(ô'ÀóÄh˜›²ÓNüo›I|f‡m >žÊFB ‰+¨>èiQ@h^”jœ;I>ÈM15‰\ŒüI,å•â&ÌÆTáÜîþa:UL™ XÌï:å³LLçϲ­ü•›^œ¤R§¸ô0‡„¥x&ÁGÁG¡®ËúêÄ(çäB– -•ß–êî«¶Ñ…ôq╱xSºz& ùAšb¹6VbIŒz°ëp÷±¯QêÌñb"X‚ÈiŽ ö@½Ù¹Œî¤ñ…zô’÷ÒÑ-JI,*Í"¤Nž¹$+W_ëHëš´­@Ñ–ø¹£`tÿ3ÀÀ—Äkê܀Ʀ`ˆÑ‰xÎ$à°±íøá 8àÐm‡}tƒ»âÔLN¾èT¡Òr×óT"aÿžÙõw­;A‹qlµëV£èU|ê \hŸ´r¾+áØŽÓžíäÝ´½´zñE4kŽäüVZ>8­†ê½00vØz”b³ågðÑ™SÞGÀáDñ ÉÏ·°\)5®ð¨ª/åMÕ@Þ2-ÌÕÛ¥‰™|SéTÍ3°,"à‚BIß ¬üOî‚ P‰ˆÃ3[Þ#}(¹d¨‰àɱȭ^§l£qnÖ).éþØæƒ˜J_ Oé¯>|Âb]n Š›) j‚¸÷ë¥ÿØ(itU‘ûc•Þö¤›ºpI黤 û£«V”Å#/ßæÆ¿#s`qâß–b‚XKRt%¹öÂ+Îbκ®}baþ¿ãªè’*ª¡^†ªvZÈ•±^LöT[v÷7Ó"^Öä]Ñ-—V”§µ A?­G:«òQã«êÌO´ÒTw†èÆ $3Ó+‚±R¿þ£R)õ]_ÀVôìõÇp%ªûsPÐ:6 ÷*ÛŒ²‘VOV0Ÿ³Ø$Š3Rš£h¤ùÎ ú¨H»*ÿ@mD8Gõšñ‰ýt‚]2ƒÝnŽÿ­S‡'ë>ƒÅ‰ iRað6CÇ€MiT;xÖ‚Ò¾µêøJN¿jðZ:À3UéÂì‡ÃKˆwaYtÅÕJ ´ËÈÌÔ¬âø0ÛúÅ NèÌn‘?üÚäÆIŒ4ÚV*½c>²C IÎxǹJÐ(! o•h¦û^™Çob­C“¾„ÕÙ©(Àk,iW ¨¢ èç-@©ã°Ù¶XÑÌFµ3×oŽ©ë%üSÑ‹&ûÚ5 _np)/ÍxBnÊ'pA>B0#¾®KJ^!¿Ôé¯õÖ c;œoãGGêuûBèôéA_i)ýI\%½&¸| Dà;Ž^‰åˆr¡&©—h$ ÍÖÇ£FlýAÒÆÍð(í×¶º05ð¤¦4˜™è‡¹ˆr½G´’ÍÕÆ$–µQêZ™*“­!íMaÌð;`H&¡áA‘£ÊWP{Öó*1ÿ‘Ã3Ýý»¾oÉVìéf˜š|.Ædê8¨ì«õWÊiíÚ”Ú1ë{k¿¹Ëö8Ž”û=3V ÅÖ„ÅLÄõ<}óŠØär¼cþXçñvÂ_u²$~üØ©´Œ{žDü˜Ü]OíØ_bœÌõs:Kª¬×mÍ·³ ëÛºx‹OÑV<¿ïdz€{üX@cd8¦ÆsžÈê´4@Ï ”p82ƒ£‚8'¤~›dNRñÀµkJ<'wQ§e0 Þ¸ä­qt+‰}vgÓÎáÉ%·Sþ5Ì~ùB]¡ÞÇdÆ@üš¿}uáÇVÅÁÆ7¼æ°5:ñb\=+%$Tbs7Œm-Ú†­õ,]O5˜©!ö˜2Ó¯»~\ìÍ<>×J¤ï=Mä¹-v\rIÒÈÀ ™·dò i¨_vLAj ]ÙøaìÓ½ZRwÏ5c¦Úx)è$ ñÎä†îºD ·3‚úMØZRàv¬Nk¦ž¹È Si£eG,;Yq¼žk¬X5,ižÌ©öœ4÷q²`þJýâCL9¾f¯åÞŠ-vpØdÕ.\p’X˜^5-ቌøFÙ¥qŠÞ¡ L~lÒ8Ô%ŠvOñœ˜2HLFôv«„¾S;þ ´©æ=%j;ú®÷ö(mŽf,q9-×Þ²µa!‡^.ØVü=Ž[±¢ËÁªÙvõÁØà!uûµA$ªo0Ybáù}õÔRˆê+q¶à| ñ©§`>+¯÷—¦]Y=æHEÝDæÅ~¾=‰jMU¥ jÝ{y е/ò³Ò¡ÊδÃN¬¸’Ï$Sl<ÂŒô+cû€¬×zÃ[ {DšöX47Bµµ%}`›•õèq"—óMòÖAíëø¿$ï fÂ2×D®‹W3Ã,+×Lײ'ݹ÷Ï=·XÖ>¶"(µ˜D/¯N.ýއMÀJdÎiœXeêaÿaL#Η‡7ÍÇgHöW͆°+ƒ—o¡˜â5™æYOgy…¬#H«wˆuq—ú Ö¬1•€†³MY¿# êï<mlP¥Ë©rÎ\ª¶ŠŽLô€î¢9ëkÃéØÙ¾Lš ?]EÏí¬_aü{`-Ldšú(?)ïõaû²­¾[[ñfë cÖ?ÈJ}U›œûXwŒ2<`½ 7O€›2bÛ¿á…îƒ Ž—ÕŸºæˆÉþaý"´…´¤ÛS 4ëúë%'7 ž3+/½îóü]hðh?~µÉÑ%˜Ø°&0ãµCôâÂG«NqS‰fhÇÿ!Z CIY;÷¹ú¯æ†‹K÷9£Üvqkëu.“?¤;¥JmòiÀ¬f8ì+ˆ ‚”èÄpÊÄ:~c!3¯êÛ‡ ½Éê~k³Ó¨£ÀòÀL·,!À™Îð,¥›”?âÕ¿,Þøü±#%„\©äÔÓÒHš·²²V¨Ö±¹n‰¹¼õó _\n´Ûa 2úÞƒLŒâ~j>! $€Æ6]¦Æ¶Kó|4ƒ§ Ýd5ªoÑ|=V@~Zñ,BQNMAQì!ƬÊmÄA“Û‡NRþ¨\ïºÆ­¼Y3*G$Îú“nâƒÕuìŪ×#ÐέìL_ó¤Fæ½Y”JÎíŒÁ¶xº ßcå¦$Ï͋ݢ©xÚ»4ÙìɈ!- à¾B´¾mo šåƮϼrD*/\úoSµÒe£Ø?AœPR†4_ëj{y>aJ²}R[ ¡,c› ,I^WÍþ%¡Ÿ…>OÀ"¤gœóg¥X{ñ÷f¼  óµý _Yß²þÇŒzÞ¸èbcÞÓÓªoÏrÊn[9ú¤… ª»ŠÀÁç¤õšhLmß¡÷ÙÛj3Œ#röº&LW ΕÝäþ K“¬á·"t=â¸É-‚tÙ èÎüÊ%³ë%]\¯/·ßÉÁK4m”Ë Âix]CæÞƒ6¨åUe*ƸÈÐK&Ÿ¡ 9SãþíWãŠP„ÞzU. €“ì"mzö¥¥‡”BVc–ç¦ÿ­ŽôR|më¥oGZþ…»¼ôtCß„fnˆvèg.]²¥GURƒZú…n@@ëaتc ùš©KYÜ]àg€ý&–9TTû‘c…®Ã‰A­¿æÉ!W‘õجtç†ö…¯Á~ôn«évåñ{4Õù¹ƒÞÊù}B]tÀ‘3åAF)¦°Ö#9òVç#ⓜb)SÅr}Üoô7˜Ú,|#Œ‰O6§ÿë>¸IÛþ8Håÿ(‚Mp³ ˜ØÚ;£ŽA«Yui- (àVõçk@‰|øA îÔoà = PVÕºÄ`¾º‰mÃÍH/²·Ù˜W+”ÉB*>Ö+°6ŽÓÑa„t×?3xΩYÃ{[¿¶ÄS0o'ùû8@43CüMÁ–ã`.K«UC¸¬hv 늴ÿ<n•ŒoÍdEmˆþNý7/TEãš·ƒ³A$Ù˜øo ?ϱY ùÙüîàÏ1ÕY¶Ë¦öLm:àÖ|CÚ/æä8ßÏÈö®ô7¸q³j ßȘC•¾å‡ÚDä&ߥcHý`ý !Ði¾¥?{’ppÏÍO¯š¦ 9 ý…6®¸ó4 ²1ûŠÛ ³WúC” 0¤í Õ¦¸¹M]v$¿Ú6BVÆmšxLÝ2tdòŽ'éêíÖ¼ƒ]°_ K7l!Ô½ªÜ®d>˜µ@iJèiØæîû¾*ïÕ`1cÿã-äÇLëéx™$×û’áX©‚k.Ží°]¿ 2œÝbìq¨þ—[æÂ7ãVîÆ Ý£Ò’š@qÇ7…Ä=kŽƒ%«C "-½Ž~=Àèh&Å;û›âzRqÿ1ŒœÅÄDB¬‘¾rÜ]âðB¨ÈhéyT‹$Yˆ×ž`œ"i‚Ÿq&ꋈñ~c žRš¢üc‹×”%¨ö³Ëb>y5°Žº“ŽõÙO¸j“$'¤õ¥ÂÉ…[rÉn7,%HÝ|‡Ö7kàzŠrrcê³? ùÜ/ÍíÈbèš™EÓêTX:^»qægUÁë7ù(µ 9žy$ñL?l`7éHVè– lŒïýµ¬×Q–aÀ]Cá)R¼rÚ‰Úû-mçenK½\ng²`ºÓ&‡áIŽ8$J¸Wؘ—Ua²V´[ï¡mšr#q7UÔœ[f"§8Õ1²½ÄU «€@$ÄHRuIBb[R‘›KÙ‰¨"›‡"|8®€áïþ€v.?ó:ßuiù¢£5ì,éLTé˜lG'kR͘’™qß[e¸xÝ) ‡†& çËÐPæqŽ…¯ˆ›/H±‰8(®Òð­þ×=Å þ³&ii ñÈ]tà<juN4EP>ö*íMdÚž©ÇK˜B†ªˆÿ±ÞĹªpQoƒ,È0ÏØŸòSKÄGÕVg¹ÎsÝ:s÷u|(¹MBs²åƒƒÔ8¡ð ´c«¯6¿u£—ƒÄ‡zÓSG$Ó—ö¿F Y—f O‘¼T}µ]ï¢,÷½ÍN˜j—o¯áHfÜ%…°†ß§Ý=?–·–Á; ­ûEÍJ/šãÜù¤ÕWKô12 :Ðæ‹ÈJd#qXb3ñT'¹û$·Š¡b{µ:r’ ª p+üÛRò+ÀÊé4ÕrÑñY× £`ŽGÙÊíêò× £R]G9)oÞ5v/&«¹è½ºžWûQhžjŒÆ^Ú9åݳrÄ3ÃöYÜQäåéõöN¯l¸Ý2Õm^1‡c ˜ÑÕ¸3k"Ô#O³U(G‰àxäÉœme€Ú)Z€@ÇArB,²æ û“ò¥ØÝŸÉ|Rê!·&áC[|ʰßé”UEP⼨蛌Sn8aÆ"wM‹ƒˆÛ—~uôÏgçüŸŸ5Ó/ÑüÑ{;–¹>,r¦ãѱp‹63Ì…¿ ‹ŽtÅÎ;¥u¹\’{‚)¾nY…t­·§ÕM&(Ãá_Wî¡/Œx¡YÂ`1ì+ëÛá–úÅÝ„- éºhóÀÈ%éÔ9R_i/H,bno'®“´ úoO{Î*4~†HÓ‡J®V–¶]k! ¡Ñs‘ÙPà ®T xÏ•K&\õ°„øJrk”Î=d¦–Ï«¤‘|(mhñYPp’Œ?ÎÆÊPàò\:pK޵sg?·œhíû;o;4Ã×Èûeø×™@Œs>Þ2å(J³ÔM#ÑW<»Cñüz(¨¥y¸ÐCÄÝʨê¨é‰ŸÄ£÷‘çŒ:d¤.&@f‚OFDÓškå?že¦-ß5·Z!ìŸE·¶WVËl ‘{ÙïØf¿Ž˜ä%¦ç²LUÛ•Ù€}<ˆ €z=K]—ˆ ÏJl"¸Z%ƒZAM…dö&yÎñòÁ}Qû©D¤9gZ±s òQº(f‰õÝau>RK.¿9Ú¢‹0ôT;”ŠÃ'ëed}„·-ïÞýz*)Í“C^àk¹*=`• b2x«mÞ™t´†Gþ÷ž9ä£Ä„”9CkT%S¯‰È0ìZN¼XÜpî"Mè‰iÚÓKIL)â܃”9.ŽÓ÷'hĪãÊw@ßèö$¨˜Sl9hÝr¡oÁŸ»õ”_MÄ÷õ×û©µ#Þ¹¦¬Yþ¡šµQW{‹€¹ér:i*ä ?*¶U¨_c-s@Yë¢Ì{°u¸ØâD:EŸêN¿%HX‹°ñ±-Ò…šËèÔ0ºÓõ2˜¹J¢K“¬íŠ„Åz5¿®¾¿ ãWïBñŸx¹³ü•ÅÐ)Uù#z_Ú{ìÚeç…ÓJ/O&¯Jˆä]IZã©{ C>4H~áZ¿¤ÿœÓÄ@ƒ[xQey:×þÐÄjÍ1ÝVy0]ìÝ£­÷ïqh{‰DýÅBüd1Ìö`ÆêæÞB|Ù˜gÛSú‚Dv‚$&Œæ¤®8—¶5¿rp=Ý„ú¡QMfRR¦±­X:obÙÍ e8•³C‰l»?©]esÕ"÷q=ˆÈµ%rè‡ÿ‡ŒðS:\óÕY¢B tD¶T9<ÇÇÉùw0÷ïerÝDrÙ¢¹ööÏ]>l>‡O«¶;ÊÜûu{Ë.§Î…¼”*#dçÚ@‡œ·Î”ã?(#` 2Þyç|9O4Ä4n¸ª#û‚±Mhan°¡òOÝIŠ …Âs*¤@‹êxÍ*ÕãP\_ÝE1õ.4™¼¤“MÌÃ~c0\öcc n§¹»jmŠèÀ®[´Ï,Nwx`EZú2O|ÝýÎ!éfœÑkÇŠøL#Þ‘(ð™ƒe˜@“t(j¿X¹a+wìÂÖÚí<*¥oiEKÓÇ(ùQýÿ—7ÿt|·):.Xÿ7¶úXMG ˆÑ‚ðò6ç '¾®Íe/hOšÃj³WþÎôÈr}Z9a÷÷@?OYßš¾Üò˜;OÙ€«¡Â]yÉ}Ãóí½'ä&ËáÏì°mS-wþ~’cgõsô8”Mãøù#ÑÑN‚ý´Uô­h÷¶ÃiÃd'KÐ>ìV}Ø÷¤bˆÇ½|öioêÛ@‘ ïH´H勹××›*TcÀx¶R빃ÆÓ\ ¯·jãRKæÂ€Û+ì„Æ¦¨Œh Ô‘Ø í ¢×mš9GȤhèGS¢W˜0Ñ~ gÒZEV³ƒëÉ +émÃóñçŽ>!jÙBC†KX1šÐ¬ÛíÙ£=È$ª¿Ü•©’ã›É©ÉõDð%M,¶ø±ÝÈ¡pâènïöB2z™Þ/¶¤žòçÍÁ ‰)uÊYébòùTìsXÃx :¬h-îï\äP ÄŸÑ~#* ®àX`kÕi9Ѫ5ü©¢´­¬¹y™»³“_$3—ÂvU´ªfO}i&Ê᡺Ðb¥É¼’[»“^ÀÀ8¢–ºß霑ÒàB°•Ømºìñ“HòËd¦ë¨?Ü#*û‚N_žJ`rUÈ÷u‚vÚ8­-Ú•Hœ¸‰aö˜Øúëd,"¼Õ÷-WÀäJC [=Ó$÷AÕûu…?£Ä|zå1ÎÕ‰4Çž<5Œ "ë´½b£^mǯù€RFm­Ù¿Ã¥¾ð0Ô’„f”-—cÁÓ“v90›àÒ›!/…†Ä7˜Ÿñ»}dÇš$¯Qã}7æ±–dSrìå¼.À1ö{iö¡2ï˜2ôxatÛmÊžÆîjBÝËÒÅì\×E|ñ„ì×Å;þ2ØnïÊéD-SÚE~Ù„t,0xö»êmA¨\­‰‰Rðù8㨈!¸g2?  k_~âCâ~)¶æ[7±õJ»s¢éc6ÆÐëpJé§#côh<áâu§ŸmjíÆ›L.&ÚÒd/BèÒ_ìÒ€ìᛈ”c_Ž‘F‰ " £±ÍrÅ4ª£d°×É³è ªÎd5oªÿÂÛ ™i(üdÛû¦º F»Mͬs] ëË;“92©°þœºtìåÓœmµ7ß(@ý¬Dü•ڇܪ¨”-t…«i/×xѽãÒÆ”(‡ºÚá×F)[EŠ6:#ã¿a+‘ð0ëoÎcB‚¤ˆ8m“_ŽF1QŸ1ÚP šÂ -8ϓˮjö&¸ˆ*ðÇcÖÃP‘áy êË?¬ß¦zÙ«€ô Ä|F`¿§ôXÀ©ú_mÃ:Õ5ëøFkêeCaBûN`µZбSy¿5à©àðø¤¨¶-9Á‡h€|÷†ÅVƽl‚9¶À§”ΓŸ9APäSÛ¾áDþ2éò› ÅÔ!ˆ^ÇÞu2‹ý†òD7qëóù™ešÓµ¡óvïh?kÈÓ®èÕ]î×ÖÞDé<2¦„‘«çhu.,XºuR·Wrk…(wÙ©.ØcÝŠò‚“#qC@Õ†¾5ù€ñs‹‹'DΜµÊ<ã¿C„<ÍSNR‹¿¥ánÁ4ÿY€³§Ašdí0û¢~aá‘ÿé²u2N´t±h÷—ʌ۬sù“yÂóµ¼? /×&ý¥ø¯E²õ#ºÄ¼ód]Ý£1S–©i¢¾—ÿŽ•¹Á¨Ï1žÄPÈ1–šâƒH%ó±X'q±T¼¾fM«4ÝîÃ=FËüF`˜`W¤Nò8’Ó´«zf k,ópáX¾tN{î,׿ê÷}àj#ŸûIR Õ@²þæ.¼ã°·¾ö!ôˆùªO=®ÖoÛÈäý–(ã_|¦µÊÛ›NÜfébFßoÃ0tüáÉ ô~g?¼:ß³Ãã‹8  DC+íö2’;eDï8%žÃ+».Ý)º–æŠkÇï`ax?S3ÌÍ%8æ¹4ˆ0ÏÏÚÔÝñ,Ôìßµ6Œ¿ÍÖ»`p%gp:MqßǘÝ!´9©•æSàê(7rtíø–U×ÚÂ\K+ 8伃SãI™yßòúJu\ФP@½ÝZUE<”s+}crY#š§$‚uh·Ëøg}‹®„óX¥2Ì+õ^±èĜĹyšÂiïAÃ0Ô¦ä½7ø>5[¤ÛуŒ±¶azf>lå;3´ü’°^´Š³’K(€ üsL‹µ½mÜ‚¿Î®´Åd„¸•<¼ýúYÍ¡1·î€ÞZcš"ú, AŒY¦Ú¢²½äÒùª~RÑð´Ã†ÖÈgU’.¤yïæI~ú¯¼û†Ù©œÎƒÝÚ’·ä˜Y$0bxݰéí€ü ¹@䍿¡9"nÙD>Üê,ǦéM]h £Å qLǨr§è…Lšk’¼+Êé<êvàƒ‚ñ.„ÕãC‹9ònܨMÍ@²U]<£ŒN6ØöE7ƒ¢î¢ˆç†òyèD¨iÓwjðÅ:€³Å,2ÍÑûçÆüjÌã[WúïzËöÿV]a2a>ÖHŒ;ýR;¥\'o§q[¹ûåÍE«k^7û¢`}ÐÔQÔCÿ?@­í[v‡ÚüçÙÑ­ö™êá`œr¬ÓFæw¦¶ÝH—†ŽW.jCŸ>Ö¯Å7„»€m®´[_`¦O­ÏªoNîTIù @)Ë6çYµ<ÕŸúĺ(«úúZV}]KøÖÜ‚ÏÈ1>i-lúbSúKÀ}#¢gÄ%pWvë| ËÈœd꽓ŠËÒ:emÜwôëQ¸Á±P×FQѳŒ¨ïlÊw*Hÿ*ïÁ’>%êoÓcO­x‚ɕŪ«/–Y&2“ߊÒp%.|`͵úS(LüÝã`4«ÅwÑœê$üçÉw… ó7MÆSü±åâyå ;¦¾Läçî‰Cî õzòÀVŠ8_@Òjþ ¼ôð߸ç¬_E@WÏ:Öº_¿BNnP\àM+}™ðõÑtA¿Òn—rÏ¡@Õ¦õÅL¦ ªö\£»'¸Æm6ˆ£H]fvèPÒñõ÷…Šf®7kþB,ÒÐ õ~OÍkyÐiÛÅ‚štå)’BÃJöȤˉo £ý¦±J‘?, €5çðp×OùHžì@óUH÷O{)øª‰pÎOe"D˜ìbÂm¼ºwè$ª(W.ÛÙ '7u®ÐIh)Ä<×QÅ:¸ þ1¸I<H+Ä·hÁí;d±ó07)'ðÚÐÅz˜âP¦>E¨œÉÃiûq1È]«sšYzcNRØðFCãë‰]»S¢XŠ ½ÓE}+M1¬É0éb^&ü”„©0ñÓÅÏviá¯ê½Øì'_nȯ¾)Éëéìbè‰ä*$ò¹#HOŒo/ï|Æ×¨>«óOèºOà”kSÄÈ-‚µ1rÍ,°DžqÍ …dÚ§$Þo'j¤5ßšõ:¦{œ ¿f¹r¸T¶O¢6>Ê ‘ìX&Špªžn›-FNãÈEijäíFI $& Žªî“ €t£˜ÝÖ¨¤E•¹ôéUlƒ+X1¿§ˆ"©’ªªSÈó¶ì¨qâÓ8~{ǾëVüˆ$»ƒG†¿'bå{‰‹²Q¾8S2zåL,!E»,žÙ^æyí¾=¼,¦³[…éRfmmø¦™äç7/±8½‰Îz×\ùÿîP„p3JèIGôÔóMØÐ{7G+l\/hMñT«>T4¶7è¬]p´ê·Ü~'f÷|ñ;ñ¤_VÆøòýZ. í}MËBêm¡‹›þ _[4¹ú“±Œ±®[:w-R´ Îù‹4 Jœ-˜|@iqÿà«F^-7TTb—aCÍ“Mo¼>TN˜†e/7™¡G× ×혲(—Ý9ôN ÷okz+Z^ í¦&i_L6ÿW-úoCÃΜ䦊áóy ÞDj  ~_(’œ¬:1ÍØdxájº²i×§ ™9ù}Ž`ŸÔ-1†8ß×âd QKÎö[¡óéêX•R‡¦ª;bǽ ÐTo¥‹;íêIMZv3Ä,›OÈœÜøÎR=aF+Ò@EÍò8n‡Ôô\Ò´ëß´ ‚}ÿ›·Yónàþ@†Â¿AfÆûUW¯Ò¸Ö¸-èJ4 bç²â Ë›‹yMÖ­ZÓü'¹‘Å>zÍà}›§Û„ºH>?ù8õah¿xÊE ë—bŠ 8ü“¿;I{üJ2ˆ:ÍI0$Nì‚0“>0hé1 íÒU?„>RWÊ,ù%¡xð,µ±N“F¡hŸ —æÑmÖ¡e[§•À<%Ë#ÃIBʯC‹ox[n4á\¬:MI >;ï&ÜF)ô¹z¾Â05Òh˜ÇUó‚–®Ž^g֙ݟB"‰òP~`φl77Q·îØß}‚ ‡%4W€/a‘þsà,b9zýU%/餽]â“® ¼ "󀈉)‡}ReîìÉC³¿<s§…ÊX«;îÆˆd8F1Çeÿ×îµutõhuZ¤ÖE³®˜õÑñÚë ç8ªâ䱸¸\ì¤rá^­Ö¤«B_'µB…jÞ ñuÞC¾H_L{y ú¨û uÕ=ˆäwŽý/œGa ªÒÀo(ÚOÉ$*‡Ž›å\?pìïøˆs,œeÅNºÈoBGî— ¶åêÕSOä‹O(Ûò×t#p‚ˤu餥&¬d\X–yH\¿Aù¢XºÕà«§zŽ ÷qÏ…ØdÆŽÑÛøu`;lGU®˜8|"»¿ì‘’U`Q»ÉõÅÜn¶H‚0ã°Ä"RGZãá“ôäÔÁš¶ç݉ŠÎN¥ÃcþAY-BR´W1p¤³âvuhAõëZßë_*Iª“þ¹./s›ÛNüælÆï€™¼õHÃCŽTª[Ç×ò7 ô¦ÑâghÎm^dË–«mß›Z·l‹èHCµ¼÷±+ób÷/ïL‘îÌä¦v>kk® âžOq/¹íúå[ÑüvØÎô:$<¹ß›•ç-YCÞìöã Q%LdÈá+ÍïVÍé$ëÔ†Ûx¢#ÛŠðÉ2Ó5P|~8®öO¡àòp~i“hÕž-»’ez8g]¥ÝH}á¹ÐÔ°fPQþí;¼3ÑÜÚ˪,Då+ƒ=g øÉUiˆA£ ÓE" ;‰*d[±!ìyðŒæé$ç€"±/ ‹£’ñë1µ~ú•NaüÒ%¹iTJ–/“9du"(xw×â^O®{‰çrÔ¤Žó¿þ‰Oþ¯© CdCã†ä¯µq)í€Vd<=u¿ JƒP>Pô»àWïÝ‘Š7À[XºVÂAx—x–x’DšÿàÜPpà=VeØï;„˜y-¤åÌ3VÍ]zrUÆ‘åzý+béJ>ÉgF©oú(9ŒçB•ÑŸ‹*58Œ£¥÷Õáˆïé^¡/A¢þãHzÖœ¸#/Û+€Avþ¼ t+‘¬ø¦ÝÈÞk½œØ{¶ü+©Ã§ ›Ô²Ͼ›îð5±$ÿÎ ¿K*fÒÍæ šQÄœ6(†U¸·M´”DQc-”h×ÌÚdÛ%\Âÿ$oŸ³Ð×@еŸKšÿj`’u`\s’~ÆÐGÙûð*½(iÉ·Òj£Îü²ó¦83³‚º‡}ÿ¼…¹)½«„ªWüۺЄêç½ sxÞ¦!«¦¸ ¢ÀWøÁAûÎâäAóÔ…–”Aß„z€ÙwºÄYÅ ¹Þäµ™XnÅ/àBÚZÛ¾¯[oõÞ7 ‹ƒD…ê?öÏdŠWÌô´4@P×*ü¸—|·ÿ9|W¤× ¬ˆ.ñÑú!}ˆA ö»†khs[c„æ›I•°ç\ í /žliËL6ø®RÖþ2_S˜ìjœ„_ñø~rFƒj¬–pÉ:_øOŸl|X³‚Ã=;õrÑ_‚ˆ{D¾Ÿ‰BqáEK LššJ„þRD6 w ã:](üTc.„¨R‰v`ùšòH‘]QöÊñŠ6Áõˆe«ŽXVËOÛ=Á˜fmñãÏ‚9ü7Ý¿d~ãÇ7 t c"¿ v-˜ºàcúcd þ¤²$Ž?nnÒS§€ÿ)¶Ænol'*zO[0v«Fíq,ºO@¼gÁqÓwgàKâ4¿ä+Ü›Çþ» ¥û£†_èîªUB©öò5o•Œ9-E6±9måûdOn-+`üâ|“*¢´×m¥}(¢˜îÏÁÙäÙ¼F§ù¡ÆkÇkÞ ]ñN??Ž4glѤěa5[¯õý#¥ñÕ~ݽüÙ˜¸NY’ëP`BƒÙіοîÓ%cÔº£!ðø?ì8ܦ "ÁHýÓc¢jÄUÖϘÓWþO“I®^ƒ‹í¯ÙÕâÓW#Ì@šDÛ#§d¶ÙápÕ-nó¼ñex>³€àäø‡ôåDiÓ5ë|•¦Ô¡ß]'~çå82«uÞr–RÇ’1F¾’<ËÛµ %„ÓŽS&²a¾Dpq}]< Êϧr©^d7Ù•’0E¨-'MØDo¡N1Ïï õ‚ÇäÖì ¥I/Ž}ÿvfìhשö½Á–q56O6TeTžÖZ¤“ÙFžŸHÏï@.à㦖“âù¸{­ùÓãV+.=î‡Aæ)N¿jõðÑ–—¯Í@ x\Ÿn@™OZ á-Ÿ­µRPáAä£ C²Gá‰.—ƒ®áøD.·ÏOtA££r‡B`¸«=êÙL Ó¯g%˜¯‡4êøH‡'Ѐ»¢sð•6äÄŒ¯:æoî´\¦¥/f¾ À8ÚÀÜ0NPÞ>q—SÓîÍLÜu.mó#aÂ?Eû7ä¯"¼áxÂUò  ž¾Mß§ÿ.¼gÈÍü z•)¯òM W†T^Åsrœ½  ÀÙ`Ú£ 1` ‡BM² ÑÔôÚÇÀ 1ÖJdËÄSé©_× ¼ªÞDe>Yêb­@m³Ûä’ÝäóF5Ý%6ÍÇÒŽg¹uÑ5ãÊTwA7›É9fƒì].[´”­g­/v' 5™ë#e7Þ‘ª% ©:‹³Ô.†‚ü¶íYz²žíHpÿzp‡…#kFkÝŠ*ç§|P߸Ëbö2¦¼ ¿Ø­7pô³.jŠÚ\qÒgZ:Ç×¼J\â%eކ!ˆ”‰¤. uE‚Fqí¨úd>Ä¡DГz}å Or;©å»Z=|:š‰”£ºh3ò&…  Ï%Ht“KùP£´àðõÿÿW“àÆÝ GG¾Ð;Åž­ oÑ6u¾&'‰I•k¤]7\ö,Ê*…|œ*Ø«W-ÁÄ 0¿Ü\wYçQÜÌ}Ç{¿ETª­Œ(RêÈϰ³ª'¤i'µEtãÃ8ê6‰uU5'4ãתÏ3ã\Vì¡iìF碼ºÏƒNÎE­K®hE™0 »ÖYˆá1RÖFÓ€¨1R…iP¼®±¢_¥5I,$‚½’´‘ürÖ‡ú!,44{d x[¶FãÌ‘pÉé)^ûd¸&/áÛ¶òçlk²Â/W {ª0Fub†ì‘ê $²‘ΉmmI’Î¥ƒŒb·81ç§#¿±vO½ÞþëE¶n¦;aøüý³KöY®«Ì¬˜KY;’)›9Ф,u®‹oÇ/蔄uÿwÉ@Òª$›ý¶ƒ*“&­½ñžˆ­P¼¸£·¥õ·Š¶ãÇ| Ò@xqóP÷@`.Žõ,Lü]ŒRyYXÝóË·.˜ŒŒn[V9Íáa°ËT–‹@ ËNìŽËÈ+BØ’ý{øgÊ4Ýo±‚&I*ZËØMû¶Ï˜Ê¶Â‡ Õ= O¥Ï“ÝÐÙH¼€JÁÆ®0nš±‘ÿb߀˜Ãò‚ïq˜h ¦m¼XtäY\>gÉÆ=TÒ±­§ !B¨ÓÐ0I$ÐÞ©à—áÄÛ‘÷O°9K}5ØŸ%Èñ”®Ù$à"Ýx"huxÅh¢”â¹q«Ù b—ã´°ôCãêTE¶Õ·-cŸ~&üÎw‘ÆÃI°?8ׂä—I Î[â-¦nN×2g`•ÖÁ¾ÿÁ&‘Û(¸hgöX ï!5ÄM¾6föcŒjÜa‰©O8¤õ‡3óÕñNÈPÆM÷Ll§®¥|œò±öÇéŽ[óìbj­Ýe1b|nTªh¤©¡¹ð»Ñ’ÖÌ>xcññ¨õ͈£χ¿U3`ÂD éã{í/[ïsmk™ŒÑæ©ß Ç|“ŽÈ=`4pýSbÀsë®^!«ù‹: ì½`„Ù׊֭jÎQ Ktülg´®Ï[â×°C8L;ïD÷ŠÚÏ*<ÿ9NæL‘dU^‡÷]αµjäþEeÇ™~tzþ´cª¨Í9²aþ;—®pôÿ“Xò6ªR# ç~]™ªôPJzÍoÇQ/Ÿô0™ƒL±=×W–‚¦>k~ï.¦Ù²ÅD¹ÄÊÖª™¥D©þò…!ºýÂÐf5†+:¼«æDÉŽ›Ó&Í%º<šºTgÂîM•´#«ê—MœÖ`•_Æ#]Äïúø¹ï}àè=¥0Ì‚Î>{B·,?ã‘Бy{Ù' ž—­Fœ7 8±ÁÜô¶pŽ Óbv ügnwwVi‡1`}CWÔ„„Y¯Óia–‘Ïâ2'°D®x©ad®Ö>Â$?­kIa"V4¬æã š´²}ÝDÓ`Éo—ŽÝ;½M®LçàQÉ•Ç ‰£gH7Ÿï6Ë阜–‹êM£&K+Ç­+Zžb(ÐÀ>Â߽͘†³I¬AÙEX5/ƒ£·»²-d‹ •fñ+Û&H’Îp™F»óPÅŸåC>ýEoˆyá%3ò¥¼I©î­Ó>®ñgгž]Õ·3l*I»äxêÓB_"íÿE9(‡ì½ÍR(‰+CMüÛvx^’|Óp½b¤€0”,Ÿ<"®Æ> Éû{`y4$ Û“k”û†VëWÛ܃ƒûEðô>r€€N ]|²Áñ6å0'jaÊYO¹ß¡ó™º»°<~ê·mŠqL‚ºƒ÷…‰uug´¡œíŽjõác#œëÕöÅû‘H¥{!´Š¿ ¦yC)Ov? Œ—·ÀñâY»i²}*åqÓW‹j ú˜!¹‡Rû YÜyÀü« ‹Ëœ;@Ω»Ÿ yR¬ÍîÇ+ÙsTñ©aÞ?d£^T³ €?ÏRÐŒ*%¢puªŸÀ/sŠ ¤D¬(Lþ³1•.·'¥7à<Á1ð÷(\"/(m ÅõÒ^oý@×òñi›«6U™LÔ~JZLˆ“S?ËÎ( Ÿ,ŽRDB–®BcûKí&ÇùáËÆ-eòHe%| 3ë|IÉ@çÉ&l)gUÙ»I³¶¡{SÇ©28ó̪ü–JÅÐLJ'„Ë/ÃVbij¿e»,/0#*XÖ!Ìæ¿~  ƒIm²öî@@4âyR¼nÖ™èË7£íüÅsäÕı½˜g˜yDªRª$m£l³“Â~~Nb"—½w#Ád;ƒêrüwdŒe¯¾öy¼AÙax!‰WåPl{éô G£Z .zîøó(ÎßrÖÍæ1Íÿºœúj oø´ "ù8OZØ0ì¿QnÇN^'s+®âºd¡1©„*&¿EºÌhì‡qþ‚I©v)_ öqCÂE²î;W¬±ÿÐÁô£c¯M ÅÀi°3Í"_£5­4*c#À„Rø†€Â§Ã@Šÿ7yovkíÛ¤¦ ÄD´éHó¢ôÅøŸ¼¼.ô¤õu-»óxÿØ­´iè€IäH9g•¿™Ê˜ôµD fšý°ÒÓ* ð£ü aþÍRYüºy¼‹Ý›ôÕ>rf½³†CÉ%ÁqRnyUÙ#.E*p¡±–š° uáwu9Ó){„ÂNQÙÞƒ¢Xò½Ï šîp;èIóJùšñ°¸ëîy)â‹ã7P)ké›"nÛ"õë¦dð™óÑ\ùî½;²ßPFEµ»Æ­åñõÀ´~JsÛŸµD¢Â‚@¡òðØ•ø¨4Åê†ýäcgR I.›ËñäøŽ„4üÖû˜ˆ´ñIô{øvõïýt'Ú9ysÚÝËÍ&Õ>õÏœ•Ù·k»}<%EÐ<„ªw3ºOq a óU5k¥Èz*Ò3£æçRó’Ipd"=ˆkÔa$uSó®t P}“,žXãêQe3ê¦M¢D‰î«úbxGàŒˆ´Ï3¹?±cèÑ13‰·y7êÄkó¼”%¼¬ÖšÜLk‚Ò¢ˆ¶ÐLˆpSŸ4~Iôiø7—ò=ΗP·Ð“œ:`ÔƇ¤bEHÙ6ªÕïçºyɼ]¬eYеy–ùõ°l<— Zœo9ŠòXP þþM!ž'ä2 þ´&¡t!Y·ÍŸ5ó-ßeD y0ŒH"u!%ïÿyû¼÷Lùûµ;nª|A];D¹FcGq 2Û˜¥@<ó ¼Å(—ì•eÍ(2žUnx£Õ•gÝüñª2Û½ÿò¤˜woÚÕµËÀöÃÅÿxëåÿ³²M&ÿ/g®ýa(íš××Zí+ |ˆ†„å±Û0|¦¬ßÝ:Ñò¤<êäÑm‡¥f϶Bçá»÷®hG2ãG¹ª'ÁFÖàm;b VĄҜÁõYG<Þ±’öAtž#ÂKriônxBíÖK®\ÐAÔÏÍv­k]cÁ-Bâdbݽ5ÁàRÎ|qP ­|˜ÄÖ™£uuÿóö9ŽÛLì¢Ö †f}ˆ„â[ÌÕºŸß~¢sx<@G䔊´QD?ð“ö#Ϩ— Rûˆåù²åq0A!q`jÏà“¡ߧ~ ÖÎt\^$ ¶R#XªÚf+)µäA/RlŽ v‰E)HGá%Ìcít«qG+j}kÞ‰Êp_yk®ä¼'«;9.S±£ÕGt÷Q½ûø–NP‰WÈ4Þ,Ì„arMŸ˜ §Q=Ÿ7ùtÊRP2òˆSi«·½‚$‡#/Ôª‘S±¾åèÆ,æ¸,ó¦*JÏòLÕùÁºø”).(‰áâ6²ÄàŽ8ëPs“¤$˜~4Öƒtá*–ÓKr— Ãi=nbîí 9Œï?èÕÈz¬8,6dwº1+¸E} Y)‚ׂŠw<}ãCsvmßÜ× Ø©ct·g|WºQá½(\“ƒÌŒé'A&óõ¯i- € çNmx26T“Ö­ƒŽ~¤èj³î7SÖs&ÄÉ[¥ ÌñòøØï9¦¯9¡¦Š$œü÷¶=Tã¯VƒÔau°Ç*{ú–hÔÏØœ `݉ÞZ·û ÖÐ¥Zs°* Õ~öñãz¹}À1‡9°xT•i‰_oºŠ2ñW…€ÖéÌE„DÕ`/A¾“*ñiÁD]I¯²ü”C#A€¦ˆråõvÝ"‡riO˜éïðd±ºcp+jžÊ‚Ϫ^Ò£.[ø ¡wG¡®¡ým T•éåÀÔI†i‚EøJ­/IjêôæÇÚh=‹®»ìúÚÒm>wbc®,® 8»РÔÚóŸ?84ºV’ñ ­S´ç¿ÎB½ûpº¹wæPU Ü÷Fz“_J¡õ5{fò†2 ‡ÉmgîU³25ÓóôäSñjé®Sž]Ð3CN7wÕ`òr‰}X®Šô[Ú»xjNçè ‡ÐÛUA½²d.õ¢àzÃßuåÈ¡cãä‡d ¢‚·€Vàh/mÛJKjb˜,þYO“}Å4ÓD'b¯™ÍçôjRѬª²p­êÏõ~F=(¢\¬iT‰Šöïž<þ°<б]诼×7Ënyë|i?µ5ã;ä~Ö:ªÉ—Î’YXÚÞ Ñ¢xXÈ»0È„ ®”Ò¶H^½ÛØ~)ÿƒÊKÉôxõ¦ç*ƒ&Ž®#úLC&&keTáÒ.…]¡¦âÓû_Ó£ý&A t‚5ˆ¢XêòÔøÉØÜO±šEé5íq&þOüÉÉè膭+´îa¨p¾ùqƒ?1ˆ‡‹]ý"•’è2!C¶øÁQ¦K‚µ„|ŠLÙ+Õ”Ò1oá–ôÄyÐH"¬åvïS¥s JxAvzx÷9f†ÄZÜr‹åç$!êx•®–1&'ßÛO{ÜsQÕ¿ì€T{£B÷Xm©Éêó?(AòvÓ9„¢ÊÊãQ¼ô•]A™©‡ÏrŽžR^81#˜W¶ÑqìóMa&&GFqYáOˆaYŸ¢Q¾„ZQ¸d1º|ß Ba‡=«”ÿmN¨°”e¾)®‡T½}ÑzÇ™²$£~!3:¹Îœ*@3,á”#p0¼úùíiØ©G³°‹Å-.ûU)ë¸þŽwÁOÒ™Þ%\ÊLNw`à™^ÅHÙü}r$ \ç%G*ׂ¸ß‘A· Wý¢u ˆën¡¯M`‰›ùïö°Æ§º6:ÍVư› }ŦÝ?Äpš 9¶ ,Cáˆz¢M Y BÑ.[ Ž‚ †Ã7Ê7ÑŠLëwú;&°Fú(‘³À…`#ýíʲè‡ñY‰ÓK)~Ì~l]«÷#‰õf~Ù~.Lä4'Ô~H–KŸ6Úý¿ò³ О´Æ¨£¢àÚÆ|Ã<%ºª‘qذ'LWŽ2uÐÙ,Üš÷¦«VÅ«ˆ\·Œ]ßÎ8p÷ïBŸ"ñ84k/Ú7eyôPb¯Ù³ÛöÎÒ…eçÃiis„ÿ»ñ,2vé'JªuAIS•’q4¼6himÅ ÷"¢-ž«%k”¯MÞBÜŸÿ¶’½ó×èᡪ2êlNd…£—'µ¼ÊZÍ ³T€ˆ¥àGLz[P >0 9 Òu÷¤ŽÚrÿ»½o$ÓÇDÕÞ¥ÖöhÀ ^\ ‡üUŠ¢“</ö•/¯—˜BT¼¦$ Û‚ÔFòÐO Ù^”x#Úlù^áßȃM„lŽ£g©”†’£Xñ11«ð¼8Il‰!E‹÷´|'è͉fm#(òò®PÅ™Âàgó’Ó|ÈôÃÎgXšœŸ»ýN~{·ÍeλÄŠ9žª^&g 4ã»:†°rî˜aòFŒ½NFwÿ2ðÂ"‘?Ĉ'ÉhéÃÂÛ¬Êçºê§K'q —ùÔRœ¢Áà²u±ˆ°qÞ8r+º„!‚“ÔØJä[$S´µžbº¡ˆ*:¸õÐ@,ћ܌æ¢ñTÅe\B¿®ì‰ìÉñ挒~I¦„%'sM#q>±NnÿL~xã=_Ï5ãX„ŽË–Á‘ ι#ÕbvmrM´:ÿEh¢·Pvtr0íΧù_Ö…}Ô»EÒÍR®-a—Ãú¹Í›ÏÙìD}ë@q„+Üümø¼IÂh9Ú­÷ÕUnˆŸ«ßño aàEá›6ßñ%œjqe#©Eθ…ªŠÞYÙíÎö¬±D†d±0AX1(yôB4ÓoOt² +‹Âh·ÊNÏù‰T«ê‘ø³~Ñ]‘«Ô9Ÿ7›S¬ÇûlFЃ«¤Áˆ¼Ëç‡îꧺmrÞ·%~ˆJ~[4¦(’_ÍrVí·ã‚ÇRóœ _Y7dÅoÒx_î˜Ñ’z,}FÇ 0þ:hV8úÀ. Q‘èÛvk[¼_TÏî;-iŽðWê¡Öþ>CÝ£b‡kAˆ_Ca‚’˜Ù¨‚ù]QÏÖãêWÂZÇØÑ+ÖVŸäd„Z÷a+¬R7c¬_jîò¸Â˜Lhg™ 4ôÎÂlN) óq¯³þu£Ñ2šÒm$7õ‰³Z0™Œ!˜¹'Y¡¦”´ŒŠ®˜Ñ:õg½0ýNÒƒOmƒëkR>À[b«M\=äØOU‰qi§Ë¬št¬ïlúUÕ>UÈÎÛ’Ø'‡û.¹{ŒâgCqÜ“I_™:X<mÑV9¸,“ŒÁ.” Mlk½yÿ.†î0;ZCÞS­0âÝ-¦ç‘Tt;Ä=ªG šxG*ò bNCwµ’`Œr()å5õQø¡@æ^Á[’ê÷hðù(™R8¼ªo¬i™#oM!èêåbûé P¾?X%pð¦ï<ëÒ“)@zAwŽe0:"ëiRb³y£Š¤ˆ³4?ÜÊ”ZsTV—d(?aÓ¨ÄâŠ1> —7Òoœ¸ øÙ Ž[óxãaFQéÆzäã‹­9>Ì}Öhùe‡‡ô|N}Ð:„ÿ§ÙAUµ­åÏ10a¯kËŸDg°k¬íNæ­n‚×ñz…J¿ÃŽÚË—¼)ì%ÄPfWt™Q«ö)JœBŸ>LÂñg‚°^ã£|wÉ«ÃÔu5D´|âÌ–X¹ÊgÝÀG~Ó¾êÞAz›¹£Éý‚‘¾ö÷èC÷§CÚN]=a{ºø¥›‘_¾Ñâ%Ö`)HÐb€l¸âŠ1ÃÜüÃÒqƒÑa´ä~z„hŸ¤”«ÂM²»è­rn€Å4ØÒ|W„†\Ý fîr»ŒZ Ü¢ˆCÛå©Fqн¢%Ó®¿Éë­‘<‰xü'lÁ_¼SUèdXc“÷ûö ¡Yù}€Ñc¢®PT<ýƒ* º›Ú[7¦€€í-ü"PÏpÿü&D£SÑ8,îlýÙ€þ£{kדÍBÿ¹©Ã»NƒGÇ(Ï?¿›DVð!I.8r lçIO^¢‘0rõ›“<ÌäÕ¤çŠîÓŒȆ`úÉãµÆ“£êÖbËÅîJ7 ˆöAä¼è4ŸÀÊš‹C2!÷<â"ךÌ>°zî)‡>o¸zéX ñÖŸÌxü·,}›ä8ø#J¦É©Î{Ìþ3ÜÁ¢ÿôçç(C¶õdú ã¿GCá±4ƒÉ’Ü®oÿ‰X>£xÞ5zLû|Ê7=~ÛdPè-œ‹ªÙ´aG&ø‹ÔZ ÷…'É)_r-ž§5 h!"Ö§R‹ØÇú§è`A’…õYø‚׳TÿRÕeclTA„hÁ‹e‹%ûwë&tdßèèîE%qÿ1Ç $e0fª[9VîHb°,ûwæÌѥ⠦Rs³§Å°8£Ãmƒ.$G!´_µýÚ«à+WZ^ÑHjY_¼¥¨X¯zìou¤»SÚ‘/aF$ÙÙW°ÕŒ²æÀ•ÊàÕ¢ªè>lŽ ·§.îÔÔ½ô8Ë2lÆ›šUZíûDŒ¯T rɵIa=sŽ 9yù_pÀ ?|F'ØEöþGÖ¡± TË ­EBÍ …íàšÝÌbDÂÑ La+Œë#ŽM*ê…¢[ü¿€¾•žAR”FŽE—–èh¿{‡Ñ¬´ö±}e!&¸k®ûGŽ^ êR;¢S&Uã¤ߘ >¢ß`¡ãŠ ÊKå±Óh F¿5›Û£àSÓ7 ¼ÔÏŸ‹€—×@&) *>i€2öP4¤eÅé1Boù~ùÕøØ Äjÿò©t"¸_þ8ÇÆß ·ùáÞ,µ£mÆöa¥`ãØÌøÑ€ŒÂ©ï-ùÕ&¦Ôm [6Ù–8SH§ÝÆXY£RÒfi8òŽs×ÂhkV¼zL.k&aúñ’3¢°Ý0tÇ’/¤”??íG'`²ÀX°~ë¨o ÷ ãm\J1Ï"QŽˆå \¶äJ.¾a~H›„=í߆d¸›nŒ¨.ÎýE’>jxFðÑ¥¶\.\åvÀ§ÐOÓ‡YW¯7V§+b†s½Ro2¯My Yu—‘áŸ<Ý—NÙÁ‘ÕOt]X§ä¾h„V³jãÖ…¤yE³á¬ŽgÕOð¡Ë¤JØ£æFY‘{ðej±nß¿‘&Ø#uU&Wè¤N|x)‹ú‡…ˆmîP$Ÿþ…é&[þˆÊ55+ñ0ÒžoÙËîG‘ì]'ÃÆè\åØŽãiñ`² 7‹=œOšNJƒFtü³Y;ÇGÊ5us¡²ô…ØÛô2~jJA»ÔÐŒó¹÷Y>àá@G%,ŒˆüýÕ}I÷`Úé^»ª&·Ñ[JEî…°ÌÖ]Œ–3áqÓ—Ï!-_|Þ;t×vfºvœ*Bˆ.¦Š’74'îR¿økü2“ší:¾gÍ«¶üKá«ÖkôDôÈß]‚ŃGÜ|ö- áþLjâ›igܼT9ŸŒÿ2{CÌWÏJ!øÜ%ü)à:M®¶çjÕ‡BÎ0]z$f9vÀ*)¼˜šwÍrC5«éB¶†© û6)pc2¶@šHŠ$\ãÝaÐYNË=ŠZ üÞiyòeãw§#¦šó*á€Ô`ø<ϵTÀHD¯£”Ûß Zç>c°E"ô^¼ü7?6!‡YÐÈóÈâŸ^CÝ bY¿(~þ!3_.- ÷2ä¶,  …`Ì”½VàâÛ¸WÈ5þv¢UiAÙ¨|œm_pƒ@|›˜¼t¤˜Ç4†%$›ÍËP@(§ =ÒZU /ž8¤},áR€â £{.ùŸ0dC&©Ë–j8¤d} ´Ç¶š5’ÃõÈØÂóø»4³Þäð[ ŠŽX,˜Hßå ÔüÚ¹e–¡x‘u7+€"é …,5JE„»é ôùÞI¿îGzü°IÏ&ˉüâÂÓÿö&d{I*\Œ¿fÁrM#õs÷¦ùQf4âIp®Z]m<þ–ʹ÷‡ù¡}ÁQo.u Yum£Ãøo¾¿¼?üp«¯Îˆ€{ßÒ/f«F 6ÐôÎ̈1XD8.›dÎg†W·Ÿ#]UûsB F̾N¬5‰y¤h¢EVuÊVW¢B9lB\õ†vÑÆ-Üklñs&Oæk³=äí¦g"-¾ôG‰ƒPthÛ`§-SÍe8!?è϶ÅÆ”»xe• –fL™îÁÒ=(ñÓ%îDòÍœæå'œ›NDiæ€c…!Z¼Ò§¨Ð9&?v²%Ü•Èä×%¹&_‰FÇQÉÙ”HßìýÑ㆕œ>,4¬óé»ùJ© ËBГ­¼æ}Âëˆí_©¬®„XzÝŽrC§ªE{œIþÏ~&cÈm~’&€q3êo'å@,F2ýö¥¸íËÐ}BIõ×¢õŸ˜·Îw4D1&ÿíý)®N.´+9å#í|‹ßQlï³F±¯ÚÁL××_… w”»Þ»×7<0qÉhÙN×.×®Õa¶·‘_QDz9]‚”= ø–×]—:öñ›òΖ?iqf"H€<·óÊ\Æ4êþ –S±8‘÷DÞV>wãV¶Éz\EàóT®§éE»Žl.íõ#jZŽ&Y¹Ø®ð¾\‡úˆÁhɧ{A©ÿÖƒë¿xf xJØt¾PntêÖ:‚Ѝt)ß3—j,©F8LẾ°M= l7l»©ýÇ&4wÙl÷'íy;);B;„*ü–T쫤RÓ< }Ü<÷ì‘™[^dÖb`# ‡ Š !óDœóMþ‚C›Úä$ Ø4¶ï{TfYËRúâ—ƒô™¶)h’õ°ý´u "NO5¤P@›J%©6…ü©L¹=íÂÛƒTY$\º…X=b÷! Q„ª>,Ð{O}þy6ÇÜbñ×Vz¯ž¡œµ™¶ 9q>ß;X½¥Æ¨JÕɾ8#õPÔ=Ë+‘/JOæ3Ù•ý?Ko©x×øgðÏËR…àÎȲùáÑ£ÙNúTÇÓÒ¸}  5[V+¼zpú´Ñ¥‰Ç‚n_KtN¢ç3fÿûžpißÁíƒ !‘–ÉWn"~S¸ÊK‡¥U–!Ž£´PÕ¯ë„M›ðÒEâü«“Òn7€R„K”?– )t§ùßçMÏæñ: ÀiY…½Ïc´K„ÛpŸÖ%³lL¤†IÍw‡,œL›’_ ˆ¢'a5 c0_®s ĪÌÌýz[± (2_¨{®¶®«$\«’k3Î$ÿò¬aV˜@£j0 Ñµ[pâ Ôïçž\¹)ßßÈ'xA ìºzÊÏtûNN›ÈÂyŸ²D7 y4nÓµÛHqóÆb‘‡þ:4õ' jÆÂ¢Û7Ym„YŒ,tGÀ ¿¬or A¾ÑŠxm?øk!íÈH\°³Ú/:õíîC¬ä3ÞysɆEûE…­ÏÖ˜®ŽøLõ=Í«Âkätþ‰ÓÓZ'-Hm²‹†^ä=¼­rû¯›ð@_Pèòó6 °šþ8íPœVm Šú›v¹½$ELÛ[>ü4]Òìý¬žA¸»òLzÙ^Š(í“ðÅó?çý,s¦Xö³äzd+«_ÍÝ_=Mˆ% Èž-í?»~Åþšgu–=µ¾õàóÌî“Æ3t œK ¬J&¼¦À±Î_5š…$x”K[+¹µ' Jµ6ú¨À¾{u©ß³Èb5¨Js ˜é³Ä0 ‹YZgmm/data/Growth.rda0000644000176200001440000006477114247643115013733 0ustar liggesusers‹í½œeõÿÿÜÞïÝÞ[v³›Í¶$»é…ìÍfI–$»a7KHiŠô*(RDD~í½cÁ, bA@Ä‚ůŠõ÷ž;ç<$ƒþ_þ•¹rýÎ}½N&Ï33O9åsÎyfžÙÍ‡Ž¦McÂ&›p„ÿFÃü2Q“âßrÚÉgŸñ c"µ”2PΘìŠÒþS¸ )EË@±2P¼ ”(%Ë@©2Pº ”)eË@ù2P¡ TUª.Õ”jË@ue ú2PC¨± ÔTj.µ”ZË@me ö2PG¨³ ÔUZPê.õ”–zË@}e þ2Ðâ2Ð`h¨ 4\)-)---+––—V–V•Ö”Ö–Ö—6”(m,—Še Me ‰2Ж2ÐÖ2ÐTè 2ж2Ðö2ÐŽ2Ðth¦ ´³ tph¶ 4WÚU:¤ thhOhoè)e ÃÊ@‡—žZzZèˆ2Бe £Ë@Ç–Ž+=³ t|è„2Љe “Ê@'—N)Z:£ tfèì2гË@ç”Î-=§ t^èye  Ê@–.*=¿ ô‚2ÐÅe –^Tº´ tYèÅe ËË@/)]QzièÊ2ÐËÊ@W•^^ºº ”“½ Á/ø¿àü‚_ð ~Á/ø¿àü‚_ð ~Á/ø¿àü‚_ð ~Á/ø¿àüÿ •‚_ð ~Á/ø¿àüþ/ý‚ø*ø¿àüÊû p7øùñ+‡^ºûïìm~Á/ø¿àü‚_ð ~Á/ø¿àü‚_ð ~Á/ø¿àü‚_ð ~Á/ø¿à÷DüþÓï#T¿àüþ³¿À΃_ð ~Á/ø¿àü‚_ð ~ÁïÉ÷“½ ‰w@ï„Þ½zô^è}Ðû¡@„>}ºúôQècÐÇ¡ O@Ÿ„>}ú t#tSÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐGÐÇãûн ñÆDÿdLì—üß)/…²Ð8ô}®„ºùÿågLò6cR‹ WAß1&Ý ýºú²1™´º bü®Ü ý ú_ú냠5ÐJè\è%ÐyÐñóŽbL˜þÂ/‚^½zô&è2èmÐÿ@ð üE®? Z=SˆóÑ×@Ï…f è©ÐAÐ Ò}DÛÓÏ¡¿ARî‚öB×Rþ#ô0ÿOA»èg t ´::*BÈ$‚BŒ!ò,™'r‹ÍBÔÅ]ìð‘vâŒ7qµÈÍ‘ÏÁcM-0&ô:ž…^] ½zÐYBç@φ.…ÞÁ‹ò !ïÐW!dú&ãCÜ¡ßAß…¢|œ[WªÿôBéïMB×AóÐg!ÆBoB”ãç ?CÈ2ôôC>…~/å{i·1¿0:~ ô2‘Õ›¡ûàM5Ô/üB¿"ð5/#Œ#òˆ9EžÝ ÑÇ(¶ËCð8Æu±ç@_wçºAx¯CèMý CŽ+rCGõœ¿Mƈ>„чðë¡/Ñ×[!t%ò|þEtu#üMѱWCï=¼RæãÌö¿]ý¾¿áêdøÛŠ$9þú¬è³Ó'z> BŽáßpM/Ç¿B?…ài} ? ºÿ'áçebΘÆEÏ“6Ï}¬b~·C§ACöB—ÃèVx9t‰ØÎ ]½-YÇ‹>†Áðwݾ#˹ahHxzt¶ô{™ÈÖ|‹ ¡[Ña±ñõжý2ƒ¿qúÿ8ºÓâŒ!î`S„#÷ƹ.îàÕ‹!ô,õ#ÝM£—éKEW[är±Wt8ò{±7GWÀ(÷FÁÃ(Ø…×QdïbèV ½Š1®Ÿ!Ãøçé ¾%àa¹†b6Cõøê“#z‘£Ã_x•ÿÊ9ç¸bþ¡Mûü !£Ð ‹ÍId˜š„cººB³èuœÉct°€®WÍS˹†µÆ´‚ÓM-ÆÔƒU5`W üª‹«¸¿ »¯C ŽAÍÐóDnß]E?ÂÈ$|‡«K°,r(t t*„­D¾ãâb C.¯#àMä®NDæa‘7Šý ‡‘ÓE>Ž-ƒY‘7öÂßòŽ0¾èG|C}‰ ÏQd¹^üÄÛ!x N4Ñß^±iÇfjD—ñᜫïaÆúGøn‡6¸ú~¯ØÓoݱ†ÿÀq™´^|µÿža‡áÍ‚ Œ-¼:Elå¹ûˆl~=Ãm«d3È'ü,×NBwÈΑ¶Nwm%êØÈëÇ®€oì|92–Ø]è"¾*fű±8zw|y“ëï“ècY'Ñëø_…Ð×8>&îØOJd¾C°Å¡cÅ>ß&xõEá£3-‚Áô™”ÿoƒÐ£È›D~ÜÁ·FàK_AïBøíбâw^åâKè~èÇr¼WüþÇÁÜðvþ£ôãàû×EǸ'‚Gc® Gþâžsü„ƒM.–ƵÂFð‰‘+E_¸? ¿#`H]ˆ"û|ŠüÐm/šlÀŸE>+ºu­èù¼è6ü‹^ 8µ[{ˆn‘8á\Á’S]\‹€E‘{EçÝ¿TÊw W¸úë`¹£s¡×BÈ=„ †N—òÇÄ7Nü‘S~ÔÕ¿ÐÄþEü,ý…»D7â÷pχãPÔõ¿¥so‘8Àñ…ï¸àJ‘Ç[¤/øz¿ÄIlàøbºpÛN8+>ýn]dÚ•ã›Jzs𨀣/øÙȱ‡Ãäÿèd„x0B\=Ü«"Ÿ½:StàLááŒÛv–¾2ØbÏ"ÿ,ø˜k²àf¶lÄW¦°§>6Í¼ÓØE–ø,+€íUôY…Wƒt¾p!„Îoá}+Å}èV\ˆÿD‘aÔÁ)xc·sŠ]+11c‰ÿ [D¾ Î%ÐÅXx£”ёěÅÎo~9ñ㯡¿J{ÜûžØ:¶gìñ—¹8PòU[$^]êÆÆ‘×ì£gO7®OÆv!7~Lâ’`D ?•¦–#çÒð#‹Žä¸?Ÿø‚<<Ë£ÆYÅܪG9fÁ¬,¾+ƒ<ÓèbzHtÎÁA憿á ] uð´Dû\-õvâËû$V8]p؉{е0~6üãÆðN¬²DtÎéç@ÁÊSÜÿ‡~ââI©ot3Ú)þˆqEÑßè=‚£èK ™Ç¸.özÁQÇÿÿB®uâ“Ür ÝŒaÛ1ô=¾ÖÍqœÜ&$h31)ù ¼N‚‘ dDæ)æ•ÆÏ¤á[>f¸>‹‘d™w;ÎÖK ü-‰¿™Cè©®–|•c¿N¬v‰ðÉÕN”¹:ØèÄ“ø…Rü…Þ„ÿWìÍÁ’Mâg“¢;e>Ü=[tÍÁ.Çβ‚W§Êù>¡£$¿zèþ+¤ äeNÑ7KÜöb±ø›—8>ÅC _ÃîbN|Ƙã?tc¸˜ÿŽØÁ#ð _œBÏÓèSf¯Ïdh'ÃÒıYr™(zý”ô½DpáA×”⌢à4>ºgdÜ ÒÁ¼Š|NÎ_ïú÷Rñtñ`~˜¹&é7éŒ^§°¿4>+Mü™«3àx†¼+ ®çÀ¶,>1_óèaŽks”³Øk~æð_9t#îäÁÙ±Lî=.V•|îBÈ>îäíÈ:þ97§KâwRô—Â†Óøä4ö—D?ÒÌ'…í¥‰!3`@~fœ~у üL9q!|H"§ü 9ö0%þÃ!®½dŸ|ÍÁó¦ß.~ã/‚íŽ 9¹O£äjæó…àOÈÉ¡°ÿÐIÒ.~1ôF‰…~)±–CÄLáã%¶pl{èïñn\SÊñn”ÃÉœ<é ‰Cœkœ\ûÑ=G×>}Ûµá(ö¥¯1qŒ¾cŒ;N]ž—’ØY’þ’Ü“ü½ð™&¾Jƒ=iÚL¿Í]{(Å\S2¾§JìƒÎ‡Ÿ#Xô±Á×KLt®Ô;c<ÜÕ±.9q$þ¦”~ÓÅäR¾Õæò¬?-u¯ £s¥|z·ðƱé÷ Æ9qãĮϓ± »>;ôs7Wuø\Š›¹€N>[ÒoóXü ᛳqØrøÔ3$~q쿽WlYG;A'"·HŽä¬Ãl…ÈKJ¹·ÿ9ñ¸Ó¿Bâ¦ÏºëäÝNœÅö£WJœ„­Æ˜_ ŒŒ¾_°Ç9÷V7>*añr÷šèG!GÏé#^d¨Ë`ãYæ—ïó`Ia¾‡ûj°éÚ¯%~ªOkÑŸ:x^‡N5 ‹xÑ„/h¿›àI#í5"‹&ô£ûnƒíð·ßÒEÿ h§çbcŒéE¯ûЙE䊋‘Ñ ñƸºä&Ÿ2 aK+À®UøUàíjú_¯Va«kÑ‹õô$_I"¿ä÷Sœµü]lM#ïô»]šA¯²ØZ ÍÁ,óÏáOóàYž1æ)ç[_š%>Í"Ó,:™=Ååg ùÆÀŒ˜¿2îúCÆ1ô9†|QwrTŽ t3Ùé®&i'IΙÄß%WºksÉ_ºëtIø–ÄV“Ý#N|'W‰¿ºEò^'…× t#n%П¶x‹¬G2ç$þ S-nþ‘pÖ±‹8}ıå8Øž ßø+$o¡í8}Çñe‰¤ôA̘·`}ÛI‚IbŸ$÷'é'_R·‹¯eið?E>‘oŸ8õt‰Uœ†œ%ò>7>uì*⬿¼EòÉ[œ7Ê5 èrÉ ž':ÿQÑ{ÆÅÏ•ÖO±©(íDñ 1|Qì3.–ÅÀâ¶ûÈ\г±g¹ü)ñ»Ú]L¢ÿIüy HÝäÆ`iøž3³ÄœÆ“A'3G»~,M{i|V {Ma«)d˜š>:ú.fÀ© üÈ ,íæm¼ÈáßòØ_¥*;TaÕ̳½­j±±Zô¦|«ƒõȱ‘97c èX=öQµ[¿Zâ(òÿ$þ1 _“èp ÜMõÈZ1öž‚GIGžØYzÒõézwÍØ±©¸³Þ †'àu¿—8ßý)Ïuô•˜3ñTY§qlqRb_|~ê"7H£K™E²þŒÍgˆÝ2ŽþÌõùp-ÇÜrð4ìsèBîD7FΡoù~—'UèO5ü¬¡­r„Z|_=÷40×Fò&xÔ„4Ò_#~¦ÜlÂg7C­ø¥Vb‡Vl¼™ñ5´ÂßVl¥‘¼¦lÂß7cWÍèW3¸ÜŒ]4¡Û­´Ù‚|ÚÀ‹vøÙÎx;ˆC;áy7º½qö3îä;_†¿ḵD?úÒ…œ{Á«>êúã⫸†¹ƒsKðƒKÁÚeàø(º?ŠÃ7­ ¼Û\I[+×*ôv 9Ëzâ¡ Èfåõø µÄëПÈdœc ¸S¨‚75ȨùÔæ븶¼l`þ Ì¿ßÒˆ šO£ÃCü^#²l öhÂç7£ÓÌ»»®'žk º[ÅÞ>'9Ô_$=ÃÍ¡œµŽ8±Yœ~âÌ%Nüœ ®JŒÖ¼^žƒüÀ]GL|Áí[ÝõG—âß•µwx“À&ØU‚Ü+q²ûŒ%ɸ’ŸÝÂïfÜ&ÏQÎq>O›âÇ}Q±Aqð)yø“G \—73Œ!K—}&„oÌ¡›90 îäÀ<¼.`Ëô&®ç}}'²Ïw}F +`£UŒ?÷7nÌ »öÇæâð.~£ä1ΚÑsܪ„ã`Tò0Áb꺕b¼)b­41kúPñ×è|†ùe^ ïbMŸ•y•¬›¢÷1Ú‹ããØtüÃn`\ ÆGN‰iñK¦S¤ácúGî|J~ÏÉõuNÚÍž#vê`Ô‘n,íäÆN,Çã‹Ü>â¿–µáÿùÝ$s¥.^§ˆåÒÎü¾,2}t‰û\,ö¨ä×o•\ücûÄŸpu«„áôÛ"øŒ-ÄÞ.8þY—;ÙƒÓŸãs?æø§4ñb ?”£“Œ?Éø’ÄlIl=ÕáŽ-…®¤wʱG¹¥±Ÿ4m¤ÁÊ 1@ýÎ2‡,òÉÜ'çЫÌ:·2Äz鈛?¤ác’c²×õ‰‰“\Ì-ù„GÜx+ Ff^íÊ4åà$|LþÖÅå4<Îbûyg­ÙÑWð§ð"76sô,K<š; øÌ’¿ÃÕó<|) /yä–?Ô=W…/¨ÂÇЭ*âõ*®«"'¬F~ÕŒ·úÓ`ØW Þ×1Ö°½@œV8Å]£ib¼-Äq­ðª9·Ónãncüí`k;þ¶ƒØ¹½kdžÚÈÉÛˆÚ°‘æÒɱ‹«óµn]vÔ‚4£ÇMȧé`‰i¿ 9ô¨ä#c‚1ÎóBç‘XàyîúFÄy.ô)wÍ(¯JÏé'Š*=Ûx¬1œáÆ»!ô©ôì¹;kÒ¥gƒJ^à°¹4€‡õèxúÜŠ¯ï î¤nbä…àg/ØÓÇXŸêÆÉNŽ]Z+_â®Í•ÊNîtƒŒÓYSí–<¢Æ¥ð_äú!Yÿ;Mâ±Ó$VsÖ^¿+<$†,=OsÚtžK8ëÔØMi­éxÉÙÞ)k!œgh%Æû¸ädN>t´»d×À³¸ó¬š8:ŽgêdíáYn¬#–ÉGù‹d=l¯qˆx½n½øvô¹¶šO#çꈫª)çWö³® T91Â' Æ—Ç6óç»k…ëݵÇjp¬@l[NU#›jø]ƒMWÁ£ìµ\ªß-ñxÐ .@OÚw uUØ]¼«eÎU׊Ÿáž*øZÔ,t}o5~¹Ž¹7pm#:ØL?­è^+ºÛö€Ø ãlÅÖÛˆ¯ÚÁÛì¸êºˆåº°Ñ®³$.ü±ÌùVãSj™Kýb—-ĨmÔµƒkä ðG]èJú¾ëaL ‘OñÌ"ì·ßɹ¸~€1â—†|ÅOÄZd Îɘgüèò¼ã ÷¹H²Jò­¸‹ßI°2õMy6ÿ²èpžå‰Å G¸k›¹;\ŸYÊe‘gä7ïvK~üiYËú€ûü±”ã‡&kÄÇK¤°Ã41B–ö²ºëzYæ‘›p1²þT{1Çׂsqx‡ÇñKÜ5ŸRÎãäAõ; Ã)ð.E]y¤?‡dÜØv€ã]È¡[ଛ;ï× çÑOˆïsÖvÞë®á”Þ»¹Søw¯¬Õ9|Äöc1±Ÿ“äÙ…ÿåÆk¥õˆ“Ýü/ñ×&œ)îäíØHé½!'6§¯Ä«Ü5â$2MÑ~†9¤ðUYl%NNwq´ YT­Õ”k°õjd^u¯£T!ûvC>±Ë%NtÞ-ƒO‘ØŸk“œK9q-qU†œ<Ü3`\ŽùäÐÅñX㩦ßjâ°¢©ƒõÄ|ºól¯´®¾]|ðWÜ|$6*zÅMäÍÿ[àK}vÂ÷.âìnbª…Ľ`qï¥üÿw  ?Ò `ƒÄƒNÄfÃè÷ð¸[hvÚÆ÷´b3mÎZóì_ºðÕÝÄ9=èq/¶Ñþ÷ÃÇÅ𦿳ÝÄv‡Û0¶6ü ÛîX6ö/Á÷-ÁFZÐVæÑú]w­¢yuÒ^òïú‚«§ ‰Õ{‰+ûÀù~ìuò]†,FNƒ(û÷P¡Ÿâÿá«8’§,åÞ¥ÎzÞÍWKëDŒ7ã óhÄ×7!çâÅfô½‰ü¥ ¼hAß:¨ë Æì‡.À¾`+=ð­oÖÕ…ì|˜øi™Œ§ƒs#È{ ¾2÷J7¯ÏS_ÀÇÞ ñ„³Î^Õàÿj¹¦Ž±Ô3Æz0³Ø»>7p®ÉÑ0ª}jÁÆ[Á§Vð§Å‘/±QV»JÖêÈ7ëÁ¤ð³‘kšÁÊì¯ =l‡Ïí`@þ¯kÎÅ“.ì¤ êƒôÁ¿~Îõ»:VÒ r™Îw!ó.úì`,ø“´ß…¾t`ïøÂž‚I¿äæ2ˆœ‡àÁbô`;@†ÃðzžŒ8ëñÌäwM~”x~ÌY'w›Å6ËÚ·×SG÷ºùuŠù¤.“w𓙜»®/ Ȭð7vvâïÒ~³0µðºîËnNØŒ73ŽV°£ kG†Ü×…=t»~y!zÖK\±=é'ÎëG6‹‘DZРã‚?CàØ :3À|ˆó†ß05€UÄMØ&S6­ø–6ÚmÇç´ƒ3\׉_êbNÝèjv°ðáݰ‹)†bGÃØò’]›[‚ü—ËÁ‚Uè~â²–~¥¬Ï;Ïþßâ>+­Ó¬Î“^íú:')åç'º¹D ú[ƒ\j‰ûjáY-ñA-¶PËxj¿&Ï¿é>S,=óš–çeWKNè<ƒÅÏ—Þûz»ä«Î; _“wÀÊöû£¼÷ôwMÆynàÄʉ[$Egªàor®¾ÏµçZ0¿}m°÷FæÝŒ¾¶bKm`q›“ã íu.n¶}Ø]7hCmø£6t§~ÚÁúNúêbœ ˜cm.dì½àK/ú¼ìé§n€ò>l™%zŠÿA.Ã\?¿–0Ž¥Ìs)|[JL± =c^Ëñ¡‹àK?õ‹™÷X7Hýù÷62‚¾-Å:×âoV0‡•èá*úÅ׌"÷1°rŒ˜sù¨Ç6Wc§«°‘UԭƯ­[x¡ðÎYç«“õM0:É5Ið1ùjyVê<ƒvž>]ž8ÏáÈ_3`z†ñfˆé2è^ÿ”9Ü}^ç¬Ñµ‚­ôÛ†-·ãÃÛG_:Ñ.îéc`› àovÞë}Ž»FŸá|ߟ#ÉãòÄËì¶ Ì¬¦Íôº–suøâz'Oo…ˆãO#vׄŽ7#³xÚJ,׊®$•lM].öñy~|³ûì-Kl“ǾsèI=«Æ7ù¦ þ×bÇuèy<«'O¨ÛwÝ]òœr¹kÏÌ· ]é&®èÃz>îÆg‹,E·ÈÙ ~r;BžCØÄ0º:Œ>,qÖèáßRli÷-[–a“cØá l"y¶Ùß»Ï>²àH'1j'ýuÑæxÔ®ö€÷°Ò,À{h{!Ç^ða:·<„o‹;Üÿ/¢ßEð£ÿXýÀ €KÄFCØÐЇ×Ñ•Ø'$w†7ñûܸ®ô,¼J€Ã¥w«.•¼Â‰õfåyâ=ò¬©Ë±÷#Kk'È;=ø-»·Áéc£ñuCé}6®÷sC)÷8Ìøº¿Áy?Äïý ¡/ß÷7„/0þïo¸pß&ù±¿¡äs2¾îoÃKß÷7üÖø¾¿¡4wü‹ŸûJïÝ8ëÍè–_ûJ1Ôãëþ†:æ÷þ†Ò3ÀOø»¿¡ô®±‘¯û}—ñuCÉv÷_÷7”°ËÁ3ìЯý %;x¾ñuCi-yÌøº¿¡´îîбƷý ¥w¡ï3¾îo°ñü‹^`|Ùßï|ßßàÄ%Qãëþ†ÿ÷{CfÄÿý %ßñcãëþ†Ò³ÇÎo1¾íopìÞïý %=XlüÝß°Êø¿¿1ù½¿Áñ±~ïopÞYö}ÃÏïûJºíØÈ§Œoû’/õCô‡Æ÷ý !xêûþ'·<Éøº¿¡ôLúãëþ†R®çÄ\SÆ·ý %¾|Ýøº¿ÁyË÷ý Ž/tâ?'wbúW˜'|CéÆ_ú»¿¡m¯ÿû’KýßßPZïržë _¿ö7Ä^n|ßßPzÇàÆ×ý ¥wgÀ,?÷7DÑ¿÷7Äi|ßßPz¿o©¿ûJkÔW÷7,2¾ïoÈvø¿¿¡þOþïoè{›ÿûªOòCiÝÞ±·Ïÿö7üÁø¾¿!3éÿþÇþüÞß»Þø¾¿!¾Þø¿¿áZãûþ†Ò»u‡ø»¿Áù[e~ïopôØïý MŸõCäcÆ÷ý ÎZ†ßû’_ôCi,_÷7”žG¼Úøº¿¡„auþîoÈåÿþ†Ü5þïoÈ\æÿþ†(}ø½¿¡ôüÉÉeß`|Ûßà¼Oí÷þ†ÒsCç˙ƿý gß÷7”ìüåÆßý 'ß÷7”põ:÷7”žù9ù8×ùµ¿¡´&î¬Ç¡ç~ío(Åk_÷7Dá™ßûJºG|éçþ†ºFÿ÷7ÔŸíÿþ†Äÿú¿¿!û½¿¡åþïopæå÷þ††Ïø¿¿¡ùþïo(ùý¼¿ûjöƒó>ßûª¿íÿþgOƒßûJë£äî~îohü”ÿûêîÿþ†8ì÷þ†ÒûŸ·_÷7dÇʰ¿áíþïoXx“ÿûâ‡ß÷7”ðe¿ûJ1ЯüÝßà`…ßû:ŽõCr½ÿûJXt‘¿ûœüÌïý &ïìm_Ñ|-¿Wãßøà­Îï«ã½c¥ßxŸÔw]ë^×'Ç9?*Ç1¹o¹œw;÷­:¶ô_w¿{\¿Ý½n´{€‹r\/í­—û×K{ë¤~Ü¿VÚ]+÷%ä|TÊFÆ•cRêSÒ^Ô×Ê}5r^Zï=_²ö—–qd=ç½Çœô›‘rÆs¿«¼G§ÞŸò«åX/í5Èõz¿ö¯çëä|‡œï”ú.›[ä¸@®ëõșȧW®k—ú:)7i½¶/Ç…žùÈqX®ûèÊ[ôs\ôgB®›=¿Ü=ª¾¬–û–©>JýJiwÜ·F®[#ó•v—ÊuËåþ”Gn^yÕéQëô¨çU.Êï©ï“ã"©W»S>)ÿÈ8Zåhåäi·M®ïóýîõ•ûÜj³ñÏî}ÿ$÷iý_å~)Ç¥œ”~TÏÅ6þNÎëuµ>i¹IîW}УÎGÇ߬z&÷w{®³v!ãPûP}móÍÚ¯ò_Ϋœôú埴¯zÚâ©×þu<ýr¿Ê©GÎ÷Ë}ö(õ½jGrÜ õ«äº¥¢—K¤]Õ_µ‡Åzœ·z-ö±BíGìdTëå>ÅéÕÒÞF鵜?@Ú[+×yqYb|½Ü¯ö¶@®ï–vº=å¾/’vTÞ–ÿRî’ëTêw·¼ö¥z÷ØeXõVõÄc/­ûËYõ~ãä:³ÿQëUßÕngG^|–þ?s¯œWWܶþCÆ¥~Eí)"ýˆýnü¡ŒCÚÛø°Ç•ŸÊgå¿ÚÓ€ô;$eÕ+‹;R?"Ç!¹^qQˤýa¹_ãÅmÕ=¯qÆ2iwTõRÊŠÏËä¨e¯ê™Úã} yÆ= ö-e½ß⮜_,õ:O¿â®âH‹ÖË}Cžù÷{Úïöè¥âVÕ/«xý{LŽªZÉÑ«ßjïä¨~XñWq¯Q®o”öwõ:;o¹¾^Îk»êÏ4®ðÆjg-žqÔª={ÚÕúj_Ôîë<óU¬|SûVþêQÛU9¬’ë¼q¯êë"¹nXŽ*בk÷¿^ã­_*Çerœ¸Ümo³´sàîõÒÞµ iGãRÞ ÷­¿°FÚÕzÅ'Ñ Å!Å ÷ªÞˆ¾Xœ{DðCqMðEãf‹k¨|U>ªÝrŸµKµ)«?ÐzµÕ;-ëýÚžâˆúKµwÕ+Õ3ÕS‡4®P{S=SýÖëU¯ô>¯}ª>ý£¸[ã ?Ôîº<ík¿mRÖóZÖ¸BíDñJý“×Ojœ­qžö«ù‰ö«òQûòÚ©æ?6ïQû}[-õcR{Ñ8Eã“e÷»÷-•£â»ê±¶£ñÈ ŒGqv¡ú)«=©_X­ýKYq]qwX®×£ÆOÚž·å¼Ú³ÆÛÇõª”ãb©’ë´ß~o½¶#íJ¿šOŒ^àÕo*n ¨ß£ÆKŠSªçÊ×!)/‘ó‹•zÊEú³ò‘úu2¾2Þ¥r^ãÈe2¾Erú]ŵQ­—ñ/ó´£üÒlœ®ãüú£üÒy¨¼4>Ðëõü9?™u˺^1ªr’~”^=[$íØ8CÎkÜ¢v«¸ q†ê™ú'Í»5ÎÓ¼\ñFí^ñZõJò¸¿¿ñ ”q¯³ñ­–õø[¹Nü„=þÃ^§íèy=þÖÓŸÿ,õòô§ãÔ£^/åÇÖT„:oÕ“…Roã8Źn™à“êñÕc¹PõHí@íPÊcûï“ûºW<òÒqêyëW¤¬þHõ]åéÍ_߬_vÛ<×é¼5ÒþÕhœ×äác£ç¼–µ½_¯_ñ!·¼Jìa©ÔÛ~=ö¾Ð£çêl|-×yðBí{LŽvÝQå$¸§8­¸§ùîZgÔ«Ýj^cqÌco*Wó5~Ñ£ÎOõKùÖ­ú&Ç¥‚#êïl>æéÇâˆâ¶§ú-÷©üUŸWìzœWu=Ä+gë/åzÕõOjG¯yó1õGjŠkÊG•÷?òÿÝþÊy‹Š3Þ£œ÷ƯãFÆ«ù’ÆÃzÔõbÉ«¼ëÃ%ÏѶï9ڼ܋Wv½FƧö²\ìÅ®SJyŒ{ƒ´¯þgüt÷º¢´³Éíh|“ŒkrÛÏfigRìaR®;PÆq \§yÑ2é_ý¦ê­ÊEõZõMñSõTåû¸u9ª^6h;RVQ}V{Püµz£x*×kþ®x¬z¤þPÇ/ë§ï—rÇþõ[oRûRëõ³Oª½é84¾Ó8Tã7Í5ï\+÷¯‘²Æišê}ZV?Ó둃ú·5"·õR¯ú£Ï[Æßê– Ç…ãåþ¢œ_/ã9@ê5¾Ñuõ•¢kå¼>÷ÑõEo> ñªúSÅso¼¾dû¶v¤ù§Ú¯®é:@DÊq9¯ë#zŒ{î³÷˸«¤õ·ºî õö:9¯yŸâª^§q«ú_ëçähó2½_ÊÚŸúG‹Õß+nÛõíGíG¯“û¼~\ãgïó.å_Vê5>Ô²ž×~5ŽÐ¸Ô>×Q;ëÛt|rTÑçª'nGè­Ô+þ,—q(jþ¸JÊ+õüöýõOïWÿ£~gDù§òððGçõ8œ’ë¯4ßW¹{Ÿÿ4ËuZ¯üÒöÔ*>ª½ôzø¥8§¸ªù³â®Wêz“®§®vG…¯êo4~ò®7­þG…¿Çj¼dÏ«K»ŠWzŸò_å¤ù¯®hþ¨rQÿ®ù•â„ÎÓÆCr´ÏY¼z§xìá›æ[ªÐxKñZ×lüéio‘‡ßú|\qLõ^õ¨Ù3>ÏÈQå¬ù¥òAó Ÿê=z¨þHõ½OÚQÿdïóè‘Æiß꺅ڹƗ:NÕOË'Ïøm^£óö”ïçjþ¼Tæ¡ëLËD_6HY&býâŨœW{¶ù—ô¯ú¢üÑøÓ>ðøÍ VI½ÚÉr—ê¹´¯ú¯~KóEÕgÕ_õóó{žz÷{Æ¥ëA—êQý–â´®{ŸËf=ø”’öõ9“®Ãè1â¶óøçÁr^× þºÿý¶ìñ«æögãßG¥,õÖ_ÛøÛ3žô|þ¡óóÕjÞ¯üPüU~iž£q“]?•£úg½Nñ¼8îÖoýX+ú±BŽúŠÆYÅ=î}úÞ‚Æñú<µx×þíªÞëû ªÿªgöù›â™‡ï,>*>ÉQã\Í#tÜjŸúÜx¹ú i_íbµ´³TÊj·júœZ×WËu+ä:ÅYÅGcÿÔÞÕ^4ßÔù©ÝÛ÷%´¬¸§¸)×÷zêµÅ Å ÅAÅUmOŸjºþ£ýêx¯Õž—HYÏ{å¤í©l^(rX)G]¯Ð÷TVÉu+¥}¿JóHmg¥â›ô§òS¼S?mÛ“ó«U~Ò¯Æù÷ëõÚŽæ‹=x¦8§óÔ8Åâ²ÊSŽŠÓ_¥¬ï™©þhÜ¡~`TÆ¥z7&×/W=÷ô§þqX®Óö5®ÒøPã5ïsbõó6õèƒêËç÷^‡Ç>F=ã׸i¹è–uþÞõ1¿Î§KÎkœn×AåzÍo5.QœP<ÔuÍ34ϵëëj_R¯ÏíTÔŸi¾lÇ#÷kûš·ØuF¹^ñ[ùî}¾¨ñ“Ž_qQålã09.ô^/í©žêøÕž'tý]õAãI4®²ïÕI:õgiGü¸ç®ö=FiÇû¡æ¥µRÖ¼£ÅÃß6Ïø¼Ï]uü¾ÙõÕ{å‹Ú•”õy‘Æ£ªÿÞõ{ý×vlü¨v-ø¤íÊxÕ_éó=*.+¾«~öÈÑâºôc×I¤]‡Q¼ÓõáuÒßiw\®—öìózÅI9êúœâäϼOÔ^÷Ô?Ø÷ ô¨øªçå>å³Î[Ÿk-U~)ÔKÿ‡Û÷y´^æ§|VÜö>ÖüNï³ï _wÕoxó:/>ªž)Y¿ïÑ7ë¿åhóféQ׳•/£žûtQñSðíF]‡é“vôù¡ê¡ŽÃµ}Åsá»®{h§ö ~^ñßús©W}Ñ÷Ÿú„ßÞ÷¨õ}híßæ÷Ò¿´^õCʪšÏèû}Ú¾–Ç÷xÆ#óÒqÙuÆq±µiüÆýÇ«|Ð8Bý¸õk^½V~+®KÙ®'ÈQõFý}.§òWý’zõç=r^ñÝû~­êƒ^§~Ñ‹_Cž~5T}W»Q{V;÷®ëhœ£y©ê‹}~ üÓ|ZãÁ•zTý’²®¯éý/Ú÷¢ÔžEšØëõ¸Ý3.Å9¯ù„/å:ïóvû~¡â‘\¯ò¶ñŸŒÏû^…®WY;“~­½É|íú—â‘ÎSÇ«|—qÚ<]úW?dßóP¹H;6’ë¬|¤ß1©_íá§®³Ùuu>¨kÞ¦÷««Üt~ê7TïçšeÜöý;·´§÷Û÷xô¼¶/e}ïC×óT_ìþ¿ÜgóŸÖË}Š{ºÞ«z´Òì¯*wµ+ožå}Úê»¶«ú¬óÒózŸê±WŸTo…/ÿéóf럥]ïzu¼«”?Ò¿Ý÷ ÷«^«xóíǾ«ö¡üv½ëÖžä¼7P¾j¿Þ|BíÏ®7+_´¹Oý»â„ÊÓûœS× ¬žéõS}WgÛ•óö9©ŒKßÓøMŸ_é{í6Ÿ•ãÅÏ|TUä:ó´^û±Ïa¥õ‹ºî£~TíVñFólµƒTOô:>ظAÇ-ç5¿ÔøÛâ¯ÔkÞ¦ùT³§¬ù@Œ_ã|Í›ô¨ëHêw5NÓ<^󻞭z®z«x¬óP¹*î¨>ÉujßjWÊ«?=R\·ö®z.í©¾Û<]ê×µ=»,ý«ß²ïª)ŸU¿ToÕn¤=Å Í'm<-í©_.öí¯ÏŠ·Ê7kÇrÞÚ¯òG®×£êŸÆevß“ŒÛÆ«*=jÿ£¼ïQi\Þ'Gû^›\gŸÓH½ÆUZÖ¸Mù¯ú£ù»ÆçêÇ4¾S}Ðqëzƒ<8­vbŸ—è|Æö—ÊWý»ÅUÞ[}Vý”qØçÂ_õ/ªGv_¤ŒSóOÝo£y©}¾)íX|T¾HÙ®cÉ}ºÎcׯåhß3–~ì{“žëõ9fVêeÁ>Ðõ]¿H½÷ù¼®×ëó ï¾ï~”´§Gå§ëµ^ù+|U¾Ø÷ðeüzTœT¾i~á}ÿFã/[Vù¨^©ž©)ÎÈx¬Ÿ“ëWêõrÞÆAÒ¯ç-ÎxôÛúsÅ7µ[-«~{ô|¹\§ãÒòJÏ|T•j÷Ú¾ŽÇ¾§%ãT{×¼GñTå£xdßS–zõš¿Ùõ_ÕgÅ5ÅY-K;ÿ¨¼íóF>kÞõXÅ]ô¾7¡ëŒê/uRíJו”‹u|:.©·ïc«\geœc<Ðøv½5N^-r³q”í:—ŒCŸzŸg(îY¾H½â¹Þ§¸è]÷V¾[¿Ý¼?ßu½Bõ]qÒ®?(NJ;úÜÙ¾G!ÇZ¹ÞîOð\çÅïõöýF¹Où òR¾éºÓR)«Ý©_Õç|?Ú—Õ¸\ão¨½¨k½”u}[Ÿãèz„µÅ3)ëú‘µw¹OíZùlß‹Óó9©4´ë"r^ǧã÷î·T°ëµÒú!Õ[›7è8¤?›GÉÑÆOr½þ½Õ‹qá¯â”æ5ª‡ëT¥_]7´ë^Íûí¾9哞—~4îV»´q½ê·”íß™zõsÊ_ÅQõ纾j߯•ûìz¥ŒGŸË+îv{ä¡ÏW4¾ÒuG‡ôyŠê•êâµ÷=jû~Ÿ÷¨ú*G•»]ï–²ê§>¿¶ÏǤ¬ãö¾_b÷‘Êužý"{¿Äû>Šâ®>R¿©8l÷‡JÙ»ÿTß›´ú®ãÕ£wrÚ¥]¯Pû•²æ#Þçúš¿êóÛ){Ÿ+Ú÷üµìÁ»ß×Ó®w¼}ß]ÒŽŸÚµÜ§zä›ì~e){×™lþ­eiß®CK½â½}ßCìÜû‡­—ûísu‡Æ{š¯9¯ïÓ{÷ ýYæ©×©Ä÷?>î=üG=íë{CªOÊï>iGñƾ'¦8)×Û}z:=*¿¤}›gÊÑÆ?z>yדÔoZ¿wº”Õ¯Ýå¶·YÊãÞH½®)E.›¤ßMڞܷYÎOJÿ›Þê^§ï%תýÉx”oZï}?YóFƒÕ?jÞ¯õz}—çz•‡÷ï©_³ë…½Uüò®+hüé}ž©ëLŠêÕn–xúS¿é}lûUñÈ]íÄ>G•ëu=Ô>GU½’£ïåè}^>$ý©ýÚ÷$õ~ÝYû>Ø÷qä¸ZÚÕxH×ytÝSß·¶ÏeÜä~ý;M6?•óªßªÿ*µ/»Î'ã³8£óW~ªÝèyç2éGó]µGûéßÎÇ£ëš÷ŸŸòÇÆKÛ=ýëuÚ¯–eÜvÿ¶êÊIË*gß4¯Ð¼Cÿ‚Æ÷Нš(Çe¾šézºþÝÕs]wÑç(ªGÊ]‡Ó<ëé_ÿÙZÏü½ò²q¦ê•êµ´¯ëK:µ;Õ «?cû·¯q·ÖëßññîëQÿ±QùpÁþ|±z)GÍs->h½ö/e}Pù­ùªý{m2/Ë?¹Où¨ÏOt<o«|”ŸÞ÷TOU/í¾•ŸÜoñEqKÊ2ŽMêo¥^ŸÛ¨ŸVÿZ¼Ñm·(ó×<~sV®—ö4oÝ$|”ãfå÷ûó]ù­ïQhžkó ÕsÕ™Ÿê‰ýûeª¯ÜVçXç9®ñŒÓ¾ þÄì/wõS:?û~•ðg£Ôës<«×Çîoïz®Kè<Õ^u\ÕŽµ}ÅKµk™¯})üQ<[%eû>­Ü¯×Y;•²â»õ'ÒŽÚ•ƒÊ[êõy¢ÅÏýŠËÖNU/<¸cóeÅ!™§ÆÚŽÎOß7°z)z¯ú©ú®ëš+ßÔŽÔ¯èû?b›”/r¶¯økõNñFå¤v­óW9(è|å¨ëþ*/Åû¾‹Ž[ñUî³ë~iý‚ꇌ[÷ãèyå“>oß4î–µ]Õ³Mr½î¯·q‚¶ãáŸ}Ž{ëþvª~Eã.µ£åªïR¿ÂÃË'•«ôcýê¡êµÊMúU»·çTõJíMûÑ£ÊSõAqAæcã.i_ãxÍGíß/Óþe\Š/6®Ö£â‘WžZVßÿhŸÔ_j¼fã=Áͬ¾zôr…g>j?ŠwÇ•/:/Ý+ÎÛ8Iý­ûþg«þLÎkþdã ÕÅ1OYõHßw³Ï ?ä¿ê£â•”רÐqëü¤=ïÍ—×x橸¢xdãq)[¼“ûíslÅ3iGß ²ù‹´cãTå÷ßÞÔÏéQõÇÆ‰rŸÕ?å“Ú›ôoíCíV®Wÿ¯rµv¢ýªJÙ>ß~íQú³wXÚñ®ûØ}ÙÊ'¹ß¾‡ ãW}’þtÜö½pçØþ÷YÒñÈQñYŸcÛu0i׎Ç#w/~Û÷a<øn×Ùä¨ó·ûoõ:©×u[Å!ûÞœ·ë…2NË9Nȼ&Æ<ÇÝþìóK9n–þôù¥õ}núÕõz(Þ>îï¥è|=ç•/ï´Ïõö9®ê›Ú‡ê­ê±âŸ\gqÜcÞ÷µµ;ïûôVß·{ì\ýˆŽKø¾Tæeÿn“gž6n“öçm|¡~@û“zû~ºm^ ò‘qk`ãÅ!Å)ÛçÖrÝ/£÷I;ö½¹Þþ]Q‡®÷Ûuk¹Þîã—ñëQóKÕ7Õ?ÅQ}Îà}ÞÖï9¯Ï3l?2.û\\å®þ@ÊòœÂè÷¼ïßTêwô½¿Jÿnƒý;äR®Ôï6xÿ.`¥~·!"ç•ÏúÝû÷3•û*î» öïs«]È8*í» ê¯*ý» *¿ ÿnƒ}¯¤Â¿Û`ùUñßmv*ý» Š“•þÝû÷Þå¼¶[ißmм§Ò¿ÛàÁ¥ŠýnƒÚ_¥·AÛ©ôï6x×c+õ» VäúJýnƒ®¯Tøwìß«®ôï6ŒI;þ݆Ç}áA)?â^W)ßm°¸WéßmмÚú)WÚwÇ*ý» z_¥·Aå^áßm°¸ GÅïQÎ?Y¿ÛðØßÓ”c¥~·ÁîãP9ÊuÁwÜëËýÝ»OWäV¡ßmø‡ëÔ÷Ý9VúwôºJÿnÃ?Šk*í» šŸTúwì߇’qVêw¬ÞÈ}•úÝ·J}¥~·AåPéßmðþæJýnƒ¶÷¨”¥¾â¾Û°IÊÅq÷X©ßm°ûæÌþã®´ï6¨]Wúwìß)~UêwÔ.öàY¥}·A×1*ý» öïÉùJýnƒÆá•þÝõß•þÝÅÅJÿnƒî©øï6ÈQù\©ßm°ïH}¥~·Aï¯ôï6¨¾Vúw4þ¨ôï6辩JÿnƒÆs•þÝ•[¥·Áâ“´W©ßm°ëÅÊW¹Îê»¶«ú¬óÒózŸê±WŸ¤Þïï6Ø¿;¯ö!㪴ï6Ø}øªßÊ_Å1Õ;)?Y¿Û /·Ò¿Ûà}NT©ßm°Ï'd¾•ú݆ÇýLŽUÊwì߉•ëûäXißm°8«xªú«çW¯õ\§ú,ãûO·AŸ?Vúw—+ý» ÖßÉùJýnƒâz¥·AŸoVúwlþ*ÇJýnƒw_R¥~·Aßc©ôï6¨¾Uúw¬\åX©ßmxÜß±—ë+í» ÞõëJýnƒý{ýyTÚw¼ï•Têw¬“r¥~·AãøJÿnƒ¾·cä|¥~·AqµÒ¿Û ñR¥·Áþ½#_¥~·Á®Óèø”Ò_¥|·Á¾_*ãªÔï6Ø¿#&÷UêwTî•þÝ»EÎWêwt¾•þÝÕJÿnÄŒ¿Ò¿Û ö]éßmðî;«Ôï6Ø¿O©×Éý•ö݆õ·îßN¥~·Áú}¯+í» —Nwë+õ» ºÎWéßm°ïkH;jO•öÝëÏUž*/ÕK9>Ù¿Û`ã1é¯R¿Û òZ®ú.õ•öÝ]©ôï6Øõpé¯R¿Û þ°Ò¿Û öTéßm°û–¤ßJýnƒ}Rç¥Gi·R¾Û`q\æ51æ9Þèö÷dÿnƒ‡ežŠŸ•öÝ­¯ôï6¨VúwÔïUúwtQ…·&«œï5˜¨IñÿDqû–ÉÙ©bP,k1^œÞ2³=(ùSŠmšœžš ÿV!¹if×Üîât1(?)ÊÙMó³Û¦¦‹]çf‚ºÿu êæ§7OŲ“Å“³33ÓAùIQÎOwNv29»y²kjn$¨ jÿníäô®Ùâö®â³#³AmPû÷j£[‹›ƒÿÿSÿOLÌ옙™ Še-Æ&fÈfƒÂ¿W˜Ü²gç® ðo’“»¶NÍì´ËVAù?[Žm)n²QqPø× ñ-Å›[‰ JOh)¶eëckVAá_+Ä·ÌOMO%J·4¼ijn.¨ú漢:dfvO×ÄLqnWPõ_UÛ69½§þ­BbûäÜÌ®­3A±ÌÅ©Mû¾ ËRLï(n.n)ÎMgƒšŠ¬‰ï(n/îž J¾”¢”‚ÿÿsÿG7çg§v§÷E— ¦‚jRnÍÔü\Pñ¤®H8C&&f‚bY‹é3{•¼ƒç'ƒšŠ¬ILwL=¶ËSŒMOm™œ ÿV!Q*ìÃâ X–âìäüô”}6ËSŒÏî.No.%_J‰¹ÉéÉ-ÅíA±¬ÅôÜ䞉­“Û·OÎ5Y“˜›q–<ŠA±¬ÅìÜÌü®­Î{ªSAÝÿºÔÜîâÞ©íx­ âI]‘ÜUœÞ»Ï2bPþÏ–£»fìK¾Áÿÿ¿ÿŸØEÒ1÷˜Û Še)Æç·ì››¥'´Û[œš ÿV!¾w¿WxƒÒZJîÚ±©¸i÷dP~r”7g77ÛVAù?[ŽO§‹Á{PzBK鉙¹]Å®}’Í ¦²jò›gvLMSžîšÜ9Ôµ§63¹½k®¸ý€u6¨ú¯ªJm™/îšt^Ö *žÔ±­Å©]SAáß*$·ÎLožŸ-Îå'E9qPqGñ±h$(–§ß1yè”}3(=±¥”?Íð âI]ßI2»#(ùSÊ윟œÝ5ãd›3AÕUUa×,Iâæâæ¾]3›Š[‚ê úÿOu|~dn¤8”|)¥Š³[&§wMÙ¿ñT<9+›f¶O2Ë[ŒošuÞ» J¾”b[§¶O…«œ˜Ù>³ïSþ ü-'&'æ÷Y3Šå)Æ·ÌïÙïOÕ¥'°”Üé.†ì ÊOŠrtçäì|ðÿêÿɹy2ËâŽÉ ü¤('ægç÷Qå XžbêÉéɽó“=5*ž”éMÅé-Û‹›'ç¶5Yãägý å ð/R[g¦·tm㟠âI]›šÞl³ß ð¯Rf¦'炊'yEtj¶8üÿŸþÿÁÁÿÿ©ÿǧæf‹“Ûƒ’/¥ØAÅV-ƒÂ¿Vˆ43»9(ùTŠm›™µ ÿZ!鼿ç1o”ÿ³åÄŽ=Åéö§ÅòcÓ“;í@ ÿZ!¹³¸mjn—è üŸ-gvnÚ>µsçÔ´ýËpAÕGUv®8¿yª«8[|ì=Š î¿».575½¥¸sÆþñ— âÉZ1;Õµ½8½­T<©+bs{ûÚAá_+Äw§v?–¢¥'´”ܵµ¸ÿŸ± ÊÿÉr¢8?·kŸ¿»ËSÜ4¹}ËÔüŽ XÖb|bÏÎYûA§ ôÄ–ò{''¶ÎÌmŸ9„ĽÔµ§6±yÒY¨ÜË[Ê ôÄ–’;gfwÍoÙ絫 üŸ,'fgvìóa³ Xžblngqj:(ü[…øÜîÉÍ“AÉŸRfn÷Ô®½®o ªþ«ªâ»æg·Mî J¾”¢ó#ÛF‚ÿÿSÿO:Ÿtnd6(?9Êé=ó[fæ¶û›˜AMeÕ¤Jïíóa÷ âÉY=pê ©àÿÿÔÿ3Ó“»»öN÷ k‚ªÿŠªüÎâÎùb×ôÈ–ù©i»³7¨ jÝZcrÆdW3˜„î3fèóÆ ¿Ó˜‘—³äÆ,›6fôc–/5fÅYµÇ˜5ÕÆ¬½Ó˜õï5怫Œ1›^hÌæWsàÇÙúKc¶­4fÇ+ŒÙY0fözcæ5¦–ºÚ×SG_ug@÷S?}ј† ×5.>iLÓ ô€1ÍÓÒ 1®ÖÓŒië‡~dLûëŒé8ΘNÚìJC·³€ºärèè-Ðp mcN] }—Þ ½ z›Y]Ýz 3¡[`Ë"èE×äŸ }Ó˜Âvˆs¡_A2& GÃP/4M@³í‡Oh7ü?ó ß=bL~D¡­Ð±|‹¼‚‘ŸƒËyˆ¶ªŸoLMô>xF¿µÁ¯ÝÐMð‹ûëámC:B[„ok¡/À7®mæž–+àÛ(ô=øv9|cŒaèËÌye>ý”9#Ó,¼È2¦ìË OA?cþŒ!÷ ˆÿçÐtôè×ðc'ôQcªº!t¦YT3§d\CßõÌ¥þjè÷Œó)ãod _bœCŒ«ùxÆÉõ-fœ'1ÎUŒ>µsO}trìú ò…ºÿhLÏÿ³ú^äÐÇÑ ƒ|Ì]Ð!æc˜»á>óWd‡2¼ 1ÏP-T1·ò 5C­ºZõACcÁ“Ðf’Pt þl}¿B^qô7þ7cèXb ÏŒ?Þ$i'~ƒÒfL^} =¬‚„.„nD#òËŽÿ±‘,ú’c¬9dŸ»º~Gï„B/†¾ ß±‹ò/|ÞsmöPE{ÕÌ£ú<;©AÏjàkíBèÕðãKÂ+Ú7œ7è·ù-_ s ÑG†ª„7O¸74!ì-´ÞåIhÄœCGA'Côbl‘[!x¡Íh ĽÑI»ˆrM”qD?1†(2Š5A´;ºb1ø_$|Ä|µÎJ/„ ŽcÇŸm̉´}2ã? 8óÏÆ<û]Æ<çYÆ<ö.dìÏÇ/þª1/‚ß—¡Ã/Aÿ®Ä^®F^¯^ƒ}]Ç=o?ÞŽß ¯?À˜¯§~/vvØaÆ<ì8:j“1Ç|ߘg¼Ô˜g‹'!×SÁ€Ó¿fÌYÌùœ?sÞAÆ\^¾þ\r¿1/“^ ½|žþvѶsÝkŒyójcÞqü¢ßs ¡»¡{ CÜ BèpÌÑnœ£—aÆ~ï…è+Ì|ÂÂoæÁ¶"Ì?‚ÍE°áÈOD.CØc^†À¤úf„À‹Ð×!l1B¿£½„„ÁÛ06ÃÈ'|.„ŒÂèw˜öÃŒ=ü vûtì¶ÂþÚÁÖ°ºûêü¶‹N/@v ¸§û=Ø/~d!6Ô‹=ô1ÖEÈyréG¶‹éw1òäÞ!îF–`OËìÅ_è¿€>°ýªÅú_E{UÌ¿šóÕ—q]+`[°­NWáߪž YUð¸¹Us_õç° 0¿æ%Øê_ûæ„é;LE¸Jø®”߇ÐÙ0ã£gad>ºFx‚ü"àM„ñEsÛŽ"xcÂ÷UwUod ȳ?WvÖ"»Zl°Žq׃¡ `I#úßä`9øÑ‚µ"Ë6ô§¹u ‹]`ùÆÑ=_iw!XÜ–u1–̯ü[H»½èC6Õ/ƒÏ:6òc‡¶3ž:Úùm`ƒÕð°šñÕàjàc-¼¯ÅNë˜o˜[FÕÓwþ¦á:ÚC–ð¶ {oÂFš©o†—-à{ ö׊-µ³m÷¸xù³ÈÖÁut+Š^D‰›¢àbôxÅ·Dñ×1t1FLãšã!û>(fÄáC|ÔYà…°‹8~!„¹& 6…á{] —†ñùüx]‰ÐodƒÄP`Fä0±-ä!^Œà"´A>‘›E'fD~çŽ ó_†M"“1ð{ú·ÛZÅØ×À‡µØÞ:ê7 ÷ˆ1Æû¦,v€/˜DV~€8=˜¢Ímð}69Ãø¦Ï9ìÿdÒÁuà{'ív×vÁ»øönÆÛƒœz˜×Bbè^ð§þ-B^ý`ìbú@†ƒø¥!ä5–,C–‚mËãñærìvúa¬‘!Áý§.—˜^FÀô6 ÷"ø‚㈠CøÁÏD³"Oì&J<Å®£Œ-Êù,>(Gÿ9üMlÎãûòèl +À¿Â­¢ÇèPº_nWƒM5ð®9×¢·u𢎶êi·û±ÇFÚjâÞfdõ5tþÛÄw7߇÷¡÷?§Í‡±§ßÃ^eBalívðénlúGØùOñ«áÇÇÜÿò&Š=Í„rgÚ=hB7™Ð¢¿šÐ²K áÖ® ó|ÁãÐiâáq<a!x‚¡ÏB`_è»âñg¡{÷ñÉNìϸCØ`»§ÅW´»þ9®džáÆÕÙfÉ'Yn¥Ë¿<±^Ì“#G©B^Õ]¾¼Õ€ÓµØRXP‡þÖãïëé¯ûÑù÷ä±û<~6¿(ì•]¨ÂŽªÀ²ª êªÁÇjŽ5ØK mÕb/µØfzR¾ÖãƒÀçt©95-uó.Cß½2è± úk8g–AÈÅgÚ6ÄêŸ`ƒÁÎ vdÀ(Cæ!Ú7ãûPÑ¥4º–FO3Øu=ÊÜ&9’kЧ<º_MŽQ þÕ0ö0µÖ¢guØM=ø\ÿÆ/áa#²k"iSZ°±­ä,S`Ú6úÛŽ.NóÿðsŸ7GóÈf7|ßKŒ{òºÿtæ}428»>9‘±‚ÿ=:‹ÿŸƒ{.מ&=ƒÄÃ`Aˆq†°Áбûè6ÂÇ„Þ aO!ú Ý,1ŸëÑO8,:ÄÂð4L¬&_?SâÎcs‡nCϘçáŒçâ飰cÁÂgR>t ùæéðèìëÉaàk–¹e‰ksŒ'ÇœòØD?”ÿ©ó…G¬ªÂ¶ªˆÿ«¹¦šc ±H 1|-<«ÃvêÞåÆ7 øÚxÐøvøv |#/j!Ë3²*UQWMQߨ!–¬ÅökÁ»:ÆWOWOŽÕ@¼Þˆ¯hÄÿ51ö&|K3úÔ¸Zi¯ =mçšæß‰/ìúŽäã?BW ²4 ýRè!ÉE±ƒ¿1Ø–ù“ÇÿAŽJ®J\oÀtó—¶0ï­ØëAèû6|ãvø8 FÎoŒ ¾%i'‰ÏJa³)âù|H©)xœFßÓèJþ¦‰éÒè][ÌÀà ˜!×Ì€ƒ™ûd-ìÏÿf‰Y³è~–¹g™GÝÊCEGä@‰ÙÁÚÈe’GsmŽÀÃxÁFî‘ü\Š{FЋ(¼ŠbQd]+q;üb“Qbƒè‹Lø«—˜ðï0á{×›ð/_f¿™ðß¾m"èI¤*l"2‘®?›Èà°‰ŒýÔDÖ}4â©&²í-&2ǹÃfLä˜sLääNRý÷Zý‘Ôb«‰\³„¡~™ƒ8ü`Æ8æîSæ÷n|Ρàîä´ßqØ&ð”¼èiøÎ#8ÿt|åÑðçä}Ø÷Lä<ñæ \úv2Ø~*˜q:6u>ó,øšûàòÉßMáRÄ.iÊiÆ™†¿üy_—a~Yb»,zšÏsÌ%óŒ—Ô,k°ÉAb‹aúa>KÿRr™Qr«±KÝTj%z¼ûYüÖ!·õø±Зqî/"‡ üá$ˆÙ D¬¶í£²6?*Ï*àW‡vI¾‹¿ 1þbÜ!dº@rÆ"n /ë÷'Êõ¡s!Æz¾»nœkÓÈ$ f¥ñ_idÆfÓð.oNc«iøœ!—É ÿLLò:p*Cnœœ{3ô™9Cž» bˆ 1O;²ì`n]ø’Œ¿}ïÁ§ö¢}ø E´¿˜¹ 7 1†ath rY†NŒ1ïÄk+ÑƒÕØÑZ|ö:ljs݈ÇÞ7?·ÁƒÛщ»áÿèãlèAìí7ðãàç_wšPŸ–y¿ ÿ†Z~fBÝø»Á>›Åøº‰)ÚVeB³ïN‡…ÇÀÂò}Äž³Ý¤kO4éþ÷˜ô37é·ýÚ¤ö>“Yý.Tå“ùÆ©&Û{¡É>÷Å&{ûËLnÑ´ÉYgrŸü«Éýé]&?vžÉó4“eÌä¿×„WT̘ðLÊ„÷^lÂÇüÄ„OÝfÂÏņfѧ]ø˜ÝàÍ^lõ©”ĦŽFoŽ#O|ñÙÉèéi§âoð-çSŸGlp!qÒÅèÍeÄ//%ž{9±À«ˆ#^¯ßDðvðà½øçë‰5ZÁÝ6l¦ýî@ß:‰'ºðÍ VÊÚ<²X.ôÞä®÷#Å8Aðhè|YbÁ¾—¯£àØrpi%º¹y­Áö×Ý!12K5©wÉsCâÚ4~$MΔáú rÍe·»1‡käΗø;Ìcëyâ–BŸûÂyFR•’<òRt]? ?w¨@ç/š6¡‹oÃç=Ç„.û² ½¼`B¯¼ß÷#z㇑~ó½³&týfäõ|úäMèÆoâ _oB_AvWšÐ×>`Bß@Öߦþð"NLÄ- xžÀ†àur@bWgMAc¸§Ëº°óüÜN`³ b³òL‚…Il!‰\SÌ3N¥È#Ó\só?ü~úÿ,lþdú;ßpqÿÙ`ŹÈà¹ôs>×^¼Þ_‚.el—Ã+È1®$–º ~^¯{%÷\Ëÿ_ƒþ¼Œºyv‡áÊL×/×{Áà>üÃ"Žý`ñbÆ0N ‚£CWÊ{7#gðkýâ?À³³¬ZÅ}«ñ“kÁÙu`áâ„CÈŸö~ Ö?•˜ôî? ÿu ±òqØëñ´ur= >]=xúül0í\tçô®kÑ^?1Ébd±^ [ƒðxˆØk˜Üyl]‚o_J_£Ü3o—c+É5WÑïâ¹µàÉzôirÙˆîiëcäVŸB^Ÿ¥/¢_¥ü-âž»ð—?Àçÿ„¸öAtñ·ø ?? ¾Šã×&Tù±Uz¹„xlM„6]ƒ‹;ׄù˜ ¹ ÷Fr6}ŸžŽ/9»;—¶žÇ¼.b\/¤|9ÛÄ4Waׯ@n¯A¯®ƒço†o'&}7þïýÄF>†Ü?‰¯¾ ¹ÜL¬t |¾ ¾ÞŽ.UÑf¾Ý‰çjˆÕjàO-¼¯#Ω#.«‡ \Óx¢äÛèX øÛÆ´cûÄI]èa71ÏBügý,áë0ºµlŽƒÓqì=}'ЉãK Ÿ‰$çsòbÿ}$ˆËÌ9Ž$À†$üLmÉ÷ÊÚàž:BgSŒ;Ř½9¿Œ¾±¥pe]z?cóFç’ï¹ë_Ëï(¸2ÖQ^þ1t•| ¾t-í®×;q z³áYÿG®µ/‘µl0¨žÜªi —iÚ˜w3÷µ /­Cð y´Ã\™é _ê¤Ï®ŸŠÇ/d½`üØ\?±À m‘Ÿ CKÐÉeÈo>aû+ðg«õêÖ‚ÏëÕzư{ÞˆýŒ3–"¼ÛD?›É/'ië@澵Ř)rˆƒÀÖm`Ôñ`õ‰ðêÆ|:r6m?=:ü|>±Ò%_Œ<¯„ÏWƒÛ¯r0½ûdö–oóN0þ½´óAdûÆ|ºöiîû,8~3:ôeç6ÓŠ^´;µ!m»Ý÷uÚÀŒ6bš6ÆÕNlÕÚ±Åvúiç|;:ÚNŒÓ~‰Ô=OŽÌ¥ýrÎ!t±°0ì>^ˆü:±íõÂÿÞ³ÜçP½ø¤^â¹^p§/ áÇú¸¾ýíÃ~úàz×w6ÿú.Bú."ÎFŽá¢¼§°YÖ²ÁÁ0÷‡ÉOÂèD[ ÓNl ƒMá§K|”ÔcOax>SÖ¿xŸ±…¯†ð+ÐÁ•èÿJ溒XgØ»Š6W¡»«ÁùÕÈ{>z5¾w5¼XMÞ²†±¯AkÏZpt µYçÄÚÔ¯'V]Oœ¹ž1¬wl>Õ€Ã5Œ½ ¨—5`p Z‹]ÖÒ-ó¬åúÚ Ü<³–ø²L©Å‡ÔÒg-zU‹/®ãú:Ú«eLµÄVµp¥w^ì®UÂÇU;y,×ÕS×TÉÿ“‚5èPÍœ<×=Ë]g«A/kÈÓjÀ¦šÝç#5NûøÊZü燱¾ C¦Ÿ„?ŸFWoD®71ÇÏ2†Ïã—n†__CÁ[˜ç­ØÇWiçkØÆ×¯û˜ß7Ë7ñcßB?¾ ÖÜþ%c¾ƒ¸™ß¬¾Þþûý1ö|øôyÌ´õSxüs0ñAlç!tæaŽ¿i‚¨ÿ5r}=ˆy?ÃÆ?=Œ¯z\|ˆã¯˜Ë¯è£ ÝbÍø°,[p¬¼O‚muãG»±Õnâ÷nüM·“Wƒ=ð¼‡ö{¾îæ” ñ۽ȳ~÷á3û=Æ×öѾ <2ÜoµÁî r1_„à_é=3ò>ã\‹Œ ã.½G…ÝðÒ|߸ïŸáŸ ¸eõn]ßþµäGκ+ú—3ëe d° ñO¿œy£<cœY|WŸ”Å_dá{–x%‡]æˆ+sÄï9ä›»sŒ1ǘ¾€Ý̵_Dw¾Œ|oA—nŦ¾úI†Ž/º ø&sÿúômtðvüÖàßèÒ]ðô»`åÝ`ÕÝÜÿ=äù=î»gÜ•ñ÷щàÇC%פ­ÐAò.ØŒyuhNrÐ=’‡2·öz¦ä›J'HÞy’‡œ|tJˆX¡ô¾™Óç9îÛ¿sn“Ðf¹žX ônšÓï3¤Ÿ¤OgøÔ±È0òÁÙŒ ï%`ÓRp`)ú° ìX†¿e>£Èn ý\^ ¡_ˉUW€Y+Ѓ•]®/\‰n­b,«ðÝ«8¿?¼šöW¿VžÛ€-yx’?IÞß{„¼òØC[) Ã…²®…ï. ‹mnpß9-`…ŸÉûLÈ¡ ™VqO6W…Ïíæ\ºÔý=Øeí÷€—=ø ||>¾y,Ÿ|TÖžˆ-z°Á…Œg!sìůô’ûô}̳qغ™-Bî[áÝVtxŠ˜ï ä° ]ÞÆ¶ãÿw0îiâ¼rŒr¦ôq0rŸeŒ³`àqï.°{¾Ï¡ÿÝèÉnxx(}ŠœÅÖž ¯ÎA÷ÎEÿŸƒíœG.õ¼Ëðµèè…üÿ"æô|r‹‰õ_ˆn_†]нìÄí´qE\ Ÿy%qÇUàÜËñQW3Žkˆ+®a ¯ &jGÎí—¹ïï´c›ȺŒë@Ï;賓ñu‚OŒ¡þw2ÇNð¤ [ìBWºàs×.÷½¶.dߎ]ð± >v!·.|A=¶UO ØYøÜ.4ro#ºØH¿àZX×Äÿ›ài3úÒLŒÔŒÿj‚ïMð¯‰8¦ ìj"6hÂG5ç4#÷føÜŒ®œœÎøÎ»ÎanÏý’<—fN/Äg^н¿|x¼¹=yüz-¼»ŽþßDÎôVêßAï«?Àü?‚Oûcø øöyp÷kàÛmovŸ | ŸòmtàväÿbÀ;ð%w‚§w‘ó~ßp7}÷u¾ènxw7ýÞ¾ ¹›˜ÿ{Èã{ðönÆü=üç÷ðA÷ ˜È\¿‰™2‘“šÈ³1‘‹î1‘Ëï3‘W&M䶚ȻÞj"ø±È'¿g"_8×D¾6d"ßù‰üàVùé &òëO™Èo7ÑhØDs›M´þK&Úy€‰öÅDGO0ÑõÄ8[ñÍSøËƒˆu¶ûmCÎÛ™ïô|]žÆWÏ€{;Ñ•ƒ¹þ`gÝ€ỹì·̣Wóàñ!ði7sØO8”yí!‡ÙƒÏØ‹>ô£—‹Ñýä>†Ð½aⳑFû\ò€¬ùa‹cØér°j¶¹’öWÂïUÄ«Ûjxµ†øp-سl]O¬»ž˜gÃò.þ´ôî&úQzôny?= ‘›„“ò>úuTð;Œ„wÉû£'Ê»œ/”µÔ7@èOøãòüÃyöA^àü•#ˆ8, ¤±Ñô_dœÈ€™ËÅWÁ£,1S¿–E·²Ø~–1 †gé/ ï²×»>,~'°‰dZÞÙÇ&“èk’9%ÑÍÔ yÆ/RÄil# ¥Áµ4²ËŸgàWÞgþ*ëÄ*Yø£½|×½ KýßßPs›ÿû’aÿ÷7„~d|ßßPõ[ÿ÷7dNöCÝ·ýßßPŠ?¸ùøµ¿!>éÿþ†Ìüßß`ÈüÞßùˆÿûŽ~Àÿý ‡&ýßßºÕø¾¿¡ô\Ëy–u±ñmCË»üßßPHû¿¿!ŒüýÞßÿ‰ÿû:OõCä8ÿ÷7D[üßßPw§ÿûr_ñCÕŒÿû"_÷Ci “küÜß0rºÿûÚÏ÷aƒïû²oô×_ìûþwÍÍYgƒÇ~íoHßäÿþ†Ü;ýßß`п÷7¤öÃgû¿¿¡ô^‘ƒ~íoX4èÿþ†ØÏýßßPužÿûœ½R~ïo8à4ÿ÷7d'ýßßoðCé“ÿÌø¶¿aòwþïoˆÞïÿþ†ì_üßß®ñCa‰ÿû ÕþïoH\áÿþ†ø!þïo¸fÎ÷ý &>àÿþ†è9þïoHîÿþ†HÒ÷ý á›nò}ƒÙy™ÿûrGû¾¿!Ôz•ïûLGÁÿý .óCõ·üßßn1¾ïoH,ðCßmþïouß÷7”æôwCÛ%þïo¸õbß÷7¤‡oò}ƒÉüÂ÷ý ñ7^êûþ†Ð‰›|ßß`fvø¿¿¡å3þïoHnò}CèäWù¾¿ÁįóÃöCWÌÿý óuþïoXø ÿ÷7|è*ß÷7˜cþïo(üÂÿý ñ“ýßßÐ÷?þïo¨þ†ÿûúýßßp\¯ÿûZ_äÿþ†žëüßßFOüÞß°âÿ÷78{ üÞßP=íÿþ†žéÿþ†;_ìÿþ†®Küßß`À>¿÷7dªüßßðù&ÿ÷7”òÌÆßý Sû±‚û†~àÿþ†Ò³Ú÷7t?äÿþ†-÷ú¿¿á¬ßû¿¿¡}µÿûê~ãÿþ†gáÿþ†¯îð}Cdòß÷7˜-?òC·ÿûì»øS¿ö7¤¶û¿¿¡”Ë|Ôßý è̯:ÄN:òÄcp”l2úñ¢‰“Ï<éŒÓÎyÚ”~ $21³C?ÿ2³srBÿ¿ç˜#OÓK¶lÞ)ÿo?ò¸ÇJ‰¹#Ïš=òŒc´qNî_“î´¡'Ÿ²OCNiÿ1§N;ùì7ɼ ƒ2æo{äïäŽ:áÈÓurZ™>úÈ3Ž9ö4î7¥wàÍ_ÿ}³÷Ÿ° gmm/data/nsw.rda0000644000176200001440000002027714247643115013261 0ustar liggesusers‹í XUE—T¬ÔÜs%5,÷×£ˆYZ–{e)niî¢fšŠ Š["n » ²=6Ù;‚Š»––ýeni‹Kåÿ¿{†¹rL÷=x ó}5ÌÌ™3gΜý¾¯F ™Ðµæ„š&&&•MªT©dR¹ŠîϪ•uÿªdRÕÄT×W™gó™îßõtJÃZ:ÈS&­¢U´ŠVÑ*ZE«h­¢U´ŠVÑ j®è+úо¢¯è+úо¢¯è‹Ó›ä×¥¡@Ÿü¾g~?(¿ïšßµUŽûæ÷½ |w2ß/¿ïAÆ®YLÎåóÏ@ÏéêFÎáðœÎÞ¤ç÷¶ ûøùŸ¥­rÌ÷%÷áøú’žÓÉÏáôsúøýyN7§œ^ÊŽÓÉÏãôÉøm•xøù¤çp|?§‡ò‹Ãós9½|ò¥;ÙÇïÁé·$ø)?’yz>çóø{Y à{xŽ—Þ“ó£?Yç=ÅÃçéûXÙ*ïCñRyåtòwìJzŽß‚Œ)=/•›Áùôp½ìIzz¾ŸÓËé ø»‘}~™ïMÖû“u Á¾áùtS¹æûù˜ÓCí¿?§“ËÇdÚ¡~îM‚ŸÓMíŸ%Yç=§ÛŠÃÙ*é§rÏÏ—õÜ—òƒßw(ÙGñPºDrDïÙô=C{Ý"ú(ÔøEÏ£ï#Â+⋯h¿Hh)Âß]Ћø¦Æ/ ×]e]ïŠz>µsEûEú3Ø–õ\ÿ¸åú dë'˜(çi\Ïíë)|žã’ßS»Åí=wù¹üê蹽ɘŸ3Œì£þÑŠŒ9]o¼4kYïBæ©?¡qùþÁ¶¬çüât÷$pœO4¢q_§ïÄý^²ÎùAé¤÷ vŸÆWCÈ:}Oÿðó, \_2Oý8çÛ8T^©|ñ}C ÍEv™ÒAÏïMà¨Ý¢òBå„ÃñüŒúa~~O*O"ûBëjv–ʻȿÒz;©ý¤zLáõõ—"»$¢ßŸóæ1ü¸ÒzCoÁ:¥ŸÚ-Çñ{S{Kõâ£vœßÆÙ4Ÿ¤~ŒÖ h~Léä÷ ~êµ£žÖO¨\ÑzµÔ?Qú©Ý ó4n¥y7‡f[ôy/­PþPÿ@í1å•zOQžËí?•7úΔ¯œ~n÷©œÐ|†ÚdÌéÕghMé£uú~|Lï'ª7ñujyܑߛðºt›üuóü¾• oKÆÞŒ¬¿–ß×!pôÑ~ מà¥tÐsš83ÒSzZ‘õ¶¤§÷¥|¡xÚxŠ—ž×FÐÓ}|žßÒ#â…£óíûDx)_)?ù<彿/}gzoCö‰Î£ïAé£û踙žpô~TÍcÊGzO‘\Šø!º=ߌ¬s>ÖìñU$úꥫéÛ àDúVC@¿ÈžQúùyfdž¾—ßEöî£zHåM ÕŠ‡Â™“±è}Eòh&˜É·¾ü£vDdÕèé'7V¹·¾vJd÷Õü¦ˆN32Oß½1Y7#½H>øù"{OÏÙáúNäçè<…ùsÑ{RûÄñP~ˆø¦f·)}¢û‰ô›úe‘­‹ì¹¡þ—Ê‘=jøE|£ë">êŸ8œHÎEøÔì]ù%‘ß6#½èÞjòFñŠâH5úEïDïGéù_Êo}ã6>}í©èýÔàÕèíS³7f¤×—.5;o¦²Þ™Ìs8nï(=4> ~\äg¹ý×רñ‘Žë±(ß ~„ÞòÁŒô¢x„Ëð«Å+ô½ù}Ú8Jˆ?|¿È¾p:i<ÁÇ\/9øôµó”¯¢ü€úi~O‘=V« P¿Aù)zQ¾.ÒG¼èžj~U”w‰âmµ|\-¤÷ q*§O­þ ŠôZ”gˆÎÕ7×׿‰â=}åÙP=Åõ"ÿÄõVßjùž™à|µüVM~õ•cQBí­(Nâëjï,ÊÔânQ>Dãz~GÁºˆ.µ8U$_”_ÔNªÝ_ÍÏšŸˆèñU-ßÉ忾õ êO÷3ɯKøý4¶büîúoçK ¯¡p†žgèþRü]»At‹î’â¡­´Ï3öý躾ôNß}Ææ£±Î}Zç•6ý"ü†ž[Úðåí½ÔöÑöOᇾtŠàè¼¾øÊúþjtÑuµfìóË‹]5cÛ~µqI嶸ø ]×—>cëmåMÞ Ý'‚§ó¢V^íjI×%†žSZvG_{T^è*k;ð´ß§¼á1–>›ž’âSÛO×i3–>‰ÖEçë»Ot¾¾t•µ]Ñw]„¯¤ã§eߊKWqñê‹¿¸tºÏXçOYßËXtR<ÅÅ«ï>58ÚJÿ¥úß—V³3ÿ¼%múÚ}×˪éëo(|Yµ’úÁââ5ÎÐf(ÞòÆ÷§O_ü¥Íçò¢Çíé¶ò&%µoƺOyã‹Z+/~»¤ïbh>d(†žchÓ—ÞÒ:_Ôž•÷/­VÒóK[¿žv\÷´[y¹Wy¡ãßÞž•wxÚvÕÐ}ÅÍ÷õ­“•VÓ÷¼²Ž ž9-n+ë8¥¬Ï)ïíYÉsž¡º‹QëÒÿ¶|µ´êLÆò/eí— mÆòãeÝÊk^÷OiO«~[ÖñNIã8}¿{–´»îTܸÓXõ´²ª/”vžPRþ›>céKYÅÙÅ•‡âòW_xCë¦åÕï”WºômÏ:ýÿÔV^㯲Ê_ž•ü¦¬ü²±âE}ã,c·§%·å%1V+ïõƒg5ß-i3V¼dlý,/vÖÐV^é2´•T.Ê+Œ¿;.î~}÷Q8Cë&%üïx¨¯/ ­«JOqáJº¯¸ç ?mŕ㲾gqñ©í36ƺme}~YÉyIÏ¥íiÉ¡†ÚÃâòG_¼jô뻯¤òý´ä½¬õH4_ÚúU\úDcµsÕàÔðªá/î¹%峡tÒucÓUÚvõßÞóö´é¨èÿY=o†Ú Ѽ¡ç•Ÿ±ù`(=Æ¢S ¾ç‹CßOí=}Q3¿õ½oqù^ÜóKKn ½oYËIé,+ûalzEpúî/-þéK—¾x =¿´ðˆö‹š¾öÉÐõÒzOµóŠ»OOjô‰à 対ç—TîMT~/ ¡ufÅ~Üc pÝû“ðH+¯ìBóÚµûšìKøLol’°43¶œqá6©:\óFÍþ | j±„~%pm\£Ë nĶ=ajÉûÖ?ôy~Y•7ä±F›œûýäÁ°sD)oŸº'Ï§ÞØÔ÷y°kðî÷“mÒ›SS‚­‚wš¶ìÌ_¿DÊcß?§ž¯Ùa!8çù^G“ÖÕ•Ç{<mùm_°Ý_}ÒüSfòüöM+~˜³y¸7ÀçÂã?ÆbB)°fi§›ãÎæ‚ÇÏÔ]¾ß©Ò¹+óÃV:Æ@bËkªµ¹,ÏoX×âfbè‡ò8ùÛ±ÉÛ&¶“Çž—^vww÷€˜ÍuýµÞ€Íg â+´];Eô{7á‹:©]Žç½–ºêá×­å±ÿƒµÆU}üü½%þAf«Ýæ¿%oƒÐÍ×Þ>u÷@!¼©&ë/¿: ì^ë´oì®Bëñyºco؃vfÿ_"­öƒS»E!f©»!uÄqiŸ ¼n£$`°ã͖֜.kþž‹EÜÏIs±á†×+šOî]P. ­ŸøÉyùÕ7Áž¦_Ž«ÚxG¡u—¹yè$´sú½©#â–Þ];ûy¥…ÎOoøÇså±Ó‰áéþwöBP³Œ×+¿ðxî¯Ñ¥íõÏ!¦‹ôlõÁ;}_ßi4‡}~:.T{ÿ×ÓÊë¨VˆOçEé „®v•öAæ ’½“ç×®²ݶQr!ø¨¶Ù:…õ„NbPÐcÛbﺼWC䘙È?÷¬ øn‰>>HüvÂ.núãœÝy¬¤\ï:Hûåù]yÒ3·„˜\yüÒ/tšhþ[ œ[ù¸àû ïµù´¿ÂõœÈ‡h÷ü¬zKç<áÛ‚1¨·G^Q:Wž?Úü­‚~N¯ÿQ§6áà;’X ÚÿZHú •l$ç¶ßÐ=$än À{¬m;ªå½¼mÖ}/Ú­ä!/#¼S³õm.wj qÛ~O>gÌÍÖµOØ^‚óšhm-ž‹«ä-¯Ÿß4é½z×5±î’½mVÊõÔ?G.táxwl%ÙYÐæM´Ð9)úJú Æ×BûVýô&Ú×÷çn× h¿““s âßë©#ãH_ˆ|~BoLöZÛñè5õZ`\XIÇî?~xBÏÀz¨wG›{â8d‘i:ƒ ž1¾¯ @¿ >76éGO^ú ã}ï09:zj¥nX–þÕÞwù²9}›éezPß5/}/ù1yÿ•¦øL…ð …’ÝëœÈC+ÇÍ…Ã?¢ƒSç–ÎÐ5^sëG‰¯àÞÄ÷»Ñ÷Oèä¶uÝy;·‚¶† ú;7Ÿ IoÁÿŽvc¯áò>§×]hÿ²è5<ý‰ÿLqû•Éá/„ÄÍUÈã†{S¦;†k¦¨?šÃ:)\õsfHöRÕ~>`—îš+ 0h"Ú{»‡'_>^}D˜F¹é½ ÏÙññ:¤;Ðú[ÉáÈûƒt@ÿ½ï{&ß^éq’ƒ—×w5é'ÁçWÀñ¼*Ÿ6l1r)ØX®9ÿõ¯ 5ovÄ} Ç¾„¥YðeËGµ]Ÿø÷À% zÔ’/D=föÇ»iI¾Ôä±x}Uí†×LÞ6>öÎçáII?e8—ß™?rÅø–íÑÆû¾´|®|¢ \ac¬iåoLNéâ]ôoà4óÊ ½FÈû=⻂]ªtÌ4°ß¾QzOp}W2«a½©SŸvË…tæã®;ÕSøqXåoÓøÏw‡‰àe8Ÿ¡õö6Gýñ]yh§ŽRX÷IøîZ W82œ»Ýq‰ÿ…ð­ùB GÇ@ü½+8ö{×I¡·>>ÌÎxwùóÖö™6ºS§»6bò3ÍORÃVòüÚ«}uï9p²Ô¡ý’oÙGx[Z|­ t óçÍwP^öÖ´-’¾+~-Ÿ[õ(] ð³æœ—uÞ}°›k¾å‡ª°öþa¬+Øù˜IöÚ¤· ûoÞÃx`SF£ ×^žÛ«`Ýãæ¡Ÿpze)Æ)N Ž’=çïa|ë2ç%Ì \–¿q§G«{xõ؉ýÞ–0¾òŒ›/åØí7ºÀ81°‰ÏåNMÃýa31Íò÷ñž‘;ÿÀ:Glåm˜ßÄ4ñÕ‰y/ˆkß ãš¸ö¬.÷Í~ôCño&Iy1$Ž\Øöú²³ä%96ÐÞÆxŽÙ°<àø!–gŸ›ý3ÆS,4Èß½nSÝüýÝv×J„mC¤°o¸Ùÿ†û=jÑ¥CCa¯?‹¿ö·®+ùCðÎiƒ÷ò™Ïèð]ÒKò36d1µ–ÄjÍE~†ß<‡üHhŠqn‚æYcZ ã9mëéu ‚„5%û ‰ÇÂQ޲>|å!3@‹ñ]vÓƒ“Ûn¡¼œû²ÞoãÀ]_é^˜×£`Óî‹oýúóØ”óŸ;:J`ëöSñýz¾n•Ñ~óï,NsZÁîÁûs‹1.ÝÍâêüúx^ftxfu¹¿ÕùÄã¦C«°~q]‡â½Ž˜³:IćF¥ÍƼ<&÷æ©«Y]/nÉ=Ä—ò€åIgæb<š°j*‹³Wg#ß2Ò–!]Y3ñ!vž÷é)¿Ez}%a‡ ‹×bÏòäsf Æ>óXü¶%n©óP›µ°ÿFÞÛ?ñÞ/rÉ|ç5N¢Ü{]Í—Ë_^ê%ð*Ó““_šâü(—?hln€G EÜÁßfK…Øu•åK'u=:Þ|ÈÞÝÇj‘pCT€V§Æ@3žñcë ÙÍ,ÛûBÒ:&בûÏ =áµî >h¶žG9Û=Ù‘ì*LO÷Ÿb÷;² ëµ×åXsF'­s6ƒß½ ˜/î¨7ó÷ÅWtä¬OÖõY·³ÿh¾a|ÇÕïuZ§‹Œ`ÓA¦/Û—0>úõ`ûwv{ˆúk÷ÊkºÃ0IQ ÙÃÃóæH8|‰å÷¶2û—öq°dLH]2;WÔO²2¹I}ÄâõŒšýPÿ“Ÿ[ï¯hJbñwªÏ~©~Qn˜ö~&ÞïÐ{¤8r ’ôFÆ›…ùAXP´¢. uPÔÁÁ{ë‹('‘Ûôów9=1Ï®‡ÕÆø<×±|3>ÙóßÜzÓðž)“—¡?™6™ÑSo#ÚÛ-˜žy|§¨CîGþä¿LÐÎhÒ¶ ßϲ÷;¸Éǧ0¿•áƒ>ûãðÜ=¾È7M߬"ëñ6X'+45jT÷8¥k‡=/¬_At§»’Éã¸ÅÓ$}“Ç3ú¢^¯¸<ü£œ©wåygïPÉž?Ù—ñ·u}Ø·ŠÙµ]ŽÌŸ?ïˆv>⿦X¿8õéƒ"åt{üU´«Ñ?1{—âÄògÍç;QÏâjt*øþ4c4ê—æÆe´ïß½±ß-îf žŸXg“¢Ž”t­§$¯òøÈϯ ß’†\×)â;ò|æP¬Èã´—Žð+É(ß¡ ÷7«v’^ËcÌÿ䮩'ÅwÅê\ik2 ÞÎ|¾˜ù˜ùÇ0+f×Òk¿‹üŠßj%Õ7dø÷ÛgÙÿ<–<ÎÙÞå)«Kg„ËÃ|"7´’lˆY}í^ä†|ù9Ú‹36ùúßj(Ú¥s·/a|~Ý]Q?r?{³¨:>ä„¶Ï·l¼öªëÏÕ¶ÆÀˆ.ŠºoÚ7AŠújŒSeE}9ÖrÚÃ=×¥gé ÏOvczr¸¹úô ‡¹&0þÇõa|Ú³9㟽Ì>…Ú3?Ôí¸Äš¢]Š<~.yéu¼Rö`´×ÉïF½O¸!Å òù¹;Ÿ/H·<ï?®Ê›_îtŻĎ|¯`}£Ð¾É.v‚Ô ü ÈÞgèä°ë9QgäùKÑJ>%ÿÕ¿×ߨ‡ñY®Óx¿#u¥p` ôϯCÆX£Ê«ï[,>ô¼§¬Ç®Í¯cÅn¬‡÷OÞü ÖYv$ÝB?™çö-õ¥ÙøN ç™Ìüzs0¤Ýýë¤ÑI™EÕ»eüé[*¡Þ§ö» ð±‡›b¼ШÖoÓŽ°ïYšaǤxD†‹|×GºG!¼Ú‹§u§uÌo¦<`qgh;û‚ï(ÃF¸"_3Ì \ØìuŽö$œ®ÇâEͼçu M…¬Î,öhÝã7í¼®(Ÿ :_Ðç8‘ǾoÄƺ£Œ?³5ãW¢CÇ¢¾WÉp‰g¬Ãñ,FOÎ$FØ1Vÿ›T õ8Úô êAÐgfÆUØ)m‹ºŠúQ\ÈACê|òøGÌ»ì쥰…Ö#Ûb´NnXŸµ±šøH—yAø"[©~ÁŸD»cwì¿]x:Â?¬3BJ=ö]4þr-”—°›ÇþŽ.ˆ¸òæ!a9¬Þ—ûçÑ¿õ÷+VT^òêÐç!Í”½GØ#EýY† Ÿçªø^’fT”×ãCðÝ ÍŸÎò€ÕÝ›¯Lf+ÏïêÌìsØëXŸÎɳAÿ[‡ÝgǤÉëœò¾à‡ 0^ ©_ô÷ÿ˯`Ÿù;Ú 8ÈêöG"Y}Њթ3¿üå5íò}ô;! ß!C†¤âwÒð+b}9j3‹‹ÜlYÞ»fîÿtæä5pÈÃïdrÇßùÈGEŸRÚ¿ãüܽf-ú%mß ˜oÙPåõ`ÆG’Ý–á’µ½ß§^cv=¥aôciÞ@ý ë‚þ’ÂSq>äìÌ£#ü¦#?3§²úbàF¬CøÚ>èO4›>Â|Íßj~7I¬ðgŽ~!Èoï©ÌDÇ~‡y²OxWŒÿÒSÐå4døg½)¥ Ë!Íã„$ßOä§7Ó×ôÀe »æ“2“Å—mv =M¯·ýïü‡¼/úruÔk¯&,¯ êªàwÊÝU˜§9ŸÖñ¶â{îñy,Î:ˆ@ØH–ׄŸ@û ÎËÄ>§ø^g·ýnü^˜ßŸü‰}WŒx³!æc>1€pG5“‹¼GÐX²Ë óª…µ¾õy=½A+”㤘·±nü"~×ClÑd>>‹üßwã]rfòi)æùÚãýAJqµ¼/öt<ÊQÒ&ï‚u7™ˆßW`ûe–·Œeõ÷ïQŽ<~×E­_Ü–áÃ嬳Ëó³²¥¸R¦?‡rÜ( ãÝ £'²ït£^E|ÜÁº½¼?iK}”cϟǸ sP ”“Ä[ìrCñã© ~ûïKF½ ´¨„ùaÎòÛhG^dõ“´&ì;Dê_J9hÎìrL ¬C¾æLÏ£7ÎÃú~Ú1äoòÙÕìûö˜õø/-e>ÚñQ.`[sÉÀÊx/j_B?³“Õ¥ÜÜ%¶W•×î¬A:ÏT[ôG¿Êâþ¨F[QîÚTÄù;é·â#´ÿQç0þÞßíˆâ»Aò(V/É­sñÄŒeñE€/»gPö7’Ý»ïšÎŽÙšùP^´»m þ~ R üAä+u$ƒ¾Kúàï$µ]΢~¥-eñRZ"û.à:ƒÕý5_6ÀuÍW±èo‚OÊnjɡ©]N€ÐŒGãCðË~7¸÷¬5¾ä÷¨üÒA–·îaö4áÚxÌÇŸguô$V÷ #include #include "gmm.h" static const R_FortranMethodDef fortranMethods[] = { {"wu", (DL_FUNC) &F77_SUB(wu), 9}, {"lamcuep", (DL_FUNC) &F77_SUB(lamcuep), 9}, {NULL, NULL, 0} }; void R_init_gmm(DllInfo *dll) { R_registerRoutines(dll, NULL, NULL, fortranMethods, NULL); R_useDynamicSymbols(dll, FALSE); } gmm/src/lambda_met.f0000644000176200001440000000745114247643114014072 0ustar liggesusers subroutine prep(gt, lam, n, q, d2) integer n, q, i, info, ip(q) double precision gt(n,q), lam(q), d2(q), dd(q,q) double precision tmp(n), tmp2(n), tmp3(n,q) call dgemv('n', n, q, 1.0d0, gt, n, lam, 1, 0.0d0, tmp, 1) tmp = 1/(1+tmp) call dgemv('t', n, q, 1.0d0, gt, n, tmp, 1, 0.0d0, d2, 1) tmp2 = tmp**2 do i=1,q tmp3(:,i) = -gt(:,i)*tmp2 end do call dgemm('t','n',q,q,n,1.0d0, gt, n, tmp3, n, 0.0d0, dd, q) call dgesv(q, 1, dd, q, ip, d2, q, info) end subroutine wu(gt, tol, maxit, n, q, k, conv, obj, lam) integer n, q, i, conv, maxit double precision obj, lam(q), tol, gt(n,q), k double precision dif, d2(q), r, tmp(n), tmp2(q) lam = 0.0d0 i = 1 dif = 1.0d0 do while (dif>tol .and. i <= maxit) call prep(gt, lam, n, q, d2) dif = maxval(abs(d2)) r = 1.0d0 do while (r>0) r = 0.0d0 tmp2 = lam-d2 call dgemv('n', n, q, 1.0d0, gt, n, tmp2, 1, 0.0d0, * tmp, 1) if (minval(tmp)<=-1) then r = r+1 end if if (r>0) then d2 = d2/2 end if end do lam = tmp2 i = i+1 end do if (i>=maxit) then lam = 0.0d0 conv = 1 else lam = -lam conv = 0 end if obj = sum(log(1+tmp*k))/n end subroutine ols(x, y, n, m, lwork, nrhs, info, coef) integer n, m, lwork, nrhs, info double precision x(n,m), y(n,nrhs), work(lwork) double precision coef(m,nrhs) double precision xtmp(n,m), ytmp(n,nrhs) xtmp = x ytmp = y call dgels('n', n, m, nrhs, xtmp, n, ytmp, n, work, -1, info) lwork = min(m*n, int(work(1))) if (info == 0) then call dgels('n',n,m,nrhs,xtmp,n,ytmp,n,work,lwork,info) coef = ytmp(1:m,:) end if end subroutine lamcue(gt, n, q, k, lam, pt, obj) integer n, q, lwork, info double precision gt(n,q), lam(q), one(n), pt(n), obj, k lwork = q*3 one = -1.0d0 call ols(gt, one, n, q, lwork, 1, info, lam) call dgemv('n', n, q, 1.0d0, gt, n, lam, 1, 0.0d0, * pt, 1) pt = pt*k obj = sum(-pt-(pt**2)/2)/n pt = 1+pt where (pt < 0) pt = 0.0d0 end where pt = pt/sum(pt) end subroutine getpt(gt,n,q,k,lam,pt) integer n, q double precision gt(n,q), lam(q), k, pt(n) call dgemv('n', n, q, 1.0d0, gt, n, lam, 1, 0.0d0, * pt, 1) pt = 1.0d0 + pt*k where (pt < 0) pt = 0.0d0 end where pt = pt/sum(pt) end subroutine lamcuep(gt, n, q, k, maxit, conv, lam, pt, obj) integer n, q, i, maxit, n0, n1, conv, ind(n), wi(n) double precision gt(n,q), lam(q), pt(n), obj, pt0(n), k double precision pt2(n) logical w(n) call lamcue(gt, n, q, k, lam, pt, obj) ind = (/ (i, i=1,n) /) i = 1 conv = 0 do w = pt > 0 n1 = count(w) n0 = n-n1 if (n1 < (q+1)) then pt = 1.0d0/n conv = 2 obj = 0.0d0 lam = 0.0d0 exit end if if (i > maxit) then pt = 1.0d0/n conv = 1 obj = 0.0d0 lam = 0.0d0 exit end if wi(1:n1) = pack(ind, w) call lamcue(gt(wi(1:n1),:), n1, q, k, lam, * pt0(1:n1), obj) call getpt(gt,n,q,k,lam,pt2) if (all(pt==pt2)) then exit end if pt = pt2 i = i+1 end do if (conv == 0) then obj = obj*n1/n + dble(n0)/(2*n) end if end gmm/src/gmm.h0000644000176200001440000000053014247643114012556 0ustar liggesusers#ifndef R_GMM_H #define R_GMM_H #include void F77_SUB(wu)(double *gt, double *tol, int *maxit, int *n, int *q, double *k, int *conv, double *obj, double *lam); void F77_SUB(lamcuep)(double *gt, int *n, int *q, double *k, int *maxit, int *conv, double *lam, double *pt, double *obj); #endif gmm/src/Makevars0000644000176200001440000000005614437370205013323 0ustar liggesusersPKG_LIBS=$(LAPACK_LIBS) $(BLAS_LIBS) $(FLIBS) gmm/NAMESPACE0000644000176200001440000000654214247643115012267 0ustar liggesusersuseDynLib(gmm, .registration = TRUE, .fixes="F_") import(stats) importFrom(sandwich, estfun, bread, kernHAC, weightsAndrews, vcovHAC, bwAndrews, meatHC) importFrom(methods, is) importFrom(graphics, abline, legend, lines, panel.smooth, par, plot, points) importFrom(grDevices, dev.interactive, devAskNewPage, extendrange) importFrom(utils, tail) export(gmm,summary.gmm,smoothG,getDat,summary.gel,getLamb,gel, estfun.gmmFct, estfun.gmm, estfun.gel, bread.gel, bread.gmm, print.gmm,coef.gmm,vcov.gmm,print.summary.gmm, confint.gel, print.gel, print.summary.gel, vcov.gel, coef.gel, fitted.gmm, residuals.gmm, fitted.gel, residuals.gel, plot.gmm, plot.gel,formula.gmm, formula.gel, charStable, specTest, specTest.gmm, specTest.gel, print.specTest, momentEstim.baseGmm.twoStep, momentEstim.baseGmm.twoStep.formula, momentEstim.baseGmm.iterative.formula, momentEstim.baseGmm.iterative, momentEstim.baseGmm.cue.formula, momentEstim.baseGmm.cue, getModel.baseGmm, getModel.baseGel, getModel.constGmm, getModel.constGel, FinRes.baseGmm.res, momentEstim.baseGel.mod, momentEstim.baseGel.modFormula,tsls,summary.tsls, print.summary.tsls, KTest, print.gmmTests, gmmWithConst, estfun.tsls, model.matrix.tsls,vcov.tsls, bread.tsls, evalGmm, momentEstim.baseGmm.eval, momentEstim.baseGel.eval, evalGel, confint.gmm, print.confint, sysGmm, getModel.sysGmm, momentEstim.sysGmm.twoStep.formula, momentEstim.tsls.twoStep.formula, getModel.tsls, summary.sysGmm, print.sysGmm, print.summary.sysGmm, sur, threeSLS, randEffect, five, bwWilhelm, getModel.ateGel, ATEgel, summary.ategel, vcov.ategel, confint.ategel, marginal, marginal.ategel,checkConv,getImpProb, getImpProb.gel) S3method(marginal, ategel) S3method(summary, gmm) S3method(summary, sysGmm) S3method(summary, tsls) S3method(summary, gel) S3method(summary, ategel) S3method(print, gmm) S3method(print, sysGmm) S3method(print, summary.gmm) S3method(print, summary.sysGmm) S3method(coef, gmm) S3method(vcov, gmm) S3method(vcov, tsls) S3method(confint, gmm) S3method(confint, ategel) S3method(fitted, gmm) S3method(residuals, gmm) S3method(plot, gmm) S3method(formula, gmm) S3method(print, gel) S3method(print, summary.gel) S3method(print, summary.tsls) S3method(print, gmmTests) S3method(coef, gel) S3method(vcov, gel) S3method(vcov, ategel) S3method(confint, gel) S3method(fitted, gel) S3method(residuals, gel) S3method(plot, gel) S3method(formula, gel) S3method(specTest, gmm) S3method(specTest, gel) S3method(getImpProb, gel) S3method(print, specTest) S3method(print, confint) S3method(FinRes, baseGmm.res) S3method(getModel, baseGmm) S3method(getModel, baseGel) S3method(getModel, sysGmm) S3method(getModel, constGmm) S3method(getModel, constGel) S3method(getModel, tsls) S3method(getModel, ateGel) S3method(momentEstim, baseGmm.twoStep) S3method(momentEstim, baseGmm.eval) S3method(momentEstim, baseGmm.twoStep.formula) S3method(momentEstim, tsls.twoStep.formula) S3method(momentEstim, sysGmm.twoStep.formula) S3method(momentEstim, baseGmm.iterative.formula) S3method(momentEstim, baseGmm.iterative) S3method(momentEstim, baseGmm.cue.formula) S3method(momentEstim, baseGmm.cue) S3method(momentEstim, baseGel.mod) S3method(momentEstim, baseGel.modFormula) S3method(momentEstim, baseGel.eval) S3method(estfun, gmmFct) S3method(estfun, gmm) S3method(estfun, tsls) S3method(model.matrix, tsls) S3method(estfun, gel) S3method(bread, gmm) S3method(bread, gel) S3method(bread, tsls) gmm/inst/0000755000176200001440000000000015044423073012011 5ustar liggesusersgmm/inst/CITATION0000644000176200001440000000103514437642761013161 0ustar liggesusersbibentry("Article", title = "Computing Generalized Method of Moments and Generalized Empirical Likelihood with {R}", author = c(person(given = "Pierre", family = "Chausse", email = "pchausse@uwaterloo.ca")), journal = "Journal of Statistical Software", year = "2010", volume = "34", number = "11", pages = "1--35", doi = "10.18637/jss.v034.i11", header = "To cite gmm in publications use:" ) gmm/inst/doc/0000755000176200001440000000000015044423073012556 5ustar liggesusersgmm/inst/doc/gmm_with_R.pdf0000644000176200001440000127017615044423074015364 0ustar liggesusers%PDF-1.5 %¿÷¢þ 1 0 obj << /Type /ObjStm /Length 5095 /Filter /FlateDecode /N 83 /First 695 >> stream xœÍ\YsÛF¶~¿¿¢ß&©© ÷S¹©ò{[¶"9¶“©b‚iæ<3L9Á,³²dŽùÒ3Ï‚µ,°²T¬dRiŤ`2œ%SÚÜÄ[Z2©™ Þ0i˜Vt¶L[ƒÎŽiz]¾¦÷3Ö8†1LPŽ)ÁLé-Sãbx¥˜Õ–€3ð¾2Ì©€~(:Üf±°³ƒ` ªdA ­ Ê— (ÐV,”è¯5+^Ò†•Æãl1+B¨«ñ‚Ç… iLJMx㢠–Ñ4¤ƒyJ$3´c xS–X‹‹àÑ•À¼ +‰µ2´\+bh1ÃY9¬)f,•·%³€¬JÕd-ÑÙ²VÎ0Bjƒ…¶€¬-òöô: ë@¯²‹ÈFºICt€lJ¢- [LŒ9"“òŽ9@¶@ [‡ù‚¤%ú‰À,óDTpÈ΢–]:ô0kéJ¬1€ÇA#\h¡FÁ=do²·À Ô’Þ2`J_b,¬0˜Ì)qA¯@«„e(UøŸd|¯j&Ç“f60þâ²™ÕójþíýÉ)&5^~¼¨€þ³Å)ûé§âÞes¶X²ï¾Ç“e5iêÅüᤩØwÿâY„¥¥ñ?ó!þÑö£wžM^VoØûº9cg€½\V'xü´úø~±<^E˜{‹ãOÛ_.Ž/§à=ÞÆŸ-VÍjº¬/VÖ]/þª¦M„ø²nf]å Q*ƒõiåR[‹„–é}•áª4¬J­?[*&¶4œÏ M]&X*è|6ùœÇy̼*¯€ÊSUyŽªÌc——<7Ufxe†Wfxe†—×ùO¢Ê¼©æ@JëÌš{Õq=¹¿ø4i'‰ñõÝŸ,ѵåÒƒjµ¸\N1;šÚ£E$n™Ùy¹˜V @ðý‡°Õ‡~ú©»,}Æð€Q/WtA<›àZùžH¬¥&¿ù°Z5kFâo~ÿ§pXcTá 4ç—³aþœÆWÖ¶Âç‘E/ñé'ÕÉ ôQ)„«„ðÐM^á÷][œuºïÊ|¿úiDØÌ’:MBgVI«bÑL¢™I$3‰b&Í3³°IPL‚bSn¡™ûbšYñ$»?YUñm~oïÍ«ÿþ烽¨Cþó|º8®ç§¤3#ÔHÒg“%šq<,ê ¨%)6”Ž]`a¢ iâx4ªlOƒ½®›3Zf§lœhúy¥Ò*›®-ÄÊ‚Øt­×ý@ÃÞyüë>Qd|£™ v¼xtN}w9T~3ÁÈ[2O!bëaæ,ŠƒÂòdnðL“Í„¥7h®%4a|‡Ú!ÏħF6Á”"º©áÈõAljîôéEcÌú¬…]ßÓ¤¶JˆÓÆåÌÏ龆æÑÐ2´8í½¸ô8ÓawÆûs-)«)ñ#á wñïª>=k›àâÀïø=~Ÿ?àùÏüÌOÎÏ'üßü ÿ…?åÏøÎ_ðçÕé„ïóýUÍåüÖ§è÷’ÿÆ_ñ×ü ÿÿÁ'|2½l*~Ä–“éÛ ¢~Ò´×K›Où´^N/ÏOfÕ>]ÌsüO#óŠGôx5?ž¬Îxõ÷ådÆOø ý«yü÷®â'¸©â§üŒÃ¼žUs^ó¿ø[>ãçü|2]"~P\|Áøÿ‚_TFTâUB†¹^ó‹ÙåŠÿÍÿ¾\4ÕñÑ,öj©cl¥Ë%_ñUu^'¼WÕ; ¾ª?ð†7g˪âM=;Æÿïü’¿ãïùþ‘ÿ—ÿ·Z.¾ê3ãEËÿh69%$ÊÚý¤L~€ÎŠO¥e¤ÒŸéù£zViCVÖtëùä¼êƒ'paêé½ù)4,š{õjíÅ8²?lªóW°Õ=Aï舑¾yõðɽÃ_ÿð°§p æ÷æ«zsc£z`¡GºG uüþuóüF9“tÎUÚD Úmy¢¢­Øü4tXÒn,<¸×žØ\ Ïc‰ÃL?{4ü²GkÕ’š¼vѧ·ˆÏ=¬®ãöhݪô5­´Ø•Ø„îVbó7yFQ~ïÍó_>Âh÷{„Ï.T‡ÜRÉmFä.w$÷õ½ÝŸÏcy7f‘Òö Õ'ØÌõ5|„×Õï­êv¼µ;4{ô1[½>bAÑÓßeO?&œF= ]I걎:q=xUWCJ©ÇBn×D"=+®{1RCå5û̲+gB}Š3±F#Î7§f‡›“/òyûP¯/3o'»»ñZπ܂Ö'¨³+0ô’¹ÿ;Ûð‘åŽ6;[웈|Àm2’£ŽŒ Øvg[­wµÕ_ï?ß@¾€ßÑR›-QBÈlÃnâaÀI&†RÁoÅ"Ž¿îè08Þº™ßšÄ5õ#¡·ºzäãuh¾¦øÈ3ƒeb;K)¬!Á ¥Ý×,‡™z÷ȱ+µ•øµËãj™ãΈÃة‘#ôœI!zY”°GVÚ‚òmˆ~@y£ÿPª¢ðZ"Ú•EiH·›BY ^t`eþìÿY=Ûb#ðQ8î’›ç’à’àRí’âuII÷Âê›IX­uöKnWoO\ŸsN>-_Î4åÄUÎ[å IýÅø{ûèoDÿ××¼~}^{ù¤Ÿh-õuÞšÒŸm­ë[+åµÐFÏ0CÈ‘APnb†|0º£­çkÛÑ}ã½¥'¤ð”U&¿ÅÒÖ@H>¨GÏt Ø™ÖÚÁÓ ¥Ko‚‹)¯Ê8§ø~$kY†­ó€È8ö:Î'·ãLr›ò`áN4ìë9ÚÀÌiïÃêÔÓAb(ûoâÞBø¤6ì"ý(>Åïmô¾CŠÕ_wbôé8þnCî¡QÚó×—ÐJ5ƹ6¬îâ‚g/úѳ¶9zN; #m†*z 6]%ýùŽWªiDçŸò\¡ÔFbºÿdÿ×71Þ^_LC‹ÙFLÇ1´²#1ÝÕ4;MlMÎ]JÅ ¥$¦K˜2xN¨5sƒY¡üLöwÛìŸÙ&]ÿksq>§ÃRFÎêMäÇzP¬PtWÉÓ¥¬[éi뇄ÐE AÛ`NÆý¿Œs‰¹p=PEp í$Fd<Ì ŽÄ6n¸P‚‚8J¡éäÆXé2Ðÿq[ *!®JyÓôÚÏYÛdZ+ŒmÚ 9»8ƒƒ[5’ɳz-‘ÇÕ ÷ —«š$•zœòÓ˜¡;¥í¦j¹öxg“ó£ã ŸU«%Â.sÜ¡lêÉì¸>9éúÀ˳ùÁ1‡·šQº 8¹„ÒkÜ-m…eÎi+¶ ž Þ@vö…¯ G’öôåﯞR®ä÷¾¤µnÚ5‰rÕÙɸÍ.~›×€Za~íÖÆ6Ùóƒ\ôö_ûÜgiW¼#Å•ÉôÍÏõõëfTÚ´{+îÝ”>e;n&*“årñ>Ù„£É’˜‹˜?®Ö™äl3pãÉdÕï&³j>­Z‘HVŒ!]×s˜¡óËYS_Ì>òyu÷sÛ/'ÇõVõy=Ø—s¾åj2ë²>U \Ò3r·ðý(];`Ë]ùÞ윯ýýчOÿ¹wxoç|mØâ–»z„Û²WþL7{›kštz½ÊÛ¢/];z{°Ì£Lé`v]faw^æWûOïÿõW7ÖÎ{××Ûq-‡ê…‚¾ÝvЫ\GÈl¶çn­-œK.«U7µ×ý_Êó]ÿóAî8o>ù)%Ÿ15"õ'[+mÂ%ïŸvºbdB5*F'¿?Vž˜-yaû¥‡=£üQÇÄFãz’ÍhÊ’fãI~-Lf ¹êDÞ)Œ™Åm¦P¦‡¥³‹§Ž§}FÚY˜]õÙ‹Ã×÷‰ §Ë¨Ñ_ú„ÔðÐlËáo¬`»³½•sC™3ŸáKŒ‘”z[4ÖaY(®ØiÈ ¾c‡ARs›ÙÁda`VoÏ' 8c~RÏëæc¶ Ëú¼ê²¸õ*«€GRûˆ‘úL`G™¢>‘vv†®ŠAÆ[9{=}ø?¿é;CjÈcK GQ‡ÜutØ@]ƊܶnÈeµPúxÐf{ÜpÏûÛkuÓÛ•ñ¥ŒG,@)¢Í6 é ¿¨{\×·=®dE1r«dZ]ý3LøMΤÅdÄ®='©ï#­½¡ûõé¦qTŸžvÜ£öYÞ‚7{í£zØì·›n«¹èmèw@·íþ£zÔÜèW4”ï6:­£~«Ù\W6×-°5ª½V·y4hvÓ-Ä¿Õ%,¹Çõêb6ù¸¹®MÏP”¶¼RGćÐVä5•D»ÍºÙbë ñŽZB‰ð•ÊZú›qBYšB–ÔKSøX¼ã Ú¨‡U/¨Šv§„ò­#TaĵHiaŠ˜)¼}œä'ëÊ 5ÂÉ«"À÷±eYˆ’ ’EAÅ“NÊBDçÒÊe'ãŠ@å¿_§1ñdÀ:#œõ‚|@‰ƒû KQ” ¦®ÁÜN¥-Lc‹“ßYÈì·ÁÉÙBK;b(#‹XMÞâ"Zs·8‘ )çÖ8iåÀäækà4æqg]¡ p¢±­KÆÉ€Ä¨ß%N´ÝF%I-NVÙBZ÷mhç|YPaáÇ…*h[ÓȼYƒæfÕî%1#$2JPb1UûMpR²,¤Ù õD3Nt¢mœ»DÉ ®Y/“Q0Šîf(˜Àr°zwTÚkE_ôäëø5O¾¦Ï8ò5V©öÚÑ—:ù:0•$¥`*´×ñëœ|MŸ µ×”ª×ÃíÚcd'ɦN¹Üܦ!lÂÆæïÒ\r}~.Ïïíáf _²‡›«ýo¸‡û¹šÄ+U€ÒÃÑ)l¥ð…SôŽ.H>` ¼ûʪm;B°H`>¿ÁÖ›Rx_§±€Ø oGzÄ@½ðŸ&3î¨`Ï”¸l‘¼¤Xw 'ÛkJ;¶NöË&×[Æ£šQ½†Z'@´|ížZú–·4à ÈhÓxÀƒÃ ÈäCgqúùº“ò!þªtȶÂ{Ѷ2þÛÁ“µ|L)Í“n}GÉ qþþýûâ¯U3iV‹“¦X,Où;mx-%ÿ>Zý«E[öIQ˜@ÅPøÂÇ@ßFqœS;–â}>ÊÊóZ¨¡¬ cè£c)¯B#ð7©tÂnWOsWŒÜ{^üp±\Ð'¯å‹Éô-\©ÿN–7Zh;A…gÚ"[e!í¨ÿ¿b]ê"xxvð²àP>r…!õn¿ÝZŸ¿kæ‹åùM07@•’ÞZÃë²›õ¦ì’ݵÄôN0Ïk®•.lÌù(b‰¬S·¼â$Q‹Å‡b2-.ßò‹Ë#¾wïðÐÜc2ï1hÊ‚ÊyE¡©Ð 8K»c"áÖº©Ï«CØju#äáÀÑß’ÐÚ¤ÿI#ªEØo|ùM±§¯ó?‹É…Æ9l˜\‚[ÄŽNéážÙ\•P)p%œ|š‚ªpæÛ (øÿh†ð{Õ܈i&òþm¶.ÆvVÐ_) ïoYo1›°òµ73›†|q,¸„^!óI±0-<…¥»0_‚²ƒ¥/o„²ô2Ф ”R"‘´EI®«/vuN :é¤J'ôÉukbŒ‡õÉIµ¤’¯ø×2Äòë;X*ì /YRñM¹ÀƒŠhùEM¹'*î ¼P*¡ $Œ×£|*ðÒONê[‚¯6~ŒÍè;}ªÔÄ‚ÑØ,}MCá@ód–†¥¯,cq7-gþšrhýÊkú êkV*Öûü™Q¬±ùH›þ6 Ëßd§¿¿r¾­Sq Ì?‰î­­m,ê`˜[·æ#)›zúk©îƒ­«An"U®M¤$j·¢‘ÒG›BFÖ+p¤Äj[µ8Äš*r$¥Gû%”"Ý”9RžŸmêX§ˆ€*Xa²\‰™*asÅä--C|SzÀ†ÕôÍf¯ €õwñY¿(€ 6õ‡ðÇœR^º_JAYè~åÁ`Ð>Øû#°ñNdžM)ëU°~QAôÑnH7ôuU·ü¡7ë†u«!X¯’j¿úõ ôF% ,~«ÞŸâh†àÁz)¬_ƒr»Jâ–€m7©Ö> Ý™iä“åÙ‚¥ÂùdCàäQÍ~T\¹NŸ¶rÞ¤µ/ôe’¾L¸¶(û¬þqÖendstream endobj 85 0 obj << /Subtype /XML /Type /Metadata /Length 1387 >> stream GPL Ghostscript 9.55.0 2025-08-05T11:41:47-04:00 2025-08-05T11:41:47-04:00 LaTeX with hyperref endstream endobj 86 0 obj << /Type /ObjStm /Length 3406 /Filter /FlateDecode /N 83 /First 754 >> stream xœÅ[[sG~ß_Ñ»•¢ïתTªÀ‡[yò`k‘%G’ ɯßïôôH£ÉÈárÉÓ3Ó—Ó§Ïåëszb`’ÅÈ”S,&f¢fI2çKŠy¼HšEgY2LɈ«EÕ XrL%¼Kžie%KiMÕ#Ó& ubÚEj$™F£ *(43ÒH 3*$,3F36 ké™ñcÉÀÐ]Äyê'1k0¬R’Ùb•Ṙ:J3çAºR†¹(©Že^*8æèWÊ3o‘…‹ƒë¥w?Sòx Zž[P­ü$ÀÒÒlrûP÷§T7†^e+¶>ªnK3 f`À›çe"èŒJ±ÔjÞÒs8üj&4ÏB6:ù-Zt…9À뵄9À·„ù¡x$Ä¡øQ<?‰'â©øYü"ž‹âWñR‰×âXü&ÞˆßÅ[ñ^ Ä1Ãéx:Áÿóó8•È‹PŒ–nœŠ31ŸÄXœ‹‰˜Œ *Ó¬0â‚L̸ú¸¨K³Üü¢š¦'âOñçåtQ|×ó]]œ‰¹˜WŸ«‰X¾‰KñYü%þÿˆªÙºúe‹Që_O÷ "¾t´OuµoM?vÔ>Ây›ÕO¼-k`a­]œÎNªY1Ù ƒ|£ê´&[œ©xûî=s‰'7€«“Ëñ˜jUúfâ6Ã*ÃáR1[ža„u\«ln–úl4ùÔPšÚÞhr–û–4Ͻt×£éx®ø®#‚bÅBEšcSr/Àç(DoÊçº)–×õ»®< !‹G—Å£ËâÑe.²Y€‹ô}Oïõíݬ 7q³7]:•"W ­»tAr,ã½äÑO-w´ š,é~åi Q2ࣖDùrߊ(ï¸Øv9•؈¸¨¹¤}¨SÜAݹj»¢€9y£%Q!òì^ˆR=¢‚Ö®«C 2Ç& Kœvµ$vaFa­¿š°䨾­h‚ÙÔpû ©¿xAJ™îë £!ÊhÉu4÷K”ÑŽ;HÐ’(ào®¹z{3æ+dûä᣿¾ç>>^‡¶¥Ï´5=GMš.¶õ»b[½Ž0¯D¶¨Ûþmzv›ŸÜkï›G”GÞ÷¸ë#Õÿ{»:ˆŽxø¹Alz}§Ý×añ³%0~ÕÆ1˜G£áh6¼<¯Ë‹Ñø¤B‘àëü€ó‡ÁL|˜ †ŸªE¿¥\ƒÛ.ª>™ŽÇ¨ß€ëêÏËÁ8Cì¾>Utœ}ö÷Å`ñHü/îq5Ÿ¯P÷äòüC5›N'_àĦˆ_Œ/ç-4^ðŽ«Ÿ¾ˆùx0?#LÞì„Ì/'°:óátVeþe3L×±éa ¦?ðõKÚÒЪu`ºéÂôueïâôÁ—ÖŠ¯£vzRP;6_Cí ¦gg^ýøîÝ›_0ôÓãc%w h×79JuMŽJ{09+ýúÓëš²«”~ÕÛn*ÜS^8‰íÊ æµ”÷Š:XjíP›Íèd]. à$Ë Hì_[$Ó˜Þ’$3™LRG2mW2»âq#ÙŒ_ßPš°!Ðöüç'??ÃØ¯ß­»ÀÔs²ÞQÊtRîck¢-_ EäÛ岑=¬––±Þô¯- U_FÕ¬ì­ Þ[+7mì]w]Öy³ãÎØáê¸ÚF¬¥š›‚µbXa­€ qT,bC¬{à]Yì ° Ä~‹çä VÂâ9¹IfG¨E‘i IÔˆ)'R”\Þ¹œrªïÚ ŒM™’Àš‹”`«Ë^Rz­7€Ã¦Ë]ĵ¶Ñw‘ƸÉ6ºÇµm|jƒ×óîiåÑñÏ/ž¼‡ä=êvtjè«g¢RÜi7õ4u@ÕF¹$µ,k¾]ªèHŒÛ©–anü/mW#ñÞRð ï›óR!×Ù²W…Üaóc’ãÖõ–âXZ£s˜ËéÄ¥¡Ý"ÔØì{C¶…&ììÉ(h•xÌ™hÏ6HÚ(;Ã[h˜á´ÓZe §Ôû·!ÊÁNRš~3£HÎ×¹…aÓ2rõýФ¢á‘²Á…(-§hþ>ˆêÇB¼6¼çu6ŠÊP]rJu¨HÁˆÎÝR¤”ã‘4$™ÀLØ=’dmä‘rôÐ6ël´.Ÿèˆá¾аn‘˜R(rÒò@§¾E:BlqíÒ¤t¶‹Ark˜¤¸D¥tðã~HÂB¿")‘ž]¤+±A>0¢šw×DT;!'J^EÝÁ±cîEÝ =ÞÁYs“C0y·fz%F\QäP~'¸·½[Ø&Í×ÇÀÏ_¿~MiÕ]!°“}lº˜*ívdð%0Ù¾~-M¾~¾£½íîÌŔۯcÞÐË.¯±a׳Ûb·F(QÂaôr·6FN‡û,ŠYA^›,§8ÈžQÓfš°]¦#yF<†,¯°éºvwìë6@^¯¹ =Ç¢|à>ró\e<°y¼„ÿ¥³n÷B¥Ö6æK¢,@6e×"ªkÖ<œ¢²fŸþn×ÔOËzÞAòÝú;µû{›u+Ú÷ôýË'?}÷¤®£á`-â×Ä¿Vè ?;ö”ÎnîfPuX?ç$™b0¦9¸åhÒ¨C?Gùè-¿íñåžyÊù+:‚E£5cR»XF¢õÛ¦îÖxu3ª”k£ÓUçõ‘ùJ=¯SH³rR.çRÓ–û‘}·¬Z yÚN¤ºN0=Í©%J+QR©>eEqëUº‡â×t”ê|4¹œ‹¯§r(…“Ó4ËÍüb0¬:iš’™éæeÔ2ú½Òyë—dmi&×ËËôeò*/G m²1jç#ŒÏ{õÓÛßVcE3ØAEúÜ-äÇÓ¾#à[ÎSRÓæ¯¬§H•Џ§ï­LÀ î6¸»;Q©S QZ•`êˆêË“ÅÖQ‡þE£@+ €åôq”vP;éY‚­ºã”Êî$yH6- IÑpo÷MHò1qíú$%pI¥†$ã·ò~I2Þs«W$Qøµ®Eàîq²ÃÑÇ|,9U2<¬ÆXÔwÝÙ²êMú£’­Ó_wÕ+@øÇmg׿ø€?a­¯>îh´;ç ˜YXa­+›ø?ׯUendstream endobj 170 0 obj << /Type /ObjStm /Length 2634 /Filter /FlateDecode /N 83 /First 757 >> stream xœÕ[mo7þ~¿‚[Ž"gøZç‚hs×özмޅØR ­æßß3WR´r Y’…ûàpwÅ—‡ÃáÌ3CÆF«Œ²‘”õ%+f‡Ò)—¥W!”AeëQFe­‘IYòR­œA Þ¡N²Ê†Œ“ü±•’7§¿áÁ+b‹ÎRPä‚´ŠŠ‚E‡))ŠA:ÌŠ²AlðÐ<[ÅÖ 9zgò€‘/@Úì{PÙ+Žƒæ 8[yˆÊ™ I92R'+Ç@GÀí<>¸dÀ”‰[<°ò“#ã”'’:^y‹0OxˆÊ–:Iù(38Ÿu¬QÁ@jH¾ E¬‚'ˆã…ù$‘ ë@M´è•lRQäL6«è€ŽP/zVdU ˜ JLc¡Ó˜! "§0©ìU¢,•ƒJÎaŠ*yŠ)¥È虲J‰Ñœ Y€aþ™ Ĥ²Ã8„zÙcáÊ‘#Ð&™SVXcDvŒµ26H;Ù°•†X-ãdJš#srX ½¬<ÕdH˜‹by´uP?++C¢L–´ÀÖa!ɉ†ò c@û䯀’ …ÇX0ŒY’¥'ô‰I¡¯¢®¨ìþòãjp1™LÛ¹úOQ/£®ŠvIù^ .§“¶™àWÆ Ê·Á›æz<|1ýõ¥vÀbF(!ê¾ÎPU¹E½«f>}˜š¹’1^¡Ÿ¢¬å··³éè]Ó¢‹ÁÛ—¯ÔàŸÍŸ-zøé'<~ùÜHW¼~ O4Rš‹B.JWK¿n<®h÷Apc…—¶ÀKG€—‚g‹-ÛrQVéÚ*]ë¶À·öþ0ء«ҵ©– hbeÕiQU¥!_ËÚÅþt“9|º”š®«ÓòuZ¾NË/¦ÅÆÔrË–Mö`ølž²e_Lg×ͬg ¨òb//›y[6È¢ëßÿý\…Î0¡Î-žaòpw'5¯š‘ÔtÖh=Ømm!OçH[«´F½wÚ‚â—ñäS‡¨ñd\ÔÂ:H!jô,BÌšA6 AõµƒÔÅsyâ%¾ó À¡té€rHW¶¸Èº‡®à骺gÆÝ³Ð°îUªÞ_;宎.êžÅávÏp ÕÎ_ ™2Ý3hÔ¢þæžWÜdrµôµ µŒµÜbøîFÓS ÿ±šá€-Ų-„'sÖð(gÑ7芎‚¡ÃXGá½ÇÇdû˜¼Ø‰Ø³&ë(‘AÅä ýNävNL)k¿¶vÞZí$(y–µsAcÀMLœY3(r¶,æTv>k!¿gäepN’'a…ý =n¾¶›£ƒgbYî9+ÙœNbføÑ·i!HxÖ'_ïí@;a/ì Ê íî6„+a®uqù† Èå¾à¿òØ~gŸÐó¹Úû\í}Nµ¬œ­r³¯í?N)+×ÛÓþçMi>øù_/._þõòÍ›×Y þ>M¯Ç“%(,ݾÏæíåíp±«;š?·ÓYÉ*”Z¿ ×*-KƵÝð‹á~_··"/3ø~)ÅùHfc .‡ŸnÆo»Wô*Óúnð¡i‡ßå*•ÿ&¹Á«»áǹr‹î_t",¿yxbéâýâÇW㻆!ò*Jùôá}³)×íðn<º˜|¼kPqðf<ŸC6l¾¼k›û_¼úÚ¼ÖDòÍb»—ºClGHd‹D˜Éq*\rÓLgëU€åñÐ`oÀ£B”܉–¼ÅûvTÙéþ¶uì5ɦv€;-v,¼ÿ®œî¢¨ò;5ø×Õëe…Q;žNŸ¾»mÛÏ? £¡¾O†“Q£¿ o§S=šÞ¾/êý¸5²ë6v]’7 Ÿ:CùÏi#97L"¥Yú®hï? uû0ú¤¯‡³ö~úÐÞêæúað[p>¸ŽîÚ/ƒOÍDß`{n×Ãvøß»ñ‡ÙpöEß¶÷wûÌŒ%ú|iÉçyñóFòF¬ ýÏË{8…¨2ÖI\-xJɽ«}à'y…’vË1]÷»Ä »ú„'{Gˤ n²¡è´¤=àÓ„9VoØu©£C¦Ž”lq„KÁVJºS²®QjÚøèÚÅ¢µR†—ôï6™MIÿvlq]‡×øËº>ïßX‘'èp?Ê`„‚ ZÕË BDðA›ŽÙå†/iHâ¸)’1u:ÜaêtxL;/Î#Dó«…Ú7)ñÄEÛ’ÖJ’”é…†áiK“¬Îr­NP_ð,Ÿ|“?Š"â3Ž+PÐ$'që @õU‰L‚7ê'Øjlnaˆ…¼Ù`à¤Òb|êÈþ1P6'®0‰ZC—Ž€i“1æõÕëR?šå{v ø´¾XÀ“œ.Œ[e-§nÇÌk›-sÇ´“´FòŸ›kF°‹v*ö±œ>;½-Io¦ƒhþ ‚Õ&Ñ âÖgÄA'êiCk8ùÂŒ\ ’Wc9O¢]gµÖ ¸%*ò3Bà…´í`‚í “§¬ ?— ¶cÂÐÉaOÒ^®1ÀxŰ‚xZ§a ÙôhšEÈ*0– "Øœ;3¦(4„V˜@aÇNiwK€ÅËk¦³$×8X{'×BŒÎ'Ï`?JŽºLâ(Ä Æ<¨íJ^Âr(k›J‚3Á‹È¡„åãÆ"»B"%¡ß!bXJû!ú†ãÝÓÙ~“Ç®¼k¹y³(ëù=×óûÚNnÝ,Êz~Ïõ>×û\“°®öçj®öçj®öçj®öç¶Ý¨ó=$yëžràéûOv=žè,h¾\¨2¡D9QOX©h ÐŽœìè œ¸¿…h>`“ŸTß$SÛ Ùs1,^Uò ÎÃpc“´26ŸB-Ûc‰ ^CŽ:Nig*ä #4¨ÕÑŽ³‚ª¼#0¢|9[ª¼£ÃxL’Â"må*&HuŽ’°€?ÅÀÚØ/AÁS8ÜgµÝdzdaÕ;Pr'<› ¶crÄåFË“œVÚxP»š%¯Á„¤åÊ¢dK’Ü®ÏI®ú¬cbã40œÓÎü £‹Prt¹›2ô‰—Ï‚){Í>-1 a£N‚iK;`tû¨…rX³r¦N:HY1ž“…ÉÎ9­@T·á 63[r |‰ 4S{9Y°±o£k ’O6°«©Ø,a˜¢ÿ1DY¾ Ëæ€ãA_\?ÍmÛ,{¹¹âl*+(ÿe´‹ãÎ 7¹@e·ñþ}ŽLëë‰Ì<ºþùt€`t²£!ÏÃ!ÕÑìJ³£!ÆÓb"Éè‡^N!QìàSùÙU(Š|j9=‚ɱÜCË+L ßìö5œ7Ë;-ƒßÆ“‹É|¼úðr|sÓÈIuS®ð*—L¤ëE_Ëšèî)³«úendstream endobj 254 0 obj << /Type /ObjStm /Length 3316 /Filter /FlateDecode /N 80 /First 745 >> stream xœÝ[koÛ8ý¾¿‚g°(Å÷c1 MÛÚ¢ëNg:;è×QRï$q`»óúõ{.Eʲ$;NÓÄ‹EàH²(òðÜs//)ZYËSÖ1%=ŽžA×¹¨pŒ,ÆÀ”L*G'’IëQÂ)&ƒŠ8ÑL)A· =‡+g™ 5:Ç´Ð'žétUhc¨pdÚÜB]:Jö’%Q¡W̘ˆ&¼fÆITæ 3ÞÑ7–™¨QwÌ*C·<³6• Ì:Bè#³O¨ ˜S5É\*sO( ô]QÁ0¯ –yƒ^«àš¢§<óÞ¢;!0Ž,pAFÔ% š`€«`"šgð.‚Ç}- QÓSŽEè)Ï¢ÒhÌâ•ÍV[‚EGl¡½$}£`0 "™ iap¦@žg¸4Ka5=éqæ=pæŸz(E4¸+aHœ¢b KJâPK˜RjtLK´!­ rhC:KåІ ‚Ê¡ QC*R JK´¡H&¦CuÀ¡‰ÅBXZ¡ XXã m€d<¡Ðú‰ZÚ@CxV¡ Mt@%83„T¡ m#Õ‚6 zm »è¹FFÔ}I#©—m bµFÆ@pb‘ÆÊð·o¾aÕÓzµf¿°ÅOXõîçC2ÜÁdÆ.`„«Oì=«^Õ¬“S¤‚¯§ËújÍÊÕ²þùæü‡ùú¢fßœÕggB/„³øÔÍÑ+a27ôùi¾Ã-áNs9\»€ÏYç»N¹³MyªŸžñú[öí·©c'‹Oø(nºè}ôJrPÛöñù|¹Z'O%_LéÊÙæª!Àí&@Y»ƒsÚ|º÷¹ë¹‹ôU-ìÝ`­‰®Ú7HØÂÓhQªs›1Ó&;×dŸñêm¾¹l1]×DÙ¤T÷ÀD2ëæûôüéÆüé^n¿`KeÕi;ø¨ÞzÀÑP´‹<:Óç(Žs”­6*Ü#2Ö¶+&’³Ø/ÂXªJ:îã¶ÃÀ²±0ÆX±÷¬ÃÝ0±å¶™‰¢ƒ1ŸHuÄ^hqùtO—wÚbb+4d6¢°ˆË#FÔr/ñ 뇘´ª×3Ûè$}bþ>öƒß¸¦è~·ÞäF«ICg¹Î‚{À–ôrHW<ª¾»! ÚCW‘ÖgÓµå&WHTœŽP@uËÆ=¶êl…!„žfŠt¦‰ÎO‡¨¡ÿxÇMè“H÷Q G(°¦Ãþ!*Ù‚šËäçoŒP¥bõ©é¶­‡§¨Ñ÷ýÏe¥uê->ŸÔe;J+;u ÍrõðêÕ]Õ× 5Ýû‰’³ ZÚK}ö&‡Ûœ§vT£î1´4šN½#m¥>ôÚôaGwЮ´#[åf0Æ–,iízŒöBåé0ªoŲz÷èÑR=2‚”q¸‰ºYGéºvÙŒ¤^Ž ¤õ¥Ë;:iHÿZ@f·kvË”Ž7Ýê}÷!Ÿ§L{Ø;Lì­ä3„žýüž¼v4§¼mjŸÃ³Ïc{›¹h±–üáÃÆjéºL’úi¾îtÙ %ë,÷Æ÷ó|ºy~‰0 ^ïæÃíâƒ‰Ù“Ó óüQ´I,&‡Ûò½”Ìý¯¥­÷è?¾ºZ¬W ‰fúÔ[šç7G› }ïIWk³b:dJ_Ö§óé“Åxž¸s˜Óû¨:þ›ýuR¯Ÿ–³zŨÍç¨'­dƒ/foj²Tõúés°M¶x|8ý󺦪ÎëžMi!g))ËiÁ¤gS5nÓ,³Ñ4ûÿ'±ÜXW6£5”æò±!’VOš£ÌG•Y*«BeU¨\ŸÊõ©\ŸÊõé\Ÿ–#êñwWVwPÏHžmUä¢3-׋ͬ÷²ýp4Ö¸A82¢?CôÛiÖv| MÝ9šÎ£w7_M@ÆaþÉ쓱…Àß”oxÑ <#“‚nð2%–9yÌè7ô=™®êdúꇟ¾{ûôÇ¿Ÿ¼œ@ÌÕ³«Ùât~uΪŸæW¯VóÍi :ù8]2Еtƒ.Ï–óëõb™sÛ¡)’Ð|õæÓ‡uÒ ©G5 ÿ4?]$or‰~—4ï“'4 V°ÃþLšäÓÚ¥•Í1­4ã[ºV†<É“Î<âX*•ï ¦­Á'¦cóMƒª©é}V32èÛÉôú»z~þ–Œ—è4¹ÎWÕIõ¬úgõ¢zYM«YuZÕÕYu^}¬æÕ¯ÕEuY]U‹êºZV«j]}ª~¯þú:©'ÕüHaÀ¨ž_LÏáÙª¡êIqêtÓG—ÚßÜ|>¿¨1—¥%úêÕô²îÛõûõôb>{|uu ÄŠùj£&0(–Z×—?ÒÒp×HWïrgqCýëí‹§/~FcoÞ¬¢ FTÔIpÚB‡‰hÄL.šm3™®™&ÚEáÜY‹Ql›vRbLO÷8}Î{4H:†ÿq·¾úôͳwï’MU‡eZ o´ä’ûäÑwQy»&õõvM$L.bû_ì©Ë§~xáwÖéœÛy¯ mýFÛ|mémN‹gó§ƒHé©ÌÀÑå–‚šËVAáê¯áäÓÙ§u]Í——Sröùouö÷ÆÇ׿/àçÕËŶ£Û":3ôs›|œÚÛÖ\Yìøù–&õsjü?·j(Àï_?ûñí[4öä‚äa~n̈ÕgMŽÕü¿ù/¤Ô6èñ»ô¬”#DAì®Ébtp𠔜‘E¦¦­Lc}Ûà¡ Dg‘^½RÙ|"¤÷ –rŸËLTßp|Û+ôe÷ªšT?@z0´Âe®ëå|qš…øÛp°‘ûÉ¢¶ LO„²/ž.U¡7«Ð¶*\,OëeŽÈ ÒIºÍEÉ¿²ƒ¤ô+òè(°IQ›Mê•Ô:pèÐDÉ£¡)·àžl&4 ü}G/æW¿¤ipgL2.£êcBÈ &[¸¦·ÂszÕl„ä: å Ð0§—ד֜ÉÂä,§××=LÆa‚a™ÒHõ8Í7h“‚AisHÁñ€X§ôè=àQE.Ú:pL²\dL%2è…¶F ²²HF²X’‚ ¤×Vx "Jq§¾¬åä€&ZÑ}PV(ÌÎB ª5Ý1AeÛ‘`–´±]ÆxϘ4ZWC‰ƒᢀ2Vpšž¦´–vpP>pÓ”%r`<¥@ í A@§`©)0a¦'A˜²ê(˜(p[Ú“S0AådîÒ0@aÄæ´]¨ÉÂý­m1x¢¤½:Ç…P€Æõ”œÞÜ> (e¬˜$Œ'-Aæ’ö_)Ë} Œ§IæCë¸ ^ÝÐÄ4;ïš@£Ò¦»|aÒ~»|ÑlµË•ï-m9+ç$ërNnVÎ!¥R)à’%Ní¦ü]}ÍP•&(. !Vá¶Žé}„Ð1Ì€Ì/› F´U@yX#Ï­@M¤‘4dñ´Â¥Ê9ŒY}í×Ìçiÿa97¬,þNhG¡ù"†²HÏhÛ×±‡cyC;‘D 5ܦý„Žkw߆ÚJ!4;¹Á´*“tŽ{C[Z!"äFÀ&UÚ«Êiéqx’1e“‘€¨o‡é0=¾†‡¯áò Ù”ÝzqswyqR‚â­^œ|¾e07& Ç+L]Q\•ÆP©¾¦ ó Ð.P4ï¡âeá– öè¥óîç:/öò :iFô¢¿€^ìCêeÇÐdœJú°N"¬Df"-}f1›á¾ÃË.PÑbžìZPVzd«·ÕÓ‹„¥¥ñíR'‡Zº9Þ$K¾aÕÛÉ÷mû³õ|qÕ|õÕÇõúúUu2yüŠO]/ÿAçøby^}l»›4Ù´!r µÒ|]ÃU üYH<§ýQq(Öå£3à«ùr n>_Uç——·.1¢ÓƒŒšëägð~ÈrWîWZ “ a¡È`náè‡A"–Ð?¤ÕB<$¯¸¡_GHÑpúÊ}€Î20—@ ¬iÌù1îKŒÿ†^ŠXÍlW0nÓ[ÉÀPnÿ0 4"¥õn`=#8í‘ɘh 9M¤Š):niœÍ ¬ ÜÒ/€îÔˆÌæÈnJjÉÓφ¬GDC¨@d£÷ äq@ÁÇý)ƒR’6Û¹EËxÃÅc…Q/È Qšö<uTLÚb$7|àJù‚¤µ€œGB¹CV°±ÅM¯Êv»@Ñ:»3-(«$Wf<í47~`=‰<›ò)il Š6iºãQ0)Òýè´`Â((Âí0õR<ÚûP~¡qÛ«]yßÞ¹åîGkî¼Bßó¶üéüì¬ÆÜ& ¿¤ÍiO1ÒPÐE}ÿO~Wqendstream endobj 335 0 obj << /Filter /FlateDecode /Length 7491 >> stream xœÅ=Ë–·uû‰O6ù^öäp:xì£Èö‘œØb²‰³hñ1dLNSJ2õõ¹ ê^T¡f$ÑÎÑBÅpqqßô·s²ƒÿÕÿ?{{e·Wß^Yz{¨ÿ{ööðë§Wÿò§o¦“³1ž¾¼âOì!N§œ=¤”NÁ—ÃÓ·Wÿ}üÍõ9ÅRò”ñüŸM)e*Óñþ£c'wüŸ3 ²ùøAz-¾¾ƒn¯o¼q§d¦ãøÚÁx_Ž/ð9“Ü2ÜšúÞÃ[ŽïñÙ˜\Žg1ýñ,—ý‘ç´ÞOjžç B¾’ïün‚¿¼ª›ŒnBL¸³NA~+ñó’_g{üŠ^'c á­€X®tw óYƒ^óžg ÖªýÝ Pæõ£ÏÃä!w“ï¢ÏYX©NŸCbìUx_‹çŧ›Ø+ÇÏiÆ’c°m«4û;Áëëÿyúû«€ÚJÂ{úÈL‚"×|&ÖÙ'»qÙœ&7ΞŠÍ<ç—×7Vÿ_®]òÊA¿A,¸ÄÖq§ú”ŒñòàÐÛˆ~fàB§¸Ðú“‰©!äO „ut8XØ\ŒGßä蛦Àpò€Vþê?ˆ "†8M¼ßúžPÕ3ŒñH„€õ߈ᯋgÆ·ñ91ƒ×yïy ³>Ã×6úO×!ÙScaž†N ¼ ´ §+ÌŸÉ oõBó ʆ0›tŒbÜ|N¼2âÙÕ3õ!ÈÏ*VÓ¡œJr‰ŽN6„r¸ °äÇÃðÕ“Á¿á3´¹†SU10–01FùõùYeö0y9¦®œÅÊ76ÇHÏ<œlnÇùTLüªÒ•M‘v*9ÃöPDÀÇ>{æ²:ä ÿpù”ðy¥)#|°ùÓT7P÷%'¿\Ãw€ðÈd]—oË×N'8º‚ƒq€Í•ìK~ 8˜y׃ـB²³ãÔ¶Þâ‡r]Ìö‚g ÎÒx¹0w"Ä, öåro¶ð5Ái:BÆùãGÆJ6 øºÑ©ª£¶ô­œú®[ä*€ ±þx–Ë¿fŽ#ô£˜GÎùœ¡I ,Å6åôª<õUN1 S‚)Ê@0U•&S./y1Wýí€.ŒlP-rÇ虜ÄIE_°éx¾ø©0ÿÁy5F£p2&'¼™Ñ@‚ă·Î·D¬ÜÂxif¬*žú°Æså„9Eà§^WÔ[¢b©×bz œ$Ûgb„çŸ+Ðêø¿ Ž™4“ŽFkΣöž`‚G0ê–b±,§™0$«Ëùñ¢–°²Ý9¿v6,Á¢Q²Ó"Á*õÅ«lˆ; K“Šmƒß)™ [IÖ jbí˜Âñ|»Z[Ém?ÁæÒ¼öm=@œàŸ 0@ eçsûît}}¼G”ù„ @ã7 p’©%%Õ¾l4•¥0zB®Xb}!.Æ Ëà´F'w¹c“ñ.ý”YLÃòáxù‡mÉ! W¾„Ãs£˜H{¼ŒËª¦–ãe´¨øvƒ‰hHcjÜ›zÞ£äk¼o+ïßWöêȪ/梘J²¦Ÿˆ1Pü€c–`’¡¡÷MŒÅ‹–¶8Ñ$ÅÄ‹~n´órÜbX–ÎËë{^3€†l•Þ?ç•P ì~ Ôør¡žïĈ»kCœQ<îÞ¢Ô^c|v¸y! Ö«>@uo² ¤1<ô Öö EûÇñ:N‡ƒ‰m²‡9f4Y9ƒdÈμè¾ÐhÖ)÷¢ªJÀ¹& žórÉÅ¡¤éKå^o_-'w/—€|xtÎÚ!o’ò€Y²jÌ_Åiœ…uðN@ OO~{ß0›•Ì ˜µkE…æ°FâÈ`ÏÄóEíE2…Ô—’ùîч÷eEïø©èWâµr$*'”ÜB<8»wÇßáëÅüÆÙ‹I ¬ ù\GäÒÛ–hàz§Åù‚à_-¯;޶ø¥ÃòÑäaïIyœ%“ÑHÜæ ¯€(ò<Ô!ÍÀ3›ˆ ")Þ¸=þù¸¸#âÏ×ü·\¯I¹"tâHRHÊ×¢€¹-Ÿ@} n£ÍÙÐ&ïÙíÚpôï— kÅsœbbl½©Û º`w:~-ƈUϽ®Zî/Rg šÙ*ààOÜâi¢¡Wbw žÀ˜Žç í:³Bp…íYÁƒå¥ÔjP|Àrù† Êd²UÊç;3R›ßI`gµÙ{7ø¥Y9ŠÀ·¥ìYøz ß³ÕÜIýÇrT…ÛžÄû°xGsâ3BmÃÚtà|Äíþá"Ýÿ×b]éþ#ð“GGWE ^ióÀ¶a`=¤Mb§£0qÏrÍ»koa•)©9K¦dã\iÁCZ³FÆî*¸ä‰¿`X€ÒX½Ãª>—5\õÓ:ܶ8½õ°d”{¹N8™KÍ”¯°°AWñ©9àµP1p˜ÁøSå$C¡¢ŒG ó›< Ü5¸ m§°ØžˆÅܾVfÀL˜·<5ÆóhsxÎ9¤F ÔXÇ6ã q¤bö º“Aëä©>¯`pã“Dº:ŒæIªOå”OósHPª¥>.Pɾ8Õç¿ µ¼Ý ëÁØHÊ'«TR2a5!—ù +.žàN3ÆR@Â`X"é̃%‰¾ÏgP“ï_ÔcAç᎟ ùo„‘CfÀéÔBtdèþá céõ1þµóFއߋá$0nó¡Óu’òåô玭:Ú7 ¼b3µÂĸ—”XÙ ÿ17’ÿ¹€2¼÷n0Ùb‹¼Ø’häÉiwœÙ£}Lõè-œ‚>úÓ¢*LᤉÕCîTt–*{[pG,2>é&‹r cÀ-ç\]Å6œ1X-t Æ{0¿{VG(¬&Cýâ"žï®·|~ØØÝ‡§_^=ýça•¨ž-9ïA‚ÍtÛi½YÕuQUe¹÷á>26¨Òç¼XZÄúÏô)y6ùÙ÷BC†ùf•Aa¾ w÷ ^Ù‡_ˆ}~5xFû¿…ýK”žÛ”`k çZB" ¥NÂÑ—ŽïÌöÑbááÜ%»w¨þ¡´åh˳wx>:ñæKÛa`õ‰õ@o“E t>ÍF{i0›Ô‰k’#x&j$¼2ë¹<öm³þ¼ˆ¨ÎŒn8“Q_… *Ù(åÍt!–¶’‹‹YˆêO´#‰P²¤mµ+¥ »pÄOà4húb8?Ïñ$8ƒr%ÛZ`ø ]©|…Þñ\Ô~Ìô˜ëÆ—˜…ø­œ}¹XÒ £æG dNþü­ØÐ­ n§ÏQ±äˆæÛF•7‘´µA‹³sKÐìùe„˜XCF ³šËP–ËøŒ1Æ3OJ0Þ5ªH{’‡L §”c+ݺÇJ±öñ©Î ‡õ;¡Í ™iêvtRB†JHza©ƒõôd°¤­ÀÕ¦çE“T™JŸfŒ_x4YaÂ=•e—d½‚Ë&–´I% ©¯ÎR­*0Žbeþš”UЄBQC0e_£ªûã÷u¾<)ò[Q ++…Pý£~H;]"QšMrn;ÅQ¾¯&{x7[ƒåÚg­†ˆa0¡§-†ùVÌ"9ø,ÃR´Ê§ØäKÉ&’žÔr–Vž±ŽøWê' ™TE«-8ù*°›TThõýuŒ`Í¥I++™ d£òÃ|Ð%Œ¦Eù/Ä* 6ÐÐÀçÉ:eÂÿÚ³V}~K!9”Íß pÞò,.ÇÇ‘v©êˆU¾P-ØÜÕØ(ÅPÖ'dGé IR~ÉdÀ81„X泊¤jçÌY…ä—Ø³ÎuU2  vz‹÷¡¥Í‡º~q}°‘A){±o)?èxóš9J‰†±oü6ÂøÇ†”© h!jÑÚÖ²¾!M2À(r*ŸßѱÁ,:!P;6¶t‰õÀÔ¿HÔçâ=f.7ÐãSò ,Hw ƒf‰]%|o:´³ÚŒ< U*“¶ D­Ê*”äµ[³‘PyÑÆ9?‡Ï1$ó‹66öShƒ^)ƒÊ‚k%X /Á’È-° ÐÄG _ PêÙ•¼²ê„(±è8Þ@óíÕRðFuhõ¯R½í–©HKuæ>è(A[¬[+”@IT2âýÛ]¡0?³M:)Í~Ô¥‹éÒ¦ó´*=atí¥ º¤I+INêÉgBOÊÌž”*ê.U"£Ñ€ûù‘1Ó mšj,ò>µþjĉÁžææ#_¡¹ ™y«`¿¶‰ˆÞå¾dp^ä‡TBò*zJPô …ˆ(!$Ót²ÖKYÛÕÔÙ¶mCd4—HùJï ×Šù¸¸ÇT(`¦>ž!™ —Á¤Y{ø¼ç³ Bá[åV9¹KI(ÞŒy€y뛎ð/UòH\`Ǟʒ\Óêž¡‹ãŠ›Mïo¹Èñh$ÌXïzX@YW­êC³Œ@ôiG”òQžØ¢mA•°mðò/GY+B4³ýù½æ°ÙMmƒ¬' —`¨ c¨_J¬K\ã§¾èáø5F¶/\ÂáLÑ5dyÚG#‰ð­.­ÌQ['M°Õžº o»sUV†cl]ªÐ¥{§RøUó:Íòq.¹¨³,Ü*µj´`Õ*E¢ý‹’šsvŠ/Z€låü7°¥êrá¸;¡4¤Ž™R óYîþX¼ÐI=´Ž–µ6•Ã’í$ƒK01ƒ’(Ts|)§0Tƒv=®©Å*àù—Z¬²QV)ÉZZmOjÙ¬˜Ã{t»¢Ž\H–äªdôÌúføR¹õ—:Û§°çT7‡¹©§©£˜²4ŒA?Û= ûR­gKþ§ÖE3ŠË–ä M£‚ ƒŽpc (`M5FÇz•lü^©wïÖáñ¬5ÄÂô¼FŽÏm¸í…ÝR©3C*òÇÓò\ÍRF”fÄϨsÉ“³…«blöìœ%µa‡tœ;o¹³ÊSËDg¬Ê>ŠüÚ ¤òÊ‹´„…¸jZªñ1ÔT0ÛqÖ²Õ\0v§Ë˜;®EH0=P£”Ù¾YLC¤éJv)¦Eréëþ©…ÂaÝKØÜöÒš"ô?1²Ñ˜ä°ˆ.ÍYæ¯Î½jÆã±1æ›ù[åDXC=ÍmO`7@I`\,Ú/®ç i'$«btSó~>Ë- ÀÞƒa[AÓ­N-‡ó·Á«rçꓞ…¸PhØ%¬-KÇḠ‰èh\sX"< ÒÕȵݤ¦ìƒØ¢œâôÌ,¨ß—Û5þÆ8ä{’™L†mY(òó…ÐÐ_égœcRï†/Ÿ½¶4¬B¸ÓÈñèyŒÃM«ºoƒe"+ ‘6aŠŒœÖ&à)Q_¯Á¦d;÷Ua^à1Mj〮èKŸ/óžŽæJgHM.güÈ0f;ìzlG5n%èä|™UÀ¶4`&L2è* ó ¢MëÚÙ[B+ž\ðöqÝ”åMrs΄B SMXÙËbß ÒnÇ‘ö¸?W¬³QŒ±÷%ôµ¨äF?Ô“™bQ¥¨Íµ®A`€tÐzM*+¼Qâc‰NŒäË—Ñ%5â¨#3`mÖÕ3ÊÛá!ÆêУäû¾žy£×¨»ãgÄ<´]:âÄ!¹–jâsjEòštX’ÍÍœøöiÿmMºûEÞÃQw9ç̤Äîûkû>¼öþé}x„(]耯°4æëÚ˜ðjVu‰€ì:vZuÎ^µiu«a™º¦8Éteb×Gg`soL2nöìÝmPáÖL1·¶¡WÔÈMp~’nììW·‡}ÒK3X¯ùû—>Ðäƒ+¦®%RZLŸ/sôý>”*4–Eó28¡+Aª IÓ5 ™Ç¹û>Þž¥ûDçFuÒ<ɲ_¼ø9‡mÝtž1Ï>µš%Ö VD}ñ†®ç‘ú¯«ïwlðòä6 òŸÙj€³cTfIEçÇXðÌkÝ}š}¥A·*ÆŒ†ú±´ Ä·^+*µ …|­ÇGK|§ ÙTI˜W½È½órŸéÇ ö8ü)3àu¥áÍw2Žƒ‘Xņ+Øe¨¥¢61ä¨ó^gbX‹î'bâÉLÖ.Y¥ã”•r°ÛZíÚɸäFsØR'gHTíUwÙ²LW=6YV,H—­dÙ+±bg26Ð+ìx™Ð4%lèA¼F[ø†”õ&Pà !ÇõRdª?ãÆ±ç8M k§›ê=;Æû~ÙG·p” w‰îÊ2ÑM6öÿt3ë®x’×׋EgLî¥ût±Ð:¨ÃÄ_¡ÒªGfܤŸ9¼î–î-åfD•٨/:ûö)Õ\2gy´šÛ`‘a;*v¦+w ÝTÜ'¸H‰åi?%<¸1uÝC8tŠž ðÇxÌcèâc•ññ¯ÔJ»ówëÙ‘·ç¥wýXïx|QÇŒ‹ÅwâVl„MJÛŠGãR“³G©@$Oе]iiÂÆçØÇg˜lâkhpÑðXx£³~å½®¯Åù»†9Í [?ð`©Œ¸˜ «ŽFïþŠÂ1)–(Wæ,ú( –—ua¢OvÕÝb™èÃWÑÝ\=” x_ãY\»æá8¶²ËÅk©0º-üÚâ(þ¼WC«šz‰,†q§'uÊ1ì¬xÙ,þz_g·¡]¦daååÂwÅÓwÚ²ÛÅ_èú¤=öÒÛwtqÙÞ„`_ü­ÊT´ÛX÷®ã=­Øcu“ ¬=ÞJ’¬=‰Ÿn€$lÉ-íCSvÓ‡0¤¯ä{?¸°Yš&ÿ„‡g»ÿ– “DY·’ íøf7­;óè’Ë>oW‰`Õ™‡ô¹¾~3”2ì0êE’uTGøÈ2ikéú¡RÔ|Å{×J«Æï–-ˆÁê=euªø.ªÿ ªÙ¨Ïo§ŸpTåTÿíëmúß„IQ¬jø ET¤DÿÇe÷r¾Q±îˆnZ%h^‹ÕU¤ÎÚ‘µ0åAs-ü n+A»­|Ò†µ‰u½%8aE6tÜÍi?uÓRVÙ¸fW4ßàZírwò`l_‡²À)yxÀ)òµú¶aôk§3—%SžÒšìºª>zõÇ+@I>üpe_\ÅÖº?d¼vÖÛÃÛ+±&3ÍoÞ\}=ü)Ô˜ñ™ä/¡¦r²ØP5eNd“ú@UÝ¿­:qš¶>´É}ãm‰Û 8”¶ã ?禒¡_ÞË(`£B–mÀ’©p|yãß3ìx•ë™ëSz‹ÄÒT¢‘Z½¥·*Ø/Xƒ¼â3 xž¿¥øô{ù÷Š›)ú ¾»cfLUl <}ºÔ`r¡]Jvõmy]…ñ"ß¾ä}â/¦¼Ýü¬… \ä†tÄ›ÍègU|ÆöS> ÐôC‹Ë—s¯/Æäøù£ï¯ÉNo:´¾•øÐ!ð \óò¶ý6%ãìž!ñÝn—9Vß?_~a•ç¿k§v[i¥YvõïdsÀ9͉Ï<–o…—<”Å5Š¥äL zã©3Ô*pãÓÿeIðÇ«ÿO=¬endstream endobj 336 0 obj << /Filter /FlateDecode /Length 8006 >> stream xœ½]Is%ÇqŽðVøà°ï8>ØF«öE $-Ú1´C䨒ƒ’#9 †"$†Ëè×;3«ª;kk<Ì <ðM£ºÖÜó«ìo.Å"/þ—ÿÿù« qy{ñÍ…¤§—ùŸ¿ºüÍó‹ŸlÂe\¢Sîòùé y4uñéáõÕµXBŒÑ«ÃKümá§÷‡/é¹RªÃÃÕµvz1"ŽëGü©–õáû+k?ïXãïrßÒÙÃñ3öÞè¦ d_³&ßâo/Dpúp‚?¨°8i56ÐX8R£ÓÖ¤ü"‚ö‡7WA-JÊpX®®­Ö0;|xºRn‰!ÈÃØ2éS_°cp ðT è‰?å³ø˜¨—‹!m`zËû磎en2Dƒµö–xÌsF®ÏáÍÓ==~Óæx˜jp°j`²GÞãÖæ=Ò¨²46Çûý%{óóÚò›Úù‰ó9f’Õ>øíTTuü'Þ†Íwê:oÕµ”K´V¦ãœ+²Æí)®§+$L´b®»ÔIƒÒ‘¿ðõ˜ªnÒ«FJ¢ªò›wSIJ"°NX C[fž/,ñ𞟊è’&÷[¶‘ßmó¸½åüu\ 2&@×HHtÆ€·Íh»àÕ?6ò†É…?^mÿàŒ… ©Ã¿ãSJ(×Ù.œö_l3Žf+¾:ˆ rƒ:kƳÁ„Ã_®®%l´Kv¤+ßõ|v:xY÷R[ÕÄÊÛ8$”p±°$§H6&— õð—PKÞf"K:¡‘ƒÔù‚­Šà5Š7NÌ‘g]sWMÇDƒÎA-ó&*è'pÑs§ÐM¨‰‡Ÿ&g£7©œßé>ý†!1ÕÏaH½áãäÑu°‡elqäJ‰a™WTÕ‘ó=|Er9š1 ïã8Ù|®>xçHý[½/hd˜v§ÐL³;ivëî˜á;¬G®}—ÎiØÖlU(ìÀO˜ó× Ý~"DÔŠé!S ÔÍó©’n#†z™;a²ÔM/r9ùÑøw⑼”Ì#1¯6T _™%n¦¼˜ü~†Mô"ð{Ê ©žíéñ=kþ"ubÜqqL'bѯÔüÔ©ž¹ÇUùX³ÖJ/¶±*‰x{šiêÉÒäT ‘ô%®ÔožäÖyšRƒA™ÍkÉ>ÁÉHÓr 9s8}_‡èÀÆ©ùAÂße$E_:y¹Ññ‰Íï&µ0`²§Ç‰n&7%˜9ü&€,"6“47‚Žœ89ÑÞ$†±n‘nå—/¶¹N¶•‹$´QÙª^ò4CmNnzŠuø†Ç}ý"-žlÇJe1>N|’VP«nQ1ÙšN:¢“Hf®—?‰(Ùkœ8=óSÁ©G™¥f7p R€³Ù²ð<©\Ø¢±ú±OL£ðð)PâÔŠŸê{à{P ð.RkñÅ)„h8*ò¬ª0ÔŸO8è&M ‰˜¯áÄ™˜íéŽfKczû4s΀ˆo_¤ èøŒ¤· Á¬ ;KlQÙªµGì!Ñ–´…?ÐŽÐG¨7 iR.Dª\ˤµH˜ªjÆ&ÛÆJÚœtÂ!I&ð©…q{Ž($‚—¹ëàknÏÛgÁ¨÷ ¶¿ò²j¦áa‚B“È@œ™Z©É·^YšÎ±Õ"Ê›qÆ™)ŽšVî*ã¯Ó¥Ób7å—¿ÎëU*í<5Qµ³‚“ô1î8ôØ£)Y4¶Ræ¾Æ 3U 4µØÐ,Š€“ªUw¸48ƒêñCiíJø|[ºnŠ­9”-•/éóvÙ›…_FEbúùÇV°8”ª rµÀ>nWA«¸8ðÈýª~þ¶Ò‰µ;²mt:"CÜë*ZϾ¦õŠž¸B™)Ëשó›ÈEè¨xœBw›Qï±²N^¥n4ˆÂ¶ >wÑg>6 ÿÐ*:γ«`ñkh‚¸mÙ~cOÒ,p¦2wd|êh Ö[³1Ÿ¶âQG8Árz·œ«· ÕàØñД_K€cVѰ-_]Y³(àøÃ#ÊØQk‚"Ì¥ ƒh„yùMÙzòÊä”Tµ~ÂÞ‚JNú(­(!OO—è\"árÊÏ |úëDg¢‡s;óÿn齉Vf6i1sšË–«=[¸¨µ1Ëù©h ®À7Ø8W\Æ(ЏMíЬŸ)¤”…’øØlJ_OÔiçQd‹´Ïwi•QâV+ï竤¥\託÷>ò2MúéÁ6ç‹%Z`³ Öç!(òšPÏêÉžß$S݈ÚÜü‘ñR º{Š€ (%W¼÷«*eƒp½Ž›…óðÁÍšpî4áÄ-ù¾ðoã»I'QÇËj*Ù‘l,ûmZøÕ­Å._¬m¹Öx -^¥>vñeõ‡ˆ®äY²9ß}°ÍïY™Ÿy;¶óÈ%·àjù5œÏeMÞuŒ–y¨}¥³¾\wOéG¶\}8Íâø(ÖÑ»™£päÌØëi3íìÕ‰óɇï80oUâ@§+q³ÅíAE€ýÉÃðœ®gŒ*,‚â&Dòî»0J-n|Ó™µ®#ûIÜhŸÅ)J„Câ¡XlózëìMrʶKŒ=X|j$ßnÂñôÞ©3®·PL:ÎÒè_<[|èƒ=|²õ¼"ÜÕZàq÷<æô–l xQ¦• iÚì‚çÙTK1à:¿nt÷fNÓò[“w$`,­2q†Ey]¸°±uqVFè&r³ýþ×®””S»7QLÌäÓ$aÈÕSÅ’å,Eè<ØÖS=*Ÿã„õª„GË{É Ïü–6¥¶Qgyl4U9y©È÷ì@®™êûq=Úø<Ê9Z{•”ËÓåÁÎê?¦¡_ûÁæSÝ\¦Ï­Ç‰pÏ)´~Ç÷BTÕ¼Oµ]0 œ?=­`”"¶ûï¿¶tëД­'†;áTÚm¹^Á·®Ð¦Ä<&{d':ÊmÜc í%n7de¶a†ªòóŽ é˜™`¸#£QláSl 5ª‚Ë¿_3©ú%£ÂWÛ&Œ½$ðWÛe!"kv]©«É•§ý*=+6Ëb×(îs&{¢£Ž99 I´FlNÙœDò(ê§^É7ŒÉxðg¦g›ÔvïrZ( qÕ,‰yqa|žI s¹Êxdês¡R½ÇÄo>‹g;Ú´±oœÆ ¾H?m;VÍômãpbÎO«‡ßóÑ7òWZ3N:‰:;Ø‘ 3Þ{tJ—âé+Ê1œ}T­?CG%kt·ô3?ÃjœY9°5zá4àb Gª„)}Ü3n½cSeÜZq1oÿÙÕ€‘5†1⚣硢ÑlÔÁÇ*­?¿²£²’.|̇´Ë.ŠXIš<ŒõÀÿ•6À I&j¿/™,RžYc|³8±½Ñj‘k´©°}á´œ‹Ô U>HwŠö8s»ˆtòTN<ÀÄ;Ð-üÖåqöî‰Oâ½A‘`EÎë^Úæ‰(§c²87¬¬¶FH]Ÿ´Ä‘¤›æ«ô|’]RµÙÎ*‹AÎÚ´˜Äžëܪ±Œ Ëñ}âýÿîƒÒ„ WÉKE³€[^!Òvs–£ˆ5Œ°þó#ÖnãÎyrõzhŒ³(ý,ßsÊK^Áƒ,‘˜3\_¸Ä•µË‚Ð| üÝñ6dUJúßp÷ÝvªYDã`7¥Y<ˆÆU2‚²`Êpøó„­¹z­i¾Š]¨@¹ÊŸ±^º´T áÀeÉëôdiÕºâ{ntTB©‰¸#߃ïڂа fq¹É®«ˆg²›® 3@7h1¬á±ë›ñˆ-d‰Ñ®N m{Ñ’Èi¼®Aà…VÁw’l]”á›tÚWá(¢Òu*‚KöDÙc+;Š5Ñ»vlp$ô†œ¸½É¦:·#JwˆfÓ3Jçy)Ždííí‡r n7ËVªµ•©õˆw¾÷r@€Þ4üSª2Û´ÞìIîœog0€tˆO@³$ñ®êÅàóy&yõ¹øh¬…õՎîDãs…Q§× nk'½¦­¢O•+!²Fü}›š#ú«Eé[85‰ÈÛ#Ƶ¡s)ëPd†¦*ßM6ƒ¯òÌàM· JµÍ¾™Øfßø>YD^%|RÑÇ0-ÞI:oš¢=¨ôÊï۠ɪ÷ ÿ8¶žjËjŒÓÌ´›¬²ŠšÜ¥¡ ׌rÒòÚ²üé(ù½#@pæè/~‘–iç0H¶Î4 H”ïc¹boñÁ•Šå±*L:^RE|ƒˆÉ—©¹íom¬ŒÂªà´µÜWé{ìÅo»ÞeÔzlþø> —ŠØ'¢hã™&Ü”]øvœ&)'ÞžFÐzÝ8Hd9 BæóXéÚ5E Ý–˜9=•tæ6ôíz6{Ù¥|%)ߤ¹ÍQå„^̽Íöo.MÌî 6)¼“swˆ#¨A'ßm9U¾¥Ð{$®vœYö¯‘f…OY¤Ì˜cÏH9çlçßq]6¨æÄÔOì«§B=8?h#6ݪ;#;¥9tv¶F w®$ÿÕfÅV[_Ö$Xß$ÌÆöº7€QodŒäƒEõŽªçÓÕiæh¢ýªÑJf7L è&…½NäYjÃS3ØÅçau_”Ó°gú³éH+ž>5Š+QWÌÙxAa‹½Å¦Ö³|÷qE9CÚädX\€›dX1Sר7kY,F§ûuMÌ mr'Õ”7&~6菱Þ¸"×ÐÍ“OK¾U—8³ƒÁ_xÛ4L“[œ':”qs¤Ùbœ:æöZN0šwäCrIâÁn«¸å‚€ì è¤ãÃ4«]jÓtϲڔɕÐ1ÜúXÍ;Ù“`ìXÇoÍlñ¥»´'(›^Ô3©­cƒbøÔ¼6â*"Ë)(­xˆìDIâeJ×c„Œ¯¯¶øÂmcz4ÐTs)1BèTÒfy¡ ÖÀ­Æ~Ÿ]¬Lò QEBæß÷7YÉi—z¶<Ýcš¿P…Ÿ-´QY]åßá=!‰yDï$Ñ—›äîAÔæ£ÏÏ_b4ã“Znoª” 018E(T öˆu'ö»©¬êJ#¾¤™r‹‡´‰à/³’r J‚ðìâù?e¿V¢Ó_yý­¶ŽIa¼«´É`ú÷®ôÊÝ…‚ G-ŠÛLé4ÑàЇ—å¾g´>»£öCêZÍBÃì"©pR»Ž^þ€`ëSºâáÉŸ©‚–¯Ó(¡1‚Öû.¶±«c+h‚ÅÉ(4¡i9„“ãüó‡ñÕ™[À-ä¯ó¥ÙäRѪ”¿YM0?o•KÏ™ÄlºJÒ.YÁÓ¦I!ÕóRa $Evý„˜Ž9h÷ §åôFDë*¼†<Øl˜•›ý )\ëW¤•ÈsïÔ:4eþí*ÀΣþ6§ä4ÔT0šoä D+{s]ç^~ÉV„øT¼)·VŽë:Î5ŸÓ%›3rï Ò ‰ÜƒO)k㧘–oØ*9OòåLÞ%¾’Ò.Æ· N”k|5VTH—QfWmOTykj²W*Òí¬© +£‚ù°ºm5s0~m\B7¦ùÈ¢±3¤¢6 ùÃc-†w¡îsûP_2]zw5äkI32âpK°#ŸP“ÿÁÇ{áÖtOö)¹2Ä.Ö¬`–ë´B¦h®pxȧ&0ñòö?$͇áª8¼¼ð&oBÇtŽ@n¿OrÈ âaÛr±qT0cE|Ȉ „Åh¸5ç{Ò”öoä|˯›”jf 0)§Û&1 |Ûd”8'"–ÌC/›ýÍéYïHd4k“š]4” g­)b·9'»†YrN*'ìª\öi1§YŒWRw¼Ùö…_oñy¼Ë¼žÓ¼ÑOÛvÊ÷öšgGn^e× oš¦ÐÊãY¯ß§ÒL®*F âçI›PÁ“ÐÖöHWæ¶ýS“‡ëD,únøõíØ¾Ÿ„fO¹:ŒÏ÷5JnƒðS:>.ΰBr Úqgo×èõßw†–»àuÀÆfqÞE¼ó +½#ˆa[eÎZnd7*úËêµéÏôJäh Ýmiñ7–Ïm›’.1ðDcÊA(fÉÁ®x92û=!(èê¦(ª‹¯Òáh<„í°ŸzEÎ:•,ZÏ÷Ió&žª$ÙŒ2…7vAœ‚TV¡ëR¬% P§X–—*·WŠ‘5ªr¾©ÍŸæô5Ú.×ýÍ6Ðà@<•äÕó?#ö/."´è †×2 #¹ø6‡ä‚¯D3OÒs5T¡u[ ü+¼mØwÊ¡2ç}„Φ»K8ù8 a>ŸƒmL¹ªj¬?ü¢Åí‘¿|¯$֓ɇðA˘¦>ù-W”Û§»¨ÅóKÜ·0±f WG©–A—A'J½…`ÓXóp,ØøH;‚ä©^ù%ÌÜyº´Aá{ž´É-¬­ˆŒQøebJtÇG `D …z@®Oÿ„iE¼Zd¿J?1#FL\C–~ÙNÍ–Þ£å7Øöƒl?£¢`…‰R¼/w$®G¾.Ö|£©ƒc^ûþ6ÓÕâ赿Íè8è.ûW6_KÙ@ð‡#þôÈñxYçû'aIòEàE JS†Õ^ˆ£ƒ œ¸_Æm—k&LfŽÕ^zx¯!X‚ ”žfÂ…Çü”ÁFKq÷–Ö*×p©°¥Ó,áÌ m“´ë:ï¸ñ‘ßË=Mtä &è”8-Üõ3Ëùá^ö>%Î̽;îÖ“úy½æJÀaW~° !ƒãbÓçÂñìÃà1ˆNrý? %r×sT½m­£¤U[_‚«èQ 5žkwsÞ$4wðÚ2OØF×:º@Ô‰LµfKÁ“¾l ´ üRoìW—{š4vA6߯Rj¢Ð:»Ød…¶¾Âô랷u€m–\“]þÉdzc¹Ê„Eg9F¾§ÜÓœåÖ^ê°Ò[TŒÂnšÛú£C0°@½žnqã÷¯o©t­X)Q>f¥ØEx2Êê]‡‰•R²&£ÆL1R‘mò«D;ÉL¡#Ât+|ŽøÂºC¡/Þ: ™g³gQ±l¢uºE$°8r©ˆm^¬p„ê*yºpmýÑ&û±r_–ñxÏçmîå²$4 P¸}˜ªÞÕlKv¸#´v7] AÛ1´.?–€­¢ÅL3U‘ŒSKº•Û¦ÞðÖèÔlÎl¬Ýee @3 ~¾IÝÍ‘NO¿Æ£ 0+ýö…³i‰z6ÎàRæ^$šÅɸºÑ{q!_8d4o‚^lvúHß‚&ÑΜïF¢+8t#÷xJ¨Ú$z’ä¢Ï ~³òóö—‰鯫Ï]ÜËqõD=ÜÎ2£ù”KŠƒ Åk­Æ¥[˜##VV/WSà{— $CçèñŒjë²/ãOºOt⟘˜%KkŒŽµX8Ç7¸ÍTm¡”M¢‚rrQRæ¤àÑnxa¬ÐÖðgJdù– î˜VGôô…$…É{fèƒ g9ô)ÕåÞÕ¡?óT¦¦*4-[nl†7-¨¢“¨£3ïõ Y.ž63"›ë#•Òè¡T©Ã¶îô²3.S•pü×+gØàc³jÚ˜æþgço l ‹Óõ§ž}ÂdLS‰Š¼59J~Ó¢&>ªrï\cdbj™’Ë\5ˆ‚ò!O¡Ê9«¼^Ùr¾^öyFIh¬²g˜j®‚ú̲-gSẰƒÆTælT˜[jùz”‚k?.2(u„VZá¯Ìà£<5‰­—.|CòÔ<Åß$ØG^þ‡¶áÁWåËîÓs˦ÝA\0«ë}Þ¥ö ñM=¾•Å5êZÌÖ'pàñ¡¼Ú=©“dé[2޲{]0öEjîdú‚®Â‰3+`[É/aï»41Õ¦ÁábSŒ÷½&¼PÄÐAPµÍŒL1’'–ÏúŽFMØGâ`‰Hv¸g\ëz•[Ëú[=ŨÖ5Uðà)gÏžmòªjÐ$O<yˆ³ }’Ê®”-ÿþÂG“ßKAe(„"B3xFy»«ƒ¬êaÿîÁVÞþ>os Š]À¤kâšq]¥ÐÊf„'çå×q~±3 ûE88¬ÝX%@ÿwívœÌöήq_ÕNÛ…xøÃÕµø@³;”ÃÄ‚aV öÏ»Cà/½}Ö„dLòwIÿ]+£‰Qý´øøÓäô‚Ü`lû‹Æ ÐÖŠŸ{|Xm!l;ð‡mF£(p1×¶]mH¹íÞígJ©¯VŸq_èý{.ûF%gßì-êGÖºOL,äQÿ±Pe¨€;l¢zíYnȱÕ+UÝ—ÖÿŽÖŸ?;)ýtv¥Ç-"÷Õ1ÝÔ)MŸ—xê]P”dª¥›÷mÃtM¬M›iuß–¿9g]IÌ U”tÍç ´%TʳO˜†^ò…­RtjPc¨LUØ&Ö5M5gúø\)Æ#Á6ñR-D ÜÃÿ!"¨9lOûšµü´ÈÍ-fÛЯ!T7ýû(f4€!å[OôÄ,½Ýë6Î÷+øµý¬D¾ ×§±²¦înJ]§ÎÑåÊÒ|KAÏ»‰sý,7F®³\BÃ.ñ8ð\R†®A¹tŠ—ÑtQ¥»&Ùº«@|„+œ&81ãÕÿ†ÈdÓG[ }·í ÓsvEzRŠh §ìWÐHÁ­wLrb<§B2gŒþÓ}šî¶þŒ¸sµ‰š«ºDÍÎ ©dÆ9 ‚¿;¹°2-£ð‹íHmÖcÞ³q!ÑŠbüàùÅï.0E~ùÃ…¸ü-è>̾ªKé[œ¾|ªÝžÜ]|2ýfomÔäoöB—‹Ç.± °Žé£½RcÂ;Т•^šßÀ ö߸,s£ejÅô™Å ùlÙ”Dº®ÞéØÝ´ìÎ X?!=¶k ãW½À8Z»½J¨8¬rI9*Ü;ÈÍ èy90‘ã‚ߢÄe™71¹,®¿4 "Vw¥íÔ}¦+aï1®y³]×#O•¦µ;>] |rT‚.üÉ÷˜¯ÚÉõ,qýÔÆæ‹éܵ'À€Vh¨]ó£~ Ò׎ãöïÇ,«Á%v§°?QPOnu•ï¶>”¦ð—ho–Aj-À‚¥›ÞÙtl( ¿‘ÐTA¢x~û2Üø»žÒ±»€†¡Æ!®Åò›4¬€ybÀçQøZ!ÙÍüí›áüý_M“òÊ’¾óvu‡¶w~«±ú¶ÕÄ6mÝDßè;+Ræzuž’jM²o`Zžn'mf™¿tAç“@bð}–IA¿ô] Þ5Oxן‹ÜžWpXn#TNp& Üø½Ö”ÄÖ t2ý¡ Ãu׆ÆÕÔÜm¢là;n,„WF”å¹ß©]ƒé<ý,Œ8+ˆöåx]}q4˜–ôíÃY%”#ûÃh‹è«­ïàjìÉ.¦'ŸâjLù¹s5ZÕýW’I‘ó•ײ˜ô.«NK³Qe[ð1†Þñ“Üñ]mÂPUÀ„ôXˆwªQjâ¶aD°Â&Îîpß—Nì9®°V¬vU˜ ÃzUEò2‚Lwßl)߉R ÿ“ ]ÇÇ”6-<úÁ¯¬q—*ÝÎæÙ*#sŽÃoø(gøyìq h½…½„ ÞY5;wHìó!híqè)…‹ŸódíÔšüÝÅÿ«}Êìendstream endobj 337 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 8694 >> stream xœµzXSgûþ‰‘œã¢•ôTpœàÞ{·ZgÝ·²Aöˆl#dn.ö«]¼=ü]ÂÈ?{‡@7°HûQs¼Ã‚ޛ81<<|‚‹è„À¯F³÷ ó¶ßàê²ÇÃÝ~I`@˜ý{Ë'X^úIÃ5œÚ@ ¨‘ÔFjµ‰Mm¦ÆP[¨±ÔVj!5ŽÚF-¢ÆSÛ©ÅÔjõ!5‘ZBM¢–R“©eÔj95•ZAM£VRÓ©UÔ j55“ZCÍ¢úR6”˜z‡b©w©pªeKÙQý©Ô@*–Š£8Š¡â)ª%£zRs¨Þ”?Õ‡  > ¬©@ê-Z@+è)èEm'E¦ºSAÁ“nÞÝ>n~Ñݧûi«±Vu¢¡¢lZD73[˜â {”÷ø¥§ÏK½T½Gõ¾ÛGo=Ú:ñ-»·dox;¬o·¾Ú¾¼ÍJ›fñwú¼ó1»êÝžïº÷ÔOÖï¥mŽØnÝùþéÖ¸8ðÝ0hÜ n&wB²NÒl?ÏþÚàžƒ}gnŒ†,CÙ¡ACåCï ýuزa/†¿7üæð_Fôa?⃷GŽ|>jËèAcðJë¶D0¡m&~¹^PÓ¶@Èi[Ŧä©2¢!”ò”XìÖþmäV·˜MjÆ.ÓÔÖƒ ZU5êc½‡šèm!Æ &[£•Aý¬ˆNàñ…V ;âüèzÝØሺYe>%’^ ±²ZølYá[VX*²æ[ÁÄ[5K 6¨Ûeäq¹Ÿø0ꆞ²¸”v’Á2r¡:Íqm%T{ïë¸÷2– UôÏWO_¼³i‡cÿå±[ Ô{ä@Ã}0ùÕÇíÖ¯‡…à¸+p5#¾ù’ƃeV~´øð•ÌäíKVÚ&AyÛ!?‚—EƒÐNC;š;í®vT…€ì7¹ºݨ6H! Tªdá¹¶˜E›Tº”LHµƒ\CÑA c¢½• Áü`µÆSk>«„>%É%h6ÙâY8%6Üç°#)Ó4kk¡NA½Ò²úš¨J(†,]jF‚øŸofßÀÂ\Éj|›Å}EŽªŒZ ª@ëè›',[¼eö0‰5¯”ùIAÕ]´÷®÷E»YôÎÈW¸î; p_Ìþ<õ@}~ú‰9ŒÙ]à\t$䜇z8'ËUi)m„CpPZ±«b¬'©ô‚MRg©“sÀ6`ÞT Ý»/äáï± ç?N­’w?•›:ˆä}½6Xc©VSaD¸L§Kðl\Š—£’³A iv¥ÆÔ܉ö"…rØ©q±ª˜þ 2e…ÁÈg‹·Ðñ‹êuWÉÚÀÕÎ2…Ñ(³HÐ&ÚºÍSjjo²9ý‚÷­ 1ö?9¾c±š'£5hæH‚˜lȈ“t\"þ µ¢-ûÎ#ÁƒIx„žÓÆ‚2˜BЕJP]2…⓸m <ë· vâ'x-žŽ§â]$Ý­x³ 3ðvÍ!F›W—QáÝ~â§h>ß›½Dãafœ_Í;KÄwÚ½»fŸ¦U[•p\U÷ÇæšCgÑiÑÍ›¾õúùÄ"ñ˜qw‰øé\pq—2:\À¢PºúøÑêãÀ\;?[á>Î\àæ¡o ’Ä–‚ôŒ5`ÿ›IpÙȯ6 ùCmÙŒòUS(‡p®}£H†¥VFQ²Ú{Í"7›,j·kÇÉ2PC’]T”r¼‹¨é­BDIX1¯ÏY·¥v–»Ï‹Ÿ ëÚæ±~åtV»ÚM„æ BëÐ<MÆó$ø_íY Xùý-ðLô \–_|p]Öt˜ X±Ök{xІõ#¼øòΜŽ1†ëmöÝE…„¡îðoŸgñÒŸCþŒ”ö˜d%~Zçzáý˜2¿‹û>R+ÐtÐPÍè‹Óà( ‹õ rfᚈF½®Ü½þeó´Í„“æ.>gBµûhà´ÝˆÆuF!x‹íŠóº¨å…à a ñ|’"»Î†Ô?¤ŸæøÎQqˆhºÌo 7%‹:ZŸ¡ñP¹±K£(I­ÚÏuõ»5? ‹–_]zu¿Ÿø&z}Í¢n4–˜qdaKñáßIú€¦E[ЖžØC/PåÔKÐB¾ÈЂZË)ÕáñÌ–Æй55~Õ»8ñÍ…àµ,vid{f‹ és£ %¹Ý6‡m?ò††Jß4aîiÁ&xÔÙ„»éqW'Aêûô÷Ù]ÏLØ=–›ŽÒ;î-BcÀèPˆ9°5ÏÉ üŒ›ZcÈG(ÒöQ?ñäŽ6³ÅWѰˆÙêí6w¦×™¯#8ež:5˜8HŽ”`[: ’sÒ4š’N«mqÉ!—ÃêR‚]úÈß>?á¿7>W²»Æ=Ó%“̇œeygû×U4ÝGÝÒ&ûi9Mb¡&ÒJ$?Y:EñqœR‘¨PzT¹B ©žµÇ2ïÂÀJ©Dü &¸Jþy,ÉQjJÔ£?ÜÂ/3x×õÇð/°yÒF·`² ³@‚ÒRÓ>%tæ©Þ¨ò'ÓjxY¦U ýP'󔬤Å-¸;'®ÀÛëÐqÂ8ÛjÏýòh2ùo Ç·fÓÉQ`‡×àx vÁ.hžŠ6HÄ1×àËŠ}“^±•ÑH= 4¡YzÁÏ—Q‚Òü0 5á¡hUˆH=Ûß~´‚‘~«iüÖ“øú«UW/rgB6ÒDÔz.ƒGE\‡0ùÄ„ö™,8h} äÛøçl½6ç&Wny+àóÁÛUgÓã8ÜŠâè?9êwj ø¹N¦æ~¸uü„ÒX[€ZqÁŸE¨Vf6“y£Z"ø·ÙÉ4–ce¢?LÊ8JÚ!|Ó.x =mÇÒ…3-­úºó›[ œ"Ácè¹0 ãħáPUõ'Œ5úµ“ß>"©9~]Îm®ÂÙ—;Ì錬‚¾­ëŒL)›Ã›×|›¬øÜî\s0=GiYs¹¬”ïo”ð±B~ºÏf•¼dV(aªu„ÂNmTǤlPç„A$ª Ùpœi‹»£y®Ö®âpù«Tanˬ¤/@ª<Ï ÃÈ6=V›’yš•šMP‘-5œEÝù=&AêŽFߢSh;q›ÃfùR8ëŵæî­¨ÔJ}b\]ûàùµ+7r$šLr»} ¿Žnʇ äŽÎ*µ#™ÎSÁÃx=ýÄyzôì«Ø¬bÿ=`‹™òêÌaIIXR'mDGXõjC$y/UJÜ$œe; Õ(rÕfPyôäê­ªPð‡€®¨ P•—&Ïñψ˜ŽSlÇ!}J.YXj×yPÀfWÇÇ W™ë‹lð/¶i±º$s´éiù?¡zÛç¸!UfþÈ.t¹QÊ Q&äk¨x8ùñƒMóÍÑÌË׈pð?îÁ¦ºUº´Sz>ÿŽätˆ½>D.Û­8œÆ!GšH¾^ î¶îföÇåþú¨Šªb}EnJíN¤²îhÑAGϸO‘xÒâƒ'¶*?T®òŸï²œ™÷ž†\àÈM.Á‘Æ’&ÍbGÒãù8mw­;vªù.šNZ±øwòÃø™,Z†dfž8~Ǭ†ƒ•¾^°|ºÔpz$BtBBR2¾„[mÑÕÿBC.¢CÁr\šž˜ñvþîD:˜e¯Ù_ÔÁþNÙKo‚=©þ‡Ì#Û<ºL‚\^#D(žÝ»ÿÑò S}SA‹Á*u0)òfm\,JU™qÄõD'Ä¥ÈWϵûS¢Y¦ÛAAA–Þr†z¹*ö€£&¨£ÈW@£nYû#fmÇã^;Ö¯Sí1Ë÷ ­ISÕЬ*ïÐR]r€¡ºöüç_Œ´Í”åEð%&Æ™uû^Íi­È†ê½]Djê ¹L5 Ïæ.Cd\]éÔqBib¤5šÌüÊF|4¬ yÛõ]bmôXD˜y.ìrŒŽ$b­E®4”€&§ðÀéS™åpêƒË=*\uÀ"±æ{£1íÛ”¸ØÝò8éÆÀ%+!…˜µˆbSj¦ ™ŠˆÂ°piœ¿S³×± MgΔsÖü:2Áj7ÍJoËãX£Ù.õ@,šùUó‰¬ j¯:ÎU)÷‡8&¸8¼ªB_´÷³y>À}'b ¿Í‰_âw~EàÚ¿õÎÎN%aP¹Š ˜°"Þ1¿yêA4]býzTg_ÙWñåw„HÞ6šM4(Éâ<÷ÖQÒAà8Ìbj[X G»ôå0h©”ðõÉí–—Fðj—ºÌÔ¤nëc«i¿*êêªÝg¯ò7høcWñå]‰çŠ‚ÔÉ‘8°ýœ-Éùw ‚f#ú„T»??œÍÊ1“—‘¾Fût¢ØÓŸ—4 nDôˆD¸¦½O¢ Tb›AÊîÁ¾–¨ÑDÜs3fHP=E]ÓÎêNÿ}Š]N£‘¿ÏÌ£üÏ„8ö£êwUl2_cÈhÌbñ£Ùáð>c ƒJ»cš€E%‹òßá$ÝepEèþà&å"´çª—šöמ$DÑcrʉ"ëŒüƒkXÅÏ®¢˜¶)lb®ZAnô_µ¨ôêu™R£†¦½äO*QD§¥hÔ©r~\ûSÛ¬x:˜|È(‘ðGéJhyAê­Þ®Ú »a»Ö»Ô/À1HÒ>îÀœ)Zh°!>9ñRˆtÝQ›ŽFß3îæE6à.Á‰{XT#2è ôÇMOI÷1‡á¾‚¹þ/ÇÃ`} K²“´)‰*y’Šó6`¸í kðÿ.A=‰ºXZÑ6³B@Ô£ó a›zÌ¢]x,1§+ðR<OîDÖMÀ“дES‘‡¿Æß²cqÿoQ>ÊE#>»û ºgáb¡©0x»#FžjÞ¹À®ehÃÙ{eÊlªZÃo …7êZÏöÿ"C!h>; ž•Ujj Œ’ÌܲªF`¾‘aAjߘIR|X ãÎB*”é6)K€yrùò͆è†2Iý¾ú´bóÖèÔ2EŠ â™¨¼øÂì’Œ²¢ø÷pg¹« çR뢓3aÉ’œ Æ=’¸˜ÈÝà͈Û(ðËó0…¯òÝnÌ¢7"kÔóçã·êbo­â6W­‡5„/wA²Æ7u H› ×ÙKŽŸ8²¿áNò±è>êÓ7-Þ6šè§b¼ùÅe#zËÀWmN?#flzñ,ÄÔOü¿ºQÈnÒãî×E1(Ùßyü œŽ3çÆãÁƒçrÒÓ¾ÇÖ³Õo—kP^„F%A³YŸ_YùÖ¤#Õ¥µN\ÖçHš"XÔóÄ×è@c™€KÜÄ S]zV5'þM6® â“ÏKÝñlÈ‘%ª“Ôù&ÿ]à!ûB/2;‡>tˆp!êöûÿoÓp&8¯ò`Ћ?”u[˜Ô@€y€ZM6ˆ»ÔÈjÞ‹ï+ú‰_£0ôŠE¹4”A®A—­­=0Qw¾$ô=>ùo±fì7 rœÞ´Å’öx:/SëˆDÓýËão@jJ®'rlo³Õ%êò€Ô'#5ÍæÓlÑ{íéÚu|jvñRC›­Apà.ʼ+äw¡k,|™ò¥÷õßÏ.tµ°0Øu‚ï’”¹ð>ÌOÜ4ïà{×ÂÂIø¢ôÐw5—ÓnÁm{ák¬;¬3F~/;ác¸—2O– ^7²ŒDq^Š,•íó`ýRÙʘqüÌûÓJÓ!½]¢–¶wØrІq«WN_ .SO.ÕZÁ´0ã's¾f5ô YðM§ò¥'+sÌšn¢—/2Ù{‰œuèPñMöÕ 4 qçNÂ3;$šü%¶ÆV &Nq8º¢ºÜ¦ŠÈ ï$¨•\Ù§'Îóèà{³ßß6gãj ÞˆwËÌÆ ÔNü’§EUh!q)þÉÎ 1ÉIË XrWu•8¦ae¨çOŸmšQ÷ÏÑûç½è'Æ:ˆ^²_Už½B2ôÕÔÓcGÌ[;o·!¢Ú¤7Ts„ÎtôR^=0M§¤“ãUëƒ%þ;|”!ê$u¨*’ÔÉjH`Äí2Yq‡D_5,. pÞ5¿ècoI]ŠÉû™ú }@¨o¬ó¤'Ë‘Y?zôœ#·†¯VÝZÚ‘`ÔßÄ» ·Ü$F³ÙiæÙ¼€ÛcæáV2›áD'wJÖož™Fò;HîÓ£•æíS|®Ý‡àM,VŠ’ïae uû­2 ~wmŽý¸Ã„"}ÉoB;YÔkö3LíØííÅ¡-´^^›\C†Õ[Wå¥+Ó’Õ*P¦H·‡/7Ø™áXš¤S‘L3 -ÁGéHHÊÏÐêÒS¹œÂ†“w Œ;3# =µî°¿&¸6Ô%ÔÝ3x'éÊ '¢>Ö‘ž ¸¶ÈX^eÜë¼súå1¨»Dü bžÄf€ð’Ò=¨uð+›ZÔ[j¸n~é'þ çï³çBïòŽ . ª×çg§“‘®Õh4Àh!V᛼hÕjI|.Æ œüÊc …{³œsþh~ &2X"~õˆYü ¿ûržxÊ÷Xȉۗ‚çÖð ê…o°«Àÿœ¼ ²¾„bæVͳ‡…G`µ‰+ö„­DGí€mQ^þkw’ùoq)ÿéMÁiãw…|ÿŠÍ2o–2X!ÂÖ ÊS”0!¢ÝØ•Lv]aÖ…¼ÜŒô[Gêé&’EÏÁéîÉxÀn,ᤀNGŽÀálî8 §t§2O¦–åß?ôIG§ ¿AE$æca‡PÅ¡°cæÇ{3ùìνŠ}PÄœÿ´ù굋[mpß±Ö›3İŸ7m…KÌã)g&|oÁ„H­S³#—ßb$êéÅú€é£ú~¾ì%²úîî/õ²ÃÍ\DitöÊFf¤ ‹Þš[ÑâkìÌß2{ú¢Oîž­¹øuKGúëß+㇙ӿâ1rh†­]e}i7tçú¨MÚ}AÜ¡Ìê"Sh4Ú?Ñ}ÙIׯK|Ê@äK² ˜þlÊÆí‘N»8W§0˜Ïà¾ßŽFÝ9ñ/ç¡©¹ì,ƒßþ€†àw!²Ž™/[J*íÏ«#§֧È)×7uìdfùŒ™Ê™ø+øJpÿ•mä­ÙP]øÖ?Qbh/ÞTŠæ›Ï´¹ÿ*‚ ù§~¶øID»%‚fîÊÉI¿òó/S­º¾© ÃiZÐh¸ MŽ63Ã\3>­ W´FÈ÷jëÍfedæ@.SY#“ÉîÇEšÉ€o(½Ó* T¨"Y€ôÇ‹#Iï!zxq±ñiÈ®Mhœn/l[„¾d J@¯€XIû÷¢Ø/ ÿÞÌæw|#iÿQLÇ7ùfó¦ïw´.\ÿÃÖ5/øó'N‡mî¢ûýÄ_óÓÚ³ÑVjGç°‚ì(.¨02#Š ÕéV‡Â[Íþ–CeƘ„"®2º8¹ˆ˜ñÃñÓ– q¯£Öp½ 1žRÔûºùÅL*FÞ­ˆªô MË9µJ­V£‚üÔšŒÏ/|"É5w˜ŽÉHIM‘Íš‡{®-w:XW\Ya&•ÿä4kÞëJ_;è?ž¢{8™%ŠVU„û"_´Ðx@ ÷£…è-Ô7¯ó" Z{VƒŠ…¨­eÏââõ…J”i¯ò ”ÉÚRÚÂ¥Dzâ™Ðì^?å3N¿¸^¶U_!l›f°DÙ+ÂCS¢•Iëy_”wÚôÑáK-W›.õ†$xêŽçâùØ‹ð 9MKZs‡–×ìÈs Ç®efJARYJEŠ‘ŒÊZ<à M³ìu^ñû6ì¢dÏÙÀÛpƒAÓP·Ï}ÿjž4?h^Àl.ÆÅw2,aöè ŒCSE–'VÏ+yaåÄGö½‹ÞECˆ$ŸÇÿÄþþ|b"¸’¿ ÅkéïÌÏÛ×Ѹ_] šŸªQçg`à³DÆÁlÇÇq»Í[4¡‰ü=ùû#œ¦?Ï:¿÷@këUÍm8…š'3Ò5Èbr’ šó¹Ãöd5®ÆÝvÆTpÇþ·ŽŸ@Ü­òÇÍ¿(þë¿à+;ö3_ˆÐP(]Wˆ˜Í?)Y3ÿ‹IÐ&®ò ºÍfeCªyÿQ‘¦LÅ.è¨-Š¡¹pÌ«Oéú1H«(Ta¤Å¿áD\ž`VòJ;Y6p|+ó2"æ#ñûKƒðd`pµÈšÿ²£9Ò Òž—ž|#ä7šÿÓ&ÒÑŒèÁG-‡ÿé†p½÷>‡uΊ. Á!-˜q˜Ù‚©¿ÚÈóO÷ ©½b‘ýô«æOVE€†SÓQªpó…\R™¨ŒƒDˆg0Mwšj ©=®ílòN^;.ãþ 6wåvEìP؉ïQ•ئnp‹‰äÂê\ôÛ€{–ûËiðú ²¼“Bž}sàãòº”ÀJÎ]• á Í.Na¬¥~Q ÈI7ˆ°k.mêy¹׳ûL}޽/ë{÷¡¨ÿœ:endstream endobj 338 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 1080 >> stream xœ%RPÓuþ|lßE³­lûrÐtÆ/9wpÐ C8 !º)ÄmŒÉÆ¯ÑØøp¬ŽRø(†Á¿CQ¤ã™(„BhBE7‚¬óJOŒ?z¿ë#W+þzß÷ç}Ÿçy ¹»!Š¢x2¹Bò_ëϾJ±»ÝØ×8Øì<ïTy`/ör¿ò÷ð‹Pþhž‡4oÄ¡(ƒ¹Vf(®0êsu&&@È„FDH÷0a!!LL¡Ö¨×¨‹¹Ú¤ÓªM®¡€Q4z­©‚ ˆÒ™LÅ‘ÁÁeeeAêÂ’ ƒ1÷­À=L™Þ¤ckK´ÆRmo(21)êB-³Í-h»È …Åf“ÖÈÈ 9ZcBȳ$ÁdVgË+ R¢tE±H†‚Ð^Žäh‡K"rG&4CÑTe£ž¸Éݬœ?öŸõ±Ü…Í)êò<Èç9ðˆ¤ueö¦Ä“—Óûò»Ë%—ÌÇ',Wk'kðQ:> WE*ÏUˆkú+' ËXtO`{ÿ¥®¶!lÇÇ®Ô5ä 7¾Û¤<û^3¥ï,`kÿøqým1ßlù޾ܣf~c©%ûÆ$WB"åjª3jU˜&Â×Aû tvƒpùv‰ú¦¤eE1Ý’ˆEäI&I "@ùÃw¶¾Y ¶\‹×&ÿ²Q› ±»à;I|Á'žø„Z%6îÁ–÷çnv[»Å•Sï—æ×äcѱê‹ã>kµL°Þ6jÙé›À¦ ¢‰»_¡KŽ«í±‹a5š<àÉ:3~ŸoµŠ3y|Öd™ÿqž¢F@í‚ùÁ°àÏØ¹à¤¬ÊâLqcâØà ë(}5¤‰‘¨xä:â£Æ{iUoäzËï§¾œç.ò¢A(À娍//ªÐÖ[0‘Õ9<ûÎØOýOkOƒÁF­,Bƒ¡Î$Áþ­$Ç5n—Ù®›ÅôÆÚ¼Þ €ÞÌ‘˜~PÒ"híº:0é©Uä%²KªJÉÉ»pÍ(á;ÿÁíN_µá`Ï»èöt!™ÄÛ’I×s?i¯¬&¡l˜P¿îA<\Gœè$·1ˆP= ðL$üžëúÉÓþ•¦î36Qs_ï§'ë…[ÈáL"\ß)Âí0: Ö[Ô£5XZæ°RXl^œÇËôB­?1ùx+i æ Œ ©dÎãWî³–/+øó€žI‹KN%ž.ácæûìÆ uk ¾pDë (/+Ûb]Á„¨ZÔCjÉpÖHÍTõ/¥Í'>7Ú>èÈÃ*:Q™’ÚPþT}‹@øð)¸Ý+¸nî’dÞÈk î<ÔRyº°§¶ÑЇG詯¿y¸z=¯`Ṗ¶íY^ä@ȶϋ\¾¹ƒMmƒ´ŽÖ.É>Ç›ötì{ºKm^ÏM7xy9ÎzíDè_JµÃendstream endobj 339 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 2498 >> stream xœ­ViTSg¾1{ADöŽXíMÔ¶Ø©.Õj[Ç"´ƒAŠ D%ì$e’û&a$aG¶H"¡jÇ–ÅNWÏ)ãÌÔÏH{ÔÚõÌ™ïÒÎÌMÔ¶3í9sæONròž÷ûžç}Ÿçù„Û!8Q–œ’ž‘éE+‰½D(F¯÷Ù†ŠÝ‡"Æk]rW³q-™ó¬.¨,nsÝ£¾êÔ Tès«U¹îvêMµ­üÖ·¢Ø´Ö©5´Ñ)ê-è úî®q¼0 /ÑûÅäuöôZ†zsšXc©¬5žj²#úeq2‰ŸÐá%Ѱ’Úv'}zêüÀˆ…É…ýŸ1C"s#´™‹E±Ô'ÀBaÑïé-A9òxõùA4ù4vyæ/ó«õ¨…]ùQ€³Aô}…¹¶rîÞ9æhQÙå°-øòcä‹Ö®žÆ‚èDmVª¸:Šî¶ì+œU5ÊÓ x©ÔÊ´©iª$H‚ts¶Ê;ÊB» —œ'Z,D—¹5tIiÑPS!¾»«ãyì³e–àÕ÷vðeË{[ÃÉ£P¬­` +˜2MΞ] -ù¶|;cÔU@‹nŸiÐëN3Þß=õ\#§@iα%­å¦,XŠe¼Â'E¬J—£Ë…2 4RÑ9ƒÃh…s0¦·ºìOCæB®±Ì 4Z(®ß@"õÜͪc•…Xj†ÊU¨€ûÚÇ0÷çµRÑ€éï96ø\Š]ËÍÜqàD¾ ŸJê§;±GòF£N·y úP2f Þ¤ Ï÷ÝWã5_ìDÂó¶]ýâxí©vwÁe`s#±hî­NNRÄð´ÊÛ õÃï¶vx£qà”½¯ý<ô€UÛ)ãOMâ¥ùjÚþÀÃï ¹ß 9Ý^cèšIæ=m=¦ý6ñø‚DAH xÇÒׇ bÓ¢VùÊFG'/@“øax:ñóüÜ¥`XtKôÓÝ¿ø?(:Ë¿¾¾2ƒ6î"BˆR¹ïé Éýq‡SRH²ö_`ð«ØŸßƒ‰Ü¿ÇðºqÎÂ¥«Ìùòþ¹ÈÁýcXðæ=d¼'ä2fý\þ_²9j&++±ïPS¿Ü‹6a·PKlgšØm-½•ÓZn+®Ö¿žÓt⩃š—·ïÆ^ÝHRưg *ça¹² Î`hieººÆc?,³ñ:ÿõ{Ÿ#ÏɼQe¿8î­Äºg-ŠšÔÚ¢ÊÔòA>ÆÚÞšF¾f¼BVŘ´U¼Ÿ='jnæÊ}ȱÿa$N2{y¦†ù'Á¨¾ÿ!S³ñLýÓþp&®¹ÌœKbd“wk‡)ü!*^¾—ú?=rþ»w¿3è[ÇGˆnr%;zÁ‰ÌÁ½F³|žZñ¥èú¥ 7-æòÂfFS®Èƒ*±;Û:ÐÓ2âH|5a:_Sß0é8þö—@ß"1ÙV )büþ/ýÏ=Cz5p¡u(ÆRß Â‡jH‡'"æ3žn[Í^ö//Dô{-0xyÄ¿þ»íendstream endobj 340 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 4631 >> stream xœX TçÒíq˜éÒ*1ö€+®à’(1F ‚¢¢(.(.3È*›ì«lËL1(»,²*² " ƒ¢AÔ“˜˜ÄDcL+$Ì÷í©3¬¢ý#ü¬Ö{‡{‡Ey{Y9…GX­Qy[õŸnVÿ‡ =‘ÞaV.!^ÞaÁ ÃØ. qس,Ô1,|yD¤s´*Æsuì.—8/oW_¿õþƒfÏ™;ïõ7æ/°s–9ÃŒgÖ2®ÌDf3‰YÏLfܘ Œ ³‘ÙÄLcÜ™w˜-Œ3“ÙÊ,cf1Œ#ãÄ,gf3+gf.³’yY͸0k˜‘Œó Ã3£˜ÑÌÆ’y•Ë “ aì(¦Œ “Ì<8JÞ´`nÐ3éi‰³É=™,QÖ%÷–ßfãØ8k®vðÁUƒŸ™ú˜^2cHšÙj³Ž¡Û†ú«y®y÷0ïaÃ…á­#LFùÚH°[$Z´[|õJµ˜iÞs hgWVHÄ=ö¼úæ@8„&]K²zŸQǃ&$3,3"‹UÊè.AÔ—P­y×l‚o)ÌÎnŽƒ•òˆl¦RÞ¨{-пÀá¾Ð,¹#ÆóÈ‘VÙ+7ï‘A4ÕKPŠÜ©x»ù=Ï’A.~Kü-’“EûdJöÃ\ðRÏ”lµî 8J¯—ÏwB™›]—W†/Híß…°æbwZ³øª^Ò€fX„fR±CxkûˆÌ&óçM £Èèû3q6οóG DGÖñdàà[ðny³¢êä©Cxš÷”…TøÂfðàá/˜÷t%ps®¬DOƒäC´³Ú¤=)=óøjÈŽ–8¸YQ³,É>2‹L">ħ’©¨A'dÑMËÒ²RR Y«H%3¿y6¬„¹‡Ü>òk ¸(ÀO%_v\¼|ýqÁ¸8TUO–éRó¡¸jÐÕ((ˆd“'éÅÑÍ’§Çš#Óy'Ü ?Ùv 8÷ü;œŒÂ¬»d¬«{tÀ.Å9–,ý×Þ¼¿‚æ$¶òÇÏt•žîãnW2†Lrp_¥TVuÒG>¡u{hàD´_Ãפ-=ïðÊg½­•¢=‹Ð‡ánŒ"3pqU×ßFóâCœÆÞÏýo®ó,ß „28a³ß6¿í+vÏÎ\¼÷G-´J9Æ~(7ñ>òM ü:·6:S … ¯“ñdÂã7pLKgÞÙs ²ãoë=«yÕ[®‰»[µ¶Çâä¯:ºoœzs;=|mø6–ž1ž ÚÓjm2†Ë£ÈTŸEäàˆ¶uËx.V´a½J¾$E5]pÅR´“ãLЯ,!<×&ÆYï¯4íûÖÊ™Ü\ÜÙßÍOŸ>ANŠ^Éësq.ö&«ã –óÔÇ>¬/kíÜ|Ôã-'Õ^ÚßR9Yðïý}¦RßBM_2óhó®MÓ‹Võ’–O1þS©8T\Ì—CnBšR„ŒÌduzfH©OŽ/pAdQãqxdCº¢6]Ÿ¶N¦ž ÷¶D¹¯.}"¤Uh³â€‹‚ÌaÙP—äê ´B8R~Êï<Ë 8¢-Ï„ê£)<›üs6¬*\[g¸ïB%JsÃtBVbž±+!»Zaþ‚£3àaGöÀC'ãøy¸Æ ` 8ÃMG‰Ö“Xˆ‚]aщ)*X6Å._z ÿÐ ž”}~ªë?7;³ÿ÷P]ù¦¤«UÑð^x6pEPPª0éä³\\!ù…ŽÇÐ6)úŠ“xœh ÑM%×ÚxØØdr»¿% ¬õ­ˆ/®µ_¼&t©6³ÎAÁkáJ5qÅŸëüƒ§ð™¸†Y4wvñŽ5‹–d¾ à–8ýQÎvÊ•àç5×à[ƒõ¦—æÈ ¬‹N(DZÕ7¦h… öŒæÑÄ ÷ÕL½¦€o?^&½‰‰’mÖ=†6z=†æ¾¶ ô=]é•"oôx‰CùܲìÜ÷3°ÁšDðOH¸›µ ÍHQŒ' ‹-²÷Œuq…=àØÒÕÍfï;¸?5;­ ÒA“©Öf.&ëÆËñšå4>< h göŽÈݯ®ËBZ¡¼ŠgxhÌ3RšiüÓ² ²srѹ6¤éѳçé%M@OãX+ñÿ‰µ1MŸ¯jŒ­­¯«l)WŠ+ôù5Pܧ­ÞöŠ],™Cæî$R;ä¢?ûúl[{…”? òÊ<¨ªL…½ŠÕYP Õ‘\ägÎ Øê©?ýd~: PJ¡£þ‚"|ªZÌÏû, ¬›Fþ JéǨ¼2]M%O£QÌ›jƒCem/3‚1WY´#ë©RLIÚ«V/K%{I§ƒ:«mp´_ c¢äfAVnžâ1ÊPJÆÊTJyî èzèˆrdû;êU=nì¤t…FÊöêYÉÿ™“ncËÊËÊK+:7]H>I9\xö=NÂq3ï‘q®Ûb<<øcgÎj£¯ bá¼Ö‡û(=¶¥Àrð9³·Š3ï™ÿ;UÓ™äúxNíOa]œ%eKƒ|«fl rîË´ŸÉßgœ¼M/~lT«¶OÆçïÏ»lì2Mí±=´Ö÷#˜Õ‰™™iÚLLW°^vÁØfË!„í ¡}a—Yt锞´78=É2jG€“#d@$çè²ê‹ …Ó‡W†GùokW]ºuþZ{­b&àQXަRÑóEɯÕ&¥ ê˜=+l)g[Î8Õ…£qtn·&]«Õj´ŠÌÌÄãTMq‡+[vÜ"lÎFâäH•›%¯ýh‹Spvr¹‚ù‹s8MéŸR•ñái.°žsýx®Æ7¿?ý‡uuaE‚GÍ*˜KÿNHÏRæ&µPöÍË*)<Èá~|vú4|vç8mßNó‰ÿå牛ýüÖQS8Æ÷bËå DñÒËÛêoìüä­Ë‚Œvç'ùŸ)îýÿ·ï¹™¤2àq=6FöÕ£­G[$ˆGðŸÔÖ7ë¿€óÜãñß¡Wö{Ažöä×—Áã¬ÅY"? IU¡õþtÁYA/ÏNÿs~·´UÀýRt¯¸l_vbŒV¨U¤o Œ U ¾²ï>g‘Ö¥¾F=%ž ž Æfz#œŸmïºA;,T˜ãÇ´_ÄΟq¢±¥Þ#å[w7xyú{{7ø·žÐ7´ dq¦ƒ×ý'ó÷]­ŽÂåÆs§Ód«)´?œ“à #‡{Ñ%í·)òh’BG?%y·q"pßöþ@¦+å7uO(|_ÁÍ~—0]üá&Kòæç“AÔ´¾ÔDô0*o£žÇ‘¿æfÆPM4U‚Ppƒˆ~Z/ªÉÜ—‘–A…‘¬$“e¢’ýkg®…Ï~¯à«liQÉÁªC8”ØÉÝGùHË¥DÃ^AÉÖè>¡¡P mýŠÑœµ7;#;¹˜ëCM|ØÙYп)bÿîö=ìê½ðbÑŸ¶ÈýßÒžÛ´õî{#&“·-ˆâo//YÓa³K”÷_¯•’oÏ}t¼*rÑèÃÿ"ĸyÒãeÚ+$¸eR|ÔÃòu–®"RãN£,—ƒ®œ“¹Þ*¢P²mº_¡ƒ^Ï0"ãúx-; ’‹h‰®øŠTÔa•óÌ(µkBÊ>õ"H £o+oiüÇW%…(ýîÓNxÀ¡©ÕU:¨²i«fl2$T74Wžn ¬ Ü/œ<ݵ¿¸ûíKmrðT*H‰NI¥ÅеŒ™tò 8òØ÷øJ[³äB#ëã÷ÔªÂ`M˜(çì&…!¿¤ê¹Öà:OUp°ÊöñFºÚóÇgÃ~$¶µÂµ|ßp×>" bbïúæ–ñGꎕŸ®NiÚ–+´µ~ 9Àݧ  ŒíÔÛ„ú‡j¼5IÚhm†&-ÔÁÅå€Ñ P8)¥?¯À¯㈳ù‡†iâRùaíþ°X $$ äaïZ™9‘(Ù:Ý#:öMðh@e§QáCxr•@–WqÎÔÿŒ²"zSæ4† ù,±”oÚWK ¸†.þWZÊÏ¢>ó%i5ž"lßšìF÷Âmej]¦.¸$H‰SvºS%Wè² ò„¢âæöÛe˜ù–L Ã|V¿ç¯è:x²ª9´Ö'hwì¶yßÌÇ!8ëá]Žæ îq»<ö(}C¢?¦G›%È£/ZHñüdxô|U[ç‰)W7gÖ„ygćÀη6¢ñˆ¾¬ýҮƓdÑ"R¼H ¦à˜Ÿn ZM»F†D¯‚7v (!—øUv>#o¯nB>‡ÃºÎu?‚…Sl`á#ÎI•b1š!q3"½ø€Ï/ öƒRž¯\ dê$]Ø~M¡¦€SÉw tD3þ¼¬0?ç;jÉ9t–SŸeèêJ¸ÔÙ°•[Ñ y‡‹ææ KéŽÐ8³M>ºÎõe+Þ¥‰ÞÄK7+® ÿî‘‹Çá÷ĶHˆå²ÅÓ6Ö®½æ%Ù–'ØLÜc7º¾ÛܹI½äõ÷£Yñ„. ©¦G¼ù\W}ûÌgÏXåúÝü:Â[úC8¸[.\æ`ãÖõÍ—§NŸ©íËôd š^Ð㿌@[ Ÿq¶Fã1ÞÈjbè›W%ÔåTVB#×¼§: twÜNÇ îOhè¬ï°ð àAO„© É8¢x°ÍÚÞ-¹Ú(ù&Þ¾ýéÜAÉåêõÉûA:ØÍ9ÛÌŒ\G•ËŽ Ä÷8QŠ1"×'^»==õ'ú¢Ëä/ÿëãÀ›¾øë{ú;û%;dü4½Œ_Òž)â>ƒL|ÌÒÍÄ´àda3D× )‰áÊùÔ†7éV·\Øñþä…«7 ªöu@=.‘‘iþd¨bòÖèß_ ×|Å‘=œ£õSë÷º8‹ýmp*ÿ˜½{½éfY£&¶VJMˆ¢+WT^Ly2gËš'‰®ù¸³¢ HN¶ç²S” LMæW˜ nÊ73CIµÙÐ,3s†ù_ªò$¤endstream endobj 341 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 1444 >> stream xœM’kPW†wYHV ˆÈV­v¦ë¥à­ ÛAqÇŽÕˆ1€ÀbˆR×bÂGP®Á 0 B°¶¨ˆu #XF­E Jk­ZŠNm½ô[<:Óàø£ΜïÏ{žç=M¹»Q4M{„GF]fŠÓiq†›øÄ<Ú>šã2dîÍ3<|0m*'âoŠ¡éÄ}YáBòMœ:V§ð P, ž£X¢KTiâb”IŠH¥.V•¨Ô¹†Å!&N¥; ð_«Ó%‡ÎŸ¯×ëç)µózEÀ…>N«ˆRiUšTÕnÅ*!I§X«LT)ÆØæáBbrŠN¥QD »Uš$Š¢d . Z¼$x¹À‡5‰ò¡&S¾G½CM¡¦R<5ŽOÉ(O—$åN TíOŸsóqûšYÉ40¸»›ÜûE£—اwŒ²úâClyȈ•xƒt3 n¿žp.òôR`I:†ðp 6LÇt@Û 9I'½ÜvX{L?”Ñ CÐ?Á-kWí`oí7Ð ãkÖWm, ‡ ƒ0Ã}à§)1Àºž§XïfD3FpqcJö:Cvî—+ Ã…¾Fò|¸ÅÑÓÞß}íaÑÓ¯Ÿøïà ù[ÚòËíÕΚ¼–Íüñ¯möÉÅå¡A‘«öªåäK’“&¦‰’±ŽN\Gý(}ÚÕHß}ŠÂ¿ 6à=nàû â¾g}Ì¢Šn¥¼Ý\qjÙÚºAHÛ5÷ÙZôDù£á'«ˆ[ÿüÔ©›p™ ì& a–l\wtÓq[£Q_›`æ[Î\«höì•” eÆö} òx•`ÚjÔ˜L_ÀšrMYePýFÞ‰¾N1ÚUåð#6àGœ“0b”´JR2BF6O~}µÓCí!Ò8È å“Õ’ æ^³Z¡×tÁ4Ve²4,Íò!érq&÷ÊTKߘþó?qàä{4NpbéƒóÅ.î;CT‹SX+òŠò]DƃrmÎ&}°ÉQŸ}ØÀ ë 99/MƒÌª²B³¥œoëº\Ú=plÛáÜšøb¬‡DX—¸C£Þ©†øÌ©é;eüªޱm{J ¦m°qÁД¡/º z,~F¼wlÎRíàßð‰œãöHK#ÝùÖ¾tóö¹ãöÔóvU¡ÒÊäÝÚØ½û£ßì³¾Vìºkq}BÌ#žÿæÊ-`†–X%1U¨.Ú|ÈTb¬fã$±$Þ£Yb¶–÷U•—ß««Ë$É^å–ÈTL!g’E°ŽÍ‚ííçál%Z ¸éôý¶NÖë%l®ì ¯ê˜Ñ2¼ÏYë Þ®ƒüõl‰A ZMXå£$™«´ƒ½^éò×þ’th´v¨tmK1×¾ŒraJ¯2ˆ£9‰"ïn%±@æñû–Ml©,ÅFs^‰œLCY4®Œ¬{ŠS‡ú**‚ì%rùÿåY¯2\qy™™_d,Γÿ¶ì™¤ˆ*‚,$T@ZZAä³ùEyå–‘>œÔÇŸù¥)Àq€~JH4*ÆÉøZ¼Õ?Ûhqa#Æ`*‡Ý’·>rÒ=’J°k5 ËàI鑾5?"G×$1€Ö^uGx/½M ·bRy™MB”RçøÁ üx÷`›lœ£T&¬’yRÔVÈébendstream endobj 342 0 obj << /Filter /FlateDecode /Length 5640 >> stream xœÍ\Ks7’¾+6ö2€·-n˜¥Â¥ˆ=жl­­ñhez½Ž™=´ÈæÃÙ’¦,óÏo&€ªú€*´Z/džn(<™_>îu­Øëø¿ôÿãËÝÞÙƒ>¡u/ýïørïË£Ÿ½'t«´•{G§â'bÏ«¶·vÏ Å™½£Ëožïtm×÷½ëM³æ¨®SV7§üÛö½V"¶Ë¾7B5×ü[S»“YûÕ1~}ÿ]ß{eÿ÷è;Z‘ö{=Í.-/HŠÖK!öŽž>8ú÷¿7?rgCKp®i÷”Rô©n¯R»°¦YÝò?ú®3Þ4›WüßuBÈæí¾t­ÊÆå讳Ö7wЃÆÔZÓ˜¢ù6íÊ*ß\¦õ Úyú-´l¾ˆ+ðÂ6g8"Ιú8ïx§¢U^—ÍÈäºÎ[Õ\Dbxïm³z5}z&%RjÕlðƒë¸J¯‰®iPá‰"¯p DáE6Á&PÍÒ1ôÓV\³‚æ“Øl¼nV7°†·0 äu<×Ó§0馲€c]Ó’Uóbš3£NÚNˆœâ5rŠ7­†Ø÷脸õñ´ªã}C³÷‚¸E×iG\~åz›µ_Âôë°pÙuÄÃÃ9GÁil¦üö~£\!Wà>ðÇ8x0LÑ0>ç_㠆ȕuAù~™mæ&ò,[s Â$9}f$7é,E5Ï"åÉ U¿ø úø¢2@íl±Ø«VÔo?¡Aè•´ðfî­$½öÝHvæÛdV‘wýl“1±„¶VჷÜá SÒ&4{É^ä¸ÂÍàš2Ãà8X tðĆ,³6Êìa“xr¾ôåøM™–žp"ù@‹UEéð±Jo‰”„7`ia"s) x:Èó¸ffíÌ ÒðØŠ¦à:~§…G3ÐNZjÔÁ îÌÇ–ŽÌ‚ÂÔ‚Y ÉÌÀnââ565³óÅò)àÀ0_;Ê`*iØŽ$ø›}‚EIJ|@¨< 8œCŸ1x‚"F"-ºp³ìduJFq¡Øo¾ÃYƒa¢WKB´ŒàaÈ^/à1GAB M*ƒç-i=FlCžØî1èèÏï*ÞŒ"´ÆŽwÚÍýC¦\‰0^G¢ðÿ ˆaé¼ú0ˆå~Óõzîë9{_úzaZOK:Ÿ¾É®TäÌÈÙÔ0«"²¬ªœFòã©Ç¿NbSó“fâ«;X4¬´×ÈÞeT½ÄÈÙw•`ÎÌûŒfÉp !ø@‚¡¦ABœ GÒoHš4ìVæb\Õuô V¤a¥æ€†È8Ó­IF"Í—¹ì‡—d¿-5/(¢e—/¨°Ïq\ +…˜‰– úºµXØ/uçH'oºvzŒÈ ÑÉÑdD!|‰XsQŠ£Ï¥0ÊD§È%8‡c—MÒ¾ùωmpð ‰D.\ Âöoy;¯K’z>=ýÛ¤â˜>Bhy¿'[#zâ¤@Ž¿Lý£¡ÔjKèN0†½>‰Iºh(Iº¨$GÚnÕÏœLaÒš‹°&y 5jˆw^M­Óq8ògHl•—Ùªq>øªÖ–´Þ_²é5OÌdÚéÌÏ•–,jSª‘gÓ"ÐÐ c’ rÊš¦HG(ÙÓóPß9è!øþ‘Ø!–yù/ûÚI:®ÜÒçĈbï$7š757çÁ¸ËP îªH9Ĺš!Ubj2Ó?`œ2²HdâeiYÎÕÊyÆ!8åì4Qà4.Ý Ãm*CG+¹Ëbc! KCè… ÏBdû¢ÄÔ¹'¶Ï4I:iØ6Ϻßñ8kêc¦³dL”WÐÇI™.+«M5Ÿ&ŒÓQ]¶£E'—'dY‘G‹¶#Õ!K”ÀÂ:,Czd×{3žÀw3`Y°a!bÆ~ÈLáÓAáèXÔÉm0Ù Ùm®¶Pµ\ph]X!u¶"ŒDè@.•.9)4J>\Zóÿˆ ±O µœÂîÊß kà׬zT¢MK0«ç§@é³JcRˆˆŒÿz $î•É2Ë¢ 5*ä‰zDšæ§¾Ù±’ÓÕÙ¸€æüööõ£‡ß¼yÓþzs»º½ÙœÞ¶›ë³‡wJ?¼âáþѯóè$!ýÉÛš MRAI¹ï2…@G§-ã‡ÈÀuÕI-²²Š69áCÈ‚Òy ǃÛ9¹‡¥!YL" …îöaqËs!h×]Äá¼ÈÞÓÌœ"™×1’ú,úRe!f¾Šq¾ s>€ô+îáÚNfßÍó#IT‚•hF͆O'§zWjÆ@å^4È‹µÈö>š+øÁ÷à]-2'±ö2%šTÛ¹˜°sŒ‘¯.‹Ç,Ì¢ñö@2JO>R z ñ…ô0L;EeˆáDDIáLÇéì"ð0 ofioÁɧw°Ò£…Š%¶AâRüâ†Áá$ƨ~ Q¸eHäp®4~nyºƒý%17‰Ú¿?ôäá*ùN”îáE'ë(k¨Œ¤®e Ó@›Á²å:8„ZNDs˜#ä;BŽÙsFS5ÛØÖ´Ú¨~ʽ.ê;dZHGfóW™¡¸§ £ï½ÎG¼ªpW“=±Ï!¶szI'%8™héô^£^'÷ÑÂVE0ŽÃ4#OÌE§?÷\s¹Pݳù€Ÿ8lf€€/itKšÂ ã$ H—"pŠéìwâLV:rªõH¼ùÕóÃÚçtb\¾A çõõæ×õqdÓ׫㗄pÿ±=3NЩ KšãÕõ:,Tâ,@¼ÌçQ´::"t‹hí9Ÿ7*dzu*e1Ä{äYBD+T)ö –¬4ò`b 1¤K†ˆ6Ì6ÉX—¬Ûñ¦Teê§à"×ý«Zî4šÊÛÌ£Ún/[˰9Rô‡ Z£{Y ¿×â9ëËn&Ì1mJÓyÑÇvÌ1–ª"×&¡KTL»™IïBÍNV†:wƒ±’:¯Az·³…”¬ Øòg8ôk©:d€N …¶ïf){O4Teʾà†gV•ï šbà;«óh`CRý1œu“K›p*œ¬"cȳÀœi&æüÒì•Vòë©÷¸)Ýü÷>)`¢åB5ÒB\½Vv4â\Xß¡0 q[ (ãtZm9GÎG_½˜›UÖñS6Ya³rËÝíaÈÈUJ£.¢ÅÚ…Š¹±3퇫˜^pTEyÒ2oc£/ÂÓèsœÀoìS+©*bÕáĽªáo¬ŠÂrμZjê_Käo‹Î²µ>æç‰p™Æ qïPÌ'£ñÕ[úNag‘ËBiç.‘ðà¯ð¿Hmƒ©ôtg˺It1R ÞmüòopZÏá¦øõ0v¦UZ-…gí ÓHWJ# L;Dà`ˆÐ’k’m†j5ÂŒþ Ê ^“¹šåkð»—< /®ùÈX«›8›MKíxF/¤¬¹®(ÁCGµGCˆk¥eU=ú7,­>dç*>ý˜—uºÜ|I_Ÿ% ÍáoƨM ¶Îu>DËM(¢àe°D<-–·ˆ‡©¤œØÍèÑõ óH «Î°¸¬/ (†KA)½¬L%ËØÀbIg-P0PPø¼ý$MJÀ›8I‡òXŸ®ø¸¡ü`žj,«‘dO†´R…ûþdƒ©Ìclf ,ó3]zJtÍîЯ Èá× tl0 ;T‰……Í’K˜A‡/õ¹CÓpÆ{Chú =Ê>D§ƒàý0Y ‰Ï0 fëJƒ”F~!éw’&êCè„’çÐ|yÏÌÝ $TÍ€žå ó"»eK8”4ÊÌ­ôn¸MÜJíCüc7·Ò©vu+“"lÝëL‘6èq^ÞÝ^m®/|˾o•³rç2Y’Çco'ßrÁF­•? oÙ¡oÙƒoéK×RF3KpÆÊ”äO‚pÔh¤Ÿåš4¤ÎzjxS€²æ%ŸÍþ0U`koÏ`9L%/™2NXˆÊÝrå·|ï¡ð®øË²8ä”N ʨ’œU¾ó஦ÂϲLâžE‘¤%Nbãt'âhäùò SùqÜXëmüŠ/Ž §QDû¹Ë¨–+ø¢ 5í$Ø[JùAȉZb7®øÙR¸Ã½û欢¬ªõ2+Tœ¸¡?àƒ2Ϫ8Øa²æO˜gÍ‹ zÝ*+Š 4f1…“R°.¯˜õN ÊsL’ åÊ,ÃGÿ nU WšQÚ·Ç”ÂU—®¼V±CÕ᧸X‘ʈ,†Sï²Ø+&A·*³^r1ÜBˆßbnÓH{o´Ò›DÃÛ/.Y*½› ¾9¶²3:‡jÜá~…×;c3I'»¢ó GÑ…ëžE…#/‹oÓ®yŠZ1Ý™àèÌîì„ðÉX剴è?àE¸°VE°×F$úuM\)[,À;L4¡0µD­–a\VbðU8Á— %CôÑôxE)Ä»w]OöåëtfÔÆÚEâ9g§´âpu¯iÙZÝ(â-Ã-•"–3">ƒÐ«¡)Má¬ìv³¶¸Ä›0»_þŠåضLaþÉ×iK¸â€æšákLPîW %ÕY‘†ª]“ÚÝK«ò¥™áX§ZÎäYBÎ`háC¹Z…ÖÂ…æ'Õ´ÅvebÈ>˜6õ-?¨4Ö!O&û2‰[æßÝÿ˜HRqÑôÓd¯ˆÂ©\£Gu›ÛÄì¢éñËâ®×ðÙâŽY$Ì4\r³C ÎzǧNM<ʸôC^€2/M½«ù‘)¿PmaÏc¹æú÷iŠ…Uô]î_Åf)~\-k‚Þ¡h¬úú\õjŽ®Üû’|¿q*Ê#v#å×ÒFÿ³þç€å PN";sVâ'Ô€ˆLw#{k«¦kí¦ov©9:~uƒüÈ© rµöHÆWºà•íWºÊ ¡áÖ®.Ò3C¾òIv ÌJ"hƒ¶ÈvDp²u¤á®M¢ˆgX<”DçYÇ,vŽßrªCàlT%6õ‹¯€_ÞN«)“ô³Üå2†ëBóÞÅTkõþ•˜sŒ1èËü©b ·ß#‚‘˜Y¨è„F^«lb¦ûkï2Òéé™Z²œh(¬ë§á*N—*Öð~ŽíCÄ÷Ifc¢?_bÀ¬‰1‘á"ÎÓ,•¢>.‡sûëµ— X×…ëóŸß€uÑR¬ã¹ÜŸSÎs îåû€[®ƒÓÙÊ?¦þ;ÖNvg0´I{o²Xö» Pp{JQùV*ŸyÌÃÍ®B†j<„¢ÚÆÝô臾"•Ï*&¦,ô“éð©û^tFO“ .m;éž ÉF…rIæÊ!ZþVI^bî/R! X¸»é¡ƒìñ¨ÕLSe£)¶½> Â•å_ÆŠ…U&Ûy6Åt±èÃ=éŠg¹£'M˦“Ü D¡|&{»ÄwŽ77E8µ`”0MøöEf¦Tž-„9öîó˜_å-”Y‰kÒËt‚± ×´FŽ¡©™®#qÑÏPÎ/c;9‡ŸÅ]Ý-â3l¢Lß}îË_NBP¹£µP ݳ¼v‘'â¸ÆöÏöç¿7øÜä¯àý‰ÏQ‚v3žôRmY1œ *Ýu¡1º(­¼—¾ƒèâøF›;.¥µj–µ„\ 2¼[-\˜?g™Ýh»Š#™…×°t|? ˜i^FÔ¡÷ðÞÔži¨½¿RÔoñå9ÞÕ·pÄmÐÔr `Ðð^‡w9?ðF["€ –¥}´ ¯ÇÑïR‡…«ÚÍØ!©s9¬¼PåeÕxࢤ§ühùzUø³˜ÕiÅ]6ÿ3\~‡×@};òõ‹ÞÆÓÇÞ~Èê$Hÿ·)=ïÚû Ÿ ]xñ=ò£uø‰‘L‡é…$¨Ûéý+íº{ÚfÒœ¿î)©Âyüв¿Ÿ‘iž¡&ûJ5)²¤M–³ä%ŸÏ:C²ßdÇ!ÈB–B'N:»õ;/C«ù‹Cl¡O ¦jƒŠ­ Ea„Îë`‡oÁ6]FHÄÐÍ,FðÎÇã3t&Gôm/" ·±*Ãçö!OT<ãÃm\ìQøÓ1KC{ÙøÓæ´uxå-L„ÎJvjÈÇtsÅ1e々=.xp?ª"9/eŒëK÷yIœU™/nS~pVºÆhÛS?†”­,ßlþ“^ÅyŸ—ÿÞOµ‘ÏÔÏ[qOÀ™¿üW^#SF[¤¼=ëÞ³” õ·ô\(õøèÁÑÿ(SâÇendstream endobj 343 0 obj << /Filter /FlateDecode /Length 4931 >> stream xœÅ\Ks7’¾+öGðXŒ0‹x?1d¯ík¼g³;{h6[4%JÍ!)i4úó›  º> ªZ-Ëò†j¢Q‰, óË'úŸG¢—G‚ÿ•ÿׯˆ£Ëÿ| ÓèQùoýêègNŸšpûè”;:{þ ?!‚¦!wä%ñöèìÕƒÿé~>>}BJÕõÇ'ÚÛÞÕý©vô—éžlxÜá\èÞ+ß©B‡£ï‡Ù®[½æq£t¶»ÈÄôÝS–B¯â>Øî‡´fŒÚ œá»g@ãÖBÕ]ñGC®»çÏQÈ »_òlud"ÖEæ°û§_¦INš“8×1£»ÛÌŠ‹Ýö6͈ÁHMŒó %¢äÙ²7Ò “éu¢ínFrøÜ–ÆU ¶º]†X"89ϯ|•w\jí«éwy!#l·}^8¤½½L{¡bˆž·…§hiË–gf.òli6R|ݼ¾¡‹¦âku¹Ñ>øj«ÿ àÃÇÿ{öÃmC¯ƒ"a;» Ù ÀnÙ/E»TÑF®`ý5aÆ’xuK¾d©T©jùk˜´Q^¶é¼HºòGO…â]fð+žhã{Ü£)ûh­¬^UëøÄv{==à¤RE…$ñYñ–t`¼èt¨õãÓÌÆéÓ`@íêIÌâ°ñßùõ±%ê‘dcË£"FeHQ’ȽRõ¯€ÍÍ=ÌšOá’´UšÞdÔëÌc…B6öÂ9ðø¤Öt^} ‡£dü ¬Mßš¤»ŠöãŒtRÑ yt¢èÝè‘L²Å2™ ´X&5‘šbPÔz‚e‰H² \äa¦¤y²é×HF£ž”ðçc`]lõ{'¥÷ÕsF*ÂEÓ=Ì$è4+‰@¡Ÿjk7 =*o®$è&éÎe-³]"ðŠÃoYÑ$ Ö&3¨‰ÒXP¤q2_̃òŒÃ¯AW-a×&oh%Û->¼Ø&á—Š®W; ­À©@Œ¬q^`59TØ>D5MÒk´»gâ"Ð{é#ÙYÒôîEüœ …ÞÀ8,4à—•¼Î`³\h·ƒQ®Û PVVeg]±âq ì$}TÁ²]n@q  ¸E_`"iE…Uò5oŽY,Dĸ[ˆ'lÄùÞ[ËúF/2Á®"h.×µè2m:žY¨æ-Fé@’ÿÉl`«m©éKïÐfç¦Áìâ*Ȇ¸ÂñW¨æ‰C‚Ìhü ‹y> ±½Ÿ ¶µˆØ^è^@lâÑ·ˆM,õÞé±å,b“““ÀöOí0cs<ÐûäÙ´ÿ»©9ˆ×˜8aÄVnâ‘- ¶#ÿ‡½ç‡oF±‚[ð_W ®ß:s¤ƒ®à{u™y¢U1v÷eÎzž_\•µ“—u¢Æø×yUT¥EÇ”g3¼C¦èJ[¡]aò²ÉÄÐh*Ð)zk=a£déÜ‚-È­I¢å(ªûðM¦:xYé5ÎÙxH®ê~e&…¥—ÝÙŽ§¬ÁNð ½EÕÛ¼#ΕéÍ Áм®îù8#û˜‘¶œFc>_ïÕÕÎÛ ;ít–9Ð*Ï¢ß÷Ø¿@šHgs—O-HmfÄ“1RS qòkàËiùÁY= ?q0qýZ¢¦-©dˆ‰ã÷ ‘"ПXI ã0èÉ}{²½†U+˜Øïe;:78ã¿‚eÀÇ}y»ÂCçgU“ÁIE,É€âà2|– ¼;¿®ô(iãc”À»Ì’­8ÇãGÖ®Ð;j­T™s—y‘=þäIÁ‚Û$Æ"”0«ØÀk $Ÿ/¬x›§¸Hqþˆ#›D:H_íÐÛ÷e²”ÝÙ1Í´6NÀ׺f—ŸÈ°È&“xfF-9æüññ t÷¦œ'9Ø*Îhƒ"vv­‘ÐE#’zP‰aò’ÿ‚ø4ÄøÉ£¼-âN:€4—â}tîpÎ]&Û„SZ÷OBßà÷ï/ e%Î ¶+s QÅÁè¶®#c¢*„Dgìc_ájkvlÔ†ƒÆ'gV´só-˜³™íO.ÿ’•«ÜÍëù¾Ÿ"— Ä7Ó7pÆE’lùØ” ÌlàTò–Ç&~Ø—‡H’þ(×ç­âa<Ü⊠Ýxê‡â* Kä\öíÖÍ‚šîIL°äÃA|½Î&ÏQœþç)ž¥'_"]qÐq˜ãÈÂA‚®eùÀ~/`ØŒÖò¦é+Ðz:ÎÝ=60!µ‚X=¤×ùho'ù¾$°¦ûn…Îà5ˆïN8¯"'Ò`H)”iŒë‚¥X€ÙVÑ&ûGرaXº*a›´ŒÜ ;¼W]’T§5'dœð²z’0 lÚ‰4ùŽä!X¤øÐH‘ïÄRp‹è°?;·ÏMgLâl“$ýÍf¥™'B¥Ê*Þ¤ˆÝF²sÛÍÍ[¢£_ÌÈ,úûáнáëyÆ›ü±ÁVðY7 »¸Y·9¹BVOé]rWo3¼¼{j3YÛD°$`B—á#õ½9Þ…ÁïUlçæ\@GP¥ë˜=kå$fO^zC8>Ùe$=q3O³ïü=¦rXÈ.tª£hûnP‹˜|A ­„Üñ– >¡{x=?{â|”íEšÑùDæm–:‘Ô‚¹5B}FÂÀ½ŠWð(R®–ÜÁuÞ[CÌîQG¾ZÔ‡«Næ‘Õű9m‡7‚`‹Õ.A¾T˨n£“K)Ù»á štfŠñ…ê½²ƒlÿ5­Ä 9­U²Ýì…Êã€2 GÉÅ»a¯Üvá¸s¢3ÔhyœU‹’uæ £æR!b€ÈåLˆêÝxJgãci‡/lSŸýVÕª.‰<‰+.?”R4Ve3&0¾u£'8·‘DÕèI^C R+ôrw‰ ωI†T:Õ“‹Õ¸âm­œé“*5µòä”ÛσÎäR ýŸYcsZàòzðë|û$§q8èâôU~ Îo>‡·àz(¯êjH=Hiø¹Ø&.OØŠHu©žG ±ps@Ùkûzˆ«´eÇ ÕR% ù.e¸5`¿2)c{=ÚóŸÇMm2 ;}‰j€R}9hœ!ßùdçࢬß$í“‚Ž Ï‰×’BÎyå7~ ^Ëx-o7…Âùbq§™¤÷% ¾MŠò0QÍ1(”]*ÓW¥µRÖ¸§@(t•ò]ætÉŒ2KG‘…žd)…C¥Ñ/$'É÷R¶¡R«®)06­º¦ÑÏU×TlÑ{Õ•Öñ_L[•Ÿj*G{CÔþ)šJÖ’Ž·VÕ:²ýT å`*ü. Jл³v€ò ”˜ó|8òô‹¸êUP9ŠQêW¨6y!kpð!`œPFS¬'Š:“t\§„á÷£–TJªÍŠldÑ )u‚ó߼⡉+'âÄ5Øu³@Ìdˆ)¢º¿6ª,×»t÷$}¶)ûw”×.S*«søc¢&%ÂÔM[ ¹Èãì½6Y@f€ÝèºhÊ“™-ATÌZSÞï*®©îÔØéÝ5T$NÞ"€ºˆÙÅ›£1‚ÏÐ4‘*2'F‚gÑ;¨Hƒã˜Æl 33u–õ]fœd¿Œ‰ ò¿äéD9›3}¤D¡!794=ßÓ;{ eÎåR$ö¸øRÒ/ÑÔ»,BŠÆÿÎ’Í¥ÎÒWž| K%*ªª`IA¬í›6›|œµ©ƒ¦ øÚüUiÉY´3»e¿ªÅ•X³´Ý±‰ÌèTqµ9ƒúWÔ”?çƒ!4ÂhŠÌSGº_îïo¾>=}÷î]w¿º¿ë·ÿêWëþÍËÓ›7ç§?>|öÌŸ½˜œ.£„S3™çïQÐᨋº³ƒø¤8Òd\d/ÈtÃ_Ñ›ÿ.Q­hÝ%vdÉei¢›ï Ð€x‡9ßÜ”ýåÎi¿ƒâ‡e sïÇ·ŸæJ³ÛÉ3åßuöœ¹]i9NL-›ý÷™ˆû* ÉƒŽþ÷ì"BdZµ+Œà²Ás¨ã 3rù°Ê£^c[çûJ~7_£)qÖ=åóP(‰­3‘gEvÊ»ýmFÛ±ä=VJò°‚ö"è7®š.@–\wœsy¸V„Ð+¥³ŽáOë?;1¹8OʨX†Š= ˜„² ¤ º56÷  t?ÏÃìÞ>B™à¶©|5(Ôüд—Þò>ìˆ0Ü âTMŠ"^äuDªkÌÆæ¸?\žäùNÉ_ß„@°¨…åóÆ%±|”K>ÿêUjŒf9;ßgþ‹¥DM­Ì?i“q"E½U,ôuцPït6²“› Xbdl|ù©¦O"â4½ØçW4Ç—•S¯ó¡ûLÚ}ÈÂR¼IÒ°ðT{),‡Uþ]HÄ©Ï½ä–Æ‹_¥IBÑÆòô4%¸åj4 V( X_øˆÛΈïwžÈÄ.ïT!J¥x3µþŒè ´$SNÐyJŽDXXÄÐîwH¯mTL¾ö Kin›´‚Œ!E­y3ÆpïØb/À—£Îî¸õ@Ë`}S0Ìœüvû7gUî2ù@ô?‘úÁcnpâè Ó ÔžÃ®mTùo£sužŸfÌ4 p[¬8À¦3Ôvl,N𣧟ôOùhh&7ûÝÜn_lÖ÷ýööòôfµ~¹ºÜüáþêÕæÙæöж`–Új­g§%«¨u•û8Aúrl.¿,Ýât`oë›3±v¦¶-õ¹{ƒïÀôŸ»0i‚TÎÒ©ØÆ›þ2 =·¿|*í¦„X’„ü}—»˜"Qô ¬~ñÃAÆÛ2x6gò›Ìcj´ '/õZ“IÿÊÆLº¾Û¶‘³¬•1ÐBv`'Ó²ÿŒW汪/D§Ã>ÒÇêQÌÒG¾nïu®¶rÕ%ì:öÄò|¢ÜþôÃÈH)€ ›kåé-UŽð {}9)ar°™Ýø÷éNQpî ýÙIS¨¦ßç£ðuÍw¿i"¿BõÜ:QÃÓ†œ#0UÙæšÜã Ì«d';8DØ3£Ì?éŽ]Jü¨tÙå„?êTïW¶Š)oÌÉ›Ÿé̉ü Á•sg¬”—½0ÂÏ1#\·Ùã×2Û¢æu¡¾ÚmбÔz¸6LBú¹Õâ¦<²_ZËæ1?H'4gúæCWåè0Œ˧\HB±|ZŠ$܇¤x¸‚ öZ7³Áºg'X²âšJX»Öð=^ÝoæÒA"‰á³z3÷œ—®ŽV/ŽV¯|\?"göÄ¥í­t¿‡ÕãrŠVpeóÓžä¶i¢–*+6ÝÅ<Àjáɺ™‘åe÷BxÊÐW5K7&E] · Ç&Rzw›¥¡ÎAozƒréúZÖ»ü>FÚ™k‰ÎËRâ¦Êºâ’«átdœpµË ñžD:©ó”W¡@îÍÜ[Dúd1EºÇTxzÖÊ¿p£¢B¹s¨HÁÕ„R¡Hq=-µ±1‰\›¹Rljb³è‰öÜ`!gïƒL çKÁÄþ !3=ˆ€2t vç3ûí±çÄ¿)>TVÌ8­—R¬‹9ª½ømwïauŽ/™Á$ã €Ž!uºß¹£íPüF¾»__o.®îî‹N˜« 3NóeÀ «¢Lj~9ÞËK@M³ÙÏüÌø„oÊ:¹ß¹± PÒ3ê¡Ôj|f~Ró¬²*J÷”ÏuÎÿ†GÀt-Ï4qÎŒóî=-0v]êã”ê² ^ØDú‡µbò›ð=ΊüB»t:Ìúócjí8Õ—ˆ$‡9’Z ²2h?éKT>ßì ì5èτɳÀÉ«2ÛHß}‘k ‰¸¨úF¼j­|.D×=¤ Ñ9fC•j û‘[ Û%ØßN°œÙaÁˆƒwŽ= :]Y–“¶(BP/ÍA6` ¡hæX·7èf[+4ùŽZ¶¿‡‚­B?èFHJÒÜ™´Ia|¹éÛöYø^¦-˜rhwO­¬Rä†÷]†p‡òƒôЙYú3>¡B+ù·9„›Vh_¤íöyÆü·Rž^Iq:ç{ÅúY97Kw}¯Wx¸¢•:æå—‚JãLº”ˆW½Îù7•øî”î^P"E„;к“øª¹™sÙá‡fß#M¸ÈÀ¡]JŸ„°ûî÷dê:ÿì’éÇ0i€s·ÒRæ©\iJ8œéÙ·N0‡±®Ú›IZ}¤_-З$¿B}Ns´v&%™šjs×ÿFñú ¸þ•ûŠxi«Iý(H²À0ó0$9ï Ä–€ÀÅᥛø±K7V ~±+¡“0œHbÒ%tÊ? s:yþµªÁ îC'wz%â2:-‡F49;s2¥Ñ WÞgù B¥KAz§"þæìÁÑ¿ÿÞ‹°{endstream endobj 344 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 7443 >> stream xœ­Z \Sg¶OŒÄë2´‚Q¢{q©[ÕªmÝZ[׺UÁ±–dÙ„ldÏIBvvB؉‚ ˆàn-ŠK]ª­­KµÎ´S;í›™Úö»xyï÷¾¶ÎLgž}ïýü)¹7ùÎùoç\ج¡CXl6{ØÊÛÖÍ›ëùëTz›~aý{N,c|L÷æ{Á(ŒÚøÂ´E>ô £Qés(øy‡ÍNÉT¬LIÍI‹Û›á?=r†ÿ¼Å‹ÎòŸ?wîbÿåIÑiq‘áÉþÃ3b£“Â3ð?ý·¦DÆEgäøO#6##uÉË/ ‚9áIésRÒö¼9c–¿ .#ÖKtztZVt”ÿÛ)Éþ›Â“¢ý>Ýœ/+S’R33¢Óü7¦DE§%³X¬eË“W¤¬L]µ:ííô5k3×e­„gG¼“¹17jSt@LàžØ-q[ã·%lOL š;oþ+¯¾6i‰ïÒiÓßšñÒ¨Y³ç°X“X¬É¬@ÖÖfÖ‹¬©¬­¬i¬m¬é¬í¬¬¬™¬ ÖK¬¬¬`ÖJÖlÖ.Ö*ÖÖ»¬Õ¬—Yo³æ²Ö°æ±Ö²Ö±^a­g½ÊÚÀzõkk#kkk4ˇåËÃâ±Æ²¬ ,‹d¬á,ÖÖ(Vëw,oÖs¬å¸ ¬¡,{+ûÑuCº9BÎwC?ôZÇõåÚ‡»Lœ?üÁˆwG|:²lÔ›£˜ßi½yÞçž›ùüÐçŽN÷™íc÷ùÁW7f˜c¼´±c¿×î÷&ŸÃ¿3^9Møî…Ï~¿ü÷ÇI5‰úÒß6qÄÄMUoLò›”2™=yçäÎ)¦)×_ñâLZãÝ+7Úå¦×:Ùg{ÇAÛ/ì­0u¬VªÓ傎Ðèód 4kÁ û4p|ÔdST8ä ×SJý~Ýf*´•]G>~(›ë`–x Ò%aR dK ÁH¶@›±šá€¦És±¹Ê‹ä ÒXkµTX/WçÒ…À‚Z¡”‰ ‡WÅÜöb¢¹ÞôY©‹žíb£==hc墕<“Ë\ir}YCquç'78lÉÚÊŒERâ8M>d¡õ{Ž~Ý„f˜ìrPˆA›§&å©©36‘¥+)®±×Z«)ç©v4º‰»Û\»¿-&ÇâÃ`2„IÙ‘â@Ð2;˜Šôz—‰,¿VqÌ D-$K…šlµZÆœÒ*u*Ðð¥Ö_ÛµÒF1þÌ-ó<7Œ³Ñ®7S×P“ZÆýÚüÚÞ™¯Of†PÞ´u°¢C®¡Âkڌܼo¯¾Ê<Ïð˜IÌÔ)6~ŽžG<|ãi$cfœ¼¨÷õ㵈³ï4ÕÒs´ñ ÇêCÃ2§$PÉÛv†ì»74ÓÝ;ÛÍF3pzOô.åé ôf( :«L«i>ùÎò°«›KƒÏD0 L&#aý4‘(±QüQ» TJÈ5ÔNf)3zÃ" ¾rM:_ƒæÝFC¨+úS×§@üxà¥W÷ª@CyÚâ0êË H|"f‡Íê?p™ó%3†'O”k¢tD>F Â¤²Pu¾õ°š4î~bøÊ@´Ìv³¥¶ ”:€¦Zrk"Nž‘h:šñSÀ±€]a{â’)IwHaÚ3 B BE u6ƒgm:ë} :í[ÂS™‘É(IZØê@l–Ý·õFƒ•òîUƒ»—íf÷®ýœƒ–ÝàÉãpkø¥·ç¸h3ÚŠâP&š¼äo b.ý 0nµ5[ôXßrïÜ aÆ&3 ¦Q»¶¬ ˆ)OAÁ»w Fæ97ÚßÄî]ßÂé} IxzƒÑzÂ! f{F>¹™!c×3¯Áx?uíUTïõ€{Å"LUèòäÉhµJÐ;˜FCyù>Jõ üB ¨Å ÉR’Ù«²3—±„ÛƒtÖýžÉG£¹ˆU“´*U+Ö¤S:•.ßsë Ê dZm¬UBðûöpse#¶€³äù:P¨ó(f|ßþ2ñ.Îu‚^o)!¼éÅÒJtõV FÁŒ{ä@í<½±ñ¾ÔƒS+Õ$€ŠÛ²kËÝÅO®wÄ¿½'8SHJOEUÄÿ/Yßhi°0ëýqm¸Q®í›-t}Æûœ{É–—š«â‹´ z𨵦þ‚èM`â{j/ÖirTdÎ*aþRîETø¤`ù?¹ßN×dk2(ÐëWkõî3ÔÊFª Û  Žh:=$°b䃒qW€µ¨ %XfI]ÜFs®wùܼ¯8ƒ|>Ø}cœïc(˜wXÖ)ƒíDBêëS27›NÉ&8¢B\”•‘’²©{ÏGhf+òîz°´<ÂFúþ˜bÞcí_]yåv]‹,ÊA&¨å© #¤æÜê²ÚÂŽ®ˆƒë˜á/‘¾áÍK¢z%Ö’âL7¹éåd-Šæ™« öÃ@9 Üª6åS‰‡›!Æ0ÏQäbHPBY¹EP¦+ÒUj¬:´ò”ÌùZiRJXÔ6I4³ˆøÝÕè•;ˆóïT¥©$Œÿï$ìµGѤ?Ýèji lé®=°€ÁltÞè>VY¾­tb59ÇAEôÞ}n™c°â¯s[Ðô‚b0àþWæšdi²,lPQÌ+^ ¸¹O }›»’iÓ©ÕBñƒ÷Gžüë‰A;“€.GNæ¼¹l ©PSjRK=5àäçݨÉÍ>ßëæô išgÝgµ^¢Òå…AAEC„& "!Âå)œRü‚²‰Ñz1uÜsH9p)ÿÙ/õ˜Ô?b®Ú¡ Ú5=Y<Ö¯Áf°Ú<€+¦.£/TÇ]Ç8Ã¥z°ôcuÿ —½Üƒæõph;=g0cV ‡¼D¤H–kÉL§*´ùb~vJrR(!ù‡÷÷ìÿ²æ e­4Cq<Þ6'–™®`½¡¸€´ºë¾éÂáÐìM—¥Š¢¨½«÷.‡ âÕîäÓ§š«ÉòíÊ hÊšwyka7fAotºL%)ÚšýN{¡®x°74o@î? 9ô8áÆ˜mWqcìýN¤¦"!\fŒðœ7¹ŸÑM ¸ºZ¬óB_»<>_¾øYù#µh-ýŒn‡vèдõ—Èckrl‹Cæþ… –hnOñ4ѣxÍwrFZF†Ø¢0+Èê½Æ™Rv3̰íÒãû»÷=¨»NšJÍ%ÏXÄ‚ØM̽˜|½Ñ…þ³»²+Óué;Zäf?üžC7ád©·X;(Á$°)r K®&B`·1Ôóá=^ÝûBn`žk]çy8¬1D_Œ_ö²-«·áÅÖÈÛ :&Ñêjª©eiáÍ5§#/>üÓ'_¹0f žz^šÓƒð±ŸGoð)º,ÒÄE;VÃÇE§k?vß%-¦g=f0óì…Tˆ³ÆÕ¦ïKéÔ”áHÑ\Ôx¢$¦-õ,œƒCí5^9ˆæÀ)â×j ‘ “r#2ƒ"Þ"XØ~Ȧ¯v¦¡7Ì.{l“©¶Am‚SëòA¬+è­žq¾Ÿ¢¿å9vµd5?3î® š§d&A§Ò‘¾Ù ¥ 3‚™ÅL]еúÊ÷Óp¸²íqYx»™Íò„K¶·ülãq7^~ˆªéé:ÞŽ£vMP¸V§Ò*1¶"¥®âÛhïíbûÔ=tü‡N¦gò"eƒÇ†'>{bsaò¶ø]Ù22óÜ.kÓ¶3ã2ú« úRU÷ÅɃm@4TdD¦KvåÇPÒZúžp}Öë tÈ-n´Új „¨Í2‹ÓÒ²â£etÝßÚXF:·“Öñý 4Þ9X¡ŠJ}-8*AjáÉjG‡µ‰r*9oÝ_|²îý#`„(Oˆb±Ñ°Pêe:Ó;7b@52NCN—KtRdó¥v(>jPÈ41èÈé}”N#‹€hþÂCkn ¡h8š‰¦y2ct\VB©éóæ¥ÅW~rØÖÜ€¸”¥ %Ά¶Ö¦Ö_Ë…Þ½Ãäá<}–C¥ñlÍfû ìž!ʼn¢y£?ÜùlòÛ–q]8™Ën¢?Ð8pØÃ M#Ñi²dÂü0†`æz2ÚS9‘ÎaÂKNÝîrÑSœR—ÏÜOî)93Î÷GÔ[Ë‹f/ â]Q«Ic—žr› ë ”¨:r$QÊ­Aíqï?úêûÿ¨!¿£I½Ù€µ˜ÿ´ñòýs²"2 pÂÿU†q¸èwuQÑ@/Që­Î©&ü>‡^CÿެÄÕÒB‡°ª®ÚÕz,¬eÓ!gúÙu@c>ÿK‰Iî«u:™œ\>{¾:ˆíiݧê~¨=J¸t¼¹ Q§¶$‚ð~~ƒóxôß+žes)f 7ÃUšÃ|ÝwÞo+Ý­ui1xù ÿׯ¹1}…¢­Ñ’-_.Õåä›5ªšŒnž Ž(Øî z µÅ ­…Œ[sA)âÓõ~znOß÷†\P”¿L³ƒÞËÆ/3û.ü£‰¶ÀAC"¤:œÖÂJ(#Šò+öæ$Š¢"[rîýpóÛ$bÓë Jp)è ’ y²šôý¦A”Z6Ê c¦3Yƒ#È' £½°¸„Šy!Ìzy¦¥8‚l×ÞtmDÃ+P5§wâœm¡œú;¸¦¼ç·ù73½ïN«Ë-_jË«¨,*sÉ‚’}h¬¥~`Áh; 7”ïsîë¸píÿcÁ¸666)÷9)Ó…;}Ê…NºÙtb7}é1|œŽ­ þµ*€¸øÌVªçlßN%%g¬b*ˆxhÒ+àqë­«Ìh’){f8xÖ¯ÿ5.¡R/Äâ:)\)Òä‹¶RÒÝjdË{.ÜküÑjÑšÅbJ¢!s·&%$‘¤©ÛWk¬6ÖRÆ*Óa¨„3±‡v–{¸5Ðþ4æÌ‘÷ÃïŽó¥Ñ=4™÷T•’S_ÂU’‹ôuµEû‹Z¨‹è-{îZÿý蚘ÙqÌ Õ/ À¶¿öëþ ÊH—&+²©¤Eâx þÅE«1߈¨êˆ“÷£™¤ïUt¯F`e¥ RÒœòbg}EÒg3ݽsëÙèÝ+(ô §wÄ߯'U:žíC¶F]^¥`)}ŽaãßÂ\;ºS)³Æ¥,Qù̳jÊr"G&ì!æ.6Ù²ªLuF ½CΑ½Rë9a­Uè4 á+Mj³Eo°[IGaí>{QËö‹y'À¬æ ñȧAѾw•æ–gÕg›m*C\Iî>h'î~úÇûë×,ß«S¨3IÈë©ug¿~ÙnÓŽþ2ÇÞC±çv>η]E×1ü J¡˜8’Ô±uúfšâIÌ6“5ÚÎã·®1dæ ì±fõïˆZºø9 ²jõÍã>0ýÌà˜–¬Ã7o_º_Iú~Æ÷.æÕgš%iéééb» ººÆUƒ± nšia÷J[8ô1<Óz0lÕ[ uPÄ£*ÑF<ÒÉ)­´8ˆ‰=ŸÅ`,7’h¾LX<Û²ÏÎàè¤Ì dÖøa¿ó|³´Œ`4’ƒ^o¯=‚&ÖW"MÜ*P+Ú|<0ãûšòcóe¯`°ìM ÁD€SÆpNkN ®¹œäQy\!$4bT@1ö: ýAWK®»v§h@JŽ\<{ÿÿ$%–z›3êÞš‡ÈS`,R‡…¼—wˆýðÜo¬ûè>œ©TÈãï8s´ÕYWSD–5_,«J * ©HË]žü_›wo÷/O•6õ ÿ`é +Žwt_¢ væ¥*’µù8?ýÓÞ6ôêÀ5üºèo÷6;w¯ M"åWƒ*B vÅǽ÷ÿðdêqnfeïD ˆ¨ÇÓ9¥ôe^ÓùêÒ£v} ì„ì2öjÒ1&b OÚˆ1a¡ÊtMJìóU¾Rʘûbü2i©NßïêPf®¸¤'¬˜f6¹AIÅ@”&¦ÿý°ò v!û¡ežÒ«A‡ƒ¼V)cD}Ëð}DºJ]Íe=¿x`ÐÕüê pã;Až®§êûþSŸePãѧU“µÕÓàçêSé³AY|+ŽjörÏtéêõÃ~]ºÀ¡OÒ ^› ex+_ßÍpÈ]s˜•0ŸH†Bã?|ØŽ^¼D¶ÝŠæÃ‚9Ù§à%„9?»û :Ú|jùä´« ˆöI–¯f„Í¢2wï Ïû` Lq#¶“M‡ôòž`ת,P*uº< ùæ”ÕŠdʃß\ËSAÉ1só’Ñ—,OË–_> \8eSÐiì€ÃÐ¥9üó:OBJÎC®^¬Ï4*õX¹Ì›¾ ÿŒ×,žÁ Ä]܈F¹Ðä8hÖǼã_Ühw—£9¾:ˆG aKO/e¦/fBÖ Ìa·¡Q­åùÍb¥J¥Äb«×’fI$DrÌÄM™©þÒ'(y·I]ļøêÂédHjxzxÐê…ÒÕ ë^¡Á€“µ¾ü@#'ê5¹µ9îÌ6ìmúРǓeŒšED­] )eì±FÙAŠ¥E«èÑ£Ûh‚û@ßór1qdV(£ê¹ÝÚ%9ùÓƒ¨”¸÷ÒB ‚µw¨‹Z „(•–fìÔä¬9p«ÏhT…úÆ[Í䣿î¯à&qý±¥ÌÈ©“ÄV$·´t5¯º%ž“kôZ£’Ôë,bÈ#âR’ÊŠò[_=ü¶åhÎJ¡Œ d<\¬D¯4V6Ò7ìÃô =ñú­‘¨M×k’”{sÕ¤FÍÌÔ)WÔ¶‹“í1Ž\¬tldbªBÆntô„'þ&+3Rd<3:”µŠ!Þ`ÆÌ_­Ö(4 ¹ $Ô±Õ\Rhu6«ÍrÊñ8Ý,-RUË\òf†Ÿ:ßQMd•©ìEŽ*3Ù‚¦šuV0¿þRQ©IYˆ j•£Hë÷5¨ôýu®GÄh‹ ‘7Ù?¹Qûu:D¿Ï3:ôz=xXl×T(ª´Nܽ!ø¼Ä¢4h4jVE…BäQ QöäR¥QmPã›j•é[Ì~:NîQ!-£ÅBž8÷¾­³Ä»}[ò¤¬H*tI`f Ö¬MNÝ·F>ÑTÑPåg)Vt¬ûÁ9hšu+ôæ«ïEä&Çà\ï ©Ê,trâ>(y¥W_óü1Î÷¿P7ýˆ' ˆ^µn»c WBcTX*Ðóh$žT‡Â"Ób“ªd•eå6»DÆ>û7ãT?¢ü–^ˆÍ°ÑçûW”vçV¦$Ä'Åï-Ï­mrïo$}{î ÏxUBgJ\bBbF© ¾É½ÏíÁ m@[Ùèk´‡C×÷>ϳ™6<• í‘\žŸO2“˜I2/·å[ŒÁq޲ ~šÓ惆ѽ}=tï8ßû½ŸW‚àÌÁÑ¥¯ƒ+€@à„bÊ÷*ýóðKy9ø¥r|0ióªÉé”ïcfhßKèýãÚ‡æÁÙü 9ÇÓ.‡u‹€PªÕ´È¨Ó‹­ `50/ö cL˜H„˨!³ »ƒ&= ѴζÁO8{Y¶T¢Š6äÄ7ö§iryÈÎ-®À]ÊÆ]b€{é Š1ûwê5ÐSxh÷S=Ù¨¥²Æ]BþÏ n¢úw¶¾C¹$#ç–íüsÑä%ŽhjÒBó!jp¹ï‰Ÿ0# !–G1SÉœÙ^ëy|yžû6Ó+Û%Ý%ÙÅWìïØž}¨¸ü– »g7yME/Ö÷Ï6Ûû±[é!"go­»»JxyooØ)ùºåè9ô;4r¿¨6á t¥Ž¾E/ 8Ãþ™=.À«Àªc,4˜Í.<îõOLž6°LеJU¶çç¬P>À¶‡üœ_¥¼w¦‹^é@é†R— w s¸2’1t¡sÔp—mÔ(ë¿ÏÞ¼endstream endobj 345 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 7344 >> stream xœyx”Uºÿ ñ!Xv4ø}W¹ê‚Ð{ éuR&eÒ¦Ï|3ï7½§M&=“ !$ôE”ª(wuYlÜu­×{&œìîÿLÂ*»êÿê}xò@B¾óó¾¿÷WÎðyÇñø|þ} —/_:ã¹Ð?Ÿ Žç' >&HÁ¦;-ýÊ0Z£‡7Mø=z(¸îAô籨áÞ|¾8G­ã,Îr_m «çà±S ÅÒ¬¤„DIää˜)‘3fÍzijäóÏ=7+òÕ´¸¬¤˜èôÈåђĸ´h ù&5r8&)N"œ<'Q"ɘýì³yyyÓ£Ó²§‹³æM™™—$IŒ\——•ùº8]¹":-.rh«Ó‡þZ(NËÈ‘ÄeE.ÇÆe¥Ç$&E§f$FDÇÆ¥J¢ã2²“RÅéäÛ„è´´èÔè´]±ÑÑYò{±IññY‰bItNvRBZhk’h7+}£xSF梬ìÅ’%9Ksó¢—å¿)Y»".*!quÒšµ)©ëÓ6<2gÞ˳ç2÷Ÿôàá+ã&ŒˆÇ[ɛŋâÍæ­â­æ=É[Ã[Ë›Ì[Ç›Â[Ï{š··‘·€7•·‰7·™÷o:oïYÞë¼¼%¼¥¼™¼7xÿÁ{÷&o9ï%Þ ÞxÞc<†7†7–ÔÚÀ>Ç7òM|3ß·òm|;ßÁwò]¼Ù¤¼á¼ýü¾a†}%h¾`D䈦°ÉaßÞ·‹J9n¤{ÔC£þxñhó˜WÆ\·†ßûþ²_|Ð÷М‡®?¬ý]Æïþ.ì|$çÑ•vF̽0.|Üáñ£ÇoïxüÿLp<öäcwè|FÄ‹|9²fâ´‰µ¿Ÿ´~Ò‘ üBîNóóѾÒó‚[§…l‘FZJêU–º+ì>ÍݬµþO"¬~«ßRm.·7ræÒÀ±Oÿ ¨G|òòÍøw…;¥Ä 9µµ-ñÀíô¤Ý©Vl‘–Î[œ¼f3Pjgu­£Æág*u¡aÐK½½³sáŒÍË$´öä®À6Øù± iaRqè(…ŒNŽó[éò˾ƒí@ù\%ɹ¬¼dÞýü?( >+| maíz3p"°–Wv™œ“ƒrÊ-–Þ;ÙhÈ‚å¦,8<úq òäNðôY9ÂÍ8ç•È7¯ßZ¸‘•”ù”º ZLè€=l{èI«|N§g”\ª ¬´Ìíùñ">ùë×Xe–sЋ@•_«“”( d.0Ó pÉÔµp†­ ­bsC¥[*f §7tÆ~¾='é{çÒ£û›kv×ô™<œ ’âÀYB§A2› ñgJÜ} ä©-¬ÍÀ8 =ˆ¡X^¢RRáÁü¡Z¯ ‚I¨NøÝ¢·fàqäÏ$G"4ÑèIûp©pçŽÀû‡ê¯ùsøzoiP§š¶¬ŒÍ›÷“¸jmÒ6 ‹^¿[âÛÁ B¯·©©¥õ‹ÈÚ]eךUt&¤³ ‰¦”»»Ë—; [£>·@’©Ó2¸ˆ¼m:¬ñ‚‘t§æ8½Nð¹BÃ"V|OcJî6ƬÌ"‹ÿȱ&9°"`cð ZiÉZû²ÊW‘_sƒ‘ªk\aîÖÕçRB¾>ŒL®Qc7Øm–ª641ÂRÉ™ìr$Àëhê`­‚/^ôàp¡*I¯Þ”Lù*‹ÎÎÔC#Y¸ZÙæÐÂYXÅ›Èjî.ô’­–ñ¤7$ QèAô$zêöê ó7mIˆMfÇwVÅbX)•&•¬ö§.÷ªâ$¬\WÀœ— mõõ ¥mF[9P5‰QQÉø>i“´hKr´Áhr3áÁÕ?î¼ó² ¸1¸Phf­jyLLzs<¶,(< Å“ñSN/¿ºoOM{;£LLÆãY¥J¬Ò­1P% È A›i†69!´³-÷œ±p 9£Íh5:Ì  •ºë忼‹ \;Úµ¨·¼³ÕšµXÌê˜"ñšekJús!€ú¨¥•ß¿ “ Œ!ôÊ”zFO¿'Ç/ÁóÉîž +=h8Çr`-ï ºˆ/Ãþ¹Yj©Ó½ÖÀ’²i͇Ýd)5Ó½¨d„UkQ“Åd,;_*y¨ia}HãhãÌÝ BÃPxSú¬lÐë2ÐÔ„»´&µÃÅY+´½jª×œ‰Œay¤rTv4€Tz­”ÁuŠX½jˆ dä?™`YX5pœ­Œfƒ¼ ùØCØñ/Ÿÿ,7Í­h‚¹ÂÞRqÉZoö4¨ñÛmuû Je¤is²!“*tå“VW··¥4m|cÇfq>­êIhÞú+°r—rÿ±[hòÕd+¢^4¥W%òr ‡*vùý5å{zvî#âŸØ1÷ZQ£†>ª8*‡µT¦ä%<ºx“лµµMÐ@Uºs3Ór·¯î¹‰^܃Æô´ç×ÊÜtFCºe‡#Ácƒª±æÌ­ÚÃêØR:¼¿>'€ŠÁ'ü#Ÿ ú¹þeB“Þ¦d ¹œ^4OéÝö.€gãœI¦~&àûЋÝ=ŽÒvæ°¯²ªÂçȪ‹we1š8Æhtî¶TU^0—[ÊÍå®ÎØËQWi#¼ÂH Š•@*Ì5¥†à© q¶ü\#è«™:ÖÂB (åÛñb< ?"Oê™I½hZÀ‰fý7z„9þÎg¾ê›Óó0ÃÀ3 ÂÄ0VXèpô7B•¢ZèãŸ?'@»‚¿Þó9@ÊÌ kG“,>°Rå2g+6$*éX¾>; ¨\…§¥Æh3—¡ºBæãlµøýæà·BG»Ñy¨:²”‹Ð#†…„Óá5Szè¬*ùB• /g*ô9«Î gKVb6Ÿ ;dŽv“ó ˆ~ã,+ÃãZ#†ÐNI•CDÚpÙ Lø ä üú* ìw¹4^ V£…³2ï úèTX4ž-™ª+R¤ä–Uô[V^¼ë2xç]}ôíI¡kW[æ>hƒêê²æòvσŽÁ +féâ5ù‹£‡ éw°·1 ˆg*uî6—Š,ÕÎݤAóÚLC’ŠÎ_°/”­7èb^^ÔEOÇɆê.2î&‹—iOJǦ©ñø åÐ|Ë,tíŸwwìÊQfHÉ–§kŠ˜øÉ¹Ëa35õbÖñ í=þr:<¸ì®úlšÉùP5v"ÞrÐ1éð*›i°À”ª¸ºd¨â]õ¯½^ÉàñíÊxVAhE2T&ÓWMÐר¦ÐCöPAŽT Æ<¢X:3¸Àn´m?ê9âŸwö¡Go ‚‰AJx<³Lš•™SbѸätUž'¨èXŠGn”´=Y5ð!m+ÿáÐÓâþåÐ_ï«n$½)‡”,E†ZÊìÀTÉ6ØIaÞ­ÂýhN/úè(™ôÿ|õ[޾oΣ(òò]è3¡Åïì¸[u™T•§¦YõV¼F©NÓKA rÐsZÉú´ÔX ^†kåÍ–=ÐXÆÕdõ‚êIîØüo-¨ÿ|O m¨Yr±º€ìf”lÄQϽønù¸ÔI×-»¨i‚w¡û\]_ëûÍh¤Nd”de‘X5í’×fÂHJŸ’¹1s}b"ñ˜[3šzl³—@®Vþƒ^Ž † ¥éõÛ‰Äà§ð“ÓN,}ßwØ|¹±G·%ü%$¼!ÚT[çóí®nÆÕet:v·£È[±5æAªÈ.ÊÉa‰Ö¹…Z)è‹ v¬ÝU(jSñ‰=‡›ÑØêýLàtOGgHBgUÔ´„>¹ 1kp²Ð`Ò…$‘«ðºìh!qûâs“ɲ²”™µ©èÈb³!b‡Ì ZöO3è×;´P Ò4‰F;0‹¸® ñŒ¨Ö>‡Ê¨ùw’ ‰ÏSa0¸™2½•%-ü I J"ÁîÏŒ¸ÍóãˆÛ¬~S)ã=ê9æ9Ц *QØïΉ½(m§A)—1Ý ]¦N²ÂqÖÿƒïƒ’ÙÅ•8ŠÚîß‘û=£èà6~†ÞþL€þŒN­uÕŸ] Ô@U‰X›—Ùƒ‚êo®è<åMZ•¸^"£³.¬s¦õøj,Ì ÁÌn2:ŒLë÷j[ªlS®J-X#™Âå‰Ñ;@ÈõÔZüUD ësìEiY™)Ûçu_h¿d5ÑîøöØ> ¾íEÂÚ”zZÏ$G.Î&6+¶°¹•!…ðþ9?°Ä7!–pݪiˆ%´L2Ìb“ïa Ù¯áå‚!‹RÁcÂj€#æ”av… tJ¥A[¨¦“žÝÃð«!+ôSSE Y‚£ý]·ºî¡Ü`AðA¡9dÄmƒS¬(Ô)i½V•ª“’1ÎÐdlÇK"tE;-j£¼»cùÙÖ[LÙ^ònßÏ‘hýíΦÖòI¦:Ÿ‰™(] [©çÏ¥ÞÓZÞÐ@—lmŽYŸ™—¾Às»©`OxÀëv:œÂQ ³B’\¸ySwüÙo¿ýh ´ûšÜZouf™ôÏÁÈàÂBòtqqa‘¾ÔTv…¼²±¾¦óÐÖŽ×ðx,Âã‰3NÍù=þÑ÷n«Æ¦Ñêõ* =g2æƒ ¨73õíõ}Ñz– ôžØw‚˜ÞÝ%Þ *üÎö¡ÎÝyþ®(¿ TµcÐW– åðÞŽ˜<ÇVë!“ø›ÇÂòå¿88X#1)³‰-1H•VÖþ³N[jf5'÷A5Ø8‹É…&»"¸°›ãe™‘ ¬.‹-8¹ ù1~|àÒ½5$% ðÉ({a»w0šÞ3‹æ÷•rŠï¤?Oÿ|a(PMUê h•A£…Š£ô Q­ Ù¼•X1ÄŽ¢ÿ;;žeÞ߀ÝCMègƒï &Êâæ*C[%ã“HÆ'‰¤Ëä{’»é5¸r!X}‘ÅÃVEàÌ`¥ªÂÀIÈîÇ…å)‡ÆÌrŽKh²†Z"ÒÈ ùêPÀû™–8•œœYeTùÁw·%ÁÚ³ô2ÜŒ°˜ô¢Ï`œ®8Ò ?zÙÏß•·ÐJT(l„kÆÃU—š,§ÁKuÇí} ð3Oã…Oœžó×_ç\åGÍn.Pz†±ú…{ÐÔoKë€:PŸžô8þØEêó1»ëÐ9R¢ÉŸ ÐItQˆæ…¡Iˆù ¿¾ô~”Á×ÿ˜üíd>ä×и`ÐÑf4íê/aËx=6ï…ØfÓö•*’–ñİð`á÷Ÿnúƒ“üü«WÁ”à!G'’Dâ1t´´#±.†<F`‰?ùLÔ•Ó¸œ gÈ}^¶ ô"2iòDȇdNìhr¸ê œ*“Wfä¤Hcvuäý=xåý¾R:mj2%Y9ù¹ÒT)!úMEG;I:U„©?´¯³ ¨ ÞÙZ-–ŽÀÿêðG7C×Z¨óîH¨ô™däâàï™üÄÀUÝ •Š$>­£¬ÌíµÐ•ïù?ºó­öÎjrN¶ÖûË•¤.û\I ëâ0?{£’h <îê«h}Êkkï½ú*Û׃î‡SÔ[1]¯¼¹%u[<ñ\ã'>æGG‰”4Þh%¼wðŽÐZcrVš\3%u `rRáë±±Lfæú”¤¦ÃÃP\‰ðk?ºÿ|MÝœq?JãÝ?LÀ¯ ?ܤ\ºIù©w¿€*F ûÃ>­X“X¤JÕHµXON‹yh4xþÔöµ—ÈA©W«õtÖ«; $@%«ý{Ö#æfÆè%ÛðÂõ% 'ކîÑ#óoôí#9 £„ÐW\&ÉÈÈÌÈ÷ÔT7U7ÑóF ÝDgêí¾_r³÷¨O®:W§`T™Û1U¸S‘Y’ ¹ÔÖÖă¤O8~ì„ôÍäïvÂê­>CPÒŸè®_BE—ý¢»±Vk0Ëè-«3l¨}ƒÊÃx4¾™}lÛ‘"æÄÊE0K4cÖ´)˰À¨±ÔpVš3s\ˆÕA!©6­Ó­Áù$×k,:«…ã¼núPï¾U€F!zàbRk‘—‘T™Ÿ¯Î±&8•Æ GR#£þóÆŸþëÃ3;ñð4´”n'g*7…üOq?8?øÐÖ`¶ºMc )þºG7·H«Ëd :½,ôâÐ-ˆÑ^j¤ëѲ='˜D!}£ކeƒÁ Î_‰_‹0èµ÷d—Z›“³”YéãèÙº 4¨=a  SiX™ 1i€$y$ˆ4„HˆFÕ7¡žÀ§‘ Á§ÁP‚I8ú œ!© 3«Œ™LºIe‚2âf,æRjP›Pß'î¡ñÜ÷ö©ÏóxzJn+mqô åÖÚVö…°Ô cz™>tÕs´©¾¬ÃO×v»Nü‹bå-l$ –_âj/7W›Ës9˜Á o½Ùø¢+PÑÜ\]sðØþ}‡ÀJök4¸õÄViŒ0õ[Xe,Ƚg]£O“¤e»x›êå…tR\lL\÷¹¿_ØÅßN€\?=Q>E¬ô §T…NôÐËÊõj(­îN<ÐRëk/§+v»Oý›¾æGúLƒfo g±TÐÞÈ©êŸH ^z¿ËÔats\9q‡ÿr'›tϬÃ`cêõ• È…¢¢0ྡྷˆÅÁºåÕžý^×PBˆÙÉÂ.Z ù\(Ïå €"Ž5¤+e'9e–‰ù`ào¦<3ëQ˜-vñPÆVÿØÿë'Yj±*c*o¾.O™ZQèÙ€µ¶ÙN{O&uì1yŽ¡Ñˆ?D·«£qXÞzF™­“ÿJº­<ÚEˆ³wn_Ù¸c}þÉࣂà²þ­BâeÍ„BP«Y Kã¯þ¾œ@A†ÁÍyL¶r+ñ#ˆs ~åçoŸºBÙÒDYuV•F§Ukh,Åöô°ªVrL$¯Î9<¨U™ŠCä0Z-F¦tO³Ù MÐ4¿|GóœfL‘ª@¯Rá쌸{ UCÚa3›i$EÎæÇݱ5kDž4ÿ¦:UÔcÑ[ –LsVa>l„ÅH›s ïø‡³¼-f"¹͘~èUüŽ…½é}ó¿9w¥úÌ{´g}KjPÇ'Êl“FͲj–‘nN ï§Ráýš*2Ãnæ6U…(R£Á.',!´ó·FQK^Μ7ONøueZÏù»Ÿ‡¸z¼² ¨/?:7V åžÅë«Ð ~ô¿×,~þ?†ýwNðaóþÁ#,üîÚ³³W‰ 3éµ1kŸÎŸgþml„AgÑØC‰ÜÈ™.¡ÆˆÊ•ÝñWá*¼ÛÖÓ}póy¸Wâ&Ôžjé= W(¼Ï–håúlvó ‹ÑÙ®wYw€GK¡Ð*!,Š[¸!5ÙgL*w —ÌÄGMÐ¥\“{w‚%•Ûf\•ýŠ86vÝ’„Õ@Í;u™QóÈê¿yÍ$†Mtª-u%9S”Ìß?£™ßuö«ú¯. Pyp©?˜ö4~ðê™ó o£‘¢ˆþ4æpRÐõ.tS~_M[“¯¤¸ŠöÉ|%¥@Õ–Ö”ù M vfƒ« I¤Iعr]ãšÀ Æ#¬ÝÖ´RKãðËÓ×îêzï\ŠBãÛ tãsö£U¡jNØÏ¯¯«­ ‰(OØRÜ5@Õ€@M¢).7Q_ÏHÜ•POùkû'eí î¤]*£Þ”ÇëñV®Ø½˜X3Þ’ù;Û³Ì,•œ²¶Qoô®øEœ¼UjÕš5ZVÇêébIv< ÔŠK/·6;ÓÄ´x„Õö*¢œµ`6Th‹ÔzCÉ•Dó,„ Õ„ ƒ˜ŒÈkçÁ"Ô$´T˜§!øªÿ­µüÎë]Ÿ!õ»‚` úL¨8»ëä\XoÎû‡?¨§C*µ¼S()¿C£ÇÑÌé_á±ø‰møq±=·©¥ª©º- éîk—­@]½´ c5sg3Xo ñ0øÖY_‹"÷ö1½‡šÏ’"òxxÞÂe8OgBÅ’ÇÐ§Ç ®Zð«Â¡–P?Š¡ÅÑŒÆÕ#ž/¦9ãYa$hOù~iï‚§'=…ÃÓôŽ.*¼á0[|&¿áÎv羫‡ý :šÓ__›Ž‡glf’ßÜ’ú~©¬*È;Å'¢=Rˆæ„yÚ¡µ&òÜ–±°ckr‚x h¨AÛf»•>t½7> ›‚¯édš\Љä¶B‹¥¡4šwˆÿþû)þ€»”,{wM)ƒóÃð3X€gáûgž^ôé`Jƒ¦—ÃÁÔn‹ñß…›cÚ¯4¹ÿØvšéºØç%ì#ûÒÔ­y³’62±Që2cC;N"/¸ÚŒ®úù]PëApK^"æ Ìàé¸ K§5bÁ§¤0SQ'’и?&,?ï»Ñy¹þPÇIÂ釬ËVoVmÌžÌ(S„òm¹/Æ-L\·y'‰–¹ŸwYOÕý• Ïñº\”æÃÑ®û£®ÜOþ’oôÈZëèÑ<Þÿ:¬kendstream endobj 346 0 obj << /Filter /FlateDecode /Length 254 >> stream xœ]‘=nÃ0 FwB7°Dÿ¥@À%]2´(Ú^@–èÀCdAq†Þ¾$ÝE‡gàYúdñss:?Ÿó²Ùæ­®ñƒ6;/9Uº­÷ÉNtY²ñ`Ó·Óg¼†bšÓK(Ÿ_…,o y÷×p¥æ½óN_ù=×D·"Õ/dŽÎáqžÑPNÿ–|¿'¦ùÏVÀ²î¸4²z ÀàE;TX;ÑÖ^ô€ @YArÊ“œÜò©€;ˆ¨¸¶Uå\«Y¾Œ9¼"87N¢#*¬I4 Â*&T@®1ò4ã>QÐ ³JÒë£FïµRÞ´|-W:]2ýþŸ²IYÆ|#^€endstream endobj 347 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 1558 >> stream xœu“klÛÖÇIÉ–ØÔQÚ¡*¤6%´XZ,s7kãÁXÚyÍ$¶&m;N£È´­ZG’­ÈŽ-ZÔûèE=()¶*¹-­·‰Ö´†óµ(Ztù²eý²؇®A¶}½”)»––!Ã0——Ĺ¼ç÷ÿŸ{H¢MA$©îí;r|qóõÛÒ㤴]!=¡„£k½u;t(¡£í½íjÍ·Ð?F·¶¡ë[IÒ1ˆýÑOoÝþª×1îuZFF݆§ÍÏötw?¿ËÐe4v^´±N‹Ùd7ô™Ü£¬ÍäÆVÇٺ½†§{FÝîñïïÞíñx:M6W§Ã9òƒgv<÷¨áeÖÅ:'Ù!ÇÝmø©ÉÆZœ­©×aŸp³NCŸcˆuÚO›œC§­vvçpØ]V“kÔät:ÜõȸnAÆH NƒÄIâñ 2N&È$™"Ó¤@fÈ,ñ6–h#sä#dyKѧp*¾Tö*ÛÚÚδÝn7¶§¥ˆFŠÈÉêÚ‘ ‰t5´»¦D/ÔOiç¡ìñÀÄ9z£K= “• ”çiôÚÚ÷p¤Š#ÞizcŽxqdñ<­‘­Õµmx F +ë[êÝÚl ¤2ÁâŒ/œ ÐWÖ×ùˆBDÏg|ÅBR(eèVriømÉ‚¶þ^‰–¥u-ÔÞý‘™t<•yàv½£å³j‚1ó£v6: ÅÂaœùØáÕ±?ÉáýG(¹_ &[e/Û¾à.%fAÌã…qæpªiöz$‡Ù–)4¬†~ali²jÏ;²~ky>¦4è½{~<ˆÍx¶~R[nšáõÒ{ÕÞ¦ärùúhU²T[èVù-˜Ï’É@ âDÈýý›«ŠÓ£_†¥Ú¯.^¸”XŒ·¤Aê`j`™»[LýX¸d~ÿ@Þ”Ÿ®)ù,|u¬x,éo)ÄèƒÙ©¥±êä…¹_òÝî‘Aø9¥©—sU©“ï”»êæ6¬iìo”ÒGÿåkª€Cê»hPȨúO@îRÉûåíò£ò^y/Â3ÚÏ ®ûâF•¼dõ]zŒú_¯ž—¾ÖŠél Íê†üáh˜n\“KS|$¼üB ”®ÉE]1€›7X.'æÂÂT–iTP¬]žÿçÖ7Ž6›ëõ*>¡ëwµò$}`>ÐÇ ¿«W7;‡?#T(Ì ¥¬H éó¢˜EÇ/é2Ói\Äó2%©í]¦)HPy>ïã¡i¾yõ7‘æf« Ãè¯ÚÅ`žñ³1:üƒG_‚Ã0p¥ÿʱ•×V~ ”LþLVÇŠ“•$“BŠÛ~ ŸÁê©«§®Þ¼…Ý%?AНtc¢*kÈU!Aú­RZ­?¥Movnº™6èãÂô>y¼]þj•B¹ÙÍΛ›‹BT˜aPbýÉôLÚ_=¦SyJî’ŒZôŽ —Kj׸ޮï_TÉ‹êÚ–Úƒô–¶G‡}Ôæ;:ðØJÿ}rsendstream endobj 348 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 502 >> stream xœcd`ab`ddd÷ vò541ÕH3þaú!ËÜÝý3ø§#kc7s7ËŠÉBßÿó``fdÌ/­rÎ/¨,ÊLÏ(QÐHÖT0´´4×Q020°TpÌM-ÊLNÌSðM,ÉHÍM,rr‚ó“3SK*4l2JJ ¬ôõËËËõs‹õò‹Òí4uÊ3K2‚R‹S‹ÊRSÜòóJüsS ÎÓƒP®¥9©E ŒA Æ L@0°0²þèàû™Ö½îûêuß®cü²îźïë>®cþ¾àû9Ñõlßµ~×L­^œ2­FrR}_÷¢nŽïrlsguÏÜ:©©W>mrÔÙâÕ-S#Ïz|gx2ûáÆîïÜßå«ïþÍ%÷û[euweÕ¼îÉG&´DÊÿédlk j®è®ªœÝ=Wîû2°Isªº‚Ú'ï‘ÿú}±èò }º·wïn=½eKÓ»œnµn¯Ü• Žß¡ßkEt¯9ù`kÆo®ß¢9¿ä~Ëü>':yn÷òî)?zÙ»§v/kžËñ­®º»¼»†ãÏDöîÆî‚É•Ÿ¿¯ÔÓß3¹{V÷ÂÞY½³Š¿3üv»gÁ¡ÅÆW>ÿ§ã<¶ßIÓÙ7smæ–ãb1ŸÏùnó20³•ÍŽendstream endobj 349 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 3160 >> stream xœ—{t÷•Ç%b£”´êÚ‡dF4…š,!$Û”¤¤ Yr²Æ@°qlÀò?ô°ž¶lIÖ[£«§õöC/Û’m âacŒ±I‚!'Þ4 M“&ïRº›žnö7îГÞ]º›³§íÑÑé™û½÷÷¹÷~‡ÍZ¼ˆÅf³¹Û _Ù¼ðkµšM=´ˆz˜ã ëÿP6¿w äp gñàC<ûƒèÍUÈÿjýk9›-–mà EºÛ$ÒY]í!… ¿r½à‰-[6o<¹iÓÁ ¢jY]¥P,(*U‹„ æO£ HRYW­hä?wH¡>óøãjµz£P$ß(‘Õn]¿A ®Sì©–WËTÕU‚í±B°S(ªÜÕ¸ñîu›D$U*ªe‚BIUµL,l”VT+„µB‘HØ(UT ¥u ¡’‰ª²X¬ïŠ¥Mr…ºyGÍ¡ºâFѳ+rVòózÅÚÅz•µ‡UÌÚË*e•±c½Ìz…õ¬§X…¬¬•lØN¶‹íf{Ø^Öß0¥b-fg·-,á<Æùýâþ%Ï.¹Á -]»ôSlײƒË:—?±ü«Êç(r.¬|”²óæK ŠÈ°©ÍW\W9ˆw™oÓYd`Ç”=†®ž®HÀ‹{þÃÞ„7÷õù’¾¤7qùF2'Ïœ»Øgió®ršÓ*$Ú%6#h°’±ú©CxgDv­ÃÒnÃÏ–WT¦²E)o¤ç4ù Îc··v—<]TZ)ÂÍìNV@9(u evX1]ÜQ§;Ýw^Œ=X‡RÚl¬1K ³ÔPaUb<êº$‹òï¥Ðt“sæñÍÒÄÄÒiAÕ°y‰~ö¤á· ÃdÎ#;Õ‚^)…°·£ßÓI„ߌ¾ «‡ªÞ ñÐjæ³-{fr{a‰P"%4ÓET¼©Uª®kÝ–û¥]ˆMþ—´vRç°×tü ×';vâˆøå åZÑ«„lß¾×K{©åzÐãêpzÿá—nr(µ‰ïô8}àÅ|6¿Ñ ¬kPã†xýx`ôwèïÒkèï?úÞöO§ÏdcIÂao¤´i´ÅG¡ÓkAiñš‚Ä0óa2µ ýïLÝàòväæ[5@êZ«ö¿!­¬¦¶ÿÃ7ТÔIbü£©Î `?m5XI‹ÃÆhÌ2cs葲6Úx=ÏèÜFåð›,­*PcЏ1<î›zo[r'½IB/m¬IÈ£z¼_Ýcû@Ÿ°Íè¡ 4íÕBY•rè¡öÃÞdd°¾–°J,VÜ5Ssí8‚<1·®§¦7ºd‹«¦£%IHõ¤“Ñt÷QW7$àœ=YÇø6†×•Yôbœý/×8ÈAüßqcHFu !6p¡|§×åO^6ª–µišÌ¸Y™8H¿ˆ­å6ë@© Cq‹»›n·ÈŒµ ËÛ=Y;8—‘ ²À«ÞAjL¸dGiQ1ë6’ z® Á›Ÿ½×(èûW©¿çP)t…ß‘O0o kš[ô n7„f/ÙV¤+&M±Gmj–‹t­jM`U¥ÑS™«ñ)¢{Ì€nl¢âØÞuè‡4÷xò†ÜxßǧNœ¬Ûß.W•V±‡^jhÄèåÿÑöVì,ÌžÆSÞV^‚Ïaìz|äÔ¯ÒhŒc×Ê=æ&‰´I4uXð^uD ˜NHýígàJwÚÓÆyó« ™ÑT0Ͱ'n£[¿á ×W|C]ýÞR°ƒTÑ‘ˆ; >,©éTHÄòš²óгã#£}xoÉÛ-Œ0´ôZ•¹×ÙV½PmßõZ`•ªôñþÞ/2ˆCįŸî ABÛ`ÔK Sv·¥RñÑóÝ•e•õÕJ¼edÿ s½¸€~@´}Äé‰x‰¾—Ol8ÝòºZ_f¬cêÞ Y ³©MÔoù¬;8ØP’NG¤U!%rBæ T uª˜«Ü­ÌÒv÷œãÎ=D‚t‘`¬¶6:ïN&×XGš¶¦Ñªm ¸}àtúcÁô5T~ÌM†Úô¤µÕ‚Wo8Hsè…&\Âmt8Nð2ƒG½E—gÂsTm­™ gØWn£/ns¨¿epü-7~Çb#7…¾í÷x“ÉD/8! 1}…Î^ËL…’¤SÇÇ÷õ•Ü'+“âêÓaÏqÕ ˜FaW¹ËúŽ\`Ujž ‡æð}'$)ªœišœ{?Ƴ³)ô€+Œñæ7ÝeUL°©UW\LŸ¼?¿ït¹#àļ¶€^ý´‰”á&‡Ö*Lc³ÉLZH#N¯¹sÙÚ‹Ìk k’¡˜7Âݾ@Ö›H}žëî _ø‚3~øðÙÉé€wï—ÿóÿ`EÕÿÿУ5šˆ8ú¼%•5åL‡çßÏÿg6Züáк>¿„ߥOJ¤â&±&¢J¤3iœ.¿ãev•UÏ|ãdõÛßPm$S±@’¦CéP 7û…'ìºÃÄÎT/Îßÿ'Ý×ÿéXö`]“Tai“2üýˆáïËSlJCýŽß1àr']^çQˆÃ%è‡>Ói­K µÐ :ókd{®¡Ãìtº\x-‰ê¡¿CseL9ͪôßç:lL…¬˜Íe茸¼1~ýp8ˆ66Êt:ŒŽF :÷NÊXc7|òj ´»Ú;3ƒnÈtf ÃŒøŒ#í$Ó€ù;Ø ÐØ¬v:›‰:7é0t€ËÅxè†Ìü² öe†€£ßD€ù¿ ÝGÀ·îíG;ò ¦'zý§ûñıèEð`w LZZl¸®¬¹ 0©¹¯—Y3îÁûyoBÎ×/X jûM¾Um¬c²ý­Ä)pŇþ2”¾ÑJÜÒP¼»Ahú䥮RxŠJE•÷» ìϵ ôwrÿ7$³ë´EÍÑ% ôZÊFµªíâC±¡àå›Íng›ýf_ô“Њwågö×Ë›RMG{»ºÀÉŒiͽR¾û1ºzC¡5ü¹[ÇP„°ˆ¬Z°+ÍxË+õ;…{ä,hûןÌü€R?`|ÂôàñÙ-›ûxôÔy"õâ”~Ñ•=›3·êH›…$´Õ=hAVôöã:þúgd¬앚±Ù(bŸû€gºû(|©­þB_9ù’·„™+«2z*ƒ¾Ì„2ˆÌ°¿^ôÂ$ë çëEÍ_³¨ËüÏžŸú;úáuô¦u2œ~âËIsÐꃼø.ß,Šç&öO”ßdvˆ=ù%zì¢ôÒaâìÐ{Óp£÷ÑÏñÛìf›˜i&¹Óì ú2Î0t;¢6 å¢|(iúqá¡f_s³9•‘µ²!ÛÔ‘Ao`ˆTuUÀAŒ^›O³hÞÖâ3¿:ƒ®þ>ŠõÞú }awq¤a'3¹oÓõ êÜ/ØTð­êÞ »¡´AÒ´³X_À¶>²°Ž\^:ÁŸ¤¹ÓôÚ.&ïaÆâÜX1¦v7é"ô!CÊ íÔÏrÍ=êÓk`ÁêiõA‡×A$ì1{ß´‚Î¥'è˜Þ`2&O7DcáˆÛ/,új–M=<ÍA3èk>ZÁmÛ¥ÜÏA¥WêSv•ElL¼To¢×Oz,A¢ùXË4übp2ú!†Vq{À!4èîb<Î@hð;+¹ 16)ªö‚q¡PžÓðáS?}—ñkØ%n#½h/½\´Ë¢3¨ì6‡´k ·¦N†NÿS/ΣRúB³lô›E;¹¡Q8™ažMÐÿÈCùÚë⦚0ýɳ/A °Ï¹ùôϬ­Ì´äIû¬þãÙ(“&5bÈ ™,šÉ0cM1Ýâ§Zùÿ^‰¯§¥H'èzÅøú9´=‰ú‘§?cÞ¦Pã§·.64§¹Êôªj3a¬åKh¶b¯h_ãk•o¬Øøó·ÞÝ#xê8µ- #iœK——^~s¾|ñæxβÞœë?VAÝendstream endobj 350 0 obj << /Filter /FlateDecode /Length 188 >> stream xœ]PK Üs nÀ§šÆ¤aS7]hŒz †…@(]x{¶Æ¸’afxo ýpœM˜Ü¢WHØX§#Ì~‰ ð“uˆq¬­J«§zÉ€H‘áù€³Ìʯòä~ m½bkHy s ¢t ŽRÑ#8ý'×Àh~œœ3.2mDE¦m¦œ‰ §¢6ÙXÀiSiV ²ê¬ýÕ2¶Ø÷Åj‰\ª-k‹²¼uðýˆàCIá ôy}_yendstream endobj 351 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 785 >> stream xœM]HSaÇß·é:ÕZŒ4×Î $,Egµ¾ÃЍì¢"ꪆÝ!·³Î9:÷áÆ¦µ¶w§éT¦s:Ó¥ôEI_D‹¢ »© è‹Š,°‹ â=útòª‹çáùøÃïÿ<äÍBmMíáãþU¥J TŒó”䟽¯<ÍG: Òå]œI-År!-ÁÞ0B‡Šõ]ªánk°ILYÝj¦ÚbÙPÁ¬5›-ÌN;+puVSk•l¬Ý*©M#s˜¯ãXÉÍ”m±I’sSU•Ë媴ÚÅJ^hضº‚qq’9ÄŠ¬ÐÌžböð‰9hµ³Ìœ½Ê¹\ÃÛM+0µü)VpXwÍ9°Šê”O«,ç¨çœäv œ䃕šj3£0 ŒAhÕ»AXp#|ÿ(çõ³¸91˜Ù¨™ñáfCÿÉJÈCÿþªõŠH†QšÖó„WC¸$‡8[¡ÊÈÒ*Y†¢”¯õ \@éSg¶âÑ“gϦîôvÊñNÔzÎÊþ8)½ìGó§p²µ"ª5ˆ|‘ž½¹=ô07Æ_©?÷þã(“ׇ”åˆû^hð;,&oÜìË&»ºãhJûSîö LÄhNá2y ’µ£â@K[S(niÙ“¨‘%R2Uì<!„(·ßçq%Ïu‡i\~ÀÒPß*Š[ÚBaÔByz} u'MØ8å%ëó›Û’*ébwør´ÇI·¿‰ cÝæ¢K.U…¨ÞT:èÅç¾WÃ÷c³kÇ4Ê[‚(RP½’èÈ‚àBLÃÚÇ7èǵýžqçóž'ˆÂŸa^°RHè5D{´¾Zo¸õòãw¬GÔÏ׿òò­;ª*÷¿ÃT9m9i¨Û·nQ7¥»¦¦§ß¿ýòub;¡¦i=ž¼=¤œÉÀ8®Áw•FCW:Ö…Ô~¤‘ 4zÏ™HÑïÝÑHØ¢Å+Á®DLN%Lzqdf{VKèäüÜÂÜ¢\F§Sc10)`endstream endobj 352 0 obj << /Filter /FlateDecode /Length 220 >> stream xœ]1Â0 E÷œ"7hŠ„²ÀÂBÀÒÄAH£ÐÜÛ) ?Ò‹í/û7ûãá˜â(›KÜ Fbò^ÃTÈ1 µ>ºq&~ÝÓfÑìO6ßß$6@¨|¶Oh®ZiþRuÈ ^Ù:(6=@ìÚÖìB0’ÿ+©¶Nôan]b+)tÐDeX¡ kÂ…aa—†…Õ-¡6,Ä–peXˆd¥ÑUWgOˆ®º:w„誫³#DW]7¼÷wA:ÂøÞ.ÝT ¤‘ãD(ˆ˜àj2MI”ø«Þtendstream endobj 353 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 2480 >> stream xœVkT\Õ¾‡™‚Ô<¦&ö23Á4&-+æE I´†TœÖ`0b¢i¤fx(Ïax¿ßÌÝw†7„W AET¬±ˆD‰1 mRu¹ò0]˜6MbVQ±Ùw87Ú3d-×êgß½Ï>ßþö·ÏÂyzp„ÍÖÐGŸY½Ê½¼_þ‘ÙOÂ÷©‚× ïy |<{Ë|Wæ<׿Ÿ¹üîæ l#—Ì¥se\çäú¸!î8w†»È]#*²€èI "F²“D“TRL¤¼LÞ&’QrŽ\"ä9[“S²-ñ±qVÃòè†ÕAAÖ¬ZdØ’h¶ÄGG&B#­qæÄH+3 O%GÇ›­Ù†å›ã¬Ö”<™™¹221me²%öá†Ìxkœa‡9ÍlÉ0› !ÉIVщfÃtq+§ÿmMNLI·š-†Ðd“Ù’e‰Œ~ÁlM0ÇX£âcgXÁ³,æ»iN¾366%ÒbNºi¸}æÙYnn»xgΨ;²L­gžfþ‰]³§#oZ·]î}ié‰n.““Lñi) ‘Ù·l«9Ë:³Ú™†5eÖqÖ™–5eÆÖÛksÖíõ˜é¼·–7"¦à8ÎÏ#na¼¿!é¾ùÚ»çÝ«Z’¼à™°uk×nÙºú‘`&L‘‘ˆé¨šÔZRGêIi$M¤™ì%-¤•é«t}¤“8Iy‘ì'Ýäéátn¡zr˸ßsILªo“-äßs=Š<ºUþªfÏ-ž&Ïq¯võ?ÊOÌÙ.WùÊëšÝ?9ÑO0¢_…ÌÒbx? ˜œÀ€~®ö•×’Aýç‚*Ì`¦3¨‘}CãÍd1û…k#Ãÿí0Aǰ ÌÔbÛ05±oh¦m³óìfyv ®0-5îÁLA£¡™{ÐÈbβÁæn!øÒˆ _b¦Ghû†#tPíëz…}t s‹®ª\6áû\-8À^^÷kä7áÂ0ÔÓ{ð£]8 r0ÈÊÞ¸ÏÅ_ ÷?‘³;ìv¨æË /Ï6›NC·?D=h0×Ò:”FKé ‚â= ÌÇuôÊßiÀê÷ŸûJ%±øÖ:hÔûºV²’ÞÔÇ”w¿PÉÖháh)~)áâÚ÷w5ƒžè8åh¶w@ /ïÔÀùã_îo­“$hå÷‰RVÒ#1 Ѻø”O‰‰›¤°‰§«4°žßZ˜Ÿ”ò|A¹(BäJbgÉw®‡m¼²N{`ûЮ®tgÆ»–没”à÷IÒ¾p”8ôaÝkr.Ùš ^äg‘jc¤2Âú´´ …¹‡è`(0R­¬cNaÎzV‰é¼Ê&\ûT Õ6»Íóºñ„9ˆMÁxd-`WŽcAÿ/¿ßvÎôÞŽ¿†Ä]—_Å¥ð!/¦À…¸óëújÖƒVÞ R:ëAaªnùî›3¨oU&”€‘W¢4»Ä¢ÿ'VQõ¶Ü&ª\ȑĖÒ1_oí¤óª­`ƒ"^¹¨úËñ¼;¬«ˆ¥Þ)AW›(•Öè£:©öo¸ÔÖMpŒŸyy³Ëãy÷å!xöQÖ‚€&h@´fæ´ÊWUòaar›dï=ò|º¯üÐoùDI½(åŸ_ê&±Â^¥ÿšògé¨ï¡¦“ ƒòðnÅ›.§çž¤ ~C½×PN´UUºé¬‡–V;Øí:I9v=0Å 8ðíÇÀzÜø ?=­7šycZmZ¸¯ij>Ø{ò•½µî±d¥ÙmÕ¢$BN½ÿà³ÿ€Ox™ÍÛ{æpzW†óÝÅ5;_N>Ë&f@eC¥þÍ\ä":–Ú‹ ¬<ýNC~còÁUvO¥šÝ;ŒÙVÚ²¡‚WذqÛ¶_E敹ǚϲ‹ûz>ï{å/ºÞž÷H'¥³â98ËÏH9Œ‰`¿@³o~«e­ÃÙ/šßŠê+l¶A>ð…·¤÷Zì ã 6´é3z-ЮX:NƒèÐfê¿aCd¤­ê¦›[Xw[ux~/G€Äû ^ÝÂdÕW‚ç·ïÁó;¶Öâ]ø™—EmºoÉcT“˜Q±"•†¯Œ«åZùr]½Ý5‹ÝÓ_֨ĸæ-’ßQã½p1¿ÁöAéW9gžíÎ=Ñ»‘ïVÓ»°×K Qj+ˆbU¹XYP$¸·HiU&Š ¡ÄÅE Сóìÿ!oxÒ¯ŸÈAÊA-EåñÒbÑ•Ón×)õÚ$Ë#Æ'Ö'åV„¤Ñû¢‡Ô8.o©k`ˆªë_y=-ïX$Ô¨Ì);E·3Xàuý”:EnÜߣ‚ú5¦ˆ3£*<3õ´ÒvA=¢žò˜bzGoù Z?Œí£tL=kÿ~æß/¸b´´=ÇØ“vÇþ~÷~öÜ…`;óÿ8­ÈK¬É ¨RÉgßÔŠMå|¸:½ò—Š´$•ÚõO+-t1¤ÉôXY S›[ò­mS’¯và\ö >‡‹°ý<y]æh,f2úQ¹ê”ÜÊŸ£= J)ºp7©zåcMª$öè†Õ=’t@?…çç³ð.#SR‚“ ß<†ªÂº £ò:*k˜à$QÊ>M-ô7[ß(9§•«Òõ >¿ºXb'Ô‚ÔÄÏ,±Ôøé{Ú (±ˆ7i@A$tòಓԂ*Z€›hTkvu»ÚÙ¶ßÆÞ VlyƒžŽÉQ÷Ëï‹® ¾¥Òu·+œJ+/û¿+œŽ÷f¶¶«Zôvדçä3øœ_×À=õ^äæNý@¨¬¢ è²eô9ºˆ¶o Gb®*!ƒšØ³ÁÐKEµàÞ7³SÍ.ÀNÍ€÷À\·gv²Ïœþ:öwÇý«endstream endobj 354 0 obj << /Filter /FlateDecode /Length 8326 >> stream xœÍ=]\·uEû¦š}޾͢Ú)?/ÉypŠ4I‘H-4šË«UbiGöJ²•_ßsÉ{Ïá%ï̬¥ ðƒÆwùÍóýÅo®Ô^_)ü¯üûüõuu÷ä›'š¾^•ž¿¾úé³'ÿòŸ.^¥}šÌtõìÅ“ÜC_E Ÿ¦« a௞½~ò?»/¯MØ;¥ÃîÃõ3f•ÙòO¥ìîõõÚ[¥tò»Ã[üŸ¤”~÷-þžRšlÜýGh¤ÍîûÜ7D½;¾Àßv?il“F’ÝÝckŸ’žüî†OÉ9»þ޵/ËÑÎìn¯o4tõ H{0eˆ˜’øÈÿûì?àHâÄĸ} p*Ͼ‚Cø¶5)¦›ŠÓK{•´2µíSØ“s°[½ûŽº)•\ؽ̋„ý¥zÚÚ°{~m&ø>il3AGjRŒ‘Ö‰Ÿ®{ŽN[ÜœÂk‰Éí¾a­é,&¸˜$uýªÌ©§zºï¥\/\Ë—×7&î}ŒûÑB”‡E/·r¤±ƒRq²Ãy^±µØL·l ²ŸäÁñm°žì<{×=ü®Wòß´’Ò»?m÷ÆéÚöpÏ.Š-éû|ŸÙÀ6ëQ(.GáÃT ´Ë-kô*_¾ 1¬b}-t:MÊMn¨ëùG¡ÌÞÚù,Þmž…‡[mkãý2!6ÀÖˆ\W7Zï“÷:·z–—˜‚1y-þøÝ>äÍ©iŠøÝzEpô†a.?¯ã»IûŒ¨+ÇzN 8Œðö0­u qçðŠ‘ GÄ?XAÄ“ÆÆΟ‘ ¹!\¸…^¯g‚€ÈKUʼ¶Þæ_å~Q…Ép:˜š,@Ç t8@o;çGÀÁÇ÷Í(ß5iÀ ŸÔTÉ:ÄeuÚ `jÙw”ˆ¿l—í|€¯ü3ï ƒ e°0\~¾ì7ß uÍB;H!¯¯486˜Ü’ßøgÞ·ƒSh‚ö3Nýl!Wœ ûdbš Ìïwó”‘í>@ý!¶&sàlœ±5!:B¢>X(F™ Š ¬Òu%ºüQây³&`ÖS3C±¤¸7PøÔ’\»¤ “\Òï¯áV•ßG@Ÿ y¨ëLˆ ¢hB„ô×N („"'¹‰pçM¸é'#*ƒhˆ!Q„‹1ÏYóÞ”6(€ àÿ-Š—ùHÑpÚ¨õåè'à>::÷ GE\ø¢Íù¸â/Ààk]°LÞõÀž0%J¨< ûÚøS“¬`¿„@ßF}à¤ÝN7öpà ¿ÄAUŒ0$Ž €¶»ãLƒ](‡.£0Îÿ‰*KÎøÑBcÎxØ@i¯‰g@ƒôˆg0ɦ'“X öþ*7º»IA–ŽM×] ¨[LR€lçxwäBßID±@ešéQÊV4Ɉؽ·ÄîQÄ´ÞÁ`A 7Çîì·XàZßêÓ œ !‚+Pœh-];÷“¦½ >Öûù岘Îý8 +°¶r?ÚöÓ> /$¿v!ò çÆ¢È6XY3Æø—Dj†Ç±fÊQžÆ%ˆõx2 ¿!FÛ  {\tA‚'×.D˜´SrD˜4£<™®MU–ZØB…ò³õà€¿Ç¹ŸüéÀI#s=ñ,2Úx¶ÀbÔÉë: ½ÎŽ[ M%õpÞãÒVˆö\dåêAA†)êVªEÉ]Azœ¶Ý·G¼\ÈLÁîʲT!5åóJË-¿‹‰ÌõP84ò1Ý(h[@,#²AAM˜:]¥XÊOBãß9Eä®X} ùhòÓ##¬Èq9Þê³ _ˆ³ø}>^µ˜¾_,FÏÖ˜B,à6ïÀ)¹ÝV÷\Zϧ€ ¸mHB%d‘CŒŒ*C‘Ãî‹EÂ86mh¹pSn‹_Í'Rmi?©QmñÞulðÈõÿ÷Œu·vG:‡1̤¢°Ÿ"ÞÝN_?û㓟={ò[¸XøôÝuõó'Ê ¤ðí¿zýÄDXMýòêÉCˬ”Še±CãVïg‘eö™¤(e«sìa²¼¼és ~«á¸¦ÈÊ=T@°ÏvâÇèsÚ¦d‚}”þûYÃ×2 š,ìH-ÜÉ„C“ØŒ?CtEcl÷aà8ü_[^+æ¹1hý¦jUd»ß1jT15„ÐP 肔ό-J3T6öBCc…\÷P?k 2À¡‰ÀGË ‹äxU~3«ðHÑÝ%M!õ7 óÑÈ+lLEוL½¿¶`ôÚ'Ùú'Ò2€ž®{Ç6ù?:èVÎÆÀÈ«½Ü’*^Púû¼8Ô¥Ž/À’Ðá3¦ L¦ (=…I€ÛE=Ô%Ø 5î¿òÆPC¸Ø¾Î/û Q\ P¶c+|¾@’[™$ë ~Î&ýõà÷¾œ'Ð\ÚÐ’d\U]Fæh Ú¦±ŸZF_™ü[¡ÏÏ2ºEWWßp60Ì=.ÐƳŒäƒ¹G—°n KÄ—¸àÄIÒ±JNÁ8!'p¨yŠ q ^yw›± ­ï Þ.(È!TaD]0 Âñ§ÜÜ¿²•MŠHȯ¾ÀßzV¹²Ü½,ÍSàÌù˜„ÑÏ'™d¼Y¾äÒ£aJ¸«ù·ú^²Ù$û½BÚ‡dÏ㥇–ì-ÂÇa$oßæ‰~½ÃŽÉø&½³þæÎ#w 0-:öñ<é‘N1¥¡Z_ÀBOÜãx9ݸ5ÿÀ¡ñCÙ²2’ö5´ÛÍ.lG[g¸š-:ölCA+Šw+!Ü©E’8´ŒÄÀòÏ©hñWýZþ­í†ÚKù‹1º‰‰Q TFŽ-hüûu´kÖëÀy¦ ³= ŒSö_•-T!7æpv$&,=6¼™»skÇw׳÷³a`4yL-Š[`çsmûž8G >y¶ej×®Çð1òÆÍY¬îinÕÐàÉNrƒ‘«Ë ÍèV‰lÈhÀ 5›vêß}¾§-®.jä[©_Œ¦”Íóæ_²aÞ·Z. ÜÔÍ7Ž›MòMAÿ[‹}Šˆ­SZNµÏh’pÍ4Q%Ú¤ìÀtÕø"‹—Æâ37úŒÍÔ𾊤ògx*¢“rôi›ÃAZÿ~ŽiZY™hB[=2Lš[Drø2tìÝ[ïàúà?½—ó«åF³•g¡Ã+ÁMû‹ÉÑ M™s—Óëå«®ïGË忯~f )LÉ7k"=ÚÃÅëØY“ÛŒj׬éúlñö'üÈ~Ô™ oÁÎæ¤–L lˆ³a)aàÛ Ž@¥á‰ðY2ð¹VŠËð›Fð»!Qáxɉ¨5c`@Â’IöSÄæäEÈÍÈ ÊCtßæ1ƒò‚<k”tãÄÜKeQ—ðÜãö ”4D²º‡H¤'û&µœ»Æ®"²¢½^Ò$Q#"p¾ÍíƒAs«®©ZàÀYU”F¸Å+6Aƒ6A&h¬|Vk9ž_Âñ†(ÉÅ…iÃÙ8”bðŒ Ñ#<6kïN¯ê!o5‡üç„°Ôßõ€oËZÛàºÓ†“›ó4‹2‚sA@€˜‰‰Ô%ðcFë÷zD¼ÏßsD‘4ôÒ¥®|6ÒCØVÇÐYƒ£•Â1Háoî¹åа9$õp]D ÚdF¢Éæ½®$r`ã5þ&3¹ö’Èñ’ÂÔJä•“ùØíáFu£Oüt0—°Õ$ñ!·@Ù†ÈίÒå1*Î’PvVŒ "·.®äÐNÄYKŸœ'æÕ2,ÜBGàr†ØÛ!ÿTz[ËÏ„rø²nå:­tožç!¤Ö…‹ÐrØ8Š|'^è‚ú.§è}KGf¬åÖÄwƒaæ¥5kvÈ[Œ`¯‡|ÿ)¸•wq¨`û®0½å ãº;Î\cµºÇÏØ9ÑQaùø@¼Ç<—›ItE+ü>i¡1ðØHö›£‡2 3AM·ƒŠ¿8ƒ'LCGékûžs‹¾)¬HŽ)¨±øü˜ #\M~ÃÖˆs´ 1 1C>äë<¿‘.á<Í\ÑsfðlD;§fòéþ¶OÝê4Ø ðn^©iã2º|îF­V,s–:ÑlmX£XÙ½D±ê"B¡S¬ôK²=Ï<~#‚ ~ÙFRàHV-Ƈl12*Ñ)c¬§ótš çä´?­ 9§ÉuæÂÛeQ 0Fz=¿ïÂ5•÷Ã`>Ìgý5~•7„öåƒdm¸,?¯ã×û!·Žã‘™_W‡¾”pÒWhš·8?nãQrc´Ê¿(ñÝf¯Õ,r òp׊³`HÚa0bîù7žÙøpL"N;™\t¬]mŠ"ÙN ’Õi2d99Ã>˜¨lQ"PQÿ(9b“\b1ó')2û áÚÅD*ÆÊ0WÎþîe¾(E6.5aÍ•Wi'±TH]µãfÀò£ýò਷éM1‘ûiQpÒ6e£Ó-}Ƈ¢Ñê=MLæõä»Ûô~äû÷Ò ¿ÏßÑOL8ä£ÜÌŸîóÀ¨Ž|ÆFqªጠÀÙ½IÉwî 0ò™èZ“r½R>ÿ±¨eeM*µÆ€®Ç&þAÞxá?X›Q@-è‡Æa˜„a‚6U¿Œ•Ãxø0l‰Ž êï/kĵ³eig¨†œ@ʼ8ï0w¼œJ'Ý=›@WÙ NêÔf˜„sKšXÇÕf‰Í‘4hfŠ@wg޳ŽoY9Êšlá28;Æ“uRPg?ì=Û(Üâë<ZIF'ÉiMM%—ôö{Ì-wdS|Éðý¶žª"æ¡n ¢r±öCÞp˜7Ÿ!¢H_A¾tPa‰’£…ŒÂÈJ6d–;añ?P¦¬ø¡ œp°Ê$"È©H&aÅЀ«ÁÐü9‚¯QÖ»»êìô³fŒÐ_å0â*´²k9á`ǰ¥‰hD-±bŒý´¬mó ´ÑoYÕq—Z8iJ‡þ38~k…Ý ó¥ `:¬M[Y÷fqDâñ5m#-= ¾²&ç2þ£I‘ˆð‘ŠÂÜq•ýKkð+ÊE3ö?\»`Ð4"Ðú}{qõ§ñ¹Öxœ"W óš'zºŒXÄÕ|tcq•  ’Óq;yô”f2Ðé¸fxlØ1v•¹n˃5âHnqÑtqYàÀ¤\Þœ¯äC6”@ ˜(„Ms¸®]² FWÔ!Ï °%¬2¯~G΂¯ò0(T#™õuÙ˜–ØÄm7—*t+$®ËäàŽée CHzQè3ŠâÊÒPxȬ789 λ©t¦¨ÓË4¸É×MÄëc¡T|.í–=´îBŽÌ•"Oè#}ǵ¶Í˜#ò‰¸<ØP±ù§%È 3Ÿœt4ªâ*Ôùe]ú28óÈn³Ãƒà™ ÄU¦jl§"ðiåƒaHM¡È«¼6 ßô( úð6wŒÑž•ò“}+‹íJúV*)ûœ/¹¸ sÇÝ¡ ìA¡J›q[PDIóð@°oò‹±t¤)k>ÐÕw\bÑñf@ÓDÙpæ}häHüî”4¤ŠSZ‘¼…)aß±;eÄçÛ&%¡?bÄTÀ¤i û»…:?Ì82irU¢lÃÝÃY¾AAR„óå¡¢ ?,~ú¶8Vàê– !ˆ'LYƒßÛ†©|±–) "͔ӆFu¢žhºzQçI28á§ /¡3›Š5U”E¢2sl¤6#­öLq›ã®Ãpq¹÷ó@ÒÖDDt¤ÃŸ¡@wjõÃdÔïCžQk•x´é"ÙÐ%q@¼j‰u“è/$mH½Øß•gG×+srôåzáÑißÔU©òùgùü»ÜÖèt‚(«æÈÐ Ì,™·'±NGã*Œa`Ó†ª\oœï~à»áÓ°3¬`£­8ü‘9n ?±‰fÉmÃB¶Êâé:ï %®1x­Êì òI È#ÊìàT ÃbiL¸% ÎE¨n‘’Úbõ« KmÞ‘tËH=¬—bÛ]“]qŽ˜e”Ï-ÍK}„¦2—ú_åÆ"Ò³ºÒ­Ìêé§$¬/%}uP/ËU¢¡%¸ÄöMdÎ| qHÓ,_XMϰi±™ŽóñD'DR¹ñÃ×¹†ªÒšÏРðÌ';’‹ñ Rc’vUÿªžŸ´gý† â%òÜH mQ‘4ÔÚ‡ð;ÚѾp®¯}YzÂÁäXÖü´6Giú8[’ÁË<°sŽ+û.·5jìžx[‡ÛÌ|‰ªåv”]‘4ˆ·9°+ª¥ìΚñqìé´‡ï1‹#™\ç+€Pwfá8\É´™ Ww³T‹`«WÉÍ|ù84)º¡¢„§10gïŽÙÇ¥Ö'€jÚTêÄðöTî9i‰·Ð§•{û"îÜàRÙ¯ärÜ6ÌÑŠ‹¥#]·…Ž®~F‘÷j úÉö]Vx®^=7"s¤%'òF _7 ö![21Ô¬°é°½¯rëIµ ÊbJSL™GntíÎ1¹ß«Ë9b(ñGªó…?kŽà"z[GÁŸÏù/¹ÅåÅ®¬§„hNç¹#˜óx®twœ+éãÀÆÿTå®,:çLV¾[—:Àu%LÄ0¹±+ÀUs·+¯¶õgV~Ê&Æ€7¬ÒÑjaø*Û¢Ì^òmS'}§’9à'pE52šd„a€ý´:[*X.Û<äiç´Ù*Å¿Íß5ŒŒ!ƒá±}ÆRx¿mJRáþǵnóªÑ1פ¸]š\I&áÏ c6Hì09kðTÌóDo®¶y:©ªèRÄ`Ìú¢ÏxÆ ù¦¥öñœMtg1 UxÞ¼£ v³Vg¡½$ÞøyÉ¢V›‘sàY,£²z©1’N­^ØÛB–v¼Ûkß:X3¥F)VÖRÞRŸ>¦ÇfC¾˜0ö|¾iÛ6´™ôlxiI•Ir¢p 0Š7Â]f]qw7´aMª8׉*)c”Mfo&7‹4oY´+ÿ( 2›Bÿûå‘£UÉûx^3<ïMS$ l·\Yx5v]·!ß°H¼"E+ÂbªÑñÕR¥ìÌ0• —Í_û@ÅëÈ€˜(;u÷XM°¹V<`ÀzúóI´ ¡Nž½±®Q +ò69¯áï—sîÕ%À,•€ WÔ”S4‘”§+k°T•PLè?¨.¯ Pq@ÉrS.¡xŸÉmtiê‰ÖØD©´71ð$|.X'qY ©Zª&‹!‚©ÖGÜ.Fgä €tÒ{úÑáÏÅ\lTj=ÝrÛæ5›½Càò¼Md¡ )|–ô—‰ÿOè/@а¸ËÅETë¼8ñeåç=ž*ÿ°nw¸Â½%¬$BŠc¹óê Fi×Ò§¸á‰Ø ¥:ÃŒ­CØ%d,ÆïSÛ]òÏñ³Ò±å]Ž|ÓM„¥«)î%˜àkä­:‹$uÚFºÅqp‡‡ÚäüÈÇ|±ôÍ÷7‹Û䘽m:6¥ÊÑŒßÖñVOÁÿ$Ñåó¹¡G¤ŠIJ}t#iŸ±‰ÓãÜíOLÎûC¶{Y¤è¾MtWW´¥ÐÇ]½úTy$ÏÌùiùÊŸ×KÒŸB7.éÛŠÔ .?¹r÷n—C$Ï,ç“}I†ç~ÇG%tëÈ”wݹ%tQã¹}~·É„áØ°ŒúRoÅ<Ǩ‚tα•"TÊÒƒoR¼ÄÑœ¯/µÕҒ̲_I_omXë`)Ñzâø`v»4n¨¤!Á‚Êö%ÒzGôIRE”®½ Rœ|„DüRĬ†ÏB±œ`Ž_ýbe8ª(ß„­Àg/5-–ÚÍ=ïÔñÙ ç»Æ\ØsÊ–h+%DãYVÖ_·œWª4¤ Gn,_ F¥Þ RçÖN™îS€%üžä‹(ÝÛšu+«ñ·®=›ë&lem¥œZ œKê*³` «#s’¨¤Á#ع±ÙòÆÙSA~o•ÿw›¤ScUðźÙQ&;Þ Y”L’£Sª¶$ãÕ LA‡Æ½8·à„‘ð¡xÀ”,Ð] Ddoà]GîÈæIìJlØ¡¯z»˜ý±õþSýüU1‚ª¬^FSÚX*âݪ\ ÁM ƈTd`€6´pp`ÅúœtMÀìNÖ—ÿ›²}zü)›bhêËÿ¤]šPº`, àTÕˆ¬ûå,{`Š.“˜Ùs—Ó6ߎò<ܾé«áÉ:P0—¦%cíŸ `-½üîÝ@Ø{)fšáh,ÆÚYËÕ›ç_ÔZ¹}`Ó‚¥k/62€¿æK#N$?~ Ð^쯌¾œ@[·Z¹}œ?\¬­Ã€ ÍBL7q ­7¯X²'|!éÓ;g½õ;9Wu¹¯NŠpqçdDÛßß3ê½-'"õ˜j™é:D¡º9Š[ª вخn `ÇLyþªEI±fÌ´3Û6|ÀƒžÉŠåm²3LKØM=HÜ™fÿ ehÐ"½R\Ü4R¬Üaf_ê»6îÿz•û¢gä1ï@£—"ê¦Ú]N‰Rå)ò¶Ï“,~ªgdèáÝÖeGö¶\ïþnG+g 0<ðµ~Ýõ“óüðAÚ@>o´Ví‰o‚§†˜\ îÀb|¸\r3—siÆqñ"ÕêŠ ^=^Ì⃴2´®önê "\F¦ÖòroóD ÇqêŸc­~×"Y=†q:Y[ ^ -Ó’ùͰËÓO›á?(&éý¾­ÁWé¬g–×ËìņŽÑ¼¬d)ÿî7ºu×çg˜ÖiIô²Œ9’²§ÍG³ãG‹i pn¸/Oii¿™Ä\TG½–Iç6³·ñØU‚gCóÑN>OÑ)Çþå „š8ë ¾7ç÷@î³y[œ¬Ec.ŸK˜œ½Bóÿlö¿@8ýaÜOzFPˆÞ\äA,óh¸AÏH´•©=´;Ë"ÇZ•cà|Hl8½Ö~©'z—Á:É‚|ê= gQ^üªõ¬1h}Ö…DñÉ·AÀ@ € ¤(÷ÿzÞÎ<Vh=2/‹N4ÅÞS\Ðg2 I í|ÁùaS êK_ž’DÙa4” W G úÀ hAàœt Óe¹ç 5«2T‹þ±{^ÉÉ-pßõåsªHÂ}.¨*ÿãmaÔ{8O×FíP…#¢ýbêºÌZ9©Í·èÙѬf^ú× ‹€4©éâ]¥—y»)u_—CÆ'}åSL' 椗úϧA:;W‰Š+*b¨‘±“©²æeKûV,—'•à<¸yœ=ŒÏM°í¸„4«ì¯Úªd™w(ü|mÄbý>ÊïbM¤”„³cüöÅà©c›#sŸ‹…údêWõDöÃòGÅ`f‡®•…8‹£0%s†£ðÓyæI#Õ•¸U?±ÖŠTòÎáöÊÎ%¤ON§€Ú€éÚÌ”3Ó2I2µë´LM}+šö}øR4ÍÀWEÓj&^Oà(%ÿs!Óðg £?.y6ÇF&NšËÄOqÓ]4(n‚TXdª.¿L™§ ùe *þ¢ø%Ï…_²8qtq¼l‘ À°X X†»‚¡@Ïv²"¦™¿<âÕW¼@Á8Ã%)íEëqÊ «]‡s„ltzyR?6£bcr7“7Žx¢|åM®EïÊ/K¶0—=Þ{ñù´LÆaã\“ý(VêÕ©âs½æSÕ'ß§Ç¿íþ¹K¢4‰©ÇÓ¦•…¹ ý@ óöëÉM$YrAb±Ý«XíÒQ¤æRôðd^­Òeèq&xËE`P¦SR¶ѧµ{`Ãù]ûHïÚ7ɵ õû“bj“GYžqèØŸWek`áÅEÛ“ÿjLQÖzx¾Å“Qs½øûï«Oß0Z £šñ:0À®óån.@é0o¶C,&Û—YWæåN4ÝC¾(Vn¶šÃpì)n”¦eƒøƒ6× ú^ñ«þa± ù8)È™šò1‡µÙ²S½êµ~§÷"Ž‚WV‘oz)‹Â[¹hºQÑÏzC5-ˆ’g=f«˜µÈéÑÔovüí†ñÁõ¼ wó¼ÌÓN*´ þ5E8ÇnÚì*k{Û»ÁÓJÑÀ‹5>çêl h¹ÒrW—¨7+ãÒù‡Ë±;Š2Ø,»Íqðšr•xÆÁÛ<f6¶‹P•£" ÉPå–NÍÕ®4Y¼ER'LlëÒãØùD{Wæ,¥og´YP˜Ó›œØ´Î³:«q4 “½™J€X бڥõß>ùTM +endstream endobj 355 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 4033 >> stream xœW TSW·¾!psU Þ:õÝK‡:"U«UœªTlU,j+ŠŠŒ!†2iØÌ &# +ÕŠ8Tªõ¯QÑ¿¾jÿVik[[µµoßöðÖz7€€«ýß{YY Nî¾ûœýí½¿ý åè@I$ÙoŸéîöÇ c$‹ÂHw“Áü‡—8KÁÙ±æÅáë]ñ® ¦ÅÃ(©D¢Ô¤.QFĨ‚ƒÔn¶¿æ6}îÜ×'»y¸»ÏuóT¨‚·û‡»yû«ƒþjqæ¶N¹=8@ã6a~Z1oÚ4­V;Õ_5U© |óµÉnÚ`uÛÚ€¨UtÀ·åÊpµÛjE€[÷é¦vÿY¢TDhÔ*7oåŽU8EQ“<ÕK"–ª¢Ôšh­¿nÛvï«v­ ^¦pŸî1cæì×çÎ[0eê4Šz™z‡z—z•ZC¥ÖQ>Ôzê=Ê—Ú@-¦6RK¨¥ÔTj5ZN½E­ VR3(/j&õ6µŠò¦VS.”+%§†S,5‚I¦ÆP5„J £&‹8RŽT õ»dޤÚaŒƒÒá¦T)½â¸À±ÔÉÃ齂>.$;¼Ì0¬8e`å ¯AçµÎ¿=8kÈëCR‡´]8ôưÃÚ]&ºœpyâêáêëªsmʇå`ýÍŠ¶;’ƒBˆmÂ6ù¨>3 Â!’ÈàNóH=¨!Í”–¶ E5”rOè²J¨©ÕB _ñ›Φ+^q ÔB˜¢Êø'ti5ÔìKß›˜Ë§ek ²Dа™ôbs"#é®M±Íêcu}hCÛù8OÎÖÁþíÜb:N%:ª#ÿ iÛҷ´Qô[«‚8~ .£åŸ<ºyôÌ9£òŽhwȺìÚí¦ô·‹§o8>Ÿ“±ü‚u:FÜ;ÍŠ_X%ŸÙ°Ê&1ŒE—©¿Gâ8uq!ò'£:üøºrÄŸÌb=½o=øîÊ•ë7.{¹Oõ^äÉ‹.’¬BUr̆W[¥‚N¸Í–@•"&}Wj:gØ †Uˆe³iƒ7+úM¼ÈÛ(þÅù¦ãÛÃQ”¬†pŸúÚ¼¨mÀ¸Ïú'âk÷~ïh=¿yž‘ÏŽÍŠ«¦Ì&ûfDeÅiV¼ÚƒWžˆ×œ‡¥ìÅ ç.ݼ°xæ¤e^ËÝ?Å‘qŽþÇ“ë€AÉÃQŽC=~~é}¿˜°`^þðÁ𺖇nw†½Ur׆e6i«€lå~¾¼S¯¡{W¸Æ÷Ñ gáÌY8‘Œâ‰¬³‰ô?ÉЩ…Œ" °† %R¾_ÖqŠ —ß!ÿU¸šýáâO8˜ËŒ*ÕVSn*9hJüxΆ·cVùðò/ŽÈjÓösÄŽW"¬Jø;¤m^ß ô~ûÁ#!ž'RúY~]oX1Ñf!Œg°€Uçh³U¹á9oÃGLu]í_‘Ù\flVlmÔ6™=‹Q»õ©n‹y èÄü½¡ž8íiì÷z^þ 9åË„÷FïÜâþÁº¢Û*.½lwQ$0A­å'Êâ@USX’ÙPÌ5ž3XÄxÔ}uŠ—?†ßkÍí©¼«Ö_M{ü6)ªšÅ9ôþ*ñüQ"ðdÎ]ú”Û]rsgv×t•ˆò}²‡žt5°ýæãe` ɈŽå2¼×îRó&œ­â»œûØá}ÖR:ü³ŽÍ 5C0ítIÔÖD‚ž÷ût¢ÏõY°ØMÓç%†Î‡ÃuVÈþjQHO²Eµsr]#4îm°Ø£Ù}oY%•B„TX,RÓ/Ž}ù±ÐDBn Þr²ë‡ž<”ÃAÞsÜÂÇNÕ=ÇëzÙC˜ÇhH&ad–Óö¾_q@w¶¢!Œÿ‡v¡NJMOÆÄg½{lïn93¢z*^dô‘,zýµ”:óh-D×S'oöoùjL?ºâåñïƒ_LpC6:¶ËþÂq½Þnè©°%¬%Òâv´©\Z‚îO`f/¹öÝ) ÊpŒù#Hæ3ƒè+©4—jüÈ¢wÈ‹3ˆócÑù#?—ö@ùŠ k¬’k-XÕ"Åã‚›˜ž.Ю¦XlÈÛU©=0ëºQ5Cß©—…F÷ r÷`¡¼™)+Ú|™øä$¨÷䦿¥§Ã>0³ó¾Æ¼âsLfg$ý·/ëeuæž<­£·Án}rdoé]íùÃg ÿyš ý1®‡ yÓá´¼ùPȶ’­cˆdÒd"'þø»ˆu ||ÐrˆÁPÇ¥²u~sÞ\úÎõŽŸ®]¹v¹y×;´`´ˆÉZá›x8>;˜·ú:¼ªS¯ê[ýÞ×ä*A™IãªÎ=û‚J2ªaTMqöQ^È«’õZ¼ÔÇUÊŒö~±‡[%çmX)èù¸Ž5€¤ßþ™¼@M/N“!Ç!‹#›¾­²§>&e.•×lXŸ¬„E°ètÜ7Lf+kl¿xþœ…6Uá«Ì³^A«WÙÞ“bž|‡ô£¤!MôíÏ¡†Û]¸£ÞTQbi}âÞxsóLî½ñ}Lü˜´³7‰"¬kúUö ŸqöÓ“˜Vôüì|+nnu=tAg§â.¬¹1Bþ_I¨Ç¥ìzÿ¦sŸ8qöÓ?xoí–-¾üô 6ÛpTÝ Ì÷í7:j’ÅZxsqeξ̼ˆŠÔ½À”VªÐÕ®OÚ”¼™+ :°˜)KÍÞZ²_ÇË •—¬QŒ† £ª0ÑÓÞ‚$fÖÓÕ8‡?mûº!æô†ƒ\B¾ùt`–Ñ))Yñ‰Æ$c† ² '˘³—ÁAä;~Ññe…uu&¾ª¤Úé@'˜µeÑÖñ=ÒU}’® #âIÖ.®VØ…‡«ÿ¦Õõÿ‹~û¨7ºKjm´bˆýëzφ+mþ6“Xùqè.4±7èïO¼ûþ&¿5îü}BÙ4°B#óÍ»d G–õ¯.±F{W8¨_‘b-?}[œ»5P§, øƒ(„–*,qG€¹Tù£u¢dg­"ƒO1ìŒUÂÉs—  ¹‚K<¤6ë‹z«ëÎ#W x´\<ßlF³ÝÒåå~õâDÚÈà~êéT_ý¼LË G‹¦\©Ôh”ÊrÅR^nážÉM׳W°äJÀ»óf<ÈZÒ¾ÒÂLÆwë»ó•›rnçüîþ8£,£4£,´G•‘]²¤<ÏXWÈSXã.‹²„¹q¿ãràljûømÇ"ò”1òÇóÊü‹ê¡iÔ™æS×qàþ)ay\¶¦(ÍÔ£Œpd·4JU¤pÛê×û‰ÒHºhþì'W~%J£œŒ‡jã;DgŽØ8Õ#º x*8öBáA÷6¾Ð9…†Xˆ Šg\8$fÅgef‹%™ÏXÔ¥ÑjMløÂŸ¡Ç~ˆÎ—9ùá׿E§»±‰íÐbŸƒ,E™S}?]8µ/÷ 4™M&;áÓîrõêc®æNý¦~ºrj_lºCySìiÝ—œ®ãìsÖ§—Q'ʺÓÂçÏÇÛUþ/tê Ûo}·õ¼ÐÿNt®Ù¿v™/Ʋ†-ÄÚ]Ffø'm²À螆¹ß}Žhñ*†oÉäÍMÆX/Žèeºg]uë9£ÿÓ$]f¤=Ûv(¹ø#ʲ`äwBA¡Ò0½×ŸÓ6üH¤p¼Àâê±8‚8Çñ]·ÕOÈñåøG£¸K)`xá@œ€ãl¿Ül¿öO&x‘ ºiEb}"ŠâtºË=w)]Zbb:0Û7IgˆËˆ6€’Q•¥ï?Иwô×9Ï]Ö×Ë9}8º÷$bUrö²I?¿À&§¤¥ƒž «‚¿§çÖñÕ²n•­Sr$™¾xæ§[Ç]j9y¾dÐyÜ-2„ œ?wöNKR©¹¢ÔRœ¾/9ŸÛ²©ö<0w¿Ø´|Ÿ27NdÄyÙûËCªô&»3cÚÞ=Ù\míu»ø»z:ÐO¡áC#U†@ÃÚÝñáŒüW{r'é_άX³zÝ*÷ío5Eò¹¹ùy`b*u&•ZêñpXç#¾þþ'×?¸¸²¦‡ˆ(«0À$9t U-Ò£¸‡MNh™ÂÔZs åÝ Wô¡çÔí¾Õ|JN-æ”t¤g$hÅËP'Ê7¶'N䉬;ú†6œ+좾ÃWŒ]ÀaøkÄ—~¸{Ž@IRnºøIƒ&ªL[]e*­iüt¢X"£É`âÇÉŸ’]è¸ø”üü/tFzFqàä‚/øÆúmcÐ4²^°£Yk5äeäò€Œ?™…§_ÉÈ7”ŠTÑÖzúŸwέðœµrí’.²4 VÉ Ûgbm‡ñ$«‚IF!õ¥¥E‡¸\lÀ)¸ØzÑT–•FÆ1G¾¢ÓJbì";4:Þ_¦Í‘z8ÐCÁ8· Ä`m¾˜pÛžèW±šÝE†-˜0 ’!03º$'77J™ÊS¤J«-[*Ná‘b–\p“-¦©m2‘Ž›#†Oß÷@1Ú‡gálé' y5‡÷Þç7ï]o¹qÜ¢Ö–ðÕ!ÅÄ9mÈ0@:¤çÄA2ã³~íœ.Æ®D ~€.R¼-8² ŠêÐUXØÎšÈº£Uµµ¹èø—ßÄ7Ñ÷ÚÝk’£‚TðÆGlña¨¯ß)ŠFò ­ß AA‡¡˜Ç§d2klŸ„@"Oþ¤CÄ' "GØT^“´îRa.V²Åõ¢QÝè:$Õ‹¯wŸïÙÙÚþz4Ñ`ŽùµIrA˜,üþÊæ—d0eñ¥ZMj|Bš˜Šÿ^ž—2*f¬©<ÿ¾> stream xœÅ=Én$Ç•ƒ9‚™ƒ.s)ÀgY]éØy x¶…ÈkÌA-ÕÜÔn’Õâ¢îö×û½X2ßË¥H6+ièЩdd¼ˆ·oõÓJ´r%ð¿òïÉÕ‘X]ýt$ÓÛUùçäjõßÇG¿ùΚ•4­6N­ŽÏò'rt ³x»:¾:ú¾Ñë”ʶV6߬7¢U1Z©›?ã³Â‹ÏÚÛVÊæ¾51¯š·øl…ÐJ7wäùGmÚCó¾1xøÃÇÿ ë2aa Êá²”lƒ’ruüíÑñ¯¿oŽÈ½R8‰hL›3|ÔÓè<‡UtX¸ÁãSØÍÅU-„Œ¶)ÏÒL‡% _¿{Ÿ6£t¶Ùž¬•àÑ4ïÖÖ´J:Ól/`ZkXIs’° „s¡Ù^ç×6„æÍz£>™& 6­QÍå6´ºÙmO'·ÕœæÑ6ºŒÉ(d6—±)Àlîû)nÉ÷dðö2öÁÅ”ƒ Ëv­BóilH†¦Å| !8ÄÎþr]Æþ{sqDl½_9£Uëõʱ[ãVÚH˜.®nÎŽÎás¯V¾ÓV åñ!VÎ0ǪÒc"_š VöS›Ë·on¶7ŸÖÇ«‹ÕÆD/Wã‡'C“Òµàp¯—®^¯{pièÞ™õž}˜Ð*Ê$r¿üåzcAä„’Í·Ào¯/Ê ±›³ŸîßÞœÖ7¡y¿=yì÷5ª0ÀÀ¡¹Ý^Ÿ~x{ò㺧Ü]Àؾ®+ÂR¥P+@tYW@%€ñ}e¾$È|(g"6T¼¶Iø©g'Ts¯ƒñ¦2ešäšÕ ·y‹®Ió.ec]šúoÓJ¥ YÐcwùÛÝyÃåeô%ÙÀö–Ã߃p©A-(ù§ô/€ó¥@“-"8;\ê¼ô©lAÂÞÌ´"hÔ{»k2æ6³’æÜÞé+cKŸµãϽº§s]’Mmßdl«º˜þᬧÓúž ¿ËŒæ¢ˆˆ ¨HÐa¸ÖøŒŠTø£…Í1É+h1Ñb¤‚ñ†›%JÂ3ò‡Û<£†mÝu|?cn»ÉåÈäáNšÉ«3O¯lˆCÀîR²íÊÛ°ÇøuSPqSÖe3cXÏû×[j!)·×EúSÒÞOONw üë|C§´èhŸ2Ì \aHAtÆ¡JŒJ!}$Ë}?³-ÅÌu² Èó!è*ct§•±4“šÊo>dŒN,ÐhVH*“E5`n:z¢×‚Hî%n_ 9Ç TzwT¶.Ȩ,\eóY¸¬d8h.cÕ†#p/ê ^iÈ#¾âG¿×„‚ÉÕÑí7¡Ñ®"×RÔŒ^e=,ŠÏktóßL°nó„€Ò—Õ§TºĪN\Ó\ 4PÝüZmå«íu/§œ ƒ©A? tÀʇ÷^Ó:%C qDhþ´§Joín2V=ŠQ~‚s"µËÔ"PÜmË4\x‚^PK€ÓÆN»P@¯z@S~ ¸•">€Ò ;7ƒÊòm^mK6™,¥Ð¤Sf¹Jþ¾p<"¸$‚ºÍ\!¤g¯)Èw8‰ ÀæûY‚BNdYuÒjÀÀ@š¿÷\(ñ‚@¡Tl¾"K%c>ÂLQ™)V’ø»•ö*`Žß´ð[µ¿(¾Œ+Ì5Fµ’–ví½ïÌÝP¢©?[„.ç2³þ±ªü±jâÊ “yªkˆÑ¦d©ˆ²ª²ŠÓ@§™0ðUó?”dÕÚï±ÎE²œ0ÿ‹Ú)n¦&åj[ {ˆÌGÍNM‘ZúŸÑ€OiãBãë;=K/:0ª™Ñ– _`%Rš`à¯ÜîÞ7E–S¨­LÇ?£œ¿¡Æj¥`«: ò1è?'öJ?%Åìò6“®s*G…>B$;Å ¡«œ& âúQ$àš˜RC( {Gåß“Ï`@Ëý&wØEfáS ~Zž!BðÚFø.-•½S+ §À +¢FetÂL†v­–^Æû£Áê£5Mvz¹ÇD|ú¦°§óÑ&Ĉ­G.HgÛx<Ä·Ü+·ì•\VQÔ^Ÿ#¨9NG7IÓLE7¢ - npªà¦VÙ3Ó(šØ%R‚ÁrÔˆÞæéžê˜xTJ®rÉÉÚʵ±: ‡vÞëÄ ŠK¶ØÛƒWý'ÀApÁèv²JámrnÔk@êðôtµ÷£b·¨Û÷Ì®o&3[²ŠÊO¶§8?‚‡gõg­ï‘ ‚LA–ú–,õ”Z²­³ûÒÿᒱÞ 8V.¿ ’q—uA05ÉZžÏp? 06¯;;!%;š÷óó–üá’¨JŸŠŒç`8gÛEÖ ïâý øy×ïtšï…éý´'AÏl/¥dð® f)–¯‡Ÿ]“3&˜ø÷¨Ï‚ µ€Z»ª’dÒŠJgõ b¦\§^cžˆê§ª}”¬IUŒr­©òØiû4ä´fõ-û’ywÛ<&5ôÜÒ§SîS±fIEÇ EZ©QÓÿ!øãh‰[yNž“hªˆ)¢¨¶8›ÒVø0T]É™ì¨F!¡¢!m]zî ]çB8E)ðïÝ£R ™KjÌxµ[:ùL~´PÊA ø&J¡’öCˆAú‡ *t¶¡GZŒÅu`Oê,ž¢í°¾ h¤jm+©(i³Œšë±3ª0¨R#†²òè¬Iúôa¾Q­íòv¨T°™SåèJ¸Ö*¹B®˜˜I_éâÕYØ:¹*Ùü±«²%¯˜26†”i)RåeÃŒZqBiõÅw*ïéx¬½`uÓ ÃÆüx–ßk@Ë{:Ñ ù˜¾¿"ïÏ:%¤ÙBéû2Ò.Å ¶–µl~^^ˬlàd¤¹¦Cn†[ŸXÙeÙ!àø”ný¶h-[㦉9éø7÷sHß‘½\ÏyD!øJ½ƒB6è#2uý‘Œxjή°¥#aƒ ­Õ^®4&³AÒOª²1'Ãd^Ê€º…Hƒ}ó W‹ S›€ €ô»»ü¾¤w:í³› ¦ë™PÚh0{«¿ÑË”®Èè ´$œC áUSWYEe‡oAˆ#ËíòX0Ã%àkû~ߦªsNŸTê\ŒŸ0J}e¬j R鄃„(’#òÚ9M¬ÙÄ\Jñdf–»”@ Œ?•E¡ÓtO€çÜ€ÂÔ«4/è6SÔžÕë¸ËÌÐldˆŠ'5—Û§¶—æZhΔ}£9wk‘Å‹G抎§Y³’ëwÈšà›(Xè˜<¶ä²19zjv~O(íPL„FïÁ:Ø”çIeOºÖòü¤:³r¹ücùÙ%Ÿ*-ç ;S²1ˆ²që&9iÄcBraç×Hþçi l5—x¿(x³%ž²9×=rB>ÛÒ.Û ÅÇ‹ûš×Ï­ õ1)׿˲í{®Ä¦’¹Ñ´0SÒ’Rû‰ÌyCÇ»­(Êç Ê-}Ê,:¤œ6©Uk®æ|N×ûX­G¹JgÉœA}m›F;I2gT‡È‹Tµëô V-å&Õ2FÖËi%Bš6=ìøSî·Ö ‡¶ã‰O2Ry8}®> ì ï“]þ3PŠ7«³|é#òôõ Þ3ÚÂæJOBêGO‰g|D/ï‘ú,í2†š>3)­ðÑ•—^wFþËLWsÖÍôk®ª˜ÐWùRå“ÎA“K¥ªÐ”Ö[K¥VéÈ=~^âÖØ¸£~ uÖw8u^d¹„A» ÌLªf8îµL*]šn¸ÉŒã°!«cœÈÍÆc„–Ðj£¬k­jâªQÏÖDyÊ*Ð Uh3þÊŽ*:Ê_÷ƒ¦"E•âÖÕ[‰e,õ¦”Ÿž­rn/®c³ï¹Îòdë6$¸³Â(¶U ¹?e]•a~á`ÐGw÷o{†îr”(›Ø†(Ü {”%™Þ¾[@µðøJIRþ×!f¸¿ô&Ä¡™ØZ@W³ÛÙ¿ìÝYz’¼Sr–H¶t*<@%b ìQÀ€òˆ²‘Æcér0J6åÊA¿ÎäÊÙŽ]ë‚B‰aߌv>è„Ìø|4{¥>MÒÌ¡‰ ›ÕÀ+öÉ“H7d ñ—×Ô„°Í2ìgY~ýŒÆNcëæca^Ê,ñ’HÁƒ£‡pbHÊl‘›ý­<‘’ÝÑ?£=K<ˆQÝ7Ò>‰}†y?‘j?þÍp«ðÂÙÿ+РT=Ø?½žRjèHöMÒÏ–ÈQõ ¹ˆ/iÄwLs†6X*;,CUz¯ê´CÕ9lvNü²QàäâéÁ4ãï0Â#¦“x±À´®ã‚ßîÕ"à°CÀú “À>²¢Êž#¦,ÄJYÂ0è™f}Ê"~A\ŠQ¥wwqºÊàùHlbç¾H'!ÆÇmD:Q6w+dqã.OTeJy`%zeMÓßuµBç^A“ŽsŒ@{jšŒ9ýÏ $¹úË$Ñ~Ey¢Úab4©›š$3o„x„/éhžÐË$²´†0±ÿtŽ_wžúuv$°ƒx2Ò“©[ÿë^We/h\‘J×g̤ZçRe´˜úuožp»ƒòà5IßÝîóYÐz»Ãáî[(·"h«[‰R~uÔ\Hr³ƒBbúóï‘ÀãrØmHAý熀/8`lÑj¯Í¥'núé—H¤«5¨óûë“»· ­‹ÜZbî<b |l4nóÉs®ož ì¨õÔÝÙÝ2»r5+ƒõj™]y”|¾«Ëì ‰A5%Õ:Ÿ•Æ’/ðR…âð¼è4P×Yƒ ÀàPŸ¢¯æJ®ux1’Ú¶ï{ P5£sµ Ôkä¶ï™‚°6 +Ú*3 Á6àë3¹À2xrA.Ð ÒO¹`frRð‚bs.D—à[ LD!}ÌL¼qC K™¶¹R‹(ôŽÞ¢¢_„ ¨áDO-ˉÒÁ8‘Ìù×…UYqQŸ¹üè öàë_†¯”p ~ÀWŸ»•îB‹/±LªEà -'Ê„Ö+άË(hæÙhÉÅâiw€MNë@0*k!ÉP.—|¢çS…x-MÔ8Ó4¥Œ68ÚUqö–luºÁø‚5L?îL¡Ã4±!õÛT$ÆP+„æËé&‘,‹ž*ÜÓm`€tßwJû/¿V£xϘöé8$ìáåÏ¿é QÅ®|Ú­k¢z cSâqå jªLÿûŽf“MT25Q廿̼ϗŒ"Fñ¸D½ó!\˹&¾$sØW½ú¢÷Í‘ÒÄÜKïWWGD;¼‚»¾¹<ú¿Yùä¸$Npö /Á"M’Ïß§ë¡CPþÕƒN“؃5â¹öbq"ÆQÛ­W¤•¶HÒð5j AG”uÌJ¼ø„5·|a£b@þàwú(¤•à ûA€'å {ÔlôhT‹ƒ¦l:†]´›SYƒTcŠÇÒá³Õ\£[óÆÛÖ87<꺿—Öƒ”‘ €õ~–àðÁnHl³|j7¤7C>€|d´G·BJéè†äê4^2Ç=#ÄÚ`Ú¸ å…GkÄÒÇ»å2èA˜¬¢M™9nþ+IÞ°O^ƒbàý–#;0h ¿Õ¦æUÛFà¢üQ9°š.–*RØ—XO$akãwÐsqÅÒÎÿ"F}ý*Û|ì°œí¨MDsžÃÀçÈ?‰áF?ÓýDôþÔ¹«Ókƒd§Úº!ÍÏ›‘n0ze7iIáRµ eß}U¦tü*_;³œÿÌ£¬:Ÿ”|zË ®ïÙܲ€q"úªué òç_§7|=oÈÐǘ\Óéµôã&ýx»¹ö.ïkú–2.ڞ⊚Á©ÉþÎhzïç å53‰K ™•ªìùž[è ›8¼ü¥›ýóÚz¢î°ÅGPZ.ÖÒƒ aòO¨üábÑ– já– ê%Zz(¼…[z(¨¥[z(¬e[z(¤e[z©ÜÒ“+ýÍ2 =93$ñ˜‰Ätµ½»y»YrÆè€û*=ÒÉB7©¯bnG‡¬z•ZµTÑ+×§ W å{SÁÄTã9ò@eT’lÒ…Êù45ƒD©ÿ™ëOµ!Çf]¨S$Œ8ú,ƒ¤F Í -TL} Ðó« ØQa›ô«ETW.ñå_m—qir!ÄXøãëõëõ«ç b‘n[cÇÝU$Eæ ¨ç×#«x“I—R$Eä ¤1U?9´¥±ˆ<ô|I,ÒMæ\ªYež€z¾yªâýR‘ÊÁúÀR‘ÊÁ/!X á…„3jQú9Zp_ZšT¨<¨Ph˜LkõR¡•ÉÁÉò˜R!õRH y¢)d9QÄ2›Ò!G\Ëoª3Àxz8¨»3µdÒ…ÝkiI«¨ñgÞæuðî#Ç ýð|o«˜E2éRf¥šEÊÌ,.ÏVÕ*¾[i‘îÞgЖq¶´°xg«gÛ•j'Kg_†,Uß/.í¾§daÑÕ3¬J‚Ïäô ©4ÿpxs}³û°`®¢°~·T·,°…£²7…y h!*•¼ ¥ÒÉîrÁ̼•’’Ò‚‡[`…é)´ ‰^'[屃ÇÂXÛòxö5Ù]³môž5ñ¸[È•—ÍÛÜÃS~.!—œ k5­´XÔÚh¼~¡vîç²sú…™×›ò8¹Î€­t²yWJ¡ÖûúÃx•É]úUúàTù ®¦ùCù•)êOR+˜[•âµÃŠ?{}S~{D6ß•gÊ/‡•5ü=?#^hM¿Lk£ÐN°ªT¾©I#? Ýœ—©‚Õ>Ã7y«o3È Bÿ‹Ýõé„îqØE©k+GÏÊÞñOÈFéêvý—iåmÿ·®^5m¼n4¶ùWc@Ï”ë¨ÛqÔ_ŽþßÖ¦endstream endobj 357 0 obj << /Filter /FlateDecode /Length 7538 >> stream xœÅ=M“GVœ˜uøÀ {èØ%‚ê um~ؘŒaÃv­0!Ú£ÖxXÍ´<#ɲ=ïåGÕˬªö¨ÝÙ„nÕdåûÈ÷Yù¾[±ž¯þ—þu{ÁzoœÕ«úÿ÷׎ùÞÚ•‘Ž÷^¯”f¬Wfe­ïµ«ûÝÅ xÝJ#…ƒ÷¤–LXüáµ÷gøî‚p«ô¿«ÛÕïž]üö\ñ•@F­ž½¸ˆ¸À‰V–™ÞK»zv{Ñ=¼Ú]=Û=¼^?ûß‹OŸ]üàI¥¼å«é÷†ÇïÐÀËî~÷p¹Fxi{£ýjÃ}ïaسçݯáÀ$øÃ³Ï/žýæ+|¶áLÁ Wüþý1ÿh½ÑB÷Œóîùîú~·{ì_¤ß‚w/àOÏ÷·ù¢»yÿÈqVÍà7BøúÙï&7€w€À{œÚpmñ­î?7o·/ßìÂð<0Ox ÿR½÷¼ûô²»¾\ÂWÁ< ï;Ñ®…á‚È0íT˜-®Äj™ßÊ­`# e·‡àL¬ °]h‡ìþªûÇõF:ÀÂÙèEÙñJÉî6üfŒ{Ým_â?ü¸ò2Ž·®~CFÜ®…é-ãª{ƒOA¹Ñcû:üÃ{iŠ7÷wqrS¼¹玌¹ÆqÙlæöŒ;Þ}GX/}·‹#ƒ•ÆÇ†1xºhYïœ3ý½_ Ó)ß}Oc^Ù>¸g ÐAw¼ s¿!ÐÉà+BÁ•Nù Q ê¿%ˆü@Ý üâ{Y`²‡×º§$cÎÈîóÈydà«Ê}øéâ n„9‘ G”H”G †J ”3Ðѯ2‰ùÎëΫŒŒ¬¿8’ÄQ:Ë“ôÄ1û{B¡>0…{ÃÔ aÒºZ8‚¬$Q HéHÓPì ÈŠ–FƒÕJF 9BÕiGVt‹+͘sY¬ôwDΩ.l)½ fŒÃ%Áy ãÅã´€\EtÊ7xWqßÅ$ÂF±üŒ¼ø´üÓõ y“E ðõQÚñ7—‡”FH§í7A “•‡c®{»†g‚q9 uɆ ï¼íöä7]¼;|ÖÚ d|Ï3Yhèžh.{*›”8§\è×%=,®èþ5à™DS¢@/‘eðiÐ3-0 HjF×ð  fÝz_Ë;RåÀÒÇ™I.[ÑÀæý0üëlã:§WßDb˜Š,DÆ35"bâ–€|–4oE^jomXi  & ¥ÁäøVÊ}X@ÅâúÁ{àˆãú¥¿ï©—xN~SOrÀQ/$-€âfII ²š&ù‹oÌN<w¼PŒ#Ñ\‰B©dMX1U˜6çÀG›.kKåŠL±}3/¨»L±x ÀÞ-Jü‘ i‡DÞ4Ë>µ2Ï0Ÿ.}Bi‚J€. ç³NP+MgþѤ‘tjß©~P0è÷R(\ g*Ø/¤…ù,UÙPÍ9†bò·k­À7q_x²r€ï¥&Þ3qxK½"A,%¥Xie Ò›êÑéê$\ÒŠBºÝ¾¾¿y×P‹…V(‘ÕfiX†¦VQÂÊ©Eš‚b •HR@OP´Ÿ´¡ ?¬º¥àD-â`aM!©–Å(–¥JƒŠ Ú¸”¬ ²aɧªp…̪@@oâ¿$ÄSl+¢Û·±‡Ñ-ÛÖ%7Íü²ô¬›»F*®4`. XdµÂ¨#&…„Ö”|ÔfY vÖ¶\4 ¾J‰^kÞßW…âr÷®¡£¿⌺sOE@Ýßíïoy*,‚À.»»–ÁQëv·mòE÷a `Ÿ4òT\C®SʆjC”ѰåjµY*!<–” HÏÛ¥°†}ž•Ò+•D5Š)4ÄJÒ•j•‚ 0æ^ø:ªà_Ý4 ¤¢÷à+åy¯ÀWÞ㤺W@úºM´¤±v" H­ÌqН)¨fV‹{Ñ © `—àÒ’ì8ÜUq:¿ìd-MCd‡@j,;RkÙ! ¾»o´3Ÿe‡+’ÎÒå!ÐV¢õvۨȑҳ’¬¬?'Î>ŽÎü›64H…Û| w?›€ä9é´›FžS¹ÞŠ’WmœHòÑ«~þR'Lgýmô ĸªTù»ìú5¬‰4“‚˜ïV¦3/@Æøµ4u©âEA]ß6ÊYbAA]v×üÉ;þ¤ m9À§¯šø%iTZ€ow¦€Õ¦r ­* µªfKÛž(é{ɪ¥Z7ÊÆ²õ¥Êu¿m”Ûk]û䟯O`„]8#PLü·mH€ð™¹rm®ö»/n®nvw¯fëNXßv®,;].;m|¯8naQ‹þÍÍö¡i‰X©xÆò 5( êê››»6kP¼€Ö*rÍnŠ€ºßÿº‡6¢˜¶1Î@Û)PmRèEEf¢'°9¨%s·òÁªR!oÓð“Ö¸R¨K`5ªF†P·\©Vn=û$ªQåÎBmª¥Z—Žý„"\¢<Ÿ9“®LÕšÛ„6˜¦·ÊŸÁAXÞª¶±˜Yý|s$ ǃñg0GÒ˜^1ws$-ïyeb[™#i5äjò æH†ŒûÖO¬‹ÊͱFÒùÞc5¡²FT®O¥Íz%M£æç7·wÛÛ]c4D™¸Ë=Ô0~α߮YÙ7ž”½¼iõ9V®Ä¶ñTñ¬Ù§Q¤”Ãåyz¸çΣê*ªaVáHªu?§¬<úÕí›_µ-  ž´!)×>F’n®o·¨Ê!Ù¸R­‚$kð`q{Ùs âVŸAör!d\¨§Ÿ·Y%ÉÀUqÑ\öðk2Ï=%é³§OÑÂ=SH^ª° ‰Õ6žê1Ö…‚Ì—ÅÎÔ)ë1ø),èµ€lX‰¦E"S€jXÉ;¤Z³zL*ýPÏo¶×M÷ Î@–ã3PPÍ6H“g.¨Šµ˜f9׾ϱjCA£ý² ›wçX¶ìÕêuçð9~.Ï‗h‘›ÅøÉžx­wP ã|<Ü¡+>¶=ÜA@µ ö£åZ¢é„WN+¨†ñ~ÚX%ÀEü9³ d5ù“}¤+Ö,êÇBì9¤0â3ˆaÎ0èzµ ýY¯™:ƒI%«]øêÊÕz 螯NH>ýâÓ¶ñ?“V¶ý& ÇÿÔâ­uüO@5Þ>§ ­õ†,ÖxC¶ ëIJdîÆ²TÊÛnÈ.Ix‹ Y«ñ†,ÔzC–€j¼!K—j}ŠÓHÉ“yÿ»³ò.o{fáÁ'_ñªM|à }>Ë-%èQeM›™n©@¾­?ƒéÆ›v„3SÓ}ŠÍë´%ÚÜr;¢Í-÷°#zË-½‚XAžÁrKï{yíxsãÎ`‹o^ØSnÅÁøsØm%ÀŸz{^… ŸñªAPz Ïèó)ÀKLË)¼ÆE@ ªy°äcÓ" Õ´¸HÓé‹€Tó" Ö¶Xu†"`±bm‹€gÂT<‡¦Ä£X¯ÖEÀöb˜‹€YÍ‹€¥¶v˜x§7‹ßôÝï^¿¹oôxºZ†‚keæS]‘‚jç¾Â׈f™®Þ4’*˜Tõ]Î éJÇý(°OZ£Á²#à¾ÜÞßlï®vmlc*ôÄFä…Ê’( •üÓŸø*( qéÂÝDøF,¥6˨¼çëÖ…à«î¹rõ[r‹'Þò*%ÞïEJ÷Õï㕞ñbtîØæ`¥%ÃëMÁú;i½œÜÊ.U¸LrouºdöÜ4:¹.¾¼r°x /Ù⪸¤4ÜP Ï!M WÞâ]ÕNEâT¯DW]È;\ªúŠ\ÞZ\ “àÕÌÖ—·£éºõÍÖS{óÈknóŒñ£@óÆèq1‰ÒxËoâàbùƒÃ5þ/«ÅK€U§Ù“úi­ùÍ€ªò°éVaΘ²‰o”W‘´T¸—HN17òGô@EX5r¹v\@ (™FßýZo‹;Úé^’{§‹ëté××ëD¿"èz ã@þÃ8ã.’ï‘ï£x€×I·ò'P»µæØM‚“_Ï#Û„ ÝÒ8X’?(ò[ÓÖåÚsµé¹–ù£‘ÐÁb@ÆêV߃áøì‚kp”è”Ñ,‰Õí…ÒF÷˜?Ç//¾ÇðtÆÀ ñ- +Éã(—+:*?!£– U!oãn†ê½CYšé`¨þz‰‘!do=Òñ»ñ¦qjœè•ý}BB‰aÉ€Óðähœ¤ï™¯púrí$häåêI›9,d2‡ðˆ·ÃÛ­ ¼ó“£ñö<”Õ ¼é5ç_ŒFíÓ²<`D¤{Àh\2jîãð–0“ €”i,¬MS~r4Nð~­O*ã½èoÅ h?Å{xr,ÞJàøú'd á—„[#Œ8eë0â4Ú =c/êÚ-}/<)FAœþ“ZŽ6z/ú¸á`÷ S¡<èû,.Fè%òtþ÷h³Åq­Ú‘8—òKqb=‘ZÆ[OMÖ€1bÇ¡’¬Ã-£à…_O­ÏˆÂhŽDû:p~·žš›b€ãBF è8D!c LˆnBÔº°Ð´cNRœ³Y±MÏ‘ØX¼}Ùb›r|=µ ÄŠ‡ny6+¿)NeZS²ª(ŒÊ/jå·$UþôÞÊo°u8:ܱ"¶-éûç뉚è•â­æz­å|ªå#µÞ­åË8!§JN0¨Ôþh%?„˜êøˆA­õTÇé¨RëÖqŠ'Uñe©üTÅ 2•Ò­âËLS|FÃG j?ZçdWTp²: *•¯<4︥ï…'‡ð„a{Gbî2"šœ˜Ç•)éw½œ5„­Btb½ánö:Æ1 ÔO§lÜB|˜3ð4ú†ü¾M35Ý6—plê1–žÓñwX.²·Š1ßîâsÉL÷ŠNtO^¦ÏoÉóÝÐöFˆÒçi!Mj²£C­=þ"=æ!…{|ê–ä(<~½€Ì7/K¨¬PÒ—&º_`í7o–˜N‘¾[Ï•f°í—w¹wæ³XàðVÚÌ …EǪà+R•[ê>÷–¨YÑÿ§êãzãyW¼ûŽŒßÒ.E0­@¸ùisÄ\½ŒÝ›ß‡f”`°'5:¡BÅC}ŠsÓîWØÂ“ÙÅÎŒoÖøXi%Š*æóø¦ʪ ¥À¶”¢[nÃ5Å[µÅªUiAºC}õ:çtͺÞT2¡ß‹Nx¡ ‡DŽ•%´iË™t.ØÅI#¹ô{¡=Ö>Ó5ÓÕ.ðÛ}aAÚ€ƒ+ÜņÚ "ïÆ-tçSà3ºóa©;Ë*[¾í(Võ䣥¹MfÒ&šdyõý¢,¼Ãb46Hä ½çvq€*ús•]¾ =f.{xAAž~ˆÓYçŠ sPU*ÿC-A°)lp‡ÏÑPRÝÜFÝ´¦lÚ§žwm|Ù(õ%Iþ÷´#²É-^©ÔÝÆ!›âÍwdú¥É®ÅÌoZ£ØSÁ{‚ÏC —t¹ 7Ù=ÇÀÍÙ°X´W¥ÑÏgþƒè<õt»=4É~k'–]õ ·q°,­Ý>aºTñE HUž.°Á¡Pã@zípk’Ô¥®Å±¨§ @%3a,€Ñ‘T2öÈNËx% ¹¾‡&ñgvZ&ñà€áøÛ»Ô4bm18Èç‰I\‘¦ËÛôv˜¦"Mi¿¾~DÓr:äy–Œø³eùa¤'5GŽØ•RÿOQ CP8h¯Ð¼Ô jô‹–­k<¶<î”äî­R‡{º!²\ô½_Jñê^È:fr•Ãqï:òTG3<‚¢»çÌ;&DëñuÏÕcržlèN>ŸÅ :®ˆû¶•×Â!˜qJø«R7à\­ƒÇvEæWlWÕàçdâCF 0÷*r…¸`®Ã)Í1õ Ô“hÓUTÁÁ »׳¤K³Rðhýp²#”„O@5ª<8ŠJ93³ìUbðŒ» /VàŸ¬t8dX:cW¨ÆZHrl%9®w‰UÜå»¶µ^Áâ:.#fàª9_!>ð‚Wd†E ²çڈЙ˜¼óËRÁ ¾J®6¦TI°W·ö"ÈÂó~KÜ÷Lè¡çs õx¾hŒMõÌà‰–ñXL²ê õ‡ô¢hF­@JÀ¡f“æ‹9ƒÅ¡Nz{MÿBÍL¶àQ ýp¦¥Ñ·øÔLÎÑ'0`j£Höèëƒ Œ÷šRùËÑ,§Áå,¡±ÿc,¢vç}Q&FÏáý¬cÚöÇŸÐ$à”4 e†ÅUÔüñ¦â2‚ã‰ê"âÁùû Áï’” æÃ•_\w¹TÔ–šûK¸ÏPLÆ*“„@œX,_OŠ£Ø—å©»ø¢öu™.(·š+ÍNÊZo׿;T*š˜Îú™CÛ.Ø™Cý£óŠž"2šàTõC±¯ëME&1xò¹Ä5M½ìË#ªÑ>îeP›F Z>YÕÔÜm “ç¸*š= 0h”¶Ÿ”_,ÉÝ¥ MŠÝ£¼[öá K4AÊFA£>p¿T(+­0žæñ¾¬.L¢çô»ÜL*—=oß­q%•¢E6¢sA¸”½°¶²§“ÜÓ›"¼Ö¿ÂqxAAÖÜ?WzÖ´pÍ9€)ÞÙ Bv …W¹p¾ÈC‘íÉC”³yèxbt&E³9“‡BÐ Ð$È?-^,‡ä‘¥Dôe”,zΘ²ð¼²1¼óg£¸ÎDOá×vý%‘íÃù?Çûφðó¯Fk2ŸUk9>®1ª=†°RU.iÁâ.IÁŸF6šé¹Ü ™ƒ¹o F8s­£´¿¨…³Nйõµp’/kØzc n ˆ¥¬mŽ7qX¬—£•›ûì¤÷Rû4_Ï­'„Iáã½y]ºåêÑçÄEƒ»•Æã‰ŠéBBÖ/z‹×õA@…ÇÒ»½N£'ÜñX|¢›ÌI[æ¹ÈÞKtÇ4ëテ/ˆ®•‹ÑÔTt™«‚•$º^ ¾ »N”ÁTزô"D¾Ôtã6¿—ãY‘ÙC³ZS—Gçê9[êRÊàDõш­xL}ô'§£›˜šJ*v3©éÜÞ8­B VÕ¶ÆCfœ8T¾ˆC–,ÝÅ!Ú¹Û¤&'ð FØZĈ>·>>Ûǧ?¥èòœnßïR8ò%†R"n¯z¬DU&Ékøÿ_&I›5`“”¬"‘*"V7¥mácçÝâW¡li¸uU86l=ÄñCXD3årY5qœæUs¿´…ÞT aÃâWü²õÕ°Uz(ò»›¬Ì Ò¨<¦û16XÆXVA„©l8íaq%`êÿZƒÑàZ\¬~T‘vXVX˜ã: ÆñEqp?*r[Ö%žá3Ð’ÈPï)~ëëri-§u—ÿ>jPbÁŸÀƒ¿Ñ¢ãû¸µµ”PÆ8s“q(³†l¨ñÌûcíbé Ù’„ãö4‡$’©©"MAä¥9ʵ“„caÆ¥BD7–Ó–BóJ_Æ= 'Ê uý/êp³ŠêŒ“âñdCÉàÀ¬w ŸC%Çf-1hw4Þºà7eÎ úøÃƒqŠ4=D]sá¯v¯„?xhü1RÎÍP«q»X:BoÆg𨸙¼×|î4 +ì Ï;6ǺÞÅlk!ר¢;%ÌXØ_'‰w“Ašù! Âý\(u`§93í4Ï8Ä™JÍäóù«ŒL9å;Bñ]EuRÝM…œ/£;`j½B×áƒÉàj¥2¼Ú‰Ç5WðNäú‚I¹ñW£þÏ@€LÃŒZžrbÈÿæ>طغ2omC¶4_6e•QÐ!sãž$ÿšªÃ÷®~Ó`±¨÷Oõq/Sšj¦ÏmæŽxD¦ Ðv>½‰±š‡Tª:¡gÅ®ëS¬KÕ^ÅY<•¥$–™¸ÈóZŽÆ¤ŠGÃf‚1jÒô°¡¿ Aðºúðgfe û‡ã½‰h” ¥ÂÅ)}D Þ}Ìi5\k«`Fþ˜âÀ̾µºW£+˜„µ‘úð…Yº–g9flØÂyÊùÉç\Žî99B7^–·Ä‘iÎÇMà´²c¶öfž!Åú‘eú1Ρ˜¬Ìùò81®ÇU ž`<PŒr±Ê/æå©ÚšEñÆ.¼ºƒÃÃ'°¼ÒHÙ2uv¸`ìÿ…Êendstream endobj 358 0 obj << /Filter /FlateDecode /Length 5023 >> stream xœíËr#I‘³Á.p =;jêýØe –å ÁÃv–-ÛZ,iVò¼8ðídVUwgUwËš·‚1‡i·ª+³²ò]™õíŒÕ|Æð_úÿr}ÆjoœÕ³òÿÝÍ™c¾¶v¦5ÓµP3¥«•™q£M-ül·<»†Ï­4R8øNjɄů½78÷g<€›¥ÿ.׳_]œýìo\ñ™@FÍ.®Ï".ðFâ«™e¦öÒÎ.ÖgÕÝj³\ì~ÿöÅöþv¹_íÏ/¾9ûÍÅÙ_®TÊ[>ë?¼3\®}-œÊ?«v˽|Jà=Þ:¹³µ†©(¼Ëç«ÍÕ$«œ×Ì”«›da‚k`‘ºZ-n¦Y—Tµa||]Æi­p9L;ï²¶6Njÿ>°|-™Ï`‰iV¥TÍx.ÏΧáC¡e­dNÁËiV¥]Í¥;Í^Q;•SM³*r¬eééD‹²¢†ŸbQÖ 8 :SK'OÁ€” ? ÿyP‚Þ`«$Øc°Í§à?‰.—'Y”¯½(Øo*”´­'`@@¸¶ú$ (%¨umO`¬¤µ^¸S1 ­^¸1 ­n}Ø\Yƒ9G.Õ³‹«³ê'?ÁŸ88ΰ°‹/Ï.ž|…ïæZèš ^}|Øô7SÕmçÍÂ;U{ç«ûåþþ“ó¯/þØûzq÷âv‘z[}–~á öthüóå};œÓá, DZ?í}ô§íÕò®Â?é~7ø~·º¼_^uØ®Ãð!èÅD"ŸH½œ ãÚâgÕ¯¯á™) °«¾¸]í¿í&üËîYõ‹ð6'';DPàÔ8RàcKJ°®Ò5?Á·µrr9gºY³¨ž|ÿ÷ÕÍfu]70xu ¤ÙÒÁ`øg;˜ýÙ®Y¦¯`ãÝŸø{÷3ÇŸy÷üÚþè4þ¨éuþ#ïþxÖ0%ÏŸ#å¢`ÍÆÅG¹l‘†JâLÌ´µö¥ç«êçs ¼oeu>Gå¼®VøÚÁJ¹¨öøZ7 @ÃkŘ1®º#C–ø,½WJV(O`sFfƒÞ¦é¹­èúí7øl“>¾N .É{|ö Ô–ÎÆ\á3˜+ØZκ"_À®¢yd²º=‡Iœåan\‘ãâ²™;¯pÍ-Ä—qfë«´.¥mÈ@µÀ±pÎtÔÒÙÊ^ [ &]öv§ÐÄ´Fîx†àvÓ‘ç†N³‹ÓXŠÊuDÅpCg¼%ËYƯ^ã[Á˜W6±¥+n‡Û ÝÅ:p¥pŒ›^ÕûUõ¢CöŽ µlè#ûdƒ©™Îè½'{B÷犌©#†Ò:‹hüìoZÍ °öZ dÿ9úbÒÛÙvÀ9d|DO†Ù¼w ÿa ¨ ïYu­«Ï% Kò‰Ó€yÞÄ]þ†A‹†3­;ßjÙu$A±lá5‡}uE¿¼k ©¸îÉó-Gb²!«ȉ 8!ìZŸFlÉóºMQÙœK††$j\³ 隯ÆðZ€Úì“rÄÊõ“nÀm‰†âçT>7‘ݬv@Ü<Œ·Øà£bVD è+ÑGFš@cka 6Ó=ùŠ£©¢[·tÐÕ¢Ð3Ø·>š ¿©n£¬¼!šˆb°ØÅiŒO ÐÙ"¾à€2ð–#Wõ—нKëd"S 3—ïö~±‰´ÐN•j§C@noFÆÐ׫L ¶Cê Ú0>‘úLu=ÕíÔ„&h›ÃŠu4PÌÃâ…ov|@å>A”Á0vÐ\5š¨ÔxÖõ“¶MBÜ„L«PmÓn€åÔ‘=Zxd¢ÔL ÉŸ#}¤¼;ú|ì­7L™ÖRj œBM₪½dü °+ ‰Ë›7ë›G5—4-‘­=Èòøasp–jœ%å ¸n¦ö|KÙÿááEa­qJÅL&Q”Ï_á²÷¹ OÜà ˜ÀÅ]BŒiºë”Œß[¢IV¬¡Œ»O8q5jÆ(Û^¢1\k”ghLôL-¬æ8ÙÃÑ0ù‹‹ °YyTjÒ«À*8ÀzT*À*NZŸIãˆ"£›D-è}7¸I EEt¥GGâ23žæ¢0‰ ß}‘NP¨ˆLƒgY£½/QZrH`¹¤âÁ)ËU<þîy†g#Ñ\ žcÚ4ÞCxnH…Âq¬Òo˜7-gˆÑ’ðÈ ºITú~98³pŠC@š†^ÑÊJ1ù¤Ãp8W輚båoã÷\¥O¼˜¡²Òá' \é‘XèfÀOá“UÔF^+;€¶©Yø&ûä³n«‡0ƒ¥qC¨"A×zp!þÑqÅv\€ÿ]°×Dárì>>ˆ]xrí¦ý¸A‚e˜'”æB9˜PÃG \ ¸ñ£×é#+rÖ£,Ù)¹¼!§AÇl˜"¦ÝØ/¶Ëº&ÍPîú¶õÚ{Œ—Í-j @›±}v¾ žFËÕW‡ÖÂ#Œ,m×aöðð“;5ó.‹Ux•s°µ™©šsLx†/¾“öÓò! ›( ·Íôo*Ü¥MM?ÃàŸË@s A‹¬qŸUüFt*7}£uö lÌ’éU9º Mƨ”gþXùá¶å×1„€M ŠM†|ÜÁM˵j-¹)öìÓ¥Bä´ çWs>»TäˆÿPÍ­C’[ô [/JQØÝ8)2 Uöˆ Ñ6€«£ ÀÄ,û(ሰ{€äÀÀ\rñÄŒ€”äÍè8æ4(8G{×ÐàG‡MxyÒ´\¿êðÏž¯ÈóŸË5æ*ˆ×h)Xáó‚4¨ÞÀkךµ*ëÓƒê ­ªãíî~—8|H’˜íáG:‚ׄp,(Nˆ“JÑΪ™Ã§=Åä3ýô€©fD‡ÿ ìjHÜ ÅÔ˜9¼ðl+Ý °ATí†Ód…”ò®>s÷Mø©z‰”€–Óãj<¦ºò@¶×ÁcX6cLþ;w憀÷Èa÷Ài@“ȧþŽ?ìÃpÂÓ–›ÈhNù!>uµUŽé4Pºˆ)R!ÒCÕô"Ù¶Õô^³éŠé%)BéTLÿ|·Ø½¦†—» Þ³jýê~³Ý­cñÔ£Wï5Kd°£°Çr¿¼¯÷Ëå4•ô\ðÚœªˆ£S6‹Âm!¤œ¨BÕ¸3[Z,yƒèÐp¦©~¬ö«›lÙ>”dy0ªV"£MÏŽµÀ©É¥ú×¹ÆÚz£ªÅMOø´ 'å”`W8°`íÅ+"÷äX“žTFí·T Iž¬M×8A 87¬(2@ÿ>Ÿã}@‹lÈ}|ʶèöÔé=§umsTó7©w µXuç|^[æ±\‡@xdÜdŒ@z3¼u±Pa1¿+;Mí:[@X[˺¢Ç pØÀ¬ˆc¨ ×Û`ýoÎçCÕá+áY;*S]7êWºî"‹ ”wãK‰Ã×áÄ»£ç5¦m®Ë‘=×CÍØaÂf6V4ñLeuŠ©2ÈËüÜš¶ÝQÌ?Ù[2eQN¬”&Ñ^ÇG×ö4>ï±¢Jìæ6Ò”yZ 0Oð±1v°r`¨.¢,ïÊgl%œT Ínú¥ mƧ|¼oïû®ázØÒ 8\£NjK‡ï F­"•lm×øÓ³O£•‚Ü×ÒŠ¶;¦W“Añxc3™OÑoo?ÔŠ E©?Âd¿é®Uoé>Ó—§qüY¸¦…¨©s<²øp ñe[|!4¯µn«/=u)J-¢W|;iI!Mœ£ ¦»¶0e—0´ãoÔÓîn“7Š^Tò~ñ_:ë¥kš*Ý”Îz)¨‰b2z§sÚ=í®›y º¥s×SÐ-»RPÓÁ4ç®Rs˜ÇÑ&äE07r‰N‡ Xž©ñ§R¢³h+ž&PPo?8¥’ kè¤ÿ™4N!½þðÜ6>›”zøqT°EG`}’ì»îÎâpÊCC 3†Öѳ’ͪ,ÊCC‡)ª×óÄGråhO C/Z¸«pE–}eÀ„Äc—U\ÅáÚ›¼àÞrP¨ãŘC–r \Žæ@qÎ  ߃'D`§‘x~³‹?ÌY‡ëf;Lv%RéhÂpUÞ@ˆcu¾s뎃ÞbÐäjŒï|õpGé’vô Y@œ»ÑÜÂ~xÍotHèô‚.ÀW÷ü2`e¯ÿ=EaÊNì#.»(dìBœâžX1|¨¢9FBYÿ ›>oÏñBpG»2iàr“:Ö]¾LLqôsËÛ¹‚|«îª/‘U“o‡Ó\#-¡eâlÈ#ïõB­ú(0`é¤Gdã~ý¦[ö@?¤‡ûÔAnVõ·=™`4µTØ¥Šô§gºöáȸùÛ@ØNŽáï_Âßý+›,ve¢0qG39cÀE¶-Ö~ð¦ %löwÞñƒÄ O¼Mw}¯wæé#ð6\vôÌŽ^ 2Ý%´½¥ý„¤¶‚ ˆåÜÈc·‹,øÁ Ã{è;-7L`AKöÉÃd áÒ$Ff~· “#”‡ÚíYÑp[lØ!Ñyü`¬"בï0ðnzÅæÕÝHòWό߃0*ÐÍ'ï$vóÁÿá™Éµï)ÐÇ_ÇÐ t~ÃûñGÖ1†ü!üP?›óÌÝÑýlc¼d<^Ì Köµ êu]Ó‡ÿóÒÿ/)§H÷áA^7c”— ×Ei=Soåˆq=o]®'®¾H¾‹ñºý›ú.Ÿø.xƒ0|Ž=iX„8vK|„†7SgkìTÕ~6×V¸XTÇ»"¿žýÌ3›endstream endobj 359 0 obj << /Filter /FlateDecode /Length 40373 >> stream xœ¼½[¯5Éqø~àq@½|mˆÇ•×Ê2 ‡1ÆlÈŒÄ0硵ټxxºivS4_æ·OF¬µ"³vwS”ÌÓl߉ؑ‘UYy‰ŒËÿx=ÞÒëaÿñßÇûËñú‹—ÿñ’ûÊï¯ÿé'/ÿáïêx½Þ®žûëO~þ‚éõoW:òë™&—³½þäýå?ý—Ï~<¡ã¨gùôÍg?.¥¼«}ú•¡Çq¤”?}mèúVFýô›Ï~œÏ·~Œô髯¤\W­åÓþ÷ÞôŸìïv]©·O¿Þð_€eMÉ„Nôu•þé+È?’ І¿µ¿Ïã½ÉTà%}úÙFò»íï‡ý]£÷9ÅUl*~õ¥ý8Êõéß­?Éñ¼æß_€KWýôå†løèʼnNŸ×£Oò©bO}»&q?Š÷õkÔT>ýóg­½¥Që§Ï½ñøŠ×ùsŸúçëÓ€:¯Ý…Îßzò¡DüÂæùÐú®Ð/ïõ7|}z7l¾Æ²?í%ýr1þê³ÿç'ÿõ%ɲæ9Ä~ò³9 ~¿ÛÿáïF݆âxSä)ÚÇg-½Ù‹µ÷ÃÐUOóñ\o'Ç5Ÿæ—Ûß߬¿!æ6âËü-—!1ÿ#¯œÃ‰ÿóO^þÏ—©só#xþ÷·¿xÇõ6uì)χ–_k;æs诣_oÆô·_¼ü|6?K/yŠx+­ù´?®9»qø¾Ï2Õô:?ÂÞëþ]Žb¨×~óY$û,?=¾úòç¿úò›Ï~òߥm©õ:Óë·ÿøW‹K©Ïy`ÜäýôÓo¿øºüå.¯”ùb¦˜ÞÏ«_¯ù×Ë«s€÷|“÷ë/þù‹_Lïú°Oë&í¯6I}´VímmŽçÃ^|룴ëß êœèqqóãú˜nó­å§—ö™‰úq:ÊìB~ýqš@žÌ1ÿé/þÂ~›ý“¯?ùÛ—Ÿüû4Ü[noGNŸþáó_ÿÌ ú6'ÍOßüá7_,ÈßϾøò!ܸ>ÍÁøÅoÿyNQö9Û‘¿‹ï_üiÿ[lnÚ¥Î>ŒOövæ”ÔOSè“‘yƒga?ýô_L­Ç¿ùf>‹ù Úñɦ…q´vL¸ÍX¥QÊw3øýg?îÃþN“n®K5¡ŶùXë³5gü[WÜvÍÉÕºj+î?lëà/·uë‹mÝúîùö[kÅï˜gk~;FLÉ>¹וëœ÷³Û4ŠÕ£Ï¬ ,Ðñ;¦×|™B1½~õÇuIsEÏk}àì~9cKÀîo"oªì³þ?}ö]“ý|¶>ƒÿ¾ÎűÎÙζß³dîôgFÝÞzº-Ü_Ç‘®ûÂõ-s¯’µYñ¥t“s_ý¥È÷/Á&¼\ý¦ßWû¦à»yïÞ7699–>É7’o Ê˜[¢ïyTò&Mÿ 2óÜÏh“gK­:m]úîçÉGXÆéèÿÝF½÷o‰·V?®ÙÖÕ6çá4'¬–øqüæé¥á«Ö&ùÌ•/ŸÁ|Èí°Íö¸oý yNäçû¶wíµ¾½÷#É/>û®mXk¥Å'ÉmØuÝgŠýKýfûû·Ÿ5›dÛy'ÿõR毶û¯Ÿtä¾÷ëõ,öIè§ŸÞætþýŸ}vˆ‰Ièí³Ï5i>òé¯?ù-§4|f²×Ǹ}éÿO?•r~ú|Ûònïjÿîö¦ó«junc1ßÊüLòäòé«ß™œfCô6I|þ%åÏÁþÿõÎn}¿“ùâö‘½>“6F›ìœØôøêçÒ&úû? £®sÍÛ0_ᨒ&úŸïê‘`?zýâO<™ 2ŸÚï1dúQûí3ün¿c”æ1Gåüó6JSJ:,Ìï´Þ–‡¯wÆ/²ÍmúùôEj¸Í‡òœJ4’Çœö¾5’¹}{$;úy$ó+ÙGò>Â]d9æLöùþõxï¦ü9mî2²dþÝgs/Öç3üôaÞœShûôŸ·¿ÿø“æÉõÜWírΕüºØ·dÉû@ØÇïçO‹i~úcÍ)³Y§ä\öeíÝ}΢¶¨äaC½>o{ töao·ÑŸ¸ï=t³Ÿ{&ö¹à«}XÍÞ̯ê÷ÏÛ¤o‹úãCyðýŒ÷'Ÿ{¿5bòùFýǦÛXAþë¥êwÍ·ó\Áõ˜]û|ûÆþçóf,w>» ikz`Œ#ÏeìxþcÜÆJúc,ÔÛtôù—÷×òã9Þü¬d3Âü°Â囹¯ÑÜ·á_Ñrr™Ui.Æ}{Ó¾ºÂuØ×üÕó—`–¡ü/ïä¿c´Ì#ìYs,Ïÿi±øëϲY‹æõ7‹Ù߯Yç»¶ÇÉŽ±NÞúú‹}ëþ'Œl!nO›ßïÿò71{9Ol—ÛßûÐúå>„¾²oaŒù­þlûþj[#´½þç'7OsxpìÎßob¶vßñçØÊ³‰”Ågƒ^~ßqâ‹g†÷ÁºGôþ—[o¾Ø>íßlø}Eùj(;Ñ7Û$²¬ûãý6VåmÔšý|)ãKÆ|ý}>ÚÛÞýiu€µ¸}ï6{"Ÿï;ƒ'hÐoóƒ}þ»-ïÇå˜g‰+?­ñ?y?ëŒ]殢¦¡øÂù­£÷u¶›zþÔÇüÔÏ>ûeË™nGÉïái>¸µâþ»mwðÍóßTö ß—-€?Ûvßuï³É:â¾ïùÿ¨ v¼ùzãtƒíHš½úbÛÈì#çgûPû~(%Uí_ðÃ6f¿þc£«æ·ÔK<“Ï÷í-ßÎ|´ß2†¯1Vå©Ú”o'.› aD¿úÉñÄv¿Þ–öÏÿiŒûGýÖPÖÕz–ûnaß,ìÃûóß‚å9·ªÛáB(Ÿû¹mŸ1¾¯Ï¾äÍ#ûœ½þg›z—+kÅ3ÅÛÜÔN¼õ¾™+ü{ÝíSYÛ\ÌÃkßYþ‘ý›?¬ŒÑLôFý¼û¢ZÛTàŸíTêíì:,÷‚Ѱ½Ç}™úÝÓãâ—Ê{©:ËW?_Ãÿ»ÌavpÈ1¥}‡ÙÉ^èwí=¦–çˆå_Xó4gøy¤ï_û1A^÷“ÿ·º6×Ù«ìƒrÿV¾Z#ë»øvŒ\&Œ_ì³ÿö÷wÍ+¶QMñ}Þ 7¸+ß^òîbóܲ?©_b§?¤í ë{Øþ}éÞñ*ÏŸ‘}ö%ÿ‘ÏÈÚÙçò|ùöçWqëYž¸€Û9ï‹ÒóbºüÎ>ÿöQdë|F5ûœþ)NAÛ³ùÛ¿‡znïz»¿®?ñòª¶>ˆ—WÍN2ùÃî®Ü¸P‹_Küæ×_}ÐÅÕ‘çƒh7a?ýôû¿üÃÝ[™õm>³]Üûç¿úòcúæ—dé&l¿¶J×tö¢ë5'Žytãáø7Iš€ù`6IÿøéG6äp×S?ýó×ë†äoþÛ[?|ñõ7¿zÿü›_}õå¶K‘?ãSÈsï9Uº½à϶§ðg·õšëÇœ9ýNòW_Î-ÐÇ Üó­å&Íî‡ô,å¹ëÌí&ìç¿úæ›/~öAÃéÞ·ß~ñõO?û ÏòœçÆt˜¯>è2ùš›­r&u™|ÍE­Þ»•?¤Sù˜ä>>>ø#›ÿÎùä‡úÈ6iÿ‘mÂ~ˆìÖ·éY«oÍ6ûK{ÿ WbGݧ^ýaÉrº³[ÆÎöÿû˜øTTo’~ÿÓÏ>læÓ±‰û¨™/ÏsÙÑïüƒf¾œÚܱ޻U>¦SyîŸÆMÐǼ©ìVÚ»¤_ó‡éT¹ÞJúA^T­óôwŸ>h‰ªãí|š^?x:ó|.ý‡Z£6iOkÔŸsK¡Ej“ö–þ—§W»P¨7¦ÿþcV™óÂ|»¤ß¬‹â.êÃö”í2{ÛMÖGí)}½w«~ ×åu“ôQÛÿâ··öƒ¦V8]^?È«ÒIcõ1‹ v÷ñ÷Ás«ßƒœtüýÅ_~Ð&9™eü&í¶Gþ3šPèA½‹úÑ7_ýæ×_üü›}줱Iü “6囤ÇÇÎ?À Ó‘`aÿûçß|þ1oK'ìïyY¾~å£Ï7~ˆÿøéGƒ(lx)› ï£ìu=¿Úˆžö9·ú/ôôoÿþƒ:ZÒÜd—â•–<7Ù5?uô'¿ý]8è_¾Aý¨Žžímärÿ(?è<[Fó€š]Öo¿üOÒü~Ê} þՇ̟õ˜ï¯`þ¬Ç<$=-¯·ùóϸ=©iN ç}X¤éUó蜾÷Kûsv*ÏÙkÜ_Õÿñ¿}L¯J¶KÁ¤Wå4/Ù¤Wµ¿Õüý³âŸ³W-½¥R~^õ©f?ÈX»Ý¼ß»õQÏ:æÌ4Ú÷Îη5Ì?ßtáTÿz¦Wy»kÿ?Hûk¼e;bn¢>æ,ÖŽ<ß¿­ÿó ´vœo=ßßÈG›[Nog­?ÀbÛòœ€ZþÛVæüÓþÈaåÏ8þZ™{°§…ýƒæºVçì‡Ym›‡v^?Àbª<–ß*ùsvªY0Ã}¨ŒI¼õú–Òó¦æÁ|<™7>hþ;;ÿ´…ýl³FMvÃ.$ûk>óõv\¯¹µsΙ¯©å­¼þö‹×xýò¥š¡{®m¿Ÿ¬ÿëüÿŸ‹Ìß¼Ô4Lµ¼–ÖúÛ\eß7L·sŽíkÓÔjÂS“9'Õd®‚WÀÙ‚/ÎbÊ?ëÆA0dx bêi V0(ó87€­”…”$ƒ{//ÿRü‰Mãhs_4{¶0©š§W±­…±l2<ç¢vü˜ Æ9woÅ5O—mc "¬1§)«æç\ëë&€°‘Cƒ  ‚h}×ßûT,ÁB³Ç0Ÿlµ>fNSs÷\Êiá›ö$ò|0Å— )p.#NÆŸë\]ÛÖ:`ð7 …Éà æq?]K@“ D ÅàÞëV>熼›îmþó¾à¹;çx>M;ƒÍÜ&<Ç»± ü˜ ºe0Ù(Îþ€ÅA°K°Bô¹<ž¯‹Aöþ–ÀÖBJˆ‚:ŠÁÞWÉL\öhæê´Wˆ9–ç,?áˬ±>¦çÁ»¤y®+5àù,“…2–¢˜ucã ØEX!æa´Í*X@ÛD¶RBRRn½ðŽõé}>Ãìƒpa.ÿ û|÷s6+Ý\² o.£5à)¶ÏÇu–®—v» k!Ìü,’ñ‡6Ýe¶ÒBÒRîýÀ¤Qç—X1•y×1²3© ãÝ’LøGÚ0S¶æœ#'oææÞ6‚]‚O2DÌåË&Á`0ç2ÇÀÖ‚:ˆ@*ŠÁ­þõ<»hã`þkSá†ðç=á“c³ûèîÅâßžã¿W‹Hß(ì©ÖCÀ&‘üq/sUšóÚØZH QHIq¸õÂß×aȶrÌ­Óá/,0ó;5âÆ€­DwÌu&óIÍ=â÷ù1Û<mƒ»5ÆÖ¡ôº˜õâÚ$¶uÿYÊ©ù]}tɳÍ/ÏXùı0S˜i•}aHS—¹¼–ãL6“ 6™–o¨o¶¾\‹A€.ÁSgÃñºÛ}l"ûƒ ¢Žâpïzf#Ù¥y¦ÃËfÎÛ‡½»q; ¾üc9ìÛÐÄ^‡³A0¢¥o C„· f~ öµóþÞD´ÒAÒQ î½À”h« }ìˆÙ|J Ìåϩϣ¹=á:Ǹ­øsYíIà”zÚd´~N>e¶^°ówzbæül[¶/SÌXìÚ\Hñü=´cë'ýñ¶æ÷:&îœÓÊá} Œíóª½ãê‹æiQ@ö`æPHù$Ïõóü'í­ƒ?‡­c&ß+¿Fû¹Ç«{ýã¥xþ.íÔú®?æÁ†E£Îeúj> 3Ÿ@²9§c™©—?Ál vÊØf©=‚âroÅA0dX að9.þ9.½tàïRQÍïð—5,F¢¡³sÃó¾aæã¶Ç1lkà/+9~Ôê« áù<Çœ°±=%EŸï­oC†µ¦Ù®g1˜ Pª›ÀÖ@JˆBJ’Á½†gò}£f²o㓲هEž|,ãÂôVzÀ>œ¾‰E:Ìuq2|[L¹ª¿:q(–d,„}°C‹  –ÁáÞ˜çäXæZøÖ»&DÇÌ¡?Z|–@Ζ›¬lB'ì]K®hP\8JÁÁ Ñ1vj´Õ@ê¹ ¨ûŒŸ¤šÚÝ•÷ïk¶÷ù¤Uçù¾al˜Û†b®Œ¾y™ký˜pƈ Û!×¾Ñ4¿âyVÏ#_qO¡cîjjß(lWS6‚!Ã[“ÌÒñº8ÌÕË6‘!°µ¢–âpï6ñsõ›ïµt;Óàä/Ì|óÙ?ôËÞô<4dd>1Ø7@ØöØ“ßn‹b¾¦¾s Ö‚˜fÞC¯Á  ³;Eö£5” *)÷^ g³×¶íñ” Ù{&ŒE û†¢û¿–&Îv1¦•­¶„Ml…­$(ŽAÁA0døyƒ˜\ü» ©c"€ýéQ QHKq¸÷ƒöšŽIØ–œ½„™[¶î NvûpøQ*û'ìæ—äk~PØöÎA0dx bæLW|U',AKF,z¡…(¤¥8Üû‰~~ž¥Ø. ³¼Àì¦ba{¶*ΧÑù5Ø)ž°pd Š9¸|_(‚É Î¡3Ì5ƒÍ»u¶-„}½s D ’Á®¿Ov…:ÜpdÄ6}¦e;÷æ:Ïë>šç&¦Ù|gËF€6½ÙºqmgÂ#‚!ÂZS|'¾8X¤^"Zé é(÷^àUMÚFó[ñT`Îg޹¾øà:‡Ž9¡› …°ïL³[‚bvqœÁÁƒ¨cfËì'UrÀÔºdhò]ZˆBZŠÃ½0𻡭(—wL ­w6ŠÊå¤ÅÖÇîÖÁ÷2„9eù¦Fs…ÉÁ&ÀÈÎn¤×h>çy7CIa·OBQ„Æä°÷À»”çdé¬ðeÈc>&1Kcpxî·½°µ¶©S‡íF˜,k,ì(vôÎ>Ëž¼€L¸\.IÐÇF1íy1 ¾÷Â$øIÍëTë:—›X¡"9a¿Òö;E;ÅA0dx Ü\/Áy«p]—à.‚ ÑC‡ø*®æ÷N`_owG¾U)Ú×c›™ÃâíòÉfÓt`Ûnf °oÓ+ö,¢0 {ß8¦ k!Ì0·”×à07¾Ÿ ÂÖ‚Z…ô‡{?°™Ç;ðöùÀ«ï3—@¿›;/¿Üî: ϱí‹/`?.ž¾•…%•µ½‘8 6Ñ ££°8{bÉìLj! i)÷~à²yJmg;Ñ$¿m¦Z0Óøó»uËÞ<$ùÊî'SÀöå\ØùˆÂB'úÆ!`Ș-cçMûØÄÁΛ» À&CZˆBZŠÃ½8µ˜ ÝGï[`:f-Ë nãçi .*±Ñ·SѵQØòÒ6‚!Ã[S1ïÞŇŒ¸­-D!-ÅáÞ\HxxŠí\³M2ïÆŽÁfíO>‚ý2ÀöïÝMÝì®`κ+ðŸ§ðÞ·æ‚!À[c‡à×ÅÀÎÈUüµÄëg©ÇÖwý±±u­Ûø´Â/¾Ʋ~á8Þü³J>?™+È™¶fî lÑ ŠF[„kÐ6ŸßØ’ÁœîϾ‰l-¤„(¤£8Ü{Ák>ì±›Ùú;®ùˆ˜oú…My»0?™Åß9`¿¸kþõEæ8"‡€!ÃZwËÁ¡ã’/dt^Q--H±ô‡§~à+K¸Žï‡oÌß7ÌÝÇo…?E¶|©øêÜã°}sœy£0›ýµq þ•cO£¼ƒ6NßýIa?‚C‰  ’bpïL¹2̸eó%"m6€Ãïuw+ÁvèÎÀØ}àé+ßåóŽÙlP%›RÀ¾>Ÿf È<ÒÄA0eøúLÌ<ŽUч2wM6®$ƒ°É A!½ÅáÞX ’;¯à7øï 3|ëq€Ø1 £¹¹íÕM`k!D!ÅaëºtÀmÕî.üu-Ì|麨´™Ê^ºÛ¤2Îû„ýԯ頨îH B‚Ÿq€0§=÷`@sK&|mûq*…T‡{'|^óIœ M¹|KÌùš;_ýʰ‘07ЇÏx„mT‹3 )æ7~äƒà¬»šÀÌ9Ð\’‚fƒ%°µ¢–âpïV²ÊÍÏaKÏû†H§Û½ÎVà(ž+&§¹IËÚbsÂk5Îk|1ì¼æbêˆ9{ä% 5  A@ƒÁ­˜:SæV“|ꦹ7Fv[ÜáXœÜù)ù©Ž°}ÖÕ³ï-йºv‚!Ã[sX‡¹O» ÀÖBZˆ"ô&‡{?°¶k(7"ùú¾aèm`>‰V*·,X9ǰ_p?u‘ 'x˜‹AÀá÷›Ä” ˜m—˜·VilÒ‘ žzÁ…:ùÅ"ÚsÆB- }e¦Â¶Ì×gqËÉ“kÀ~f:q Ũ¾• ðKY˜–xì ‡Ê Éì˵…´‡{?зŸïfNÞnúŒYæM3‹‚ògø‡oÇ-Ÿ¼sKáïWWÇûÁá§T`ª§ëz [¦%ƒ°??hÔ28Üû›ÎüÔ/_¥û˜IÁLVw¾È°A[:ùíp\á‹ ßz0 ÖB˜Þq=!VUoIpÈÈ¥–‚j}ï>´|á´V2©ò}:ý¸gwÝ'lË~ì–І£(æÓ󥆿Š%&c£½§$°µ "ŽbpïøÅA¶ÄwØð83'(\ ˜IÒåv×Õö…— Aa±ž‚à¦E æTîÖ5Ÿ½±Þ…À¾  ¢ †bpëý·±`#B$„¹äpb0¥ƒ1…Î$ ‘°Íé¿Û?pGAó€Gx=bNF9 ®K@÷à†ú]ú©ù½˜7j‹¹p±¨Qßž#DÆf˜Q<†† ?iMy˜àF\j8Áá-ˆ±ýT],l´,}F¤"Žbpï-Á§3õˆƒ‰iîíK…1~«yhRyXÌ–}nsk“àk ¢¹ ¶¡'E(€ßC?6êc?.,öÝbÒŒEú¦±»)enœ}tŒùlüJû”ÉtXÛ±Qt†ýŠƒà&£i`È+8ØDCF‘Õ4´…´‡{?°«jØ•g ºÂ®J˜ÓíȪ[$ÍÄÜÍß­Ëfž0#ÎGÖý)(l£½s ~¼"¦yûâPÝU`Él-¤…(¤¥8ÜûMþÔÇ/ÛæGâ][w¥˜ð€‰øô@Oߟû]Ü©HÐcî?30Vk™¾ÉpÆXð0³@X -›6Ä`~%uØTBRžzžÜÕ?Ewyß0õÀÍBŠÜõ„3Í|à ô{v¬ýApÁ¼ ù§&Ëœ“‘tÚ D ÅàÞ \×\+š0lò…i¶èæàkbêfãËæÏÔôஃÖ$èŒÂá-„Á]Ì5ÝvðAÐCu¡C„Ödpï7‹ðZGå?0sÆ*¼š·£ÅÈlEôŒ'ì[??]ëƒ`ˆ Ç/^þUˆÁiy6 „/¡CPPÇ`pïΛÜ%ìjvüÀ 8¾ž.lëz¸Y.„ÝÄ § °'¸s Ö"0‡O ÁaN›gÛdv«0µEèM÷~àê¥VlUš/‚僚“Y>­f®mBOOpÏÇ ÂL…à(QØÕOÞ8n\zÆ’oÔ×Åa~œî”,€ý¶†ZˆBZŠÃ½å%€°5  A! ÅàÞã¾ /Êû†é°ëÈ’aËUF„.Y | ÐýÊëw&åˆÖ‚åý²a`Ó pGY䞈 T&ƒ{ЭC÷Àƒñ¾a*.!Í{7Ç—/Gf<)`Fk`ä‘¢ÂÍ58†ŒÇËÂ䊺8ÌuÃ7ô’Ø×v]?“BZŠÃ½Š%#dædׄ(MÈÈ·3÷hõÕ}û¡Im‚¢»;ãâ "|hC›Bp M!d¶ÒBTR nÀ;3[}èp˜ä—j~pÊ ^t~½bÖ…°{žágç&nã0dx `Ú…7â`®÷ç&ƒ0Ïð…´‡§~ÐH¯lË•Òq1-Ì\Ž| ŽØÐ~ÁaМ¡R€n#@ vX÷Ö^ ¸‰@˜ƒÁ\«Ü¶M ÝB@D:“Á½ØŸ¢e¿¦˜^nvCPuµÝ>ÝsÀnÁ†Ë$1sȳƒÎ`Y[×k1H.y‰HYFr)!ŠÐšî½@ÏjÆýwÏ÷ 3Ÿ‡-ºW½¨çôåbžah&lr;’ 8i|Áþ" ÀºàPý °dÔªi! i)÷~ÀéDôj¶pUF “’s±ˆxÄ&_%íѬ˜¯]+wþfáø}k0X{ 50é*ècó4,K0hþG­ß©™š?éÎ;¥‚Œ.èóû†A¯Ý¯ÙŸc¼cºxÁ3hÑ´8±Ô7Š‚ åà 2ÜöD "‡¿›ØíÔBÒRîýÀ8LŒç°«k|a¸EÏJÍûúiÀcX Œ„}ˆäàE;è˜!Á3çÆá#½DRŸ²‹H´î.%D!ÅáÞ ôÌ.’}ßÖpä\˜ù|’iV Á]^¾å· ;[&DJˆ r¤°½@ðÄ0B ÿ.›U`hrªvws-“”e Hy·Y)kCPœˆ; ‚Ãh%ŒlNâ ›”d,«•´ jîýà&¿û(7¿Î‚ba(иpzéÀÌ}13(¢–¢!18†ŒÇˆáE°8ôÌëFÊÌ~JEèM÷~ÐþqÒœ1¬»¥™<•qrS=¼òV•G¡ß).$ô ‚#¬J;Z¹ï9˜§-»’A˜UOæ' jîý çNÆ8:üÊà}ØW¸:.x¬Ã„/ß L“»DqÁ /8† ?÷S¼rÕkp˜¯ÞÆAØÐ"(¨ep¸÷ƒ‡ê~XsÁ€³D`,o\ZyBJ:y­Yuà°/:÷d¢°@°±8 —Àd»§÷c88X¨_Ûd~¬ô*¢–âðԻǒ';ˆ-&ÂÒüÙÔÕ-éäüfŒ»ŽD¥„Í_åt?«E1ðN‚àS9ù„èXãƒÁVî£&€­uTƒ[0ØÌ ­a_,ÌÜ úàJÈhiþ@Ù7¥ž\” ïZaþ‚š}¦‚!âñ²0˜ô l¼–‚FO‚@*²ý½˜ú3]J‡†b`ì’£âØCÒ/¼l»¿Êú¬ŒèÙ °¹¯m ÄÀÌß³%‹C¯HµB­ttƒ{/0ñÏ3£-t– å@ü>1–-Åwxs÷mÇ„ƒ9´ò*øhü‘E· ÞÆ@0E¸s1Û q`–†AØd@‰ Öbpï²¥˜eÐã÷Æsì[¹¤»´eç´¨u³Æ,ØBšÍá®/Š4ÇLBà0dx41óY[ÆûàüëZ2[à´…´$‡§~0 ÍÀKhûÿ¾aìžs»7÷ä& c~ŽEpA=óbØ(`m‰»nPØV§l†ŒÇËÂØ–âu1˜Çæk˜¾LðB%…”ƒ§^`/|ò4QO¦Ü LcŸ¸!‹Ø€a> „ð±p34)Ì»¡.ë©°;"Êx£¯ƒ·.Hü ãY–¢†`ðÔÏ`~8,žAL•“ÄAŸÁ ٲϋ~Øòñ+Š ¾kÁA0d<^&—ˆšAÌpÇš-€­…´…´‡{?89zŒ¡:\¸$b`Ö:¼£|­ô[£Á:%‡&Ë?`¶¦ `X[p >7Óáîz_—د©…(¨¤Ü:¹?޵,¾o˜Ú‘t´h|$¨ú´òί˜ß)¼Ž‚ƒàXz“aLvÇ×7€/K QHKq¸÷ƒ~› S?”ý90 ‡ï‘©ž8N™ïB Ø\*™Z$(¬lÉÎ@ð¡ìÏÂX·{s)9É8]AÉ l- DPÇ`pï,às´ÍGB6šÀx‚l!믞­Áóþ$æ|ì¾ó«‘(ò…ÕŠí¦B%‚±ØZ:íÐŽÅû(¾ø^ü…ósWDqì"|M„Ín­%ƒš¾vDq ®ÖRBRRn½à«ÒD™(za²ODîÒ”è#sÂ¥Ê},û<ËUPdDÁIÙ—s âN¬æá¹É ì[,WB¡#<õÛfò°; „ÔfÀUòê‰ejLü¢K!aÛ(À©pQ Ü—Áa-„éH/츺‰pÐ襗Šj~ï£ê°Nš)¡ã• ÃpÑ–aðeÒ·—€ˆªk;ʼn-c0 —…±Úòº8˜Ÿô.ðã%”tƒ{/˜Ut:æÏ÷ ƒt¥Ú@5 ›!ì'Ò<°gý¤´ j'œ[â9=Ø<§Ö Î.ÁðcËM" ©&wåéÍx†–iÿLO‘žöŽÁ@0¤È$ìþ ØÎEGÂà 2ü¶™˜ŠpÜàPp EGöÐBÒRîýðÍGcÆÛ!Å|`êY™Ç?r¸+|¶ô9À¹3ðz%e#¸:²+ˆ`ˆ°˜3ÇëbÐ< 0$4z© ©Èö÷>à’¢°âGë̘˜î9f¼€gãv°'”ø<[À¬"Šø:RL5êÆ€ $x õô²AÑÜ •ÿÊ‚ˆ¡—~j~ïÙ«=7Ô\˜âbtzæÕÄHÅ tÀ´ÿÁcŸ×žee!Áó×(@­Á`(‡'z^"ûÁŠs é(÷^0iÓÞY®&#o7Ÿœµ”l–ÈÔbÝw÷Tݧ³xøÚ¢¨'£‘É@0Dø ¦6Ô)ûÒ&°§ƒ"ŽbpïV°ŒÛŽ{À+VyówE²SìÀí|øÀz‡$Ä$èz‚q: ÐNéu5çÝŸøôs:ƒMñÛÞ_FüÀ\°~t°²L¤ný8+ë:«0Zô¥ÍÛAqÁ©#8|° c`‚ÉÄ¡ŸŒÈ¢Œ.ÃuhÒRîý@߯ÉËL¿s}ß0ôÀ¶Í rDcçÜ`6"hÛ·„P¢ È%ƒ€“.z†.Üâp q¼Dºç H°´ƒ§^Ðý`¹¶N'«À ËéÉ0yâÔðcZƒ‚&N4î›4W\|E󊢎!°ßESQH?q¸÷W…¾ãDÀãÂ\8zX`:{ºoF«tƒ­Ì4`éù=̓(NzGˆ`ˆpÏbN–&»Ò=Æ’Aøñ"%‚@:ŠÁ½H)ÏúÊi¹wD`ÌZgeÔ­€ÇÀIg~C5£ÂÁÇËD\´ú’ ùe`0"¬…0æzþº$E Í™B* TDû§>ðäÌôùí¼0¾ÕGFÒÆ D ÑóäJ€/Æ 9þÅ€ $<^âðsG4÷Ê{‹?+ó-ø{hÌæ÷`¦3*bø‹ sº¹¼ÓüÜ%½ 5òP½8sZm£hX|Ä@`W½¸…A |0h̺,Ž„K Q„ÖäpïEä4ñû*¤{ß0£°Þ&"Nlÿ7˜;ɱYX.,‹¢°¸·´q2/ Óëm’ܳm…€ÝŠC-D!-ÅáÞ:Õ²2öqÐ{}a˜¦°œ}眵·“h,7òª’Âò’ì >hÚFCô0$ƒ°ûÔºA@ƒÁ½ïˆòíCu°…™ï8(óé{sÛ[ gz ¼gIäωáêl0øûÁêd©8ZÁÉà`/e¼¹¢Äƒ§.pòÀ¢oŽNp8ÙÊæ·ÌˆªŠJnñ„}î¸"ÔÜ)ZPtç ØEøäA¬xlͼÁ?)Çrh ièÍoúÓûJêÚM-ÌÅS/Š©Ù­Wnh'Sâh7•˜š"(˜Ê"8¬ÝT`ë.ˆãÛCFÕn*´…´‡{?`Ùfis“(0m S܇-_õ¤ïÜ…ì!óa£~:w.ùêX¿E! dá•Ü-Lª¨ÃDV›b—Øï¨ƒ(¤£Ü{$WfÞ³ nÝ “Ý1\Êy¡‡áqMØì—VóòÜ(2\Ä@ $øÅ+1LÁà@mÀØZH R„ŽäðÔ šY;Í"HhJæB4ÊÁjkWgx&ÞX‰<eŒCÞ$öDzô!7WPØBºÚ{ÞP6yzz¶Î,¨"þ9J®Hƒ ÂdpÓ¶©—^“ÕÕ…‘eÁ, >äçêk¡Œžô ØLG(í¹(*.[‚aʰ1+T00kÖ¹‰ l ¨DP„Ú`pï‹k æ¶jY{“ ¾ó멌‚[Lä6½T¶ãRe¦YQÐ)%8¦ŒÍñ…y€‚ƒŽ/› Â~æƒA!½ÅáÞ\±Èˆ¹¶ c`*Q –±z¸ë¹è[•LŒIAaK;ÁEA~±Ù³¬˜&sÀñŒ³’‘™O~i! i)÷~À–sÀd0º¶UAá©ân”¸à²#àÅ7a7·`Wû…` ¸kg˜†TåÁ¡aã2š¶VRBÒQ î½€áãªoÕáv!*®| á=\!Ó[V "°ç¸3ì/ ‘•ß— ,®£- Ý“* TÌÊλ÷¦ûùlä„ Ï£…çPAù>Ë à—V6o•€m¡©µŠv†+mÙÁp=¦ÓõH ]$¢…ë‘”…t‡{/aÐQ;¢ äYX$Qòˆ³/_î™P¢õ£´,„6ð[¾hdCÁ`ýxYKÉçŽplžü¶ÈÜÁÇËßC36ÒNø…ã¬3ý`’—ò139ÆUŽä:î ••áÝâqàùK ó­®‹@Hx¼Æö)~-Fr–—P"(¨cp¸÷‚“ýůپÓ÷ c'Á6æ–/3LÄ~IØ'ûÆÊi¤h ÅA0d<^f¾ {§U)ÕbÝ(‚0í{~#& )I÷^0š ÖÊû"}£"¡켈ºígö[°ûË0¿–(Ê`Sr Þ‚¤v[æäÖú&°{¶S Q1œÁ­¸f±Ð†Wøû±%Lú_Ý{ÇÏ–³o4œžûbP‚bž·<;¨8|èÒD˜Ì”ýâ òrKF*ÊÜ--DA%ÅàÖ äf©XÒ,½×ðK–À˜ÑÉK8!ß•ù ØñÊ<_Ý‹ °}6‚àdáE1|Òdaú‰šOdÐ}wô2TÔ€¿KAµ¾w¯kêâN.…Gè s! Ï*ʼ3u¯ùú¹¶_3⯣uÀ<Ú. ʲ¹¢n̺>¥“ ´cë'ýy‰>ݘ;-0逛\n\ˆ®¤*jñ:ø€;kù{Kœ Ù\0<^c™2=½Tf'—ÂÖ‚*U ÷N`{xð>ê 'ßÂÌ-: ˜¬¸ã~z¾P;Ó“Ï]GÚ(Ìsç ø —]`,ÿˆY8ÅÁò“Øp’ Â~¶¯,ä-þ€­…ƒ{0ÏgDˆšéÔV&ùa&Ûªpмk9”l]J=`›‡+pEC–¾à 2|æÆ<0QbpÐ"œX‚2{9@hÔ28Üûá÷b{äzXÞ5»[/³Xmá¸<×/AçõM³`»åBˆÍ¢ÀÎ~qøxÃEš³y¾.ðX^"[ )! ))·^à ¹P}#‚(ÔÀpzíÜŸkzíÙ,/å?ê½E²c”pg>Lq|*R`:‡†G!£)Rh! i)÷~0•ßtf¹£o ãwD¬]¹ÕëRDa{J‹ˆrp­ãÝH€YǶÀ$¯qíççâÙ»$°Û©‚(¤¡ÜûÀbŠê·vÂÌ}Jö±ƒ34-Ò5æ ÏÛ¯½Fu^o-üé[ƒj—Å+!ŠÁ\ùj{F-éúYÚ©õ]†Å!½‚¥‘fˆ•0&'NÂæ3çTÃßÁbÜŽ…qïí¶1¸)½l`,nÎ=ÄÉ¡ ö´D§D ÅàÞ ÐÈÜÜqõ}Ø÷Ÿûa‘ug’k]œf(ÇQ7Š i´‚ƒ`Èðpbtxq¨ð €ý:‡ZˆBZŠÃ½Ìj€8ÝIcù,àíùaçqǾ¹b»±ðu KÝ(ŠçV{Bƒõ¤q"òQ­'ã7ÔÂÖ€…‡{pCk¹¹Ü¦r0&1Ÿ¡e÷ò öƒÕrð;Yô}Ã\ƒÇÿî{ 3’ÀæÝxMÕYÝÏ®ÞjÙ(ìP¿„D<ã ¹9n¦Ma÷ ‚ ¢XJ,@pë§d94¯Sø††åõÌÓÜ«×+ñ~Ò-ê`nþ5¼1 —…i¡\I9Í=öL?^–¢ŽbpïEÜ!:±q˜&U»¼ñ‹—›Äd6é+Œ.X#†^ 9¦u^9“>?Ipˆ×G×ú\CAµ¾w!òÁáøy=„±d½¯žÕdÀðÌ«ÄHMÊö³´ÕXà±RzãîytŠ÷`ÏÁŒ}I_©[´¾kOÿß‹¡¦Ã…)nLG)?øî+78„Ý·Ún)Sé‡.Ì…r3–éÀïE/ó„[Âît ÷À  p î MÌÙàKNVH®,ž:èw©¨æ÷Nàþ™&,G¸gÈ\˜ëdelÁΣ2ßó`ˆWe.»õ75ƒÂ. 7!ÁïI€8/£¢¹ùc¶Mawà.PRQî`ý¡¯jñ¾aæˆñû%šfÌÎï£N&,lëÑ ‹ÉT‚ƒ`Èx¼ÆÌK> )­¸¹=o2?^B‹  –ÁáÞæ‚ñê8Y‘aa†Mg+Ûd ßa©àq±fj$äï9c£¹à“õ6ÌJHbÐ3Ž ’ÐY’a© ŠPšî  ø@iãÄÜ “ÝÇÅóטmÐ}úYv×np+í\Ú(xë&§¢ä×Âþšƒƒ àc“AØ'ÓIæŒ"´$‡§~0ËJ£3)ÌûŽ—Eú‘üPcÆ¡ýPÑY>~¬3‡€™Šfaë’ÄŒHZ…´‡{?®ËŠÕm©|ß0VHÅ+É7äÆ+žÌ°š[ »g¶ülWaV3P˜§¡'ð#‡€!óÆs˜Q‡Íù(Ba¯U‚B*; î^ø$Ã-"{ëû†)Nެ6¨ú«® ŸGCîé è8DÁa-„IÈ",v—Ø}©…(¤¥8Üûó—n[Þ7ŒM áfWl\n–÷’§ˆJƒâ‚ƒàØ+ædþr0o/?þSa¿1…A!-ÅáÞy½);(¨ž$„[.=>(+ØÃJ#Í™ÈxöÂFn,Z–æ+m>p`ZÄn,Ë/IÆš¤…(¨¤Ü:óÛßhnÎß7Ì……ñ°sžm6/,ŒÇÜ»ÙKØâÓá’ÂŽ’yã 8Ž9 âÛÅ[ô%#Ž¡0KKpxîfÇCuË“r°£¬nÉÓpÚºˆ29®‚Ó=ŸXÛ) +•‹à¡2-¡/[pPV7Éì'Z(!é(÷^ÐÃC‘NCa§Â0“yÖRÏLæAã·Z­øeƒmRpC"¡úçœÜ“b¸`‘ —ÐB¡%9<õóþü§]¹ú´/öÖõ0±.£šÊ3ì„mV. É¢HP3Òp‹`—`ôB” Û·I.€­T…Tƒ[)ŒÅÀ,BL­.LÉŒÚeAEêô†zšˆú­iýÏÕ\p^©Õ‰IÕçÎ`p0qº$‘Z]*ˆB*ŠÃ½˜õ-ÖåŠË»÷ “”ê Æ/›w|ùË”à–ð¥O|~E‘Û9†ŒÍÓÀF=‚ˆ,ß{KáÇKh¡79Üû—Æ´AV‰uDc ZìLŒM3&Ͷ°-GN¥ÈU$ ËE]¢½ ð÷¢)WëùU}ãØøKQHAq¸w–¸Ì:xá¾+Kqa/ëó8?‹^½ôbd.‚Öx¹G‚Ã{;0æÕa¬Å¾ÙÎÛ¡ƒ¤£Ü{5l0¿X9áÖ·0çvËNî&s:>ÈKد)áö Öbp9å0*Lv¯·`u¥ˆ|È “: TDû§>` &Ìþ–2'²Àt\LÙÞ÷Œ`-XaÛH©ŒjEC¦1 ØtG&DEý¤h^°‡€¢;²PARQîÀqì¢Kä`‚ø…Á W¶§Î;1¿??®¸pæ{q‹æ gƒƒàÑuÃ.Œ_qƒ‚ê !‚°5 ¢%ÁੜÝvç{P&ÿÆ–wy±äÊÕž³§öIž}Y0çï#oůâÁáÓ"1¹Ç’àè82?^B HG1¸÷ÛŽz"™^MŠÙ¦ˆ¸™oøJ(7è9ÝjE ôl/^nŒ§A1\“Bv„qÇÿÅ {±ÌÐ襂¤"Ûßû@OL7ãaq Ãt0æoÛy¥Ø™Y}H_«l¥j‚`ˆ ±c˜&8Ôe >^–"Žbpï 4¯ÕéŽ$Íó a!¡Óß±pÃíOìš1mó쀨߇—P‹ÖÁÝȉ°pR3ΩõÔ¸íì[ÉÕƒ›úxWVÈŒ–8å1|£âŸ´ÏÈ76¿ø“÷Ò•»–õÖG¢¨•néä "Ü›˜>èŒM=¡LdæÌäZˆ‚JŠÁ­ôVD<ÖÆyß1×m¦¨år"Á~áÁ°»š-!/6xpÏKR8–9RI %‚°µ A!ÅáÞ Ì‡….ú6E!u©0µÒ[9R¬ÆŠ‘nöûd¤} |V‹ƒ`Èx¼,LFâçà`of—øñ²´…´‡{?°Še¿±·' Û›y ÖyÆìZðÖó>¥3M‘çíEÃ'V‚ß0~Ô^lì› À¼c¼úF!­ÉàÖ ú‰!ÑÐY ¦¦”ÇEÍÙ^ ¸¯A·À_;.ÔŒ‚«ÒȦ°8dDÏIDVùÐAÒQ î½À†ŠÆäR*ö SK¤êô3IÃluÐßåƒÃ9µ¶Dô“)tÐïT1šß;͇Ç÷¶ +4Án§ÒŽÂEvHµüÝäÅDÛŽ`‘HÛ)¬P‰Zóo0~¼,ÄkOµÅj·¸ökOÊEk[óMkú”f$´+Š» ,# ×lN¾µG‰€¶¢(vÏß=$¸¯æE݆YýƒÂ+—„¬¨ûPAR‘ž:Áøt$Û¨ÈÜó¾aÒ½fË£çò6׌#à¦L¿‘C*ƒƒàȘŽ!8 wÏ’ù‚B QHKq¸÷ÄJ‘,Ÿ¾À:éWO‰áQ±~?3XØŽ>}v¡s®ŸÏ†•R­g¹ô&Á!OíU+Qz3¦…ƒ{˜•¤²6bAª·…1ë±]Z¤7Zí³4÷Q$¹×•¿•ºö¹OVcek/pèÆ?0§›ƒA™Kª•Ü”Â~¯â:4ƒ{¸‚9Ù]°}""³ô«q>ã>ÔÇWÛ.LçÃÄ%¶&¢`°Kð’ˆC·Ò`Ð/¸ÉQAß{&yHBnrpNQ󥤀éuÒ6‚ÖMµ<¤ùÝ0†Ø^^Ï@Ø·(ôn5 ÷>øÛ²¤¡~?b“(D¶ÿ|Ãv&–O9'^ œ2å“~äø½0犚¥Pa™È…™_¤Uƒ„ò‘!°µ8äDB ª,·°ªÓás§×­¿ƒ0VÙf ñ^ Çê+ë~Ø?kO¼¾nŒé7"C†Ï5ÄXrwŸÈ¡£BuÈìJÔBÒ’žúÓI=h(<áÙ¼0§—ú¶³'JïøGíìëù3Ó}Òó؉â,¬ !Á“ôÓüÃX š;,€i~me£ŽâpïÁXs9ß7Lö=©9ay‘âÒ0&彘ž`×F ¶þ‰sÀq[íçòô+˜~`®‚(¤¡Üûq:n;ii‹ÓqÌ<(›º™–­0TM£ñ`Fê q8(N+ P‡€!Ã7Tİ|pP¼ˆb œèBPE6êV±Ãcmóu•7,b@¸Oª}©VUγ^ô=[®d[" ûT¢1AQܤ ¤Ÿè¯ ¯Ö^™«-þ‚}¦(+ ©,·.ÀÞkvþwd¤) ŒU,xõdÃ8'##´ÝŒâ˜¬œÑvy 'RœH£!u[±0,Ø pw°Dè.a)! é(÷^0àŽc5¹æû†±[‘¼Nž%®x©³|±0aß(pÂEaõq LWŒ;Ì/‡m`—ˆûÙ¥~_J£ùS'8'fÄfÕÊÀ4¿P›3Ôí©òÂ%/¸sdæØ6ŠUÞƒƒ`ÈðY‘\\-vm¹‰pÐ/˜¨—Šj~ï"Ô‘ˆµXºS” LH\ƒ¼yæ ä10ÌKØ,̇‡œ‹À*Á»•í„£ r§¨}æ4"€m¢‘ ¢Á੼-‚#]ÊQ NKÔØªQÒjS¹-‰6á~lɹE„(‹ƒà¥à„aÙЕ-ÛwK`…‡”BZŠÃ½0á\'ƒ‰OFÐÆŠ0¸U£F8±É&ìà û}0%E9xsK‚«ìí)Hp,Õ[Údv¿,Õ Oô¢–äðÔ=3kè.,0n²l4ûeãÁBŸæ8Ýö;g€©X¨ïß‚Á¡›°@¨Š'š[ßGRaúì )Bir¸w‚GMV˜C÷÷ c帣õšt¹21Y‡‡rl×y³%3‹.>b¿NŒ‡è¯«Ëë‡üÍ•ßR¤o† "ŠlÿÔ~hË·û!ô}alãñJýÄu¡í²íèt}¿˜Ê}4†Tã÷ÉÈ]çÕœ0<^fÀ©^ ,·Í¸–—P!(¤´8Ü;UÚJœ¸]Tµn`TÞÃD^£ü¼G‰”€ýRöð:(.õƒaŠðÄ Åñ€ƒLçMa_Ø]‰ Öbpï‹Î0#ç<þ!j@ˆ´B^ÏúY>¾9à‚ËŽ¼QTÜÌÁ.ÁMºDpw ,#Ó.0/ÚÎ@*ŠÁ­‰-ŠJc) Lñ`–öɆ¯m¡‡;Þf¸æeË5×\¿KÅÐ B‚7P}q³c0à#DJ—JˆB:ŠÃ½ÜÃFfÉà¾;×Ù.3Áª–ì\g¹È“Ï<‚0ÿì[!eg "|'¬ùuØÈ ó\eÞy!°5¢Žbpïz–X§Ü÷­ï¢UìÏ=n¶.ß¾gV>L=ϾQTÖNÁ.ÂsÅAÇ`@ÇØ{F%D!%ÅáÖ L÷gºr%­…iô^RiK_ÜD`ÙöAÀáöb f‰à€,î’ਖ‚lýÔ‚K¤‘î4Ã2‡Oèsœ= 7Ó®=áâ[Ï "ü$†7“.d#ƒ”|± €/KQHG1¸÷‚οƒF e‰ Ì…Xø”˜ÆlÝS·­*ÈŠ Û]PTäT ‚‡²DÆ #Á@µ±*VJ‰ ’dpïÓ xDþ”u2!p`¼0ª;°z CgÊÚÞ}<ôÈikW ×F02«²£½@ðôÄÏÞ¶÷ë é…ë*ˆ@ŠÁ½p‘=`Z9ó‹ô…)žì®ÚÓð ¯ÅmEÕ,GØÜXi ;þŸÁa-„In¦XP_2ÖXZˆBZŠÃ½øÎ*^¼ÙJN7¦Vdj˜'·6Äÿ#‰°;»!Â3(Îq tsÌ<â{®q°¬a» ÀŒ8€í’ÒRîý`ö°È _; pfv)²²@½‹)"ý”²Ìf Tm!;'^åâ@}u1°m­+Ra÷3€AAÅ`ï_ذ9É’µv삉¨¸í±zE^¿ ¡zY‡@ôôÆãïæ•¶æ‚ÁßwÄdd³ á”!ðãe© j(·DVN'äïX˜ÂB%[`»2tã+«]~ÀûÔ·À¢èŠæ&ÁAßP‹ß'GÛ "´ÄûOR-­¸ÀMyzJ5:œŒ| ÌÕ‘²+e:–œƒç6Ì·ê€ýÌÌ.¢(¾-d@/8G¬ÍîsÈeF„ÝU*³2/)¤¢8Ü; pïotG'Vœ!ùIµºvvÆê~2Eq^ÂnEº:QœÇÁc8¦·3yy…;r0o¯€7T¸«6µ j-·N "=Á-lÀoð}ÃØÌcQ/ØŒîÆ»jnžv¹3Fµ)È«!Š¢ÁuM V^[Œ=f ²—ü]"2K/%D!ÅáÞ *âbÓÌÞJÌϹ›!=!óLÚt“-~·hIxa yÀðxY‹Ä1Cˆt7À…7ßgý.ýØü©ÜfT^kav¦ÜÂ_î]gé&‘~b †°oÇ[Új£×Ú lasÆ–ßp¢}¡AMJ˜Ü¤‚(¤¡Üû3é‰ô3ž²ÊïûcQNî˜øeÑ]ðâÍ `·“ëçqœ¬8VHîÂtÝ÷‘Bz— ý.-D!-ÅáÞf¼Ið[Eeá÷ ÃêÖÿ§ÒÀé`¶Øž»¶¿í³‹¥l CÄãeal'§Ö,iì£èq( )˜óë³þèSÕœ[TÞH˜†³ÙßNîh2 €(ФܑMR¿ŸH>ÍCÀãea ¶ Á ûY0ddóZ ðwé§æ÷D2)X¨™°ba*R²'žGMMÏãaŸ0ü¹2ÅbpQ C`¨'ÛÓÒ¥¨g© ©Èö÷>`ç=…]‰y ø…9OLIfñ[C<ΗÝ|\ÛtkOÎÂŒ-uã 2¼ò31s‚><9´ÄÚÑ”ØÝs¨…(¤¥8ÜûÁ°‡ág3Ëj…™#0–«V7•ó†­dŽ$ü@­L¿×!…Y2òÆ!`Èx¼,Lí,ØBåb–kÊìÞ÷ÔBÒRîý "k# ¨Ëû†9+3ÿ ß‹íþýæÇ*ø@¨ÐqmXƒ`ˆðC1 [ŒàPYü™"ª¼tBHG1¸÷‚©Ý¹h yëû¨H¨•TÁMÌñê¥ÿP‰áðgÙK£¸©×KCâÃÆ2h%=1YúÕ¼2Ó—¸öÀ ©+wõOÂøb]4!®±G?eæ€+óPŽ,qï®Ñ¤¨G¢¿2S„÷šæ`bÞ«]„IáÇ ° éL·.`êha&”¢ÆlB¥#Í‹©Onh˜Ÿ=NŠ„½2œƒÂôÚ9¦ ßÛ3<$%8§g „}z‚A!½ÅáÞ?DFY&e “ýâyÂ~*7ëÜü6ka^]Âó|WYwQØê²3 ÖB˜Ãk¶/H¿d(}|(!é(÷^À¸v$Ö–V~¢ÀXüAÃöÛOd:ÒÉúKÊOä–ʼQ”5‚å'ÆìÛ¾Ã's0ñeæžlÔ1Ü{›)ÄC]4Ö¦ÃÏÃ@Ôüae‰sÀë 0]ÝsÙ(8yÁ¡E½ãBÝo+ÅanPƃ2sGu¥"ô&‡{?¢Œ‡/ɯrÞ7LFj>ËOr0àªòÚ9"~ ŽÜfHÑÎHëCÆãeaŽÎ58äëâÆ;1â]Ål¥EPHKq¸÷ƒ}Se儘„m¹ÂùÖ~¬ˆ˜¬™Å;<á[=³ŽÚc:õÙ{5?«­…]y}ÜÏ<9ÛKrP(SîºÇl¯ºÓØ*†Õ{4×2s™ÏÆîR§Üf1_‹ÂâI݃‰Væ±…Aù1`æ²¹ÍVD­(¤$Ü{ÿèdâÎïæô±›¯Tãt´¹-c‘!À¾‘Cå” (žž(„§¢¹SÝj^Ý%m ì¹þ©‚(¤¢8Ü;Ó³lÁ(Ûú¾aP¸uÎâé ½ä‚[$ÝQTæaÈ-Šîyu‚AH0z"è^Í-£JÚöÓ6U…T‡{'à¶1NÌ5×…«”À˜û%jÄ4ž ut3ÌÚ¯"'^êƒ`Êpï{ad[#3þ]yÉ ìÅZ˜//>!½Áá©XÉ:ÜÏzq{˜†ô –ÝßZ,Î_ðÙXm…¹QøÅ%9\u¾0‡î»œAA©Q´Å %DjƒÁ½|kÛï„gžþ„¼¢°'þî4‘­Œ—Y¿ûûJp§EB5{2ö(¬j÷‡I­0‰'9Šæ`Á]0ðtJ›Â>iU…”$ƒ{/h®ÇWk§™„I102)2í“%G=ýܬބ8·!‰ )ìÝí>%˜ ?¯àP¼ó’Ø'^j! i)÷~`R´Þ†‚¥‚ó¥,0‹ÀœLývøùÉ“zÀ6kYV¹kQØ•‡'r2< ±0¬CVŒ©ä%ƒ°‡üA‹ ½ÉáÞº.wfg·‰í}C˜“›h®U/H؇?Ôã!E¹X‡gN¦H™Ë#Ë'EšMC QHIq¸õÂ_šy®âU$¦šÌå¾gÙ¢“=Ú{>XK$mÛ Ëà@x>ÐÙØ¾‹b4„¤‹ƒ`ȰÂ`õ ݽK—ÀÖ@JˆBJ’Á½+µƒæ®©Í8Å"o°…ÝµŽ¼cÈ{ º(“…jˆŠƒ`ÈPj³ôªÀ¼à`x¹-„·äfA!-ÅáÞÜ·Ì´—MYïÆv3îÓpDèpr'øñvOÁñÖ6{%;Áa ƒ3›🠀­tEhM÷^°ªŠF[®+ÜÔfN=¨ùïïÂTù Í#ìó2 ˆEA,Bp îî\"Ýj+“ƒÙìv€ýƒZˆBZŠÃ½Qâ¤õwĨ±»ú»>” ߨ“îþæXñDžü]¹–Õœ0<^Fm2 CÊ1­ß¥±šß{¿½Æo$7¾±À˜³,Ó„æ/ó“»tǺŠ8W˜åÎ@0DX aRgæ@r8à¹"­ttƒ{/°  ø´›; Ñ™ø‚Û?“Áe³4«&ú‰;! nP˜­yãpƒ_ÌBTwœY ¶²Dö%“J"”$‡{/°6o ·'ðй0—g|ÊŒ¸0ƒ‚W´5ÿ¤ð94øwùÒ‡¢‚bpïf†<%–n ±é(VfÆuLT‡§ cie€~™ßh:Åï6íh.ü µ`PPì,$¦}ßU…T‡{'`®˜þÍR×QpPs¬5{SE*"ûp<“FSmxÞæËnYÛFq6&¾$Áá÷„-*0T³ ŠCoð1“ ÀÖBZˆBZŠÃ½þÒ.ä +VcÅÀTŒ®ÔêÞVe"_ÏGz•ÂT¾¤¨ðt‚‹ü“qÛÌ,W6€­•tƒ{/0SE:¾+)q1•fBÙ“+mb ÈÉÌæƒ³Í¹Üx6)A^6¯^#vÐˇØâ$€ðã%T )-÷NpúȬ§Ë£Àœ%œm‚š :Ò| #¹­¬A0T w]#¦á6&8ж#U.ttƒ{/ð¡]º:I¸ò[Sô„#¤ð&$ËNÔ|¬bAAÀJ=Á@0D[Lj]¶-E¬!Óµžªoꎅñ£z[ 6U„®ÚVV¤÷ªº!_ÒÏ›ßÕ[gÏÑUƒ]—§ü+¸Ðó£H‡YÍA:|z”oð’Žæ:ª“ý Û-L0$D˜`¨ ©(÷Nø(ôŒ«îì@W…9ý¢?Û)±ø–²¥‰õÊeƒ®Ùòßy¥@QØ}ÎØ8î¼Æ_˜æi‹CuG€%£ÒU`i! i)÷~ oéú­@ »x,Lv7‹)·ãxW¸ôFߺÆ@¶º`ÓCŠAÇ q ÖB«™eOCwõZ2ûó£¢–âpïMÂHl鑺30JfcwS+n¡ýöLû:S™º‰vAÖ7‚!ƒ{ÛÂ5×c9› ê§/i’AØ- Ð"(¤¥8ÜûAŸ`¯7Ða_³ˆûÞa }_ÃáÝ6¦g ˜>ÁˆU#ʼn‚ÁA0Døþ™˜‚}pÈ+I‰ËÈ‘òIZˆ‚JŠÁ­Ö¯dÞž™exgÆK¤‹»žlU™‹ÁÌõø1[48XEǶ=8|ày. Òñ-ÝMK`k!-D!-ÅáÞÚrN7“1ªì}Çð[NLe„¸2wîñôAÊfIÎü K¶×Ý9¬X¶À4ÔJƒZ8'SDÕ%D!%ÉàÞ Æ—bæväE ÆéluÏØ—Ù­¬ç},p™'lëMÅæ0(6Á0eX aNÓâpZ:©²döÍ´ é-÷~ðÜyÁïçð¨á÷ ÃÄ÷öØyÃLäÉ]þa%ìÉæ9(®Ì²^ä 2/ÉW{£gº§± Z€Âüœ=í§(¨¤Ü{·fu²Ýž•X§:0VôõBÈ.aa:ê…È¡DKóCQH'`‚Ø` "/ c¹í˜ÉÁ’±\KAÌ+é(÷^`Uë¨î`ÞœÅÔ™o¸Ù‚ÃÜw™Ë¯…µÕ€}‰¢—(.„<ˆÀƒ)CG½ÿ¿µokÒä8®{Ÿpè7LøÅ³-ÔU}$m‰!ù j#ü@éa9Ðßb).H þõ®“眬êog`С`pñuMuݺ*+3O^²d†Ù†.ô÷¦`Z×г…ë,h §DÉ‘¶›–Ï.Aæ ·pPá‰î±‘;ùÈgØ©­µ×PrçlÁÏìo¸dçær ë>ÉG›}èùù!G‘5 %ˆÉNô;²$íGæñ ÅçÈôË|¬YrÒÜðã&»ž^< 6Nn`‰,½ > ÄA¸†Ç讳 ÛóF‰ëôa(‰…¬áòš®XçŠÈAÁÁð.ʘ‹^á$ â÷ýÈPß%üù>×½wÀg¼à!¸†Gè®s ¹)l’VÙ ¼ôàÈÚ´3cu‚mã|†1è”WYvSc ~f‘vZ%[ä¡ê-PMØûà3Þð(\ãt ×yȳ ¶88.&šQÁLg™Y'QOÃlf—eŸq²ÛŠGš×8Döû|Róá… ’ ¿ ÃbÆ};ìØïd X \ÆOo¿v#ðrÛL5ë’0{Š[Dæî Ƀ­=ŸéxÌC…&hÍczVxÃ%‡ƒS²ØMÕš]è/h YÁ£v×YPŒû3 ÿÛúÃP²†ž!TµL$½á\Ôm¯tá3tµíB Ë×€;Ô2´àgöo¸$ø÷Þ‚8—¡>KãDú:J¯H`óÐB>OÎÂç¼ ñÒ-lGäzÏ>ôüü£Èe¶pq—C§h«ÍC ¸“À8—0ä гI:2>ðBN!kTJj¡?³õUì-Têµ³>àô(XÒGÉîça´ ¬Äb|ªÌ ˆŠÏ ϽQ«çÀÉv%S#xîl@E9Åü<1!˜ßF~· Ô¾ž’…Ý‚kÄpüz¹âí²gRL·^‚}<ó:§´7ñ[#aAÆî ¬ÝklEYÕ‚Ÿ«"®õ¥ÖÍ&&­Ë>&§µËQ¸†Gé®óàܦ©…%Mh³¤Ð¹p“i<èRä‘kœiÀã|ޝí®qÈ Ö-ø™}¶qD H„[‚eøêÈ #›¶’3®¦¬!­u¶àgöü¥J£.8PµÐÎ% ÃìƒÏÏ}®áQº…ëa<ª ÷Ù†.ô<ͬ¡1f ×YWÜø®Îïá’êҌֿ ãIâtºˆ:©˜j,Êvéò™}€QrÉ$Â8$.‹ˆÓêCÏÁ’rYãt ×yÐNg¡ë ,{ •e ÕÖþ ¯@CÁpânÉGØÐlÌvj³?³ ¼á’RK6P#ÓköP• ¶ÁOÊTî(BB@åûrzÐsè9†¬¡1ªë¤5• M IõÃP23é nG癈àÿpu.ùüLgªy¨°¯ `¤ü\-÷æÒrvÊq|^@cpµ¸Î"¶"24„¼}7{ /Ûó®XÑpþ©Ó² ®šm› šçqô¿o³Â-éu?Û¯²—Ðó²7@ÏÌÞƒ}7û\ÃCt ×IÐV}U«CÙ¹³à¤Ôž ×)-9ŸAÂ¥oÉç$“iµàçCé¹³@FnІsèBÏxCƒÈ¤[¸ÌÂ1" û4~“ª¸,Y'†8˜¯f…o‚¬#¬ ŸŸœfÉ‚ söú3»x~è%seî35P#¨UïÏPHcp ‘ ÜÏBÞIB>€kÈ;‰%°šcÄÕE†zäE·ˆó1™Uݘ*µÿ¹}Ðúu=«ƒç‡^rÌJ ÆàÖA;#¤ÜîˆQî?ÿìûõë ¨>ȉ#5ÖÆh˜*‰äµBy28Täºd󅔚ϑ~Šþ²F9˜R„ èQ=D}¡?Í×#hÂÞ;Ðs(‘«,6Y#­î&A@†Çoÿ$´é’Ø-í×ð=†Ö%T©›Â ÀùÎ+,w#G¨k¬L– ðÙ]à,™"”•Zˆà¤GïÃÏx#Ñ+ä¨ÕÀu©ˆ‹í†¯z2††P±3©ô6)=“éú96Ö,§ŒS)ð&úÍ«?³ ©á¢€¦¯ÙÀº ÊWz–n_†µ[¸Ì‚DÉŸO‚NràQ `§™ò@¸È$sçb@±àÒ8\Ç•Œù¶3¦Éu‡±Gv%úâ{ ñ -ë98檤ªáÁº…ëàéjµ.£êýÃPRÃO¨à=©}:å´;Íù ?¨cÊNQã †œ øÙúþ^21­[>ûH¼‘?Óé,lÀ\c9¤mU ~fñ†JÚ=†Új 0òNvQ›'ᤸÎ3ûç‡(2>Ù2:Cæ$„ý«5ÊÿøÕãÿzüîºÌµ­×_Êã?µÿÿþazüE#Áã AÇÆ" bËn½ä÷É}­,ÐÐ\Ô‚è6mc-—Œµ ¦‡ÏS¯å’¡–ÝS{­,kžšÆ¶²d¨u?£aŽC­éñk,æôˆÿé?Ïþîáï~]§Ø$•©‰[ï~×/ê<®â€D‡"ʼûðð›§·õÍÛ{t;ŸÞ–ö;âdOSû‰x4ËñTâ'??Õ^:¿ù·wÿôð÷ïÚG-ǺÈ9ý ­¾Kpèm­0â—jµÓ—•Œµ"[g8óº–K.µà•š˜¬¥’K-àÐeìÑ%—Zc§K*¹Ô:¶ îC-•Œµ<ï^k\‰ŸøÝðÍÞâãÁÌe@%í3Í»>ÜüæmYÛ_Ê‘¿çøˆå›||ĉ|Äöæiøˆü9¿øµí{!½ bÏíXþèÀGhÛïwÙp`ÎGXs6 Ã~úå¿<þùÓã/~õ«Ç¯>}ÿí‡÷ßûñ»7ï~ÿùä{ñœj[€ õñÝ—OAÕ¾,ˆh?¬'Øþ¨ó ZóÿšÓk¤¦±æë#öQmÿ6rùë_´C¸Ó­‡¢IÁ ‚×b˜ Ü"#_°@úF‚Álçµü!!6XªÛC´ÇpÙÂëí!\"W ð›•:aÙ‡³‡†ÃZ¹Hp=X²; Q°Ÿ7%:}”Ûõí!ÔÏÄêVà´7æ)]îë€â=íb “»G»téÐ@ª¶« )kÕÐSè=hÔÚ C×C—A” š„sôÄao2í,ŒkçшlìÌ"¼CÌBɼÆýÁ#Ö‚¡;ç“kZäq”Àè ß=5$7šƒÓ©]4\05œEðñb=¸/ñéQ eÂTµž±‘gr5ÖfklbŽxiÙ¤]pÛß"8#D–ÐÎÜÂØ~‘µe‰€þ«¤æXÐprÀ‘“¦†ò¡¹X#Å i ™¢ àØqÆ´ô:Ø„°‹Å:‚¡DÄñ™íœ2ÂÞ ó^d`pÆ…MPg*öcRm Ä¶aÃ}ƒ¤Û‘œo]¹Zqˇá9kšð­ K‡#ls±'hþÔö„v@„–¢‡Õ©Õõ¢áé­p“ÝN*•§ã©T_×À°\Š ¿1FX" 4Ügl­=ô/ÜTg¡yÁJ»t»+Žç¶ySÁ&JÂ_òO[ÄiDÐñ Jëϰ±%oà&¥Œ Ý"Fu)þqQU¤e`xQ7Œ”ËÃc¶þF_ᬢ5ˆüÌà»6p…áßcÄÆÝ£ÊºÈǾF´ÄøL§, eDÆ‹]‘ll‘P ¾$=‡ð%Y OÒ„X®D9p}qÎs—2nç—d˜Ü%Vt[‚ƒÒ9sNUÖ¯‡¾5P»ØÀëÎK6$ °+‡?NáúÅiÇòqiÀZn%yî8ÿ›âO9¹Ñy#‚@m/-gŠ$•´z"LjBÚƒ“…@즯{ØŸWÒ•ðt§/H©±ë¹0äÄ4¶šS÷…[¢ž$ÂD¦öÑu «ŒV:Ü"s~ï÷É[¢2ä˲zKÀhˆÔ3èÃÁxÓ3¡/Lz›$k(43®¼|Øa‡]°jßÒ,¤íÛJVÁ0Ïø E¿Æ-I?ÆPdKRhkΜ.Xs~ßUÎá8‹\½ÒÚÝyÌÈÒ)¯ ¾JáMVS"Qô~ÅÇ%•…—\ÜçÈúXù™&’нò2©È.E«Ò„óF_ÀCÂT¬L]>|MŽxö…8«äØO^¾Ç¢«4b…Ï`ƒnC¶àYŒU8/ÚvH¸N˜E˜RÞî:v~ëzçÄmQ22 yà-?¯'ͬ-^oò)\«8æßgÀ‹íP ˆ`¤~Päm²¢ã©C×"ÓP^v8mÓE}ϼ‚X`-±kö‰1T×I´©9B©TôØKqÕ¦>:*EÝK%‘ç”Ùêzðš=eÓwˆ'5>“LVpMQ'¹&èÿ"ks­æRwšIųL¯ÜØ#§.˜áv‰¬áÁ‘”M”&ìmºÑiaD~X§x{¼o蔆®E 4b 6ÆÙ\ÊBa¢"òŠâÔ“™÷ê]ô˜ã~9”B³œ¼ŠÃ‡ÐbEÑW¦ mdl(Ïeü÷S¢†uó,­E¬RÜ0u ƒÖ–ZR1~ â{„íÑú]ÌT" Ë6×¹1ŸÚ4é–‡{ó(HÖëŒ tè-Ôá=p¨eˆMÁ»´U’`‰ù‡éä>^.'¨TÄN'D cbºˆzNapÖhB†ç}|ˆ¤ÆkŠÁqž×]uÖ-X«y[ú&ŽsW*WkAʯ@'æÙ²U›%íœÌ˜*Yˆ‹’w(·yWž©Fò1³öqØŠhA).ntc[´ÏáÍ\ÈæMœÆ‡„âÃüñjCýí#΄¾w}§®’`øÃmŽR'¢3Fæ2ia€Ðâ¤癩GžÄ±*ø’/>{Èõp â~P^„µ3ãÃÛŸâJ·óH[G^6ÁÉ4ñ(Ì!UÉ?x‰3PN\¼%²^XÒx'b{¸ºWä†eh§Bò^C+²Ñ§ž;ñdÖð¶‹4)4ÆÀeÎ}†“±ëdpÿ‚r0-•)Ô2¼ÍLÞÂgâž©§ì«"¤&—|=b¥-§I¢ðE¼R±ŸõÑ×–«ÌòyÜ$uŒ)ç4õ6‘`áHJ¦}Sø½ˆ›M¨Å¼êSA¨${Qõ}'çÃZN] WyŸ¹²ÀÅž.ºe»m,}BÇLEZK6ŽÍásÛÊmniž·`jÂ.qTÒ !IR#ÙjÁ”Žá«,᛽ßá#të_wnB|§™¾„øàRõ8"|Ìø~ë8ÝÓ¾e–iÀƒD»«Døê"*¾ fƒ÷ŽÐæ6 |uNµ>\•äGðÑ{á6|‹7¸> Ⓢ ¾*eJø¸Åˆî›8.£{ Û<ЉîÕUgÑèÞvJ}˜èÞºX²º‡ð<ÒXÝkëNmf¢{MŠŒÑwt§‘¤ÂèÞÂm•àŒùBtHp¢Z'A#K±B.±½M%Æö”“,±½:ûÊ4¶ÇP·Û;½2ÆöBtÅö #%¶‡üK‚íÕM\â{@ê¸À†÷v‚¥Þ+ó²Fº·,ÖK݃wÒc÷fzOÞ:¸WOwlpoWÚÁ½SÚÅŽííb Œí!”…Kc{s•êÃàøÀUJiç¶ÅŠ6¡{3ƒÜto)F„îÁZÆ:F÷Î*´<ѽ2[Œ3º·Y˜7º*L®Î螌‰n#ºW¼!î­ÃîádÍwè^jÆîµõ9µÿ„îM3 hG÷Bd5(¢åÌ õ/T£{`Uc ÜÛŠô¦Ü+"Í îmÖÅ$¸·_Üa{“6Fb{Ǥ/al¨Ïˆí­›7†°=8’¤'¶·À¸!3xéÜ›wï{Ë¢ÓÁ=KA îDšŒíÁè™ì“¡½ê (±½ ÙíÍÒx&²—ÆYFö–ÙŠuC{õ $ÈÞº‹)KhoÒÑMh¯ãåÚ+âìÕ]ß1‘½uÖzÙkTÑòØì˰Ú*DÈÚ»{@¯d‘"`o±ðÚ½C*‘½s“e["{Èdã®C‡ì»RÿÈgúº‘Èh€Ð5!{á %Ðɸ:óÚ[ hÚCÒ6ÊÒ†ö ×Þî ½IºÄ„ö äehh!ŸB/’ÐÒ$Mhï0p’Ð^YEâÚƒ€/ÕðU½b$ÒÈÞiz™È^ÊÄFö:ÔÈÞ²Šötd¯ø:²·Îú¸DöÀ¦lW~‹cOˆAÈÞ)lº#{‹èû€ìÍ6?3²· ÐLd/Ô€¼ „ìÍô"¸ ÈÞzÚHÌÈ^ãðDâ쥑ƒ½Táw`LX¢zX£ëi¡‰&Œéâ:¦·oZËŽéUÍ11½’v‰éْИ,€ö#ñ÷¤Ò[,!&¤—\CBzŒ2qëÞ.ªczå°0jLc0=¤?éroååЉ3¦,ïÞ/´+Åaz`éµk„élÁÃdL¯«Eé­«¬£Ó£²éÖ1=º¬ÜLo^ÅQ'¬·mf% ëͧ8TÃzÛbÍi‡õ’í6¬çOž°^•­fÂzKXo›x &”À@É·×;uw×B"ØTÀ¢Sï ì-“uWöÎY7@{¤ÃÏK\Ïæg®7I“Àž9–ìÕ’Kc`o– ޽b¤¼#{ÌG5"{‰MvdOš´öNQv`¯Z‡”ÀÞºËÌE¸^„Ò¥é q=¨b¥£Î Âã& ±p=æ-» ¸Ž:i¡q½ÕÔß?ɱ‹ ^÷À1ª‡UmToÛ DÕëN$‰êÙf"Q=47IE싉ŸÎn{öz4®Î*´ë¢ºæt5íw‚cœî׃¤t^¼ö–óâ²™ï$)P/íÏÓÛÄž&¨—–V êIƽuT/Ä8’£z©TM½ýÔŨ^7ã3ª‡ÆÈxÕSøÛà±×ŽƒÌšêíp¢z©q4¨Q„¤§¨g“õ°šõâ°‡áìä÷Åц‹ëè°ZV›\RÃW/ D_Y+Ô³èà³—Z@ûì¥súìQw|övÛò$ª×ˆ£ì³7É©+]öVûµÔ Ï¿+¨×Æ'Õ·A=KÕÔ›u©Ôq\övßÍ ê1uêmpÙ;}ÿÔç¤?‘]ö¦EˆtÙ[ÑÔKŸ@½<´õFn»  ^›ÓyqÙƒU/.{ §ëêaùêÕeH˜´Ýõ6㾉ë¥`Ô}öR«ŽëmúëaA©*4¬çÖ[etŸ°/¡¤†õ ÊQ7kXo=äp`XOÀF¢zÁ‘ä_e?E†=Q½m7 ET/» ª—þÕcº¿[Gõpõ©2ª‡yKü3¬·KqÕa=´Çá×CàmqHŠ-z®¢öÚ3·k`o9,zØÃYX¨°7¸ÊÙÛw«$ìµ·š34²‡–ñbrÛ[û HAñT¨A´ßž¥«Dö°¥ˆ´ßž]ÆÙ ÁN{Ói(B¸΃Èq=°¼:ÐNíQfT/Ýà ëAïGj¸Þ6ÛWI¸^$7¿{iƒcd/(N‘½¤Q‰ìuÃítÚKgYC{̽v¼öpo Dî^{^Œí¥3Îàµ'•Jb{¡«oЯDràéµ7ÙÌ-½öÚžåq¸vRÇM鱪3epÏüZ‚{8ãëÕoÏ`÷Û“þ«»í­‹  î®–ÿ¹üö Wãí›~{„ÓoošÉÞ¤ÛÞTlŸnloSˆ‡Äö’ŸLl/LêƒÜۃ퉯±=ì.^c{é=I·½u¾ºí¥TÒÁ½Ývz÷âªME岬VwÛ›lðopÊVø÷:+e·=ëÝÜË륻í®Nç›U\½€{8§ܳхÁ½ 5d ì¶}þã€íVŸ²Œ¨ZLy|'¶yûw²–&Òko—EJ‚{àTe0n|o«öªæÇË0Â×öðLdQ4‚µðí‹Ùm!|ƒ ’>›¯&·$$œ_š¸&Äg©©C|ÉB$Ä—^˜‚ø`ãÅëÎ_—£âÛä×Ýö ®¼B| Þâ³CzíÑYßð^„6–E¨\öÖj°Qð^7Qê.{gRüC dzÞë«™ðŒ‰Ä…ÞÃm-Zãð¼Wl» i³B×ðÞny<á=oÃ{=üAzí­4ÉNt&’ˆ…)àÐÙƒÝðž}XÓmo?¹ÓÞÃŽ¨âÈ7ÆÅÛtÞ‹s'ƒ(é¿Í6%¾—®C£ß™Wã{¸(¤k¾×˜Ä÷Ìw&¾½7c:îÑÒ2ѽHO“­ûålÞë¤Ãð^¸p‘PÞëzÒooÝSmsÊDîaK¿½”#ßkû„¦–Æ÷pö+¾ç 4|„ÂøÞfÛ| ìwüövéÞßK?ã{ñ¾‡’]{ˆïµ•VØøÞfé)ñ½Ô%Úm«æ&Æ÷ {!€ñ½-½–„ïÖ‘&ÞÛÒÑ^{{šõ Þ Qñï¥{tÂ{Ý,á½ÍLzÂ{ÛœN òÚ³—×€ï)ªTâ{¹uŒï©#u¾Ùc½à{ý,ßëצñ½!z€ñ½Íšã{PX‘ΦÛÞ4'8`·½S.‰ïí“¿púí¥Å»¾¼à[à)˜¯~{¶U2ÀAw/_÷½M€Ïá àÃvØeö"€o«v97À·Û\ºû퉯ØoïàÅz^z·ßÞºñÆèF»\>ÜÿÇà3ÝýöV¹]&À—ñ}ºÛžF“ø^ªýßK‡üî¶'}mwÛƒäxÁ÷ºWwwÛ–ÞÝöÖ)…#)Q·]H«ñ½4[în{ºVÞKa<ñ=E ¿ _²öÚ›Šc.Øk/ãŠt€o¶ÊÝn{i`€\×![rËPŸ|)€¦ßÄ~á{Û¡ëTøÚ•¢ZøÞ¦˜, ï¥ã€á=ØO²t¼·mvMK¯=pÃ{`Ôi"jxbÜVv닸qÒqO¢XÂ{°Ÿ­WÇ=ÝÆö¶99la{¬ÉØžùÓÄö t¢N1Ýö2$Uºí¥Ã˜á½¾Òmo’Úèµ§àLF÷ö *!·½$ëF÷2"QwÛs8¬D÷ ™âÎ2º7!œÝwÊ” Ÿ$qñ{=çÝbðÛè^Þ ÝKÃFÃ{}ÛÞÃI|3ïÖ)“£{ ê«"ÀÈk©à.ð^Ê0¼×-R ¿=$Ç»óÛ+áEªI9&‰’‡„|á{Ë¢˜uòÚ›2öLzíÉߨç8|¯è\B÷ºžÏèÞ–vìÁ؉Ç´«# Ó†öpµŸéˆÎãéÐ^ý2´×yÞŒò|¦uçà²'ûTA{¸“c<€öè=þbˆÎz„š5n÷ö½âà7ÚqI_Š×Yú¶/ o䣘x¨ä¯uÌ]õ\ÆUò .ù8¼»[Þݶô‰—Э<.M¿?ô‚5bh4Ž0$²eÚbó6‚Td6žŸÛ kXýôa×Ð#;ˆú,˜Ã³­¿^CÚ;à3^ð\C#t—<ÿF,ž  À°à9sF„×ÿùþû÷ò i¯‘±Çw¿|xú‡o¿ÿþ«/VöÍ˧_þK{x÷ßóôî ®ÄÖöÓ&·ýZëÓŸ¾z¼}ûÝW=þí?ÿèÛ)„HV/ÃÃŽ-Hû áñT@åwoÞ¶o7ز?}ƒßkã·õé«7oçHd|<µ‘Åïõ,OŸÚ¾8³£FÛ¼Ó´mÇå÷üFžŒsEõˆ&_–§ïQ|œç¼=}déT¶§ÿ2tù=‹ÏkåÔ6MóÉ1!à%û4²£1豦º?}ùŠÖ½DMä)Ý–ùéÆVñôÛ¨ºÇÂÏgãRgvvNå(ãϱÆñ{Ÿ¦c›Ÿ¾@ÐéÒÚ[ž~Ι–¥}Îc?Ûà¾ç0ÏéÛûf¨ñ•—¥<}m7¡¸ħ÷è“æýmtDZ==ÃÊÕ}µö——Å»q>~|„Xêzyñ»ÞÞ—±Á¶WÚþz÷eÛM¿Ví¶ýU»7ˆR?ÊS»j÷Ö?ÜÁŽyúƒÚÏóéËãÚü-ûGÊA|ØœçØºö6?V#!m¢}·vÞÔãéÏè¹âœoÝ埆ßos÷a ÞBþXÄ~Æ%ÑXáX oás¯žðTJõÖmœg}z¯ÕŸÖc½œ!| xg´òúQÍksÿlÄúV·¡ÓKGãùøÝp? #iÁøÂØqû8Ÿ'½Ù¿è_ð«¼ý2.Íù<†í„…ü»_·«v [ ¢W-ÞZ?äacÝ ‰?Òöœë¾¿ìÓÜK/u%ðYüâ_r±_è¤@¼oœŠê¾BMž‡òþO}Ÿ®KŸwª—µýžÅG#œÃÁC¹|v†c\NåG·q¬ãÑ~ÿ¿ñ)ÊYÖì»}•±ïïP¡Ýûø5l§[Ï÷SÔÿø§áHF#íà SÙŒÂÄø§ß M¼×/'3¿ÖÊ{Uiõåˆ~©ÀÐÕÇ¡¯8¼Ð\Ôµèð6Æü³Ãë“Пh QJi¶üÇñ Š ŸÛ|½??<Õ#qÀqœacÑ–|,ÿä~—K¿ïÿtzûËJ_òݵ-âo߀¦õ b ¹æÇ¿OdñçÁöºlƒïY㜎û:*éì!—ø\^;àCAdÔÆO!¾>6Áy<‚tœMØç+±D¥ÑÒeá¼6.§Íõx¼¼òW‘V¯‘†ËðÀ×&LüµÃkÓÝðFv@Ÿ  »&6Ç}G>üÆâ?Gÿþv(³nÞ¶éïjUyiW{\yÐÆ[øý+ìSîÀ«ØØç&Þ±ãBmÛšàÛ_½¹/LÃÕà¹|Æß½ðjg-_ܰ¼Ë-û_ïêÞmÙ­±É/o Ìÿ…-¸[îòÊ}™‚bÁ¹²6a”«ÉãQëjq¥ëýðŽßj¤F?¼9*d€ýéoÛRÁçÆdá%¾ÃŸU˳ ã‡yT#¸^ž`ã¦&Í/‡Øµ¼Ñzù÷:${}éb‡rp^«[e‹+ÏÑÙ›‘s±p°+ýä…ߟÞóçtáŠ>Žçá®Ûí4åøü»¨‰ó B¢UžîîËŸ*B€*ûÂâ„ÂéXÿóW'Xú‘‚Ž*;NÿB!tþ>Žõq9Áê±í‡¿þÐ÷fÆáÜvî·]lXï &ØîÇq8¯Ìí[Œk§ügKq §UŠ/~ÖÜÓ<â_ôõæç[—Gv¬ fMŠ}÷ËÐ ÎÑÊÙvXá›kk¼Ö§¥9@ïëÓß«Ÿºî\³½í*ÕþvøýA-Î}×îÒ©|¬ÿÝ×mÍöÖÓ´\ê|óËçi{úˇþôk–NíÂyÆuV 6³Û~~ú›¡õçoû«cïÞ«›öý2Ol|?62*þŠ_<ÎX¸òwÅofyúÕðûgld9ŸþðÇ¡íPp¶Ò©e °ÖYÌA£šk¹ü~I×áþEßu&é^ $"p§=ºÒÓÏÍ‘‘úQåÚßÚ—2à#P°Qì_þøvhö59øë×VŸ^ù}'ÌÀ¥$ÄípZ‚¤Æ&ÛþÜ6íJºÔÕ‚„ tÕrBRjùì6*¡Ê×âgÞË>skhf?û×§þò¨Ðüۡצö¯oØêYΟÀ¡ŽóøŒ¾°ˆ×…Ž?ûù*I/A ?WòÝÓkí° úî;¡ 쵌Òëkb×xa||Y»¼û¥?Ah5À Ö™|S uïø)~Å—f¥¸å \%)kDA±_cÐÆ£4jPÇ}þ>äN`ù5”Ï#1GùÑhàχórw¨PåœÖû®"]Ûó¿NWe„RÝëxÜ¿bÐüC¸hF¾DkF¼ÉyäÍþã V·Ú~ÞÅýq½òá¾Ë¯u=\Øng#µ¸åšžçgºÑ­öÏv*& â3pd×e¹ ø"åøœðŒ{ —ãdõ¼¥ñ/Õx“I~+›ó ß [ÿbX²_]Gç|/àaêç #e‘¥Yؘ›Ï”-:±Z@h_E],áíÇ«Ç÷§Ixùû^§á¥ºòNÿ8V|øQ5ùy<qÎ ä÷‰u¡ÏŸ@y¾6õ\WÒsâgƒ™X)—ÓûC'ïÇÛr”¼F’ðÛY& °py=^”È‚òç².uyÐ X‹w è3Œóí º˜K7š~#¤£dðЦðO{èËŸ ïÝ׋ÇY^Ñ=x‹4ÒõÚåöÛ;Z;î!µ=¬IŸåº+¦Óvö_xá‹1ç!üf¸f®°*±6ŠoôéÎÕ>ÆîÚ.PãÇ£w‰‰lp‰X/¢ÏEÆOï+úÑ;ùílû¹Žç+JÄçœÀFþîám™àúÐÎØÆxôÊ<äŸý¿å¶Ÿ9endstream endobj 360 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 855 >> stream xœ‘luÆ¿×[{7¨Ì ªAåz‰C‡ÂƆÒm*‰ gŽ…M’¡‚«ë¹ôÚÚ]Æ(mg»¾°†±D·tL:ìÖŽp!.!º‘!aÒ˜å'qAƒˆ0ï-_mÉ›¼yŸÞ÷}>Crt„a}UmýÖìP¨=ËhÏé´u,xµ…e‹Œ,sRËsØœy¸çIÂ2ŒüI Êéêt·µ´*bQ󱴢²Q,Û¼¹B|S–ÜmÍV‡XkUZ%Ùªd„]lp6·IJ§Xôz«¢¸*KJ:::Š­²§ØénÙ¶a£ØÑ¦´Šõ’Gr{%›Xít(ân«,‰ÙÏŠ³­Ê)»)’[¬uÚ$·ƒ¢/-ÛòÊ«„Õd 1‘§ˆ!ã†ä:2Ïìa&uïháU‚ªÅÔ™ƒñßY¤Ëy¦Dd¡f×zhÞÆÅá¸z–tœk‡£…B“Í0]„Ë™Z„ùзÆçÕ&®N%ÍÙUâ_Hn©ÌÏHpðV›Æ:Ódž^_°æHàÓžJðOë ¸æAjâVz2±6‘¸0 ?ñ˜Oõ×é3ÔTºuSã×=Cg¾=wúèÔ‡ÂøìÂùxùÊlk—™© F"`_«½mÈœôž»†š“È<¼”dîþ‹~dY¼ˆ7M7¿«_Ou5ïowž¸rÀ¬F¿ˆÃŸòŒ·»äÃÖ"Ô5àj,ºw÷Þõ†_èÊ„ðk<†ßø›®R#Í-Û[Ñ”ôÇâ£Ó“‡F]QáâÜm~6í)×þò‹=æÎˆ«·‚‘@Žð]C0"<†‰ùªÖc4²Äj XnRi®¶›‡~wgoØß-Ðûöëm˜~-ÃÓo6CìØ¾˜;áX$ËÓÁY`8i¾Á•k/˜èô,÷اÆOaÞíáLPO«8vŸÅmÚœ)Ù=JÁéLצG¿¿cü©`_(P@hñ5úvÂ{°wàÀH0:Þ.3½Äÿè@_ôd¿pfj&µ<æ}é -ÜÐþ‘Ù.íïj‚j¨>ï¹òͱ* óÞ/íÝÖÏö•ü± pòc>JPÓÎ}^×Aa•wL«úƒ'Ç Ô:Ä©+–V +r,1cîT¿Ñ¸4b|‚ÿê~Sendstream endobj 361 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 6405 >> stream xœYXçÖžueg4ˆQ2Œ™%ö^cÁ®ÄÞ Š¢("EYú"E¥ïîÙ¥w¤/mÁ"*6¢M4Þk»Q“˜h¢&Æ\Ï{ŸÿÛ]Sîÿç&÷¿ŽÏª<;3§¼ç=ïû)aºwc$‰•óò5N¦¿ ß’ˆý»‰oKè*댷k)Xw¯ïï8µ¯hÛ³{ãÚשDâ¬r Š Ùéã«tæ9Üq¼“Ó”QŽÆsrœëï²ÓÓ#Àq¹‡Ò×ËßCIÿ¡p\è¹ÓKé8l†¯R4mìØððð1þ¡cC|f å¾Séë¸Æ+Ô+d·×ÇJÇþ^Ž¦ØÆ˜>œýƒÂ”^!ŽËwx…0 3{n€ë¼@ç MïÏYºP¹(lñî%áK·Gz.Ú±Âk¥÷*ß5;×®ósQ¬÷ß0nü„w' ˜2ÕiÚÐaíGa˜ÌJƉȬb1«™éÌ`f 3„YË eÖ1Ãf8³žÁl`F2®ÌÈŽMT«ã“ V²Þˆ£ â›Æ¨š¾/o`ù;Û͸¿äëß?WÚÜ­‹ã Cd 'Ns÷ª:*)¥í)áv¯ ®ßõpøvûí°ïðïHO—aÛ½ä¶:X2ÌÓkIÛä]¾SêÅ%׌âJ£T¼ÐÙÏÊ£oÉàö%@„Ðå!S‘0«Y&Ú»‚cè;'˺wu&ÄRø$:DfB© Êöa©U ,£È*Z±Ù2›Î*Ëà€ÏÅϤŽüÿ€ÈÉp.Àa¸ ×9Ž'ïÉ)Pdÿ‰bk†×KÙ¸èZíslu¡3 ƒí0`Ýü w/£}¿*ã Iý,¿!áK~óí5Ù¾4’žD2’ØÛçðGS[á¹f9yïk"ža·v"!eoPÈ6¿õÀÍ]öZ!wí³›·ŽO]Gë¶’ÄE#¨— ãCÜZ#;;gð]×ÁyÉ/iÞ—5c.-OìLfÓ e”—÷¦Pa”¹«×Âz­ws˜ìÿ¦ò–ÔóÇá<´˜q dgi(–°;HÁŸ}K,gÿäM”+U†h:M¿.vѧó6.užšÜ|VÀ{,iùMPO“,,ã¾µáGÀi8YQ{‘³ÐM•á°\+#»0>ó¤ñÕ=æÁ°e¯Ëñ¡£Äïù}ÞÁ`”mW/¤«w,„íæ  ì}¬)}Mìt!D!;šzšéuŽš a§kòè#«ÊÄ~FI©¨‘â"D^¡É€L®ªJhû·i”Ú­à àc~0ÛJuP¡7!hŸ±'K Ù¥Ï܇ýñ®}YÛÁ¼ƒ&%à¯v§Ü ] E?Ôjöí…`ˆÓjRTIº½h3´##F%'P‚c?—Šé¦>Zз.VÇ‚ xÂÄW «g¿1®Ë»aýO·’C!Âᔊj0rua•~ŠÐݾó>Yñõý/O[ 貨F¨2mËW™u®–vöÁÏùòަ•™ÂóTk`xA4xZ’jÖA8Äi’Ô‰cIªý¬¦N…T‡*£¾Š´E½*E-ƒŸåŽýPªÍÎVå{‚4É*Ð:‘0ûQ˜ŸTHWSšCõasù¶ÐòyÒwÀNË}”É5y;‘'·OM+¤¢'-7-ç9°ÿH‹KS‚C!¤æ¤åp&Ia?ˆ>†ÚÇ }›o¹=Âé7®×ØÙº©ð›ë|ºG­O3p†ŽÂ;òóëXlôñE:³TEÔ»õ>àV¸¸)ó=–*Ê¢ªªKJ«š·¦kåÕU'²kkk÷š(ßÅÚV1.š9É‹üœ}Ã6€'çô$èŠ@ßñ:WPÃÅ̘ã¿uóÖúSµ‚£3›Î#¿âPüVü‚ï¸pI·ÏTÚõFˆ3黥ó5ÚÂªÔ É‰äùÀ?þ³Qú†v6‹6tÙ­$%Yªì½ rØ‘taK=E%x¥û6ó÷ÙeíwœÄà}{Ë„–‰Žôn¢; TÌ bæ:O1­Ë-ª?~²è‘#¡UUÛô«À6'­ö óÛ®XîàUÚ]:ªoâ×ñÆ­§Ïþ*pú ûžX/Üì*§ éÊ+…cÛ®G5¬IxŠî‰J ÈûR\‚ç“ ô\U=”Q¨lÒ„k=(˜ÖÀ TZ D“„ºXûUaë.‡ˆÊWë³t5YPÏUE)C"÷¹7yŸþ°ùýs•.§çV½¬w0²~jÄüvkÀR@•—’0„ÄÓ‚lVÕ`ÃWµfU²õ‘TdÄ·ùÝ)áÑÀ—†WU•í3^šh"yc¤#éCø金ïÇ^Ùy* {A•$(¬ vn©Ói*’ÆÝi½˜VBWôOs_Ad°Ql¸+ítïxÖªëÊ/½½˜z…¢?n@ÇïS-™!Û ‰I”]—í‰R¼˜PªÕÑý5ã—=ŸÌÛ`ÚùsúÝI*” µg-Bi0œ¨–‹ûYµÿ`ÒcÒâóÀ!Ò³Ó²édm†ÑÆ i6âTSòÙ¹¦Áå ã J £šr*Քϋъv¹»ŒìââiÎÔñíɺ™QI¹†j¦8"q%}èöï!ûy ¼iÆ­ªáŘè€Ö&­¦Þ»[á¾%“µgepSp³æì‡ÖŒ3UGÊ÷5ž&8²»Ú+oh!‚Ký€OÎp?3“>¸·ãp“>ß Á×>mi.+¦[ÊâÈ_Ö„øâ·xûÅØ¡£Qœc”¢ªs"Ÿ§¦cÈÑÒÕÈN€!Q\Wß/Ÿgv½ÌŠM§8u(‚¬¹ØÊVÉç®.š½Z7Ø1ºµt=· y×lŠô-U¾•8Ëì…ÔWMb7’/O‡²Ö~Cõèäub÷ãP:Â6'_Vd&gÄÇ©S´òƒ'hö‚l¯VÖï:©½Mœ®’ÏÆ1‹jà\ôБ7(¯ «í_kÒ8ÊçÒÎHüšÇmd,uŒ+é5–L![ÉVGmÐ zà ¸U ß{ü"ÿsQo?y‚g5É%ò‰Cä6h1å¨7HÐêÚ£/¥xƬR^C+ö«“-Í99š”!.…*Ø.¸$ÒP]\RuÈëÀº™N«† „¡¸Mþú'¢o°IÞPƒ&Ïñ/>§[¯ðèøüóoŸ úœÈå„üŽ ýw,˪ñ”ÅÝ~+C' ÝÜÝé™&'ƒ(sÈöj\Öq§«%µçqïÍúóRìƒ yÃãÞãÀ}sýêí“õ1ò# ­É”¤!’SeFT唕D=ÜU>¾BPùŽj@G9/œáZçS´W®ÑúRâÞYäSéºÇÓ6sï=[ƒo`ïÏÜ<qÌ»RX{p5Œ¢KÕôÕéEé…ÙùÜó~<-¯Ö×ï«”–VÖ÷F(•0b¼ü8™ÅïÇÞÚikf¹>·õ\QÎGä–SóPI,êJŠK;—ð¿%Ù¢?¶L¸„ýïXÁ&_ÅZNK-£:Ì 6b›ÛŒ}©oÄÜï–?j´³ÂHjé0‡…*}vIZ¶~?”÷m Ÿ´ué/GÉ^þ›ê[à÷ŒôºK(p®ÿN³[é 3…D?ÿ¥Ù©¬m=ZË®Ã~¯´”:oØ ïѸ¶žðjõ¾¹ÀÝ+ü¼2ò¨ßOÐÊ“—)B´öªÒ„}ñåñW“Z.Å_!Òƒœ 2½ ­‘é{™¹”¡á2qO™ªy9;‰Z°²Ÿ+Ô¢/Óƒ"JM'-uIJúa¹í÷©â!¾.´Ò_¡ ¨1ÖUVÖ™ ÂŒãj$í—ñîe)†‹“ùé¥Á—áwåìå{ÈŸ9¤@HÐ%ç™MK™]ØRÈŠMÔjc…˜Èœü]¢J7œmbÓ)ãÏluKO–·l-NÆ×ïFdkìÔië½a·pÓœq³]FN ‚–®‘³ —6ò2uº¼L!¿8.æÀéK+oǤns÷»—í»…ÖÉ·sÏ]Rh’ˆ´¤QÒhÄ܇˜c49&'þ¬,Ü|¢P€N¥µ8‚ªƒZÈѧèâèøÓV턽„ YC*v3‹júGƒ¹:;ØPXoÚaP¥Ë.MËÑWC1ÐG”'³™ùí®é·íIó).2·N]=:=)²¨oÌ(Á™b†=ÎêJ͈Ò'€C¤›Z²^¼c_ØRm¸`ò µ˜¸,P·ÚÒ’m…?(!Y“¢I4¹¤0Cg_ƒäÈ,x ýñS®¨øþeëãÙ•«a%ÌÝ:ÒoAò4˜ Kô#šg´N¾µç8|7jš¾®»’ñ¸Ãù„÷ç‚´Š»¨}º ·óΕ~y¹¸AGxɘÜu0–Ár˜·j7‘Í 1íåP,4Ò¡ ¿)ÅöN[Þ Õ‡Ë—Ž¥s ¶>€³U\Wë­Ž#ø™8ãK:ÒGáá+ÆðcǨóMÇÇpŸ¿xn”´½@¿¯¥âvœÏÊ41 «bö$Å/ ›”# dÏãTðÁxî€ì¨;¤7agŽ·±99»¼®¨¡nw¹"Y Ú¡òÚåýÔ“´,6bÜCé |ÊG‚ª +5-;C¨k:Qy¸—@Æø(`y »<ÈÃ#b¬‚E¡7t¡É®´6¿ÁV´;0ÖsÜ_Ç¢{ýð_Ç^“ž‰Û¦=>¾rÜÈÖÅ“¦s²ë_æg'¥%k¨8—oŽØ’àA'Ø=kKI\º†•‹…¸(S‰Å·Ëp2´3 ³AŠãÄûüåà&߈   ’ ƒùùiÙ¹‚^¯ÓQy¬‡¤øÅþ+ÜÜ©ÚQ«!™KÎHÌ̹ÿ´pó›ÿÍm¦ú”‰—îHÚ©ÛŠ_Z\GGô2ÒÔùê.P¶‹xÑå“V‘u>?/3ã6нe¦#â7ÉÂX2Ö8¨Àýx+ËN²€CË‘;sÝ<5½ }›ÏŸÀšÁgíloâqï_¢¬†FîÃOtó’Ëüù+·x*ÃþTå§§à.÷xô¹ñœæŒõ*ð© öEd¤¼ïÍÙ~9ÑcÕ §~ïü¸äžþ ™'sO„ ª¬€âØLnð6ûάØ{ÊV8̘î6ãÝy—>»~êâ' æ4k¿¾gß¡Á|µäúegûRìWù0–ªné¶5J×´ÁBš,Ks8 ŽpU‘¥‘þ‹Ïl{ˆ<½ì1P°ý +‰õGã ;xŒé|üë‘h%ؾüŽœ)næÿ¿‡.¿K‡à¤E2•ªk;œß@g·zAZ\wÚ ØÆÍÛ5p´yÓÐ<ühó¿@FŠ®b/³|ñ õ¨ ©««4Ô ä¯ÝkC ôg!þ•¡uIƒ‡~AÍr ïÿ26â@¢/ûiM©äû)>&ñ<¤«S“2 =wà"@gÀ±u¸ ßÁ~iiúTHç2“ÓcÇ;‰‹à62t×À9›ÈΑžHïGó¨„ÊLOË”ÿúdŒ¥nÃ4þ‹Ý6 töói,pqšu²^™(v#ý20˜¸‘Áä-:%…†ªÖ'ää=¼Ž’³Â‰o+´æ~þÏJìó¸Î@å î¢[ú¸€ß£¥¿@£Õ@4e%ì&{ñø`{Å¡¤Pƒ°#%I±œeTEuYIMûº–éC‰µ‹£@¬þÅ&ýþRÇ¿â$¶5#f»œ”ÿñ7Ån,í qyÕ›%´7KŠxØN³S’!´æêçÞ½‰=?nÜñsà ……ºWpx…³üÿízX¹è¼r3Êed{kìyã5¡g÷)¥Ö= ÖÖ7Ь{1Ìÿ±ˆÖøendstream endobj 362 0 obj << /Filter /FlateDecode /Length 6624 >> stream xœÝ=Ù’Çq ûmÍðƒ¡WßfœVÝe9LÛ²(…,‡$Ø/”1\¦ `–@jõõάªîά®êévˆÁ‡6êÈÎÊû¨þz'¹ø_ù{óâJìž]}}%ÓÓ]ùsób÷/®~ôvqˆN¹Ý£§Wy†Ü ÜÎKXÄÛÝ£WŸïv}ƒ‹Ñé°ÿüí…PVæßåùÇ×#ÌàÃþÛkå%¤Þ?Á7B8öß±wylrÿ!¤Tû/ÓÒ1„àö¯ñw›‡µcßà*Öyfyü*/nàçWäñKümc”ÎîO×* B*»ÿŸ*!¢ñ8'Úè÷Ç/®0Ä ƒß$È…Ð!Á!QÚ@äþóÒ>ˆ”Ž½Î‹´O °öí¼Üñ&#H†hè[–WÐ0‘.RtBí?šyIv/xZ{ºÞ“‚piÐEcôþ؆ê9Ùò N®ÎZ8í°DP¸À¼Dt@?¿ô‹+-Ã`¢*{ôˆj$¤à4þNoÔrGŠžŒïbµÐÁËý‘Œ9}3/›»0HåN Nø ìÿçúÑÿGØq&nw C?ßÿÍ .W1O¢wrzµr Ú?“£^!™é- ÔìùÓ|<*þjyàÎá:è|@¤¢µúÇ…ø½÷S ëd”R¶…™Zkø VJŸÌ¬…ûJ ÈðjwˆƒVÖ‡3øõ@ ¿Qôæ¥èõƒWRĽ?æôˆÏìE¬Äï³ëƒ ‚òZÅ=åóQíÀ¬î Õ`<Àä 虋„Ò²•T@xðªdÎþ‡íWÖ ŒáG8ëÙAŠÃ8êwû‰\ò`¶¨€®ÇÁëø‰eÍ4öw×ÕÂÑPL”…Í ¥V!OeyѤ~‹E 44f5a.x"…ÛJ<õ ÁK"3ò{&AÕxO cÄÝ}O{yOú_f`´pâÞóìlÎy ¥ø6?(/á3?üè7Ö‘)nPʆÊ[ÈÄŽÀ`24a 2 ²A—°LzÏ„e·“•òŒ®¤^"Ýco*t‚f…CÄÈ(c Q$="ƒJeñ8 #­AqËeígTÿSE|žgû¸Á*yEžÓÕO/ó"Ôçé)þް¾^ZEóï4ѪTPnQ‚!³Äý‰JU*ò;í¿@Ð H!4œpi ûÏŽDb/Ö(¿ ä¬z"2…‡’ÖŠ‘½í«üXƒÝuCm…çÄ´èYÇíÐ)æÈkrãþ_ NN”YFé õ†ü¦Ž3´@} Š·‹ <Á‹d†6.øJf,Œ÷r|H“&ƒªêÉ„U Þ(\mP3Ú g-x´ÜÑ0"s>XCKË x{©±°Ä½Ê‘‹ÐpO"õ -BGÄ‚aYh@Ý#7e1¡il¥6Us@ê³Osˆ wDÙ¥¤cdMCTXP±ô‡ÙâJ.¶F ×;ô¤ð„¹“åœÉNVù§Ë¬à/—Ôë6/ X¸XÓœÖé)”º—åøôû8>0 çEVì|a’#/êÝ"Lï1B¨‘f»U;2¥/8`‰{€¼O[,2óPŒ¬ (ÉSþ‚Èâe•Å—=Ø\‚X3ô»GM.aè¦7p>jÂæ,‘Ú:®â9TlP U€I›œs¦m"Ò®ª(Çö®šŽºçÞŠæR§*C²¸­Âø °11¨Ýõ˜Hº*2’Æ&.OùiŒ]'ŽŠbºæ\xÂì@_;‘â·×ø×Ëȼ-95 CÈÊ43B‡¿·)±›".Yþg@¶„ >Î3AlË r¥a_ìJÍ!„¯Â8Ñ¥0NšcÙn$ ðìYõÂ4§’æ <8)†k*а9ˆ’Ò<#ݶ9H ,7Cò='ë”Îl¦Çä!Ð`JÚ1Vè„™h&¥FÙ |Ž»¸’kø¨P(Æ](µŒñ?LŒ¼Î¶Pq ÿQÉ®è"ç˜MsÞ¶ óxŽ®•—AN ìKÓyê­œ‚?ðö”<>Ù*2ß~k .Ÿ8ý„£ÐÀŠcáIþ˜» fB;;ZŽvKH­ŽÁ±ˆd"Ü)&1–¸Ã€af'uÃÒt”ƒKô4eKöÙÄœ}.dLæ“yt&cíÁ¨¥û&`ê·&ÈD‘m§lGe\´ìŒm%Øà䜪͌· <¨ý[g àïu!eO·$z„0ïæ@úý ¯â~Ó´ï.$ÀKÄìÑRHt‰Oâ›;¹î‹¤ƒÕ)E±¥ßCla%yãÍ ‚í'oºØ›c)TÄÊö¶‰ØiµŽ.è1DµAºF¥k™óÁJ×*¾EºNè>Z0Q¥Ð‚Ñ›# ªd—Ïã¨äþ|^Qê;òy/†Øˆ9 ‡ rÞZn9¤Z–snsÌa®þ:sHI¶spjª?he©X”I¦j9†Û ‘©4þðt.º•÷AŽ[í¿j € Ò™ (Oê¹n†2ؤ/á”÷ƒÙkà9ý —HôUh×%zXð‰õ Ú&BÁq±2‘Â?Î2ÿ59¶ñŸ¯R PƒÁj¥²^§ÀãD=.:æã¬BC¡e'hJýà•))û¡µS=…Jó¾µPhuõ™r-*›>´KbéC †îÌ¡]øÛèW›ý³S¶D/e=åøTÍ>YB `E c‹9m‰„ø·ÙZ¢ V—š‚€‘_Ôyë™JZX¶Z³e@-ð 肈I0‘ c0áH¥Uý±’”ÆmËyÄÕ%Ù!@æl¬×+…@œ)}±’©m©_@l`;‚R…­.,——ûŽOÛbf¬õ`…Âʵ¹äëí$p… ~Þ7„ 5T„u¤­âLø£È¯°üIƒlöŒ’^çQ¨t>i4€Ì¹£}Ïq¸LEdê\V÷$ÿ;–ÇÌÐT›åUÞæ‡P±-¬Ô«­º©±ÏËÍsÌ­’ǘ‘+%4 Jé{U‘tmZYW•êLcêðå¸&Vi‘*ŠXr<:±£HÍÝøJ^Ê•PwŠJ°”üd»œ zŒ ‡!(ͺ;Ò‘I³ìõ Ðxc¯]‚v)ÐmXTíŽk®ƒGÚ¤D-e¤=U}ªeì%ŒŽµ#m-`O/âëðØ0¼Q8;µ,À`;Ò;É\Ù¦îO_ N‰JDIG«Z#mÁ \×à&Q™¨rvå,ÆÐ{Ëïqƒ ‘'ÄWºyL$…£GÊqi Aà‘ò=ž÷'¶òl-¥3×¹ˆÖzÅu7שˆ<Ä‘8RJ‹Zo˜ùÓÞd}± ¥EÎ÷emãš……Ë5ò‹XXÇ8¢iqï2¿UøàT q¨÷޹졖©¤u›Ò@‹î·7S²™S9(õ(cE kZ$0û—S4I.‘}FÎ, ê*|\¨++Œ®­´XæÛk .pŠ cß/ ë¡ý>qÇ·H è?u›ñPF9$b½(è¢ æð6©ÕjÃ]àfÃ)#_*¹“½à¬“F‹âý‚œÌ{¦ÄÆiW(ÕxÔpÝ©qὩÞÊÍØÅIk/j5HÕ®â°úùUþgËí¿ôP2½zÂð¢ù8áݳ}ã0.¹f„°/]“…}'ÔšáäBgƒÃ•€@Y8)„ƒ -‡±¡¸ÿ÷ë d@<û„Vá{àP G[ üŽ B-*â(Ý/—î­º_p® ]¸žLÖ Qµ¢Ë©ifµo§Ù‘sŒï‡AÇŒå2Ú2[sE™#‚P=PKîaD×jSÆ”•)Îáý”°ËŒLì’êc²8j˜LsÏ ±v`¸<£V¹kÓêt™b® Àÿ㼊“ K×k±}òŒ(TJKwã[xÊ-'j}P2ìp3Ô(£Ì6fî½ÎÛF¹lŠH/7 íËš"òø®«Ii·ç&}è|êѪBRÎ&áw“XNEà*ë˜V JÂ8-xùõM2`á§È§Û¢~æ1uŒUzŒÏ |`Ú‚ZN¹]ÃíÆ7Ï ”5i—Np¡tËù1r^$‡è f F7),8’¼õه ¡áüôúɽBŸ³j½Å#k¶¥Í—\ÚB†:os Z–r6’¾À.P·ÿtôFTŠg¦³·¶Y)'C<êÉÀÖŒb(ÞT$ƒ>€\6T.}«†;‡òh¥Rï‚àcA¥‘†2 ë@ŽCn.ñB¿‘•,ÑÊ;Ž”¤æ4^   Ïh]8NÙ£øªhSØÎ7b‘"Ýüežè¸P¼|ÿuX‡«JRÓùGˆ4ÜCºT_ÎoX ±ò;Õ~Øa²ÿF«Sè Åxϸ󯩣e¶ÙÔ绚­D½AÝò°á³ݪSuHèúE©3 =0Qä'³úI1m€ÇèR•ôŸ¾*Í_^ánÇÚ ú]§9½èS¯ýå¼V a|££d½2‘Z¹!Hin„†X*¼Ç\Iª«ÌfgË’éÜw³h>_ÞT’ÕÆ!8Ÿ`ÝÚ›ó!ÞN‚¹¿-Å3év’ vdΟ]sðJoÎúõ$ÍÞ û`ñËÆ d.¡Á¾¥{/•ë—\tsË}T»HŸ«]&1w¿6sµË=uؼ‡ Yàè„@ØLÞ¿­:he,¶j…íŠlÌÇ^ ÉÒ³9ÇW³ìïî¥SÃ:iÚÊÎä8#ØhU¯> ïP¹>p‹âà@[*]ºJ ,flõ™VnŒ0@ê2òø<•ÉûUh9N5·ßîòÞþH_ÿq~Ž TÕw"àÜÀ‰tgw˜‘.ÎI›a¥*ÊH{©ý§ä1Ó©4Tp—,Q©¹ñ-®‚±5]olé X˜&(Ûê) lA·A¶,bLyά2Óä“ ¿®¢RÔ& I¤1æ$7£ä g5Õ"§µðt¢Øq î@iLøŽpLÏF£vÙ©Ãm¬¡Ñ]2KÄ”Rí×d[¼¡(eªêƒÊEA߯ØÊEñ7”ûÿäl‡p¸Õ{ïp„Žvk¶Ý`÷—`&jœŸãu”Æ¥"¶6–(H¼šéÉœ ºË+cÆ•ÊB%B4h….(ŽUýX8r!¾W\ó÷Çï]þt3EŠÕ¤À*Îèm)ØÆYš‰6Ós’ßg[Zd@ýëNÌ÷&£E'·»¾Ž29ú¸+B‚°ÎÕÊ{r¯ºý¦tüh5Óº ÉÍ…4M²¦~rʾ›©Ç°ão[ŒÁ7TNYü¯‹•4›±Æ@˜Z¡!v-0‹ó0"ðż2‹›¾_"v[ãê.YijÀ€æG7x,(ºYØ'ãŽÍ<¾œ‹1ÿÒìlåñ!ÛJËí–Wã+ljº&WŠ›„¨ÓU¿à]¯º¸%BW•U AÅÀUöŠú[7eU2ŒÂp×ÅïxEÇÈίÊù¶Z%)©k¡Ñ ‰cLÕÏx ‚$êœE)’ÇûB*©Á˜ÿ«Ô";Û[îQKùÚ¼É'”w+±´mmâȘ0ôþšŒ§F•AK½-Œ_„ŹÐ"FÁ$±yŸG8¦ÇaE¯aS™‡ïú3ÆK$>"‹ôRM<Çb`¹y§65ð»—j/¢¦x©¶!N’ÂUÝ=r0§ÊoÇ!QI^·°pŒ”‡¦r"¥ª¨ÔcqY±¶¯Çì7;þ<ƒóÀÜ» ¯u<2$±Ïä!ʧr¹“Çû2X§ M. »;”4K³´ Þ¨ª‘ÇrŒŒIÓÁsaX˜Š³?ò“L(—ˆòIw§RÄ”_¹iãj•Y¼5 å^É_EÜ)È$1±ÔSõf»=5g„{e5–ÉBi£ë/Ú,BáÏ´/•©”ìAáå·—õ³¬ ÖIélv÷|goÚo¤<—IhäˆëÍÒeÇšY•žmùg踾Êkª,:×k¾½J·6¯†uxwïýd½SaXº X6‹ª¢«S² ­]ß•ƒ ‚wr®©F…ˆý¢£H)v}0Ñ5}mÜœ0¢uU©U&U „IˆØUë‘ßTñOT£u¦_&ÙfíˆS-Ú[\{"@t”ò©ª¼PXÕý=ôc##x5‹4xŽøÆ‹ é2øÆ®ª¾~P·`zPÖRâî ˜ÇÑí¤·pÊO]ý/λï®ÄîgWÊ%ÎÝ)¹øô<‰x•öôäùÕo»_rá ¦ò%XrÀ6&%ü`µËŸrY§L¬™B/°]1ÈZbë ‡Ó7ú=l»FyŽ I¨nËä…ýµ.}ƒ_>@š:›þ°I×Ô±)ËüÍÒ‚å±;;€'Ê”_ÙÒn{ü¨4ö@&íÿ¾Sà¯çƶ_%Zú–¾<È¢VSãá­ØÈ­ SêíÇÀŸÎ'}ûߌh—)fcu?qÛÏÅý¸>\Þ(˜ªËE S’÷^û¥Òå3) Æ¦ë‚×ß!žpº¡£wÂgª[Íþ\#m‹ÃQ¸öâÍ‚΋Í󲉫p›RÚÿŒEù@ƒÊ÷’§V´j{îtív‡y=^žU‡Ãª†~¼ÚîØBz¨Iœ1~žJ¥L!þe¡kZ'¨-]àVD›Ä²2^Á*;é-Ù(‹C°pçøä-d1Ì&±ìSYLlhbÖ-ÚÃ* Xƒ($ò·5‰°ØPÞ‘3ûÚùN¡D-óÒå¥IÝ=bèþJ–R95†\¸µA*ÃP6e{ÁjŒ{¨pX1‚t`üÙ²|¿\båð¶3ütxoÇð2^ÌðùÇk¬Œá>s¼{í±·æxlÙsxyÑr<ךh-g% l³/j“Þ„@hOÉoÚ¾Ûh¿ÝÛÉü¥Þ%FÿðšÉÎÜ(Ÿ?”söJ¬î¦´š|O»úºÎWˆTB>6Y­Üû”8½è'û8¯î$sŒSøÀ»dð?ÔÕßÑk®nÈG ú.¯á«ˆöÓ‚¹ZÚ€øcݯ§ZþM}Fx3¾IT+iÑfÛÛ¶pm>RËê˜hpŽŠ¯±ÏDé•»µsfQbôrt² M#Gò¼ 5bAù\ØéKÝRÖËּʧ”C£&l®5Â[çç¦ÓFýLt@ôÊd÷ŒµIÝ~ZÐG*X´™sêÄÐѹȱ—1 ß-ì°)åê,ù:ñèøD^EzX y¤Ê‰IÊ|6ÿy#²ÃvÍOÖ÷ÐØžºq•ÚlRs¼æóÕ…Un yÐ;¨Ú]“ÃãR?ú‡ÇÂ×îñþNÃ=Þÿ>÷¦ Â m*¢´)ÅP—ìN\Tu~ÕÊoÕ=(!¨ë6dAÜ€…Rã•÷ٵΠ=÷÷aâh]áz»´®ðÏÓV´œÝö‡ºÔù¯ÈÜç·w‘\9Ó²Áç2øž|‡÷":½°Eíì ÅU#e0%¥’ç1ʤ-yŒ¿Ëv"Vm´Kã§Ø «úÐðˆ s=~yõè>Ïuï ý¥QMŽŒéÃÕ¢C;ÕÊJùYÝÖƒ+·Å¥ÈF}§ àB¹Áa|N`àij~IăÑïÊÅÍ‘˜‚T¿¾ú—paÕendstream endobj 363 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 6845 >> stream xœ¥yxÇÖö*²V B[°!Ù5¸ôPC(¡˜n0¦ ¦÷†-Û’å"˲Ü=’«Ü‹äÞÅtˆC1ÑR=Ü\O2Ëç¹ßÈ2–r“|Ïÿ?–ý°žÙ™9ç¼ç=ïó«÷G.wܺu¾ñ¸Ñ<îƒ÷¸ù‰(‚;ðÖA¬ùÀÚªþƒq²¡Üæ!°ñ}¸m0Áçñ‚”˃E‘¡~>¾b»‰“ìfΟ?oªÝ¬3æÛ-Ûïêç±/ÈÎqŸØ×kÿ>1~´Ûìáç%Ž´›¸ÈW,-øøc©T:}ßþ°éÁ¡>‹'Mµ“ú‰}í6{…y…†{yÚ­ ÛmØ·ßË®çpÓ{~.Þ/’ˆ½Bíƒ=½Bƒ‚ز,ÈÅ>x¹hEÈÊÐUa«Åk$käûÖE¸¯ôpŒòÜàåä½Ñg“ïf¿-þ[œ·íß>c¦lÖlùœ¹ŸÌûtìüŒ¿hàg‹'Nšì:eêîi{¦ïý˜ ÆNÄ|â#b#1ŽØDŒ'6ˆ-Ä?ˆ­ÄD™˜Dl#&Û‰)„ aOL%vˉiÄNb1p%V«ˆÄjb&±†˜E¬%fÄb1—XO|B8óˆ ħÄb(±›F 'ö41‚IضŸEX£ A¾C°D?¢?1€XDX‰ÅÄ b ñ>±”LlÅÁÓcˆò‚y߼Ǿ—òÞ¿ùŽü#V3¬ÔV¿ ²Èad …»…5ŸÚJ=î·«ß7ýÝúw2à õ 붃Fü÷ -ƒ:ÞÿàýÈÁó†¬òf躡o†yû|øªá5ßÑôÓ‰#nŽ;ríÈÃ6«mjlØ&Ùþ{ÔÖQ£GŽ®ø`Ç9Úxù³ƒ)bž² ÙL»‰vžv­v?™0fݘ¼1GÆÜ;”+Ä•ýs=7DË«å|ùð&÷-P§ñTéââQÿîR›HR&þÕ ˜…<²¨46I€/«{.€ÓIÝ4¯øÔ€¢¾ÑÍF’È Î¦Ñ² ›àôŒ”ŠT‹s]U TÏÆð‚ž÷Ü7øœ=’²3„2 j,*Ì®ÍaNùžZ וOœ l+b=ªB3§íËZ›Ú¨êÆÚ'OÒý$*F\c™ŠÐóÚ °CȺӺê^*BÖfÖ9HÆ/òµ_¨%KÞÀÀI÷^ßí<æº8UËÔ‘M½ BVØcS¢,XŽ3Ö»r+¶ÉzÖ"Ç7È8°¡¤„ ÉUÔô8Ñä• ú×Zœb艳„{†³ÍT‰fCöÂBÊ,üTØã_cª}‡däG·|oþp½ËÐÑÒÔ€4&ÕÕC‚‹É2ÐYÉZ2àwè‰í[ÏuЕæŠ‘Ñ!½ëU"ä ]þ8Úge·$Ì¢& 2Ï ÃLè‰Daæ•þ0J¾³±¯òp¶ttþPo¼Éy‹cýƒº– ÿ<žAN»r¯å`f}“±_ Ju÷¯ë;ñ 6€ï Í)OtKañÜa1è9®¾Öó*9Ÿ[Š£Qg®*¨¿ù½jrÒƒzAµÅûý-rœÄý0ß»è€m¨ù­”‰ãzÞÕgpöÅ|NËåÍN|Ð=…\ôY¢4n¯JÌdd¥g‚"ª2²<88B¼öêžW?üpúÓwÚbNÁçváÓÖVbôög ù¬§ÈÝè° L®‰>_/û<¹Øø€dYô&…·|Wr(•F§,,hŠv%¨eÏ£²½Êq~Ùv€ôÜ’®ŠoTZU!ÖéÁ%ŽY’§tù!rÝN™¢=SÿTÏûÜ`Œ·¶Ò_9}…ìvOKƒÊ#+‹²N0µé¥@¨ûÇü7°ha*²Ù&QK®;¾Ì¿ t2•~gU¯1ѲpÁ.»N1÷úDíÉ€ÚèÞtðä«f8XÅö…ÿaÖ÷4œ 7 Út½¾˜i ‹@=ÈeÑ^¤@NØçõñäߣ_Ì‘ ‡ÃÉÇP?@‰EPïø=ÌÅ`?kæêÃ0”†<8NpÐ"‘¼sf»fúªg‚ʾ °¤(@ãÐ4Sàa‘LÞíÄÞ‡ÔD BÂþk̸‡Go¢]ÖÃ÷úÊú"îÊŸs­R‹k™¿-óh˜Õß—ñ_ PŽW_ƒÑû…zç›mií–ø˜Ÿ8kS{ŠëI%f²°f’ ’ê¹*=¯Ã`JµP 2Õ9êÂlPNaŽbIFá“˳bŠ’ÙÌTìv°M¶ ž«ðK’bØ&ÔIÏ„Üp†¬MN±,uAfvNìÜHã»Ñ8VùJM ~·c~ˆÞبR3bAH/’I<‚¡©Ùe6©-¡¹‘€ ŽŒy4¸¹ÑùåufYwÓŸñ9oÈ£u2D$J&Aµo% >Yuýáé†ßàð’Â4ÁjSUi*eº_M\1 j´Úš;ãÁ.´l= úçDø>ü¨áç"3b¯uAí îCZî!wO ÄV%ÖÅ5(š «M\Rã¨M&TV‚z¶["ÜÖ[c˜¼†Èù¨„™aš MÐ5äi“›RêANuFþ%èlT·2+)U·ù—/K„õ•½Uoé‹û¸ýT¦O`Ðnø;]X€³6j)ƒ‘~Êx_ÖUØ' ª¡BˆôèdŒ(ÂsŸ–oMÁ¿hV7^H#´DksZqeø ´¢©¤¤Ì£b®¤ü)SÐôþãÉ8ÚQÒÚÊ‚¨ÃÌÊÀ28‚vÞ³`éª×>Ò_½Ýuj»³·Oº`(võvîÓ$€Zñ·ް¨aœŸŠ,ÜSâZ² Òˆ³)ôÒJÛRÕ!–SU ÍyNX.ç—FÆ:u¼—éËp fů`.V÷ à&Z[Ú!åð¢Ñ ©ãqâ~3‡£Ž>®4‚ ;Q¢dÃ6¯ ° Ìk–¾ T—é²—œ7À¡ÝÙK©¾žá¬I¾øÂó4<^‹‘c¸u Ã:t„ëËkàa¹£ z.—+ŽÅž_ƒ.p ÜÊ?Zpº.ÿ+p4Ä8çí+€X VÅn‹s “Ì}(€Ö˜ÜpÖkŒ¥Ÿ…@éwŒ¼}QUϤ‹tÑ:@•U6~9È/pû˜Ý1Å,X^¢Î e·Î\l&ôÊÑz˜ÐÃ]ʉHc%Ü|öôéŠgh‹$§ ¡™†‚HNbõ ùâÐêíÛÝWOÅôn䙈ËpÁÅÓ—á¶Ë<½Q‡4~ɇ>—èx ñéѸš(Œ,Q @ÈÏÔÔÈ” €zn¸ýc£²6ªš-Ï/Rª²Btñ…Xz—•VÖH¶‰÷&y³Á¥~y¾€š¶Ì~¶wA@i«ŒŽ !`Ÿ.Xºz_ŠQ³u‚ÃàÐߺ~„¡õ ˆÙ …´‹[GggGÇéÓnÎÎnn.¬ ­ CøÔc¥Û„Iö‡:ÚÎèéuXÝo–žXº¨è²RÓEgXÿ‹Gµ%í9wm³:ÁUpŒú3Jt@UÀdËOMÌHÌQlšé)@Þ¬<—pÜø8njN䟫,{Šð¼ÍW° ¸4!ÑÑ?tÏC›„¼4¨2&„'ƒÔc3Ùw÷ ½EçÏW pÝŸÕ\·äÿ~» Ð;÷ö/C±¶µ76Œ&‡oàYúLSýiðõÃâoQ?­þÓ°žä°ÓwÈJP™VRŽù ¸ƒ À¿4 , .ºP_Ö\êhJXÿ461ÕG ŽŸ¼ ²”•Ì0¹¼)¨"9Bƒ[èUÐó:ùôàÆ={=¦÷É‚û¿ðz0­çÎÓµâŠà`±88¸B\[[QQË '4ÂÜc Љ…·Ù=æ< ï‘=ÎìËû37`Ó >·éÝ]R¾?q>¸1=˜·dÑ\ç:ïÜXöÐÞú„;Q¹)×£ÓÁei™ 𣶻mYæî¡)1»k“¥•¥§i}{;+D÷´Våš‚úlæX€^qKêÖ•°žÇwåùä„gÚ—È3öäù´‚#ÔÉãǯAaÑôÐ ¬ÎKÒŠ-®z*óó¹>„ÍÏ­íuŽÅuL%¹ÝLØšè–²Óè€/÷Âm ùq¥R‡¿ðÏÈDÅ䜺.å@âÑÄcnðG› sšÚƒZ aUäî¼tizx°Õ`ïäV¦gŸ†âôC”šLw?…Äñ¹â ÚAz¹ºÔ¬N`†È¨6Œ‡#‘L@£Úð  ­ àäUÃgް?Ç_ûéÖ­k“q“2Þõû¬WÃ’ú_´<˜€Ñrµ—†¢â•òD&p¥‡È#E–&Iû)i^låÁ:mÓýñ,¡¹»É4Cg–9ŸêÞŠñj’;|øù[[:´{{ïXæ÷…Bðjkmù½Å)pæÌ2ð¯~IÇ'$$à*kô,4éͺ^½.b°pøü´zTUÑÖzë*ø™‚&~…£~ ~âÑW¤­.©×Åg'e2%GŽTc²ûÚ°{îJg'ÿÞ;¢d=Ü®çÝ~Öce>²ÒeýæsC–dœðeJ­ª, P«sU* T•T&–ùÍ~¾#>yñÀñì.sûLöPCÝ[r Ðûkw®Þ_YU¥ÓVkRò’²™¦=ÈÔ£ÞîÁrwÏP6$*2mWŠgªÔèD#ÔŒÖsÑp0[Ëè’è’(yB‚RÉÈb‚Â"0v’3£Šcó¥:?D±ÁỼS1¤¨¨â¨’‚ìl†)*¬«,¹ #©$:?VÞŒµlM~­îx›:äšr뵞£pnuñÛ! •r © Œ˜vªÆ9Y§5æªìôÚ`~X!›ã5¾L ÓäášÝMÒ0> çB“W+Œ¼ï 0‘ʼnbÖ”Z#7{±¨@(!-¥™™¹9L}ëÑ“€ºdåºÅ«6Î÷Ø^Ü!aUªì\•ZoÛë÷j©6$"8ÖkÁƒ¸]ò¯ÇÿË¢—¨Ÿ¿{’Äë{·®»çêy§¾3^ÖãÆ÷9]m}û77U 8:1!E””¸D\W£-«?pa<¿ÞG®HùKß@Þ¿¾…¡`ÖcdµÕ%ÂÇÉG“h IN¥Ô¹s§ï}}Æaùüu›—•À᪞r­ãšô¼#†“ØT6uãZ×¥¾ ¶£Añ¨?ØMÉO[«ª¹˜É€ȃ«Þ¼*+MWãFÖèÝ#Š£pKe,ÏìžÞŠ¥«a,áyÞ °5Ïá·ôù’«GÁuêÁ¼ûÈYÏ[4ÕÕÝ磇ˆe[ÁjÒÓÙØæ¡O½üvIçúRfQ#½Ò©ã‹SÇïÞíܵc«“ÛJ]´Z°þì•‹çºþxÖiÃÚõ[s~£ƒYZÞÙ â§o~âÿ‡øú?w“n *ˆŠñQº¦3h—:T[lµ ?_UIÁÍÐû7™0÷ ¬Ã1À²ïøÜXXAËѰE³fay¤’edfd‚ª"ª"H$Ø¿»q#$áhÜJƒÛ¡w& ÁøyØVÁ£™Ðêèç‡O2 8™V%¨£Aµi³Ëâ…Ûõ·¾»þÅ­«Í[·•0êDU¢*‰@•ð¿´NYdľ²wà˜;ÐNËãΚ>J´ö6/ `}ÈBø^æÐ¼(`ë’Ò2ÔÏfÊ_þv9©À¼åÕ Ø‹d#ØKʽ—w;Èc[ÈôTJ…G•7fT·f0™-é5­ÔO?t›ÌÃì×î ä¬#‰>†cèŸzžÒ#L°'VÂa<è ‡ñá}ŽO·Öx‹¼Dm5 :gõ§ßõn—z?‚Nú.-ïüµ7ú×׌—ÇèöăQ…î‰Iññ8÷«¢uÍÙG.y|Fæ1í=÷Üþ@Íx ![·,;fÁ1¡’}²ÜJeg²25ù8ªÃµ¡ÑñnËÍCcñüh jj5Ex•WE7î‚ïÁEÏGȦšr( ·ä¹6ƒ µñÈ­JjéK1Ý€+Øí¹{Ü•O¯ÀQW`Ê8ú ¯¥«³ëWüÍçV¾DghÔY ›ªˆ(Š ‰—*’TóûòÔÄ”D ´ /‰ÒUekó2øàú¶O—Ëv_'{ûV§ÓgZºn3sÖÐñGŽ”Uä䔉÷îdzÈÓëŸ x¦ãŠ¥­ñÌ/og*#+;Žâ™šÒ0·=Æ™øTgtœ·–w–›Áç6r‘tnvaÐQU’Yª(¢tyÆÆ+€ÔH •9Ið °ÉNÉÁŠ–*ÕUh嚸lµÀÕs–Á€—|x‘»@gœõÊÀš7T"±êˆ|´Q©–ÔÛz .WõÔt¸õν;¼6nŸ[_Òy­)žXp¡—dŒ'†]+†|¦Ò…0˜}0˜G*|0B`…ùG 6ÁJ:?'/h6FŸ‰Á›„ôxU "¥2±´@™›À´C“SâA„mh¹¼(3S­ÎbÌE| 溵ÛhP—¥ÌF6Ý!Þ€»:}ùуwO:Þ™e¼MÊMÔŠÄó6ÍÙÂlû4jXÖN½2õäòûNU€ªÉÒæ²¦57œçÁFÌp¿bͮƼXûbç­ÏºBœ$‰ÎMÐ$²ÐfÎ!Ä€À’pÉ%;ÝwÆ'&Åa6Wf…×T|wþáYæóoK›F·=Öq£ðA¥7øð8üVW©ô¸8ç€__à/R9¨C™qåÙ©j%°•Ê¢#b3cKãØ_·ŒÙ,ŽQ É6¢(Z«ÉLOÇÈ".§†¦9‘m<ðmi-95iúÔ*¦LqLTU)SﻵEÅeÚ˜œÄtVR°øP\æ¯çl*òA1È¢ÊdÅReRJJRõoGc„¸ám?ºÈôÇqNÒ_ïÇþ¾V蟣lf i›üv´‘ ¾ÖÁ…wà¢w" êhȼ£ŽX1%ûTb+ì%ª|â'ÒRe!V‡u_«LäÃF5$‰ôr‰R¦D& %+Š,k|ý9´:ïݱË3t¿¿w¨-;[¥R[˜€[xá­]oá`!&ÿ>M¶xv0É6#˜uwLHœ‡o>ùU²ï¤F ä«Ã²óŸð>lå^ѱÙÉ $€äTE"²^g#OP&áÅjbÒ“5±¹ @ëë6¿n(ˆWƒl`[£iÒ”(3åÙ¬‰<ßç¥?ó&ž0ùÝy&ÿíqL?ÏMås{ÞZÓ9Å@ÕKFaʈ¸$}õ»}|4H ¶â2Yi¥¦“Ñ E·<.Í#QD–Pßß0€éolÝOomm°Hÿ”%mendstream endobj 364 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 2480 >> stream xœµVkpSe>išÃ¡\œãXÄsN ŠV¼"¬»* +J‘Ò+–B/iÓkÚ4i’æÖKnÍ›4MÓ6i“6½&„^ ÜZR©]ˆàHa½Œ»3*êŽÎ:®ÃŒúÙwö´eÝqüåäÇ9ç;“3Ïó¼Ïû¼¯€ˆ!µ%y{ZÚ£ëgï¹{ÜÊî>!`5÷îG,ÂâØðÊE×—£®e¨ò”y'!ÊU–-²r¼°@ª¯Í}@üèÆ?¶~ýFñó¥yanv™89[!•”f+ø‡qª,·P¢Ðˆ×>#U(Ê7=òˆJ¥JÊ.­L’É þôÀCbU¡B*Þ-©”È«$yâm²2…xgv©D|^ÒÍëYi¹R!‘‹“eyyAt™¬|k…¼R¡TeïÈÍ“H ‹ïßD¯»ˆ"ØM¤iD:±‡È 6[‰ˆmÄ‹Ävâ%bKÜKÜÍ ÁßDˆë‚ZÁû1;bÎ S„çbóD«Eý¢d>ùÃÙ‚ï)ûBzátœ.nŒëYÊõ˜}œ(*àž˜"úmÕCX)•ß7užñ2]ö]¼Ø}täìPAo]QV6ŽÓe³fEUTf¤pü«#h…ßeï+,·ê5F¿·ª@ ”Ùænõµ†šƒ¬§w -‡7©K’c;÷çjòóÝX~[%H¡¢JS¤-ѧBUïRE6zÃÍLc‡?Ê먮7Ú­ õì,È(z?*@wÍ óe!çD!úûíÓÏáE8¯ÂlÒô Ÿ -A+P"ƒXFg(NO¢Ø/‘àƒ©®Ý°`=Ž}vö;XEÌ}‹{iFø~’. eP$ ƒEëÈŽ …ËÁÀf‘§Ñ«Ž†AÉ$Ä#â{ô žùÎÖÌ<™\Æ*Îf9 r™¾B]¬ß5óðÞ°‡q÷úÎuóR9ä[O€ý÷ÓûM—»‘è´è‘ö¬,¼`¦^™ÅQNÏcq\¢lѺ J/ ²xé ½¨ŒÔ•ÿZ&‰ž@kQ>zí3)Zˆ—°øîô×$zðøALà?ãxvù3Žh%/V„‹§]íàå7xu*³N¢e”¡Ò  p,¦°ߟM¹:xÄ;p˜}§‹¶Þ&ÉÞyId¼$©$^ŽJè Ý_F‘è {åè€d&Ã1ÛæJÔ~ä©h¿¢gyéxÙnóQýï…ÏŒŽN€Ž6•f‹±äTe—¶§sÈ;qyëÐfÍnHIg ‡ŠBY@­“ãDë¼/|=n¶åHÛx¨– Uf†«î6_¼E/F…Üfô#m2Ùj –*~Û14Åœ <'áóÄÎLOjk~3„–æƒCˆZ¡¡GGÍã¹E›‚”wEˆþÅ‘ôW·ü#9„VÌ•¢1¾K×®)3•×50ùøAQÒ-Å?"_ÄâŒãû/¡˜«hix¶¯*­µšz¦*sö~ tUî°œ®ŽÛø?>#äLÜÚÕ1Wç~•Ç 3ÉL Ìü¼YQ€žƒøÂúP$â9v-d½Ãà5•ÎMÊÁtÕ¬Æa—Ïïbú¼y›]f£ÑZk®f%«´é°zz²`zxØ9Ä´©ú£p úCíýÝG¼Sà¦f!Êlz…™±¨t¯¨ÒÛüÞÛ!þ'(øzš‹{[ˆÎ£×i‹ªß™T2L5w7w²ÍÁÖ–P[Ä€AêÓÎú{±ÝæP†lŒÇÞj‡2PYkjËê*ë•fu½ÚX¨„•'GüǘÐÞ3¦^xFÞêö ŸFà å‡É'"–Þ_5~ ®ñ92áOÍ]õ$mùY9¸5<÷7~½ã|à߀ù ~.Éݽýnf­=ƒv"¨·²23 ßa,bÍZ‹ÔÔÞ!驯O »ƒóë€Amat{Ê‹ùuÀØàñµÓa[úO!Þ¢ÞËINÏ‘ë«ê³k_a&ÄK“Î]¦ÔÊr¥\¡nW†»CçÞH§-› TÔž1é›×§QÜàͽÃheªwå—òù¨k6·öð=Ðj õ]¸ÇŒç &NOØÍæ]SkGÓ÷ÁX T§·¶ºÎZg·ÍëÃÕóúûÌ™ ÑÔ/ä¹-©§)O§ÏÛãfŽ¡'†>«‡ÌxÐZì|?%Ožùf1Þ[+’U§~Ï :¥+h²Cˆ…KáWÛ½­—Þu¹Î ;¨¶Ûm6C]­4s#rÂu¾¼jx_Ö^ùödf_†B­4ÎŽèF«¯ Ä+huPð)×$ä’Ðúc’Ïø‘)Ù4r %ŠÎ‘Þ‘¹ƒVFýÎÝpÇwêªJ“ªÎʼ†ãEk~íÅ’\©t¼ì[dN¥‘FéÜA>uŠü$0RÊ€¡³¯¹Óëb΢xÑ·ÿœ®ÓßÎý}ÍlþZ†¶£eTŒîrOñÓkPÞSš/-Ì/ë‘2ÏâótŸ&")Γ(kBG‡0KkÚ¸-n´¹Äj÷‚hÜÌ"&.V#[¼0êY¼˜ þ áwâ endstream endobj 365 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 535 >> stream xœcd`ab`dddsö Ž´±TH3þaú!ËÜý»ï‡ÏÏ&Önæn–•?&}üÉÿ=T€™‘1¯¸É9¿ ²(3=£DA#YSÁÐÒÒ\GÁÈÀÀRÁ17µ(391OÁ7±$#57±ÈÉQÎOÎL-©TаÉ())°Ò×///×KÌ-ÖË/J·ÓÔQ(Ï,ÉPJ-N-*KMQpËÏ+QðKÌMU»ML:çç”–¤)øæ§¤å1000¥¥30Ä0Ä10=ÂÀÂ0QãGß©¿û–üHZú#i ã÷ÂKÌßwü¸"Ú}¼{Ñš•ó笜½³oZÏäîîi3šºë45clºä~OaïÊÌóTjkêjoín娟Ú5YîbÇîä©1½MÝÝ¿£Ø»Sº£.¦È?Õ5­›cÆ”îÓ{Úäõ'æìmYÖ1½»·{=Çw-öîÝ{º¦4ŽZ<û;Çšî ÝÛ9.J^ uÑw Ñžµ+Î~™0¥§b÷DŽé=Ír®}Ñ›wuNšµŽã;ÐÆMÝ{\ׇ-óémèæ¨oì®k˜Ú5AþyëªÈI}õ@w¥süÚ˜×ÕÓ45xOÖáêßYÝnÝñ¿=Ø»ý»K²rË«r«c;ºš»»8ê¦vOûv×­9¾âÅ?í²ý–™Æ¾k·‹XZ=ç¶9<<@ÌËÀT·â^endstream endobj 366 0 obj << /Filter /FlateDecode /Length 2769 >> stream xœ½ZKǾA޹ä2°&â¸ßu QdÁJœXƒÄ(r–KcI®É¡´ üãSÕ=ê!¹’"-À;œé®W×ã«îþ)cÏþ«ÿ.6#Vxã¬Îú÷«‘c¾°63žÙB¹LiÆ e2¡UÁ4Ïöåè¦[i¤p0OjɄů½7Há§ì²úÏb“ýe:úò®x怑QÙôzeá¼0Be–™ÂK›M7£|_VÇýv<ýqôt:zì¤RÞòìôáãÙ)_8ãv³üÙlŒÜp’&“&ÂB  iŸM—£Wùj?v©Ü¨wùÖ†ó¨¦ñNqPsÛÍ\v»MTYxÐX€ó30&a²Á—ÆsB€chçTdW¯Æ¥Uºêäïüåý(,XM·VÀ‚}þt_eé×q(«âP–Ëa"0¼NÎrÂÊ8­jÄ4¬)Cãhã$„ßÿÁËRŠ„r Ìë¤{z…´2ÑR©³4äÛDc¯´ü‹é!dÁ2£Da¬ Büq2q¡Bh&^‚±AŒÛ8©ääÀêžcAW0çdÂ*±àg,}\r?eµßîö›aüSêÂÂÒSn³|û˜2“ÍQ5c¬7>kß|<3€ Ú¦Ì6å| šn]Áyºd_äó|VŸ²RÃ(åÁ9dêòì•`uÂè0Lê\Ðþ" %„.¸KJ £”DéSûżÿ‰ÈŠ…V>"Ãæ c;xG@ÏM‡R( ÙÆ¡†‰› ž#رɗ?8EDÒW ²j{7ž’ç81ÑE(xä¶™I9Rd´&)Ê\¡à¢ðÎRHxGXÕDhO=<Ü žSÄGÞS0Ipà*š †äoÆ\ǃ©L{Õµ^¶0Ð+ü–ƒÁpí„U!‹Üí×Ûj˜âRC7Êmm÷aHÅ!3‚§'܆‚T b׈„Õj3 Ò¾,©Wüñ=<Œr5$  È ªïC¢Êis}PV£Êj˜Æ¨A”Ó@0QúBò„Ña½F)½„1ÔJA}b¿ÄJYˆ^“p6h­œ)d/5­öóaP½„Jé“”û «º†`\ç]ÍÆP{„À¸ð`í Ç)5xô¨,xþ]YÝì–ap³Wò÷Ñô÷qg ¦¹¼z»{Y•wa~ø];¢!óÏ×?–‹jý¦ÄW€Eȯ[x³Û6‚ЛßË+$«Á“rÀÿ$³FÚ3”…S|@ÇÇ1}¾bQúŽ$tÎØN'PŸHùT¿†Ê“ÝöM¹_•ÛE#¿7ùb·lq´oä!j¤ôi½†2€ÃM\º)Â$Á˜W¶Üý]ã3­‚@i[Äèš“œAéš®!l ÓnbŽÏ ð«r<éUÞõû“â©‚Á‹š'ÝŽ¦[Ü5ú‡\¡…va—ÁŠÇb L~×á|ÚB¤}n¤EÓ°z_\xçm½_^ÓhZ7æòÞëvËšv3GÒïÐuÒøjŠf×çåŽ»Ú €Š~°—B"ýƦ•Šp¤v¨ ΕHŽÖDÖÝ–0ºê>|瀞 Õíƒs Ê pUp¥°¾˜Xߎ›Í|ÿnˆÅbq—ð ¼Þ3ÆþWsÓ¤J6!KÀ·2Çüööêlf<Ã.7&Û%h…ˆµI“*¿ïÞ“!÷tHÅ:²$Ï/f!?’ Õ|a» ˜9kJ:ƒ¥°ý€µêMCË÷«ÉûòéùšKËU—8ß_FþVî·å-™òâ8_îçÕzÑÕ•—wPiöóÛYþv]Ý4‚æ¯ß&Fj“uaÁŸE3Läµ6çù?Ù•××ëźÜV‡tY%æé¡ZoæUY×¥—Õ²è–àé~ù‡kDª:+†Ò‡T° úüûý,ÿÓÏÕϵ0}1BÙãŒÕå­üc‘*Ô7Á¤('Œ×/$6( Fð¦\û|_0_N„dgÉ—¨Éa¹$ô„—Eú¢~! ®˜Néó‚{iKøÍ;ÏÀóˆ÷WÚç“iy¨®Ãð|Y®ö%´Áí‹Ýu7ú>-1¡6‹·>ty¢›ƒêPo@äßO¢ÝÏaF¶+±:Í@20]ÐR@ÒÓÂRW‚'§pŸo·ëj›X2}#3S‰rÕM –€ÎñR¿3Qt‰Æ/!(ÎB(ï;ÄÅ ç¸5æ™ÑÿúËyªàõnî_ú”öå-IJӺÚÕRçÛã¦Ü¯ÑB1šwwOëÿF’çQè§ ´>oN@i kÔ†°¥ŒÃ„S©$ûHÛ2ßD@ ²#=ÙJÅ=¯[è«Aû‚ `‘cm£{y+Úè´¶;B’kCm<ÒàYÝÄnGéKáurr#CqInÑlOBJç>Îuü„|¼Å#.y5êˆIgÆò¾ØöÁÒ=#Mçwésà^÷rá<Í?|,Z)'÷ý€®¹7ד-v£!•ÓÍë°j³O¬åy¹ih»þ‰WìsUþ„d€=m !j-U¯i¡wIi`Fœõ<»¡qpz‘Œ¸9&­³|Á•¸3íõ2õÝÛç¢Ó z;Þ R&½•÷P)‹lÀþyç>“.gœ^ 1prÏ.”½Xjb:: ûxKozWÁ&$èZàm8Å…œÎ쮞]ˆ6Òç¿#Ô:׌|È–áIU Äýá§çª;"@»Ca“™ómç É1+•f×ozWî¢ëÐXº'!qA‰ùÉ#ƃg¶gÝËæèY‡º+|¼Ê(ARKOƒO´/¯‚Sï?Ï®Ó&Þ“L Ç…‰ˆ- ˜ˆš ÎlÐvƒ«’ ºHîn¬ñb¿kŽ’ç‹éØÔ*/$à«.ÈÎlÇaž<а¸ê"¿ŒË`H¸\Ë0ã¨l˜êpf›v³åÅè”uqþendstream endobj 367 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 530 >> stream xœøýCMMI5$øøø‹€øOùJ‹ ‹ ¨ø|÷]÷X÷Tns‚Copyright (c) 1997, 2009 American Mathematical Society (), with Reserved Font Name CMMI5.CMMI5Computer ModernilJMŸøˆø©€¬ø¬÷ò÷AÝ¢Þø ù£yžqojpjo¡| ¬¨©©ºüt™z‹‡zЇx…zT^RY‹t„›¡ –“ž“Ÿ“Ÿ™”²è™¯“Ÿ‘˜‹œ»]³L"Sûp}›‹œŒš¡Ò»·¶‹ž–€p‡‹~…x„y^ }i„z‹‰|i†}…}‹zW¼gÇôÃ÷¤øt€¬øüÃõà÷áù$™‹‹‹œ|‹{Šû‚zЉІ‡‡†ˆ‹…z›‹™®“ˆ‡–‹ƒŠ€ˆû ütˆ€‰‚‹}OÆfɬ¦š¬£¨´šÇ‹‘™z‹‡yŠ…{†{Js[d‹o¡ª–Œ–Ž•u¡øC–––÷o¡ûa™¯ å  O ¯“ Y ×ìendstream endobj 368 0 obj << /Filter /FlateDecode /Length 4837 >> stream xœÍ]Ýo#ÉqÏkäÀyK†á¡³œëïK..AÉÝêídÅåê,Š:ŠÚÀ|ªº{fª†CI»§^†½£aOWWuÕ¯ª««Û?ÍD+gÿSþ]mÏÄlsöÓ™LogåŸÕvö»ó³¯¾7aÛè”›¿9Ë_ÈYÐðÊͼ„N¼oÏ~h–ó…Ž¡Â7oæ ÑÚ½÷Í>;!tln‡Ç>!œ ÍŸCŒÚ5×éQ)U³»Åc뤤MvøÖ·1˜f³&½ÜšÒÙ†¾ßã³"8Ý,I?ëÜ»Q¼íæjïn›¹S16ïæÖ´J:Ó,Ó§Æ"u7v©µo–—¹áÛòCÐ5F7÷¹KbX¦D»ÌT#P5غÿ2±… òQJØ»‘†ò‘écÁêñˆ§ÒürèåÐ$_ÒÖ;2Ù÷™ U‘o~}•Gh£¦Jr?ÿãùïÏ´Ò­R ”íü t«ˆÍÁÌ ½Jû-RÏzÌ<h­—ÛÔ‹ÐÁËnúá¿æ¤OwT´cÖ‚Ð>øÂZù /œŸdsŽoet¸<'ñXäy8Èå¢cs!e­•™Û;b>K¦…!#Wr¢×´=5§ÿ%Í—™œÖ@;°÷”l784Ò†²I‡†Ó¥ h9äÛa¾éX¾ú@ñXøc¶P®uJ‡,¿&D/šþcüà«ïƒ£h%ñ³N‡þvP÷Ü–!›ù›¾í? 4&ú5€Þ÷ÿ!O|ôª4ŽjæÚèqÀØX[+€ ÙšÖ˜¿I’0ÑJN F·Âú gì“‹ù|¡|… Í7ãñEÆ·o¥ j&M+L’Zó÷óóÀ\µÆj9[š'"3µ„ žÀ¼KçM‘z2Œ"Ú­‚)´6vÒù+>§#QúÖ› l×ø‚hï„,m륒r¦q¡ÈååPÈ‰ÑØÖyílÁ¾ùñÑA©Ö;ýh~ƒ‰îÉù…Ÿèüzë'ÆÄ¦·ûâhHŒJ’ŒOpnÆœ£–rÎ f*çƒKOR>oúÒSèç:a€‰¾ù›'BÔ b¢Q&ˆ·˜’‰l‹6D%CA‡¿¸±ÐÄØ±x¾=?ûDÃì=<ÿ~¦%àƒ…AÁ[ø| NM¢qsöúdTdHi‰qÁÌyøÇ©©—…2Lˆ ª,m#1ôNà˜¿¦ÁƆú£cÛdšÚ Û| Í>.úcg>°óG€3ze󛣂pDOéØ/a²1€Òú V`^m”/Џ‘]È–\üî¶ ^õºþ‹£¶Ü0$@C |7Û¢ëè›ì0$É–¼Pšélªæÿ%¢à9#ºx..»ð9Ê;r»§a¹›qyÊQ°©9 Ž­ ÝÔAÄô¸vKxTöYÊAG¶:¢R°KŸÐ&AžÒ 3Àšœƒ=D¬ý›Ç€Ž›Q:è² ©Kƒˆˆt¼ÇªêZ) ¨Jè&Ý1Ô!ZÞ†¬ñÄmN~–«>¥'Q@Náó<Q€åè ]þl¦|+DûQûø|mÿÌ(äÙê~d>Cße?þÁãS:³dŒüù®?uh‰j£> DuÃÍÃ1æ<…Óã‰Ö-¬fU ¾¢ \oN‘IXÉne”¿9Ö!t…+]Xôi†²v{ ÏŸž˜é>-ì éÀ>,Ô;îWi("‚øpMMÿ KÕ×ÃägºJˆh<ÿàò† z=ŸZÆ…ÖÂòžiZ6ÍìP.iÒ„&6®˜{œºÊâtÂ1È•³„É|ˆi•;²ü]Íú`èîÏC^iÓMZÔy¬…ø®¬Ä¥QÝB<õMò'„ú8ÙT¨?=¾4&ƒf ušZÒQI–|Ë©DÒ‰¼m²*c¥.5R:=4è]ª$mÒ Ó'ùºiFR6ðn–·ÃûËœ’$oHJ ûäbâ 'hت³2å&)R3Ð5xi€3¬-Zgo`Á)Í q£ºWVa³½)A‚·²à£Ó™ù‘mçöd]+†ÌÉø˜\`J ‹+¾†;å¨5 +’Œµˆ‘N:Órpšý”ix; ᯠæY3í` Á_M¬G2¬µó…±˜½s€žAÁƒ*¢7˜zô€8¶wh §cGõ*OjÉ ¼XÒ°€.Û¨nºÙnÞ죄¸Ñbc­ÇØ1!ÚÉÝÞ F“uÔ1I»€ð†¡ÐÐKÂn(Ei1ë«Sƒa\Qè¦jø ;vàëÆƒGrèï)(P° ¾†Œƒ¶%”¡?#ìc`ãè2Þ¶9\ÀörC s8‡Œp±ù˜»Aý@H(€[¶Wú~þyâ•r† …Í1›ev G9â¨[k-‡Š‘cËI[Ó}› ‚&aiÒvùØŽˆÆÜ&ŠMòR4¡Iá*áptÓCQg‹³~M=OxiêSQŒA±÷ÔŽRúiE }¹P>þw¿9 àd½Ÿç@g€ÍБ›IA¥Õ³ýúì |­‰ã=ìOocJX~Ü´4û˜NhL»¤uìFáB¶ŒÓ @X-0…ƒj ŽõÓ)áÎhÿGIýó‚’~HÁC”¤,´üdR`¨-æT(©7·«Ã5LØ@ðå¤( ¦=% 0:Ö:XкÄc~óé” B¶ž“:¼]–uø‚U›„…#¥öª_{(>eêð~Tä“5Ϲ03òÅÂáŠ4‘9ž7Å_„XòÙ#2¥é`^Úóò†$Àm«™…%0Fyc;òÚi^„RR6K)ݯª3+. ¡¤˜¹`mˆ×cÄìjk]Êä†Rk œ()YÑzØD½ªÄR²JÈV´Š£„ƒ®Ùݬo7‡·íîáP…5Ü9Ђ)à7u&K Ó?Q‡'‹‰gfS èRKU°5wÛ*n_bàŒõ)ms W/iè†]`´*™¶ÂÊ 3³–i£ŽÜÕKÊO¶¸‘E)Õ‚{£¥“ŒÔvyØ_WrÄr ó_³‰ ©U%„ÌHÌxúp1¯“°@káÊq»‚k4QwxN¬D*ÐÄà©:†¬@FäšÑ¤­âùáO«Ýö®Žw±–I›AÇv^!Ö"¤ˆ‰§VŸoͤÓ_ÿö×u̹ C­Š@¥$¸0ø’»Àè¾’Qw‘!w»ß½¯"2j.ÊZFí ×JFí@?¬âVŒ|\Tñ‹[µö­x¡¦÷ªé½£k#&%WÉ{÷^ŽÂ»YW²wmpoá4klºÞÖIæt«BꛬúÕ“˜vXÉjá Öâ®㯊>js½¡”W§u3€f>ŒÔÜÏż€†KÑjëíÝŸVË}%T‹­\µÊµÆÙª¨Ö…„ÔúÃ]¥Ø½„„ÖEÓ™H’tÊj¶…é×ú’†ÄŠºÉ¾NÒ„Ôêírÿú°¼¼Y×]Q’9ý ë»m{ m°<Ûlu$Þå+xn Yu½]Ò„TäVC8Sµ–ÚÆÖgÊé6F9bªà^VœŸ³ QR¶^Ñîë¬Çz3 "Ë BõßÊ/©ÿ7ÝÃÄ6ÝËë?%µº¼¾½ªšn¢Ô*§›(©ÿØVõYœ«Í‡t/‹« ô¾¯ëŽY£Æý’ÓV’º&¨ÖåUã~}xØWÜ3Fí¨eî&·"Uh³­È ߊ,h÷‰'™[¢jfp<9iþÃT%Ew®Xk•ÊÖNWÔ<ãŒæ©óÉ÷¹e³{“ŸñtÆ2? ¦Î•šX°$;oM‹¿–'êB¯2¬¾“ŸÓeÇ q„£A Y¦*1±º:‘ ––¤*K|/bªòÁºÚ<À$g5.a•­ Øe)8Uf縺ó8»2WvÒšÖÅì¨ÔXIªŠÖæ½rÒ·ïı“ ¥d–Òv¨£ü˜Ûâ,‘zPVMD«]—ÕK©¾K„DÔ`° îx¡8/[7ø›œ±Ã8,ȩɺu!ɉƒq™,?Œ2>Ä¢ñœDùò—¼„íèЛ”`Z=™…s¹NVãÙ£Q ÂT•>c™¡ƒX—bQ–à "ÊÑ'ñø ?~5"? ©Ž ú=¤…uOU¸¨¢OöJÄ?¤Åv ; «£Jsv&’ßKÄ Û= &}?Œc¢À>¥b)°/ú7Y`¯"&L"gaÚùSŽ0ÌæßæAs*>0á»K¦K5f*íɪ¹%)Ü$µó£âÖþS,Bõ.U·>MóÔ½ Tt'êðFnŠ’f½›[‹ëyé/Áejð÷y$F˜#CÈïã3x`7Lö³CÁÌpL§0:%g—ÐÐË–}’YJ'qz¹YLK×tšî/J¥%î´Lýñ‹&’‘>ã΂'ЧSq|_5*›Sž:{%€îàL*Õ‚®®H…æY·œ:«Ð—J®ú?‘o©XÒÖ®Ì60õÜÍ'@5{Ñ»¡wÉ䥯k†ô°ñ‡á‘ÌO‚=Nà}^½§ÂèS_ðé¬{ï >_.È-ŸFàE5¥²æúr¿Ü¬åJð620zÍ}Ê)]]ß*…ò…K¼'G”õýúÐޯו—ié`Ájõk…I2ZÚT*øJµ†#¾rNôÓ„0Ú¯[: 1&>TÍI”ÒaJj_/škn£WMqr•2%hS17ÇÙªUû˜ö=##%ÛZõ¹JùK°UÊ”)©Z\¥­\s’«®W²\~wuc9Qå©Z… ©>„ÃS­j%€ãÀå×—tºb¾3%ê±ìRñ‹À1!U©p¯*”R­3]°@H-oîÞV:ÃQPŸ«dyëÇ0Rª®{!”*¢§].¾ËjÇmë/1WJ8LX1R•ŠÐ; ®>W¡Ír[«š=##®Sn’Kô"£t_g“ çÌ?¨ØM xSïù«:⊠o‘»ŒûŸöuö ñJ„ÕŠÞs}Søèšë›¸j×*KÙËñõj}S `µ“­êK¬ö²…˜â ,Rjd´uÓ*δ2'îö×·•Ê;`‘ДØE³_ßW=eCˆU=d£(¥Í¶Rm@Q nÔ«êÕAôŃ 5:] ˜¯aK9Ì_ýj¾°Ê¶BÉæ¿Ö‡·»«”Uζžÿ67’¢K<³ÿ~÷ú°¾K­ð‡ßô-ºnþûòÇõêpýn¯ |£†#Ù¹‘ Í»åÍÃúkìÖ ¼D“´Ñ ':–Öã%%r^(—zÍÁÙBI‘:Èþ¿ü© ²`_c$?wåE0ª#ƒ ”q!ý 5þGh2®$öéxå?×ûÛõ ù仇åÕ~y¸^ Ìë;iöË›‹æýõám7Ðæò=~×*1`k×L5Œ›ã˜aýæÍõêz}{¸çB–oï×ÛåabÛ¼>\µC<ñí~Å 9ZI ØI1ÅXxu>±ùŸýEó/9ü¥ æ9aÊWØl¬[åõÌÀOxã(ªZè•ú»³ÿž*!endstream endobj 369 0 obj << /Filter /FlateDecode /Length 3563 >> stream xœÕ[Io䯾+ù û*™¦k_Œ8@bŒƒqË€¢¨nJÓNwSî¦f1üãó^U‘|Å&5ccZ@4qÈâÛ¿·”X?.XÉ ÿ¥ß«Ý+½qV/Æ¿wŽùÒÚ…æN–^-”f¬TfÁ½µ¥v‹C}q ¯K¥¼r'@áÇ Ø-Ò¯Õnñ׫‹O¿ãŠ/02jqu{e;o-,³¥qðdwñïâ“O.—Zè’ ^TÛû—Õå’k ÿç/¹a¦þ#cñܰʉzÉxºaJ#&+$¼b}½äüò?W_p¥×~qõ‹«?ä¬nê6q*–@Ö)1PAÞ¢Ô†»ÄŠ)_z_, ˆ¯\¾P–Ây‰‹ÀqÌç®Úí2•„2„BáJΘ>¢×QKÂðþ†3ÂÖK-•¬›Òi]oÛŠJƦæcÜqBZ ; ±'øÅØhàÜ|Èç÷' ¿Z^ÕÇö³tƒ£w‡º>7šÛaõ-Sƒj\¬—¡“ Kë,°>0KaC]~RƒZÝF9`E)qGs~ú3´úW:©º`úhàg­‚-íTÔ½»È=‚ùþþ1ЬC?šÎsñHäÿYŽÙ@Xp­VòFå…4¹ÚW›(\ã&Ï)&)<©6'$“öo¢XXËhAêÒædm꣖–œ·Qý­>sõîŽ"œÒÉ@nPPs¡9þºkÃŒg>£:²HTn L™jTëh'Ø¡6…¼” Šàcíš”‚X΃ IHçs0Epý6§à›¼ƒLŽò0åwô‰dJ°ã3Úõ‰P*d%í¡PÅr„ÿHB: ͆d²,ÙDXÍø³ôÌ÷:„cÖ¤4¬c ã êËHD{swÊ?ãèU/ÁêÈ7Ñ€`)´A¨w\–L‹.óÜe¥…Æ&YZ¯vQ(½>5OxuŒÒ¢k,Ûdå<tÓYÖ_ô ¨„wôBLAõƒ¼ü¯X~=-xIá G]àaÁkºIËZÛOZ0œ5´€'¦¨4oã ¬ÿ´<Ò‚ö›XtB¿w‚Zѓϳª#µ„ ­Ç½d'JuÒ^œTÚCÞ!EFoÈݺäìUÑê!C Äþ¡¢¡v99ñ4´çÔ¬(É:šq¶XÕÑŠÙÎ<0òQc£††ºI³¯S·CòM^>78Ö3âç- Æ` ì‘&,54Ø‘Íõû¨‰öü¸ÂéíFÆ%2I×l‡Xlî‰ ãù5a´Ù';°¡©A|§®FÑ®¦.-û”Ñàb΄ˆÜ“ë–$Î ‘l7è?Ñ Üö¼'Nt ðUH¸Ó9|)bi$®›ÒÖºŸšƒ¯8v¹"‚>adG°³!•®¿¹œPDJ¨õ¼Wd¢Œ ¹¡¯¡Œ@œ½ ;‰!.ªÑfÏ0Þ`õ".£}ŒoJ¯ihâ]X]qò‡®ðLÌŽj3Yç¤`LT´÷Û‡ATjú¥ÁýX 0!1R…²¾³šÙW:Ï»ª‚áåLQJj4ª;Zð‘›ÚÒ£òJ¨’\ù G'¼ÑˆÖ°ÁÓ}Í6}è6sýç:©Äû‰FŒåã \|Cˆ<S`€ÁÞ{¤4УH“C”FqŸ9q¦¤×ãm!A¤ñMÃg„ \n½17y%RÄÕ>é¶ê1ÞUü;–‡}™ãiô-lz‘b»é7J~£†3¶S„‡0:‹¦5´´‘6Ö¸j´ù2·×››ØÌç§Chrg¶M~"¢œ”ÊEˆÁýFõ(@Є0ÛÏÊ7;á‘´S5v*ÅCc*ô¨RA3•pͼcBø¬TÝx¬Èƒ©Zåï÷ЍSÐ!|G-H ùŽ½Ð¸1bD·—<Ø'Z®Ã& ÆŽö3ESÂqì!ŠU¥M•’]*ÝOsƒœ ƒ\6ž8“´á°Þz¹ÞÖDl¤ÑI9‰}¼«W: T/ålª¡¸6¾Â؃st±8¸ Ú¥Ó¦4ò6Ùã…~ŤžÐQ“•»”n5ï, ìI…OÔ(³êIÿÔA™¢]:3ÕÓ-?aðëÓEÀïr¼ÂE椧áÒ)3ü¥ñ_‚XЄÙîK©ð>?Û— Ðs•Þ¿Ú]‡ú(.¯~èD5 ¦–ÈBK&,^@Rñæ—óâ º_&3fZVV)Üa%xi HPVw»áõáLÈaz±Êf¼®‹;ñìxÖ²g”§„J‡úc¡ \ôw~9O˜JaÉx6÷í-·s¨3OhN)»Ï 'Óc+/ø+n¯¸_Å ŒiEÆéãýv·Ùß||ÕpÓAæ–<Ï„LKŸqÚ6¯ëÃyôr€;ﲇ5áðk†Ö«óèä! ýhYªÒZ¡* ê gÚ8 -Я`¥±(f¼ØY”ÂæW3=€aÕ¯ ªa0É}²<¥$T&Ÿ–ŸÇPÐøag‘ú!uâ`®ŸÂûÐ]Âÿ½gè'q>v¢:÷É‹ýíyL¥ }°#ø_ž'WKÛp*ãõp¦\-x`TƒÎ“«¥ƒ¼ ÝäjèJ¯Ì“$ké!3è<Û‰óhå!1˜ùná*…,˜qOX~k6Î¥‡|aÕ“ä %p³ýiÔ’¢”ì‰ÔR†î°.¯/Ï’/Òx(=ŒØ\^LJݮ:¼=Ï(Å!ír—ñ»iÔo‰6iŽÛÐØth±ñÓ_x4ÿ%ôÕvûÙô™€Ýîÿ`J> ŽŸ +߆OÈß ÷É’7tI˲ý]àü×S´þ“u²®qž Ÿ[ÇÑ ÉæéâÕud©BfÝbɉ<‘o”g ¡ˆ…²g+Ù é^+Ï N¨ò$§ ß­O·ÿϺ}Ù¬?¾¸n_7ß·õýû}ðþ÷ú°¯·äíoªõ¡j7«ÁNßß׫öPm¯‹×›öe§Jqó:sHG·!º†³I‡w°ø‚œ"ȃIqŽ> stream xœÍÙnãFòÝ ö„ÌC¨Å˜Ó÷áÝ,°Gfs9 ä!Î-Ñ%²ä‘ä¹ßªî&YM‘š#V° áPͺÏîê<Ÿ°’Oþ›þœÝž±Ògõ¤ÿçvqæ˜/­ÇléôDiÆJe&RËÒ7ÙÖg7ð¹TÊ[wðžŸñ€n’þ˜ÝNþuyö䮸Ä"£&—7g‘x#ñÕÄ:ãà—Û³Ÿ‹G¦çZè’ ^<½_ÏöËÍ:½á®¨_T«ÿªJïmñÙôœ3Í`qÁšþrùà/½ö“ËoÎ.ÿšCûï¶š/«?L;€õA|[ïvÕ¢¾À•Ê—ÞÛzUí—/êŽÄÙfý¢Þ.êõ¬n`›âªPWÓòó˳ïÏØd\TÊM²†Jʃð8c@dBDQýs5=g¥cÀ‹(öáÙ{iŠgø¨½çF›{ü‹aLúb?(”5ÓË㊸RqST×ÓsáJí$™ W¼˜ [ Æe»ŠÙâ¿•@ÇUøcÂóä­wΙâãîõ,|è½R²¨1„Ç³å¨øH¯)vdýfE%²¤C²8’+æñCíTA`'¹XO9µ)€À)æMÎo»`P3æŒ,ô‡õ0o tº¸ëV(Æ;Ð7×Ý{ÊÄ,¼+} …Eâkà¶ðÚ«ŒÜ–eRFS”`±š0°Ë9˜Ó. MzÐÈšÀÛ¼„mYÒ cÒ:µ––ÐçûM€5 Bû@Ê.wž÷_7b¤Ft92Ìf¯©^ÿÔ“—szE–4Ô‚ZH”¡”P<óˆ^;M bÕ:—û!AyGLàɹDÐðíŸ@RÑ “ñ\ Éj¹”ã$x˜ôš¸à(L”\ëÆQ¨hˆÛ/"„œ¨—ÎòVbÆ3ßhR[Ð*ó*~kýÿ4ŒU+òžÄƒMOÑ |f>âµ1ØñÌñ(U3Œ¨Òsˆ5ˆ>€æ¬SV ¶yÓÙØ¦·¹F£®""ˆãÙÌ”€ÜÿÑlÙüƒÉ,a¼ïIȆÇLVh`‡ª¥qÀÖ€-)(ý( «? u$H¦×4Ø${°›¾Ä×P€y¡"¡+£F3pÕì0Ÿ‡Œ vlb©úо¤6€…†@­so¦.Lr°öwIÿÁ7ÙEßT–Ä0êI¤ZëU”è¾ô}E‹žüà )Y¡tÛ‚ˆè£ÎKâ⬾õ¥ƒð/šÅ½ü« Ø…j¤©D¡šCÍ+åBuˆÏ†ùñj@ P ²h( ¬“Ä^QSz¡X°Ë.uT»6§0°TÂ×hÏU|d¿ë¢I/d$$˜ë<臌$/»êÕ›Á„€” ñß!KàJÚÈÇýÅ=KŽ„¨Vlω”)%YJ'ŽÜrFÉïÕ³Ô'ÑC¬… ÕoRZ¢¥m?E…åØŸÉ¬k«~Ckãv»´ˆ°e^÷|˺‚1æƒøÁ`MñëH®œ‘ç‘|2”à>ÁŽI}Tr#ééH‰ ÏXȼ|ž±F$²iòn‹Hž¸ÁHYxPH“Õ©à|3îÎðÀÇÅÕ44×SyB駤Ç81tR,@UÁ;86~©G²X|fuë;!Ïåü`Tu©vn“þ˜Îä& Ó²Ñ8‘4¾ù1y}Ðÿrׯ¾ðýR{,Z£r£áŽ<%ÃÌ•b†=²â |ž[ù.½ærÌ]LwñR¹wXW¡°ïŒ2€®‰æ à ¾Ž; YÃiêí¿ôv š|(ï»pï5P‰`ÃFaSkÉ[N½´pz=…ÒŠ;åÐø”¦|ñψ›¥Ò¬ðì5bPpaOùpn›Må«ì9î®@’ñ“Q+ê,”ª:*Mñ£}l0?¼’Ñ„},B|,2R¿^$!wÖø^í+­Ïïã{1¡Ph¶¾WÞµiâX¯'L¾Uzä>K0ºœÖ2=Gê<)ÉÉRmˆ|?GBwìòòs9Eó°Ì†,Ë!Øs¥³Î:Â’R°þÀgÃõ1…7«IÏu˜CC ù%DuÜ(t>{MðPÿ9Ðf‚ˆ¥¡¢è‰)‰Šù±8ÖÀŽÐu»ƒˆF¤Æ±¼h˲ õ¦Æ¬¼È<.5Âè ½mÝ}ÒŻĩÀ´ ;åý T—³£µ-…B£Æ=ñű ò¾õ3Eîh=Ï‘E5‰yívLWÉIìd|ž´O¢ùéÀE>€ˆ<øâ§ÿE$€Cf¯ Ìöo²³ŽÅ&ôãÌ;&„ϤKMíÓH’ƒ°<ÊaØÓP?ˆ]ÞÆï„‹{-ÑÅ2÷°ZØ@¦ÊΟ èL),<ÒQêì'Ì{‚›˜¤™0j<¨ÑÆv¹è¦u ³×*3Ô²ü D ‹ó]íWT5½z¥E´"¡#!ÅÐu6Tè »i<(p‚ž×Œì^”Îc±¢W0’Ùu#—\ѽ¬¬'¨ÎzXµcÛw»z/øákãcîbx¾ FNšŒ)õ †õ«†gh½c¬ì¡b&ÛOi{l 8Ù]ø-~ êêŸ !¦a–.§ÞwÈvgt9îì³È7’GTC!U{•ö%csO¢U;­æIøK½¾¢­Ð hÙöDOû!,5ÑY:ŽÄ=A𙽟M5Ãv Ýn`£@B­%ÐÀ–íMS8uNæQ`,ïš(îFKTÚ<‹à ;¦ºñä-ènÐ,L»o׋˜éq€yÌW¬å:^0Káªûmª†B(”³T\v¼téofà«Ù 3x xL3Ì`¥‘¸kSJ-¸‡ºþàa-=/ñ„¾˜WûjzùkCêà Np&@:CvU<]®«õ¬¾š"Îsî@E^Rç(^IBŽ¡jù¬{,`© î(tä(þ~~Î9¦Iž¡J|ÿœñ¦q«j ¾…'¥6ÐÉúÀ¨ DÚ !ï0…Eï,‚@ôâDäk0uýà §ÑŽu%ç9²Ç÷àÉÈòPoBšÂk݇`‘ÛÁ'?}ûõ''áK0(_´Ë°ý½Làq˜È¼lÏNãe2t›Í¡Z)xÙÃŲ_L†jv¢HÅ“¦«‚¢’í)c€"?iß|xè ¨ªÕݳ…hGÑ:ÃöÙi#ÁÄKw®pþ Ìâz|mY‰n¦ëúTù74•ü3”Õ†‚ ãi–t‚²  /1°SL‹êöö4Ú²Œô£Úz¸8( D'ˆ÷Ón~¦ðŒÆø^ |%c5÷GŠ †r3xrqñ@0ÏÓÆîùvIM¤Ôã™ã÷ÙX/>ˆÓp†í15=ÓJèD¹ÉqÍëÕ‰B¬4a‹øÏ±Òj¨ s¯=M½ õX‰»¹ºBa( ¥h£³Âp[ïNÓ€µU&iNZ ^¡3T‹ÛÛÓ¤e ‘ Œ‚âº*âñ+ùxÏNTw4Á¹¹ÛßÌNÅ  -™ÉÅIÝâá:&ÁA˜Vd˜>Y¯n—ëë5M*8™K’f³‡ß{Qé$´Õ!ïÜCù±}}¢M<ÿv¾«àÝÉù¥Ÿç8ŒîK΂Îÿèþ4~=åßÕju1xëÜë w´ÉE•xÙD¡/¤÷Ná}óž,_é–@kÚ‚m×è)­ÑµÇ™¢ë{IwYÊOÓ¯ÊìŸmæÝšýËÍûú®»¤cŠ!~=.–¯ëíº^8ßßWómµ_ΊEñã]=Ûo«ÕUñr¹Ö[\¿Ì˜m  íše¢È¸9P˦¾¹YΖõz¿ËÕ£¸W€>ßí—·Õ3´E ?îçí½"^|¾Ýâ¦}Dö[q¾¨V÷ð™à,öÝöªøÇïûß=}Jb¯ŠHw¼ôÌq­;´Ä/º7s!-xkãg¸"ü3?tW‰^„­bi,¯¸Ò^Á†œûB”ÜrãDN‚wnAìZn1«²¯¹tŠuo$o…—ïÂ@,^|f °=I+(J*Ær!B?-%1‰`œN YçWç—õnÑÝ)›×‹m]ïºxFÓ¬¾Ÿæ¸…ßXÞr×ýÈÍÈý5íч–XñÝy´œ!!\†e­íAบ~Æ.:–%Æ. !$#xÁ—ëå~‰² [€]$(›ûguqfÄ2‚TMKÜÂÚ’ š¨±ø+‚JêâŒDKgëBw­5Wô•FKâš¾¯4`sGXDÿDÿåúf³…n*Æáb`=ï$¶ß JHëûÛz»œEéFº /o—o"È^À< P÷ g›yÝY ~o¹$ù>W.…Ç{ܳtþÿóš%—&œ¯„k–OãѼŒG¯ŒÍ¹‘c»þÝXŒçøôx›Ñ“Ëçä\.ÛI{8¡¤x?ḡ±ŽtIjœc" ¹·„ß‘¹¿w¸µr0`ÕãÿèL8,ðLJŸˆƒ›1‡'ñdB‘®^GÚ‹ì >×ëž7aÊ¢QvŠXÁk<áÔÞ©ÎÍ,TׄUº½¢Kˆ´r{%(@Ý“f¥Çètºˆbì Ceciù¦‘IoHi×Ñ56ð+9ÔO³ ÆûÑY…‘Ñ¥0ډǓà—ñàÜ0ÕizÖÅXÝ Í`$<)¤Dgc¹ÔUòq0æ!B½NW„˜7;ò‘Qê¸QWc>92È´I7†LoN±jnùþd¼4&Ò=ÜtBNí”úë"åþ…Ë€ qô†Ó‘k¢ï0‹“ Ú ]œ”Õü½‰m&PƒXép­+…•MpÈã^Jø„©‚˜øÐ­!pQe˜ŸdŸ\äã=’ U·$}Ô…ˆÁ»—»‹Ñ°¤ö8/†¸…fP2ÝNúþ ²–‰×kºËiH¸*¹Qî kïòܪ}›*è\‹2¥1-'+â„Ëþ O vu;¾’\Â¥7ê\>£&€wø¡ÃšoC>ÜÐÏÈûz¥»×CüƒÍ  *{Ðb{¯çèÞ<:I}d¾%äPKùhQã¼8N„%5”éž nÌ5¤iHÇd¶ˆFŒ4öÒ{ Q⢋$ï1öb™Ã)—vê)Áýɦ^ðˆÁÆ9”Õòz{ê‚UÛ÷;$‘GþW$ºtÊøÃÿÉ7(j׋®¨ÝÖÏï—Û®ÀwÅ]5û-¯Z²ÿT©gy—‚m òÑÁ‘ŠÒ?á¶_8Üí6³¿?û´yÜ2endstream endobj 371 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 367 >> stream xœd›þCMR5'øøø‹‹øáù.Œ ‹ ‹ ­÷Ô÷Z÷W÷Tmq€Copyright (c) 1997, 2009 American Mathematical Society (), with Reserved Font Name CMR5.CMR5Computer Modern2uøˆù<‹ë÷á÷ Ö·æ÷ ÷¦öøá÷UaˆxM{ƒ…7‹|‹ûXͼÕÃdzåÉßÅ‹õ÷û Òû#û(=)W·‚™¦«´¯qŸlާ¸Å©Î‹ìÜQ)7QK>Jûˆûc‚ЋЄmøeu¡øCœ÷n–‘¡ûb—· ä  7Ÿ ¬– äœ mè‰àendstream endobj 372 0 obj << /Filter /FlateDecode /Length 5084 >> stream xœÍÁÎn]1%Ñ]`RÛ20ñïN5àà Hç-Q©·s^¡FeoÂ×–ÓÏôBŠŠ–8R*Z–yoŨ[L›q+¡Ë;uNJÀX‚†£ø_¢¡Õå7êÈöEÜùÑÖ:‰7¡Ï&l²Å›q ,,ÜÃNo2KƒòSdï#V$;ð6Ôh¦î 8£ˆŽ´É®1#^Ä€ Z†·(÷í®¾Z$^-²¾µ¬– “„†KY ÍóˆUÔ8 É"¨âv¨¡¦Ÿ?ž‚Ï#œn?¢øp¸y”âYz|”ž?¢Ü# …’üç­˜¼ìº{s €1Ãê 3\¾Å#€s¥GTšEŽÊêdönDÐZ\ç™ÿŒÆ‰XpÄu#ˆ|žÙ,dÎ7Å&iá¨Ú «DBq´èÙm2x˜Mo4®Jݨ3>!ŽÝH_x¸!"c^ÙÃ.ÄgC…ÊÏ'½ˆ–ZÀµÈFKýÏÊ1%&¼}éðÜÇq L¨»wÏÖp}È„>Rd¼>Ûd©·ÄåALþ§H”]-å+‹úîâTƒbÁX¡‡çÿ¼{EìßĤÝÔ¾Cú|™V h‘“"^Q Ý‚ûÌB¼å5œ¥3 ÊXKŸD§‘\ s¨ƒ3»ƒ)•³p!ÑtKl5cÄY`(ظo2Ww&ù¼@¿ãd¤Ì+—­peéGy"$ÜÔ%ã ˆCIed–=$dhâyia‚õc¢[ ÐfªRKo¨²S_» ·âÞº°ñ w‘’ó!®t€÷v dÒ!±¯áìÊ•…¿ÁŠ}áŠ/ÏN¾=€Ì­ÞCLýÕ „«p½‰•1ÂP¿º;èà`M#·'¯›‘·3 ‘7lÙ»°¥è¥V1ô¾Ÿ„yAÀòqЖ(Ý?œžýO½9žÜJ"WÄÂNà'1Ô¥YçÝÈÞÄ/º)ðÔJ7ÊÞßO§'{µ2½‡)AQÁN(¹’`¤ Š+þ6 ³K¸à…¶dµ.ÖœŸÆã„›fÝ#ÔrmgW|¶jñ‡à=Ž‚ƒÊ”)4âå4;xŽÜ+¼ÔVkô$œpÁÿ †ðf¾ê5÷\%ôÈI%öíY`˜N¯`53öIÒÑs^سøW¬Aþ Zô^UŽŽ‹‚×[t©I B/¬ÞÙ•‘ 5Ùéå¶ð¢ÞÅlŒ6Í8ÃhºãLc”ãàU’o ‹ÙS§þF7¥ˆµ‰ Dƒ,Š ðZYïˈw”KÛ„¤ç¾vØGð:‹Ä 2uí\)õ¥ãl¼·u†šx}ivðm>FÑs¬À“ïÎÏ©[E×ÒXä‹´ ZcÌuŸ9áh G]¶úŒ³tc² ¤Rjº}!ìu"sžžâ¬ÛÄg¦ Ò¨·9Ä) üQLrX$çr…QU¦š&1ê¬5Ê¥“EÈÆÎ£Ï:¥ÉkP{ŸÓ4™•£kcrÂð·?§Ñ£äS±F/‹#…šâލJÓ æZ &…B"¿¨Î5Ì_«JÎ0X÷Áô›ò3nb¹˜EqHc-¶&Æ¢¢æ£Å¢m•À¹(ø!’4ºuÜì:bi8o&fÙìÈÊþ™Lûm#WÑ®;ŠÑî¿ml$Ød¾¢&aqoÔ$´¯ö®‚©ñte‘/ (•ì#³F<²( µcN2i¬×… ”#TWF¤äèi¦£ÈG‘Ð’Î,Z ¨QLK 9õdLÙÜ+ŠBQ¶¡WÕ¼¦áL´üp…€ÂÆÑÄFu9!Ts°šÐ*ëªë¬Ïî´)70a~žJÁ u![ô–hÉžiÀÖÒý„<ñ˜’3!őۊ….4 #I~Œ¬^x¿zSÙô×ùù®ÏböÚ#CQÀÐgr~t_'Ú—G²Iñ)™#8k[‰³_OwÏÈl`ßOzU%—RœÞºfélzAÒ½¬æå ÐÐÚ²ŠßCä ëMpÞ«wW' Ä=ð‡./z!Ka|u}~ !`Èn1× Œ(Á`€¯‰¹ Ÿþ¿ÉwÞoOÎþéO8.ˆî™àݯ†ÛÛ—Á€ÕßÀÙœc®1[í-æÂWàê!#㸠uè„?¿œxF¸!„¨~Œ74zÎVæxãóùx)Þƃ»<=ʽ•/cÝõ<*Ä]DZ… ûXu,dþDNï¸Þ /i~çqÞ<<ÌkÈ8˜ì˜’!øgˆ÷ ÁKŒ÷¸¨BcÛ·ïhGgYýY‡×c’ÞjU¸—¾ml3BuÂR_(í](ÐTÏzðÈ0ÅM¡¦&n+ZÂÛˆŠu‡BŒ°3“­.ï¿«";úüm¡8Ò®È & ³dEï?©q† „˺õc¡ßá>b­Ý¡§!á* Oêr µÞ2ä˜o˜Ÿ^.U "ø>³(Ò0iáÜR¶ñ[Ìñ´ê´æ9«.ÔFègJA*g‘OWú²*›ˆ1Øë×z83"5ªTÑ@¨o¶L¤ù7w„Œµbû‡ê]ÅQ(Ú嬜Œ´t’*QMªŠ¡â<„›´ ¶gÏÔ3³„†­•…À=±E¡µO:W¶3ÑÙ/ð!­Ú†’ôþdÒ­VÍô2‘çM•ÝÀ~²ƒ§Î4pr/Ol :ð @p8¶_Ì:RsDضtÖuH ˆ¯æmxÍŠ)}bq y‡”‚Ï?˯ˆŽ³Ó…&7sç¬/•ÿ&ñÇr±ÐøizîÃØ9av¤ï³êÊU»…ç]ÍäÙ°’ ›IÃûD×ToSßTÂ}ð SkŽÀla~¸u‡c«+ËÍK»¨…6åÇ3°g\Ö‹¬™]? [««·%ôU5sóôXöÀÓ ÌÙé }‡’×[JNéº#ú<ÄR|¬òa2GK=—`„-%=j]ãayw=P‡€Ú“ðþ›f ,öwpƨx…‡°XéówØŒ´Œ4žaÆ|ñÆ›Lo$¡wioW¥|«V¸ñ‹ñÁ§ÅwNq)ð •{ne‡«a9Ëþ…áV+by uå¢IÛeæÛ¬ÿ Ñæ®™.óËcËé#J¨ØvWO|^ϰK˜:1Üù'œd»E*:ÉZ7{´’ІDøÛÆ6³Þ)ùÑß*•;_„Ø0•¥Ø 5åŒmùž2÷–0GÙµüäâg!}ð+úE?iNùâ“4xʉk<ëå*•¦ê§2p!*\«Ç˜z7fYX‰»È™þkd5ï~Eÿ/'¼ú8k!ÿš¦ãÛ®S“äÛ´7s³JMow*×ڌщ?øp>`[†?ê­µ™6bO©¾cu-&Ä'ÕYyÅ º²ù7(,þ=§º,ñ|Ș 1gEêB¦Õ¸l«Z”Ô yþ©oôƒêtØÒrWêËgêñ—.¸ÊºQåËÎA2yΠÂÍ=Õ¾Š;‚&.0(̦÷påeàJ¼ªë'ÈaX7†¿'òL¼Ù™ÝXXzw7¼è.›{à ëÄRãèì˜Àã—AN¢Ü¯ÂCÃÂÅÁ¼Ÿ®¢YÑ͆Vt¼ÔA¼wT%©è5Ò OhFÿ«Ü`VÿH¦4JÛÕo²Tôz,þLz‰hÒ*–>eø-)õ3ë…²>°yÑž¾ÀµÁí;<%†n»Œ P±äQ(x~6€^8í~¨«þT•“Yåòá'«œ ¿PTw„Æ_³h5&lrý×þ‰ô“Û殮Զ<¥¹yta°Œ1R–>ðÔ'Ãi=‘ùÀB|ßxðÇœbÿ#_OçYÎÝQ†\]r½.pLŽxxϧWxß“¨p†'9ˆ‡˜ýŠ:‚‡]‹ø&ý íýÙÛ·-žÇ„“27„#_h'IvBBXœdRÛêgèïíôVÏh!°`ˆµ-;Œ?ZÇ—¾ÎçyÔ ìÈ=Ô±›(¬ãwPhÇm# ^]¿(;LÂÜO.åÓ­v¤¦X˧°*w¤Úº5u°£µCéÁ&ùòxŽÒ!ÂÏá•§u¾»x{´VvB—gvÜš-g %ò¯S²%°BÅVþ*¶p³ÔÅ+¶ $Ç>¹b‹›3ü„ª5AéÜK…*4\LváÛ“ÿ®˜â8endstream endobj 373 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 2906 >> stream xœVyPWïf˜î•Iƒœ¹< ¨rÈ­¢ˆ„CD1bD¯‡º±D<ˆ¢€÷-.AA0^‹"ñq@ÐѸ‘D“¯§Þ°»o Ù-ÿݪ©šî™zßû¾ßû¦ôõ(š¦ ü“3ò’sÓ’uo¶‚9-Xè ¢Í8K³K“/ÆRÈP„ õXÞ2Õh@#!a%¢éi^›l¢ÂL²³³÷^±2?;-%5Wîâ<ÙM¾4_þç?rŸäœ´”,ùò—œ±befrVnHZæÒÕ9òˆÄ¬œÿ-üÿJQe—µb¥oöÜ¿ÜÕk×. ÊO ^––š™‘é<ÙÅÕmÊT¹Ã$G'ŠšO…RaÔxjAM¤")*ŠŠ¦l©…”CyST,åC9R¾Ô\j2åBP”5… ¦B¨Ñ”1%¡>£xÊ„2¥hjUIP#©Q”=“Ò' ·S¯hk:ˆÞBŸÐãõ6èÕèý!2͵ê³ú úkõëÅÃųťâÌt¦œéfv{ŠmääÜjn§–=¬bØ¿ ² VTÔü:\o¸œ0Bœª0nmóì‚]¦’S­B¯-ë‚ˬ¤ù¦Öîû—“‚¤øß‚œU6MJœ=PLæ"/®‡%ësÛû6ú¬ T"8*œå±«%öÁ¾=V0 lz?Âlœò¶—yòoš<± æéäÞÜlSËtEÚÛvÒÄ ]×U¦’æ8‹ï‚cLMT]N âÀ¢hpŸ) ‡e2‰³ O_Ì=g%öžŒ¤Öñ¯oÏÄ£ñð™Î.á0FÞìè•i–#¥pWI÷ª µ¨×N3bÂÈÃú ÅÁ2|šQ˜óÂ]˜Í‚õ3[‚ƒgÙbk²Ú )a— ”tŸZóà%/ŒSjÇAà¬ÔîÈJ­cÝǵŠ- ã‡dSɇ‚x”-ãú&`·q/Y‰ó΂ÍE…æQhizœ' }K¤€æOv€~%î';ȕڂL5c$lB¹V›T"!P0âwüæ÷#8lôEíŒËË`T¦¬}íGòÐr³˜Ø/½’÷—çI7üöà–sœ+³h ƒ±É1]Íý틯X•¹ö;”UŽÎ™Õ]9ÞòàlVäN©‘æp®¶)¡RIרEšçñͽ†Á&vØlŽeíi\}– õqà÷+€#ˆ:ÂÒ·Ì*°Áú/±;rAÞ¹áñS}ã±aSçÂXì 1¿´Ÿ{þHzÄV|‹ v"÷±`ÛB¿ìÁÖ°žŽ­}ü~ÒÊ™¨³YÍåßïü[•´•ýzÛÆë—RX|N¸{rct`7wÂîÎÙOM%k…<¡l:¾ÆJ®Ê#|Ý/Þ“ ìt­ ëÖñ«TâÓ†ê*ër^,Ù)4¾ úýDk2y‡Ì7ŒŽ%º’´pÝ…2±c¥µ·ìÅ㙉Z{9ypbÀ^[&~Ãô N¿kă}Îmôeìï 'ëøMÛ·¢-ˆËúêà1ÜaÕ~×0ï²zYŠ4geaæö\³çç«Û÷ìÒŠÙjvGJþ¿MØ`}þÖô !«2!?Îþþüßï7¹qKú]Ô±Üè{´ogÕO?­Ø¼.;7-céW±ˆ L>ÑÐx¶ªo¿¬·¤twÕ~ŽÐdh¸ <øÀwd lŒÃ±1„‹íðÆÅ8+Äo0ƒE`Љß1ƒê‹TÁíàIØ õD~ñ72™O°’º‹ YÍAÚGDãžS?b™÷•¡ËdPÍÂ\Æ«‡tü©îˆ³(5¬‚î ¤žCª9­Ä(±VÇfœ¥Ô|FÀCÎ×ÚãÏ À. øj¤O/Xž_ŠÌ Ð×;óws³ÙÒM¥[ËQ%ªÞ[~èØáƒå‡.CÆhÌEq‚("úù­×7žZSRTíEdaéŒõ±ö蓃%Œ~þ>+–Mc ¦-IôAœSäS¦7Ú_µÖ,õ*&22DЧ=ž‘j‰ˆµÐ£ÉÄ<‹SFâ ÍH1&˜ôÀ@’ÄŸwLõÄ<ë„ÍdÖ鸂Imˆ¨ò%X¸`»cßnLƒå£k•ÍWe’õsŸ°xÈ÷ÝüË$§ðˆ0wW§`Fw^¼’ú*ø·€[-Ãu|ûÃú iÝîõ“HÍ X„çà9}–`†-À”‘árçÆ§ø£h”P±¢fÍñoŽï¨çv¶ð{Þݺׅ¸—·íL—•°WI¿RCï;±Â»<ä37ÑÕÒ gj.®E8;³ [Kñ ùæú°¢ÙÞ†8w÷è8Ù Eï”ôyµpXμCÛÃg§Õ83Xƒ-Á/Ç)xvÀùx-ŒÃŽòøÖ¡3·¤ÝÄ8JùEO‚—ä}ùÍbd†¬ÁüÁÿ7òå N°!ö³ÏöŽ ‘ýN^ºµ“d“¨u•$™t¹Tulóz…´"¿8-ç†ò©'¨qüœ„Üy±R8N¤ž=¨‡÷ïwM'‡£[B¸éÚ²NFÒ¯ÉÔŸfÙE”ïD¦éo§[Õ­ j¡„G×¶øêtR×Ô<’ íf3q©ø9[qêœKËQ¸Z–—”±:+oQa(rGчS/¯<ùõ™¢«Äj‹ü÷/®Nº6·+DH…~RÔž­»p¢ÝC}÷m*ñ¼º1®éU¨‰{Ò\÷OàšØIs@Ã*i8@ó‡Æœ‡ÍJíÃðÍ€9IÝ(:‚¶t‰Àž 0m S×û“OÒ{¥îc*9óW€+YɃgwZîžIö—âݺ×;g—è^…±ì›¨úI¾‰y¡±ÒŒ¦Ä ?ÄIœ}QüªÅAÜ“!£^·ÓWúà\Ÿª4)|šŸ¹k]I~#lTêyeÞùл$»Àü]?Œ³Éo°Ä7rUt²¬üš~éGwÑ•¤¾ÉwþeÃ|;çù3g?|Ýwû®JFr—\’41 úñ l$žˆMƒp “v¹‡ß“{B#ã V±ð7®uàÑ–µÓJ —@Ák‹|…"æÏBàýB¤Æ©ü{0½&#p ‚À‰`ÊE2=ت°®±ØÃ“B$ú?Ъ"ˆøÀ[Y}ÐÝAÖ(+å“~pá@¡¸‘BM¡Ø(÷ˆPvÂKÒ0/ ÔÃÛwvŽ ¨ÿ:ôeæendstream endobj 374 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 1964 >> stream xœUiTW~ÕÐU¥´­Ø–ŠhW¹lÑà~<*Jƒ   ⸌"²JÓÍÒf”‡Ñ âÖ¸/¶ƒ Š(h”‰²™̈gÌ™ŒÆ9¸Ë-æuæÌëÆ˜äïü¨sÞVßûî½ßwƒœˆa˜¾A±É™±¦Ä˜hocòZû’Nvgä¡ y˜&)ÝÝf¥¬U!¬rÂ*çCùûýá¶+ìî ¦~È™a&†yF†Gñò;Ǹޜ–Ÿ`’&øŽ÷“Ö˜¥;ÒÜØôÄøɃ2c“ë ±)¦ÐDÚŒt)":%]²ßý›ù¯@ÿ4BhHŠq}Zz )c^fôšù Ö†ÆÆ- OŒH6HÞ…¡…hŠ@‹‘'ŠDKE!4y#€Æ!= DЧ(ù¡(1èr§ùBÎtc7zÉŒe*C‘Š…UñÀiµÓ §g7çΧ”œÒ_y…Ì®d›Ù§\ÎÅÁµ Ø"û€KŠ¥ÿ]`Hƒ4Õwe«`³Òi;§i¿wùbkÃé¤-yBWþË*¸™ðþËRCc´š)9 a%¼%SIäõòLõ úWÁ4˜úøCˆg™!¥Ð^«î0kæÒÖ·o¯·¶‰(zÁSèEÉܦP¹2°Œ²!ÒÏt¥’­«1¶`þÔ0¥-£-¾VÔ,½T숗§$†ñšvèm¾½8Òsy`ÀÜ-¯^_mù–»ÑP‡€Øê õc˜UœfaÇ… û¾,Ø~H #¸Í…ÛñV‚G'ûóß÷4È:(à(˜úÃb¤¼SV;€ý´•ŒÆ<6IOœâ‹“eЇ²äÝØÓ£Wæ–äâ >imVpœñ‹R³6§xkñÖSô4»‹¸ü Sðßð×û«Ê«ÊÖâ:|?¡zn ¸>xmɦ=ø0ºÜr­ùlú²B­ºû À$p…TÀ\S÷ry¨@ª‰+Lcó×lJËH_‘N¹¨'¾Œo{n\ÎI9+–™‹7”Æò0œ5ƒQÿDæÒSqd$G²H*Œ"áM•·Äƒ^ÊY5<1½‡÷ϘÇ=Å<.Ì€ÑD"£õ>­äì¤êè§åG>ßs\ÛÉmܱ¥ óqyÅ5"\ý /‹¼½§†AeËçÊV ð/NS²*Zf(oÔÂK"ÙTQ5ÿݵã­Òj–†pj8ƒ-Ýæ.» ÑýlKk€í¸ ¼¡²UIñ\®§zn~û¢¦å¾Ý.\åGæïöDmø(9\ÉkVÝý '‹M £þ¦’åFÙ•4• ÈJ[#èº $“#líd°Ü®$¦‰þ=ñS¿Ï‘Ç ‰þ#‘6¡~å±@*#a éG&ûž}.JÔ.±6%5¹·à+G+ëyM2é-÷qXQÔœCݰ¼õÅ«šÖûTÏq=üZ¨R¯P~òž::Û–\”·(‰Œ\h¯»© F<ƒ¾]Œ\™Ba Þ‹pýæÚÏÎÇwúÝ! ¥1n,q&óÈ쟆ƒ7ôélé:#-»aõÚ´¥8¯;uú3˶£;®ó;Ÿ {WV}ƒïá󫋦óTìTV~Ïž¿c~´Ë½‰Ê}J{²ä«ã¥¥;¾Ò¾ã ;S 31ïóÇU¾bðLŸ6[HrÐc®Ç”ònjÈK”ª‡Ý”Þo„k g^âNÜû&¤u~ËÔ³#ñH<5i~`°>n¦d s~Ô=ýÝÀޤ·¸ ßÜw»’2­í‘’„]ÂÌiíþé^·ev#ªé¯h0>ß¿¤½Nýý'd\XndÆ*Ñ‘;8L[A8¸Ú«UÀEf±'SÃDWÍ8ëweLàaâ ÖTš·—çúü/}ª:Ê/ÜÁÿÄ·³êÎ%U¬Ü¿ó6wGñå Îvw È?”s0goNÄ yFsΆÙÛVcÞáOjÃ{ôîf{ÂbýÙÞHÙnƒ3Õ†+©»Ç™@EË¥bªa$ÔÃ('ù­¼K(/Çd>-Ø©dQ~½<â묗ÙV#^èç;:üÒÃmZâÁî%¾]“ ”¶û…àôŒnQq?.ö°T–\†ëÜnÖT>|^3{·öcÛèádwƒd÷«þOŸO\ˆD¤jžÕó ÒT6üæ©kn>÷—†«§Ö¯Ö›c¥é\Õ¯s¹?÷.ä‚ô+Ò#¢µ©×#Oaú6,À1Y+#øf{ßRË(¬ #_€Ë‚­¸ØÕk:"[÷BPɺ},ôë ¢ (‹U*KT}úON¤endstream endobj 375 0 obj << /Filter /FlateDecode /Length 6560 >> stream xœÅ=Ùn$GrûL~öcÃûà¢5]›÷1¶Ð.dIk¯¡ƒ€ õÐ"›®Èî»çZ,üíŽÈ£*2«ªÙœa!@S¬Î#"2îÌŒúuÁZ¾`ø_ú÷âî„-®O~=ááí"ýsq·øÃÙÉï¾Wná[o„Yœ]Ä|á$¼2 Ëa«gw'/šÕýé’µ–1¡y³>] ¯[Õìð¥ô^)‰/Y«3Æ5¹±3²¹ ÏÞ9gšÕ-þáã\4ôù}Ñ:ß\q¶t ðl¼7Ò“ÑqVûð‡÷Òm.q|Ól²Á÷Ú{nt|¯á/Ѽ‚GÛæx³¥xÁ3o¥SÚ*@'­—ÍÏƤó'ßÇy—4ÆvÓ7¿ì)iàžq˜ž¾†žˆ„’J%ê¥Ö\JÛÏÊ›—ýèk¾N$u àH+hãUóË)ÐBÀèEsúj7>:…ý‚ Øž.•­k¾Øþtö§á=ŒäßÎ.½¶¯ú¡q!,•–åB¬Épݺ9:ùߪ¶|ù2.7,šï—›æyM¹ŒÎò† sé-­³Íö5鼋cJxOû¾£l‡ï¼m¢h^píû¬Ÿ iµÌÄZrÞz­y¤YfuoEFÒ[ RÂX(K™¾§B`úí ‘*n©§t¢Ù^Åg à™*ÙKJ}óõ§°bŒ7Œ8àP–0‡×I*=cÚé‚ç(·¼‹}·(\0žvNìGùiKHJI·ËPÉ è…’-ËÅÀ‰<"Lá"죀(Hop<Ðt¸JWtϯÇ%±íå×–»*Å«¥ýñÔ‰VpÓ/fóÁ€~6A¤,û>b kCiû0® 8ØØš2Ã䢥ÅFÁ9À Èt áß'.`¾T‚AÁÑ züCœ’+Q4¦¦ë›pW–®QZtÅäoG•@•7W=üÛnÍ EÈÃÒÀ ß0æ•-ÀC -=&Ù\G@`ʆ(´BË&"H^ [±ð(C=ïçDÎùÝ÷ÎPã­`ŽY,0w‘þ; *‚ÛØÅ‹`c¥Ã.®ZI`=h¢à—^«€DÅc—ÂE0­ èð9bäÀt«†ŸŽú¨“5ý‹b|ÑZc½uzö—z8Tï ¨ópO¡W¤ P‘Íãÿ&q»å"6Ö†4rûÆ<Œ ÒÆ]"DI;ã„ã%õ~sø$…Kä`“ã1Ô5ÿOÔuÞ=‚º^çÆ¢¦®zºŸ™‚º/¢ÂžDœS&ì㤌‹hµ€>¾UÂÆ.Ap´ÚbÊ[›ðtGPë%”Ç£ =O(põHψÀ³c>Ú—)–äþ–Læáþ©7Q#(ªÖ€£þ†Ç’VªãüSͬ+©÷›ƒäp­“üM P\‡Qï£Qqu0—`)D‰z¶“Ì4_oSN/%žÀF–©QhÇqÅ[ÊžA'a˜ S²8¶³ +ZPL¥ØÔ€ ¥¤ì½všÄ"bDµÞÔöÁjjµ±qàØim( ò+™Ÿj×bã”. *g2êQƒtr!y…#¦MéÚ*Õµ˜ÈËSeAÔrEMSγʱdcò}Ì \ÔŒ@uSö…¦ú±Ë„ê’Uê,ġÑ5^¨1š95EÐ’aOÂò–9UñTØaÙR°šÚÞŠ‹Ê™=¸¨ãÛ‘Ää¶DÎj]Йú|»r2^økGìCÐæƒB§˜Oa¶'õ|ÊxÐãÆ†Å%Ï‚ïRReE”¦µ—Ù˜î&ÑÙË9¸KH5®²˜<w6Rœ1=Ð ØD±ÂGÚÆ·`¥š¿’ƃ=Ñä­â0¯oþ=æúyÄC ˆÙùäåÉ /wIÞ‘&#OõŽ%íú GÄ]ØC çc{¿÷cJ= wió®ŽnžJ·ÅFÛ¹Êk9„×UÅI¸2h HÔ:bË`QAññŠœâA NWë ÉÚ#¨RAàpUfP!ææ öÖópá°Eåq~I ©‡QÉÎ͇¡ÂGS½³8jzn\º4ÑLëBXÌ<ìtÖ,F5Á((…ùqUzƒUp¶øÜ/)oMS•±¢Û›Óê0îcÊ¡’¬D¥àÃŽ¹/uýQî Ùµ„nÁ"¨ {ËU©|CöXÑ×aÊ– œ/hRM<#˜ ¤sØa÷O·9i¥ëôY~}]ªÛÒmq„«ˆ¹a²Lgá<ÖÛCÖ ùÀçÞàí¼vA*Ó*Ýñ-ÙĽÀ?àˆNlÉÇyHT¨—KÀMÚÎ\Å¥ néÐsŒÏwdtÚ¾%mv50ƒ®c§Â¾Xµ]O–|yvòÝ jf;ªï¯O@s´ Ôxý­Z( lªÌB*{Ì‹ûõÉô–J;x¸Ÿ>w†Y`‡§ÌÔÈÁ3í#ñà¬ÀÙ_2˜ØEK_K&,>€~ðæñy4ñ éLÿ¶,¦ÒZá LƒvaHmœmôè¹ h5“Å\ ïd2+ã6äÇã• h<è¦Ú­÷ín½¾$ó=Ý‚qÁ[#t1!( yè( Äü¢˜K*=^`‚P9xâTK jÃ9K­SóNÏÃ¥ÂQœ£q—W¹ôéìÆ ÅT«û›»U»»¹›‡ÄZµt ±`Ðí ÷~ZclØNÎo>`*XEÐÐtªÍ8 “ FðÔ'ÌvµÇƒáL€Pëb®óæf´87ád9ëf3æ#SñÕ\‚D»H}>ø9~$SÍ£i¹ p?S̨jpÎðê ’êy:eÁÃí¬ǹ”EuÈTó)‹.)d è8O×yÑdºÛÕõ¬^t…ÙÇ{Ò`»¸,ya9£s[*×›ÎX"£8÷im%,Ÿ4î“ØJ2ÕfÕnïnf ö³ö"ó8»VóP9'j9î{ø!•ŸÚ”šb*Àùų™LDNg’éø¬Î™è§'³¥dÐÿ—?Ѳ„ /¦3å “ý|Òu±áhJ1ègó€Ÿmò§Y—h”‹Ùæ ǹÔn¸,K.è1Q¨½ûõn^ëÂ08ü$Æ¥Ÿéún¦”Þ-ê§:ÃñlvqtJõÑ2–·s®0"H³¥3Ùy…¹Ÿè§"8~ò-JåE«@φ-Ê×ww«û÷ó`–‚V:ß9Šmòe¤Yæ&\ÑôÀ›(Ô¿ý-þ”¥ü?OÎþ律!•àÍW··ÏCPUÿ’sŽa´PÀv¯Jæ8 ø<½w¡8D~ÿyª½xÖwVÏ-.΃“üsú=çŸ×û—ÛËç§ –40höo·?ìׯBÇ€û€Eì=Õ¬ï7ë[2Îw¯W—÷«ýÍE†[4?¼Z_ìïW·çÍÛ›ýË yóóÛ~˜ÏûÄJèÜL46ªn×WW77ëÍ~WR×J屨Ǘ»ýÍÝjNË Ýü°¿l3µxóåý=žŸäÚöyVݼYݾƓ GñÍ·÷çÍïÿ¶ÿ[¦ã¼ùf³_ß_¬_¡»pi–à@IÅ×KÆû<2Ž„w[űCCÏן1Ö7sV…ÎcV«.~BP‘€¼á­PÎö£á Íua+ZàI­õRÈc&“?%Z.5F5†ÎrQaˆéïj ¥’Úëû§åÙz·ž×…7—ëëû5ØÏîž”έ¯à§K,ò‘yçf×ÿ(zÞÆ3Ã=/;a˜heèDóí2r@×g”ýÏBŸŽ›@ŒÏO?g®`>`ÎZa`Äž½Gkwåo67û<ó•1 €dD˜*0Þ¿\÷ â‚Ã1l*8 00þT¾÷¸îtи¸®‰Gʰ5á7P¼NÀ+¡ÒXmIZ鱦<žgb!À#ÐÑÉz1vì-šCÙu*œ±Û¾"ïi}’‘û‘XQi}w¤ ³é9¿wäOF{ÓZÐo‡`ÖgÛñ¨N«‹ S%k6x•Ì|,0ƒÐzñQÕ£ðø$øºè¸‹cã¹õúº^~OjZ ïi“x3OË;<•'^Uí0Ž«h'OŽOU9ê…E‰‡`é­…(0£^Ö]ÍZ`‹rEð\þ‘Kµ&aÝ% mMÉeïb_¬X–*:¦ºS¨á aY ‹‹bÕ]º¥Tºõʤ@x>¬¾Ï´Ã“¡2˜J×Çs™§î¼éß±³ÌÐp`Éâ%¾Cé8êôêk8 Žmd<ÙÿSmÁµÀ!?T&VxjŽq>R¢,uDqŽ—m—¦ Œ#JÞÛ ÃgB[³R¢E):î*Fèé‹㫉; aOø‹`^eÖä{v< !R#w|Gdi…“L†g!˜¼\”K®YW–«.g„EÎÙú+àùuA(†W…Œ7îXTì–ô›Ð^S· )-_O Aû¶åA×e¦DYgë« ÕŸY±” T!VÀ•wÔ¶Ô²¬c X @Kí¯É•â®îàpp‚&Ýtá6=Žîxyÿ©ªhø¸ò‚ÔBt0¢ju<ÚÃW¹z°ôжXŠ‚… ‚IÙ“Võ;T^0í§å«‚Œ´²êد¤9e'*ÔoNµÆ"l²,¼8¬FÆ*‰ o ëýt§­1<1ر»N‹"ýEÚ:Uz‘嘓"Ö-ÎkBeô>Á溇o®.ˆ­îF!kã­ýNUY+®úeÒDT©R6\hÐ+–Ü¿ìʆý>¸‚t±ÃÑ;–A\FÕ ©fš´Ä”ÓY_(U ipZ´üjn`sËõLF-’¬XïÕi*Å÷>M,:¡9þ®ï5e—ñîJàswèîJy\ ZƘ ÏrsM íÇêÉ@8cz',ý*9\ «9Õ åX5«ëÄXsª¾É š¤ÇzƒG²y… ?-h)‘4t¸åÁ¼cB”Õ1Hų¤’‘ƒ€}Xš¡¼ýšÏ(7fÿ°ÜmE BlG3‹éÒ£[¬ BÍÃl”˜6¥U¤7~¹½¼xéù¥GvDØÖqz•¥ç Š]œBŸÉÉÈPqU{¸;b žÚ(†IEϪˆ5dHW5¡7|é’Ò®«‰+´U,!¬nUáK,%ó?”E¶e¤VæT›]^Ü¢äPÒ¤$œ^¢C 1(o›†$v¨ª©ÌÊÈ JuÈÑL V¡«o2¸f@ùŠÏ øyB2ºñdQuz¨$†…(þ“u|É¥Õ-åCåƒã2ZÄK¹nÊÛöEã~YSªo"‹†b“ëÛ~ª[!n9v—’‡¥Úû‚¸˜<á¥>D‰ò>( r#¢€Q4YÚ;I•:¹pZ—úÆ~‰uÇbÝ1 S£U6a3ÑÛŒ3'w;¹ÈsW=§³)s´èAU³×ZÌ? `‡q,ÄD½‚‰ú&îÄææ˜ÙU,vûCX“èãQÀÉ T=QHšÃÁÁB#ômg%бÿJY§¢ ®ík.P¢P ˜øP®pÅ~ƒhĸìPŒ ÃB„•ä}Mï‹×ÅÐ*bí:®ð|@É›ïˆ8I±KòÇ0c1t;.â\Ò•‰Ð\Š\–±儉tÒª²1Õ}‹‡å¹0G£œ"• &+kækúË¢O(áñ§Œ\þûH¢*€ ŒJªl¯èJS¥Lh…a.‹V÷PŠq|GÖ“ˆÒŠ€2u¿¼J% x³ž$'Ó’Tr7¨c2¾"!–wù(‚@aÑÊÊw‰Ø‹£B˜Ž…mÈÁä®tÄ}úl¤Ømœ>a™ˆ%3å>FGkLý Mì5ø0CU¬Qî­,E5+øhéµ5 >h£úbQâ²B€‚_LäȦª‚U¶ ‡A AÓÛ`VC‘Ñ0èÛã éakºÉê}#:áÐwöyþÊìJÍq·n˜§ìwcJƒ `9<ÝdPË‘¿NÜ:—qíÍ$ÿ(îÓ:²Rõ¤F3ëv†jö}0•Ò…S ¿ót «1©æ›À˱øB"пØÚî†ô÷ÌN•nÒ9‡*½va>YA$‘-„)‹ 92ßbägQ¹£¨LÆß°@xì'kâãS'#ö¤Þ#“4ÙãˆOçÔJr$ï6ÂÞÆ1¬Ð_w:r˜à®ã`.‚΢_„¸¯3§Ä9MS›òéŒ%&ËìÁø(ÉÛ ŸúÉÕ€iMHóÈ´9.Îdeå^T 1÷¥&!tJ%ð@wÛ¾tIeuc[SqDšÌpH- gVù¸ƒÈe°RaëÇ´uU N»NРN.dϳD«tYÂ^ºê$À‡{Sªê•åëµ üˆ$Øêp8VD¨þˆ–Ã\®®êaâÞ—3Eµ¤j_«,œÞSÞì“›ø " žé©œ³ð,V œ¶q PÍŘ&kl€{KǦf‹¨ïþ;îË”3Úé¯h¨V÷5ä ®·SÙ• ï"ÿÒƒ"u9eŽ[NV¯Étd6Ž]eþlŽ…ÏH%gj×iêƒ^«‰,WÏ…“ç"¨xm"Œæ€À3paÌëþ4ø˜nñö„-¾:Ña_aá­k…]Üm@™üâöä‡ÉsKxœU[œ[2¾åX®ÜóþËx2z ‰Í #çPUËÃ7 – `ݾ>1ÔBH‡ðx:F´<å€`"‹¡Ð6ž EÃ’»c1ùÂc¿û@(p:ÿ†A€ôáÇUª1ËQzC cr Ãé1ù’¬a?J÷¼‰­Qh_…é¶Iƒ&øFâ1EÊ_v¯)èëÈÎÆ¯ılðЃsîbâþGÒ !ǃ;ùkN¸Œ¶P>n5¸䀯Ó8ÓøþÑ…z°BÙX/0¾)ÞØ/!’ø!ƒøå‚ø–‚ýIg¹ÅÛ÷ 4?«~ä Òè2þ„9¯O]¸s¯Š¾åÚF«‹gÄ»m¬cëm$½Óà„hġˢÇÁR¸7"ÁáE*§fs·‘âZë$Ñ#m61û'LdË8W8-'yÒÓ‘÷–¼@ƒíÍ—Ù™«Î[³Zªy<‚¼î”ŠŠú¿˜‘ô+ ÁT÷mŠ{°[Ú_t¤Øe>Ĭq“ùŽòaPhà›j݉ @¶ºAìté$ÚzòØÉÊÛÞ'Õ6 :ž …‚Eø0‚OüǤ.„ìYlª¹É@g‘äñÃ27•îË8ã…^ÙÆ·¾ùºçÂp`N8Ìái&Š¡Îå¥HƒÑ¥¸ëšÂ> stream xœ´½ËŽ5½Î6ï«è¡ *ªjI€À@f±7à‰g >ÄH'ñï¾ýˆÏRµG`ll¼ßÃE‘ÔêZ*‰âá??ç÷ÿÓ¿Ÿß¯ãç÷Õ¿ÿþûOÿáë>žŸëú¾Žóùéã»õãøiãûlçÏsÕïúw_ÿ~¿ê¨åžãj¯G¹â?žþ<#$üç¯ê¾õÏç÷ûùÇ×ÿô¯¦ˆï{*íûÿþ‹¶LJ ÒÔ7~žz}ÿã÷ëŸý?ÿçÿýÿþóü§¯ÿõ_ÿûÔU[{®óû¿ýÿn]çQ~®§¿”ýÛöOÿî¿ü‹]]­ÇSãzÆó”ÿ~ue̱å¥î¿þÇÿãóÿÇL®=?÷´w×ö?ošÆÝ{‹¿ÔÑïç9âÞÇ]ûóÿCUo?ýzÿÑÊÿ˜Iõç§Üíýûç©j »Ã–ñÝÏ»Ìgð»ô~ý<ñmõ'×ïóý}µŸQúÝ¿ÿëü¿Íÿÿ§¯ãû_~•+x®ï>Úýsÿn„R¿Ëu?÷3ñèñ#(W™__Iü™#Ê4ðÜ8ê˜ßô&Á*b€ õùç÷PËÏÙ7Ä1ÂF˜ÃFZÂkŸ¯ýUžùu•9äšCJ™%eþŽ[û.ÏYæ~â;¾®ò”òó,<Õ>sìqm¡n H à'¡ÔŸ³~¯ágŸä%0Øm€>·}þžA̪–^ŽöÝŸùÝÞלբŒëçßµŒùPO¥Ï5Šx.f ?sÀuã»L†{þDû&À˜*b„)ók¸Ú÷’0¿‡²© Œ¶Á ¶ÑÞ³ÀÌZ=âO9Ž6Ÿš3KÊ䚿ՉcÄOoâÉ9×]ã©·õ?€Å1jÈZŒ©#F˜2eÏ¥}I¨gÈ^:ˆc„­0‡­´„÷<0·Úø%ŒãêeŽªý¦Þrý̬ÖùÇèwâ©·^ãç:7Ž»ÅÈ%Á˜:b„)ÇO¯%ÌwTû©wê0Ž´bqØJKxÏs»ç÷оGió7?³$ôþ3W­ o|ÃeÄb4mºãO!8uÞóoSÚú|þé®m´ ¥»õÂÛ£çò<›xâ`ýæyð23ºZÿ©cÒŽª¿VR΂ïúêåç™g<øó{¿óÓpê¼®óçlëóù{ûpc*ÀRΧ⋶€óÆã˜„c„LH™˜Þ“ÀÄæûêgÌ//^Ê KJíøÁôùÈœó1<Û5¥ö)­–ÄSï|Þ~ú¹8ÆÜ+=÷’˜:0B”óŠelIˆ]Ö®ƒxŽH+Ìa+%áÏ<0·ó¼Ãž~_O|“¿åž«Õu–¹ÊηÊýŒÐWÏÚ°ú O½1§§m=ž%@‚ß„¹ÚÍg5‡Ï'yþ±–â`Ì‘FKÂ{˜Ø1×Ò6¥wì¤â ½(ñ•Ox‡°ù¥M¡óé:®+žã©öˆg¤nó¿ž%Àb€)ó ?ç:âR6Ä1ÂF˜Ã6ZÂ{|—ÍU¯•iÆüYün”Ö~Ú4´ÌïãŽj~Ñó…SKY ^3ó7|>ÛÇsØüÚÖpc*ˆ¦L3ÎùH SÑ\ ¥€ Ø­ßÛ>Ï€û޳Í=hlæÛŽ$÷‹õÞ†Sö¢Ì ã[ ˜kyßâúÜöyø{ÜvÌeíümpÝHJ¬°sT›ðüùx„cUhóŠóé®'Ÿ£ùNøÝ(óÕûÙùs¸ã9¹p€(ÏÜ—=WâxNæJuŽc®T±‚Y€!5ÄSâos/±ÜUãá•æ°–ðžþjcœ1ëqì 7ÊUâ';ñ»Ùù~œ Ïü¾Æ|ÃâBï¥ûÂ7–ìXŒ©#F˜ÂóÖ’ÐâËX*ƒß6ès›èáïIpb÷ŸçÅãØ"L]±}‰“N<ç3[ó¿N¼ßcg0Úñ,%Çü1Œ]‚ñųÒ"LsGlW,`à°µT Ç–âH#%á= >‹s²O,8ñªãq,)OLâ'ìí×´³àÙ¢kOÊü¯ëÚ8,»KBb舦L]-ŽT–OᦻlÐç6ÑÃß“ÀÄÆsÄ»GÑ+Þb‹§Æ9j<NCƒg‹ùk$Æñ¶óxkŽyl›„ÄÔ3·)À–0éót²t¾–âXvSŸy`n½síq Š3Ù¢t Ë<ôÇvÂ'¹¹@wœ „§ÞØÌÅwœOÁÙ/%SGŒ0%NZsö)¡âT¹tÇ[a[i ïy`n×Í—AüHÛ€çÔx*â¯póm2æîç,?wìmŒ§ÞûÀ¢œ÷É÷•%$¦Ž‘”#NKB¼*ÎMqü¥m…9ÒnJø3ž7ÆãÝå%HÊüAÄ!¿žð)Ì“ÝÜŒÇ鱜|}ÇYßåâˆï²mŒ©#D‰sfÿ¶„?ªðH‡qŒ ‹ÃVZÂ{z©á…‡×AyøRe>­Ä[­„…óWZÃßRb{›1üÉ[‹{“ ,1”{1çßÀÆ\AîM‡pŒÉa»-á=½Ú.úéæ˜óæ«M”¹rÕØM_øîûõ̃ ^C<) ãEŸŸ)õ8x) !5ÄyK„ùÀbûîás¡ë»bìËd‚9l¢%¼'ÁWÛ<ÀO;•aõ_„ï©1w ñõ#±2æá!ÂñÒ™ß9Â)¶ H0€„yäŠU-̳j¼qS…pŒ  É SÀk|/žné\±ò/Ê!÷}rËÉÍv¹ç»±µÄñœáÄ®xQ8ôΩå>u" È_*a â§Uóórò§ïá ôH/JxãˆesMé»âðØsØDIø3 þÂæšYñ¸Ãÿø»QæNìÀ±¢óñžKSÇέÃ!¬ÀÙ6Ž«Æ’`LØnŠ2×ö«}/ óÀRîMqŒ°æ°•–ðž÷!åàŠ[NÞÀ˜«W¼á <ÉsMâ›äªðfÇa®c±ÍMŽÏÌ H Á/BÃ! ůn ˆ±Ý” æ°‰–ðž„W76&÷|–ÚCÇ•( n°z̵&^÷5¿ØzÜ|Ý ËÛ#sœ·–˜:âñ2eZ,WÌß彫 6Â6RÞ³à2·b±ÝŒcÐ…1)ó¥{âEóä‚2â×;pd?ï Û€Åqú$SRbA¸±PBl>ëX:„ñ¤É!+SÂ{ô‡ÌiÏó:®/Æ ˆ)sÃ~ÁWÑè»MW$5~ÐÆ¸†©ÛÇáù»¶áÆT€;Qæþ±…ûÃëŸMJ'†·EúÍaû<ü=].5øÏùCì¼[aêqkÓÎZ}ãDÛú‡¼0&õÈ $Ž·Ä’` ˜–s#XÃcc^„¥‚^!ai ¯YpeŒ_ò¯ñß2ŸëÒxfÅ»lž.¹lZMŒ•Qï2sÜzSY‚1uè Êü›>±_¶„6Ï › ÀÏײAŸÛDO‚G˜¹²<8 ÜØ{,‰%¨Öùðâ©) Þù:uŒ„¸0;âé_ ó÷RvÆÐ€7O#óuྠ&>ã. †¸.ƒ ‹&.¯9è&ðÑ Nã¿¥ŒpMàV®Ê5ˆ{½ oXcÜëuÞšãÁþ}I0¦ŽaÊ4$\B)anŠž²é Ž¶Â¶ÒÞóà‚êº{ža.ø¢Üsnbw„Õ·Ç{ª…§]áXë¥õZçå” ,1”û³Õâi®›á!+’Ãv[Â{Úzâ-ùIáz01–üjF¼û¹ÞãWÐå=ÇÜÂÕ]‚1uà—)J8þë÷’ Âµé Æ©SV˜ÃVZÂ{Üd¼Fžß;œð|Ä~&¶%ÞØ>`ŸOûƒQtÇx¸¯·cªÀQ.¥—„ÊÃ]ê¨>þÙ3dT‰¼gÁ'rþôç+OOà SRN|Oói)ñÒ˜oû^±ù4øþ„ãi™Ët|ÇÉ1PåÚ$SFrÍe vG»øyZ…0z‘2ÒÞ³àßlž?Cè˜{Ò ®ý¤Üp]Þ8ïÌWýÉW?\ÆðDÝüÂÍQ7}oH1À”ù7íq4¶€Ž+÷¥‚î1Ú †´PþÌó: w‘˱UjòtݱË-¸47F`ÑÅ 9bm87 ÆT€‡J”Æ{°”0Ÿ¦^7ÄðÕÊ sÀDßÌç¥Ùuò2.îî%îPãq¾;cöÌi¬xsŸ´pœ±ä¹5Ç|¢~Î%Àbí¡ß¸eó莓ó’OŒ€ Y`[(ï)ð¤I§q¿o¸™7ÊÓã71Âdp̋ە“4Ƹ= 61Cq–cꈦ̙ǖËâú|SAü²ÁŸÛDO‚!§ƒ ·Iˆ«Þù"ì…»Š¸ª]E/Žn¸µïè±@G„–ÔM‚1TÄâ¬Û 8nœLR…pŒÉ!#SÂkü‹Åæ&æóñÅÄ’0Ÿðø«Ÿ±'Ð]6¢â²ëN ç@Óu*9âö›·­”*>_á€Y) \Ê» âø‹Ùs¤Õ”ðžï(æú¿º~ÝŒ[”›ëÎüáà6|>и/„Çtâ‚xqÌ·mS¼N¢ Ü–, s¥)»bÜjÈ q¤•’ðg˜[úãëhýŠŸÈïF™[帵©ÏÀ3ß.|µ¥Íß~K8Õ¶³ÁA“ ¥`·€ÄT#Li7È)!BÆRA8¤ bH%àÏ,xÏY°æˆ©>xkÊ|]ÆN/B= nK¯ ×19ë•8®W{„ïnñž)›cêÀQê‰^Jˆ¿æ¦0ømƒ>·‰þžƒÇÂOëÁüsãO–”rpÃ0÷1 ÖãÏQc›sÖÄq·者qÄó±–Œeî_‹‚Åðë¿Ïø{¤áÑÚö±-öð÷ té^œØ*£ZL‰—ù`,lÑ%ËQèêžæ²ÿ"‚=¹AÇèÜf[‚1uhÃÌðÛ‹QKˆ+ÊsÓA¬\†ÌŠÃVZÂ{<’ÕÊ5åAðãoRæa’ówÄ<;a¤56ÓwâqǵqŒÁ“¼%[X¢Ä·~}[B§c¸I¤Ã8FЊÅa»-á=Îí‚7‰¾LÍ„8àtùd´*ްë–GÇ‹=1à×ß—€ÄÐN ʉ_e ˜ËHÙ4âØ-Ì`-à5ºsÆ`TîÜš œ¢“r ¹ ÷xhÚ’3Ïzˆó%;,s‹cn¶û.Á˜:0B”8ª„{ÆzgHŽuÇ[a[i ïypeœïî¿Ö¹ya@mRæŸàŠu­ ~®s=ŽqÉ„XµNŸŸÇinL1À”‚+ä% îdS<¯lS9>³eø¶;ªÜµ¸ïÿÝ(Ãpo¢ðqõ¸¢±Ót¡ð³8]î):ØM‰-ð4ÈÃù^K8ákëÅÏm™‡¿mçßg¾H±á‹³8ž½¤\5æs¿€HLj}fœ°FKŒø™#vÉqxS¦„ÄԢ̅+\ƒ)a.ñô¦b„7É sØJIø3žVŽ.åH\è<­˜Òøc,ÓFö?ð=•“ûáØôÍc"“ÄQ¹SJ ÆÔ¨=Q ¼K oÇÒAŒ¨¬0‡­´„÷<è´]ÈE‡{ipÚ›Rp®.ö÷àWí ³Œ•¬&þ0”³÷#¶J÷’˜:gIÊÁ[Kˆ 7‹²BËnJø39KµÍyßÞRQ©×IÇ"{ÿt¦\Fr÷Ýu}6ûhcÊÇ¢ÝùgøòmˆÔ(–/„wÔçÇ´n~Û¯€nøÕçiíÔ³˜”Ïþ+À;ž´³\qëfŒà@”Tr8BÛŒ©cøõ, ½áIKÄŸ-ÌܶÒÞóàÅzø.^‹303)Œ,.p(„od¾Ôq"éUWtÕg–ØåcpÃŒ©#D9àãXÂM36ĸ—æ°•–ðž‡Â„/zÑÃ9Þ&,Ê@l1‚°/ÅLzNCDýÖŸs}|a¾FS>ü¢Ì¿]l <~þH#{ å+ üº7hï)ðq ´] ¿æ¦ù<»žMøjœû¿œ±F‹#’Èv ‰í½0!.‹ãîÞæ|ð`Y1EaÙhûð¶ÛâÞÐÿn”øåϧ¨GVÀ;6,% ±4 DZ¶`7¿8*B#—ãCQûI‰íSœƒ% ÝÕLÂ1€6˜Á&jüŸ9ð÷ut.‡7Iix·LL/ؘ¿Úv2FÑÄáçË(Mr\±=}–„ÄÔÏ (o§-óä/Ò:ˆã÷e+Ìa+%áÏ<¸Ö‡[I×P'EQpWSbµfޏŠ43ñ‡wiׯ0Ÿós@lºéåF¬ …°Ô`hÃâ°Õðž…r;„ΓOÎ&Œ¹›äϺX _AÄÈS”¿ÉØEh4þ›¢Á xà§—ãæk(~š)™XKt›Ã¦JÀËr:²ÔÝÇÜž /5%.uÂW<_Èz½v é%ÞC?Íñ CsI0¦ŽaJ8ê§·]qŒ°æ°•–ðž¯U\Îaaá#¸(¸±ˆˆÂKÍЛ±ˆÆÈ„:yïbŽùÄcq“€„‡žÀ¤t-f0×ìxÀR1"Ad„8ÒFIø3 ¥$Õ ˆbê_R·æ¥ ç+^ ¯ŽƒñOÄÊA‰ûÆäh~¹H€1Uàu$Jô«Z©¨+ë Æ›F˜Á6ZÀ{<†E^ˆŽ2ÌsIJ©ôSG+aˆ6è/}cÆ)Ìqᵿ$SÇçkQæ67‚ˆ-áˆ?€¼¶°~~fã<ðm>3Oç ‘»Yð&e>¼q¥WV±°•H8º"Ù“ûXáÈT±kãèe@éÈT-Ì=hª5"–hâ`·rsØ8Kx›¯XÅ3ùT¼Á7Ê\yp=4ß«X‰;î4J9-×µkmí^ŸG _߆Wo’2·qNLsO‡Í‘5Ë­Ìa-á= ÅÀGÙïù.GÌ墜<̈·êwÁ¾¡±ñ^XýHž4G¯Üx[€1Uà"–”ëáOÏ"äZ*ƒŸ&øc˜£ßSà*ßšR;¯ò¦ „ú•ÈÓ=PŒ›p¤ÄcÖšlŽëâ6ÝŒ©#L9â÷½$´¢5\:ˆð'+Ì‘vKÂ{|ŸŠíJ$É4–0e~AñNŸ;®ñ1=~¸ó~àl$Œø~¼Gˆ_°Cjж”H¿©ÌÝeSAŒËN!Ž´QþÌ‚[¨ù¾oˆ¼Â1¥Å‹>Jzòª"lnÞÈÃËɼssÄ%Éu¥€„Ô€ü=Q rÆ–€9CØX1Â|d„9l£$ü™ýô!+È_b˜}Ræ7à§ÌpºpIRⵄTQbøé æE‘â’`LðÓ‹2Ÿ¢ ?½$L]÷®ƒ~zYa[i ïyðV|Üoz“ò ç±Fú@h1Þ$óÕ` Œ;çÊ碶K0¦Ü9‹ˆ@•„;eéÆAV$‡­´„÷<8·"D.úÃXRžãAØÉ9í‹.oF±V âꔫY2D!€± – Db›"oˆ$Ü7¢Å¬BßmH†´ZÞ³àõlQ2×ÙùF[”‚p†Ò‘¬Û{”vÁ¯Eã8ÿ5|‘‹£ã‹\ŒO¥@.ÊDþ”О¸{H„8ÄÂl=ú=.Ž±ÕªtœÌ]7…AµðjÖÁCÁ…lèª30Ò¡ "–’#6¢m“`LŸ¯E©Øà- WvK±|«ul¶ÒÞóàíÊ<¢öÃN7JGᛉo¦'÷›!¸Q˜¡&Žå® äxqL}÷.Á˜:Ö+Êü÷D¯$DÍ„]1Ö`Ya[i ïy(3α4S:ãDxàˆoœ{ööàÕâ]jüáí`íG;xSg ÆP¡‹;°CxEž*ŒñÞ„‹ÃFfÈÏ> Õö8pKÓŸ˜„Ф… îæ®šj§ëÁ8ößKΘc Œ(RA\0ˆp"Þn ?°mY ˆq æ…ðš'Õ!†]ÿÅY™RXæäuH¹xÀ@æ11´Þ´Ë÷`6³%SGŒ0%Î(?B ó`·‚uã`”'rØJKxÏC×΃¥'žGu=DAŒmx:kœyà…¼« Жnû¨H·qÄñj—@laÊÁËVJ˜ORAõ"é0ÆÓ[ù"2GÚ- ïyð®åiÊŽbªŸa,à(lEoRÜÒŸ8óÖ~¨<Ë…$G䟴M‚1=„ O$ ¬ásŒS1²±d‚8ÒDIØg@—b¼²{Ü3»éלþ‰|Èø7vLƒ ’q8ú¢̳qpIZ„¥®wSdÜXÂ<þ Ú:„ÄÔÆÆA‹5|3ó)á Šw^æ°$%<)ó+(Q/ ü®ÊP‰@Â3aÄý¨>™ªÊ™Y@âLa1%âÑÇj U *RQ3ƒÅ6$ƒl´€?³àÃ7Xãåb¨úïFa°úZ˜8ZbÛ²ýú’ûŽØøã`ØGS>øIóû-ýÛããNöz¾·L¾“'è4 9d` xOA¡ï¿x{0”C„†˜ëy(ª2ò𽩤wb¼0Oå“#^ ½- ‰«js%a¾¡NZ$à|/YÅéò\i„9l¤$¼g¡Ôàñƒ¹ºÜLˆLÇÈ^»ãTùeŒø²xKcüaöUëG$'œ›a©Pfs¥ òR-a°´@êFb0­HYm¯IðV­Óã‡z9‹òà±”¨=¿°HÍž°!!H0žþÆü-ÞY%‡ßÊQZ” »®Pãîq) ~àÏmŸ‡¿gÀý|TPСЗùïFixXK¼ìNxè(šØâ^c»}ñW›—’~-Á˜:0”ëJ˜[ó±ë Æ@V˜#í–„÷®(îM0|—nâ°‰ÿžŸÆ›WQ÷—I¸y}µDb} ÓÃäp â­É$¯d8™f‰¡DÌ;Ik·5ªœ LCš(ï9¨î ªÌ„â9¥#"£@´FÀÆ|þolxõƒ '„?¿á€[Û.¥"$c (±±Ft,ô¹íóð÷ xÛ|!"h³+KÍšÐã‹nq/‡lÔZž+W×Mó¡Ò)ý`ºµ9ú‰ :Ž6 ìà6áFþŽnq„êK¾pÄ=Ó€d°µð¶^ÎJŽ &8›€|Hü˜ð|Ñ…|£V`<ò¼Kfxn¦¿Y@âƒ:I(L&H'¯`¬P?B˜`›h¯9з1¦nŒ+fþ»Q®[’r :Ú¢<rJºê´4ç”Ìýs¤j˜£…É’˜:àÈ%žØ¸Ä²„xd˦ƒq²Â¶ÒÞó £­9MåpTQRÅÝJR¸3õYFb¸Áœ¸"Ž‹&)!ñá¨"SÂe‹Ô'®TÆ I‡0‚?”vn[i ïyðq¼á÷ÀÃÄÈѤDž ®<]y6&ÇÃ`ÁƒZàäçç)‡ †'¦Dœˆ¹×ã{ ˆëÁOO*Çg¶LÿØ®üÅĹm#/JäÞE’|ì&·}|c¶äÚÇW•=Äçáô=ú87ò¦Ä><<°vò-³è¼“ï?;ƒ Lï)¨`}ÛYq; ®—}¨*ëi3öUq»au]·ËOÝ®zÛÄYn[£]LÛòW¹m[ Ž´PÞSà÷bñúá>â¡„¶H@­Q¶÷ùÄÈ‚<˜­eŽ—DK¡|¾’\røuñVÒ „c€LH™l¯0Ö·ÝŒ² -f•”Êl5*¶Åyæ¼X†‘’øÃèàÞ7Ž»0žÚŒ©C‘»¤à7¾$¸ä®u¬¢¼¶Âi·$¼ç¡U^÷Á2¿%ži¬òŒÄ²çgž¸ž¦ohî5<’Ï£ú]’`<2DÇ”Ž[ç%!*”M1VùšñÄà°•–ðž“pŽÊ°ƒ QF¿…]óÀs2R=2¾JÔ]S=)*m§„ØŽM‡°R~îgã°•–ðžçVp2y†¶õI™ËÓ…ÔžŠ«‘¸ö)(ÃÕ©±âÏîsã˜/ËÚ6 ÆÔ¤DöPSj݌ذá!+’CV¦„÷<y¾6í<Ž%%^+®ûfú>ªä?‰;9+iG’k¼!ì±ê‡`K hÈŸ]*ˆ?+Þ ¶ÐÞsPbØ{ÔˆÜ|M)|Qô(år2½,¶Ñ»Ç_ÿ`À¢?•M<ܘ ðh‰ràÏ"ô±üµ¨??¶}þžÖF½WæšÂ3‹±°J×2ê^2Š®D°>ê„c¡+¬%jŽñ¨î¥$Sˆk5Në½ÛUq×¼=6"9ò}) ¯YÐ{s³v]{†îÀL ×Êè»Ë†eç–ÃN”pI†rÒ+ÆRÀ QnÖ“€ðÅÜ×îÔyt ¶Ü:äH«%àÏ,”MZ°/‹º¥,&ž” »‚Ú.IF”BÀPø‚Jâ¸)¹‘,½8T¬]Œ©#F$EõÃ,!²µ7€Áoôyíòc¯IПýà2sipú»S¯‰jyñ»Œµ§4øÖŒ? Ócãˆêje“øPÈIRn\º¦„ÈûŽx(뎲B”e%%üg'z€ôH¡@ðï¢<(30·txV®¸<Ñ˾.ŒAQ…Yq4•$²cªÀQæçÀ6ÇeN¤°[‡06¦ªœ$ÛhïYèâP¡`TûÝ(½1s¾Ü¼QˆRKZX£ÔàLÍ1Ÿ˜{nÜTxnQjWb½$”›!€ÖAŒ÷{ûØöyø{ZÄÞŽ§sµOJ|ýQ½=\?L•dŒð:•+1Ö9TžH† ùÖK€°TèžõQÅndxRÀÐý–5ãL’ÃV[À{ô|ôBŸ×Ü70²Ù„ð…?ê}‚·'#&â' ß]sÕX%DhŽ›†¥c¨ø|-‰Â%àà¡8U¾–æ°‘–ðš·ÂYš±ÉÄs˜„Š:…%¢…]³+J¨FÄòu&Ž}êŲÙÉ¡ÓDJ0†ŠÏ×"œØ&. Ã,Ä1ÂF˜ÃFZÂkÊ_Qã8Œž”GûeÒ]—õ4z¯.÷ã@øPß8ª³õ$¡®ì½ª@Q:n¥—„ÅX—ŠÆÀË~ž&jøŸIèQ<2h^øL©¨.S£Z¼ofèÏÇ‚•ï,p\HGXsÃ~µM‚1u|¾åD[—p0Þ5U#._F˜ÃFJÀ{ ´¿@e3goÂ…o eð±ôÞè7R¢–Zë‰Ul´ÅqsËj†Pvb^©¯Ñ U<—|âÏײÀ¶PÞSЬ7ŽPÿ[Ó庘?-‰]híŒã„•K?¿ÏS%Õ8<ñ¸½ó3¥#Éh ˆ ÓMAgå¨e€>·}þgªI:>3²@¿bPKø…lƒŠ*‡¸°&F}Ñ 9¢2ã&@pdu~ú¿lïØ-ĈX— æ…ðš]ŠÑIù=(ý»Sp¯[ÃQ}!ChàB<¾épÞ ##·¯‹#Ò<Ç&!ñÁõsQNŒ^T3-udUµ´Bi¥$ü™‡Šž¡Î5ts'ådZT„Z³«ê<Ï“Oc‹â³ÉêÆ0¿Ø{`L8½‘r<ðn[ÀÁÒ²©A8Ȇä)à= úà"ð;|ÅÎ:2aT^öΪtsŠÊ|.PïrÖQ¸Ÿû½qÜ*-a ÆÝYGIi0,¡"­(UTe¥ ú\zôkŒ…EM ÊEþn•DŽÊµM]-[Ùªü+¹•iÌñ F°Z‚qq‘ʤœÌ$K ¼ë þ|-+Ìa+-á=ÕÇ­tþœEMÆ’RNz0ÇÉsRAuÂÊÖy‚–‚]új¼‘µ¾F7øÌ8š!ÞϺoóóŠ}zŽ,v &á€Ã,G‡{z“ˆ“Ò®ÏÓ\ ›¯@£›YáWö0e> 'ÎÜü2©‘m5ŸÄ8¬hókŽrª $¾²É€)tü- ÌÄY:ˆq`¡bH%àÏ,T ÷a&n}Tš9)Ù£ÚagQÓ(T9â*!v}~'3ôÂ+{7¦ÄRš‚œÞ%¡0ÆÑ*Šƒ m‚?O›5þ=:ÜøÒFäÂy%e.ÓQ¶d>³ê%ÝQì9ªÇ£¦7ñ‡›å~.Žht¹À–˜:°Ÿådú_J8Qæpé8]Y1­0‡­”„?󈹵£!X|DÙg⢰ùr‹]^ኌWÔ`Õ`áx%]¼ÌNŽ›wß)Á˜:b„),0µ$D~ó½é Ž¶Â¶ÒÞóP ÂØq°êÃFyÔÞ'%Ç#» wÌŽB.¨G×7ŽkpA°„ć¶`I‰Âeèè% ¬ö°tã>WV˜ÃVZÂ{˜[<Ÿñû«ìîð»QæKÛÅ’#}pþBæÂ契\8µžÚüübNe7vC‰E)ð‡/èî°¨ÙÃ2@ŸÛ>Ï@åG„ô:”Ò—”xÁF¦Re×Þx%¡‘FM¬j"ýÙ8 Ë¥cꈦ¦„‚õ¥ƒE1e…9l¥%¼ç¡táï·v¶ôHBc!®è²Ê’ÏHÂÂR/bwß(4ɲÒ⨋%WòM–6fíÏ%ßÕA—æ°…þ²ŸÅÂ^ñ:q! <¬UÑÉ,lŒsܬ£ä³U‡ÿÚ8"+bh à'áfv Ç{¨Ý©ÀXÑGß8l´%¼'Á›¾!NèÐÉtûEy˜…ñÔl爆“6us×êÆp*dY>Ü1ò\AÍè÷h ñ o*Q>R6ˆ!m”€?³ÐKc²W'‹‰0°‰.ý wõ:˜*Öa­qì ¢ÎYÛ8¢Ú/Dñ¸!!êïEñ9Žªze“OŒûY`Ùg/ûYXðÀ:3J½Ù†{QXH¨FÔÚvtøGjDMÔ’8îá*˯'Ç\œâ囌«|2‹‚bDK@Áûi©(zƒ-#Ìa#%à= ÅÉžìX›÷ߢ8XhØò$n æP6DQ£¢‹‰(ø¬wöRñ@ã¼æ7%N('nø8<¸ãYÒ…qÅGåÉ!ãRÂÛ|s8Ôôf0ú–円+R"*Þǹñ`]+Â=‹Í ŽÌ€ÄT±G4tíØ2ãÐÛ-U~¶1¤ðgª‡{(#!;¿e ²¼F¾?6ÝWe…Ü ‡ Â3$ÑDŸß•W -Ø#d‚ÑÝTåÖâ‰ñÓ’~sØ>KxÏ@ámvoªA΢D{—Í¿Éî/lYX#Àíañst¶þ³Ã¦6‹Âö0KÛÁ,Ur6§«8l£%¼gÁ5~>jÿ¦æuÄqÍ£™(ųuféÝ=q,¿…Æ’£¡,Y ìJq6f%ƒ5:‚õÚ&Ÿiƒ²À2ÐvûåÝà}Åü¡²* íf/ÄÌ/í¸ ?9B‰?lÇÜÚâ@‚è`HŠ#‡Y©´^Јi) þl)®æ…ðž½¢ãdP`¤A²ø¾),ІBùŒíA…ÅÑ4¨ŽIŒ05¶¤Hާgy~H0¦dlˆr"rpI8r’:ˆU®V˜ÃVZÂ{ÜLÔÒÂú¾(ó€Ô‘×ÿ° YTÖ‰Íl\ûÖĨWÀj6æˆÀºö, ‰©#,5…ilKBWî½utEÿ/+Ìa+-á=e˜²-Èõ šM‰8²‚Èb$œÁ¿Á_ÕÚ»èÐ|èv89æÿÜ%KÇçkQæ®ëùNל@ÛT#`šF$‡Í–€÷,Ô¾ãBO”íºÕìX”ó‚·¬vÔ–C©»†Â^ÑÞÄ{¨ìQeŽûdž%SÇçË\µÃçC hÊ÷]ÒaŒ2›•$­\Þó`ôÃÄÏ;<Ù Ø” gV‰Ò|ˆ(s^Ì#þ°²L¹6Ž(¸ ¤U}¹ét‘—"7"^V@Œ+>—­G- ïI(~é;¥« `JíŠ]FBOÈwT¸XFGX òÎð°2O 0¦ŠÏ×¢D‘öŽG¡ÍÒ@üùZ6˜Ã6ZÀ{*Ôìfµj.°(ƒ+®(F„ÐïPåØ\œWâCNëÆpë^ÊŒÝ`£ñl¥6XÜf`Ù`Ž´ZÞ³³^áÏÍK¾¤ÄÀ¢¶]ÎzVa>¯–ý”T=ÃsÊW/ÂR᫘ÊH… o>Üœpjþ|¥ Éa«-à= ú7FDð 1‰Û„ Ív+"wÃÝíÂYq£Ñ»q¸žJ÷ƒ8¢ X_ © ¢Dè®\oÈÎX š²5– æ…ðš*¼ßó~jSÖ[RZezô9TÒýd}é 9#þ°¨ëØû•§cªø|-Jé<óJÀ‰%bi ޶Á¶ÑÞ³àƒøÜŠs_‹Rj¬pK@˜-› bý,a„9l£%¼gÁ_;*À‹PX§Ó”µKJcjaGÅøÑtfì ÇoýÃuòɰcª€ëB”Á?DÔß®8ØsØF xÏB`'gOQ<¬)ÈŽ(ÚƒÉu‚ŽÝ7[<ªšW¸VûãÊÆ;9ZXò.&ÊШ¹;,m ‚8ß(»Ï 6ÙÞSP}_Õy¡)[õ9¸HP2ãäu'nTKQߢ+í,-gŽùí=m`LŸ¯Eaç€%¡ábé þ|¥f°ðž…n—|çº]„C9¡ªK9¨¡7м½‡«ì©ÄFr8«Ôœ#J[Þi´êˆÛÍ,ÒéVK¡.Ùéîåç2Ù£_3àÅrUÙ©H™a@›)s·}D[½¦BUãfyŸŽò Æqï;XëÊ Ò+–cªø|-JcÜ¢DPÕ¹i ޶Á¶ÑÞ³`ÚC41‰"ÏPR½)5Â#稛ý ‘ø»[lAkâÓÞ‘wgŽÒ™—' ÆÒ#L‰øÈòÊõ0q^:„#ïAV˜#í–„?ó`Vb ›´vT¤$²D‹_$i¢•eŸ8Êæ×ÄŸ¯†Âfcã(¨a¸$ãå*¯ïŽF–›á!Ìa-aŸ·¿×ŶÁs‰dEº¤„Û )µí¡oÍœ{èªÎÄÀÜt¯£öA¯kxb*@š°(C›ÞÑž-…wº`—v}nã4öùªxò”7oÅ¥ûM·¬K”°1V¶ä³¾U#ÄŒGs1S:j(¦€†D䥢)1ya)ïYèŽèRèlWOmQF´AÝ*®ž§ºJ«.>ñ‡…ÐʵqŒCØ(€ÐpE,Ê=xóC8C¥Â™m|ß%‡­¶„÷,¸jD,büu ‘žj´mˆ ˜ÂC…6ø„ã÷|cmXóå%)ÁøÒ Á¢ 楄Îb¬©£»\kZai¯IÐïû¨¨zv%IЛЗº¸º-É`íÚ­- ãW[\èmmI„·¶$¢¸-‰d[©XmIl„8ÒH ø3‹˜Ù×ê± Aoø˜Ù¢4ܪnþ®ÂvgäËÄn@ø3GtÔ \Œ{YŒ©#F˜‚†KÀ‰šT&5)ªœ"ÃmSPœðs›—õƒwóq‹Jg±£mlRø»Q:œ<ǽֈ”¬©ªU¾øç)*}ÌÞâ¨Ø¹/ÆîŒ¸(‘˜ù½DC£¶© ޲Á 6Qãßsà»ù¹X 3®†±Ì'¥LT9´ŠÅ:!^´Èð] !’K€1UÄS\uÈNv^´ŠÓqpiƒÒF ø3 e=T&Ü(îù»QæË·…ÇºÑ w 4?ìcûÔ£«ó­Rü1c]Y†‡ÇC,ÄpeÉsØFKxÏ‚Nœ8ŽßÌúc«Ü¤° ~]2•î’£¿2ØžGŸ³zÄmLùÈ3¥²d À¯Ñâ lÝü¡Ê.{[ÎÙ̧5QZ“1)Sk$e´ÿ87S?Mˆù”· ª³µU 0¦ LH”Ê)a®N« Ĭdƒl£¼g¡xQ¥=NH4%*˜²(_UØ‹‘Å@bQ=˜Hþjmã¸ïŸ]¡4€Ÿ„([ï±q)A£Ý‡3á¤?9l1†¿ÍgÇ\ìõƒ…û0 7iÖ5½ã—J§c½KçuªMŸŠ*–I@bû8“0.¦ÁYÀ<32Ž‘7'Ä&JÀ{z_¡ËŒE/,Q:¾ÃùÂhòjÆ÷;ÿܨ-ˆç£2ÛEŸóg½†SA 0¥âO°”ÎgÖˆñÎ’ âH%áÏ$ø÷Š&~¬D'¶ ƒ‘º‘›y¨#px"7G âÓë½q\]]^%Á˜*”VÈ&Z§„Š˜ÝTü¶AŸËB~Í€+Û‹ã»f@@R3þÝ"ã/nt¢\²0|ÒlN›®o ÆÝAIi¼ÓO ÁK1Ya[i ïyðÔ5w¤pRÖÁþ#‹Ò˜Q £öŸèy„|‚ª0¢µÈ@¨çÆ=]Ú&Á¸úæ')EΖp2?#uœÎàH+Ìa+-á=ž•Ï“]+ïCï­EÁµutèŒG9ÖFš~ؤ]2ôykÜ7xxâC/®¤\(u±„Ï ÅÓ³‘Êñ™-óÀ·í*qs3ótªé°(Ü¨Æ a!n¥Qþ wF—n©ágŽ Éñœz ‰Ýi<)›Õ”ÐàÏX:š{§æ°•–ðž÷…ƒ{ÍØa5y4¥²ó@ô>­òÀœ½Qûþ”‚•Y ã8Øç[`åŒãzŽ?p^ ‘— æ°‰ÿžƒ*ùêwpŒ’ÊwðÉ"/JA‹'›) Î_XÞŒÀKއ FS‚1u èI-Hvgµ„pý^›aܪ¤’9ü¦¶„÷<Rß/5 k|‹-JEÁŸ!»±Òì[<.ÔLDi˜¢È]1<‡ZžI€1U 4Œ(gW“4I8PÎÈ€PFècèÑï)¨¸¾ î˜~7JeÑÑ"^Qy«àhú$F]oD¡.Ž•²—cêÀ7!Ê©&Å–;¯{ÓAŒ^n²Â¶ÒÞó  ³ÙS¼™‘V–øQ³ÓÁF\8‹(Öý¼·³è†SÁg+Ý×á¹YZµôÆP}kÆi¨F½ çoëê¨ÀuâFîw ÂÁ‘ÍT”Àˆ³Îƒ1cèô2G¸'Ã;a ‰Oß&å@Ñ%á@DùÒA¬œªR6)ïIèŒ|þÈí{³=³ — ötn¬ò5zO ï5N ‹£t¾j$!1UàP-JAÑ’p²ÒKê8] &­‡”€÷$¸9 ßNö•]H…Ý?”ÿå£07cè®–µ5¼BulLß_Œ©JQ*˜KBwwé †»@V˜ÃVZÂ{ÜkŒÆK´pp³ã™)ôq£ HGÅB¬Íh²`ìnd/†Øt쌩^vS`PJ8m‚1@6$CZ-ïY¨çŽ/=\úh •¸ý«4õ{»õ9ÑÙÔ—8 §D3TäN/ÂRñùZßúPÀÅv ©Aøó•6ˆ²¬¦€¿³àbx\ô46Ô’üÝ(Qç(j¼áGÄ´M,MkåÐmø5ßSˆß7G;”ôJ†ÔðùÚ(¨·º°îRA#l„9ÒjIxÏBáôU½q.VÚO“™j,Ë.œÍ•*Ü’!2;ÙáGþ(I0† œíí‹r*è­?È~JÂpwÉei ¯YðÝ\zv8m|7›r"~ àkÒK/^¼­ðN¸8mô“!‘æP Ô`LŸ¯¤Ä­DE“TJˆJmצC·0"dc xÏBùE·ú\üà@¬jrD¤Ù$¦ŠÏ×¢D=¬¸J°„~ñ`lÄ8òÑ3ØF ø3 †4¤GÑÜÎê¦DULµ­JòA¥²þÐ÷G×`X|r\ƒžeK0¦ŽÏ×¢ÄÏ"^ ’¯9–*£áÏWZ‘¶ÒÞóàõÞñ°Îó}²ÖÍ¢_‹â;]K`ݬ¥£©¨Ã²Â¶2o…_ó «ç`Ï ¨’7ª(JÔöF¨Á9ÔE¥k/e5‰ã¯PÝ©HèѰJCð›€ò9ü¸‘e“ „c€LHŽ4ZÞ“PÚóÃÆyщµ(qcs1 yÏ•¥Mâ`‰Óñ‡}7ñšG¤ú£l¯$$¦¤€‰r ŠÈ’p `é V¤¬0‡­”„?óà!´²ô`Å×ßrU‚Áw·2*ÅD‰ÿ‘øóöƒÃ~jKØýÖ/Ov4žB)IGJÅH‡0N¡´"9deJxσgšQuÿÓ}ÇdJœâdt©ÅËüA ¾Ø­0Ä8p<Œ›4Çs©«‰$žwL¢D)¯ï% Ñ_”*šJi„9l¤¼g¡3 :×ÄΚ &Dqhœ`:” ‘2–r&ˆ ã|Ryâ6‡j‚§ãSÁ„‹r\¬ f Ç¡=¨t‡w©¶Â2Ò^“`fQh鎯äw£œH@(ýdí^KÆýÔ™¦çÏØ¢váØS…"RÄ“¢º§´u€Rƒp É!SÀ{òïWU£ªiœ”Âu'œ=Ê}w4Á|Tð¶*(;:*³ 9¢œ-Üõ˜*>_‹r ØmIˆ»ÂM Üû4AÛ@~OÓz”šRÙ8aÁ½H[Ñ ]LqoRÍ\ØçôèÇ=˜üd Ƭð Ë<‡t™§‚Cµµ– æ°‰–°Í€«FxY‘ Ó¦“„»±· Ï"(†_qíÌf„p„ÜY-¿"еo, 14À"ˆ¨ôøþð+„Ÿ¯el!Ç¿gÀÜàÎà¹v^W,ÊsªæØ=@¢¤YD¿¶‹PáHܽ‡¶8bK×7ÆT²i¢\HmY¢‚䮃éÇ4 ¶ÑÞ³àɬ¢!ÊaêdF štTÊì>-®+[*\-ñ‡%5ïkãˆ:l»bëˆI9x²¢„þ°4—uÇZ±8ÒnIxÏC;*Ÿ¡[ASòÖþ¸“q2",V¹+aìw¢?qÝjã‘K…¶` Ü ù²4Zì©B;gÚ ¶ÚÞ³àÙìfÓ“(øÅÚZI ×ÉɶÖt U6ì`£ãS|¹lŽr2¦\݈r!G3%DM©V—áÏWZ!ʲ’þ΃/çL6¿Õ•5)ódu^ôžxÙvzSZ“¨ËÏûìZ7ŽqÈù' ÆÔ!—=8è›"á¼n:„?_iEr8ÇÝÞóàqfxœ /c’P=Œ 9¶ü…)º½3òK8—q¨ 8†o(%ÁøTsæ$Ì¿3‚g$ {ƒI…°NÜÄ‹ÃFZÂk\FæCŠHÖîûÝ(Oƒ?6^-sÖœ‹öÍ Dâ;™—)ŽÚY×ÉŒ©#F˜rápi ˆ`:—c¼[T¸Ø¶ÒÞóà®jpÝE)i,þI‰e+jIW4Šæ;ᛊ£5BǃF/)À0@”†SÊPÑ™p© FùjaÛh ïY¨Dß _sïz«™0ðÃ.WlI!Ù½\­Ëu­ä÷õ[Û8¢·Ò&@°+½> m°0°‡W\ß,ĸU• æ…ðš}‘؉`HWOHJGÃöÚNf„Ço>âJ#þõº#B¶0ˆ] íà³kÆÍD¹*ĽS@Aq„TPXœèsèÑï)¨Æ ùa»ªv%¦°ÔX›{Ô¯jèÞ]K¤æ•Ä–ÚCŒ·9ÊàM—%Sb-E9/6ª³„Øþ•M16ͲBi¥$ü™‡ªå–”n|7%n´Qÿ®<›Ÿì{u³‘‹1îäÈ6Ç5TøL„¥sDa—­”7)èò)Â(˜C+’Ãv[Â{ªR²ÊðñÏfK`×3ê¼¢ÅÉFö'.Œ?ìOz+æˆR²»c¨@f‡,~¿4´ _*ˆc„0‡´„×,Ôóˆ)ÖñŽ=èC5¥òpÞ-ª=!$zKã=MìCñµqĹ ¤ð“07üØ€xø|ÅÆî,ãµ.Ìa-á= N,¢Òð]`ú»QzãÑûVzÂ(j”Ù 'üá®^qd¥DIHœmÉW-Å¡r’ÀáKG¶%O+Ìa+%áÏ<øž>MY{,¹mÊÚ*µ PGoV¨Ãö•ø£óAÙ8ø‚YŒ©B{wRp0–°Wª0ÆN€i3¤Ñÿž‹)4z.£F>s¡“ráR¥Þ½¨ÊÅLçÈ ã©… ûƉ–»cêÀQ:‹¦„Ö˜ mÄ1ÂV˜ÃVZÂ{tTô±7¹è'0ådwŸ¨‚÷à”[ÊmDÏ+qã#Òmc¨é±SÎý¤”'r <~®€xqXpðË„ä‰ÿžŸÅÑYj§DU©ß@g%›7ЉÎë·ÂÄIb\Ø¡)Öâx¹¾²Z„C %$à,jö'§Þ¶Á i³¼æÀýǃn¯Ø=wÞQˆr²þù<ѵåúŸèvS}èÔ]J)ìn†ÊŒ±/, 8tŠr£ãHJˆ;ŽðÒH… 0ÁŸ§Íÿgª"öÐa΋ßE $n§ØGÎ!õ`¹=¬CöäÈ IÞÂ:L¹o“U©IYÇ&­HÛm ïy(–L/öC•=å¬lÞYTê›—Ê(çþ´ÄF†ŸuãèŠÒ¶ãCu7’ÒãÞù;DúÕ– a$ÈÓˆä‘ðž…¶VpªâRƒýû’r²Ÿa"CiµóæeÌèªïz;äjêˆ(Ž»0¥ÝŒ©ãó•”ñ ¿8%D#¸èÏgÂØZÑŠä•)á=yA: ãvù‘D”EÊÈ·ù|Œ=PˆÃX‘oØ ™#6¸K·*i. o¢SÂņê©C^Z!ʲ’þ΃·çd¿Ôrèö}QÜ£šW܈]¼Zθ»x•nŽÂNL°ðáËwSb/¢æ˜p°;oê F­!†´QþÌ‚ëȨê2‰ÃúïF©¬Cv\ÛÏ6¶½ŠUн2›ïÞo8QÇ3Ô,SŒ‹Šâ,ʉÐ)à<ÙÙ*ˆqÃ##Ìa#%à= ]RpÛ}†¹Õ¡À»‰{'Ž[ô¥Ÿ±ÚŽÄNÖ¹G‰êÍç’˜*pÜå+ѽ(®™:„?_iErÈH xMBar,~2x‰÷»Q4w-·*‚ èD<›…åT„'§¤ s¨ IJ0öÕá¢\8t/ j*”:|±·¬0‡­´„÷<ø¶Ž$!¹g;Ú“ò°‘-«ºÍwa8×ç«Qy䀈•xH¯ÏûáW)‡/|hŸ™” >% ¶\ϦX! 0Á6ÑÞ“cÿ€óáþ›tºêhysñvÿDˆ ¶éØËØÁ =~Í1õ]±uà*@”Åh-!ï¥c»!–‹Ãv[Â{\øáCïW«Ü'!²NO„E–q0\š>yc””Uûisô“Î\K0¦ŠÏ×¢ è0ÿ¨ülÄXLe„9d£Æ¿¦ÀåãtõÒîö ¦Ü8lU$pŠ˜¡‚›TaÄÑ«©9\ÂŒ©ã³š¬c”úÞ¶ÖAüÙjY˜ÃVZÂ{Ê9õucš¤p†Ž¼GÕÃxÄî„xí^ë舰Gß·ã5M¹ÐS†£/ô0IÑ„jÆ1’°,»ÔÑæm»6ù•…'â3Y‰rÆm=B¤PDyb^×GdÌx#–æf6’9"¬m„¥C·Q „÷wI”€VÏ›a¬5´Âi·$ü™‡Îe¸ÐGI^›- rµçOv¨(*j`ÍcrpŒ?,ºs·Åâß×’°ð¡{³¤¶N ±¨M1Žf²Â¶ÒÞóàEî ?Ê%£‘Ú¢Ì?~\ÖÂZàjÓ*’¯Qo¬ÅÑPh)5+]”ˆ\ÍÐpç°T£X»Œ0‡m´„÷,XJ=ð&éÀ¹ýwxr-qÔ¿TÕ<¶fõT0â(s^˜•s3QF)Á˜*b„(ñ·B%tIˆ@ÙHmµá!+’ƒF¦€×$ªÉ| ÖžþÝ'ù‹šCoãA…HscT–6¶ã¬wEÞH€ªO§Š½à5HŽ´Z^³P,ãÍ&·GÍzȦÌ×yEéˆÁ{tFRÏ×|ç­AFo·8-Gœ–v ÆÔÝŠ(MœSÂü¡†Ó0u#–QV˜ÃVZÂ{*`}«±pÕß”H}½Çªðª+:®4æ®àƒzæ¸UÒÚ„¥ãóµ(·ªwKÂ}Ñ»`Ÿ­LŽ9l·%¼ç¡Z™ÒâÞ\÷Ó¦œL¥Œ—’Ë¡J–(nd ¯–ÝÅѹ¥%Á˜:”ÞÂÛåÆ8}I¸ä±±a½‘ aY™Þóà…`-ì{×;Ëd,Ê€ûI‹Hꈎ€¨þ&ˆ Á›ÁŠú|¨…€‡w7LJc/ÎPŸM<©ŸÙ2|ÛÎå¾!kóŸS"]&Bb¾¶è-|ãa?jbTèï¨1—ê0k†—ÒÇe fÄ)k» âa#Ìa-á= kaÔÝëžß‚Ò븹còê´×¨ðy‰ãèØH› ½(²’ã © øM™ZÔÑøóaù+8± æ°…ðž“§¯J?ò©€ÚE)Ȩá× {l¸#oùÁÍ•ñ'R¡º®ÅÎŲ H|^.‹lJDŽEî³%»^ [=>²i÷6^AHw¹.°åd¶n2šÝ%èÐwÆrïî ••ÝS€°T`€(÷Ãx} ˆû²gÓ Œ½(mH[mïYèžL-°ªÿTIiü¦q}xëÅö_àÞ²A˜ºˆ#ZÆc —„ÄÕ«¤~ç)¡à|¶tÅ>/+Ìa+-á=0 —×¹¨à¯¶(­ã¡Æ‚Ü\"ˆÑç„ÞÚä˜?¬ºK0¦œJE)•áR–pvk¥ƒøóµ¬0‡­´„÷g7Á5<±v–‹Ò+²4ÄŠ, ÄØ€ÉsØDKxO‚.‚ëVý?ÎLÊ”¢Ó:Nïlé€Ó<‹žN‘:Ð)&9ÚÉ»r HX³n¦)¹SK@xÒÆ¦‚86¶QþÌ‚7l@Ô£=óɪ¦Tì£KtËâ7ØXÑG>ãøF†¡˜—âà›$œ¨'´†Ç%Ê®€¢d‚9l¢$ü™„²yN5taäEyxÙÀ®:ˆ…¼Ùg„É>ªŒmbP©OŸ?-‹°cxbWFNJü‰c±€VY„] š*#§úÜöyø{ªžyð^Õ,LJ¬àá£~nÅî¡bëX³ëйEãȶq$6§„ÄÔñùÚ(Þ)+øÒA ˜¬Dz›þ̃{ÖYÇf ½ÇïNyØç¹3*+ÒiŸ¸¯….gb4u9y5kŽ „–„Äîlž”ù­Ãmm ->–ŽæÎæi…9l¥%¼ç¡zÌ`C Ëeˆp1˜©ÄF¼–ŠÒƒçHŒP^Þv›£Ü,YBb¨ÀÐÀt è•]¬‚O¼Œ0GZM ïY0X˜`éK?\Rnd‰–Î&c¥.=aŒU©¾µÅ1ظ&%$v7ÚEáõË’k÷ØtG¸°­0‡­”„?óÐ}ûÁíçÝԨ܄Ø(¼ÿÔu;R¯Ÿ‹aøÄÖfj#)õ8«ê(TFÖß eLÂÅU ˆ»ô³/Ÿ¯4"9l$%ü™…J Ÿ,–u5¥L'åF9ǨÛÛPôP©a¤MDZY^ÉRM9Þˆòƒ]„ ÷5º«N°å¾–æ°’ðg *Ö­°ãÈa­n:Ó"ÂóL÷`å:æ(fbÄ5?Lþ1G¿T7VŒ¡"˜PQx (ÌŠHĈµ–æ°‘–ðš/¤O6wq}Sž“W½h‹Î•ˆX©G]-#¨Ü7¶‹ÌÑÆwqmS˜Ä·$D ˰´: zjë<úm¿~\5‹2Ý69æ ·M€1U|¾+·‡³˜ÎR@ŒØxZ`ˆÑoûù Î3ž™‡Yí¢d£ì¨wˆvÔ.Â3ó<‰? ź7‡bY‚ƒ¢¨ã&U$ Qø]8¤CÏú¡0 qØnKxÏC-¼ÆOáÂ^ÄQý!ÌŒŒ±Â!‘ ~GPO¬VGÛ8æo»o!ìÄá=¿×è~óçhùÄhÞ% Äa%àe¿’ÊŠr©›’B’râ"¿ŠBN/‘¢”7?Rà eqôÎkK0¦]< U *H"£„pÉ_K!\öª0 ÏebO‚ç•1X>ªt-I©lõ<®FzÔz8y~i=1êè²#˜9®ã@C 0¤œWHˆµãú^ÃQˆ–ˆ¸2@ŸÛ>Ï€.»ìjün”Þ—oùÂ{x/ö”›áôü8ê”B0þ°M+Ìq¡€Å’`LŸ¯Ei7S¦,¡qkž:ˆ±3‘æ°•–ðž‡ÊÔDjwd9ýn„{“8áØÇ!MªFDïÙǾmzÇ<*Ü›ÃCÅKM`mÐ5¼²`* Ž6Á²Ð^3 /ñVÍõh}ÃZ ¦TF…œª¶×G™eªë^]8¦œªünŽ¢Êð’˜:p$Jøþ&  a.‡êK/+Ä‘VJŸy¨óNG=θÀ­wD‰ Ü{U}Ð9úU\¸>QñØk–zœ1%¾¶ª¦=9ú"S4ág«»ªÏm—G¿-W*ce„^” |˜Ê(JÇÎN1ö A@t¹é áªxÔîJ õR³ 0¦ŠÏ×¢T4,[ ¶©‚PŽ9ºÕÄ`-à= ÞŠ5¡¢b4 9%¥€»"OU}•p5Ø‚D8î¬â¶­o7¼ÕK‚1uÄSæ³%§,!îŽM‡pŒÉa+-á=¹Oæ–áš@¦°ÖãÄ3cóŠWò¤a¼’/H0Ç5Wüg“˜:ðR…Ø%À—÷_V˜ÃVZÂ{ª,yó¾é€Ûàw£œ5Çআ)(ÏwKŒä€¦[.qÄ&q xØU!B…|}д}°[±a¤‚u‹‡LL ïIðÞ¹ ùʈøiÖÁ0%&Äž_=R·ND^ ¿’»ˆãPß+Ó¿Ì1N¦‡Y‚°tÄSnì/Rš2l:„c„¬HÛm ïy¨ÜQÏf­+¦DŒHlŸza„J-Üiú“„OçXr„¨lŒ©ãóµ(±—BÑWJ¸ž›®žáê•·"mErØJKxσGˆҵpÑ1')ƒÙS¶€"lk7ò0ÞL—2–ÍpÝÌP¶ãî`œ¤4œq—zøSEu,NÚ`ÛhïY¨eÍÊíMQ˜"ÄÕÔÃ?t\”“«_$ʳD¼×ÇH¦o÷Æç (z~Aabi ˆŒc,Ÿ¯4"9d³¼¦  çúsò­‡†I¸™úýÞùº>Ûe.¢”0Þ£7³ÌÌñ?›A*@€3 ,¾¸†÷›aeV@Œøf™`Yh¯0̨\ªÔs¨±ë¢<ŠñDÂÿ|] [`‰>ˆ±FPß=9››¤„ćn’‚VzK[é-Ä1ÀF˜ÃFJÀ{|—=ï€ÖÙïjQ"ü¯ò¶ otÖ-‘ùSÆ{)úî=ëóXŸÊ6ܸ¹ûlR*êÝ,¥—bÝ·‹²L”„?“Ð&ø@ÖÖ5Wµ 6R°½,7£“¢Ä[Ô\å&Ù¥~Â-4Car XøpÄ•)ç¡t ˆª«»âÏײÁ¶QþÌBµv#ÁàUj—„ÆêXO‘×…vOÞem•vÝÑQs?Àncã§opÚ5EEÇR@Ú•Š­Ò®ŒÈJ»4QÞSàVø†Ã<ToxMˆwzg¤+kÔ4l×£jjc#ÄéT®8Î;ëà<}ÃT¡}PâMÅž$o²kéVt­sND×ÒH xOBÍ®Jl’çòvÕOS˜Y6ñ£pgMÏëfx1¶µzç›ãA¹ì%Á˜:b„)n––¿Y:ˆc„­0‡­´„÷<ä8˜Po5:DÜ Ed<¾¿C~‚‹'Ë!++ÉÓ’ßTÁiY`[h ¯)0@^ºÖºbU’ÒQw¯ÔÈÐBXjÁFwÖWâxÁÐ1h\ƒ×% 1U`€(L$H%³–†¢ZËsØF ø3 ̬tüà{4©doSÚ1ý3p±Ú£,Ot¯ˆ’¢‘‰-ñBQzbl7¢|–aéÀQn&ýZBÿÞçÒ!#dErØnKxσµh"œ¡|–¤D«ø+Õ“¹"­ÄkµD$[¤M ÇwÚÀ±8ú`fˆ%SGŒ0%jÆ_Áxƒ–ŽCþ¢e…9l¥%¼çÁ¹5ýU`µ`e>](MxcCßiª4‡Òð´ÜGÜìŒ Ÿ-ÃȪú^£OhLùÄúâ`9l¡löóö¸¨Ž $%"_Âí¬Î{»›ÄâŠÞ|—ûŸÄòŠƒ™9ŠºûI€awû“E9äù–€†ÊÓKEsµë4BËjJø3 ®ôgUIt/ÿÝ=£ÿ©’[lr|0Û‹n§ë°ÚEo­CÖmjcÙn/Xù–âÏ×2Á6Ñ^sPåê¡ø[ ùÝ(ôu-_FøºÂwÁâV„¨[ÍZ¢ú<ömç6<ñ)+…©öKÀq Û‡îÎ}nû4üÏ ”ò|ÑÑ}7å©'eJ¿OîSàSþÿh{—^[–Ü"ø8ŠÚtN„#ý+Û½·¶%F9¸/„¤—úsmަÞ!&k¢ý‘V¤„­ôÜÜ·áÛŸÿøo?JÔOGînô9@ ,Û"çóæÿé_ÿáëoÄÆÊÃ_ÿô£|ýÙýÿ¿ùq|ý»%ÝÄ©¬Ê²éŽ3+9¼Õüs©Íaœ?ò¶pûJ©ÍyHE:x\v¶Tr¶v#Õ(¥6ç! ÛÏë)•œ‡T Œ½-•œ-õé÷ó/±¥Ž¯¿Œ?ùñÿÓ~ýþú·¿ûñoþ}¹¯3wÔå~IýݹÿÄù*WÌ¥¾•1+º~ýîûÇüùÛö›ßö˜ë¼æÏßÖøw<¡œ?[î÷²uþ<îžØ0ùYðÏUÜ£þl¿ùÏ¿û³ÿ×ïî0(QPµþÕ]DÍ™QD{ÿÞñ=þ RɅ漤â¡Ïè)%ÎKjvd•>¤ÄyIÅ)À5žRâ<¥Øöi—9/©Hl«O漤ä÷Cêñ—øŸüãûûm|‘‘ftFbq¯ý%–hˆ1*¾Ä­àK¬Ñs®àK¬QPz³ þ™*ñ%úŸÿÒ—¨¬càE³Ä‰B»ÈÿCÃ×W$UÜodÏ໯—÷µäúŠLœyö°ûçÿýwúþý¿þü·þ›ßýÍ?÷ù·11»ÞïQ¿†ß·M_¿û‹?÷Wø»?ýáþú×Xù¿ÿÛøë?þáïcýþE&²|êýØvBÕßÿÃïÿö/~¹-ÿû¿þ‹üý±®âÂõÿç.®[ÿ1 ï Z@ÿçûßñ/]Éîëc«™©]íû(¬@*Ò‰ïþÏÿ‡¿âû‰3æpÔ×oxE!ÀüŠ1Ç}·Â·ÿwúÍo[‹\òõó~óÛè_v]gýùwd_Wûù_ƒ=®«Ìñó9îå·È?Hæè¹´ÍŸÿœÇý}ýüîÇÇŸ÷ÿþÝ®«÷2·ãÇíëúù§ø÷}O\³ýü3"'nµŸ¿€ÿõSÿß¾!©ªÿü§`×ã¸úIù;ä×}ùÌuÜ?¤°Nåƒx÷{›L†L?ÿ·íÏÃDz?<¼ù —Œ_î¨<þþ+þñõçÉ…Ï?ÉÿA'ï[Å™¿ªH4^÷#Öçÿô—?{mxýïÑ=à~Úˆ–×¼âw_~9[´—ÿŠŽcÑä8þqE B üwŽ÷7<!Œj ÆEäçýã þUÄ‘~¿î‡²þµ®rD¾Óx)ûO?ï_âÿþTIÁáÚýû“ó¯We?÷ØSÝ?ýÕ_ÿúWÿkœëQh½´ýŸMsÑã›:¢íÏ_ú˜«Ý’ÿzUqI9ß_Zû_ãT nZýýý&Uí f Æx«ü/˜ÿ½½Æi°qÛÅ ææœ(G-WçØµµ›|Ó—æÚ5XÝ}ÈêßqD0¦©#V˜s¯<®¯0PdºuŽ¶Â¶Òo?~Åã©FÇšA×̈™“ÛB±¦Ì°bàV4ììlàe íL É»ü›ÑQ‚¿×7œÅo ¤á—l°„M4Â˺ïñ'›óœüÊÌ™ècy«Açò›FýëmFá[ÔÔ´˜‡|±Q”'že7‚iêÀ q:Jd7BC§Ë­£©æ¶Â¶Òo?èÛ8xÊP&ålNÃÓB5uQ¦ZC/[«ˆ†¥jáe‰Å!Õ‰`š:b…9s¾6ÂrÜ­ãPÁî¶Â¶Òo?~õÛRÔ_EñæÈ·%pNÔŒ#\**„0Ñô¶ƒç¢aiE’eJ\«Ö`š:b…9‰¡_ü±ZG¿üs¶–°•FxûAß°ƒÚ:ço÷1-ùhh[¢:»¤á&‡n‰èº_¦©ž‰Ó+cØ p[iE=¬°„Œ4ÀË ù…yŘݎ†þÉ™æì!VÂvà0¼`$pOv²é85:#!iéøõÇæ°k" &I¥ÑŠù°"%l·Þ~з³³Šƒ“X¿œÈŠêŒ•¨Œqϸ*\H2 K1°Úœ³N¢q¶6ÝUL´9L"ßøºU€TÄÃ}n¹üÓ :¶˜¶*£Ñœò0*Q¬7:G•(ŸAi˜‰msjtW*aÓÔ+Ìi˜@°*Ò^¶Ò yXa [I„O?äk£«SçÒ–*!XNQÚTcˆu z K‘ œUÅwØôô9Íæ »4⨽<4VÈÃrVàÓ zõqՙ앓Œ1ZG6kñKâHïÅéj–¨á’¦©މ·Š¯ mŠÎ‡ ÒŠwANÚˆõ.È+ôœoѤâ¤Wb ÄA˜àL»"›½F“Á™$lìœeH=z˜>˜a™ 4©€ÛˆæfÇH "ëèÎDNš(€°%©×(¥9‘÷ôýààm†1‚l¥ …“5Ú‘¡ÇiØÉ 䔸uÚ@$5@žŒ‰3­½üþ…–òP@Zq,aðv‚Ž•C5WysÆ¡+Ïâ‘5´r†8IX99ðAŸwœÄîÕ¦»Fo{Sl€3“·‚ʙʩ_Û:¯~Û/ŸN=5ÎÌØœŽ„,„H—•ÛÆy¤a%î)0PĺLSE,0'Š„¯¯Ȩ•ÒŠóþ”°x{AϪšöt÷±KΨ¼Ù²ãúM³M]› Ke¸‘]$¾G/¾”ˆïçÁtw'»ä´Î?†*ûÔ¥ŽêNvi…%l¥Þ~È7ŽekìƒúýäÄehlJäE^à&aggŸT LM’Óú$_@ÍY8#J€èþ~ÖÔ RÁ>·€-4ÀÛúÕ “£$'ìÉ9 ¿ÕM£¨¤b¬xO†"ÓpKLtEݦ©+ĉ£ÿX#h‚Nê ­ˆ‡–°•FxûAßzáD™£±îæÄëÅXA[ª˜áP£È"ËHÂLìEçç' ˜÷jÓÄâDr!î@Ž­€¤¢}îe]®~Û/Ÿ0…öI¦ù LØVœFÏÆ óa§NT&äç µ¯{¹é•N™ÎáÏ̸^Wj0­PÇI¢%l¢ÞNбbÞŠA! Ž™r•AÂ1;ÕyÓxDõëññb ¶\nš à˜8q:ð¥Õµ¥œ´¢ê-aó´úm¿|BqÙŠÔ˜J—ȸ:&«Íâ<Hî½N&v¾ç[âb5\"˜¦ŠXaΉ÷ÁDX“eÅÖ!ZQ>χ„Œ4ÀË ú5›·s/áûÁéhB9Kî&Ôƒérs¡âé¦=.M&€io`lN|Q3BAGŒTAR,` ðö‚ž³ŒbMz&NtùÅ2²Ð¦ê?¡–SEÃP –DìJ`"LKG´ñ1g|YB4ûCbë ßhEJØn#¼ýoØŒ«'g?8ãäÝÐýðÀw ¢ahÃ×1¡i<’ö äÍi1ït4íRÑ20dƒl"×ø@¿ª­o_Q ü½9Ñè´âïy°¬ø@σÀQsq¨+OŒ“² +ÐÀ´TÀ/q"êÂsœgçcŠ4ˆFTІ”°ÕøðBžÁéuxÀ«Ñb'”ðÙ,ZÆ.FΚ†•ƒû™–¨|@@’‡Ç»šÉ}ü‚ Ù}ã¡B4ü¢–‰x»@¯¢ïÇà•²"õ-9É]q,‚þûj|^JMFleæÇqàý\nºiýæh2{"f~QAqbXê×ÇiŸ–xÀ–a‘t‚K\e2ðæ°¦¤FÊ®î³^ja~ƒi˜yrBˆ%b¦å@$5„¼QO‹¯SË£žö©€4“ –°‰Fx;!ÇN<Ó`º3çÒ¯>zn¢Ý0b¦vŽþ" ;ö0S¢šî*€¤=£59U %öO¤é´Á6Qëß>ЯÒPlïèÌ·9xÛÅšŠîóµp›Á4»ÜœÜkئ”ç~ÜB¯#°AÌÖaA+¶­Üo?äç°.$xp0+±ÆQ Þ¶UŽP8ÒÒ4,ìi‰;&ð i„¤•æ¿9Oó…[‡K ¶–°•FxûAßêàQ‡‹âÌÁ¨_\¤bÎåÉoá€ìy¸(®FU,Æ0Yb`²éF mX!KÚŒPc[û¬©Ã4b£ª»•$l·Þ~з†Þií¸ò{KŽþêlÏ€føVr6b~oq0Ð˜Ž¸Læ·&ÆÒŸ\ËS_‰ˆVÐõ!aðvBŽ<çˆ:Ý“Ž™Ã®,| G#ÊGºåJ†b#qKD?€'BÒ‡ž’’SØ50¢Ø¬>tVÐà KØJ#¼ý o‘oÞ©gÆä´É'¾Q8Ó½£Ådœ²Ú’†ohä¼%N¼9mÓ5ÿæLRÞzàJ¤ô°Â¶Òo?äê¬[Œþ»ø½™sbñ‚ÞhçÉËÑý6QWÒ°tðLË«âe?LSG¬0gàün#t$Òo¤ô°Â¶Òo?臷èïß°õ‘æ#!^âü‡0c’AÀFÎ=åg ý~÷BÓ„†89Ñ~Nsy´œ×F­pÇÛ¿%d\"¼Í§K1Ûqp}aw 9sd˜||MŒ"É–$ =¹§nv°Ü¦©â×›£[["4¼8ZCSç•´@Û@¯~» ·Ø`\¾S‹19Vº¸]Sja»Ó°Q $NÙ^N‚àQò@ò`°zao¼ûš´ÂÊ-acð²îœ}­b€žë“Ó:C#.¯ýê¢y‡½«_Ý͹ØIKQ±wyu‡?âpzÖ^gO¤ÛO[g€·ýò‰w¹¨Of_Èäœ<|T18î‡q_T1¸iŠ“·”PÉz"$M±ÂœÁãËD謀IÝ Ò KØJ!|ø¡nÀƒ\|Mÿ~pJcÐ^ì¡vé½¢•Ïh¤aiÑî…$ 7[Œ4u`…úÀ]ƒ!+„µ0)(uˆVc«B¶Ò~Ð7AWô:g§csÔè¿püoç/øò‹ò™1P¸‡hA’Ìàɨ„eS™¬•ïE)q?Ú”€ÈîC43Ø`/g?‚­ ú-M„-ÀÛŽ¨“nûø~?8¬ÓG ¨•A(C…Á•4¼Òü)K´“x0̓“3Ùãû2{ùnÙ>ØFXÀ6àí…<»pÀŽ¥‹ž‰Ã§-„IC#Ÿ¥@›z[Eâo‰¸(_ÓÔ+ÌAæÿW" ?@Û:D+Üq °„­4ÂÛúV‹«¯“½Ã’ÓŽÉU•Óq+~j45iXŠfE["Úì=DKVˆ³0Ì+¢ì¯o$îíÚŸÛh/;AÇjÑ fU~?8÷"VxJ‚a—5òxÇLfžÅa‰ûÖoû‰`š:bEr8ìÕqg::D+æ{H¤ÝBxû!ßNtxÿ>9‘ç¢`áa ·çãñ“í\½_£yJHÌÂ÷}#˜žÞÀOΘ‘¡²:ï©£{?­°„­4ÂÛúÖ§ÔEÂ/#ÉÁœ ßÅOUzL¾]h{™“åʉô¡gZq0½.œ'@EÆßVaZ1¦½–hú*€·ò ç3øfy¾¾9øn-C­ pqbÿJÓ°sðÔß'ºDm„¤mÝ%‡Ó/7ÂD_°­ƒ´¢VXÂVáí}ëÓÃ!5üÖ̹/x±ÑÒñâ80Ï£NœUš†¥¸Sn‰óbû#˜¦¬ç¸G‰pè¹Ì:H+êûõ°•Fxû¡Šh´_h#Ü"HΉG„ Ó-;o¯“£ÒDÃRÌíÜ÷^ø`š:b…9‘ ‚¿††—mCãͶ–°•Fxû!ßg•ÖC§ì›Ã<‡ˆÏ?møXê/–¢CN \³šIzdJNa"ºÎëK ¤õ°Á¶Ño/èÙŒÒ&PÇ•Œè£Šrö–ï¼/–ÒŠ†_~Ã)Á†#Á4UÀ1q˜F¹tåL¤ó°Â2Ò/'è׉ÔTôN:fNÔ™3P0ynv)L±íÖy?uÇ&†â= °\Ÿ¤gÚmNãye´ƒ?ihJbÞ&XÀàíƒüÂ<÷Ö/׿›30ÆQ2@1®0„Ä4ìDú”hKZ`R°@œ…Mšp!«­XŸã!a«…ðá=‹é¡è†Wüˆe‹Ä' O>?]Üô Cv6-ÃpЕEIw'ë%ç~ð_ø’…éS ìçþØjõ‡ r -­oכʽ“sUý`††:GŸŠ¿Þuš„}¨1¤^›4ÑCÚœhZ‚KÎkc‹V”c µ$Ò8-ÿ0Ÿ.];iq*Ã;´9j¿‚<•oåQä‹ý´ª^×5®AÈe²D=™ûdÓÔäÄw‡wF¨Ñ¿n¢èa…%l¥>üÐÄSœßwSuHÛC¬!9 ÷ù°¹¤aéÔf¦$ZSâ«L×îÀ5'šÐ…/F8x˜:HÃ7Ya [i„·òmògå%çdnL¼šÄ“AìùÁŒŠ 3IØY™á¤Ï£=Âs¹i*ÀqbÌþèÈ…ÛH#Úe‚%l¢ÞNб¨ýÝo«ßNœ5h0-'\rzNc«—Ç+ò}¿â”LK°ÙËF0¯ÈÉ)K£l…Pt¿u†o²Â¶Òo?äÛâ!\-l׺9y]­½›Ÿ²ƒc¹EÃÒÎ4K°&c#˜¦ŽXaNÁôê8˜–*H#æe„%l¤Þ^¨Í Òäî¿ø¥m¹ä <´¡ÁK¼‰G†SeïôH +ìsêãQ{{µiâCÞ4mÈõ­2ÔŒOú×ÛK¤Éx»@·^¾/hþÞ Ù¨˜M¬&vµ~* »јJ#ÄlJÆ‚iªˆæpO>8z« Çd„%d£Ö¿\Wh—‰Z=N‚6cB¨FSðKSq'|º$a#÷ÄR`¢Gâ0=tаl%›ƒ¬SI„¹L°@Ú,€—jT5±)ë#9Á‡~<8¾³÷âc% ;Qcµ%¢zì‰`ºùÕ19ù P˜ÿ‘*Hÿúca )€·¿ºOQN$K?9ŸŽ4š*r² Ì@é_Ù¬ é–`£¬ôÁa!›3˜ˆ’q.úÔAÁ.+,a+ðöƒ¾E‰.~ø‹ ¹7çÄùn\{Z4±qÞÁó®h…ÕüQlµ?×™&2¤Å‰ AÃ2®Ž¼Ç4HƒësÛåÕoËéMxö…æÿCPŒKãÌÕª&² ñ[=OXÝ¿æûÍ¢õ‡„æ$‚é¥68›s.>’!òÛC‡hEùÑ2Ò/'ä×®ml#fN4}œŒLÈ¿”¨ÃLf¢iŸ8Î'LSVˆSQRºØ2U¸kä¶Á¶Ño/ÔÖ5·-6…PW»9“éí%†{5¼µû½c¤aèbÙƒ$zdæm“Ôðëdt<÷íå÷¨•‡Ò s˜` ›h„·rL=%ÙYíûÁa÷5JÑ+)ÌÐÜDÒ°³³¼$¢ËõHš*b…9Ó¦6BÁ¤¿­ƒ´â½<l£>¼ g M¯Q«Ë¡9§^¯Ê·í5p±CƒÑ3iŠÎ[¢œ|ÁBÒÔâ VV$B¤5ü o1LzrÿƒeÉá;DKÑ.Îq²ib”‰†¥l‡µá ÏIŸd±9£.6çØmSÃ2¶’H+…ðá‡|S£\ä¡~?2ˆ-8#Qfkt\C iZµ™)‰†Ç`*bA2Ø%Á‘ãY¨V!ZAFXb[M„·lï|Tõ¢®Ê¢MNã±^Kusi„gbІg E°)«¼ÑáT9){áÁñu MZ±å–°qFx›/—¸×Zb¯|Ñ%qæ`¼$Üâ§ÁŽÁ«3é_³ô–èÛ´LSG¬0§ŸŒT#Üœ¸g¤ÒŠuXa [i„·ô-f1 ³·÷Ý’£Ž–%hb Ž—qõA§¸ì‰—@œjY¢£ÛÆF0=1“£Ž–‰ Ž—©#{b¦–°•Fxû¡Nã‡j$xÔú~r.÷vž-à&ˆo¿š{rj?§G>FyH0c#$}h erÔè2ŽÆÆ®ÖAZá+,a+ðöC¾aÓ52­96ÃŒŽ!ˆ–†ïóFÐà¼Ï¤eõ9¿é`š*~ý±9ó<6BY|²OýøL¶yÝËvºÓ0µŒ#õðΜœ²ôƒ9ô¦Í¡˜•˜$üÁ\¸-0§²k`š*à9|JÎzª0©@oå!V àí…<È؛ç‘ÝÉ)¸‚—íÊÜÑ(lŒ¤ahÏÄ&HDRÞ@äáÝdÔ¨ÈFŒr9'Ã9?¤åm>$hâFx;AÇb# q¬.ã›Ã\ÄGüdY²[£sþhIÃNv4O‰s©ïªLêžœ(ºŠ«¦¢æ O.R!ZQÞ2Ñëß>Ð/žŸ¡š²ðjoNÕ/…'j8™Äh€©\Ò0“¬RBóÁ4uÄ s"mÌËc¨~0R ZÁÞÇCÂ&OÂèð Oà ;^…^‰3D¢ÇCLçóûã1§ÇÄçƒÐÅ>)0I ±ÀÍÆK€ŠžÌ[Euè4¶Ño/èÙ¬ljÏÚÜï‡Õ»ˆ’®¶2¸ÅéM’ðŒ}ORàšìonÓT×ÄQe{"°Ü7UT¨Ã ØF¼½g U-’\_&pf‰h˜È¡%ÆÁÖ(Hº«éØæ4Tï%@e¹^ª¨ê–6XÀ&rý‡ôk•xiQ»5x½0§|ZK¯Úz£¸8j@4ÌÔ(EIŒÒùl'„¤©#V˜SÔþÒѭ⩃´‚Üíª a+…ðá‡|SÉhìÐï§N^‰ÓqÕ¼ÅbpبiXÊj¸”¨h6ISG¬0§¢ÇÅF((%Ø:H+Ðǵ%ÒJ!|øAߨš»žÇ¥=äNëhA-eL›‹áãÇÅâVÒð)Q™Æ¦©¾‘Ó”ãk„˜ JߨC´‚cÊ%a+ðá‡|»T_y±MÎæTæ®E¼ÄöLüEbÛkD]I–2NR¢UU¥ ÁtѸ­‡¹k‰p_ñ×O¤÷°BÛn"|øßâ8?âfñ"b2*ÑqQ+ ³ë¸nƒ¢aæ¡ÚuIDZÂAôÒ%ÄäÂÖb.:ô8mµÑŠø9¶ØèRa#ŒÅV4ßÛÕ LâBTW'ž+HÂH>1ùóûþدÇrÓT Ì9XÞk€Ÿ»¶’Št Òç¶ÏËßÈ+ÖCEÉ4w“ $$JH¢j-)cŸ% 3‡["2ñž¦©#V˜SQ*º Šþ¶Ò tXa [i„·ô­¢K)`L/MN´‡› “8 9ÏCà /Ç¢áÛ…Ó«”h“²LS|§³.?ôå¥ÒŠuXa [i„·òyqV4š3:ï»Qz‰Þ •ÑÆÎY$a'ºùãɦC¹Ú4ñCÞÖ-çúŠÑ NI½>·u^ý¶Ÿ>ÅVÄú€º$Öhž†î gçOt5'uµÊ˜¸=n‰-R6€H*€<#$öòŽÑĉ`ú\æyõË|zÔ¹9°îŠÇÞäÜßiGäãσ™æh½ ØÇ üøì*E×jÓÄæÄå "œ‹$ @<‚úüÔÖyõÛ~ùt²×@ëºm%'6%LO>uŠ6&KDÃʆ7½”Xº‰ÁtsÞ|r*_‹!ê×CiÄ·¬°„­4ÂÛúoI'ûà)1ï’cò‹ˆ–q'wñDõ O)±&ÁôòEÛŒ“·Yœ“·U«¨ )a#ðòBŽa¨¶Vy&NÌøû~´ÛY¬V‰$éóXΤaèø¥>®ªâ'˜¦ŠX`Nk´S•µç©4“ –°x{AÏî7dš¯ªãØäDš7<ÁŽ'Š˜ã&4‘*C†b<Å–¸V¶³‚iêÀ qNVM%BdÉ?T€DTÈ}n½üí; ‹¦S“½O¸ðÜ4€ÆÒ†íéœÄSÞ7§Î(ó!ééœÄä ¶ßJ„ޤÃTÑ•“˜6ès›èåo'ä˜÷o»9Ìa¦Ï­¥³K!n#:ëÔÝÎõ6·Ä”ˆ‰¿çµ’vÃØÍa®ÐF`rÐÖAñ.+,a+…ðá}‹¦THýº”å–œx8ŠËÐÅ1æÑy. ?4ƒh(­hœ£îÏ#ù®ïåIS3Ã37@eÚPj¨N,J$±&‡rìR‚¥½ìä\ØnD¤HÂbUÚ<¼q;ô,5Ëàª%4.ÙISG¬0'úÿÆg" ¨ž©Ã´">j=,‘V áÃúvq/7Ž—>'šÃ&ËáѪøV*¨ª6Ìhâñàû×0]üV•œƒE)ˆŽ8xõ”Ñ ùcm‰´Q^pØi´‘gÛÂÛs2˜Ò±â¾ üû.옉†™göM€„fK%BÒN|MN)|%6)·Ò yöM„Œ4ÀË ùub£+Òø ’œÅl舔‚B]ÏH–Kf6îbZ ³_K˜¦ŠXaNT¦Æ7f„û—VARñ,` ðö‚žÝßÞDÞ>:®?8l#@)ƒ£&q!»µu% ×.ncZ¢O&Á4uÀ5qؼr#DBïSiÅ;¬°„­4ÂÛù†(Ô rò®÷¯d(T 6ò=RcÜ-NñIÃÐÁs~KD§õ@0 ±ÀŒø Æ¡øC…hE|é i„—t¬æz÷…ÎÀ›1Ø{-Bƒ·'{?Íû 4¤ag¼RlÁ6) ` /Fc7¯¿ÿÜÇSiÅ;L°„M4À˺mÔ¸×Èz13 ˆRª¬S¡FÒ°’OR"ú@”‚iªˆæ¶­J„‡[iE;¬°„Œ4ÀË ù…‚Ïýò/^ÍYÚ¶‹Âðx4½¸‡=žnDÂÌÆbC œlÈœ¦O7‡I:æl€Á ok ©Hgb§l¢Ö¿} _1[m™¿0s¢Ûëb” ¨r`æÀmÅÁÖVCS nÂaKœì’¦›&lŽŽƒ¡°¶,u·éM+,a+ðöC¾]w›qðb>U¥xÆ„ë!0O6Ç2@Ò‡r»“ó,ð-×ï™*L*ÚÑ_ß¶Ño/èY\ÆôÃ[ü‘™3˜¡‚bÉɉçÀî­i:¹ƒi‰[ÿxˆì.m4£±v=—W#Yi…;L°„M4ÂÛ 96Ù£®¡Çí÷ƒÓu0†¾Ðü™§WÜ[ ;+αRâDÙê0ÝÜ<9uê|K…©£¸&ÁFXÀ6àí=•3{ :ê|?8i‘rvîóÄÖ£˜hº°™k°ðܦ©#V˜£©ö‰p`¦üÖAZ+,a+ðöC¾-õš€”œèx–Mäacÿ,Èx¨&agç¦>ç{Ê^núpƒ sb÷,f ½+ÊV@ÑNü¹ìËåoèÕÄ ÆÖ®¢Ü_súÁŒó~Í85?‡íåÑãÄ4ÌÄrK\ê†bÑÒâ,îŸ!JøFÝ:D#*hEJØn#¼ý o‘fÛª=Ç¿œ¦´’“%¡‘÷ø-5ºÛðÄÌ47ÑÇ‘û3ç^ž4Ü+’Ã}ùP4s HÀ+é÷ǶÏËßÈ+xF5LçoÌœ(/8÷G¢-î‡1}íJV"M:%b°óÜIRˆS&{? êkÆCiĺŒ°„m‡ôŒÙ°Í}ññÜ^ùDý…‚¼x)ˆ#æ–4 ]Ø¿´Ä*jŠ"€¤©"V˜S™-‘…»©ƒ4\£°øðBž-B“£i7çV;}d™]DéI#ÇJ†vöo±Ddáµ4uÄ qZüڜ؂ßÈÁD逧–°•Føðƒ¾]˜åÛ¢ œ}K“s2ýi…3¥ô[™±]&–â}KÔsÿÚ׃¦¬g úi#tŒ*Ü:H#4d…$ÒJ!|øßbê8Î¥W—äÄ6|ÃøßÝw {ò v²É‰?o3b½Ü4ÄsØE|L ÕI á– Ðç¶ÏËßÈ«ù‹šŒ[Vý¡ƒ´VXÂVáí}›(jQÔxàaqs”Ë6 lõ‡FLØÆ5 KÙë$%.´]ÙIʯINWR®š7­£å©­°„­4ÂÛù†ÍW¼²™çæ(°Ï®ìÕG I/IÃÒŠJC \†Û$À¦•T˜Ó4PYM•­A´¢¾®‡„lL€·ôìÄßVN¥ô™Q•ºÇJUagœ#¬ûÙ†É7])’+Æ*®-{ìÜ. BÒTÇÄq“~#OÌåKýøL¶å§írg±»Àéñ‹ÉYºáF£!4ßàpEÌ6hIüÎÍ’¸ ï§FHúôøÅäL%a0 u 礖°•Bøðƒ¾Å`Cì9yì]r4·."‰1škwEs =ùîºßØ9 Ze©"’ÎÉwÉÑèºDÐh»Ô‘ÃïÒ I¤•Bøðƒ¾E‡Ì¸±ÔÊ}?8õâ&ƵØCº±Ìuœì=O–N~±Dtö\!iêˆÉ94C_RÇáÔ˜´BÛîC“2^~È7ÎðmqáÛ˜9ãÈxÁxàÌ#FzàmèHäº_XÎýq$NÕÇjÓÄysZaò½Ö×ÎM]ãWf¤’HðáÜŠ*¼+ZÍveš7Œ“xçŽB!¼™†Wx£Ý{½¨îAf´žÇbX]Oö®7~u ´À²Ï/ûåÊÐëö¢Kbø0-Z ¡s±ÏÞ¸…+V¢{ç–èxdܦ¡"ˆÎ<0"ú Üf@_Æž4 =”k%‰®f(F0MX!N>|m€©“3« ­8‡–°‘x{AÏêÁm¤›Ç¶8ɹÔù?Z ñúÁƒKNš 3'«]-1:w `š*b…9Q€„?…n»çSiÅúxØF¼½gª½–¶¸Í‰ËRS¨àl¤tïôÆÓ*Ò0”;)1YÁ˜¢¥#V˜³X±h„x>më­ˆÇtbKØn#¼ý o ÂüÀÍàäÄq cm oð4>š†kK%’˜nÎB“ÔÏĉRe1,~ÀÁºUVÄsΕ$l£Þ^ȳ8ÀUL+5ƒ0*Cý1hê~¢fòóÁܹgU–“LSE¬0G u¡¸/¨tVÈ3?S2Ò/'èW|Äaë£ï§°n7"‰‘l~„¡³8¬Ê†Kq¸~µ‡Ä¹ÔA¦³á’9ÇÅóA#Ä(±+8ìäÐ =GEóëñqÕ«=3‡åŸÈ‰9јö’g'ÓqHÃГS¡ÉAbÎL€$©!äŘÊ÷òÁáW©`x½g—†®‡—OÓØ­+2¥Î6˜Š¾tǦa§:žX"r0Ï‚hªˆf,îé ’YæØ*DÃ3!ζš^„c,EK‚ÅŒˆÍ)ºÖs¬ªÛ}­IÀJÍÌÛÒѲ–›¦¬ 'ÎÑ9뀑ÛrZõçDz/—¿= W…âñ¬Á)3fç0D¥0Et3ƒo©®µðà€³!ÚC¢£{ˆþ²DßOEó˜ µr²7w¬·–°…Fx¹ ¯Níu´OØŒ©g·h6ô…Ù¹Êt×ÞÙb÷4bGa­%ÆÁI˜††X`F׳ŸÚ·Š¦”÷´Á6Ñ/èV¤/` ›Ú¯&'ºúÒ‘¯Ý•S0¢gUž{„! Hl²Ä}=>Ÿ¢‹‡mÎò@ÍKík5>S:Š¥)a»ðöC¾]<-Y‹•›s±n7ª°.Õï^ø ó×.–V^I`VmÀ4UÄsNþ †×Í­€$üš9Üoy2Ыß.ЭÈð‹'ÿhnót‰ê|?í£Ëá»cUW{%a¼i½w@b2ŒDSˆ±ªþ ¾±²àÔ*D#$hDJÈf<] S½°çèÔIíæœ“5þýâ‘Õø>ésIÂÊ“C^,ræÀôÔ!êæŒ“O•Fˆ³è­ap è¶@Û@¯~» ·N¶]“T›sqRgÄ Ä«¾ŒžcWõuMEæ’XÈåÞ¦©#V˜£W/#DéNl¥C´Â9š–°•FxûAߦ·˜ûwñ’hμøŸ1õCq•Î-i¸vñZc‰59!Ö¦©®™sèËBoüꬣ«ñua´Zo/ä/2Õ5:’ÝùãE™ù?cõL†ÆÐ·-pñÚ˜Iþɘ³Øº^58© ªF'-ðç6Ыß.Э©"ѳøZoÎ:²g ß´OíGT^ÙIÃÊ’ W¸–ê…`š:°BœYu®,„Á†û©ƒ´VXÂVáí};K’” ßÌé¸G  ý¨êw³pa' ;Ov³—ÚTÎ\Ÿ$Ü ’s?#_ßåRÃj ©h/çCÀàíƒü:Á×~rz|2‡D”$sh@ìa4i˜Ù°ga‰Â”e.'Aðe.›NƼprz@B“VŒC¹%d›Þ¶Ó…™½x©ž|â5giðul»ólW9âþ Þ\J¹âç…Y¼:éSó³6g"ùcÜWm&FHi…øC À‡ rëâÏšlå¹9H4Dl ò¨~Z¢^ÚðDÑO±9ÊiH„©£ëpoèm…$ÒJ!|øßây5E‡ê›S‡þ†¯¦S†vI}& KOÕ¹J"½ëÁtÑ÷æ L!¸zÓ:L+æõÛJ#¼ýoÜek) 6'¶ƒ/ ßûÚ£úH–6Þ•-Ñq'M‘ÔòbL%)y9:‡}d´C9>³e^ø¶þýÀÞ¾œ©w…Ê©¾‘j€¤²˜µv% .f.Y¢óÝ!LS<'æÃ#uN÷›W}ê ­Xoó!a+ðöC¾]j¯tôÍa29b„™`lË)cÓ°tàœÑ‘úz ˜vªxr&“'ÂÁdóÔq8=­H Y™o?è[Y…Žv›Ã$;„ ÎBâ('ÞTwhcÑ£||ÉRÏ…¦ qs¾òxyGž@‚ƒT€Cµ>O[µüm;ý‰¬ò©-›äœ¬XŒéz a;¹…ë²h˜yêA^÷¥>LSG¬0'Zp ½FýÒOQ:H+Öa…%l¥Þ~È7î°Æù–œxm½#½0š0v'žÐ6 K›F¼Hâ, #ÓÍ/aÉ©|J„zHWßçoűÞËCÂVáí}Cžë59£ŠdìOÐ ˆßÏeÆæÁç0Òð M¶Ä©~, )xøDÆ:yð§åñV¿ D+ÐýÞ ˜ëåÛåÌŶÑÞBþˆsb %Æ”è2'÷—†‹ßŠØÑ$%–îŸ0M±Âœ¡!>F|ãHÃOþ6¶Ño/èÙ˜*ƒœq¬ùýàÔÁ‹ËlL@ilÙ‚ÉP+iZ4‚SÑ­=LSVˆyQØmBd*¶‡Ò sfkIÂVáí}»__عÇ™ßÎ`2 Zë(Y?ÚÈ–«IÃÒg6)›Éë`º_´Íi—y²ux(ʶ¶Òo?äFô¢ànñfNQ¬°pèÐ6BgZi˜Ù˜¥*‰z(·³< ¢ßâæœ<þÎågŒÔ­@´Í1,!ûáí½ºñ‘¦>P‘4'(#L¦F àžš«ªÖù7çRžqÕì´ó—¾L6vRMºNM[Ñêrñ7i|Ò õY2ÐOûéQÌMëŒÞ“ùëÈmœ…yqz¹hØ8Øæ@qIôì×óIÈñÛ 6èÛë[ãæ)í\o’ /αróv€H¤2Á¶ÐoèWÇZ¼1ðàÈŒSǯ‡ Å7Ò±_Jô¸—b§%N¿n ÁôTWÍÑ“s"Œƒ…ÂÖ14e[a i€—òk²Ϲ8csÖ©F…‰—ž#Y.I˜Y™»`û…ð|˜¦ ¬g^šŒ-„EøPAÁ.,` ðö‚žÅÃè@iixpÜN¸Åù™:Kh×ùX&a&SrýùbþO.OúP.CrFa7`ôöËx( ·d€>·}^þö@^-6þ^ê¶9WÓYúס'ØmÁ`†C]ÃÐË j,i2ë±Üôê~|6çÚÏ'Btª¢êmŸ—¿= WS=YJejìæD+.„Hag©º 1p¤:Ý"Ž ÛCà:™ÛdÓTâÜß@´lekx  )a ðö‚ž˜Ò{{»8Êts.µ]ˆ^C@JB>Ù‰4 UŸI´Ðwn„¤©ã^‘œóâ¥QÑFÖ‡¥µ$­X?êCÂVáí‡|›lŸ6‡Þ “s²m4§žìhÂ^ (( C9HÏQ¾•Hz*»{s"çktî£ZIÅ;L°€MäúèWƒ×¦Ê ~›SŠš‘s–wTäñï{±J­øÞ‰Âȶ%ZjŽE„¤JHN¤bÞ¢kS‡hE<Þÿ-!+ðá‡|ÓÄßÞ”`”œÁ®3+HÌ(,úš$ìì,qÐç•[,^ž4ÄsX²P²€T¸ÉØöiù‡ôêêj6pü›X2.U ð¾VÞ2ï'@½EžÚ—h•Ù)QOõHBÒJ­IÆA(Ä 3[¥B´‚G3’H#…ðö޵ãP¿¨KÆätLB˜ L†-Ü- CÑì$[£$€iªˆÉ9ØÇCÑW²<4V°—õH«ðöBžá÷ƒ1”0WÑ7„é7­ŠäÀžIÃ̊ǵ”è˜(½LSE¬0‡6Âú ¥ƒ´‚VXBFàåý*Ly:9äûÁÑ ‘”øïÉ7}ÅSÙéy·Y Š)ÐÙ%LS#grO1æšš "íµ<dc¼½gÜ"Î"ÏĉF§ŒTvn9 ±ÊJvrë>%î‹F¼êÀ$5ÄsŠnu(¬K¤í®ò„„m4ÂÛ zÝCP­¯69›s5UÔWÝ P<¯œH*ÜÛù°‰ZÿöA~-xN ÚœëÐ5L¾Xkˆü¥ž4Ìì*zÄb?gˆ¤†ãÔ㥗Oõ­³Ò u˜` ›h„·tl°ýfÔœ0•*9…—Ù“Î?fÕœE“§R©Ú:ØÀkéR¦©+–Ú^±…-ÚC‡hE{ÈÆx{AÏ"Ï¿jž™³¥ë׆óÌs]<îr!,q?å Aʦ©#V˜)È”Â}AÄÅÁ:H+Þa…%l¥Þ~È·É¡eúúaGR"Zpn/#±sS“†¥¼MZ"Ú¹ðA‚ISǽ"9 >!žîχÑ ùÞ¶Òo?èÛI{¢ŸÞ…§üä°(áâ"øá\9o„ºkqˈŸG»ÆöXž4À1qXS”qÆR¯­A´b¾Ÿ ›h„·rl1ñzåLNgí] NJFçW ¥tu½söƒMO,Ñc˜û0I !/F]jÙ£åµè”ÒŠø<¬ ›(„'èXÜøðƒ¯ñý?8Ñ(KwBôP`~_Ü q…& ;µãDÜ®7@’Ô€âDÖìõ•Ñ#çš[…hÅ;†K"m‡ôìÒßÖ´’œ®a`×bÛ‚ÁN“Ñ-§†Í£S*ZX¢vvÙBÒÔ+Ì©]ít…PN°tÊ¥’H+…ðá‡|ã8ߨÒçNAr: 1"X†ª²&¾“¦ê¡+¯jÿc‰8{˜lšŽº9õR¿WTاŠZóš##$‘6 áà xÖ¦œÜ|?8'Ow"Z°­4ú(¦fcÒCÉãQNƒXK´É%#˜¦ø&ŽÒ衟Y° Ýgûi…%l¥Þ~È7¾ùvVD}oΈ¾ …Ñoâ£`ðJßW’0ヷ@Çxá Z*b…9Kß+ú:Øs—D*äÑÛ6Zëß>Я‚¼ëQÑëå{3Ú©˜â>;üû³[4ŒD=õ–ˆûç@$@žŒ(¡_{9GEnÅi‚%d¡^ЩŠvÜ÷_öâ̪ÍÑ›\I¼V¦ëEÏ€Õ’†•h{²%³„À4UÄŠä0]Ï}ÈRGw§2‘iµÞ^È3¦õtNÿý~rXÒA‚[ΚlZݾtÂÜGã-KŸO&"åò¤}ÀœœÉÞÆ 0;Yi:L°„M4ÂÛ :Ö8ËO.ðËŒe5‡š'p§&¶¢®š4<»Ôu›" ;ÕòÄwüÍë`*b½Ñ.4–e¥ Òua i„—t,füàávê}39…Ͳã:ã.÷àGÚ4<»²Ë=$®©G/!˜¦¸FN„*®ŽBˆßäñÐ!aqd—{HÈÊDxû!ß.Þ šÙQºÕOæyQ\Òу4³1úÅž'–GU³T$É_K’U7$¯.ƒó Œ_4òv[` h„‡ýtH…ÒñˆÅ’Óä°Ïî­¤â¢ÛëkÀÅ6*$a$_ôùˆâî¹—'MX@Nì48­7¬…]4+ ‰X þ\öyù‡ôj1O-ž¿Ÿ¢Ì©L'0p˜`Kù•4Ì<¹£*‰xJ<IMTÝœƒW#Dó |)Ò!®ÑŠ”°•BøðC¾ñ9.&^_¼p˜s6þz/»F_âx¨Šà=iXÚ8›Óõ`ÿ-!$M±ÂœèTŒ‡=!tLÊÛ:H+Ôa…$ÒJ!|øAß®¦<í¡ê±ä\êé>È­ žr0‰4|»´‡!‰H9z"˜^ÎDLÎén”Dˆ„ô£l¢´Âi¥>üo—@]ºÞ›q±­ãPYÝÁ¹<·ÚÉ:{Ïé¹ ê:%‰¦ŽºF0­I@›q¢ŽØá®§ ÓŠzöD¡D)„·plƒãú¥ÓäÜwE¼‘•Æ·îÉ!+£V5° vZ°­i¶´U ÓTkÏœÀágt§Ò y¾ùKÂ6àí=‹œ/ôE>ñáC`^|˜4€é©n›3PlǸ5¸aã¶Á¶Ño/èYçßzUŽ-LŽæˆ N8isDn½K“H.¥àyfB$N¬A´tÄ s‹Rõ¡C´Â½Ö‡„í6ÂÛùÆ£ëOÐtÍŒK__¿£¤‘0;+ IÃPî¤Ä*ª$BÒŸ“1Y’Ÿ£ëK” Ò xa i„—tl¸àÁÁ…GŸE¬`ØHáV帣qMñfæ8/õ‚•ÄbÔDHúð×L:(â‰åô=V×Rê0­G«jK u4ÂÛù¦¡¾Ñ€ž?5sbûçb° Ë_ãÌÓO3+iX:ثЗû- Á4uÄ s¦Ü'ÂQÒÒAZA+,a+ðöƒ¾EâŸR‹™P›΋pÁ1IŒÚ ÔUÙí€4,-š-‰ûI‡Í…`š:°BœÈo‰˜5BGö~ª©˜o}n½üí‹FŒ(¼9~á£U2.¾cœ‚-ÜJþj¢a%º¡X`C=³°éc´6còÖë+fRÁpMMš` ›h€—rëäÀäÐ}oFCÛ?D *;›¤ŒûÙ;HÃÊÆBIÌRÔ£ŽIW'î%§+þSiE;»qHBF àíýZš ¼ºÞ¦“£ ¢ˆÔŽèY1#©ÎÜb¶˜?.§FquÒË·Üä({ÈQܼŸ”‚i›ü8­Óêûå¯ÈÚ|?8±?!l¶Ö°»4ãüd% +ÕÅ•[oFHš:b…9ʆ7B´ÍY¤ç°Bi¥>ü o‡&F÷Ož¶ˆƒ6¾È*Œ]…Ϋw¦ÌØÖhIÃRvØH‰8¹]ÒÖâ0yÞh<·ÓŠöÞ·DÚ-„?àÛ<4ÙwM%çâƒ\ÄJ‡¥|럑æ|& KÏ_žm¨%®LSE,0'vLñEŸè';•Ò`Zߟ¶Ño/äÙbe¿2’×ä_»ðå úòðÛèly@&ªŠ%:ß`Á4Ä sNÎÝ6B\Ð"¬C´â}´‡Lôò‡ùô§tsLe›&Gý7"BxR”ÕÙ°“-NéØ÷³M{˜¦ 8%Ng¦DP«h~•M,` ðö‚žÕ#J˜“iNìpáŠVO&Ù”¡›ÚUŽÁ˜°x>$›§$‚è–ùeÉYìîi„8+9:D+Î}` Ûm„·òmf–#3i7çú¥1L¦þ>øJj&3‘vŽ¢)0’¼‘z}’‡i“SyD›S· Ò öù°…xû@¿î;(F¦Ö®Íä4¶sŒP™ØëcMïŒQªgÒ0ô侦%bÉÁtÕ–Í)l™÷·<*@*àaƒ>·‰^þvBŽñ÷îKމÃE ÌÆ-ÉxC˜ƒ¹¢a&/)ÑçÁ4uÄ sîwãiéBh¿d©©pŸmn½üíëè[ÕëÈœ®aR(<.‰ê<4“™¿–˜®xù³“õ^¹P´ áÏ™ïq5ÓòH_>78I9Ó%ù¹mõò·íô'ú ]|åÄ&Ží`2u²ènŸÏÈBìIÃʉgê”X(ÎK’Tâ¢óL¼:ŠââÂn|Ñ pô¶·„M6ÂËy5™ûÒn‰SØ„'B)’…©eQ àîol8cRÏzH¬‹ý¸Œ`š:°âR7-N×5B´Pœkë­ÇtKÈÊDxûAß&gùFýS¨’3q°Q‚F“ç^3²çZÒ0t±ó°%®¡Ž§09\j•œÎ£­à$ح³b·–°Fx{!Ï»ÔKyÜÉiLNˆ’H¼kw>‘FI¤{:3‘{^™’8ÝÖ^IWŸ m;q'Ba:Cê €—–H»…ðöƒ¾a{éÉìN• &©FEŒ÷XÅs°S iØy(ç¡YBØ´+;’s Óy"¬«ñ½M:D#04cÔ2Ro'è×:xÊQ÷J“ѸM1×É7íÞU0s&§Òÿºj~ßëµi„¹,DZH€èÖ¥zòtÂ’œÊûÿ¼X ØøxpÆ„ˆ–4¬\9•—-$M±ÂmÈ%ÂÁƒçÔAž©Š’i£>¼gKÝì»_SÌ)M® ¦EVž3œE[·Õ'gUóK4”+lÓ‡»@šÍ·Â ³Qß*D#Ôi„%l¤>¼€gç¡n™÷õ•ÙÛÉY<¨9Kaƒ‹oAgeš™hØÉÞ()ÑÎlgÓÔâLöJ¥º¤Ò Ya [i„·ô­059š¡0o 9ƒÝj΂Ä$Tý‡þ³"3J$ e”èµ0M±Âœ ûµ¢ dßHÂ1™`›¨õoä×T‹htsø~pX‘r¨_.­©©Q×ëÙ+&X¢_¨yÒzQć8**ÍÕ‡'>iûQ6ÐoèVe=Oìè³)9'Cï0á7Œ{MŒa»’†• ûå)1Ô,ÅËMO×"%gPQ"t–¥Žîb$šàmŸ—¿=W,…Ÿýà—•œ ?‚‘ß«GÎÈškIÃLu?±Äd}P"$í”ä覔‡Ê§ÒŠtXa [i„·ô-†Nžl²õqrËv#RÐMèbvVôÑÀùiXÊtÜ”˜l¿™¦©+Ä™ÜBH„Á-„ÔAZÑ+,a+ðöƒ¾õƒ½£º»æ$g s)b—ßÉÎVç8Õ~±+g朾GJâd–Lv·ÍINc×›h슓*šû椖°Fx{!Ï8Ç7¶VyfO@#ZÜnz"¾óIÂÐÊÓ~ œ" éCÏ÷â`ö]¼! g^üšÛ/›TÌã8&#ij ’Þ^гQU÷sê¹*9ÙâèÛ†uEãt:è.EŠÁKØFµÄj• ÓÔ+Ì)È7O€8¡xª ­‡–°‘x{!ÏžrptPéÙÒ³ #Vx Â*Öh×áÎ\z¼_½ë-#Þ®‚héˆæ°x9bŒ.¨Ò!Z!Ï&Ä’°ÝFxûAß")(¸…?8 ,l2]øL·ØÃQ4,=xŽe‰HÎ{"˜¦¬gÒ‡DØ$OCg-iƒ>·‰^þv‚Žê"¿òÒŸ=µœî+Ï”ýh+Ê®÷n8z^ìno‰5v<6íKrNzÌÌÛ:20­°„­4ÂÛù6Õ}ºª„,9W‘žÎ÷ï‹¥çý•ã°êò¸˜uT6­“ÄŠ)Ðs#$M±Â¥ $ÂÄ+ÚÖAZAÏ IØJ!|øAßîÃðCà’aOÔ&÷`™gŒ1C©+ix¶ÔcŸë¾N–§£"Ý#ߌÅÇ^/WëúT°›äÛ„”…x{ §–ušÕ“39ý2¥ñ‰:NÌW´6º’†•,#H õq2@ÒT+Ìé˜~¹k²RiÅ{Ûi£>¼ g—7]‹Ž4“Ó™ aÂÓ‘“?šÂýPÑ0ôྦ%*‡$‚鿿ɩ V#^éRGqs‰´Bi¥>ü€o+Z‚ÅzÃöûÁé¨N‹PÁ™úà`¢(•5i:™ál‰ûauO0Ù|sONeJoTæÜ§ Ò wa Ûh„·òŒ3¡"í‚3­Ì™ `i:<^ Ûh¢a):î§@|Yõ Z*b9‹Ï'@nGßD+äyü, [m€·ô¬p6ްõ‰ÇÜ‹¡Wà8c‹·u?Æ#‡4[Üç³Äý°S÷zR‡[dpŽg®F’Õ_´Â½×‡„M6ÂÛ¹µX€8þms¤"PPO9™ú£Yæ<€hY KÜ:(0BÒÎ INgÿDhLhK¤ï}>$l¥Þ~з¨ì‰ëõåsÐ}ùd¤à= Zt‘F“·žŠ2h7„…cÌáÎå¢wç©ä,æŸ šþÇ»Š5ˆV´³¯ƒ$l´ÞNбV8vx ed&çê0´±¯JÌÿý•À•4 š#‰HÍëÀärƒ¦äœÈ/õ]žRkÆuê·?ÙMá`1eT æðœÚƒ×C".ÊOÓËsj“srûÎñuœ¢íc>$d¤^NÐ¯Í dëðñCä Å‹V„ ºNØ‚t3 +5ºUŸ/ÞMs9i+Àq”0*ü^ðìA ¦ê8¨±„6ÂÛ :6 {H—¦öäÔÆo}jŒqã.ÊQm+iX:™€e‰K“’`š:b…9{7 s±[ª©Pç¾?·‰^þvBŽñ|+6Iø3£óÕhq(Î:.ü}.¢aeåN»$¢Õ~_!éÆñõ›qß‹ þ¨¼Ì§ ÒŠva i„—tìähߨf%Hr& ¾"RâòÃõ" **Y¸—¶ü*‚z,}|l›âåISü'z)D°byc±X¢“V¨C¿%lŸ—¿=W›OWÕA›=Î ˜7×9 >¡¬‘4lä6Ÿ%¢ÊÈõ¦âbT4íØ« $¿L?­ÙKÈ>¼í§Kq#ÃfkñOË5#øàó.¨q>z™‚E]Úùñý4À×ê¤ÏÂ6§¨©¢¦­^©@¤"œ‡8°øpn]dÐp¿ø~p:ßÛÖÅ~;±U1éÓ`HÃPöüI‰Ú5fI7wcÚ¶L„¸ÆÌ‡ÒðMVHbÛM„?äÛ©Þ¡5.,ß›£¿@‹Ç<1²(÷Š+NK†ªñ‰%ÚÁ-uˆ–ŠXaŽ”Á Ê­C4bFX ­À‡ð,:¡3ùDóåïçä8„8ŽGÆL$ÇwRX>(®±[HJ4v6MÓÔ×ÄK® a _~ëNÐO+,a+ðöC¾iŒeÁ¾ã÷ƒSYƒZpám˜Aߢ=ÞQ’†¥l°•½jHžLSG¬0çÀh—p°6 u†o²Â¶Òo?è[¿èÑñhZÝi¢µP4œ^ú}QUÞè!ãWŒ^º‘–W¦=f<³2€ à» Uä}2°„l4ÀÓ:U ›G_Mo™æ`¯/~Þu±¤²p&É…=i˜‰c²-1:›fA´tÄ sb·ð+°9øP!~ш”°Ùx{!ÏÎgûéï§rêpP±3x÷M¯MÃNÜ·ÄD>ÐF0M¯“£ùC‰ÀÔ[G6½N+,a+ðöƒ¾5VSjæá÷“ÃÓ«@·ÉkP2µ»xKb$1'&ôá ‚9 ß„Ðÿ—*šzû§ úÜ&zùÛ 9Æ!Ë1Q†i~ɉÑFqê“oÝq{3¯¼¢a&G?[àÔÕ¦©"˜S¿ÃPÄîq<;Yƒh8FRÂ6àí=‹ºÛ/ÔŠ°¶Å ò‹@ÁùÈŧ¤˜š=MÂÈÂêu}yå±Ü4ñ±@œÉnÀ 0–bW†[æ§ –…xy@§î[êU+}?8Ê0‰Áª†¶“-¶j’0óä? Üa“À¦«;]&§°©z"lªn$é°Á¶Ño/䙿ûÖü¾Ì¹¯¸ýÅa6ZÎê é,*ìþ²µóóHÍ{¬YóëCÏ´¹úÐwaø#¿-ë·„í3ÂÛz5¹áÉAÞÉ©Ì̈(Á¯ 'Ð×âv]ºTÁŽþþ߇t :*»-4|*6% ‡.[I¢žjd/ÓЯNuàÁþbDD¬TˆVˆo4Q>È-õ¾zÞ5§âeñŽEuIëÌö$Ò­çC¢q"t"˜¦ŽXaν¸ _HbH¢ål¬Ä4‡4Ÿ^„gmm´vpP탃ç4 I:†wÞfT·›¼§Üv²¸?%"Çþ| $­éâ›Ã,ÛPðp·uVœÃ KØJ#¼ý o·Ÿ8ÒXjb·9*]+qÉô¤ ßÁDÃRÎO‰>8ÀƦW÷÷lÎÙùÃjº£Nô—²Ó ùz>$l¥Þ~È7Žöç1ÈzµiâÇsî{¬4ÀP YÃÈP”°x»@·Ú`‹ø¢¿Í©8D˜0A|=j}ð0ÖTâ×£àè|LŽœJÓEx›s E=¢ý¡à`ß¶@ŸÛ@¯~»@·zQhç~ÛæÌÆKNgQn µÇ%é~Âø4BCã­£©Ƕ¶Òo?äÛɇ›xt“oçýÛ)˜è÷ü};FJ‘„¡-§ C`ñÜ>LSE¬'ºÖãö%„ØäÄ]‚*D*ÜûzÈÆx{AÏX‰7¹¿!²`o1Â=—ÓŽBñ‘4œâ 2%ÖÉQ·Ix¸tò•‡_¹<²þÆChÅ9î/–`™ –?l§3Só|ÛÁì½犭„G(è'ŸjÏÉáÐýô{ އ¶Ä…v( éÃ7ss*Îv6BAGŸ­£¨çOaÛh€·òŒscJ;Yï¼9=CYjøÓWÍ`m¿ÈL68ñçwL ÖËMSˆSQ²¸âñë©´B&XÂ&áí;«æå^wº9÷«, &q²­Ä­—osÕ'z$‘^eK”ûѦ=’¦Ž{Åæ „b#t$“m]éfÛ K¤ÝBxû!ß8Ñ7k$œ‹×Ñ…‹+ zLF>7 K9&Ë(¢­aÓJbÝV8lV@l®‘ØVXÂV áÃú¶:Úº²Á6gLýM‹Ò*‘ìÕãV‚+óT:X/jF’åü¥o“]ÉZÉhŸ°—7d{mMù`ÛI¤‰Bøp‚ŽÅ¨ÃÎ[kg@šÓðƒ ð{\p]& C7O-Q›†Æ!éæ§õäT´l„‚® [GQ_†m…$ÒJ!|ø!ß8N9®0zò0ç¾ò‹Á‚$ý…·U]DThÑãÜ5–¨<ÊIÓS‚›3Фm#ļª‡ŠÎ>‚Û~ž&jù‡p,Η0Ö÷,œÀ»9Q¦ª`Á“(k ¥K’̃ï¸F•ýy|ÐórÓT Ì™•Ñg€½­T0¸õµ Ðç¶ÏËßÈ«ÅC>81ys&+!LN3Žwýøò±gKfvn_Zâ~ºõ`š:b…9 “þ6BTJ=uV¸Ã KØJ#¼ý o…­+Íñ3Ûœ /q&xö>p0ÓQ T’†¥‡fÂHâþsfŒ’>ذ~sV¡¥Bˆ¤”òÐ!ZáŽaW–°•FxûAßêñ¨PÿÞŒ8] ÇÐ%üÖ¢MÃÌÉ´%KŒÆ4'˜,:Çpp°¾Xÿ¼U†_2¶Y/äÕdÉA;õœ®ûn´úB«Ó‚??7oEÃLŽüM‰Á›j˜n~KN]¼E¡ø™ZE=ô3– úØzõÛºÕ¸ÝZ®ÉTÒäÔÈâ¯.%Æ@WŒ ¤Ð’†‘‹;™–˜ìŽb‘Ò ÌY'¯™ˆ¾’³n¢áH [m„·òl±UüÌ84çdÅ6Cl,/‡Ù`÷ù•ÓÇC"v:Ö093“£82À@JÊV1”´²°DZ-„·ô¬ó„)ÆJ¡ßÅæDùâ¤h|Êló'–œŽ#“çb `š*°@œ ý•ë´ãØ H#*d‚%l¢Ö¿} _YÖ1u®ðš(F~A§¦—4>LM¾4ö´G‚ ævZ"¤>DžšHšŒ¹xðòÉ-ÖT0Õ¸b›` Yh€—rjr–øâsk=µ¥¦$ S[ßq‚°›>˜6¿9,×ÞµñÁÌ:H+è¹5 [)„?äFûNÿÞŒ±x½º*×±{o³l¾¡Ýõ¸=GÙ€?/—öù¸ÐJz …×PCƒáøLÕK‰©»(–› zˆ›S/Æ—Wäím¤à0Á¶ÏoèUó¦kaFØæDúwc|4á"<ØR4 =˜Íd‰É¹~ `úP©MrpyŽ'!De%.§Ò!ZaÞ²1Þ^г~dûQž¹$çÄH,Ä®ë z´«A³QÒ0trdŸ¸×²LOÍÃÝœHpþÊõ§Ò[A×ÄÝm‚%l¢Ö¿}_lõCóF£_âÌS×€=Û« ¶Ø¿V4ÌÔt2K,¶(IÓÔ+̹ŸþBèx™Ü:H+Ü[}HØJ#¼ý oƒ[®73îŽßŽÿ ‡^º‘ùÞã$e$ ÏS™,p?Üð=ŸëMR§M>m bÎZj ©p‡ °…xû ¿û„žê”¾9«ò.;9ö7Î/ü};®ËOGËMZ$q!(DžÍ?K1fçÅÀËï«$n±V@ZÁ,aðv‚ŽMŒ©Ñ­YŠ›Ãw 3õyPZ¹?i†¢gÙ–ˆy\ý`š:°Bœ‰ü®ÍU*@*ÜÛµ?·‰^þv‚ŽGüÄkŒ9¼.šo]åct ”"çL$ŒœØÈÔç-š>–'M÷‚ä4L™Ù º¶†¦^^ÛKØD#¼ch‘R1ppÑ1s.ì¼£¦hÝtãÎ}Ä®hXíNR ÝúÈz2À¦Õ{s ÆÐ$À‹ÆÖp¨KѶÁ¶Q^гès:òýR¤^ ïéçãË»ð¸d>!z?‘F]6¦ýr™œÆÄµDЫcêÈ—Ë´Â0QËŸæËŸ‹EÝ÷kÑÁ ½9 ‘‚q*Ð1¥¾Ç«Á™$LììþjŽ?O€¤©"V˜ç|_ :ެ­¤¢&H Mäúè×5ÔapL9,"ft ôëâsL+Ü» +la¦Dì£÷ @Ò°@B0@E‘ÄVaZ1ŽòxI¤ÕBøðžEƒìCM,&¾±ä ©@Œ`¢TÁ‹ælJziw6rt×|H4öHÓÔ+ĉëˆ*#ÄŽÔun¢ëHQ±„¬L„·áÛŸÿøo?JäÙÞוui ±pEŠó÷›ÿ§?|ý‡¯¿ý-‡Æ}ý§åëÏîÿÿÍãëßý¸aW¢Fst4øë±³¾Vrþx«ùçRɉ!ÕB5Zé\å!”œ-„ÖøS(9!õ.|™³…>í|X¾…ޝ¿Œ¿ÐñÿÓ~ýþú·¿ûñoþ}¹¯Tÿ_q×ÒcÇq³V¼Îzï$¡ë]å #ðÂA,l/FâPbD‘´(ÙÒ¿Ïù¾ïTßêË;CÑIÝ3ÕÝõ8ï:Sü¶`„õì…íÇÜ0 bËÌ«Ðòͳo?ùÃi»½kŒÍ«§°Ù5³ÀÙ)î”pJëÙþøÓ³ÿä·Ïìÿ¸j4ô…qBØ{²6íò•QA•¤rÕ¥°ŒrÈ:ŠÞ+–&ä0Êg±ŒZæõ37›w‡]¤›gP4íà]°Í‘ó3îŠýFåò±µLÏ«ãT®í˜N×CnDO2œ~r^ý‹lBv9Xhí©f¨•S5­ÏÞ½}óúÝÃÍ_nYµi«§ûï^Þq‹4—0N¯nî_?¿yñòûïžû ÃéþÕï8Ý÷öe.#ö FqÏžÛ‡~÷úùÃmâ'ÊéG>yÞ0[ÛFF ×3gvúéöÙi+HãÚŽÇHüÓnÜ åZâ k}N’jF†Ë—Db.E|X)د RƒzÞޕµ,$o| –<ˆ<êO•gCºLÖcdì·J] €¡j`èÚ¢¾Ôè²üŽhË£V ô§¢ºM1þi"Œþ!ºëðÊ›šŒÒÔϨÑþ¯¼1¥eEˆª¹Ã/ÔôTf3h¥›ESÐȸêSE ¦QÁ2€@Ye$,Ëï2öBèCcì£jôºAsRoØÈ?Õ›”øÀX@sÚ#Jo Íx0<¢²ö.â›Ãïä;ß‚×íÙ4 twưª QL HÚ¼râÜO\þ³ì)|£äsJ¢¦<‡[Onè=¦I[ª2ÈP\Ö¿]u›Žk }ºE…°f‡&dO+ÒÖÁ³¼‚ƒ¯ Åü’o¸U)y½øaC>fo·Ôâ)©LH`30î]ä§(-=˜ Aªä1úî”9ƒÐ Q¸0”J Yi…¸WpìÁìÁ±èÓ¥î-^›Þ[£—# CBº+ Í. 6ç(4„éŠ_KË1ÂÂh)³¨oàwÊY¯1­óÞšV‹A`ßtw*ˆ`ðfSmo ©~0Ñ ©bñ6ýéjmˆ©?Ã{*ozÄ´]dRo¹ûôà^ٿ˶M€  ¡\šE!-Ž™ìVH-ÒQ¤žDÞR²nMÑó:™Wmë 3¡ŽÚ_ö@âô#H¡0ç6BÀjyÑš“øYdóèdøéT–çURyûæì ­ÕYÇh°¶ Éýhuò•œT!0¨Ý 9í±ÔIÏôx0„E€¢æÎà»Úጞ¯ô|†O Œ« Bs€à£Óœ¾ 5<1,ã@/ò!!pK“AΗAW!® Ãd⢊R‹ÒýËdÐPFåßBüdŠ™;ôP jñYØÄ„¦¼Ž‘K’R×eœØPëHO¹ž‰BVB­R‚"˜*›ÊÂ)v«ÇŸ¸¾g ,fÏÒX¶1¢Ì¨˜úxnÀá-ûf¡bð;U_eïŠ E&Mf¨5+®Ýµ7hX<ì]´[…ÎÊÑâŽèg}9‡¨nÉDßèÉ–t.œ@?3¹JI%„(ÍâSljþYŒ]F|¤“©â†QȆMYRAk@Ç#öæCrƒÖP• · ç»h²“=¢Y"µz57¸Út.héR<÷´êKC}:"õª²·Zrh›w#d;Gº ‡•îß»c»a :€ÙtWZ¬º*@£ IÑëMfÓ[[DTÐ>4é%tÕ;ñÚºRoz¨¨mDŸ~«Q•"ZšÓ«]…Èúä­1¿'á>Qr·u•*O¥ûöµî‘5Bˆ6¢÷ bëB¼‡Gf²¶AP8–ùÓˆ"' oưsBšn„; c@¢zá–]؈ʘ’Þ4·IÔ“ì©É~ૈî¥w€×ýÙI uÿ¢—¹•$î^´‚:‘ÆT×4íé®éxyå Ÿ!L5'z8éɃŸæi²÷<…kùrKŠë(|óð~_ˆ`U浿«“/J1–=ßSÄ0²m^Îoªëjv$IåL¾iä«HÖÖ6Ã@F[ȦŒ„¨~ òÊ5¤ïµùÒ¡v³›Scx!R9©ri-)nó¨)ÁÒµˆ^Ô±,ç\P½©¹ò2Š€¶×nãØ“h;†èWí?½%'U2¬Í£lÉ$A+7Ä—ªG”}ËÍwHr³ñR,×,zG®¶$ÜheÏõâHýQ×tþ–*¼xÓ—šëœ)K…cv"i­::æ©tÖ:Ù¥tÊ?ÀÒéÕ¡ ´6F.EéýL#b{¨nÜÄ¥«­L»f•"jS¥`°=+Xƒ( HR:C¾p–|ΊÔ$ªSâ³ ¶M§A¤ub/HV<Äê÷›>^=ùª%)z ú7ÃLƒcI‚^÷ðWº¢—'kàÕ^k*†”5fl^,2J¬3ˆ”R9›°Ü+p %¼G1èÀ•Cï r¹©Lã fÑ4&Vï3Ü} ¬ç9¤²}ic~+Ëm„7¸µd•©pß·ÍÒvzyœÊ} óe*Ðæ)®4j&!¸ °Pˆ[E^ aÌ’QÂ;LpŸSÈ¢8 ‘¤%›4Bþ˜’ Š'+“åM œÕ.Ñ6E1wðÇÖÓR&ÔkOÏ$]B™\€déœø–å[qn®E1¤–"9ˆS®©Â<(ß,äŠÄNš.uúœ¡|ëB%d¨Ï_v«9AýöøJ*³b’(%È”oðÓœ£”o…â&§&(ßÃo|i˜$hßäiN¦ª|‚g|ˆm%8BkLc‚9o®“>Þ7¡Þųie¨Æ—B”ê?¹#tátR: ©NË–=€hó£*AJ'n£è 1H¥÷»I@bT´…˜6!]ù %9V”¤^tÅ¥MÂeHì2_à:gÙ¤5¥R¤sÂL 3HWhV+¾N\ÀÍÖi"(à™)šÛœq«j¤œ‡SÌÞ>%g}{ïö8œŸ«»ýºw$/ÅIhï´Ýºo×Þ;LÙ±wŽCêù'rŒÎ¡¼¡0Ò ’þöN¿©:Fîík•‰@H߃Xðv©ð<5Mf6ݲïÍÞ 4Ç¥Ùä®ɤ½9$\F‚ìm‡³]oEÈ|UŠ…õøioQV†Ž‘½um¶\ÃÓv:w*S, !ìð…-gÛ»cåæX²·•êà͘j"âØ[Åá²p¶þ‰j°ˆ7å‘3„÷nÖÉïÏýY‚Álk12ôšÙ¤û¥ÇÒ¨£Ke?÷ŸØÅÑÞ¸asïÄÞï ô -tï@NÉuïõõSsN±—¦W: ³ {«>ýÖú˜\t/Rž»|´çâÞaøÚgÍjø¹´®sAéä}®“œýðö Æ*cˆæ…4é=³ªíœ÷’°m8çšåTá7Ðß^ƒtz>Î¥;§@Ù ^nîÎÛËD2\@¯QE"qwȬ~Xä™Ù+ú¡·Žf/„§f„x‘ºì&Ö^}-è¤T³LˆÃ%Îz_H‹‚¨Rm=mç^fJm Ùk@ç$^Ý~5Òå^Î…¯ªƒXVÒáRŽ& Õ÷*.9¸˜?×W±u4=¥š!¨õ:4dÖóMq®R´Dˆ×w€¼ê5³./Oñrmã_žŠÏh3½cæ±W«{8Ôͪ¹xÞ4t3ò„s¶ñpõ9MY IOnEž ž ÎA5Ó)='€=ÓÑ ¢s~ fEˆg½Å&š›ùb4žºC¦£Ó=u{ŠRœž†=·Ç°E¨w΋i›Tï=ŸÔ\4Ƴ0éz³ç. „© 2¦§ÓÕ¤=T~SðWŸ,qìî¬;hCÖkª®È NY×{(m”«… ’ì¢Þ3Ã7‹¸ñŒ~$+‰úø  MBi·óêÚ€x”ÒIª^¬ð®“ãÎŒ‹Âšòp"D¢êÄtò÷zK™¶Ô9ø#'G€`Ѻ÷°†ůÏ7õq“–„{÷§n–m¢¸ü1-h½D°»}fØh~ý˜nï@ÃÄÆéSü.f„Çx*·wÕûÊé·±m¶k§wømR(|ú~ýrùý­¿Ñ0âtßc3ž=4Þáëø×_ÝÞQ)Û2Æã§1·Ã§~âoãPEàno eÎ šBszÀoø.G øB[Ò›åC/¶Ó9Œþóùå?¼7ß+ë[_ùzίoï rÂéß–!¿Ñ²{ì§ß/àÏðÛ8‹11ÝášÌFÚET(‚fŒ<ûOžýÓNÿy ŸÑ0&ž åM^Ÿþz Oµò§û×ø‰Š˜tFlhš¿ídª˜r†¢1Ú }ðƒ­µk[M(˜áÁwÀÁúéÍkNÙ¸Oå¦áÝ& Oo^hL§ 7~Ÿ¦ê¦±OthïšÙݽ 7p#ÓNoÏc_-#^jCÍn§/—Iݯ£1µæðFÍPïy7'ØÞŸ žüZË1Íj™a<~h™Øú¡å-÷¯ôh†±ëέ¿íõPñbˆû†nq]õwü9L«J‡}©c1Θðe¡_Íó,ëŽÚ)ÊÚb;=?¬–˜g2Œ0`YÁOç•qè&O¿ùýò’ÏœHzM§?ÞžøÔßc´ûlyÖ>ÛÖzX–ùã2äþ}ákV$YžÅÚîæâîÄô‚ÖèiÄihòíNñ“ú÷#c~dOà;®·¬ÏBqãâ! 7Ÿ©¿óíòðv ó>Ø+gÆVñ_í[ž 현G¨íÈý:úÓóèõ°Âõâ,Šqìf¼áw€š•`úÙ˜¯5E÷€  ±¾¹ˆó×e}ï|Ù£^Û*ƪ&e\‰Ã°—}Héí‚L¯ÿI/™üö*™®gm²-.Œš¾é>ŽÕ_-¸sþy…¿Ú6…–÷ç–yÞ ^ZMÃ;}sKÇ´ —û¯4MÓ9YëŠÎï–ß°-GX>Ö6¿CÑÉS»xßOGjºöu*ß:zo,qR“¨ÓˆŽT3QTóç]ÓË-,ÿå‚§i±¢šafE9mˆK8‚}}!\`ˆü¸@Ùß÷f%ÏŸð¤Y¿[×ß”*V¦÷|¦áuüO$ÓÁ7Í} …—.š°à"®Íêq\ü‡ýëkÚ1Ô5ÓiÙe”gZà XÆ`ty)Î;Jv\ï+®Ãv5ËíæðÈßveJèfÊÿœÒç¤t˜ê“ÓdSz!õ&”xM½1;vÐqJÿúô”‚)n9Ì)ýâ©]B­Šb+þÀ.¦d†…Ù+Ç-ÌöÊŒ’"ÛþæSûÖqذÿÊ©!0¾nããOíÿìÐŒeÖlÖäq‹n/·h}V¬ÛœR¿Ø¤ÃìQœ(›á½-§e|g„ë§…Híf~e$ózsvE qŠ5²OS¸—VË ¿vÆb/ñâ”'½ṟ]nêËó„®,‘Õ Ì÷é\‚â.È¿IOé±à?£?ªS½¥fXC;ùrÙ“GäÇ¢HJf4X޻̚ð_n ò!Í>9è-‹È€\㕘4Œüz¡í§²Çe˜zöëfð¬Bç—Gˆ¯ØÜ ÜÃåW`áÙL~xDá[ÅØ¯>@r¨Œ^4'…ð¢ ?B^ˆn3Ó¬¥kÒ1ϼ¼\sæÎ>Ž Ü•ð¾Éáâl=†e›_œÏo}rÝÎÎvëK•FŸ7;ÆÍ,ÈçOñ‚@Màç‹Ä3ÕÍz¹æ æÚGÿßÙÓu7~\öô-bî:dµîŽ›k1Ún"z…®È~¸œˆ´¼c ¹äÓ$±ž\ªèJ߀ôìëùÀ¾×ÊG(¾ÄºA^\ÍÓ›uâÃý—G.w0èt%|iÐÍoÄjêä£Å¦[kÒ4ŒÖÌ §ÔQmkýÔÑÄ[ü¶ °þ ÷ /¦&~Aöf¿²¾cÒì=þ{Å£p‰.¨˜Ø….ˆ¬ö^ºâa+Y. ùõÅOG‹ÇxŸ­ðÃø´p7«þ¼þ–ò8œO}1§á}6äHùÕcÿ8zg€[ ¯Má g\ñ%ýúÈ÷ïàˆïИØ$¼´‚øAu zÒ‘‡É9zÍ×t+Ôù½™ØÓ "~¥øóÕú÷§ôâ|""†PñÑ$©íÓûûó‘^•‹5¦mßÞþ˜…_ùh.Ó»óûá#^l·Ð¹¯ªx@ EøØóŽ}ü?œ7¢WÃøÀ© ¯„]¹þǧ¦‡¬Pƒ>D!ÇÃ)[ë;ö/—S²!Û`ƒ·ýÅ'wˆ‚1ãôæ19¦ã©S¨K.Ô˜´½endstream endobj 377 0 obj << /Filter /FlateDecode /Length 5239 >> stream xœÅ\KsÇ‘öVøà«/ëƒ{,N«Þyy°líÆ*,[¦àÐP8FÀ¤E`( )K:ìowfUuwf¿ ÐTè a¡Y•¯/«3뻕¨åJàåÿç×'buuò݉L­«ò¿óëÕ'§'=3aëè”[^žär44¹•—0‰·«Óë“çÕþf½µBÇj‡?uŒÆèê°Þh­á_ºzÍ6FélµMí¦6ÒUo°=Ĩ]µÏ½…4Õ7ë µ Áà|Øjªí™âb|Á‹ÜÙI;2qŒ¬õ›"­!Bu­*†è«mê… ²ºMc0RW/ñ·!WýGé«ýeÞ•ëu!gs ¤NLþ¶ëMH¡{¼Y+¿étÞ»¥¿>ýìäÓÓ“¿Ÿ@?€Yýÿß^kïWN ´2VˆÚ¸U±–.¬nw'—0\=HÀàÇí´ôH‚²bFÄÇEQ !Q|ªÛÝA­OÿÕêB4VãV åñG®»û¯…{òѲÅþ{C–òÚie)©j%ÙJW××d©Ç;A©\­µbkU?uk¥~÷ŸVÛÚϦýÿe6`@Ä,gËOצßÄ:pò7Lª¬5ÈaaAýê€uAÛø+YS[Ï—’Ëœ”µ †­ôdùuh¹Póòæb™}_[{"¼·"h¦vïƒ[ÑÕÑhέŸ®ÏÖgëExÖ˜RkaTZîðzw~º;¼YÈðÈÚ)Ë­”6 »§˜Œà¥rX>zåL‡vÉ=…j?®:}ªšc®žYˆÄÍ*dâF‚¨Щ ð•DÍx^é4qŒä:ÍØ ĦrÀ ²f«ÓuÄ}€fìÚ#Óíy8æ `-d> ‡€ ñ ¶åa0Híó·ëA†ª?‘ÿ˜{ª/Hóç¹/pòí!°°Ò6ë¤.yƒ}^§´çuúËBoKû ,—µ•@ÖùZ Äç¦z±Í\ÖÇÉ‘”‚óB¤ðÕm§cÏ÷äooÖ°,üNÁŽ…3«.‹O²\ Ât^–bª}wL·eÛ!d έ…KJi[½.í¥œþFÛ‘¬Š.Gé í/»÷ɤí@U  ´¸7hp{fdÓtÙÈPëVÿg’°iat’É<²Ä‰ ATùw¹Î&$õq.¤x®ÕÐùáôfÚXä Õt‡væ=µ?æ‘>ÈFùÁ=¨nÅHCDj©SÕ>Í^‘¡eÍŒû¦›|¸û.Åîôg_N 1 ëöý¯Ð’C¼à“„â¸(4#ZÓ‰}’=ÐæýD8 ȃÿöà¾"Ý'XÑl¦FD+c–¿ç ÍÀˆ5ÚOY3mi³4fç·òzLO ¿µ5ÚO2#íAGtÆéº!\e И÷¸óMÙú&›û¢W߯,€ƒ¾Ú¾"ÂÙªWfšk݉bÂÝËÍüvbÆéÒK—Vwy Ø ¦°{ê™éµ U¥¡{MÍlì!ÏãW6£{JÜ • mþtGáƒÒЦnî<ãé! µyâ^XÔ ÎÍ „hk”"´˜* ùùK„<@6X)Øcp^zŽœBõª>ëŽu{µ%DÜdUR Às£JÛ rpž€Š·T€/švW}Iš¿Eƒ¢„ö½%ͯˆl¿%gÓÓÛÌc ÑZjv"ìoþu!ÓÎuáuïý$Øë‰lû;ÙA‹Ä :¥­¦œÍ÷yBn$x›™!lLšˆS …ûßì¢ÓAýº€AÙ6t£˜¿î¸ôjxJI-¨uÞÞÍöâRF„›Þ¨dpÎ7íçü$Ò|‚w¹(ó…þ-¶¢_Nº”W¼þ½.¥cT7›³ú“z¸)ÕÜ'û‘° ðÑÆ€KÖg'[•(ãeCÕ."{ÞÑsAˆèI5wý¢úQü°òü¹ŒP€ ªíH<’ b•$JLi¨Jþ¹kþr”CÊ»/Jó!4ФlëïZS6Ú§€²Qý)Ê¢B,#ÊÂ;emÐÒè,CCˇ ïw” õ_%ûœ‡:OŠ 1Цä¶Sj åªM˜W,Ì¡°/™˜Þǹ¨p*èKËúÈ5€x¶yc‘BÔþ7ƒÎ/LÑBÌIŽ.ƒ#Ð^Aâìõø´;®Ü™Åì3¢ÿÒ™ÈæÈ̺^]ŠëVàa¡{¼1V~…æÓâ?¤=83>«€ÚWlȳöøò¶ Xg•î­’ÎSbÌX°¨¤4jdÓ®ÖÊúÀWD=SàÑØ‚§øÓ Ö¨dæ†èéyRg`»‹º…tÉŒìQ`Э7/ˆûo@=´ºéüWt5Ðôï¼SßZJDíˆ!,Ên½³´}ìü&¥Ã®º0ÊÕ™3Ž#öθ"’¸‹Ê%ËgìZêœø’X¸?[½Bªü óv€n“6Žœ¤|¸$)̡ÿÊÇ¢ñÓç¨RÁçWy¸Wlø´¹„Íh8¡+Š™öÔvr'4JS¸Šoòa‘Ú³›< ^QOz ¬IЇ³"*„‰[’0qt#hP6>œ=ÀY=vÒó$þ½ìd@å› ƒP*Ü#úeÓš½Lè}pN窘«$ìÙ¯©€ô–Ì6cÈ“‚— Îã²9éZÐ#þ3õyîâ¹§ÖŠ‚! €žt}Æ<Œ€áç-—íInúZXÙçæÓYBÀ¨([W÷kn0>zf“6ç]´=šD6Á€ðÇL0ŒNÖ‹ ùpž&8¯•$n¢j“>Õ³!ww—íëb"Ð&Î3fcð:<"äÄT°Ài‚â…ÀÖ*ª <J߬¥ßhz—Hã«Á%IQŒlMÂŒÍ,ß.ÉÔ{:7ý•X:Å”xpAP¦¹Ê«ƒÒßûþ.¥ž¨…¯ïzY}!U@ÈÄÚ^GŠÑ EÐ9ÑD/ú#z’ꪩ´C~`wÓZ¦­ù§´UöaPÒV ")C™žS䂎=Š~1ï@ñ—”÷ ©± ír@Sß‚à§ýûîé´Cdé`~Äos¶ðs! ×ÂàÄØôøKú‚ªó·üžðš:„ÒT_ ŒŒ—"ráŸÀƒoPlwÎ%ý¿$@€ ©±“ÇóŠwCÿ®nò¬\i³…i\;ïÖé]釅²ÈTœ‰„#,ñžÝ·ö²Ø—‹¿äóo@nJÊÐfsnß,“¤Šß5pFX0v¨K)Ú4S²¥ÎPäž/”z—„.°õJ½óˆü-[éëg”&ùbs~¸Pâ ¦(rê—ãŠBˆMt5µÈ¾ ïåãrE)¼TulÒß/C>&މ¾º,Ä!0ŽSªòŽGeEB|5%Á—Šª,}Í5ý|¦8PJãzLùáù“…ø~|&LEPK._cê0M~71Ð䨻¼ÆàÇJaÞÆ€äÖÞ<²Æh rÜ³ËØ|m ÊöœüŤXCÀgD˜2ùïxV^Ðã^ëã…è÷<17dn¶ R÷µðÁG…ÉJ–[‘…ŠS€ÍàsͤÁzÄ’á1#u…yQ5—R®7ooŠËR}’cË%X>Žù7 `Â3ÿ¯qg‘?ÞazÑ ¹„ávwøçáârô_j*0WZ@|¶d‰Z)ð¢K-V£–ƒhºðÊ.q"øwlµe#NºÒÓe˜Uª–èJÛC Ò~ûr™½e´«z<[-"Øå§¸\­YÉjø‘àöúIWÃs‹0ë )ÏyÄͪǤH¶Ùe\Wy“ |L‡,ƒš›-± ¤äç·JÂËgëßǦ Ìd¬Z¸LPƒHóþªÉzg «´ág­Ô?o ðÞÒ:AÎ=B òŒ³¬ìUÖä¯)fº”€~+éå_n†wÊ«¡ûd.ãä—yúõçE¦Ë†ÙÜã¦ÂÁtü®¶ ½ Oš¸9(*ÄO>ª±Ïò$»êÐ,gïYœÐ+ÇyÄŠBßV~ZŽ\Y?¬#©®».3bí ­Ì©;$õ€°’0¼êî&Wø5é#KåÁªê-+×;䃈Ǩ,%xØ[ôæ ¼‘béj¼©Ê¼–ü\Ùùcnt!ÌÔ RÙ-á¤!OTù]óYro^øú5)CÜ#‘&˜ˆÏ“d9ajÂé*ÃTZz¢B›rµ'nÖ£ÉCe/4<=þÑËÃ"yÃ>Ên{Ê‚}@+˜íøö¿KŽša!5(+ТkíóZ˜Éì]™_Õ×iìŽS’jGpdê"ý\Pûs;žç„½qS¦«hŠ·/r3V~Ùõ.…?Á3EóÒÇê~’}kvsÏþ ‰ âMõU.wr¾=`õbjÁ‚os…ðªedì—€à(8è™"¾$•¿²·y<ƒD|-tð’ÕáìñšË±äùÇežÜÈBªÅD °¤j ›E$Š•öaá…hÒ2Ç›'’D)ÑÑ™ã?sî[/‰-Ñ'“Õš—Šˆä±òšöy¨© xz]>;å§37€ÊÐ/Œâ’DߦBÚ};1%:Ÿç¸ÁDGЋ@geÈG|Q°J”¥(Š%=ÒķŽÔ³ô·Jzn;€$pÍg#J”—Øz×_vÒ<²P€°Ê¨6‰áHZ>þ mŽH…EߠÑ}~ÚºûœÀhEÏ0ßo2 “ò'­”_$O,„X¾Óü„¿å{ü‚ À@ÊÅñY#$„sÉ;`Ú·‹V¯6à_!n¹Sí–\çWª6¼Ãpû_qÔ§ ËíÌQæ/·úCÿàz‰Bx­¹MäýQëT¬q*¥%n¬¤ `!w+ø_u“Œj%xíx_­œ+õh=RÎf¸¹-þMޤŒÖÏ×n2ÌåÅZ™M3Ú`W}B+ßY'¸úe#£§I’D£NSÎïÌ›2B¿D€X(¢×éœ30$KWc.˜b¤ò<êqÙi "ÔýÔ£û–û’-@ûŒï"W ;|ÉÅ÷°k#ºýR|ª—¤úiÔT,K®ZmSgp‚NIâmXÍ]ÏE廊žc­•¿GùžK¦:#{]ØxRûJÎÎ*ë:Ö]ªØ=*2 ¶3ð­þFXL”êôÄeUªt覓Âhô6¨:cg§ßâ@ ‹ok|[8+'o½ÞAéTHÕØG Ç«‰Ça²+òû! ,sZޤ†þѬRÔge¾N(ç»!ÊÊj ©¦_Q K+q©­ÜS¶Ñ¢<Âågd¤×Eö~ÔÍñ&dŽO‰`#>§ã!.û_% €Ë²MÁá“øV¹DgÏFÐWˆ!ª|+kØ­ ·Ž\dFôØé”¶yĺ8Ö%ÞGfä˜}™.¦hª¡ŸŽ<…µ$°T3Ÿ¢¶"9ZÓœ–§´H|3¨7MÅ q´ìaäE”ÑwîøM^DmY”Ùl0 hüDp9ó ÞŠóB´q_Iëa2Þ “|á`¦¯ôû‹O¾D©§§{¤ºØÜ,q“o_8l_w×l¯Éô°Ã‹aÏÌu××%ü‡ 3¶â4x*†øÿv»?Œ½JB’¾¸~‡RliŽAjV{†w¡ ÚàkXuÆB5|z¹_uv‡J¸.`»G1ùƒBõ¹ºI˜Lº»‡çZØ÷ž»~Åäü®!PXÚÆ†Ü‹¼ÿjµb”θT9w?Q"÷CQêÅ£ÃEX6|»b޽[†¯Bt¨iôöÅÔJöà°,–—Fj›‹\•îw ÁÛ!w?ÓéUp¦Rç’¾~õ}‡ÅïTx8s‹ª bÛy$„n¢•~0¡@ɺ{®ãü°ÍÝôIMÞWuïH-&wo}æõkt1‹Ä`”p5ô™Á6ádªê-YÊg¬¼³oœ’Fwoߣ1è> stream xœ½]Ko$Gr^øH ìƒ/¾4´\”Ø¥|?d/à]awmCZ`×ø Î-¾fV$["{´üÛ‘YYQf“ì$t˜f)+#32ž™_Fý´­\ü¯û÷ìæH,®Ž~:’éé¢ûçìfñ›7G_þÉ„El£Snñæò(¿!AÃ#·ð:ñvñææè»æëÍñR´!ÆèUó>ýBJÕœáo#„s¡¹8^£[#}Sšk׼ßN›Õ&·ˆÂ4ߦÇ1:šoð±i…¬¿Õ-þac”Î6·Çø\†hšõ&·Ò6ß/•‡îƒÌä 6j>liÞãO#ÐÅýïó<sHƒÐ·lÖøTµ1ZÖø>*Q'SKŒð1„àš›Ô\H£Ò4K“ž'‘’y—'æ# °ã¦ÔÚ#4ir*†èqP*w”‡Q‘à¡7èXBC£»Î4d;À~•~L<ƒ=¾%Íé«÷eö.õ^žã¸¿üSpDMü´>–iœÞe9¶ðÿsk¦T¶=Ц´n—ÞdCóøž**6*ìciÈ€u‹¥”@ÌÊüò6xï›U"ëAcœIä„j~ štMڬ¤K££ý ¶ ø+öø>?6ð“ê階¬S !AŸV×]?B7—3/Lð[*Ù*ë9¿ikÆïØÚ ‚,­u¼Ô!IÛˆL]ßÙ•,¾çù±¡¾Z#?WÐIAåqç×ç¨Õİ£j…ê)Ú ;ÂBØFOñÄÁ d×9´~ó—a‡`؆ҡÁs`X‚PYðD„Ÿ*–µÂeqé”9„sF û>i’Åå¿è–\ÄgZG4]yÌjÝu¨ËR8 6Ì¥›Ed\ó{"Vßòßé­ ™é c¢¶ƒ€«%¡dÌè¬M«ÀfFýv;ó %ðmT!öºûÝNªõÜ`i ü’“æhû­ñhúežèR·ÁÕ«ìß'¹7`/þ:wš‡ÄUƒd7ƒ¨®^‡4kÛ ÒM4ÆNØÏ&ó;835BÝZŽ{Á^9=ÞŠãÛ¸á0"ÄPÆüw»Æ!]i;ZŒè ¬ÂB90™: :˜P04Ò·V›Å›oŽšßÿ3p8Âðuÿ·ka>¤¿ÿ þò8¶Á.”jƒ.BñYoi'8jZg´L,…Ød†¥à+·ïh $·<ùB¦Ö  2L0UÙY‹$`eÝÞ£2ð$ ~ì9*"ŠŒ —*Xµ`’‘àq¢›TnúÈ øäµÇa…dÐ4HÿôäÙDpòà1üB—ýËp2ld BÁ‡ˆ2ñæó=ø$ÅVJ'øÄ’•Ä}¡Zã%—é_ìÔü±h!ð[›61º}dª¬sXаwF 7`¨÷JùÆ…¼ìÃQ0ñ^ b"àšUÙM™V9ÜÒ·J£ïaˆ–B{è3©ôj_‹©C¶<*6¨]È«ô>³ÜKXQª‘°-ì¾Â¢¤Ü_XÀ˜?2Å‘¼D0kKöÎãò‚.õ‰ò"Ãn 9 ûLlµCEnÂÀÍÈs=./O±g³¦Ÿy-ð×ÉkóH³€š;­¯NëëÓúurZ|gF ,+ÄÀÞ÷!ºò˜Ðp‹96Œ%ƒC…qW·.oò#h8Ý ¹Èš†°æK4ªfiZŸ p Ã]Œ®ËŽÝ†F?²OÓY:‘ÆiN€„4-£# iX¹3bì^=É…4|.ì¥S#±÷„ø8‰ " и2¢IŸÏmJµ#Œ¢‘p!1j §0…žQŸ@ÒI?dP$8üô¨ƒ,'?õ¨ Ø ŒÜ}ö¦(cO5ý;„à8×ÎUêfÏHdjR±š‰ž¯òšãÆ’ù훣?ÁJ&ÿ7ü÷îêV ×ÂJ”ÀÛYÌ$Ýr;dÆâîâè^×XyÜèÇÝü6¦4rpÓÒLìcZü7à6fsåp²Ý8ätVcÿV P[œȬ{:!Ð{ÀªJÿº$” ýð:)YÌÃÓ)vµè¥Ë·g›÷ r[z‡c¡4®{JâBHk´sÀKç!…_ôOžNòjë¥Í»‹ÍªÎ¬ÀnKP Bì¤Î¬¸)¡‡:3Š šÉ:8â¤ØÒXòBÄíºm8sÙ9kÕÙDÞû2šV'SHº‘ìŦРI F@Âú{ˆè”ë×9f?)-[pdŒÔõÅíÕæ]¥±±Uh¹Óæ!/ÏÒÊ6"™Ã!ÝÍÅFÖa2äòRyëpoRWe²Ûn #uþþò².‹ ±žÅ™ZjùLC¡XÇË:SH–Â3JÉ~Ç$ÁÚä> TŠQ sµ.hPÌ'ÓSÒbÁèÉ*3S $ÁGFéíKFA¢~……ô‚¡ðz 1œµ\äT™yðîÀëdë Ú£~^gô¬xàbõðÝKǯñ$H˜W+-,Dr|7‰ÿK!ŒçN÷]%¯«`¹LÜ„i œƒÑ>%ÅüÜ3-4¨£Ñnÿ\G3sˆÊg ^<~äb´ÓJ ñ‡|øU ™ÀmuFN×±?j륗²äë´’!S¶Å9Jéå†LiÓâîs}¹Ê’kàæåü7¡õËôç:ã· šsŠ%¹‡Ô pëå°&$G æ5DÕC’60¶5MHö‘šÑ3•¼ql#—·]†¦„‚œ90w|U7†Ü ³ö×È ©³ïßßžWÊÏ:Ϲ¥†KàþOŽ—Vð€>Û]ÄgªúçUI$0QJ°Ú%W"½V²ÚÉë –dó¶¬îÄoØ/_›d<ük¬M€€ ¨C¯M„ˆ úWX›> !”6oénÆ!-ïdÔ(Hìîbóá®Ò†rI¹Óæjz³ri|+¼¦¦7}Ò™Sˆˆä!ΈH˜µ‰®.K‡ÇÆk<ï2à3RàÔ9»Ë­½dè·u~*„bGÙñùMþ°KÌ ¾º4€ðaP¾ nwtÔØð ‘ºQîÂ'*¾ù‘g`võmÌñ€¶ºc”ƒ®×—ù·-ÐÞ— l†!òhßuÉ€½¾b߈‚£Í'p‡¾ˆÍPô “¶ „†¨a|SÚ&[H…pBÛï¯g!(q C󦎨˜­g Ë´Mw6œs—Az¤‹°¢ <˜LZÉðÚiؤÍê8‰Žd©×OIØ‘NÁµÄŒÖr¥íVUúþT¡’TÎúÔOt„g.U˜‘ö‘U=.õ’1ò{ò?èi:%FAe¨ïpw>æ#üîàú·8ˆ©:Ò¬f°“—déÇx鬵ƒS~¤d£_g ÒÑ úû™Õ¾ÏÝ訇gºùñ‹8 é¢+Å÷èŒnQ6ÀÖàu ÑÝJ`’Fo6ü/·LâÐÍ,à……}Ì@V1„~öÈæ; ­üoGoD ¬’‡±Ûp&!î“>~ £›ÓÜ‘=šø=Ð>$å¤B”Ój‡r‹¡XŽISp6ì¼CSgJÅÌøêŠvÓrLIJ°–éæÜÍY÷ÐŽ86 qÎíísa‹F„`{d!†…²“¹®5…­ ü.ŽÍñj°f ˆ‹½ÈC§c¤bLáãÃ¥„nlPswFHïÜä* Ù|ó_¥“8¼¢s/·æÆ"q:f÷11 : zÃ%3¹Ý@‘E³â-în»Ø)‘'× Â«c±sôtQHö0±9žÏðŸ;²‡¶ð‹QI»•.Y1ïNuòÓùuÛöOÀÃh# ºÇÃH‰xY c,þ›Aw—Uv2 †’ªu®Ð!b(©ß½¿]Ýž]|GA2Âr!ûMsäùŠFxön†g?»»ü¬ê®=%–O‰¸(ˆXÙ¾Tíå4 nû*ËIHVÛÍç_–Q{ù‘î­Öç•ÀYˆ±à£§ðÌágÈíôËšjœø:¼ràR¸T‚ºÉ{­sFQk\+*­JPmPŠÑÒÎV‚­AniÌ×%FÖë²³ò1¡™3 ‡†ÑyhV%c½~^‰UxN8°ìµ´½D:Záí„=»«ë·”jØÇHæT O×y°-©Sp,ÝyXÞ“åçaw÷ÿ#ª08O{aÀQ‹XšÀ‘Š‘º¾©¸ î©S&¢ÏtHÅ¡’~ÿ¯’;ͰJéîòÅg,½Ÿ#ÝVZíäåø*‚€óq¡¬8êT9f–œŠ9ÔúbSI«Ò¡˜­tDÕU«.Ä&¤ÎêÈdQ`Béthš^¨X¤ëªe6œ]ë‹Ë“JZ€¨ãÀ¨ý¼º«£ÆW§Ê¬rüÇgbðþüÃêú¾ÎÜ:´Ò„蟟TŠ¨Ò‰4Ÿgx*ŸH¯ÎytºaYȯ¥ÛÕÍEu+vR;Mf—Í·»A Ö2¼Í¢¤*ÞÎfq.ÖÙ{ëâJê³ÕõïV•¶ßR‚ʤήb1Zlfß_l*M¬àC_abX×Mr9üìþýÕÍêϪÒÜþçU­³ÀlnW«››J«–*¨Y®hJÑ`13ÊÊ*íg¨"L<HÏ&.ªh¥«›Jy™±­tŠÒBœâ•ëaŠbög’§8!ȸ^-t:ï³õíæn}]-½_Õ Ï¤^Æ1J×ïï7•lA¾D‰UºÂÜ›Bêfõð¾ÎÄ´ë0·^‡„aƒ´ ÃEQŠJ[wÚ€tx.‡u.gk« lárxwq½©¤`:ØVéð* A6ìÀŽÉ‹å±hÓ±¼‚åcVë`“42´·€ PÕLÁûVY×gÊ•¬d±ÓÎÁ– œÐ¬ŠW \W>é—¿Ä’š©xd“âTøÓ+ QƒÁÝñRãyüÏ."Bì%©LÅ(oVi•ŠKÏų°G)òKKe9\ÅC¹ÑÂÉ`„Ǧ[aSáL2ªü°ÁL$¬ Ä€÷ó¬¼%W=TSË€XNó’¬éˆ1H™AV b€´„ÙÛækòtM~ßtoj2ý}Ñ£ôؤQM‡ôèZ$Ø–M%3i‡·ù1Âkn8Mˆ+¬Í9íúš4é: ÂçÊ5Ýã÷ä÷†ü~‡Íó ü¸Êa}lnýcŸVÚæœ6¢‘çÐ3±¬X–†ùêB‚—œQ^ÿå¼”N¤Èm}É)\™¦‡åPØåêì8¡k¢i~8†ÌSIg2ØLK¿nQdZ¢·ô¼rì&7‡0k®Fåú—BC þ‡‚?æÎA‚ó¢%C)DоY†%† Y|*¤cPÍÕ¨ðh×ËäXæÙ Þ*ÝVúæë&uèlìÏ Óô×j8uÝ‘”<-Z½­5:£kÿ=.ÞÁH`‘Âxio?# T7VF  ]S•!Óx¶õ”iƒû¤à%ýZ­°ò"{ýÓ^À'&°VcÜ;'2(Þ¹t;˜ˆJõ¤g'2Y]ڦܓ½²ÇBïížPûx/Þ–Þa1Z>¼ý+ØK‹EA™¢bÀH¹¶¯H cN×®·Ðë9@ê’Šâ´¶ {ÞU[„¿š?ì‘‚|ÄÅžÌO¿£È$äoÚ«Þð&wpöɉk‡5_{òøt.>U'§°óaY_‡\Â`°ñ€Ïa"<.'±l—åâ×–x—®º¥•ÁþBˆÅÀAJÜ_IS°4l¾g8(Û]ZË1ô0„†ÇdL›éöÝÜÝ¡³Ü£ŽS¥Òã†AŒ;º Çd)…¾´ÉÏä7}NÃ]zIŽÞž›ª6œJ«ý€TJgûõL -ú«Ènד*ÓSö²jiÕþZ²X-÷Sç]Ì&‚¿B6§4n:hþ • »0t“%»‡f"ÝTïvÏ/ån]Ùã~WE?ϬGl+Fê#ÆäÉ€Ž'—úU!à­…x‡ß‰OŸÝÝc©mÊ;Uœ)=P¸+5ÿŠ€¨Y*b×èŽûpÙv@«QUƒñÅÊáÇ…Ò×5ô®û¹xË> '4±–7TµOXA!Ouvï¤0VpsùY)ºÝE%vø¹5vÇ_zœ¼ä±ëÉI@ü”îÄcÖíåÔ•ÔQÍ ªQýcÉ´…^ÃXô´]§Ó–(‚\À÷\qó˜##3QÖ¼ØCïÛéÀRÉ”»“¯b¤YÌ|å ŽÏîuÜaÁ&¾ß“áä±×±3zÃxÌ”(£ù‡\x…&}M‹ÄŽ/œÃ¤£Ð©ñaã> &O9¶p/={ùô´ày 7£%»®Oqf×ÍÑLìÿÁ¨—.[ºÒþ”o‚³=ð{l+GÆô“v4ÂË74öÙÍÙîä=m7gû™¸gîæø§Ì=Ê2ßä='øž÷ß ‘OÙÌѱM¯ÍRå;=3›9¼vNrŒ6}bE[“¾{;üÖnz)РV‚5黨¬bÕ÷s^ˆZ® ZzB*TÍ^>¯H)‰ÇÊJ@¯Ûh ü”½æwÇAûT>A°§ð òˆïŠç¦…×£ó¨îUŒ_l„‚…)ºY䞆Å8B¶Ÿn¥Ï”ˆ¾ÀŸBòóç¡ÊŒ¿Þû3ù=s– ý(•~î·b²#µùà¥3\)Ÿ‡• “œ‡-ûL~×—<-¯y5NØ©ÒCÊn±†s2Üû¼pº?½†Œ ³tŽb¯H‚’Š0(£C¿Ÿ]¢WLÝ»ú3£ÿ>¯gÊ*¿å¿S( Šô)‰{ˆ² u*•BaO?™ °g¾çTªÂ„0Ô{ Å ŽÇ¸Ü Ÿö½$(‡4vœM”ƒÎ´‡Œ7ÃOˆîY0%AˆÌöBÚÉ7 jL‘KmdÐ×Ç:׈uâ)£T NÝ]y¡¤îêU~ÍàmJì´ùx€ªï¨Mû­õ“T ƒº¹XUªéjüˆ[§Ç'_^%:_¿«Ï±r‘åUXVЧX¦eúÅ,ëªu¿Ï:Hüëð¬ƒÄx¦ P7Bð«Èuí~g9øW±{„T}»Gˆ6‡³{¤ßºvªo÷Ü%~8˜Ý«Î±Þî½Ëz»ÇY†vïápv¯>ÏŠÝ{ž»Çy¶Ûîݱ›Ç‡¬À/H™oàW­“ïêQRÕnÐuv:¥1ósÍ^.tF»­TפXBéáä+´ôÌRãx'‰¦ »0Õ)ÏêŽË-ßu¡ÇŽ'ù¾¦bt×e÷ÙªU¸™µE ¶uçýç—;\ËŽ_!‡«Î¥’É#ä?n†õè¡ëètî ì:t…þ/?t…Uþ9ÛõSL|¡kÿzYz¹{ì\:&¤\ø‰ìcí<øÛ¦ê©;(L¹¿³òP|·A`ÈÚå§«L] î}ðÍßq |S Žÿd»_7Ö’ö‘â°# ìv¥p[ï0JY0*ªtÇ qrVÐè"äCÖ(Ò]3~ÈZ´OU 0?ýA[Q@à-ùŽlÂÆƒlŽ4RØ´ÿ´4áM”Ìý-iOµ–vOÛ³¢Þi$BÍ<qTÅWà¹y)œCŒ1GSÛø(5Û#¬ VQ'2{…œÞ0”ÙMfŠ\NX-vq Õ.^çÛý}—ì3E­»Ñ˜üA‡Ôa©†?s#¦¯%Ìíã²­á™Âø»` y|Ÿx“ɆÀg4·=L_¥õÔïËT©¼¢-¿0ÄU8‰“©ä×å"_<¤K‘(æ½aÊØÇ<úå> stream xœí=]o$ÇqFiC€óàçüÙävÒßçèAlÉÁGÄ Èì‘{$¥#÷D.ïx/þí©êîéYîò¸Š~ÐzØÕÕõ]Õu?ÏXËg ÿ—þ{ryÄfgG?ñðu–þsr9ûãñÑ¿ý§r3ßz#ÌìøÍQœÁgNÂ'3³±zv|yô²Ù̬uÞ{+šsü­á§µÍj¾PÌ´Š›æ" aŒsÑ\¥!ÜèætÎZé÷2ò[ïœË§ž’©·ä÷ò-tƒ¿¥÷Jɦ/Œ­ãº9ÆÏ‚1¯l%Î Pj€RáOÖ*ÆŒqÍ5þÀŒL¿-cÎÈf}Oe÷Œ;§¦MÓT§¸l.ÃwÆU€ 'Jïš%’NÒˆÖKßÁÅWˆ=\Å0ßœ=×—saaE8èë°6c0ñ‚àñjøL=ëi§ÊÑ”v¢YÆ€– µkŸ½%ßWÝ‘u³žØÿÉ0ü‡ãÿ²s†d¶R韥ý6 (Û8Ø‹™i½•«Ö+ãÜlÁa ÜgœàäÞk¸÷M„AÀuÄù9}À¼Y6ý‹ùn hË4¯šþz* rÀ–ç¤Ëáú*êV£©aÊT ƒà@¼9tÿé^Zgk ²”LŸðÎבg”¶Ó݃;¸da¼Î¡{5÷—KI{i ¸µyG¨<ã²ùÈÙ;­;žCn À7õÙÑÓÙ.,„Ð/¤…½4Ò o½Ö<LÅÈHdÚÁ´ l,¥l l–q¤u€ÞÁù„Í$/t§§[Y€;—Ä H?ör¥z—Î)Û]&üiOFð- —˜Mÿb;”pJn:J»Ûʧä”å±xÙ^Ö¸ÍÁûME•¹÷c„g&~qþ‰³?¢B¹6\”j¹Â5qþqšoEÒ§D˜:NZ :C=€ðp¦msB>/©V¦¿×›8Á¹ YPɰ¨ï‘ùúe>#ÀÐå)ì0ž£QÂÿ •–b"c[ª@>¦Í™È”Æ Qˆõ!M¦’ˆâ]’!T•žÆÅQǽîI;Zß|KV{ö]ül@_çXèa:!+/o‡k£:3aMq»EN…LFÙ×(Ì…ÃÚˆV®­Hk†‘²£ä“íäòÖoðw´'q}Ù«D6ÇŸ.¨ù¸žjÞi½<_™ñ°é:‰}m®É8„PT«¶eV²Bq-Tm¡$,Í%, _µƒ2MpÌ/f5¢NI”âð3J©˜–”1:¥'y°Ê¤ƒ“æ#(¯¿#L:Ò[ƒýÕïzáêÍÌÈD¾6©^„l™øå,1P…­Cvñ£˜-y '2Ê3tê’Ê—„8.¥EÎG9 Ü̲¦vɈÓftÍå„[p: Ü«= ™oÖó ±­i> ,6›õ~.LËRc#Á´2 “Í×äóóü7¢Ër7Æ"’„ؼ@UXÃûLüÞIq1X£#ßh¹¸Ša²´á+FfîщË~€u°J'¶<¶É ‚ŒOËõ ÃÚ:O¥^b{†ƒOlï’z¿!|ÿ.(\í Ãp¤ú#%W¾¥ëUÄ…[«\9œt‹[+|0‹£îÔ9ôw¤CnDDoZqää¥eÎpu²ŠO/híäñ%U¯zJÛ½°·.Æ„£Â)&tf”¬`9Ê…”t©zêQ'Qß„s‚ÞÂ[álÜd[.‰Ï@öD­‰‹K†gŽÔ[É^Ãýz°[FÖEúMPx7/82âàŠy¬Ï3EÇt.{Pà…“ò§ã£¿Á$ãа.þ{}vä˜GÃ&ƒ£ÇgJ3€È̤á`èºÙõêè L—J¡k3þq=–â` 8 B©J\Êß*ŒJ4?þ±ƒÒ"´ÄÕµÛ€÷fÿm<àŽFöù÷ÙÈJ#…{Œ8ÔI²Ñõë‹«S²×ãáŽKÞ2E¶zÕ|äO>Š'%xð°ãB+ÉQû|¼ƒ"˜ƒv‡¿ŠÉV‡G2ÙìUsǟ܉'wÛÐ|½º9¢9ÀÎ8L{TmD4Ë»l«³Ë˃¢™lõŠÊ‚0lÿUH/#ëo‡_é”p¶ÓÝ“óH é{†ì=ì·=ÓNƒ¢‘1f¿Ål΄hø;ŒÍƒ~ð¨«ZÆ HÁ-‘ÍOs­Ñ ³ÍòŒøњׂ¢ƒ{Ù)¸)¿…ø65dSo±¡Ó*à4 ]š¯ÈA¨cµ.Â%ÔgF?8„áSüÁ T“¸¢+°Ð܈Ãy¦óGvNr!¨Í³Ž+²Í£plI¢É^yœ°â”ºÏâF@ç’ð`ô>pùyk€‡ÈÑËî˜2¨Ð)Dí[2!Ð/óŽ Â|¸ÜeŸ7 œEO¿Np£c¸M@%Z»‰Ã1Ï·ìf†˜•ƒ{X‚s"«ðÇM!#i­÷ò±\ÂñC¹EßPº%iÝsîeô®Ä/7íË@¹Æ Tx™ï2éܵpn@ ]õ*Î6,¹šI=ÏGW³ÈxŸ­§‚]Þv:Øu¥SÒY>s¤3q]½bë3Ò¤l2´ã‹Äñj¯AÖµ!ktüìèø_^62¬ ¾6Pl˜N£¢ñ ±9ø‘:"&ØÉçpQFÄXе‘þÂîZ»æê4}ç)W‹ï6ç«ø]v)mÀ£æi.^¦]xú€pA~ŸÌÑÆd6ˆ Äå}óŽN ¼ƒ_?Í ÆÅU¾ôYl5œ¨fõO9Ï;%5Ä”ÅÿBxÀù˜G¢"kA¤X]ÊÚXD\]Ø* „š1×ÅÀ2À6T­àp3™mˆ¡Já1î)¦5ž ¸·Œ]çÖȃó:àŒcš>š°2ëtH*Âú‘°ÃÄâDdK öq`D×ãªtâÙy0BÁºžt´º"®Ìx¶™7%ŠøUyxH ¤—E)¸‘ ±~MB¶ZRÓí*òæ9ÍïvHÎãÄT¸_DÛKŸ°ž¬ Šk°Ü´Ü’Û/H]ËÈžÒéÖKªé–Išûq*Øœˆ]бU“§xCÑê§cT9‰qxÉ6)²Rth›ÈÿE$eŠCÒåZ«cnR´VtN"Aƒ*2èœtÿ?Å™ïŸçN`Œh‰œ¦ïM/ uÖ)Jƒh P o»â¯?CáÖèÐîÉQwq&â†ÞжH´Æ7hÕÑ„Ïz1ì=’u#«PýðøLÏ3[ãÓ`WÄäKU‘ÿ†|ŸòNÓ­ƒ*]R&z:ðD%kÆЫ6³(k•‰°~_rl®õ“˜Æe…'¾RxÌźÄìù`UUkAx‡,èJ™u©^Ér.à>Áøä³…†oV*"¼bš3MRyjTô>æ?%V³Àª|ø]K¿‚¥„õYÙN/@4 ÑJ´ä’¾Žû;åkùn0)© ×´ÑÁ@Þ¶œt2Ây"ºzœ„˜2ÐìÛ} TÉu T©œg—ðÁbÜ©ûðöè»ÉðU~⾂`¾uXHÃWW[ψB¼‹85¿ÃÐYI—¡ãl¨X9À++i¶ÎÖak^`+#EÄ–Î(ñ¢¤Ä‚,Ü€+Q.]ê¯J:îæ`Ç‘<™’†»jy‘š] †Õ¨H!Ê%¦3χæ8'’“Ôd\R7hˆNͲßJ~JF’ŠÇ‡å¤,šoÈI)‡…³ËI)ÓΧ°÷æ0Y@”>¥[,˜fÏOu˜PxJP^m˜ ƒ>`kÞGÆzNÆ•È2'/š RàúM?t‚ŠnõÓêúê›/¿:Pî!f9è~ r´”æ [½=PFHGû;?eÄæ„ Šœþ–‘œ ¡è=CI¤“@œûo…bœåGà‡A(YÎUެùº—™4`ŒËpÖÙ/ä-rÞ,¯N?\œœäpàæ Ééû‹ÃܘÀ²*žíôç/Ÿ}÷§Ãœ _¸¨‚ÂcF™£‰grñûÃÊE’Ùþ"µt«›õÛ÷«Cf”éf¯š‘•^µÜò¨Ø*i‹Q¶…æ7G)×̾Jgø¼‘`w TÌDt\-ÞHÌÀN3p‘#KÑ`n(¤M &‰bL>:.ò|ÒTáÚTRv\¡Ö ùY hI ІPÞ¢A¡Ciã—ᘾLXÁäÏ?àÍ< Œˆ3íÃb8ÓËQÔÎ6Œ+*5û1 1v\¿‡ç&ÞI‡µ“Ðy zLU®š7šA¢¡µ,áPæ!ü¼’ûIå䤚¼ÄuÂûot•rªqõ9+2¨©Â°àØÏBa|1¶x Ñs¬Fà‚WSFjô^#@êÜ(¿«Z%)|^>—;ú±/—H£Òê¾ædvu叿xd¸pÙ´ôÕ`…i¹♉VsëmPnÿ]söÁcÑàj££ÝL3Û¡¾¹mTt*˜C}äƒ%€Y¶Êe.Àœílν¡.¦œÛíÚ™{ŽP¼lp pû¼ø8vAÏ‹± |*˜Á~(ŸüM…"²M7*¢¨:ÐÌ£Äcé(’»ÂÁ ýТÅh5ŸÆZ¬×Y¡/›.¡Êµ^+™çÔk©Ž8;Tüܳ ·°Íð {—9TYàD„¡ýý´I‘ž^ÜG ®¿²ÑÛ+¬ê“X&Dz в©r¨ÂœÀb<훉ǒ“»„oE¶*H÷§µ-)Ë~3Ò_ÎÑæg*–ÆÅi2ttÃ95€cÝm*±G€UZ!¦ú¤E vœô\ën´ÙÑc—I£7% ­£â¤!G¥·R¡¤Áµ1<Ÿ½á¤fy—z~ˆYŽ‹+¦d—‡C3¯ßS‡Ê` ÎJ0ã…‚C•oã³âË©'“„è&œO´wayÐŒY¯‘)w¶|S±ìïÞÊÎ 5;ß*7ô’x2סâ(Ôå!yy¡Àÿz¿;µ‚Þ$@ÜIÊ’uX» Dk”Œš…Ôê†ieôm~·Hpq|Ž]_¼RÜá°^Ÿï^1 Ñ ™sd_|dÎõ#=ǔʆ'zŒ7êHÙùÇø·^ïö2.,Êmxƒ§—ªi†±{b{ODÈõ’•ãZí™Á²^—ÌŠŸå–êŸ-ZU°ÀÐ- +f)—ˆp1elaáb¼¯'ÒøÝÝ¿]‡J76•©¸ižî"»:ß{§R£ûÁÂþÛEx¬ËÙY å)ú“Ë`Th’poLöÄ:=|Ï«±S/ IÐÁ•÷í©¦;0•ʨýfxÊ$!©OØ›T ´µ9C^B.Ñ`Þ飔Á^óXÿ@#²ªiî Ás‡ÑY·…q &±dˆIum[êy܉–Ù{1ÿ»ÞõÿÇa¶xµŸà¾ÿͰäIípYèø`oQø¤«šÚ¾˜ ¼¥_‚&*MöhÔ8º#ÜyÏ;B—ôÿ,CéÇŒ†ƒD#òÍ‚D£î^E(¼Ô™€×4bŠ:hùð¡Îë빆;ã•Cp¼ï*š£{pJÿÏØUˆ&FÍW×Ôï€(Äx„\Ý ;Æüc zE=£[·ä÷³‡cãZ0pmæØ$KŠO»x6Ä&PÊ iû4*r\á=ÉdM½]ÒÒyú‡´.;ÑÕ¢‚`ŽE Ck²èîPÃa’ʰâØé><‚aÎ5Lü±îŒÅnQ _¨ææT@S“Ï쨹™êÚulŠuJkN%‰°Þ¯´Lâ]”oÆð+˜†“©Þkˆ:¹MUY»°÷â-ñp“ÁÈ“:>Ýzg,ï}àK%dòUÀÍQ&ͰM5ƒ»Z9Þª> ú \tåvŒ>ä>5`ŸÛžX²T¸šþR.ÒïJ0*@˜LZÝP‹«ao¿OË9Á}UÞu‰|øO‹½|*ÜëE23áရ²ÎdHH26†ûöZºu¤ûÂ1xѦð¦H.wô¢šD¼ÆªÛR ¯°HœG¿?Zø>6äÚAû®u×N Ëþ÷±³9;X$0瑎ð©Iù¯âx):¹¯E‚µLÕ\‹oEìý™m@eê9ÅüÄU¤,|л룫Œ@jrÒó¡³BiC83(¿Qº¿G!ûŽ­ÂäjËÇV>ÕÃmw)·‘QPDcŠ ‚Á,DŠñûûü‰N×PA=J@ \YÖ-çÒ¬Sé{\ŠjWħ¤lµÄm=cçÜp'?—+d¤™fµ€ñŽA!<$¾jß­-@Äò.—ÊÁ,e cpÌ„z¹kÎÔË>e*bÆ?ÀæÆÀÜ=í¹nuè )»l$¹€z…"³.væî§ŒDÁ(Kz_ÿÝx«h¢„ŠÉ®ÍïH&ý É*A>ÜõVh¬ {eèW2™ý‰ôÙ{Y‹„à¾ð¼Bä’F+'8: Ù¬K°žÁØn¹4E¤Iì@útvªaw87>Dj‰EµŽ_ÛÎ7>„Žé.ÔvŸH^Ç™ØfkýÍ%Ïlgª@¶¼ØÆÕegÝOTœú`ÆtñF6Öâh0t¦ìPzd,•Õ"4d~V¨Á° SEx=Àê‹–ªyç jh¿L’ XÀí‡Ï_'ÇòÕjÝÛÝ»=SÆRQ\.¤7`i=¢=éË¿êVÑY_(Ù’Aа^D9o᾿[ók©•;’⃻õ”±þ5a"²™²™’Š"µˆ!@1$BIá÷&þÝå©,™œÛ¾Ph¸%Ò–®º™£„gË—åîaËÇ©næ pž§Rkž3ûz£Iò—Á/ÚÖ„“Væ&åuâžSµ1Eº'âÉ”ÿnFDNÞlM8îЊMÆŽ~G¼Ç†ß˜·SÿRÈ’.ŽIO‡ R‡®è ¾a)Ùg!ÒŽ¤l9†hŠ‚ýn¹•mÈ‹|U*P©ç»¯Óåðüógºéî”©Ž¢ypʱçw"¾¥ø*ëFî/½BŒHÈC?S­†–)ãèÊÚîJ“·®} fñµiL{À‡ÑN{ú‡fµ+%å±±…Û¹oBhü'Œ»žXås‚Š®ÛžôVÁÉ}xÒ»(ì¢.ÜA ìXÃ~oT¢OÇíX4‡ä‡bf—ÊdXƒû¹czнo”ä!Õ½=ìÝ·ˆy¿xÖhÇ‚pP•;V„ a xÊŠðÑ?kä÷"7:]p¨˜áž¤ÿ”Ôž™ï=±­OÄi{ºr½ß`¯Êu\2Ð#ö ³. "nûwi²ŒOí†"ý³%?.›‘m®?±t3ìâseút½cÃËÀ 6}¿ojEßÃüñÞµ¦~\ƒw ãKÓw×W›Ã\l·?}Ÿ¶á ŸFu?..6‡=Q]>; ÿù¨k»0ÀÐAúíëW'›‹÷=$¢ys{_Ð…‹‹»æýòííê)®ñA~ÃÐéÑÚkaÃ~q“réWÍ_®6«ë“Õ»Mè4nÂúM”Ié8¶?„?éü»oäaég:v8ÅÉÖÆY ¾lÖÜøÂ¼Oƒ@|;fæƒzÝÞÄðýiøjV+=Sð'¤Y$ÁzjûëÑÿD7 `endstream endobj 380 0 obj << /Filter /FlateDecode /Length 6249 >> stream xœÅ]ÝsÇq‡õ”‡<_•rˆ‰Õ|(Qªd—bÙ%J±„*§JRªNÀ„Eâ@)?øoO÷Ììn÷ìÌñ@IùAËÅÜLOOüúcÖ¯Vb+ÿ+ÿ={y$†è‚·«ú¿¯/Ž‚ˆƒ÷+´¼[+Ä`ÜJ«€¯_ožÁ¯½vð~¦­ÊãC´1:œàÕ‘L«­ÊÎ^®~zôéwÒÈU€uœY>;ʤÀ¯V^¸!j¿:}y´>¿Ü\ŸþíèËÓ£¿ÀZÚ˜èåjùpﵤ€-EËûqM–z¸mIá+4[êîlw÷Õxœ­8%kª­½ÞÞüxüãñ¼b}ÿÉ]\ lòÿ!ÛpÁZƒÜ6Ä(Pœ,Ȱ”׃\û(,+’r?©p¥Ö¿ýíñ‰UvJ®?á?̃YÃÞB€y¤ßEx'ƒR>èé]ÀwJ L8þéôÏ#×û¤™°Š@†r”²Ä‚¬®œsƒÖ1“ööøD Jˆhüú9>ƒ`Jg×—øìaíàÖgÇÊN#´Ö0ÆåA)Õú_›è[¿Ä×^G»Þ½N“œ^oóOáÖôõîg²èò|“&ŠÑÀ²·yþ,</„‘&2r7Wyˆ:Ïâ„БMBwt5¹Hýô»`(ËŒ‚‰À¶Ós`ÒÝÙ±•¼Þá/¥Æëõh92!FíÖ ãsžr0Ÿ›&ül¦e>Ï÷O©ÔÕd<†}|4ë錀¹ÿ7Ö“.öÈÖ“.…Öóqö¥PËTµ¯‡2Ö Öóz,Óiã @9èRl:sÊ,-çë?]Ýn_Ÿm¯oQÙh!¼_¿³·?<)¦2úµú)ÿ‰¿kýSRt jl8ýúèô_óôÒz6VXmK¯ì͢ǧKíHà˜`-ˆ³Ë{üfw‹æ 8¬Úð³¹0ÚO [¼­ÎóÙmŠ•ŒÀjœoó£´K›öج¯ca°Ì÷ŽZÌ…õ,?¥&ûy?‘¸.’A¢ÁLËG½þd&öÙåv^¦˜bÇL1X^Tûl9ÿšY"C4 3+qRÆÁÃñ‰1à7„\Ÿ.ýœ@,.‘ 2€[‚hvå­t°ñgð›7Åb1Rg¢3ë'Å= ¶ñj•Äl¹Þü\˜­Á?«°¾C窄,Ò F­{ ÑyC¦/‡ µfÂ11œüùeö•àÅll^ƒ¸!Kí21HR þ_ÀžfG¿Ö'»ÞAÑ>øìfgL“Ÿ¨õ†.úÙüÛ{¸Eœ¤™Übôaà1Ë-‚1ˆ&“%¼~}yuûH~1;+ºZòÌÌGcõƒ¸Æâ­èjÿ~ò8^XÛÁƒs¢K]¼|ùHñKÀ.ÎÄ ó$;‡Çq˜Q Js11íð#.†KÃgC?Ø¢ü<]É?Êá(Æ´ l¥Ÿè¡€Ì +8ç#‚çñÍý—R`øJo·—Ïoožnn__¾{œýØ_àgöù£¨”86œ­ôh€]K³”¯têqdQËgÎ…ñq°§˜( 7€ˆÔ3V?Q¶€ûŠÊfxAÁéÓííóÝyb °C1¬¿½ÚαÿÍíöz„§rýǧOç?½½¼}>ÿéÙå»íù dÿšÁyÿeZ`¤âÛŸÿ¶=»½¼›–Qëgo®àÍîª œs·yñfûReaS“BY€¤Ú¶§®Ñ·Kó7Ñ·S¶ƒ¾kÌ}‚ ¯¬Ï¿Bj$`0+…+c „{àmb/™÷ŸH,_0‘Ë€¹¯/9PJù€,¢o(>Û’?<É“ûØ›|JŽDL¶8 c fï(þn…pÒ»gcvÆìÅ¡&#tÀt¤áW²¯ë›Ž±YÿH7OùsJ¾€,Ø„»âQ Mÿ,náE´ôýݱµƒ’Ò Š…áÕ˜:(ÝsÆ(Ç+úýüz‹«AHŠ¡/iô”й$ô^Q3ðÄÔ§ßY ¬ªS¨N'ZfŸv¢à'óî h´õµ2‹ˆŠÑŽZž¯¶%1¨Éû1ppx›û*X=: x­Êù•ç¿Ó™0xÈÖƒ¾}I†__æùƒƒ#xMžéœgÕ ¿‰.ÓÐÿ˱ÆÅ¬,ò(à|írty~ŽñHHJqæG4ÆçÇ­Ü „pEÊÑa¤HU‚1Uª2 opf!óØRÐaý%1_ç!`ÐêH_k8ö >†1.K?¢vêm †Â^’\-‚)bâGcÿ_óOð->;i™îÒH˜A:#YhS)Ú…ž,Í.ÊpCôL Û6v—Œ›/>J|ÉgæèV*<é ]½¤Â­¥3ƒy7‚bô;Mg÷ËéùI gžÅç\KIðóæ0—R)pæH@î$ëdÁ3Z'jXv/1·¡°ª@ç`ÖgÃ4ÒMÐ`þ|Éždyj{W^Ãoq@ñ)¯/qg0ýÀÃý“‘j0&€ l1.ÔK#m˜J³€zóSI¤Ï;t¦ (žîMžS_ß’ÁÙÄKÁ}7Š[:s·¦é~˜J"Jä´…';îüs;*uÜ¥aHJÙu€»ì—A츓<ÏË[€a}ÿ…«ëJ¾:” u+ÌM*œvÍÝ/ÑÏ뎆PÞÈe&Y¬RîCÉ]•1tž_r^,ø.[èìK=”=ìG˜byµ„¿(=êÅ^ë´tÛKÿ¼³›ä%‹îÂLô%Û(­¥ÊrÄ¡"c@Œ6“œF¶ô–˜Œ‚tô”è¬Le«Y™ K;M#x—,š¨¹‹làd®SæÓËÈ vI¥À—±Äe’²Tžk™Þ4 (»î¨ú„<×ò9«â4¤ÂÀyý~æô¢>#jÿ44µ•Ð àªÙŒì‡¨\°„1 ŸO¶FA44 É̤œéëäF¨AìxÁž8fõŸO'IöÅÓÝãÒ›ü°ˆºž×E{Ü 1;b^oÀâ–çËíhf^‚ª5 4¥MÇÐçÊ­$b±l2ëÝk Q MW:¿D;‚Š"ØžŠšežpŸUT?a¸ª3©5Y´ðçfc*v_°°ŠÿBŒÞy!¾xqGõp7ÕÕÛtÌ.¥úqAŒm^&sTd¥ø’ZTþçy¨“¦çC¥ŠiaV… 04vëÿ< aHŠƒ4;À ›™:»í:®¨8º¶yr”öž¹¸«R#l¢q\”`¶CW1˪ð—õ÷3ëˆå ëÜNÞ þ2ê[a¾ª ‘¼³Ü|ÿ=Ò×ßPÑÎ~%˨úµ MðZÔ¬ïÉkª7 ÜW‚Uâê&G„»(óf&,…ª\4¸…\§ŒéüS` ‡þûѺ¦JéŽÇ EiÇ<ÆÝ˜x15x³¤_c¢(4)Ñ¥1²ÒéFq  rƳ¥nƒ…2£§zeØÎjg•;[ºõ[Zíæ ¦šó­‚«— m5 ÿrÞ}#„¥¼ŸpÏÆz³C&Z@ü!†Ã'4Jƒÿiæ@Éa¸Xo¯Nz°++tX|AþH,NÉÐXÀ tH­ØOþ (w>é|ʬeY˜cç%voIY­xY\„m2äIºÀ¤JžR°Xçiù)«Yx@ס™äº ¶ÀŠbyNÊ ‰££ÜW,‚ù X²ÿ ã™ø•ËZ‚ ÜV^©ÙT€ˆÛOú´_ß-¬ª oÈ/f«: ïeäò °Ùi™ÊšØ[„n´[^';&O³ôa9L0žuÏ«Kª yx§}:k!ËSѧÕ`úþ__½`½ªû#s±*‹sÊgî.JÊih÷˜Ç¡0“e)šÄ#%…Ä4´®J° ZD±òçÓÎ3nÓG ¾W \XbrMÅc’Ð/4æN³4Ü.ÆŠk©k[•£Þ§¾.ï|#’™*š89ÆK\ˆP ÿ¹…Ÿ°¯Äþ€°á#päû‡¿¥ÜÅÜCè‡>bcÑñ‰ "广Âfç\}•vY#Rk žiD®íËð6M}¨¤?%ªÇ,c:HÉÁµ$ÄÀìñÑé Ãò %6Ký£y–^¢8=tˆ]ô,»=ó‡œ%ÂøÄ ÔVÙ™Q/Yûxå·‘-F”$A²¢ÌkSSC­1oÙsƒÜ+ë*Ï 7g³½ ;f_-êgdh·{mCÿA Ás¤XÒwèÚ‚Š¬ÞB«åÖã¾P~HrXȆÝKxÌì' cÆ)ºœtÎþ³£t±6Â~ƒýB°¬Ú„’)ø¯ ËÖé2‹Áå‰Än~Ñ ù«Ø9mV˜ZÀ¼Ä<”^Ý‹‡XQcÆyJ+¤ºÄ„ÁRÛ‚ˆA(S(ÖMžÍB¹¡ˆ„p3øA¤¶.Æs„yÒEv=¥ªâÁQUM:7yilÛ\‡ÙlGC’ÈZèÄDÛ<7´«)F«x¥žä\eŽV)ÂÒä¬Å•)yAŒ¼ßl2wß­!ÒÁû,ôp¬ºÎŠ/¬M+#vnwT¹uÌDÇ{@è0 /êZ¦hɹ‡SГ3n^©¦´ú$¥eðïfËÕòBؤ­$ +öºpÂ"P—ÛÂÃ¥V„"¥_ˆÝä Ã÷`y NGÜæ¨R¡š€7. Åð8uØUR¶Úpiô38zµ×G¦¦€9¥ø;f•Ò[2Ó!G3¯¾£2{@x¿èhPâ2XΣ¿ÎìQë–)S(‘TÞùµ±,¤±tþåúäHâØµŠ«Æ½­GÓä/óh-üžÑyͪѬ¼½€8#ÚÛl0_5†¤~/¼H[×úöè5B9 Î;¡Sº ²[÷®` :÷[´ñ”A{ÖƒÌf`ÍjR1 ´ìÄ)`ý;}™˜[Ypf®#j˜ÄÂûÇÉÿ5U¤™·²aÒb$ÖÅRLµã Su® ®mŒDw¹Ojñð$׫‡ì’ƒõYtº-–V#ex<¢/aÑTëÇ*G²ZÍJˆ=šJGŸ_͹uÕ”’*YÙ¢– ø^‹}ÿ÷Ú5ÓÞõÂ_LGb(6é€4JŽ'V}dÕ$G½w1€ïªa§O=yïöv:K‡Ý !JÇ£çéÂÞèªÉ4e=1™òTB •{8𽬲éYîky¥jLM¦–W‡—i¡4¾F3† ;I°‹Z"¦ooè„3ÃOEVy$££ª…k¤¿ßÓ=É»p32,fDEejë™Ö£úØË^–…ßsáÖ¥ 7ÙNˆ&¶î1A.ü(NnÈ•S]P@—leÓMìÈý´ñxu§ä¤°¤Ñ]Ó¬æaÌíù»=‰¼¢+˜Œ0YºX£ªd=Q ÜÓ̯ߔÕÇìnû› xÇ8T·äÇðŒr9uÁëh,ø…‹Li\¤üp6±·Ädb:í>ÔÆ'`F#ÍNŸY—qn¬æ¤;_&& z˜,=±Õö¾Õ³GœO½«t™ßÌû)©gl˜ûÚßr˜ƒÔÛÌÿl*’RAVrܾüRxi™–ò¿9YÊñÞB²!…ŸNî½iVN¶¾­Ñºâ’Û¾ƒ§âS”+ó§ÒGW„mêºF@ÛC±¦›]˜éru†éõXt~Ìß„3½/˜ló„xÅéƒJÕ(Î @1…‰ÿP5mJîÇz¨¤†¸pº­…¥J¼WËìÈmž2,ž6»7‰` é¦õeï!Îο[òfjOµsq¤jo‡[Ã$lç¶i´^܃Íñó¾ks9‹‘að›åzÇ5žß‚¢cîw¶;ž6°õÉÈ,îõÞ_¨×íÎì˜m&}š}FC¾ˆ­Š¥PŸ+ {¹Ç{½­ …OruÙw/Õmó]õÝ=dó+ñ©Ò ²‘¤ûÃ÷q'ÆC³wD©II÷€x;×÷Î3ç°ûîØ`§3üK:o‚ 3Ùé‡Pýç™mv¯g£R¥Kª£ØüúUÙJ¹bg%ì2é»QƒVS¡j³ˆ{Šþí¹i(?ìªs’D=Ÿ¾Ü‡“XYcíGט“}«°×®Á<­ò*õ·uo1}¤?„ð Ú§üI©{¡Qy GïËÍ%P$^÷ïí¯ÚL×éýô<Ïîxœ’PµH_´xY뢩PÞ£z=·½^£uì±ìç „nn2ì ªÏG¾›÷´jao… á…¾{¶Ž;¼ö',÷‡Üm®ÒĸwÃ=ëÕ¸Iµ¯¯èžoPÐë'8!6Ô‘‘õ‹´ß‡ÝѤ2äƒÉ:(ü:ˆö•u+“oˆ¾·®dᘋŒn"ÑL¯i9HUe׺µü† )Ç 9§‚ÐùjœJþFñ ÷× ´1’ÛTÆ ß[ÐÇïÉkjù6kn~425ÌW¡|oÇ`ÃwûìFn¾Q˜naãX÷V“S”ºDeö&6m8Ѽ…¶éàrB: M˜JY(éâß÷`Pƒf}ƃõåcì¯ %#ÇÊæ; Ü»h]¦ÛÇÞ­‘CI©Xyh’4øŸóé[°;»k´ õÊ1ôC]kbyË/<Ñæï¾¬0kþZ¢^~ó¦ý±ôD±4]ù£2¦_ЙÚâÞJC:«ôYºœ¦ëëtì–ûžhuU²Çh…F´ü¦‰ÕRâ‚•àí}+\ÍûÍi2ÄÞ{ ¢Z&ÔQsð‡Áïó]8¤Î5Ýäט–XµÔÏ–þ¾/EËî[ÂXìaë}„áE¡"„ÛÒÓ1îõ·‰|É;D7<ÎÉTVÚ÷Ø…E¿O}³Ãº¹öÖïf£×žîÛÍVÝp‘qÁå~wzÃ%^à÷Ö7“PvÚ\r“«ÀO9Ô¬–™ZÝ›Iº *ª$ó3J"ö0D×5Kä½Pö½àRh²NÁµo§vܦS³æ&gÌ•ˆÉ¥àëô]–‘ØüB|n•ׯ†þõ:´4gÝòkþìDÊxá"ø‘å^ž1_T ZÖÀ¢Ùo^ß‘O·Ë÷dýsBôj¾@ˆN‹×ke.óF¯Ý¸a—f¾'6>ñPw7•>C\uÅÂ[ì‘gI’–¼&æªÓøµ(èp%£Ýò´Ö‚UãÞÓâ[{Ý«´J[5·î-.wló”8ÎwUþß)©ûY1Šf(ïÆ j±ò}W”¬J0–vöRdÀßáv‹¢­"dÈ÷ô;ÁÁ¤Ð…wŒ74þ.u6cºÐ Ó—Ö¥ ºøÄ']“š7i¹êdÕ‚èÇL«JgàlX£­5)MÇ?‹µIFRÆüm´Qê ‰ •«ÔAj“¾ì^-î™àÿ1EåÍ,C”(Mž“a…Â$¿è9™’rÿã÷»&_‰*Øå”ÜŽà`¶ìâÿm†Ë÷•œ¾Óú—£ÿˆôendstream endobj 381 0 obj << /Filter /FlateDecode /Length 5934 >> stream xœí=ModÇq:3>$âC.D‡¼ v^úûCŽØÆJ±±Il‰h`–är)‘œÉ•–|ªúã½ê~Ýáviøë Ñcwuuu}WuûûùŠá?éß'WGlu~ôý_Wé_'W«_ýÓWάüè0«ã×Gq_9 ŸÌÊrbõêøêhøÝúø[îÅÊŒÞJ‡ÃÍ(½a~µá£òŠ«ÕñéÑ7ÃõzÓ¼×Êþáø·0G9²„•L¯Š)/œã¼÷VÄ9Zrt°¢ÔÂè¿^o„³Ã/ÖnÌè˜þ!HƸ×U¹Œ&/×ë„OΉáõ¦9eÕð!hï¹Ñqé÷ þVŒã†;üíÓNøÛzçœva‚aLúa\o”ô€”~HÁ˜²§J®FÏÝp5ã:l¿[ ȩጬsÄ=¦%¹ãÛˆ•õ°ÊY'aâ—Æœ‘ÃóiA§†p Á¸Œ‹Hï•’Ã ™w´€Ñ›Ä6 1¦žFx¦Äi'2n ‚„- ï¼ÞÎð.Ø0I*ˆ px±‰ÜÁ¶®iØŸŸHÇdü¨˜ñ¸"؇ó\ ÷‹°°Êáqæ…·—〛ç\i †£àpRá,œ¥¾X#_ ïG"‘8iw“ˆâdû{x»¤…¤º8û AzøKq°ïÈÙŸŸ7è$ôð*†vNé9Ÿwþ€{ÛHÃAžùèµæKáŒFñ ¤Ad•£ã~ø1l“1¯læÔ ?À!pÀ!|ØR¹ŠÔw pX÷ä÷ÿA¨„Bq§Ë#F »žÅ•$0ñ-ásBÿ¼&—Òâàp ” Ãçy·8.ÄD1OGP†Û½ ga¸îËÙÄA‹¿Í;òÃîu\ÔtåýYìr j5 §rBé`Øçn &BÞÛÁv¹iú†;C(…´D&“JZ»,@0 LóNO:7piZÜÀ*¨ƒ&ù¡AÁ$.PÌ @LC¶·5Ìôû4¢¨üŽ¢¯wóÌ(3i?¥Ìü{æ[¦M!„`¸˜0 䕲<µ7•L °Ã¦žf(lQ weRh¸| û­'[i¬Èƒ )ËK¤Î°?WÏKÝ;óö4Ú"G´6˜ –3Žàd ›4épžai“‘ãwd™Æ}pCqò‚=ÚQ[ç'‚PA¸Î‹¤;M¦4*OÄ3™+YfYRðd ¨ãÒ` #àì 5EÀ 8O0χŸÍàNÛœ‡:ߌXªžî#ÂÖ•FØž ˆ=m#õyá 6l©HQ‰Ú]ù§¢Yª\ SfxÔQ5­'/ý D×ñ6…‚¦Zp/†ÿ¤ˆõ Ÿu´á^ wgùõñøo‰øG# ¸ðM4Ò ™¿ƒé"âMÅ8‚7Ò%Ùð!“ÒÔ*úv&Ëh^’¨µt¥ä…J=%SßuÀ,|N§££H8Ѝ-I°M=`&¥v¹ô‚g#é°ß öÁŒõ ñ«® §BÀ:=§ì¶Ö ù¼KÜÈ•ÈÇp$üCvÑ0ACͪÜ*áÉGìè*[ ’jj²é˜óq½˜˜Þ•`"/CËšd™LÈî.~÷pf=“LM>ÙIX©Gc*{›‰‰.*¢ƒ^:8?Û¿èØÕ*¦ÃñÎÙ¾[‰ Zí‹ãÛ@íµï ~Oî+ŒK¹àïáoóV¦–™‰Œ5‹ÑØ,ò)åBt„hÃÒi™–Ä;0Bwͪ6$EæÏ¯« d„-s×PjTív1%DbhÓdŒª™‡ LøÞQ¨|½‰1ÎbBµ ´"”g%8)ÎZDY ¿ø¼vð}–¦ [Ð}ð~ø×-Õi×q¼ëC¯¨0ÀÔyE^5ÅAÓ黵pèrâ8ÑÓ›ÛÌôànœïúz ™Hp~Àý+E.“afx4–÷1Ó '=–D)q¯È“¨èÁêŽ`:ÀG Æ ÿÌròŒ)MÈÜFðöúÃq€àú0ÙcZc– ‰ÄPצ qršrpИopö€lJa欠þº±‰Æ€)î÷™Ï@™ÚŸ8«0p%÷†-|ð=ް” ˜ ¡ Ý·•5žÔq&¿)qNÒ6Y9s¡ž1·®ØÝ"ÏÐÚœYåišjb£ ÃîÁ@>HÙ…A•ór7W1JfGMùÊ-¥B´QpÜRÔñÑ–2¼PáQ€RáÓÁ‡©pgc–wš2β0‡ÀYÙÉ”+G^_¦‘‚«Ç‘¢=e›(¤hš¥|¦ÐùÙå ¾_vRMYÁžÛAùí6îíw¡-ú{6æU!Ö®Þ™?Äk¤"Ì&¨Yпš%ì>w}‘%Qö–F\Ô!%¸<›?·ØMª,ùA1’“ı€öDø€ƒ–s„‹»Œó¹€]¦±Ùì–Rž¾ò4ºa˜–Euà¬ÜÇ6g0; ¾f:‚¬îÊLõl®ÑÚcB"¥_¹kö;öà;1[ I{3žù®zÜÖújNõW“¢Ö ™EôA}ùΦwA;3^¸gŸÍ‹4N~ÕµÇW) ”0#]”«2k•¾èeäFÀÚì(ªbÕ祫y6ÀMÕÇÇê0yy0¯5¶£5XÃð®Üê'3×%dŠüþü92gNþ~ŸpC„Ä,“°cc‚­¼Ž¿=z~|ô{g@àÇ#¶úòHHQ†\i`èѨÕÕ‘PN|þryôõ㊚rTÒ>”«šS±Ò)o'. ;¶Âäðr§¾Ye£¸ppª˜ãsÂÅIÿÄ TÞ⠲⠣aÏÃ9’„¤+Ǫ.|2:Ÿ/Ï·ÌÓŒÊáùòÑjkøÛTÓUSrÍi$Å 1ö XÁ(ÓßMya^p¿Úó¥Å2£2ÓC¯A\3êà¿¿8:þÇ?©|™FP Ç`„KÂú˜70Ò|4ßEåèŒj)*™D«¬«¯¢û²|*Ü*~Ånœ¾|Äyâß 8F52m)_#åJ²òY¾>}ˆÓ”c°nµ8Âë.GA”¦Ê­ øªOÐ^]õ©Q_Q)µ•¤3$ûÿuæÃ:ó.ÎQ€M }3T’Jˆƒ•$²îŸ’Ü#K% ÇóÓ•äÅì7•¤ä¬Ó#•$1N†L¥‘2„–ñ¶J*ÐßH%GnAñ)‡,¿¬UÑ<ø6»ïBóFÜRl4hNÁö ÑŠu£% .G7Ȧ1ñïW–—‹Òëç˜1s$¯ñ,·%¹n{I™ Âú‡øõMŒ£L™À¢M¯2J£g’6ÚÕ‰æ9ØFªaÀVwEñ9£¹l©Ha&FKO” Ð1@ÓØüflÙîDÒÜ4îùLA¨V"††¨Ò‹âyAÞŶS|ù†DµW 3½ÌØîmƒÔLÁ.…HƒÍëu#Ì‘Ø6Á’¾‹Ê¼ÒAž¦rŒàxÈŒ/"886fy¥¶huƒ]!%ªÃCQ nË~”j…nØ´~éq«¢ÓRè ”6ÒnQÊ$2™è¥fÃŽI^C™Å¤ŒVÊîš",ªÙ{ö¥³ŽX¶ïÕÊ{uºmÕ=€p°„–Íh[/Ù·íe'Ï"šàÉtKÏ©yÓ• \4Bk°õÁíÚ´ÎOIõ{ìâ‡f"#•ø"­æco^¥'=6§·ýyFšV€+ÜÜŽÔBì· ¨æ*áÝì²&¬â¹\Í‹¶ü4?J3&J¸"!—NHNÇ)÷V\°ãdO­âÑg=¢T/xúŽ»xúªÅù;v 2Z•·ä€ŠÌã²J¦ —„ÁYkU îù ò»EdøÛìL¢Á—àsõ‹ÞRb‘/Zæz.ÍEEc<í6m4e„1©%¶€‰~ À&Fø…h„MéYlráéªo†Êc—Ž4Ëöæ}™©.£;Šý©„N*Ɔ؃ˆ£ñ"#afÒ~—ónÉ‚"tø«u+^IÀ7ú U. M˜Ÿ¬5Öoœ)(D‚çxÚQGPZQ®ú¸"ð‰ó8^$­Xxè 8ým5òÊËŒ¥ÓÑr«êq¿)‘W˜àv¥+M;Í* ˆÃMÙŠúÁ‘2ETRÙï" ìýßÓôÍB-´>wrí–4knð«K¯é­€N,ôY‚Â.Шì’:Ùý‚ËŠ¦Cë/Þ»ƒÂ²ážk}R_¢XµBLÃQù b«¯+Y‰­ÈåMf^÷§( ¥ñlýÔî Æœ×I&â9£Ø\“×b×óIn#<Œ5:!JpŽlq¥$3EÙ9¸·Ÿâ“ÛÌN¢d4"N©Âq;¹…Y‰Üÿ±-­ÛÝeÁšFVftKc{zE‚3j¨ïºë3ê~nuÅ|#Øa¦ é¦pÉ©Ä<¾«1^ª€ÓÙàn'û¤]7Â#.Ìdþ®VÒuøRØä¯\ D@ˆ&jËܳŽËœû>±„åXJ"¯ÍáÑ"9VA.b©·õàª~m˜ùb`”œ#Åx])4ö{] Êé,§ú§:I$¼ç¶Pgñ3†S×e_# Ãoç“ÞÌJ¦sËdÁ\qÈg o `„°ËÕF`i؈"õïËéÿb¯Ï¼ÆŽM3Ù>'Ýi¢=§ ¬µ“ºcÍ–ÊEïù²KV›WÙÙR…Lö»€Ëöb{ÞM0Oa²ÍgDqZýŽ×عùØh\/‚Hw¨FGüÐÈPòÅ¿£D¥ëµï€ù¬‡BÒ˜T6z1o €!L;!°$<—ö­U´÷ îí¡Ã!¶:d6~¸Õô÷,[|ÝêàÑËÄiRÆ9ofñoë-DÆ>„¬pŠ¢=¢ltqa’qß ùê4mܨ„—û%\¡sþUçÔe°PÅ:…#ûPgíišÆ>"˜æ|H$`#¿T*wùnÊû½œ‡/{X“ œ^toÑdaR#iž±Œ+Z½-ƒüUÍ ®ìà¡ijSÌYô«VeTox¾Æž9µê^ŒÍËæó‘W/Úòc±‰£@碖­Vøf‹ÂØ-‚xUù1&p’†³õ v8^K@gžzxN.¼ìb‹SÓ7œž²I¥ƒ€K‘Bæô[!¬ öâ«ì1=#N!!ñ6ω€ƒôÆ\q(0‰F¤ÖgSYÑ*9¡–F_ßW "P ÍXLUÙT)Hßéøëó¸SÍT1æÍYüŽo) 2™~¿"ߣÂÜ™,¥ßZ(˜^‹=(>&òÓð×é3^’Íçá‡k:ä¦Þz³Ë´C ñ)Ýúm2Ö:ß1jÀ¤ã_½ë}Görݪrp¼Ö8¥·¾ˆßÄn ÇŒ…¡¶ŸSg‚B®¼ô9ÞSÿ§—µìUб.ˆGk¾,x±Âë^E&ó:"€ODôàmÃe PhÕ%¹nO+V{§U#rà¼ÎÄí „©vk Ÿ´EéˆÖˆÞ稥ƒA˰I¡Ÿ†[pÞô’ú·‘ Ò‰¼l€™×vNÚ…§k}^ô²zX'eÝmh' §ÎË-$»ß,ÔÁáÏ÷‘¾Xƒ¹F·oøedÐPVAðøšü~¾Ï}”hõÊjvà[®Ë¦–R§ ïÜT­nƇ³A|=ï÷Ðj,l6˜poiÐЂþFGßÙp÷xïUwú²ÑO¿ê. ÍŽ\ËmØm0^ï—$¢.<–Ê„›pãp®q*tú/qA|4Ççä©¡*Ìq.¼pÖÉ?¢u¢Ÿø’K˜yã& •ÞÆ#ÁÏ…Öê>ãæl¸%ß¿?ŒU7'¼ëlf›nºÛ`t'—¯NÏïîà2øêYï1ª1δÎí-L¬Üv^.«Št8uKñ.Ð|‹Œ*€ê©¸¯ub/`vD)ä§A™òª (‹,íá =´ÔGšo‚@BÀ2r¼¹FrñT~Œ KV¿»„ßuY[¯êâ K§ fN÷—-­ña¶ª¶×çÝ2ÿ„Ó¨–>è‰P¥CÃiyAWÅž¦3GŠß׎‰ä?‚m¸rnѧæú~†ý6ØöøPH+)F¦æG«*M€ÛÁVã^½ü.îdzâ-§Ýº­„ CÍ}6Ïg⿈ˆwzèñ)¿ÙµóAñ}ÍgˆÓThlÉÎY#­ÿ(wG ¥ö<€AÒDË-tJ›³~oé¹m¦,«RÎË®ÐÉ5Ö¹øòˆ&9îZNH /uP÷Óeø¥[1ßHUÝm»t£Cr³j(:ÓbÆß*ÝšPg-²-dsá”-mø,ÊCŒ/Ùù2Oº£¹®Å¸fdf¾“Pа©b¨ éPqœw„+†kX€sxû þ÷Íù‘c~‹¯XbyM‡®( ¸­nÎŽ0Ù'öS¯–?núÏø¢;ë`!£èõ­ôޝ0 ï…á3¾wgwlº0ư"‡WbñFdÂâÐLÞ<~)Îðvš¤kýó†¬dehZø+ jVÓ•NÈBz&`ª–,ôr +Ép¨hlèXÌ_¿h¦èBWïžfK`j1 +}þ4‡¤è:[lé ”Ë“lʆ×>ËszÏ_®Ÿ=Éa 0f¦`¿Û‹ó'Ù˜ ÜÓ–øT¹¥Ó§Ù˜@ :xqT/×± A’T2™áTS¸9»ýÐíO¢µ¸ 5>¡Ð4‹'U[IÆÉJçÅ®ž@ÊÉZ/‡sþì=† û£ï0[›ð2Rb£wWWÛ›û§Ùe>;²ÞËÄ,‰°LŠåú ¾7(tðê†O?Å¿åà+&=áÛF 0‚¿Þ^^~ü‰ —ƒàè^bx¯Ñ3ƒëóô'ð8Îé» çïdÈ{:$˜äˆÏ£ôŽjÂ$`‹XüÃãû·g3 ̧ä¿|ùüÅgà2͆ç/¬6„_ïÎ^¿¾8¹8»¾»»Ï#$‡€ìç·wWÛ;t3!B±møîtœW}~sƒ‰\jŒ½Ýy#?l/ßá4mñ¿‡ßݼþåwL{«1Ó·œ±Ÿ–J„‰Hª É9‡?3]Ôƒý@Yâ|Üþ׌j: 1ðÑi-%…Ì<“l† šW++Dr›„/¶W¯N··ûx‡=ûü h¡ðcý*Bº»5º;á…´36ëTiÁÖëÙÑÝó+f+tící‘<'€ ¥W þd"+ä¤~ô ÓãQendstream endobj 382 0 obj << /Filter /FlateDecode /Length 4274 >> stream xœÍ\Ý· ¿ýÎCg‹ìDßAóà ’ êôÚ<Ä.°¹]¯ûv»óWÿú’’f†ÒÌìÙÎi]äÁ›9HQ$IÍï ÖòÃÿÒ¿—gl±;ûýŒ‡§‹ôÏÅåâ«ó³ÏRná[o„Yœ?>‹oð…“ðÈ,,‡I¬^œ_žýÒ¼^®àÿ¼·B4Oñ·cŒsÑÜ„ßÿÐK3—[òø:¾ÊWÍ‹0»vNeC.ȧdöu—ÒfdßFRŽë({Á˜W6{7ÈÞÜ|½hþFFÿëëA|m'a×ü¼t¢ÜXd߯"‡‰Ûõ>>×ÎgÏIŽ\ \(ÊdშajÆœ‘qNX,ç‡ýðíòÑù÷gRŸÜ‚Êo@ÃÂÆ;Åe·AZT¸/É$IpÒ:Ké\§urãÞ0Õ+L9šÈmC榣? òôÎÛŒ­bÅýóýRXømƒiàÜ^€á™öNfR8P*T?’Þï3ÕB™­:¡­À8½öQt‡ÇHJÂĪy†´‹BD9Ø·«%¼å˜ö³ÆBmñ‹AÇæ×çg΀šqà ʯvgŽùÖÚ…Ò úø¯2 !=p$WÛ³ÇðºTÊ[ð0£WóÞ‰+pFè‹Ô„{2Ö$ º§æj{-–ç¿u¬––HBK0sü²òæýiqÇdFì¯+BÊJ#…»RôPèŒÔîò’к;r©[«lFëa³SŸQj`Ä WfŒ]\ôOÞŸšV­¶9µ7uÖ¥=è¹Ê(}ùFßüòY¦Z+Ü)îÍ3ÔemœczzÔ¶ŸÒS©0êý'õ"LF'ý¢ÿÞ´^åêm«ì x¤ÖäÊý¨Ž¶ Î[Vz¶½ÚoŸ×Y˜ôˆ>W9B‰{ðݸGÊK~!xYë>„’R-ã¹Ý;¿z¹¿Xßl7÷ê¬ÎkD‚ÜQ,«ø¿Î­0apHéâ°}\ÇO0ËÒ±‡EââVܵNx 0 á±Ð.âì§Ÿ.WZè– ÿÛßl¯.¶/nà•à€zEDg£`¡mÄ£ø§ü¹oä£wbqþÃÙù_âüœ) é ]˜ä†[x¦­nx+ؽPÝc4ƒ€‰ë8[ÇëåòÏ$¤Œ= RÕ‘Ð: j•‘P: zwˆdÒÚ€HHÕDB¨6 R§Dªru‘Pº÷Õúêæùö概,óµU†Ãào݉à p(§á°&ú#`(œ `h½2¦Cï¸õ ;,T5±PkÓÂaýXHIÕÆBJ«>Rju±R:Rzw†…tÒÊXHIUÅBJ¨2RR'ÀÂLåªb!¥tïë«ÿn÷•ÐÁžÙ\ëB¡Vè—Nƒ„„VBõ„ZO€ÐÃ`™!çÜi6„ºF©)˼*rpJÜe¤ê!8 •‘ªˆƒ°yf Ôjá`D\J©. 0žÑûã8˜À•NZ #äRRup°C\J©2Ƙ˯**,²å:WäRJ÷Î_>Û¾]}»ÞïŸîwuðh‚ŸÍEU8Tø¯§ÁCJ,¢žD#>" ç$ADg”¶"2 £–CAó}»"¬Æf»C׸{ xÙüXv>ÛÉ¬ë ‰NµTÄ×qJXØ(`¤ÜlÄÚ-Q°7ç™:àÜwÈ|ïmƧDÔE“Ñ-Õáqm7pV"Go…ŒŸÄÔo#ÖÏV*© ;Òm³€•Qn„s¶Œ*âF5Þÿšpªb½Eɯg ð0Ûæ'¯8 Ë;ŽJ™z‡ãÑ5J§×ð{Ö Æ/nn]pJ“<õRb0C#Šøö×…<4:`÷î~a;ESJˆ;©Ý‡Jœ#ÿ‡Í H&Óz÷ÿ¨?ЇOs´ƒSº²ÁáÛˆ‡…±…IÙðõîùЀcš£$9Ð!sñ1Äbçj¡Y‡øÓÅG.z` A¶ä9ñm·q v‚èDRБfšrœ\»ñøõ ®qúšy©œ8¹ö±ù*þ6ÍÛ˜-ï#Þq©ü„Ù뙳8]ÎÏÝ$&s@˜QË¢4A˰m=êvCôvXE³Èjn k¢ìÝÁÉÚ\©mì–­R@ìÖ·üúšÊ—Ä€²?üH{ûšîÛÛ£Ÿ‡ ß^¿ƒÄïMt7€$}°Dm2Ìò¹MG‡\%…µ¡­íw¢dÄa>Eo¢ŽÏ©¤f0Òæø{b0³« /w ÌN²ôù»Ê9z¹ 9ãIY÷b” ÝTëø¼8B F3AT )µ‘Art¯qP€g<‹æ~GSÎâ µ²[Bù!uãÀ>Yn¡s©®MdAcÙ·z&¿,t$_Fûr@wž‘ ¾³H9d>á”b¹ÈâÊu:iw …Õ,´¼;Œ?V&ïRbÀØš6Ý™×63–2Óô¹‹¬`‹èíg Ã‹‚‘ã±á„E ¼Âbûî¬ñ¯–?ÔŠšÙc`¥;?%êqCœÖöKò2é„y*p÷¡ 9Çd°Ø.?9¡ðsÉ×âÄ+5Ôùë–L&V:å›Ñ)03ã4¢øoÇØµŸÍüBð5@w:ß¾ý¯&M6 †à(Q±˜¥ž ²WQß1ÈŽXÌ|+„î,Ù“£¥¼µêÝÔÊë\ zZKÌ Ëpœ&þ{Éy4Üû?- A §×¹ë`¢s'ÑMnXv¶F·`16´€äëòdÆØþ#ã³_ÆÃÄh8…ñâl†ï!$ü±³Y·´!çƒ •Éß ©‘>'®çÂÒþÅ\£¦B,aAñe®ûfÉoa»N’?GñObã_÷"™°HüŒž÷=Œ·‰%P’ûÄ{R;Š£©ì‘¹‚QÈ‹G=8rýL“¡ú:…`!‹À'}÷ ÿɪ –¶ÀGŠ‰Ù¡%~’‡¬žTØ’=†‚›”¡.°ÆŸý™«·Š ™ çÄôE͹ÏeŽ[©‡¶ö¬*ÖÛðbxYˆ£•’w(ÄI,}ÞU xp‚ÈUñáËéÒ'Å–º'8ŒëÒé’_ÈÈígÇÛ:Ht-Ü…îl,óÕRAÍ`|<ã 8P(©; ¹îèÉ쓌O¦^ ¡}˜çòÀGÃÏN ã*Ñ\•/Ū³.Þ÷ˆÌòùëëQ·ú8wC«žôÝă±>å=ÉáSð„ž¦ù ±‚™°mªÆ™¾%‹×(³ñ{­X˜Íé99i´em³è…oQ]]¼Oq Ï?ÊO÷ñ u©ÒÈ;o\Xu•»M·a÷Cð9¸Ä"Ö–¡Ö¸&¶Fwúùr2³š>!*½£o$ˆƒ°C¤[æî²4ûa÷K)üÙ—è°­1©dšß¦a:F^̳¬`—ŒBº=ÃSÆeYû‘øñêÀ–Wƒ§zô‰é¹~Cרú¢Â±dÉzç sÎîG…žÎfiµŸ¾\\å’ÇèŽíèÑü,œÙW"d•CÜÙð¡¡éÁÙÿåÄéÙendstream endobj 383 0 obj << /Filter /FlateDecode /Length 6216 >> stream xœ½=ÙrÇ‘Ž}ØX±ß0A?¨gƒhÖ}0VŽe]¶(ɽR¬©p fš¬D>øÛ7³Ž®¬žn„Ñ =¨ÑS•WåY]•üeÆZ>cø_úÿruÄf¯Ž~9âáí,ýo¹š=y~ôè™r3ßz#ÌìùùQœÁgNÂ+3³€X={¾:ú{óëüþòÞ Ñ\à³aLúæ ¥÷JÉf^{o¤‹ïcƸŸžÿ9C ×j-ÝóS€ý,aŒ{{13­·Òá`Ûj£˜™óVyÅUœ³Ç9Ú;£xœSqbZm°PM¹Ä)ÐpÑìæÇ’¹V:×,ærÛ:ÎD¦ð›ÀŽ÷žÛ4¸’k­—yì¿4”ùVià3æH‡m½“ÍÿÍ…mã’ ¯Y’çÀ³ó^šf³ Ï›ÍyÒh6/ËÚìÈÚ ¬“S<#åœã70à4Â3ŒWcYÜyEI¹)éèuEH¶o6ëLªÁ·@i¡y€ Ì̬½ƒ·Ë9<|ù3þn½sÎT\=Œ¢æRÚ¡54¾5¢R2á·#:¦…ìéØ* …•4ú-úÆe ?¸}KTŸGh`†0{‚òKBÉÀì»,¢Ûä¶Z¢Ó"êmbÄ™UŒ¤ÜRèðu&Ütz†z±(z{=2s33gLóæ7"ˆ×žãA&È¥:}¶ÜÝ" L8®À±°­ÖŒ·^kNÖ,ú’f‘Éu4ފʼø€)>Ð/9§5DÕÏ#9bsEÜ u-›‡…Ý­´ìÍû»ø>ø)Ì d0ðÅVšAä³jJß÷Ϩi„·Ú©KçQÓ*^£S•½Žé$U3ªÃIàžÕÃ5H"Ѭ#)Ú‰ !Ö ìÚÁšØÁß…ÂUD.¡yUtµç…é¼bN®:§’ý'H* JtÜuÍK¢^Ä7žhyaо(–Ób(CÁT«V*•êÁm~N›2Ì÷ôiŸàs5±l+! Ò&i-Úd‚SÀµuþ0˜EXF‘BŽ +Ÿ¿-÷ÑÉÀ³aµ7£+PyòGp Òð–»¾[ÈŠùKgÿLψ‚ºä 䪘‹™`̃âSsH 0í4ê#7ÜV;©•Dyh ÞttA©§ðD+–µ×GØÒËÚ¹Qü4õ¢0‹s-ðß ŠÁ%x e‡c4èÉV3Ëd­íüX3›#ã:’§}mm»áDƒ2–$›‡ô㟖!r½Ä)ð!üÀKÈS²_Ç×΂2¢Ì@w¿þžXáÃøÞ± KÎ %*) j»>?陯ÙÄ„t–Rµ©ÎŒžº’¤h}È¿Í1Yç\é:¯t­¬[%Ù|¹AáxÏ h:-ð'*j’)1j‹;&«ù½Ê&è£f|At4ËÍž…ðHÆÔ+"ÙhH–„Z#‘ë€_ÔàÛ¥w•_ ´µc|×"²^‘AsôÈpV §Žð-¬CÎnB:ÑKJz ^WnêÔÏÔÙ.öbWBÑ'ª¹K$ ÇøÊ FZ ?LC©¤€Z93ºziU1MGDŽÕNnsàßò s1áO õ äJÂým•ÆP_„ĨœÁQç£úuL±_Ž(‹ùbÄ’t!¡° “ã©.ª›ðPþØ®€ù@aÊ4'ß‘ÑÙ€ï渦b R‹3ýø~Ò-›éG×ÃÒcë’<÷JW±{f +¦ AÎ\ÿ¾x¡º ÖIº®ô>Hs®~3¼ŒèhÆÞ÷©}*oÞó;0ÅäË6èL½=“ÍWaDÒ”ûi(-z*²ŽàÍʽîåÐ#ª¹qðRú:Ra¿§KUŸ&©¯†'{[…ëbx$¶ R'i©íRZ¯â<Àù¶\,a â~Ñt¯7¡ #Sü÷¤±oÛ 9´ëípáe*~W‡ÍþîEˆVîÉ;Ü"Ke¢5 _À½QNÕ4½˜'aû˜Ã î@Â^9–5F/˜çyc÷Sî¤Î nÉwp¦€eªÊ¥¼ŒTAûù…4P ÕsB½À%%æ~ºg XËiÕa„0²^×Up‡U^‚ûÌ›:ƒ…4¡ Ø­»ô†8¦^¾ƒƒ!s¾ $w¿çý !šÍÛ½uÅK/I½%iâš·® 9_ö”N¢t^9.tT:,ÃÈqò! QBbH«¦<.Ô PÕ³ÖÅ~_rƒÁ­n¨<øL» ®ËÚË1·hФA’9G}NüK/;C¥Ã-‰SòºªÁX XÜÚªý…Sç\ Ú+\ª´¯‹Ã©«¶oÖ-6i–E­nÄ«žf¤®¯ÞÒ²P~‹îé4ŠŽš-–¨³†ÃÛŸK©ûŠ*dæ·èex‘M[à€Š)ÊÌ‚’M£úë8ÜññLò½·¯‘ã¼;R ©4n(rW•M¨ZR€ç&÷fÑEŠƒCîH´¾vtcÉîÁÊáþmZÚ°µÒ-­>Œóe— ª&,›Æº··Çž9¦›´UBœytÊҵʘڞþœäŒÛ.‹ ºr•ʃ\TøôÒp`$þá³btU¼m>Ÿ;ÑŠÞÇ’—î@Åg¡¨Fk}ˆú‹Ûòa(¤@2ú°ôI¶WWfE¨Ç%ç9Ö9jžg‘2éùmû¤qˆéÇ>$QßjOyâȈƒ‡T4¿ííëu”ï²äâ+c bIëf´ éÁG˜°öÀ[°*+A/ººà4‡X@Œ}ç yjtT¬fx ¼xz,èĪPoyþXEHàÁþ 8ªæ@à1Pɹ\RwAÌŠ˜n°$äK0][Rþô`­-Ÿ¬PU ‘¶IŒM”iðèÒqK>Ä‘2«ÈN<÷ßB†£è÷“±R‰è̆¾'¥F?Daç9õÒc‰M?1SPTô]‡Åˆ‰,—¸¿¤ c„D«F'2ø`fùŽ~8À•ÃÙ­ [î(ÕF·™P2£^tÉXÆdû¦RAvà„ ³¦œo>`F  ]eúnfD‹€ÏžýõÄh¤‘ýÿo_9áÕ΄t@‹œ) ÀÎ ûÿ>ÛžÃl+Z¶RK(@ñ Î0vbKF‡ç#ÔÀ‘ ÜÀÌóÕQsºØ/æÏÿ7S*2<;|xo\“Áºè Ù‹æóËõb½<ƒpƒ§uXpÊ[¨1°’ZÞç^BêÆüL;¬o| æ¿Ž'bl„É UbûïoZ+d Òr徭RxŒ «dQaäUu œÚ æã‰ÈÇ+¶F…v:Éâ­mÈš€‰a\Õ‡±.‚ðþ̋ھªÉ ŒàšØÂ¦Ûó‰MŒ ·±7Ó&¸x˜*¨»¿*¢Kp ªÅ®]-öÛËߦp21‚ð-îjdIó Ø‰œ’ÁOŽ¢Â´=OE+Ú%±.*ÉÄIN:û2¥–dGLPM¬%ÙiŒ/ªtñŽz’kw"=Éþ‘`*zâ[?Ÿ’x}¶›DOºâ]rÜ ŸVOÀyrW¡zµZM#_ÀÁT…êõÇwÔK×ÂúçD~0úÓ› RÆÏñ­1Ö?ëÞÜÝ;\ÙmöŽÓ'oVQíï]ßò˜I-7gçÓp–͈ {F{°ÿå[/LÚCüÃæ`ãò¼k~xú—¼h¾Zï϶˳ë=|ˆ?ùæo_}?öÓ·ÏžŒýôôä9ýi¾r-iðLîXç‒¯žÿg¤ÓÍ1Ãjn¡F—á•BhøVAŠi››KqÀK©´b°BõKæ™ÒbF(ð¬*=E¨…ÆÍÉ““û|€ƒ×¸‚ ‰«, ú7Šq'FSÄÔg@1Îð^¡RÍ pÄËKމ©Ðà=}©0ãæBvü° RÄÈq=K=r$× Y9(¥½ÎP‚À§[ ‚‰ÒBMÑ(­æÙüãéòHkàý@©vÿy$Eµ|y¹>Ò÷Sd/šiø’ºµPQT§—‹W“&=£|MP*R\í†Àk®æ§Y®t™ ׄYª`¼5úƒ¬—`Äã*\Óöè·¬ÇÕDÛꂃuYù”P@ÔVîƒð(¸s„'¨¤}Oýæ9ôÒ _UËIw+,&“d·‚¢Úž]Oã¶Ø6DgŠk2gŒ'5eÍ×D»v)rDS}2 ‘ÓU¨& 0©f«VjÒ‚C1Õ27à¯.×g‹í—¯¯7û‹³Ýån¢Ì Ek‚8Ÿ=\NS3æØF0îÏvû‰<úL[!ûd’½d!‚ª0=øôâr÷Ë4ÛÉÂXl\ÎLA`¾)Tý³<¦j³(HM+×±`~v¶kÿt^ÔŸó.DPö‚ó»í‹æá]Ú4èÓµ³`>Ô¯z‡ˆ2DáÞAÞŸÐD²­QR”͆ 7ü¼œ«zߎ1K‚! ÓV[Ò2fø0ýàéÀÐC„;×ã×)`€ôª¾_KoIÐ+Ò×)=©¿9¼Îæ.>îKžˆ#§y))¿!äÜàÁw@ƒ×üû·‡w£§ûg#^(4ïr ²ŸÅ|k§H‰»<@ÿÒ»ôÂ>²éj(#WQ()õÁêpxÑŠ–³îì/iB‘ßr÷ÙÁã’c§=–-¤×žÆï²?.ÆŠœ"›ªøO9Eu½ØíÏ&ý-7õTt¢Ób¹¦¨&Ú©”Âp­á?Õßî±Ä× à’Sd¡K,Et©÷tà ä>ùuàªlmcíÒ„cÖµÖï&Ú Á××xš]É4Þ´¬0=˜fƒFr0pw•Û˜tó[߆_>ôæ!E71- ÌÞµçÚ†¯¤ÁÔNè¿ ­‚o)ɾÖBòœn6ãmÈCïüřޅ\|ËØ»¦òEÏôx‘à$¤‡Ÿ’#v¼‰Cž&:D¿šŒo:ÇKŽ÷Öb, ¹5㎄ Üîaµ†¹_Ûë†H SïÝ\×½­]Š—>€y×> ø)q nöebÿ¡p­7÷>¸;Ð1!ê Rm˜>hy”T¯Â¦Æ¨~¬uÙÆ¡€NÃUQ£K•fB»ºF%”Àá½Å¡mäª>oûJWWâý۞ơ¥ö¬ïø³ƒZô}/Þ"ácóFƒwQEwk_š¼o€Fð®=SâlФQž®9 ­È ÆQbl'¨ªh‡úç(’àYŸ~WËð ÉˆUŽU-–c½nzA~·¶+2.ÔÂãÅ< Þç}¨ÃSI¨x¯—ì •ªlU`³7Ò‚tAaO†··É¸sÛVZ‘Ó1¤Õ‚‹Œå$o©óU貎„¦ÇHE‡ŠÜ€®T(«=®ƒæ!ØcKw2õ¢šº¡r¿!†HD¼ŽSMº§‡tY'v€œý ‘÷‹#m½a.­Áí‚ՑІ·Òto®Ž¾ ÏÚͰMš­Â3H„ƒ%žÂÄð¬¢ØÀ”TÔdÃHΠZ®!gÛÐ`HÞ‘&5a-t9 ‹iÀ@@ 9®1¡mÜOÏAS–N]Ƕ@~e¥5,ēȎD·u{Û¨êõÓŒñdŽÁ—k•¶†½×ü€¬Z@áE ~ ÜõZ]< :w’bµÒ‘ªÃ×?–G:ñ!Ñ瑦­¢,—B mÕëψ2#ŸYNno~õ2R–ŽJ0ã"õƒÓ ¡þ;„tußÎss‹R‘ÇE;Vª·ÿiq`%"/|0;úü´£«%èËgÑ—°zê±ÁXĤnß¡A?±ãéÛïRW [g´E&]Ü B§ý4Ðy”§Á’ÈgÅ‚žÁŸ‡gl(®óè`M}JêẬþ ÊÃÑßD°Ý©+‚ÒÃù ñ_0ßE·Zª!0šÕ¯O˜æ6«Ï«–¬^š= qš Ÿ]&ÈÜWc5ͿݛЄó|gUŸ°ß£êÏ7„´ü1U¼ŒWán_Œ+ƒèHH3¼«¢]5â²0]ý°‹”¨Ðô%d,žk™£\À¼NÐ}ìŠrY©¡°ò‰…¨ƒuvÁ§q~ä'c@ƒbIq³M¤d÷Ÿ¨M¯5ôŸÏQh"œª¨nD¹,ü›0ñÕÇsd”r™Uìšv,Fh çef”W‰ Ý>Ú¹CÕîzÜSkå:ˆòÁ[«:ʯ^.ÚýÍòçöt±Ý¯67û‹öìôæÑõâÕÙîÑùbysµ¢~ì Ûë£òóÙº=ßž­—°ÇË?®._nÛ×íÅ~uS…:ÿÛ“–ÏŽ!)ФÿÍ=d}àx@ú9Ò:þ@•˜Š’ª[žŒýˆë_{o*Ŧʹ"ðÑ]ÕŠÀ2[ͧT¼éS!qÛ¾²‹PPo(æíKDwYÜöžøž-M·¯Qo5ÖZØ} r5•Jòò= ^gíþà:Y1ÔŠ¨Švò‡”ã \IÝÇ&î¥çi­Â—~0×_Dç‹-…~^ÿYŠ_#`þi‹h*|+ßZ j¡‡R\rdšÅ’Áy|´£ËñCx]Bà®<îI,HîÛU„†+FÝpÔ»L»a*Ɉ°n²<$”R”¡ ³&!‘,ͺȷ°ÌÅV¶E¯6qZn¦—ÞRBk¦†É=Óè BF‘EL€ŒR¾N€?%¦~RVâ»~¢º-ñ>+hMŽ{ÑfЭžö#5W õ ÉzyNt]Bzâ»0öxÕÔù¨?ºrtV¼úǵ!! ²ì•DXú¼Š’PÞ÷KN„þùÐkUÔåuœ&óî çáŸ;pÑuPKÚÆ‘Þ¹[ ‰Æûd ‡Ò Ë´Ð)i²ìÀçe†÷a§ÈawÄÎT´Ð?ªlA(¸°ˆ 6<$’ ;áž9ÿy‚Ö£œNãXeñ„•C]Æ”2ùQG’†„”7‘!Xß 2~''•žkO‡ñÖKM’ƒˆÜÝf„H ÷ä$fÃfy%÷`èWÉܘ)‰:BJqçb^%,çôÂØ‹Ê%v¶ûˆuãºJü:¶=WvT”+S$m?5Rn†~†áõ”s졳Á|%$9JÆþÜv_ þzôÿŠQ‚endstream endobj 384 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 471 >> stream xœÌ3þCMMI9(øøø‹ûVøÝùVŒ ‹ ‹ ©ø@÷c÷`÷]ns‚†Copyright (c) 1997, 2009 American Mathematical Society (), with Reserved Font Name CMMI9.CMMI9Computer Modernbeta ŠØøˆø×ûV ÷5¤÷ø¤—¤÷¤ÿ1æª÷YÍÿIÿøÐÐ[Ì!rv‡|o2]V*o)û1ý Š…’†’‹“‹ŽŒÐ÷¥£M»j͋؋խÁ¿¹·¨É‹Î‹ÉpÈ]¨¼¤ÌÑ‹ØûNû?{‡zˆz‹|‹|‹}—’ž‹˜‹œ‹œ‹›†÷÷P‹Lf9Wrr”rŽn‹Wmƒqv¢„º®¢‘“ª«u–\‹e‹Gv@]YebXpS‹@_ÀÔ•‹–Ž–Ï÷¥¢à×÷*÷‹»¸sGu¡øC–––÷o¡ûa–ª Õ  7Ÿ ¤‘ Q¿endstream endobj 385 0 obj << /Filter /FlateDecode /Length 5597 >> stream xœå]I%ÇqÖ¹Aè`ƒŸ¤K5ì.å¾X `Ø ´˜°$² dàqz¦¹LÏNÏpHüÛ‘KUDVe½÷zz †óTKdddÄK&¿Ý‰Qîþ¯üûäîBìn/¾½éë®üóän÷ï׿øÔ„]£Snwýì"÷» á“Ûy ƒx»»¾»øÓðôòJŒFaÝpO~¿Æß!ÆèÕðUú-„”j¸Ãß~G;ìS£íƒ¯ðÿ¸8’Ær¼Ã¥òc aøÿîGý^àßmŒÒÁ¯s¿(á7ýÃsBAéD¾¸¼Ra´!\~Õ0x"×Ã4ÁåÑ:æ•êÑÃO3QÒ¨á ùL{ÆÈ Œ¢´±rx§‰Âãäuè7™Táwi`¹ié*†èóÊ…ðž.}ÿœÌwOè8d6ÀØÃ³y–=%ïy™Fš¡ì‚ 2wDËÄ=áçÓºž­ö@˜sßáÂ=áB Ý2ç/׿½Ð6Â)³ë+BÓxyåœÇ‡ÏŸŠ\èàe0탯³¸ûûe^Žføó0¨@þ|™T/ɰI ½ÁÁ_&Øà7] /G#,ö¾È PÇøˆkºª‹º’rŒÖʼ¶:…²’ŸBFô°ËU`½2Þ2½vp.à¿ø48rRµÁˆ¡rñèÌ­Øy¶£¤UfMÚ˜µ!Çàš6æÒ‘AQø% Žû ü>‡]úk•Í §æ£–‰lFeG% 7ðÒêÂû"Ñ jKºÆ+=:Û])Ð,Q¸Üé› & Ø-ÜS²Ìtß.Í:Ò ¤?M”ø ë¸þ‡5¤…ƒ­qb’æŸÉ÷R­ÐàF¥\œ¶g±J¶“aô`¥ÂŽº±žÈ 9Xl@|ù ?͠ǘ§­w³BÄ <°üä‚ᜱ Úo柳“}£Ÿ¸4õh|¨iž'dÇÄ‘¡È¿Akh"œ…mP zÙ%#þ¿Ùˆ3Õ…³9úMÓhútÓÅqn]˪FŒ¤8=I À >»¿XØ7¶Z,°õ ôÂä¨dź CÊÐR'v}$“ü‚TÉ5£„+H+‘£^ä>I„¥ÆëUL-¤ 1[w„U¢šmÑÒe<õ8²›Ê§ã=¦_±>Ç…À¦‡:DÉ-Üäõ·|hÄNy4€f@‡5’ëR»À-Ö‚"‡Ó¦áPC×<äu·äó¡ªŒØ—“ã=¨ž¼4¬Ã}Uw&ë§YNÓ~@¾wpÄMÅgö÷y"TV _U,7ª…‰¡ÜÐ8ŠmH%ù¬ñ„ÙI°†MN.f³h[Ó‰YÝâÐ*ÅsN‰ÜØŠˆ]~“™ îly2ð"ŽÒO*º3ý¼0Ýå;û–ùżWæãüÝpγÝo-ì4%¢±Opª5Ï{…è©(ÕáÖ TÕåu òdǨ0µÜóè›pçgaxhÖ‚\ÛÆ™]:HŸ—}D½d£Ä"¯ïÌ }°½Áú°»‡Gh²& c‚ND˜‰ðe¨&Á²”Dè†Ѐ°—©!B2³VmÂ4h1|3…Þ“í”!œm;£4gØÎäX¢í¬aÁ/çã¼jܼ†]af˧ÞòTÙÂ+~»‚ø·0€оÆ# ~¾ŠxÂKT~À@}Ý(ÀÀûºè|ѺƒÀY~±öF'x­¨<0ÕîŸ ó© }pþÇ«Ð?Š€õãÔãÀˆÓœ §>?ÊoË„>Zt@†W_êhÀ|1q ÛâzeÇx‚¶^Ûž–Úׇ98!üøäu2QÛ[åÀvêdœº|”·ÊƒS-/×hÕŽ«?M)Ʀ^®)ÅØ*E¶\ÃbЦ˪;˜„ŸâÉG­æxŠìKCäI«c8fõj—3b¡Ç·XzÕPtFÆe¢ˆ$³h¤¨º+F:êS7€ÿ"â2‡N''Š%¿ok׸â—¹í—`æ½ú%ÿAZS°JA95ʽL3•Æ«Ý~:ï&r_Øòö`LWúCæœÇ ܳÙçëDÓ(½ÔkçÓwdñNŠ/¨ç“é*ÂFIΧA¶ÜvEòAQõœ…áëj+|*7±¨rníT>•òìsY­$=–Pž £ÆH—+P<µÿS«z¹²½Ã“_U/ÙôUy‰¾Ðù©¶#û1ª©í?ò!‰úþÒ®„gÙ$p45”îäÚ˜¹mrm kÛDTMÊ1]± w°ÃÎÈɪY `$«Ü¨¨Ì¸BT©GAW}\TE™-.ošðŬ°Ï@[ IÖm[¢•ÀÑ ò©áßO6þÂ:…ºA||}ñ °ÉÛ ±ûÕ…r¶Ô쌔ffww¡€#=y~ñY·¢Œã—RQ¦1€ Ô `ܘKÊÈæ!T$€ÊÿM´. =ØÂRÆH;Ä¡vTΔN%O” lh\’†‰ÉÅê!çS†‰À‡)d¦¹‚§†<DØÝ659‡(q›9'íÿ:ÏßÍK>Í /A É ÑHÍR;Kmµ1˜kÄéƒPÌúÑq¾Á…ïÌ xNZÐÑÉÚ €ƶé',2æx/Îè¤:£uN•Ñ„w3Synõã¹ñ'yï°ÂÎ'û§”Ãï©oS‚4ÄÕJ5íÀ¬Æ¦Â²2ü §IXÉu¸‹þ¾†âQ¿Ï¾Ëå•29bFµH‘¦Å‰-ˆWW—3¡AÃyñMÁ-Rû>s _DñR˜ Žo.²‘˜‹è¤PÊ´E)ëf(®í4iÑ®ÅTxÌèÌâ ÎYñÒz%×™2˜nat¸Ÿ×%m‘OÓ”E’%Ȥ®çþJŽLãJá$.¢øTËZk‘93dý“V)ÙE¥SDõD ¾™³«¸Èe†F"#ºBâ[ã319¯@öRÈT¨\×-޾uh,Én)Lä,f±~EÐ÷ǤÉ'e9y)[°p ø~N~ÿ»bú/ ^.”Xø1T•-³H4ÿ@“ê/qJ./ŒÒ¸æØvÜöÞÙBîLCÉá¿èÎÖK*æØ§ÕËéñó²s2·øù4s‹#:ᘹj’°`õySÃl@2Êã²±-0g« 'çZe€´‡…‡5VaoÔß`O,Mß=ä®àãû¦¬GAƒN=îW…1Áœ9ø!k_Ð=i›‹EháÎá ýûlô¬Ý\JA[¿,§WÛš°˜\§l…°H'ÂcÌ —N¯í)„óÚž_ïÛ hkª0¥hq=<éý:Qy(¢<ÂwÓùoà¢ËÊ.Õ‰DE,›óòduí­œ TA]kaÑ:sÏë“¢IHîóu<¶/‹ŠÒ°[ àÉÈõò}n»O§ Ul­`8Ú fµr4”€ŠÁ².¹©ÅEº©±,¥š cj—z¾€ÀIiJ…Ôó—DšïÁ©~ÙéÕHSÁqp$Ђ6ùÇVoBâ¿ç§2‡ÊbÖä"ÃЭT°Ø±-ò{Q¢¨Ízà<ÅVù]ª<~ÔçÞÕ²g_Ô:v4þfiüjó—Ag׸ÜÚfÚìç›ÇW€ 6­d%ùb‘1Æ;Ö¿ÏŒqS¤™¤eOr=¤m\˜BÚ š1X.׋·>­·Ã'û[ ˜ƒ{;µCDFŠÝ£ ð¡í:•þ½ëïòŒãe½Th £Z½¼‘\ɾs÷&\7uËqfn½åêf¥€r½( p©R"&^h4–p^aá´’mÅiº¶#R´GŽb¥(¥þ¡ ÌñzÂF®°D«Þ\Ó‡ X,©U8+]ùÝ,@«ÆYçÙ•ªºêyL2+艔¶ç`P‘¥ƒÇùhJ ˆh júá7“¡’a@ÑŠŠ—Ëï_‰ˆíu'l zl#(‰•uZ1ÅZ¸†?×”}E#juDoIyÆÅ'‚lŽë-/{ŸrM!˜¼È·b,/{ªÚÜ µ lO»•+é×LÉÙdã†â.2}GuÈ¢@g%¼Cˆh2o$€šæÒíµ¬)=Æ%Ú„[©J•Rqø—Y­^…†ö\¡Øªˆk…"ƒµs„"äCÌhfxÓM¾ÙïÎ%ä<'ÂÑ 7½Ìâ æÈ ±BÕηªÏ^a­Ý[Öè«9¥d™®ógLH´YC …Þu!Z÷ß×Ö܆=[-Õté—!' ×Ùá[ò™ã= E±ÌgÉR59ÎX3ɶ7)ÀE{났щ4ºµß’­¥1=JÃw—g²Ž›úSƒŠF§ZËÍÕ©Ü£ë hƒÜc‘ì^- =ÏCû¸Ì›¼§„í4·2õ ²7牑ËJË}^ö\˜¸D ¯å`ù©­‚ÈD¤kÍ#îhŒˆés+_a¦Æ‚Y>`—˜¿a✭/¦ äq¾_öÃÌbJ2Š«P=S' GÚm~xb´™A„éÂfNn¢¢({ëý²U@?=YQgªÂ ¨³yN(Éõ† —€}÷útQO=lz¸X±q®üuj»ò—V##s»Ç¹HoÏi‚ªÉTì;!óW%úOU½Ó­÷^€½õ;æjT‡/y~Ï{ºáëÝ’óîíÄ÷sbqYíx1.ÆÕŽã.ïíÔ‹aµËVÀÏû¿³›ÞW•<ßÝY‚’¼Mp윭^‹øR¹ë£°ÚE—œR<_ÄC¼þq|Žu99G=„vád3ƒm‡9,~I¬žSet-ÍÓU6¥ú®<˜/жú·tåvÉ÷Ž!ºÊ¶)j>ñ‰4^}æÃ‹¼Ÿÿݱ{ª)W¥"ÆÜNÚÒm1W¥Ý0}9?W%RÕ –w6Én,o‘7÷«4V<+,¨(tÿc»ýme©•q‡EwUˤháD%c@s­ÌJÕM'(—R‹tÂZõ¯JX£Ö+,öHK”õu¶Ç¿PBïòM]Ú̧S²|=ó©@i3uÁeiRKKç/gK 9zRDà¬\Í|úô +VØ•±þÛÖFíµk¬sÍ.9ðG$Ö n“´H-bꙓô¸oÊLÏ@ž±äG»"üŽ—yr±[¨ÆÃ1Û÷1$MÛu”¡ûÌýƒIH¬„.R z `ò€=/³¨¤r6],`¥y)ÞŒ‰ò¤ yœåòé¶¹ú=Un0oì>OidHÅKΤ`óÆˆ‹B+¸hAôÓ:ŠŸ 0Ñ©\wÖêOí’'1b z¾Œ’ƒ##X ~®Äs¦*=gS‘໾©ƒu¥&Wϧ/ÜѼ'¯J±ÀÈä±á›Ì ¢±geq |Ô4¶ý!i+˜àZ±”Ò¼<Ç' p¤î†¾]WãG)´L·ç@ãFt¦&ÏÇ^´©Œà1º×©êSÆéQE•³_^ MLÏ"O~{ïÌ‘[H½œÛ‘óŠ…•ñHÁÞÊóÞôüѨ•B”_+ÓÈÏ/dWNTLqúEd ‰Ù’äà2†ŽÒÔp×ö‹Fb1ÕvatZ‰É+m$l}”Çkìz¡Jznè•Þyº]Æp–€ž²œ~=›höx/¯Ù [’ç+Y1p9Ö×;Xùû ©ò6¡°5‘$Ív.¿iYÀñ7Q|£äÊñ¾¨Ík=.¦㺚ºó.)ˆ&ðJnvM–ú¨NÉ9öîôÐWA{’¨ZqÆaÜ26(îÃéÌ®eýÐ%l<ŽÜßæ!þqë–o-ëG…ªLI©~÷hóO4¯·<)”4räX®-2fħ[ƒóÝ‹³^E‹‚Ân!"ZѵúÐÅ3ÒþÐR\&©lL*€9pº¬—IÚºc –ê-»êLU©¹»=ðÜÍü—ν+äK¿ò_ÔHÊ«)H£H{Nwæ,)‚*šÿNEœÑ.n2ó¼Õôt›çÁ{'=$׃o7|“¹'^¸MB…a4e¿pLñÍD|@H#ÔJ®ã ÔøÃÅÿî)‹endstream endobj 386 0 obj << /Filter /FlateDecode /Length 5980 >> stream xœ½<ËnÉ‘{¦  ìeOñÁÕ u9ßcgÍ€X¯!Û°Ì9­ÖF‹l¶8#²5lJ3گ߈|TEfU6[2[˜Ã´’Y™‘ñŽÈüqÁz¾`ø_úÿåíë½qV/êÿßoÏó½µ m™ê•_(ÍX¯Ì‚+m{¦÷›³kø\*å-€›ü?žñ€n‘þwy»øæâìw/¹â ˆŒZ\\ŸÅµÀˆÄ¡…e¦÷Ò..nÏþ§ûÍo–+-tÏï^¬o_u[ñj¹\q¦Œu+؉âÂȶª÷¾ƒaÄX®óVd4+¦pçõò/þëë½ö‹‹gÿ>‡MVØç¶Ä“56åç±!‚ߘ8Sðëþòas¿º¹ÚÜ=Ü\¼¹ÛâàSgºûÍþáþæòáfw·O+c¼{€Ñýïó¿UwµÙÞo6yÝîzÜÆ5üéjw;þñf?þ‘‡U‘³nÿ°~¸Ù?Ü\îã¾»w«ë·ï7#µœ˜%×Ëñ7.0} ¶oG !%•öþ1òÿ©OÃùrVÀCúN¶¾ücÞ·Ê@´Õ‰§TžUøíîNk»¹»Ü¤cò¦»Ü] ÿŠïîó Ö7ù/NüÍõõÍå œ4ž_f+6ÃO‚ïíúöõÕ:¢Š¬Q_œýõŒ-Úò©Ü¦a¨xzXÎÄBkÑkn¢|ž/A(à4¹ÑÝúr)lo83Ý ­ó½î|±ÞÞã$˘3²[ß‘/¶›8OÁù܆ðÝ{2ã-þvÀõÀ¹á·÷Òt7dø]cz˜c½s Hü­3Æu÷imÞvaaèÆežl 2Hë¥ïèâÃé½R²Û“ßôÛ;܇â6.×à ǹˆ ­ã/XIàMXc8㊠ß-(C/»Ÿâw‚¥ÅŹtýt»—d=›Œ[½…ÓÕ½YÂdÎ7y²bap6°ºÝÝW ¶Â{éÀP÷8 »¦ ¹ÃõKÏÂá!d-VzR‚0ŒÏââ ˜*íPÀ¾â‰¦ÉéÔ¥u¶ ùU„­áçë@}&d÷1!d ¥™t–w»ðáA'M fºRJÇÝ9zFtU„$t¸ø¶ÿ€{_IumübÅ9(Ce# ž/WÒ(PK°¥ÈßÖÚ‚ï© ]‘ß­9[„(óž‚(ìèß)?¦}0 Z$1™7ҥÀãá:øô!€õ Q¼Œye‹ƒ¦’ù6îÍÁæ¨ðRIÛÁ!XÐX’ǽE~OÈXñ8” áÖ™n2k`¢÷#ÊOõi¥áwó³gôI@~1ZV,õ$‚{ÎR÷ k¦ú£¥W¢þ`ÞDª™¼î®çe™ /»ÿ+´ò•°®wRe™*>ÔÒôÎₜ×BÞj‡³jòšjªÉ~Ò¤g°D+z'¦‚â¼ñÌ#ãá>U÷a©5ÍØný–Ì}_.+Î¥À¨t]ÝQ¬‰È}F( Ðã(× ¢™–QD5$ÍöÈ[‘I„×2 ˜nhb™ –¡¬ \ÀáXP!4 ^ êï -¼¦ÆuÒyŒÃÜ]ú‡½QKXè<üÎ$\NŠ„ˆ › •!ú%p›RþbºÿÆQæxè ‘®KkTžvؘ•×A#&E²ŠŒÔÜ`‘/ÃT05-v¥ Ž|:.¥¥d¥Ó‘‰§(í¥fDÈ2¿{é ñ¡„5 ’÷¯/ƹ…¿ÇÀa-yîo U÷y“¦[W»Äqy{X5=í]u~Iv~ÙP©ëÈ^‡CõabLø¯Ðüƒu-AG½çftÉŒšHn“à#/ f„+FanÓ~yR|£À½t˜Ò|mÁ°ì–QŠ5d*)‹£ *IÔÐ%¿Vkœ ˜ìÑ,þ„Lhx»–Ãzgæ CãØàÝušâ¹<ÀÞ8-,Q¤xhÚÀUš1aÊfÜž¹è•}´Ã¾±ùäGx-‹Ü‘_™¯¼Êñ¬çš‚XÁ‘«*Ù çë ¤&P hˆ HœÆDr«à‡¨]ìSRYLÕ¨ßi4ø) Ïé¤Ãˆ$z¡iåI8<ñ€S…à‚S˜tcú}ع .$ªOã@Z`„A¼ç“‰ÜÌ:ëÁÀ[VD|ÔnUܳ¥øTL †øM§ëû„×%"Î!N à!Ãfè, B¶õ6íסz!Ñ õG…>” :Žà¨–ØG*HgšSˆpÑ­§ãjîÇaÏgðÀÞ!p‰~¨Aç7‹Äz7§ßWqyàº!GºŒ¨Ž7·ÀæŒomç…µM¶'DEÎÆ*ÿóŽ:¯T$'! ÆÕ‡”nO;PdY -³¤iYûÆûÚx ‘4Qø;»YÔ~0C‹¨‰Díç¶”@æKPŠ…I+}()eJ‡ GËAN‰åÄ? PÏ®Kƒ3ä-<šÂBÒméÑ– œ†ñ0$Ƈ¦Ä“‚LøI¬PËŽ>‹{±èPßÓq…ƒDi;âžÍ0(·`6!Ô?†C}Ï¥`ƒÑå=DÜÈ»&KÚ5ƒ}"E,ØÎãW ‘¦"¢üëÞ̲pô÷#{FgL¢?ã€ÕR 'Í£ƒy4¶EÂñ"nBŸ¬«lNGËó-øÝy?dMæRŒ…“Öpö‰ÇJO×qzLäyq*oHu»×#õ à5uPhx[÷Z­±Ëâ<+;3ºQKX C+L]J‰ß—çSÆÖŠ7¶‡‘\+Ò ÿçOV+R`% µ",–ücó°¼ø>/Ö‘´D$Z2añpš7ŸŽsˆ7Y)ºÿXTV)Ü“ ’‹kÓvó– z:r¥{nDëU·åÏ~æÏ6ìÅ*ýqƒÆ@ïÃÈ'ceÞ] }øøns’ ìWÛ×÷XôøÊƒçiAÞ­û,D@Jg L__|uš])Ìo–ÌÿjynÌb¦áàˆ ‹X§aG&иÈ^%©ŽÛ[q< ;#t4X.“Æb½¬»ïHr|ýýÍö‘#lÎeµ¹Hè:$¸wÒ,?AçI¬‡ƒ¡ûR:Oz,–ÈAç]¾ßœFé Š2ßÉ”žî-¸CÕé´žï]‰êK(=ˆ®µ+°žPëEeD‘Jíy,O˜¾úö»óé= “äåÑTïIëûð—/¡÷(²WY²O¦ø3JPÅ à}‚þ8‚Þiû°Œi1Yïsµ(¦@I.„seà"­ Ž3u¿©çNSL4\ØGx:”W2êÐïã0&Ìפ˜0Tkf*—1Wøšøâô›>Ô£¦?ué/©ö†ÓmÖI”°À*œ 8©C_ÇÖ†´óôTXm™‚ù™Á½áÖÕÑn•æåÓªwÎBL¢+XQ+ˆSÁ€Íë6‹8W†Ö¸ ¾×±Y@Ä‚&ê3 èßì(•$Or;iRò ¥A{¼I{ñ%á€N–Ð%Ô#%¸nÀˆ0jŠ1+EÙ(埤«³'BiPÀvÌ$9 »,ꤩ¡J2EbÛ®Jâà° ¡ô“û¹+§ˆçc‡ŽÒJ„>^HÖ.Bb‡ ­¸.ŠúšHÑM,p`™äŠÄÇ#äÇ®óÁ»î¿­ VMr“4Êð»¡Wh;-|IÈ*y[vܤMBJÓœ1ec^M‰Ù¦•Ž-0$ÎÐu.I*R˜-ÐdÌ6ë5ùÅHÝ&8`¶±©öe¢ª²weNsltC-Ï´žÖ=ÒJù_΂«¸lK ÅA€–UŠVv•r0]! ¨…¿€œ4 ÀFMµ!âKÄJ-Ð1èX“ …ÅšPÞ…=rMˆOUõ‡3 †Êt°‰¡Šç],c½ zÀ™Q¨<ôDô1Šáç†I©d&@t>•ðA­Їö6Ò)"ÚqEôÁ.® M8Kª ˆ÷ê¨Ð,X*£TŒtZqÑ+ð>¯Ÿå$ºž‡.œµ‹îïè=WYlÅ &PVdêcélGÇ]>Ö§©V„¤­ ݯ1ûWÚš#L(3Öë©£S7#èVá¯ê¢ÀéÒÛcjƒ­Šê´Î7SžDA×(èÉ4Dðç&¿ˆëBê‰o•kz~rÇŽ ©ÃpÞÝ¿E6ª ‡\s>”p¨ÃÎ"wÙJšSË=â( ¯ uÚ"Í•°ÔÛ [lö¼Íl Üh£6~¬î4”P×´q•vT7@KZ`ym²Ѥ-³îÚæ¥¢°ñ]›Ò°W,äÀ®;N… ©ݓݴö8YzQ´‡ÉnñžÊg΀w‚ê®Üe´têXy AìÒ?|‰{AßÎ_Æ¿yÎ[êÉuSúSèâà’± äiÆžè±EÈæH¡g£[˨á àb˯fÔŠq]¶ˆÇ)*i´$’´?ƒú£ÔºRÁn¸²— <7Í;;‰·éÐKµÝwÕs!2oå”8àœ(‘ÑF+pç¯üdj¢©]žtÃ&.¡Mj¹^ÍLÁµëÈIFÅf§™.Y¿ÛXéd¥¡ÜEÈxѧU¸Mí ÞÄxµèn µØcnz”“ô½|1m:ÐÅ{b¢%Tt}³k$/¶è5§wªëCñy7–…g­ÂûŸeîæ4|ï,ðUžü¡Ö5dœAßEL"š™´²o4cþ©òNdH\rO‡¸J†Tè'XSDZwå(kZõU!é‘¥&Á,8fâ‹•š„†Ð%¥¤c!®(ÊNQf€”X¯:H’jp\à'uýk¼c5ðÍ£ñ"HYžû ðyceò º¥hé©ä“u6$‡rãÌuê‚  ó=ñs³ƒ¨AŠÏ!Áã)‡!]}D‚•Ù L8 á\Õ‚sI‡íõˈ´-3–°Ú´…#Ç*ø÷ãMV‰´ÎÚ w`9ðt>_494EAÕ*•÷C°É³R×J·…/ =QažÜË9¹•0H.QÝ:úõ¨z?Ãp7 I|èãH·Ç±¯‚·RxTWÔT ¸˜_šÍ®×m.DÖ)I».*†Ç¯EsU¥yGs+±Gd'ÏiæÁ1+E,*Á8$ Œ¹•ަ¥$VÊŠŸ°´.6_Cºàqßlau ¾ ðUo5`^àÒ½âŠÜ‹ÕÞÅgÎïÙcÚyúÉX%øl“ ™cRuÝU/+ÄSñE¹`hø™¹X;¹n³0ð•­??¯nâULúóÀk§.ÞÆ¥P§qˆšÌ$mœ·tD 4€óZ¯!ù°[c›˜‘kYx" j±¤H§|l }Ýt€Ã²´ älvÔ²Pÿ§ñtÇ« œôE¬µX‡Êq”©ŽDÙ1Ÿ»Ÿ¹å¹RÜöZºJõ].5žóá‚$¶¿€¤°äsûÏ<ª‹‡¸ŠP¡ªã/ä÷»G¨\ú¤œäÈÝŽB0G_‡mMƒóòÈ•øÜ“«˜QçzpI‹/SrÆMµz0ÄšsP|OK5ËM”'²ÜS(Oi±ìû½Gáëþs\ôÌyŒØ <’‹Ø÷/Þád,—”¹Î祦A™™ÊLWªÞ%ö\7bùq^ëMzsg>½ŠÐµõåS_¸¼A_V”q*÷öeÀ‘µóKIÍâ×g(–Çë^š(¬eÁ×!ŽÞ¡‘…£MÛ8H?®12tЦ&LŠ*³g²¶;T[ÃË+JßcÚ å割?ëzŒp“g±"hüŒé°˜tV~´0¹õb¸›®Zå솓¸žÔÿ§Hê×ßR³nÝ›Jb«ú<3[žõûë§‹°p}øÍŒPo ÑS6Ë© Åùý´ æ§Š€8Vû›Tì%Ä\ô`ô¯"$ô¤ÏÕüb°ÍLÝLJΊ¡%¢|ulîf9Ä Ç¼‚Uv=Só¡Õ"ʈˆaWÚ#tC¤Àãàö¸ WæÇáë4Œ'Bv„‰þ:W-04–£úùý<Q¾{wO`FAP.Ó…IÁ€`k^üžQ»ÜaNÂæ$û7‘5ñáÞ"·wI~¯i­Rwà3¦œxÃ…Ã (ןüä~ˆ=öÍ')[ÏZNõ]pþª¶¡Ò ¯JŸÄ •:alþú¬ÞàÂñ ®Fëx#[?ÿÖÕä"•t¢ˆ0ãgõ*í•£q)«r–pvõÀæwTHç×'½Ú"’i硇„7=]ŽJkÇóסS­JNžÈ©ŸV¬"¼E`Íà1goèó6»¥€NÚ(”£¿PZ¤4þ5ÑÁdå/—¨Îœê¾‹¬ÒV烟KL˜^ÈÑ™êãÖñÑ7•ú§dÎ=FY%I!døÌoåÄlw™é­3QèÒ€ŠÄ¸/ÅOVâCƒü€°Ê˜n¯úÐËŽM_@I1^Y!||Ã鈇]«zhÞ"/˜&7[^ ŠFAq• à‹fÞ+(¡GÛÕ‚ÈI}…É+À³oë”×f¼·R)Ì ¯e,[ûÅêÖcp7Ã¥Á‡P£Ò“eWíÌmÕw‘ñ½ŠFÂèB3©ó¨ò2$Éw“Þ¯ âhÖ¹xÆ9].¤õ&Zë…ØmÄ~ý³~=¾–KF’eÃm$p)¿õ)+'™Q¬1êîÏã£ßàæ|GEU"þgäuG3'èÏs4Ø¢8.AåÇóŽq˱½¦[ŽññáʰÉàãkØÐ¯soí¹’«g®ûQq "ŸTz*+'ƒâ¬ž¿|ü-qÜ«7´Lx\É&Wš!ÚÊ‚;)Þઽ­ãc¤MóÊÚ:­Éº™èègÞü§ @uüL,Z?†„%ÜW¼é½ ©kºÎ;¡†>—¿žý?÷ÉJÃendstream endobj 387 0 obj << /Filter /FlateDecode /Length 7677 >> stream xœ½=ÉrGvwÌ„o¾#x™†ƒ]Ê}™°Â¡ÑH³%{(ØŽðpM„ h¨ÑäHúz¿—KÕˬÌFSB1t`«•ËË·oõÃ)ø)ÃÿÒ¿·'ìôú䇞ž¦.nOÿp~òÉKåNýà0§çoNâüÔ»Ás&N-‡Y¬>=¿=ùûjsw¶fƒöž½º<[K)íôj‡-cÎÈÕþ–Þ+%‹çïÈ«wÓ«{|ì¼—fõ-þ4ŒI“H©é|œO1fŒ[=¹É‹7á'cœ‹Õm¿½^mÂÏ.Yïœ3«í¸³ÚIÿqþד/ÎOþvo§®ÿÝ]Ÿ8ækOÕ~êTiÛ3§B©Ywº»:y¯K¥¼XÎ~ìú÷À?¨£èE8‰Nççïaµ»z8;ÿ.ïÔ8¯´Ä´dÂâp6¾Ü7@Hkýûš,e¥‘Â=ÉRÜ~¹b©ë«·d­§ƒ —|`ªXêÕêZ=§‹I@<˜1Ö:>ùðÅ”Ì@Wûq™ci5hÄD²Ò§—›½þûó;´VxS@Þ3DemœÔþ,èì EygjZ*ŒúðI½„,ñû÷ íßÎ)Y,e¹Áø`J:úÇóýÕž-ƒsX¥^M×{¸Ýn÷ß.s:#§Ê+ût™+PN—´{þò¿¿Xæ\΀Ä)ñc¡ ƒXÉ#¾¿ÚÝ-Äÿ$÷°SÕ½/ŒØã-)לYjÝ/YI€T%">;ß½»»Øì¯.Ÿ-s:«'>Æ­eIìdÚ…….¶wûÝææn¿œä’¾Xr!JãRæ¢K-FiQt• \H¸4Ï‹•ÞÝ,s(+AO-õé»›…ÎĦ.‘q™seaFuqóêŒHþ_h[P¡9³Ñ¶8G5\0æ•–@²®¢ª.íšÄD fÆ[¢òSSày´.,<þggUj-\œ ; }›Ì_ÚÙZQ`Ù0Ü*÷«í>>í¼°[ˆ•³=n` HÛH6–õ0ây\ÐÂáéŽÆË^Ô¸âî_q¥3¸ãÕ§Á›·ñ`\J»¦Ah)êþéù‹“ó;xaãmüå8ZmªÛÃ犻qÁ Éq'Û8v[uá8Ò+ŸgQµÚÞ“4ûàà PàB`Ô1n HlÓª\‰Õ=¹“w”É$éDh±’§ß’÷\@Ö*æ'/¡´á\Òìôü¨á_p½ó6Ž-è°ç±¿+1c<Üž’À£W_žðžIÈx§x< îY†¾G ¥÷íw:s¨bЬÔz<Ê!"sØ7“À›¹lCmõ&B^ÀîW›wdPB!‰t²·‰ã7gkŸ€ø F÷r\Â]^oé±nÊW[;ÁdŠÑ³ûTäŽÓƒ%9ÝÑ!–·åÜ#Ànȶèø×g XA$À:P¥q×gkà)šøM˜Nž~o(‡¤— /‹4iK§ >bÊò<@z¾º¦ Ü\ÖŒ3QóŒ IgNŽÓxægþ$|ŽÄùgzët=$ĉ„ñÕÒˆvr£óé«?Æes‰E¥1ù´AHÀÞòz"Ê(߯ÁÖÉþíôe„b ä1Žà^£Â{xÃlK¼¡Bì§¹@kòò¹âÝjžpæ–ø hyP)V Vq+ÉA­q(¾¯K*@.FÈ‚j  S¾c@×¥ßûé1Ýÿ6I.Ákþ”P,¼¦ 2ô¢³dB Š(ʸ7L^ÈÀdÈp„Ç:d%!‚e›©Å[1 Bkm!w‘+Yf#µw²š[J8äÍÊÿÄ&’ÎLŠc¬ $°ìò-nq,ó^(Àœ{ú?{‚ 7‘¼‘IúÆ5²Ê\ž~<©X=£ˆYÊʦˆy–1Ós":¤A],pk|ÑPÙ\±6ÊBp¬B:Ëë‰Õ×{þÑÁ)_l;H:†LF"L80TªèäµÉzÒãš§yØé\ÿ¤)õJ± ÿeÀ‡¸uH’â®"×IgÂ<6 LP ÉØ“»¤u-VÄ:¿œ_Y-aéãâÝ!ñ3(k3Zý%\ €©¨Š#5- søçÙÞOÔˆ%X$† d_Á¤·´¹/H¸œÚ@ÀÇÞÌÑ2‘6Þ‚° Aw³/ÂÕ³ùŠ’ ²ïYœ±ŠÊ“Ë8c¥×èãДÅÖqî ÁsŸT ©óP ›®Í“ðØ®¾˜V~ñ|ÒR˜Ð"“v6t7æLdXœ ’Ô"J™‰EPÁýŸ‡/|‹¿…në}•ʶ ò¶:ÇlÜj Š”OPÂ!WY‰Ê€{ð³Ðu)ãH œ…®sX!HDW)¸˜õªfLû>ªœÑƒe«¡a_NÃæ ˜œ¬Dm²ŠPÝï€do¥£fQÀn!ã+,AÖl,ƒñ.åÅéºxç˜{š_ù?aˆ:E•ÍKŒ‰âàýˆá-éLLi{:c¤o°¾ Ëš» ZŠð:ˆ(j÷"¡i8çhjÐŽa„6ƒUÈT.3Õ@Yû8Ä3Qs®8w©LVv:A)z?a{ŒÚþzÚcE)8#¼Pac„U­AM‡‰ä:Oá±¹x²®ßä ™¼am†›À† ÷ý™Ö¨DÚTïÚ“?ä3–ž|Sý]áÈÌËO¼rÝÒ™;ôˆzEØ3`å7õýNd†£œ~“éŸÁº¥^“Žä'‹>þ™¬O™b:‡¶FG…\Ê‚Ð(¸DRß@/q¹ÒÓѱf˜èÐ É@8Þ·$#¿~zBxnBÖÔLÞ|N„Ñ6Çz~gÙUó6Îî@óµšÆô”ܹ¯‰ŠZnÄL$£ÏtìÛ` Íüicú)$Þjjäw[¨ãúÖÆ¶T’ß—ŒkÜ1õ –ênBKžWzu6Á¤R@•¢r v#böU‚Òµ²%hžË—7ö#Áþ ~O.a¦M4„ðï§y> ›E$AÎÙ,ÒðäábÙ,Ú*ô§}”lºÖÂÙ,t©…³YèRËg³ÐÕ–Íf¡+}”lºà“e³ÐIÎf¡K-šÍBúÙ,t½ÍR Þ²Ù,t©Å³YèbKe³h”JÅJÛûý[c‹$|¤$“Þ…=e:‹¬*¹ß38ÙÍíB©, u-ã%oqöµ°äUa//8 h2ò£¤Ò¥–œd© 8Éj N²ÒÇœdÁ§œdÒ¥'YjYÁIú(‚“¬÷1'E¼…'YjyÁI[Xp’•>†àì\Ø‚“¬ôQçâW–íW-0±ÃB/6·¯/7Ÿoïö»íB’KéQ¬úé2ò8‹-²ÒÛ›‡…²[C¿I±Ø«Õ2â8åfÒ¥ö»ÍÅÕ2˦s_OÉ“ˆ!+-ǰ["ÇBŒQ„®b¥ûÍîábóv™‹¦cOGa‚¢cx±Òîê~YIÖ#°§DÄXÒ@—ø2ÌžÁ.™î¢áJ2=^P/t(?ø’ÿ:ÃÿŽÊ²Õˆs° µhÉ*ÒC¸˜e­ƒ«Ø{ÇyÌîÕŒI!Vòl͹Ѓ^}NžnÉïÛô¦d¼øؘ›‹Ñ æ©S)¨’F<¤ l¤á‰çh ´VšUœ‚7U8F©õò§èo©I_Ê‹0¤È®,ÆÀM‘x4 è„8 .bø<{Ca(HÌÓÐrr׸£Ÿâ`„oUñÚJL"ážM?NÆÐ’ËÐB™WEÓǤ©i̦Ja$éN §Ä`ˆ5èhå{ÏR²ò5¹"zzQ9È•!êü&l·Œ„7ó*îÒM1[…ebt “1¼•²‰‹ðÓ,!-M@Yç1»"e[5Ýp7Ÿž¾XÅi2th°s<Ì¡Ôñˆ¾›ˆÑËr¼Ã$9é…­ò-ÖJëÀ9€a`Ñ|ƒ^ì”fçí3Pb$• (¥J5€ZÕM ¶¥JêïøëÈg1Ü#åTâÕÉ=”‹³ð:`ŸJ±œRÍ褤-j7ˈkȨ½ Hz X|JŒÑ×kÂ+¦Ÿ­œN¼rJ¾Ì¬¾?Ó „£Q«Íu¾§D¼HmUšg‘…ÔÏÒŒ·íŽÅrŽJ½‘„1îãþ`‚”,Ñ[çÈ\*DÇ4Mã~O†ß„¢#T?™ò`¢2Iau ™¨Uóx3¹ð )‘õúºWE(±Ÿ“$›yÖÂ’àÔ¾fø¢Ïz@®ËÁ;€s#Ä¿[§V4q¢HA HâäER9B œ¹@‡NÉuõӌ檪„YúI©,™9VQ·­ÒiRrr1; 7[ Lzó«Îï!í×»Õ¦ ¦¨´T'1±V$Aøà0Ltþr:uKOC5Ú¹U¶PD%žÎiB!Æ–HÞš“¦!þ[z Ý.ðYŽÒ¨Guh·3|–بH©FØET˜1óý3s¬YãH}c5Ÿ© {P(¶¿éˆ¹ Ûr f°KLòu«ëLTZgúòL]Z¼)¨‘Êy ±7ä ¿M™-lV:O“Bœyàì54ÒþgÔF´ÙeP‹q1‰7K%Bƒ1ËkÝ9”w2Od‘´Ï®½öݳïS²;\'/“ŒÓšuéBºxÌÄË÷zD>nO;¥W_s´qža"‘ÄDÀÜ„m(ÌÍuÞ=È  0p“-JMJ0•¥&%ˆí‹X Xzõ»‡¸I ‰>Ò°â•VŽ™àƒœ?nBfH¶T%Ž’Ü—ªR"ÄI´7ƒ˜ê6)%~:S– b³;šÜ•AÛ£( ÚÖ9bäÔŽˆ¼J:YWà¾4Œxµk‚­ lФ¯ûˆ+¹2'K0‹´Ï&é6¨4ÁB1Qq!Ü9@ý‰„{¾¢éîµ›¢«ÊÑ\Ú¼OயGÆóSœmé¾ Øææœ'œÚT•5ñ–T1ÉdÈ‹Š`×T])Mö"º¨I}œó9F&+.ð> ëÔÅÔɱDãGQJ¿ë’ððÜ * ·‰™$§šJÌ¢ææ€ßDðPÕ/&«þB-™É2C°`-Îë'^ÇÂ}àn85®Âg©êØt ULÌrk€I«Ýñð=®(ï×ÙÅûƒÓ((£ ÞÏð÷ˆÆŠÈª ×È DL›ezî™7“¬w^[œÀM:[ê»lJ].í¹Ôå²)‚AËÀ~¯«NºLzVê ¶@¬E—méÕŸ?‹8ƒÂŸÞõ{ÕQb£xÿc|]O½b£ÛâµÍ˜q4@´ A¡L$§^‡4wÒÜ oZDZ±‹;6nVñ(±"ò©¬±«8!VSp¡IŒÈ@ ð÷r§³’¬I+Â-ËJ— ÎÔ)Bᄨip,f븕 Tù2()‚ÍàB™‘7÷Ìí„€‰¥esŸëÊ(:X·Öð¿dä>è¶GS;‹l2™VX}˪GiÄÇÜ“U2é GѤuÇI0‚wôñÏdƃÝ&U¨e·“F'ŠŽ"4u‹ÀMƒÝÞ£·Xe¸uKZ5/ÿìø1¯q».•aë&ßïÐAîÞ]œÑxÓ[Ÿnü"ußTQHfª‰îmsÝPXÜ—q‘œ†¶‡d º.=öÁµŽŽ~ˆ ¢ÎW+6*í‹®[órJ墩ã¤tzml,ãüC8kw· â–ëE^gP•9~Õ½E*Á¢ÀèÁU‘o*-ß|±){‡Ð Á4Ò2ô܈ThñÏ::U!‰JÒõ¼Ùj|„5§}ª ÜfÊõTĆÏ-|ŒêZá%¸K'„,}—¸Žúuý†ŠèA¯«¬ Ü7Ó.ªØÅ-6s(:‹J&3é,ÐÞ»n¹{õHŠþWêB ~ßî@ö¾€ÆfÉåñëžš¸Û¶gàah«oJZ©ÃÛOÒžOÔ×YL ÑZn=6Ô£M\~K~Óç¤yZ+îa?~OÐ÷d'”Ñ^ ÂH ÊY¾B±:\•Ád4x‡`—Øšñ*a¤A6í$sI6BQuÓÙøÃÈ¢›}{œŸëŠñž`¡ycÓ\ 3á›/IJ•ûß„lÚáî æ1©(ì?5ùì${ 6^ÇŒ‘j’Ï7gâWÇc±ºÁIFQ+ÛÅç¶Û½,¿ˆý§`î £Ã|ß#úaœ1Û1®uLøix7a·ò‡™7Z[¥ÌÙWì\^ù‰‹¾+Á¹bKZòŠôijǬG^ÝT‘h«€•—è±õÖDôGX–”q?dÔòÝL®*üŽÃÝAå­iÐç´ƒxü‹YˆdEÚfÌDжô2ÅÛnû´=&0ëDáDiRƒèöJý¢Ò¾‰ïÈÊ/NîÆš3åêpw¥¢d»+k4èZ9V£‘ uˆ’~ize‰#“£¨ç¥ì;\æ…l3““¾ìãéf·äùý»}üKH=Û$EO¨Ä„ÒÝ,4úBPBaaP½ÿu/¯Gg$ÒÓÅœ›$§óu+àf"]eC2á:énÀð¨˜2hKÌÀŽš›÷µ5šÁÞÈÓ¥;ÿ¹qzFÕycQ•‰ö*%ÿ^æA­2ì?b ¢pè!^G>F/¥Õžï¨n ›µÝ˜eƒ•zl¨¦ã¤øE.|GàÒ¸Í7ÓSI¾­ð¡Rl–RºÜOºìë§23 4>0וòP@uˆÎé¢cô>Έ• ³L„öÑü<òÐÐDòeDœ`í}öõó‰¢âÉð#-T»b´ë·ûýýï?ùäó—Ÿ}=¼U{ûúÕýnûÝÕÅ~Øî®?iÒ€„uºª¥U6$U(ª|YKøAäÊfÏE¡Ž‘!;qîÀïd)‘Xjè2£æÈ.•3ÌWZ…æ:‹Gà6_ÆÓ:-øEÌ¥ ̤z…9 ÆW ßÁ"šE^óÀ}5ìÒ=xcŠ{H¿>¹¾½mg•ã\:W‚Qd¬\,rãRFTÌjÆÒ/‰²cV©OöH¹-É+˹)lH¶íºå*/ž°:@††íã$è$>ªz-çôÒ/b) K‘ í³Û‚±ò„ãÊ*¯œœïý"’pXטÔ>ƾ|b‚-Uã/¾ìSŸ"½,Hºt5Ö\Èj XË®ðæÂ`Ú6tW³I²‡9r% F Â§ö|á¯z¬6\Bì—mMHµü? é^fd/ìP±ø|ÅT„Áù eRÏñªâ׿(› Ÿ·©`²z ™’Á0’rÈz ŠïZµûܰS¶qbü8å§ÎŠšò€MTg«˜>Æäß"ž‰ìô?þ‹QRû™ŽŠ¨Ðú½¾ WñÀ¨ÍtQª6Ðê¨v‹TÀ.fš8˜Û¤²«õšR%8I)ž‘ÄBôþÁˆ¥6zŠñw¼¤™JUìq€Ô™¦<@Å‹Êê}ĊࡺxÒ%êZÂQËÁÉI'¥{§©4ôF#÷?'D«Êïö ž—ê×ï½ûCi¤ßÊQG¡´–ø]gªØ4•{B9ôæ_}^"™Þ$áY-Nß/Ø1Äe@T¹ÕÿF2¶™³ümš[_üyb|U¼y¬gº ¥q8 Z±<èä3ÂzT¦‡»÷˜”:¢ÍWŸM€ú¦úI¸uïyÇl óþ?Ó„Á{Òõ€s²!<·ôµ!*G²PÉ;ªgOý”a:£vñ˨›²š©zh?qâCUÁ 7þH~o*E¦ žGE·Í¤m3ºžô2¹1ÇÓñ;žfހϭ;†ƒ’þ-håðº#D#Œ€„Þ‚¼Ç“ûò‰[ôâö¹zp¤Ýô|GÞ%ÍëŒZWâË'ðT¬ýÛŠéÙ7'ƒß6H?ÈQ„ ŸZ•d¿Éa½8t€c(î˜EnäÑ,ç-mCL ÊQ™ÝûGmI`t…%Ijý{>ÓC à¨tù²z–F2h„#G X•ƒJ³´SH®N«“ûŸÐˆi7fSL/õ±Âw0Õ°Åa*|7q² °ÃZøPQè‡Jà÷I¸*àíw[|¬™M®ï’ €«¸ûH¸¼&¸í º;CµÙ2~ñx”Íÿ/É£÷,@KÀ£²\£ã|§Ý*çÁ6&•í\Hbé†|…ÚU;¦T¹>ÑÚÖ"~¡“ê˜+aþ&’/*A©TÙ–ÁË0 ï¢9m’1“ñµtùn:Z‡Î?b‹”SøW¿yw­éc<Ógc#äóW›X¨ì'J ìÁž)sŽs Ÿ4Ö>“ÉsF=J[êQÎZ/uvÖªyåT¼N9 ܆³rÝp‚ǽìã Ï«oÊÆ/wŽß4j¤Ø¥ºá2x['°æCViwµ Ÿà;ºù'ÜWZ6dÙd×Ê\ … -ÙÅo‘öiÈÅ‚†ãã®™í)lÀÊÜÄìáï¿9ô‡a®·ŠŒLÅ׊Ù@âÚ™«¤Iä]ܧü}ßWþ,E0M×…z Ía§K#Í 7fd ³Ü½x;‡œz…—ñ÷RM½ÌT>\“^ë«©éº]¿yÚuA\½´ƒâ3zéúöŸ‹Xæ(ýæ…÷ñ`¥Õz –­Ù–k®íq>Á‰hÆß3§àøíå@‚Æ£ÚNAÇy™Äö(‰Eu¯ppm²næfAȨá¼]Ì1ï[i;×ù¢rw\·™¨]EsfùgpoZ”Ç¥9aR£ífÁXü¢ ó> stream xœµV PSW¾1pïU”¶Ö´°Õ{Ó©­ø´Úõ­[_hˆZë ) „$Q BÈý A ÄrQA‹ÕºŽVkÛu«í®Û—m­öu.=tfO¤¸ngÛÙ;s&wæÏ=çûþïûþ#£üúQ2™Œž5n¢ï× Ò32ih?i˜pLwKw”? ”Ã@¿æ¡ƒžŒêŸ@úÇÐÊÇ)¹L–’–7[«3¦&%$”!ª‘Êq“'O£|qìØÉÊ—5êÔ$UlŠ22Ö¨ÖÄÈK²r¹V•¤6•!Ó Ý”ððÌḬ̀XMZ˜65aÆÈ1ÊÌ$C¢2J¦NÍPÇ+çiS ÊűµòÁᬳµ]ºAªŒÔÆ«SS(Š NÑÎÖÍMM‹0¤gÆ.REnŽW¿ž˜´)YCQK¨¥ÔpjõµœŠ¦B¨ÔJêUjJ͡¨5Ô\jAͧP¯P‹¨Hj1õ;BåGm¥>’-“½ÛoN¿y„ü?¥ŸÉ‚ÿ›t­ehfsžÉÚûÇõÿpÀøï ¸(J%¥[v¯ÅwÊ¥ñ茢#÷0-n±*›îÑxÀ6=s®lû71ê‚…œžÙek‘O'0Ò(³Õ^½î‘í»ˆb¯ÉÑ}ô"K ‰Ú:s…£±xÿÕˆS3ž7’lüùØûèñ&Ä”—Åùù‚`*äõS¢2c€]8ã Ac®wk=›]K|òÉÏç öJÂy9Rw‡+¶W Ö bÜ!}ÙRΤÂÚ©$ êl‡¡ž<‡¡®7 ¦Â±z^:iePÎOO”äÚ¶;}]±•–¸Pn÷SAÖžé^ÌnÛ)’"ÐÕ—#ó¶‘‘2GÊá 'â'™ „íÙ8¹çVA _‚yÑwrô“4Zát¹ØØ²B0rø ‹Ãý=´ãËòÈÒ¸µç9ÿ­V¡†CWèJ4ߟðå¦Wq Ýk›ÝHÛâãÏÝ!GGºÿéSh»8…L“±8?ýYøWµ_rØùY ôÐ ÒÌÊž« Ðu†Ì$MRA°)ÏU¡åèUÄî¶èêøŠ (#ÛÛ¯ô*4çÁ4Í'cü ”£()!0\ìýÇô³_Àá§îŒþùµ_ãØ–|Á’SÈ¥GD'­"F‹Ù¥kOî€spŒµV)høû{Ûà2]hÅ~l j !(¢¹—ËݲûŸ’ õ¥ -b ¢o>ùFEUa¾‹3[2ò!•Õ{Í••^§x0¡N5}í³˜ªÿ ¾ý‚^"c¾­kÿ¸i/ȼmeuh͹ãmr$Gó;³÷mnöÆå ·Ds›¦‰oØsÈU%dèÙ‹ …¼BØÂšÊò*ªœ®ŠÙÞMªÄ,½–ÓìNpm"~Œˆœµ/Á•Ëk‰ú @ïR7g¬2Å©`5;ûÎJD£~w;ÞECqÌ0ĵ€K¶g4À~(ƒÚµì§ÙŠñ€(¯—,·o¦23É2žñwŠó:»šªkÝ|}µ‹pÆÞD4„EÏ[9‚ï»]ù¤ðË ÖdÕ‚9œ€ýèôÿ_Þ°lЂ†Ý$=“£p©GQkÜiHÏ2Ò=Yuµ;=uî—ižvþb¿Aœ· õyo©¥ô/Ñ$²|n!ã^N†VµB°”B1Ø®þ©Ìyæø»èËo¢.â¯×„ -Q–ÎÆê™JÁe†­lff–À‘ÁÅùK&æWaû\ßþ¯°û1WÞº‡Gí0º¶‚)Øh,ÊyxJ´Õƒڵ°­LÓÎ>î;¤uh÷+ŠßØIüÅNÝ܇}ëžD:v"ïC±vÝ #'Ð97—z–Ã4ˆ[œ­bÿ7äÿª×[ÜÒ¬Hë¬tÓ8¦ŒtpüŒÚýÝ®;+¢¨#¯Kvendstream endobj 389 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 332 >> stream xœcd`ab`dddwö 641ÕH3þaú!ËÜÝýãúO/ÖÆnæn–µß· }OüËÿ=J€™‘1¿´Å9¿ ²(3=£DA#YSÁÐÒÒ\GÁÈÀÀRÁ17µ(391OÁ7±$#57±ÈÉQÎOÎL-©TаÉ())°Ò×///×KÌ-ÖË/J·ÓÔQ(Ï,ÉPJ-N-*KMQpËÏ+QðKÌMU€8OB9çç”–¤)øæ§¤å1000103012²ýèàûñ©{Á÷='¾Ï8Ìøäî÷÷˜È}¿%ÚÛóÝà`Ïw†îy’¯´ŸýV‘û]ò×ëî÷‹ßÕÙ¾»þ¾ÈzŸíÎ>Ñï3¾‹³Þášlÿ[Bž¯zö§éß³f/œÍö;a*û ®ëÜr\,•ù<œ‡'ðð\ŸÌÃËÀ÷–zÿendstream endobj 390 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 2108 >> stream xœuTkTSW¾—˜ÜÛŠ¨uâøê½±í¸”:(XQG«•XD«¨Õ¢"" @HxIÂ#$Ùä•ï>ð¯ú~–©V«¸¬Úv\NµËq쌫û2ÇsqlûkÖYç¬sÖ:çìýíý}M ó£hš–)WD‡ í¦ia’Ÿð¶ˆj0z0V þð¶oûí[X6“Gâ§£( Mk2 •Ú´]rb’^15~š"xÞ¼9Ó!3gÎS,NUë’ãUÅ •>IªÒ‹‡Åm|²ZŸ£˜º I¯OûÓŒYYYAªÔŒ ­.qá´éЬd}’"Z¡ÖeªKµ½b¥*U­x•\ЫU©MM3èÕ:Å m‚Z§¡(jôb¥.CoP©W%&%Ï ™=f:E½C½G­¡ÖRë¨O©0j)L-£"¨Hj45†K©Ô["lje¥:ÖOê·Ì¯S$‰0ìi¼´O*ke"˜æ”` Ôã|Â-ü~0J^ºÇê4‚¬óN¢~ùã8ãVmúr«a.Ûí°Úá8´ZOú¿ëcTe:¯xv‚ÝQsÇC‡ì, väƒ,ãS7€ŽÓ0Mö»â•v¸ Mÿ{ 9`<È·Ð(G–|/%y²€Áõ™¾Á>úÌc!£U2¨üPÄf5•p1 bÏ…ïWÂx²œ„ˆ#ŽÄãd&FàÜÇø.úUƒ³Àd3ó…da?öc²æöá!\uäæ‹BH ‡·çƒe°µPîá„Qà&yè_ýù®ç£Qþ´÷ïw*í`«àŠJw–@«i*¨÷xjÚ:ÒÚãB53²K8JddBTÃ\«,ŽáI3[‘"ÄvûYè€6¸ñbf­> Fkò¡.zoæôIðž0ÿ¨¢lCa1—Ò’âÜ,‘±¶’º“ø.í¾â£…g j,-Y®¼hÙ8ÝÆY»ÿiàJ]Pf+³V”‚Ø0çðd“æ*GY™·sì{]}÷¦3Ðì‹SîžÙÞU¸›OìLsFìIu†ÕÂ%¶Ó{ôÒΠ;WV´ Ê­‡k¨¦Fa‚çzèõã¤F .¦Èq²LÆ(ƒÌöÇÄ÷gZجïH'C$·ô§¯î½r“;oXË(5ª„pxÜÀ½"Ò‰½4Rýß/ÁóB˜ü—2…3¡±K>šg>všÃo2y¨€` áI + ±c{súÁüÏÅÄ»`ŸÏ×Tß À3ÈWI}LxÑ®< L€`4y0×Ýrÿ’‡>r·ö4J„±è‘ÃèÝÒ“Ô[»ØyQë"5ncSsCCóÁDÈã›wõ{ô¢:˜gÈfË¢9°]ðƒîò_ûz»¹ŠM‰‡õ^¨ý†?gXÃDgÕpºŽ‹$yòHã‘Ý,7mã_¶3Û,&%—¥‘5ÛïƒO$à}h~M?¥¥ê/¼ÝrlaöÚwážÈ0ˆ4NDFæˆ-7•×ýÞ‹v/}ù:*¯÷œ“Æù±”^¨öÑ—×î´uïèâ{z*†¦ÕTj6Ak¬Ê¯¯t9]Æö¤õ ;òŒ\jSRZlÛ²ð…Ñû·Uòùy9H€Õûu•æÍÞœëÙ¥O¢q$vúvKñÅO:¹ ­ŸÁlX Åã6UºÁ ΆªZö™I> ž7¹à`CïöuTy}Sô9°57ÏÖ¤š2m#(?ìÄùJ{Ow+ßX»¾öJáƒU‹6Í!ö¾’8&ÿÊJa5–Êá1´ë»4=›œÉbðøHc kßd„ÿU×/½¿èºÕ~F4.ŸÈ°×~·<úx\ɼ¸~îÊ•Ú +9’ÿï~:°u°Ø,æÁö☧4ÒW%&¼”·f6iÓ ™i©ÍÙÍ®ÖFŽpd¥(˜/v—ˆ‚iùÍWŽ‹?VÃØ÷Ú:K=¼0lH ™žÁ úàVH„µøµ˜ï%l} lŒÕ°D³m®6Ê´»Â,éYt=ãü Nžò]8pÛþ¸ÀŠ>zK¾"]ÙH\‚±€÷áVýņ/?ï< ÝpAïž] –Á¢‚•ÙÄ/*; X1:ÜÀÿñ·?Q¸@nYr‹ÖšvFCªhl+dmM—¿hmÆé¨8w ž±È̼CX†O½á”{[ëºÚ³ÝÚB+Ø,œûÚ™£ýÀ~wjɌИÅÑËyK2¥ý l21Úƒüíjý¼ wÞ“àä‡ûm"YQzóAm…ÕQbµBi!¿&c£%"`“'»²¸\üØ"(ÙÉ““¢c–ìsO×~ðXóI8 ­1•¹ e1 G´6Ö˜ Û">ŽìËùÚauÚÀ6´UíkÏ®×å¤&„^R·ôÙ8‡ÏöNô6}HPB¥ÿà¡1#$‚dp¤¼¾FlƒuåÖåKŠ Š82œLw”8ÌNóx‘d?9¡ôHp}©üJfO‚&Ûžæ6tÔUïr8¹òò2(ÖFóòôO¶læK-6kv˜»n!Ãdzej«*½2¢ªf|oöçÞ6Çãÿ†g·¿½ÿŠú/êMéendstream endobj 391 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 2917 >> stream xœ¥V TSg~1$ïI•4E´MкÔ*Э޵­ÕJqWDA ÊZ‚Èb "‹@’ûeE (² YÄ'nرµ¶Nµ=VO;K]æT§:´ö¾ø3gæØv:Ó93çÌy'y/9÷{ÿï~ß½Ÿ„qÆH$6`å¢ÐYþ®ÇÉâ8‰øì0ñ9)°Ç¯9×ÈÀC n'ž‘í…­£Ñ07Žb¤IRꞀ¤d£..V«÷}!jªï¬yóæN÷}Éßžï› 1º¸¨ˆDß•zmLB„žþØî».)*.Foô}áu­^ŸüêÌ™iii3"Rf$ébߘ:Ý7-N¯õ ŠI‰Ñb¢}'%ê}WE$ÄøU7c蔜ªÑù®LŠŽÑ%2 3îÍĤ€d]ŠÞ•±*æíXmP\püö„é3f³šYÃLdÖ2ë˜`f=„2‹˜æ-f33“YÌ,a–2˘—™åÌf³’YÅŒ`F2ÏR\7&“ùF²LòÞ°_k—n–Ö¸ÉÜdsdÇäò­òÇl{‘Û?\6ü¸»›{¨û×OMxê²ÇSÓD“§Ó NÄå5’NçQ©˜èœ­Ì®ß …y{Œ„øË˜üLSÞà y;ÚÚ 'à"7ñx^`ͳ¡’·•ñ¼ú Ž¡]þ.™,›Nm :à´ ¾È’ûb¼å¤YFÒåžN/š÷a_¬Ã åô G¹·¢‡>ˆJry±†màßµ¶Ó<—Ì ×ë‹1†Eɽï·Ú T¤îC*ØïÖŸ˜2mCPè•â&2,Y˜'Ó°ŠžËÅ­ö?ÎïI >Åø´T<‰J;ó!™K^ñŸD|Ȩ;Sð5|ýöC©"%d…’(™Û=p¶ºU}èx—½º kg­ÖÁ°™› „! •§óz¶ >$WÐO4õJÅÎYÊCPjÌ5›÷ä©–­YyféÁ%àC¬ÄL%$é­¸e8Ùš=|FV.dXÔ{ÈX2bÕd˜s‡¼§yg÷ ¸mȼÿáÞ.?G« j‰Â–Λ뀫ç­õjŠ%Ù àx‡(é‹êðzDQ܇O{+>Çdüñ|UUGç¹¢Zà®^\E¼ÈøÀå»4pa‹:»x8ÄÍwëÖžÌé_üë=T£ÿ}2qá†íF½ZñU/K]0RcÔü/!ÏzŠ ´•ø|pÍÏçÀµ}?i-·‰“”xê'­D«¾ØÞ‹2Eo^Iïú»ãpjÿ ¶üî“IjEæ"ؘ˜’Æ‘>%îe¶nì£úó2"PÁ;¢Õ‰‘aÙñÑröSJèðÙ!i§ÃkŽ–â€8MiÜi6@.÷Ãñ¾ \ÑZ²äMÂRøÜŸŽ3ðWMÈW킣ŜU Ú±bsz(pD 8åÔEÀgPÙVm2UÒáX¢Žâ"©¸žž¥¤Üj…Rξ Œª©|7 ò¢þfG?p”“–&7›ö9×'µ©ÄQòƒ¸E¦‘çNÞ¹ŠŒnŽÜóñÄ»‚§œ ÏKñ‚sŽ2»ÂÂhF 2ïÊ!ÕbI2¥˜Ì”¥œÎµ5Þ¡s¥ ÎAçàhбFÈàÖD›…ŠŽkX±`à÷e»yËð9 ¥ujñ µ<ß`;l³BpGyŒ)¢)#C Òæ"Úö T[šM¼R¹©TZ‡©§«6Ö×W:\—Ñ–©1i·ªŒö7éy'ÏŒ¬‰l¡«ÁÉ$ió·ÆBæÞMñY¦EXÁ…ö-D Þºt­)ïý¥­ªu «aì€0(à·g·ÃA(Û»¿¨„»LÖ+q¸¼e°26z=y†xl:uêÝ÷¡J=èD©Kb’ìå‚’"öÿ8ƒÐµQz>øg‘z:û³:›Ø,xá|sñE_ôûÓoÅNtàߔ،î,αÎ_µ^!óÔ8Í D®éìð!÷ ûY£åâó]ÂhäÛ×.“_?é(a°™Uœ¡'­‚“Ð×¢…tXË`Óµ¸“©— éÖý¾òÞZªQƒ%'Ë¢.Ð$¦&ÀjȽ^Ðe.ÏûȲo×µàæð”Lz¦»èïu ý‰;2ÞŠ/Q/nP’E®]úqÅ6ó½Ö.h‚³fÇ +YÅÃ2ñªòdlkD¸VÞ¦íîr´ž¤ã¬“"}ë7’.÷à̤N½s®òosä’O_в²Ïq"p7n?ü‚²ë½èíÂÐýÄ[Ÿ²ø<”½VFäœ §‹kˆô!}j¤¸ s”bØÏFãZäq^°:à8œ~RäbT°u%E5U‰×˜ÒÝÒÀÄå!K¥a›¬·ù}à€VK«Ù›Á†ƒioÖ2cÇà3«ÊK‹¯‚À&›t”lɰܪZZu`3•ä 'Þ¥|â\%éË•ÿZO õ¶ÝÔ´ž2·þ`kýÌ¥F }¼½ÿÍ©Þî½ÜZ›öÆâãÿRÁö‡vNôÙ¶C•öåL[ ̆ -ñ+8êë~t¶½CÆÉ»û‰¿ÝësœëðVˆ¸­F™G$Ë7Œù&ÕtNyNÉþï¾@õoãÞY¶]u4ælKMµn5ŠúK>G)æÜRvkÛ Iw[[·Ê3»\\SŠ[ìeårVÌ î(Jåî6·ÆcxG±‡Ê> stream xœÍ\éÇu×çE>9A>äÓXÞ˜Ó©û Â¶@+‰À¶6H-,¹KŠÒôîR¤ÿñù½:z^õtÏAîÈ’ ±ÙSUï¨w׫þóBôr!èßòçË«#±x}ôç#™Þ.Ê/¯¿=9ú—?™°ˆ}tÊ-N^år4^¹…—XÄÛÅÉÕÑ7ÝýñRô!ÆèU÷-=[Hßݼ³#±ûûãç'ÿ ²ƒãd‹€ ¤ŸœƒÒ×ÇK™G5Ì"ÄPGv…!ÚM,©ûÆ•¡¿8^ªÎøî ,íp „ ¸~JŸçGµË¼ %Û ,°”b¤ÉëÜg¶g¦ðÓ½>ÈE3åô8óJÙ="–0¿+¬ ¦Ù‰Wùµ×ÞÓú VÛÊÆç:;/c…m–(› µö]"»›«™wí_ÁÛcÑ«DÉŒé•tfØkŸC¹¹ž^¯Òˆáï¦GðUØ"¯3ª›Ä@A÷#iTô8¬÷!Ó«}ðÝ[6œ ûEÖ ¢É¸L<¿aëu$½´YÎ2:rJz•†ô*Y…íÉÉjÑ )‘ª—Á£™|)tÈ[m¼_øÞxÌÊSþ°"rJpñ›t‹¥êS!æ)'…ç2šâûƒ€]yvtòÏE¬PÒ¨¼ ulAÝ )°b%Of½ñ²êMKÀ¼Q0>Q8®ªŒÈ5U5jBûºLï-ŒÏ”FrÔG+[T«lyÖ¯ÆEcÏM(cE3Qw²Áˆ~³^h¹XBnaŠõ N“V¢Ùn%ú•~Ðô¥ÑØ}åA¸ì£µ2ÏøŸã HC“ý7šN7€l:½·Q6êW]‡t–œD€ÆI cƒÁ²Wjp3Ú5Ææ>¯FJËW{ÇV›±0lAîØk¾ÈúèdÈ®™b¿žt!Ü û ~ÚÁ¬{l m¦¼8†u0ÆùÇv»ÖÐ"“"· Å…Ϫ"Fïóf*m’ƒVÚ‚Ô» ™Q·Â62ÀYHf­.ß…»,%l§!oé5â]#ƒ`du3p(º‘9¿pSð‚—imw+† áK† “¦»¢&2Ì1 kŒ$Ðýûo’jÈîËü»D4çDfß1“ÄMŸð–džÛàÓ›yO¬‘BÛÊé›ßùóu&*1cË.28ü×ÈPÚ/¯÷ŸâE„" fµhS‚. ¸ßçý’- 9kùD¾§—CèxšŸWÊô¦)ma$è`Yqtz­D”\=/ó!ÏÁÆ“tU}„*¦íìí Ñ·‰°e-éMþqWP±‰¼®3év:#‘+)M’a>Òû%%00¸ zlɤa• EÊ…­ @ZEDÉOr7Åm@ñ~`šÔ(ÏLÔ Ca±¡@íwÇÈeÔ*#é`°¢n4à×Rž¸½e ÈU™O¦Þå¼­Ñ5¢Á Ràù! '!€ˆo=ÆÈ4”CÑji U}ÿ¤`¥å¢Zˉ({”r“L`„'§ÑX\ˆø‰ Ê|ë|î*Þ:â9äÀäïH²´B:Ä Ü¿ƒ àëf$wËÅ—ƒ>gsëæªM¬f^WÆ»ß𦓔Rb$äm+gÂØð$gYø±»'3hùÖFà¢òFÜwˆTgKIÿ°&R¡fm«‰DàTðûm^¹B1%Òa ¿ds1F7a0&ý–ݳÑh‚‰UF‰9½5Ðñ—«m»a®áŽ^ë`¬‚l,ö$ Õ@ W¢Ü…ØJ ™érØHtlì;'œ!wæEaGnMÎ ìóLŠ »yÞ¤d^j•¹Å’Ø?žq0®ÑV'h´Cèt–)©zÁmW[®’s«ó1ë,8œÊj ªîTÚv r®_ü–M»—_½XåC«]s:P½i÷"ܲTLÚÉMÉå|%Ϻ^ë8lÒû•N&kê÷¨²ÂÃÝŠ–‰•Mï`f¯Ûh À'åQΈH¸ R}òÝúⶨWÚÌ1¨ð+æ±yâP ð5”éMp¾®ñÝj7¦Pê½ó26¤7ƒ7þ]%×v7ž×&ŸT@5P5jÑè¨ÜBˆËç|ôZâSŒF müØÁòJç0—¢)äF¤g'lêGظШúuÐ8Å@N±q…ZoWæuCv ³ë©l¿)I&@È–²œEOn)¤lôÛÅ— h94€©ßthP Ôç _eXm¨œø`ÉÞÒ’B÷©Ò{U˶Ù©J¯éµ {Ee^è«Ú0¶Ä ÎT3‰ßtÿ›ø3w¦G‡)ÄAvlºî4.ó³šõ¢qË*­^`“^¯ÓN•ÊL]«­M¥2ë­ŸC¹±Vt*D¼i ³[+×vÇÚÞavps•ä‰*wF¸Ae­Æ=>­ðÖA`š9OZ-UdÔá¡]ïØ™‘Þ(¹ñŠ) e0 ¯,X²—´äƒ(K=ùy)‹…½’SxUe¡ Ürž³‡ iû_B»ˆð8ÈÁo³„…KXíC>Æ£W´äßå4Zã'É÷n8`ªáQ#¶ÂçVôY 꼜:¯r½R.Zu´TŽu( –ÈÆÒüÖä«2¸èjmZï$º“'ûX¡]„{‰äFÔA²'Ä€ ¢ß°ýTCúm‡Õk2‡¾0ý×…éjžék{ØH‹é•¥“6xe$³yÆ/é¬ÝC*ÕñDä‚ýiÄS âiL™“rav§pëíƒ §ô» '*‚«>Ÿ qÙŒ@Ïc z÷ ¢‰°Öú¸›h‚†9Ñ«’˜mv9Ǫt¾µÅΟRï"Ï—¥ð"üÔÁ;–³©—#BšªûåZþŧÄÞÂôC ‘ö]KÁ_¬"æ‰ÆXI.–ÚjÚ¿Ò‘[ ÷J'Ž3î•Bi²Up¯&ØáÍÞ‚PÝ«…öJb?AIh•¤!lÉ^ÕùõΨèm´ƒ[‘¡x•zû^ź=×#6V\ y!K&‚»3ÂïÍnÀjÇÁOnÁ6¸×QÿŠÓ‚bê˜#ÄéXeÜ@"£¡Œ/‘´fö´Ó½áâ® ’1,±‹¸OF“X²×¦÷}¢Éºs;D“­ï³Ý)š„• þ–p*š\O½0q‡Ô‹Â …1?m˜¦ÂŒzªç1µi/ÏC>î$޵uÍñ…äwê~ïãwjn³ÍïXMG¼ð;"´ŸÉëÈïx¿%¯³FxÈ)w<åÍÇ;ÖqGdz½­ŽghMÛÇñå+;«¢]õFl>e .«TTÞÇ™ÒÅ GÉàèápâÐgSTµÆ¿öˆ¿Þ’+#‚^ó¢žEOSR+Ãäút×GŽbH»ÑÃÃû¥¤öÙ>7hÖ̶F–‘ï¥l’+Ò')Ù|ä<«d»¸ñiA(5Y~¶*©'PzîÚN°rg‘5LÜÔ =0ëû\  vœ™ÝvnÛÛ%Ûw›xÙLÙãjãþMK!¤Œk ‹Ú¤Ëˆó ‹˜GÍÚ¥a1wCòÞ6s®·˜ ¹¯˜È:Õ—³½£šŽ’ílШu£5 Ñ£¢[püÖRÓÑÂ:çÖfCZ¶3éÞUêÐH \í äLóä߬^Ï|^Ö1Ý m+”xÝkº4¦|/)ªñûœb†8e‡µ|„­¤ww[¥ƒþh[… ¡§¢u_Ú˜jŸc+äÿ"Ÿóuÿ0Ñw—\-±Éƒ Å‚äðqÍ<Å „Þµ…o2&ô©°ÍÝØ_! ›'g©¢ˆ¼],m€ ãþüƶDCTÔ»ããÂÇJ’ˆr[S¢ëá»ß'Ä:TF5™Q7¢öúнˆ“n{ò…ø¿§kñ Ç›}a »© _høºø«Ò?´Ç%{Ž” ܪC¨Ã”7ÍÍÜÕLÖ-$‹Æ¼Tzeö.Ð$ÖúVé½Ã¡¹w‘×ÕɬRó Ïþ ¶O²M8ã\½¤Òæ›fµQš¡­£>¸Mí™~ꚈòsFú2Ï#9›ÛèËö™È [1'ň]iXaÚjL""ª¡%2˜úÖ4K¬Ý…b÷Ú»Pé*r‘[奥#üªn熭· 1iŒ’Z—ËRcµñmMíñ³·€*úë÷m&nÒÁñ~7Íù"¨t½m(ÿËH¹ÄèÏÛ×GAÄž>Ûý¡Að¬o¡}8 ]–ÎLéÉ{ÿÍ£F:¬5´S01 ’cëµ`í/x*.6ðÌ Tµÿ¢QõJ·þø@øG$t¦o[î]+ÜÏ#m*_µ Þ_Ýß¾9ŒÌ)ƒ¤Q¶Òý„A’p©ÀHåàÃ\.=|$êìl¥ûó7ç×÷Ÿ†² Ó —ÆJªx«@ߦjeîƒ?µcGÙ@zysñ*%qt”A 8‹¼ºÛ‹»yJÜ!’®‹ªvô"HX!p ¨×—ò"à¯i@Љ˜Ø¬À¡ʉwÅ Ö‰¨ôA0ïÓHõLlу9‘⯨Ã8‘Á]1HÏQðw W¢"ÝØmàÝ]ÝÜÜ{{eC/u+âO³gÊÑa»g'úï§ô#-´aÕ·0Hß_Ü^È’cqÊÏnØÃ¹~-#˜ÒÚÚÏOnß]¿<»¿8?Œû×^R_RkßâµJJ§¢îRƤgï®®În<Œ)N’Ã;%ïŸÉ[ÒÁKú*"¸U:Âï~õ+ú© ùKˆx·´ô¥%»/Ï./§¼{ü Üï)²¬B=E¤Ÿ¤IÎ2¿†êõ=’Ï0Ù<öχ¿Æœø–¿ÉÕ4Ûe£8Œ+V+ÿ, IjÏFu™ùù1ݶ‡~¹n%z`MÿÓé'?¾½X¡Gç[õ—¯ž>{|ŒØËŠîé³I¶ý>!RáÍ &½B4dO»Àyé…¬(‡n#v_"|õæåÐwyëÒöROÿj=}€$†îéÝýdxè¾¾?ïW°ŸÞÞR±5#›Š_•m?œ]¾£iÖÓß»?Üžvÿö—û¿Ôæ%ê´ûëû‹Û—oï©VžxE…A¡­KËàNЇg­?F@OÒ‡„|˜‚9S\‰—z~L5ÉD‚ì"ˆP× ´¢0¹E¸CP‚þÙ”(¸!K4Y‰§f•(œc4Y$FËi@Ó»ûììêÅùÙ]«“ÒÁ?ä^’J„ #mÛÕ dNÛ ¥- È%Þ6l†Þ S¸æ`”å?¯`"Óûš~Ûƒç×g¯O± –o2ôªDËÄKeæ:¯ã? Ñ®áÿ ‹Ô[³ c 4¢fDhT‰Ý@?õĸiúƒE°¾ ò*†TER E*%¢5}všh¯¬×Ç«Úèæov'M¡M¼5¶Ö. ~J­ˆCë[ZðGÿ’ëcüendstream endobj 393 0 obj << /Filter /FlateDecode /Length 6383 >> stream xœÍ\Ko$Grö™0ö`À÷ÆîÁÕ†º”ï‡`¼†°°ÆzG4öàñCöÌ´@²gIê±ÿÞߑՕÕlrõ † °;:+32ž_dEÎ_Wf´+ÃÿÚßË›33ÖTr\ÿ½{wVLs^ÅbÃhÊ*DcÆVÖ‡<ŒØž½Åã>„š1Ý£˜á¯gV–[µ?—7«ßŸŸ}úÊ»*X(…ÕùÛ3åOÒ*›4VŸWç7gÿ3üîwëMtq4μ¸y=¼¯/Þ½~ˆŸlv¯×¯×ëM»ƒ‡Ñ»õÆÆÆZ@°)¥²Þ¸äøs´¶šþç5vZНvH¦d·þßó¯ÀíXc]ÿñŒ ÿËÙù¿*քʧþôÝön³»ÚÞ>ìÞþmwû޼qº4Ümïîv—»ýí}ãØØáÔûϦïa¸Ú¾»Ûn§Ö û·óöÞâ§«ýÍüãî~þQù›øÕ ÷»û‡Ýå=áÃæ»‹ëo·Ç#Uv¯æÏäª=eùâ“H © ”@8H£8ŠãÑtÿùät)¸åtÑæ#áÏöÕ´ç0ME¥ÏL¦úYjišî?ö·ÐÓ»ííå¶)¨¦áruøYïï&ÝaÍ÷Ó/%bÜöíÛÝå:¦æ¸-L:˜nÕºÞõÅÍ›« ]JL—úâüìÏgfõ´…²ÂðäRïB^e[ÅÆd“úÐùzcFgL yxÏÏ,¥8lù97ኟ¾*¡›Òá· ÷7«ó+LrÏÁ¥VŸ†kȤšâ0C´0lèæò|ÓèñiÛ|àG[kvNWX0í«2ÛV¸á`·© Öt’jL„´zö÷°Œ‚Õs®Öï=~HÃÅIÆø:|Ðѹâóµpe¬÷yØñs†÷Ãqe%ð «Ÿ¦I¦N‹Úb¹©ØÐp±>!¨j¹\™¶ñn{­¸ì‰]Û4šäì4|ÿ¶ ¬2|ƒäKÈLÙ1Øž.;þ'6/Èû&9ÜbËßÎ’è¶Óoø^' ¶ð¾›bËa\ÎиQbûÉ/fÞ¶™M|fí~cÓ‰¥ÅNúUöDïÉæ2ìïÕNl©a²ù|¥|'k‡7k‡teœþ¦S皇¯»É›ô|ÉV-£œ¢L•š*Ìá}³ ›†×ÃÌ,œ*0Mß?› köá¿›`CŒut‡üZs‹wS~Í>yW˜V£7.óÂhM?;¿ÆÆX$¿÷¶—ç ¹çßLÜ~¼|n<ÃÅÅ‚¯™)!2¬·ñyLH çÀ0¸Ã*~yœ)¦Øøÿ%+È `èdbøÒóÇÈÉ!ÿ²D|-Œý—%*¸|ªÄ ¤$à×sÜ÷ôçfÌΤ¼”#Œÿ‹dÄ"œat†ÕzËч!ßvŸ/ƒ\½1»gŽº7ºÉº«o/‹&àë§!˜sx×Óûäù~þØçËËc˜õíðýcÕM äÀû­®Ê üK³4§÷eéLÇN§³´ q,ö§déK=™¥Ï-vדˆ:¾F<˜±$”/Å范 J*»ØÜÐç°^tûµ+*W[“ é ïZžK×¢W¿Èd|踿îÜ?J¤Óì”ÒÆ[$' Ú1:ÚÎç`­À‹Kã~Ó™ñEÿ¥­g,Bëu÷¹Wzï`Œ¡9!d¬À¦á¢ŸG¢-†Ôj1ÓÂNŸZë½Î‰²pÂIJ?lu@°K|†í:›îí‘W5‘sd·`W‹E+uÞÞèå†\-%ìát;뱋¼$õ^‚R?š8ýü|žý”“¸Ñ–t(z$³O¨´ Ó—x‹qN`¤ì Ì\ôö×;þž¬ÓzeÝ냾”#¯'µ*pìÑg1EY&?¤–š@8NPÏÁ f3–ÁOAÎÞî¿[Lj…hô³½Mö9M©öäl]¬%D‹ùÀ”3¹§Ýô¸u‡3‰Ï:ˆ3•Ò|6dÕVï=½¾ë|¹ÏB½×õÎuÙÑû1ŸÍóÿ, ¸C„>6ðœý‹aØÓè¥N>\ï_‡-Å~-ÁŸô«y/5ä˜R†æVÊÏ@ÝÈÈ@TÝjß¿ß]¾™­°Û-öy·P*‘ö÷ª† ©x䟾RDdÊ ……—Ù‹1@‡^[ëÃJ˜«Xf†5yåb~£(Œi¨«¿¬nÏ+ŸWßcÞ¯ðÿ7€È8³À©HÄxø¢æÕ (ˆaåBH# ç5ÉòtÕ”EÑ)%Ÿyd)”ŒZ4aLÎ@ÏJû®ƒŽ)< T`¥*”jÉ(Gp JÁÜ °ÞC ° •MeñX¡ ÜXP’Gô J¯c0_Ñ1Î!2¯û¬ò øð+WSPŽGl´2¼*%À@HñnZ<0Pqb°¥KEºXF y „”„%•’ ÌÂðk¶07RœÊH‰”¹7RàýŽÒ2eÚeÁ>±Á>\Nük,çsÞì¬Rd+ÞúÈ­‘b¡Ȫh,+Eš+X±T¡8Â>LÌ@§óxËE@žêÀS¿§Î XÚ ðW-4-:}&åUŒ +d¥PH ¤Ê좔2B ’Žò=«ZRI”)jáØORB3 `³-TEc.V£†Šc<( e$MË6Å8‡ãS–Ž%(&qãQß9¢£ùa•¤ …#Hî" Jw ¡ˆfbMM›Ô4XÅây,ÊN„´Ç„f³øLÄ}:z))Ši¥ÈÜ È¿’×Ö=d”«ü^é‘$”¦˜eQ8 %ª¦J£$nBÏ´P»èô Ö¦ûö0ÄH×3ú·<¾RÙ¨$¼”¥”qQÅxDzÚV`‰)²ñt%R`´í! ®”Væ)žPBŸ5P¢x-Ø ´:¡d25¤æÏÈ»ô~Wœm2öI“bóp©< ,O{Ñ!Å’1 L(;p úO\¢“a‰¶…F´¶ë eFJVµÈ!‰ŽAæ÷â®yƒ©–x¤¸!ùJÆ1²Át¼OüΧNW ±¨9D€Ø(²„D 9ê9jØtR Ì O,Kð­ ò~Š„NR䀛"Oj} ˆk`Ò™¡zCáaM16Pª“Œ€+èÚl“„t/¶å >yzGR=‰Kè@V¹PˆK˜ì“>“ª#Ái¡á cÇ5?€+¦%i$+E€˜Þ ÜTߊ«~ç­1L߬$¥ü,ø pâô±´ôJ‘b„áßéd“˜¼&$$#È$ÂŒŽ,Å´`bAäe±"“·ÐÅ¡“CåY‚˜„Ò°·°¦ªy@ueSÛ¤Ñ'¬š`YYÁب¨©m ÑšÐDRˆË gÍ‘*,”wM¨:ÜÀ‚^+´Š ”*Є‘B¬Ö£ Phr0@çT1 íj&ŽÁþ¶‡o˜%×N¦ïl"Ö)EÜ ü´Hæ!I',†}£'PbÔ™SÐjO©I:â&‡ªá„$à$;ßlÝžKrÈ Ó=Ñ$“Cž¬AÁ ¬CÅ Ç"8W-‘‚’¥aZ äaNXØY} !›°Wz>t <Œà„å¡Z‰uÙ2 ÕCq'R©!xÁ&µ!0iq|8wÐ¥u‰MôЀ¤~&fÔFÈ‚L˜TZ>K¡V›¿xfaº‹mÎJà-ùOªðXÒåmjØÙ¸—@ÍMsÌ„Ä%θ¦0—¤4ST-5O.À&qIåqT _[¯$S+ƒ©ŸN_!VÁV>£¸—U¨>Åšœ©2 JÕ™„:yQ¨ój’ñPH"†0l³-†æ@–ëmuø-\Ûó<¬í‚~ ‚qZ£‚PÇV­ þò<+!6©¦¡AO¼QÄ_K µ†$õˆm½§V™ù.Ì ,!è÷4Ò(“n´Zò©ß Öu¼ž—Î!ÞË”@ŸçYú³äA¢CÖ{ÊKÔfS5I…¤ì˜(…RÔ®X„èRtuÛ€²gD`RÍ$N&|º|¹¨·ÄÊ6ÑzbR&_a¥êÆ¿b{h¥›&w“P'/LjˆA+Pªx|Ì¥…ÑÄ‚Â)œ€èöªWh/‡1žÅ5=žÜë\ï¶W«wk«2û7ov×»‡Ýö¾õþíÓ3Œl%Á§6Š|y{µ]{žÍ˜8ü OÎï†qÐæåÊÿu·³æÙng ü¢·è£öxà9p{Õ½U—îTõ©uE}ê³UX ÈPà*Sb⇪‡'Y†Q/Ù>2$˜('E¦;ÉH{’Y´{²Q~¦s;{ñØ’ï‹ê,æµ—^Ý8Œ*lÞ¨@5Û„Í,cÙÜ }Ô_Ü|ØÝ­7,3j Ãîòâzõz°Ÿž¿^÷ºùÉ­¼­“ÆzI¾ÂÚ_Ö…/É“ôz¾µËF¡ ö‚"ÅÇâ—}{}×^Ÿ5ÒÛÅS—ó¢Å¨°oëWêûNt…‚wüp¸‚²è^4?t­coµ‰¯&_†Ë®“xø¼kÂúíq»|7Ô4tm§ÚÁÕ7`¶ÎÀÛn±7¿{ÌN´à1,ºzhÁë»|Ol—ETL‡ÎæK½¾ƒÕ÷ÚšjÜÞÏÝkùŽ7€¼v¸è|j ý“½ˆÎ箳WkÇŽ^Ÿ†ÿž{U¿8ôSžØ8Ïr;M×ÍÖD9ÙV¥80j3›Z)0M5…]{´18j&æ '³ì \öÑJŸ`®.¡ðI/B:jÉkCúáÝšýð¾Sð‰¾âçš É=n«»K&?×1ÌÑfÙƒ:u¥Ç¸`ýmg¼›ƒ0´Éš0¶ÍtyƒZùÇ®—prJÝÆ%ÃQ_+ïêØâúÆÛ>$°›º:¹ÂóÌeÏ—|¶(»SW~/ß§Zúi rÞìOµ“ÎñäÀ˜®”‘h¯m›î÷ÜN16/6y§¡:’08å•®kì/nõÄÓŧèÏG¦y׉#bìØw“ôæÑ¹ÁÿƒbåžÄ4ŸñÏ 4nyMŒZ]î÷77æè)ŽŸhèn—äºVe3 ‹E‡?4$÷Zù\ð¼ÖÔç{[WñõrAnÑW÷#¶xÝyZßÎÛß;jæä©¦gšð%‘ÃΟ¼Áv¯³ ¦]0pç–7Z”¸h¦sŠê›*Ã…köîxh–§Ÿž¸—']ñw]Ö»X6‹ÆÒ±kfc‡}ñ£fõ¾ÿxnÖ÷AïÆ>Ù}ϹÙ½€ïö=‹ýôOèxÖÈ©TɆ—ò1RåœÿÔ… †.]ñË[ í œH‘¯™¾‹ˆ!óÚ@\¶Üߪ¼xíæ AônÒ4–—5èF\<¾Nb¦«;‰úÓ2 ±–ŽxѬ'‹–ø{µÏ^oÍ//Í|»Ô©,Âì÷V?Çež}bÂ¥Ÿ­ ?&þf–A?úÑ5šæ’·‡K¨Ë«®=²ÛØŠ€“ÓÒO'Àœn"óvÞéëá¦íc4<_‘Y^øåëñ§œK€9oº.oÐôóó:§5_¨—#·©G×ïšßh>}]|‡>¶ß° ³ØSÐñ¹ }ânÉtYùùT©¨™ò‰ <Çppº–·ŒúbJ¡‡t'nòð­v˜7)>`F¶Iª}þ§¥ b’£[pÞÌÙ÷ŽÅºX‘Ÿlv³êçÑ­xjþëæåÿùYåÉà45 z|ˆH{>º‚øÄ5™°}|—¦=ü‰ê’¾ÝXp³ûÝ>±ä©kéÆWâÖþÉËEÝèŸwÇœïQ¤O×j{¹]}±Ã¥°Ëx{âÛÝâFH ÑŒ <<ÆæKõwO>ÞU!žP ô+]¾ÙÝ^½è­ù~µ×à ]ªiw…úµÌËìÊËëß~¡îú“ ús²‡&/&ݼ Ü&.•b_èF0HrYRAÞzý‚’×Çù×T»§µðý‡%ئk—gÃå E¡ؤ#סÁÅ¿m~±¾Ø'bíbÖ—Ro¢Ù²XªSØó‡Á”¯©¶Ÿw1Ä;½²8ŸK‡ ÿ|ö½ö»^endstream endobj 394 0 obj << /Type /XRef /Length 281 /Filter /FlateDecode /DecodeParms << /Columns 5 /Predictor 12 >> /W [ 1 3 1 ] /Info 3 0 R /Root 2 0 R /Size 395 /ID [<32856dc54f98b5e21400f87cdaa4fa1c>] >> stream xœcb&F~0ù‰ $À8JŽò?ƒHÌ: ›m/(öŸÙŒÆþH"ÿ3¨W­ÆûÍ[£ñ>’Èÿ F7Añ~®z4Þ‡/ùŸÁþÑF [®Ëò@¤r ˆd™ "¹dA¤pˆdû"Å‚@¤H=£lˆ-Ë "™ÜÁj ÁìÉ»D2²‚Hæl°¸*ØL°z… RPDJ®‘r{À&¿™<ì†`Ù)¥"¥%&0ƒíâ–›œv§>ˆä7™ p,~l¦ˆäÙ "9ØÁ¦½©yôÄ=©“`7K‚Iñ ° ÿ€E&M3‡€:H¯Ä2°›Á¾æ³Á$çt°]Àö2‚õZ2U‰3Ý endstream endobj startxref 355922 %%EOF gmm/inst/doc/gmm_with_R.R0000644000176200001440000002263615044423070015003 0ustar liggesusers## ----echo=FALSE--------------------------------------------------------------- library(knitr) opts_chunk$set(size='footnotesize', fig.height=5, out.width='70%') ## ----------------------------------------------------------------------------- library(gmm) ## ----------------------------------------------------------------------------- g1 <- function(tet,x) { m1 <- (tet[1]-x) m2 <- (tet[2]^2 - (x - tet[1])^2) m3 <- x^3-tet[1]*(tet[1]^2+3*tet[2]^2) f <- cbind(m1,m2,m3) return(f) } ## ----------------------------------------------------------------------------- Dg <- function(tet,x) { G <- matrix(c( 1, 2*(-tet[1]+mean(x)), -3*tet[1]^2-3*tet[2]^2,0, 2*tet[2],-6*tet[1]*tet[2]), nrow=3,ncol=2) return(G) } ## ----------------------------------------------------------------------------- set.seed(123) n <- 200 x1 <- rnorm(n, mean = 4, sd = 2) ## ----------------------------------------------------------------------------- print(res <- gmm(g1,x1,c(mu = 0, sig = 0), grad = Dg)) ## ----------------------------------------------------------------------------- summary(res) ## ----------------------------------------------------------------------------- specTest(res) ## ----------------------------------------------------------------------------- sim_ex <- function(n,iter) { tet1 <- matrix(0,iter,2) tet2 <- tet1 for(i in 1:iter) { x1 <- rnorm(n, mean = 4, sd = 2) tet1[i,1] <- mean(x1) tet1[i,2] <- sqrt(var(x1)*(n-1)/n) tet2[i,] <- gmm(g1,x1,c(0,0),grad=Dg)$coefficients } bias <- cbind(rowMeans(t(tet1)-c(4,2)),rowMeans(t(tet2)-c(4,2))) dimnames(bias)<-list(c("mu","sigma"),c("ML","GMM")) Var <- cbind(diag(var(tet1)),diag(var(tet2))) dimnames(Var)<-list(c("mu","sigma"),c("ML","GMM")) MSE <- cbind(rowMeans((t(tet1)-c(4,2))^2),rowMeans((t(tet2)-c(4,2))^2)) dimnames(MSE)<-list(c("mu","sigma"),c("ML","GMM")) return(list(bias=bias,Variance=Var,MSE=MSE)) } ## ----------------------------------------------------------------------------- g2 <- function(theta,x) { tau <- seq(1,5,length.out=10) pm <- 1 x <- matrix(c(x),ncol=1) x_comp <- x%*%matrix(tau,nrow=1) x_comp <- matrix(complex(ima=x_comp),ncol=length(tau)) emp_car <- exp(x_comp) the_car <- charStable(theta,tau,pm) gt <- t(t(emp_car) - the_car) gt <- cbind(Im(gt),Re(gt)) return(gt) } ## ----------------------------------------------------------------------------- library(stabledist) set.seed(345) x2 <- rstable(500,1.5,.5,pm=1) t0 <- c(alpha = 2, beta = 0, gamma = sd(x2)/sqrt(2), delta = 0) print(res <- gmm(g2,x2,t0)) ## ----------------------------------------------------------------------------- summary(res) ## ----------------------------------------------------------------------------- res2 <- gmm(g2,x2,t0,optfct="nlminb",lower=c(0,-1,0,-Inf),upper=c(2,1,Inf,Inf)) summary(res2) ## ----warning=FALSE------------------------------------------------------------ data(Finance) x3 <- Finance[1:1500,"WMK"] t0<-c(alpha = 1.8, beta = 0.1, gamma = sd(x3)/sqrt(2),delta = 0) res3 <- gmm(g2,x3,t0,optfct="nlminb") summary(res3) ## ----------------------------------------------------------------------------- library(car) linearHypothesis(res3,cbind(diag(2),c(0,0),c(0,0)),c(2,0)) ## ----------------------------------------------------------------------------- library(mvtnorm) set.seed(112233) sig <- matrix(c(1,.5,.5,1),2,2) n <- 400 e <- rmvnorm(n,sigma=sig) x4 <- rnorm(n) w <- exp(-x4^2) + e[,1] y <- 0.1*w + e[,2] ## ----------------------------------------------------------------------------- h <- cbind(x4, x4^2, x4^3) g3 <- y~w ## ----------------------------------------------------------------------------- summary(res <- gmm(g3,x=h)) ## ----------------------------------------------------------------------------- res2 <- gmm(g3,x=h,type='iterative',crit=1e-8,itermax=200) coef(res2) ## ----------------------------------------------------------------------------- res3 <- gmm(g3,x=h,res2$coef,type='cue') coef(res3) ## ----------------------------------------------------------------------------- confint(res3,level=.90) ## ----------------------------------------------------------------------------- plot(w,y,main="LS vs GMM estimation") lines(w,fitted(res),col=2) lines(w,fitted(lm(y~w)),col=3,lty=2) lines(w,.1*w,col=4,lty=3) legend("topleft",c("Data","Fitted GMM","Fitted LS","True line"),pch=c(1,NA,NA,NA),col=1:3,lty=c(NA,1,2,3)) ## ----------------------------------------------------------------------------- t <- 400 set.seed(345) x5 <- arima.sim(n=t,list(ar=c(1.4,-0.6),ma=c(0.6,-0.3))) x5t<-cbind(x5) for (i in 1:6) x5t<-cbind(x5t,lag(x5,-i)) x5t<-na.omit(x5t) g4<-x5t[,1]~x5t[,2]+x5t[,3] res<-gmm(g4,x5t[,4:7]) summary(res) ## ----------------------------------------------------------------------------- res2 <- gmm(g4,x=x5t[,4:7],kernel="Truncated") coef(res2) res3 <- gmm(g4,x=x5t[,4:7],kernel="Bartlett") coef(res3) res4 <- gmm(g4,x=x5t[,4:7],kernel="Parzen") coef(res4) res5<- gmm(g4,x=x5t[,4:7],kernel="Tukey-Hanning") coef(res5) ## ----------------------------------------------------------------------------- diag(vcov(res2))^.5 diag(vcov(res3))^.5 diag(vcov(res4))^.5 diag(vcov(res5))^.5 ## ----------------------------------------------------------------------------- plot(res,which=2) ## ----------------------------------------------------------------------------- plot(res,which=3) ## ----------------------------------------------------------------------------- data(Finance) r <- Finance[1:500,1:5] rm <- Finance[1:500,"rm"] rf <- Finance[1:500,"rf"] z <- as.matrix(r-rf) zm <- as.matrix(rm-rf) res <- gmm(z~zm,x=zm) coef(res) R <- cbind(diag(5),matrix(0,5,5)) c <- rep(0,5) linearHypothesis(res,R,c,test = "Chisq") ## ----eval=FALSE--------------------------------------------------------------- # test <- paste(names(coef(res)[1:5])," = 0",sep="") # linearHypothesis(res,test) ## ----------------------------------------------------------------------------- res2<-gmm(z~zm-1,cbind(1,zm)) specTest(res2) ## ----------------------------------------------------------------------------- g5 <- function(tet, x) { gmat <- (tet[1] + tet[2] * (1 + c(x[, 1]))) * (1 + x[, 2:6]) - 1 return(gmat) } res_sdf <- gmm(g5, x = as.matrix(cbind(rm, r)), c(0, 0)) specTest(res_sdf) ## ----------------------------------------------------------------------------- g6 <- function(theta, x) { t <- length(x) et1 <- diff(x) - theta[1] - theta[2] * x[-t] ht <- et1^2 - theta[3] * x[-t]^(2 * theta[4]) g <- cbind(et1, et1 * x[-t], ht, ht * x[-t]) return(g) } ## ----------------------------------------------------------------------------- rf <- Finance[,"rf"] rf <- ((1 + rf/100)^(365) - 1) * 100 dr <- diff(rf) res_0 <- lm(dr ~ rf[-length(rf)]) tet0 <- c(res_0$coef, var(residuals(res_0)), 0) names(tet0) <- c("alpha", "beta", "sigma^2", "gamma") res_rf <- gmm(g6, rf, tet0, control = list(maxit = 1000, reltol = 1e-10)) coef(res_rf) ## ----eval=FALSE--------------------------------------------------------------- # y <- rbind(y1-mean(y1),y2-mean(y2),y3-mean(y3)) # x <- rbind(x1-mean(x1),x2-mean(x2),x3-mean(x3)) # res <- gmm(y~x,h) ## ----eval=FALSE--------------------------------------------------------------- # y <- rbind(y1,y2,y3) # x <- rbind(x1,x2,x3) # res <- gmm(y~x,h) ## ----eval=FALSE--------------------------------------------------------------- # gt <- g(t0, x) # V <- kernHAC(lm(gt~1),sandwich = FALSE) # W <- solve(V) ## ----------------------------------------------------------------------------- print(res<-gmm(g4,x5t[,4:7],wmatrix="ident")) diag(vcovHAC(res))^.5 ## ----------------------------------------------------------------------------- diag(vcov(res))^.5 ## ----------------------------------------------------------------------------- print(res<-gmm(g4,x5t[,4:7], weightsMatrix = diag(5))) ## ----------------------------------------------------------------------------- tet0 <- c(mu = mean(x1), sig = sd(x1)) res_el <- gel(g1,x1,tet0) summary(res_el) ## ----------------------------------------------------------------------------- res_et <- gel(g1,x1,tet0,type="ET") coef(res_et) ## ----------------------------------------------------------------------------- res_cue <- gel(g1,x1,tet0,type="CUE") coef(res_cue) ## ----------------------------------------------------------------------------- res_etel <- gel(g1,x1,c(mu=1,sig=1),type="ETEL") coef(res_etel) ## ----warning=FALSE------------------------------------------------------------ tet0 <- gmm(g4,x=x5t[,4:7],wmatrix="ident")$coef res <- gel(g4,x=x5t[,4:7],tet0,smooth=TRUE,kernel="Truncated") summary(res) ## ----------------------------------------------------------------------------- specTest(res) ## ----------------------------------------------------------------------------- plot(res,which=4) ## ----eval=FALSE--------------------------------------------------------------- # ui=cbind(0,-1,-1) # ci <- -1 ## ----eval=FALSE--------------------------------------------------------------- # res <- gel(g4,x=dat5[,4:7],tet0,smooth=TRUE,kernel="Truncated", # constraint=TRUE, ui=ui,ci=ci) ## ----eval=FALSE--------------------------------------------------------------- # res <- gel(g4,x=dat5[,4:7],tet0,smooth=TRUE, optlam="optim") # res <- gel(g4,x=dat5[,4:7],tet0,smooth=TRUE, optlam="optim", # LambdaControl=list(trace=TRUE, parscale=rep(.1,5))) gmm/inst/doc/gmm_with_R.rnw0000644000176200001440000017532214271746755015434 0ustar liggesusers\documentclass[11pt,letterpaper]{article} \usepackage{amsthm} \usepackage[hmargin=2cm,vmargin=2.5cm]{geometry} \newtheorem{theorem}{Theorem} \newtheorem{col}{Corollary} \newtheorem{lem}{Lemma} \usepackage[utf8]{inputenc} \newtheorem{ass}{Assumption} \usepackage{amsmath} \usepackage{verbatim} \usepackage[round]{natbib} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{graphicx} \usepackage{hyperref} \hypersetup{ colorlinks, citecolor=black, filecolor=black, linkcolor=black, urlcolor=black } \bibliographystyle{plainnat} \newcommand{\E}{\mathrm{E}} \newcommand{\diag}{\mathrm{diag}} \newcommand{\Prob}{\mathrm{Pr}} \newcommand{\Var}{\mathrm{Var}} \let\proglang=\textsf \newcommand{\pkg}[1]{{\fontseries{m}\fontseries{b}\selectfont #1}} \newcommand{\Vect}{\mathrm{Vec}} \newcommand{\Cov}{\mathrm{Cov}} \newcommand{\conP}{\overset{p}{\to}} \newcommand{\conD}{\overset{d}{\to}} \newcommand\Real{ \mathbb{R} } \newcommand\Complex{ \mathbb{C} } \newcommand\Natural{ \mathbb{N} } \newcommand\rv{{\cal R}} \newcommand\Q{\mathbb{Q}} \newcommand\PR{{\cal R}} \newcommand\T{{\cal T}} \newcommand\Hi{{\cal H}} \newcommand\La{{\cal L}} \newcommand\plim{plim} \renewcommand{\epsilon}{\varepsilon} \begin{document} \author{Pierre Chauss\'e} \title{Computing Generalized Method of Moments and Generalized Empirical Likelihood with \proglang{R}} \maketitle \abstract{This paper shows how to estimate models by the generalized method of moments and the generalized empirical likelihood using the \proglang{R} package \textbf{gmm}. A brief discussion is offered on the theoretical aspects of both methods and the functionality of the package is presented through several examples in economics and finance. It is a modified version of \cite{chausse10} published in the Journal of Statistical Software. It has been adapted to the version 1.4-0. \textbf{Notice that the maintenance of the package is converging to zero. The new \pkg{momentfit} package, available on CRAN, will soon replace the \pkg{gmm} package.}} %\VignetteIndexEntry{Computing Generalized Empirical Likelihood and Generalized Method of Moments with R} %\VignetteDepends{gmm,mvtnorm,stabledist, car, MASS, timeDate, timeSeries} %\VignetteKeywords{generalized empirical likelihood, generalized method of moments, empirical likelihood, continuous updated estimator, exponential tilting, exponentially tilted empirical likelihood} %\VignettePackage{gmm} %\VignetteEngine{knitr::knitr} <>= library(knitr) opts_chunk$set(size='footnotesize', fig.height=5, out.width='70%') @ \section{Introduction} The generalized method of moments (GMM) has become an important estimation procedure in many areas of applied economics and finance since \cite{hansen82} introduced the two step GMM (2SGMM). It can be seen as a generalization of many other estimation methods like least squares (LS), instrumental variables (IV) or maximum likelihood (ML). As a result, it is less likely to be misspecified. The properties of the estimators of LS depend on the exogeneity of the regressors and the circularity of the residuals, while those of ML depend on the choice of the likelihood function. GMM is much more flexible since it only requires some assumptions about moment conditions. In macroeconomics, for example, it allows to estimate a structural model equation by equation. In finance, most data such as stock returns are characterized by heavy-tailed and skewed distributions. Because it does not impose any restriction on the distribution of the data, GMM represents a good alternative in this area as well. As a result of its popularity, most statistical packages like \proglang{Matlab}, \proglang{Gauss} or \proglang{Stata} offer tool boxes to use the GMM procedure. It is now possible to easily use this method in \proglang{R} with the new \pkg{gmm} package. Although GMM has good potential theoretically, several applied studies have shown that the properties of the 2SGMM may in some cases be poor in small samples. In particular, the estimators may be strongly biased for certain choices of moment conditions. In response to this result, \cite{hansen-heaton-yaron96} proposed two other ways to compute GMM: the iterative GMM (ITGMM) and the continuous updated GMM (CUE)\footnote{See also \cite{hall05} for a detailed presentation of most recent developments regarding GMM.}. Furthermore, another family of estimation procedures inspired by \cite{owen01}, which also depends only on moment conditions, was introduced by \cite{smith97}. It is the generalized empirical likelihood (GEL). So far, this method has not reached the popularity of GMM and it was not included in any statistical package until \pkg{gmm} was developed for \proglang{R} which also includes a GEL procedure. Asymptotic properties of GMM and generalized empirical likelihood (GEL) are now well established in the econometric literature. \cite{newey-smith04} and \cite{anatolyev05} have compared their second order asymptotic properties. In particular, they show that the second order bias of the empirical likelihood (EL) estimator, which is a special case of GEL, is smaller than the bias of the estimators from the three GMM methods. Furthermore, as opposed to GMM, the bias does not increase with the number of moment conditions. Since the efficiency improves when the number of conditions goes up, this is a valuable property. However, these are only asymptotic results which do not necessarily hold in small sample as shown by \cite{guggenberger08}. In order to analyze small sample properties, we have to rely on Monte Carlo simulations. However, Monte Carlo studies on methods such as GMM or GEL depend on complicated algorithms which are often home made. Because of that, results from such studies are not easy to reproduce. The solution should be to use a common tool which can be tested and improved upon by the users. Because it is open source, \proglang{R} offers a perfect platform for such tool. The \pkg{gmm} package allows to estimate models using the three GMM methods, the empirical likelihood and the exponential tilting, which belong to the family of GEL methods, and the exponentially tilted empirical likelihood which was proposed by \cite{schennach07}, Also it offers several options to estimate the covariance matrix of the moment conditions. Users can also choose between \textit{optim}, if no restrictions are required on the coefficients of the model to be estimated, and either \textit{nlminb} or \textit{constrOptim} for constrained optimizations. The results are presented in such a way that \proglang{R} users who are familiar with \textit{lm} objects, find it natural. In fact, the same methods are available for \textit{gmm} and \textit{gel} objects produced by the estimation procedures. The paper is organized as follows. Section 2 presents the theoretical aspects of the GMM method along with several examples in economics and finance. Through these examples, the functionality of the \pkg{gmm} packages is presented in details. Section 3 presents the GEL method with some of the examples used in section 2. Section 4 concludes and Section 5 gives the computational details of the package. \section{Generalized method of moments} This section presents an overview of the GMM method. It is intended to help the users understand the options that the \pkg{gmm} package offers. For those who are not familiar with the method and require more details, see \cite{hansen82} and \cite{hansen-heaton-yaron96} for the method itself, \cite{newey-west94} and \cite{andrews91} for the choice of the covariance matrix or \cite{hamilton94}. We want to estimate a vector of parameters $\theta_0 \in \Real^p$ from a model based on the following $q\times 1$ vector of unconditional moment conditions: \begin{equation}\label{mcond} E[g(\theta_0,x_i)] = 0 , \end{equation} where $x_i$ is a vector of cross-sectional data, time series or both. In order for GMM to produce consistent estimates from the above conditions, $\theta_0$ has to be the unique solution to $E[g(\theta,x_i)]=0$ and be an element of a compact space. Some boundary assumptions on higher moments of $g(\theta,x_i)$ are also required. However, it does not impose any condition on the distribution of $x_i$, except for the degree of dependence of the observations when it is a vector of time series. Several estimation methods such as least squares (LS), maximum likelihood (ML) or instrumental variables (IV) can also be seen as being based on such moment conditions, which make them special cases of GMM. For example, the following linear model: \[ Y = X\beta + u , \] where $Y$ and $X$ are respectively $n\times 1$ and $n\times k$ matrices, can be estimated by LS. The estimate $\hat{\beta}$ is obtained by solving $\min_\beta \|u\|^2$ and is therefore the solution to the following first order condition: \[ \frac{1}{n}X'u(\beta) = 0 , \] which is the estimate of the moment condition $E(X_iu_i(\beta))=0$. The same model can be estimated by ML in which case the moment condition becomes: \[ E\left[\frac{d l_i(\beta)}{d\beta}\right]=0 , \] where $l_i(\beta)$ is the density of $u_i$. In presence of endogeneity of the explanatory variable $X$, which implies that $E(X_iu_i)\neq 0$, the IV method is often used. It solves the endogeneity problem by substituting $X$ by a matrix of instruments $H$, which is required to be correlated with $X$ and uncorrelated with $u$. These properties allow the model to be estimated by the conditional moment condition $E(u_i|H_i)=0$ or its implied unconditional moment condition $E(u_iH_i)=0$. In general we say that $u_i$ is orthogonal to an information set $I_i$ or that $E(u_i|I_i)=0$ in which case $H_i$ is a vector containing functions of any element of $I_i$. The model can therefore be estimated by solving \[ \frac{1}{T}H'u(\beta)=0 . \] When there is no assumption on the covariance matrix of $u$, the IV corresponds to GMM. If $E(X_iu_i)=0$ holds, generalized LS with no assumption on the covariance matrix of $u$ other than boundary ones is also a GMM method. For the ML procedure to be viewed as GMM, the assumption on the distribution of $u$ must be satisfied. If it is not, but $E(dl_i(\theta_0)/d\theta)=0$ holds, as it is the case for linear models with non normal error terms, the pseudo-ML which uses a robust covariance matrix can be seen as being a GMM method. Because GMM depends only on moment conditions, it is a reliable estimation procedure for many models in economics and finance. For example, general equilibrium models suffer from endogeneity problems because these are misspecified and they represent only a fragment of the economy. GMM with the right moment conditions is therefore more appropriate than ML. In finance, there is no satisfying parametric distribution which reproduces the properties of stock returns. The family of stable distributions is a good candidate but only the densities of the normal, Cauchy and Lévy distributions, which belong to this family, have a closed form expression. The distribution-free feature of GMM is therefore appealing in that case. Although GMM estimators are easily consistent, efficiency and bias depend on the choice of moment conditions. Bad instruments implies bad information and therefore low efficiency. The effects on finite sample properties are even more severe and are well documented in the literature on weak instruments. \cite{newey-smith04} show that the bias increases with the number of instruments but efficiency decreases. Therefore, users need to be careful when selecting the instruments. \cite{carrasco07} gives a good review of recent developments on how to choose instruments in her introduction. In general, the moment conditions $E(g(\theta_0,x_i))=0$ is a vector of nonlinear functions of $\theta_0$ and the number of conditions is not limited by the dimension of $\theta_0$. Since efficiency increases with the number of instruments $q$ is often greater than $p$, which implies that there is no solution to \[ \bar{g}(\theta) \equiv \frac{1}{n}\sum_{i=1}^n g(\theta,x_i)=0. \] The best we can do is to make it as close as possible to zero by minimizing the quadratic function $\bar{g}(\theta)'W\bar{g}(\theta)$, where $W$ is a positive definite and symmetric $q\times q$ matrix of weights. The optimal matrix $W$ which produces efficient estimators is defined as: \begin{equation}\label{optw} W^* = \left\{\lim_{n\rightarrow \infty} Var(\sqrt{n} \bar{g}(\theta_0)) \equiv \Omega(\theta_0) \right\}^{-1}. \end{equation} This optimal matrix can be estimated by an heteroskedasticity and auto-correlation consistent (HAC) matrix like the one proposed by \cite{newey-west87a}. The general form is: \begin{equation}\label{optw_hat} \hat{\Omega} = \sum_{s=-(n-1)}^{n-1} k_h(s) \hat{\Gamma}_s(\theta^*), \end{equation} where $k_h(s)$ is a kernel, $h$ is the bandwidth which can be chosen using the procedures proposed by \cite{newey-west87a} and \cite{andrews91}, \[ \hat{\Gamma}_s(\theta^*) = \frac{1}{n}\sum_i g(\theta^*,x_i)g(\theta^*,x_{i+s})' \] and $\theta^*$ is a convergent estimate of $\theta_0$. There are many choices for the HAC matrix. They depend on the kernel and bandwidth selection. Although the choice does not affect the asymptotic properties of GMM, very little is known about the impacts in finite samples. The GMM estimator $\hat{\theta}$ is therefore defined as: \begin{equation}\label{gmm} \hat{\theta} = \arg\min_{\theta} \bar{g}(\theta)'\hat{\Omega}(\theta^*)^{-1}\bar{g}(\theta) \end{equation} The original version of GMM proposed by \cite{hansen82} is called two-step GMM (2SGMM). It computes $\theta^*$ by minimizing $\bar{g}(\theta)'\bar{g}(\theta)$. The algorithm is therefore: \begin{itemize} \item[1-] Compute $\theta^* = \arg\min_\theta \bar{g}(\theta)'\bar{g}(\theta)$ \item[2-] Compute the HAC matrix $\hat{\Omega}(\theta^*)$ \item[3-] Compute the 2SGMM $\hat{\theta} = \arg\min_\theta \bar{g}(\theta)'\big[\hat{\Omega}(\theta^*)\big]^{-1}\bar{g}(\theta)$ \end{itemize} In order to improve the properties of 2SGMM, \cite{hansen-heaton-yaron96} suggest two other methods. The first one is the iterative version of 2SGMM (ITGMM) and can be computed as follows: \begin{itemize} \item[1-] Compute $\theta^{(0)} = \arg\min_\theta \bar{g}(\theta)'\bar{g}(\theta)$ \item[2-] Compute the HAC matrix $\hat{\Omega}(\theta^{(0)})$ \item[3-] Compute the $\theta^{(1)} = \arg\min_\theta \bar{g}(\theta)'\big[\hat{\Omega}(\theta^{(0)})\big]^{-1}\bar{g}(\theta)$ \item[4-] If $\| \theta^{(0)}-\theta^{(1)}\|< tol$ stops, else $\theta^{(0)}=\theta^{(1)}$ and go to 2- \item[5-] Define the ITGMM estimator $\hat{\theta}$ as $\theta^{(1)}$ \end{itemize} where $tol$ can be set as small as we want to increase the precision. In the other method, no preliminary estimate is used to obtain the HAC matrix. The latter is treated as a function of $\theta$ and is allowed to change when the optimization algorithm computes the numerical derivatives. It is therefore continuously updated as we move toward the minimum. For that, it is called the continuous updated estimator (CUE). This method is highly nonlinear. It is therefore crucial to choose a starting value that is not too far from the minimum. A good choice is the estimate from 2SGMM which is known to be root-n convergent. The algorithm is: \begin{itemize} \item[1-] Compute $\theta^*$ using 2SGMM \item[2-] Compute the CUE estimator defined as \[ \hat{\theta} = \arg\min_\theta \bar{g}(\theta)'\big[\hat{\Omega}(\theta)\big]^{-1}\bar{g}(\theta) \] using $\theta^*$ as starting value. \end{itemize} According to \cite{newey-smith04} and \cite{anatolyev05}, 2SGMM and ITGMM are second order asymptotically equivalent. On the other hand, they show that the second order asymptotic bias of CUE is smaller. The difference in the bias comes from the randomness of $\theta^*$ in $\Omega(\theta^*)$. Iterating only makes $\theta^*$ more efficient. These are second order asymptotic properties. They are informative but may not apply in finite samples. In most cases, we have to rely on numerical simulations to analyze the properties in small samples. Given some regularity conditions, the GMM estimator converges as $n$ goes to infinity to the following distribution: \[ \sqrt{n}(\hat{\theta}-\theta_0) \stackrel{L}{\rightarrow} N(0,V), \] where \[ V = E\left(\frac{\partial g(\theta_0,x_i)}{\partial\theta}\right)'\Omega(\theta_0)^{-1}E\left(\frac{\partial g(\theta_0,x_i)}{\partial\theta}\right) \] Inference can therefore be performed on $\hat{\theta}$ using the assumption that it is approximately distributed as $N(\theta_0,\hat{V}/n)$. If $q>p$, we can perform a J-test to verify if the moment conditions hold. The null hypothesis and the statistics are respectively $H0:E[g(\theta,x_i)]=0$ and: \[ n\bar{g}(\hat{\theta})'[\hat{\Omega}(\theta^*)]^{-1}\bar{g}(\hat{\theta}) \stackrel{L}{\rightarrow} \chi^2_{q-p}. \] \section{GMM with R} The \pkg{gmm} package can be loaded the usual way. <<>>= library(gmm) @ The main function is \textit{gmm()} which creates an object of class \textit{gmm}. Many options are available but in many cases they can be set to their default values. They are explained in details below through examples. The main arguments are \textit{g} and \textit{x}. For a linear model, \textit{g} is a formula like \textit{y~z1+z2} and \textit{x} the matrix of instruments. In the nonlinear case, they are respectively the function $g(\theta,x_i)$ and its argument. The available methods are \textit{coef}, \textit{vcov}, \textit{summary}, \textit{residuals}, \textit{fitted.values}, \textit{plot}, \textit{confint}. The model and data in a \textit{data.frame} format can be extracted by the generic function \textit{model.frame}. \subsection{Estimating the parameters of a normal distribution} This example\footnote{Thanks to Dieter Rozenich for his suggestion.}, is not something we want to do in practice, but its simplicity allows us to understand how to implement the \textit{gmm()} procedure by providing the gradient of $g(\theta,x_i)$. It is also a good example of the weakness of GMM when the moment conditions are not sufficiently informative. In fact, the ML estimators of the mean and the variance of a normal distribution are more efficient because the likelihood carries more information than few moment conditions. For the two parameters of a normal distribution $(\mu,\sigma)$ we have the following vector of moment conditions: \[ E[g(\theta,x_i)] \equiv E\left[ \begin{array}{c} \mu - x_{i} \\ \sigma^2 - (x_{i}-\mu)^2 \\ x_{i}^{3} - \mu (\mu^2+3\sigma^{2}) \end{array} \right] = 0 , \] where the first two can be directly obtained by the definition of $(\mu,\sigma)$ and the last comes from the third derivative of the moment generating function evaluated at 0. We first need to create a function $g(\theta,x)$ which returns an $n\times 3$ matrix: <<>>= g1 <- function(tet,x) { m1 <- (tet[1]-x) m2 <- (tet[2]^2 - (x - tet[1])^2) m3 <- x^3-tet[1]*(tet[1]^2+3*tet[2]^2) f <- cbind(m1,m2,m3) return(f) } @ The following is the gradient of $\bar{g}(\theta)$: \[ G\equiv \frac{\partial \bar{g}(\theta)}{\partial \theta} = \left( \begin{array}{cc} 1 & 0\\ 2(\bar{x}-\mu) & 2\sigma\\ -3(\mu^{2}+\sigma^{2}) & -6\mu\sigma \end{array} \right). \] If provided, it will be used to compute the covariance matrix of $\hat{\theta}$. It can be created as follows: <<>>= Dg <- function(tet,x) { G <- matrix(c( 1, 2*(-tet[1]+mean(x)), -3*tet[1]^2-3*tet[2]^2,0, 2*tet[2],-6*tet[1]*tet[2]), nrow=3,ncol=2) return(G) } @ First we generate normally distributed random numbers: <<>>= set.seed(123) n <- 200 x1 <- rnorm(n, mean = 4, sd = 2) @ We then run \textit{gmm} using the starting values $(\mu_0,\sigma^2_0)=(0,0)$ <<>>= print(res <- gmm(g1,x1,c(mu = 0, sig = 0), grad = Dg)) @ The \textit{summary} method prints more results from the estimation: <<>>= summary(res) @ The section "Initial values of the coefficients" shows the first step estimates used to either compute the weighting matrix in the 2-step GMM or the fixed bandwidth in CUE or iterative GMM. The J-test of over-identifying restrictions can also be extracted by using the method \textit{specTest}: <<>>= specTest(res) @ A small simulation using the following function shows that ML produces estimators with smaller mean squared errors than GMM based on the above moment conditions. However, it is not GMM but the moment conditions that are not efficient, because ML is GMM with the likelihood derivatives as moment conditions. <<>>= sim_ex <- function(n,iter) { tet1 <- matrix(0,iter,2) tet2 <- tet1 for(i in 1:iter) { x1 <- rnorm(n, mean = 4, sd = 2) tet1[i,1] <- mean(x1) tet1[i,2] <- sqrt(var(x1)*(n-1)/n) tet2[i,] <- gmm(g1,x1,c(0,0),grad=Dg)$coefficients } bias <- cbind(rowMeans(t(tet1)-c(4,2)),rowMeans(t(tet2)-c(4,2))) dimnames(bias)<-list(c("mu","sigma"),c("ML","GMM")) Var <- cbind(diag(var(tet1)),diag(var(tet2))) dimnames(Var)<-list(c("mu","sigma"),c("ML","GMM")) MSE <- cbind(rowMeans((t(tet1)-c(4,2))^2),rowMeans((t(tet2)-c(4,2))^2)) dimnames(MSE)<-list(c("mu","sigma"),c("ML","GMM")) return(list(bias=bias,Variance=Var,MSE=MSE)) } @ The following results can be reproduced with $n=50$, $iter=2000$ and by setting \textit{set.seed(345)}: \begin{center} \begin{tabular}{|c|c|c|c||c|c|c|} \hline &\multicolumn{3}{c||}{$\mu$}&\multicolumn{3}{c}{$\sigma$} \\ \hline & Bias & Variance & MSE & Bias & Variance &MSE \\ \hline GMM& 0.0020 &0.0929 & 0.0928& -0,0838 & 0.0481& 0.0551\\ ML& 0.0021 &0.0823 & 0.0822 & -0.0349 & 0.0411& 0.0423 \\ \hline \end{tabular} \end{center} \subsection{Estimating the parameters of a stable distribution} The previous example showed that ML should be used when the true distribution is known. However, when the density does not have a closed form expression, we have to consider other alternatives. \cite{garcia-renault-veredas06} propose to use indirect inference and perform a numerical study to compare it with several other methods. One of them is GMM for a continuum of moment conditions and was suggested by \cite{carrasco-florens02}. It uses the fact that the characteristic function $E(e^{ix_i\tau})$, where $i$ is the imaginary number and $\tau\in\Real$, has a closed form expression (for more details on stable distribution, see \cite{nolan09}). The \pkg{gmm} package does not yet deal with continuum of moment conditions but we can choose a certain grid $\{\tau_1,..., \tau_q\}$ over a given interval and estimate the parameters using the following moment conditions: \[ E\left[e^{ix_i\tau_l}-\Psi(\theta;\tau_l) \right]=0~~\mbox{for}~~l=1,...,q~, \] where $\Psi(\theta;\tau_l)$ is the characteristic function. There is more than one way to define a stable distribution and it depends on the choice of parametrization. We will follow the notation of \cite{nolan09} and consider stable distributions $S(\alpha,\beta,\gamma,\delta;1)$, where $\alpha\in (0,2]$ is the characteristic exponent and $\beta\in[-1,1]$, $\gamma>0$ and $\delta\in\Real$ are respectively the skewness, the scale and the location parameters. The last argument defines which parametrization we use. The \pkg{stabledist} package of \cite{stabledist} offers a function to generate random variables from stable distributions and uses the same notation. This parametrization implies that: \[ \Psi(\theta;\tau_l) = \left\{ \begin{array}{lcr} \exp{(-\gamma^\alpha|\tau_l|^\alpha[1-i\beta(\tan{\frac{\pi\alpha}{2}})(\mathrm{sign}(\tau_l))] + i\delta \tau_l)} &\mbox{for}&\alpha\neq 1\\ \exp{(-\gamma|\tau_l|[1+i\beta\frac{2}{\pi}(\mathrm{sign}(\tau_l))\log{|\tau_l|}] + i\delta \tau_l)} &\mbox{for}&\alpha= 1\\ \end{array} \right. , \] The function \textit{charStable} included in the package computes the characteristic function and can be used to construct $g(\theta,x_i)$. To avoid dealing with complex numbers, it returns the imaginary and real parts in separate columns because both should have zero expectation. The function is: <<>>= g2 <- function(theta,x) { tau <- seq(1,5,length.out=10) pm <- 1 x <- matrix(c(x),ncol=1) x_comp <- x%*%matrix(tau,nrow=1) x_comp <- matrix(complex(ima=x_comp),ncol=length(tau)) emp_car <- exp(x_comp) the_car <- charStable(theta,tau,pm) gt <- t(t(emp_car) - the_car) gt <- cbind(Im(gt),Re(gt)) return(gt) } @ The parameters of a simulated random vector can be estimated as follows (by default, $\gamma$ and $\delta$ are set to $1$ and $0$ respectively in \textit{rstable}). For the example, the starting values are the ones of a normal distribution with mean 0 and variance equals to \textit{var(x)}: <<>>= library(stabledist) set.seed(345) x2 <- rstable(500,1.5,.5,pm=1) t0 <- c(alpha = 2, beta = 0, gamma = sd(x2)/sqrt(2), delta = 0) print(res <- gmm(g2,x2,t0)) @ The result is not very close to the true parameters. But we can see why by looking at the J-test that is provided by the \textit{summary} method: <<>>= summary(res) @ The null hypothesis that the moment conditions are satisfied is rejected. For nonlinear models, a significant J-test may indicate that we have not reached the global minimum. Furthermore, the standard deviation of the coefficient of $\delta$ indicates that the covariance matrix is nearly singular. Notice also that the convergence code is equal to 1, which indicates that the algorithm did not converge. We could try different starting values, increase the number of iterations in the control option of \textit{optim} or use \textit{nlminb} which allows to put restrictions on the parameter space. The former would work but the latter will allow us to see how to select another optimizer. The option \textit{optfct} can be modified to use this algorithm instead of \textit{optim}. In that case, we can specify the upper and lower bounds of $\theta$. <<>>= res2 <- gmm(g2,x2,t0,optfct="nlminb",lower=c(0,-1,0,-Inf),upper=c(2,1,Inf,Inf)) summary(res2) @ Although the above modification solved the convergence problem, there is another issue that we need to address. The first step estimate used to compute the weighting matrix is almost identical to the starting values. There is therefore a convergence problem in the first step. In fact, choosing the initial $\alpha$ to be on the boundary was not a wise choice. Also, it seems that an initial value of $\beta$ equals to zero makes the objective function harder to minimize. Having a gobal minimum for the first step estimate is important if we care about efficiency. A wrong estimate will cause the weighting matrix not being a consistent estimate of the optimal matrix. The information about convergence is included in the argument 'initialAlgoInfo' of the gmm object. We conclude this example by estimating the parameters for a vector of stock returns from the data set \textit{Finance} that comes with the \pkg{gmm} package. <>= data(Finance) x3 <- Finance[1:1500,"WMK"] t0<-c(alpha = 1.8, beta = 0.1, gamma = sd(x3)/sqrt(2),delta = 0) res3 <- gmm(g2,x3,t0,optfct="nlminb") summary(res3) @ For this sub-sample, the hypothesis that the return follows a stable distribution is rejected. The normality assumption can be analyzed by testing $H_0:\alpha=2,\beta=0$ using \textit{linearHypothesis} from the \pkg{car} package: <<>>= library(car) linearHypothesis(res3,cbind(diag(2),c(0,0),c(0,0)),c(2,0)) @ It is clearly rejected. The result is even stronger if the whole sample is used. \subsection{A linear model with iid moment conditions} We want to estimate a linear model with an endogeneity problem. It is the model used by \cite{carrasco07} to compare several methods which deal with the many instruments problem. We want to estimate $\delta$ from: \[ y_i = \delta W_i + \epsilon_i \] with $\delta=0.1$ and \[ W_i = e^{-x_i^2} + u_i , \] where $(\epsilon_i,u_i) \sim ~iidN(0,\Sigma)$ with \[ \Sigma = \left( \begin{array}{cc} 1&0.5\\ 0.5&1 \end{array} \right) \] Any function of $x_i$ can be used as an instrument because it is orthogonal to $\epsilon_i$ and correlated with $W_i$. There is therefore an infinite number of possible instruments. For this example, $(x_i,x_i^2,x_i^3)$ will be the selected instruments and the sample size is set to $n=400$: <<>>= library(mvtnorm) set.seed(112233) sig <- matrix(c(1,.5,.5,1),2,2) n <- 400 e <- rmvnorm(n,sigma=sig) x4 <- rnorm(n) w <- exp(-x4^2) + e[,1] y <- 0.1*w + e[,2] @ where \textit{rmvnorm} is a multivariate normal distribution random generator which is included in the package \pkg{mvtnorm} (\cite{mvtnorm}). For a linear model, the $g$ argument is a formula that specifies the right- and left-hand sides as for \textit{lm} and $x$ is the matrix of instruments: <<>>= h <- cbind(x4, x4^2, x4^3) g3 <- y~w @ By default, an intercept is added to the formula and a vector of ones to the matrix of instruments. It implies the following moment conditions: \[ E\left(\begin{array}{c} (y_i-\alpha-\delta W_i)\\ (y_i-\alpha-\delta W_i)x_i\\ (y_i-\alpha-\delta W_i)x_i^2\\ (y_i-\alpha-\delta W_i)x_i^3 \end{array} \right)=0 \] In order the remove the intercept, -1 has to be added to the formula. In that case there is no column of ones added to the matrix of instruments. To keep the condition that the expected value of the error terms is zero, the column of ones needs to be included manually. We know that the moment conditions of this example are iid. Therefore, we can add the option \textit{vcov="iid"}. This option tells \textit{gmm} to estimate the covariance matrix of $\sqrt{n}\bar{g}(\theta^*)$ as follows: \[ \hat{\Omega}(\theta^*) = \frac{1}{n}\sum_{i=1}^n g(\theta^*,x_i)g(\theta^*,x_i)' \] However, it is recommended not to set this option to ``iid" in practice with real data because one of the reasons we want to use GMM is to avoid such restrictions. Finally, it is not necessary to provide the gradient when the model is linear since it is already included in \textit{gmm}. The first results are: <<>>= summary(res <- gmm(g3,x=h)) @ By default, the 2SGMM is computed. Other methods can be chosen by modifying the option ``type". The second possibility is ITGMM: <<>>= res2 <- gmm(g3,x=h,type='iterative',crit=1e-8,itermax=200) coef(res2) @ The procedure iterates until the difference between the estimates of two successive iterations reaches a certain tolerance level, defined by the option \textit{crit} (default is $10^{-7}$), or if the number of iterations reaches \textit{itermax} (default is 100). In the latter case, a message is printed to indicate that the procedure did not converge. The third method is CUE. As you can see, the estimates from ITGMM is used as starting values. However, the starting values are required only when \textit{g} is a function. When \textit{g} is a formula, the default starting values are the ones obtained by setting the matrix of weights equal to the identity matrix. <<>>= res3 <- gmm(g3,x=h,res2$coef,type='cue') coef(res3) @ It is possible to produce confidence intervals by using the method \textit{confint}: <<>>= confint(res3,level=.90) @ Whether \textit{optim} or \textit{nlminb} is used to compute the solution, it is possible to modify their default options by adding \textit{control=list()}. For example, you can keep track of the convergence with \textit{control=list(trace=TRUE)} or increase the number of iterations with \textit{control=list(maxit=1000)}. You can also choose the \textit{BFGS} algorithm with \textit{method="BFGS"} (see \textit{help(optim)} for more details). The methods \textit{fitted} and \textit{residuals} are also available for linear models. We can compare the fitted values of \textit{lm} with the ones from \textit{gmm} to see why this model cannot be estimated by LS. \begin{center} <<>>= plot(w,y,main="LS vs GMM estimation") lines(w,fitted(res),col=2) lines(w,fitted(lm(y~w)),col=3,lty=2) lines(w,.1*w,col=4,lty=3) legend("topleft",c("Data","Fitted GMM","Fitted LS","True line"),pch=c(1,NA,NA,NA),col=1:3,lty=c(NA,1,2,3)) @ \end{center} The LS seems to fit the model better. But the graphics hides the endogeneity problem. LS overestimates the relationship between $y$ and $w$ because it does not take into account the fact that some of the correlation is caused by the fact that $y_i$ and $w_i$ are positively correlated with the error term $\epsilon_i$. Finally, the \textit{plot} method produces some graphics to analyze the properties of the residuals. It can only be applied to \textit{gmm} objects when \textit{g} is a formula because when \textit{g} is a function, residuals are not defined. \subsection{Estimating the AR coefficients of an ARMA process} \label{ar} The estimation of auto-regressive coefficients of ARMA(p,q) processes is better performed by ML or nonlinear LS. But in Monte Carlo experiments, it is often estimated by GMM to study its properties. It gives a good example of linear models with endogeneity problems in which the moment conditions are serially correlated and possibly conditionally heteroskedastic. As opposed to the previous example, the choice of the HAC matrix becomes an important issue. We want to estimate the AR coefficients of the following process: \[ X_t = 1.4 X_{t-1} - 0.6X_{t-2} + u_t \] where $u_t = 0.6\epsilon_{t-1} -0.3 \epsilon_{t-2} + \epsilon_t$ and $\epsilon_t\sim iidN(0,1)$. This model can be estimated by GMM using any $X_{t-s}$ for $s>2$, because they are uncorrelated with $u_t$ and correlated with $X_{t-1}$ and $X_{t-2}$. However, as $s$ increases the quality of the instruments decreases since the stationarity of the process implies that the auto-correlation goes to zero. For this example, the selected instruments are $(X_{t-3},X_{t-4},X_{t-5},X_{t-6})$ and the sample size equals 400. The ARMA(2,2) process is generated by the function \textit{arima.sim}: <<>>= t <- 400 set.seed(345) x5 <- arima.sim(n=t,list(ar=c(1.4,-0.6),ma=c(0.6,-0.3))) x5t<-cbind(x5) for (i in 1:6) x5t<-cbind(x5t,lag(x5,-i)) x5t<-na.omit(x5t) g4<-x5t[,1]~x5t[,2]+x5t[,3] res<-gmm(g4,x5t[,4:7]) summary(res) @ The optimal matrix, when moment conditions are based on time series, is an HAC matrix which is defined by equation (\ref{optw_hat}). Several estimators of this matrix have been proposed in the literature. Given some regularity conditions, they are asymptotically equivalent. However, their impacts on the finite sample properties of GMM estimators may differ. The \pkg{gmm} package uses the \pkg{sandwich} package to compute these estimators which are well explained by \cite{zeileis06} and \cite{zeileis04}. We will therefore briefly summarize the available options. The option \textit{kernel} allows to choose between five kernels: Truncated, Bartlett, Parzen, Tukey-Hanning and Quadratic spectral\footnote{The first three have been proposed by \cite{white84}, \cite{newey-west87a} and \cite{gallant87} respectively and the last two, applied to HAC estimation, by \cite{andrews91}. But the latter gives a good review of all five.}. By default, the Quadratic Spectral kernel is used as it was shown to be optimal by \cite{andrews91} with respect to some mean squared error criterion. In most statistical packages, the Bartlett kernel is used for its simplicity. It makes the estimation of large models less computationally intensive. It may also make the \textit{gmm} algorithm more stable numerically when dealing with highly nonlinear models, especially with CUE. We can compare the results with different choices of kernel: <<>>= res2 <- gmm(g4,x=x5t[,4:7],kernel="Truncated") coef(res2) res3 <- gmm(g4,x=x5t[,4:7],kernel="Bartlett") coef(res3) res4 <- gmm(g4,x=x5t[,4:7],kernel="Parzen") coef(res4) res5<- gmm(g4,x=x5t[,4:7],kernel="Tukey-Hanning") coef(res5) @ The similarity of the results is not surprising since the matrix of weights should only affect the efficiency of the estimator. We can compare the estimated standard deviations using the method \textit{vcov}: <<>>= diag(vcov(res2))^.5 diag(vcov(res3))^.5 diag(vcov(res4))^.5 diag(vcov(res5))^.5 @ which shows, for this example, that the Bartlett kernel generates the estimates with the smallest variances. However, it does not mean it is better. We have to run simulations and compute the true variance if we want to compare them. In fact, we do not know which one produces the most accurate estimate of the variance. The second options is for the bandwidth selection. By default it is the automatic selection proposed by \cite{andrews91}. It is also possible to choose the automatic selection of \cite{newey-west94} by adding \textit{bw=bwNeweyWest} (without quotes because \textit{bwNeweyWest} is a function). A prewhitened kernel estimator can also be computed using the option \textit{prewhite=p}, where $p$ is the order of the vector auto-regressive (VAR) used to compute it. By default, it is set to \textit{FALSE}. \cite{andrews-monahan92} show that a prewhitened kernel estimator improves the properties of hypothesis tests on parameters. Finally, the \textit{plot} method can be applied to \textit{gmm} objects to do a Q-Q plot of the residuals: \begin{center} <<>>= plot(res,which=2) @ \end{center} or to plot the observations with the fitted values: \begin{center} <<>>= plot(res,which=3) @ \end{center} \subsection{Estimating a system of equations: CAPM} We want to test one of the implications of the capital asset pricing model (CAPM). This example comes from \cite{campbell-lo-mackinlay96}. It shows how to apply the \pkg{gmm} package to estimate a system of equations. The theory of CAPM implies that $\mu_i-R_f = \beta_i (\mu_m-R_f)$ $\forall i$, where $\mu_i$ is the expected value of stock i's return, $R_f$ is the risk free rate and $\mu_m$ is the expected value of the market porfolio's return. The theory can be tested by running the following regression: \[ (R_t-R_f) = \alpha + \beta (R_{mt}-R_f) + \epsilon_t, \] where $R_t$ is a $N\times 1$ vector of observed returns on stocks, $R_{mt}$ if the observed return of a proxy for the market portfolio, $R_f$ is the interest rate on short term government bonds and $\epsilon_t$ is a vector of error terms with covariance matrix $\Sigma_t$. When estimated by ML or LS, $\Sigma$ is assumed to be fixed. However, GMM allows $\epsilon_t$ to be heteroskedastic and serially correlated. One implication of the CAPM is that the vector $\alpha$ should be zero. It can be tested by estimating the model with $(R_{mt}-R_f)$ as instruments, and by testing the null hypothesis $H_0:~\alpha=0$. The data, which are included in the package, are the daily returns of twenty selected stocks from January 1993 to February 2009, the risk-free rate and the three factors of Fama and French\footnote{The symbols of the stocks taken from \url{http://ca.finance.yahoo.com/} are ("WMK", "UIS", "ORB", "MAT", "ABAX","T", "EMR", "JCS", "VOXX", "ZOOM", "ROG", "GGG", "PC", "GCO", "EBF", "F", "FNM", "NHP", "AA", "TDW"). The four other series can be found on K. R. French's web site: \url{http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data\_library.html}}. The following test is performed using the returns of 5 stocks and a sample size of 500\footnote{The choice of sample size is arbitrary. The purpose is to show how to estimate a system of equations not to test the CAPM. Besides, the $\beta$'s seem to vary over time. It is therefore a good practice to estimate the model using short periods.}. <<>>= data(Finance) r <- Finance[1:500,1:5] rm <- Finance[1:500,"rm"] rf <- Finance[1:500,"rf"] z <- as.matrix(r-rf) zm <- as.matrix(rm-rf) res <- gmm(z~zm,x=zm) coef(res) R <- cbind(diag(5),matrix(0,5,5)) c <- rep(0,5) linearHypothesis(res,R,c,test = "Chisq") @ where the asymptotic chi-square is used since the default distribution requires a normality assumption. The same test could have been performed using the names of the coefficients: <>= test <- paste(names(coef(res)[1:5])," = 0",sep="") linearHypothesis(res,test) @ Another way to test the CAPM is to estimate the restricted model ($\alpha=0$), which is over-identified, and to perform a J-test. Adding $-1$ to the formula removes the intercept. In that case, a column of ones has to be added to the matrix of instruments: <<>>= res2<-gmm(z~zm-1,cbind(1,zm)) specTest(res2) @ which confirms the non-rejection of the theory. \subsection{Testing the CAPM using the stochastic discount factor representation} In some cases the theory is directly based on moment conditions. When it is the case, testing the validity of these conditions becomes a way of testing the theory. \cite{jagannathan-skoulakis02} present several GMM applications in finance and one of them is the stochastic discount factor (SDF) representation of the CAPM. The general theory implies that $E(m_tR_{it})=1$ for all $i$, where $m_t$ is the SDF and $R_{it}$ the gross return ($1 + r_{it}$). It can be shown that if the CAPM holds, $m_t=\theta_0 + \theta_0 R_{mt}$ which implies the following moment conditions: \[ E\Big[R_{it}(\theta_0-\theta_1R_{mt})-1 \Big] = 0~~\mbox{for}~~ i=1,...,N \] which can be tested as follows: <<>>= g5 <- function(tet, x) { gmat <- (tet[1] + tet[2] * (1 + c(x[, 1]))) * (1 + x[, 2:6]) - 1 return(gmat) } res_sdf <- gmm(g5, x = as.matrix(cbind(rm, r)), c(0, 0)) specTest(res_sdf) @ which is consistent with the two previous tests. \subsection{Estimating continuous time processes by discrete time approximation} This last example also comes from \cite{jagannathan-skoulakis02}. We want to estimate the coefficients of the following continuous time process which is often used in finance for interest rates: \[ dr_t = (\alpha + \beta r_t)dt + \sigma r_t^\gamma dW_t, \] where $W_t$ is a standard Brownian motion. Special cases of this process are the Brownian motion with drift ($\beta=0$ and $\gamma = 0$), the Ornstein-Uhlenbeck process ($\gamma=0$) and the Cox-Ingersoll-Ross or square root process ( $\gamma = 1/2$). It can be estimated using the following discrete time approximation: \[ r_{t+1}-r_t = \alpha + \beta r_t + \epsilon_{t+1} \] with \[ E_t \epsilon_{t+1}=0,~~\mbox{and}~~E_t(\epsilon_{t+1}^2) = \sigma^2r_t^{2\gamma} \] Notice that ML cannot be used to estimate this model because the distribution depends on $\gamma$. In particular, it is normal for $\gamma=0$ and gamma for $\gamma=1/2$. It can be estimated by GMM using the following moment conditions: \[ E[g(\theta,x_t)] \equiv E\left(\begin{array}{c} \epsilon_{t+1}\\ \epsilon_{t+1}r_t\\ \epsilon_{t+1}^2-\sigma^2r_t^{2\gamma}\\ (\epsilon_{t+1}^2-\sigma^2r_t^{2\gamma})r_t \end{array} \right) = 0 \] The related \textit{g} function, with $\theta=\{\alpha, \beta, \sigma^2, \gamma\}$ is: <<>>= g6 <- function(theta, x) { t <- length(x) et1 <- diff(x) - theta[1] - theta[2] * x[-t] ht <- et1^2 - theta[3] * x[-t]^(2 * theta[4]) g <- cbind(et1, et1 * x[-t], ht, ht * x[-t]) return(g) } @ In order to estimate the model, the vector of interest rates needs to be properly scaled to avoid numerical problems. The transformed series is the annualized interest rates expressed in percentage. Also, the starting values are obtained using LS and some options for \textit{optim} need to be modified. <<>>= rf <- Finance[,"rf"] rf <- ((1 + rf/100)^(365) - 1) * 100 dr <- diff(rf) res_0 <- lm(dr ~ rf[-length(rf)]) tet0 <- c(res_0$coef, var(residuals(res_0)), 0) names(tet0) <- c("alpha", "beta", "sigma^2", "gamma") res_rf <- gmm(g6, rf, tet0, control = list(maxit = 1000, reltol = 1e-10)) coef(res_rf) @ \subsection{Comments on models with panel data} The \pkg{gmm} package is not directly built to easily deal with panel data. However, it is flexible enough to make it possible in most cases. To see that, let us consider the following model (see \cite{wooldridge02} for more details): \[ y_{it} = x_{it} \beta + a_i + \epsilon_{it}\mbox{ for } i=1,...,N\mbox{ and }t=1,...,T, \] where $x_{it}$ is $1\times k$, $\beta$ is $k\times 1$, $\epsilon_{it}$ is an error term and $a_i$ is an unobserved component which is specific to individual $i$. If $a_i$ is correlated with $x_{it}$, it can be removed by subtracting the average of the equation over time, which gives: \[ (y_{it}-\bar{y}_i) = (x_{it}-\bar{x}_i) \beta + (\epsilon_{it}-\bar{\epsilon}_i) \mbox{ for } i=1,...,N\mbox{ and } t=1,...,T, \] which can be estimated by \textit{gmm}. For example, if there are 3 individuals the following corresponds to the GMM fixed effects estimation: <>= y <- rbind(y1-mean(y1),y2-mean(y2),y3-mean(y3)) x <- rbind(x1-mean(x1),x2-mean(x2),x3-mean(x3)) res <- gmm(y~x,h) @ However, if $a_i$ is not correlated with $x_{it}$, the equation represents a random effects model. In that case, it is more efficient not to remove $a_i$ from the equation because of the information it carries about the individuals. The error terms are then combined in a single one, $\eta_{it}=(a_i + \epsilon_{it})$ to produce the linear model: \[ y_{it} = x_{it} \beta + \eta_{it} \] This model cannot be efficiently estimated by OLS because the presence of the common factor $a_i$ at each period implies that $\eta_{it}$ is serially correlated. However, GMM is well suited to deal with such specifications. The following will therefore produce a GMM random effects estimation: <>= y <- rbind(y1,y2,y3) x <- rbind(x1,x2,x3) res <- gmm(y~x,h) @ The package \pkg{plm} of \cite{plm} offers several functions to manipulate panel data. It could therefore be combined with \pkg{gmm} when estimating such models. It also offers a way to estimate them with its own GMM algorithm for panel data. \subsection{GMM and the sandwich package} In the \pkg{gmm} package, the estimation of the optimal weighting matrices are obtained using the \pkg{sandwich} package of \cite{zeileis06}. For example, the weighting matrix of the two-step GMM defined as: \[ W = \left[\lim_{n\rightarrow \infty} Var{(\sqrt{n}\bar{g})}\right]^{-1} \] is estimated as follows: <>= gt <- g(t0, x) V <- kernHAC(lm(gt~1),sandwich = FALSE) W <- solve(V) @ where $t0$ is any consistent estimate. As long as the optimal matrix is used, the covariance matrix of the coefficients can be estimated as follows: \[ (\hat{G}'W\hat{G})^{-1}/n \equiv (\hat{G}'\hat{V}^{-1}\hat{G})^{-1}/n, \] where $\hat{G}=d\bar{g}(\hat{\theta})/d\theta$ and $\hat{V}$ is obtained using \textit{kernHAC()}. It is not a sandwich covariance matrix and is computed using the \textit{vcov()} method included in \pkg{gmm}. However, if any other weighting matrix is used, say $W$, the estimated covariance matrix of the coefficients must then be estimated as follows: \[ (\hat{G}'W\hat{G})^{-1}\hat{G}'W\hat{V}W\hat{G}(\hat{G}'W\hat{G})^{-1}/n. \] A \textit{bread()} and \textit{estfun()} methods are available for \textit{gmm} objects which allows to compute the above matrix using the \pkg{sandwich} package. The \textit{bread()} method computes $(\hat{G}'W\hat{G})^{-1}$ while the \textit{estfun()} method returns a $T\times q$ matrix with the $t^{th}$ row equals to $g(\hat{\theta}, x_t) W \hat{G}$. The \textit{meatHAC()} method applied to the latter produces the right meat. Let us consider the example of section (\ref{ar}). Suppose we want to use the identity matrix to eliminate one source of bias, at the cost of lower efficiency. In that case, a consistent estimate of the covariance matrix is \[ (\hat{G}'\hat{G})^{-1}\hat{G}'\hat{V}\hat{G}(\hat{G}'\hat{G})^{-1}/n, \] which can be computed as: <<>>= print(res<-gmm(g4,x5t[,4:7],wmatrix="ident")) diag(vcovHAC(res))^.5 @ which is more robust than using \textit{vcov()}: <<>>= diag(vcov(res))^.5 @ Notice that it is possible to fixe $W$. Therefore, the above results can also be obtained as: <<>>= print(res<-gmm(g4,x5t[,4:7], weightsMatrix = diag(5))) @ In this case, the choice of the type of GMM is irrelevant since the weighting matrix is fixed. \section{Generalized empirical likelihood} The GEL is a new family of estimation methods which, as GMM, is based on moment conditions. It follows \cite{owen01} who developed the idea of empirical likelihood estimation which was meant to improve the confidence regions of estimators. We present here a brief discussion on the method without going into too much details. For a complete review, see \cite{smith97}, \cite{newey-smith04} or \cite{anatolyev05}. The estimation is based on \[ E(g(\theta_0,x_i))=0, \] which can be estimated in general by \[ \tilde{g}(\theta) = \sum_{i=1}^n p_i g(\theta,x_i) =0, \] where $p_i$ is called the implied probability associated with the observation $x_i$. For the GEL method, it is assumed that $q>p$ because otherwise it would correspond to GMM. Therefore, as it is the case for GMM, there is no solution to $\bar{g}(\theta)=0$. However, there is a solution to $\tilde{g}(\theta)=0$ for some choice of the probabilities $p_i$ such that $\sum_i p_i=1$. In fact, there is an infinite number of solutions since there are $(n+q)$ unknowns and only $q+1$ equations. GEL selects among them the one for which the distance between the vector of probabilities $p$ and the empirical density $1/n$ is minimized. The empirical likelihood of \cite{owen01} is a special case in which the distance is the likelihood ratio. The other methods that belong to the GEL family of estimators use different metrics. If the moment conditions hold, the implied probabilities carry a lot of information about the stochastic properties of $x_i$. For GEL, the estimations of the expected value of the Jacobian and the covariance matrix of the moment conditions, which are required to estimate $\theta$, are based on $p_i$ while in GMM they are estimated using $1/n$. \cite{newey-smith04} show that this difference explains partially why the second order properties of GEL are better. Another difference between GEL and GMM is how they deal with the fact that $g(\theta,x_i)$ can be a conditionally heteroskedastic and weakly dependent process. GEL does not require to compute explicitly the HAC matrix of the moment conditions. However, if it does not take it into account, its estimators may not only be inefficient but may also fail to be consistent. \cite{smith01} proposes to replace $g(\theta,x_i)$ by: \[ g^w(\theta,x_i) = \sum_{s=-m}^m w(s)g(\theta,x_{i-s}) \] where $w(s)$ are kernel based weights that sum to one (see also \cite{kitamura-stutzer97} and \cite{smith97}). The sample moment conditions become: \begin{equation}\label{gel_mcond} \tilde{g}(\theta) = \sum_{i=1}^n p_i g^w(\theta,x_i) =0 \end{equation} The estimator is defined as the solution to the following constrained minimization problem: \begin{eqnarray} \hat{\theta}_n &=& \arg\min_{\theta,p_i} \sum_{i=1}^n h_n(p_i) ,\\ && \mbox{subject to } \\ && \sum_{i=1}^n p_ig^w(\theta,x_i) = 0 ~~\mbox{and} \label{const}\\ && \sum_{i=1}^n p_i=1, \end{eqnarray} where $h_n(p_i)$ has to belong to the following Cressie-Read family of discrepancies: \[ h_n(p_i) = \frac{[\gamma(\gamma+1)]^{-1}[(np_i)^{\gamma+1}-1]}{n} . \] \cite{smith97} showed that the empirical likelihood method (EL) of \cite{owen01} ($\gamma=0$) and the exponential tilting of \cite{kitamura-stutzer97} ($\gamma=-1$) belong to the GEL family of estimators while \cite{newey-smith04} show that it is also the case for the continuous updated estimator of \cite{hansen-heaton-yaron96} ($\gamma=1$). What makes them part of the same GEL family of estimation methods is the existence of a dual problem which is defined as: \begin{equation}\label{gel_obj} \hat{\theta} = \arg\min_{\theta}\left[\max_{\lambda} P_n(\theta,\lambda) = \frac{1}{n}\sum_{i=1}^n\rho\left(\lambda'g^w(\theta,x_i)\right)\right] \end{equation} where $\lambda$ is the Lagrange multiplier associated with the constraint (\ref{const}) and $\rho(v)$ is a strictly concave function normalized so that $\rho'(0)=\rho''(0)=-1$. It can be shown that $\rho(v)=\ln{(1-v)}$ corresponds to EL , $\rho(v)=-\exp{(v)}$ to ET and to CUE if it is quadratic. The equivalence of the primal and dual problems can easily be verified by showing that they both share the same following first order conditions: \begin{equation}\label{lam} \sum_{i=1}^n p_i g^w(\theta,x_i) = 0, \end{equation} \begin{equation} \sum_{i=1}^n p_i \lambda'\left(\frac{\partial g^w(\theta,x_i)}{\partial \theta}\right) = 0 , \end{equation} with \begin{equation} p_i = \frac{1}{n}\rho'\left(\lambda'g^w(\theta,x_i)\right) . \end{equation} Equation (\ref{gel_obj}) represents a saddle point problem. The solution is obtained by solving simultaneously two optimization problems. We can solve for $\theta$ by minimizing $P_n(\theta,\lambda(\theta))$, where $\lambda(\theta)$ is the solution to $\arg\max_\lambda P_n(\theta,\lambda)$ for a given $\theta$. Therefore an optimization algorithm needs to be called inside the $P_n(\theta,\lambda)$ function. It makes the GEL very hard to implement numerically. For example, \cite{guggenberger08}, who analyzes the small sample properties of GEL, uses an iterative procedure based on the Newton method for $\lambda$ and a grid search for $\theta$ in order to confidently reach the absolute minimum. Using such iterative procedures for $\lambda$ makes the problem less computationally demanding and does not seem to affect the properties of the estimator of $\theta_0$. Indeed, \cite{guggenberger-hahn05} show that going beyond two iterations for $\lambda$ does not improve the second order asymptotic properties of the estimator of $\theta_0$. The function \textit{gel} offers two options. By default, $\lambda(\theta)$ is obtained by the following iterative method: \[ \lambda_l = \lambda_{l-1} - \left[\frac{1}{n}\sum_{i=1}^n \rho''(\lambda_{l-1}'g_t)g_tg_t'\right]^{-1}\left[\frac{1}{n} \sum_{i=1}^n \rho'(\lambda_{l-1}'g_i) g_i \right] \] starting with $\lambda=0$, which corresponds to its asymptotic value. The algorithm stops when $\|\lambda_l-\lambda_{l-1}\|$ reaches a certain tolerance level. The second option is to let \textit{optim} solve the problem. Then, as for \textit{gmm}, the minimization problem is solved either by \textit{optim}, \textit{nlminb} or \textit{constrOptim}. In order to test the over-identifying restrictions, \cite{smith04} proposes three tests which are all asymptotically distributed as a $\chi^2_{q-p}$. The first one is the J-test: \[ n \bar{g}^w(\hat{\theta})'[\hat{\Omega}(\hat{\theta})]^{-1}\bar{g}^w(\hat{\theta}), \] the second is a Lagrange multiplier test (LM): \[ LM = n\hat{\lambda}'\hat{\Omega}(\hat{\theta})\hat{\lambda} \] and the last one is a likelihood ratio test (LR): \[ LR = 2\sum_{i=1}^n\left[ \rho\left(\hat{\lambda}'g^w(\hat{\theta},x_i)\right) - \rho(0)\right] \] \section{GEL with R} \subsection{Estimating the parameters of a normal distribution} For this example, we can leave the option \textit{smooth} at its default value, which is \textit{FALSE}, because of the iid properties of $x$. A good starting value is very important for GEL. The best choice is the sample mean and the standard deviation. By default the option \textit{type} is set to \textit{EL}. The same methods that apply to \textit{gmm} objects, can also be applied to \textit{gel} objects. <<>>= tet0 <- c(mu = mean(x1), sig = sd(x1)) res_el <- gel(g1,x1,tet0) summary(res_el) @ Each Lagrange multiplier represents a shadow price of the constraint implied by moment condition. A binding constraint will produce a multiplier different from zero. Therefore, its value informs us on the validity of the moment condition. In the above results, the $\lambda$'s are significantly different from zero which would normally suggest that the moment conditions associated with them are violated. As a result, the LM test also rejects the null hypothesis since it is based on the $\lambda$'s. Notice that \textit{summary} reports two convergence codes, one for $\lambda$ et another for $\theta$. The ET and CUE estimates can be obtained as follows: <<>>= res_et <- gel(g1,x1,tet0,type="ET") coef(res_et) @ <<>>= res_cue <- gel(g1,x1,tet0,type="CUE") coef(res_cue) @ A fourth method is available which is called the exponentially tilted empirical likelihood (ETEL) and was proposed by \cite{schennach07}. However, it does not belong to the family of GEL estimators. It solves the problem of misspecified models. In such models there may not exist any pseudo value to which $\hat{\theta}$ converges as the sample size increases. ETEL uses the $\rho()$ of ET to solve for $\lambda$ and the $\rho()$ of EL to solve for $\theta$. \cite{schennach07} shows that ETEL shares the same asymptotic properties of EL without having to impose restrictions on the domain of $\rho(v)$ when solving for $\lambda$. <<>>= res_etel <- gel(g1,x1,c(mu=1,sig=1),type="ETEL") coef(res_etel) @ The type ETEL is experimental for now. Although it is supposed to be more stable because no restrictions are required to solve for $\lambda$, once we substitute $\lambda(\theta)$ in the EL objective function to estimate $\theta$, we still need to restrict $\lambda'g_t$ to avoid having NA's. The solution used in gel() is to obtain $\lambda(\theta)$ with \textit{constrOptim} with the restriction $\lambda'gt > 1$ even if it is not required by ET ($\rho(v)=-\exp{(v)}$). It is however sensitive to starting values. That's the reason why we used different ones above. \subsection{Estimating the AR coefficients of an ARMA process} Because the moment conditions are weakly dependent, we need to set the option \textit{smooth=TRUE}. Before going to the estimation procedure, we need to understand the relationship between the smoothing kernel and the HAC estimator. The reason why we need to smooth the moment function is that GEL estimates the covariance matrix of $\bar{g}(\theta,x_t)$, as if we had iid observations, using the expression $(1/T)\sum_{t=1}^T(g_tg_t')$. We can show that substituting $g_t$ by $g^w_t$ in this expression results in an HAC estimator. However, the relationship between the smoothing kernel and the kernel that appears in the HAC estimator is not obvious. For example, we can show that if the smoothing kernel is Truncated, then the kernel in the HAC estimator is the Bartlett. Let us consider the truncated kernel with a bandwidth of 2. This implies that $w(s)=1/5$ for $|s|\leq 2$ and 0 otherwise. Then, the expression for the covariance matrix becomes: \begin{eqnarray*} \frac{1}{T}\sum_{t=1}^T g^w_t (g^w_t)' &=& \frac{1}{T}\sum_{t=1}^T \left( \sum_{s=-2}^2 \frac{1}{5} g_{t+s} \right) \left( \sum_{l=-2}^2 \frac{1}{5} g_{t+l}' \right),\\ &=& \frac{1}{25}\sum_{s=-2}^2\sum_{l=-2}^2\left( \frac{1}{T}\sum_{t=1}^T g_{t+s}g_{t+l}' \right),\\ &=& \frac{1}{25}\sum_{s=-2}^2\sum_{l=-2}^2\hat{\Gamma}_{s-l} ,\\ &=& \frac{1}{25}\sum_{s=-4}^4 (5-|s|)\hat{\Gamma}_{s},\\ &=& \sum_{s=-4}^4 \left(\frac{1}{5}-\frac{|s|}{25}\right)\hat{\Gamma}_{s},\\ &=& \sum_{s=-T+1}^{T-1} k_5(s) \hat{\Gamma}_{s}, \end{eqnarray*} where $k_5(s)$ is the Bartlett kernel with a bandwidth of 5 defined as \[ K_5(s) = \left\{ \begin{array}{ccc} 1/5 - |s|/25 & \mbox{if}&|s|\leq 5\\ 0 &\mbox{otherwise}& \end{array}\right. . \] See \cite{smith01} for more details. The model will therefore be estimated using the kernel Truncated. The GMM estimate with the identity matrix is selected as starting value. <>= tet0 <- gmm(g4,x=x5t[,4:7],wmatrix="ident")$coef res <- gel(g4,x=x5t[,4:7],tet0,smooth=TRUE,kernel="Truncated") summary(res) @ The \textit{specTest} method applied to a \textit{gel} object computes the three tests proposed by \cite{smith04}: <<>>= specTest(res) @ The \textit{plot} method produces one more graphics when applied to a \textit{gel} object. It shows the implied probabilities along with the empirical density $(1/T)$. It allows to see which observations have more influence: \begin{center} <<>>= plot(res,which=4) @ \end{center} We can also select \textit{optfct="nlminb"} or \textit{constraint=TRUE} in order to impose restrictions on the coefficients. The former sets lower and upper bounds for the coefficients, while the latter imposes linear constraints using the algorithm \textit{constrOptim}. In this example we want the sum of the AR coefficients to be less than one. \textit{constrOptim} imposes the constraint $ui\theta-ci \geq 0$. Therefore, we need to set: <>= ui=cbind(0,-1,-1) ci <- -1 @ and rerun the estimation as <>= res <- gel(g4,x=dat5[,4:7],tet0,smooth=TRUE,kernel="Truncated", constraint=TRUE, ui=ui,ci=ci) @ The result, which is not shown, is identical. They are also many option to compute the $\lambda$'s. From version 1.4-0, the default algorithm is \textit{nlminb} because the gradient and Hessian matrix are well defined analytically which speed up convergence. The other choices are \textit{optim} or "iter" which uses a Newton method to solve the first order condition. If the option optlam is set to "optim" and the type is EL, \textit{contrOptim} is selected automatically to restrict $\lambda'g_t$ to be less than 1. It is also possible to change the default values in the control list of the optimizer with the option LambdaControl (see \textit{?nlminb} or \textit{?optim}). Here are some examples: <>= res <- gel(g4,x=dat5[,4:7],tet0,smooth=TRUE, optlam="optim") res <- gel(g4,x=dat5[,4:7],tet0,smooth=TRUE, optlam="optim", LambdaControl=list(trace=TRUE, parscale=rep(.1,5))) @ \subsection{Comments} The GEL method is very unstable numerically. This fact has been reported many times in the recent literature. The method has been included in the \pkg{gmm} package because recent theoretical evidence suggests that it may produce better estimators than GMM. Because \proglang{R} is an open source statistical package, it offers a good platform to experiment with numerical properties of estimators. \section{Conclusion} The \pkg{gmm} package offers complete and flexible algorithms to estimate models by GMM and GEL. Several options are available which allow to choose among several GMM and GEL methods and many different HAC matrix estimators. In order to estimate the vector of parameters, users can select their preferred optimization algorithm depending on whether inequality constraints are required. For the vector of Lagrange multiplier of GEL, it can be computed by an iterative procedure based on the Newton method which increases the speed of convergence and reduce the instability of the estimation procedure. It could then easily be used by those who are interested in studying the numerical properties of both methods. The package also offers an interface which is comparable to the least squares method \textit{lm}. Linear model are estimated using formula and methods such as \textit{summary}, \textit{vcov}, \textit{coef}, \textit{confint}, \textit{plot}. \textit{residuals} or \textit{fitted} are available for the objects of class \textit{gmm} and \textit{gel}. \proglang{R} users will therefore have little difficulty in using the package. \section{Computational Details} The package \pkg{gmm} is written entirely in \proglang{R} and S3-classes with methods are used. It can be found on the comprehensive \proglang{R} archive network (CRAN,\url{http://CRAN.R-project.org/}). It is also hosted on R-Forge (\url{http://r-forge.r-project.org/projects/gmm}). It is shipped with a NAMESPACE. The version used to produce this paper is 1.4-0. It depends on the \pkg{sandwich} package of \cite{zeileis06}, which is used to compute de HAC matrices. The packages \pkg{car} (\cite{car}), \pkg{mvtnorm} (\cite{mvtnorm}), \pkg{stabledist} (\cite{stabledist}), \pkg{MASS} (\cite{MASS}), \pkg{timeDate} (\cite{timeDate}) and \pkg{timeSeries} (\cite{timeSeries}) are suggested in order to reproduce the examples. \section*{Acknowledgments} I am grateful to the three anonymous referees of the Journal of Statistical Software for great comments on the paper and the package. I also want to thank Achim Zeileis for his suggestions regarding the way the \pkg{sandwich} package can be used within \textit{gmm}. \bibliography{empir} \end{document} gmm/build/0000755000176200001440000000000015044423073012133 5ustar liggesusersgmm/build/vignette.rds0000644000176200001440000000057115044423073014475 0ustar liggesusers‹}RKOÃ0 îkì!@HãR®\öc`Z9p›BëmyTiJ'~9ÃÝ’ª-b•’Úù>;ö缎<Ï ¼¨ç{Aˆf8Æí×.ß‹¼!þÏ×B,Kf6ËÅD˲ƒÆwJd…arM@‚¦œ}AJîEÆ4K('Oì8Û(•*ÓifƒÇjEæJ€49©®!‹#d骃ž6PŒ Æ{ìÀé[NˆköŇ‘J;w”úÆ!e¹qÜ„jkFóÛ8¶öÀ0SjÀV~ šAŽÄ QPORQ¶ªìO!™VÇ?"O,…¬Ò@­¯õ³¼ë&OÔŠƒ„–ty$ÁU¢$¬PENŠ,Åžð¾û¡F¹ÖÇð™)‰ f0ŒW¶ÐMâÛ=ø_Á{aü]G‹Á#lK¥-Ž#š1î¤î½0S;áótfMÑ÷OþÖ`†Z•7œ³êåã¶Ã¯;Á„Ó¼;Á E'+ñUÝ¿HQíõ;gmm/build/partial.rdb0000644000176200001440000000007515044423053014260 0ustar liggesusers‹‹àb```b`aed`b1…À€… H02°0piÖ¼ÄÜÔb C"Éð ´¤7gmm/man/0000755000176200001440000000000015044422052011603 5ustar liggesusersgmm/man/ATEgel.Rd0000644000176200001440000001517315044415476013216 0ustar liggesusers\name{ATEgel} \alias{ATEgel} \alias{checkConv} \title{ATE with Generalized Empirical Likelihood estimation} \description{ Function to estimate the average treatment effect with the sample being balanced by GEL. } \usage{ ATEgel(g, balm, w=NULL, y=NULL, treat=NULL, tet0=NULL,momType=c("bal","balSample","ATT"), popMom = NULL, family=c("linear","logit", "probit"), type = c("EL", "ET", "CUE", "ETEL", "HD", "ETHD", "RCUE"), tol_lam = 1e-9, tol_obj = 1e-9, tol_mom = 1e-9, maxiterlam = 100, optfct = c("optim", "nlminb"), optlam = c("nlminb", "optim", "iter", "Wu"), data=NULL, Lambdacontrol = list(), model = TRUE, X = FALSE, Y = FALSE, ...) checkConv(obj, tolConv=1e-4, verbose=TRUE, ...) } \arguments{ \item{g}{A formula as \code{y~z}, where \code{y} is the response and \code{z} the treatment indicator. If there is more than one treatment, more indicators can be added or \code{z} can be set as a factor. It can also be of the form \code{g(theta, y, z)} for non-linear models. It is however, not implemented yet.} \item{obj}{Object of class \code{"ategel"} produced y \code{ATEgel}} \item{balm}{A formula for the moments to be balanced between the treated and control groups (see details)} \item{y}{The response variable when \code{g} is a function. Not implemented yet} \item{treat}{The treatment indicator when \code{g} is a function. Not implemented yet} \item{w}{A formula to add covariates to the main regression. When \code{NULL}, the default value, the main regression only include treatment indicators.} \item{tet0}{A \eqn{3 \times 1} vector of starting values. If not provided, they are obtained using an OLS regression} \item{momType}{How the moments of the covariates should be balanced. By default, it is simply balanced without restriction. Alternatively, moments can be set equal to the sample moments of the whole sample, or to the sample moments of the treated group. The later will produce the average treatment effect of the treated (ATT)} \item{popMom}{A vector of population moments to use for balancing. It can be used of those moments are available from a census, for example. When available, it greatly improves efficiency.} \item{family}{By default, the outcome is linearly related to the treatment indicators. If the outcome is binary, it is possible to use the estimating equations of either the logit or probit model.} \item{type}{"EL" for empirical likelihood, "ET" for exponential tilting, "CUE" for continuous updated estimator, "ETEL" for exponentially tilted empirical likelihood of Schennach(2007), "HD" for Hellinger Distance of Kitamura-Otsu-Evdokimov (2013), and "ETHD" for the exponentially tilted Hellinger distance of Antoine-Dovonon (2015). "RCUE" is a restricted version of "CUE" in which the probabilities are bounded below by zero. In that case, an analytical Kuhn-Tucker method is used to find the solution.} \item{tol_lam}{Tolerance for \eqn{\lambda} between two iterations. The algorithm stops when \eqn{\|\lambda_i -\lambda_{i-1}\|} reaches \code{tol_lamb} (see \code{\link{getLamb}}) } \item{maxiterlam}{The algorithm to compute \eqn{\lambda} stops if there is no convergence after "maxiterlam" iterations (see \code{\link{getLamb}}).} \item{tol_obj}{Tolerance for the gradiant of the objective function to compute \eqn{\lambda} (see \code{\link{getLamb}}).} \item{optfct}{Algorithm used for the parameter estimates} \item{tol_mom}{It is the tolerance for the moment condition \eqn{\sum_{t=1}^n p_t g(\theta(x_t)=0}, where \eqn{p_t=\frac{1}{n}D\rho()} is the implied probability. It adds a penalty if the solution diverges from its goal.} \item{optlam}{Algorithm used to solve for the lagrange multiplier in \code{\link{getLamb}}. The algorithm Wu is only for \code{type="EL"}. The value of \code{optlam} is ignored for "CUE" because in that case, the analytical solution exists.} \item{data}{A data.frame or a matrix with column names (Optional). } \item{Lambdacontrol}{Controls for the optimization of the vector of Lagrange multipliers used by either \code{\link[stats]{optim}}, \code{\link[stats]{nlminb}} or \code{\link[stats]{constrOptim}}} \item{model, X, Y}{logicals. If \code{TRUE} the corresponding components of the fit (the model frame, the model matrix, the response) are returned if g is a formula.} \item{verbose}{If TRUE, a summary of the convergence is printed} \item{tolConv}{The tolerance for comparing moments between groups} \item{...}{More options to give to \code{\link[stats]{optim}} or \code{\link[stats]{nlminb}}. In \code{checkConv}, they are options passed to \code{\link{getImpProb}}.} } \details{ We want to estimate the model \eqn{Y_t = \theta_1 + \theta_2 treat + \epsilon_t}, where \eqn{\theta_2} is the treatment effect. GEL is used to balance the sample based on the argument \code{x} above. For example, if we want the sample mean of \code{x1} and \code{x2} to be balanced between the treated and control, we set \code{x} to \code{~x1+x2}. If we want the sample mean of \code{x1}, \code{x2}, \code{x1*x2}, \code{x1^2} and \code{x2^2}, we set \code{x} to \code{~x1*x2 + I(x1^2) + I(x2^2)}. } \value{ 'gel' returns an object of 'class' '"ategel"' The functions 'summary' is used to obtain and print a summary of the results. The object of class "ategel" is a list containing the same elements contained in objects of class \code{\link{gel}}. } \references{ Lee, Seojeong (2016), Asymptotic refinements of misspecified-robust bootstrap for GEL estimators, \emph{Journal of Econometrics}, \bold{192}, 86--104. Schennach, Susanne, M. (2007), Point Estimation with Exponentially Tilted Empirical Likelihood. \emph{Econometrica}, \bold{35}, 634-672. Wu, C. (2005), Algorithms and R codes for the pseudo empirical likelihood method in survey sampling. \emph{Survey Methodology}, \bold{31}(2), page 239. Chausse, P. (2010), Computing Generalized Method of Moments and Generalized Empirical Likelihood with R. \emph{Journal of Statistical Software}, \bold{34}(11), 1--35. URL \doi{10.18637/jss.v034.i11}. Chausse, P. and Giurcanu, M. and Luta, G. (2021) Estimating the Average Causal Effect using Generalized Empirical Likelihood Methods, Work in progress. } \examples{ data(nsw) # Scale income nsw$re78 <- nsw$re78/1000 nsw$re75 <- nsw$re75/1000 res <- ATEgel(re78~treat, ~age+ed+black+hisp+married+nodeg+re75, data=nsw,type="ET") summary(res) chk <- checkConv(res) res2 <- ATEgel(re78~treat, ~age+ed+black+hisp+married+nodeg+re75, data=nsw,type="ET", momType="balSample") summary(res2) chk2 <- checkConv(res2) } gmm/man/growth.Rd0000644000176200001440000000166315044422052013412 0ustar liggesusers\name{Growth} \docType{data} \alias{Growth} \title{Growth Data} \description{ Panel of Macroeconomic data for 125 countries from 1960 to 1985 constructed by Summers and Heston (1991)) } \usage{data(Growth)} \format{ A data frame containing 9 vectors. \describe{ \item{Country_ID}{Country identification number} \item{COM}{1 if the country is in a communist regime, 0 otherwise} \item{OPEC}{1 if the country is part of the OPEC, 0 otherwise} \item{Year}{Year} \item{GDP}{Per capita GDP (in thousands) in 1985 U.S. dollars.} \item{LagGDP}{GDP of the previous period} \item{SavRate}{Saving rate measured as the ratio of real investment to real GDP} \item{LagSavRate}{SavRate of the previous period} \item{Country}{Country names} \item{Pop}{Population in thousands} \item{LagPop}{Population of the previous period} } } \source{\url{https://sites.google.com/view/fumio-hayashis-hp/hayashi-econometrics}} \keyword{datasets} gmm/man/nsw.Rd0000644000176200001440000000303614247643115012714 0ustar liggesusers\name{nsw} \docType{data} \alias{nsw} \title{Lalonde subsample of the National Supported Work Demonstration Data (NSW)} \description{ This data was collected to evaluate the National Supported Work (NSW) Demonstration project in Lalonde (1986). } \usage{data(nsw)} \format{ A data frame containing 9 variables. \describe{ \item{treat}{Treatment assignment} \item{age}{Age} \item{ed}{Years of Education} \item{black}{1 if Black, 0 otherwise} \item{hisp}{1 if Hispanic 0 otherwise} \item{married}{1 if married 0 otherwise} \item{nodeg}{1 if no college degree 0 otherwise} \item{re75}{1975 earnings} \item{re78}{1978 earnings} } } \details{ The dataset was obtained from the ATE package (see reference). } \source{ "NSW Data Files" from Rajeev Dehejia's website. URL: \url{http://users.nber.org/~rdehejia/data/.nswdata2.html} "National Supported Work Evaluation Study, 1975-1979: Public Use Files." from the Interuniversity Consortium for Political and Social Research. URL: \url{http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/7865} } \references{ Lalonde, R. (1986). "Evaluating the Econometric Evaluations of Training Programs," American Economic Review, 76(4), 604-620. Dehejia R. and Wahba S. (1999). "Causal Effects in Non-Experimental Studies: Re-Evaluating the Evaluation of Training Programs," JASA 94 (448), 1053-1062. Asad Haris and Gary Chan (2015). ATE: Inference for Average Treatment Effects using Covariate Balancing. R package version 0.2.0. \url{https://CRAN.R-project.org/package=ATE} } \keyword{datasets} gmm/man/wage.Rd0000644000176200001440000000147715044422075013033 0ustar liggesusers\name{wage} \docType{data} \alias{wage} \title{Labor Data} \description{ Data used to measure return to education by Griliches (1976) } \usage{data(wage)} \format{ A data frame containing 20 cross-sectional vectors. \describe{ \item{AGE, AGE80}{Age in 1969 and 1980 respetively} \item{EXPR, EXPR80}{Working experience in 1969 and 1980 respetively} \item{IQ}{IQ measure of the individual} \item{KWW}{A test score} \item{LW, LW80}{Log wage in 1969 and 1980 respectively} \item{MED}{Mother education} \item{MRT, MRT80}{} \item{RNS, RNS80}{} \item{S, S80}{Schooling in 1969 and 1980 respetively} \item{SMSA, SMSA80}{} \item{TENURE, TENURE80}{Tenure in 1969 and 1980 respetively} \item{YEAR}{} } } \source{\url{https://sites.google.com/view/fumio-hayashis-hp/hayashi-econometrics}} \keyword{datasets} gmm/man/bwWilhelm.Rd0000644000176200001440000000565015044415701014035 0ustar liggesusers\name{bwWilhelm} \alias{bwWilhelm} \title{Wilhelm (2015) bandwidth selection} \description{ It computes the optimal bandwidth for the HAC estimation of the covariance matrix of the moment conditions. The bandwidth was shown by Wilhelm (2005) to be the one that minimizes the MSE of the GMM estimator. } \usage{ bwWilhelm(x, order.by = NULL, kernel = c("Quadratic Spectral", "Bartlett", "Parzen", "Tukey-Hanning"), approx = c("AR(1)", "ARMA(1,1)"), weights = NULL, prewhite = 1, ar.method = "ols", data = list()) } \arguments{ \item{x}{An object of class \code{gmm}.} \item{order.by}{Either a vector 'z' or a formula with a single explanatory variable like '~ z'. The observations in the model are ordered by the size of 'z'. If set to 'NULL' (the default) the observations are assumed to be ordered (e.g., a time series).} \item{kernel}{type of kernel used to compute the covariance matrix of the vector of sample moment conditions (see \code{\link[sandwich]{kernHAC}} for more details)} \item{approx}{A character specifying the approximation method if the bandwidth has to be chosen by \code{bwAndrews}.} \item{weights}{numeric. A vector of weights used for weighting the estimated coefficients of the approximation model (as specified by 'approx'). By default all weights are 1 except that for the intercept term (if there is more than one variable)} \item{prewhite}{logical or integer. Should the estimating functions be prewhitened? If \code{TRUE} or greater than 0 a VAR model of order \code{as.integer(prewhite)} is fitted via \code{ar} with method \code{"ols"} and \code{demean = FALSE}.} \item{ar.method}{character. The \code{method} argument passed to \code{\link[stats]{ar}} for prewhitening.} \item{data}{an optional data frame containing the variables in the 'order.by' model.} } \value{ The function 'bwWilhelm' returns the optimal bandwidth. } \references{ Wilhelm, D. (2015), Optimal Bandwidth Selection for Robust Generalized Method of Moments Estimation. \emph{Econometric Theory}, \bold{31}, 1054--1077 Zeileis A (2006), Object-oriented Computation of Sandwich Estimators. \emph{Journal of Statistical Software}, \bold{16}(9), 1--16. URL \doi{10.18637/jss.v016.i09}. } \note{ The function was written by Daniel Wilhelm and is based on \link[sandwich]{bwAndrews}. } \examples{ data(Finance) f1 <- Finance[1:300, "rm"] f2 <- Finance[1:300, "hml"] f3 <- Finance[1:300, "smb"] y <- Finance[1:300,"WMK"] ## Silly example just to make it over-identified ############################################### res <- gmm(y ~ f1, ~ f1 + f2 + f3) summary(res) ## Set the bandwidth using the second step estimate ################################################ bw <- bwWilhelm(res) res2 <- update(res, bw=bw) summary(res2) ## Set the bandwidth using the first-step estimate as for bwAndrews ################################################################### res3 <- gmm(y ~ f1, ~ f1 + f2 + f3, bw=bwWilhelm) summary(res3) } gmm/man/confint.Rd0000644000176200001440000000724414247643115013552 0ustar liggesusers\name{confint} \alias{confint.gel} \alias{confint.ategel} \alias{confint.gmm} \alias{print.confint} \title{Confidence intervals for GMM or GEL} \description{ It produces confidence intervals for the coefficients from \code{gel} or \code{gmm} estimation. } \usage{ \method{confint}{gel}(object, parm, level = 0.95, lambda = FALSE, type = c("Wald", "invLR", "invLM", "invJ"), fact = 3, corr = NULL, ...) \method{confint}{gmm}(object, parm, level = 0.95, ...) \method{confint}{ategel}(object, parm, level = 0.95, lambda = FALSE, type = c("Wald", "invLR", "invLM", "invJ"), fact = 3, corr = NULL, robToMiss=TRUE, ...) \method{print}{confint}(x, digits = 5, ...) } \arguments{ \item{object}{An object of class \code{gel} or \code{gmm} returned by the function \code{\link{gel}} or \code{\link{gmm}}} \item{parm}{A specification of which parameters are to be given confidence intervals, either a vector of numbers or a vector of names. If missing, all parameters are considered.} \item{level}{The confidence level} \item{lambda}{If set to TRUE, the confidence intervals for the Lagrange multipliers are produced.} \item{type}{'Wald' is the usual symetric confidence interval. The thee others are based on the inversion of the LR, LM, and J tests.} \item{fact}{This parameter control the span of search for the inversion of the test. By default we search within plus or minus 3 times the standard error of the coefficient estimate.} \item{corr}{This numeric scalar is meant to apply a correction to the critical value, such as a Bartlett correction. This value depends on the model (See Owen; 2001)} \item{x}{An object of class \code{confint} produced by \code{confint.gel} and \code{confint.gmm}} \item{digits}{The number of digits to be printed} \item{robToMiss}{If \code{TRUE}, the confidence interval is based on the standard errors that are robust to misspecification} \item{...}{Other arguments when \code{confint} is applied to another classe object} } \value{ It returns a matrix with the first column being the lower bound and the second the upper bound.} \references{ Hansen, L.P. (1982), Large Sample Properties of Generalized Method of Moments Estimators. \emph{Econometrica}, \bold{50}, 1029-1054, Hansen, L.P. and Heaton, J. and Yaron, A.(1996), Finit-Sample Properties of Some Alternative GMM Estimators. \emph{Journal of Business and Economic Statistics}, \bold{14} 262-280. Owen, A.B. (2001), Empirical Likelihood. \emph{Monographs on Statistics and Applied Probability 92, Chapman and Hall/CRC} } \examples{ ################# n = 500 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n = n, list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H t0 <- c(0,.5,.5) resGel <- gel(g, x, t0) confint(resGel) confint(resGel, level = 0.90) confint(resGel, lambda = TRUE) ######################## resGmm <- gmm(g, x) confint(resGmm) confint(resGmm, level = 0.90) ## Confidence interval with inversion of the LR, LM or J test. ############################################################## set.seed(112233) x <- rt(40, 3) y <- x+rt(40,3) # Simple interval on the mean res <- gel(x~1, ~1, method="Brent", lower=-4, upper=4) confint(res, type = "invLR") confint(res) # Using a Bartlett correction k <- mean((x-mean(x))^4)/sd(x)^4 s <- mean((x-mean(x))^3)/sd(x)^3 a <- k/2-s^2/3 corr <- 1+a/40 confint(res, type = "invLR", corr=corr) # Interval on the slope res <- gel(y~x, ~x) confint(res, "x", type="invLR") confint(res, "x") } gmm/man/getModel.Rd0000644000176200001440000000251514247643115013646 0ustar liggesusers\name{getModel} \alias{getModel.baseGmm} \alias{getModel.sysGmm} \alias{getModel.baseGel} \alias{getModel.constGmm} \alias{getModel.tsls} \alias{getModel.constGel} \alias{getModel.ateGel} \title{Method for setting the properties of a model} \description{ It collects what is needed by the method \code{momentEstim} (see details). } \usage{ \method{getModel}{baseGmm}(object, ...) \method{getModel}{sysGmm}(object, ...) \method{getModel}{baseGel}(object, ...) \method{getModel}{constGel}(object, ...) \method{getModel}{constGel}(object, ...) \method{getModel}{tsls}(object, ...) \method{getModel}{ateGel}(object, ...) } \arguments{ \item{object}{An object of class \code{baseGmm} } \item{...}{Other arguments when \code{getModel} is applied to another class object} } \value{ It returns an object of the right class which determines how the method \code{momentEstim} will treat it. For example, if \code{g} is a formula and \code{type} is set to "cue", it creates an object of class \code{baseGmm.cue.formula}. It this case, \code{momentEstim}, applied to this object, computes the continuously updated GMM of a linear model. It allows more flexibility this way. For example, it could be easy to add a GMM method which is robust in presence of weak identification simply by creating a new class of model and the associated \code{momentEstime} method. } gmm/man/vcov.Rd0000644000176200001440000000313515044420121013044 0ustar liggesusers\name{vcov} \alias{vcov.gmm} \alias{vcov.gel} \alias{vcov.tsls} \alias{vcov.ategel} \title{Variance-covariance matrix of GMM or GEL} \description{ It extracts the matrix of variances and covariances from \code{gmm} or \code{gel} objects. } \usage{ \method{vcov}{gmm}(object, ...) \method{vcov}{gel}(object, lambda = FALSE, ...) \method{vcov}{tsls}(object, type=c("Classical","HC0","HC1","HAC"), hacProp = list(), ...) \method{vcov}{ategel}(object, lambda = FALSE, robToMiss = TRUE, ...) } \arguments{ \item{object}{An object of class \code{gmm} or \code{gmm} returned by the function \code{\link{gmm}} or \code{\link{gel}}} \item{lambda}{If set to TRUE, the covariance matrix of the Lagrange multipliers is produced.} \item{type}{Type of covariance matrix for the meat} \item{hacProp}{A list of arguments to pass to \code{\link[sandwich]{kernHAC}}} \item{robToMiss}{If \code{TRUE}, it computes the robust to misspecification covariance matrix} \item{...}{Other arguments when \code{vcov} is applied to another class object} } \details{ For tsls(), if vcov is set to a different value thand "Classical", a sandwich covariance matrix is computed. } \value{ A matrix of variances and covariances } \examples{ # GMM # n = 500 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n = n,list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H res <- gmm(g, x) vcov(res) ## GEL ## t0 <- c(0,.5,.5) res <- gel(g, x, t0) vcov(res) vcov(res, lambda = TRUE) } gmm/man/print.Rd0000644000176200001440000000206314247643115013240 0ustar liggesusers\name{print} \alias{print.gmm} \alias{print.sysGmm} \alias{print.gel} \title{Printing a gmm or gel object} \description{ It is a printing method for \code{gmm} or \code{gel} objects. } \usage{ \method{print}{gmm}(x, digits = 5, ...) \method{print}{gel}(x, digits = 5, ...) \method{print}{sysGmm}(x, digits = 5, ...) } \arguments{ \item{x}{An object of class \code{gmm} or \code{gel} returned by the function \code{\link{gmm}} or \code{\link{gel}}} \item{digits}{The number of digits to be printed} \item{...}{Other arguments when print is applied to an other class object} } \value{ It prints some results from the estimation like the coefficients and the value of the objective function. } \examples{ # GMM # n = 500 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n = n, list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H res <- gmm(g, x) print(res) # GEL # t0 <- c(0,.5,.5) res <- gel(g,x,t0) print(res) } gmm/man/Finance.Rd0000644000176200001440000000275514247643115013457 0ustar liggesusers\name{Finance} \docType{data} \alias{Finance} \title{Returns on selected stocks} \description{ Daily returns on selected stocks, the Market portfolio and factors of Fama and French from 1993-01-05 to 2009-01-30 for CAPM and APT analysis } \usage{data(Finance)} \format{ A data frame containing 24 time series. Dates are reported as rownames(). In the following description, company symboles are used. \describe{ \item{WMK}{Returns of WEIS MARKETS INC} \item{UIS}{Returns of UNISYS CP NEW} \item{ORB}{Returns of ORBITAL SCIENCES CP} \item{MAT}{Returns of Mattel, Inc.} \item{ABAX}{Returns of ABAXIS, Inc.} \item{T}{Returns of AT&T INC.} \item{EMR}{Returns of EMERSON ELEC CO} \item{JCS}{Returns of Communications Systems Inc.} \item{VOXX}{Returns of Audiovox Corp.} \item{ZOOM}{Returns of ZOOM Technologies Inc.} \item{TDW}{Returns of TIDEWATER INC} \item{ROG}{Returns of Rogers Corporation} \item{GGG}{Returns of Graco Inc.} \item{PC}{Returns of Panasonic Corporation} \item{GCO}{Returns of Genesco Inc.} \item{EBF}{Returns of ENNIS, INC} \item{F}{Returns of FORD MOTOR CO} \item{FNM}{Returns of FANNIE MAE} \item{NHP}{Returns of NATIONWIDE HLTH PROP} \item{AA}{Returns of ALCOA INC} \item{rf}{Risk-free rate of Fama-French} \item{rm}{Return of the market portfolio of Fama-French} \item{hml}{Factor High-Minus-Low of Fama-French} \item{smb}{Factor Small-Minus-Big of Fama-French} } } \source{Yahoo Finance and \url{https://mba.tuck.dartmouth.edu/pages/faculty/ken.french/}} \keyword{datasets} gmm/man/residuals.Rd0000644000176200001440000000261714247643115014104 0ustar liggesusers\name{residuals} \alias{residuals.gel} \alias{residuals.gmm} \title{Residuals of GEL or GMM} \description{ Method to extract the residuals of the model estimated by \code{gmm} or \code{gel}. } \usage{ \method{residuals}{gel}(object, ...) \method{residuals}{gmm}(object, ...) } \arguments{ \item{object}{An object of class \code{gmm} or \code{gel} returned by the function \code{\link{gmm}} or \code{\link{gel}}} \item{...}{Other arguments when \code{residuals} is applied to an other classe object} } \value{ It returns the matrix of residuals \eqn{(y-\hat{y})} in \code{g=y~x} as it is done by \code{residuals.lm}. } \examples{ # GEL can deal with endogeneity problems n = 200 phi<-c(.2,.7) thet <- 0.2 sd <- .2 set.seed(123) x <- matrix(arima.sim(n = n, list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H res <- gel(g, x, c(0,.3,.6)) e <- residuals(res) plot(e, type = 'l', main = "Residuals from an ARMA fit using GEL") # GMM is like GLS for linear models without endogeneity problems set.seed(345) n = 200 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n = n, list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- 10 + 5*rnorm(n) + x res <- gmm(y ~ x, x) plot(x, residuals(res), main = "Residuals of an estimated model with GMM") } gmm/man/gel.Rd0000644000176200001440000003312515044417156012656 0ustar liggesusers\name{gel} \alias{gel} \alias{evalGel} \title{Generalized Empirical Likelihood estimation} \description{ Function to estimate a vector of parameters based on moment conditions using the GEL method as presented by Newey-Smith(2004) and Anatolyev(2005). } \usage{ gel(g, x, tet0 = NULL, gradv = NULL, smooth = FALSE, type = c("EL","ET","CUE","ETEL","HD","ETHD","RCUE"), kernel = c("Truncated", "Bartlett"), bw = bwAndrews, approx = c("AR(1)", "ARMA(1,1)"), prewhite = 1, ar.method = "ols", tol_weights = 1e-7, tol_lam = 1e-9, tol_obj = 1e-9, tol_mom = 1e-9, maxiterlam = 100, constraint = FALSE, optfct = c("optim", "optimize", "nlminb"), optlam = c("nlminb", "optim", "iter", "Wu"), data, Lambdacontrol = list(), model = TRUE, X = FALSE, Y = FALSE, TypeGel = "baseGel", alpha = NULL, eqConst = NULL, eqConstFullVcov = FALSE, onlyCoefficients=FALSE, ...) evalGel(g, x, tet0, gradv = NULL, smooth = FALSE, type = c("EL", "ET", "CUE", "ETEL", "HD", "ETHD","RCUE"), kernel = c("Truncated", "Bartlett"), bw = bwAndrews, approx = c("AR(1)", "ARMA(1,1)"), prewhite = 1, ar.method = "ols", tol_weights = 1e-7, tol_lam = 1e-9, tol_obj = 1e-9, tol_mom = 1e-9, maxiterlam = 100, optlam = c("nlminb", "optim", "iter", "Wu"), data, Lambdacontrol = list(), model = TRUE, X = FALSE, Y = FALSE, alpha = NULL, ...) } \arguments{ \item{g}{A function of the form \eqn{g(\theta,x)} and which returns a \eqn{n \times q} matrix with typical element \eqn{g_i(\theta,x_t)} for \eqn{i=1,...q} and \eqn{t=1,...,n}. This matrix is then used to build the q sample moment conditions. It can also be a formula if the model is linear (see details below). } \item{tet0}{A \eqn{k \times 1} vector of starting values. If the dimension of \eqn{\theta} is one, see the argument "optfct". In the linear case, if tet0=NULL, the 2-step gmm estimator is used as starting value. However, it has to be provided when eqConst is not NULL} \item{x}{The matrix or vector of data from which the function \eqn{g(\theta,x)} is computed. If "g" is a formula, it is an \eqn{n \times Nh} matrix of instruments (see details below).} \item{gradv}{A function of the form \eqn{G(\theta,x)} which returns a \eqn{q\times k} matrix of derivatives of \eqn{\bar{g}(\theta)} with respect to \eqn{\theta}. By default, the numerical algorithm \code{numericDeriv} is used. It is of course strongly suggested to provide this function when it is possible. This gradiant is used compute the asymptotic covariance matrix of \eqn{\hat{\theta}}. If "g" is a formula, the gradiant is not required (see the details below).} \item{smooth}{If set to TRUE, the moment function is smoothed as proposed by Kitamura(1997)} \item{type}{"EL" for empirical likelihood, "ET" for exponential tilting, "CUE" for continuous updated estimator, "ETEL" for exponentially tilted empirical likelihood of Schennach(2007), "HD" for Hellinger Distance of Kitamura-Otsu-Evdokimov (2013), and "ETHD" for the exponentially tilted Hellinger distance of Antoine-Dovonon (2015). "RCUE" is a restricted version of "CUE" in which the probabilities are bounded below by zero. In that case, an analytical Kuhn-Tucker method is used to find the solution.} \item{kernel}{type of kernel used to compute the covariance matrix of the vector of sample moment conditions (see \code{\link[sandwich]{kernHAC}} for more details) and to smooth the moment conditions if "smooth" is set to TRUE. Only two types of kernel are available. The truncated implies a Bartlett kernel for the HAC matrix and the Bartlett implies a Parzen kernel (see Smith 2004).} \item{bw}{The method to compute the bandwidth parameter. By default it is \code{\link[sandwich]{bwAndrews}} which is proposed by Andrews (1991). The alternative is \code{\link[sandwich]{bwNeweyWest}} of Newey-West(1994).} \item{prewhite}{logical or integer. Should the estimating functions be prewhitened? If \code{TRUE} or greater than 0 a VAR model of order \code{as.integer(prewhite)} is fitted via \code{ar} with method \code{"ols"} and \code{demean = FALSE}.} \item{ar.method}{character. The \code{method} argument passed to \code{\link[stats]{ar}} for prewhitening.} \item{approx}{a character specifying the approximation method if the bandwidth has to be chosen by \code{bwAndrews}.} \item{tol_weights}{numeric. Weights that exceed \code{tol} are used for computing the covariance matrix, all other weights are treated as 0.} \item{tol_lam}{Tolerance for \eqn{\lambda} between two iterations. The algorithm stops when \eqn{\|\lambda_i -\lambda_{i-1}\|} reaches \code{tol_lamb} (see \code{\link{getLamb}}) } \item{maxiterlam}{The algorithm to compute \eqn{\lambda} stops if there is no convergence after "maxiterlam" iterations (see \code{\link{getLamb}}).} \item{tol_obj}{Tolerance for the gradiant of the objective function to compute \eqn{\lambda} (see \code{\link{getLamb}}).} \item{optfct}{Only when the dimension of \eqn{\theta} is 1, you can choose between the algorithm \code{\link[stats]{optim}} or \code{\link[stats]{optimize}}. In that case, the former is unreliable. If \code{\link[stats]{optimize}} is chosen, "t0" must be \eqn{1\times 2} which represents the interval in which the algorithm seeks the solution.It is also possible to choose the \code{\link[stats]{nlminb}} algorithm. In that case, borns for the coefficients can be set by the options \code{upper=} and \code{lower=}.} \item{constraint}{If set to TRUE, the constraint optimization algorithm is used. See \code{\link[stats]{constrOptim}} to learn how it works. In particular, if you choose to use it, you need to provide "ui" and "ci" in order to impose the constraint \eqn{ui \theta - ci \geq 0}.} \item{tol_mom}{It is the tolerance for the moment condition \eqn{\sum_{t=1}^n p_t g(\theta(x_t)=0}, where \eqn{p_t=\frac{1}{n}D\rho()} is the implied probability. It adds a penalty if the solution diverges from its goal.} \item{optlam}{Algorithm used to solve for the lagrange multiplier in \code{\link{getLamb}}. The algorithm Wu is only for \code{type="EL"}. The value of \code{optlam} is ignored for "CUE" because in that case, the analytical solution exists.} \item{data}{A data.frame or a matrix with column names (Optional). } \item{Lambdacontrol}{Controls for the optimization of the vector of Lagrange multipliers used by either \code{\link[stats]{optim}}, \code{\link[stats]{nlminb}} or \code{\link[stats]{constrOptim}}} \item{model, X, Y}{logicals. If \code{TRUE} the corresponding components of the fit (the model frame, the model matrix, the response) are returned if g is a formula.} \item{TypeGel}{The name of the class object created by the method \code{getModel}. It allows developers to extand the package and create other GEL methods.} \item{alpha}{Regularization coefficient for discrete CGEL estimation (experimental). By setting alpha to any value, the model is estimated by CGEL of type specified by the option \code{type}. See Chausse (2011)} \item{eqConst}{Either a named vector (if "g" is a function), a simple vector for the nonlinear case indicating which of the \eqn{\theta_0} is restricted, or a qx2 vector defining equality constraints of the form \eqn{\theta_i=c_i}. See \code{\link{gmm}} for an example.} \item{eqConstFullVcov}{If FALSE, the constrained coefficients are assumed to be fixed and only the covariance of the unconstrained coefficients is computed. If TRUE, the covariance matrix of the full set of coefficients is computed.} \item{onlyCoefficients}{If \code{TRUE}, only the vector of coefficients and Lagrange multipliers are returned} \item{...}{More options to give to \code{\link[stats]{optim}}, \code{\link[stats]{optimize}} or \code{\link[stats]{constrOptim}}.} } \details{ If we want to estimate a model like \eqn{Y_t = \theta_1 + X_{2t}\theta_2 + ... + X_{k}\theta_k + \epsilon_t} using the moment conditions \eqn{Cov(\epsilon_tH_t)=0}, where \eqn{H_t} is a vector of \eqn{Nh} instruments, than we can define "g" like we do for \code{\link[stats]{lm}}. We would have \code{g = y~x2+x3+...+xk} and the argument "x" above would become the matrix H of instruments. As for \code{\link[stats]{lm}}, \eqn{Y_t} can be a \eqn{Ny \times 1} vector which would imply that \eqn{k=Nh \times Ny}. The intercept is included by default so you do not have to add a column of ones to the matrix \eqn{H}. You do not need to provide the gradiant in that case since in that case it is embedded in \code{\link{gel}}. The intercept can be removed by adding -1 to the formula. In that case, the column of ones need to be added manually to H. If "smooth" is set to TRUE, the sample moment conditions \eqn{\sum_{t=1}^n g(\theta,x_t)} is replaced by: \eqn{\sum_{t=1}^n g^k(\theta,x_t)}, where \eqn{g^k(\theta,x_t)=\sum_{i=-r}^r k(i) g(\theta,x_{t+i})}, where \eqn{r} is a truncated parameter that depends on the bandwidth and \eqn{k(i)} are normalized weights so that they sum to 1. The method solves \eqn{\hat{\theta} = \arg\min \left[\arg\max_\lambda \frac{1}{n}\sum_{t=1}^n \rho() - \rho(0) \right]} \code{\link{evalGel}} generates the object of class "gel" for a fixed vector of parameters. There is no estimation for \eqn{\theta}, but the optimal vector of Lagrange multipliers \eqn{\lambda} is computed. The objective function is then the profiled likelihood for a given \eqn{\theta}. It can be used to construct a confidence interval by inverting the likelihood ratio test. } \value{ 'gel' returns an object of 'class' '"gel"' The functions 'summary' is used to obtain and print a summary of the results. The object of class "gel" is a list containing at least the following: \item{coefficients}{\eqn{k\times 1} vector of parameters} \item{residuals}{the residuals, that is response minus fitted values if "g" is a formula.} \item{fitted.values}{the fitted mean values if "g" is a formula.} \item{lambda}{\eqn{q \times 1} vector of Lagrange multipliers.} \item{vcov_par}{the covariance matrix of "coefficients"} \item{vcov_lambda}{the covariance matrix of "lambda"} \item{pt}{The implied probabilities} \item{objective}{the value of the objective function} \item{conv_lambda}{Convergence code for "lambda" (see \code{\link{getLamb}})} \item{conv_mes}{Convergence message for "lambda" (see \code{\link{getLamb}})} \item{conv_par}{Convergence code for "coefficients" (see \code{\link[stats]{optim}}, \code{\link[stats]{optimize}} or \code{\link[stats]{constrOptim}})} \item{terms}{the \code{\link[stats]{terms}} object used when g is a formula.} \item{call}{the matched call.} \item{y}{if requested, the response used (if "g" is a formula).} \item{x}{if requested, the model matrix used if "g" is a formula or the data if "g" is a function.} \item{model}{if requested (the default), the model frame used if "g" is a formula.} } \references{ Anatolyev, S. (2005), GMM, GEL, Serial Correlation, and Asymptotic Bias. \emph{Econometrica}, \bold{73}, 983-1002. Andrews DWK (1991), Heteroskedasticity and Autocorrelation Consistent Covariance Matrix Estimation. \emph{Econometrica}, \bold{59}, 817--858. Kitamura, Yuichi (1997), Empirical Likelihood Methods With Weakly Dependent Processes. \emph{The Annals of Statistics}, \bold{25}, 2084-2102. Kitamura, Y. and Otsu, T. and Evdokimov, K. (2013), Robustness, Infinitesimal Neighborhoods and Moment Restrictions. \emph{Econometrica}, \bold{81}, 1185-1201. Newey, W.K. and Smith, R.J. (2004), Higher Order Properties of GMM and Generalized Empirical Likelihood Estimators. \emph{Econometrica}, \bold{72}, 219-255. Smith, R.J. (2004), GEL Criteria for Moment Condition Models. \emph{Working paper, CEMMAP}. Newey WK & West KD (1987), A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix. \emph{Econometrica}, \bold{55}, 703--708. Newey WK & West KD (1994), Automatic Lag Selection in Covariance Matrix Estimation. \emph{Review of Economic Studies}, \bold{61}, 631-653. Schennach, Susanne, M. (2007), Point Estimation with Exponentially Tilted Empirical Likelihood. \emph{Econometrica}, \bold{35}, 634-672. Wu, C. (2005), Algorithms and R codes for the pseudo empirical likelihood method in survey sampling. \emph{Survey Methodology}, \bold{31}(2), page 239. Zeileis A (2006), Object-oriented Computation of Sandwich Estimators. \emph{Journal of Statistical Software}, \bold{16}(9), 1--16. URL \doi{10.18637/jss.v016.i09}. Chausse (2010), Computing Generalized Method of Moments and Generalized Empirical Likelihood with R. \emph{Journal of Statistical Software}, \bold{34}(11), 1--35. URL \doi{10.18637/jss.v034.i11}. Chausse (2011), Generalized Empirical likelihood for a continumm of moment conditions. \emph{Working Paper}, \emph{Department of Economics}, \emph{University of Waterloo}. } \examples{ # First, an exemple with the fonction g() g <- function(tet, x) { n <- nrow(x) u <- (x[7:n] - tet[1] - tet[2]*x[6:(n-1)] - tet[3]*x[5:(n-2)]) f <- cbind(u, u*x[4:(n-3)], u*x[3:(n-4)], u*x[2:(n-5)], u*x[1:(n-6)]) return(f) } Dg <- function(tet,x) { n <- nrow(x) xx <- cbind(rep(1, (n-6)), x[6:(n-1)], x[5:(n-2)]) H <- cbind(rep(1, (n-6)), x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) f <- -crossprod(H, xx)/(n-6) return(f) } n = 200 phi<-c(.2, .7) thet <- 0.2 sd <- .2 set.seed(123) x <- matrix(arima.sim(n = n, list(order = c(2, 0, 1), ar = phi, ma = thet, sd = sd)), ncol = 1) res <- gel(g, x, c(0, .3, .6), grad = Dg) summary(res) # The same model but with g as a formula.... much simpler in that case y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H res <- gel(g, x, c(0, .3, .6)) summary(res) # Using evalGel to create the object without estimation res <- evalGel(g, x, res$coefficients) } gmm/man/summary.Rd0000644000176200001440000000565414247643115013612 0ustar liggesusers\name{summary} \alias{summary.gmm} \alias{summary.sysGmm} \alias{summary.gel} \alias{summary.ategel} \alias{summary.tsls} \alias{print.summary.gmm} \alias{print.summary.sysGmm} \alias{print.summary.gel} \alias{print.summary.tsls} \title{Method for object of class gmm or gel} \description{ It presents the results from the \code{gmm} or \code{gel} estimation in the same fashion as \code{summary} does for the \code{lm} class objects for example. It also compute the tests for overidentifying restrictions. } \usage{ \method{summary}{gmm}(object, ...) \method{summary}{sysGmm}(object, ...) \method{summary}{gel}(object, ...) \method{summary}{ategel}(object, robToMiss = TRUE, ...) \method{summary}{tsls}(object, vcov = NULL, ...) \method{print}{summary.gmm}(x, digits = 5, ...) \method{print}{summary.sysGmm}(x, digits = 5, ...) \method{print}{summary.gel}(x, digits = 5, ...) \method{print}{summary.tsls}(x, digits = 5, ...) } \arguments{ \item{object}{An object of class \code{gmm} or \code{gel} returned by the function \code{\link{gmm}} or \code{\link{gel}}} \item{x}{An object of class \code{summary.gmm} or \code{summary.gel} returned by the function \code{\link{summary.gmm}} \code{\link{summary.gel}}} \item{digits}{The number of digits to be printed} \item{vcov}{An alternative covariance matrix computed with \code{vcov.tsls}} \item{robToMiss}{If \code{TRUE}, it computes the robust to misspecification covariance matrix} \item{...}{Other arguments when summary is applied to another class object} } \value{ It returns a list with the parameter estimates and their standard deviations, t-stat and p-values. It also returns the J-test and p-value for the null hypothesis that \eqn{E(g(\theta,X)=0} } \references{ Hansen, L.P. (1982), Large Sample Properties of Generalized Method of Moments Estimators. \emph{Econometrica}, \bold{50}, 1029-1054, Hansen, L.P. and Heaton, J. and Yaron, A.(1996), Finit-Sample Properties of Some Alternative GMM Estimators. \emph{Journal of Business and Economic Statistics}, \bold{14} 262-280. Anatolyev, S. (2005), GMM, GEL, Serial Correlation, and Asymptotic Bias. \emph{Econometrica}, \bold{73}, 983-1002. Kitamura, Yuichi (1997), Empirical Likelihood Methods With Weakly Dependent Processes. \emph{The Annals of Statistics}, \bold{25}, 2084-2102. Newey, W.K. and Smith, R.J. (2004), Higher Order Properties of GMM and Generalized Empirical Likelihood Estimators. \emph{Econometrica}, \bold{72}, 219-255. } \examples{ # GMM # set.seed(444) n = 500 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n = n, list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] ym3 <- x[4:(n-3)] ym4 <- x[3:(n-4)] ym5 <- x[2:(n-5)] ym6 <- x[1:(n-6)] g <- y ~ ym1 + ym2 x <- ~ym3+ym4+ym5+ym6 res <- gmm(g, x) summary(res) # GEL # t0 <- res$coef res <- gel(g, x, t0) summary(res) # tsls # res <- tsls(y ~ ym1 + ym2,~ym3+ym4+ym5+ym6) summary(res) } gmm/man/coef.Rd0000644000176200001440000000200314247643115013012 0ustar liggesusers\name{coef} \alias{coef.gel} \alias{coef.gmm} \title{Coefficients of GEL or GMM} \description{ It extracts the coefficients from \code{gel} or \code{gmm} objects. } \usage{ \method{coef}{gmm}(object, ...) \method{coef}{gel}(object, lambda = FALSE, ...) } \arguments{ \item{object}{An object of class \code{gel} or \code{gmm} returned by the function \code{\link{gel}} or \code{\link{gmm}}} \item{lambda}{If set to TRUE, the lagrange multipliers are extracted instead of the vector of coefficients} \item{...}{Other arguments when \code{coef} is applied to an other class object} } \value{ Vector of coefficients } \examples{ ################# n = 500 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n=n,list(order=c(2,0,1),ar=phi,ma=thet,sd=sd)),ncol=1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H t0 <- c(0,.5,.5) res <- gel(g, x, t0) coef(res) coef(res, lambda = TRUE) ################### res <- gmm(g, x) coef(res) } gmm/man/momentEstim.Rd0000644000176200001440000000412514247643115014406 0ustar liggesusers\name{momentEstim} \alias{momentEstim.baseGmm.twoStep} \alias{momentEstim.baseGmm.twoStep.formula} \alias{momentEstim.sysGmm.twoStep.formula} \alias{momentEstim.tsls.twoStep.formula} \alias{momentEstim.baseGmm.iterative.formula} \alias{momentEstim.baseGmm.iterative} \alias{momentEstim.baseGmm.cue.formula} \alias{momentEstim.baseGmm.cue} \alias{momentEstim.baseGmm.eval} \alias{momentEstim.baseGel.mod} \alias{momentEstim.baseGel.modFormula} \alias{momentEstim.baseGel.eval} \title{Method for estimating models based on moment conditions} \description{ It estimates a model which is caracterized by the method \code{getModel} (see details). } \usage{ \method{momentEstim}{baseGmm.twoStep}(object, ...) \method{momentEstim}{baseGmm.twoStep.formula}(object, ...) \method{momentEstim}{sysGmm.twoStep.formula}(object, ...) \method{momentEstim}{tsls.twoStep.formula}(object, ...) \method{momentEstim}{baseGmm.iterative.formula}(object, ...) \method{momentEstim}{baseGmm.iterative}(object, ...) \method{momentEstim}{baseGmm.cue.formula}(object, ...) \method{momentEstim}{baseGmm.cue}(object, ...) \method{momentEstim}{baseGmm.eval}(object, ...) \method{momentEstim}{baseGel.mod}(object, ...) \method{momentEstim}{baseGel.modFormula}(object, ...) \method{momentEstim}{baseGel.eval}(object, ...) } \arguments{ \item{object}{An object created by the method \code{getModel}} \item{...}{Other arguments when \code{momentEstim} is applied to an other class object} } \value{ It returns an object of class determined by the argument "TypeGMM" of \code{\link{gmm}}. By default, it is of class \code{baseGmm.res}. It estimates the model and organize the results that will be finalized by the method \code{FinRes}. More methods can be created in order to use other GMM methods not yet included in the package. } \references{ Hansen, L.P. (1982), Large Sample Properties of Generalized Method of Moments Estimators. \emph{Econometrica}, \bold{50}, 1029-1054, Hansen, L.P. and Heaton, J. and Yaron, A.(1996), Finit-Sample Properties of Some Alternative GMM Estimators. \emph{Journal of Business and Economic Statistics}, \bold{14} 262-280. } gmm/man/formula.Rd0000644000176200001440000000206214247643115013550 0ustar liggesusers\name{formula} \alias{formula.gel} \alias{formula.gmm} \title{Formula method for gel and gmm objects} \description{ Method to extract the formula from \code{gel} or \code{gmm} objects. } \usage{ \method{formula}{gel}(x, ...) \method{formula}{gmm}(x, ...) } \arguments{ \item{x}{An object of class \code{gel} or \code{gmm} returned by the function \code{\link{gel}} or \code{\link{gmm}}} \item{...}{Other arguments to pass to other methods} } \examples{ ## GEL ## n = 200 phi<-c(.2,.7) thet <- 0.2 sd <- .2 set.seed(123) x <- matrix(arima.sim(n = n, list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H res <- gel(g, x, c(0,.3,.6)) formula(res) # GMM is like GLS for linear models without endogeneity problems set.seed(345) n = 200 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n = n, list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- 10 + 5*rnorm(n) + x res <- gmm(y ~ x, x) formula(res) } gmm/man/getLamb.Rd0000644000176200001440000000756215044417256013471 0ustar liggesusers\name{getLamb} \alias{getLamb} \title{Solving for the Lagrange multipliers of Generalized Empirical Likelihood (GEL) } \description{ It computes the vector of Lagrange multipliers, which maximizes the GEL objective function, using an iterative Newton method. } \usage{ getLamb(gt, l0, type = c("EL","ET","CUE", "ETEL", "HD","ETHD","RCUE"), tol_lam = 1e-7, maxiterlam = 100, tol_obj = 1e-7, k = 1, method = c("nlminb", "optim", "iter", "Wu"), control = list()) } \arguments{ \item{gt}{A \eqn{n \times q} matrix with typical element \eqn{g_i(\theta,x_t)}} \item{l0}{Vector of starting values for lambda} \item{type}{"EL" for empirical likelihood, "ET" for exponential tilting, "CUE" for continuous updated estimator, and "HD" for Hellinger Distance. See details for "ETEL" and "ETHD". "RCUE" is a restricted version of "CUE" in which the probabilities are bounded below by zero. In that case, an analytical Kuhn-Tucker method is used to find the solution.} \item{tol_lam}{Tolerance for \eqn{\lambda} between two iterations. The algorithm stops when \eqn{\|\lambda_i -\lambda_{i-1}\|} reaches \code{tol_lam} } \item{maxiterlam}{The algorithm stops if there is no convergence after "maxiterlam" iterations.} \item{tol_obj}{Tolerance for the gradiant of the objective function. The algorithm returns a non-convergence message if \eqn{\max(|gradiant|)} does not reach \code{tol_obj}. It helps the \code{gel} algorithm to select the right space to look for \eqn{\theta}} \item{k}{It represents the ratio k1/k2, where \eqn{k1=\int_{-\infty}^{\infty} k(s)ds} and \eqn{k2=\int_{-\infty}^{\infty} k(s)^2 ds}. See Smith(2004).} \item{method}{The iterative procedure uses a Newton method for solving the FOC. It i however recommended to use \code{optim} or \code{nlminb}. If type is set to "EL" and method to "optim", \code{\link[stats]{constrOptim}} is called to prevent \eqn{log(1-gt'\lambda)} from producing NA. The gradient and hessian is provided to \code{nlminb} which speed up the convergence. The latter is therefore the default value. "Wu" is for "EL" only. It uses the algorithm of Wu (2005). The value of \code{method} is ignored for "CUE" because in that case, the analytical solution exists.} \item{control}{Controls to send to \code{\link[stats]{optim}}, \code{\link[stats]{nlminb}} or \code{\link[stats]{constrOptim}}} } \details{ It solves the problem \eqn{\max_{\lambda} \frac{1}{n}\sum_{t=1}^n \rho(gt'\lambda)}. For the type "ETEL", it is only used by \code{\link{gel}}. In that case \eqn{\lambda} is obtained by maximizing \eqn{\frac{1}{n}\sum_{t=1}^n \rho(gt'\lambda)}, using \eqn{\rho(v)=-\exp{v}} (so ET) and \eqn{\theta} by minimizing the same equation but with \eqn{\rho(v)-\log{(1-v)}}. To avoid NA's, \code{\link[stats]{constrOptim}} is used with the restriction \eqn{\lambda'g_t < 1}. The type "ETHD" is experimental and proposed by Antoine-Dovonon (2015). The paper is not yet available. } \value{ lambda: A \eqn{q\times 1} vector of Lagrange multipliers which solve the system of equations given above. \code{conv}: Details on the type of convergence. } \references{ Newey, W.K. and Smith, R.J. (2004), Higher Order Properties of GMM and Generalized Empirical Likelihood Estimators. \emph{Econometrica}, \bold{72}, 219-255. Smith, R.J. (2004), GEL Criteria for Moment Condition Models. \emph{Working paper, CEMMAP}. Wu, C. (2005), Algorithms and R codes for the pseudo empirical likelihood method in survey sampling. \emph{Survey Methodology}, \bold{31}(2), page 239. } \examples{ g <- function(tet,x) { n <- nrow(x) u <- (x[7:n] - tet[1] - tet[2]*x[6:(n-1)] - tet[3]*x[5:(n-2)]) f <- cbind(u, u*x[4:(n-3)], u*x[3:(n-4)], u*x[2:(n-5)], u*x[1:(n-6)]) return(f) } n = 500 phi<-c(.2, .7) thet <- 0.2 sd <- .2 x <- matrix(arima.sim(n = n, list(order = c(2, 0, 1), ar = phi, ma = thet, sd = sd)), ncol = 1) gt <- g(c(0,phi),x) getLamb(gt, type = "EL",method="optim") } gmm/man/specTest.Rd0000644000176200001440000000244514247643115013702 0ustar liggesusers\name{specTest} \alias{specTest} \alias{specTest.gel} \alias{specTest.gmm} \alias{print.specTest} \title{Compute tests of specification} \description{ Generic function for testing the specification of estimated models. It computes the J-test from \code{gmm} objects and J-test, LR-test and LM-test from \code{gel} objects. } \usage{ \method{specTest}{gmm}(x, ...) \method{specTest}{gel}(x, ...) \method{print}{specTest}(x, digits = 5, ...) specTest(x, \dots) } \arguments{ \item{x}{A fitted model object.} \item{digits}{The number of digits to be printed.} \item{\dots}{Arguments passed to methods.} } \value{ Tests and p-values } \references{ Hansen, L.P. (1982), Large Sample Properties of Generalized Method of Moments Estimators. \emph{Econometrica}, \bold{50}, 1029-1054, Smith, R. J. (2004), GEL Criteria for Moment Condition Models. \emph{CeMMAP working papers, Institute for Fiscal Studies} } \examples{ ################# n = 500 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n=n,list(order=c(2,0,1),ar=phi,ma=thet,sd=sd)),ncol=1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H t0 <- c(0,.5,.5) res <- gel(g, x, t0) specTest(res) ################### res <- gmm(g, x) specTest(res) } gmm/man/marginal.Rd0000644000176200001440000000277014247643115013703 0ustar liggesusers\name{marginal} \alias{marginal} \alias{marginal.ategel} \title{Marginal effects Summary} \description{ It produces the summary table of marginal effects for GLM estimation with GEL. Only implemented for ATEgel. } \usage{ \method{marginal}{ategel}(object, ...) } \arguments{ \item{object}{An object of class \code{ategel} returned by the function \code{\link{ATEgel}}} \item{...}{Other arguments for other methods} } \value{ It returns a matrix with the marginal effects, the standard errors based on the Delta method when the link is nonlinear, the t-ratios, and the pvalues.} \references{ Owen, A.B. (2001), Empirical Likelihood. \emph{Monographs on Statistics and Applied Probability 92, Chapman and Hall/CRC} } \examples{ ## We create some artificial data with unbalanced groups and binary outcome genDat <- function(n) { eta=c(-1, .5, -.25, -.1) Z <- matrix(rnorm(n*4),ncol=4) b <- c(27.4, 13.7, 13.7, 13.7) bZ <- c(Z\%*\%b) Y1 <- as.numeric(rnorm(n, mean=210+bZ)>220) Y0 <- as.numeric(rnorm(n, mean=200-.5*bZ)>220) etaZ <- c(Z\%*\%eta) pZ <- exp(etaZ)/(1+exp(etaZ)) T <- rbinom(n, 1, pZ) Y <- T*Y1+(1-T)*Y0 X1 <- exp(Z[,1]/2) X2 <- Z[,2]/(1+exp(Z[,1])) X3 <- (Z[,1]*Z[,3]/25+0.6)^3 X4 <- (Z[,2]+Z[,4]+20)^2 data.frame(Y=Y, cbind(X1,X2,X3,X4), T=T) } dat <- genDat(200) res <- ATEgel(Y~T, ~X1+X2+X3+X4, data=dat, type="ET", family="logit") summary(res) marginal(res) } gmm/man/gmm.Rd0000644000176200001440000004174015044417453012671 0ustar liggesusers\name{gmm} \alias{gmm} \alias{gmmWithConst} \alias{evalGmm} \title{Generalized method of moment estimation} \description{ Function to estimate a vector of parameters based on moment conditions using the GMM method of Hansen(82). } \usage{ gmm(g,x,t0=NULL,gradv=NULL, type=c("twoStep","cue","iterative"), wmatrix = c("optimal","ident"), vcov=c("HAC","MDS","iid","TrueFixed"), kernel=c("Quadratic Spectral","Truncated", "Bartlett", "Parzen", "Tukey-Hanning"), crit=10e-7,bw = bwAndrews, prewhite = 1, ar.method = "ols", approx="AR(1)", tol = 1e-7, itermax=100,optfct=c("optim","optimize","nlminb", "constrOptim"), model=TRUE, X=FALSE, Y=FALSE, TypeGmm = "baseGmm", centeredVcov = TRUE, weightsMatrix = NULL, traceIter = FALSE, data, eqConst = NULL, eqConstFullVcov = FALSE, mustar = NULL, onlyCoefficients=FALSE, ...) evalGmm(g, x, t0, tetw=NULL, gradv=NULL, wmatrix = c("optimal","ident"), vcov=c("HAC","iid","TrueFixed"), kernel=c("Quadratic Spectral","Truncated", "Bartlett", "Parzen", "Tukey-Hanning"),crit=10e-7,bw = bwAndrews, prewhite = FALSE, ar.method = "ols", approx="AR(1)",tol = 1e-7, model=TRUE, X=FALSE, Y=FALSE, centeredVcov = TRUE, weightsMatrix = NULL, data, mustar = NULL) gmmWithConst(obj, which, value) } \arguments{ \item{g}{A function of the form \eqn{g(\theta,x)} and which returns a \eqn{n \times q} matrix with typical element \eqn{g_i(\theta,x_t)} for \eqn{i=1,...q} and \eqn{t=1,...,n}. This matrix is then used to build the q sample moment conditions. It can also be a formula if the model is linear (see details below).} \item{x}{The matrix or vector of data from which the function \eqn{g(\theta,x)} is computed. If "g" is a formula, it is an \eqn{n \times Nh} matrix of instruments or a formula (see details below).} \item{t0}{A \eqn{k \times 1} vector of starting values. It is required only when "g" is a function because only then a numerical algorithm is used to minimize the objective function. If the dimension of \eqn{\theta} is one, see the argument "optfct".} \item{tetw}{A \eqn{k \times 1} vector to compute the weighting matrix.} \item{gradv}{A function of the form \eqn{G(\theta,x)} which returns a \eqn{q\times k} matrix of derivatives of \eqn{\bar{g}(\theta)} with respect to \eqn{\theta}. By default, the numerical algorithm \code{numericDeriv} is used. It is of course strongly suggested to provide this function when it is possible. This gradient is used to compute the asymptotic covariance matrix of \eqn{\hat{\theta}} and to obtain the analytical gradient of the objective function if the method is set to "CG" or "BFGS" in \code{\link[stats]{optim}} and if "type" is not set to "cue". If "g" is a formula, the gradiant is not required (see the details below).} \item{type}{The GMM method: "twostep" is the two step GMM proposed by Hansen(1982) and the "cue" and "iterative" are respectively the continuous updated and the iterative GMM proposed by Hansen, Eaton et Yaron (1996)} \item{wmatrix}{Which weighting matrix should be used in the objective function. By default, it is the inverse of the covariance matrix of \eqn{g(\theta,x)}. The other choice is the identity matrix which is usually used to obtain a first step estimate of \eqn{\theta} } \item{vcov}{Assumption on the properties of the random vector x. By default, x is a weakly dependant process. The "iid" option will avoid using the HAC matrix which will accelerate the estimation if one is ready to make that assumption. The option "TrueFixed" is used only when the matrix of weights is provided and it is the optimal one.} \item{kernel}{type of kernel used to compute the covariance matrix of the vector of sample moment conditions (see \code{\link[sandwich]{kernHAC}} for more details)} \item{crit}{The stopping rule for the iterative GMM. It can be reduce to increase the precision.} \item{bw}{The method to compute the bandwidth parameter in the HAC weighting matrix. The default is \code{link[sandwich]{bwAndrews}} (as proposed in Andrews (1991)), which minimizes the MSE of the weighting matrix. Alternatives are \code{link{bwWilhelm}} (as proposed in Wilhelm (2015)), which minimizes the mean-square error (MSE) of the resulting GMM estimator, and \code{link[sandwich]{bwNeweyWest}} (as proposed in Newey-West(1994)).} \item{prewhite}{logical or integer. Should the estimating functions be prewhitened? If \code{TRUE} or greater than 0 a VAR model of order \code{as.integer(prewhite)} is fitted via \code{ar} with method \code{"ols"} and \code{demean = FALSE}.} \item{ar.method}{character. The \code{method} argument passed to \code{\link[stats]{ar}} for prewhitening.} \item{approx}{A character specifying the approximation method if the bandwidth has to be chosen by \code{bwAndrews}.} \item{tol}{Weights that exceed \code{tol} are used for computing the covariance matrix, all other weights are treated as 0.} \item{itermax}{The maximum number of iterations for the iterative GMM. It is unlikely that the algorithm does not converge but we keep it as a safety.} \item{optfct}{Only when the dimension of \eqn{\theta} is 1, you can choose between the algorithm \code{\link[stats]{optim}} or \code{\link[stats]{optimize}}. In that case, the former is unreliable. If \code{\link[stats]{optimize}} is chosen, "t0" must be \eqn{1\times 2} which represents the interval in which the algorithm seeks the solution. It is also possible to choose the \code{\link[stats]{nlminb}} algorithm. In that case, boundaries for the coefficients can be set by the options \code{upper=} and \code{lower=}. The \code{\link[stats]{constrOptim}} is only available for nonlinear models for now. The standard errors may have to be corrected if the estimtes reach the boundary set by ui and ci.} \item{model, X, Y}{logical. If \code{TRUE} the corresponding components of the fit (the model frame, the model matrix, the response) are returned if g is a formula.} \item{TypeGmm}{The name of the class object created by the method \code{getModel}. It allows developers to extend the package and create other GMM methods.} \item{centeredVcov}{Should the moment function be centered when computing its covariance matrix. Doing so may improve inference.} \item{weightsMatrix}{It allows users to provide \code{gmm} with a fixed weighting matrix. This matrix must be \eqn{q \times q}, symmetric and strictly positive definite. When provided, the \code{type} option becomes irrelevant. } \item{traceIter}{Tracing information for GMM of type "iter"} \item{data}{A data.frame or a matrix with column names (Optional). } \item{eqConst}{Either a named vector (if "g" is a function), a simple vector for the nonlinear case indicating which of the \eqn{\theta_0} is restricted, or a qx2 vector defining equality constraints of the form \eqn{\theta_i=c_i}. See below for an example.} \item{which, value}{The equality constraint is of the form which=value. "which" can be a vector of type characters with the names of the coefficients being constrained, or a vector of type numeric with the position of the coefficient in the whole vector.} \item{obj}{Object of class "gmm"} \item{eqConstFullVcov}{If FALSE, the constrained coefficients are assumed to be fixed and only the covariance of the unconstrained coefficients is computed. If TRUE, the covariance matrix of the full set of coefficients is computed.} \item{mustar}{If not null, it must be a vector with the number of elements being equal to the number of moment conditions. In that case, the vector is subtracted from the sample moment vector before minimizing the objective function. It is useful to do a bootstrap procedure. } \item{onlyCoefficients}{If set to \code{TRUE}, the function only returns the coefficient estimates. It may be of interest when the standard errors are not needed} \item{...}{More options to give to \code{\link[stats]{optim}}.} } \details{ If we want to estimate a model like \eqn{Y_t = \theta_1 + X_{2t} \theta_2 + \cdots + X_{k}\theta_k + \epsilon_t} using the moment conditions \eqn{Cov(\epsilon_tH_t)=0}, where \eqn{H_t} is a vector of \eqn{Nh} instruments, than we can define "g" like we do for \code{\link[stats]{lm}}. We would have \eqn{g = y ~\tilde{}~ x2+x3+ \cdots +xk} and the argument "x" above would become the matrix H of instruments. As for \code{\link[stats]{lm}}, \eqn{Y_t} can be a \eqn{Ny \times 1} vector which would imply that \eqn{k=Nh \times Ny}. The intercept is included by default so you do not have to add a column of ones to the matrix \eqn{H}. You do not need to provide the gradiant in that case since in that case it is embedded in \code{\link{gmm}}. The intercept can be removed by adding -1 to the formula. In that case, the column of ones need to be added manually to H. It is also possible to express "x" as a formula. For example, if the instruments are \eqn{\{1,z_1,z_2,z_3\}}, we can set "x" to \eqn{\tilde{} z1+z2+z3}. By default, a column of ones is added. To remove it, set "x" to \eqn{\tilde{}z1+z2+z3-1}. The following explains the last example bellow. Thanks to Dieter Rozenich, a student from the Vienna University of Economics and Business Administration. He suggested that it would help to understand the implementation of the Jacobian. For the two parameters of a normal distribution \eqn{(\mu,\sigma)} we have the following three moment conditions: \deqn{ m_{1} = \mu - x_{i} } \deqn{ m_{2} = \sigma^2 - (x_{i}-\mu)^2 } \deqn{ m_{3} = x_{i}^{3} - \mu (\mu^2+3\sigma^{2}) } \eqn{m_{1},m_{2}} can be directly obtained by the definition of \eqn{(\mu,\sigma)}. The third moment condition comes from the third derivative of the moment generating function (MGF) \deqn{ M_{X}(t) = exp\Big(\mu t + \frac{\sigma^{2}t^{2}}{2}\Big) } evaluated at \eqn{(t=0)}. Note that we have more equations (3) than unknown parameters (2). The Jacobian of these two conditions is (it should be an array but I can't make it work): \deqn{ 1~~~~~~~~~~ 0 } \deqn{ -2\mu+2x ~~~~~ 2\sigma } \deqn{-3\mu^{2}-3\sigma^{2} ~~~~ -6\mu\sigma} \code{gmmWithConst()} re-estimates an unrestricted model by adding an equality constraint. \code{evalGmm()} creates an object of class '"gmm"' for a given parameter vector. If no vector "tetw" is provided and the weighting matrix needs to be computed, "t0" is used., } \value{ 'gmm' returns an object of 'class' '"gmm"' The functions 'summary' is used to obtain and print a summary of the results. It also compute the J-test of overidentying restriction The object of class "gmm" is a list containing at least: \item{coefficients}{\eqn{k\times 1} vector of coefficients} \item{residuals}{the residuals, that is response minus fitted values if "g" is a formula.} \item{fitted.values}{the fitted mean values if "g" is a formula.} \item{vcov}{the covariance matrix of the coefficients} \item{objective}{the value of the objective function \eqn{\| var(\bar{g})^{-1/2}\bar{g}\|^2}} \item{terms}{the \code{\link[stats]{terms}} object used when g is a formula.} \item{call}{the matched call.} \item{y}{if requested, the response used (if "g" is a formula).} \item{x}{if requested, the model matrix used if "g" is a formula or the data if "g" is a function.} \item{model}{if requested (the default), the model frame used if "g" is a formula.} \item{algoInfo}{Information produced by either \code{\link[stats]{optim}} or \code{\link[stats]{nlminb}} related to the convergence if "g" is a function. It is printed by the \code{summary.gmm} method.} } \references{ Zeileis A (2006), Object-oriented Computation of Sandwich Estimators. \emph{Journal of Statistical Software}, \bold{16}(9), 1--16. URL \doi{10.18637/jss.v016.i09}. Pierre Chausse (2010), Computing Generalized Method of Moments and Generalized Empirical Likelihood with R. \emph{Journal of Statistical Software}, \bold{34}(11), 1--35. URL \doi{10.18637/jss.v034.i11}. Andrews DWK (1991), Heteroskedasticity and Autocorrelation Consistent Covariance Matrix Estimation. \emph{Econometrica}, \bold{59}, 817--858. Newey WK & West KD (1987), A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix. \emph{Econometrica}, \bold{55}, 703--708. Newey WK & West KD (1994), Automatic Lag Selection in Covariance Matrix Estimation. \emph{Review of Economic Studies}, \bold{61}, 631-653. Hansen, L.P. (1982), Large Sample Properties of Generalized Method of Moments Estimators. \emph{Econometrica}, \bold{50}, 1029-1054, Hansen, L.P. and Heaton, J. and Yaron, A.(1996), Finite-Sample Properties of Some Alternative GMM Estimators. \emph{Journal of Business and Economic Statistics}, \bold{14} 262-280. } \examples{ ## CAPM test with GMM data(Finance) r <- Finance[1:300, 1:10] rm <- Finance[1:300, "rm"] rf <- Finance[1:300, "rf"] z <- as.matrix(r-rf) t <- nrow(z) zm <- rm-rf h <- matrix(zm, t, 1) res <- gmm(z ~ zm, x = h) summary(res) ## linear tests can be performed using linearHypothesis from the car package ## The CAPM can be tested as follows: library(car) linearHypothesis(res,cbind(diag(10),matrix(0,10,10)),rep(0,10)) # The CAPM of Black g <- function(theta, x) { e <- x[,2:11] - theta[1] - (x[,1] - theta[1]) \%*\% matrix(theta[2:11], 1, 10) gmat <- cbind(e, e*c(x[,1])) return(gmat) } x <- as.matrix(cbind(rm, r)) res_black <- gmm(g, x = x, t0 = rep(0, 11)) summary(res_black)$coefficients ## APT test with Fama-French factors and GMM f1 <- zm f2 <- Finance[1:300, "hml"] f3 <- Finance[1:300, "smb"] h <- cbind(f1, f2, f3) res2 <- gmm(z ~ f1 + f2 + f3, x = h) coef(res2) summary(res2)$coefficients ## Same result with x defined as a formula: res2 <- gmm(z ~ f1 + f2 + f3, ~ f1 + f2 + f3) coef(res2) ## The following example has been provided by Dieter Rozenich (see details). # It generates normal random numbers and uses the GMM to estimate # mean and sd. #------------------------------------------------------------------------------- # Random numbers of a normal distribution # First we generate normally distributed random numbers and compute the two parameters: n <- 1000 x <- rnorm(n, mean = 4, sd = 2) # Implementing the 3 moment conditions g <- function(tet, x) { m1 <- (tet[1] - x) m2 <- (tet[2]^2 - (x - tet[1])^2) m3 <- x^3 - tet[1]*(tet[1]^2 + 3*tet[2]^2) f <- cbind(m1, m2, m3) return(f) } # Implementing the jacobian Dg <- function(tet, x) { jacobian <- matrix(c( 1, 2*(-tet[1]+mean(x)), -3*tet[1]^2-3*tet[2]^2,0, 2*tet[2], -6*tet[1]*tet[2]), nrow=3,ncol=2) return(jacobian) } # Now we want to estimate the two parameters using the GMM. gmm(g, x, c(0, 0), grad = Dg) # Two-stage-least-squares (2SLS), or IV with iid errors. # The model is: # Y(t) = b[0] + b[1]C(t) + b[2]Y(t-1) + e(t) # e(t) is an MA(1) # The instruments are Z(t)={1 C(t) y(t-2) y(t-3) y(t-4)} getdat <- function(n) { e <- arima.sim(n,model=list(ma=.9)) C <- runif(n,0,5) Y <- rep(0,n) Y[1] = 1 + 2*C[1] + e[1] for (i in 2:n){ Y[i] = 1 + 2*C[i] + 0.9*Y[i-1] + e[i] } Yt <- Y[5:n] X <- cbind(1,C[5:n],Y[4:(n-1)]) Z <- cbind(1,C[5:n],Y[3:(n-2)],Y[2:(n-3)],Y[1:(n-4)]) return(list(Y=Yt,X=X,Z=Z)) } d <- getdat(5000) res4 <- gmm(d$Y~d$X-1,~d$Z-1,vcov="iid") res4 ### Examples with equality constraint ###################################### # Random numbers of a normal distribution ## Not run: # The following works but produces warning message because the dimension of coef is 1 # Brent should be used # without named vector # Method Brent is used because the problem is now one-dimensional gmm(g, x, c(4, 0), grad = Dg, eqConst=1, method="Brent", lower=-10,upper=10) # with named vector gmm(g, x, c(mu=4, sig=2), grad = Dg, eqConst="sig", method="Brent", lower=-10,upper=10) ## End(Not run) gmm(g, x, c(4, 0), grad = Dg, eqConst=1,method="Brent",lower=0,upper=6) gmm(g, x, c(mu=4, sig=2), grad = Dg, eqConst="sig",method="Brent",lower=0,upper=6) # Example with formula # first coef = 0 and second coef = 1 # Only available for one dimensional yt z <- z[,1] res2 <- gmm(z ~ f1 + f2 + f3, ~ f1 + f2 + f3, eqConst = matrix(c(1,2,0,1),2,2)) res2 # CUE with starting t0 requires eqConst to be a vector res3 <- gmm(z ~ f1 + f2 + f3, ~ f1 + f2 + f3, t0=c(0,1,.5,.5), type="cue", eqConst = c(1,2)) res3 ### Examples with equality constraints, where the constrained coefficients is used to compute ### the covariance matrix. ### Useful when some coefficients have been estimated before, they are just identified in GMM ### and don't need to be re-estimated. ### To use with caution because the covariance won't be valid if the coefficients do not solve ### the GMM FOC. ###################################### res4 <- gmm(z ~ f1 + f2 + f3, ~ f1 + f2 + f3, t0=c(0,1,.5,.5), eqConst = c(1,2), eqConstFullVcov=TRUE) summary(res4) ### Examples with equality constraint using gmmWithConst ########################################################### res2 <- gmm(z ~ f1 + f2 + f3, ~ f1 + f2 + f3) gmmWithConst(res2,c("f2","f3"),c(.5,.5)) gmmWithConst(res2,c(2,3),c(.5,.5)) ## Creating an object without estimation for a fixed parameter vector ################################################################### res2_2 <- evalGmm(z ~ f1 + f2 + f3, ~ f1 + f2 + f3, t0=res2$coefficients, tetw=res2$coefficients) summary(res2_2) } gmm/man/getImpProb.Rd0000644000176200001440000000247514247643115014163 0ustar liggesusers\name{getImpProb} \alias{getImpProb} \alias{getImpProb.gel} \title{Implied Probabilities} \description{ It computes the implied probabilities from objects of class \code{gel} with additional options. } \usage{ \method{getImpProb}{gel}(object, posProb=TRUE, normalize=TRUE, checkConv=FALSE,...) } \arguments{ \item{object}{Object of class \code{gel}.} \item{posProb}{Should the implied probabilities be transformed into positive probabilities?} \item{normalize}{Should we normalize the probabilities so that they sum to one?} \item{checkConv}{Should we add the attribute convergence to check the sum of the probabilities and the weighted sum of the moment conditions?} \item{...}{Additional arguments to pass to other methods} } \value{ A vector af implied probabilities. } \references{ Newey, W.K. and Smith, R.J. (2004), Higher Order Properties of GMM and Generalized Empirical Likelihood Estimators. \emph{Econometrica}, \bold{72}, 219-255. } \examples{ ################# n = 500 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n=n,list(order=c(2,0,1),ar=phi,ma=thet,sd=sd)),ncol=1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H t0 <- c(0,.5,.5) res <- gel(g, x, t0) pt <- getImpProb(res) } gmm/man/FinRes.Rd0000644000176200001440000000161714247643115013276 0ustar liggesusers\name{FinRes} \alias{FinRes.baseGmm.res} \title{Method to finalize the result of the momentEstim method} \description{ It computes the final results that will be needed to create the object of class \code{gmm}.). } \usage{ \method{FinRes}{baseGmm.res}(z, object, ...) } \arguments{ \item{z}{An object of class determined by the method \code{momentEstim}.} \item{object}{An object produced my \code{getModel}} \item{...}{Other argument to be passed to other \code{FinRes} methods.} } \value{ It returns an object of class \code{gmm}. See \code{\link{gmm}} for more details. } \references{ Hansen, L.P. (1982), Large Sample Properties of Generalized Method of Moments Estimators. \emph{Econometrica}, \bold{50}, 1029-1054, Hansen, L.P. and Heaton, J. and Yaron, A.(1996), Finit-Sample Properties of Some Alternative GMM Estimators. \emph{Journal of Business and Economic Statistics}, \bold{14} 262-280. } gmm/man/bread.Rd0000644000176200001440000000342114247643115013160 0ustar liggesusers\name{bread} \alias{bread.gmm} \alias{bread.gel} \alias{bread.tsls} \title{Bread for sandwiches} \description{ Computes the bread of the sandwich covariance matrix } \usage{ \method{bread}{gmm}(x, ...) \method{bread}{gel}(x, ...) \method{bread}{tsls}(x, ...) } \arguments{ \item{x}{A fitted model of class \code{gmm} or \code{gel}.} \item{...}{Other arguments when \code{bread} is applied to another class object} } \details{ When the weighting matrix is not the optimal one, the covariance matrix of the estimated coefficients is: \eqn{(G'WG)^{-1} G'W V W G(G'WG)^{-1}}, where \eqn{G=d\bar{g}/d\theta}, \eqn{W} is the matrix of weights, and \eqn{V} is the covariance matrix of the moment function. Therefore, the bread is \eqn{(G'WG)^{-1}}, which is the second derivative of the objective function. The method if not yet available for \code{gel} objects. } \value{ A \eqn{k \times k} matrix (see details). } \references{ Zeileis A (2006), Object-oriented Computation of Sandwich Estimators. \emph{Journal of Statistical Software}, \bold{16}(9), 1--16. URL \doi{10.18637/jss.v016.i09}. } \examples{ # See \code{\link{gmm}} for more details on this example. # With the identity matrix # bread is the inverse of (G'G) n <- 1000 x <- rnorm(n, mean = 4, sd = 2) g <- function(tet, x) { m1 <- (tet[1] - x) m2 <- (tet[2]^2 - (x - tet[1])^2) m3 <- x^3 - tet[1]*(tet[1]^2 + 3*tet[2]^2) f <- cbind(m1, m2, m3) return(f) } Dg <- function(tet, x) { jacobian <- matrix(c( 1, 2*(-tet[1]+mean(x)), -3*tet[1]^2-3*tet[2]^2,0, 2*tet[2], -6*tet[1]*tet[2]), nrow=3,ncol=2) return(jacobian) } res <- gmm(g, x, c(0, 0), grad = Dg,weightsMatrix=diag(3)) G <- Dg(res$coef, x) bread(res) solve(crossprod(G)) } gmm/man/smoothG.Rd0000644000176200001440000000672515044417733013536 0ustar liggesusers\name{smoothG} \alias{smoothG} \title{Kernel smoothing of a matrix of time series} \description{ It applies the required kernel smoothing to the moment function in order for the GEL estimator to be valid. It is used by the \code{gel} function.} \usage{ smoothG(x, bw = bwAndrews, prewhite = 1, ar.method = "ols", weights = weightsAndrews, kernel = c("Bartlett", "Parzen", "Truncated", "Tukey-Hanning"), approx = c("AR(1)", "ARMA(1,1)"), tol = 1e-7) } \arguments{ \item{x}{a \eqn{n\times q} matrix of time series, where n is the sample size.} \item{bw}{The method to compute the bandwidth parameter. By default, it uses the bandwidth proposed by Andrews(1991). As an alternative, we can choose bw=bwNeweyWest (without "") which is proposed by Newey-West(1996).} \item{prewhite}{logical or integer. Should the estimating functions be prewhitened? If \code{TRUE} or greater than 0 a VAR model of order \code{as.integer(prewhite)} is fitted via \code{ar} with method \code{"ols"} and \code{demean = FALSE}.} \item{ar.method}{character. The \code{method} argument passed to \code{\link[stats]{ar}} for prewhitening.} \item{weights}{The smoothing weights can be computed by \code{\link[sandwich]{weightsAndrews}} of it can be provided manually. If provided, it has to be a \eqn{r\times 1}vector (see details). } \item{approx}{a character specifying the approximation method if the bandwidth has to be chosen by \code{bwAndrews}.} \item{tol}{numeric. Weights that exceed \code{tol} are used for computing the covariance matrix, all other weights are treated as 0.} \item{kernel}{The choice of kernel} } \details{ The sample moment conditions \eqn{\sum_{t=1}^n g(\theta,x_t)} is replaced by: \eqn{\sum_{t=1}^n g^k(\theta,x_t)}, where \eqn{g^k(\theta,x_t)=\sum_{i=-r}^r k(i) g(\theta,x_{t+i})}, where \eqn{r} is a truncated parameter that depends on the bandwidth and \eqn{k(i)} are normalized weights so that they sum to 1. If the vector of weights is provided, it gives only one side weights. For exemple, if you provide the vector (1,.5,.25), \eqn{k(i)} will become \eqn{(.25,.5,1,.5,.25)/(.25+.5+1+.5+.25) = (.1,.2,.4,.2,.1)} } \value{ smoothx: A \eqn{q \times q} matrix containing an estimator of the asymptotic variance of \eqn{\sqrt{n} \bar{x}}, where \eqn{\bar{x}} is \eqn{q\times 1}vector with typical element \eqn{\bar{x}_i = \frac{1}{n}\sum_{j=1}^nx_{ji}}. This function is called by \code{\link{gel}} but can also be used by itself. \code{kern_weights}: Vector of weights used for the smoothing. } \references{ Anatolyev, S. (2005), GMM, GEL, Serial Correlation, and Asymptotic Bias. \emph{Econometrica}, \bold{73}, 983-1002. Andrews DWK (1991), Heteroskedasticity and Autocorrelation Consistent Covariance Matrix Estimation. \emph{Econometrica}, \bold{59}, 817--858. Kitamura, Yuichi (1997), Empirical Likelihood Methods With Weakly Dependent Processes. \emph{The Annals of Statistics}, \bold{25}, 2084-2102. Zeileis A (2006), Object-oriented Computation of Sandwich Estimators. \emph{Journal of Statistical Software}, \bold{16}(9), 1--16. URL \doi{10.18637/jss.v016.i09}. } \examples{ g <- function(tet, x) { n <- nrow(x) u <- (x[7:n] - tet[1] - tet[2]*x[6:(n-1)] - tet[3]*x[5:(n-2)]) f <- cbind(u, u*x[4:(n-3)], u*x[3:(n-4)], u*x[2:(n-5)], u*x[1:(n-6)]) return(f) } n = 500 phi<-c(.2, .7) thet <- 0.2 sd <- .2 x <- matrix(arima.sim(n = n, list(order = c(2, 0, 1), ar = phi, ma = thet, sd = sd)), ncol = 1) gt <- g(c(0, phi), x) sgt <- smoothG(gt)$smoothx plot(gt[,1]) lines(sgt[,1]) } gmm/man/estfun.Rd0000644000176200001440000000521415044421564013407 0ustar liggesusers\name{estfun} \alias{estfun.gmmFct} \alias{estfun.gmm} \alias{estfun.gel} \alias{estfun.tsls} \alias{model.matrix.tsls} \title{Extracts the empirical moment function} \description{ It extracts the matrix of empirical moments so that it can be used by the \code{\link[sandwich]{kernHAC}} function. } \usage{ \method{estfun}{gmmFct}(x, y = NULL, theta = NULL, ...) \method{estfun}{gmm}(x, ...) \method{estfun}{gel}(x, ...) \method{estfun}{tsls}(x, ...) \method{model.matrix}{tsls}(object, ...) } \arguments{ \item{x}{A function of the form \eqn{g(\theta,y)} or a \eqn{n \times q} matrix with typical element \eqn{g_i(\theta,y_t)} for \eqn{i=1,...q} and \eqn{t=1,...,n} or an object of class \code{gmm}. See \code{\link{gmm}} for more details. For \code{\link{tsls}}, it is an object of class \code{tsls}.} \item{object}{An object of class \code{tsls}.} \item{y}{The matrix or vector of data from which the function \eqn{g(\theta,y)} is computed if \code{g} is a function.} \item{theta}{Vector of parameters if \code{g} is a function.} \item{...}{Other arguments when \code{estfun} is applied to another class object} } \details{ For \code{estfun.gmmFct}, it returns a \eqn{n \times q} matrix with typical element \eqn{g_i(\theta,y_t)} for \eqn{i=1,...q} and \eqn{t=1,...,n}. It is only used by \code{gmm} to obtain the estimates. For \code{estfun.gmm}, it returns the matrix of first order conditions of \eqn{\min_\theta \bar{g}'W\bar{g}/2}, which is a \eqn{n \times k} matrix with the \eqn{t^{th}} row being \eqn{g(\theta, y_t)W G}, where \eqn{G} is \eqn{d\bar{g}/d\theta}. It allows to compute the sandwich covariance matrix using \code{\link[sandwich]{kernHAC}} or \code{\link[sandwich]{vcovHAC}} when \eqn{W} is not the optimal matrix. The method if not yet available for \code{gel} objects. For tsls, \code{model.matrix} and estfun are used by \code{\link{vcov.tsls}} to compute different covariance matrices using the \code{sandwich} package. \code{model.matrix} returns the fitted values from the first stage regression and \code{estfun} the residuals. } \value{ A \eqn{n \times q} matrix (see details). } \references{ Zeileis A (2006), Object-oriented Computation of Sandwich Estimators. \emph{Journal of Statistical Software}, \bold{16}(9), 1--16. URL \doi{10.18637/jss.v016.i09}. } \examples{ n = 500 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n=n,list(order=c(2,0,1),ar=phi,ma=thet,sd=sd)),ncol=1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H res <- gmm(g, x,weightsMatrix = diag(5)) gt <- res$gt G <- res$G foc <- gt%*%G foc2 <- estfun(res) foc[1:5,] foc2[1:5,] } gmm/man/plot.Rd0000644000176200001440000000502615044417655013070 0ustar liggesusers\name{plot} \alias{plot.gel} \alias{plot.gmm} \title{Plot Diagnostics for gel and gmm objects} \description{ It is a plot method for \code{gel} or \code{gmm} objects. } \usage{ \method{plot}{gel}(x, which = c(1L:4), main = list("Residuals vs Fitted values", "Normal Q-Q", "Response variable and fitted values","Implied probabilities"), panel = if(add.smooth) panel.smooth else points, ask = prod(par("mfcol")) < length(which) && dev.interactive(), ..., add.smooth = getOption("add.smooth")) \method{plot}{gmm}(x, which = c(1L:3), main = list("Residuals vs Fitted values", "Normal Q-Q", "Response variable and fitted values"), panel = if(add.smooth) panel.smooth else points, ask = prod(par("mfcol")) < length(which) && dev.interactive(), ..., add.smooth = getOption("add.smooth")) } \arguments{ \item{x}{\code{gel} or \code{gmm} object, typically result of \code{\link{gel}} or \code{\link{gmm}}.} \item{which}{if a subset of the plots is required, specify a subset of the numbers \code{1:4} for \code{gel} or \code{1:3} for \code{gmm}.} \item{main}{Vector of titles for each plot. } \item{panel}{panel function. The useful alternative to \code{\link[graphics]{points}}, \code{\link[graphics]{panel.smooth}} can be chosen by \code{add.smooth = TRUE}.} \item{ask}{logical; if \code{TRUE}, the user is \emph{ask}ed before each plot, see \code{\link[graphics]{par}(ask=.)}.} \item{\dots}{other parameters to be passed through to plotting functions.} \item{add.smooth}{logical indicating if a smoother should be added to most plots; see also \code{panel} above.} } \details{ It is a beta version of a plot method for \code{gel} objects. It is a modified version of \code{plot.lm}. For now, it is available only for linear models expressed as a formula. Any suggestions are welcome regarding plots or options to include. The first two plots are the same as the ones provided by \code{plot.lm}, the third is the dependant variable \eqn{y} with its mean \eqn{\hat{y}} (the fitted values) and the last plots the implied probabilities with the empirical density \eqn{1/T}. } \examples{ # GEL # n = 500 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n = n,list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H t0 <- c(0,.5,.5) res <- gel(g, x, t0) plot(res, which = 3) plot(res, which = 4) # GMM # res <- gmm(g, x) plot(res, which = 3) } gmm/man/sysGmm.Rd0000644000176200001440000001337115044420034013354 0ustar liggesusers\name{sysGmm} \alias{sysGmm} \alias{five} \alias{sur} \alias{randEffect} \alias{threeSLS} \title{Generalized method of moment estimation for system of equations} \description{ Functions to estimate a system of equations based on GMM. } \usage{ sysGmm(g, h, wmatrix = c("optimal","ident"), vcov=c("MDS", "HAC", "CondHom", "TrueFixed"), kernel=c("Quadratic Spectral","Truncated", "Bartlett", "Parzen", "Tukey-Hanning"), crit=10e-7,bw = bwAndrews, prewhite = FALSE, ar.method = "ols", approx="AR(1)", tol = 1e-7, model=TRUE, X=FALSE, Y=FALSE, centeredVcov = TRUE, weightsMatrix = NULL, data, crossEquConst = NULL, commonCoef = FALSE) five(g, h, commonCoef = FALSE, data = NULL) threeSLS(g, h, commonCoef = FALSE, data = NULL) sur(g, commonCoef = FALSE, data = NULL) randEffect(g, data = NULL) } \arguments{ \item{g}{A possibly named list of formulas} \item{h}{A formula if the same instruments are used in each equation or a list of formulas.} \item{wmatrix}{Which weighting matrix should be used in the objective function. By default, it is the inverse of the covariance matrix of \eqn{g(\theta,x)}. The other choice is the identity matrix.} \item{vcov}{Assumption on the properties of the moment vector. By default, it is a martingale difference sequence. "HAC" is for weakly dependent processes and "CondHom" implies conditional homoscedasticity. The option "TrueFixed" is used only when the matrix of weights is provided and it is the optimal one.} \item{kernel}{type of kernel used to compute the covariance matrix of the vector of sample moment conditions (see \code{\link[sandwich]{kernHAC}} for more details)} \item{crit}{The stopping rule for the iterative GMM. It can be reduce to increase the precision.} \item{bw}{The method to compute the bandwidth parameter. By default it is \code{\link[sandwich]{bwAndrews}} which is proposed by Andrews (1991). The alternative is \code{\link[sandwich]{bwNeweyWest}} of Newey-West(1994).} \item{prewhite}{logical or integer. Should the estimating functions be prewhitened? If \code{TRUE} or greater than 0 a VAR model of order \code{as.integer(prewhite)} is fitted via \code{ar} with method \code{"ols"} and \code{demean = FALSE}.} \item{ar.method}{character. The \code{method} argument passed to \code{\link[stats]{ar}} for prewhitening.} \item{approx}{A character specifying the approximation method if the bandwidth has to be chosen by \code{bwAndrews}.} \item{tol}{Weights that exceed \code{tol} are used for computing the covariance matrix, all other weights are treated as 0.} \item{model, X, Y}{logical. If \code{TRUE} the corresponding components of the fit (the model frame, the model matrix, the response) are returned if g is a formula.} \item{centeredVcov}{Should the moment function be centered when computing its covariance matrix. Doing so may improve inference.} \item{weightsMatrix}{It allows users to provide \code{gmm} with a fixed weighting matrix. This matrix must be \eqn{q \times q}, symmetric and strictly positive definite. When provided, the \code{type} option becomes irrelevant. } \item{data}{A data.frame or a matrix with column names (Optional). } \item{commonCoef}{If true, coefficients accross equations are the same} \item{crossEquConst}{Only used if the number of regressors are the same in each equation. It is a vector which indicates which coefficient are constant across equations. The order is 1 for Intercept and 2 to k as it is formulated in the formulas \code{g}. Setting it to \code{1:k} is equivalent to setting commonCoef to TRUE.} } \details{ This set of functions implement the estimation of system of equations as presented in Hayashi (2000) } \value{ 'sysGmm' returns an object of 'class' '"sysGmm"' The functions 'summary' is used to obtain and print a summary of the results. It also compute the J-test of overidentying restriction The object of class "sysGmm" is a list containing at least: \item{coefficients}{list of vectors of coefficients for each equation} \item{residuals}{list of the residuals for each equation.} \item{fitted.values}{list of the fitted values for each equation.} \item{vcov}{the covariance matrix of the stacked coefficients} \item{objective}{the value of the objective function \eqn{\| var(\bar{g})^{-1/2}\bar{g}\|^2}} \item{terms}{The list of \code{\link[stats]{terms}} objects for each equation} \item{call}{the matched call.} \item{y}{If requested, a list of response variables.} \item{x}{if requested, a list of the model matrices.} \item{model}{if requested (the default), a list of the model frames.} } \references{ Zeileis A (2006), Object-oriented Computation of Sandwich Estimators. \emph{Journal of Statistical Software}, \bold{16}(9), 1--16. URL \doi{10.18637/jss.v016.i09}. Andrews DWK (1991), Heteroskedasticity and Autocorrelation Consistent Covariance Matrix Estimation. \emph{Econometrica}, \bold{59}, 817--858. Newey WK & West KD (1987), A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix. \emph{Econometrica}, \bold{55}, 703--708. Newey WK & West KD (1994), Automatic Lag Selection in Covariance Matrix Estimation. \emph{Review of Economic Studies}, \bold{61}, 631-653. Hayashi, F. (2000), Econometrics. \emph{Princeton University Press}. } \examples{ data(wage) eq1 <- LW~S+IQ+EXPR eq2 <- LW80~S80+IQ+EXPR80 g2 <- list(Wage69=eq1, WAGE80=eq2) h2 <- list(~S+EXPR+MED+KWW, ~S80+EXPR80+MED+KWW) res <- sysGmm(g2, h2, data=wage, commonCoef=TRUE) summary(res) res2 <- sysGmm(g2, h2, data=wage) summary(res2) five(g2, h2, data=wage) threeSLS(g2, h2[[1]], data=wage) sur(g2, data=wage) randEffect(g2, data=wage) ## Cross-Equation restrictions ## All but the intercept are assumed to be the same res <- sysGmm(g2, h2, data=wage, crossEquConst = 2:4) summary(res) } gmm/man/getDat.Rd0000644000176200001440000000275614247643115013325 0ustar liggesusers\name{getDat} \alias{getDat} \title{Extracting data from a formula} \description{ It extract the data from a formula y~z with instrument h and put everything in a matrix. It helps redefine the function \eqn{g(\theta,x)} that is required by \code{\link{gmm}} and \code{\link{gel}}. } \usage{ getDat(formula, h, data, error=TRUE) } \arguments{ \item{formula}{A formula that defines the linear model to be estimated (see details).} \item{h}{A \eqn{n\times nh} matrix of intruments(see details).} \item{data}{A data.frame or a matrix with colnames (Optionnal).} \item{error}{If FALSE, the data is generated without giving any error message} } \details{The model to be estimated is based on the moment conditions \eqn{=0}. It adds a column of ones to z and h by default. They are removed if -1 is added to the formula. The error argument has been added for \code{\link{sysGmm}} with common coefficients because the check is only valid for equation by equation identification. } \value{ x: A \eqn{n \times l} matrix, where \eqn{l = ncol(y)+ncol(z)+ncol(h)+2} if "intercept" is TRUE and \eqn{ncol(y)+ncol(z)+xcol(h)} if "intercept" is FALSE. nh: dimension of h k: dimension of z ny: dimension of y } \examples{ n = 500 phi<-c(.2, .7) thet <- 0.2 sd <- .2 x <- matrix(arima.sim(n = n, list(order = c(2, 0, 1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) x <- getDat(y ~ ym1 + ym2, H) } gmm/man/fitted.Rd0000644000176200001440000000266614247643115013374 0ustar liggesusers\name{fitted} \alias{fitted.gel} \alias{fitted.gmm} \title{Fitted values of GEL and GMM} \description{ Method to extract the fitted values of the model estimated by \code{\link{gel}} or \code{\link{gmm}}. } \usage{ \method{fitted}{gel}(object, ...) \method{fitted}{gmm}(object, ...) } \arguments{ \item{object}{An object of class \code{gel} or \code{gel} returned by the function \code{\link{gel}} or \code{\link{gmm}}} \item{...}{Other arguments when \code{fitted} is applied to an other class object} } \value{ It returns a matrix of the estimated mean \eqn{\hat{y}} in \code{g=y~x} as it is done by \code{fitted.lm}. } \examples{ # GEL can deal with endogeneity problems n = 200 phi<-c(.2,.7) thet <- 0.2 sd <- .2 set.seed(123) x <- matrix(arima.sim(n = n, list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- x[7:n] ym1 <- x[6:(n-1)] ym2 <- x[5:(n-2)] H <- cbind(x[4:(n-3)], x[3:(n-4)], x[2:(n-5)], x[1:(n-6)]) g <- y ~ ym1 + ym2 x <- H res <- gel(g, x, c(0,.3,.6)) plot(y, main = "Fitted ARMA with GEL") lines(fitted(res), col = 2) # GMM is like GLS for linear models without endogeneity problems set.seed(345) n = 200 phi<-c(.2,.7) thet <- 0 sd <- .2 x <- matrix(arima.sim(n = n, list(order = c(2,0,1), ar = phi, ma = thet, sd = sd)), ncol = 1) y <- 10 + 5*rnorm(n) + x res <- gmm(y ~ x, x) plot(x, y, main = "Fitted model with GMM") lines(x, fitted(res), col = 2) legend("topright", c("Y","Yhat"), col = 1:2, lty = c(1,1)) } gmm/man/KTest.Rd0000644000176200001440000000345014247643115013137 0ustar liggesusers\name{KTest} \alias{KTest} \alias{print.gmmTests} \title{Compute the K statistics of Kleibergen} \description{The test is proposed by Kleibergen (2005). It is robust to weak identification. } \usage{ KTest(obj, theta0 = NULL, alphaK = 0.04, alphaJ = 0.01) \method{print}{gmmTests}(x, digits = 5, ...) } \arguments{ \item{obj}{Object of class "gmm" returned by \code{\link{gmm}}} \item{theta0}{The null hypothesis being tested. See details.} \item{alphaK, alphaJ}{The size of the J and K tests when combining the two. The overall size is alphaK+alphaJ.} \item{x}{An object of class \code{gmmTests} returned by \code{KTest}} \item{digits}{The number of digits to be printed} \item{...}{Other arguments when \code{print} is applied to another class object} } \details{ The function produces the J-test and K-statistics which are robust to weak identification. The test is either \eqn{H0:\theta=theta_0}, in which case theta0 must be provided, or \eqn{\beta=\beta_0}, where \eqn{\theta=(\alpha', \beta')'}, and \eqn{\alpha} is assumed to be identified. In the latter case, theta0 is NULL and obj is a restricted estimation in which \eqn{\beta} is fixed to \eqn{\beta_0}. See \code{\link{gmm}} and the option "eqConst" for more details. } \value{ Tests and p-values } \references{ Keibergen, F. (2005), Testing Parameters in GMM without assuming that they are identified. \emph{Econometrica}, \bold{73}, 1103-1123, } \examples{ library(mvtnorm) sig <- matrix(c(1,.5,.5,1),2,2) n <- 400 e <- rmvnorm(n,sigma=sig) x4 <- rnorm(n) w <- exp(-x4^2) + e[,1] y <- 0.1*w + e[,2] h <- cbind(x4, x4^2, x4^3, x4^6) g3 <- y~w res <- gmm(g3,h) # Testing the whole vector: KTest(res,theta0=c(0,.1)) # Testing a subset of the vector (See \code{\link{gmm}}) res2 <- gmm(g3, h, eqConst=matrix(c(2,.1),1,2)) res2 KTest(res2) } gmm/man/charStable.Rd0000644000176200001440000000324014247643115014152 0ustar liggesusers\name{charStable} \alias{charStable} \title{The characteristic function of a stable distribution} \description{ It computes the theoretical characteristic function of a stable distribution for two different parametrizations. It is used in the vignette to illustrate the estimation of the parameters using GMM.} \usage{ charStable(theta, tau, pm = 0) } \arguments{ \item{theta}{Vector of parameters of the stable distribution. See details.} \item{tau}{A vector of numbers at which the function is evaluated.} \item{pm}{The type of parametization. It takes the values 0 or 1.} } \value{ It returns a vector of complex numbers with the dimension equals to \code{length(tau)}. } \details{ The function returns the vector \eqn{\Psi(\theta,\tau,pm)} defined as \eqn{E(e^{ix\tau}}, where \eqn{\tau} is a vector of real numbers, \eqn{i} is the imaginary number, \eqn{x} is a stable random variable with parameters \eqn{\theta} = \eqn{(\alpha,\beta,\gamma,\delta)} and \code{pm} is the type of parametrization. The vector of parameters are the characteristic exponent, the skewness, the scale and the location parameters, respectively. The restrictions on the parameters are: \eqn{\alpha \in (0,2]}, \eqn{\beta\in [-1,1]} and \eqn{\gamma>0}. For mode details see Nolan(2009). } \references{ Nolan J. P. (2020), Univariate Stable Distributions - Models for Heavy Tailed Data. \emph{Springer Series in Operations Research and Financial Engineering}. URL \url{https://edspace.american.edu/jpnolan/stable/}. } \examples{ # GMM is like GLS for linear models without endogeneity problems pm <- 0 theta <- c(1.5,.5,1,0) tau <- seq(-3, 3, length.out = 20) char_fct <- charStable(theta, tau, pm) } gmm/man/tsls.Rd0000644000176200001440000000501715044420075013065 0ustar liggesusers\name{tsls} \alias{tsls} \title{Two stage least squares estimation} \description{ Function to estimate a linear model by the two stage least squares method. } \usage{ tsls(g,x,data) } \arguments{ \item{g}{A formula describing the linear regression model (see details below).} \item{x}{The matrix of instruments (see details below).} \item{data}{A data.frame or a matrix with column names (Optionnal). } } \details{ The function just calls \code{\link{gmm}} with the option vcov="iid". It just simplifies the the implementation of 2SLS. The users don't have to worry about all the options offered in \code{\link{gmm}}. The model is \deqn{ Y_i = X_i\beta + u_i } In the first step, \code{\link[stats]{lm}} is used to regress \eqn{X_i} on the set of instruments \eqn{Z_i}. The second step also uses \code{\link[stats]{lm}} to regress \eqn{Y_i} on the fitted values of the first step. } \value{ 'tsls' returns an object of 'class' '"tsls"' which inherits from class '"gmm"'. The functions 'summary' is used to obtain and print a summary of the results. It also compute the J-test of overidentying restriction The object of class "gmm" is a list containing at least: \item{coefficients}{\eqn{k\times 1} vector of coefficients} \item{residuals}{the residuals, that is response minus fitted values if "g" is a formula.} \item{fitted.values}{the fitted mean values if "g" is a formula.} \item{vcov}{the covariance matrix of the coefficients} \item{objective}{the value of the objective function \eqn{\| var(\bar{g})^{-1/2}\bar{g}\|^2}} \item{terms}{the \code{\link[stats]{terms}} object used when g is a formula.} \item{call}{the matched call.} \item{y}{if requested, the response used (if "g" is a formula).} \item{x}{if requested, the model matrix used if "g" is a formula or the data if "g" is a function.} \item{model}{if requested (the default), the model frame used if "g" is a formula.} \item{algoInfo}{Information produced by either \code{\link[stats]{optim}} or \code{\link[stats]{nlminb}} related to the convergence if "g" is a function. It is printed by the \code{summary.gmm} method.} } \references{ Hansen, L.P. (1982), Large Sample Properties of Generalized Method of Moments Estimators. \emph{Econometrica}, \bold{50}, 1029-1054, } \examples{ n <- 1000 e <- arima.sim(n,model=list(ma=.9)) C <- runif(n,0,5) Y <- rep(0,n) Y[1] = 1 + 2*C[1] + e[1] for (i in 2:n){ Y[i] = 1 + 2*C[i] + 0.9*Y[i-1] + e[i] } Yt <- Y[5:n] X <- cbind(C[5:n],Y[4:(n-1)]) Z <- cbind(C[5:n],Y[3:(n-2)],Y[2:(n-3)],Y[1:(n-4)]) res <- tsls(Yt~X,~Z) res } gmm/DESCRIPTION0000644000176200001440000000232715053352521012545 0ustar liggesusersPackage: gmm Version: 1.9-1 Date: 2025-08-4 Title: Generalized Method of Moments and Generalized Empirical Likelihood Authors@R: person(given="Pierre", family="Chausse", role=c("aut","cre"), email="pchausse@uwaterloo.ca") Author: Pierre Chausse [aut, cre] Maintainer: Pierre Chausse Description: It is a complete suite to estimate models based on moment conditions. It includes the two step Generalized method of moments (Hansen 1982; ), the iterated GMM and continuous updated estimator (Hansen, Eaton and Yaron 1996; ) and several methods that belong to the Generalized Empirical Likelihood family of estimators (Smith 1997; , Kitamura 1997; , Newey and Smith 2004; , and Anatolyev 2005 ). Depends: R (>= 2.10.0), sandwich NeedsCompilation: yes Suggests: knitr, mvtnorm, car, stabledist, MASS, timeDate, timeSeries Imports: stats, methods, grDevices, graphics License: GPL (>= 2) VignetteBuilder: knitr Packaged: 2025-08-05 15:41:49 UTC; pierrechausse Repository: CRAN Date/Publication: 2025-08-26 15:40:01 UTC