vcdExtra/0000755000176200001440000000000013163540741012036 5ustar liggesusersvcdExtra/inst/0000755000176200001440000000000013163461153013012 5ustar liggesusersvcdExtra/inst/doc/0000755000176200001440000000000013163514377013566 5ustar liggesusersvcdExtra/inst/doc/vcd-tutorial.Rnw0000644000176200001440000023435213163461153016675 0ustar liggesusers% !Rnw weave = Sweave %\VignetteEngine{Sweave} %\VignetteIndexEntry{Tutorial: Working with categorical data with R and the vcd package} %\VignetteDepends{vcd,gmodels,ca} %\VignetteKeywords{contingency tables, mosaic plots, sieve plots, categorical data, independence, conditional independence, R} %\VignettePackage{vcdExtra} \documentclass[10pt,twoside]{article} \usepackage{Sweave} \usepackage{bm} \usepackage[toc]{multitoc} % for table of contents % from Z.cls \usepackage[authoryear,round,longnamesfirst]{natbib} \bibpunct{(}{)}{;}{a}{}{,} \bibliographystyle{jss} \usepackage{hyperref} \usepackage{color} %% colors \definecolor{Red}{rgb}{0.7,0,0} \definecolor{Blue}{rgb}{0,0,0.8} \hypersetup{% hyperindex = {true}, colorlinks = {true}, % linktocpage = {true}, plainpages = {false}, linkcolor = {Blue}, citecolor = {Blue}, urlcolor = {Red}, pdfstartview = {Fit}, pdfpagemode = {UseOutlines}, pdfview = {XYZ null null null} } %\AtBeginDocument{ % \hypersetup{% % pdfauthor = {Michael Friendly}, % pdftitle = {Tutorial: Working with categorical data with R and the vcd package}, % pdfkeywords = {contingency tables, mosaic plots, sieve plots, categorical data, independence, conditional independence, R} % } %} % math stuff \newcommand*{\given}{\ensuremath{\, | \,}} \renewcommand*{\vec}[1]{\ensuremath{\bm{#1}}} \newcommand{\mat}[1]{\ensuremath{\bm{#1}}} \newcommand{\trans}{\ensuremath{^\mathsf{T}}} \newcommand{\diag}[1]{\ensuremath{\mathrm{diag} (#1)}} \def\binom#1#2{{#1 \choose #2}}% \newcommand{\implies}{ \ensuremath{\mapsto} } \newenvironment{equation*}{\displaymath}{\enddisplaymath}% \newcommand{\tabref}[1]{Table~\ref{#1}} \newcommand{\figref}[1]{Figure~\ref{#1}} \newcommand{\secref}[1]{Section~\ref{#1}} \newcommand{\loglin}{loglinear } %\usepackage{thumbpdf} % page dimensions \addtolength{\hoffset}{-1.5cm} \addtolength{\textwidth}{3cm} \addtolength{\voffset}{-1cm} \addtolength{\textheight}{2cm} % Vignette examples \newcommand*{\Example}{\fbox{\textbf{\emph{Example}}:} } % R stuff \newcommand{\var}[1]{\textit{\texttt{#1}}} \newcommand{\data}[1]{\texttt{#1}} \newcommand{\class}[1]{\textsf{"#1"}} %% \code without `-' ligatures \def\nohyphenation{\hyphenchar\font=-1 \aftergroup\restorehyphenation} \def\restorehyphenation{\hyphenchar\font=`-} {\catcode`\-=\active% \global\def\code{\bgroup% \catcode`\-=\active \let-\codedash% \Rd@code}} \def\codedash{-\discretionary{}{}{}} \def\Rd@code#1{\texttt{\nohyphenation#1}\egroup} \newcommand{\codefun}[1]{\code{#1()}} \let\proglang=\textsf \newcommand{\pkg}[1]{{\normalfont\fontseries{b}\selectfont #1}} \newcommand{\Rpackage}[1]{{\textsf{#1}}} %% almost as usual \author{Michael Friendly\\York University, Toronto} \title{Working with categorical data with \proglang{R} and the \pkg{vcd} and \pkg{vcdExtra} packages} \date{\footnotesize{Using \Rpackage{vcdExtra} version \Sexpr{packageDescription("vcdExtra")[["Version"]]} and \Rpackage{vcd} version \Sexpr{packageDescription("vcd")[["Version"]]}; Date: \Sexpr{Sys.Date()}}} %% for pretty printing and a nice hypersummary also set: %\Plainauthor{Michael Friendly} %% comma-separated %\Shorttitle{vcd tutorial} %% a short title (if necessary) %\Plaintitle{Tutorial: Working with categorical data with R and the vcd package} %\SweaveOpts{engine=R,eps=TRUE,height=6,width=7,results=hide,fig=FALSE,echo=TRUE} \SweaveOpts{engine=R,height=6,width=7,results=hide,fig=FALSE,echo=TRUE} \SweaveOpts{prefix.string=fig/vcd-tut,eps=FALSE} \SweaveOpts{keep.source=TRUE} %\SweaveOpts{concordance=TRUE} \setkeys{Gin}{width=0.7\textwidth} <>= set.seed(1071) #library(vcd) library(vcdExtra) library(ggplot2) #data(Titanic) data(HairEyeColor) data(PreSex) data(Arthritis) art <- xtabs(~Treatment + Improved, data = Arthritis) if(!file.exists("fig")) dir.create("fig") @ %% end of declarations %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{document} \SweaveOpts{concordance=TRUE} \maketitle %% an abstract and keywords \begin{abstract} This tutorial describes the creation and manipulation of frequency and contingency tables from categorical variables, along with tests of independence, measures of association, and methods for graphically displaying results. The framework is provided by the \proglang{R} package \pkg{vcd}, but other packages are used to help with various tasks. The \pkg{vcdExtra} package extends the graphical and statistical methods provided by \pkg{vcd}. This package is now the main support package for the book \emph{Discrete Data Analysis with R: Visualizing and Modeling Techniques for Categorical and Count Data} \citep{FriendlyMeyer:2016:DDAR}. The web page for the book, \href{http://ddar.datavis.ca}{ddar.datavis.ca}, gives further details. \end{abstract} %\keywords{contingency tables, mosaic plots, sieve plots, %categorical data, independence, conditional independence, generalized linear models, %\proglang{R}} %\Plainkeywords{contingency tables, mosaic plots, % sieve plots, categorical data, independence, % conditional independence, generalized linear models, R} {\small % \sloppy % \begin{multicols}{2} \tableofcontents % \end{multicols} } \section[Introduction]{Introduction}\label{sec:intro} %% Note: If there is markup in \(sub)section, then it has to be escape as above. This tutorial, part of the \pkg{vcdExtra} package, describes how to work with categorical data in the context of fitting statistical models in \proglang{R} and visualizing the results using the \pkg{vcd} and \pkg{vcdExtra} packages. It focuses first on methods and tools for creating and manipulating \proglang{R} data objects which represent frequency and contingency tables involving categorical variables. Further sections describe some simple methods for calculating tests of independence and measures of association amomg categorial variables, and also methods for graphically displaying results. There is much more to the analysis of categorical data than is described here, where the emphasis is on cross-tabulated tables of frequencies (``contingency tables''), statistical tests, associated \loglin\ models, and visualization of \emph{how} variables are related. A more general treatment of graphical methods for categorical data is contained in the book, \emph{Discrete Data Analysis with R: Visualizing and Modeling Techniques for Categorical and Count Data} \citep{FriendlyMeyer:2016:DDAR}. An earlier book using SAS is \emph{Visualizing Categorical Data} \citep{vcd:Friendly:2000}, for which \pkg{vcd} is a partial \proglang{R} companion, covering topics not otherwise available in \proglang{R}. On the other hand, the implementation of graphical methods in \pkg{vcd} is more general in many respects than what I provided in \proglang{SAS}. Statistical models for categorical data in \proglang{R} have been extended considerably with the \pkg{gnm} package for generalized \emph{nonlinear} models. The \pkg{vcdExtra} package extends \pkg{vcd} methods to models fit using \codefun{glm} and \codefun{gnm}. A more complete theoretical description of these statistical methods is provided in Agresti's \citeyearpar{vcd:Agresti:2002,Agresti:2013} \emph{Categorical Data Analysis}. For this, see the \proglang{Splus/R} companion by Laura Thompson, \url{http://www.stat.ufl.edu/~aa/cda/Thompson_manual.pdf} and Agresti's support web page, \url{http://www.stat.ufl.edu/~aa/cda/cda.html}. \section[Creating frequency tables]{Creating and manipulating frequency tables}\label{sec:creating} \proglang{R} provides many methods for creating frequency and contingency tables. Several are described below. In the examples below, we use some real examples and some anonymous ones, where the variables \code{A}, \code{B}, and \code{C} represent categorical variables, and \code{X} represents an arbitrary \proglang{R} data object. The first thing you need to know is that categorical data can be represented in three different forms in \proglang{R}, and it is sometimes necessary to convert from one form to another, for carrying out statistical tests, fitting models or visualizing the results. Once a data object exists in \proglang{R}, you can examine its complete structure with the \codefun{str} function, or view the names of its components with the \codefun{names} function. \begin{description} \item[case form] a data frame containing individual observations, with one or more factors, used as the classifying variables. In case form, there may also be numeric covariates. The total number of observations is \code{nrow(X)}, and the number of variables is \code{ncol(X)}. \Example The \data{Arthritis} data is available in case form in the \pkg{vcd} package. There are two explanatory factors: \code{Treatment} and \code{Sex}. \code{Age} is a numeric covariate, and \code{Improved} is the response--- an ordered factor, with levels \code{\Sexpr{paste(levels(Arthritis$Improved),collapse=' < ')}}. Excluding \code{Age}, we would have a $2 \times 2 \times 3$ contingency table for \code{Treatment}, \code{Sex} and \code{Improved}. %\code{"None" < "Some" < "Marked"}. <>= names(Arthritis) # show the variables str(Arthritis) # show the structure head(Arthritis,5) # first 5 observations, same as Arthritis[1:5,] @ \item[frequency form] a data frame containing one or more factors, and a frequency variable, often called \code{Freq} or \code{count}. The total number of observations is \verb|sum(X$Freq)|, \code{sum(X[,"Freq"])} or some equivalent form. The number of cells in the table is \code{nrow(X)}. \Example For small frequency tables, it is often convenient to enter them in frequency form using \codefun{expand.grid} for the factors and \codefun{c} to list the counts in a vector. The example below, from \cite{vcd:Agresti:2002} gives results for the 1991 General Social Survey, with respondents classified by sex and party identification. <>= # Agresti (2002), table 3.11, p. 106 GSS <- data.frame( expand.grid(sex=c("female", "male"), party=c("dem", "indep", "rep")), count=c(279,165,73,47,225,191)) GSS names(GSS) str(GSS) sum(GSS$count) @ \item[table form] a matrix, array or table object, whose elements are the frequencies in an $n$-way table. The variable names (factors) and their levels are given by \code{dimnames(X)}. The total number of observations is \code{sum(X)}. The number of dimensions of the table is \code{length(dimnames(X))}, and the table sizes are given by \code{sapply(dimnames(X), length)}. \Example The \data{HairEyeColor} is stored in table form in \pkg{vcd}. <>= str(HairEyeColor) # show the structure sum(HairEyeColor) # number of cases sapply(dimnames(HairEyeColor), length) # table dimension sizes @ \Example Enter frequencies in a matrix, and assign \code{dimnames}, giving the variable names and category labels. Note that, by default, \codefun{matrix} uses the elements supplied by \emph{columns} in the result, unless you specify \code{byrow=TRUE}. <>= ## A 4 x 4 table Agresti (2002, Table 2.8, p. 57) Job Satisfaction JobSat <- matrix(c(1,2,1,0, 3,3,6,1, 10,10,14,9, 6,7,12,11), 4, 4) dimnames(JobSat) = list(income=c("< 15k", "15-25k", "25-40k", "> 40k"), satisfaction=c("VeryD", "LittleD", "ModerateS", "VeryS")) JobSat @ \data{JobSat} is a matrix, not an object of \code{class("table")}, and some functions are happier with tables than matrices. You can coerce it to a table with \codefun{as.table}, <>= JobSat <- as.table(JobSat) str(JobSat) @ \end{description} \subsection[Ordered factors]{Ordered factors and reordered tables}\label{sec:ordered-factors} In table form, the values of the table factors are ordered by their position in the table. Thus in the \data{JobSat} data, both \code{income} and \code{satisfaction} represent ordered factors, and the \emph{positions} of the values in the rows and columns reflects their ordered nature. Yet, for analysis, there are time when you need \emph{numeric} values for the levels of ordered factors in a table, e.g., to treat a factor as a quantitative variable. In such cases, you can simply re-assign the \code{dimnames} attribute of the table variables. For example, here, we assign numeric values to \code{income} as the middle of their ranges, and treat \code{satisfaction} as equally spaced with integer scores. <>= dimnames(JobSat)$income<-c(7.5,20,32.5,60) dimnames(JobSat)$satisfaction<-1:4 @ For the \data{HairEyeColor} data, hair color and eye color are ordered arbitrarily. For visualizing the data using mosaic plots and other methods described below, it turns out to be more useful to assure that both hair color and eye color are ordered from dark to light. Hair colors are actually ordered this way already, and it is easiest to re-order eye colors by indexing. Again \codefun{str} is your friend. <>= HairEyeColor <- HairEyeColor[, c(1,3,4,2), ] str(HairEyeColor) @ This is also the order for both hair color and eye color shown in the result of a correspondence analysis (\figref{fig:ca-haireye}) below. With data in case form or frequency form, when you have ordered factors represented with character values, you must ensure that they are treated as ordered in \proglang{R}.% \footnote{In \proglang{SAS}, many procedures offer the option \code{order = data | internal | formatted} to allow character values to be ordered according to (a) their order in the data set, (b) sorted internal value, or (c) sorted formatted representation provided by a \proglang{SAS} format. } Imagine that the \data{Arthritis} data was read from a text file. By default the \code{Improved} will be ordered alphabetically: \code{Marked}, \code{None}, \code{Some}--- not what we want. In this case, the function \codefun{ordered} (and others) can be useful. <>= Arthritis <- read.csv("arthritis.txt",header=TRUE) Arthritis$Improved <- ordered(Arthritis$Improved, levels=c("None", "Some", "Marked")) @ With this order of \code{Improved}, the response in this data, a mosaic display of \code{Treatment} and \code{Improved} (\figref{fig:arthritis})shows a clearly interpretable pattern. <>= mosaic(art, gp = shading_max, split_vertical = TRUE, main="Arthritis: [Treatment] [Improved]") @ %\setkeys{Gin}{width=0.7\textwidth} \begin{figure}[htb] \begin{center} %<>= %mosaic(art, gp = shading_max, split_vertical = TRUE, main="Arthritis: [Treatment] [Improved]") %@ \includegraphics[width=0.7\textwidth]{fig/vcd-tut-Arthritis} \caption{Mosaic plot for the \data{Arthritis} data, showing the marginal model of independence for Treatment and Improved. Age, a covariate, and Sex are ignored here.} \label{fig:arthritis} \end{center} \end{figure} Finally, there are situations where, particularly for display purposes, you want to re-order the \emph{dimensions} of an $n$-way table, or change the labels for the variables or levels. This is easy when the data are in table form: \codefun{aperm} permutes the dimensions, and assigning to \code{names} and \code{dimnames} changes variable names and level labels respectively. We will use the following version of \data{UCBAdmissions} in \secref{sec:mantel} below.% \footnote{ Changing \code{Admit} to \code{Admit?} might be useful for display purposes, but is dangerous--- because it is then difficult to use that variable name in a model formula. See \secref{sec:tips} for options \code{labeling\_args} and \code{set\_labels} to change variable and level names for displays in the \code{strucplot} framework. } <>= UCB <- aperm(UCBAdmissions, c(2, 1, 3)) dimnames(UCB)[[2]] <- c("Yes", "No") names(dimnames(UCB)) <- c("Sex", "Admit?", "Department") ftable(UCB) @ %There is one subtle ``gotcha'' here: \codefun{aperm} returns an object of class \class{"array"}, %whereas \data{UCBAdmissions} is of class \class{"table"}, so methods defined for \code{table} %objects will not work on the permuted array. %The solution is to reassign the \code{class} of the result of \codefun{aperm}. % %<>= %class(UCBAdmissions) %class(UCB) %str(as.data.frame(UCBAdmissions)) # OK %str(as.data.frame(UCB)) # wrong % %class(UCB) <- "table" %str(as.data.frame(UCB)) # now OK %@ % \subsection[structable()]{\codefun{structable}}\label{sec:structable} For 3-way and larger tables the \codefun{structable} function in \pkg{vcd} provides a convenient and flexible tabular display. The variables assigned to the rows and columns of a two-way display can be specified by a model formula. <>= structable(HairEyeColor) # show the table: default structable(Hair+Sex ~ Eye, HairEyeColor) # specify col ~ row variables @ It also returns an object of class \code{"structable"} which may be plotted with \codefun{mosaic} (not shown here). <>= HSE < - structable(Hair+Sex ~ Eye, HairEyeColor) # save structable object mosaic(HSE) # plot it @ \subsection[table() and friends]{\codefun{table} and friends}\label{sec:table} You can generate frequency tables from factor variables using the \codefun{table} function, tables of proportions using the \codefun{prop.table} function, and marginal frequencies using \codefun{margin.table}. <>= n=500 A <- factor(sample(c("a1","a2"), n, rep=TRUE)) B <- factor(sample(c("b1","b2"), n, rep=TRUE)) C <- factor(sample(c("c1","c2"), n, rep=TRUE)) mydata <- data.frame(A,B,C) @ <>= # 2-Way Frequency Table attach(mydata) mytable <- table(A,B) # A will be rows, B will be columns mytable # print table margin.table(mytable, 1) # A frequencies (summed over B) margin.table(mytable, 2) # B frequencies (summed over A) prop.table(mytable) # cell percentages prop.table(mytable, 1) # row percentages prop.table(mytable, 2) # column percentages @ \codefun{table} can also generate multidimensional tables based on 3 or more categorical variables. In this case, use the \codefun{ftable} or \codefun{structable} function to print the results more attractively. <>= # 3-Way Frequency Table mytable <- table(A, B, C) ftable(mytable) @ \codefun{table} ignores missing values by default. To include \code{NA} as a category in counts, include the table option \code{exclude=NULL} if the variable is a vector. If the variable is a factor you have to create a new factor using \code{newfactor <- factor(oldfactor, exclude=NULL)}. \subsection[xtabs()]{\codefun{xtabs}}\label{sec:xtabs} The \codefun{xtabs} function allows you to create crosstabulations of data using formula style input. This typically works with case-form data supplied in a data frame or a matrix. The result is a contingency table in array format, whose dimensions are determined by the terms on the right side of the formula. <>= # 3-Way Frequency Table mytable <- xtabs(~A+B+C, data=mydata) ftable(mytable) # print table summary(mytable) # chi-square test of indepedence @ If a variable is included on the left side of the formula, it is assumed to be a vector of frequencies (useful if the data have already been tabulated in frequency form). <>= (GSStab <- xtabs(count ~ sex + party, data=GSS)) summary(GSStab) @ \subsection[Collapsing over factors]{Collapsing over table factors: \codefun{aggregate}, \codefun{margin.table} and \codefun{apply}} It sometimes happens that we have a data set with more variables or factors than we want to analyse, or else, having done some initial analyses, we decide that certain factors are not important, and so should be excluded from graphic displays by collapsing (summing) over them. For example, mosaic plots and fourfold displays are often simpler to construct from versions of the data collapsed over the factors which are not shown in the plots. The appropriate tools to use again depend on the form in which the data are represented--- a case-form data frame, a frequency-form data frame (\codefun{aggregate}), or a table-form array or table object (\codefun{margin.table} or \codefun{apply}). When the data are in frequency form, and we want to produce another frequency data frame, \codefun{aggregate} is a handy tool, using the argument \code{FUN=sum} to sum the frequency variable over the factors \emph{not} mentioned in the formula. \Example The data frame \data{DaytonSurvey} in the \pkg{vcdExtra} package represents a $2^5$ table giving the frequencies of reported use (``ever used?'') of alcohol, cigarettes and marijuana in a sample of high school seniors, also classified by sex and race. <>= str(DaytonSurvey) head(DaytonSurvey) @ To focus on the associations among the substances, we want to collapse over sex and race. The right-hand side of the formula used in the call to \codefun{aggregate} gives the factors to be retained in the new frequency data frame, \code{Dayton.ACM.df}. <>= # data in frequency form # collapse over sex and race Dayton.ACM.df <- aggregate(Freq ~ cigarette+alcohol+marijuana, data=DaytonSurvey, FUN=sum) Dayton.ACM.df @ When the data are in table form, and we want to produce another table, \codefun{apply} with \code{FUN=sum} can be used in a similar way to sum the table over dimensions not mentioned in the \code{MARGIN} argument. \codefun{margin.table} is just a wrapper for \codefun{apply} using the \codefun{sum} function. \Example To illustrate, we first convert the \data{DaytonSurvey} to a 5-way table using \codefun{xtabs}, giving \code{Dayton.tab}. <>== # in table form Dayton.tab <- xtabs(Freq~cigarette+alcohol+marijuana+sex+race, data=DaytonSurvey) structable(cigarette+alcohol+marijuana ~ sex+race, data=Dayton.tab) @ Then, use \codefun{apply} on \code{Dayton.tab} to give the 3-way table \code{Dayton.ACM.tab} summed over sex and race. The elements in this new table are the column sums for \code{Dayton.tab} shown by \codefun{structable} just above. <>== # collapse over sex and race Dayton.ACM.tab <- apply(Dayton.tab, MARGIN=1:3, FUN=sum) Dayton.ACM.tab <- margin.table(Dayton.tab, 1:3) # same result structable(cigarette+alcohol ~ marijuana, data=Dayton.ACM.tab) @ Many of these operations can be performed using the \verb|**ply()| functions in the \pkg{plyr} package. For example, with the data in a frequency form data frame, use \codefun{ddply} to collapse over unmentioned factors, and \codefun{plyr::summarise}% \footnote{ Ugh. This \pkg{plyr} function clashes with a function of the same name in \pkg{vcdExtra}. In this document I will use the explicit double-colon notation to keep them separate. } as the function to be applied to each piece. <>== Dayton.ACM.df <- ddply(DaytonSurvey, .(cigarette, alcohol, marijuana), plyr::summarise, Freq=sum(Freq)) @ \subsection[Collapsing levels]{Collapsing table levels: \codefun{collapse.table}} A related problem arises when we have a table or array and for some purpose we want to reduce the number of levels of some factors by summing subsets of the frequencies. For example, we may have initially coded Age in 10-year intervals, and decide that, either for analysis or display purposes, we want to reduce Age to 20-year intervals. The \codefun{collapse.table} function in \pkg{vcdExtra} was designed for this purpose. \Example Create a 3-way table, and collapse Age from 10-year to 20-year intervals. First, we generate a $2 \times 6 \times 3$ table of random counts from a Poisson distribution with mean of 100. <>= # create some sample data in frequency form sex <- c("Male", "Female") age <- c("10-19", "20-29", "30-39", "40-49", "50-59", "60-69") education <- c("low", 'med', 'high') data <- expand.grid(sex=sex, age=age, education=education) counts <- rpois(36, 100) # random Possion cell frequencies data <- cbind(data, counts) # make it into a 3-way table t1 <- xtabs(counts ~ sex + age + education, data=data) structable(t1) @ Now collapse \code{age} to 20-year intervals, and \code{education} to 2 levels. In the arguments, levels of \code{age} and \code{education} given the same label are summed in the resulting smaller table. <>= # collapse age to 3 levels, education to 2 levels t2 <- collapse.table(t1, age=c("10-29", "10-29", "30-49", "30-49", "50-69", "50-69"), education=c(">= as.data.frame(GSStab) @ \Example Convert the \code{Arthritis} data in case form to a 3-way table of \code{Treatment} $\times$ \code{Sex} $\times$ \code{Improved}. Note the use of \codefun{with} to avoid having to use \code{Arthritis\$Treatment} etc. within the call to \codefun{table}.% \footnote{ \codefun{table} does not allow a \code{data} argument to provide an environment in which the table variables are to be found. In the examples in \secref{sec:table} I used \code{attach(mydata)} for this purpose, but \codefun{attach} leaves the variables in the global environment, while \codefun{with} just evaluates the \codefun{table} expression in a temporary environment of the data. } <>= Art.tab <-with(Arthritis, table(Treatment, Sex, Improved)) str(Art.tab) ftable(Art.tab) @ There may also be times that you will need an equivalent case form \code{data.frame} with factors representing the table variables rather than the frequency table. For example, the \codefun{mca} function in package \pkg{MASS} only operates on data in this format. Marc Schwartz provided code for \codefun{expand.dft} on the Rhelp mailing list for converting a table back into a case form \code{data.frame}. This function is included in \pkg{vcdExtra}. \Example Convert the \data{Arthritis} data in table form (\code{Art.tab}) back to a \code{data.frame} in case form, with factors \code{Treatment}, \code{Sex} and \code{Improved}. <>= Art.df <- expand.dft(Art.tab) str(Art.df) @ \subsection{A complex example}\label{sec:complex} If you've followed so far, you're ready for a more complicated example. The data file, \code{tv.dat} represents a 4-way table of size $5 \times 11 \times 5 \times 3$ where the table variables (unnamed in the file) are read as \code{V1} -- \code{V4}, and the cell frequency is read as \code{V5}. The file, stored in the \code{doc/extdata} directory of \pkg{vcdExtra}, can be read as follows: <>= tv.data<-read.table(system.file("doc","extdata","tv.dat",package="vcdExtra")) head(tv.data,5) @ For a local file, just use \codefun{read.table} in this form: <>= tv.data<-read.table("C:/R/data/tv.dat") @ The data \code{tv.dat} came from the initial implementation of mosaic displays in \proglang{R} by Jay Emerson. In turn, they came from the initial development of mosaic displays \citep{vcd:Hartigan+Kleiner:1984} that illustrated the method with data on a large sample of TV viewers whose behavior had been recorded for the Neilson ratings. This data set contains sample television audience data from Neilsen Media Research for the week starting November 6, 1995. \begin{flushleft} The table variables are:\\ ~~~\code{V1}-- values 1:5 correspond to the days Monday--Friday;\\ ~~~\code{V2}-- values 1:11 correspond to the quarter hour times 8:00PM through 10:30PM;\\ ~~~\code{V3}-- values 1:5 correspond to ABC, CBS, NBC, Fox, and non-network choices;\\ ~~~\code{V4}-- values 1:3 correspond to transition states: turn the television Off, Switch channels, or Persist in viewing the current channel. \end{flushleft} We are interested just the cell frequencies, and rely on the facts that the (a) the table is complete--- there are no missing cells, so \code{nrow(tv.data)}=\Sexpr{nrow(tv.data)}; (b) the observations are ordered so that \code{V1} varies most rapidly and \code{V4} most slowly. From this, we can just extract the frequency column and reshape it into an array. <>= TV <- array(tv.data[,5], dim=c(5,11,5,3)) dimnames(TV) <- list(c("Monday","Tuesday","Wednesday","Thursday","Friday"), c("8:00","8:15","8:30","8:45","9:00","9:15","9:30", "9:45","10:00","10:15","10:30"), c("ABC","CBS","NBC","Fox","Other"), c("Off","Switch","Persist")) names(dimnames(TV))<-c("Day", "Time", "Network", "State") @ More generally (even if there are missing cells), we can use \codefun{xtabs} (or \codefun{plyr::daply}) to do the cross-tabulation, using \code{V5} as the frequency variable. Here's how to do this same operation with \codefun{xtabs}: <>= TV <- xtabs(V5 ~ ., data=tv.data) dimnames(TV) <- list(Day=c("Monday","Tuesday","Wednesday","Thursday","Friday"), Time=c("8:00","8:15","8:30","8:45","9:00","9:15","9:30", "9:45","10:00","10:15","10:30"), Network=c("ABC","CBS","NBC","Fox","Other"), State=c("Off","Switch","Persist")) @ But this 4-way table is too large and awkward to work with. Among the networks, Fox and Other occur infrequently. We can also cut it down to a 3-way table by considering only viewers who persist with the current station.% \footnote{This relies on the fact that that indexing an array drops dimensions of length 1 by default, using the argument \code{drop=TRUE}; the result is coerced to the lowest possible dimension. } <>= TV <- TV[,,1:3,] # keep only ABC, CBS, NBC TV <- TV[,,,3] # keep only Persist -- now a 3 way table structable(TV) @ Finally, for some purposes, we might want to collapse the 11 times into a smaller number. Here, we use \codefun{as.data.frame.table} to convert the table back to a data frame, \codefun{levels} to re-assign the values of \code{Time}, and finally, \codefun{xtabs} to give a new, collapsed frequency table. <>= TV.df <- as.data.frame.table(TV) levels(TV.df$Time) <- c(rep("8:00-8:59",4),rep("9:00-9:59",4), rep("10:00-10:44",3)) TV2 <- xtabs(Freq ~ Day + Time + Network, TV.df) structable(Day ~ Time+Network,TV2) @ Whew! See \figref{fig:TV-mosaic} for a mosaic plot of the \code{TV2} data. \section{Tests of Independence} \subsection{CrossTable} OK, now we're ready to do some analyses. For tabular displays, the \codefun{CrossTable} function in the \pkg{gmodels} package produces cross-tabulations modeled after \code{PROC FREQ} in \proglang{SAS} or \code{CROSSTABS} in \proglang{SPSS}. It has a wealth of options for the quantities that can be shown in each cell. <>= # 2-Way Cross Tabulation library(gmodels) CrossTable(GSStab,prop.t=FALSE,prop.r=FALSE,prop.c=FALSE) @ There are options to report percentages (row, column, cell), specify decimal places, produce Chi-square, Fisher, and McNemar tests of independence, report expected and residual values (pearson, standardized, adjusted standardized), include missing values as valid, annotate with row and column titles, and format as \proglang{SAS} or \proglang{SPSS} style output! See \code{help(CrossTable)} for details. \subsection{Chi-square test} For 2-way tables you can use \codefun{chisq.test} to test independence of the row and column variable. By default, the $p$-value is calculated from the asymptotic chi-squared distribution of the test statistic. Optionally, the $p$-value can be derived via Monte Carlo simulation. <>= (HairEye <- margin.table(HairEyeColor, c(1, 2))) chisq.test(HairEye) @ \subsection{Fisher Exact Test}\label{sec:Fisher} \code{fisher.test(X)} provides an exact test of independence. \code{X} must be a two-way contingency table in table form. Another form, \code{fisher.test(X, Y)} takes two categorical vectors of the same length. For tables larger than $2 \times 2$ the method can be computationally intensive (or can fail) if the frequencies are not small. <>= fisher.test(GSStab) @ But this does not work because \data{HairEye} data has $n$=592 total frequency. An exact test is unnecessary in this case. <>= fisher.test(HairEye) @ %# <>= %# #cat(try(fisher.test(HairEye))) %# @ \begin{Soutput} Error in fisher.test(HairEye) : FEXACT error 6. LDKEY is too small for this problem. Try increasing the size of the workspace. \end{Soutput} \subsection[Mantel-Haenszel test]{Mantel-Haenszel test and conditional association}\label{sec:mantel} Use the \code{mantelhaen.test(X)} function to perform a Cochran-Mantel-Haenszel $\chi^2$ chi test of the null hypothesis that two nominal variables are \emph{conditionally independent}, $A \perp B \given C$, in each stratum, assuming that there is no three-way interaction. \code{X} is a 3 dimensional contingency table, where the last dimension refers to the strata. The \data{UCBAdmissions} serves as an example of a $2 \times 2 \times 6$ table, with \code{Dept} as the stratifying variable. <>= ## UC Berkeley Student Admissions mantelhaen.test(UCBAdmissions) @ The results show no evidence for association between admission and gender when adjusted for department. However, we can easily see that the assumption of equal association across the strata (no 3-way association) is probably violated. For $2 \times 2 \times k$ tables, this can be examimed from the odds ratios for each $2 \times 2$ table (\codefun{oddsratio}), and tested by using \verb|woolf_test()| in \pkg{vcd}. %<>= %oddsRatio <- function(x) (x[1,1]*x[2,2])/(x[1,2]*x[2,1]) %apply(UCBAdmissions, 3, oddsRatio) % %woolf_test(UCBAdmissions) %@ <>= oddsratio(UCBAdmissions, log=FALSE) lor <- oddsratio(UCBAdmissions) # capture log odds ratios summary(lor) woolf_test(UCBAdmissions) @ We can visualize the odds ratios of Admission for each department with fourfold displays using \codefun{fourfold}. The cell frequencies $n_{ij}$ of each $2 \times 2$ table are shown as a quarter circle whose radius is proportional to $\sqrt{n_{ij}}$, so that its area is proportional to the cell frequency. Confidence rings for the odds ratio allow a visual test of the null of no association; the rings for adjacent quadrants overlap \emph{iff} the observed counts are consistent with the null hypothesis. In the extended version (the default), brighter colors are used where the odds ratio is significantly different from 1. The following lines produce \figref{fig:fourfold1}.% \footnote{The color values \code{col[3:4]} were modified from their default values to show a greater contrast between significant and insignifcant associations here.} <>= col <- c("#99CCFF", "#6699CC", "#F9AFAF", "#6666A0", "#FF0000", "#000080") fourfold(UCB,mfrow=c(2,3), color=col) @ %\setkeys{Gin}{width=0.8\textwidth} \begin{figure}[htb] \begin{center} %<>= %col <- c("#99CCFF", "#6699CC", "#F9AFAF", "#6666A0", "#FF0000", "#000080") %fourfold(UCB,mfrow=c(2,3), color=col) %@ \includegraphics[width=0.8\textwidth,trim=80 50 80 50]{fig/vcd-tut-fourfold1} \caption{Fourfold display for the \data{UCBAdmissions} data. Where the odds ratio differs significantly from 1.0, the confidence bands do not overlap, and the circle quadrants are shaded more intensely.} \label{fig:fourfold1} \end{center} \end{figure} Another \pkg{vcd} function, \codefun{cotabplot}, provides a more general approach to visualizing conditional associations in contingency tables, similar to trellis-like plots produced by \codefun{coplot} and lattice graphics. The \code{panel} argument supplies a function used to render each conditional subtable. The following gives a display (not shown) similar to \figref{fig:fourfold1}. <>= cotabplot(UCB, panel = cotab_fourfold) @ When we want to view the conditional probabilities of a response variable (e.g., \code{Admit}) in relation to several factors, an alternative visualization is a \codefun{doubledecker} plot. This plot is a specialized version of a mosaic plot, which highlights the levels of a response variable (plotted vertically) in relation to the factors (shown horizontally). The following call produces \figref{fig:doubledecker}, where we use indexing on the first factor (\code{Admit}) to make \code{Admitted} the highlighted level. In this plot, the association between \code{Admit} and \code{Gender} is shown where the heights of the highlighted conditional probabilities do not align. The excess of females admitted in Dept A stands out here. <>= doubledecker(Admit ~ Dept + Gender, data=UCBAdmissions[2:1,,]) @ \begin{figure}[htb] \begin{center} \includegraphics[width=0.9\textwidth]{fig/vcd-tut-doubledecker} \caption{Doubledecker display for the \data{UCBAdmissions} data. The heights of the highlighted bars show the conditional probabilities of \texttt{Admit}, given \texttt{Dept} and \texttt{Gender}.} \label{fig:doubledecker} \end{center} \end{figure} Finally, the there is a \codefun{plot} method for \code{oddsratio} objects. By default, it shows the 95\% confidence interval for the log odds ratio. \figref{fig:oddsratio} is produced by: <>= plot(lor, xlab="Department", ylab="Log Odds Ratio (Admit | Gender)") @ \setkeys{Gin}{width=0.5\textwidth} \begin{figure}[htb] \begin{center} <>= plot(lor, xlab="Department", ylab="Log Odds Ratio (Admit | Gender)") @ \caption{Log odds ratio plot for the \data{UCBAdmissions} data.} \label{fig:oddsratio} \end{center} \end{figure} \subsection[CMH tests: ordinal factors]{Cochran-Mantel-Haenszel tests for ordinal factors}\label{sec:CMH} The standard $\chi^2$ tests for association in a two-way table treat both table factors as nominal (unordered) categories. When one or both factors of a two-way table are quantitative or ordinal, more powerful tests of association may be obtaianed by taking ordinality into account, using row and or column scores to test for linear trends or differences in row or column means. More general versions of the CMH tests (Landis etal., 1978) are provided by assigning numeric scores to the row and/or column variables. For example, with two ordinal factors (assumed to be equally spaced), assigning integer scores, \code{1:R} and \code{1:C} tests the linear $\times$ linear component of association. This is statistically equivalent to the Pearson correlation between the integer-scored table variables, with $\chi^2 = (n-1) r^2$, with only 1 $df$ rather than $(R-1)\times(C-1)$ for the test of general association. When only one table variable is ordinal, these general CMH tests are analogous to an ANOVA, testing whether the row mean scores or column mean scores are equal, again consuming fewer $df$ than the test of general association. The \codefun{CMHtest} function in \pkg{vcdExtra} now calculates these various CMH tests for two possibly ordered factors, optionally stratified other factor(s). \Example Recall the $4 \times 4$ table, \code{JobSat} introduced in \secref{sec:creating}, <>= JobSat @ Treating the \code{satisfaction} levels as equally spaced, but using midpoints of the \code{income} categories as row scores gives the following results: <>= CMHtest(JobSat, rscores=c(7.5,20,32.5,60)) @ Note that with the relatively small cell frequencies, the test for general give no evidence for association. However, the the \code{cor} test for linear x linear association on 1 df is nearly significant. The \pkg{coin} contains the functions \verb|cmh_test()| and \verb|lbl_test()| for CMH tests of general association and linear x linear association respectively. \subsection{Measures of Association} There are a variety of statistical measures of \emph{strength} of association for contingency tables--- similar in spirit to $r$ or $r^2$ for continuous variables. With a large sample size, even a small degree of association can show a significant $\chi^2$, as in the example below for the \data{GSS} data. The \codefun{assocstats} function in \pkg{vcd} calculates the $\phi$ contingency coefficient, and Cramer's V for an $r \times c$ table. The input must be in table form, a two-way $r \times c$ table. It won't work with \data{GSS} in frequency form, but by now you should know how to convert. <>= assocstats(GSStab) @ For tables with ordinal variables, like \data{JobSat}, some people prefer the Goodman-Kruskal $\gamma$ statistic (\citet[\S 2.4.3]{vcd:Agresti:2002}) based on a comparison of concordant and discordant pairs of observations in the case-form equivalent of a two-way table. <>= GKgamma(JobSat) @ A web article by Richard Darlington, \url{http://www.psych.cornell.edu/Darlington/crosstab/TABLE0.HTM} gives further description of these and other measures of association. \subsection{Measures of Agreement} The \codefun{Kappa} function in the \pkg{vcd} package calculates Cohen's $\kappa$ and weighted $\kappa$ for a square two-way table with the same row and column categories \citep{Cohen:60}.% \footnote{ Don't confuse this with \codefun{kappa} in base \proglang{R} that computes something entirely different (the condition number of a matrix). } Normal-theory $z$-tests are obtained by dividing $\kappa$ by its asymptotic standard error (ASE). A \codefun{confint} method for \code{Kappa} objects provides confidence intervals. <>= (K <- Kappa(SexualFun)) confint(K) @ A visualization of agreement, both unweighted and weighted for degree of departure from exact agreement is provided by the \codefun{agreementplot} function. \figref{fig:agreesex} shows the agreementplot for the \data{SexualFun} data, produced as shown below. The Bangdiwala measures represent the proportion of the shaded areas of the diagonal rectangles, using weights $w_1$ for exact agreement, and $w_2$ for partial agreement one step from the main diagonal. <>= agree <- agreementplot(SexualFun, main="Is sex fun?") unlist(agree) @ %\setkeys{Gin}{width=0.5\textwidth} \begin{figure}[htb] \begin{center} %<>= %agree <- agreementplot(SexualFun, main="Is sex fun?") %agree %@ \includegraphics[width=0.4\textwidth,trim=50 25 50 25]{fig/vcd-tut-agreesex} \caption{Agreement plot for the \data{SexualFun} data.} \label{fig:agreesex} \end{center} \end{figure} In other examples, the agreement plot can help to show \emph{sources} of disagreement. For example, when the shaded boxes are above or below the diagonal (red) line, a lack of exact agreement can be attributed in part to different frequency of use of categories by the two raters-- lack of \emph{marginal homogeneity}. \subsection{Correspondence analysis} Use the \pkg{ca} package for correspondence analysis for visually exploring relationships between rows and columns in contingency tables. For an $r \times c$ table, the method provides a breakdown of the Pearson $\chi^2$ for association in up to $M = \min(r-1, c-1)$ dimensions, and finds scores for the row ($x_{im}$) and column ($y_{jm}$) categories such that the observations have the maximum possible correlations.% \footnote{ Related methods are the non-parametric CMH tests using assumed row/column scores (\secref{sec:CMH}), the analogous \codefun{glm} model-based methods (\secref{sec:CMH}), and the more general RC models which can be fit using \codefun{gnm}. Correspondence analysis differs in that it is a primarily descriptive/exploratory method (no significance tests), but is directly tied to informative graphic displays of the row/column categories. } Here, we carry out a simple correspondence analysis of the \data{HairEye} data. The printed results show that nearly 99\% of the association between hair color and eye color can be accounted for in 2 dimensions, of which the first dimension accounts for 90\%. <>= library(ca) ca(HairEye) @ The resulting \code{ca} object can be plotted just by running the \codefun{plot} method on the \code{ca} object, giving the result in \figref{fig:ca-haireye}. \codefun{plot.ca} does not allow labels for dimensions; these can be added with \codefun{title}. It can be seen that most of the association is accounted for by the ordering of both hair color and eye color along Dimension 1, a dark to light dimension. <>= plot(ca(HairEye), main="Hair Color and Eye Color") title(xlab="Dim 1 (89.4%)", ylab="Dim 2 (9.5%)") @ \setkeys{Gin}{width=0.7\textwidth} \begin{figure}[htb] \begin{center} <>= plot(ca(HairEye), main="Hair Color and Eye Color") title(xlab="Dim 1 (89.4%)", ylab="Dim 2 (9.5%)") @ \caption{Correspondence analysis plot for the \data{HairEye} data.} \label{fig:ca-haireye} \end{center} \end{figure} \section{Loglinear Models}\label{sec:loglin} You can use the \codefun{loglm} function in the \pkg{MASS} package to fit log-linear models. Equivalent models can also be fit (from a different perspective) as generalized linear models with the \codefun{glm} function using the \code{family='poisson'} argument, and the \pkg{gnm} package provides a wider range of generalized \emph{nonlinear} models, particularly for testing structured associations. The visualization methods for these models were originally developed for models fit using \codefun{loglm}, so this approach is emphasized here. Some extensions of these methods for models fit using \codefun{glm} and \codefun{gnm} are contained in the \pkg{vcdExtra} package and illustrated in \secref{sec:glm}. Assume we have a 3-way contingency table based on variables A, B, and C. The possible different forms of \loglin\ models for a 3-way table are shown in \tabref{tab:loglin-3way}. The \textbf{Model formula} column shows how to express each model for \codefun{loglm} in \proglang{R}.% \footnote{ For \codefun{glm}, or \codefun{gnm}, with the data in the form of a frequency data.frame, the same model is specified in the form \code{glm(Freq} $\sim$ \code{..., family="poisson")}, where \texttt{Freq} is the name of the cell frequency variable and \texttt{...} specifies the \textbf{Model formula}. } In the \textbf{Interpretation} column, the symbol ``$\perp$'' is to be read as ``is independent of,'' and ``$\given$'' means ``conditional on,'' or ``adjusting for,'' or just ``given''. \begin{table}[htb] \caption{Log-linear Models for Three-Way Tables}\label{tab:loglin-3way} \begin{center} \begin{tabular}{llll} \hline \textbf{Model} & \textbf{Model formula} & \textbf{Symbol}& \textbf{Interpretation} \\ \hline\hline Mutual independence & \verb|~A + B + C| & $[A][B][C]$ & $A \perp B \perp C$ \\ Joint independence & \verb|~A*B + C| & $[AB][C]$ & $(A \: B) \perp C$ \\ Conditional independence & \verb|~(A+B)*C| & $[AC][BC]$ & $(A \perp B) \given C$ \\ All two-way associations & \verb|~A*B + A*C + B*C| & $[AB][AC][BC]$ & homogeneous association \\ Saturated model & \verb|~A*B*C| & $[ABC]$ & 3-way association \\ \hline \end{tabular} \end{center} \end{table} For example, the formula \verb|~A + B + C| specifies the model of \emph{mutual independence} with no associations among the three factors. In standard notation for the expected frequencies $m_{ijk}$, this corresponds to \begin{equation*} \log ( m_{ijk} ) = \mu + \lambda_i^A + \lambda_j^B + \lambda_k^C \equiv \texttt{A + B + C} \end{equation*} The parameters $\lambda_i^A , \lambda_j^B$ and $\lambda_k^C$ pertain to the differences among the one-way marginal frequencies for the factors A, B and C. Similarly, the model of \emph{joint independence}, $(A \: B) \perp C$, allows an association between A and B, but specifies that C is independent of both of these and their combinations, \begin{equation*} \log ( m_{ijk} ) = \mu + \lambda_i^A + \lambda_j^B + \lambda_k^C + \lambda_{ij}^{AB} \equiv \texttt{A * B + C} \end{equation*} where the parameters $\lambda_{ij}^{AB}$ pertain to the overall association between A and B (collapsing over C). In the literature or text books, you will often find these models expressed in shorthand symbolic notation, using brackets, \texttt{[ ]} to enclose the \emph{high-order terms} in the model. Thus, the joint independence model can be denoted \texttt{[AB][C]}, as shown in the \textbf{Symbol} column in \tabref{tab:loglin-3way}. Models of \emph{conditional independence} allow (and fit) two of the three possible two-way associations. There are three such models, depending on which variable is conditioned upon. For a given conditional independence model, e.g., \texttt{[AB][AC]}, the given variable is the one common to all terms, so this example has the interpretation $(B \perp C) \given A$. \subsection[Fitting with loglm()]{Fitting with \codefun{loglm}}\label{sec:loglm} For example, we can fit the model of mutual independence among hair color, eye color and sex in \data{HairEyeColor} as <>= library(MASS) ## Independence model of hair and eye color and sex. hec.1 <- loglm(~Hair+Eye+Sex, data=HairEyeColor) hec.1 @ Similarly, the models of conditional independence and joint independence are specified as <>= ## Conditional independence hec.2 <- loglm(~(Hair + Eye) * Sex, data=HairEyeColor) hec.2 @ <>= ## Joint independence model. hec.3 <- loglm(~Hair*Eye + Sex, data=HairEyeColor) hec.3 @ Note that printing the model gives a brief summary of the goodness of fit. A set of models can be compared using the \codefun{anova} function. <>= anova(hec.1, hec.2, hec.3) @ %Martin Theus and Stephan Lauer have written an excellent article on Visualizing %Loglinear Models, using mosaic plots. There is also great tutorial example by %Kevin Quinn on analyzing loglinear models via glm. \subsection[Fitting with glm() and gnm()]{Fitting with \codefun{glm} and \codefun{gnm}}\label{sec:glm} The \codefun{glm} approach, and extensions of this in the \pkg{gnm} package allows a much wider class of models for frequency data to be fit than can be handled by \codefun{loglm}. Of particular importance are models for ordinal factors and for square tables, where we can test more structured hypotheses about the patterns of association than are provided in the tests of general assosiation under \codefun{loglm}. These are similar in spirit to the non-parametric CMH tests described in \secref{sec:CMH}. \Example The data \code{Mental} in the \pkg{vcdExtra} package gives a two-way table in frequency form classifying young people by their mental health status and parents' socioeconomic status (SES), where both of these variables are ordered factors. <>= str(Mental) xtabs(Freq ~ mental+ses, data=Mental) # display the frequency table @ Simple ways of handling ordinal variables involve assigning scores to the table categories, and the simplest cases are to use integer scores, either for the row variable (``column effects'' model), the column variable (``row effects'' model), or both (``uniform association'' model). <>= indep <- glm(Freq ~ mental + ses, family = poisson, data = Mental) # independence model @ To fit more parsimonious models than general association, we can define numeric scores for the row and column categories <>= # Use integer scores for rows/cols Cscore <- as.numeric(Mental$ses) Rscore <- as.numeric(Mental$mental) @ Then, the row effects model, the column effects model, and the uniform association model can be fit as follows: <>= # column effects model (ses) coleff <- glm(Freq ~ mental + ses + Rscore:ses, family = poisson, data = Mental) # row effects model (mental) roweff <- glm(Freq ~ mental + ses + mental:Cscore, family = poisson, data = Mental) # linear x linear association linlin <- glm(Freq ~ mental + ses + Rscore:Cscore, family = poisson, data = Mental) @ The \codefun{Summarize} in \pkg{vcdExtra} provides a nice, compact summary of the fit statistics for a set of models, collected into a \class{glmlist} object. Smaller is better for AIC and BIC. <>= # compare models using AIC, BIC, etc vcdExtra::LRstats(glmlist(indep, roweff, coleff, linlin)) @ For specific model comparisons, we can also carry out tests of \emph{nested} models with \codefun{anova} when those models are listed from smallest to largest. Here, there are two separate paths from the most restrictive (independence) model through the model of uniform association, to those that allow only one of row effects or column effects. <>= anova(indep, linlin, coleff, test="Chisq") anova(indep, linlin, roweff, test="Chisq") @ The model of linear by linear association seems best on all accounts. For comparison, one might try the CMH tests on these data: <>= CMHtest(xtabs(Freq~ses+mental, data=Mental)) @ \subsection{Non-linear terms} The strength of the \pkg{gnm} package is that it handles a wide variety of models that handle non-linear terms, where the parameters enter the model beyond a simple linear function. The simplest example is the Goodman RC(1) model, which allows a multiplicative term to account for the association of the table variables. In the notation of generalized linear models with a log link, this can be expressed as \begin{equation*} \log \mu_{ij} = \alpha_i + \beta_j + \gamma_{i} \delta_{j} \end{equation*} where the row-multiplicative effect parameters $\gamma_i$ and corresponding column parameters $\delta_j$ are estimated from the data.% \footnote{ This is similar in spirit to a correspondence analysis with a single dimension, but as a statistical model. } Similarly, the RC(2) model adds two multiplicative terms to the independence model, \begin{equation*} \log \mu_{ij} = \alpha_i + \beta_j + \gamma_{i1} \delta_{j1} + \gamma_{i2} \delta_{j2} \end{equation*} In the \pkg{gnm} package, these models may be fit using the \codefun{Mult} to specify the multiplicative term, and \codefun{instances} to specify several such terms. \Example For the \code{Mental} data, we fit the RC(1) and RC(2) models, and compare these with the independence model. <>= RC1 <- gnm(Freq ~ mental + ses + Mult(mental,ses), data=Mental, family=poisson, , verbose=FALSE) RC2 <- gnm(Freq ~ mental+ses + instances(Mult(mental,ses),2), data=Mental, family=poisson, verbose=FALSE) anova(indep, RC1, RC2, test="Chisq") @ \section{Mosaic plots}\label{sec:mosaic} Mosaic plots provide an ideal method both for visualizing contingency tables and for visualizing the fit--- or more importantly--- lack of fit of a \loglin\ model. For a two-way table, \codefun{mosaic} fits a model of independence, $[A][B]$ or \verb|~A+B| as an \proglang{R} formula. For $n$-way tables, \codefun{mosaic} can fit any \loglin\ model, and can also be used to plot a model fit with \codefun{loglm}. See \citet{vcd:Friendly:1994,vcd:Friendly:1999} for the statistical ideas behind these uses of mosaic displays in connection with \loglin\ models. The essential idea is to recursively sub-divide a unit square into rectangular ``tiles'' for the cells of the table, such that the are area of each tile is proportional to the cell frequency. For a given \loglin\ model, the tiles can then be shaded in various ways to reflect the residuals (lack of fit) for a given model. The pattern of residuals can then be used to suggest a better model or understand \emph{where} a given model fits or does not fit. \codefun{mosaic} provides a wide range of options for the directions of splitting, the specification of shading, labeling, spacing, legend and many other details. It is actually implemented as a special case of a more general class of displays for $n$-way tables called \code{strucplot}, including sieve diagrams, association plots, double-decker plots as well as mosaic plots. For details, see \code{help(strucplot)} and the ``See also'' links, and also \citet{vcd:Meyer+Zeileis+Hornik:2006b}, which is available as an \proglang{R} vignette via \code{vignette("strucplot", package="vcd")}. \figref{fig:arthritis}, showing the association between \code{Treatment} and \code{Improved} was produced with the following call to \codefun{mosaic}. <>= mosaic(art, gp = shading_max, split_vertical = TRUE, main="Arthritis: [Treatment] [Improved]") @ Note that the residuals for the independence model were not large (as shown in the legend), yet the association between \code{Treatment} and \code{Improved} is highly significant. <>= summary(art) @ In contrast, one of the other shading schemes, from \citet{vcd:Friendly:1994} (use: \verb|gp = shading_Friendly|), uses fixed cutoffs of $\pm 2, \pm 4$, to shade cells which are \emph{individually} significant at approximately $\alpha = 0.05$ and $\alpha = 0.001$ levels, respectively. The right panel below uses \verb|gp = shading_Friendly|. \setkeys{Gin}{width=0.5\textwidth} <>= mosaic(art, gp = shading_max, split_vertical = TRUE, main="Arthritis: gp = shading_max") @ <>= mosaic(art, gp = shading_Friendly, split_vertical = TRUE, main="Arthritis: gp = shading_Friendly") @ \subsection[Mosaics for loglinear models]{Mosaics for \loglin\ models}\label{sec:mosaic-llm} When you have fit a \loglin\ model using \codefun{loglm}, and saved the result (as a \code{loglm} object) the simplest way to display the results is to use the \codefun{plot} method for the \code{loglm} object. Calling \code{mosaic(loglm.object)} has the same result. In \secref{sec:loglm} above, we fit several different models to the \data{HairEyeColor} data. We can produce mosaic displays of each just by plotting them: <>= # mosaic plots, using plot.loglm() method plot(hec.1, main="model: [Hair][Eye][Sex]") plot(hec.2, main="model: [HairSex][EyeSex]") plot(hec.3, main="model: [HairEye][Sex]") @ \setkeys{Gin}{width=0.32\textwidth} <>= plot(hec.1, main="model: [Hair][Eye][Sex]") @ <>= plot(hec.2, main="model: [HairSex][EyeSex]") @ <>= plot(hec.3, main="model: [HairSex][EyeSex]") @ Alternatively, you can supply the model formula to \codefun{mosaic} with the \code{expected} argument. This is passed to \codefun{loglm}, which fits the model, and returns residuals used for shading in the plot. For example, here we examine the \data{TV2} constructed in \secref{sec:complex} above. The goal is to see how Network choice depends on (varies with) Day and Time. To do this: \begin{itemize} \item We fit a model of joint independence of \code{Network} on the combinations of \code{Day} and \code{Time}, with the model formula \verb|~Day:Time + Network|. \item To make the display more easily read, we place \code{Day} and \code{Time} on the vertical axis and \code{Network} on the horizontal, \item The \code{Time} values overlap on the right vertical axis, so we use \codefun{level} to abbreviate them. \codefun{mosaic} also supports a more sophisticated set of labeling functions. Instead of changing the data table, we could have used \verb|labeling_args = list(abbreviate = c(Time = 2))| for a similar effect. \end{itemize} The following call to \codefun{mosaic} produces \figref{fig:TV-mosaic}: <>= dimnames(TV2)$Time <- c("8", "9", "10") # re-level for mosaic display mosaic(~ Day + Network + Time, data=TV2, expected=~Day:Time + Network, legend=FALSE, gp=shading_Friendly) @ \setkeys{Gin}{width=0.75\textwidth} \begin{figure}[htb] \begin{center} <>= dimnames(TV2)$Time <- c("8", "9", "10") # re-level for mosaic display mosaic(~ Day + Network + Time, data=TV2, expected=~Day:Time + Network, legend=FALSE, gp=shading_Friendly) @ \caption{Mosaic plot for the \data{TV} data showing model of joint independence, \texttt{Day:Time + Network} .} \label{fig:TV-mosaic} \end{center} \end{figure} From this, it is easy to read from the display how network choice varies with day and time. For example, CBS dominates in all time slots on Monday; ABC and NBC dominate on Tuesday, particularly in the later time slots; Thursday is an NBC day, while on Friday, ABC gets the greatest share. In interpreting this mosaic and other plots, it is important to understand that associations included in the model---here, that between day and time---are \emph{not} shown in the shading of the cells, because they have been fitted (taken into account) in the \loglin\ model. For comparison, you might want to try fitting the model of homogeneous association. This allows all pairs of factors to be associated, but asserts that each pairwise association is the same across the levels of the remaining factor. The resulting plot displays the contributions to a 3-way association, but is not shown here. <>= mosaic(~ Day + Network + Time, data=TV2, expected=~Day:Time + Day:Network + Time:Network, legend=FALSE, gp=shading_Friendly) @ \subsection[Mosaics for glm() and gnm() models]{Mosaics for \codefun{glm} and \codefun{gnm} models}\label{sec:mosglm} The \pkg{vcdExtra} package provides an additional method, \codefun{mosaic.glm} for models fit with \codefun{glm} and \codefun{gnm}.% \footnote{ Models fit with \codefun{gnm} are of \code{class = c("gnm", "glm", "lm")}, so all \code{*.glm} methods apply, unless overridden in the \pkg{gnm} package. } These are not restricted to the Poisson family, but only apply to cases where the response variable is non-negative. \Example Here, we plot the independence and the linear-by-linear association model for the Mental health data from \secref{sec:glm}. These examples illustrate some of the options for labeling (variable names and residuals printed in cells). Note that the \code{formula} supplied to \codefun{mosaic} for \class{glm} objects refers to the order of factors displayed in the plot, not the model. <>= long.labels <- list(set_varnames = c(mental="Mental Health Status", ses="Parent SES")) mosaic(indep, ~ses+mental, residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, main="Mental health data: Independence") mosaic(linlin, ~ses+mental, residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, suppress=1, gp=shading_Friendly, main="Mental health data: Linear x Linear") @ \setkeys{Gin}{width=0.49\textwidth} <>= long.labels <- list(set_varnames = c(mental="Mental Health Status", ses="Parent SES")) mosaic(indep, ~ses+mental, residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, main="Mental health data: Independence") @ <>= long.labels <- list(set_varnames = c(mental="Mental Health Status", ses="Parent SES")) mosaic(linlin, ~ses+mental, residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, suppress=1, gp=shading_Friendly, main="Mental health data: Linear x Linear") @ The \pkg{gnm} package also fits a wide variety of models with nonlinear terms or terms for structured associations of table variables. In the following, we fit the RC(1) model \begin{equation*} \log ( m_{ij} ) = \mu + \lambda_i^A + \lambda_j^B + \phi \mu_i \nu_j \end{equation*} This is similar to the linear by linear model, except that the row effect parameters ($\mu_i$) and column parameters ($\nu_j$) are estimated from the data rather than given assigned equally-spaced values. The multiplicative terms are specified by the \codefun{Mult}. <>= Mental$mental <- C(Mental$mental, treatment) Mental$ses <- C(Mental$ses, treatment) RC1model <- gnm(Freq ~ mental + ses + Mult(mental, ses), family = poisson, data = Mental) mosaic(RC1model, residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, suppress=1, gp=shading_Friendly, main="Mental health data: RC(1) model") @ Other forms of nonlinear terms are provided for the inverse of a predictor (\codefun{Inv}) and the exponential of a predictor (\codefun{Exp}). You should read \code{vignette("gnmOverview", package="gnm")} for further details. \subsection{Mosaic tips and techniques}\label{sec:tips} The \pkg{vcd} package implements an extremely general collection of graphical methods for $n$-way frequency tables within the strucplot framework, which includes mosaic plots (\codefun{mosaic}), as well as association plots (\codefun{assoc}), sieve diagrams (\codefun{sieve}), as well as tabular displays (\codefun{structable}). The graphical methods in \pkg{vcd} support a wide of options that control almost all of the details of the plots, but it is often difficult to determine what arguments you need to supply to achieve a given effect from the \code{help()}. As a first step, you should read the \code{vignette("strucplot")} in \pkg{vcd} to understand the overall structure of these plot methods. The notes below describe a few useful things that may not be obvious, or can be done in different ways. \subsubsection[Changing labels]{Changing the labels for variables and levels} With data in contingency table form or as a frequency data frame, it often happens that the variable names and/or the level values of the factors, while suitable for analysis, are less than adequate when used in mosaic plots and other strucplot displays. For example, we might prefer that a variable named \code{ses} appear as \code{"Socioeconomic Status"}, or a factor with levels \code{c("M", "F")} be labeled using \code{c("Male", "Female")} in a plot. Or, sometimes we start with a factor whose levels are fully spelled out (e.g., \code{c("strongly disagree", "disagree", "neutral", "agree", "strongly agree")}), only to find that the level labels overlap in graphic displays. The structplot framework in \pkg{vcd} provides an extremely large variety of functions and options for controlling almost all details of text labels in mosaics and other plots. See \code{help(labelings)} for an overview. For example, in \secref{sec:ordered-factors} we showed how to rearrange the dimensions of the \code{UCBAdmissions} table, change the names of the table variables, and relabel the levels of one of the table variables. The code below changes the actual table for plotting purposes, but we pointed out that these changes can create other problems in analysis. <>= UCB <- aperm(UCBAdmissions, c(2, 1, 3)) names(dimnames(UCB)) <- c("Sex", "Admit?", "Department") dimnames(UCB)[[2]] <- c("Yes", "No") @ The same effects can be achieved \emph{without} modifying the data using the \verb|set_varnames| and \verb|set_labels| options in \codefun{mosaic} as follows: <>= vnames <- list(set_varnames = c(Admit="Admission", Gender="Sex", Dept="Department")) lnames <- list(Admit = c("Yes", "No"), Gender = c("Males", "Females"), Dept = LETTERS[1:6]) mosaic(UCBAdmissions, labeling_args=vnames, set_labels=lnames) @ In some cases, it may be sufficient to abbreviate (or clip, or rotate) level names to avoid overlap. For example, the statements below produce another version of \figref{fig:TV-mosaic} with days of the week abbreviated to their first three letters. Section 4 in the \code{vignette("strucplot")} provides many other examples. <>= dimnames(TV2)$Time <- c("8", "9", "10") # re-level for mosaic display mosaic(~ Day + Network + Time, data=TV2, expected=~Day:Time + Network, legend=FALSE, gp=shading_Friendly, labeling_args=list(abbreviate=c(Day=3)) ) @ %\subsubsection{Fitting complex models with glm() and gnm()} \section[Continuous predictors]{Continuous predictors}\label{sec:contin} When continuous predictors are available---and potentially important--- in explaining a categorical outcome, models for that outcome include: logistic regression (binary response), the proportional odds model (ordered polytomous response), multinomial (generalized) logistic regression. Many of these are special cases of the generalized linear model using the \code{"poisson"} or \code{"binomial"} family and their relatives. \subsection{Spine and conditional density plots}\label{sec:spine} I don't go into fitting such models here, but I would be remiss not to illustrate some visualizations in \pkg{vcd} that are helpful here. The first of these is the spine plot or spinogram \citep{vcd:Hummel:1996} (produced with \codefun{spine}). These are special cases of mosaic plots with specific spacing and shading to show how a categorical response varies with a continuous or categorical predictor. They are also a generalization of stacked bar plots where not the heights but the \emph{widths} of the bars corresponds to the relative frequencies of \code{x}. The heights of the bars then correspond to the conditional relative frequencies of {y} in every \code{x} group. \Example For the \data{Arthritis} data, we can see how \code{Improved} varies with \code{Age} as follows. \codefun{spine} takes a formula of the form \verb|y ~ x| with a single dependent factor and a single explanatory variable \code{x} (a numeric variable or a factor). The range of a numeric variable\code{x} is divided into intervals based on the \code{breaks} argument, and stacked bars are drawn to show the distribution of \code{y} as \code{x} varies. As shown below, the discrete table that is visualized is returned by the function. <>= (spine(Improved ~ Age, data = Arthritis, breaks = 3)) (spine(Improved ~ Age, data = Arthritis, breaks = "Scott")) @ \setkeys{Gin}{width=0.49\textwidth} <>= (spine(Improved ~ Age, data = Arthritis, breaks = 3)) @ <>= (spine(Improved ~ Age, data = Arthritis, breaks = "Scott")) @ The conditional density plot \citep{vcd:Hofmann+Theus} is a further generalization. This visualization technique is similar to spinograms, but uses a smoothing approach rather than discretizing the explanatory variable. As well, it uses the original \code{x} axis and not a distorted one. \setkeys{Gin}{width=0.6\textwidth} \begin{figure}[htb] \begin{center} <>= cdplot(Improved ~ Age, data = Arthritis) with(Arthritis, rug(jitter(Age), col="white", quiet=TRUE)) @ \caption{Conditional density plot for the \data{Arthritis} data showing the variation of Improved with Age.} \label{fig:cd-plot} \end{center} \end{figure} In such plots, it is useful to also see the distribution of the observations across the horizontal axis, e.g., with a \codefun{rug} plot. \figref{fig:cd-plot} uses \codefun{cdplot} from the \pkg{graphics} package rather than \verb|cd_plot()| from \pkg{vcd}, and is produced with <>= cdplot(Improved ~ Age, data = Arthritis) with(Arthritis, rug(jitter(Age), col="white", quiet=TRUE)) @ From \figref{fig:cd-plot} it can be easily seen that the proportion of patients reporting Some or Marked improvement increases with Age, but there are some peculiar bumps in the distribution. These may be real or artifactual, but they would be hard to see with most other visualization methods. When we switch from non-parametric data exploration to parametric statistical models, such effects are easily missed. \subsection[Model-based plots]{Model-based plots: effect plots and \pkg{ggplot2} plots}\label{sec:modelplots} The nonparametric conditional density plot uses smoothing methods to convey the distributions of the response variable, but displays that are simpler to interpret can often be obtained by plotting the predicted response from a parametric model. For complex \codefun{glm} models with interaction effects, the \pkg{effects} package provides the most useful displays, plotting the predicted values for a given term, averaging over other predictors not included in that term. I don't illustrate this here, but see \citet{effects:1,effects:2} and \code{help(package="effects")}. Here I just briefly illustrate the capabilities of the \pkg{ggplot2} package for model-smoothed plots of categorical responses in \codefun{glm} models. \Example The \data{Donner} data frame in \pkg{vcdExtra} gives details on the survival of 90 members of the Donner party, a group of people who attempted to migrate to California in 1846. They were trapped by an early blizzard on the eastern side of the Sierra Nevada mountains, and before they could be rescued, nearly half of the party had died. What factors affected who lived and who died? <>= data(Donner, package="vcdExtra") str(Donner) @ A potential model of interest is the logistic regression model for $Pr(survived)$, allowing separate fits for males and females as a function of \code{age}. The key to this is the \verb|stat_smooth()| function, using \code{method = "glm", method.args = list(family = binomial)}. The \verb|formula = y ~ x| specifies a linear fit on the logit scale (\figref{fig:donner3}, left) <>= # separate linear fits on age for M/F ggplot(Donner, aes(age, survived, color = sex)) + geom_point(position = position_jitter(height = 0.02, width = 0)) + stat_smooth(method = "glm", method.args = list(family = binomial), formula = y ~ x, alpha = 0.2, size=2, aes(fill = sex)) @ Alternatively, we can allow a quadratic relation with \code{age} by specifying \verb|formula = y ~ poly(x,2)| (\figref{fig:donner3}, right). <>= # separate quadratics ggplot(Donner, aes(age, survived, color = sex)) + geom_point(position = position_jitter(height = 0.02, width = 0)) + stat_smooth(method = "glm", method.args = list(family = binomial), formula = y ~ poly(x,2), alpha = 0.2, size=2, aes(fill = sex)) @ \setkeys{Gin}{width=0.49\textwidth} \begin{figure}[htb] \begin{center} <>= ggplot(Donner, aes(age, survived, color = sex)) + geom_point(position = position_jitter(height = 0.02, width = 0)) + stat_smooth(method = "glm", method.args = list(family = binomial), formula = y ~ x, alpha = 0.2, size=2, aes(fill = sex)) @ <>= # separate quadratics ggplot(Donner, aes(age, survived, color = sex)) + geom_point(position = position_jitter(height = 0.02, width = 0)) + stat_smooth(method = "glm", method.args = list(family = binomial), formula = y ~ poly(x,2), alpha = 0.2, size=2, aes(fill = sex)) @ \caption{Logistic regression plots for the \data{Donner} data showing survival vs. age, by sex. Left: linear logistic model; right: quadratic model} \label{fig:donner3} \end{center} \end{figure} These plots very nicely show (a) the fitted $Pr(survived)$ for males and females; (b) confidence bands around the smoothed model fits and (c) the individual observations by jittered points at 0 and 1 for those who died and survided, respectively. \bibliography{vcd,vcdExtra} \end{document} vcdExtra/inst/doc/vcd-tutorial.pdf0000644000176200001440000204017513163514401016673 0ustar liggesusers%PDF-1.5 %¿÷¢þ 1 0 obj << /Type /ObjStm /Length 4616 /Filter /FlateDecode /N 91 /First 766 >> stream xœÝ\YsÜ8’~ß_·éŽ “ÄMLLt„,Û²Ú–Iîv÷„cƒ®¢$¶ëP×ácö·ï—È"‹UeËÓ¥*$H$ò Á2&™±L1a ÓLçŠÆ3n™ÅE;–3εbŽqi3Ë…Þ ÎÐÀAFÎ5ÑS8¼d©€®ÖtƒAk@–Ô×€, (ˆ?.ó t!êg˜È+(ÈJ¢2€ã&Ç+@V ˜E7T•Å1ý $ð¤YÕD[“ÓÃ+G“’Z?û‚d-0dŒÜ°µ¹¬¢\[Lœ5tƒ&– ç ¤%ÈŽ²ÉÐcÈF€n91–´ç€lhŠ€6n@.°&7ˆ~q“£ ›sŠ{p¨šË98€l34q€ ,0¿€l%*c¹UŽÿ×?þÁÒ“rUŒ‹UÁœ„,œ²ôåz5©fåbà˯ŠKT(œ¹.Yzˆú“ù%ûé'â`½ºš/Ø?âÍ¢,VÕ|ö¨X•ì‡G¤Ì :JadêoYö·ºµy^œ—oÙ§juÅ®{±(/Øu1ú€NQíYùåÓ|1^zØ'óñm`_-æãõ¨Ü£WÏÙÑÕ|¹ZŽÕõйDpT8[¿ÿ£­<¼ój5)é.ãp¾ž­HÒgºü'´9÷¿`dº8ÿ«Ã3Ì„ê¹@/.ÂKÄRhÈUÝ2‹W¯"\M¼f2–c»ÜÆklŸGDâcÁ¹NHË\œ[îâs]?×±»º~ƒ0¡™…«ˆÕB«w5Ψ™`6›¯ˆj<à #udSñ*ãUÅ«ŽW¯6^#< UŠODx"žˆð"²RDx"‹ó½®2“žŒðd„i(e„'#<áÉOFx*ÂSžj¨u8Ÿ­ÊˆCÄ‹ÑI9®Š‡óÏ =1 Å\½#š.Pµ·Ór9_/F 2‘øñçÕÑÙŠD@ƉOŸÌ‰e#?ûÎÊ ¦¯=Á4•ŸWúÓOíë²»Œj±\±ðy{™…±ÆšÊ媇ôío¿C{&’tI€ùl=™ê/¨7a]­6ü@¢‰²v<[y %ñcsâŒÆ ó§1ÎbœÄz³$æ÷Cb¥ZVú«Üà$î'ÝÆI#Nò~p2²…Sx7¤BÅCg2?¾ÀC!²_Ôžÿ¤K¼w6ÑÇšOIµÿ>Lž@Ë()ò`ŒKÈÛR2É@ÀwÞ¬<ϫه!Ï…=–ŒÜœïb4uO²l:²l¾F–ž@üæ‹i1ý~uÊ¥çÏØj±.ëÚ›þê&Ųô=¦‡þùõ“ÿ>žæãjvIÞF@‹ÔÃáU± Íî‘Ä„x£ û•·W¡Ôó†ÀÔ7¯Q½ýZWWÞºæ2Nøý|tvó{ëÆOQsWÈónAê½oTN N¤¯§Œ‰}SIXëëÐ5‡!¡gõ;´“0ÁÒ›kíUýÒ÷]ÃzË1E+øY×OËêòª.‚ÀÄ?¤éãô$=K‹t”ŽÓ2½¸H/Ó«´J'é4¥óô:]¤Ët•®>ÍÓéçôË^Ž<èðž0“â’Ü:?-kþõïàVùß…—OªI‰@Áê ›¾(¦å6ãÃ'¬F³K(zOªålä§ÛÓ =[•Ó_(Šj3D‹—zŒùèí“çg§ÿ-„jÙÑÛ¹…3[-r¦È™F›oæFrwˆdÁíÁÃïtZx†÷@Ò bü‹ê#ªÕ‹‚ˆ«rÖ’„ù¬„4 ²¨æãôO/Ëò#ª,«Ï$ W‹²ôb²† |"QIÿuaë¾SZ̶´lqóPqA,%K°P²¾˜$åxþo¸x\¤çWóéõr>ûŸi1ƒ£‘\/~ô¬²¸¼=i%àqe2É0•éÄP¸†Aq£ÿMƒÀ7¹ZM'wÂ<2ªrà ò¦ƒLT&hÌ<ÑxGoN~ýí% Åùq×#¨}“›"‹˜Vÿª8œßG†(äÊ÷¼ËºƒUÔ£¥ü<cC$°Ð"ø ¹ mâ[ZÀ4·©Ûïð"ü"ˆ2ᎮÔs]¦û€‰ Š„Ù³oj+ZFìB‹Z†–ÆBU´ìŠ7=Ïûû=àóÁÛ?i¼ü_|>)8º>«Tç•ZÞ÷&ÁôgL1··5ë=þ.T“ Wx´ºnJÌ´¶Ü”-îìðfCÞ£§¯Sç]ö® ÒMo\%j³·êðz†¢ßîÝ}}ûï¿ýýfAÏÛ½o …ëž|¬¹5Z"b`§î¿î„Æo#4^ŽªjT-Fëi¸_U“q‰[ &—(`.Ãæñ|2Aý&vöÑr7áuIëèå¢!OÊåò~Ãä~P|S(¼žÁN-GóEyST 嵄݊ŠÃKr¶iÒ¶tï鈮oëˆâskÆ»ƒžDA··6JôÕÇééÉÑѳ×k«úì`¶¬66ŠD™¾"1=E2t%%,‘k«;߬õ»yÖX¤ÆRå2Ú>ž(YÛDžùýíð9¿)H×Î×!£å¯”¹ÒÜwLå€fæ¯Nç¡N|gÈZÊ,ã«lóÌ÷›í¬»ÙdfmÙ=„Ìû’:ÜÉ“ü´—Kº"RgŠØ0ûv¶çKŸ£7Ùb¹']¬ùŽtq.¶Yz‹­›½Û£sø}øûÁÛ/|o[NÞMll\ŸógÙØ~õ²‹nøTË€³æâʈîp³búéRãºV!ï¤KƒŸÊ˜¬ùÜžxÚZSO¼Øžxzéà.ì­yo­žGU¶=Cg^ ˜ùêëÕÙá£7¤:Ï~ëz?nÛûÉzã–—¯šìV~>ëÜu?ûžßÏç{A¸ýUJØß-OOŽ€êc1)g£2V3˜ééz²ª®'_Hk]—³q5Z“Ç@&øõèÍCêàù–‡~“nÚá««~(z_½ëeÞ¯Gû=|õý²ÿ?yéÝå«â?Ë+Ÿ­§ïá$W—³ûqÐïß+w®ë–‹,¼6|§_ÞÛ!ÒݶvxÀïà›ßnÜ6¾9±%ŽC†¹¨ÕÛ®ö¹6:£;\N+²Ì¯±Ö#=)À^I&Œ]p,Ú÷ïZê…Àu´ -åÏßÌ* P§˜>Û©p”LZ^ï©£M'ó”¶ó:¥Î9í#¯ âæY*Hw΢à½ðº ü©€º XÜèJÅâ.W*·¸R!gª ý€R÷žÛV"¡Ä˜–*¡=øV'”S&ÑCÓ¯õ† Ô‰û¯; õÂæ7ïÎlùðt„á{m„¼ýò,Ñ”÷ ô™H”uw"àPYªEÿ¯’%7X–6ŽÅÓ³ã×'o¡Nº!ãÛƒÙ‘×Ö½•îáymI9cB¶n´kü~“Qò¾ë5jZ¼C×­ûþh×' ûWrf¹ÚQßÙp5*ìœq·ö (3í3iY ôýØ´W?9…?>y ýÙEÊÃïuÕt(ˆöàpÊQÐ1¨å®mrB.»˜\_!ì/W>}UmLw9Y>±}YГ FùCq}]¤“bú~\À4O×d×›DÀ5Ú“aÝ·=ŒÐØèá-ϔڑèn616Ær‹¿9âj;:æb¨QèØÇTc!hM#ëô¶£ÿýæê µ»ÿU'¯ õ»Q<ϸJä ÙVÒ$X&‡6æpÚ=Sªr“äv êl,Œ»§3 Æmx†ÕgIîda{¨ùu€h,žÅãFñìgw:æÒÝÓaÎ[;Rè”àW gà¬ç›Y·.ѹ?ò—Ð>™í“Ö&°—PN‰Ñ´Qù~e`AòoÚȱS¤pàDƒnòý¬Þ³‡:µ0Ût–0Khë¯Ô.Q| 3¶¬^9ùX®ªQѶxqNo0_ûáêžÊ3½ãÙ¶ß°Ï(çtÔ×Ð>æÛÝð‚+CÛOß¼Íétj߮ɮ]Ûà†øqÚÍ4ßp¤—ݾ¿ŽÙé%»”œmT·dÍ.°ãx¥>’r†¶‹b¶ô1îèK éh1__÷Yã÷âúâx_@ý´\N¾‡g7^¡ëœZm³â9žAÈjÕtß;"³Ër 5VËÙŽ 6ö…œNo7Ú¸ ‚ÖRjJG¢mãq—]%Óbu•|™/>¬“Q‘žž¥Gãô.ÛI,¥‚÷è\BGêd°‰ßaÌ·cÌ…MèÐzuŽIРšÓå~[×0…Ù:Rµ‡?ŽÏØE1YÒQ˜D.¢_3…š¥;[ðÝõ±ÁÄ VŸ;¥M|õ¡½SË\@Cö£ÇBŸgltº©LZÀ4%:Žd›¬Šk:¤à\SB°’u=»ûN;´Ê½ròmÒðm²ZÎ/VÉ|q™~ä6­2y7i `š¢h 3 ƒ¡¬7ª•H´ø{Û¾e+:^­yž8úO.K8ý猀g˜ôñ fùQuqQ’6/ýð,giaùæ¬xS³>SÜ©OÑ=åöV¿[﬋ |+|'t_ùå€ÎîjßçfK¹@ô_*X<¨Àâ¹…ûÅï/¡ÍÖ¨ÿ=º†½™º¸ð]üšÃD×ëftü½Y,£U{Ž6wÖÂæ÷÷>!(qq-¬zƶÿDŠ0Ÿa>)ÂBöƒ…t©éa1 BäaélMÇë)ýÁ(‡BATL‰°OÊ¿JׄßÙàÿOXuŽendstream endobj 93 0 obj << /Subtype /XML /Type /Metadata /Length 1557 >> stream GPL Ghostscript 9.21 2017-09-29T15:03:27-04:00 2017-09-29T15:03:27-04:00 LaTeX with hyperref package endstream endobj 94 0 obj << /Type /ObjStm /Length 3071 /Filter /FlateDecode /N 91 /First 836 >> stream xœÝ[msÛ6þ~¿ïæ¦xétÆvâ6i¦rÚKÒÉŦc5²äHrêܯ¿gAÑ"Eɦd¶™¹‘%‚$,vŸg±XÂÉ1É’gÊᘑ8Df‰yo™Â•h5ŽŠ¾TÀWy*¦L¾‚j6ISÞSÁ3­òSiçPQF¦£AG2¡—ˆ:J2£ 3&àÚ2N£ e˜ Ô²²Ì$QpÌ*G·<³O(˜šnE†‡èñ„B€,Z2gé–VÌ…ˆv´f^&ˆ¡ óšäÕM· ‰®xæ@4UŽPTC‘ä¸>’ÌF¡Ñ…Ñ,HC¶ ‹‚(8,Ά\¤B`!R[d!ªœX”èYY¨Yã¾² )Ô-5h ‹VQÖ²,Õq,&GW< ¤:\Md*e#KÚQ!±d¡‰lYK:Å’C‡Êi˜F³ùâèr8ÃtZöø¤˜ŸÍF׋é,Çl¹ÖOêæqzó~‘{¤~UÕ}ÙÝFç‹KR‘§Ð+‘btY»|t¢àÎäounÀH::LÛ¾²öKué[o7 Î €[TTË'j=PÛQ/k¤ Wµ1ïÓ艶"÷ZOVG=U׳ìtîË2郮W÷é:Å%£ñVרW::_ökèlð‹ó3ÂR@('ކ×?£—Õ),EHù§8Gâ‰8?ˆgâD Ä©x%~Cñ^œ‰³éx:ÁïÕÕPœ‹BWçÃù¥¸ô7ùïs!.RñA\ŠË/×—ÅDŒÄ⣋+1“ѤS1Åﵸ&t‹‹EYš‘@⺘¦çâ“øt3]åµ™˜‹yq5*%˜ŸÑê|L/ÄârVbñçT܈ÏâOq+¾ˆÿ³é¿˜ `Ò¿¡XV‡@ºw‡%ݾÁTœïfU‘.Þ•÷Gã‚À¤VÄz1¼*ÖØñl1Î&ÆꉓÑ|®dD3‡@Zœ.Ы߲ï«a¾Æ—ù¿ñã4zZgX29˜ÌG« +"mÓ0¶h(;Ò0äg³ÊñÉæaIÖm—½[†¸÷Ô3žèc"ËÆ´üm¡k·:jóéjÄÓŒÕ!Pz´,€„ÛTe„‹´Ô¬ƒ@æûN”š]€^@ÓB] Õˆ×Ë!Y€¤‚Ãtv^Ì–³A†äM!Õ ž&§®ªpüõ›·´€å´lóÉq‡%Þäf<¦ºƒâ,×’§@NrMëTóX a©Îaåw5lü4š|¬dÍÓ×nR¹¸*$îh£7ëa±Å=b“J"m§uyÿ©*”®ëɹÀ#ŽëREŧÀ½¤¹ÀrEëséyÄüßM¨;.ÿòË“7¯ÉU„ŽT†@-*+³Fe(­“–Ÿ]¾ë\óˆUj\+O︖/Ö}éÑñ]ºåÑmÍ%·ñöÎùØ¢!å\\îq†f† åvdaÜF–鎞¿yr|H­ûǘÎík:‰§Ë÷ï4ºßtM¢Üd:»nº†r»š.t5ÝoÏ_¾ôíVkÛèèØh‚>WF:\µ_ŸÕ1¯—(ì€hpþ¡èûáL¼Ÿ Ï>‹=.Ëe¬¸!Nýt3S˜ZÅ¥£»ptÚ! Å,2B{9=?ÎÁhvÞˆ› ¦ ùÙtV öüÒ ŒŽcŒiF®¼›}²F˜Zú£$LÒëŒÞÖ”Ùä]yÀõ­â«*[ (ÓYzÜê åàU•áPþ]UùŠeÈU¼;1LËêsé2=‡+Y•NÐlïîКåîJ¬Va× §T+2ÈÉTåë’­ˆÿóÑàð·gYKx¼£×–mækÕb¾êÊ|Hê²´U lÙfžöû©ú-ûkÿ6k4k6¯¬®=À\Ô¨3—n®˜[® mÔ‘h%©nhõÖäPÚÆ!UÞ%A6p¨–+9´†Ž­$¢q5IDWJ…m“PD¦FÃVÈÍ/26f;-zúÕIbþŽô"eɲ{8»OŸí–ɬ”J*³NC“°X«(ØHUkåÕ¼š&ÔªóŸÓ•¹N†ðÚ9‚àRi¹dK­ü®ÆCj®AC âÕô×ÉÂÒmd¦5ܯ ½ÑÁu³Ö=Í26ê´éÑjM¯Û´kè{•ì+÷ª¹W³ì|§üà&•ݩͺØiÝV¥_ó²ÍN)L<<×Z¶yɉzÞEn½Å¹ç Ëc$Áv[¸µqNùµ 8—à"@)ÿœWØéçÍY½«¿hד¼£¾979_Šë=RË5(¯xðôÞr#ý:B å"àL½Ó\£/¸0î`"ï —A÷™oH2 :òÛY8½è³IqGo ƒáÑÄB°¢ìÛý…Ú”–!¢IO‰"ÅuÞš€ÊP—„Ú€¤>%r+‰> stream xœÕZ[oÇ~ﯘÇEfçÌœ¹AÉŽë4VáJNâ4ÈC­m¢2i4ìüû~g¸K-oÊŠ\E-”xw–³3ß¹_f)Ee¥¤,%\³b²Š²QÁ\IÅ(W«2ËÕ)"’VäXfzEÞb‰Œ§!ce=ü^Fæde øÇ`’RÖz+•e§¬÷øÝ°²Ñãã•Í^^ Ê‘B•³>â&)ÇåIV.æÁ‹²¼rÉaAárÂòä@““Vl˯˜Ë“€Ù”¢âÀ¶0c–×Á‰Ìø ´y“冔·{Y«¼“u¬SÞ—ŸXùÄNhß@»µ„8UpFÞJ*°c3n" :09퀢p˜BXÇ9Ü$ù‰U4Yžx­lႊ,üµ‘£¼•Tô0.FPŽª˜±eÈ/3öm‰À Y"Ù€•™UbÆ^%ïå­ R³-^HI¶à¤²pV™„Qø cŽ'•="RÌt[ïT˜hñfŽ"sïU.»{è†!!ÒÃPÁ (/fÜ1xhôÈø˜@ˆ;D€™ùZ0†LrÇ¢†¢ êGƆ w{ìAc°y’•±ÔÜ£ˆ’H6’Ü åà"AU±^Ï"‹({d! *‡=Ä XòF”»ÿòõת:›NgË…ú¥èµQ—E­WW×\¹¹úæšk,×_Uõd6]ÖS¬Â›<«.êëÉè|öëü(4„*s_Žæ˜ªVkV—õböq>®J°|ûyù÷«åhYã(žaéb!eôr>_ÕK¬Z½|úLU¯êÏK,úÍ7¸ýýC-«¿­1Ü¢Œʨ¡Œʈw)pf ÈoPN£ 6ˆSs]±G |—ˆ»A=»G[œ iÚ@šOBêöñÔ „´%y…´]öX¤iR)›.R¦“rcyÜX7>…ŸÂOáFϹÑsÎ{(ôQè©K¡·'QèÊü>Ÿ†B6ÇÓ7<öûxB6´(œ¦E¡Ñ¢Ðð:ìãu ù†÷ Çx¿KÉ"WXÛ'˪—ˆÂªñèÀàZ€òƼ ±£v$§kQ’ÛŽ ‡õ<¤Ã ¿Ú½%àõª’pkä—’ðz‹u™íz”%çlGÉ`Ä›T!W îA«V³z“x€("©gÇùl~]ω›¢e@«ÁÓz±,Š¿ZàõÏÿVÙiqmH¢tD–0ýxs#3/ë±ÌíM¦¹R8@¯ÅuaæÕÇß–Eî/&Óÿ´:P4ö¡@¡Ñ ÛmA!¥×NdyPgEÛ¯TõÃåwë ãåd6]=úâÝrùáoUõéÓ'ýañûøÏæÓúæF×׫§£ùÍdúv9›Vãùl±XŽ~«^¿øÖèç¯.¾,Öp˜lê“ à‘QgÖ°jÏQ'Te´D¨Äÿ-DQ“dßÉêŒ0@´ AE”A¬-&k ´¥-ÎF¥œ°¬%…s>jFjŠRL'zp>Ê'ÍRí´ ì,Å!@Q;h@QŠ·¨È‰É”JL'¤Û°’׆:° œ Ox/XOÞæ$VñdçRhœâËt<»†º•Âl^&óÅR^gX‚Íù*j!å·å?_¢–ür1ZÎ'%¢çª.3÷˜÷bÔYn W€:\f?L'@PKë`µÿŠ„䪟&×Ëwi~®DµƒalÃá)Œýz$%ñÚËŽwˆÔníØ~R¯!%©ŠÛ}-\Šô]þ0TìjOºÕ”:J÷ÆE-ùÀ¶öpÐFzÖC[Jî´1ÒA Zúƒ*uV„‰%©ò ·Q1t:I #$8Z„®Œh]¦çdOrJ;¨š†ÁÊÔ(iª)-M®mX.ëK_pH±·š“ô%À,~,TòØ‹æ|é<Á_¬@FדW÷4–ÿóØH±¤Ù]?>û®ŸM " KíÜnã<â" *¼´õV*é É1Ò ÂÏt‹*°†ÏgŸuN»ÖKа˜"—tf­/Á›˜˜¾™Èý9e±-K«ÕÈö;‰ E$€¶Y¸I:‰–D£EAõ)yO ¤¬w£7¬Bš¯„dÍ@UÄÑI4.öÞ£E½ :—Ïž_~ÿú¯O.®~N’·¤åU âñ|òa9›—~v™µ{2i#ìÑ—l'Ä-ÆR:ØLõdôáy=yû®bU©ø¾¨“÷“›Ñü˳2ß`Õ›Ñ[”¡[8+I—¼ß„îg“`æNÿÏÑûz›üï–£›Éølúö¦–Å/&‹˜Q0òñäjY¿ÿ±Xr‡ªCvØûâêìÿ:Çú篻ì¦gÓÅäöÁ-Ÿ‘ïò™·øŒl·'£‰J‡¡ÏŸÏ¾\¼•‡Kr‘;¿:–»mËùòBiF¯^Øf4Ôf€Ñw„yQªëª®ÞT7ÕûjVÍ«Á~[É“ÙïWÒ¥G* [l 8n xS} íÙ/àêuC{?xÌAB`×ñãŽd•[¯v8õ@Žl9¢jG\ާN AàX= Ì„ü# ‰ŒÂ¹]B, !— àèPT<ïÇÎàžœëÜŸG÷I‡\6Z:·."ÿ‘ÓUÃ(l# š¨Cߌ¶w=Ô6c«j5D=Ô·Ç·Ñ ëŠº#]؃3嘵³ßm»“‡:ps­ewLk¹·¨í㊺µ¢!D}”ûìˆÚï¶±þ{¦¶öÖÛÎÎÅÝ6f´Ú$9\vZ¾tHV{iÁ¬ýÐ] ÷È¢¥áDÛ7hö´é¾µY§ƒzÈwM<8µG1êw;HßåôŠ¥)Â8ÒzÅ(1lßH88¦´|ÍÒ@òˆS&§? Tý»)ƒµšJ«¢Ũ uLÛe-g/»šªÕÊhT«AËg9.–³ç¤òŽ‚É¸Ê·˜ø/7¦Ýd¦“Ë ¼7é`ÓA~öøÙ‹Ü%–漇¥9rHº~@Í4–ËWi€,ÅRU8ë4Ô/<QL:ó# ¤p/D½#E[(=V¤°'EŠÓk¨^y¾ó¾|·vDzе&ÏGy½Ï`a„À #Õò ¥$À¸ÚAOcrÞ%ñÇÛƒuewùZ3i+ÍB$-\” §Š'ÄÞíÓŰ­‹2·«ŒâÃâQªXV»Kù.]„£:V¦‡OÕ](³£q/³NH÷´e«åó8ŸYsi·Ã'#Ñ &œzPz|XNw¤åé½×^ý`&Ãû œ1õlµ>L#fàóŽæÿ†&lV¡Çž ï$„ü¤|n´Ý|¤0±ÐdTò5¶KÈ3 U!Á>›Ány7•©’V ^ÂG¯å[ ¼ ÑýÉAž|Þ’ BÌ}п¸#ží×ËS™,•mk I8vuÆ€Ã7cä/¡§é|WÖ|uö| ç‡úÚ06–TÜ»óÇ|£:4‡9"^­¹‘ ÀïÃÅ]ez:yó¦K„¿”4L²À_Û·×a÷O™.!wzl®&gGÍ!Ïá-™@ÉD†Xÿ¿#½qŽendstream endobj 278 0 obj << /Type /ObjStm /Length 3796 /Filter /FlateDecode /N 91 /First 830 >> stream xœµ[Ûnܶ}?_ÁÇ‘x§X7nÒÆm`;MÓÓ<(cÙ:9#9uúõgmJ(­S8‰ºíÅ}YÜÜ¢¤óŒ3™q&¬ÄV0i ¶’©ŒÚŠ®±ÕÌ …­aNyl-ËL†­cÞR;c‚;ºÑ3!•eÒã‰*ì&´Á½^2aê„y:¥™È$í&|¸ îð`ï˜x˜ôh(M× ¡cНÁS'°¸UqBª3ì„Ghìà'è?ŽÓ¸;–)e8vð 9Šã+è~z‡;˜ò‡…`šCJH¦¥‚2pZ+‡[…fÚH:e˜¶ŽNY¦3‰[…cx0]“1#Â=3zU8m4=^ f õBJf\@gP%tÌ 3žn…A¡KýVÒ1«5ídÌšpÄ3ë¨;Š3ë9íæ8]¸NÐ5JÁXáˆfNC‘PsV“&,sŽ” s^†ÊXÆaófRcGs–):¢Œÿ”–,ƒ0ì(–áv4óÆWx„§ód*¯É(®á8ÝE¾ÁÉ<À"8i !xÐú)¸‡ë(ÿŠ O‘¶¡!<:« \D E÷ZìiÒ3ô(¤£'ñgÑ!Cè©×–<“¹-dÀ^doI^,*øüEàbÈ…Å<¨ð$ìA· rŠìc3Ú#Ìðx¡¹Vÿùö[–V·«†=ËXú}Q7ì,C@°ô·÷¿{–àÓÎ%d’ÕírÉ>°ô¨\ãRŠºpéë<´\Ûú¹¸kBD…Ö›|]àù¶k­‹Ï¬»î¬l–{r¸.ò¦\]²‹uñé¶X-¾°&ÿ¸,ê§ì»ï—ã^% ±ÝãêD‹héü@à/ëób]œ³‹|ÑTë‘ÁÇr”“‰v#9rZNèäF9ÔºYß.BÇþxòÇÓQß̽P«ø¹qY¾ˆEªEŠÈ4–¯ÐÝuY¬ÎÇZI–\% ¸H°~P°¾ƒäz²›v·¼Kˆu"yæAyjèBÕr™ßÔäDÕçb½Ë²Þ¥[áðu$Ü>(\ï¾,>˱Ø,Ö0H?1zä¶îA©&’ºBO)nFâ&âãWb@ݽÀ"ì@Ä[T×7ËâŽw9íl$(ü8E:ÙŠÍž3ü6gôÑjÀ5°ÍqŸApͪ öju^ÜÀ±AÅ8”§`i›H–çCå÷ôµ!©ª®Ï(žÆ2ÄȼŠ'LIÓ:Ý»wxU>«?ÝâRÖ@Î ¡>¢(/ê‡uTÖWˆžwv6)VNDФ3{õ à!Q竦X>{™«úïb¹£Ëj,[BÓ”E²õƒ²‡\uxü2È«¿aÕú¼\åËC6ÅJ%™ÙË› Ùë¸ÈëÛu¼ù ®«E‰ñ°ZÍ/‡ÕGÊ·Š×»Å_®‹â÷Œ…HLª,AÞ8â”HXL[kˆº©ÚxŸ”/¿Ôe3‹ÚRö0»;½çÅ.ÀÆ%8ú@ËuÙ|‰Bà`µBÂiNJ¡i»•ݶµÝRÆò®Ö¦xŒ®äÏ«;ÜOw[ÌOØäÞ&ENŠºº]/0‘ÌwͧMÞaZ.8ª¨v£ÂjqZJÒ7ß¡_Ô뀌Ý/7=ý²ˆ”G3çØˆ(±5€Çêê,BäÙǼÆ4r¨ªçÍ.»®ò Ðm£7â™*Ÿh­F¹Ä¨{'Å"LÆ1 ´pf‘HI•†,±F0­À \wzû± Ý}]®þì» ¶±^Ðç)Kßž¼Ú\° ƒ·‡ž\5ÍÍ7iz~ž¯“ó¼É?—u²ÈŸUîîØB)•€70‘b$÷WÁœ–$úF§¹! k!cµY•P†’b¯©RåReÆ)\¬ö hFµ$¤UÂaháaËŒª@2ñf¯JšQ)é0Rª:Q+€ E(”¢Ü'¤9U•“¤Ü˜´C€F‘h¿W5=Zt‰qèƒ.Á·TP%„F‰RT"’äC>T I•= 4n¯þ=§d"ΰ((?„ $1ì>A=^ɉ0aKU#b O •¡­²½bz´ÌAª™ ­Kï(ˆLìÒŒRPÊÀµ©Œì|BåiiÑÞ3<^*¦Dβ¤W-‰¢w+m±WšU(Š4%1q>˜Ïa’J'4GØ+ªGKK&¡àåpr€ñ$šj¯#ËŒÂS‰ÓYd Žæ-ôRL‚©ì~1ͨJ QÑhH·‰Lc"”"Qûv©õª\ [‘”å9BÉz3´OTsjX}Rçi%7I”h#©I”ÙëH<«¶ÃBÞ$rc8½§w–’Íý¢z´ècÒ"x¹F¸£7hËýò眂X E(§8z-ì(ã”ôÎ5±j¿18£&ë ¹ r;¦9÷»€’˽fS3Jg(ÊU¨¦¥1+#§Ä„1¡û{õh%-•!»3´¢³CU«„^ïÓŒZÛЩ¤%ÒRô:S2,ží×Õ/¹Åššpj0š5:äX|¿“†5¹2ƒDOLfºËï—Öç”äbXБ7T(Ëh^%1-õb¿ž>§.סR”½Ó:)ª€ÔiÁ‚ThêýfUò"O瘦Â|ȇ Fïyø{¼¦Ì–hLø4R=åMH`´Ü+'̨•Å 2žhZ:óH¾.Éë"øÒ7?ýþÓïïÿ{x|‚Œ,}±ZTçT Mß•«ƒU]ÞõÖë|ÍÀÔ¡:è‹uyÓTë°´gS†  $^÷@‚è‘´‚ß•çÍUfáºÓÍöŸ6¤͈vÛ£–Ô4*kÌaÚ½pEwއe@´°Guϵ½ûò\½…a]z˜ß¼,ÊË+41YM 6•8Ÿ¤ïÒ<]¤çi‘^¦Wi™þ™.ÓUZ¥7é:­Ó&ýëi°[xÒ3äÜèÒ2¿¬1¨¶{Þ×_ÃIZ0eÛ¹ <*—Ò;Óå`tèçüºˆ-ñªÉ—åâ`u¹,paz\Ö´è!(L ÒÓ¦¸þ¢Z·L’þÖuŽÊ9#³ÿðûÑû7vz:ßîrÂî6²;]4Ëì–ÛGÍš³œl©÷:·H’³HíðPDŽγXç‘f*Í´ÎGj>>ûñ·ÃƒMåL-k=¡å,Ž.¤nóÔ,˜ˆ®Çþ¬òƒvj@÷Ž;Z-Ø·TÁ:äÍ:°˜ Û`šnÄñ¢º¾ÎC4oÇrÉŸÓ/_Ëà ¿ò±_ í>7–1¤ü“X>zñ뻃wöü7?×ÉüØÉ ÿÇNFÕþ©ïþþe!'Ï´¯ukl $xpƒT;=Õ8°+È”VÙ†--+µ½xZË7ÛpÇik)GÄ”y"~í¹Î÷¥ù›#+Ýû[ÛÜøÛAz˜¾J_¯ËÓÝØ.M/Ò‹òs‘^T·kŒ&eç}ÕªÀhò)xa]ÞÁ›«uQ¤Í_Uzû¤xÈ'é5•³ÑÐ)»—‡[N9ô“¹Ni`Ì„S¾x{|üþ%EÀlŸä>¹µÔ óɹ㋢E;%ZÏGëÛu}‰÷œ_{õé¹ð¿&LôÀºû£íŸÇhLmçýÆ=û;v¸/Ér<è°Ý¶«Æûãâ{Ö]ÕŸ¥ãšÓÏ„¾öÇH¢é~ôä‘;›¡;ƒa#w~} ‡þ9ý%=ݸõ¢ZV«-Brñ«ôêËÍU±ê¸öþ¾*áî½ÓßÐ+ßeqÑ´{ëðˆ›b]Vç}@Ÿqû8,>§¥wé—ôïôïb] b$ðÙ®ÁÉ 3@êt""‘×ÎŒ~FˆšÈ Ž^üprr¹!2•™8óžŸYóïe`Tò”Ý †í‘r …™*Çlfvðöí/ïÎðø³Ws•l² %ÇiîWð&G.è€R&ç̈¼Üº2¤}mrÕÞ5Á’ÎÒä4=Â&zÓîeËp7½¥¸¿^#R4}Тé#“`²°°œ>ý ã#Î$ñg|¾8WüÚ¥]ÄÄÝ4ªú>u©×-¢±<êÌŽ0~†Ñ=œ\dAväW*ö«¡áç†2Æ›G30Ù+æ„^“µ‚û#"|ÂÒÕ¬N譄ЛFû™KßrLðû é[—͹°(ÀoZ*|÷Ò·Løâ¥o¹ð­KßÂSlß ,Ù¦Aï9º‹:‚=¸XÓ÷A›yCßÈèÅïöóŠÒn¯Û°"±œ¾À±áU^\Þà<¡ü ƒkBn(¹Kè“$›‰°bg5—¯ &Dµ–ëÄ úB0éû*a½×—~Òlƒâ ¥.Þ&Ú–%HÃôU“׉P2¼ÔŠJh&ñ~¯ÅÅù h„d™ ƃ‡'ƒ€÷ô¦t¢ZlÖ®KËé³±îK Z µákúªªø o+?  Z þÙ“ÅqÞ¬Ë0@%\HÞñòÖþ‡-†Wqý S̳êíª‚ŒÓ¯lœâ|­’­¤k’'¶™awÜC‘^Ãm`÷ a ‚‰?•;¡äPuPç¢Pô}ã#húUåÅE±¦•æmúÝZïC¯¢épùÿY‘Ï]endstream endobj 370 0 obj << /Filter /FlateDecode /Length 6488 >> stream xœå][sÇu~gåGìãnʘôýâ<)–TVÊvÊ2ìT*ÊM„(ê×ç\º§Of ±K:•’]-zúr®ß9}ºçÝJ z¥ðŸòïÓ·/þåÇœVw/ÔêâÅ»šþº*ÿ:}»ú·cl‘W:F{·:>Á¯êUʃvƬ‚³ƒ iuüöůÿsÓàRÖë›Í‘|Î)¦õŸUÎ9æ¸~³QƒUNg·~ NYÜú›$h®Óúbsd•‚ŠëÔB“ qëð½ìð’›§¼>…Ö:ñz}2m­¡hòjƒç¤36¹¡qœŽvœ¬‰0>ÎÐæ¨’…Îð9g¡w|4 ¦u…£kè3¬Ï6ã0'ÔŸ¶6Žs¥þNx®ÿA Ï„ð>›¾ù%þGVÊ›ø?Çÿ¬ÐJK^—ç2ðãø ¨ÿc›"7ŸpÎA;][‹9^·™Ÿñz²Ê0¤ªQÀ’Ë­„“Q´²r‡õ8ÌÏcOÁ3^}ran¦VBìgj¢÷,'Yôxîƒ'5;‘#ý$mWG¤ÓîùàQéõw‚/¿À¯°~KB¤†œ‘%$B ›XΘts ‰vˆ!šº[)!m˜Ó ð@7¼LÖ¾ûóENAûõÝC>ø•†•zohù  ?2y°ÁdžÃñ5£”6™äN)´–g‘üLÜV*Ãbi‰$ì*Ö—wQðá ÉúûM2ø€š ÌA9Wú6Êm»¯å gÝPWòÄ^èV{\ß‘¢»´:ÒvÈ®0ó¿6Iãë†4¸ÌrÛ^ŠÂ,=rký×ënJE|77œ¿1²…¼#h–Þ%«QEñÍäL†7 Öñøµ>Þ$>÷s C–1(eSZ sV,_TÆÆUÀ iâkƒ=IµFí'JüßO6)îŠ ñÞ™2õò¤=`€Ðø’ù3 fÁÖQ²ÅXP-'´˜,TŠžÕF£Ú%fM5ƒâƒbðƂ޴Ÿqa&Gm5AÁ{å¬IsK´0¸±I˜ dŽ‘yUÆÞ²êQwwâùµxF„4ˆ&¬Ýz7صDDÛˆó×9y·>Ãh~–À“.o^“=ˆ,ÆYûyŽ’Á)%ÛS2Âÿæ(i7sÑnH.ÆÃPD/PÄÍSĉæÿÊ}(xüV¼yBoªžý½x³¨šÑѬ»90híÖç¾@ëŽ9š%”D'Ò´-Šn*²yÏGèºÀXƒ,šiî±üT2Ì“4Ë‚õ´lŒh‘ú4É9í% ­2>}G}}Ï«Ò.WðVÔ"B !SÆ µæuVÎA{?€ƒ‹ƒW†'q\‘ /¶Î^+ʃg`ƒeÁ48@"¤s)£×÷ÔZ+ã’–¿ßTÀ•ÙjçdЧ2Ê‚öŠç¯aœÈØCßhQÏd‡§"À ª;J°HZ.ÈMDg){‚éNÇÉ›3¿,ã˜t(8ØL-+M±î[ì%­I9QúLÀ2€óÊFãIòK”ü`Àð€ÂR霭֤d÷-½œK²ûk93¹¼[éÙîe««Æ¦nžãje74Q Ÿòr¢búçL ëóm1DìwB@Ëø`¥ú¡¬­0¢‘ë øN JÒÑô#éªÌ¬¹P甜%01ö´5 ¦¦ãÒòØÿ{ç|°„mÊ‹,s&£ÍN®h]1Ì”(;ïôÔ_õJ ´È¾*ŠˆírdtRÔ¿ ã§ pTVÏ·'5EqÃô‰œË\QàtÌè”·o*µ¦ª½íÔå†ÜOvàÍ’eQ¹Ï¸KÆ)¤¹1 #l"C_ˆÜw”ÚYEv^ ;ã0»d ­qæÁ; .iLˆq„r’$xz³ñ˜C ÔÁøçλŽ:Ö’%m<ë vó]ÎŒ›¯90¨®ælª N§ q¹k#‡¿¤‰GîÄLøpRjëÓ¢-JLÕjRMÄ×Lúl•éu±áÁX°1½î ïV¨FèÍœo½*ÌŽÃ{9ì·ó“¨K&:è¬Úš˩qJוkŒmÈj gU„ójÃ3;0Xåçj B§n·hì%÷$ýÌâ:\·ö¼t|—.DÁ×`z×gëcžZ g .a§µ˜èO'¢0+G삌–q8Gß5q¥DžÎ þ‹–À±AÉãÅH’‚WˆËgeÛ‚¼DûÉÚØÓº×F$¨ž€_¦()Gèý 8eŽGSóTÇ3)ÎN‡äM¿°îå)b-Ä«.£€jõÐAŸ4UîÜÚ^Áwî]ƒ°~lËi‘Î1»¥vqèg-`wŒmAÇò“µ•Ý]bÐð„Õ6!”ŠÁ¬ÃÙÎU×äÃg¹&˜"fâ_Žá•pûsê¢æáUŒ”¡7MºÆaÂb5,Öbšê2ÅÁ¥»§™èN•Љ_qg:¹‰4ÐϘSê$íf„MÊ,ÌœÆT\»E™>ø ÆJÙÎðÝ“¾š ä rÈ Db÷ØGóäwÙ˜’|žú®ó7zÛ mž;,E}óe7Ñ:ˆîÞ§šÂ¦¤Òvà žqË5õ næQíQ7Z¼!Èø©¤Ó ü½7+q¦Ã±‘t?m„a@ î®àŒ½ô+FF%Ç!8ë±ÐËE–|wïœkw/ÔË.ß[ß\òA#…¸Sb&&Ê(?¢Ô'ˤ¸nOس‰ØÌúê­Lµçãç“l%¥³ˆû€j e"¯mg‚=_Ÿ ,´]ÊóÉw·%*Ha¿ovìDNkT{™Ê!ÂË")òJ»^çê ƒÒï ~±àñ@;šµõ´ÉÉæÖqšàzƒ%+91¦‚hÖšZô€V}-Aùã]qÍɆϰz†k.¬_Ùà‡ì˜ezs”ÁœЖÐVéÄ—ÕCÐi܉¬YrýàÊ ID÷t&¢±{¢^ Áò~fÂ…Ä®ª—4côF×­“'JåÜúh?Ĺé ?ßí "¸Ø a‘û6œò½^1Xsâ«’I`@7%IÈòV¥„ò²h­ Œ8•ct‰w©s@šò¿+B(cKånŽjX¸²E¹%¹9Vç=\J;ã®ëæuÒwìQ[Xm˜}]çUnÀ=›—á¡÷‹&"XS°«+k)õZG}ð ˆ ’¤•%^ÿrGÿ€gd±`­.t^p;Éê±3óëóáäDºaiãýî“5cÖõ‡0@Õ¯†Ø­ùÑ h‡¦¿>Úà3_µG¸Ä5#S¿´Þb¡’÷;ykFÞŠ¼¬Lêlïçy%@ÍSöU—>zMÅù“Y qBP¹_æ—áÜgö[Æp ‘€“] ·#Ã÷ÈL„"àe?K9ÌsTÊl¨H…áÙô….k¸K€õˆ>õDúâ2RD L°@ÙÆ"àFø…ì¹ ¶OäÎÕŒL÷n¡Ð€¬jvkøšúùýšÝ&C¿1íd·Ùý;J+;O»°/JÜndàaIP ¥ž è{JN$•9ª¶É–d d–‘Øl¢¹/ˆ=bÅž–¥"é`2‘Yy§L„¯*ÓÒ³" ´”ëj'ûu/üŽX”{:}UX ¨EÉ{Q®&!j7BˆÈF©bôK ÍjÚ:·Sö‘×C¨M·¨<°›Óú`œö KÎw²:¬þ¦€8e¥å…™oÅk¬û*耹«mþ¥d·rWÐ,·ñÉšhþäÞŸcOŸ/J‚yŽ$,'lŒ€¦JRȶ„Íñ&G¹&1f¸Ûˆôacæ'Hbª;8ÏKÚXüƒÈšË$Ë!L×'JnñDÐ`c ËÙ™³–Q{JdC?+ébíRO,ÝVê•1xÈ)úÊGò/Í o¥EE¬²Dä7‡¢Ê¾«¼d@¯¢7ðÏÔ«Úd¹°ÿkFù`)£`ް“Ýf†Ý“êÚæUñ4Æëر¾?í@GXŒ‰l_±Éש/¹£\»¢õ¢Þhlòìtœïéð"3rà&ÊÊÇrÐ25ßÓî¸V!/B_nZ!ôBtõ ‹Ç”œpr¡óŽ adEBª¹üæö-X+ªLOšÿ/ò 3Æ$i§<´´MÙ€7Œ{Ÿs#dnºò¥¾ÚdÜ¥Ce§å÷Ô:Žû„uº?—\ÄKrþ׆¦çŸ±ÌoñðϤ'¥ÁM ùePÝŒœ,ëà›wÊÉ\¾§“˜ßWåöËü…t wÌáùTpë½;º%w„„º–¾k¾¸ôS÷†6ã¡L±ïÿœ]}¬ú‹¡'ï^ÅÞ³=ÀJ"w[¾¶7Œ¦†ÃÙq¾;nE'yŸzÚ y生œ¶¢Ý>•èh°I¾¢ÓÒ¡¬Òàz•”\çŲåÃTO:n…唿Rà°Š^ÅŽ¤‡r*f’ëá:¸}ˆwCŠz§pį+²Ü^ö!ŸßnÆx}a˜ö£ªgežÅ~ßSmßPãËòßa-§ÝÉÿ4ñtbFÖDo¹ª?ö2!5ù–ªý•N“ÃÎB>>ãØ(—¸-Ÿ—ìÎWÈJÆîlg9o3)¯ç¶ž--ïe4iE¤& =Ù=å‰ðdjT6Ò/Kq-Sô‡ÍXœ+’?cJ„è_~¥*'eŒ-º…¯NÏ·œ6À6§”ÕÒ*_JÚê«W­Kb]»÷æ3“=Ž!›~½f3&›Ys²…ÔXîç+…gROÂ>ã¬DSâ®ïÔdæàV=ˆŽz”R|Ú8h,µIØ6Ôt´C4g?~Z·¼ÿ~öƒŠXDÜSíËæ„„™v‡3ÓJãN©0_[*.¤¹m éX˜11ù" Ò¸(„;+Ê»Ž»?‹gµÖ€.´¨²UŽ|®p9L\§žøŸmžû ÁA¤·§K;¥¤¥ŒþÔ Ö’O¦Ô!Ž ×t­¡¬V?’U‘%…¸b?±ºç}O¯™Bœ‘‚ø,)Øå£q«¢`^µo>ú”Öíˆ ycá/ùÊ%…Wé½nœ:eWœ’¥#xÅ‹WÑ™¸Ë©TºâñàgûaׯIøát0?Œ{ÉêJÅGüp mÁ¸*‰Œ.—N@;ÆDãi‰út^wîÌâ^”ôm<ógá9èá„a9ö·ØlÃÿï9°×p:Î;À#Áì”óe¥¦“§ 2Úäïoö¢íÀ~»ž¦Æã±–hP̱Ã1Ϻ|ÑDKú‚R!vdßqLˆ‹ˆ…õáÄÏÙÖ<Ò‚¸Ø}ˆ ×Ui·”%ëî^"õ0[¸‹ŠŽŸã‘¸þpR׃že'aÆýò=cϲ&º'õóQ¿³wñlŽ7:®ŒñC0áñó›C#)z&ñz1™³é¸±ÄÚ…ãiÝí`c¡žXûÇsy)ï zÎTê=WÀÕŽRÜÓiìo–Ï,Û‚ûœ&–U…éAªä¢—૜L‹AÂ0J¶€Ù^üý¾á¸É©çv÷ a>¹ÔªäYfV•¬³ÖcÕ̽|±¯öizãç>S7¶'žmÑ 2âÆ¢Q©²ëÈø©t· ¦q ÝËx»¼t·äRºu!®ʵuý6õŒë¢—GÍ…(p•üû{ôsóUoèˆÌ„@ó@X^i¹'ã<à©Ð3öQ lþLî=.Ø"ºÜ‰±¦·dµÀÿeÏì^´«ï®Úy*ÇØTì'ðÇ;æË=Ü£ýE!B$&)Ÿ‡Ç¥- é´tLΧ¥É‰è‹K'!–3>ö`ñæY»Ñ[bžõ:”,¹ËÈëûÓ÷ÌUøTÅ¥ýõ¦LÊèUš¢Å Ã`*@Ý?¼8þçr1‘VC•Dà° Œ/åÀ;7ðrpÂ6seI):@ŸbSðåå×›-§žPS_£¯ñlЉÎx,/ÀCdž4aS›"~8B¼‡—ýà€X`pÎÏxx^Œr9¾÷ªWã}éV¬Tk ~žÜõ„·Xœ¶÷ÈÃÙ Ö ðÕhX“Ë'’oF³å \}â¤y¸é~†YC:ó-_x‘›vtÏø·rB•,åA]i—ÚiÕiÎ+Ñ‚îßKôÙ ûDÙ8{¾3âŸÆñEâQjÙu“ä î/”å¯zý=yÏ– ¯™â˧tŽn¡±„i `ÕZ'úŠ¥è:y‚ž$DÖtÒÛøKÖ —³,…‰î#+·2a6‚‘IdU­N¯ñZH;^Ëùc{qN5ªo»è„ÉdŠÇt&ÃzèóeJBs¤ž±Å3oÉã¤Pîo+­­_·þÚZ˜-x%ñT jƒWL<8Á:¥Ç#6¥ã;¡î÷#u¯Ú,DÇw¼÷¤£ç\ŒUžÜ@³5ÒβI–mœ•CÛFÏð—N÷¹gx,Ÿlb sI¨ÄÈˉžI™¬“˜°Çð½$Šn„"Š>Y¶ešÉt†•Ë™Ð+—t> stream xœÕ]ëoGrò‘¹?‚px7ÇÓïnå@wgîâœE؇`ER”Îä.MŠ–ùÛSUÝÓ]ÝÓ³\Rrr>h4ÛÓêzüêÑ­ÇAø'ý}v}ôÅ·Á_ÞÇ—G? úõ8ýuv}ü›Sh!¤8C°ÒŸ¾>ŠŸŠc¡Ô0yl½Œ—ǧ×Gß­¾Z{5ˆ äê~}2&aÍêv=Ò['Üê<ŠQ VoàQy'ÂêÛêq´ÖCÛeÕ`ƒ^Ýák‚Öªjrˆ£¼»·øè• jµÛ²Áï°C=hh}^˜º¾ƒ·V8£5ô\^ÇIki-ïúÕúDºÁŽ^`8SL5ÓÝ5þCÜÔD¾j‚ýIáÅ(V¬ñ ¼…•íVWeÀ‹iúfUuŒsE‰2H …vUXíhžã(=[u"…òjõš^j§W»ÛøÖ"e•Tvµ¹ÂÎ\ðÞÛø>ŒÆû¸¯q šeB›Õ&ÎÂyi8µ¶yèË8H€—lÂë?ŸþËPy.WžÝÅ­XÛŽÁR²Ù›Õîu$¬ " '”–«´í0rˆ«r bâêðèbq¾z½-­Ï[¶lœ±×ywV›ª zkî/q4‘w·9É{‰8‘! Bªã!†`Œ`”H]ÎY˜óí¦Áïˆ7@øÌ‚H¥Òœk3¥Æw–¬Nb°Ë¸˜1Øj1¬ïª“«¸_óêÛrýЂa•ÔÂÀÊÞV+Kk0 ª»µïéõ̾ÏVöDHXnœÎËÎÎðKÛ- q@ú¿Ç×I?½é[©>'¶tI’9¡Ú RÍjÄ ®c£”è››¼œ7 =m"ãØ d¤ºwI¢úÙ4ü„o¡Ú %ÈÜpÅTl×®6Låù÷+8`b︂'‘‹Rn‚a´wYyÇ9YN¾ƒ¬EZ`ˆ¢ÏĤ(ê[ê"x4¸Œî?}uÏ¥5¨myË`çèØ›±Àcw‘ˆÀ"«ïWkš MÈ÷ßGŠË:$4¨Ø9ÏWØa|Û¢¹¨~ˆku!4ËŒzUôíàʈí¸Ï`ÒûÂêûuéî.KA'fíÍ%ÉçRsÖ„ 9)ZŽ­qgÞ¯=h“ /MW$4U)b²9š¢ÌœÎÐOj¬oB‚‘–µp26ß]FŽU&˜J$ㆎ¾h²j˜N¬Õ#ܤÀo kDæ[ɹޤ¯ d¸i¡$(2RÈ'ŸÇ·– Ãµâæ=)u&}ŒeÿsÁ58”†šzD‚Ǫ<÷X­´ÏìIêMŒ£v´oeËß—mIÝTޝ­éG1u³IË’“ЈÕ]¥e`¨»7µB¡÷èõ-è™Cß´«C™ñ˜pàîIÙðØ Ó[°é ŸàêH±&— › /}yQTw·ãk@w2ÎÜÀ‘ øh™ Ö8¨»ç}¯7)=‰ZŠ:‰…# bçB·ž-ºŸÅló½:+:·Lµ¼UÆ8©ùتð¦Ñë†×Ö«Ãà_ÎÛ¢f(‹&®Ùæsų™¶8Lšƒ´xž¢­´ZôNd=6üËm¡k‚K9Ár¸DÒ:8 ;ºÓ0-4±Ù«ä¯ÃXŒ¼»Š'ó¬†ÿ­ðƒ'ãë†×·Fð ÄŒ±˜(»w…Pgø»Ôjý-éN:)Ý œô RúÜD­‹òظ8l­ò‚´]´ÞGCœz«>¼*Ü‚þ•aÔžOò-c t%­% AUP?ÞuzuÒ4@0Wß2þyû°Ðâß ßSôþ/uÃÔ)›rüL€;Ú¼²äiB[F—Ë82P·ênËæ|>‘ίþ€Œ2Z;±ÃÝ%”$.}‡‚¼3ц¥˜ÚÃCŸ®Á#ÞG%f†/¢À1ˆþLÏ‚ð9¾%M©,¼þ‘˜#Z»¤ãĤu¡l˜#L™ˆ¹‹='Ý·ÃËÕo#]ed.\±”£­X ¦Œ«À¨ ;ÚôÆb,eFg´R–‰ðÔø* z¼zÍéÆöã·Dd|ÃG¼/#òßÅ%¨ü(1èBòQ¢ty[Þ³È'Ó¾DáÖ.n eT  ”ÈÜ}yzôÇ£ö1s^ß>:Ö-‘¶êØŽúµm¬›º¨æ òÓßÜBÆÈ$8aUœq·wSHŠ·¦÷èFE¾¡#ÂiPKËѬêXFü,¬$°ášÑá±k—€C` õÚ9˜tþ2 cµz±f[‡ØB‚UkfpŽ·.s_ÔþÐÀ£ÒgцQ¹aÞð˜ 4|†ÖïeþìÅËøÈÝx²Ó|ºR ʤ~YmIß5*KH .yµµzAŽî½‡Kj¡S­p` -‰neëöÉ®à•9‘0Ê,aØO$³tÂèeWf{"KÈd.²m2áª4fñµW  ÏÕ5¬Â•øãúô/O•>`»atêTñt%|AwL RC†5 .º4ó¨g¦cAËIë¤ ›‘ÁT—ÐÅý!­#T•>(¸·»ç~H—ÔE8’Š¢-ˆQ§²Ôf–BLQæm¥¸„Pj‚²š?~¥r ïÛzóÚ©†ÁjgSk³c¦Cc]ŒŒçød•à«r‘±É³¸4?†ºÃõ䈢Öú°—‰¬sk‰b<‚Ì ž£:›å=*0xž³2ªCŽÙ÷ã‚„ÈÓùx)ds1èXö€yÒÜ·àéÅW]̽ÉUí”T!•–¤Â-°2 BiœŸZqHÌI}SÛø„K ,¦ÐÆ7NÉ®$Ü–sã̹׹™c ðQ| ZŸï£ÖN– Á¹Š{þ‹ éÌ™Q`nµûýÆQW²Ê£1®R”˜B«¦ ³^2yß0!{Ó•±™œ"~‡dè Û/eÌS@»½FËUŠRj;HçešÃƒŠv@,(Ê晥6vs†zAš}ò;Ñ­Iã®Sç6´É*8=…Œñ«‡s{7<Á’-;0“ɯ—’S]£?µðL?]°ƒÈ“Òhò¾ÆG3 ~,v‡¥+˜º:É{¤È^X¢žÇŒ·3ªLm¥¡¼p™--êl¯^îg „ÏŠk-œ]z{H6 ^0¹MÐŒ†ðžÍ–ƒÌUþ&%œ©MP¦ ”ákëMSLpÒ‰yNµ„¤˜êr2RÞ A6¾DKsU -ìºÕÚOzdAé¬Ï£Á,(†3®å.—°DSŠÄm'ž3\Ä~.”ni7:wXõÌyé™íÅŽyWÞ3r~€òéÛødc݆2¶qù¹|¿-¸¡2|'Rãä¢9œ„¶§Ô•ƒÑ z¼d)wf2Ò’áN†i zS)ÒaŽ'Š)V Ú~¿˜Æ[ÀŒ¼$`!Ç2KH-çRFŒ`eËÄÝÒ*—²%(h­š»¹1ÜÏ¿Œ>êèMå3›Ö£«Ö¦™m·•á™:¡ Ê3¼³„¦ÖóÚ›ÐÄ&Ê×U»Ö R^¡gnPj^§ö§—=ÁÉïgQÑÿd“£5’€ëk^e¹§­zR¦UõÍ<¬ål;®Ë¦êÒøl^#CþŒ4wÍj—‹>Ú`l§èc/uj5Q0Rì¸@¸¥ .¥3GÈ“U¶Ä%¸&ÀAoZpÏ¡@î;°â†Ž:ÜØ‰äµ3ߺՋK¼]p·ZäžB$ŸW°»Ö‚ÌXÑ‚á—ΩpvA`hƒ¥òc‚¥@ôzX—ß­þ±¸kŸjæÀ; ºzæB}ÌÌ­¼5õÌArš·P̃œ‚Eœx3æ`§÷äŸS’0˜” ð”Bå ùÉš©8"ER6@ù”|ÄÊJhŒy¥\€k’·”Æ0E ébS–ö†BL¡[öÙc³Þ,i0#®Ð¼úŒ†žC­ðÕÚK0¨0cÄÕZb†è+³…&¬çÔË ‚äŠï"ö Av·Þ¬sÂf&뀄.Q¤â¶ß0Å‚öhL¹Òû‚­ï ‚üâ7ü d¯âÜSùGç.!—¬F™b"HööU JJk"`çV¿ßÜ—XMb5œâ&~¢q÷¼4vǵ8E%%À1ëZãëT;ç‡qß!Š] š?ë©à+Ÿ@B¹O-Ô@‡¥j`­Çf´xÉQ†ßä¢w§›üôSôvÇB ïb¨#ߣèû±­ Î^ÅôóL/› y®¨>…@b¾Qûc0$þ2ðÌß•/þÚr»2¿"®Eí\³}„ˆ(DÓ–ª}fvØ€ÒVÆZ!ªQqÇ'èiƒíÖA®ÁZ$1±ÆJƒp’Œ“;å’™†Q ÕB o$ÕTNg¦ÈhÓ‡[ÌÅ:¿Xj=°Ôà8nQjÂÊÛ€áÒÕMÞÆ *>¼¿Zçþy«vÜôXÆÕ±^¤øjjBA¤\ýTÐ Ø _»Q¬x8û{qcôìªñ Á®^]µ]B72A‚ÏXµÎ¹Š+Otú)6^Œïl£áGSi{*9÷©È!îR£ÅƒAÒ¡e~¼» |±/YT<ø;~€÷:ƒÉÙÉÝ}qºîâxÑãëf 7é/~‰SuØ*N¸N ÎÎèêsŒ)ëhœ¹”ƒõ£^¡MWºdX½ÀPkÎÌ6dsÖxžmà ;;Õš¯~³Ð¡c`VG²Jåðžð<’žj ûtFñ˜8ŸGY`󛢯 X}Gfú¦U>a~ŸÒ)`¶4–šë=÷°¬sé¯>ƒ™çŸI=“Á}™cÁIû¿úÓÿ1­³4MgÏ‚ˆg«U¼Oƒ3n]ÙYÆöW,øá6YûY©Äˆ²e~°Ói]*O—6¥Üܦ roJ¯þR¼X®jØùYÆhmZIa9ŠiÓJËÉáé:˜_1)dlW§1 uLùÕE ùCVZ4»FS§r´s¼ ¥*QáÙ ÎDçSãºZ4Îcz?´ -á8TÀDSs—.»éÇQ7ïâÏ¡F•{λ$çУ 5KäíÜÄßñΉjÈíD¶šfÚ?îf°Iñ he«º‡‰ëœ9ÙTŒ>«lòëçe3›¼táM—¼ Sf‹Z¸æão9¦Éµ@ ÝU:›EÞAôœ9×HJíòäêAÿürŒ=^‡esÓ€|ƒ\ü!—Þ¸r®Ü‹4Ý,ÒÌ$-Ìu“ºAÿ{îðÀ¨NY²H\©Aøt¥KÓ€K7Œ}ùµWuãeŒ\×9Mƒì­Ùd[±ïXͲg.ð ¼ Jüož‘¸}\%œ :ÃŽb³ØJ‚—ýŠ.n ŸM犢 R"`„>ž5HJëmq&SÆñf’i4œ›£ Kî7$6O7‚%ò3 lD”8Pí+È Ã…pQŠ»¥®yä¨ï„¾iDQŽfHßDC,ûnXhªSn;I¼ëütÑIû%¶¢Ç>[©0x[‚9 [%g0’ŽÎŒŸõ¥w¯æi1-ù zcñô¿8>R”¡Ée S]–*‹èf¼9BEØ”O0bU8¦Èn×¹Àåºxƒ7MÈLZòß-ùï3S„‡ýÞE™›&8«¾“Ø0¹bˆ„i¤ÚÒ§IôqÌy‘Ö`…°Rð¢Mcô C¦šlnž&ãºPçÏ‚_”wL=Í¢Z–âý‡)ee©ú³:ø‘ˆž@4õ'¦Ý º«ÊÔ:å+#šØÌΠ餭¼‚úb®èW;¬Žž8‰Ãk¢šóJÍ.ÝŠkˆ7zEfiW¹jÆcb³:#.—,òa7/.”t¿f‚þ¡`„å{t-Åiö·sUÆ^ˆ&µ¡£­_£<@nHª:†Òêµ–rª­ÇŸ\Dæñ?K)æ ´émBÀF~a‚¾YSU¿stÑZJèÕW·ò o'ʵQŸèËosÑ3è­™\:’=Ý<7uÆ}éÞAçÛtqž£²Ïêœg纸n€‰ö b©nŽ-Ž;Ðß_ºñõO»‘®9z!ÇD:UàÅ5›_¯Ó]m“NËwsçË%ësÎ}¥v2å£ð<|Ê;\±÷*ÊÄ;¼” œÙv°a)${ßËz9‚z”KùšÙzËn$Ý §–X÷ ½ˆ6ïËîUÁqE¨v¤âÊ"³Ì,–$y<é±^[—SªÛN‘ÑŽ• =™ôÙh® ¦ú#²yüþh<þg˜ñ®0-§¯»à5%ùÍÕÑKl%±•ô¬ÁFé42S£åJ5Ý­(ÕÀLÛ“îÒÿ’¤@bϱ)HÔcBž*‰¢$ݰרÿÒýXÅœw)<–‹ŒŸ×‘T"PVnÙjÓ ¾Ú‰pÐoâc„›Þì¥I¿$¯D‡¹U49ø`^ë„U9~ö¢#âï:…kå×·våÝ]—óð¾—Uûr†<½0¿f::ö wddÁ¼:ÈØwÂÛâÚ5f½J¤ës¨9/s[Èàº> Oµ‡ð—ÑY+<í° (uir‡œ†ävU£SpfÙ®.-$Šmú¯ßáZðØA*F¡S³SŽr@D©©Ù÷ÄŽÊbø <óOàa:°cॠRÛÌ"1·[H]>Ÿ»”\Ø@\áëlìN;âTjW{Åé×v[Ö®+lð³~Â!E‰n®Íšáeg蟻 3XæµwÜìŠ&È-ÂÍÐ%°;L@s½pÈ”„ÚÄ"›ày‚Žc¸“NZjnê§stÔmý_7,@ØâÎpÑqθÈùþT‚\C©Y6èë½¥ÿ=ÄöS·bºk M¬ ®hø­½ÞÇ”¸&¼øßv°úŽˆÛBham¯¨ã¿ÖT ÒÕøsƾ’ÚþÑÐ…Ò‘óØò‘ºᬉÉVyÇýÐúšiPKiÒ÷ª®Z¨3={A¥ƒ/YöÝ{†0Râ«_—Ç—o®øæVøñ‡ƒ¹PŒX<ÊonÁ“ H¤/ãnRž2Ó'»Ä­"£R‘ÿ…Ä?)ôèHAÓMê¬Z5{x*Ü-Vï"2›–›BVÌn™ÕºbÌrlk ¼@…|'Éë„d^·åƒÇ«¹2ÿüM‰çµIð4n­dªíª:‹t¨;ÜWì,;Õ©Êt)¯îæV£ì¤ª®ycÁSÁ\U E€vHßJôôb=]‘øØÿË@ÓéžÜÛg¡«ªËG?V¹÷Ôû~QÃ&À_+è¨q‚q^õ…Ùta&Ö’‡Áú1ÿ§IÕ©Ë‚¼ÈO·sPOq¾Xo¼S˜­þ»¢˜KƒÞ Â÷¥aÏdº¾Ð»§àßv~}Õ‰]\°‰w©ÀË;ÔtPæ»L‘Ÿþ\vâ³üòëüT®!ÿ /´ "µÂ1?½äMOŸB•áúRÄ>QÖÒ=¶å“¾™Ùûɧ2HO,p7˜ ôö†~$¿×¿>†ßifߟ+/ç Ï:íî;_d¤`¸bZ}¾þ|Ž;À :ï(›ÞYÊr‚òu‡¦}EÕ›“‘ìº.˜`2èÍ e)+xžvÁçÅêBÇBûWz _¯Oð4 ò¸ù%µÏôô< ŒÃd[yö÷…㹆ÉRø¼Ì°Ìë¶–é¥+ßèüÒ–—®×Rô^ªÞç2¿T½>MïóÐ{É>7½Ï¶å ‰™Ý¾¨ÄO;¼Y fÃ{†uÛ´‹Û˜'öU‡z‚Zxî¶|\2_pšæÇ3õôiiw×3 ÿÖiØ›méðUgÞ¥¿gw·v}sÐ!±|ÔcLRƒ”µÚã¾?€åNRÿ‹âû²ÃG?O€IþcŸ¯:Ÿô.tè ·Ÿiz~üþ^žÌûø€²ª£€ ¹d ÑcƒôyQ³ŽOn¯F{/Ù‚ºæ è^ß{xi{ú¼ûò)ú¼Å¢§ËEÇïòÓO9*Ô‹bã•W¢( Œ~²o‚“;Ô7)šµ%ú+Òb¼Äs—åu§£ÛN—︠°/:=öto1=Få ìøJ¥Ãn´xÃúÉ/?.=G½Ýø.?ý¼³Ce×þÜ $Ð{x> stream xœí[Ys·~ç¯ØŠó°ëh'¸—í*;‘í¸ì¤l1®JÉzX‘E›äÒ©SŽÊ^–ÖuÝÌS¢=Ûtœ¨neΗ¥uΰ¸®ü+aìV2×¥U9÷š‘#KæGF¤Çû®|ά|É’Aá¡BïÕãÒù–sB\B÷b-ǧ;†ÝìäoÝø/3ù|;™¯™£œ0-z”=eõ„Hhh½yxY±A=à÷ ŸÎ÷eòWÌÀ‹½ "ŸðÚ äÚ]JÑ#gKXËk“ýÓç•7…Uw¥uÁùî}ÖØbÚMþÑ_Îh«r€ëçj•ŸT9VµãŽáÓsfáëîüˆáò Cç¦p™7só%É…&Ï*Ëm錥¥Xá!1VVUWLë|$‚‘$ùÝçšíIO9‰3"ƒ¡õ¸´LI†& ã8+©Ç;-­[Æ Ú tˆâÝ2|Îlò¤NÙÕp;»!3P-TTÖÂçd½ôj+«Ü0«pî™×ßå°Dã&v™<øã;±ióý-9üü„ÙÇ›]ÚÎ9%‡ðIæuÍb©D4]Œ´™MäñáÁŽ<{ ÉÇ×ÒhÑY3>‰8»8ÑÆŒ¡çüà ŽRi”ðª“G =¿ £¦Ã&2<«šY $µË‰Ìã…èTÐ"_=¢sBè샦À ¢‡ýÆ(‹­ÝçЭ…Ó*qˆ›ŒïeªëÝéàŠ6|‚£ƒRªÂ`ï¼£ÇzÈq 넇m8ʺ¾g+SšM¦D Í&Stð2‚Â*MÁðè v¨° •v¨S:èÎE ¾'Ÿ\702‚Ri:ö7¤œ2.Âê4 ZÇÐ{T{ϰ×Ç‚ë©Y<ØEZYƒ¢%ÂÚÃ&.s/Î[aS¡"^$ùÄý|…Ç ¹^iØ'Ê^é(‰7d=ðM:ÚÎp—x¾Ê¢®r“¶$|°¸% œ * ÖtZÏÅ™²„¼:J3¨ÇCåT¬&JP)‡¥†ÓE5?=«ÂyU›gYWµ†“óŸÖ­x‰cm¢P³p`Ú«…µŒxŸùm”ø-µQóÕs2­JF)-4`w—EWI–Zàÿ’68é­1™õFÈ | ËtF‚{”ºSZô~j,Cðv6š¬lYGWù@ y٪΅ùé:í=á"¸L˜éº mÞnV<†ò f­¨Ò“T=ºÅÒ€ÜAùæÿ\§5¥•Þ§C£›‚þ†µåü·H$ÂASJ®Mñx¸Ñ`i#@§ÉæŽY#{A¼Ü]b‡‹QÓ}’…U[`U7‚¾É¢O\à[ïßkF±ßp1 Üä¬òbꆵî K([Oª$aÐy9ž•'Yt ­¼ªÆçnÐR?ßT;åŒæÝcY8Pµìw¬Tˆùé5þªªäy5 bôLÇ™´EÞ/ývCHšÆ¤îMï ãŽoYpÍë|›¥=Ÿ¶$×ì!$pùêË÷ZöÍ};*€ÞˆÈ j©]QÚ>H0>/<òÒügDc¤‰#ÑÀï¦÷ÐÂÆÆo’‚wýnûÛ9³Ú‘RE+A¢2ä¢iNQ¤›Æµ¤Íˆœ3Syw—É:)É‚XlÔ®!wDdxVW¬Wæ¦ [C3pJ!Öö-c^ÐPA´ÃÒªÀÍ¿IœÏIÔ*8µ.vÞmn¶Í`!6õ(»öAnä°• ó£ÚùEmšÚ|Ãô¾ST?Ä´_0y‡hÜ0®ë¬®LüTÅÓª¿’çpÿ}o¿é í+2¤=›)}ËhÍóúóf¿”-›L{ÁÌ8"sQa´1Æm>cöp™.‡‘mºA4éÛ¢IëÒªï`¢Øw™üié$(ù]bG\§$rãôdÛ8Aú˜W„:ës;a_ ¶MÙþ«a~ÜrܾüVÒªO°!r³¡²›k‹ÇØw$C'å@²_©}/ 8 Î`ë’ÑÀúk…=+RDpÞ8‡ÖvÛ¤Fþ‡°ðO(ŒÒÖ5ù8bvÆ¥!Ÿ13銣~Ê©`u^¿2SˆØj'7yÉØÚ{¦“¬«T%ï¾Êçœ2n§Óè7§´½ÛüËb¢ðDÞåÔ7çr¯9{+ÃOŒ&UÇ[Cž¿ïËTîáð–iqøãÞ‹|ÏþxëI8ö×qOö]x»Ò•ë!ž5dàh5œß0ëA¯_.¾MÀÄøÑƒCÏë{õÃëi:rÙIsäêÚ«C?bÎN]ûRZHbœ=¯ËE²÷ÓÖBæáõ±Þ×@Ÿd̸Ðj²ÁO«Õ7—úÐò+¤WÒèN(,¯ÐÎ¥^/@>qT'¡ :K0åíei)~åLìýVödå8¬ÌÑ®+×rÁ­LÎì6V&û8.žXêÄ‚“E]Ö6^&ò‘[IfCáž)ïáš|S›.ÄPó ¡i!“Ö Þµ6löÃw+¨û&Ñ .&ª,& ÅÖ·™vȰoj;¸8ט„ôü—¸–a¢#’óD`™uÝØì^ÅN[šM çÞ8Ù½êÏ÷ϑٻ€“‹] ¶Á{î­ûŽóÁmŒ þµF4É ë]ÆvAŽ:ºI°ØÅRKÁ“ ìø ,…ÈÖdtQ×Î'yStÕ«« ^|茯3 ¯æ¯t0”àr Ï!™ãRqt¯œj€Á+”Mh dD±¨ÙªÀjržy Çjc䇖R)MŠ*mûÀÒÀÕà(¹u‹K> ôtã}¤ß7úÀhQ|\!¾äPïΚ1A¸«dœe-Sù™W‡N*|惘Xc'®¤0Ü‚ âåôNh!pR)‹‡ÀÓtvþ¯›A"Sìg ö¯qœcB!dÌ1¡ N¨!TÌݘYøþÙT“Š|°tÂ@Ä??ºÍËAôˆL…&ln¨Fêw!´7ã9:ó2ÓŨ¶}’ƒh ëštãæú&³iƒå6}K(¤µEPs>?¯ü¢³ò»PÉrAŸ„Œ4 ˳µÿƒ¦aô-Õøq=…Á¥Î T e¿­:H¢êct.v ÖD5>ÀžlÊ÷¤îÙég{‰ Š¡U?ç5ݾ£CÉ[{Ú‚† el;é7’©!ÙÎû•MUG{º&Èï7TO·j“tÔÐ37œÚyéË–€+ÒMèÑÑÇì±Úä³t×¶U;‹äd­JSYÖç›×=-ÈWúƒ‰ÛšfÝ$K&”Ïzá»àu)S½¶fG…_Ð-A¡,Í«^¦´Jºè™{G2™¨ë*I¦•cíÜ!ɨ_f¾@žÝTù$ÖKM4èúDµ)a‹h´±Ñƒ¡X`ÁÇZØ×q…ˆU"+0*Ø!Ö`âÃ{‘a­C{Yq6gÀÂO¡6Ó÷­›Ü|,oô£Ê¹m`4Á ?ôcwLÅ´ªß©]eø¨È©Ö¶Ò¬Ñ"°%H¸=ù®ob‘T[šñƱϢ˜T^Ÿ•œ µú¢Tn³nÕÃJ›ÞµnŽ,¦­%³XY<ÆÕ´/x8 ­%»Iûó.ﳺ½[R@†Ý}1ø5f¾D¥{£œlÂG5¢F@®bÜèUbêÚªÈ\–X®¢xÜýjÛÖ¨@1×-ÎÒˆaÐÕ®™\#¶zM‚Œ_&^À?O7>a ?_WchëØ  Æ»-þwÉD%ò«ÂNZ|^ƒÂae4«ƒBsUŒL¬§16±Zs<­?‹JÑOþDYM}AZF‹à½¢;ic¶öeÄ:œ; ÂV07‚ JL¹#‹Aþ¸MLR„K1%µÕ.mÀ°ÅÓ® Œ&ˬÙÅùbd æÒ`éè¦÷Ê7Òhô+t¡"¹\J¤ùb†<ó¤xÇÖj¹úýÕÆ{Yñ )€8²¾ÅüZÖ|¨ÑÚHϬé,«y(rF*àRS*Õ[ã*nT­O*âpG.ßÙy,p¥yT¢á 1‡6êÉY£.F;r?@_‚h<µºúÕ|œ:)iáÒ§ÕÿW‹ÑsRRê·™˜kv0}=‰ë–„Fs˜äö³õ+Á†P#œœþ|aGn“ÙÍÊiúuÑý«'>ÐÐpã””¡¦0¬ûÍŠ,sDÑI–”^o!'”Ù§aÛn~|TøMi!X’/Êòèhƨd±[bPæ_a!cŠtÒ‚n4%¬HÔ–µ ª0Úu‰ÃÒ#¼N ÌT ÀædÞk4>ðš¾’çÛ|ÐÚNXERQ›ƒîš‚ÏÍof¨}…?êH±7n½ÉwKã/~PKDƒ¢¬ùŠ3…Öá:m^êPa”W¬JÈ­wdõážÚÇqµzwa> stream xœÍZIoɾù äîÄìÔ¾H€q0žq0cÈÁö¢¨%C‘Qò’C~{¾WÕ]U-–+q ,–ªßþ¾·45¿6¬ç £ÿ†ÏÕå쯽kÎö³_gÜ ÕÞ(éDïU#ã½°ÐVôZ4WëæïÍvÆzÐA‚P?7ÍÕÙL ãzÝ(ÇdÏEcXÏ„•U=³ŽƒuvJŒÚá‡sÉfmŸ¹¸±=ƒU#—M"ÕL©^êD«˜èñHG¥Àï¦$ðcg`1~j²…Dg3´S¾‡µS;ÔÈ(1*Écé¶aÜô>3:ßÁ·U¯Œá®ù8ãÍø÷k¾»+`\âw8@wFi k†Õeób¤qçÛs¥ñ¹8Åt"wJöÆA.j•n—³vÎ{ß-H­…hil³8 ·6ÞÂvKn·‚nEÏF ܆Yô^³ÈP^ôS)M’íz"nEðíböóLIÿe À%n,ïUºØ ™(࣠ ¿—$#.Íx1!‘‰†‹ Ñ€‚L4\”DcÆÑx1!âÌÐg&.6OL®½÷ >…ð”Ü7íOFb˜ízyµßm»w‹g¦wÚ[$ú¯³öj½¿8¹YnöÏ»ß͸‚•,ƒkH7ü‡š8š!ö…k¨eLYçxû‘.T*ó„æ t»M‚Ïb˜¼ÓÓ0Ï‹Ì]YÞ![KÊ­¤U9É2˜Ò(ÀäÍH ¬Ø,!ÆÍƒuè4'¥tX†f§ÚÏ:dñ4ª7ÈbTÃ%RMáó|"ä}Тt“ëm¦>)„¬;šKn'²WùºˆC™Î”ˆŒ9@4GЎꋈÖ<æqÑ¡ë;êP¡&ÓÉcfŒ²Ë…¡"ÆO)¶ÀîÌ ’4—[abêŒ$°|´?ÇB†Ü qgÜëX=Þ ”c°Úx‡e|Ä õ}Ò!îb8¤&F)¢ªG°Ý#í7gePŸE厉vÐ\e!¥­zÁQË c¹-;H ô¢B“Pø-‘/ÓMÁ4ð}ùôDã:ò©Ió*äîÊè(ôWÁuàåÊL®ûܰ†ºÇ.1"¨ðU¨>®ÅRš4N¬9èùZŠ‚»êäeçðêå¥ñG#§7ºVÎSI|$ dÙ©Ï‘’Ž·‘kƒ‚ÝðŠùÑØ…Û³'A½ºš².3üƒo˜2€í#£DW|ÚœÁ~è­É9;œy¦Ü¤ÓªB·L§ãtÚ¤Ó:©tìªsȘ^Ñ7ƒQy¬Þd¿b¬•F?^ñˆAÐÒê²¶ðZ8¤C”×Û$n°‚Éé4LqÉÍ©Tvr›ºÀ¸Þâå‘ׇ»ú_šcŬ¹cLì£?JðÐópD@£mCÉïbφI“ÎZEt혴ÎNglèñ¡í;jûü.¿½ªö£OyèNNm²Ûc˜fÂÒ÷|˜fÊÆ8Æ £j=ÕG²¥\ý¢L¦-bôe—c,ÉE&)W¿÷y¸–e[T{h‚ú¢ƒy±E±äœgTÒȶŽZZ¹½5vò9x Å·*ä!>ír’„1¥'Ö“S°$å4YO!Ká0ÌǦìî°‘l,w¯aƒ/Ñ0•sµÜóö™öVèZ{1Í›¼jÅ]ìzÛœ‘73òƒ‡ #üƒñ@:^ÆN svóÐbèEÅëTÏá¹Có›vr’ƒE©Ø ËÒˆÛ®Õ*ìw5W…•U‰ˆ ñU èäøòGk§{ܨaÁôÒLµZãù¤Ïý¦~¬oð‰r›8Ž &ÛçbÓÉÛT¹É´‡t0b“EœoÅ ÙeHÝÞ/U¹Þ÷¼ÇðZו?ws-è%Bó u˜KtºJ§›tZUè–étœN›tZ§Ó0¿èø}…ùâ–bšsߦÓçÊ@üK:í*ƒsWÄ´àx£Œ~;zÏ‹é}^aÿ˜ k¯ŽëÚãÇÎôç™ù¤òø´"ð¦"ðº«í(ud€¹@58?T÷)Ο‹$%$¼¨„­ŒÅxÚfË¿¯XùÏŠ7›Ìò]E˺r*´¼¨ø}S°„\ž4€‡à6Ç\´=ºÿQ¶Ÿ‚ä$/–΋ Ò³ì\%¿äÀ¾ªn¦ž€)“SxqÀ  ¯¸[ž.µìHè.’¾Cûy~ÈcÓæ½xS™FáeʎŒRØB ‹LV®ó±°ÓÇ<Ô-xQÄ|<íÒéc:mŸ5l”Ù‘.u-”zˆŸ.‚{¿qzzüÌ$~¶ùšmj°M7•Ð{]1îdñèVAf'Ç[xû:ñâ“xUíf‡·¬cW ¬Ôãñf¤V!«•à²V‹ÐV‘6üÑ‹ræln¯9b4,,úÿ_ þNG•yñ)Ï‹åãýkijû'Yþò÷ëî%!mÕ¥ä}EǪbßiaU³ªS‹ÒUD*ýµ#íûÕ-èC%Jµ/лí«ÛËïR½”îBœåÓEæQ#:4/X|Æ|]qæä@\-õÛ‚žœ§¢¾|<º…G¯_VLº¬d¤ÌCZ^=D8"àIZF毫åâË=ƒ¸¾^9þ‚ h™L—blóþÖF6ÌÂé 4 „Éîq0½YuäBŒä*Œ;UýÂQЊœB‰z|Zþ“þ0úóìߨúˆendstream endobj 374 0 obj << /Filter /FlateDecode /Length 2411 >> stream xœ•YKsG¾ëW°ö°5¬2Çý~øæÇnÇ»Uq”ï¦V,‰²(ÙåøÏï‡FOfHy½åƒÚM4 àÐóq¡z½Pô¯þÝ\=~“Óâ|w¤çGtùuQÿl®ÏŽA¡^ä>ÇgG|T/‚ë­_„ä{ŸÌâøêèîŸËd{­én—ª7IùÜ],WªOJimºSìú2–×´ísŽ1v'ضÙDåºK,u²:u_–+ãcŸt÷/úYeÐvh3ô1[ì‚AÈ9ØTøm‚³8fbo”¶ØS*„u ³Á/˜EV©{Û‘4å´ËŽ„·Kü_%ç‹]?êýŸãWt 6É[0¾Ï¸:µ8>å/–ºW8ªa/Y¢ŽÝ®°ÉønC¿k¨ U<ÌÑ1@ÁaïnI¢´µ‘´6)öd󌸋5颕rÑ‚–Ö9¹hŸˆµÖÝÓ%Š*e¾Y£T†å ÊI“q·tcÊ–_Ê•g‡ZNt/¬S²Ýçâ5Pxvf€p]©rÞЏîÍè‹'¼K~ý­ç!e%$ßÉ΄01¶ª¬l®ÇtÀuü5ž{èfˆÚ‚ظn+H®«žÉLN^‹“'£)x)F&ðZŠ)ÑïtÎd°¦¸ËêiÖÔgDfU«à'Î8'ήÕÝñ’bÊÝ';„ÐéÒkpÃÍr°À9¸WHMÆN87DÄ[°ÝÇ%d…¸ÃÓÆaG¤Yé¤IâÊFø4„ÅJC–÷šEŸ±ß‘ |[H¶TIgn»ç ×½uCPáœ2 `!±¦’üdtãWˆ *d§B–¡¶aÕžr¾¬Â•îWgåh 9ƒh´ÊUA¸e&ó^àÄ5]'Œ Œãƒ)ŠÔýFšµ,ªX0ED¶Ö Žr„h{Ä'”šäH‹õ i«œ…‡®ÆÔXÃ6§2¨t÷w,5Á‚é^® ,æ \’ ð±dÿfi€jr÷3Œ©{F»ViGÁ¼"ڜޔ]‚5KD8ƒÙxkB'…`‰ÆvÛkÆj ‰·!ÞÀ¿jöºÚÄè~ÉòKqð.°.lò¿«Pí»ÓŠ•´ùN¯F*—S¿•Pn½†ræË7>ÙJY"Ñ9¥ʱâ6­q +ãùÔË5‡&ÒÇ4®UA¯sòŽuð¢êUù!#Ç»§x;!{~*ÎA…s ûuïbtßP5p‚)ÀŽJçÊl:Ïx…Œ¡RÕRT–RU1Ú?ÙÛ§|öNàk'ý|1j¹ar‡4Üžñ:|J™bÕ­±“È;ì>Ü„qR`Éu 2|{]…] ²ùq8ßU¿µ8³S»¿-Õg“-Jy«ÏÇ%øð¯æa ;lf?P@]BæjL¶SQ¶ÇÊ'JѦöZ„:‰ 0È€_D2’1"ÈÌ„ZB(×-‹ü‘ÕqJ1bÿ¦aÿÅ^ý.í{‡;zÄkcˆÄ1›Ú¶“–‡y÷ÎFýs$=*¡I³BdÍWë‡Î+÷ŒÝž$"X¿ù¾¾.žÎÊܯ\ óÝK+Ÿ3Ü{ <æ`FIr\4êöŽNæÞRB•UÈfÖ:*?orŸHÈ©G\êy»6™\!¦õ5ÁdŒ­7c´ ê øŸ¸Õ-ýoQÌ! ëê„Ý͈Ûï§M•Îèu¨CmzìøÞP§Kg`“‹h nYHRÿ&ÎÉD¿/ÔÆ  å’µ Í`A§`Ü4ÍYŒ*%}tÔöZ”Ï“±z Þ“kj»“ƒëKæž`Ú‹¶”þsL S>¨-ªäŽç²ë¸¾»;Aý®˜ub§È»›öÖ•v¢Òn¼÷~$aP®C»hgYò+a¸3w7bŽºç¨w"€ÙXç‡úýè795¡»²ÖP7/g°u cô÷ÍhÉmS&€ˆTà¬`Ų=.´»içD­ \³´¦æýýh»ÌûÚEeÌ¢5‘«í¸C‡Ræ©dÛ`{*Â2{ÑIî ~ÜxdïkŽt—̤\àÔ‹^¦ï Ë¢úùL"Ò- Ž*ÌúA22MÇÝg­ˆby©˜¥¼òŧ"[Jš,óȩªïu»÷sÖ•ÆL©S)è :ÆI?ßÎZ÷D“öñ²ÚYzêqB9°¥þ "ƒ­ÇÅÆï©‰È¶9&Ï~†0û 5ïJlä!ùŸêVÑ«rÅ>¹ÄÙV·ï…wm°øŽÂ.yï ëu_Œ“ Þ! †ré"+<4Öºa`)cÑt¡XÎèü[ë`¼æîýº^J£z!†%J㜼®ÑTzÏC…ÞRá1óB/ÞMf]„Š!7¼Õ­ÁÒwA¶(„H€c¿Z;‡ŒhZÎÓN¡B—M¾›òax(šn–GïDI†ð#ÞMØý}Ò<–>ƒL‰ÿÝ÷︡òú´±ÖèhÄ "ÁæbÄïyÀ¨Q—v€³8ˆ†άpq3B=CœïáOÉ£Sæ…¶‰CГޱ@™j†j’{?”–!#3jbð®ÈÉí]U2»ámÐH^k1ûòž‡£ºÏT`pó¶4­‡ù02~BFÙâ‘i/Òx5åôü=­Èˆ‘¥A€«P~†C•,ô—#ƒ½6¤àœl‹}”“HÔ2ˆ«òÜ9xö9grÑ·ÎÛHÛÍ`ë Yp³æ`2½©òµõœË£Kð«Üߌ#()“(fŒ˜×ïFhž‚µaØñ9\jy4Œ)k'C´(¢õ®™¯¬ÈüŒ q¡aó§ö®»™v(3³‡ôvâ]¦)}¸`ÂØ8¿*Ã.xò‚x?ÖAQ´¿õ抙 A9årV·Ól>œ½9I×@ëú/|'ºÈendstream endobj 375 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 2679 >> stream xœ­ViPgîq »AÝ8 DíÔ“xálT³(Š1ÞÈ1ÜÌÀ0œ¢Ü03o—\*Âpß3 @+èï\êÖš;e­ÆRãf“Tj¿&{|ŒºWòckkÿtM×|Õï÷<Ïû>Ï+¡&Q‰„ö Øà½xâ× q†Dœ9Iœ%ô“×X ã,ÊmúYÔ–­«fm¬£t)¸8ôÏtIùꛆrÜPäTJ*‘¨“ó|4‰Ú˜¨hb^ø+ ïU«VÎW,Y¼x•b}‚JªV„ê¢U ¡:ò¯Ø¡ Qé2óÖFët‰«-JKK[š¼P£úõ+ói1ºhÅÛªd•6U¡ðÓ¨uŠ­¡ *…ý² íOMBbŠN¥Uh"TZ5EQîëÕšD_m².54=#< B³#>aÉdŠšMm£©íÔËÔêj'µ›Ú@í¥öQ©…”/åG½Eý†ò§–QÔVê9‰3åA¸¡¨,êk‰¿äÒ¤™“R¤nÒ]ÒÓË.9nq´Ò¾ôæyæCv [ääí¤ržäœ1y¶¨w+ÍÄ qéØ2Y^K.ŸyI8zü[÷¢,£^­O3è X%ÓÁÛL'áœ34G\æL¾/Ô𼩤RÞ¤ŽHC x†ãëJÚZò°A?Ü«~âà _ß–=ÀuŽx+í*žA”Z%V"Z*¦£k²ÛÉ`©Ïž°mjÎø×å;*™k!Ç×ü”L;ÅÔG¾xÙÐn/í‡T ’~"Œo(ZÊá6%cåϘúa®ža\E}¡M|Î*iGÓÑ64]*Æ! y.}€_IJEKðKxîÃÕè9äþõ¤àpE‚ª1g0¿®‚8SÙ]Õw²é$œ†ue\e,ì‡ PÂîܘ܈H]° Þ% O«8Õ&¹Cí&…{Éü/Ý™lÉî9ÿ)z y-ü;‡RGÊO3Øwb=4@µ¡^ùÆŽ—ÿªŸ8$kéêmvÔ¶»aÅî]ûBâ;:Óå®csˆ€U—„³‘“ cP›LI§âçÓ^Å«‹,ôýÒ rõ*é Å!‹¹Tõ-AíªcØèxTè8Þ¥|8þÕ¥8Hôyó‰>wu¹JÑN”*»} 9ÞãŒÆŒƒ fúÓ::ºÍ¶w[÷…&¿•ÂïÓxõ¿Kfá‡LÃÐï=m‚eD®Ð*Îè–ôÜB·¤è®ø†¬*3òÓr s™Uš#*`}–j=÷µ%ÕdË»ãúò‡ólùW³!€ Î öQ drÅG%éÀ&C¡NŽ%L:ä¨0AÍ ®¬ôheY‰Ms ŽÆÝúnÞ¾ßu¨VÛUvDYµ³ βWoòõwË6§š8>ËTÜl3”5É]Å mÈ­% èÍÉ—ˆþZ¢=¢§ Mð´CIWj0ýr1u·1˜ú"õý;>ºÂ]PîdvŒVÃûõœ«Ø\l§ q“8YV~¼¤ü*°£5¨ZˆƒÓ!ž0¡d*9F0çÊ—áÔ×P‰ãEV7 v›b쇮1hθSURµá8x´@iuùqÒk™…VÒ…Lô´Z21<ËÑyÙ£õ—°[,~ÁàbIoëê6ŸêN©3òœ¹ôˆ©Ø+­û7É£wŠn¨æ†\S/W–@-˜YLÿV¶Ü?E¦èC“{M«ii· ºˆL²‡hzUmYÅçà!0q†0cha¯Ió éiãñ\ßWÝ×!ÊqxÜ^C¨ –2MøˆAÓñVOüÒ¡ôX2MJ¦Ót–'—€‹óoQóŦÃG¾@îw°›c’î䯛Z¡ ½ö›H?5c¥U¼$Hê#}E'YE©âú¿rd:üäJü(d‹rÀ(_µšÛТü&w‘÷…ãžù)ÉþÙ©šýQù™ ¾˜?VÞx zÙžäæ:=BÙ|áã‘ ÂK ±7›¤‡ ¡3š&EçÄù²¼üœƒ cŸéñÁúÖµØ}ùb¬Àó¯!Çfv!‡ê£‡ 7«Ø˜]̤¥l^QߘџÑo‚3ì5@Sœ¨6ès®?ù“‘¼G¨Ÿ~^¬!%D Z"Ã;Dc^SQI"x`ß+4¤€1YŸ¢O…`Ó”ôI^0Yá$œ1X휥1©j*àµ%ÀC#+öò Òß);\šm(­-«A™âwîüøg ”´¥äKbDýðXì#¼@¼÷PÀä#$"´ÖGæyºm¿`Ǿ›´#Û {fL°%¬$X<Ùk ž‡çÿñM$è¯m1hs¹£À@¿\ü‘Áôøc™.*R ‘ n9Ü—Ó —‰Ûà|嘭§eºÀšÕ¦"U#ɬnéA«Ÿû#©ø*RËZ*øö{QÈ /Â2籠ç—þ´½€^úÞ|"Š2 E‡ŠäÉ[·§†“û8švYÞãÀ– æ.°ÏÑÐã +šÿÍ-+Ò HZI4pH*EÞ_É’LI„Ȧ#Õ<_[`͸L,ÆõÒ­OF-¹Õòš#Õ¥GKzCä²ÚºŒÖÖºúæ¦,KXV”>FÅ…·ï-ÕªS}ß8PÞ¡“LJë#I¨æ˜ööÇ,|Go³û†|‘÷‡Wo …x6`–«ø³Ì®ŒÂÕ7 `ïØê'Â!²ONgGÄî™ã¥¾2uòg‰:!ÄÏCuãþð=ΈîÒO"Ç®ÍøÈÿœª¨“ØÀwî¡Å’Gˆ’¢ñ/²Á¨ÞÐÑÑX¢OŸ²örx ö!}pùqÐÅŸ5‚N}ÚŠ¢às9‚ø×Ó’w#Óc©?æm„<£1EÇ%&FôÕm'Í=Íw)v4‡´ÅÊmû­ùwSšŠúsË gSêöB»/mÓêØ¥) 8ã (Kyþöp¨äùÆ&®½ýRÈÍ‚~ð@Ïø r¾’>¬í•‡^Œ¨\aÖTÄÉ)©U÷‘|8Ó|ñ6òjÀžª2®$«Œ˜ïÓ„˜@-Þ»ðrȘ¿ì?Éì&L&{ÂðS"Lýž0õ7Û3M캌™þ©KÄîD.ýîš’8ð=aêmìÿióù/—#2ÄÊçXÉšKâ„“Š;‘–lúä‚}Ù‹òW€‘t­ ÝÖüÞõ¶æ†úsƒ õgGÌŒ²h’÷gØÕsõ¦•g²š»m#I-Ñå\[g_Y°wª_Û‘î·~‘/Áûr³Áé¢)û¹=DP¢}¿@kõ;ä&cHÖ>¼¾§û©©¶¤Œ¥\nÁ¡LHbXÒ[;»Í¶Xa]ÂöÂÄp.¼'È €i<- 3r¯íÊKø‡_’ý€ÖÚÿùù2‚-¤â·ÌÍÑûU Æüz.¯Hwb ³"õD6ëɸæT‹•(Ø\UMã  FpFÔdÎÙaeƒ‹“­ÂÅQ½.SxWŠú;~Hûendstream endobj 376 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 4764 >> stream xœXTTWþ~ãÀ{ODlyQÖõ ˆ½b‰%ÑXP¬(ŠË€  ¥¥03¿Þ;"eQ  &˜¨kªî&ĘctÝD¢Ùü=rùïùßR6Éž³gÏ…C™ß¼{ïw¿û}ßodŒÙ F&“qN+wÍ™kúu4V&ýy4NdßY=Îæã˜a+—c6nZ1nU!Ë[¥,ÍÎÿyô´‘Ò X4 }†3r™, Dåp4"ÐÇË;ØfÊÁ©6s/^8Ãf®½ýb›~ž>ÝümœÜ‚½=ýÜ‚é¾6.}<ƒ#l¦,ñ>úúìÙaaa³Üü‚fz½9u†M˜O°·ÍVÏ ÏÀPOÇÿ`›Mn~ž6ý«ÕÿÃ!ÀïhH°g S€‡g ?Ã0sWø8=¶:Ð1(xmȺÐ0·p÷"=6yn>äå½ÕÇe›¯ßü×,\ôúŒ™³æÌ7„aÆ3›gf³…ÙÊLb\˜ÉÌ6f;3•ÙÁLcv2»˜•ÌnÆ™ÉìaV1³Wf53›qdì™5ÌZf3YÏÌg^cœ˜MŒÀ¼ÊŒfÆ0ÖÌXf(cÅ cF2£˜WdÌ< 6cÆÄ2Oe«eï Z4¨tPÜEn0s2;i>Øü€y+g«9;î-~ïÁ=xÓàG‹,j†8 ©³c:téÐ:+'«ÜaoƒamÃ÷?;bÍˆŽ‘¯<7òëQF-µ{TÉ+#_©¶Jj«žÇ`D{£´¾D&íïY,¨Š4'ƒ 4Iª¢ë}9Fšu :B€P²§ô7 ªà(×¼eigäBÀ²S2A¯OmÇWÇ`){“Ø›ÏT²uú.h€:ø'Tö•ÎàÈC)J@ž4š“¬U9% ƒ å(Ãý(“K§±CøÆý äàä½ÜGÔâ –,M0Wr弄‡‚|è¨äÊõÃi:~šß}8d:Ûo+ Z&’ª?*ᬤŽÄzéOY-ZbZÊ¥* pììgdY8ߎ¼JF?™‰spáÃg8R$z²E ¯¾ßo×+Ê.\,2ÂÛP´0 Ä v‚+¿OѪ§=ƈ;Kp})ºeï¡­¤k’÷ÄõÌÊ!%D\îàrymÅ °& d™HwœJ¦¢‘C+´(LÔ‹ƒX­"žÌ ÂÎ9°æ¹Üôn8r P„oó?i¹~ëÞó¬Ëpp¨[ Y¥Ï„àËA_¡  ’FœhF×Ë^PÓèîð,™.8â6¶2¤9ðð8îå8 ÅYÈXç]aG*Ú8²âg\{3~š£Ô(œ»Ü^pøÛÎd ™è°kƒRYÖâ«°’ÿ2e4Ítt…Òá»c‘w랺ut»¢Ýkd<±{¾Ç4´f\iSýx4vX.¸-qŽ> ü†Í8'ýµ¥ãÓ‹¯ï£[ ¥Ül2bÁeNÀ1¸˜»ÃTΆ’©‡–’•À{lê`Ÿf¹/S4a»<ÎmºèŒhÏâL0¬Ï'ßÄúcœyïÊîÞ'¶ÊsÖJ:ÐO¼/¾G^Ž"ÒqOĪ"!‚w7DVV žv]âãèv‚RQÎ’E¿¥â=º•jø*ú63Ÿ³êÙI—\M¿–<.Ù¢g©Ðû¨v¶Z,œƒvxWúÞD¬±ú÷4ãÀÊþ¬L80]܆ÕhÇâ<¨\™G†ÒXI› ’M¬á}Œz_. •– Å~X³†ÖWð¨Ç™½#Ò#RUe`M™”QÒEcºIAfˆéOë2HIK/â­$c¢Ý«p¾Av†‘»In”Ø&&ãœ÷†qWœtÎ^n+j¢¯ 2jÝ~ØtHáëº7î¬C—O”QŠ, ñÓuäûôWåEa] ¼Ô…ÖFvf.ì€C@Ø__Bé…É^ƒtÛ(+3Z‚4LHËK͸eb™&ùßy¬khµ:Q«6ñø$ykL<¢ P x|‹C§Þ‰I1'ü“b¬C÷q\ ɱiz]M4ð† R?ß PŸ½Ín7î_½Û\BOQMµiÀ¬Ö …\òÆÁBN,ÄEiµ1I¢*üèÚÙÔKlg\ìÐáhÞ¡IÒjµ­B­Ž>¼Û™ÈÊ҆ܖû„KÛNWSóçÈŸ¿™“qN-òéT½ÚpšÜ"½0MRƒ£¢’ÎÆS’…RGô¡¨Üa!D«=–¡ ÓF¬dËõŸö9âgŽÌ™ JñOÕê Œ—®é8¬ì}y26-®¬ !-ëd6Öô £3AÌVé_À:^@U¿šI]ÝFâÃ…B<ćһ£†zi >Òb¹´Qš&dfƒÒøâ8É¿ØX²Ü¼‰Mû¦¬åÀ?fIm/KW ²Ë„QÂâs76Á6Ä…ŒÞžý5õ~²|Þ³ògÁÜÏ…¨v¤ˆŽŸ4Ø»g ÐúZK±á¼b'‡Ž¿ˆ¤ÄYïcÁmÓ®hj\d8ਿÐÀUÖ ¨½Þ‘VÐa8À÷C?[¤Ä9{qÙZ]}Ó±ÿ U'%ôþ#;F§Íë H+WHŸsP¦ÓJ-Ò—èŠhº2]ߥ”y°‚û˜WÎUQì*ÔÕÉ:-„ð½ (½³ëÑ¥ßì–ñOäÒ.<&”§AeêÞ¾…[“KîSNY0‘Œ$6ßÛãtœ~±»0ÿ8$FiÕщŠcë×…ºÒЉZ{]Qm¦oÔ/­/®;[f¤•I,u»é%ü2¦¡gx}_¾\¶òžrü‡€ÞDÄI$‘qd2ñ"^hClñ(6h‹^"é%w„õ°±Öóm×O’¾¦©ëÝ”ï7øqeM-7”ÕûÏmÉq¢^4%vÑnÇ@ŸmÊ¥À[áÓ>Ñà ÍÍ_<2E$?“•³F2ä{öÅ_ Ùz­&M< ¡üÁºˆSeõùÍ^M«ì‰Ü“ðâÄÊ»äÇ?ŽHRÙW‡sŸÜ«Ãè3²ÚûÔv_ޝâzáò±ëoÒ©kû°Õ¿>&KQ]P‘Z¨W'Ñö#šÍ ?u*¯¸´"¢ÎMé*ºWxdî¢Ø‰ÛÞ\í›åÞ䣈 ð¦Þ¥,?T”è [yçÛkq#¾þà꽯¶T戮`mq@’N™Ó@½"C—ŸËã4a>|pé|ðð!8îÛŽó׉«ðɇÑ;½½·Ð<Æëz훯ø©y1]‚?è_&íYåg ß²¿äwþçþ¥§3ÆjÄs¬7ޤJ< Ðvô¨ãÒ)ÆÖ>†«üóñˆ“ØkÞ¯¾4j¾è :?üÎq£®àöÔÅ”«ñ¡ÝZ:Ü[}Ú¼ïkË€ÿgÎã¼Â„”èp­*Z«HÚí@u'á ?êxb»ê.a T&K2ëéƒpaÊbçmÚùä …Þ¦¤‘Z¿Ã &ÒÈ1¨G.4®õp÷õñô¬õiÐÿ-ïù’Rð‰G1›´wÍ‘(1êË5ù`:ìt uåÿë~ZöyÛÍse!Kÿÿ‹Ó=§„`û>YÀ¿™.Çjœ#t§I+ØJmj`„ŽÇˆ¤«w³¹Ï–ÈLÑüåëx6`fÓ¨?Hùˆ€éy1>Â8ÓðšçЇ2—0À”¬gI™„*ª´E¿ÓX[Üv‘ƹ“£Õ@Rœ¸/jO¬ l×ì½…*½Z¯>â"¤™tlI–^—•!æäÕ7 ü˜¹ÄìȰCóÞöQ´ç^(«?VuÈïpÄÞùŸ-Ä!8«ëG«Eɸƒ® G”о…!1\E‹Óõ²Ð GÉq%þK rûÞé«eM­çoS‘©WWz&GÀQÞ«*¸î”¡°ùÆÁwÇ“déR’·H>Æá˜o?¥âm3í.¶QFn ðjrÆ tB&ÃÚÛ:žÁ“§Àv¢iúR)-L¸˜r‡Az*dfQ§¢wÕ‹U™z2F˜ªÉÖdñnìa”¡tÇfg¦}A“/ëXgl££| ?öðk[!£2/7=]Ì- Q¼pfšÝ¼Ç÷íVzD7Ú‰7:M¨™ðÖ©ëçàSþûÙÍDF¬W-›¶½jó]‘˜ïw„4(=w¡ÝÜ”‡4@Ë^kÙuZ$k¿²Êé;_êË¿¼¼{„rëa‘ ~M@q‰!Ü»¬ßx#ÐaŠKûgŸ\¼t¹J4]*}öTaà9†K|ŸÂxöuw7øžïW˜v³ßý¾‘ì(Å}o´—ãÁbA´hñjU" ñI||N\Zî³ø§[~×·8zØÃ«Ò«¥º(?5•¢;W=‘å`»oã*á iŸÍõ/d`ÿ~ }Ô—RŒÍôDN ¹ŸõpBµŽˆ+6¹és´pe1{ÚÓ.äò½eD¡äšô?@ ÝÐÔ¯ãú.Zà-´n5Ê®£5:ã+rIGhþV‡ªœÇ%¨–Âqê~³Ù†º¿ÿ5?å_¼ß Oy´°ùˆz•ù´ 3v—×Ö—^jò-õM/\jO-¥o^î°t‡ƒ»RAüHX\ÚŒó9¦ñÔÄhôɚD5¨ ™L£ÔªgÂÏy„¶:}‰D*ü/‰ï¶d¿=âÁ3[ ’BeA&y‘!ŠI¿ä“ßÈ*þˆÿ!¹é”Ï©Yÿßv¶"A!Æ&…EÒö."-*/–Òœ³ŠÉ‘œ3ñ@IVKö¥sF ” -Ì–X>“ii‰²rË¡:K+†ùoþƒendstream endobj 377 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 8356 >> stream xœµz XSç¶öŽ‘ìí„Jºà°ƒS­óT­E­³â<"ˆŒ2O) $+ ó<$0Īhµ‡ªu¶Z§ÓÚÚÚÚoãÇé½_Øž{{NÏžû?ø <<|¼«èøÀ¯Fµßæm¿Æ#Ô#d—‡»ýÂÀ€0û•®þö–ÃŽ·¼Îô’†y„دt÷  (j½CÀ¼ÀÍóƒ¶,þ$da袰ÅÒ%»–†».‹Ø±<ÒmÅn÷•Žž«¼¼×ì\ë³Îw½ßÿS¦™6}ðŒ™³>þþœN#çŽÚ:¦çØ^ÛÆEÿ`ÂÄèI“{XSÔÊ‘J­¢fQèÕÔ‡Ôpj õ>µ–A­£FRë©QÔj4µ‘Cm¢æQc©ÍÔ|jµ…Z@§¶RŸPP © Ô"j"µ˜ZBM¦–RS¨eÔTj95ZAM§VR3(–ú Nõ£l©.”ÕŸ@ ¤DMqCí¤ºQöTwj6Õ“ò§zQÔG”5Hõ¦úP})JI‰©÷Ý=¨ ¤¤TWJ-è/(üÚE-ì&,¾éº­ëWVë­ÎŠfˆžÓJú2ÆÜè6²›®Ûuÿ´Ç–ž}z^íµµ×~ëeÖ ½Wô®è3¥OKßÑ}cú>·Y'ž+>ý^â{÷Ù¶õ/þýz÷S÷;oë`×Ýî\ÿ!ý¿à4àåÀÕ  Ô<èÎQb+ya?Ú>|°`°ÓàÛC&Y=¤hè¡Úac‡-æ2,lXư‹Ã Ï^ùþÀ÷ŽHÏ+¬[À„6›ø%:Au«ƒÒºœMÎSfDA(äÉ1Ø­íÛÈMnÑëUŒ]ª> ©´(«UG{5Ñ›CŒåZu¶Z#9ŒúY!ÇãÔ)9¤Øù¯… ήÓ^…}p«š”æK"iGˆÉÕ0ØÀg³È ß¶ÂR‘5ß&ÞªIj°A]® +ýćPô’ÅzÚY‹ÉjÕÇ4PNž½·ýÙ‹epH‚–Ó?];yñBÎú5Žù§ßÝ! Úàµ4<“_]L‘n5̧í+ñ­Ÿi@…„Ó÷ø>çX¼è_ç?%¥=&X‰_VÆî¸ðáÌ ÿ‚û¾|1¤WaÿC£E~,Lƒod 446Æ7È ˜y+Ÿ õ¸úàÆÝ¦)‹gŸ5¡š½4ð ÚbDcŒZ£ ¼ÍvÆyCÔ„rHŠBðø°yøc’"»BêžÒ/s|§IŒ¨8D4Uæ7š›ˆ’DíÔgEhT¬+ÄÆ( DR«¶³|·æt Ù›KoõßB¢¯XÔ…Æ3Ž,ú">ô›¬5¨›5•ÐÍÐháÄ.ÚA™S'Aã‰<ù2C* §P…ÇA0¨6s««ýª¶sâ[óÀkqÌvBd{f³ ns %¹Ó:›m;üN†ôïHX«½M(¸žuЇ yµ¤zD›ÝðôxŸ1ÜT”ÞpO Ƶ…˜#[óœÌÀO+±©1†|Š"aŸö?AîhkH¹ó™MÞns¦{ú*‚Sä©R㉅¤H ¶¥wCRNšZ]RÂi4 ).9èzH¥'Ø¥ÿõ‹ãþ{âr%>Õî™®™D‘׿,Î;Ý¿¶|ÿ#Ô%m¢Ÿ†S'dÉaò ­Dòƒ…ÐÉJˆ‹å)ñ ) ÊMªgí±x­wa`…T"~R\)ÿ"†äÇ(5%èPˆ‰nÑ—iO¼kû‰£ù§ØùñúÕ›9u&yÜ^†_Eïχ ä‰.J•éΓÁÃxýÄzyôl«Øìûþ =a‹šòjÍaIIXRgMD{Xu*C$y/erìœe;U§äªÌ ¢täªMÊPð‡€Î¨ê¡P™›&Ïñψ˜Š“mÇ"]r.9XjçíyH j¯ö+ŽB®"×Ùà_lÓb´‰æƒßCï…üo¸“ÈK1ÜÒ8©Ri¾Xd¶U®†â‡¼”ÈÑC!?™Ð¶üÂ1u!¼S£HpÒHÛëcRî&ƒO¼J™÷>VÙb!2¤d+S Ëë¡„k/Ðî߫ѧ$Ö|o4ºM`›ã#•® \¸ ’ÈZ±)5Ó…LyDaX¸4Öß¹Éëè…ý§N•qÖü*ÒÁjž×ÍNoãó£y\ê†X4ý~Óñ¬ *¯Zn‡Bî±Lpqxe¹®hÏå¹ á¾` ÷áÄ?ã÷¾IàÚ¿õÌÎŽQP¹’ ¿4Î1¿aò4Ubývd¯ì+ù²{B$oÅ&äp žóÎë(è pfhj ŽtúËaÐ\!áëÔ4’ÿ½KZlÁ«]h3Ssªµ—­ºíš¨“UõÚË–Yå¯Pÿ{VñeÌÄsDÁª¤HØvÖ–dŒü{A“'ÕîÏg³rÌ‚Çä%B$‡¯Ç>•8öôKêQbzD"\ÝÖ+AJH¶‹É å@ˆaw$n4wßå‚TwQgÇ´³€ºÃGGó}>c—ÐhÄo=óÿÓï!Žý耺íåëÍ÷2 ³XüÝHdFöE8´×XÍ }WìJ°(e»ý·:K·›\º/x¿â1‡õš³U‡ô¦}5'ˆP4E›œsv“sFþnjXÉϬ¢èÖIlB®JAô‘_•¨tªU©B­‚¦­äO*QD§%«U©r~lÛKÛ¬8µ*˜|È(‘ðGè h~Mê­Ú¢ôØ¢ñnõkp ’´Í¥Û1ç¯Gó 6dNN¸b]W”Ʀ£Q{y¸KpÂ.U‹ Zпþ%aóÁ0ÜWB0×ÿçÑd`°>„„%Ù‰šä¥“ïævÑýÊtȉ•AB’J"_à[ ñX̋ػ¶‘·Ô¬+!n}eZF½ÞØÜ|óÉ¥~âè~%‹íÍ~ž%'BfèLSµú„¦ŒÈQmg𖤤5HÄ?¤òûØê} Ÿ4(8Èlª4ê«IGj‘šZ'TÚœ,¾ˆ._Ü@rýÈÛ²ø=ÓŠ÷pÚîê‰s·Ñ‡<Ã9SËAžÓ1^¢m–ù-I¥ŠKâ¼×øWíØ?‰ÌoÂé“Ï8>ïv$'åXÜÕÝŒøEEÒž¤JO}p6Q(fÁæùÓæ¦ÙÀ­>¡¸¨ªSe'‚¢søu² ¿jMNj­¶æ¸içE¯'¤t»ûê"_‡WKîOב”œ&¾ðk>Ô’·ç-?/½d.¦§9/cÍy¹•eö0ö´sÂÔeÉa̤êò¡‚)ÔF†o<xêæù‹ÏH_mßµ*´, 44  ,´ªª¬¬ŠØSO’õ™zäP/¨3¢Ì'(Ó(lß:“mC¢Ë^)M-¨Ac ˜À2[©O„1;!fçæ'Ú¯ý“Ÿ×°ß‚ 7zäï5wG‚±üâÔè 7(ðÉÇS-Kƒvs‰‚›o. ß¼#J Ñ®*تFåoDyM_üìÖ঵i±;%‚|¯R}F£'±JUÕþ=O‚Çq,F‹lUšätâ»R¯Íμp¬I£7{ÏE,—½BÓÞ8JTÙ2ˆ…ð˜8â²ck[>ì™”@ӯݰµšÌ†-ŠÅCF8Ö w7¢5ß÷OMy/F &·@ôÓHTÄiu ºN×À!s?óRmUîoتñú]?³¡U@”DxIˆ>i]ÊþÞ®Ñï’u\c#i£õ¿SóÉõ­ƒ‰#lÝ*lð÷ÙüúŠªËêŽU®Ù²¿[嶨ôÁÄ.%*äJ9^ֶ˯å5ò|Ë„Zu Œ–mîBÚVØÞyÂÏÍžÐE·½µÍˆHSäAhÒÒòI[¼ñK{Ë<äÿroÿ¥š1rœßíûï½Û÷i/ýÛÜö¿ ¾LX|ê@vƒÙSû+ýTd¬…e ü«JçC&|¹"‘$ptÛ[<…U¥*ÓÉ0¼ç4”[¸È‚Žíšíל=äz>±Å’6<•—©´Äéhÿé÷oBjr®'rjkµÕ&hãó€Ô'#5ÍäÓlѬ¶ôö_ÚµÿÖ< “ÐßÄ»ê¨ù–5£™ì³|;p»ÌTm!ÝÇ;¨Úa¿2à·Ò:HJP˜UølÛN’d4±X!B ¾›¹³´uù-7 ~lŽ~¿Õ„"žNίGÛXÔcæ+LmÝåíÅ¡´N^“DÆfÔûê³¼tEZ’J ŠdIx–ðÅàÛ2œô‰Z%1‘L<$FIð:ó34ÚôT.§°þÄ=hã¶ÌÈBO;lfÄo GC]CÝ=ƒ·‘­9¾ûŒ–¤htM‘±¬l·Ñ'Æ;iÛÔ+£QW‰øÄüø-›í/ÑïB-ƒßØÔ žRà óK?ñ¯hÿˆ=zh»wdPppQP.?;¨¾F­V£˜ߤùËWHââÈ LRZJZ΃/QwmøËr™%süÒÏ6MO‘ûó¥Oû‰Û-ºÇ–Ѥuö©;VÖ$ßià¢ÝÀÙ¯,ÚP¸'«á¬Ë§3ð{x±ëÁñ[¬CÌ‚HøÍßÈuâIßb!'n[ž›Â·2¨¾É.ÿ³ò‚Èf¸KfÎÛÕ¯ž†&®Ø6‘V»6ïöòwÜFZ„#R=ÿù-ÁIã7Ï…|ÿ†Í2¯cô NakP楔0!"¼ƒˆ¿¶0ëB^nFúmÈ#õtÉ¢fcÔg"àƒ%0œÐùða8”Í£á3íg™'RKó\CÛVu´:1ì²A›¦.bðöw( ðþz ‘ì ;ˆÊ†5ÿIf:¿Ý¶'e/1ç>oºvýâÆùkÜ·:zs†hö‹ú#-p‰y>éÔ„³ÆGjœ›œ¸œ¸ú#iÕ¯WL];²ÿàÿŒ¬¾yðKìG¡Ê^ÖÈŒpeQï9•Íþ°ÒÎáã3§Î?ÿàtõůšÛOR÷üa)?Ì\ȥϑûS3„èëKãØ¡ÛVï^¯ÙÄ̬*2…H£üÜŸØñb‰·ˆ|I‘Ó—'­Ûé¼Ûáæ 3¸ï×£PWNüË9ØßTzšÁ}>b£`ø]ˆ¬‡£pN@³¾¾âè¾¼ZÒekv9çú¦.‡mÌŒ£'s–D£À7‚Go„hoý¿ˆoÿAÛG{ðz=úØ|%)PáÅ|µøED›%ÝsÎ2‡¤Äß ýÅ]ÄT)o¬ïdCšÔj.C£ÉÌ0wƒµ„ `@ZÃcƒàäy´Æˆf™_…­­Nlz–F éLNB–,>9)IÎm^´8h|+ŒëZ"Švø3wÁtÜ ÷úbÖÍë§ž ®ùм«Eò¹çÑä:øŽ™ÒSÕ*Û¶õŸ˜ É ´“e'dgkRs29#?ž _’‡f¡?ÏÓ?æ0Sço”ðŽm#þðRŸE,¢W=^ð?žiÞÞIMü*$Óüƒ}Ð#:[)l]Žz±ŠÙË|œa=l<æ»ïðñÆË‡ï4\Ò×>1·ZOåz•œø¤é»®Ãƈ¨J!Y7ÚyÑö ±Ä rAô‰³—¿$@=õ`Á ‡uƒ]Óç×Nâ&ŒvÞëí¦Ü_þýWgÔÔsWÏD}Jâ¾pJóˆió@ûÙ » 262«kÏÂPå’YSBƒ¹í²§ÿ+Jë6Ö{.[³")˜a/OÜÛ[w¢…؃bd׈Æ7 N¶Ž¶ÎGwÙ‚Ðé" FÒö­(&ÂÃK @ƒ7°ùíŸDKÚþ&Šnÿ$ßìNÞÉê£vEåÑŽÿPQÍͽŒZÂu6h¬AŠzÞ0¿˜EÚȯeËwWx…ÇÄËåœJ©R©€QB~juÆÎKrÍŠ¥e2’S“e3æâîŽeÎj‹+ÊÍ"ýŸ\fÍÛb­þíZàûÓBô'±ÄDjvá¾ÈÆ­4м}hêúæuÜ$E+Ÿ¸‹Wq›p—8,$þbNîõ¶:‹{?Zš Lzª6ƒ4"?³¢+Ï.lN>׿Yj|ù Ÿ£8¤aÙ¼¼V[^¡Ó×AsÚí…©‹–î,)5¯¡¹B g:~º ‰È¯¹¯ŠŠŒŒ”,™/ýV2ΤqÊsó «˜ÿ½Oý#õìŒ E‘Àš–…Û2Ôk²rD½§~Œ<%E®eI¤ä»‘§ðdÀ€çz⸶‰m¯˜<5%=ûée$>ÃC]r‘škߎm îB”‰µlãè·c_6âA¿FÓh3ôèP#:¬ðVB~ÚmÞ¼äw¢ W£Q>”„‡CD ‡kp Ó‰1DÞ‰b B§ƒ’|®ãÿƒ˜ü úIùo©hùsót=à‡±‡3bwHp鿶żÆ6 Q*3ìÄTJ%Ä ÀÝ‘•èûGç÷ÔŇé¸Jy D1>å±%¥¥…{άkr{l‹þaÅòÇ@·èvU5sCO¸Ñ­.jöbè~·Njæ¦gfs•Z­ŠY¸ÒÅI"—“S¥Xh‘{ç¢ßEŒRt2K¼^$Ú—¼˜Dûg§à¥½äVâ{) R§h|ówb†`ÁÚ¡’7 äE\ŸIÍ‹/ßȪ€„J.!%*‚!23ºHþûÂ÷ …ŸÜšÀâAn/;Ú Y£GÈñ´  QrdOãâÕ4á9Ê4€Oò)(“5´¥HiKçjo[Çÿ°íw÷$ßOÈ/%M#-ÏÌIKÓˆI‰Wpøo¿ÎKŽ%ƒ5Qÿ¬„œüÔìl­¥KZm ‚†(óߎ®³p7ù®÷mßÎ,tG˜¼c¼ïÂä9ð!|œ:qÿܳ®‡›»ì—úƒßT_I» wì…¯³î°Êù­ì,<‡3p.ež(A=nf¡.E–ŒÌ^ sa9™ûÉ–E]ãï íö›¯7$t ºAÜwë{lh¸Ë&ii~ÄcŸ–r¦­€8ý¸‰œ¯y¯öÉÏ㎽š/=Q‘SOæhKÊÞF ßF™Sök”ÔœG¸…~þù“ÍÑŸ‘˳ö5§øûæ5š€¸³'à•M¼‹­±•ÓÖmQmîþòÈrïD%¨\éç'êOóìÀ¬™nž½n…¯Ã>2óÞ9ÔNü3O‹*Ñ<«Q²’K|tRâ2W[übå5Ä a¥¨ûç‰õA]§ýˆ>|>÷5Q1 @?³÷+N_%i»?ùä˜÷ç:Îõ1DT™t†ª£NÍ™Ž\Ê«fÿgÒ‰qÊÕÁ.ÿ­;!ªDU¨2 UI*ˆgÄm2YqE÷ë¸lÿ¸èŒ·¤6Ùd€}L]. Ô7Æe‹%H€¬Ÿ=û‘#†ûËoÕ“Þm§KšA@ÈâyéÅc!¿Îü'sÚD8Έž|Ú|è—¨uÞ{×®rñHIæêצ331õïRyþéÞ•ÈÜYÑ/ï7}~¤2ÔœŠÞ­ 7¯#äêŠE,$@ƒiºc‚¶ÈÜóšÚw(Ý1ÿÓÿƒÔ¹+¶¤$ÂÖ;ñCêß•ºÉkÜ¢#¹°ZWÝf`ÆánD÷þí4xýª²¤CT^=n8SV›XÁ¹+B!¤¹QÅÉŒµÔÀÏ/B9éÞ‘K›º_éÁuï:]׳›!«gÏ+ºž½(ê¿,p4endstream endobj 378 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 1333 >> stream xœ]“{PTeÆÏáÀÙ£á‚ÐIåœÅœÄÉá&‰²\b•› ”‚q[ÈÊâ|j€‹b»\¼!i²¥¨ÄÝ»ÀxYBÔÄÌ,++ß³}Ò´Ø?M3ßÌû=ßïüžç{_’°´ H’Iâ£Ýݦ¯Ë„y¤0ßBp¤®6%˜Ê­ ›GbUxŠc`#ýºYSÈÚòø|«;¨Ÿ Å6iKP$©PWHùÅÊÌô •Ä9e±ÄÝÛÛk‰ÄÃÍÍ[²"W®ÌLIÊ“„%©2ä¹I*³È‘D+R2åªb‰³o†J•ïãêZTTä’”»ÁE¡LoñIQ¦*C²Z¾A®,”§J‚y*IxR®\ò/­Ë¿EªÈÍW«äJI˜"U®Ì#B¼"ƒJ™¾:3:'w+A, ¢ˆb K¸K áI„áDaKؘ ,ÍOŸ‘b2”,%ÏYäX,îPaT ÕI=²ä„J±Ééae¯à¡# /(a–IÆ.¥qÔÌê&}­À@߯‰Vi¼´,ÞI_Pý@Ûý=;éªÒ=ÊæPF,hq¬– Á“Ëä]#l3Ràù5Û–gP "槇?Ã|°^–oGD¦&ð]þSvF‘ÐbÙrâds'bn¾ïº00È?5_w¨ˆßºU¡æžöšKð|€l° ž +ظ–äO?DŒ;!Ñ{‹¹²c%=Šqäpõ Îcm-§P'êI;¾£:õDþîjcêêPsqiuoÏìãÄ0¥¹ ú_èÈç£H…7XpêÇN`Œí=£îaÑtcø|«¶•+((Ì.ËFi¥G»y³}M`«#ǰÖH !B$ë‡;iÿ茕;9¸ç‡ФÍë&ºGöwvqën‰Ä&;sðú~0ô’OÀè8Àilÿü,_8ÿˆ9WMÉ&á4 üÆ‚Åõðe¾«B±ˆ73h±P‰“ûà­ÑÁ>Èè#/CȰnØÜD˜Ç6¡Ã5-è:YY…4—桨¹[ 3÷¥ æ©[š6]WÀç%£T´ Ôæ5Æå”% ÆÿYˆÁæÙÀàñ»(•[JoYºëköï©åõ›ÛK;óhøæäYu‡fíq2›×<1q¡¹£óŒÂÏ/^‘È…£Ø&°BÁgyû¶©FWŒ¼Øä`¶Û5ÚWvÇÆ)Á ÆØçG¯ qæ&¿Ãorø“)Ù"ñ°Õ$ý2D³ ……"`®F®ŒÄ3y1”#½°H‹&Éñ[x™|YIº/bý_FÞ2GsA}Cøå*Ù;_ŒQ‚ÜäÂÆ´Ç4 óž‹’N%ñ§×ÊJïÖUVê65e¡D&4&Ô-Ù½nÄ‹Ûìó0æÛÒKÈæ>þ,®äœU·ðñç²j]›£JjrVìVAfà›ÁÇ÷Îfåtp¯¨LNºWHnæ%òŸ’™QL'̬3†¦'cÝô¡L pž½VòÏŠˆS¬Iã níÎFRŸœŸÎŒJ_Æa‡èú¡ñ›?ì’É9lø¿™?^ؤÿ}‚lŸj‚Fػ˗·Š”x¶Õ—4hÀî¿zzNªõEëÈ^a%ÈLñì.-ªFUL}y­¦tûÖ­•~òwÀ¶-¨ípÐÔ–×7ÔÔÕUsbu“ÙqMû›hœ¼W44Óø7ÓÒKg=c¨ÚÚÚXo=‹ þZþÑendstream endobj 379 0 obj << /Filter /FlateDecode /Length 13406 >> stream xœÅ}k¯ÉqåwÂØßp¡}ía+ß-¼ÀX’µH‚­áZXH†qÅ×P"yGä<4°¿}ãDDfEu×½rŠ- $“Y•¯ˆ'"3£ÿ|åŽþÊá?ýóñ«?þMoWÏß>pWÏüùç½Ò?¿ºú§GTçvÕ½„rõèÙyÕ_yߎø—Òò1·põèÕƒßþ×õñ·xøïËãÛkt®%x:Ÿ¾œO7óéÍFÙWæÝùÁ?ϯ7^yÂO½$ý žnæÓWóéÅ|z<ŸÞ^ÿÇ£Ïxì9û«G¿|ðèïW£z>[y¾1”—óévc¿?ÌÇŸmT|½ñôtcn>Yz³Œê©ÁxÒöÖþñÊü¢7o½1‹ßlÌÝ7|²õée’oçÓ˲7Ë+ÿ¸Õ±¥™¿,C½¶ó¥0«ÙÇjþÃõCŸ«]À§«ñj>ýçÆB/õ^l¬Úæ’Ÿ¾¼^«e>·äsyãõæä|¹QóC¾=žþs>ýñÞ¯,OË¢,ÒbäËŸoÔûjs nþûq>-eaKì¾5 B¢«ÇÞ;aØ£'„XO6óÅ;Ú¾WÚŽNA;Cê,\¤d í©[ÞXdòvãiy÷‹-ù{µ±_mLü"O6§áG˶¨ñÒʶãC:q¾â š½ÙèÌÛÍn/]|ajž÷Áí³_m|fùôw›MÿAäŽìeŠSî–×_oŒz«™›$šY~¶ñÅ7ßþzã‹7›C0û÷ÃŽŒ9ý‹éï¹Ú®†#ʤ³v¦L%´jg±¼ Î~ñŽAmÉ&„lAémHò?Þÿ™ïm­Ÿm´÷Ò<ýPûˆ™ýù£ÿÆ­jºJ®¤c,WÁ;G=ÿ|óôê·W¯‰&"…?»•w’ëÇT¯Rô1ÁâÑõ+ß=ý=wèýˆV¾»úƒgÂ1ךóÕ·Ä/ÿHÿûÅòZÎáòÕ«¡Çp,~–¼|ðùR«WLÅÖÒ[+ÇîŽ5šZ£dU«ÑŸÝ¶8J¸VÉýçh^‚âc;i/¸âð¼T%«Z©„£K¶––¬jµ1ÛZZbkåcp‘Æñí2#1Qß“-YÍ[-þØVó¦%«ñÍ}°3¢%«Z…d&ÚG‰­UœïGÛ HÜÐC3¹ZbÇê[q¶ þºš±@#ˆÁΘ–¬jšëj+I­W©¦Œÿ/ýêÍó“‚ßü↗CO,¸ŸÝéB…æ¯h°Ñ»¼ò¡Høz%9‰å˜\(ò£Þ< ôÇ©¡ïõÁJÄ”º£#iØø ŒÔÿTøÿCâaÒÐÞ·éTŽ-ôv a³éwÎæû6™Ó±åÂM~í2„¦¾ñ[ºì† /^ë™ÀêÑ•\:ú¸ÙæûN éP“ùÃ\ìðÁL³‘’¬HÙ”«÷î"}±BPGy‡/G+ Æ„˾ìòE6×ôEÂ-¿Ç¨‡t¡DŽ. •ˆÃ…=¦gEqÄ$Ò.K3™3²²ÑÀ}$<,ÒÇt”ì®Ï¥C‚¥'K–v™iÑg,]Œõ²RC,¯%`I%}½¨AÁü¹°Ùæû®H%fV v™æñ¹öÆ@×·eú#Ì h&€+”í6?žÀÀD ã}©ñŠBb0î‹‚rï cÖ¶h½7MäK6Y¼‹Mà‚ÉK#ŒRV(ç6}ܡհC'wáG SäîY‹S’ùÏ‹®Hkô• Ta&5FÃV$nóžµÈ`ò}Û û’ H¿,{KÉ)Q—]$LGjTâd+ÄN =ÈÏ¿Œ÷2‘-'hYv¢Œƒ~”cÜÅ3±&1— ³þBºì„ö‡}hK³»òæh>œÞ–;â%»¼F“Kþ„¿ï´0f»´]ÜÏB¬5‹Ëû °tˆXý.’­ÒDLûX ûéu—`È”×xÌn•>¼ŠÛ ‰vpýè¢Î ßE?݃{~Qz­£á`jÚCB#µê™­TnÛ£Ùnˆh:—™_—mÌüx2ر7•˜9‹¸ J,Žun.àO”cŽ—Yße¼`!» «ò=€ÍÛè¦s¾½}¯õ2l}±$áè¶­ÝGäÖbs˜[_l¼ÿ†ÄävaD0xžÊ.ÒÚ Ó(Ƙ¼ËqfÒ^ÞȺ hYç÷‰;ÆJQ”EÞ&þÏæm“ ’ŇctûÐG%þá˜/*„\ùzáHרIG×/k˜ ê¹ïÊ} e‚>õ.ëvpèhr‹»lWÎ/òîöè£]™²‹<}1Ëvåç~@Ø®ú}¼d!£Xkï.BÓ5„ê×};êZr`gÛb|+5f0’mÙÔhœó|Ÿ`k*}ª4þ@& Ç{z9†ˆ’õòÏ>+ä$±r 6Á© áI\ êGá‚Rpx-wOÄ;sI‹Ø4Ì,‡”À&|Æéá1_<?*i´:…Û.¤c8kЈTðA)_2þJÝŽÿZÉœRAé$•\ÐÈ2”©â*ÕQpˆ,#¦Ä%P—úÛR¡@ ¶ eœ€âÑ>z4…ã|‘Kj9â34NªP7ŸE‹¨J%Í{t#×ñ/(¡9œJÃR‡¦ºRãµFt %…ÝåZâ±KF`…’L~)"iÁ:Q f¶£)Ì”%”„zlüV§w4îyo%4O°x•:Ú¸l·:>møòApô„9.¤èÜTp<%Ž%j´%Á7BIÑ–"q?žÒEPBÌ8ü’¨ÓÉñÀ‡1×ÜÎAzÌqp"Ú!¹ á$öÂÇ%ø 'Å ¼‰Dº¢qòÛ<‘".!Åæî¤Ê'L©$‹èиó‘«$‘“=Q#4NÂÞ¹J¦Ç"%2Ç9ñ1*ÑCŒ–âÀ%^Þj¬=T¢º¼IIJ·UX&üXž}/% ¡dšþ %ò‘§\¥€U1”ÖX£ Tãvªã-D.©ÜßXM¹DFY 4—d)!¹ ^JO(YViSå®¶¤Q?j³ßi$ ¡Û¶in[õGJŸmIŠOóVŠÝ~™JX‚–Ö©¤é„iSìŠocTRtÖu¤)ÑzäbfƒJ2Ôt™±”¨‘[U*ÉýcæIî¼¾¥«C%U!E0ÁNõ`9Qc:RHýµ¡x“”ÈÉÉ݈RJ…iK‰€&v#‘ô•$ð;¤6%X”`$;eI]dŸ¨ k´¨N‘ꜫ ¥œ¢ÊµªYÊ{ÎUL­ú]ÑÖDÆJÎåNˆ„ÉĈÒ'¨xŒ0ÉGƒ4ê¢0 ø’ Ù# ¥B«ë²Á©T}QÙR,K56Õ_A»T‹j•â!Q‚°ÆÌÔȂʸWÌš…ÞÔh‰K3ðœZÅ­á †^pKP>õPUÕ¤>lô0‰8§ÊÕ0(kmŽ#sžªµKŽäȯl—£e‘öÍuíñ´B]¬™ôˆkJýX†in=–ae’A.ãÊláÆ²Ó ½5þd+UK }O0ˆn‰FhQ¬Î$#â\ _‰´xnEi"³óÈí‰8‹¸¢F§?äÃJŸbò"8“bÅtÂÂ"¶K–'4.–¢ôeP½Ho ©t0b>…U*eŒÍaø†VÆ¡àƒyFR²*ÌS¸iì^G4èkì<²— ÅÊJ2(t“d(t %iy1*9¸ù’rhO-¥nlSF¬c¿2Ž[¸hléK(ÚCƒ±•Ùs m1§ÙQSB„‡úû‡vÔ”|fphÛ Ê¡±fÁ2=E§k„N-(«ÑiÉ„¦Ñqä£Z6Òõ–°ô0Ìý 5^¦¥Ð$‘£ÎàF >¥pfXÊFB²¢Ðd”ÓqEÀáˆ¬Ü i-µÛƒÈ±k)tª%ÌÂJ¿&ðfÅpb‰[K²˜ƒ|VB7%ÞJPÙy¢«$¨Â<·ú Æjra€˜TQº Ó{eK¸ã4Ö2ñ¹8D{ŸEsW'‘ðÙ ‘20]ô¤@F,‹p@èÆ~ Aý3¢wóU€?Yõ¹3vefÒ‘'„>$Œ°¤iÜVIl’ÆB2‹5ÁLÝòp­aNøTƒ¦Âx°m7Í7Ô8¹ ¶¹† \Ú'„·ùl´wfʼnrýÛ3zFRpÁÓ 5Sà¬D ˜ :á…i<ÌGm½ñQ*Á"1W-ËÕ´Ð5à!É}Ån"UÇ:t€Òˆ)í4`¸{8V”ŸE¶‚Q~au´dÐôó!ožP‰ÿ™°<±ZÂ*8;¦ ë¹=<¦pQ¡Ü³yÅ•QQGŠ  4×V) ê •¢´Mù¡ØŽC*“]oÅ“ôW&yRg z:êž”"$)$óa ’¨˜ìUuU.ä-™1'!¸$p莛–HA%Ø z§/Ñd;n¼ªyÂúÄ(cŠ‚ÿDÄ27>@×ñÙÒ€°HìT¢ÈlIÔ'&‰6à. ïæ¿qVõƒG‹5h%-à¾H¨—h)qj± å8rµ!¼Üb‹Ì—t [à“ȈÚ F2e×àˆ“0$K]Q.ì4UžZB¶§^18ÃÙ)ÖÂfDƸd„Ž2˜®ë? ¤™"}ìÖ„¬«P"c&tC±³@kÕº‹”’EÖ€!âcª\2M¤ÊePr>pÈZT:ABZæŒ3bÔw!l‘—ŠÄn 2-I¸}*”ˆpà„|˜Ã ZÒÆ©¤«FçÈRÁ°Ñ{*)r}xŒ0„ÉÚê,$ì8ôlfŠJº¸2c6 k:BI¯¡ê¢Ý‹#¨ ë–p7VÜm][Âê¹¶²þTÒuÂUF¿1¢†¼ª·Š•T#‘Fh¤`6e70*ÔTÒF‰~Ê€‘h”#ñmu«@ $¥£‚¦ò JH >UÓDÖPy®ªr"·@ù©ª{"—EK˜M6¨‘2ýƒÈ«"K‚s#QŸÄD‹O©ÐØ$𪖰¨Þâ-Bá †ÂB” Z¦Âj*5+ŽqS¡…QHNˆ+·ئ•i#>.Оx—¾)ü'Hv·&‚XQP¦§f„Ä!èüŠ¥IP™š1"º•Ô{Eäfl™¨MKp¬4è/v/5—ÆæØFæz↩ýLØîÈÆ¾¦HÞ°ÁTÒ†-;à‰ùT[Îá oí=•t í N@¢G˜Dx ùô …[PÉp•P‰Ú‘ÁQ˜'Ëcv‹4*"\‡JúØ×>Dt:(CRΔ°ëä…w ¯¢’¬ÌF¹•TñR…š1)W[Ø•Xz‡¿jç”r‰Ä”&Ú¡’ò–Fxé¹D\sa¤Ür²¤•û&3£Ä–JÔª òË#ªø° pœíBNª ûAE†*ôénØ Ê!;a/ƒ*ÃNÆhúÎX÷A•Éq\Ï󎶢ÊÁixsRex¤ÞNvÆí~¡ÊðbS0+Æ^f^±e¨¼î*[ö“"¶ Y‚zƒ-{‚o 9 ¶ìC‘6eËÞ«ã9Ù²wº5ٲ뚸b²eG¨·"Ë®iœt’e‡®'£`ÙÁ!‰F ɳ(²7É2#L0ÊœsðÞè;•8Õw¥Ë.xíß ËÎG1“.#¼!Q–?$-[ºœ:ÜÉn@¬ÓÒˆa8Çù@¼ÅÂŽ}@ —Åf!;Õbìö8¶4÷vaìî!(+àŽD'b=þw§Q¬i" `ºá¨V¤uöQ¥i²8LÐ*c­heа--Ó2ŒVT¢Ã.ÂÑÕ=@µ,ºx×þ–uË@mp cŸUÍtóYpzZòÕ­±oð¥-[^ÌÆ$ ܲ"µ©Ï:É쑳ü¤ÒÈC¶û º“£4»Û.X&TÓ°©ƒ-!¦è,[& št.ëªpĸ-%fun¢îVòƧü®Î]·Á«SosòD˜ø&ÌQ¹dé~ìW)ß1ð“-'„PÓŠ±–šÔ©¬–ƒª=æ‹ýïbùrÂxl.$Á…ûGâØ! luŽ$Â'M|à¶t>-t ‡S$î*!I2?AiȦ”ú,¥yU·kiݨëõ»|:Å$ɹ¢Š´vd}–÷ɤË9M Ã)gF­™.ç4Ž­µ¤Ë9M ³ª5Óåœ&бµfºœ“ü9œžFF0“æLôn’æŒÜ6&iάbs謓æŒ*§itVIsf¥uéÕ‡IÙÀ‚›ÔØCìšOÔ]?l1"ðzxF̺·C’g„ÓÅ-ù¬îÍbŠóW4»×níÑ5,2»ýr'ãpóü©9Eõ½?‰Ï=Äw9"›Í±¿"Cyûõ›ë‡Øãm1¾yñÍuȇ§O–¾{j(tÙX"[ê æ»dØÎERe}À #Ns „\¿–´CO¯¡ˆÄYºž‘éëû†ûa9"Y÷Ò@”À07O²Ýª,ÂâcpD¨ Q}¢tü'Qž´ýµ;÷ÖH|ã‘ `ózE8XfÉý&ž§w³Ûk~_¸w~Nû2çÆ¬ÕÜП÷ÎÍé8æ¼8)ØyqdøüÒ¹¡ù`Á%’ØäÚØïÿÌÂéZ;<}uóòé÷×ñû‚:±Ã¯ðÑ1°y;Î`?ï΄—G&¼Ÿ#çï› ïÝÕ73áÍ×8ïÍ;­¦° )À´Ï:`¹¼ß0+i‰­ÅYI¤«:Èyçì—F מ ÜLŸ6Jlþ´.[mîºQ²ª•;ï%šZZ²ªÕ;™1µ´d;Þœ‘(ûkfF´d5o0ã¥ØZZ²š\Þ-ö[£dU‹“;¬¦WKl-²ã ¾ÀRk”H-òK«Ía7Jìhcï‡7µ´d5s±sXÃÌœ–¬jÕÎû¦––\2#^BÐHOi·mkð¾ø‰C£|ÅŠÀëŽ,ëÜwäÜ|ïeĽOicó>è…¨¶ËYóˆ`dгøí²'ç¼mÉO·}ïg÷ù“ƒIÒðí’C GŒ‚^6ª};!áî,iRÚöU¢÷î#}±eIÖ°ËÕíÄ™{åzÛåbjBÅ×û®OtaÇ&o¯ËÇCƒJû¤®\d"2ÙÙavp<1£È§wö2œ¬“ÔÄ}.£Îp’\Mò»ª3ßãs—½AÈ™ÙË\û%quÁÂvme½+NE‰Ã¸AzÏM¬0¢™8î¶ï²}¼ñV\,/’xeûÎäG¯)Æ{Ù”8«”CÚñ†hrí{±±Ê Ê 0üm+'ÒºÐMáa×á$ÝÁò>˜)„cÚ‡ÍL¿Äûv2˜'Õ¸õ®Ta;«È†¯òÿ y:¸ˆj!6qi3ÒŽœ]!}߇½ OÑíÅްч+8ÈLzG&ÖÝ'¢5#éi>˜|à–Ð.b£‘³âÞ‘*ð㉠òWp¤çX·oP¿ï‰ÇÄwL‰·Ð[€Ã…_aòböñ§ÝÎÇ´zí½eÇÉ%“s܇ky׸îPùJ|$m< #öÃ Š¶³&| Ä"šÔóe£I ³Â.ã>áÌ(ñÒûbX?ÀðÅ;~)â£q‘9šÈ»ƒÀÃSœ’¾Cö'#à…‘å ^8n™¯ë.äQŒXVíŽѸ€G¬3—/n›ïÂÏ„ðÝóó1îä8J¹ó7}v·ÇÓŽêô…1zî6ô‹¥ÄÑoùI˜¿‘z Ë>´‚@ —Š`h·'ð½ûˆÄâî¸íM¹ýUظºKRã”á”ȼ¸]òû.ñlôÛyÌ?º!†Èæ°Ïhâß/·ñìzáÜ›c †‰èÛ?Ë0ñnDÙEªg` mö0u‹æ!EÄ>Á³E—Ë.9:íÃ. 3ívånï{k¨hã_SÙ&¶¸ÖзK0dQ>nqi3îoÛE×ç./C–ô$7/"}²pàÚáA¶¯ˆÌjEïFÄ¢¹¯p~žo‰Å¢¹¯îF.iòË{­êµ[²Nrw³»$§Š‰ T¾\ŒTz|Æ>VN>–;ò(x.Èœ wj€/²ÄŠ+-ðWj•³æ±jú+d¢àl±9ÎEA¦#¥@³_ሟ˜ŒM®îÓjëϲƖ9ÍDÁý¾ˆ›\Ý/H|À7¼È½åü_TR%ZìruŸJšÎN—»ûTÒåÂDì‰ï½“ž8¹3»þd*JxT— Š𥫈«$É›’„ÓüƒŒx)s ÿJ‘”ð]’ä2Ÿ„æÆùê „Yq’˄ĜÓ$q‡ùô\ò8Td˜|®ŠüQÎÃSѤNäìTäFWB. Ü×é]/%,<ëYîâ$Ï—{2§‰Àl¥€AX)Íþ—pyDÎòð€p©˜ q\ L8õ°¸Èw—¸¤6NÄl%Ünè…ÓŸð±E~ îq@pd ÛÁ—m2’&ðQ»åj2Il”+) ƧÞk—¿"I±çIIޝ.sª¾”Á×qqƒª(Êp8Õ:çÌÓóõ‰ïˆpBÕžDFŒ/1áª)_GJÈ·â5Ë ßÁ nÈ·3·ÀS™CÂÕ_N+¤NòœÈ.ô>ÞÊ]î(㦊¼E$ŠokUM¼’2b›¸ãåƒ$FâK¥¸‘îR—»©>rަÒÜ4¾LTäR_å4E/¡¥"#LÍ]þÑÀ•¬ %4<ªwº¾¸Ð‚¾Ç÷˜ø(Rq m_uI…|\dõp+‡yÙÉ•¾Ý‰LH:!Óƒ»œž“ŸE¹a”N¸®…K:2ñ-W¾ƒH•ôír‰C–Ü8®J¾]u ‘Ä—GÒHÎäÅ>S’$õ”ø®H]’ÜX‡õÅŠpS¸†‹@ íÜ¿q¢4"ÇÛºdœ]^ZJƇ—’ÑøR2:¸”È–‚1LS¢SaJtºL‰N©)Ñi7%º4K‰®ž)Ð6%*¦DÅ”¨0-%CàL‰ ¥)QÁ5%*ܦDÀ”ˆ’˜U$["Ê&%V!¹D]•–KºUl*‰ªHªü\ÒŠ*Érgl€•h¦Ë4D½Ü÷`D%zñmVBÞ1…0µÄÚÁ=Ni(­+0’{§YAxRßõŠíXÎî'= ¦% ŸŠÒ©jÚ­ã46¯Z£XO¸ÓµŠš×ôvÿ´@^ŒIA 5^bt‹2ÃÃ.!A› Á°]}hƾå”4SÞ° ÊÁ˜IΗ%CMiÎAò=¨­ÍYnÐ-æ8#•™ ›šl²nh1뙯þzcú³¸¡†púž’ ƒàŒ¤©–!©N“a"\Ò«a+œü´BÃIK-ç¡–ô¦éàEœ£LXšr'CS¶7øUFäJ>¬,gM37iZŽ*Ä“Êe§w·'Ý#mW¦¤Œ0t”“4"H–…4*±DRæ Ÿó^õ$¨ÞI^µÁa Põ&ù๠y‘yà„ù®$[¯÷O‹7Èt›Ì†Lã–j¦#¹2ŠíjΚ¿d§ ŸC®ã뜖ªY¯ÆÄU˜;µ Çx¢~$'«%Bú¸5“Æ}ඦÒ>ZYèa²f•Ü©öÍÊTOnÔQ±c™µ<ÚPÿ!½ !VÞó ¡=kVž©)}\ŸÚ„·”6«Æá~uXi%òÝ ¬ Íí©) ¨rsú²dÕ¿Ý –I³C#å¢iî2pú¼ž zлší LâtoP(ñoKDTœ.Oèžb§ËÓˆ0Ã]¦(þ "&üøŠ@¿¢fb­då„yò–¢¯¤Æ‰¡ù $®§(žpK<zþÅomA5Q©½Hp쪵) ¿‹“ƒ±; 9¹Ä³QÛDò›Ä‹ö‹³ÕÅÂq=Qj5‚œ I F eâ¬]ÍSN¢§6Yì-§b’«Mf\W»MjêuFL;qGMS9¬?§Ì‹Á0Î'%œJYóK݆¦‘๦nØ•Ôát cá€S3¤†óÂyu0ñáÜq]v…„ÍÜwƒ?ñKE}d %€ÛJÿ”‡qÛÞr5Æ0'%Âç|ġʑzoE=AW7úÅq/Ù>Ô©øÄÌêŒìú‰uvM„“:»-›ÔÙU'Ë9©3²×‰+¬ÔÙå5²ç“ùàìyô¡ÊK3}ÿP'>^§Ï«—HŸ·LФϛ“cÓçÉyßôyËÄŒüyËÄü üypùqWÿ£åÏ› Ü—?ïŽ/}!­-°,&×'\'²âˆÛd¹NƒCk+ŽÃ ~DF¹Ãó¯¯"=O'RzxsÃFøÅ±ÃS‚ÎÖ~è?¹~ˆÛåä}~yûœßíÄ0Ãá-^^§xøê™{ Nf?ÿx~†¿M¨ØñéùÚs)¦6ëªü-ò£RZµ‚/"ËÔ•Û×òi²<‡/¯Õw’ºÙô-ºä‰à”½ݨ‡gø\q$g‡[`<3ê³T4€Ãók¦GX_šîäìtãG¶ˆ´:Xõß~v ÃÞK:ÜΧ×OOçÓ›kýìj#aòã³Ofwn̈ndDPæ ¿€H]_wHZkþðí5¶€Hs0ôï•J[¡á»Ô x.#q^Íö×\#ÐòèêÑÌ—Ã7æ²Ò%Påo®‘6½Ó:Þ¼”i­­®*£Ÿø==²0G©‘ªlΨ9m­§Ã'ôjÞd>üášø_r!¾[¾÷ÖÔµïýeéˆùö/Ÿ¢Išƒ8×=4¯™Ç)ÜMeÇÃÒqñãkÑ’(ÈÃüÜÍMYaÙxˆŸã!‰l,Ü\ÝÄâYyTÕ!”¿§êðÏ6^q]ré½[š ,k F.t"\)c¬åÿ ÆÕŒîYÍÿB_hVø1Øg'¹ùó2§_/ruÃíJÏTXÙ´RöX:A®Œ#t"Ç÷ C¿üÏ„81VÌ.3R úªÉä>â’*¥*JÛɧ«éÝh„¸(ŽøqCêÉ—ódÂè¹ÄèŽ×•cƒ¼#Ý0ueõa:ѳñD]Èìë9˜õBÎq‰åB¢üÒÌQи’Œ/Tu&1ýVš >~Xîæ÷×ÚïD³1·¬'W¥ºµØ·¤Ÿ1ÄB 4ê*X•6¯¬þõú¡Ç‘yW¬ ¾“_¶PÎ#O~ïÑH¨v'eö[­àxG$¿Yk¿5SÇ¢™:-`E“ˆStš."%â¨5PLóLŒEVSæàÉRc‘Ë*¶Ù×9ÑíÔš`IÔŒ¤ß¼d“•IÒ,^MùЇšy O¤¸Ðž-ko_¼ëÓs¹,B2`-*0&ƒ!CâUPõÕ[éUé$Mgê?Y:k»eÆ`¿¢*¯HâD ÙeBÿ°X¸×F3¥©Föž§ ¿‘C¨d÷pûõRgÑ=Ì"vÖ›J·C îD/òkm;Ã'ç*ƒÂLþ¢w}MVžÈ÷ YœW°µŽú#PŸ“vÁ¨¬!LTókÍšáÑ7ÓÚt/g2­õÕ¢Z”áEŧ»¿ tî  ,Û ß0ªµBKkn©kíÄÞáöKç W±âcДxbÀÄŽÖ&ed’LÈ ;_([ A$âÍv-t†UžáÃm2Ÿ2s³Â¾ÍŽÙâ¹ ¶ÛÀýö…X­¦òšé ‰ý´é%N:G’ 1‚‡cHk qºöÈúF"áŰv! .Õ$›ï›H6‘[—*Ñf+ñOÕ8•-Sz»4bdXxŒ÷)Ÿ©Ò¸xÝÑg³úFRŒ ðÍB9Vòµ¥>¦ôišd(ÉþÄ—¬—Å5oaÓª,NçØ^œð€ITf¿»nøÌ:4F͉ïWœ‚¹!Èžà,[¡åýåƒGÿ»ÃoØP:—œÎgè=S3ÏØ Õ•U1Fˆ­¹âWůc4…x~iÃ@ÁÚ´½ˆã^|*^X'Aw±'#M/@ÁwÊáS<äÊÀ"†Z¤@„A#©£(†ó>hÈœ.ýiˆK»>tù§,®ô_…›#sS£tAõÇieÈwñh¿ƒ"-ue@%ü*ô[8üþs>Œãê¤eLÄha¾Å…ƒöÜÕðTªzø§„ö)BÚ÷^.ú©»$“jPÜêÕ±áùÝ$KG–&ÿ¡éDtžæ^¯h ghl§z9Q݈³14Ü 2Z˜C£¸GY\¸Ÿ-qûÅRãµÔ(´ü¿]0Á¬q©,|~'E7þÇxl‡Ï ͰôÃ2Òt¾TF ÿ¿É®<äRò"nÿ$ÚYÂ]´Dß$h8üÚЋ| Â×9Ê (%ÊŒ› ¸Ã XãêD›5(,*ð·Á_Äi|mæ4÷Fh¤ŸÊ#àI¸´Ô^±&ªÄ½Kq\(ø¿¡"" YU«ˆóßOõ°r‚`è!(=™öׯŠq‡yL!ô¡‰ìüw½² ﯊صZ‘.üö4,"޾ XüN– ëóMØHÖ‹íO¼•OÀ…05 ’Ø_#‘ø×ùaþý+‘ê[KþoÀÌTUVZ»Ðáù *æiIfçó©¡M/c2äç>gÕ[B!²f¦Šu)Œ%‡Jcƒ‡¦yÅr—šÞ›9x²ÔP•ÇæÜièÆªŸ¸"är”5„N‹gJo­sTfB~ûBrí—‹B*)¿[:þЄþÅh–L²°gÜõ\„,8ÙÉ\‰ÖäBƤ<^øÖÓúÝòÒÚ.Ì·^+$E¾gÔª,&$òÝàqv!ªf-Âd­“Ò\ϰð'3èõkk¾•Ïy2ÿˆL°èÁ–§Òÿi™p¶{ãÝÿbAäóe€ÿ!ìDiÃá׃"ÇC¯í¡ø&Ø®KÓ\»‡‹g™*ýmú¯Ø¤ºÀâ!ŸpGò ðoe˜‰6€y:5ȃÐO™À3¡F?Ãc%kV4Í÷âNi6„–~ÿ©¾D+ðx‘ ãÕZYûo“§v×LÚ蟭­¢›Ò¥gä»<&оފ8 †BÐ&Nؾûãk ¥¥y2¸5οl2 ¬²ü;¼Y¼Õü*ÞnÇÊᘠ·ÈÖqÑ€]oÙ_\V"b?ççK§žˆæ³’sÐÃÇR§%‹zc*|eÄãÅRùÖT93•ZþRâ¿Pø›Uu T#¶+,[ôѾѿ·,Fäa»¢d—Ã$™Ð<,ë×n× .¬GãIמÏ÷^˜pŽ–´îfû dâ´íÎuÍŠÜ&þ é·’ÄÞ%6á}I$ÚâRKçúϲqã}á/5öKÙka'¼3öÕò¼Å!pÿÊìð|rFÊR0µ‘—ŽÀfÔ“g¹Íí#s]Õ~\1Ö¿Ó–¶F`ý!Žf!°¾‚—ÒÃjç`ÓPìbôÕÒ€:Î-Ótÿ ¸KSOp¨§SÂ1µ†×ïeØ‚þ€ £K‰™ã ɈUŸ:ûåãFcof|ݘ޷úYçIŒÊ< U«5Ø€†ù¬1ͦ7ÆüS€0¦òLÀ°ºhƒZ¶,A ÝKz<à¼õ“ðÓ€¨á®;ñ¼·Å²í³êOd@p¸~19ÈÀ5¡L±Ävû 븞PhÄëb: ¼¡â²X* ¬c%g<†óñž{o¾P–;bë¶ñãÙÛºÛÛĹr9¡ÇXê»ìCdýñœo“ŸX{5‡ ¦Þ­¢£ÓXD"Ëñ¯8c¼n›Üq\Jj-¨ÇZðù%±8{—iîÏüöû¢ºøášçù…?qu§ÿ¦³ðõPwìf¼XœóÿZÛ…iËÞ`gGsÿñdüƼ¿s_÷*¼¬/ÅÛ‡U°sLØ÷„±„hç>¬ŸØ¿úÔ(ÃMºc‡®‡eÉVq( _r$¼ëcdæl¤±+Æ-1'} вìòüTÖmÛ8f‘ë8^"öoîgð{÷ ~ýÓëi“A\8C=üŸß@ßIÐÉœÿÒo%糪oî9,»>k9Ä7¤OT?»£‡q¿˜†ã«§/çÓOæÓ7ž¾½÷é¨{ôÅsÔµA}ÅÕ°»p3Ÿ¾šO_̧ã|ún>ÝΧ7óéOóéëwo´öãùôùìòOçÓçÃýf>=žOOL½ïw¢yóüqÀ½ÀrÜq¹Ëè N&7r“ø†Yåƒ@‡XÌ™ñÐøÐ÷¿=øÿxƒÕendstream endobj 380 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 4875 >> stream xœ­˜ tS×µ†¯—!NÀPHîu%$…@BJÌTðˆ-ò K¶¬Éšµ5žmÙ’å0³ 5˜¡1ZRHbš´M“×öµz¤\¿µÞ‘Mšæu¥}]Z^²e霽¿ýï1qÁb±&­ßºëµåË"¿®Íe…™z”Á$}- 'F=J<øÊÚG‰×·•=º¡† ÓØ0mbë#>õ¦õ=ýÀ»Ñ™Ñ̃5휾lúà W óñÌÝ3?æþjÖ£³vÍN é£Ã¢ýÁЫõ¬ËáØhŠâ&¾œ¨Ë0(ŒÆR0’zS¹‡òÃQK‚Цo…sÓ·U@G*­1¥›ñë.›×Uóš1‰9fu”¨Pž¨ƒXî Õ ½–è„.}Gd±Ý µ*$èæ}F‹xЩ5 2›as™;QL':tYá -ñ±PúÚ:ÄF¥h=׿³7Ø|çÁTs´²©ÿ—·€T¼õ03UšBËøúr(&·¤Ÿù}Zds«@-C™ŽR ‹vYl¬ªô»Î&º~ M„Aòƒ]¾ý«ĉe”ú,¿5ÒAT$ɧȶƒTºÁVa2ùlTí[ugƒ@ W!Ñ‹uz-3`е ç)œeõÞGƒŠ9ïíxÂ[Èû;dGAîgÛo>Ç<Äp™yÌ‚ù×¶~ˆB\4=A1v¦ž›z$øÎ¹b·]¤»‡Î´òlËáí‰ÂùYqt‡ö>, †—YèÉØáóá5\“Õd+i6:•-(Ê©××%ÞÜQ½xL2“Å9óü_ç# % Ê<ãV‚V£•žÞˬa¦oyÈ•Ï^Bó®úÑò;h=üÉ'§Þò«®§žË×fžŽÛc1ÕX#'böÑâ±S…âÞdÿ–™ÉUe«ô©F²gXmÓ:èfŒG ´A‡>8–aŒ‡W Ò½`wÛk5Ý…8JýÉ€‡÷3-D‹þw6nb:?—–òü€¨k@¢Î£/qWo¿d¿{gV’€™š»…–$nÜ äå]§Åd1;éè°‚aV~õC6Zû.WÅר6)Q`&]PM£+´Å#>¢ÇWÿ…áÒÌ ÎØPl†R=tRúö,W§ÃÔä8ïß>ÄÌÞžËÄ­x‚Þ¿sCÒ @ÎçdƒÅa·¸Mv::”ømœ1Ì•‡~Ìl̤KeÖ+Ò3ò$”üh~ç 63…YÌ,Xqjã/ûºÍ-t*³\‘hÔm1òdxåv½ƒnƒNK;´ãºiûnTfþN²âTl{aU—[ÏÑäÚã´èÔ¹>pB£±6Ï-qç‚ôFAƒ1ZkýJµw°Â›»Ùá—‘œk2[\`"=rÐã‚/*§v0TÆfæY¼ÇèûŽwµD}ÄvHjc™*—ƒAzRî»Çb®µRo Áøay^9èd /ÖPâ báj Ws†ÑÙ‰#Mç Ÿ³A`é i£ÖXùX+MÖZ3Õˆ6Z(Þh:§T %2Ô‚ªÜj]Í<<Ú>žIÞß2*äÔƒÉä¨"£C« èæí* ê¢6ò >œ€~Äý5Ø ê¥ }hI™K¨ V»°Ù“ù“ô}B ¥H­Ëü7 ¿ÕqÔiÆe²XáûÙ´tøœoÆ­»êKhÆÏ¾;;ækh÷¤²_ »É,Á ó…;lç%ÔyÇ\ÔÉ*Š‹òJmL=Ùƒ¢O}´¦6ÙEÅ|•gOwžx¸©aøNs·2ÕCeéTP’ {iSMÀ{âTò±×˜ÉÉO%P1_ÃK7¤-‘ìV ƒH ýèžN¼ŠÒ¸öF³û$¨uêlåt6ðqá%1‚”J"…ªh'§jŒÆ½ÓRÐA•'|Æ ÈÉKLÝ%OrñªŸ#Þ`zö}Äþgr1„tãÚÄûgÚôã/Có>y÷T÷QÚUèK÷€Ìv‹‡ŒFw±|ò‚h}=–‰+lTšÏ½Ë©ñ`òËAB¿ÀéF ­•`¯¡Ô¦,PceOežZÁ)ýˆ;œõL¯Q§“€”·¯=åŸÏßër0–¨¨’)kw)µª-¸rÛïiöÓChù;äÍåší&XHªJ”£ÎU¨W˜~m ÊeöCèE®:ÏX RR‘ýüžð‹Š‹_? u¶zÌ}Ì#à;ù¶¼~} na­ç«Žô .Ã8Þ×úvëð1´Èï‹1¤dÍ+M&d%ÿÈ}’¾ã.S“ç$}½h÷¹{q DðÉU 0>ãÆ`n½Á77ôòÐì˜÷Ð_.s=û»‹;pþ‡‘ÿ>ú³Ô|­‘Šé?!ΪΚ{K¡cÞë‚“ ½gÉ‹ÈÁ=ÈìPem]½ xÿÐ%È“þ„$ƒQkÐ`¶J˜¤ª;§|¡ùõ ߌ?Ýͪº4;æ+pÓ˜Uëïò€´ÇfÂbk¢ƒ6o3T“ O‰÷ÿ[7XOA°©¢‚«Ôs»Ün%Ýe‡6…àîÓ`¨ ¤Ä#ilnòõœMìÞòûZxùµÑÌÿ»Ê¦ò¨tF£RE­[òŒ®ÈÝýƒÍ_ÎÐ]7Îuöcbû UÉdô×3Ç}ÄÏ¿~‘ýõôPˆëls:¯Ùà3±¥43“#=(J˜ß^4ø PˆµŽ÷=v‚™9ê•Æ§Éx*…±d¬Ë·B‡%ïžwÂîØ«6ÉéfCªÀvk5â…Zæ˜8C£Ÿ›KA]¼ °9ìžÐä0 ¿Ì<6zíþ®è Ý Ç°çî>}wäSÏ­ÑpºLà°VÑ‘†uÝãC^Ÿç:ëjh…ÑqîåÁ3·ÀAPƒ”k¨Ì2­J@éäÕ¥Y@®K\"*wœ^“Ùk£¾£þn¨‡Z•OÙ^†{ ”•–+$æyfåœø]Öᮥta—`2;¬TeCsøÀ©¨-¨U‰q_ýÃE4«Éí‡Ü¨—éè’}ø‰ ‰CäÊ©[E hgƒ¿¥…üÆ`£#ÇYã…Ä] /ÿ›ÑÂJ¤Wzj¡JnTAÌS¸¡òŒY­Ôëe`ÄîFV4'áÝW/]'ÎöÒ?¼«ìMôö¦É{±*ð¡0¬œ›HÛÃF*$á¶Â›–3u·Ûà8ÈSGŽm™Ç<¾€Ù¸ü̊ϾÔÍŠ·õ¸ûÍ´ÝÇ=žÿÜâ²·QÀŒY™t4ºÌ$ùêî„ÄA´äW >VÁºŠ½P+n¤pjÜ÷¨]éü…¥¶ê-g·g  ï·ÁP!ÉÑJ AD k•Õ~}ç‰äÆäEÌ,¾š*f&D½ÁY„ƒYE_å¼2:Y¾U™¬)æ©ø²„tP‚Ø¢¨jµU6A=é+qK D‡NwчŽtá±hÌyöÐ5ìÈg}ÌFo¢›\ô4Í@sÿ‚&lºÎ°hfàû*fîhïwÙÆT[ŽA/fûØwÙ¶‚ÙZIãré纺Ø;“¿æ¼u>‹¨lQÚa:#uŸ8Ûœˆ[OòyßûÔŠöaM ÏŽùýn€[¼R¾OSœÇç¥LÁýJÙ ðÔ;½ PCV”×å—dKSSºKF¾¼õó[ub…6[«°$[Ç$Y^¤ÊÕQ1Ÿ• Zç2™IÌBæ±§7aIþò<œèózɨ’{ˆÙ¬ÊÚ¶KònÃ-Ï™V4¹á4í¿x²W\+Uôßq~÷R-~/ý ˜3 G?2°¹6𮲺†Š…²Vµ¡YŽ–ñÚÕe2×¶Õ·¸öÖb€~5##'ç9GèÙð¡ AV¨±:~ËÍŠ¸…V§ÙmPäõ?ÂnéîÝtNî¾"<–-à 8)ž‰»†?‚wÉÛ/ßd¦SLÍÆ!r½pìÿÆáªŽBçøÉz©¾\O+ê¤ &× e]iýÊé0Øe2£V®§Jãs²²ÌÑ7·,M–mi´„¸”q|o-v§ã׈÷þ 4óÒé7’>˜B#èqî}QÊ<…£¤’ššíÝôuô²û(ÎZï4ÿ‘%|f‘öÛFçjü~¬Ñi‹ ¹j1ó¼,SŽ*š³7ÔÕ:`rWFF@è©#¢O>dX2#Kq<År¼ósæò3}etÔ,6i«‡#eòà’ ~{ƒ¸míùûKij`ö6=wbð&§`o™@k(§73ÿp±×‹ž_Ãû‡Eÿ¾pŽì¨?¸qßáÃ9”êfBÝ!ˆƒý™üŸþn!Ñeø å}5E!ÃB{ü<;æÏ¨à×'iÈËÊÌÉ̯- tÛ[©˜¡÷ѯ¸’ú<~vVvQµ¨¥#ØÄŸFq—X¿Fv6úŠãÖa -^³ÝîÃÂ2Æ5“€¹6€NfÐhÅ‘k'Ôb9 ë?Þóù§FГ#hÁ;ô6v!ïóJ<οĩÑ*jGÜ'=ç{1±¨.)ÒŠÅP0æBšu]'7&&%åSÅÉêtlGHFÊiG\ÕÎjGe[ë9l8êëÊúÜ2È[ÖÜèk;PÏ<’µè)%š½K©òÞ4ñ.Sd¯ÊtRž¦Pš·.«ýŽw9xRÜy÷šžn¥†Ð;QHÊÙÄ|.]WºEQÈÓ*Ó·íÅUd-«l´ÕÕl H˜µwPÒssðrŸòŸ©pÑB_h½š«}&É3)8ex*5eâÊúi“}®iÓâȪ€Vendstream endobj 381 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 7169 >> stream xœZ\S×÷xy*.ÒW@í{Të¨Zë®Z«âjA±Ü ¢"{…@Â{—öNØ Jq/œ8‚Ú\¿Új‡¥Õ¶ÖÑjϳ—þÿ$þê |àæÝÜ{î9ßó=ßs£ˆ2ïC‰D"ñB'gçÉ“ôΆ‰„á}„·Ì’q˜°÷¥£Å[Ô óߢ–-ykQ)ýÎ@di†,͆:`%xÓƒÀk0e&„$,  “zyxÊìÆn×nò¬YL°›2iÒ,;{?w©×v7;'7™§»Ÿ›Œ |íVl÷r—…Ùã)“Î~ÿ}¹\>ÑÍ/xb€Ôcî»ìä^2O»UîÁîÒP÷vKüevËÝüÜí ÖN4üZà"s—Ú9ìp—úSålï¿~AÀ†…‹‚K—,û$Ä!ÔQî¶T±mYØv§ðËÝ?ݹÂc¥ç*¯ÕÞÎ>k|×ú­›6}Æ3GÌšýΨ9™;vÞ»ã6Ÿ°å=—‰®ïOš9ejE >¥fQ#©Ôlêj%5ŠZE¦VSc(gj,µ†z—ZK£ÖQã©õÔjµZH½Gm¤Q©MÔbê}j 5‰ú˜šL}BM¡¨©”#5ZJM§–Q3('êj95“b©7)kʆ²¥Ì¨¡Ô0Ê‚¢©·(Žb(žêKÙQý¨þÔÊ’@Í¥Ró¨AÔ|j05„²¢¶Pê Ê•ZCBJ™SaÔ Q„èyŸúTš 2K4»c¾Ðü¬Å‹z2}X,1ý˜ÕÌþ¾f}‹úé×ÔFÿêþ/,c,o ø×ÀeOzsPØ o8ÄzÈQ«‰V·$S$÷ÞXñÆv ›ÍÞ|sÆ›ßY'X?¶ °¥m7Û¶}wèá¡O‡5 _4¹µïÔ™¢€O9,ß!îšW«Ÿ÷žé¼HúéúÃs8É—Ÿ /…‚éÚ[PjEû38_±`2‹ ºXÿž ɤ#¥ÈÏ·Uò‹ðè¹0ÚB]‰v7H ‚È£TÌWÒ`‰C°-.µ¡#‚ÉÞÕäÕ? Ë#ÿÖ>Æk-d!¯>Òï ëö@¨\´V?é R·‰8A€iÂq†ÐßU8yoóqÇ¿CÚöJőӨ¹¹ü6ãp(­ÔÛPJøÚN%-yhâ¢T,.uÜÑmº:»œwÀ,–@(pôw \6îX1''jáK­è¼ªuf‚ø’'>ÆæØ|âX<K~™} Ï/¿‚‡ÝðtÖÞéöß®^½~ãŠã¤‰Nóíù^‚‚,°‚˜¬¡‹«»NÑã¼ TÅÀ1ÊS³af¬Æ¶ªÂè@ýã*b¼††±ø¢^*÷¾ý+Ia©ÞM1Z!G+: ƒk—Ì…p‡-AÕ~aI± I\J2J‰’z×mQ¯'cɆbG¼Èox˜ï;~ÜŒbâdÈ?…Oxwvð6ÄLšþÆÁ»÷_t\:·evŸžÑ€˜&¤QëÏ…¥Zx_ ׺‘©"A¹ ³¡”½xáÌå[L¿ØqñGž?œàðhóƒn‡ã¢G¿€MùííM.a¾^¼äQ}J‘û+8î¬ûŸpL÷zóžÊtf—`»góʺwÛh°þì`:L›ã°-ÅÇYAù«,Z Üw%„ÍLNr«Ëu•$N΂[Ž*¼h×}}ûè\΀ã-¸v9ÌZwT·™8í¯GZÖÏh¹ÀÓ{²sù±Á‚P9·«Å’Nkñ6Ñ´U”s»Glã°5¸BÊZZBöêƒÃÝïNÌóâ:û3°Á”Gà=,¹o-9 RXÎþ|ñWÀ¥—Ê«S¡.©UGš¹~iØ2g^òå^ÃR¸‹wqÛl㊌©ŠÍºwñÖ‹žé`89£̃:†Ó¯µ}8xCÆklbôÈžN¥·Id'½ŽîIQ«ZˆÖúj­%Oáä°²Ly†4Ë?óã´Ÿ©iÜýË7E~A\ZxzøînëÄúìNV&¤p®W¤ 1þP6ËãýçᔼäaKü×Qk‡îÜé=ióêü;R.©,9?1úˆðãÄHÚ[’Ö\ÀµxœI©#ÐžÒøÍ ^òýYOÓ•Ž0U G}ê †Àj ¸³UzzìrV£ÑÑØ|ãæµˆ™=(ï}ýû¦Ë2L•òéáªèšn“ï²94)Æ+’ólØ^ìFl¶|ÃxÙØúŸ?ÔP©æCÜàiƒ{®iW‹ôAŸB‚!hf}ˆgÞ£{‘6kš¡4è9ø¼‹Íãæ­»‡;ÊÆÓ;54œKuZ+CÌ\tºš7áeÑ}x’:‚«6YÛäŒÕØ{‚Ë«O{kAu§Rj½Jm¡à‰C¥ôë‰Ï€g=²{ê£þjdÓ|4¨ 17é’ê.‡+y¸Gwaë•¢éEJãäÙÑ>sОF*Åÿœ‘K×ßä$Šƒè`^só ¥ÃI}¢Lr`Ÿ1é€êTbÊd¼Ïd¥·<ù4ÜÖŠª„@3a‰Kƒ‰sú\Gc¾Epۢk AŒx©‘Ù–£J”¿Ç@a—Vt½–ØŒ'–6ô¸ƒà`4ì^çDzƬ$Eü&TãÏefef¡¦J¡–ÊB"}–}¾ùIÇݳ?éÁ”¬{˜iE%BŒ™°Eo²†¬ŒÂ¿ÒaÁ]R œo¡7àS©trMÔ±èƒa—wå{¢X’`1!»¶c.YΤÒëà”E‹AÉ¢÷íÔÀ_™²ŒÈÈöÊÈ-ÓVÿV”QÁ¤ÑYëšqßÌø‚ð2T€šQFIfÓÿÉÚ'Z«sº­àD0*ôid+|O¤îCL]wAu‹ÍʈçÁW|}Í·ø­£Ð?MDUq]ns‡rÒÒÔéEi…(1ÿ:í»ˆ—Ç‹wáAÆ]rÙñ' mFmÙÇw31„2?‘º#Ì#ÜÅ#ÆiÛÞý-ß7èLc½ïh5ƒzø‹… àlñ™ÞIƒSõ>7(/ìË- ¼B\ÜS߯¢+]Ŭwô„î¶ôCè v8âWÚÜ6<'µ›OÉLV¥¦3=dOtª—) DAX ‹j{ð×·G|Ôòö£ç=µ¨é‘¸¯ÑÈ4ôÃcñ`<Ýb»ñUèk@a(ò忇AxE@r#Ïz÷ØÞœ-Xôª„H¢_mXpüg®uªh9 %Ô¥®æ$-n‰ff"xIä&äæåËà æ7ÅÿÐÆÌé ‘$çùV“3Ï0fLs§r§q$Xö]’A„ 塤è&ŠõÚ@¡öjE‡ïÿº{EíL KðµM¥“Êã4 å!Á¦úõ麛†øef,JA©¶(E:+Jêïì‘™ŠÒ²rP!SV&  rÛ»½åÉQ°ÉÏäzµèMw˜ ;AÄÖù£ÄˆØy—¥Øl˜ Û:Qb¦Ùâø4Ïrebª4¥õ7G"<ÿS<|*¶üyX¿÷·Òn©ƒ­¨½ª f l´Gœ¢?ÉÑØãS¡ E[%EJĬ6œWƒùN¥Ø'´û0÷zÀàà Aiâü-W°sfTެÐ=+A_P€ QQNNQ†ê;PœaÒ:ƒè×¾Y)nÔtcp5 Éʸ #[tp°«„¾d5ê´ TÊì A ?¤½ãã½ù-â^Qéb|·„ÊSw¡p[¯ TÎ ƒè R£bæ‚QV¹ã q{\ëèg=ýë$ë”8Óò!¨^S>¦Ð’–zïm%[‡aÑø.}øã¸§­èPm]=>æ‹Ä«]fÎ]ôéõŽ_Û¯¶_iYílÄéÃV%>_%Ü`£÷Df"æc“‚©/‰½£&5OH£aYç®BÏ’ÔdÛP±TÕâÞo›Ö€Ôn•4ü¼¡ÓAqäXÍjÊQ3ÐKÃoâþãÇ5<ðéh`ÁæøÕzh…ÅïR$ð!ë×Ä ùhþɈKlÑÍ‹çÚ鱤¹ï0½=G«jɪžp…C¡uØ ;…âÃø&œ`YE-âð6œÃNERMôÉèSè>:‡Úеüc…wÜC·QS\±sÁf²Órä„Ŭ‰Y:õ"†¼^¦{tß Tz 3Ð$µ§ï|¸äÜMêÊ’ºK¢ˆçn™Æ­cÔ3OqÛh“Z\e¬Å£»eë2-rIX*…pÇlHÑ·—ÀÜüñáCû±˜Ç2=Ý©4ÑìÀó£¥ù ú×ÃK֮ݺd<‰…žEÂ.ýùs—`Ë%«ú ½NÙaå kÉŸ1 „Eì·ãgÎ9rúì¡ÍkW¹º®ã'{²)ûd-ˆypóFGC\}x¯)¨Ê,LSV&ä!¦´² ¾R±{MÌÆT¯-|D¾gùļ·hþŒ­¥ÞÅ ^‚©˜ˆ¸¿¡È³Hšmï‚>F1ÌôçËá xãyÛwÍa'××rQÙn“³˜ŽGQ(>=%¢è¢˜¢T5JG™éE™y ôÇ—Ù1ó)ËmlTóÕ%¹è&b:ÀMw¿uLW÷Äkþ4ÅÀ‡˜g‹KQ¹JþYL_¶7¡ðHYYSî]ÕYô:Æü …(-ŸSE“žœJˆ¬}V²>ôd$¯‹=w ]CgQ+ú2ïDá¹ õÔêÂósדc­AØ]® ‰Ütß&®85MŽ˜Z†Ii»2’º)®+½E]úÐL˜ÇXý=Ç'ú;–¿¦)ÿ—«”ÿÇEŠ¡ßÐÝYqì sÓ© µDÀ$R“nÐŽ¬Ø´Ñeå$~! Î>ìçÍåZtù~Þ=ÜËÿºØ›–œ¼C´{j ¨E!È 9"¿rŸR¿ºˆ½ˆ¹\õù¾FÒÓ½î—Êǧì @ÇÎ\F9q•\t½L£ÌïÍ­»O¬ÀýÉbßCh†²†žo„I¶Xà6<Àð'Œ€AKž )æu!!!!uuu\7X¾ %WݯêoZ¶.ñ9šÆ¬ÛºbNÀÆÌÚíœ[mò¡Ô²ÔÒÔ2Ÿîö ÛtõWª¢Æ\6â 鯘?t\ñ8]Èo;˜ýQ#y:»Ì­À½ ·=Õrâ:ô+~ÏWÅe„ä'ª{ú%C—îÏmkZSàB€e6ÎŒOŽ9|Lz¼ÌÔG²¢]Lo'Ž8ÑÖ]®x.˜÷ºb ÝK#ðfç{4 GîAžî¾Dú£èôÈtR’ÒQ6S'+ •…„ûÏû „QŸåNòPøýµÞ1Ô€ØêÀ¦§•°yM/<ÌPãkܰíT®3¹}°ÿ¦Ø@™ý eöiÕ Àآɤ™h"†h<O°€ç† e§¥S¹Ñ¤ó2é½7Þ¥¿¾Eèá}Ã]f—kú—÷¥{ûÛ:ˆÑ—îÿè?X6Å?ÓvëÖ­u<ý'?7è5ñëf=¦ñ ûï?yrÿ âþ³ûâÕXu¥î›¤cMÆ÷LŠÆ›z ã5zuW&Dìëêú}G¶Õ &~.¾µV1/a«^Ù5%—Æ5lmšþ1ó3·ÎÒ<÷TlkQ9ÊmîM0g­Õ™Vè¯ÿÖß’‰„QlJƒ…!ÆÇøvÜ&öí¦“ ž E<|,–´t<~°²(Ü‘ÃJ±¢‡sn¿2éÿœ’DÄ-ÐSü¶Eíà"÷”y#Fr×ù)ƒC˜._À>,ÕŠò…T3!€¸#¿G9Ï1a$Ú¿ØÅÆE-g¼svs…)'¾üĸ¯PwŒ£à9½'!Ç‹  ƒQRj”œÁv2,Há™~&îÂnë·Àia:!°Žµ:H%”ð'4¶¬'rMØÂÀ*H0J.\A{'Äùð¸\†de™Yy9\mãÁ² ÄÉÝ;ºÀñÃEŸ~°4ç\Ÿ••“ÔL½¬$4Ò/nÇÌoƒ9 yü#˜q’?ŸÎýÔÃíQsÌ"Èù ”ëºî¼ÃOl¼ýó½Ð^T“•D¾Q\&¯©V—6´xœ‡‡’° 'yŽcÁ|Ácýö-X=µ÷á$Â:´.ÜeCZȃ¬#ÚÑ"צ¨RU)*¡ÔLD¾Òr:i925;¥”0ZÛ¥“ÿº{æûé«v™¥šµ¢#ºó$¥Œ`{¼Ì "܇Q"ï¦ÒÒüz. šá=X ½¨.KOGEŒÞCþ†N, Ó÷Ë^(DÁ»u³Yyw¥€t-lÑZÑ9ëï`õ÷8 :ÙNG[O·žøêÎùMëV9¹.à¿òe¿®=Ú‚Î3÷fÜÁý±åsf¯:¶á« NòûüÐ¥+æ÷p* «?<ºíxÑ^ÃÍÁìL‡VöÂ埜_¶|‰Ãê™zawá¸èßTÔ%ê—{¿˜ý›ºóoJ¸Âf÷ʉB ä“áZˆ]mÒ"3BëP5Ê*K+Õ}ëÿ>Ãp˜YmÐL"§[×QwÈQ„w †Åƒ?;ž$ŠGZhIfV–Št1Uaê ©<Üϧl‘f6„¥‡ÀF:HÄfº ØlôLKú‡)@B÷è4:]zô ƒßÉdg¯ýâÖýë­7×Éä%|wÁz"ÞRRSPJÊŒ@qŒóšU3»8’,èCD°†˜ÁÁœmö«ñq—úúîljÜW½{7‡/šÿã5òΙš‹Ú[ß‚ó·m$V®,íøNûøÖÒ+Ö’¿5p êÙÏ£» õh½Ëæ¥1Ìrˆ`q¤ØEµ­"ò”*3/U2uÁ¥ò8ä-3/Ì&†Û©¯ê»Pu ·O ËO}ÚQKÀc<ª–‘üר!Þ“|RY²>6>)E1¤”Öjª š¿t¾ûåqÍ]µÜI#8w1îµt²GÁ*LÁ_¦Ý|UÞÚÏÐ)tâĶ<}f¶Áˆ6ØÕ#ÛDM­m­/È™°ðå63/=‹»,²Tœ ˆNâpÍß‹“cS’P‚­¢$¢¼*»Œ4ÜðÝì­7lعÂÞ¾yÅÉ“ÍoqӱъÚ#GÊk Ê]]Ñ<^#ÜaïìhÝ´Ñ}åüö¯>ѲïÂmÃ̺#Ë Ëü¶¹*¢õkÕ®jQ«0IÿaP›Ÿ]˜ƒÊ™JEADÜ®Ô8%‡ïáíq‰QñHa‹¢rbòà{ìnS—‘š‹˜’ò‚Êròj6€ƒ…ÁëÚ﵋ö Îf‚'Iä·bœ‹û›Ü´;³àöê*ÄÔd©sºÄÿ“o5ð‡Öª­ãØ“É_¿è"ñÈ¢ã%瞪/i®G§™fé}Ò¼½ÑaµOU¤FSYZW”—¤âªëŽgïG̳ý>¾A!ÑŠXúMAÛ™(ê@Cžj/wyxû:u×}ŠÏˆ6yYý·üèyríÍù­Áˆ‘&Êã#ó²vñ0‹Šñt¢ÜÇìÄ<6)’ E2ñ*yu=l%Œ«_íGM¨°?¼ ƒ._W[]l°üskÉÁQP±Z¯êv·Úˆt”ƒuaYqitnJ/-ýxOPPm ² ³ Ê2Qýv^˜^“ö9ª'w·7òŽóOsHâ0uY™›‚­<*,R‘ŸXH‚$Z©Ñ Q‰Hn›Œüp0øFJPªò·CÞ{ö =¹õ©Ÿ§Ôt±…öåH¢Å¤$jŸ¾À £ƒ¡0þ½T쫊Ý˃‰›ðr¤^…ÝÕÀGí0—¸KÞN¤T±0‚.h"õÔCvŒ¤ P“§'òŒæðH; }´;ö»n—ùù¸×K÷¨²ÓÒ³M¢$;÷òínQé`"‚'ü=O0;îõÙTÕnH‚YäFû¯Ó½ö’I‚¦¥CxÊFg'¢]]¼—„:ØDÆÇ'£p&!ϧ*S•–]À½h{±¼(.©mU^cvi\ft6o õBoû'Ÿ“ z ™ð?Ú1SóR¦]&˜ ./±Ù%iåt±bHBdT"©×/IˆHE(Þ6¬8\]‘]\¨êmêD'u°Ÿ„Ú_á³|Xã>Ø|L×gÊËŸakèæO`(1ê‰sØ¡Œ…ѺǷn¶¿‹Ç౎¸ßGÝ—åµÏHà“ˆ»¯tb­HŒ&¬ì>c]Œ"%"540Ò²¤âòƒª}G¸ÎÙ“ÄF—iÓ$º÷P½Æö¢æïybS” ®Æäôm¨×™}q‹OLBJF™ÎÁC8kº/ Ž£/žúõöá—[]A_3`9ú6ˆûÍ™5cg]L©/ ã²¹âcÇwŸC̽/·L]¼q¥ƒ3ß­l.éÿ둎O; P·]¯l3¸Ï¶~Vwen-ø‹±åâMK¼«•jýbE‰y»2¸Ý»¯ë?Œ¸vÒÃÅ/Ò30„÷ ’¦x¤¬JŽDÈŸ‘ü®wŒ~|ê“•ËW/›´}nþñ "-³UDZö|v3åÑJ¢(¬¿{ð¨t}óE‡}m¢ÁÒñªÕ}øÝZòà ü‹-‰,‰ˆILJLàü‚"”DÊÇçFDÊ5^(DÈÂå‘þ!DiFG”䫲TÙ\C]MIÊC9q…Ñ…Qêн¨Õ•T–ª‹ë+P30&_X˜ óòi–-ÖöÓõçú™‡XöÕZZê,PÔZ_Ã:endstream endobj 382 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 3954 >> stream xœ¥W T×¶­¶¥«Tl•NEÑX…³Ä ‡Ä˜8+hpBie2ƒ"2K÷i@A™q¡e†'Dãøcô©¬˜—¥â‹æ™§11§ðòþû·1fX/oý¿Ö_wA5«/÷žÚgŸ}öQ0}û0 …‚]â¸Øyºéãûò…üNy¤È–WÆîÕf#™A‹ŽdV.¹´@Õ_ æJ0ï[÷Ž…‹~5÷ B¿ÁŒR¡ OZ´#:ÄÏÇ7Ìj¢§µÕô9sfO¶šac3ÇjQ€wˆŸ§{ •£{˜¯w€{ý㫵Až~ÞaÑVçú†…íøpÚ´ÈÈÈ©î¡SƒB|æ[O¶Šô óµròõ‰ðö²² ³Zéàmõ:Ú©¯K‚v„‡y‡X9yy‡2 3.0hÉŽà»Ð°ð#Ü=¢=½¼·ùøú­]÷IÀ{³?˜³`ªÍô3f³šˬaœ˜µÌf³žÙÀ¼Ëld3K˜ÍÌRf*cËØ1˘åÌÇŒ=3‹yqdV2o3ÃKf8cÅ b†0Œ†y‹OQdú2‘Ì ÅjÅÍ>¶}ª”ý”õJ컩ï!3…Y†J¥Ú«ú–gÿÅùr—ûõïßë€Á¼ä›/40pÅÀÿR»¨¿VÿsÇ ƒÅÁrªºÛ$|Þîc´@]n¨ªi¡dž\±Ó²G Ÿ¦ÕB\ÒÕŸ6#Ù¡7‹ŠûmW«‹#¤ôO·ä±?­¯›0iƒ“s° ¹‹ K&šiYMË•,ðÕòÍäzYcTTá[˜…o)å: äqø´çd6ùÀf±$ƒNÀpîƒç8H û‰O4€Ìƒ8SP-žl,nƒFhÜUâ[ìëÀ…›„!AÝýeœ$¿’×pŠœÚ¦ìÎêžÎBvt‚N—”(|¼ÚñôòÃËÀ’¤‘)Äš¸w¤OLÃUh†ƒ-J2ìŒM€z1‰ 'Wއy0»lãemÇžÛð.¤#sõóÏ>{–Fß-(!šô(ƒ®¸ CZ…H±$$e”ížõ/)Šûð­¡š{¸¿æ¡Ð`Èϯo8›YÜõ +‰e»Ñ~·Îo〠¹y}›|›ã€Ãwx„"ZÚü%`/±—M5†–´V¨…k`L50 —°š»7ÊQýLÐë£wA çQyìXUqý¹õÇ6 š[ðÜæÀ©»}è-¥–œSàÓ$))»‡uÏå{~Ð"ÓóWÂhUµéÏ¡ÚÐÜ{¥§Ù¿g»-%¬Öª%º½+l´Rá{P:ÿ éÇI*µ¼)Ù(¨²¨mtÿ÷Ð_C)m}Pâ·eÙç‚Ä]y …·k×í 1†”bàÊ`_©økâÖ^Dy Á|3= S õ,ÚQ´S¬÷­M¾»›Ót“oÇ®¾p„ ¾Õ´GHÉOÝÀ…Br˜Hl$ÎJƒƒÅ‚J¦§µúŸ…"à*q*àø¦Èߣ¢æ®×1ÿÌy8µ¼6¹ÿ%á‚"Šʯþ’ò…<Š'GX«á_Þié¸%œÓ:³ËüCƒÖÀµRáx‰ŒÇµZV?aóĉ©œKišlô9G‹þw%%Ûßäç±h›i&±ŽIP'jZdZ¸+~KàZvŽó†ÅóSËÏ øäßìf­î…ÿCÐDUÀMhí¤Q.Iî’zS¥ÄÆn *Iå‘j [è²tš-ªzºˆJË6¤­t} ¯“¦2ñxo±á2×e‹èµ±|2§âûÇ‘ËÊß ñÑz]lŠìàåLÕKœÐzðmäk RSˆêWw~aãÄsò?éùrãI’\Wá´„˜äS„ëõA©¡©:ú*\ˆIM:¨Ú5ÂÙ_xÂFÃNCtZ`ºžV*'KË{¾ÏŽÌKÉË"Ø0#«d2Ì@yMÕȘþèzò‹ÚFîBF"Ál8ÄA\÷:ã’¬ìÍø¯b#¿ßÄofqůµ(ÿôû쓬¦­nÓÆÌí#ÈøÑ3É(2ìÛÉ8‘fþ2œ*«,çðj_’ÎÆF{¬[@Q¼q ÎWԉǥœJè€zÿ’@šø2ª%'ðÃ×­8ñ;¥ìA;qI©L¾µé´¸Í¸á°=aÖ¬±dþì]|g4½,8);u)»’Å'{ºc”-ψ'ú¦Õ‹OTÔ×çìid¨Á—SãT¼_H¸Ã¨À~Ètá %ö“Wó µ·É ;­êßÛ d%2ð9‹}îd—ïÛº,!&.1 ¢8·š¨ŠãUÅ >-ó¬ ãNÌ(n’-Fœü÷/Ö`QõÄ•¸•ðÐöþPÍÏq¸ø™p½¥®?|Ë·n…å3E¢Žç[CÏèÊimŒmï¼!…œˆ9$?Tœ–kÐ¥è ™ )ˆ®¨((,+ÝY³%F›êë*Dç¹/¢ï;vÝ<["*Ú—»™ ßäíÃ!&cs½lêâpàœÛ¢m;/Ý8žxuyµ°öè*˜ÁTüR ®Yqµpr2fîç®õ<öSU¢8úx­'oóÍ­­Ÿ^…|±×𽿆âP+©Žò±ÿ-s^ã&àw¿g”oöºÀ.´±èDÒ™¡šÇ&oàÉbÓ®K¿¶ðJC[Z#‡3:cﱎ¬æyŽ|oö©vwóõus«ñmj4V7›äŒš±ÿn²¸Ö²õ)~ñt~ËPÍ-y¼<ƒ'ã*=³×VÃe˦£Wž}õܼS*d„‚îMcŽ•ªz|œ°ÕuM«}¶ -ÓÉK&¶?±¥8\lÛ\„ŠPNóø!ÕÝ’ˆƒ1å[Î%ØnÌ2¹­ýaØþ¦ÉÚ÷6Ùì´´¬ýBkëç›'7€%ι‚fÏEÍ-@å‡÷l i96PÀ;/*%L¢#‘Œ°îÙü?ßSEd3I•f9÷p,pw{:É­ê|:%Ù%ºèãüëŸ"wÞaq ä|”CT&kaBVî¢äQbX·=ÿ{?M­2œJk‚&hÕU¿qq¯œþ`ÊÝ_ÿ Õ´ýoÆüAÛ•ê’Èùÿ‰ÿ‡-yì 熱67m "OK÷†Yà´Õߣ6öW#ßöÆÈÇ5]Æwp¬ñ¶èÀáœÃí8šÖœœ[xdç>&7¤úkÆMÉÕUÓbÿ-š•gér÷¤êaoœè¾!Ù À†Âà쪬©À톸]"i¡r -c–P×z>‡Úâ.˜ù‘¨yµ&óP‡}î‚1íX4pç܆ÆlLXù~çû8XÐüŒ³¾„êÞ‹åL¢¸&}mRþù1Ÿ“ i´7U ûvgDdè³u9œVåK²h–sî8‘–›y h´óTëì'‘¾KÈ -m/@æ!ißçB+ §ÒOeIÙ׋/Õ4p½PÈ÷ê,NÝ_ n¯¿or.r(¿+;¨ª¹²ê¿9í±ls°Ÿk ÐÂß0V6ÃE"¢íÜI®‡:]M×ìm®N‹èôŒ3ò8ëÞCìƒÌlióQ,XÆ×DK;Ú?i9AÇžÙSv9Y;éì<×TS$ôÆPŒCÚŒø’‚ïXQ7³ GšÒ1;øp'¬ŽñÛéå¶ÕŽŽ>iEyé'ó¡…«;êåì½ìŒ•ηqú7˜-h^]Ì#ñdôÄé8$üc!Æ4?×€”U{ƒ#J~2Ü~ö=|õó¥Ò ñbÞvX~œCð¸Ù4–z@›ÅxW-®á=h3^úÒdhV¡‘‡/×–å:W9ä‰l#â*‡.'w•WU–4Hnt¡¢°!¯¸o ^Z×Ä]‘‘¢£}Ð|°ç¼i M)-ƒÌ2Aóò<±ã›|kÜÞÈOSMM“`*<´QÜC%ÆwþÉ÷-£ü7£¢…¶¾(ÓZ7ù§zL$É$YO¬'È.LƤ4ÜùL )Dâ}`mIì­„³ð#´S§ûÙþ³y]7rÊà$4DºlƒQ0Ö‚MÒÆÝSvÓ‹^B«?Yj.Rà ÄŸ»•| äÇé )A˜ï4"€‹Ô©A»áR!דOD-Ûœþ’1mðòÍcÅšN¢éìÀQµ’¢Ç¢ JZe!¼V•™°bwR\òBØE[Ó•Tw¶©åª¾¹&Áw²cþB-ÂÀ)ŽÓWŸŠ)«ª/9}ÖýPL†p¬¬6×ÜýÖe í¼—Ù.Éâ»ti-/b{ùT…SЬî&ެ5Qj ¥¥ˆޤÏ)4“ þ„?ñ_ßê€ûÜÝWè=æ­øÈ­rGmKyÕÉ‹ës÷„š£Íû+ûkúŠí±z2tL²è¡ÓéSôñz]> stream xœ“mLSWÇo)Ö«4-.é•{kæ|Y â–Ètĉ¨q ç[”Q¡£µ´¥¥ÐÞ¾ñÒBÛû´¼´TZZJåE(UGªv*Lq[öe~X2³,s~X²·ÌS¼~Ø…ÅÏ[Ηs’ç<çÿ~ÿÃÁÒÓ0‡³ª¨dÿ‰£Å;ó–»Rë9© i©.0Ç^|ºxxE&Ü¿/û°´(ç@˜—&>øéSV_CŽ5È DêLŒËáh(g‘ºžÒÊke:ñÖêmâ»wço¿—·[\¨”jåÕ•¸D¢“I•{¨W¨«åR%ÞZ Óéê÷ìØ¡×ës%ʆ\µ¶vï¶íb½\'—K¤Ú&iøZ¥—J”Rñ+½¹¯6Eje}£Nª—¨k¤Z†a+ëJ Ò:%†}„mÆöc'±CX VŠá¬c,›ãô¦½™ö„[ν‘îM9©Ÿ˜c DÆSYqNÊòœû ³Vä0Ó´p³TÍþö^r &;ÇaâÎÜâ¿á„ñ°ÑÝj1ÛJ΢½“ê¸c²QÁohZ‡2ò¯½_zZah Ó•¡8 Z¹U¯‘X+À[Ð?ìvx ßÔÀw1Àû¼6Áes˜È?M":Ý¾ë ³ý€Å,åËa‹š”¿uÔ¬¨/´\Å‹NH éòßæ,:‘Mäéîì|€‚ÖF 5ÍÄ1FXµ—É| ¯ ì`w‹bYIÞÝ`«ÒîÔ;(Òe§;À…Sà{a¸‡˜GM¯»:³—»8†‚:Hi —ð"_Ðs .²`£œN I;\VÖm1êöF<Ä:å ›ÀÙ/<#;;* AòÏS®`6¾|˜å°Ð4E7Ó&ºðFž‘Z.ê'Sã¼(tƒ7Œ³ýp«Óv;Qøîf*^¡šº7â鑽qQð–ÿÒøäXì³›àÙ¡zÅ&6m¥¤Ýdé¡yÎÓß¹èéâV‘»ÛÓÃ(BÑš¶gµÐzÕRîzûhº`D/C›ÑöìÃ÷ÏÍ'Fbƒâ¿é ^ø7&‹—ÿà¢/PBäŸöðë 5hÆ–*R´7´Vƒ=»n¸9Ú7Ó í_xö3d/Õ9éJ²Mël3^‘c8UæCR÷p˜ÇTúV&V?Ï V§Sjþªøy>ßÍ`Ø?¦M¬›endstream endobj 384 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 932 >> stream xœU‘L[UÇïmÙk·UæLjÚL^ߌ„1Å  0þuŽMñ¥¼.ôÇÚÂãQ¥­]é]¡-mY­%„9ŠÑ,Í̦ÆDÑlš Nþáj¢d1&þ±û‰¯l#˜{OÎ=9ç{ A™ @­zCw]méøòÏÈø 9ÚêÖÏž p ¥¹èWîV´eˆêr¤’#UÙÒû^<ˆcpðI@@höE2 ­Vg7tR‡UT]CÃóÕÔ³µµ Ô+fÆn2ÒJO;3픂!Ê`5š'Gnt:m/=ʲl mvÔXíg_®ª¦X“s:Í8û3@µ[-Nê$mf¨‡óÖáåB6–çõ9ˆ+yVÎkqA-ö➉¹©i''=n_h*áÖá^±O”ÌížšB>­;î‰Ï†Bss$î‘–Q4†‚Qiçfc‰š „uÁXëº?ŸCޤ.‡#—B33(ªLxcÞ‹Ýn²œ?%Òóƒy¸¶‚ +rœ¼¯þ&÷átýM}D_i_2¯¢ÛÊÏßýamauòôy.pÞÇ]Ï¥³±«_¾¶úRý«-'nô|Ë莳o@¨òAïÑ5óR'2(»¬ÍíŽÎøŠƒ|oz1š½b]œÇvò‹ÎŸ~ûêÞÏwèÛ®ëLõÛÇnýõãÍÏn&í«jê½_'1AEþŸä÷åüA¥Î¢,Ë¢‘qrów± A,›G)¿Nd²(—ãИN\&$ÝE\SÚð¶pD.´ðêÌ•G›ëÄ' s(­ãÿÜÔ¨·¥£¥Ì„‹•2Y”‘Zâ7Ô;MkJ©;+Â_x,*ñØ®¢ë»‹¾ßêÊorðž—¾óßûj±×úÒ~@Ú ò#··ºUsaR Jÿô;ÑH&9› #áT"1‹»7k¢®0Š¡šŽD“ü›ÍšhÍ 2îOx}~——Üæƒ‹øÖcDøk5.ìbQÀ׈4ÊKÌ8).‹ËŠ ñ²¸(±Ú‰ÿǰ@”ã%qzûxQŽë…þGð9ŽÜ:¦à—Ë¡l–,w,Mó„x(¡(î+î'÷•=}fBµ·˜V©${€ÿ…ÇÙÜendstream endobj 385 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 7150 >> stream xœ•y\SWûÿ1×kµÎ^ǽ(÷®³Žº÷Ö:6 $$2O!!0 nm«o[mÕØÖ*j«vØñj­ÝËžÛÿ¡ŸÏÿ$A“¾¯}ßß òArOÎyÆ÷ù>ßç„GtïFðx¼žKÖmY»uëÔ)ž?frCyܰnÜp~sþ±Z0œè»ø¥áÄÚõ‘×Vƒú€Þ|л{ð°ä\lx¾/ö#ø<^šX³D˜*NOˆ‹Ï =6dêœ9³&„L›2eNÈ¢äØô„訔uQ™ñ±ÉQ™ø¤-Âè„ØLqȘyñ™™©s'OÎÉÉ™•œ1I˜·`서„ÌøÍ±±éÙ±1!Ë…)™!룒cCÛ;éñ–“S³2cÓCÖ cbÓS‚Ø´(e±pç’Ô¥iËÒ—g¬È\™µ*{uNÔѾµâèu¹1ëc7ìß¿9aË­‰Û’^NÞþÂŒ™³fœ:÷ùQóæ/³pì¸Ýã'„Oœ1yÊTÉ´éù1’Ø@„‰¹ÄóÄ&b±™#¶£‰­Äb1–x™Gl'Æ;ˆÅÄb'±„˜Hì"–“ˆÝÄ2b2±œ˜B¬ ¦+‰UÄtb5ñ±†˜A¬%fëˆYÄzb6AƒˆÁDLð‰!Dwb(! Hb8Á,Ñ“!ž!zóˆÞijĢїx‰èGô'áÄ@â9"‚ØŒó‰ß&æ x™¼W»MìVÍÅóënéþ«@F#OõÛãj¨{=gõüâ™ðg¾èÝë«ÞózŸ{vȳçúLèÓÖwT_E߇ýv÷;Ó_2 ß€·ŽøÑs›Ÿ»Hï ß4`PÖà©A}ƒ¤A0XüéµCZ‡† ýrXа/‡‹‡¿Í˜õL9;`QÈ‚Ê;#†X<"z„jÄë#©‘á#KC{…šCãjúp5Àý«›{¶šÇMå>¥µ {, d.ÈùÐþΆ  R’›@ûˆ,o›3€=ñ«Î -¡‚lR’ SA9~ìh‡š2@‹†p=èòÃΓ%”>‘ù@¬—•!y &½‡Â«n7éŸëî¦#<'$¦4à-àDÒYHR6†<÷ZŽë‹AUðí=WlÈîg€Á`fʳ‚Yì9ÛgÚ1rº,ذӴÀêáôâ#öƺ3çN»uªD”R¤Ë×)YŸ-Ùn¸›³áÎà ýdÓrn§T°©’„ý>úýswô«ìí‡n€×©÷w]à v= û“w­rvîYÆŽ"w£ãô˜-€Ãɯͳ%k·®žâ9N醷ݼû°õ*Ÿ3B+ »/¸1õG½Q?Ä"êã•_Á¾°ìC$CáôÊ÷Íß?»ûÉû¯n]:qAèTŸÍœ|{n 8èBφ>6ÔÁ.EüØuiQ…¢`©u鬩_ÅCAPSÝ“•ÒÔ.—*ÈïÐÅ$´FBæ=Î3ìåˤgiÙ ×ð™r7Wâæ}ŽSv‡»MëMz 0PŽ‚òœ\U¡¼ˆQ)å¹ê¢„ƒ‘{…&¡h=ÚŽ˜ÛˆÄîŒüöuÛ•Î…d*ÙTÄŸ›¨™;îý çaW'ÀnÇÞÉ 5°2}Ö!Wiqƒ‰Á'¢t7œìCÊ^œšÅgèšôÖ”‹€‚ÏÂÞ°'»¿tsÊæð4Q[ÊcþЇéMyéOÅ …ItUÉ+$5E Ê²”7¼ùö±·uùÄö%ˆ˜:mf@¬“:øpj 8*¼àH Ç^Žƒà:¸ù÷]FýYÔ=ý ZÐ4!9Í@Ù ¤ß³ܼ›8ž6n0mt=0Qu"³$Q+(bd΄“Kp$ŸC}PO þí…ßß8{ÂQÉÆ£8ÁËþªƒ/úª2;I¢ùرu»ÎÞ~ôõ½¯nU½´O¥X©”²©k6n] (ܳÜ0Â̓à1"u*¨¨t—º¬É\Ved*~6ذ1†`«¦*1I+,bÐ ² F †ÔdÅ5•Q¥r´ÊjMŒý6݈W{+²³§?âõÞÅûıäa 7[ÏÊì ˜ò"Y+IW1òeTÕE —±i³ÜôÌË=pô¥ninu~o´×Üÿø³G@ݬ¶ª¡ÊÕpa>_¹V*e å¦ÚÇ ºù¿ÕuÀÏ?L†M4Jªs€(©”&`i18ØýŸ ¥^÷=æ%js£U J¶Ãhø¯î÷÷#°Éë~l>&û«*œ|B°¿æ»ùÜJn&[ó2à¤ÏŠš¿cÎý¼ïòØSEï€HJ ^Daá;K¯ËµKëHÉ@[¨xY›/[Ѝ\»¬²ÒX|¨šy5æ"páYxä»×Žæ4&W³‰5Bó|ûNëf+8HÕ6†ägú$‰žÑ©1§@+0Ø‹›êK^FÊVT™™©”io¹ÃénxÚÍ»‹ÚÖÒâtp@fÍwhØ7ÉÜ•;'/ÔŠÝ· —BŽÿîÍö½‹õlÈi³¸€ËÌèMp&N²—!”ŠB£ÓI34ꤖýÎìâ4 ­Ehôí1¿¾z¢ª¥Õô:½F§Ój@j;°‚*P\­7×ÀA%ç(Ì\žüçu1èU÷/Õ<ø&…#IÃy~| édù§øMÖæI4LÞdÁf>%c\“‹ó. žûŽð_®pÍ“ììeºš‘¾´T¨µà£Œí ˜r›ý”‹zù™´š<€ö Pœ§m þvCµ"ÒÛ/âÈý(ûï×<Ž’832®ÿcîø÷–˜hÉež$nOÌj@­Öº*,%¯šXs“©ÉÜd¿„ÿÊS߹齒ý 4µ*n>ÉSj€P7yÔÏÁ80Pà§"tÔÏ¿x½F‚€Z8æOÙóž]Ô.øï"Çéló….ÈXÔÓ/>jI4 }¸~(8è·À·À·gy=(ޝµ ¸Tg+åk[7ïËû|N-l÷è}ëj¿·É“äkÖ¥íÁ ÒkL¦ ¨¡êEÎÌ¡,bÕÕøÏ¿¿sùVóÄVtžÏ9=öÖ>%ѯ‘£Ñë‰í(SÜTÜÐYãAP$­Þ‰Aøu›¾ ” ƒCøGïPp¼­=L½õØx]«.Ö©Û¤Ï4¬+ÀZ Wü ‡8Ë›í¤­í”>— Çö¢åe}Á1à5ÀÞFù€2Ëý£›÷!FJ ¼@7H ÍŽB|0?Õ%u•3ejÌm P¿™6JÔÚš]¦AÖ€ÆQkÎmyøú%{Ãq¦2õê }úA*s»ùcZ™ É‹ËIËßW¨íqu‡ßƒ}Úa£¿oÂÂÿGã–¸UàòccºÏ³F`eQ ª: «~<„‘ßcò§>—¬Á°;ìù=2 rÒŸ¼wŸ É>i'\4ÎÒiÒù¸†åë-  Š¶ñ'AKíSàt˜üAãÑ Á^?!åKP*ÈaÁ¾×P_A|öcL÷ôçy–ÏÆ‘ap†À—ÜPÐ%aʸ žj5`²ý÷š›Ý©ÖäášS{Ú‹ÃΘ¿<¯£ûÿ“€µþ—<ÖwÕçLX= ¶Õæ}¤TØõ×믣2Y+Móuê"¯„Á\’›WyXµÞøöôo]C¢^¨0y¦”MÁÛ ø 5¬l ?EÐ(ÇÐuä,…tßÚýÚ\Ùb.VÊ¡¨NÉM‡ïkžøäò盘'ôÑ5hþ'ŸA]®(OÏR+óLfldÖ^@ÍÝ}éZÇEØ2ö ±±z^«T.ˆ?ál-=~uØV½€&Œs}*Žqð¹‡.ÆOj±œŒ¶V—™Ë(=9¿3EëçŒO#!›ý„“ *üÞ៤ÂYh^Pæž”MÒxJGjÎGµEQ• òµá1kXO蟾íN(ГÆ$ÛÚ’Ýö…Å ¦´âÆì&,7꣕ò㻌#h»ÍVꩺÇ ïÉP¨V&éŠXÛ¥4ƒâCÉE+ƒ%L'…Þ–æQpF p0çÚ`54Õ–œÆÍÀÓ5 DçZ[´E‘ª%| sI¥ÇOï:©hF^Õþæ Ï4½QŠº¸øÁ™Ó®Ö#¬N¥Óåq•ùK'A;•ô¦½gnAÞWÿüìfÕŠlµ*¹p7+³lÅcáêMIõ_ðˆÀÞ:gHGŽæÔE‡$ÆT¼ä?4ÞG=õ#.GúÇN•9Ý,; ‚-$Ž 0ìâÐa—àsnÞ½k°û ·Ð-ð6¡m) BÏ`„_΃8è{ ^jT¹lھݢ(@­[pä¶ì]×Ñqá& ¾³3d@«+bŸ UŸÝ„ï_çsÇ8 ýhŸ £îk“˜ØÐ¤+©Ð£bÈÀAÿ(ÿ¹i}pJÀu Õu¾LT´¸?<W:γõ_7U|¨/Zd‹V@Ë "ñEˆ’0Ó?Á šLvÓ… (y8ì*«2Y§¼ErP?Üûòa²FQÉ69jŒezsº«Ð(—«¢±BÖ¶93J•¼‡Ms%X5ê…i“ãœI5+Ê”ÀR[ 2™3Âe‹Aµì“5pæÚý3ì‹Æ™Â™؃ޓÜ~úò…KîW_·iÃz6}OÛpìp¥ãØÉz¶Éi×õàX¶vTô$öñˆ·1Lò7†§”'8ÿóÍF4R þ§‹/wvGàïW+:”p#àyúÎɶ«à©R8ð T_ñ”é_®?DJ™JÃÐJ…ûŽmµÆàá¦Ç„Æm¯Vg³M© Ê;¢wó‚^“ÕÄXjý†%ÓsPoðÃ`Æ3ï+$@RȨéhÐxg¹j“õhs2îŠüuÜûé/ ñÕ»ñíùelj£ÈZ/2'Y”†${F=8B]|ïÂm(п’²Ûê-¶ßÕ‰È͵ùÂRsâsûž˜¿Ð*aþ¬´¸è¸iQ± Ø¨ÀÆ6g”‰s“ò£—?Çá”Ó=aa®s‹|ò¥+RÓü0Ååý/ÃÀàÿ:-%3p¹*ãä+‚¼Ó†Æ·Ôè¨22x.=dpP0È ±-·‹üB2üÉì…t–ÐX$õœð£MòKŒƒ$ڀƿ ¿ L÷j?,N{Ï ¼[™x·b°kJ^{ðAÐC49ðÞ>ãGS*ùdæx„›ÅÑ™9’ÞfôtÖò%«AðÎ}§>(ƒ½ þÓ2}¬ÿ„%z;ÕâÏÊ<¯ÁYJ‡Ž÷±ÚX]¬.ìÓÆh÷ê¼ßšpA‰€¸ÝZ \eŒÙú””¡ôbä·êã ûžŸúüà W›Ÿ¢õ=†‰çK§ó ¾Ìn®g5çÓØXL^æÚ:1ÈÁB2 ©½ÒI䫤 ϧæ ň¬ Ïþ-¶M\­'7¡(ÝW$°Õ gàáô+X‡[ ÁÚeâ"ÓL§ŒDe>zPS«´Öi4Z­Ìkg.›[ÕѺmëø‰‹§Dítž–²6‹ ÏÅT¥´6e7Ø^0ýîŽ¯á Øÿ'øÜÃÕ·CS#ÔY»»®¶Ž\†sܼ‡À|òjø }üíÁ7À\:‡L©RL*Ó!n¬i({åÒ΃KÑ(´ ÿÔ£±·W~ {À~ß‘nyˆ¨ Û2’c™/Ð+´¤X³›%5òc ‹(ƒÉŽ·ž>n6Ÿ:vK³Ú¥ûC§.™ç•}.îˆw¼xVkFO›‡†LAÓrƒµzV…ÌÙbªs™cpç¯p9œÙ¬“Ñ`4•[WaL9Å8É@“§keEÑXU{س¼¼fèòÒà†ánÞõxSQ¼Kߪ¹y|H}¸ì ꆆŒ÷âÜïX´Ì¼ìé«ÁDjÊGËa7øüï°ûÃsqçç50‹Zèu›½u¡ãܧ·OEDl\²m.®›wèE[N¿{óÎÕ÷OlÚµøÅåÞÏ5à'.XŒ›ð^(x øð=ÎMë+0SùuÉ‘ê(¹ŽÉA»´R¤¶ÖN3aä_åsfÎeˆc{_ƒ |n#¬¡cа±³_ÀA8DåV«ÙŽõ‚KZ-ÌJ“ïYz5û  ç©ÐC—]‹çz2 [ðÖdHýãÍŠ#gVB¿u£vûöý;?ÔW§ÖÊtjV§R*Ͳ }âÙGvÀ°8²šw•Ëçs œ…«ÃÞÚÛãŒÝHž„a‚7ý/$‘­¿Š ` ®”Tä¤)³djf?zNöwæ²xßìì%2 … ¶ù_8H&†h‹´JPœU™_Qoq•z°1XðÓÓ,…?Ò?y·óöF¬TWÁþ7ˆ{ƒ~CÙ’U+Uæ %.Ël>cx½²cÏ…0cn6V£W¥jrÓÔŒV®é´ø[—¯E=r–¼æ ÇWABƒ¤A/ØÀ%¬"LzƒÑP^f*ǨnÈrfï›$ÓÞÝuWÁ¨„¸Ï©)±]^Ù`ª¶êSµÁf¬†ÝJ¿þ|>^qV9µ¡œÞbm§ÀÛï½r¹”r¢fú òb&X6nÞ¶TêÉÝìËpäe¨¹ C/óà€·ÞˉøÜá?úÒ%åÀ‚m®Êsв %²"¦3øÏ…R 'M°Ø‘[í,-µ[ø ¡ÝÉ'#·m_0cYû#ÏÜý˜F**Î:í,7%Ê·%î³Ó8޾œ|"jÓ®¹³—·%½xæÓO™£hEaUýÛm'eæÉˉábO±¼åâ"0Í6qy´­¸Ü*©qY®²P[”Ï N%’É4ˆËf5<‹bƒž%¨ÊŠòêÊK‘™EvßÇrp{ǧ¼[\ Ÿ3àÜ—¶aÆzúùdžB­žÖÿ1š@;=ðLò<éÙ뾋RÍû•›ã¹!#h{q¥ÔRÕÒ©B£UÊÔku-ìc(Á¥`.+(¤«âUŒõ›ðn¾U……^pVnŽXZ\TVÄþ°õÚ$SæÊAv°Â,ªµYpù0°ïÕ,Di¤¸;ª‚sË ¥µ–&3£7[Îê­Ö+Aß®Áf¬¿*ª+Šb…™}É›Ùî?Bqñdz4œ 7Wz$¹N•„¨?×iòpöÕ]í}e€Vôr &üùpA5ï·•¿…u4í'A4Œ¬ƒ#õfîŸ1 LðŒ¶¾€•+Ù¹Er¥š‘¤‡/Z¢”è²µ ™’ç69VYÓÂ,ïœ)˜0^ýì3É'nA!ÜÈçÆþ1„Ö[õ¼¯Ç¯dU2†"ÿ\¬•5ЋËs=‰/ÑwÍ2Ï´}Úq£ï#hH}vþ.øœúfá Ôœ‡º¯O¯Î­itÕÖÛ”vµ™©=qöè;€zÿýÄ—F£ž3s1D+Ñ˹ª³¨¾È P Šë ¶â# ÄûA»4 ìÁfȺԨá’ç¾+÷á1Ì8×àôׇ6‡/_´`tîtÃé¬ÉXnÇåÛãÈÚ vŒ´ÿòë£_žÛpùúŠÏAuqÛUÔß¶tG|kQim]U³]ezæp[‡Þ¨K'RcR²#cSØý©ÉÚíFµT‡ƒêµÁÉ`o\ ;ù\+w‹.˯¨ÔE F"Û—˜ ÒRh˫ʭJÀ’gº¬01+SÄ”´L\i+.6š˜òÒWÚëðT¬²Éª2+%Í ´:ë*Zª°ûTy)·Ä –Ê$[z¸Ÿ¹Ö‹y¦»XØ»'AüàYÒ±endstream endobj 386 0 obj << /Filter /FlateDecode /Length 150 >> stream xœ31Ö3µT0P04W0#S#…C.=C (Âɹ\… –`1#S a"€rNž\úž¾ %E¥©\úá –\ú@U\úNÎ †\ú.Ñ@3c¹ôÝôÝ€Ü}o] Œs~Nin^±‚¥—§‹‚ÚŽ¨zqµ•õÿàßu ¾ÂàÀÂåê©Èø,#endstream endobj 387 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1Â0 Üó ÿ m‚¡êR– | uœ*C(M~IS†³t¾;ù,ûá2° ïÁá“"Ë:ÐâÖ€#M–EÝ€¶ Ëgå…ì¯Ê¿Þžàk ³ñ›šI>Çs^Õ[¦Å+¤ x"ÑVU×Ó bý'•Àhг9u fÿ®¤h*±ß\C ޹in’ X¦ß3Þù”‚/ÄCÊSoendstream endobj 388 0 obj << /Filter /FlateDecode /Length 3830 >> stream xœÍÙŽ[·õ]ÈGn€H­uË}1Úi§-š©õ ̓<›'Íš‰ûõ=‡¼$Ͻ:’F·ü`‚â%ϾÏÍTtr*ð_ÿÿÑåä·ÿŒazv7Ó³ÉÍD¦_§ýG—Ó?.á„4b»è”›.O'ùS9±sÑÅ© ¶³AM——“ïgËNˆ`äìÏó…Uª‹AÏ–uó¶®NêjUW÷uuÉœ»"çRß 9{Ãü¹½üeÝ<#ûŸemt¦W×uu[Wëºú©®NêêxþÃòo©3FNRvÑZ9]=dyLÍl…ÀüÖË Kûõ¸!ù óÉE·KQ”“>ï)8Äò’`‰¸e”âtù÷Éò×ßÃ=)S¯tŸ©ØúƒÔ‹”,H醔ÿäœÒuS†bb&ò„‰ê!ø¦‚¼bTíß[ Q³ µvñ§’¼MÙÿPÔÐÄ*bbÊtQ€Â-”ì èdKNç ÑéÑp+,ƒ06A$:cp"Ìn`_Jå•›=àR87í|宬˜‹À²Æ`Îi={GÔ£½Ò Ôd×>(‹MGB2p?Ø…˜ŒIæØ%îy¡·èšPZü@Þnó÷^*dtZ;Zò¶ ß]“²“‘Qëe‹HíÄ-1eMŒqèÕ›ÕxD~õ•&‚ØÉAÔù_‘h«Ÿ¤/U)`íá+REé>ÃE©îjŒ:OHã(ªNÛÔx¬{£ä˜H0€JHMaããäá½ÉnO¡¤áD¦JÙ5ú oåpš§b9OlÏ”’Æ¡9n%ê@Š”gB)¢" ¤³ƒpó¦áù@Ù“¦l‘³”9úÊÈs$W†¤˜VëØ ëÁêNøÆìœõ½#ìÃ[DMIoBÒ“òŸÍYãê; Ö» M beuÍä¶-Ä<®«Ž©›µxøœù¢7ÑiÉf¯:i´r¥8Z£"ð]mwKxÀHG>[QEÌÙ¤PÞåtLçe2ø!(ô8“*Üä iJè#7´G°”Ñ´x|¨FC¾-uÓ¼çDµ2—„÷„j6Éï²²¡D“9†QîÚÒÊHj`Dô7òØ^ÞÁ㢣@ûL>*¿wÃŒZ£ˆjv~nqÌŠè ˆ,UÒÄ€6½ÄÅ*c 4J¸ÒŽÌa‘ØÕª™,×h #&;Íu¢MÖ§Ù—Cz½# ÊzTI¾3×ÖÁ“ø §·; ê[ ÆŠÂâäN‚+Mð˳ §Y‘4#y)•EC`×Á°xß9tƼHî>Êà&¬µô™°@Í—?~(ä ³´u¡A@°¼uF3Ä^Ê“¯««^?õB®êÉ}Lh·…Œ’1ÉîÓ‹ýO+ÚÒ/«HìQYµF°«+Ëœk·èѯˆ›©+Oð-+Ŭ,s®Ù§Èì ©Z”¯©â«è¥ô°‹Êkq5jO—òúÉaœE…jÐÔ"ƒ"§ „— Xæ4Š×1ûeþ—”BßÒœþ;¼¤1¦E?ãBþs†Ô€ Br©ý@Êm5O%Æ3@0„ÌtÇËô\oÈ€ñŸwÉÀ®wš 4[$ç\lW4s# ¹b€à„¦ÎÄ*ò&¢œ³Öèc¢LÎàê‡Æç̳\,ð¢§Ã B™§šrÇæ³cS3ìú5ÿwOÜGö±¹ øö9„-ñ8*dòì‹ù„·cÅÚ±9®„ÿøQ%…’ü<À·ü‘Ò[Ò51磾5n´¿õeIlÞžÒ©ú{›/ä&ˆÎ™_ù¤bˆWýH…g¦?çDaà ¨¯ίb1,[öÔëæoÛ¦jKnöò‚`Ê„r+ÖÞÜ;ü–W4C«KuÐnuV6d›Ixw¸‘ùŸ²Stv>=OØÜ{^šÐTysÅ3–ç&Yjº,>˜póð\póŠùä’Ÿn®º~ÍÇpõçühn@Ð-›4%`*4àÖm°£‰HÖ…Q„vaŸouµûÉ¿%®†8{"Œ/…º̈Þçªjô* å‰NŒ8e*TÚ œJeP4ü¢¹YOoÔÞ8„èë(Ò Ë»1±–Nð¬rÂùè]_,Ài54°D¬°D¬Ê°¡5©OÆAsÏ&W/á½ëé8¦ðı8j„N3iÖâ8•›ýX‹ó› /‹Cäf#í½kiëaͶ¾¸ergØÙÆQ#áܰçÀvϱ¹ ðH¡Ué ¤Ý0*§“>zùýYÓ£}U˜«úÞ'ýËx"µê#¿ ÐU¦”…iEŠø½ä¹Zᘉ _î|±Øãµ§» ÆG”¦ b•+š` ũ֨/ÚÈm*Œ Ù m!××fÔ?XÍ"£uÏ5©›ó¶Ièp[Œõn<÷ZîhƒÍ«Ñ4]CndFZ4ˆ¨‘&:žýä2âÚÓ"Ù6¹œmpe£Ë¸mÇ“œ@ÑÖÓmþÎ =O÷öÕ7œtئZI‰ÄlPŸä}œÔÞß¼ÊtµaØ w-ò ‰y”Ö ‘º´dÅüÊ…ó‡ŒÄ¥éRS]»žGBïèÐ1Ú5%ŒrU PdcÆQšÔÇ^”ž•¯!8¢(àJV‰ÎBGE³ÀKãã¨Öä9ÏᣠۜÙÚl?õBŽSU1${HulÏ,£Ý6ªÎz<9ùŽ †Œ§—"aB‡=Ó%ïlNðDL“¬”U¹27¥í§ø„ FvàÚÚ$ƒ`är/R`±Y›-­OÚï«ôVýs±s[=UÁl§%.f¡~úD*ˆ ŽÉbs‰JSγºj–wœJþ¿´}â5dðU’^I ïZÝ40¨j¤Äü¿kÓju˜ü\þ$)W޼p2¥ «m2ò¢ËÙX¸˜õ°Ì2SwX_>#ä3š?¨ƒÞ}K5 4±÷Øo d§­žnLÙ7ž]3«&@ï ‚A\Ïí-ãnR†ö§^ܰu™–ÊÀetR¸ìÐ k°Ó²“žŒXK…ó’{æ‰wXƒÑê°¾÷ ±nØÖ‚m¥œ¨Æ”rý·ÌXƒµ‹vH”í^¥(ßÀNUÛò¦|ÔÚ"­ õS…j•Æ?U‰[3ÕönYqt—•ôC'PLHžÝ²"’£õ04¬SŽd4gïßQìû3‰ÁøWMÖ.óÏZšÑ ýU=¢BSvb,]ã¹fóèDìñø†‘ævQ›-Š»Q|Qä͆阕”;ª/{SL†µ™ f€‰‰”Vù63$‰}ÇÈØc;È™Äîž{"¡eÕú3G̹æ‹ÞØHNÁF@Ý%Ìt“üvò_éi»¢endstream endobj 389 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1Â0 Üó ÿ m…€¡êR– | uœ*C(M~IS†³t¾;ù,ûá2° ïÁá“"Ë:ÐâÖ€#M–EÝ€¶ Ëgå…ì¯Ê¿Þžàk ³ñ›šI>Çs^Õ[¦Å+¤ x"ÑVU×Ó bý'•Àhг9u fÿ®¤h*±ß\C ޹in’ X¦ß3Þù”‚/ÄDiSpendstream endobj 390 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 4941 >> stream xœWyxSeºOÔ”"b´ñœ:ˆƒÈ2ÊU‘­²ÉVËZJé¾'mº$i›}}³ïiÓfiÚ¦ ÝYÊ&Ë.È*:>ƒA‘ÑQÇçK=Ìõ~-8WgÆ{ç¹ôI›çœ~ï÷¾¿÷·0#G0˜Læ#IkÖ¬œ=kè×—£O1£“FDŸfÐ…?„wzš1~É¢§k|î§ßðÄ0â –±#Û&=™ðXtätt’JŠ8弬ÒÄ5%™Y¥Åé…œÜôÝY¼ôŒÜ¼Ì¬B^:'7/'½¨(½ ÃI/L/Ú™^\Î`0&/.^’Ä)-+¯¨¬âg¬dfe'oÈ+(KÅ‘¯{õñ'ØO2¿e¬cLf<ËXÏØÈØÄHalflalelc¼Á˜ÉØÎXÊXÆXΘ͘Ëx“±š±–©eSÇÔ3 ŒqL#c<ÓÄp1ÍL ãYÜnÆHF€¹™98BÈz‰õöÈ×GþyTnÌܘÛ41Äg£ëÇ,»bìg±ïK›7~æøþGmæLèzlÕc扱§OüôqóãƒìÔ¨:nP a”Ž.õ2Ñrô;cEfMšZ¤•T “ÒJ¶@ØÐ { C£±“Í.htHurЧ/³Bœf‡Õ}=Œ1ú 7= $DÑfue*ÅNГ{a¿a/~ð¬:8ü¾¼ Ô®KHM­¥ñ‚Q BPÉ‚®º¿’¶¢úƨ¸èWbô?ýpªîëÖI¶ºZÁ%ÁwKëœ V¯…Ôé‡ÌAÿgñf¿Ùo =ÖÎXîÿüO@ôÙ²ó×l£<­ÄÄŽŽÜCw»ÐsV»ÔPW+ÉÊåù·!ÛA[“ÍO5öïC#`€¸°«7iö¶U<©|ww8R¡*SÂæÕ$ƒŠØAo×éüfÒsÙ{¸¯£6¿B-®ÝLÅ jqKÏûQO˜ýë,ôûØ’Ll>>d¨'6h Ð` šƒÆ 1¾²þÄtõ]ã_¯ÁŸˆk+Ïѱ$};¦êÁƒúFÐx¨M µ ÒªÕ"ÂCëÙgÎÖ¡7ì‡Ä޹åKN›ûÒ :Ž¢7Ò·Ø5 …N¢SëÀ Л.¢Hôò7h4uúòÝ£·øËÉ4‘¡Ö‚„’8Àè}ƒ‰Œ‹F3†‡}ù «ŽcËò4ò4 Db¨’™TVªBÃÀk<“k85[ÁbsîCó,AÊUÜšû6ˆ@Ðsèù»Î/ܺ='3Ÿ’¼³Ë—=<1)ŸŸW»Ôÿ<1[–ÅS‹Uꜘmiiim¨ë‡, z›r““óéGøYTÞÒíùÙ@¬¬½aÖƒÞà¤â¢›Å>tú¶ ƒòË/þ%$õÆv4ÉØ`ÝÓpÉÜbtµjò[-¡æýÐ u")ËË€KUøÌ@gGAÛ–7Ó¶•T‘²¾œÈŽ£è‡0‹æŠýý·ÐÔ«‡p) hÚ+úTt »D[YåD­ÚïoòôôíÚ»ŽÏž’öúµê‚<&9&†M—7Ž­Yý<²[lƒV" tVp‹*vnȸ‰^îAãú:«‚"'Éi-6¥Ùr¬è#BM§oÊ3ëð/>\MÆ9Ç)º÷.Û±»ƒ»: ¨x:]§AO ¯™VU£&k6V-O¢BâÚãw²vP­ˆa¨³wëL{7˜‰\±’«Í“‘UKÒè$QŠV•ñÚ 6KúºÞmýCóEÊÙfrC#ÑWеuF6=Eú ;úzü´»«[½¶ L\¬¨¦²§V¬mÄŒ‹¥ïœïìó{p½Uëež³ŸBOÞeEs£ûn=¿´Œ[^kR8Ĥ¯ÒÅ…"}-ͧGoáu{·åjøiñüýܲ~qnèÛýV”J8r>•Fµ©°‹ ·„ÑüôÉ12n0í§£ß³úË9”Œßî°M~{×Ë‹ø²J9©–ï 7ʹò" ø NÉK)*Ìâ5¸æ‰˜z¡ñ”£Íì/Ñ—ßµíºÐòEO¸ãAJÅ%r®fŒ(²ˆYr?ð‚K½d󪋊6øœm>ÕþQýÇ9u‚ÒRܳÂ&#â ¶C^ñ4înJn.æÆœ¶VKØèÆ {_ìwÝFéCøÝAî°Ð§è8Ûܸs LоÚee)” ãÛiè=–ìÎ[Ÿ›Â‘¥çß²1yÍ.*ÚjÐÛôTû÷çƒí@4vH× VdÓ‰”pMnz(€® ÉïÃàl)·V•r RV8ßyÉl Ù™§€¸7€ØÁ!”I5*¹†ÊO\^VD¦0ÒŽåÀq\K§û÷ÝÚ÷3¬FÑ l£ô`î½D¨¨–’¥¬PÅÇÍç(8;éñªê]KW±E| ë çLû-ª~¯Ñ1Ô÷F_ËÝÞ¶öŸPÀ•WQÏð7ÀbÎÙ¢“=ížÖV²v³h[F 7/«8j`§AÐS.p;í6;–S"Ì1JxùÂm[dŸ¹wïs4ªãõ ±µßêõ3÷^CÅ_°¢‰ÑGÙBüvM°Z#9QÖ n µ4õÙÑõý@O¦Ÿ™}bþwhò'ß;Í &hLAΟJ3AÄjΑS{½_µŸ¡ÂÇ÷‡vè®usˆ¸vBxfþ0'zmëÔÛ?"`æZEÇÇpA ¢ªûÌûâ_žU4ÀÅêðtLæi~ÎÕ Õ:©zY;¬QôÄûæxI† àU5¡¬Õò¥fµ•Š@‡s9tª÷üÍÅ §6èÄ^€Eg28Фè¾x]ÌÍ¿1ô"6 0;L–èÔÁqøkzòýK„4A0t¶ T+¼5[ õÞ‡æÄëBUv 4‡´ K|ÁêGN¿õ¦ÉìwN„‚aÐéïìüÓþÃ`b)\­‘*É¢íÂ2(Q]µ§* ðæ±(uf…Ò °Ùô&³žjfïeVÔ5¸­{mõ ©F@Ê´ %Ô¸ùu‡õr ©®\GKø]Å-;ñ`FÓÒÏÓϽp|åGÞ£ÆË§(kzGþá_£eÁ›üäÝ@pÔÍÞ&½Åè¡V¹ÙJ>hªi›v K€ØZs¼çh¤Â'ûºzxÏýŠB¡Qh•ýõ#޽æg¼‚<{Xh²CpMÔw©ÍtÜĬ½Ëhý»étÒ”“óÿü+Ø|]úIÄ© צÌ~všq¯®ˆC-Åy“éßÃn|P ÞâÆ[QNM¹ÝàgößAïßaEYÑgÙ7c¼ö‡`ýŠ1úÜ—ÍÎ[£§ÎF< F~šJ’ Õ¯AÔØòwL ìÎ[ÍŸZD**Ô’2ú1bÙO`n¤ÎÇ,ºOñ§ fm†”BUÃAƒ­êˆßYÅ-.ÏÚy lïñÞ“Ý dälÓYˆ¸Á'°áënFgÃÌÁ©Øð½‹.²Ñ‚ô[D}‡F~¸òý$Eø¿-Ìoî·ÇÿäB»õ²Z)øµøü–Ÿ¡Û¬7éÌúM´ýÀ _ÆÜèç,ˬ|)3‡Úºs [)ú™˜¸A(÷ãòúýèö£¡Û1•ŒþÀ67ì›.^ .#LÍ.Ë̤¸Ü”‚ÙøÕ‘1(«–ý­½ß7gŸ£Ÿ$éî˜Jé¯9ÓŸ_‚/}`þñ~^ÂûÙ —Õ­Cûiuþã Σ†QhlÌç s«e… %/‘fC%P ®ÿìøÖíƒVª‘Ë5déâ]ùrOØü¶1Béݸ 7|¸"̱qèAX@OÜf~|j?ö$]ƒ£ØÞšz‡ÃåT¹M¶@¹ð~,ÛÂÔÝbõþš¬ÿŒÐ+ä* %ãî¤ á. ·6*ˆí¹‡qŠ˜bûŸÁ_¿vËÃavNSÃŒN}æÞa´ÿ ŒÞ‡kÌ["Ópñ‹µÀÿùÓSî_U øžWisÕÙœfÒdëCkÌÁvõ…°ä RK4"ÍPšIîÍ8ÖÖRßå'ƒÇ±×•Kx[0ªjcÀXO=xìðÞêÐËŽpC$h:Üpÿ0ƒôZ§§…^ Z½f»Zš ¢aýÆÛÒÓÆÛ³³$UöZ™—•™‘UcÄ;38þÿÊJå%²"ª„&*ª*¥y Lz7lF¬¤û]œÏl=W?ŠEL ö;òr6¤Ó1•)”´L%ÆÐø!®:¥àM¿qÇöaHïeì[´hKZJ6©u©ÛKšJ}¦R˜ ÛV xÊô¶?ùíÜÚäå·ëëþî·'>¸6–ãCXQ.³Ñì™GÓOÓñCÒ;óÔ¿œ½8}t¥ì)ì¢?|¼Þ¢0`¢Wc®çoËÁƒ% á£&>ÄIÝEc|ˆ¡×ZÅøüa-Ü‘L¬XÄž»`8ˆuE}çŽ4#FämêÐÕw±¹ûè'_φõz†Ø:Ň^ò£ýn?2ù™?Žƒq–õãˆ?2¢çÙl:óêú!7—Ü”±izÕBzî߯ÇkU&…u(êu†K(߸î@öU¸ tô8|0r®Á•ìc9Á{ÎÀ‚^A¿Ê®UŠ5eØzTj%&½½SãÀ h—’@¢õlXš•´¹c¯Ôaƒ¿Q—OiKŽ  êtmÎîS¡.U¿¾lQIfæ[+r6±0ëÄe;Ô™©sÈì¿yÍÅ»žp¢£pîò(ÌôÑop*ü8ú<ÛáÀ.Í@˜Uf™B¥”+HlÌ­h¢,ÈëOÊÎb=ws¤N›ÞlÒSu=£Ú m¡'-2?B ÆþE#“ÑÚ/ÆžYaV˜-&‹ÑH">²G&;3›6&¸Šü[q^=ÑrÌeÒ˜%x$j ¹RalåHYþ!–åd‘pv„¹ïÌ7-ß\d!Ot%›žP4~èqÄïÎ%ÝE£o Ùˆü<ãh^™Ó—Œ­6á÷6u´ykk|¤Wä­­"X×Tïr¬ÔfG V|âõœ]ëÞ m ¯¥œåì`jÛ[D¬Ì¢_›¹i÷¾ëg»P2zª[Â;tº-ñF_ô¡j/óÈ{Qò=Vtåà6ަŒ$ØD­Z "×ÓBz® '%ÀŒKóQœ\V(/ŠvH²† nÑšLR¯ÇZàîFÓâÑK&4=8ü!åMÙ ú©ûKÙë Ið  Ø›(MsÍO\W ¤Â\ Ç¼„‘?]„¹·~ˇ}ÏD+ý,”¦cŸ/íÉ̯,-áx¹{œ.£ÍNbÀét@ètښŹëòó©‡Ãc?äøøC4_ÐMø¢§¯0ÑY{‘[GSçÀÿw:iþZ¥@^I•/)Âcåb“Üê²ú»dàû€1¬wé]FWäëshZï,N§ÈÚI@£V€\Mnž‘ñ|ÆsÚ°özæWe§3Ž,Ã7a)=£Pb;…/ß3o~ÏB›úÙ¡²¦ââ²²â⦲P¨©)„ »ŠÉü“ÛLôõûÖ;¬“=ìa––S^›¹ Ó–½SQµÍÿDý©ÞÄ,­ÔHTB­dÃ,inh÷’Á}ο`éªe¼M;ñΈ]uÆ&ì¾ô6ƒ p—&-õ{"--ƒ=¡¦^F©aŠYxáˆ?´©šsÖ&oÍN#—¯/«ÎΣ&¹ÙyL„Gd¨K GŠëþš.ô¡Õþè‹^&¢ïØ&ÎÚ?=¢Í•’ô³÷¯ã­MBµCäñ9 zÒhsëF3#œ>•.AO_S«mgWÉÃEþŸá~½N´Uœ®–$h•Ûg½ŠfÙžwB–VK+&´k}ƒ3½Ìþ¨”…Q{ïoÓ/ÔóêËíЄ?s³ `ÐèÕF‰CQ'÷Ò»¢çã+N¼åÚ8k€XfÕè¨=Ø`—EãéÝ")P› p‹êëm»·Eþh–e^4é×Y.Ú!çkiªZ^‚û#¶kͤ;æ’1{î}‰âÑø4 lÃÍ”ñTÅ2r]IÏ GÒ1ÙD\¹?šäp¸P‘?†Nw<se,9fä> stream xœÕ[K“·‘¾}Ùƒ}®‹#ª%uïÇ:66HZ¶WaVž(š3Cz¨™nsF2MüÛ÷K…JTWÏtHEl(,¢@"Ÿ_&Àwd'è¿òçÅíÙo¿‰¡{söîLªà§;£ƒ¢é´rP¾SÖ«Áªîîªûk·=Ý"NËtå‹Ûîù9–’!vaÐÚ׿>Ë›ÈNé0á;çͽìÎoÏ^öÚ\ßu/v7»»n³½ì¾ü°’bÒêþ*¯¾;ÿ*­:?Hcñ'­i‡¨ÁÍZ);XÝ_žõ¿¿¾½ÚÞ_ï¶ØëÛ>ÄÁüæÛÕêüím¾&.¤vƒÕqÚl’¤8Ø2çËó³ÿ=3ƒqN†î=Îðþ§¥þx&¢dg‚$™Û3#…‚ª#7uÄ9¡°|SgíL³þ‚=ßí+ãâàT'}Ѓ“£^Ò°‰3¾ûß—°kƒ„ü9»ãm¬lƒ·ìPãÈÄpaï89ÀLZdÛèhÝ t#…}ª:â½tÌ,++Îò8©¬¡ŠZ "rªq„Qiåu»cáTNšÁs¢2Àif§fr`T;YëÅÉ$lÄÛÐAøƒ*;ÙZ fµVA¨Áú›Z­µˆfЪ‡·á/štúüKú‹îÓ$üÅÂÆðÇ É-““Tû’ÚB礬ÑÔ­À´l9 T6Z‰:ÒPï¯9Ui¨ªEW*fÑ• ü8ɨƑ†ªðʨ÷G*a 4VST²Â ÁðiJ€ÜeRBõ Ñ![ÃÝaûQk ƬÁíÇÙ¼^º!*bä D¥ k»µQƒÒ>¦›~уŒø ¿Äu=ý œ£VêÓaŽM6†ô¥Ù­6‡¡ºÎAT&ŸåùÍæâûñù„ÜiË"J ¨D)b–4d< 1¥V©ò˜ÿ=E€â²iV É-FœD6 _¬,"½ñ€÷€-©…ø>&1ëÇ)ˆ¬õ|’žy¡¿HB'UÀÒûMŠ?À @$ëì!¤ƒìu0€° ¼/"¯3³)þµKÁõ´,¦‚0Tƒ6bsqj“K ñ”ÀJè–3„c¥m]4Ó¤¨ö&“ÄàÊÑŒÄ.e¥%DhŸ¥‹¬Gîäu¢a1tsÁ(ÒzRyåFÏN‹p×A¬rŽ"Eìÿ‡ÛÅ6/nƒjÈiÅ‘d³œ`69aav¤aZ*î·õÀ;+“EGœ–ÍUê$à¼yŒ³ì@£‹rÑQÜÿ9e.–¯²ÂBÆÅì|>€<Í—ì"§Qk¤ØïäÏ{*ÓÞMVý㪠'ŽE(Y‚l˜y§¾Ÿ¾‹{9{­œ×%‚ €Ù ’±±s„ÍŠ ƒ7pr-™k&ºfŽøaT¨óúï³=CSrÉ›áVHB±k–Bâ1@¿@{¼x/;Ûì<ОŠªÇ–ªã‚{,ËOùSØØäÌŒ…‚´~ÒvØÇBųî3àØæ¾·KÚ–Ò£ŒXK85Ó/7< À±ñ#‚±ê¿ík–]Ò à¡þ}D1ÈSËŠ‰¨ï«b`ßçä­?¤%t«]E‚óM@l‘Y n;%¢_ϳV³oài$þüá¥é åÙHý«šH3q+*”¤:•¨ yʳ£t2ÊTÎHTÏ>;ÿ¬Ø2·9Kw¨]]òžì8žgš”Ëæ–)âómÉݰÁÅD9Àç3OåóI²­|¾(‘Kdw$>÷,͸fKã|ô-›ÿ±Çf á{Ò\>[ħ Â5„Ö9T È#¥Ê|ÙŸ™{¥r†ˆZçàŒ?SÊ;/ù*XøJ›Ãy®ÍÑ!á$–9îcYrI=¾€®ž¼ÏCú¡ˆUº±ýè}Ž2¸+¨Ã¦ßìŸzmÂAW6B¡yÅß!‚9ŸÊ„G‚Ú”Ž¡eèyÆ~fv¤aciµl¹ØÃ98/ƦÔe?EŠ3+‡\´M‘â#9_Ñ%Ç ŠiÅR}„Ð&ÿ;Ûk*Ü\³2xKæHÑ5r’]1.EȈ…dE+p~1Õ'd«Nx•lÕ![ÉÆ„öXw!±o)îQJŸá.2ÌËv" ­w[>LZfïyá÷Twa¨ªIÅð܇¼@>F^ ê[òÜÉÿÞ$ D)eXss3Í,xMQ9Ê”ÃÖ»i§¶*Û9¾‘:BójV‘„ ˜^ ÔûµB Vµ9Ãf=ùS`c™g_”Þ 0ÌóeÙ$tBÙþ¤Ø÷@ÙŽ¬qzÙ^«Õͳ {<ÎóO+Û?Ÿ{²=¦lŸ‰3]|Z6éÞ÷  Œ•£¦VŽ3¯!9k Å%0€‚nv€cúS{„5>{jãÁã/ί bŽ‰Í¥ÆBã¡æ²¶öÐʱ[§ý»ü»¢$SóɆ] ÏZ”S­ˆšßûå*б¦é#ýŠèe>¡Ò&éìWÌ{©wÃBÈÊOxm¹…à†Ü?¨Ar^—¬¼W'ÙîÕßôDr¯üVþpé]ô1áÀPÇÑô¡ò¿E’i…09"¼Ÿ47‚ÔNO‡€ÊRqðÀ¹JèˆQðP5k‹NÊb6ÁC|Š\&6pÄàóL}k¬’jŠ Ô®#L{%V›ÆgÁØXD[2J1~Št¨b´Émõ²ó·èu|‘Ƭ&Êìîš´@æMñš²˜³ÕFpTÁXne„  I•_üMDyäA+Y|íU­Ä¢Dò–Y ¢v°³:—ÚpByW‰ùªáft‰öEAŽopmw EÅ+Ìi´™¶Xó"ôbf‹,b%EïÓ+Vúw}õÓ¯k¢EX­+¿|t">ÃR€=Ëؼyc´jÛê;–Û_fù‡.Þ‡Ð?Œ0¾Î|¤Y7eú>'5£¾{X<ô/F‚™Ë§EÊéÚ¦Þ:Ô7ÛÉ|é Œj–žè-¹œË¡5d9h5o”5—]Nz_ÛUFdjj.Å6µŽuÀI–_…Ñ÷aþÇ[ðO÷®ñë³úõ³ø úcÍZ˜±Šú‰öO‡'ØÿD} ’ í±6´ñÔ6á]hÓ‰öˆ.´gå#ºÐjzípzúA›§'mröøå+'Ë#ºy¬¦L¹&Ûzz&Ê ¶.àys³.Úÿ—N;Ø÷—Ï™¿TÂÕ‚;}|ÇyÄ á8ZL×þ':N-òx%÷à–!4>ìú˜é8¸Úü9òÉ^Š:=ÒgiõíÃÙ¶¤Ø]Ó1é'x7K?ÏrýФú‚ãòedºo‰ŠUÔ»5ëÉ”ûì¸wŸK£c»ÅoØEHó> stream xœÕ<Ùnd·•ïšÁ|CÁÈC)pÝá¾öÞ:vˆ[FxŒ ZRwk,©dI½þø9‡ë!/o©zIÛÅËåìùëŠM|Åð¿ô÷ôêè?ðnõìžýzÄï«ôçôjõå ŒàÊ­üä0«“§GñS¾2j’zeœž´«“«£ŸÖÿu¼ÑBLÞ‰õÙ1ŸóL¬/J몴®KkÛýê_Ÿ—Ö]iý÷º4ÿXZ»ÒzRZKk[Z÷ušãÒü]i]”ÖuiíðÓÒÚÄ“•¾ÅVÚ56miM¥¥KëãҥſÊÁ¸ýó™Á|…ŸOþxÄù䵿«“?üþÿ-öîö~r1÷tðEÅÙ}i]”Ö®´®KëÓÒª˜ç¥õIi©kà"Á)©É+ÅWÁ'çÔêä øçѱ“÷R¬w·ÇÉådfÃ&ç½·býü˜MŠ3eÎЭ3Æ¥# H5î,°j˜úÛÁ™+lnKë›Òz3ÀèWä]ún‡g6z²VæÁQ¤³Ü¯·÷Ðä@Fz½ýÏm'ë-œh“0âGxÔYÈh@Œ)×»Ë0„s¥+ôÖÛë:ÉYìÔÎFØ¥ß ; &öJÀ‹tÈ J!MšØ9ãUš ëÖÛ[èÎXné‡;Ò]OH׋”0¦éλƒƒÓ©ŸÔý_ÔF€ /TšŽ;nð;ì2ŽŽ»ŽÖãa… %ÎÀìr_=\¿ ß¡òL\*<hŽ;˜õE8Ðs9ý2ú¢Ž¸.`yÏŒ'¹‡±<Ñ87ÞdÇ 6œëIJ³ÚÑU)ÉhëR:Þkar[…i40µpÄ,—ÞÁž±ã\¤M¿rÜUþð* Fx *üÒ0kád[ÝSîë›r°K2ó.l‰ + §V/*ùÞ rà.'áÍ””„ÑîEý=ÍêI„Ïv•*Ú3$ä_ÁPIÖDñƒÿ`ÎÈ8PøLd;÷‘ ­“R7¨{Z@_¦s²¥mÉt <Úí©a«‹–«âJŠ+Ÿˆ(3ØF*àY ì'¢ ƒ 4T -WâÛR^»# `pŸœ CÌŒQÂä i‰þË2<ÀªŠÊ!ýž”ƒ–L­g2\) ÷~2<]$D*«VÙ]¥ýuýG‡üGXÏ;\ãîP2‘愞=G(fý ¥¼50™^ÛëP\dŸ- ¸‹kd€åƒgPoOéƒ4%-ŸàÿÌ_/êa «'1)-pÒ›zîY¥ƒZ£‡ê>«<è:o"¶²û›(žXŠpâW(@KÂÆ‹\zv(ŠѶ†t­”{ëŒÐoп`l <Ã9£‚êñ‚ÁÉ‘€$éÄ\XÏ{Ú{Ç*æVˆ¢ .MFÓ!wéinßöŠ$”ú¦v‰xÏêQ:•°Xþ ¦¨ÅßA±JéàO"Ñ Iq0iA,:CÕú˜“_WühmñlŠôÌ CzþâY`§¤œêð‘%.üÖe±{ïžÆíÈcL^ZoEÃt  ’§k­³žå!Jœ%·ñw˲žIŠÎ‡\^~Œˆ‚‰Óñ’*"žêè|l'ØH ÷®uqˆ“ùíÀ¬îæmi}SZïç’¤…eã³—ÎŽƒ„x¼¸‘á?Ÿ=m†#œŒœÇjŸÚUH>t·àÏÔº(À]üý-yglŽÉ¶sIâ#†Ó“àÞ‹DxRd±?'·'¥uYZç?••xi}RZª´>®dô^Eiý\Jòs鬑Y‘X „Õ·~¯”¡k§­MýÀTä÷ˆ¢­iò9Ò´© ®¸¸´fÔ†`øýœ•γ ÖèTðm¿ŽbWuf~ú¬´þ´7Ôt_¡S™ÿ))‚§lc€Oåðµõ»eñ¹ÄtŸT Õ æó‘\<”ÜýËÏŽÂk¿”ÖG •Î/"²BîUi]ïŸæ‡Ò:/­³CW ìëá4ƒ(êX+³Ï9²*T{0”¿-­miýcùËýÓüa€óóAëÝŒøb0ÍGE×=€«ÇD'dؼþð¸JýŠ«ïìTYì|?L Î5ÀÚå"œzósÃÁŬ. «49ƨcZz)ƒ-ljœÛ#£¶“zRÔAÙÅ!pڹ㖌ù0+N‚V^¬©ÓMY°pCèG[˜­zè»ÐËCܳÄ4xÓÀÚ‰=Qé8Å^Ÿ:}\:t³Ö3iâÒùˆ{â%ñ\"D– ×ÌKº×Us¼‰ª]§¥Û(ÞuœÌ°&1ù‚÷`µRkX#°HAÅ'®] Â=­ o3ÚÛõ±› @BÃ_71Pçñ˹c¼oBakâˆÓ€F:Q…´ëM±jpóÉçʆG &¼ §V‚¹¥_ 7€™ShüQÀˆUÆ7£ŸEœo›@À´§þ=nò›“£¿ÅT/|f_ÝîÉï¶™©”ßÏobÊ®”u“ò3¼ks|ò?yî·žO:…§óý„&U‰mô¡á­¸¬`‡†§êïüà©TgMýµ_•?DÙ(­Bè¥ÉLÅÜPÐ6€5{^IŸ:ÝR v‡¨VâIu–Àh+îDòÛi#˜㕎j ZÃ0ÐFÚ!ñDíÁ©!pªôÞV1YâNñ•edšú»À  ÌÑ£ù·}ã¾ÊšÉFñe1ôjÆV*p~ÑŠÝWÌá­œøbà½Ý=y\`\3B©æÈ5¥³r ëDH²ã±€ÌJ¦DšUV„ÔPË}UAžÅ¥{Re8B8“Wa\µ²³L÷š„üïãâ ê×ÿ^ö|IÄï4šê@§Î ‚óà=†T?X)AE ¨öΆºî)‘/ª-I2sÝf©ØYÊ€{I†ýwßífo äå0þ1B³ÖEŽLK-hžËše I» yY“˜ižlÌEFÇÕ~š·Ê= Ò©š§Æ$‹v'NoTÏ QÄíÜæåe›#”K!Íok¦¸)›)ø¤šâ#Ìp'«EçÈy¯lùËÁX’|’B ðñL7›@;ƒUiüç½ñ¬óq–‰¡søzyäãÁZWoµÖoÓ Ô*±`wÈQÒ[7¶Ä9Xo"ŠXø ×WhM8µq$>IÔ5I–—ÂÖ Û `8ì=ʃ]›ol3²¬»´˜sg X„¥­k£'L€.V?当|- YðXgd—ºÀ#,n`P€¾slw% V£ñµ>ô|ðëù`‰ß/ä5u¿… ¨'¼"_.•JÌÜÚøÅÊN'`¸åýëd'1Û9£Ñ «E ­{WbD/´+‘#Óœš”i†A[#)¹/à®é’çƒ_Rýû³¢UV$YMµL]¤¿Í‚¼¬¸¯dðÑ`ªÊjnüù@Ñ] ÆÕ¾šaŸã^újÀôãÁºU8×€éÙà׺¿ÏJ뤴jJâÇÒú¦´,Éþð˜Ï­ZJý]iÕ”ØÍ@oÕTAÅl¥³hj”ÉÝÎQ‡$Æ¿|òaèçw¥õ¡L°Jg%Z?н¿Œ5ŠÎ‚{hkb®¯{­47¬¨$þ90bŠAjá!Mà4ߘ«‚sl!ͧ&U êe#À¤çàD5å(‡D¥ù¾à·µõX8TÖ¶”ˆk&¬B¼~÷´Ú°CßÏb²k}÷çÑ¡/¶+)rBEƒAµhúœÇSȶÔv©|ø¦ªÔQè}¨TªPÚÕVq…n¬zSc,,x¬Áì©eð+%†t9zîèŽb™×&Ü6ˆ›“Ô-lÌ¿’y¹f=šBœË%ìf {‚ãv%'©§û½,yMÆ-¸þ’W'¦3¿öÚŠc¤°s¿+:©^"ºh¬jj¾,-jLiVO ,ý‘Ùø@QÀà+ ÀdJ†ô ¯ô ç° 2_HŸ”Óöy¿ä§œ À€úM%å2m¹Ú“Ã$íÒ”—ˆÁÌܵòBùeð³”]÷1ïyÀ²ÞžXˆ‘)¶Ozþ ‹Þ%ü—‚¬šÛCvQCz}ZÉÆ:Ònh,+\±Ž÷MR@!äÿFqP)J“ÞYH_Ó"~ÍlÈZl‡ým=Ê1ÇKR7¬CƒíPD¢v²Z3¼îà–¼ð¢ÑùÄã4~N|ájuÂ{êO3ÊtŠ–0qÇÖõ „‘evQ”×ýÞTß›LÜÈò¼D=Qõ)5ߤ«'~+YïÞ™¤y¤dÌvé\ÃÒ)à¡Bel –v¡øŒáõ˜þ*öú¶ô—&¢6DâÉœµK&•âÒ‹‰(¤¡ª¹'Y¦$ÚesÓØM\è€Þàµ&e,^^@ o8¬$eÔÜ 90•£ `dàM f¢p2Þ ª÷S»Ú8Õ Ï)ß¡j”“R¢¨ÆP²"'\Û#UÒ§¦Qå`Ì+åó§dÍáBH\ÕìÛÄðt¸‰m2½A˜B¤$ B àÁijÁÄ Ég‰å)%mØ$miMüy\Mºåð6þަe™\\#%t2â.}Ç9¥äY®÷¢"¢$EI$ÂE»‹±î V+®']^lºMJÿ)…mDÓK.¹D½,Œ á«D0=OŽ1ŸîÁ§©\r/°èo`æ^º^ºõ‘îUÄ»’ ú¢ÒW„ÒIÀœæ{1 NÞ±Zd[ «98ÓhT8‡†[“n&_¶‰È¸ í¨·Î”¨k¦hzQ€á&—\‘OöÇ­ÁZšŒ.¶c5ŸoFóíÀ¤~ø¶†B°Ö%nÆ×‰¦¼ QêB ì´`¶ÜeP˜e¶œ"›~áþèuA{¦Ü’±ä¿¡8a =¹œÝ¼†}¤|6*ÃJ€…ÉŸ]÷ ur{o8QW£Ø†8ö˜ÇäUêÎ¥‘‹4ΉŠÐôÌ1z×wZÇh¼Wð><‰§/Èô²ÿgÝ 1p±Ze.O#ÏäéOÓxzs¹3çfZ+Î/ÄÁ,ÚÍÄoúî%­’,½—ñ3'ÞZý„Mp~@aãM­%³Íʘ–¶É665OdÇEð·wô]HHý5öEg,Õšà„O¨ªO€°é^ô({‘¦ó­ ‰v*VìðcŠ5$hÑ”¸’4qW«7O)wÏè†yIî’ËæÅL¥ñ¤™ÁÌΠ±ö€ xˆXhÔ—oþX¸é«Òú²´¾ Aˆ9'^Nœ‡×kȽ†ä®É¸ïZŽÔRx·Ó•)ýõ¸·ÇëFGÉàýl¢L[®1S?D¤Ã(íJ€¨ÆÆÐ„œŽK–>>ò®’Þµ³¿ë3¾0I 6„„í)¼l·'ˆß¢0iÏ K"\bž=°4&Øò󧥓d^j¢­FŒkL¸ÆŽkd™äë~,­zQ¯Þ#©5IgƒiFi§C8¤Ípôí¿wÁÖüÅÝãð_ üš ­y²jàŽŸŸ:õ׃}UOi”¾{»4Ú ³Ö›‘Ì&hhB:O I~Û÷]0=,‚’s 7ÚÚ¯ÔU…ï¨(¬&ëG”yMÆm¸¶(yãß/Ó߯Ò߯ÓßoÒßG†e› û«nØz}ŒAàEøƒ¥{Ýûç#Yúý)ÍÅë°Ñ¿ r#V…¥«ÏØYE»‰öîHÓ‡ö?×yÜ1ÆRl];M™;™2ÜÃDÉÀ&:nãds¯›6:zÐîÿY~®ïú¹†<’ÏðpÔsõ®3Ï. ;ÀÈÑÍgxóCá!ö˜×)}éT#¼ªCñê™%KùDÐCÈzºù)ÝÑüg?Â*=ZQ1@ñ¡µÕ+°Cÿpdð•¢Ç·œX] ©Ý6hî¹W~­G PBTÎ8/„ˆY­t€f2ü..ÉlxÍ­ìéߎó€s²Ìm«Ñó þX!| ¿„Óâ£sdñ›83FPƒi=“,¥u9ƒ™M3üºOú!ư®ï¬5§ÄH±>ÆÎRï9YÓtÑëϰ‰e íèm™ã¾éƒ‰‹SˆT0ßtó!`ãå.Â`ø6 L2-yO¿Õ&F¥Cºw§­s\Õ9¶dÈ}ù4äœ 9›A»£çÈо'ä°‹ØÆôô¶îð’`õ²Á°pS¬_¹ŠIÎçak†k×ì>­cB˜æ‡ÓJD‰½ ]‡v¢UÎBþÆãž[Ð\ÒÉ«)bèÍÂQÜ¡‘oô8Õ\–s³ø^Çl¦£‹Gµ†¹Fpì7~RF13¢V“gNœ4¶°œ³#¾M?œÅ'Ô€_gqBàk®§-ª–—5N‘Ø1÷Q‚i¸”ã”íÓûq4´sK7•l¡*¿bY`”=6ëߦ×é:¹tKjÆâUJá§<8Õí‘2Ä(·»‘ _NiÄèŒŸÒ‡Û –þÌ´D©ŽÇ ‹ÊÀ¾dXêR7¦Ã»ƒÍœ»NŠ$$''\)ùº~¦giTŽÍÜÓ*‘Ì=dp+ª,÷Nþ‘¡RÊGƒ/MÇ`XtÁ´“&:ŸƒS–…äûP§dƒ…´¸ì!ÕQzUt@Ö¦t³¥™& 1Ϥƒ1Ñ’N&Z”G…_E3#rS4ÏÒˆE:»ÂÕ¦z¤$§æŠ ¾˜S–ÒÒõæÑ&ÓU”`¹¾õf`LìI éâ;¿^ !~Qqk–cïäi|zÏ!'„ërØ^ ®Ü?²ÙÂíæZ_:°Ù0ÃñÍf'gªŽ{5L„Âò±ì¹ÊxÞ<"ä&"z²C­ÜAJ²vµ1°K§°äÙrž¦dûæ2ÐÞÙìð·÷Cˆìo ¨köÊ¥âaÅØšš@ÜAê'ôÔåsz!”s¾|xfoÀ´ö9…ßá µQFÏ}5 ä¯FÚÌhí–4ÅC ý%C¢S,t÷ÖØC6ºœÊç9œ®8HÕ«W¯01 Êd×Ú¡ÔÁ Ì&3ÓÂ)<âk˾´LN“_‰z"­”Ó} »ƒãæ‹Jذ0kS:7“Å£ÒS¢|Ç‹sZ-”^ /þH¹é@´ œ­´Ãåîüˆ³ƒBW<|øPWèì©3'/€iefc4‘ª®¦¾èÁÈÑCõ‡·oÙ!*ßI +ªkD’û”™ü2+U}Õy,¹Ã³¨Ó#R6t†;^åñ¶yz=ݾ­Z#­OŸï8{‡NŸ¹x È ®]õ›˜½_H7ˬƴÛWŸýQ Ö`oŸ'þŒöŽü ŽVÆ5º`9,…meR¡énqçNX[wÖV˜¯”ED eNIª*µ[ÒX/qEۨЗݧ‡hSÉ•f9m–D•Ï™kÕµ(Àöÿ+Œ8Ûã#±#'Ï]rÔ±³QaÓi·Ò\uO²¨£Ãä=,è˜24íÎtjZfAOöÔ!wv6Aü (êÍÌendstream endobj 394 0 obj << /Filter /FlateDecode /Length 160 >> stream xœ]O1ƒ0 Üó ÿ € b¡ C«ªí‚ã  8QC_ C‡³t¾;ù,ûá:° Áá‹"Ë:ÐâÖ€#M–EY¶–'ÎÊ Ùß”<Áf ³ó»šI>/u“WåB§iñ )(žH´EѵÆt‚XÿIG`4‡³Á.£ªq÷ŸJЦçMÀ5☛æ&©€eú=ãO)Ø ¾qSÿendstream endobj 395 0 obj << /BBox [ 1205.14 4061.76 3311.29 6167.91 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R409 73 0 R >> >> /Subtype /Form /Type /XObject /Length 82 >> stream xœÓ21°TH/æ*ä2420×37Q0103Ñ36S0240г4…ÓE© á y\† éD(åJã2Õ32067V(jñâ, v'Fk0W 9iQendstream endobj 396 0 obj << /BBox [ 3012.73 5149.19 3101.97 5238.43 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R409 73 0 R >> >> /Subtype /Form /Type /XObject /Length 95 >> stream xœŠ1 €@ û¼b_“Kây/ìÔ¨ þ<¬-†™bšÉ¥`¿é$M’9;\Zgk‘T„K|¾V,8¨¶v2Ñ`3„†r.èœÅ"½.ZwÚH1Tú?óL#=ýpendstream endobj 397 0 obj << /BBox [ 3012.73 5065.16 3101.97 5154.39 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R409 73 0 R >> >> /Subtype /Form /Type /XObject /Length 101 >> stream xœ‹±Â0ó«â*xŸôÿ®€gà€ Œ3f0í£È1ÁÎn²Ó=4sÿàRÕ­C-Ìk‘lÎÓÇÆ_É®’æÎT[ã%Lžõô±á‰´*ïÎ/ —ÁõŸmÅ ?@øsendstream endobj 398 0 obj << /BBox [ 3339.22 4061.76 5445.37 6167.91 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R409 73 0 R >> >> /Subtype /Form /Type /XObject /Length 82 >> stream xœÓ21°TH/æ*ä2661Ô³0R0103Ñ36S0240г4…ÓE© á y\† éD(åJã2Õ32067V(jñâ, v'Fk0W 9™Qendstream endobj 399 0 obj << /BBox [ 5146.81 5149.19 5236.05 5238.43 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R409 73 0 R >> >> /Subtype /Form /Type /XObject /Length 93 >> stream xœŠ± €0 {Oñ›Ø!ž‰(˜ Ð!ûK¤¢¦8Ý×-&ã¡‹R2åÒÃ$§Œ^E8üó]±á¤Öæ8ÈÕ‚MáêÊC K ý|WÚI15Æ?óJ3½=ˆendstream endobj 400 0 obj << /BBox [ 5146.81 5065.16 5236.05 5154.39 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R409 73 0 R >> >> /Subtype /Form /Type /XObject /Length 105 >> stream xœ‹1Â@ û}žÀ±Ïö]îHt@‘@ºH"ßÏIHÔ£™bwº‡v®ìp“¹0´†xe1UéùóûÉ…FGrEZt cjmÒ*çõn_gs¼RÔ›ó€ñ:¸üs{à†F°‹endstream endobj 401 0 obj << /Filter /FlateDecode /Length 3953 >> stream xœåko·ñ»~…|Ùk}Ûåcù’–á6-‚MEÚg^©t'ûä‡ú!¿½3$—3»7'b·Žøƒ<.9og†~yصê°ÃåïÉõÁ¾áð|sО¼„h ÑÀ¢GŒ{ôš!ZÑ#Ž*B”ñ©¯8u´»áˆªÞ+†z7Æ·,c¤e[iÝÑŠéŽÖ­uº2ÿ/³9¨b×Yoš[ØÅõmˆÍâj£!tA5œa»®ë]³ÆÐð +ö]§{¤w Ö:kàrVÙ¬}ŒÊõ0»kup>\h‚WWÆÕLe]eØu}³~œ´-ÐcšÙy•nN NÎ…먚õÙ°HÍÈÄx뀰MZ¤ó¡Ï„™­5…Qc#³@¨;;^‚]}VÙº©Ðm…^Uèu…N„y‹ ½¨ÐU…N+ô™(: MG«”Þ"þ>°¾¹@¦FÛ[Ý\&jPÐù¨à÷ÌÕ\ÏPlA«°Ÿ#¹¹Ë?¥1P%`Qð¨Ñ8 ÂlnªÐ®’ÜcÔ!I[uÚ»"xµ-‚2Ú¸fY÷åØ&£R¶çŸ]à ×u&ŠâÅUÑÄ_W^­ÉŸ/yü³!p&rÚö@µŽÃfðAÅ’Ôw}[¸†š ˜+í¬ÉTôØ32ƒ{Ÿ©WÑuÖ¡C:5øÜ{¢j£W‡óêqÏ$ÑàzPÕ³!YíÜhè@ }Ð}Óâxè:¥ôÀÈÀ¬kì`ÿD¯¥IÍŽŽï*ôœ~þ‚ÀùšÂstu·zU¡×:æ-*ô¢BW:­P‘‚_ _ ÿ¾Bß æõnÀ_5?ø¼þ~'|ó„&~u¯ž‘/¸Ág‚_ c´9 Ÿ.’οa» ¹ÛB1:´! ºþ~¬n·&þ(po’­!c36¦¥×U7×ÚAÏe…HõDEú®BÏibv à×ZW9ø~s¯Äniâå„<0F3m­q^ö˜Ù}T!o¥Áš-¬ñÙšmŒè›Mò-à!õ`Ù±ã‹ôBnÀ÷6 NÁɾ\+8í¸¶ÂcûŠžÖÑäìÊw¨qÃ}ë·Ê¡E:¬éc³BÏï{Ls† *kCÈQC¦ 6‚›RÙ•pYph­–›âÀzÕKÎyîî‚3´]îýYÀøÈ€êAàa!BêáŒ?a>c ûÖ5çù„ ÊøìGF4N×ð&8à;æNóÚ&"YÓøfä–1òx™Îb˜‚ÑÐpnð é4‰°‹E,wy¨/ê ™cš|ž°ãv“)4 yg4—_ëë•Gø5AFttup™ùÇœœÇ“ £K6|üù%‹èÙæpLŸ¸)È΄›öKrOW&dÑoÔˆ¬L‹ÓÖLüØ#t¦âv>o_ûN¸9ýSŒu.Dt« µïq§{X;<`ÎÕ‚h‰ã^ŠãçV¡§÷: Wú*±ø6qf¾¬P_¡ŽAB²ƒ­ó”_†æ4x&XŠ‘ ²l#,s-˜èW1ÜûLX›Ò+t¹"Œ‘€µð+ÈR¸"¬¤AÒ2Ò Š$¿¬Ðq…¾­Ð÷­‚m]4j’Yc2<ú5ÉÜð¾2$ÓüÅËp"¸g¿&ÁÑÏû Žî‡Ÿšàè¢zW¡e…è¢Ê㌚Vù¢²\ËR‘¥P CÊñJ˜Gòeù´BO*t$ëÙ”€®ÑN )¦Ï ÔL™èïöàŽwDÿŸjó—z-üºt-},°GNJÉGå@ÕT¬cˆ~=©Ð…Äúk†ãí+ýGÁvåî^_TˆÓ¿Š¾Wro·Âº¶²_O |[µ·+Õ‰tœJÞÒ8´"åΘC9úP{Ÿ{Ó×”£ÀuÅ0cžç 8ÄÎËÉ®W„ K{íß‹>¡J•Áó Q†Œûq×·š(Ùßã|BvAATãaTÖä3ßtf½4ãTiÕ¥0oCÛ±{ÉFXûZ€h¥”ñ}#L|Å gr^ü䯒ÿÑoGþOÿ uÊõÓ¯íĹ}TY£X‡ºüç$áâÉp#òl[˜’¨%o{.|±y xù$¿¿“5;Nù&)8ûÔeð€£3?+ÚùÚîi£ÅšÓ(Ù¶OTEMLÞïÈšØzÝך/ ¥.Ûž:t\Êòcݯï0œ' ¯Î·ëL)χYÙ*¹Û”YÇElg@ Úµs«”‡eÅVŽàYä%åzY®šåŸOÆŠš§o7”æ«­§IiÀÍÃ"ãZ«=ѧ9¥Ü»ËHTêpv,í@©È“ÕËÏi÷ÛÜvSY ‡øzFÔš%Õ y¶{RI¬ã繦mç"¯ñ:Q&=´Aá ÛÛV+Ôe&Bö!0$•ö-HO™A—FU"žl¯›´³¹E qè +M/µÔ”jG0l&µ£Q¡¦–‘x]h“K XpUf‹äƒ¥ê(Ž?*nÔL–gë.ŽS¡`{ïI¥U²iïÛ¾÷TŒyÓÿA«¶Xƒìj9jÍêËr·6;ÕBÑ´Ëfì÷>LÏÙ¾Õ«øA«W¥:5*í®³P±u€úβž+Ô¶Â ý“ :¦6@,Ž;ʘ&d=aËŒ+Óè|oGÄ0{a«²âk¦[ó²øiFÆb“«ï²EFý'ä…¦5×Ààµouç47Ü+6áÛT÷ñPÙHwèûá$‹[6,Igê öé%é>ŬW]ñH|6u2ÏEØÎ¸äƒ]ˆ>neÁЩ˜Ê|Æ9où)¯ë'9hä^¥Î‰ëڜԗºsçí¸½]š6zÒÌà±™Aó>ÁÉ%óMþ̪Ìû·6©U×ã™X¢¬éGgut±åÉR»ó0à*z§ÀY(°­€‹j¼/l€§ëVgÀvýþŠê÷DÄ’6~ 7GwÎà |S §¢LA{Tm³YTþ8, H?:޺ˢöqJpîÁ¹ò0lý:·ÇÅÉéÇ:˜€K¢:M5Ìõ_Ñ Þ6´¤Ðöt Ø4Ó¾é2™5ÿ¤@M!Æ!P‹[Z‰œ•õq8OR¤VLƒqí MõûŒÎñÁ¢$)ƒÊ‡*6²áw‚»'WBÄRp/_Vè› }_¡¯´3’¡ÎNFÊY¦h¹§ï(sÆŒ›âj[ [mgÇ^‰FØTU 2 ʆs_ûèNèZÃô`` o-hžË|C§AQ }Ø$¶ÞïCÙHÑžî{þdú,XK¡/÷ví׃˜0Žh©7°Ç3ËI±(4¼AŒ)No™S„7íí&Î|ñÊØh¯] -æ Mà,ýäþÃèZÑý‡_hÞfdT¬Ý—°l!½>g©›eÒ5Üà®ûÏ(¦r#Dˆ–µÄ­³z[¡ŸÓW0‰’ÐÏ¥¼6]I¤ ‹ØúÞ?Ó-fY¡÷Ù„÷ø[íe¤|¡@ßRWlÔKðµíˆeÀrѰGaÙå¦r£@ÉŒºÊïm*·{5•×½ßÍjwø-ëÇÖïÜ?þ‚nf{v•¼‡-sƒ7¢;ç74Ñk]Þ;£<·/9€hŒ.­ïöˆ‘Q™éƽWÆÎgÙq/.&¹"Œ·ÌwÿÒŸó)µàÇr>DP„ëêëøËªMþÎvýÔYvs„S/ssR¶^*w\'‘.gR½ôázWlŒ‹ 'ÏÙÉwîÅNæÔ7ˆ¬Óë¿Ä³ÀЂx¯¼OºPüÈVäò‰vtØ^#«ÝÐî½M-ÀƒÃ[üJá‹…»œ¬B¢åêr4Èø=Š"oXøšx3ôÁó·øê±XQŠC.h2[˜á˜Å]€/·“} ûˆ÷éz(˜Dï‡0(僖`zOÐæÄ#µ…£äû6åœÜux–¬ùwUý Î5˜€„¹¾’pÆŒòŒLq.òd—nUY\ê,¥2gw=6\|ìLj\ç50“®`\WyB§æ™nÜPÔ$?¬”»¢—DÊøÜFÙÆ÷b–åH-Ý‹Q’¾²éGØ+Í×àåÙ´ûÞk³#W÷®h9˜¨æGDØ 7—5sû=kAsi{&ïø›ˆÌÄÛ¬]Qõ£¬ZU$Å©ÝöVéâ þ~œ™ïrùYh¼ »îxwã“{>p#Û–õ¹Ÿ)´ôǨ´)è ¢P^éåæo~q… &n­¾íày"^0ŸÒòºwÅV«ýñ;Ò:‡¨é­vÅÔe·\¦÷€O£øF{T†&×ÎaFÙE.+£Í¥t KLóav`Jβ?+qµå LØ”ÔÀ]óÑî¾q%»QÙýø¨©çB291ØV¤@äží·Øl¿äÊ ²‹œmò:*Iôü3Óÿå8ù0ÈãAw>Ú$ÃÞKMm:VÛeÖ‹‡rÝC¶žHüÊrÝÉ{8ónö¤üÍýéa1–þIH€Ó«í#aŒÒß,'N×.ºïIµ/ºŠQc½ÊÞwÞ[‹Îz‹Ý¿´¼Ô“$=…£7ô¾Ÿ“^gÓ–r-@Ón¿1ùÿîŒû‰ 3Å 5óJ'5ÊIíxÒ»²S‰™§Â‚Œë”£8“^”“¤¤ÜƒÔ*#Í“:Oå$¨}˜Tˆð×§\rãgÇ?øüû/ΗGuendstream endobj 402 0 obj << /Filter /FlateDecode /Length 4785 >> stream xœí\mo$7rþ.äG Œ7sÙ™k¾“FœÉÙÆ+ç|èÕJ#m¤­FZ[@ßž*²›Uì©Éë8wûA\6›¬wV=dχE·Q‹ÿ /îÎ~ó¯).¶‡³n±=ûp¦òÓÅðçânñOç0BiµH›äµ_œ_•WÕB³é’^øè6.êÅùÝÙŸ–¿[­aò®³Á,¯Vk»ï첇f€)b\~\9·Qɇeÿ°ê6::›ôòš*šÌ²‹s¸””wË[ìO6êåe™ÍÄĆÞxïn…ƒJË œË¤d­)sE¯¡ù„ݾëLZ¾Ã±É:›—¨ƒßŠ}g–û]YÄEµ|Ä9:´]^×Ei·DŸíªŽ},cS—€v A¹ÆÞàØâ‰¾“·ŠO¼¿"‚"*#u°–³Á.÷EÆ>¨°¼[i¿ 8”±_¬”uËþ Ì€ŽØèã±,<+÷þÀe<±Eî°©S„Ñ—y@pÖ£ÿqþ/g:l“:ÄxÛ“Ñ€Ií.ÆÀe‚VÕ¦ëX•Ý)“pÑhã‹òm§b²Ã¼ÀšC™ĉI¬í Ö$_Äfµ÷|ÂLMOÙnt—TÑô°ÎŽºùò7d{l¾@Q¬M7ÎØÅZ©MrN‘|·Ì´[eSY5—ú@@ç}toœ²lðíjmÀ]#WáöiPF„“¸fÎwYF@?ºCíîópåT¨’RdþУ¡¥ØÅF-™ÅE¢©)«¨|ã-=óÉçò^Ói€1òtºwe°KžÛ”ª“F$U—GºßÑtMpäæG‹Ag~Ɖû@ÌÒÒ|ÀŽ¼Çªç²HT®uf\/ð) _d|·"f7Ù© fÛÈcvÚt¢vþë?-ÿaµv” Ú:šü×µõ­Ðz¬­¾¶ÞÒ<_;×Ôùà oc 8S[9ؼ¨­}m=ÕÖ®¶ÇõÔò¿©IS^ÖÖôøï¨y_Ÿ÷µõÀf[ϵõ†^~'¼ü(ô}Q[_×Ö·B 4Éš£&;¦I«6Áë!bP)Fˆ{A¼‚ðŸó”kšƒ‡™CvY[ 2£,ìóz`µ}|Gº¾©»Úš¾‚lÝ“ðÝg D3«&ô]‰3~ËÆUòtí µ• žtÉBð:‘CôéÚI-WèCºlýéE*” j©j¯-G$غF TíLµ¥D_W™1»`ŽNÚ}¤'q@ô\[?1\ÔidÛΤ«¦þù¤X©õù‹8Ø:H‘í!ÜMˆª¸e®^uö5ÿ_Ä0‰±£Xõ—Fu~sÒI7—LøU±{¦¥Úy!èXŠb‡i˜*Ò¯ÜHû­ dRÃçô29f¬­Žñ¯~>þ¯þI&ä’9$^t¦‰=ÅHS 4ì%º_³-`ë^蛾Ѫ€¼ç‚=­¬Ã]Q'©÷–µêã+Áhr²š›œŒË…(ï±Ò®këFð’àÔ¤]jÃLö…“Î}%:²–^¢p²®­‚lHˆO’ Ø2ÝIÊÐG-7î*-Š0uÞ𪄡Ā¢7 kH…-l(&ÀRX³@v®5n´Ji»q êÝçÁRª”C÷-u÷Ø,—÷XÜj¥MT¹°Æûån ›¸Ë«Â„FÁ,P|\™\Ä\P„N#Êñ0ŒŒ¥>q·6ÂX|XãÒò-'㲌6¹ÐÀ¿0¸}ñ˜ µ*9,–ë‹«ÚÌUbì:…Äç4dhê¸Du§7Á©±ïsÕµºYnÇ:ßâšJ»ìæµ÷²ön鵞 x¬.Yo.mK7lµ_T¸P‰ÙJÞ›¸ ›NAÕ8¿“Éz`tSïMíÝÑk‘>ÙÛÚ{K¯}*»1!škcPûîÝJÖŸÝX—¢ ¿{FÃ=#’zŸ‰°%½'Ù@k“Àº¯ˆà‚]* s v ëÌ&)Ë¡¡}„Lסårð¢TíùMVq·£qhfÛ¶xîÇî©Ø¿GlÈŠWõ;*Ú…:g µ~ !‡ï3zuÖ(v‚ÛÍa(/uþ8°äâ€ïåw(Å´ÈN…ËFQ%ˆ¦Ap´èYà»–” &€‚@¹>¨´¹‘¥7]p¥À7ÊŒxépUQ–u°Tl0µþ•L—å½FTáNýÍÙXœV{çKGªq l„·+#<„àA76p‘S' ;;Ñ$Äg,àÛ@@»c÷ Ðëo‡h [Iö˜‚L3Ãæ¶œÑTÃÍ€8Â~Õn‰ ,žCóê$søZ…[„j´àm^¸Àw¥g_WB²l׉{f4lˆ¼‚lÏ¿-S  Ž¥9 pEšÝÜ‘8+²LÆVê‰[Ž”2v³¸Sç`çäøÃcÉÙ Ì>¯shT>eœp@ïm3Û@§O@óÀj‡¢½”£‘lé]Y°€Ïº×káÖJíG,?ƒÒ‡T´†ì´GÏ„`Ù!ÈþqÐ@“Ýw«8&‡J‚*î[—átïPÒ0¸ –oŠGPEv¢‘ïy'BPúÖâJg…2õs,z_ðv°ÉÈý¦ÎÖ‚ÃFiÄL¹³ý@bÿ)‡B¸ˆ‰¡² ÄÆ} —Ö@Ù–!Â|Sº]ç¢ÌŒz#{ä§BÚîi~zSƒÂ3±t¦UX`fõé”±e…ÞÓBÎ{6“pÏ–;dq‚õ„ùpƒSC¾ŒyDu¡L|WNžFæ¸õ±ó¡›²tv]Ý È”·˜’dSA»–Ϫï<”¡azdÖàðÃXF$zÖøM„œå«U4Ì5öl.Ù¨ú‰;ÕªÙ0ë›SÇ9n‡g>û"Å`¢žmÎèܘŜۧU=ûÂ;ÄÀ5"Fª©KêòA-;l`îM9ÙÎ'>€Çïa `¹±_q»§PÇíŒÙ/7û—%BÒ<†•0 AÃ:ôù€Õª>‚Y2ˆ, ,ž:`žûì -oá5ËûiHÍÛK—H¾O"kŒ‚ž—9o„´Ñ¢ïÊúCyÕ'ËÁ²üʹ€NP Uì`ªƒš|Ô(Ç鸠n¦…Sé‰ÇÍÆÌ]mPV¼LÊÿ©c® 8¼?ÅÑ„ ³BÐ¥ð´c×`fÙwˆˆÔXÑÖöÆfe@³S¡qÀžaÎd!Ïe0g/çÞhæxâ›"·¬ùZ›W×LcKÍçÄ&¡Û´µÖ¶d1Ó=Ê”òùOOå7lÀþFÉ}UCæ¿ÕÖ7µE7éF®ßÉ1šFË7¨GæyAòDîsWFèdOeé¶Ê_ævb*ý¢õ“èÄ3×£Ÿ‹t\Á®&JÌD1~@àZY°: rGúø2.X² °ý£b•‚ß&S¼>ßÀÄ?hÐðÛ{XÔëï\#[)ä; ÐîP˜o*Õ’!¨¸ÑŽÎÐfÎM§…MÞ/™?ï™’¦èwF%ìrZÄݧƒÙ‹Ì{ÄÙ¯&y—çg8ƒ|,.¾?ë_Ÿ)cÀ÷ôÂX¨¸:¿¸)€¾”©=·gßâ(½IjaŒëÒ8jì±^£f?€±Vü²­`N'ŒÈÀ|™unºry¶V(Yoå`ûnUãâ=ëÎW:oô ô‚ƒHZ¶&KvÔòç!Uù¨ù>Š€Ø{»UtùÌÂpÑ ='…"™yðªÊ<°£!uÏ[DÈ0¸üˆœjõìü呇5AØ'<›®Üo…rŒÒZéªÝ7ûVÌTòÑE©K6³¸£:„´£î ãx´-D3Û^­f±2KAð¢qçºI~É"*ihH-S >ˆ(Eê¦0¡sáUë$Ú'3“*_„ö‚ —ÿIw¶ƒX=·«¼©™U»Û°s:ŒÊÂ5©10=ª(ÄìÂ3°’ÜÆéÖŒá-ÚÂË+¤ ”èž ê|ñhݼôªÂ;Sc¡"Ù²Câ,Ö6gÇõLl@œ8xDÄÉ`’¬øºCIÄCò€ŸÍ€CSdg€¨KNÑ!4UÙ0 ÷M]¨ Ö'>CQñT¡ ‹aX¼sÍXvƒ‡£@ ×nPÿ 39"ã^íp–åwßñIg‰aš1ÁŠ9Œ·=—(KÿcÙ®#^ ü,¯‡NˆG˜Îgs$Phn.D°šiÏ‘üÛU¤˜㥒¤ç®6l{V3r&‹ nÜ< ·"¸BÍ\§õå€ ó'/h`g ͱÒ{V1t°¹BPï s †i³²-›@LÏ]‡OîªÞ÷¬Pn,ÛìŽ|…ŸÐlǃ™®M¨óÉKƶWJ%ƒ©Àþ–¾çš¹.Ä|ˆ/Éa_öæ›ñƒ¾¾9Ÿ vô]ÞÌm‡¹Ï6yj÷7´EÌ$‚oK„Õ†¨7sƒç’ÁÔ’†'v•¢¿˜x×à “lðÕº„ò÷A*+TÌÏÂËtÝÀ=úþåIXäã¤îž@icÙ Þ=e¨ËQ,µúÕQ^ã6XÕÍö@½4 y%PJãîÍyÒÄšiçïE “aÙ*ThŠÝùÖÒ¥}éãÀÍÌí~ØÞ0Xzz.ÝV§íFxzú{éíþ5‚ðV×ëÃÍ׿•¾M!Xc$ÐæQhÑSö¹ÓW„/}°Peô}íü ÿ 6%9HÉí-“R}ù³ÚùÇ“â¤qo„¾o>cª`ßÉIü\½ô o Ÿü0{>a#¡`r-ôÝ’ÿUù'•/ëKVè'ŽWþDã§O‰(*¼6 zc'ôýÕ^k?FÝ?‹¼ðQõ/L“WÂ+wÂf*-wZÏ?~–OÔýO·“Wë^ʧ.˜=¬5”eøé÷/Ì h™ëÚºh|­î÷»×Â,+ø‰€4*¥zD+  Ï?®IýTÿJRøÑó²…C}mEé±’S§ø“âcË(ÈÔÒ:ó²>ú(6TJzög®§QõôX¶F.„“ÓDáwþ÷‹OÌ[[wfŸVKñ]*ÈÞ×Ö“ðÆŽõÕΘù™„ÑY„yÄ/Æ¿£?”ºOüe•¯äÕþ(Hø@>.§B“·ŸþúíWØ^ŠÚ/Ä™1Æé¡’§ß5Øÿâ¥ße)qhäͼ¨Ñ1ƒã ê_{£‡¨Ê—•töËc¥Þ©-e¶û©þÇd2°ÿçr9yC ÿ Þ÷—Ȧš°9=MÓ*·¸X넟ag  †ž¿/Ý8ûq†_Ðendstream endobj 403 0 obj << /Filter /FlateDecode /Length 5078 >> stream xœÅb·ŽÑ Ýi¢лlà °» y3U0ʹå&÷AÙå-ñ1Xsw;ŒÆš!šå%Û.”ËWe™àûêjù0ôr‹¿Æn'„†ó¦n8— Ë;ìV1DŸÎè¤îª®—ïû¼Ç„­´ƒíÃI‚öQ/ïû€›ŒÓtÁ£ŠÊd<ædoÿ÷ò?‘ŒF2*'〒/¯€n¿ZÉIˆè i=´Ö¶µ>µÖÿ.{sµ*+ Œ¢ã$‘1ʯñÁx“i#(.39”ˆ2SÁ¢YÔÃãK ۑщx j:â¾#sÀ .íeAJ^£lYh²e'«U¨[~hó¶8/íèŸp”ÒX yã, ×êèà6—«Ê¿?®¬pæ†ìn9­ÖFÆ)Hµ|± jR8h—"¸DÔµöz²p„µV°Væ… á\X~LÈÉ‚Uï í ïܶ¡3¿Y­µ²p<¿ü8¨m|Æv>þSÂJQEÆÓÊDsX¬Ó#-²|åÒ,§ÒJH Îøç8Y>º D†pR–9 ½Dñm‚\D^H_$> À»rýOy9wX„'F\ujn-g ÓóÇ3ìi+à”ßä£yÐ8}»3Õ7ãÚQ”Q‰u6»j‚yõœ²ìû D´vù¼TvwÛÎ!Û.š›"öD½]g”òRƒËŸVÊÃÉáì(n‰Dݰ™Á{Ý0ï¡áUæØ½2±™6˜0è:廙ÎÒ&@©ÒçŠmsIÄ'3¶òÊ ÖÄѤ!áµÌÚdvš¨tùàÈmAÌŒ‰]« í°ò¶µ>µÖ»Öú–i=6º¿o­;¦µi­w­uÛZÊõ–úáåÅ_.²÷‹\¼;ârø°Vð:P— ž1 ç,`ÓgC'¼Êe «|®3£@ó+¥´œ@T3èÍ#ÒGAõAk’¾¶màˬ.‘Û#ðZØgÀ bDLœà¤œ ôÆ’+â\yUø*Ryd\2[m»%º”ê¿«ºœ9°•‚j¬Àô½©ó1f·Ä6fß Þ*tO}teî@²S0 oAÆ`M-ufò߬ÖV©äíü®qï†á÷§ÖÚµÖ}kM­õ}ký¶µþÄŒ»j­×} ÿÞx{];åÏS»¿cD­|j­]kÝ·Ö_!îbú#œò7}ÿ·¯KFÜoŽª…kf×OÌW²ò†ÁÒ%92ò 0褵¯Jï ƒ‘-û\öw‰Ù½ÔZ_/™‡ºÇlŒÏ[ýõj-•õö3ôM}Ë´¾ž¢&¸zq”œo[ë»3·Eøè\Щݑq‘ ®JÞŽŸ¤ ýÇ‹—¿úÛRf1 <ªfK •Z:@ºDÄ»÷îÒ`!¥Ì}éÞöîMñh­ ËtR”TºèjÐÌ,ï!òÒb^á“ò´Âð!6ÉÜmãò~G«hò:Á UGÞ¬ ÕìY´Cï6&ûJà{K!à‚›((n¤HÍM»L&œgª= Ó6€% Jï¦õ>ôidÀu0‘Þ§Ö»éÓ^±«]“^`Ú tnýœë¹.‡Z˧„ïÁ mq°…£—l½qÍ?\ Qò;¯2'ì w}÷j$Ep³)Œ;„B06d"±º£.¿ç-E¯!'3hlõ¦''Æp¥lUÃhàÿ dÎH ˆ‰G²á=è:p“‰ÂºÅîo^u7"Ó ãò@ˆ ôˆ“ße–iïcUf…ÆR(!>ß&»w˜nÕÉͶF­Éè †dùǨ¶É*ñ̨†N~§š`Þ†8ÿWÅGa(î—µ²Hrð9Hinr’ìÉÁc¤CÃh ‹pMŒHäY G£WG¢ˆ]ÁÈ©*žaª„ö 1> fT‘5žšµV‚£F Q‰–€øõÄ驱ààE@uFæ=e¿Ê °4}\i5‚¨ÉªºLOJ‹¢Ën(»e­±„¸pÿêHN•Ϧ—FT5aÆÑ¢H6u³ÿ2 æ\dR+—qéV/ÿ¾/Q=ªÏ(›^Ð}·g¥yJ9)E§p\‡Êú̧onñ]ž*o”“÷Ýi)W€Ô5!F‚­ägÀ é$EûŸgî¨FÂojDzAÍÑ=“» $®'{£BIìÚ«¹1Kžõaç£ÖòRn¯àù` ’Qîr‚B \§¤ “4l<qp¥`û)@I½`4ÀŒœªª °M®9VÕÔ^Qí—½yÉÔŽ¯™šÛùÚ&?2å·»ÃG7òŽqF©CÉ”Ú6¬_ÚöÒÇ}ï|}f-â=󵂺‡ý‰½c¹Kš£¥rã7ò:¾?Î«ŽØ¹æ¨UÜllþ‚)ivÂm™e~ÁÕƒzç \щsÿ;”âýÏNÝ7vÃ1Ê—œZ¶– ³÷¿Ff“ìùUk‰ÖZ3_—ó>¤S¨OÜß¡9 Ð|>@{ ý|€ŽLöm´ly“53ƒ]˜h]§wq_uæî5ü÷­Õ9ˆÓh·Œ»ÿ\ž÷ÝE£«†,î²WRQ'[a ±8 ÿZªâá»êÞRÎ.ˆWï t$ÞõÉf>?2-ž L§i§=Mm~5?ï–÷Èy4?O¡tzÅЙS#Ä:å3\À¥Šß3ˆx"‡ž Æìü$˜¸cPñck7ï–`ŠéäDgG}ÈvœÎóÝ£øÀ`€xíç§´™ÃöýÈ„ýÈœ¡c¥»‡„°‡5Ëa*µþoòÖ }þõ)sÀüzzž°é¹œý³™ã3 ·¢Õ‘¸¬îÝŒ #Ž˜ À©)GµÏ­p2T}B˜ðâ$?9o³=:A8NëwQé_cŸr~øÃˆá©§M9?„iSÎRÚwt =’Éï¦aÆH]‰h¦¿BèÑÔ¶›§Bûl¼· ¸«µÇKD1+” {Ø#ßL?v$Ÿù΃Pd ÝdžÝ̸cp}ÅÀá dßl èØ?Á±|3oZöÔÆ›ÕÚ—Àñøö~‹go=r®„ÒíÕ«}Á ³æËUy€ëêÁ¹£¨½cʳð~ =–CÛöù¯"âÿ;‹ã–`åØ©Ý‘o¬€9¢&bU§¹QêñȦ%Åþ„cô9FWÅÞè¼J Þ•wc!1™¶X;ÑÃR»Ã[â{¥»RÍ…„tϑԅ8£ëÒ[®–¸âó—Üæ%¾12³Wtõâ‹¶xßдÒ݉K]ËC‚­µŒw=[iã¼;]NOá@õ‡½‚"ñ†CËð<¿WÈáÔÂùÐ>ó/+N± 2ÞÑ£%ÕrÛcöž|(kwu,^³Dþ!—ñÓ‚÷…,ãåêñµ˜õé"dž^oºyß/j¨¨£µï…ZøT—\ £uÛr@7öŽÄÝáR°^ma 7·ú}Þßd¥xoaîMí±XIþšÓ»Yì¹ý,†Å4|¯‹ÞÜv‰™Ýˆî×õûëºC—|sÉ•²Õfv7Zã¥/c×[Dõ f+nóð€×ð‡»Â¥|}èFt»¤;.Mï !\æ¥kÍòø³ÈϳɅ%²r½‚@–­ Üô²0E ž¤×Ê•¬"ñ®¼¾•ƒT×FÖÜËÆSÅV'g}8ü„‘­¶v6Ü2­ÎÂ=H£sËiô)! ÖZ“«àqOû¸žQøª©_&[³£êÏ:^±)u"Åó\Ô¬­gºvÙ'wOëÎYLI'¡ðQm}ÇhY¶ìu<Ûp²rº_Fü™?·û•Ógx<Ëq¼\úó¥«8y×\Ô“¥Ø‘›ÖuúA®úR‹}X/ÔVÏûuíÂrd¼}ªy&6ŸàçƒÙ6{X›ýÓ²™ª>+›ž!ù¹É/H4žÊÊbÖðüTa·h'sÖC´!ø”mMµu,÷ˆÞûɼƒ)˜ór\–z?çx,›7ßÌ,3Ÿ!3öLH® Ãþ­ŸÂì!ÙÏøó™‘\“XŠÙÌþ71ÛL¾µ.ñ lXH!d·‹» ¢Ã(«=G/¤±°Sô~¼®â¤´†ÁÜC~.&¿Ã¨7i…_x˜Œ·…ñÞ­0!n0šÂ;º 7ÿS£…±GŠ”DîorÀ!d~(i€6¿à C…aÄm^0^¿¼úê]h²H’ó"3ÑÕ˜£•üÖ,„öÖ$šè¹cF•zø)6K‡– Þ“öe¾¹¥Èwˆƒ°N•ËÛV.ÿ¦@ŸŠ ¤3·dø† Ç0ÀÆ…0tM¦>fˆ"„|‹6ã“®ÿDF¿©ë§ØÔ"–G|˜òÊ?åB²9goÅ89[Îiþ©·À~—t~oSôhiƒÂ¤°rW:ùá#¬ÄMaoÙø88¾A•8ÏtŸYs‰‹Á.'ëS¤\†ÄÓeÇ7¾82Âhˆ®¨¡zˆù½“€¥ó‹ /ŒÕ!ó ËÏÇZN|ºv‰?'¹üê« [‹ü£HÂl©÷EªäP ÃH©s© Á·~ ¢(´.Û~_ßGL–C^“!÷+˜èBæ´Du%1á—H-ð‡¹n)I·dSÛ<¤¾0 “¶$-ʈJi_Þr”Å1›â•ŠÒÓ=•g–éáæb?àˆHì¿ùW/?Gó Èò;£‚Š-U@˜X Ñ;Jì ¤ðÜÖ`¢x†R$¹.êãÕ¡¥0u”ÖY®Ë{ 7JÞîÐÔ,dÂÛœu²™î÷6“øÊ•Q®Å%ص§¿ð5KÊ/ ûÒËûå+€J!çÔ‹:¤H( ¼)‘¥û.6j4xÀ7Ñ”t›.ÊYÌšR¨v¢‹Ã¢”§}QœýHeÄç “|²xK)ðêd2͹øHšèöendstream endobj 404 0 obj << /Filter /FlateDecode /Length 5429 >> stream xœÕ\Y¹‘ö±÷Oœe«ryƒÝ4†Ç˜…×X[ý``vJ}©½Ý]­>4Ò‹ûF™d0‹YÝ-iö€Äf1É`_dæû•åJà¿üÿÉõÑ?ý%†ÕÅý‘X]½?’ôë*ÿwr½úîFHWRŽÑZµ:>?JÏÊ•3£¶+ìhüp}ôã Öb4Ñ)í‡q½Áv R~½‘R™Ñ¿Ã^+„VjØÑ`!aÀÍZ‹1«†ØRÁ…á ~öB[i†;öÔt[©¬7Ã%ë¾¹Xo´7£ qØÂ)¶q¸ÎTh!i=+¥µ:Â…Ρ[KcBhÖIË;ëýð»½ÆØí¢R–<ù”fqð(‘%L×{5›Ñ ÷éÁ 5}sšºtŸb¿ˆ2D;›œ¶‚ YwÈÊà„ò{›/‹C[ !£ÆÕÅ„°Òþtüo(`%Wq¹9’oƒÖŽóãðF¥F#ÜðóZ¹Ñi/6è`âðk\0h@vð³ ÎÏàVa×Ò X—h0B¸,]\”ÀG"3Féìð 4fôÑâNq ¸¸¤5`¸ÁEü¨E|º›4ØF?œ$Ò”vÃî)ÒÁ'Ja$œ&Ó(ºAÙ_$ÅhLîŽÆ³ÕîO¸ˆqƒØ ºä°}‹Ìâaí«:5cŸc{“&q°·ë ìÅ ˜„FYª3»ÓtK¤FgäÒˆûÄP3É™y b抪ùý4-í„lf{_YþX¸ÄX÷amí(£#%÷1 Rž­Q¥´¨ǧ ]¯2c}Ðɦó\…U‰DÁN‘ÿˆÝ*`=Y—ðdv— #8½Äš8·ÉúØsvü¹k4 ºå¾fý " o WqU1›‰›„Ÿ2ñ#Ï:J+ ïDÕhèŸ S3ÃTh÷×™«I¬Šo=3ŽˆKªŒN5 â‘\Ru­¥Ojší)É LÝr™ïÒh¸DÚkD,°ÐšdÒR£ii+A×øØ{¦xi=e¡û‚­Ç‡Ÿ¦ù,$o ¼Š{] ¡lñ[œÃ¢€ ò" Z£”l³Üj«ñ1µÝ²Ÿcµ¸x4ZÉ·N:,½5¦o2ÛE±”}¨ €Q –ãuPVþQ™-DMK£ýþøèÏG)<çàƒ\݈ Z—‘C`_Ö¯bƒ“ r}ü·iî—Ï€–f¾¹GyW7¾C†nHPG0‹E0EI¥l¬xw¼!*bÑó®(ç–)Çôø°:næv丸õÃ[2°î´`À “TXkR¨“êa‹“÷+*Äq+iˆ…0bÁÏá j„¸,ر"s ™ÛtÏL³&ª¡<帯ÝÕå¶LKw8DŠ:ÀÝÊun9Œ¢±úä£õcðzØ:ÂÒ>qñ9ʞȗߢ‰èMµ>µ#_,ÀÐ&’ë»ÏO‰0GØ•—¡A ”!¢kB¡VNú˜Ó•Óöºz¹$¢‚8­œ¥É"€ös܇¡Lº"½Æ-ƒH÷í<¸ FMæŽèqÇ·‘möì[ˆ¼Ø1(^À_ 1pNó ·ëa@{Xý hóÀ܉^–C„³º>‚˜¶ì§Ž«£7ˈdT‘„¥ç) H€t¼ŽåàIÝA£làA ' S‰Ëy¼FR#t©ÞÅ:|séc\Æ: q1x²:XP…¼ ŒÑƒ6§]Œ6GTÜÏ&“ï… >‚e…8… ¿c[ØÞ“®ðù$¨j…DÙG­€|¤NÊ™ƒ,j÷ž…{3gσñäBgÁxÒ{€ Ð5;ê ÉÌ£&(0™Q 3—“^JŒ?™Ã(zV”Ë›‰%×®ÜsP½ºO `›—+ëðuÒ¯ˡİ P-›+Cô°»mt•^pªlý¤ ‹Ò=”Ö¶´Þ–Ö}iýçPš/­×¥õÛÒú®>²F…† 4†'–¹*­³Þ‚u™Wýezú¿Iž[`Р'Ze¶]S¢¬M`”˜x+µ)?-˜4g£4 ‘ÓËÅ¿íä{ñT ÕJ”W®)²áYœô¨ö¹œ–ý0žáH€õ(S° 8ŽbÜDÒƒ{¾œ9âuôâùI™¢Ú³›Yi?‘<¹ì´òÔMǵD<¸øÚ7ñ›UŸnÒá²À†ËÎj¡ø4…HåmÙÆ§´0ÈtOY’žÎ ä{ÁzÜñd~wÇ]ý ¹köqÓWŸýÓg ‚Ãí¨’³oYÝîbnÂÒÐAlÕöli^q~à‘—Æ2¼i*ágµÁ}ô6UqÐg£"ƒØÃ׎™„„ŒmÐÐcüôô¥«Å)SŸñ]i•Öû.Ã(=K0ÇT†0$ùVO˧d—ÌNîS̵ŸÕŒ›è‚U¹Sz×xšI)Ÿ¬ð'HršÏ´c„n!G¨QjצÑK1ÉiôB¬¡~ŒnBZ²§ÀGˆ  “1 ŽÀMÈŽ¤† JÚÒƒÙ‘)”±JŽÖ ©Ã¹8Y.ÙšN$Ù¶CCö4 Ÿƒ”ß“ahAiq‰û‹ðÚè–d%`ÉyIXšÔZ §UŽ––\ïE)¢¿˜«zX”¢% ®gQŠÆÚ“ý’(EË~ÄýߌRê~KÛpa8©2Jåç¨çó¢ÈذåK¢Ƹ—ŒP@Àà'6½^_ŒFÞuê:—ø¥öÝ÷+<@»u%¬nê¥åúÑ6Ý`‚œwÀfÝüÆ}M+ÎÒtcéÜG(Š’6ì¾-ˆelÆAo08AAdP¯nÓ,Ò½)H"exf´¥"ÝØ{¢*/±’S¸sÜazM§{%¹ÞA?œÖts‘” ûjU¨K Ì¥T¡ðMg½ÝUð\›g®B·YÁbÁyëQi‘÷Œ¾íäÈ»ÒúÐÁ÷Ó®b¼¬¢xÞ Ð‹·Q‡?íRXëLÂd¼Ñ®K=®“ çßùáyÒQ0eÉG<•åÐ]ë1?»¼œY(¶ª…>#·…ä„mvã{7"¡UÙ+íX½5_î·Í7þ+$nØèˆç7âTä«¶ˆï ÇBùöRé†ßBÜ/ @¸c k­ }¬M­_•Ö/f§­à;çÇ~嬉|e ¶ù°ˆî±ëæM•°½?Õ³&/WŸÛ+¡éÌtuóJð6—A½Å× Òͺ Û;–=¹jœÁʪ³mèÔ¦'W‡3Ÿç—|¾îédï.ð>£WÙ÷Û—ž.5)gûî’òz&¦‰YÚùº’•c§3µßÿ\:7¥õPšq9Òh¹;8uí»/­W­r¡¤2µž¸lõõ zj1 û¾³3ð‡Î*· kþ©7+V!Ú{3íõ0„™«ØLO% ž®#3eªé|U—ªVžֵò;}+ ª $ñ 2N¥¼¹zz4ü¯æïŹ#¾ ³ØzÛ;z˦¢Bƈþz€ÊÖÝ´¶ßbQ?²Œmj}[Z5u|Uù¢ºˆû§ª_²tÆÚéz¿‹Úé»M[›l¬zbZÖYO’y‹#4pK3ײé×:­=E~ü¦Ç¸oJg­,T9^w$z¸äVñ©Îœ ›ÿRZ<n•KÕÒÏkgªª»hÌ·B|mýªrî¸c©gÝ=ÜçMgÜ·•­']ˆ±)ü4²á?:È^aº^nªX}Aeæ7•™uê—û‚Ð~S+qýæJÜRyðhÚýGˆ˜¤/¯þ§Xû}g_×ý÷\j—¡ÿþì‡;ÅË'X6ÏS©¬²djíJëCG‘O+¾“õa&×£ñÞaú }Óy¸<[UÑÿëJ½ ϨÈ5[÷"ˆZu}¾3.ÀùÜ8òÅñÅ,÷£7øöJ³‰:øEúõ”à;Ñb_Èeàç”ükÙñìrwëÚZJÏå^wÆuƒŸ7Ö|¬…k0gWR?Z½º>RÚâ»Ç¥ç`í5ˆ•…üË·W-(¦U0ÞPìvŸ„ÜÔMåáWžÅªÓùfq2@8LWÜCÖƒïòC:ˆ@âÛR†’ï ¥u©¤MwÚó=ƒsù%)Xm¾çGuߺtdKãM _ÒQz;š¾ü {tRM3°ƒa¶ ^à°èMߌ ó×”ØE¼{¬èb± B]¶#E~È+/ca÷¦vÓtøpÇW¼ì﯅€–=Q¾8d»ä)cÛ«L6}$£LñŽ®ú¨8]]élò2Iæs¼{[÷~–Ìs+éÕ$)<•¦á^:ÓU²<ocØâ‹—T5†UTœä ÌT.UJònvëü~>ÏstK”¼›QâBT*ß0Ê3ž–EϧSqzª ¸Kô•Mæe8Ug™=¡™ä1åR¼yš}—|óœFvVÞ ˜ò*;'JŠÆmç¾Hal·>ãŽÑ&Zò-Â=§Â¾,·:ï«VŸ±!|–‡uþöIh¨}ËÚWuŸûc½Ç˜Öª½Ì+R™½9Îm.FãÁóºÑyzÌî!\Ã|¼fé›!\ËJº «ð0͉eѨôù“,„´Á wOm0NxÅ΃¸Álð,G× ´ >1y —>@# qŸ¤J¤\q´hþØ%Trë¤Z¸QGÏírMŽ–•ÐOÙBŒ]M÷ANH|QÒ©Æò“/.hòlv+yûé– W2] Êà‡fÞâç‹S8ìò†•I‡­9ÊŠÄL%ý¨0„ãËñ,Í£»¸É«ûØÈÈ‹ ©ôœMþö‡§tˆÆX7£c†'ÜZÁEOw?ëÜHvT|ëœì›¼u:áKUlÜå£OÔ é£ iºxUÞ¥ ÜÏíE^†»”ü¤ôAšˆÀê¸ãÊ É/B1´m''ø†ý‹z á(µ›2AödN-ŽÝñ©'š¤mÀ¾¨Nz C‰êÖí’WzläXœ"7ý±î2*Ø@óM3ÂMÁxQy|!("CU‘ Æc3™àd²}D"ñó,­|H©DÆ7vÑÀÆ •hvÛ¨d" üë‚Hïâsb¬f’]yræG_v&ÑÂMyãï8ÀXAïi篞Uì9!º»ü‹—ÄF¯Ú5HŒÂ íe»ûÁò~˜Ø' ¢ñÞ÷°ô!]:ôÊ/F 'Õ—qšü ḩŸ¤ rÐÁ»“wLã»É9™bÂËa²î@”7 ˆi4À×m+ÛùÊè[²lµá©Cär4Á]`ÄÙ†BõQFbOr £<{,¤@ðKäöd<¾ËÕÓã4ænÉqaòÝŒô!}‚,—N<†”Ö[Žÿ4RÒ$œ‡yí2_2Ôo¾ªW‘ôQ2I×:Œ™¦Aõ.d ;²›•v 5›rEkÛ§,¡¹$F%4·üÕóì Ê9.nl¦¤ûŽB„(¬\J3rXàŒïEÞó”rJ4e¤o-9Ð ß1 nìÈ&ï곦JC¹ÁQè²%Î4¹h€Õ,m/5Éz›‹eôÜ´@}Èù#]J)º êS)exØÑl™O¸g¡ï=ïì°Œ˜,·/b¢óèÞJmc?5ATèd&9"e&ÏÄ^ü|бŠøŽRÌàñ ›KA31`Á»,ïÙŽÊ;FñMJÛUT{ñ*¥ûž•Ê(×5 MŽŸ³dÎIÝC¶ºá£ü_ŠÙq‡ ŒìóÔë…?€«©žÑÄ“¤%¤Ù–+øÎ;¹uûêqаVÈÌ> stream xœå<Ûr\7rUydüS©Ty&¥9ÆýÂMRe»¤7+Åk±öe“‡oÒšäP"%‹ëŸO7€ƒnœÁ’’wKñ( .Fß»qÞ.Ä ÿÊÿã˃¯~ˆaq~s çodúuQþ_.¾9‚ÒˆE¢Snqtv§Ê…”a&,\°ƒ jqtyð§å÷+9Œ\^ÔÖ¦¶Žkë´¶^ÕÖvµ–ÖËå³Î¨ËÎjlÜÚ©à—2õDg–qµV"š!Æ¥_­µ4zréVÿuô»«%t\¬¥¢µrqt?¿o¯5¬Ø&‚6´‰L›¬'»Ä¼ËQ]ñ]ç”´ómçדÏÂOu³…¦ƒKlÎôy( 6" ÏsB-bHûE)]^ÐÖy0@y… ^ÆŒ r"„Çë€3NÆuýrsE{ðÑo -ïërì VÖÎ#ašFçM¢°!ÊÁ<—åp1ðbóŠ Itµ4Hƒ]‚ÞE ]ÀH´ ‹ ô o$ç NÙ7ãze‘MÅe³ˆÙa wy¢—í5ŽC‘“ðwä¤>—×s›‘ȱ7_”óÒ—kH ñ–1Õ{:‡'£H²e«»¼K¨ØM`<1C"låà­ ùÙ*èeØ–ŽÊhô#É¢ ãÐDhFÁ½†²Ô$°ÓÌ'°^§CêºeÂ'\µ‹Žî1¼â%=î°ï.©¹ê2° ¨*3ž|ŸéÐ3Î3w85ÓÎ:í–Që¾¢õ2ÿl&êõšëyè",$8gl]Î#4—‡ pdå<Ï¿¦½dŒÆÚ=¬h Mëô¸ÊöŠSÌ  Z¥J4§à.¡¹½ÆCgp[³Ú”3aÐuJd É+"ؼM£o¥P™eû<ú†qÛqc´€ÐÕL™§­ÜŽWÈ(k#@Ó‚è~ŽK‹à½B‘Tù8Ãç›§kt™ÈAÅKý„}p ÝJ4vпd0 ,pÍ‘¡s‹kÄ ÈǪi½°iYÂ@2F\“ÜÙ&cGæ&ó fªŸ«éñ¨Š ˆB&§• ?ÖÖu‡[¯j뤣¨O:*ûöœ¾ÖÎví„áäà*á†l«¸˜.Àß‘`~Hê]ãu‘XðçM›¥¨to8Ñ´G]èbôg¥74¯Ñ©Z€1 ·Â„7#ü›¤©­ˆ:Y[°[LZ0Þ¤‹J‡ ²¹í¢Ö=ÒÎhvÎpmßàí¸Á¦šÔó*ˆÛŒi4ðÃò3‰ 0 H½;°C»óÈDL}lËΠvû¬Îr£-ª*#ÞktÊd¯ftǬlrI£[c³ÃTìÏׄv&nÊÙ£m8šÙWLr=DÕÌÓnÒšªVÞÉ çµ¼ª¦L§Q»†èß³I5D­eËzèÞh°¨™¥{úÒJ4H+.‰Ž›•«ùü”Hm À˜U;.‚–-Æeaê ¶Èh¥bWÍ‚‘@¶Ç@̃Ÿüá@!/~:‹ßH Z9ª…õœŽ°¸<™ÇÔµçâà%ŽRàr-¬3nPG=œñ2j>ÈczA ž†°f,˜^)Èó4‰ ÒAìË$äåªÞØ5ë¾HºÓiU,ÊìÝöÐct"Ä?‡„ÙŠlH( ãŽ=ì¸uÎ0l8êJÏ^¤4@U¤„0€[Ý"å[æ®o³¤BùBŒb :3ö 6­'¬AµU͵¯;Ò…äÑëίo:ã¨ï¦+… èxe Af¯UÁ¼ÉÒMí5è1õ@Eh‚ÅÝ‹ ×rD;`ܼ‡)XÇ‘àÞ¨a¯0ÕE pŠ5J€)R/öUß²Ogó¡Q´Û<+°@W½c ؉ø«) ¸Q°TÆí['g¼Y®=&Ú·Uh,:4Ëô ´TÜ#Å\"É’t@<é#B‚Q¡âŒúD¸7ÿ•4À›è lJ1ÆËþuµ¶à¡¹øu',O!ø…àÇÖMþçÚ¹¦N ͬ­ëNüür|>OœMܧUüýÁÑ?5xºé`‡0VàÁæ§ ´.3:q`TqÏrLüþ`ùåêË]M8`¨C¶‚q.«²éÀrÖ™ÆQ†å´ ^c™€GîB5³iÀÅ—ËPÑoFDKÆt;7ì>;"Á@O~©é÷ ï)ÿ9\tø‘€8,y.0tPíµþ?Í<>åuÙwÕwH›<ë,HQA¢ú-CBÅÑOµó+êTÔìá¡'«.–êä¨ßwÚ; ¾êÀMë=éôu.y¿â¢vw½C:ˆ|TiC+°“«Š|ÜüÚ$ÙÅ[Ù|Èë·YFy/;tôqÌGªÿmäó¬3å²sµ½íöÍóG¯òÉdñ¹Íyj¨â'+-Fßu ©bJÛÚúÐ!œZð¯J8ú'œÞõ’tøñ’ G8/:ç»ê¬×[¥w“‡]þùäf’AM¿jÜ;kžÀ@›Ë¸Èd¦rTüàÛJÅ>Œ[€k*_ûl—Z0—•5rJ0`­µ·0F/·88À—&F ü#%V «Ç=4è$½<ø±ÁLº¡) EoÆhó‚e÷² ²¾™9cþkðèFFúŽÅoÀÄ06ܺ¦f¡ú_’£ÿ]p Sga-síX‹õQfƒò$“@®å¤äðmîÅÔsòXv\%1[¥ïœƒèy58k"9› LÅ)wytÓà2öz™RX&UÈxžiJ^§–ÚùfCÂÞfç×Zçs@-B.(ËB,Mì;{·ByÇsͦ9ÌGÊ&=2µ ”€ƒ‚D$ñ`ùî5 8gf>Ó7¢é –‘º è&Yå}n° «ª^émG’öü-n*õSJ _5ZÆÿíj00ŽomÂf<,ÖYÙ¥Y³Â§Ä<YcSê. rš¢™SÊ œæ=4’çYnÛÈYŽÑÛ_ØËøeŸë ',µˆ¬£ïZ™¿Öì7勌pº”œíÏ/Ã`­)‡ùWÈ/g³üþôò“œñ¡Ù:cnDøiÅ3ôµìã¥4/&^ÂTŒi8~ËU£lnÝo°Ñá=| DÍò)8H>|/Ú¼?&Ì#ú»q‚^ðò¤³ÛþX{Ϲg}гÞuúzá;Šõr}oº1ùêû™{dàóþa{Ü–êU9ëˆËjÕïÛ¤W¢FGÿ±óëyçèÿÒÙãCg=‚ïé^´õòµ´2<¬™äÖˆ›6BÊØåõ‰šev~{R[–6YuÞE¿¶ÏC¸9UOCª×©{¦×™ýzܹ± %¥YÓt½Ìš¢ßÕÖ•ÐÐ }ê&ôHqCæ ¨º€¥NûIçêÛÈ™*öJ¡ºÎ•êIÁÉ\ˆUO)VAú²T¹ ×Ô¾âK!›/˜UœMZh·öÿ“q´[þ¹*;^ž–U(y™Š2qlŒÜ¼Cì±B…Å Ð(×ö§sz¹ãÏR?3…´K̵n«?ƺ´lt}ßmb­'“!Ì×`ÖRû, "NyW,ÿ× AÿmmÖ™Ú?túªo¿êüú9 u†ŠÒU95òñ?I˜Tª{ÂX¬D4­ô³R \2~?…º&5Àb©\´±¤Û0ÇvÿÉÏö2PÓ' Ë ŒôýºkÖÉbEo’ÏTŒY¬ œd2Ä3°›‰!³èIóª‘\§T׈2¸‰A5´_®Ö)]‚ˆp¦É#‹-{¦úx%mÑìîÛcO«ŒË³™ôÌå‚9%ðiñígOÜ”u£]Î?ÏÁ‹ç~œ²npüa@0‚ÅbG¾ª5ëwyoŒ üŽ!¯‰Ï–ߟÖ3ó(|Ž +gtE¾¡î-½¥H„Nª­äpÒ†;o‘0n§ï”Y¸™ƒò]35Sªñ0V#“çµ®Oi\‚Šëaþj÷4ÉEŸD¢½aÑo}eSVÀ,Ê´<={OÛä8(Á©r/¸¸ó“7O,¶½¦Ðlÿ‘pû^Ö`¬2dN1.¥¸ZNAà½Ç€-{HrLG=éžsXÃ*tVF‚ §”K˜ÔÕ¦âëü°­O1 Ä&?ÚÂÄŒ4³üwŽì&«ØóÜÉÃé2à]^ ¥µ‹dÅŽýØ£Z|pf|{Ô›.÷=‡c•Ô‰M@[wú"ŸQlû.¼Ú½ï)¾Ä)!?nƒMƒb±î"É£°Ó¤«ré)ùì3)åð©˜Z¶|\)}F.ŒF¹b{Ňï-¢ÆÁŒè[Ë þE~’™m¬—“?ÓÍçãQ„oЕ<(WigugÚEc6ú,/â„Û¦(ÿ±l>}µÇ“6õ½þ$û½ûlà&ïgvÓLãC½äjD]$ÔxA¹v]K|5R•§}övhsÏ3Ãü@XŸHþ*ú„Þ„{eàKv;étp#Î.ÛpmB€‹ˆZöu€“.ë²'ŒY½•“¶|3y&Ž¥Ÿ%ÒÓ<5Ý –àõæìå¯J¸Ê3l6ÏÉ÷ßÕqÒKj<îK(L\«ä%ïá -cªt™I9ßæ%¢PmBy|×úw\1õ:‚šgŠ¦Î¦LTߦóØ&3ú•Ô ?ÝЩŽøµl˜kž#ñ‰Ì·´šxG€‰G¯Ðù[‡J*œ:'obú£äÏÍwl%’wð┑þÎ'@Š!\¿ WGdï•[>é=Ü”ÂÊrÄ–iøØöS'<ýON“Vü!tç=FŽßhtÿçâbÈ ‘‹ÝdäÐi_ÈÖ+ÈÕ6R)=cû1B>ƒIrù¢$/RM5½ébB†“^•çðÁO„w‰QÅ¥+X<\2(u;y¿0‰síñNÚ/j°r€†¯ÒÐéç4 Ñåçp×¾¯F”‹½ÉÑÁ´OØYö±µ-Xúº7EÞ$Z¦8_óeš6©¼ë€°è“<¤Ñ––ãž=W83é°Ö­W¿X§»Á±-m9ëŤ[c²dÎT/'b̾½šŠ¶lÉôVøùÙ*(p½ÛC2–èC@’ö7YJ­}ï¾ÛHúX+ÅÃõóû–r\Ow.<¿-üÕ.ü-] yßsŸ‹`2¨ˆQº›VÝ27Ê ¶ÚãA_³*8 ‡B|OK?Å›ï®Qr¬òò¼jà{qÈ"8–¯õEŒ²Ccî'¶]á¢ÿ¿ —¯¿ÉN™¥ñmòG¤×Šð§Ëß¼LØÂzpùâ›U Íói53¢Ø÷ƒª¨øÈDEiwør 2„ù1דl,&v‹ZYÊZ9l|SÓQ ßX#u+^8rܸNõ˜=ù( FW±õ Ó þŸ¢àØ.›Æg¨ÖCßfRjZ.%(»aôÑ…É7‡x)cŽOðŒž>ÁnäeEθ£t;!ŠÚcÝŒlÌÍD 9l7ó®tG­'WGä¸x%š1JËÿ`û÷«š€x’Ë"¢+·z)ôM¢û¿S×5Ö{<Ãù—Gž™!È#òŒ}ªìÑ;É£‚š ßg÷S·‘©é7òvÖRG¦G15÷ÇC~`>=³‰õÌĦxEeõÔ=¾ ‹÷In†‹'n|=ÐÜíÔE¯Èœñ¦w\ˆ~Æb­ðsNAçl…e©)ú‡ƒÿa endstream endobj 406 0 obj << /Filter /FlateDecode /Length 2788 >> stream xœíZ[s·~ç¯`/3]6Y÷‹§íŒ­8™¶IÇœø!íu±ì†%1Š¢O~{€]ìòˆ¤•¸v'=è ˆË¹}¸œr&¦<þuÿÎ&ù:øééz§§“ˉH¿N»GgÓGsè!´Ÿ¬´Óù‹I*¦Bx±Þ0ãåt~6ù¶ùû¬5R²àU3Ÿ ƹעù¦H¬HÇEzCþZ[l\”ÆõÖy°ß÷I VwmQbEzQ¤+¢ßY‘Nˆ±Ô̇EZcÿÝq^¤oðçÙì?óN„`Á1>™ÿyàÅe±ê¤H?mKÂKÝÚQÜ?½ôGbì«­n[:µEÉØÚbãå•+b¦ ªãï‹ä‹ô HœÚ­#L‘±ÆÇEÒµ›?ßCÿ°Cÿ˜Ò 惤ÍWíÖÁ¿Ð” {Ø" öÊö±š(£eT‡«¶wc›²7w°”ÜÁH?Ž˜fÈ>‡Û‘øi‘(o_b0~Bñ‚vnñçPDŽ¡±Jù’Љî¦H+"ç¿£Òˆ":äÐc‚‰»p¸oMøùªH×E:Ú3—V‘ù„|‹Úü´¿—{é£"ýZ>î%ÊÅrìθsózçæšiî i¡EÃŽäú‚Ðíœ&9Õ¼&°ÞwšçÄàcB›{L¾$t½zó ?%¿Ú18QF§/btÆ›o”Ϊ÷ðäY¶"f  —Aêº<0yR…q¢$† ]>rЇ*bb:úǰ¸òN„f‘õp^šfÁаn¦B3¥­LY(…eà„iM¶ËB›¼šP(€u\ªlˆ Áˆx,Þ9$« X*ck«œaи‚V´².FE9Í,´þ#NÁafÏ,ÇõŒ‘"ÀànÅÁZ“ú€þÑ ²9?ªTÖ¤!Y# ý®§’V*×$'$eÄjgÙAÕ3ßaŒͱ9™éµO'G0*úDAÄyL-èÌ~8\V;¸êŽØ)ÏC¯ê“´Œ·0õ¿Pü°­"Þ狼™´@Iã›ØA0)A:è(\ÛæOˆø„lÏMJÇ8Hy=l,£)“‚6àñÛ^…x”‚. OP©• Y‡ˆËs¡“Û¬0øeU!®_k‹ó´°&ÈK´ù6‰VêÎá7½~‡ÁäW§Àpœ^AFÊü¸ XRáqqˆë]¤,+Ô/®²UX±T1e­ÏEê ¥Í‘C£äÐŽ¢oŒ¨1 orE¿Ze&EJ(fœéP²x¸]’Ë=>Ýã•A*[°üÓ+õ=Óí,à»ÐTá>ªbY1Uö´Pu«ó bs› 9å'L¥!@@«F¥až6Ô°éÙ~;#¸ªñè˜ì-8=+Ø`ke¬œ©’"R’Ñ>GvD ”:A(G)Ïœõý*H×G1kŒW¾ùnÙ!pæâ4íÛú„zÊpE5`!Tí»EJ Üx_oA‰£a^ÍuŽDל'Ö@Ìyc4ÂbIß"’£¹çÞ‚vChQûåb#Ì P«*)ŠVn°EÇ,ª"š4l‰³Õ6çÙ 4/rBB‹s'Ó#‚¡˜´ä-¬) þUÄ×EzR¤ê#v|\¤§$˜ (euI’ £Ü „ç„è>+ ñ0ù öi(ŸeB˜µ¸Ý]V;Ü"$“ñÊsƒŒ(·<#$$£‡EzTõ»Ó?öÍÝÓj ð‡Šgx‚nú*¥$]EW2é/ˆ³V%X¬+8}®rçÚ©¦+ƒ–ó‹u_dÚº<½‰T8QÃ9c)V2Ã3ž¶+BasJ]] ªªÞCÜ,5Ì]Åó]U>T\â>Wavw &±O«0hÅý3[L¬×Ý ÜÓ,ª¼æ1X:»…CE–Òr/²Â`瀥»éÒfÕUUÜçã0×¶é'b¼õÍue™¼[ïe›y(LÙAçeJp%}Ì>‡–˜Ñ ×ëáW¿@h•JÊŒOÕÍÑP”ÄWçE"ozðûÞ6­Š´®¤Ñ5â]ÕÅ5Qg`?¼Â{‚ú®eßxApH° Ún‹T]ƒ錰Ÿ Üýö ª˜FÊîïÙ^šì}£÷YõygSÚþ˜ïÚðœª5ñWê™Èߊ„["RþçÄÖðøWÑàêkp”Ÿ18Á´)ÿ%ÊÜq©ÙJ¨]0#2ÀL£òtI¥$&âù(G†ÓP¿®»;<Ã,°ïh«}MPÐû'E¿=éu¡ƒ×³VHnâ¥e©¿nFæ%Áqmåû^º,> xïÁ·w‡˜öHÒÁï"sÒä÷#Þ}z(Ã<”¸Ý&‡Ô¡¡xT`~R©W8aL…ˆ q‘W„ øÜ`ÑÄ`ŸSo öýäð`Ÿ¤òõýB/åk!ëÀKPµÉ]ñã_&™â…‹‘MÃüÅ íÀñÀ{>¾å¦Üö#®ñ¯Ž CÎêD/7 µ—z‰}1ݯC¿N«{ö"Ãz7T\V„WU„Fhmu¬“õè£Ë»Gè> òvPûö™á}]7U7Ö28Ûl>Ò|ÛDÙJ¥á\­G›øz1‹E|͘ØB4uOHK¯Žêç5`ë ½dÈž†XÐѧ,3Zhª°K}ÛŒU(u ¯HnâÄhN¨åÉÑÔÚ–0_ùª{/Òµûzyo`þ¥©™vãÛ…ûP õÎlR >}p$APOÄÙ“zW¤ÆT"müª62o•XB…ý©„z(ÉÑ’M½+ù@%ÿgþRI·òð“õÕ©OïH9çH/÷8’1xж•tê×™ÛI§æ;ª'}¬ÑÓèkïØ~€íÿ¶öŒejØPd`Þ»ãFèòŠíéäg:­ËÆendstream endobj 407 0 obj << /Filter /FlateDecode /Length 4014 >> stream xœÍ[[s[·žé£~ûÔÃŽyŠû%“t&NìzÚ´Mc>8“ô¦®±DZ¦dGùõÝÅuqJ²“N;~ðÄ‹½|»X_ÏØÈg ÿ¤¿×WGúλÙÙîˆÍÎŽ®xøu–þZ_Íž.a|æGo„™-Oâ§|fÔ(õÌ8=j'fË«£†å|K{o…Îçl”^Jn‡VŒ1m†w0,œÆà°´&‰aõ§XÆœ‘eÜ™aû6¬"˜çà œ)®üp³cœ‹a» S´5vØÅ/»…1ÇFÆÓÖLûÂ1nx;_GÐÖ2qƒK¹ÑÁ8ΰ£aŽÃ‡ ¦J!ñ uò: kãÒY¥÷JÉa3‡ïS6ñ#¼PÃê,íÎW™c Ûü8ÔI‘Uî8ˆ`.Ìðþm1ROà ÏGëÓ®™íeàÆ;B» QZ‰á ‡…wÞ(&nºKPÆ/‘gD!G9ü8¯ Ö%v0h¸f0¡/¨5Yø¢.|:ÿ÷ò¯G°Øè™…ZƒÝ%©35ãg†1éãbžiç-FιºŒ"µÞA{œå>ƒKÉjMäŸNm<°¿«<—ã ²F¸„ѨTpRˆ|jã>I>hëd“xÉnø*ZˆaÊ$ ³/*{‹ÀÃ-Q@‹,¡ç£×šGAíˆI]×ÎhëФP‡´‚3PÄ£ï +ÃsbFϹÒôFÏãá<‡3Ç+…ã‹F÷Ëz~Xmꉎã0fø;Ž2nêÛ?¨^ÅC3îuðLÒ jÆT=#ƒV+E kwT lþ4.§H¢ ¼©*u\m„®œAôBÄQ¬Þ(? êp`ÕPjéEN’Z÷Í£+2Þ)]üËS³ú¹jÿÃvŤ¥ÅkŒv'Ggm¶»ã*¯—Øð\öØMÌQÀƩ퓖‰ánx?×í Øk9È-±¥¸ ¨µ-7Ëyã$³¼œmtÖf3 ã1P„«{\M6¥‘lÌïÛ TX(q’xÒ¤GÅà¢UÔ„ÚvPÀ_ˆÊSõ'Ú¬ mAFà¨È"?…}”UYFÁ$£Ç`À{¦RV^éÐn›»ªÇ“ Ï"j"òqåÓÓt4œf2òÉ›Ì;R}mt"Oh"B8˜uýQ™UMT: £Äœájvô H^vCjëˆpîa•hø &f!BM,é¹3 !r~r¨ ƒqï¤ Çë ûmã-['ÀN ĸ‡g¤à™S%¡–dòyR‰ìˆrD[òŽ9X!üÎ=Ol$ƒ,ª$'È£k¾W5Œ d•hnI2mHkð]ø 2xI¢UtÆI­}¬Y‚“80€:²mI|lÂÏM\Ä3.5 9G(, ˆ`¤.èåeQÕ/£ûöF:-›× pZð‘1Àài B{[r7£üþ–ßI¬~YUçe+ø½­Í( —y­ÖA€Ãå\r€d=,xE b$žã·Ól‡Løý|ûÀUqÊ0Az'{¢WŒð*¤T `çó Fƒ`+S—…z[(ˆ™üªPï µ-Ô®C- µ*ÔëÎv•ð=Á;‚JñÌ|£”A®¼YiH¾º˜^MÀ€4æm >¸ŸE«ÈÈBrž}Ä(•I¦(1Wf!ö†”÷bŒ˜ƒÅ 5êc5\&zȆ—ã‡Å.úLPP`©„é€toW!ÞKÚE𘾠ÈXûF݈IQ ˆN?Ðoê—=©rZäfËoŽ–üЭ“`+R© ªx#ÄÖ?Ìa Ëâþîâ\Ç›,rõºªùaG$¤ÉkX¬²DóË8Ë)AˆËèYã¸"ŽZ ‰â=ý¯€ÒÊ*´*p1YR…ª*|]¨±P7ÅÝu~%¶s@¯½-Xû6žÖƒÌÉèALoÀÛß¼÷< YyÝø¼N \`:.‹7>îœhž;ÉÜIHU踧q ‡1?*¢‚! ‡7æ~H )­|4·®qY³¼“ÖM(ÎBÚýwaÚÔâ€BØð2‚“š„m‚€iš}šß3¬n«€#;‰q…° Y?H’Œ¶¹çÔ 5 ¶ \‰¡!øpÆ”•=-Õ¸KQfS‡¤ Ñ6pí–^4rª¸§ª¹‹£è®È=Ò¨éÇxñÝ4ž$â9.¬h2¬ã¸6ff5LP%»Š á9¨›™e gh\šÅœàšÆíÈÖ4¸“œ{]—^ÏCžèRÉ ®@j ‹ŠÒCÒÕñz¢éñ³¨È^1Ó†¾ ´¾ÆÆ ‚ÒûušJæ_—%n»g&S!sL;sô1O#UÇÕu%ß8½6{N&x·Àzc<†0!Úà 3xÔÈg@—-µïþìè€SŽªä,¤‡íÍÄ¢ÿYõöm²AkmG6€Ï·$¦ O5û.bm+>oí¼Á9‡­Ü@òQAÓ#Œl⓱í4`#ï†YPZxˆsûhmÏ%}z;î÷ =&)ï«Y¯V)/ñ.UÔœµ¡ü ‹¡§DVœ Ꚁ¨­lÊé“)DO ‘Ûº2É›©e]¡ytY{¾<˜÷ªï ¶$Øm4A<ãÈ}-,œ]K?[ ÀTBÄ{ûó|¡°#Ã&†…Zê¢Pï õ¬  »Bäåðy\ÔÁ«2¸*Ô»BuÐÕ¦Pc¡n:«¼.Ô%á&S]½èälåY¡î:®æ75«¹ìŒÕõžd)ð„''¨÷f ûd¿J0 yÀX^*Þvÿ¾¢«JóOõ,@óÐêٿÊm¡>«,zð¢sƒ¿tîí²~ò—Î.'Šìò´s5·ä2ñì€Ì”ônR|ZÎyّºPoªLáÀÕA^5º5îpH®-½\YÉÒ)ÖS"ëLm õ¡P›´\Ý«R¾²¢Ë ªƒ‚Î «¸Â” L!'ß•½N …•+9Y1½“«ƒ„iÜΖízÂß’ã²HmoÖ4ic?9çžË,ŠŒ(+q™U Î;ö±+Ôu¡z¬*î®ókòHöL¦ºÉj}3/ ö†â£5Âçæøðmgžß®œW³ß¤}„ÒûŽ)à†áóåO{œxžVͲ1uo¾è0s]¨Ûƒì·^ý¸zÞ‡òöà7zi,Q¢Ø<ÖSÒ¶ï2n:cŸ*ÂÂ×ó_w¹Åy™Ç«²Ëâ^•¼íðUÝM½öãz_T²š•$Þ/SUÝ©›ÉÔ“ºÌq<ínÓùˆ“²ã¢Pï Uc{ßõ—e>ïÛZ¾ä›ýýª÷2ón 0þ´Í·r ô'Kéïy|Áöœ¹5dÓ©€#0?qÃùIAÛC‰OsÄ+O H~9àw¶˜X(©! mzT«­^âÃU.ÇåvtzC †ÝV¹Èç:½×žOô©Õ±qb#å_u[“©X¿k’×Ô~ƒú—jIâ@npܧǒ$Ò¬Xpô¿ðöé1>#×>5F”:sìlßcŠ Æ¬äÜä䯫Tžöõ}g?`V¯q>ñ)mNö«Êô¯ E¢Ò÷ufß!y’â¢SÛZ½‰¦ó n@DW£ñWjÿÔDW¤žDÊ gÛ½z‡È '•Õ°¡õk)é©…rDòè+]z—Õʧú¾Öòªilêm+úÈ}R…ÐÚF)ñ‘bÈYóŸÝHÖ]çññJŒ9}†„8Ív¤ìGØ}ª‚¨âÂe)åц±3"0‚3T}ß/Û…Ò,‡nAìn¨ Õ#Ì,!îw{ÖÓêß(¯Õ|WgL¬†Ã0“ Ùh”ÿ¶÷!¢ÔÂÇDh’Áê1U-#ÇšTüXxòhÔy›t«Ü2XšuöJ˨–5Ý'¤Ðv—îÂÉ6¾†Æ7!Ek&´Ã4v”åI»­¶I Lèÿ v¼Iã LÕq®hó á\N˜Ý$Xœæ°÷X Îsqm×p¹¹ë „’ ÈäŠ&} Œ’¶l’§õ~Á>¥¼­šöfN:/c}8Þ@x&Û¦“Ôÿ[$Ùt¡\ÎK/i»÷ã­”Òiļ¤ð©ð÷PUóÑð·ÖÄ^v¨ûˤªäÚáó ²1gíæÿÝ  ±Ñ`¿P“ߟ;'[“s—O–È´ÄíÓ·ií¶üÜ+÷ÖtrC€BùäëN.zCÆ‚Ž[7JÔ²P@õÉ„OÑ„NÍñm§lð¾³s-ÌÕÂI5’»³Ž¦×±Z6èU/u) öv¾!;ï§•›ƒ2lµø=ÑÂtUèZfy[¨mG®=Øu¶«cŸÕíê2&›´ µÒú`qÜ®Ç]ä °Ùc_T|â¹Ð"áù6$ `}OJc—ÈÁ¦‹šÚîâdìÀ§©rÆŠO¼<ä !Ÿ©Ý©Û²Ÿ¢Ø–F 7iŸÀ…DÖÑžˆ]Í»hzL7ZQ^Ý;QïÀµ§µ+¼@<Úx‰¯…NuZ%›Ç[ñ@i^z«HNyÓãCc²[=ìõ™V{‘öwM®Û´Æ^Æ´^õ›&(4º>ðüP‰!½jcuháäé‚cÆ6†Kn[õ+–[­kòTj¿ÆRÿsíˆs›§ÿŽ-ð uGÈÿ Ùçü|,+{†¾¨£V‹vx×Јö¤’B»uì^ÛVx6çæqfŠ“¥›4˵}qňBC¬Uþ?¡ÿêIô™¥ýñY'âô^éÞU¹TiÊ+^´çøŸk¯·àÏ*ù¼ãkµäËBUTµ¬÷ê4½b‚Î ‰•PóM‘ü×…ú[¡êU~ß»¬]/Lo U~®ïu9\ªsý["ûΆý¾ÓÎ[HOR÷wQ_M„‡[v”ú®¯Ê™ZO>iµqE¸Þ?Iíñ8»ÿð'=i×u~éM¬’=.އ¦½íEžºú–©7@Vo ¢…uGÚc@†ŒµÔÙüs,ט=[ý þüègµÿendstream endobj 408 0 obj << /Filter /FlateDecode /Length 4102 >> stream xœÍko·èGý‰ ȇì¾-ßKM Û±c ЇHŠà|’l%Ò­“íÊèïÜå ÷f÷ΈÝþ —Î çÅʯOE+OEü×ÿ\ßœüáoÁŸ¾ØˆÓ'¯OdúzÚÿXßœ>:ƒR†S)Û`­:=»<Ék% ùVê¼m­‡O7'?6z!ZœÒ]Ó.–¢µBh¥³XJ©Lk›ïã¨B8Ó¬`2 kÙ5›…mðV5wq‚ A Ù\DX !ƒn®›Ò°nù,Ž;@í]B"‚ >Ím'” ªÙìâ/„•¶y?¸_ÍRw60°SÄèŒñÕF»H•AúHKœlE×Ój´ Íæ< ©›õ62ì5ãW„åŒ=tx‹ll2é°—õ”.²U¤EYª·0WÂTÛú ™X%Ž>áí* Ô+Ÿ6ÏÐfñϳ?dzUò4´p`.mhmPp”ßœ}ñcóC’¡ÁE¤lü-³&…‘&4/Ô¾“!‹ÐÀ¹:ߣ6‚ Ö¦uÊМ³sГ›…l…ðFÕ´)Ð]. t] —ÌÚ‹–òm tW ‹혯?5ü.XY×§äÀÐe„éLófPNД8Œ5ªYQfùÁ1Ùæ*)›î@ï¶›,a Jg¨ ‡ Ý‚A4¯ìµÏšn½=jnÓ~xׄŠímÔï:P’›ˆXÂo!êÈ€íq\L—6é€_ÐŒõ 0' r.ŒddA:0„ /£½Ika8ÙuGàtœ8tè e¤WPw“ŒYz§@ G^÷l•í0¸H_^?}G–¾'0‘Âu6wÝù®?-GÕO‰®õ²œÖï‘Ú~vެ öq¶k•“>Õ­4*¯QÙŽ:);N`µ3Â.«E M¿{I,å*Ÿs\-žÂ e÷.O°n{™ÏIº.’  Am´I'`ÀÏdu쿱äµÔVATX&?oºLüË‚ä¾w¦à­“ºèñ]³ÍtayGÐP’U]Ù<¬;@=øZ`:"0àU:0ë»ü=ˆ(! "jÞEQv¨‰þ¯óà³¥]·7Iƒ]ÚIIëÀ%£®®®3ÒÎ7oÖ¶2De&V’üd¼Aå9Zî5šèa§çÔ$bA$jÞ« èq±¢¦Õƒ"®`˜°grê)0šNg²%à9'ã‰1o”sYrý”8¬…ÓJNbY‘qDt3¡·Cðã\1ðÊÕK¢g'J8«IEIÔÖà;\¯€Ì{PŒä[%ØnÞ¹w-i•€ð¬TèÝo!®Ÿr‡0•@)…õƒÈÌ[:˜|êMópqöKŠEŽ"Ú!þi!ì&?êÓ“P¼@…:€?t…Ô_æ1›Ö(Ü0ûñžSªÄLÃt" ºä¥ÎçÕÒ&«¥õµÃ]%·žBÇËl=§œÂIåJLJå–ºò<fgÁ“ Æv“Ó¥øK>Žl6½#ãöºw%VfjW»Êtâ¶V„â ²XÀ£‰–ÒѰ¸!3^éH@Äç Éqh ÞÒÔEv³àpTÅçE^­™ì½Ëûœ©§ˆð{nx/XR‡‘w÷×AÌ+¾*y5/àû*B`'>™"Óõ®¬¾kVÄG ê„à¸ã8(GGg²à¶>Âqê¦$ÐâzŃÄ([T¥›¾µÆù¢÷¼W(8à9‰‹eqæT‹SJ ƒ0û‘EAŒ…€°Í)—‚´:{ü^Žk‚e›ƒ#pPˆQAòb¼ŸP„å¼EØÇ„%Ç©ªøy€¾ËÎÚ ã]R9±-Ñ;é@I‹“üŒªUp‚‡´TJ¹ú‘M’ÀV™g2"ÃòÌá¤Ë’­‚}>pjætj'åÈú¢¶™Íåž­º UoS"]‰’ÌŒð£,9"aÊ(/ò„h5¯œò@©«ÜÞjdÒ´°4æJÀP”…yFâÙKLÜŠ\ãà:6ØÍåó¸@ ô°@çÂ{ÔUv tU®3ÛmÈ]ˆ 8 Òåʶù2Ȭ1Tjo£‡(Œ’Y¸ÊÂB™´,Á1ÃÚdØISyÉ%=ƒC"^àZ~e|:%Ç_Jv XG€`§\”–ñŠãJö»=ÿWçÀ«±]ïÿ—Q@lîH$?¡+‘Ü!¹µû(Ê›Xξaä>ŒË ê>®0Ëçï œ&:£fÑÄo˜ó+rcæØRì<}ÁÓÞ·I¼¯ÄFN§"Åå­R–D¼Ke—":ˆíùÆ«âÝöùžræ‹r×nä÷n6Ö‘®Èͧ÷趃9ã[Aï[œ>HÙ×öe(°| ™8¸æxÙþ¬˜êg8HýA|Ä=n ôëléÇîáß™JÈ›3š°!+z4ƒ{ª—ÜèŠ8›}è÷TkOvŸº˜¡vßnôuTLú¡@ ô¨@ŸRpûE¬Êê„ÌN%ÄÚ¬’½%|ÿä¹,г–)B"ß3ó®QÁÖeð%Ãÿ’AøšÑg$æ–Ùî·;T4,ß1ÔÜZ¸ÃºcÐl¸Mîç7¹åyMF%ÏÛjÝa2ÏQ)„-H„ܸ ‰ÊÓ17Aå¹cƾ\,cµ'Vqþ·iOö+að+ŸFù§Ü ’úžÅ\|Ö£?º¢… ,ß±xn ¤ ä  ôÑ »<° YôŠáð-ÃÍ5Ãë»`¸A4;ÆDš99=À ŒçËÙªì[‚¥œ4F¤û½*Ðv„°v‡œ§âÈ—œ›»,çXnˆZZYŒƒç ´Ã‰Hž8ç·¸äŠÃ³aè!>5o:×I!´Ô1µ<âG…ì]b?µÎ-C95‹‡ÊéÐzÖoÜ!œ¸e&^2j7ŸÒ­ÉWFÆ\ÀB4x|ó¦ûe’V¾0¨‘°D!“®ŽdP¼‚™gPV’AK3Æy ôy>:Ïúvô¸ðôiz^(Î-ë³ärßù 5‹>:6NsnˆêU!Ïh‡og)¼"»DºJFšD wôI•àÝ0g˜1Ï^`—u:,ºC¥Ói¸¼:Ršj~åÚ¹ p‘%7GR£ÞelÚ›jÝK¼ÞnK}ø]œ ™Wß$-eéˆ@Œê)o±žBJxç´0¯ã¦w]ã(áó'loã°j],ìÖÅ}„û¶²´2#Ì!B­³Ñþ[,FìÖÛæ9Ûõ¾‹Ò€¤3–>à³r\m§eÛ‚lE8Oý%ŒóÕÞ;¬mÍ”1£HœHíðR^8Šeã8 5ö²TUHý+«”ÒÛŒ¡“ØŠ|u•LJaƒÛe¶ú)¿ Eز©{Ê´#‡ãçÈ)«lI¥6‡¸ž©ÚÛí1”làU¢–º¾KÚW|Ó¹K¡:—AĶa]LN½gG[Ä-Ðgeªw=ë­!]¿†¾0Q·q¹²îÙÞbÿ4•Þ]‹• ÚÖPm]mò  ²Âµ¢½B+#æ}ŸWÆòþDñ´lÈ­s?ËÂíªk¦ ýyý¤ ¾ÁƒÜ+§¦¾ït{hà:–V3\»–×È4}dÛ.¦-áJHÞó6`3|ä#RÃd¢ðG)¤í.ªeÛh˶ó0/÷ \|1/É)C©K‘ØX›ˆœÔ6ñSƒ¾n :—º˜zIz)É‘ÉÛQ+ÐÙÔZšj¡NV¨¡)• (K°3. ôÆgÀŸq83¯›Qóv7"½ú`ú›[Òãš(Xß÷˜Û·ÄË‹Ù^“š?-Õµì] åŠÀÒšèmótáèƒôÉIÅm:9Uó=0cÜñ%zÙujþFJ=ÓD¨_D„øþí×"+ömY+uðõã¡^jÑÉÅp2îŠÕM½deå¨mÕÚLÞÏÅž2âøðǧHG5vÓìÔ¼ R›7±›ªèªj<…ž¼£¡‡Ø\|ß“¶ý{‘ìò«!éB….7=mhµ5#“$Ö‚ëœ4w¹Q0!rÄ¢FTÚ‘£Çw I'}Ô†W.½Iˆë #œÐ×àÒvGêk‚P½Û­;Òh‹¥7ê£Çd˾âљҸ:t©ß¿ºÓ©O<¿´`~J—Ì f6ôÂ÷Õ[Åó¾õ,ö%V>5Åå‰ȱŸ—×÷}»ÐQºµŸ9z^°ùà[ 4°õŽ‘ Bx¾,Ðϳ×ó^'d f⵨zÝ‘¼Æf¯©(týt ”½è0&eëÚôG8jOlÚ ‹+lÑ•pýµ¥²µntûúø[ë~æÐ¥u€z1FðÃêÙuõìã´{HYõšYò¢@_èi°õ]°Gø„Ñ–J…K1$^:¡©¯ø¨åfñ%óù?Ûi#…Â#gR;ÿ–1² ³ ­•ÅÏË éßÓy‡<Ö/×e ?Œ··h-!>RÀÇ~ƒàŸæ¶®»Ì•â4SmÁ‹*dÆž‡Ú3%\lG_£<êæ»8¦ÑØa•ÏÍ·p¢b¶Ã1Ïlìç©V³¨ƒn¥nÜb<É ”Å?F+áØã’·ùtOÍ IáƝjLÞTòâé‡yi%Ð ¢oŽÕ,óŒóµ;JXóŠpÀ°ç5Ë0c4ËΊMqo©žÌ+ټʔ'NnTg•,ÌBݱJ&ɤñdOÕ7N2D;ßÓr›¿q· s¬¾qÎ(Gê›fÆB1S”…†¯˜œ`fó¢@æë%‘Ñl"‡9ÿÌ‚(WÎ9•­J}”¹áo›Ï‡蔢Bé0L‡Ä1ꔨÈnµ­1¡”æ¸àÌH²ûiÙj¥qâ±®yHJ:]jª¶*ËÔE¡>9;ù+üûu‘‡•endstream endobj 409 0 obj << /Filter /FlateDecode /Length 5858 >> stream xœÅ^ò²`6áŽûýàzXr I¨ üö‡åÝòH‰Ü£î޲hùí©êgõLÏîOA`îUwWU׳»ªÚ°‘oþ/ý÷êÝÅož{·¹¹¿`››‹Ÿ.xøë&ýçêÝæ‹K€à‚oüè0›ËWq(ßpîF®ÜÆ8=j'6—ï.^ ?o…ãr8lwlTŒㆻ-…S˜á–Þ+%‡7ø3wœñáö¸Ý)&Fíýpû*~&†}üôÎïÂHƸ×Ã-fçÑFW\17¼GØôÃ[ø;÷œ+ø€ßÌ:¡‡gÌýhþ‚À‚1¯ì𿵷F¾ # S¦Ê´w“@³†OÀÂYž „QÃÍëéÒa‚&SÌŠ!â#¼PÈœWzC­KÒß]š6Îu œHÜ—~œÒ,Xïœ3q àìÏ—ïîÞ©°…eÊoà¬0À ø¶ Ìó$£ê'­‘vØ1ˆÌàRÁâ/·¾üÃ…€—ö Z—× HdÏuSÿ4TƼ'ûùa,Mw˜°1‰#w^ ×UÖfò¢d ‚óÞTßV„íÚâŠ|ô gß“˜}¬$ v™;°ZóÈŠˆ Ӿѩ¸”UÆûˆ¤gš YT)ÈŒŠ+Юø«w†þš„Þs“ø­`1¼ §,PyEÖ¹°$ bg,·Qº` ׸QîC(YˆbÜnA†€;tøäŒ‘rHƒº4 a ¯˜iÔì¯DÞ€Ö P­êì©è“]úX¥v¦€ÛÑöF —=Àe”†= œÑÀG8sw¢—‰@Zj\ÞT,Ž•Â›¨.vꊶF²k³ øÙ"eµŒ9ƒëí€ hù 3š¬Nˆ²\µ)µÐ—Ý“ÃוADèoè$qÛA^šIpÝß]^üçEô2ŽòÍÝc]‹²èÖcaѵ r{ùCžûÑó9>rkè|/‚ù}CކJfúuµh &¡èn-l²1 ?R6¡ìkë¤ÔS6í¤‘#.+»EüHtò:`%%·ø%@ˆofÁ’ »wÁH n kŸ¸ü»þžÈW¶<Ô¸?¤IÁܼª2¼bå%z¢ƒ ¼aZã-5 ÈØ-Åèn P%³È~¾å#(A{¿Þ•¯7åëa›&k¶ö0Ès¡87¨Uæ„ØBxHîDS½¢žµ‡®C ìüë@ÆöH…˜c‰…€8UøMGøÌê_ïúø Ÿ0eÚ¾(ÐÝ‹û+ÀsMt2ÿIÜtN å’¼L54ù‚Cœã÷öçâŽil˜r wš#²T àUbÁÆã«F:Ö†—4‡ Ë3í²¹ÐüJsˆÈ(HDmø¯r„_oçnÍAHg&ÓQ¥Ùî@À”)âà0Ä^Ö|¥$æöeݺ=ù~3 i¼CÐ ¡ÿ,ߺ‹Ã&ñfÁG¿Bb<´=ìyNbÈH9JÀÁN*‰rFK€Üü°vƒÐ/u^j‘â¦hãìpßC?æÜøm€¯è"E6Sú Füc„É!¸))Û$˜ÌœÙ_‘|‡´p­M+ìK9[O–HDzÌHËá«:Åã}ÝðÄ4´Ÿ§­pÉš £šíÞ“Ä,ɨñѨ´!\b@«P·šEÑc‚Ë™¤XH¸!ÑÄðñg¥ Ï~u1ÛSGí)bïÝæòÛ‹Ë~1ü+:3Làe2ž ú ‡ùëCùzY¾Þ–¯CùºîüvU¾~ìüõ®|Ap—??ŸL¸ìcÖ|øŸúùUÇú¿ïù—úù)þ$=«ÓTÀ}'N«¿ý¶|ýWùú²|}Q¾Ö:ÛûÎWýëm‡¤ ÷"¨¢ðvÔÕ1‹òçÏÊ'Ï¿þ\¾ V.i¤YàFg7Ò012¹‘Rdk7‚+9BJuwØüqs¼À°^@^Æ1Õ2b£a. AÑG)ÔÆh4È .  •1Ümþ Úàÿ?@J÷Íš¡ß_  ¢ë`bŽhe`ÎŒ‚¸õà÷{ÁÀC»Ñˆ„ þ% {Û ž[©ê¬++®8£9¸'ˆ泞$¸r[Žì/ÌüFZÜ'¾7¶È’YÏ\pF!sEG¦ç³ž$¸‚­9ßh†™­‡½PxÚB'= G0M°˹1£àó9—ážFlú®#ØY™‰²”§ç ¦ 8cq„2óYO\Á}€;„GBƒhÑ ÂGeõ< Á5g20x¥p‚2›õ`ÅUJéQ°•æxìº\n‚‚QDVV\3paWF×FÎg=HpE{\Ò°Ül„FNqnO&< C0Œpe]€ƒ`äBÏæZ€!xAÈvp£æRzÃH³ð_ ŽQäÎ pÑŽ¬@=wóYOV\ó®’Zœ…_¡Ä$8Of]XqÍÀ…U XqH¨u^HpU,°IAš‡ê!´D¿À Ø\ªFçᦠ¶nj„e¼U¢ópOãØÈ ¬o4P³bO‚§!.&Â~Žà™`ËúV`Hj瓞üþô¹,¨†!ys®9˜…àB¡±¸Ýñ ÷«Ãû<ëeÈm­Æ€føæp¼>Ü•à˜dÇY ŠŠAÇNB0pÉŒçÀ„Ó`Á‚Y@ +q³&(91ôá®G8ä¶£‚ˆžþyŸ>kÁÐ’ÀÅ[1 Ó¥Ï ¼Ø©‘Ó#è[ú4úH´ —ÃÃwû·‡†Äï$K¼Ó’ØÇ¼ N˜çÁOÅ\0ˆoµ¤˜¿¾Þ‚¯dºáð©(r #‡BI‡˜Ðó©$.‘'˃ŸLh&$ë€}Ñr¿¬¹_"¸¹€ÐÄëBž°CŸFßI˱_"1Ç~«ä+Mˆ§±OÆÛà™b‹÷×-Ò9 ̆M› ­’¦<6›«8öéHƒ+‘¦¦/ûH'N—€¸#L41¶D¹+„éüЧÑwR˜rÄ›H,qô*iʃæyð“1÷C`¼ÞZ•`8‰—†Þ—’ƒ‘$ëi!aY VºÄêÉXéÂoŠ>ú•óR?õÿw5Õ«Ô°¯iË]ñu½Á'w¿ûvׂäñi fSÔ€œS¢­Õ)7¾¤úf±Ì:LÀb 4~ƒ,–\gÉ›•Iä*r;LJ¢hÕ䪪*«^_B÷¤Ärzñ,Ø+n'Ïk ]°¬ó™ÀÇÖk”ªªPŠÏ,³²© Ç2pÎ5U Bá¤Lc¢^¡§_†ÒTiÔ’Ž¶ä!P ¸­µ ] 2b”ªé¯Yø,2¤àæM-Òø·¥1O§ë¹`†ÁKš÷Ümy3,õµõU\/fM)*þÕª¸ ÒÅnÍ*!JPæî$D ,•ÉöÕT<­ÀQÁù¸u(ÿ"TÿjT"¿\þ‡x9µø÷^{ 6)F“ÁŒƒ è2ï}aAÝÀ·S[+R͹µŸ±VƒGR¥.€Ô¹‚³bÈ‹~EiªndÂ…† ÀßLœU®±êQ¤C#KÓÛŽX\w\K-ËèU`PÇÓ£VóQòD°);­Ð¥Ëá‡nQ-’Š\Ð`ïC âh%žýðž…{#GО]7ΩL,À2˜ý‡®— |Ïâ(8õi§h¡õ†É¶ë¶†=Ë;Uˆ@ãàõ?ÕÒ®¨qÚb¨<ñ.´$ÙöPXËÅâ*ª•Wõçâl¾Rø…’BZ,•‹¬o¼Ë$œJ•ç‹Î= @+„KÅü¤£¢¡1ÊÑ}…$©)8\ªùÅŠ>H'7óþ þÏ[^&Q߯Ð^T ­6ÂzLÌB¯žÐ^!!UÔ4ù¦µ¸‘QnÉùSN€¨4¡ñf¡˜±Ô£ˆœ²äîžÏj@Ð+¢Ûáù±b è‚ʺä[H1ÝûN¹\-«{è¿õk•ܳ:÷/!ûN¡^-6ûÇòuÚëï;æ´Ñw±Ç\]TÈ}ìx¦ºÜËN…\æÛŽ/»©SÿûQŸ¯öeq”為N5©9üï_§æ¼õœ?Õ…s¿‰¦WD1Mèìðxd6b%11Q_B& J#Ä ·`¶!ÂÓöÕ_±×ÀÂö/¬â÷`Œ_߈=LǼŊÓá¸û.u`°Àeø‹'#±c;—†?+—,væÊØmš¦Üý>h0öq„I –¹‹1ïCÂÀ8(î_Ó€ÂÐï†I‘f"¦iʵx#s'1˜@ÍcѰò [ƒSz‡ÁpÀVeÖ]ÊÿÀ¡·¿–E®ßlË÷1ò2/í"R8Tѹ iW WðU*Ì®1¢HE¾ 'GÑî…0b+Y {—¥Å€RÑCâ3÷ Y[J*µò‚ºë„‰¡äB;s’üCtqÀ{— ñnrôV†$3 )ð¸0Ä›*—¡F¦BÖ£Õ\k†rM5ž¿†‘N«•š%ó!Ó¤`¡9¶î ·`ØäM£ºìq4§¹.ñY{rûËaCñ¡;¤M;p9ƒö;|8ÇÎC Åþps»@†®‡»Ö ÁìÖ7-4&íä‡ AÚ·‘„†ü)4Y˜£½Ò¯i~¾6m².-Ùyñ&Î]î,aÛ}êC_UcOZ:ni§Amío{>d8% †=t"H òg7¥Û5ÃÒBþÝ$Ì!ËœVï÷5dEÉ>… ‡»= °=9*‰ W "7©Å™Ép¶Q»pâìî‹iN¢$ÞƒØáuS&ÉiíCD€é(§ö rRÒ .bO‰üݬl÷,xÇAOÚ8æb6( ²ÆµJ’šÜÖ( ?ë”w¨}›¥çŸªÑb•šý±wf×$fuåxÆFÜk lóÄ@^6òøà-Á¬&í;ô$­ ŒgJm‹nºsÂ|À#c›W bÇJB³m:~_{éaaí;ìžuTƒKÏíQ9žlëÓ–ÜYãéÉá}']–p¯×ðúóºÞË8¹ˆÝH ãôhŤ²íÍ”1ÃhŽ*© s½}£iÈ­\HãÒ¨jáhZHºíǪØÓ®òhŽC_Ca˜HP:Ö¢L¾Ô|VT‘àÓ´Ze|“¼3¼Ú¬ñ¥°É*³•tÄ}K@¦9~>ÿ±•~©}¾mû*… ©~!l¼ìêiWW7±!±ç©­+Òvâü#Ú «Õ ÔcfŠ Tj¯js;÷"Á¹ÓêöÕŽÜ«¾&>JÍîÄ^ÇÓn¯úé3‘eÈœ}ÿ% ÇÃŒÖÙ1«rHÏß váíï&'”ÌZ%–‚òUEyÒ¡Ì4` O!JH÷ž„bL Í-‰†f"Ìð±è óõü·\[Óv)¦øiáü(Ü‹‘+Ó6ˇc™÷ Y"S€¸ 9 rt·`ØãyÚpõ”Š‘¼Ó)f£i‡o݆­“C9€ÙkÊ|‰,¶~G}_êm_ªÀÜÍAFé)ô¡oïüÒµny˜LžÎ‹©I«˜=Ëlôˆ’‹Pè±\í{]^zÖ% åÏÕµ[xM9×>Ô Ã[­£ èXôª1h…K†G%¯ÆÝ7d#b’›(i±á9÷¾RÄ 7cS*ûO+¥gR`°žF/Õ¼·Æq!:üKOŒ*üzê$ªBÞAilþ™Ñý ´á7¨Ž X ¿ÖvDËwŒsæçÏg¡¡‰êbäKŒ`aÞf8Œžñ‘ŸtÑ4h_—šÉ÷±o[r0“ýÒ:/3¬ããU‹‰Jµ·¸¸þ¥Í”Í[Ô¶k*Ý  ˜Rù(J:Ö QßÓàdþšU6lAfb–Y3ª&äV††-s‰Ì¿¦¤Ö”i›"óˆa+çmŠ LÑ s÷Éc Ú÷¥Âk+Ÿ¼k_t+ÅÍ;_1@O“[äÖÄW”\¨¥gR^8 7û̓EH$c“¨·2&ÜLÊtF3yùcz5áDˆXï[ž^´7áHs‘§Ú[]û­Ÿw¯(ñÁAHù÷.€pÁý©¿\^Qnæ¯-˜ˆÉk#<3.¾H˜D²_ürÜö¼Ó~p#©†Ö á’Ne”ÿ¶‚fôñÑIÖ>ØA¤j'Ùñ}–I¾BŒñÂs‡” #©3 óùê䓱¤áðO‹v”M°òq¡¤’‘¦3ÏþMÅ'í¤ wŽzªÕÐ ã4oÛB•Ÿåbpµk/z8¶ø\_Xߺ¹áKÀ¤úLM\'™$7Ô®„5<›½ŠÎ؉dø;Þ³ÿG ³¨Zµ5\uoq`s6ØX¥&@ì>å™Î«(ZvÎâ"Ærz°G¦ëø[*ÕDO2¾LŸeÄ™AkuÒ 7q/Â!"õð»-=..à†6ÇyÓÒÎò¢ffOÔ9#gW ôI ·ÍÉëÛ ÐûÏB°^ÎB0ZrÊŽ’’gñfËq{::ª)õ Í ¤ƒf~ECJr¯hÀ,­º¢µóèuvÍ ßÆüÏøæoú5rH T­¨ ¶ |§ò¦gøA‚Â" úof>¸5ü¸µ¶‰c~upØá¶‰åWU§¨4ÌR£µBØG1 8gõ„Ye¿ÅºhÅδ tÁ!Þ}ðmLƒêŠA¸ÞâCxø8+}l™ ¹€„;åX„Úöæý©bFï"N–/<\¼?&ªØYyÑ`¥LÁâyñÂ=iÁ8ÑÚFXND ÈA-NK;7ÇâóµH+ˆgH„ÄÈ*Q0ùòt,„O˜z¯ð TqñÍ=]8Ô2žSbÁþ±¥˜›æ‡8O‰µQO, jóšO:‘Š^ÿ×}vK°pɹ?‚‡…ë’;Ò ñ¿ìX€endstream endobj 410 0 obj << /Filter /FlateDecode /Length 6190 >> stream xœÅ][o$7v~×ûcZ‰T[¼/F&í¬ƒ,Ö1ÖÖ"XìF¤Ñ(–º5ÓÒØó’ßžsHyÈ:ÕÝÒÌxásªY¼žËw.,½9î;qÜãéÿwG¿ÿλãëíQ|}ôæH„_Óÿ.<‡B»cßy#Íñù«£øª8Âuø‹qC78y|~wô·Õ¿œ RvÞ©Õ7'¢ë{§Åê*·Ö¹õ[cnÝæÖïrëî£ä%ükxè^MÅê«üðï«Üü&·®rk[¹5æÖmný.·î>`”Ó²ÂÒñ-3àÈ xèÄ?9ùáüGBt~ÄñùŸŽÎÿùÚ–eÛ ><+?á±l™Q¶ÿ@Â{†ðßeš|•[‚‘‚Mn]2Džñ~CäkæPï8½“o Éþ¯4ŸÃýùå)Mþ¸¸ž…=xà6ö‘Dt7KÁi—wƒ-8ßÛÕ+fº²Àf ïË^”æ}þ}ü½eZfód{—;™~dñ!²šÄ£‘ Žë‹j)*è&·.r+96?–h1tâ„cËA¡çãN:”wd(÷>·î™y_äÖgŒðn™ñÊ ÖÌJGf”KfBŽ[æå—ÌRo­H':ç-ŒóK€× =¸u]“}r¼Y&俼Ý1K¾µÞ€ j´^ÃϺïqÓÉ*áäH.úW2tÖét†&Mghà EdÛ$J‘Œ–®û’ôˆœ m?^XÇìðLg{¿ïðDôÌDí?äÓù…1cžäðòÑt ­îœp«¿ž8W‘ãJM¯+¾ç(A€C—I¸¢žÎ&ž?¾8‰:lÐÔŠ¦È5·sg;§Ý´Ò¡Åf^3œ|ÅXŒæWBªÏvXŒÄ·Œ¸”ß2¿ræýgfÞSγ+vî"·~b~½f&á\ƒÝ›ÛkÙd/;!óy¼" ”½¬`TN-LçÔ¤S[ú*éG^œ³J‚rW•À‘A°=´W#Ê¡Â[ù2“-4zhvE ¤ 5Š [ÅLÆ|€75CÙøæœ+åJÀ>AÇ€˜Îp îˆ(Pòa!½SJ¡µÄcƒBƒúŠ:(qUPŸ´ÔÐ[mê÷Áº¢x:Èú2½ æêtG’o˜«‹úÎaõzMG|…Óõx•‘…Šbîz9 vá(×g„sì?pXΠ¼´¤ 8 Ò+ÊLÀlFŠ"ܓɥ “ñÂBÜÐg†,†v¼8™ìÆO'À É®Æk48^u°8-|Á€ùl„° UX  è¹h³(T€c'C\FÚÀµ½¨Fû¥Å†áÅ]ÉOMÖù>Nh´]GypuÍB'ŠÎ‚ùƒå¹~ˆÔO3n¢.>–èqæ H¨»YÇU  Ø‚i mPEM÷3$±‡W}&@Á‹bgãjã>ïNPë ¶TÄÎÒñ@ å*„æ’<ÞFÌ%]'•>@¹šÊ>“9|±.ÒÀ±,Îß–% Ìp‘ÃC3~?Œ¥+" ¢d†œGåÛ@¤7…ªT  J݆‹òט O•êØ—AheåŒI’oã‹høŠ¦¸)¼EF{ .j/åLdy± $MXÕíU‘R½U{žÞ¢¬UØ>A7±yˆƒBl‘Ï‘(Ê€?£6³‡ ‚öSÙó)Ž Ç-jªýIö´™Á©Î”`½8º§[zÐSÞ‘ÚÆa•«Ôá&J†ôÃ&–rV(€ñhUÄêY¤ÈÔ"Œy³ójÝI“¥„óÒ· ºaðÒ~@¬Egô ΧG'|… þߪÜI·…-¢P""­PŽYF•eÁ´÷²²‹ñìB#}á[µ+=Îd‰Y¼%>O<!­ î,vÆKç„‘;pׯd«]BJ»O9†[«¥ÓèU0S”\ä9Œk>wxHð4h‡°;8˜ËL B«1Ú"Pwà÷Œu4_Ôý^ÿÚzÐ6“…‹dι‰,ãt’`$—J5ãÒä\ßb_°]à×Ñ÷ˆ†Ý^d:èo|B,`˜'0š3u‹0Dn¶ƒÖK*‚š»ü0ó¤Ýëƒå 5'>—Ú}dx’ ,–àùí³ŽËÚÓšU*]!Nô•z‹Ôl4Â2ªb&×$ å@ˆa¤AŽhÉŸâaE½‹°<¹8KÚ¨—ŽœõvZ‘oLvíØWà_ð«”_ÿI®g±Y˜UäÔË‹)ÝwÖnzÄ¡•Ç){c³K8¥²qÓ} 9÷ÃjCì2ž™(×[ÖŒÓâ‚¡Í2 §Ì ÒS`Kª™ÿ£Œ›|Ð= ×›  ½µZVÔ Ý§QÝ펠žC·GYçã #ë”ÖdŠö+]# a%¬DÔaºAB- w“á¤ÌƒÚt°@1LÍM¯Q Œ4‚ù¾â÷‡¸€ôé6®!±I†˜ 1õuÞk§~[$•F„dö½vtwy=omâSï‡öpÒb_óµÃF— ôÌÀ‹{¤;rµ_B¶c^\°Y¯I1L#{ `§&Ñ£o<íDå÷Y“„°åurqºƒ£4åÉ4E0n—å½9åz1Pö¬UM1nD ¾~Ÿ¶-ô™`Øu”’K~0]tÒšFf’ Ç)%÷z–«Bj=ŽMŽ#GÊÏ”²0q¿Yý¦:}\ǶáM“\¥6¸ÜüÍ]\‰4gçÁçG½1—_,¶ú)q%:%K:ì³F –ô d-å‚qã½w«*|JrXÛJŽbçF¹à±÷€t¤¬ÅޏŸ§qFœ½aæp.â1øð¢Ùƒ]œ[0û{ÎE ¦øWÒ?ãÝ3<²aFÙŸÞ(«™YnŠx¯‹Èò˜§3ƒZx¦T`;F5ùh`ôFSHìß&ƈ1eUœ~ЉÉF¥„­Dy!¹øÁ1©=R;‹Æ~¹i’Ú-*‡µ- sýRrª³S8š‘ ÂÐ|T‡X³Òa–—VÁñy?ÜÒ£2C|²ÓÆÑNnV"Á$>€íZÎ×Yæ#®?žÛB9ÌÖtñ_¡Ž´=-5а Öm]rÂäHRžA9tƒbÖÕñ-ßÉÇûs|MövùªQž'¬G¤éeÁë䔯ÂÞ¯âŽd,OÉán²ŠræÔdÓ‘˜êc}xÄ?›µ(s¼¿¥œ3³ºfŠâß ˆH!"ÒU‰Çxü-kvÀÂVÒ¼™^å]ŸU˜oµy“• …Ôll Oã.æ©aÄg4¿ŽÐ qóØ4ÐP!4tmt7 exÜû>Æbº@hYg$(B ³ÑHÈR¤!¿ÝÁL¤I«‚µ%ˆŽ Æø^R¥ †Nª “È®s„:„|U(ò:p¬VEõ*âp×yà›2BüÙirTmRa÷n°vòÁz§"3O–þ%FŠ:pf“úI3£4~Ð.¹¤*ø²¯ 7lòrÞN¿+à'ƒ€ N-`xÝj‹ªù²¥ë@i…ÔC5_ÚXÑåä‡.½÷3š=Hcz$pUßí¾à‚Mçø?Åǯ\:šŒ–ƒæî’ íä@Žñ÷¾ ËÃS‡é“ )o"‹ŠÚv¶ÎTð˜É‹‡C$)ˆbŒÝ¿³X&r7CƒW±/ØÿI}ƒ@âŸw©¨O}lcy7ˆ²W6'B…ñM±çE¿ÖÒš/5ÜÓžœ›Ó;Fè"0|Éq× ñôS´*¡ßgó±nš®uÒ“Ú™|I…Å)F7Eâ‚/yÏ›O¤a1ó² ‹Ûo‰'ƒ™SݼM­â¹$¹KÖ–¼Ç¸»ºz£d™Fj£šLŸ´}ˆ˜7h13ôï[A§Y½Ý‘ÅwÅRB1IUÇnÓ¸ÎQJÜîcÁí4 Ûƒ–]¡ÐÙiÂüã s‹ÖTqÆ&Ä$A»á®öçþ(”̳{š#"¼JªRe@Âìpuù$>5ÎUg8Ò7ß—yâB¤ °‚úŸ¦šF°}6W±Qx`ÖäŽKtZ.ÈËz:B‰žm^>íMYý#\’#K G”ôçk~ÀÆó;›öZ#ÆŽ–òQZ„]ÀGÑ@C¼¡cŒ$®s1A°‚ž“º™g—f£©[ÞÊàÐ?ßI«RïÛ ±âaáv6FÀñd³UE8¢\ØÕТ=¡;¯U{yãk¬„^Ɉ…õ€UU³‚Õ©’4œp}È»­W BÉ›ˆ²µ  žfëÂ$X7DlXLrö4åõëŒéb_/\[e<·Å¯OP뤌Ϧz‰”¸Ñàš@ M¬‡5s[´ͪ."iw­=™Cüy4¤®jV¢³ó´½‚?j#+"«%9"rJc>v;KçÏêpk%Æ\¢IßïÌŠß0¿^1ý¸›Låwð¢Ô•~Ï„»¸Û„~%,ÆÇ€/J’ó´qœ5V[ù6±+Ÿó£Q‰Îº¯êkÆYÉtÔÉôÅÅJ" Ù§ÕnQe:n+E±;ìÙ#Ê|Ä–8”óÿ†yFj~ËïEщ!/ qÃSLê¶–È)½°°åø8”ó‘o¸HÙE‘äë¶<Žtù“IÅ]ºz:!¯ 9tè¥(pÓ ÃG'º‘`Õ¬'€$Zƒ“£yjõ-éBò h-–ÿ* º64—{UpÒ¬Pyügƒéyt¼×!§j{…µØ9&7˜bwC$Ä B:jŒïy ò½­X„Ø›„M{XþJ‰ƒnãR1 9VDòÈÄQ}¤€œ…ú PŸ7P CÔÂ<ÑqCæÎgêñXw·+;¼&íSVΕÀ0·8PÎw'qv_ÚeoÇsW3¹2Ákfº«= &ùÅqQ†Œ¢F¾¾C^nXr—×Ë'ÞæÖÓ*#Ÿ–òåášyå1·8¢rZoý0²1~ôcÿ.Ýtpë–jíDÔÒ4°M.ê¸D®*óÀ´0OߢÝvÊÅY#Z{êçZ‚ˆ¤ÀhêÃ?Ã`l 0íc µ*rÀs÷b= òU—3¯–S­Â¿ w«§%93s©Þg‘î4c£‘ÓùÖWHrqi¥¤ø8žYd˜å@d60a3Vx …àÛœ["Ë­¬ïɤ´ I“Ìx̨¨OûØ_qÙîÃ9gj€;ÒȒ˜KºÅŒ®ØeðÛr¢JñI‰t åèÔC{VHe® Éæ9«\U¼à™jeÙµ„¬À´Ì­ƒâÞàät¦"£Ê~ãÆþp~ôç£ø]FX Åã·;>ÆX¾ô1F ½VÇ@cD,ñ[Œ’Ðo%NÈLOlÌ ]=|@!¸})¢› -H$…A$V@­;Æ$¯ª¢'ZM¤1Â4´8Zã]f[ÛÚ*X›­?ÍÍr¯g`›ªøTÄÎe׿A—Œ©Í4ÁQP­¥¬,[œ&éîÄbËågËÊGJÀ´Pý%K~ùΗ¹õEn—™ s—º¸PõšôãÕ ^¢tCSŽLÌð,EÉK¦‰ÕÓô>$IüÕ<La,—Ǧª¶Êη±;pvºv`ʵÆ ¢5[LÁÒpô´ u¸Ýè •…,nó´÷¹ü̪†ÕË%°J/f-‚0ˆ E d[‡…Óxçrˆ—ö ŃøÄËÚ˜…"•üaºÞT}÷-mé¢[]AΧ€îâƒÝTÔjΘÉcSäQYº³´§CoMQÅúbPU1šovCg|Ê—û⇕þÅ)¤PKâ–õb¨QÊWšB„Ø-Ç(¦…Y‘¾¤š‘5®ÜÖ£-gã6mSÛç–½á)ícÖ ”W“€'°—žýŽÕc9"SôÂÉ9/”c¨¬U\ôÖÑÔzˆ›€ã¬.gZÞ'‹KËi– áMö õÇ‹p¬ bm€8λ#*ºôf¯Ýjʦ뒣´öšýQôÕg” Öt¡R•<.~¢)I5´O)ƒXC,ª.F°Ã©ú.)iêr²WµPæ°P4ÁÔºº&¾B¾cG¾Ñú—c.WýrÏWˆKðûž‰Z—0xù\!ùšk;ᯇf˜H6›Õ\OöÛ“Šúd{¶üîêRÑ/n3äãÒååg ß3ëf°Ÿhß“,*Å Å þ…éǦå’ø÷C‡þf]ܼÞ2“ìû.u¼¥Š@1%‰eΦux™°Î':}N.£"ÈOróoM«™Hëƒ˜ä¯ A·L?ö$ÿ›Qûër|¡ߠö£6ôzU½WªúHúú%ÍLJϿhÑ^¿¤á-/›4vc<­.™f$Íjn´šŸìUePI÷åÆþÏm-I¾Îio"ÐHW6'xsë#DLø›tg¯vZðŠ3"¢É@ÉWIŠ]—A¡  póF“`B‚ƒ°6Ö6 ¡¶¿³kmó4|p^k?tz`È@àu]Í|´„»šY>].xrÚmÍüÊé9>P`Bê+—æÍKÈ—JÏ4æ\g%ëÜ¥(ÐâÇÜzú‡¹ùõcÌLŠ\7OtÌ>=š2'ƒˆ§ŠÜ]‰´ºàÝ—¼Oü¥¤¡ë]ñ»Gr­„Œ«üïYfíHõàç»ñè™Ð d¾ÕµÄÒ½{†UÛZ¹±oN"¤ûµÅˆûô={¦OAÜ'D?ØfÍóî„H— "Ùn\‰ÍÉÉíäˆr7ØÛM¸O Y¸2Š‘L‹@±0û?P|çÒƒÿׇåíìËEþü ¸âxcm/OÏ}ûœÍZF{0>¤ –›ä€¿˜3‘lŸàqäùSþ½|1ûœi•_Ë_(º¡ø+Å›ÿ<·Lný@·õéþ®Ínë‹Ü*Nß3ôGŠ„|º?:²ïï½Ìí;fä}¦“ÙÒ¡Ö{ßÖ™¿kÃý î¸<}ýKÉ}+»¿ôª?ÓJÉ“óÿIá?ý?áD±Yendstream endobj 411 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 5481 >> stream xœ­X TS×Ö¾1’{EDÑ^…ÖÞ Õ*ÇVë@ëˆZGEDæ1Â,„„$;„€Ì s`@QpÖRµTÛ'Ö¡ÚZÖÚªÔs顯ÿIÚ÷^ßëßµÞ:,¦uϹgïý}ßþvTß>”@ °Z°bÍtó/ÓøWüð>ü«B¬û¹¥k§Õ«ÔÀùs_¥\Wº½º°D4Øl„`Ó·aø+÷óvv¨t òD ‚Јä’ð8i €Ìq¬·“ãä3Þvvœ2iÒ Çy¡¾Ò@o¯0Ç^²ßP/ù#Äq­Ä;ÐWç8Ö%@& Ÿ9qbLL̯ÐÈ ©ÿl'gǘ@Y€ãßH_i´¯ã"I˜Ìq¥W¨¯£ù®ÌßHBãd¾RÇ_iEQ3ÂæK„/Œ.Š”-‰Zã»=Î{…¯ŸÀšÀµAë‚ׇ„¾5õí3gëdã<~ÂÄI“wLy³¿-E­¢^£Þ£FQ«©ÑÔj-5†ZG­§œ¨ Ô8ÊÚHͧ6Q (wj!5z—ZD-¦–PK©7©eÔ[Ôrj*åJM£VP+)–JÙS/S¯PÃ)†êGYS¡ÔÊ–H ¢ì¨Á”šB½$°ô§f’ÜS}©Dê©À]ÐÞgUŸfáxa–ð×¾µ}ZeY}%š!j =é/˜Š~öýŒÖ}­+ûïŸÔÿžÍ“ü`»ØöÑÀe Ú2è;[»sƒ_|bÈÈ!Æ!w^šúRÄKﳎìIö硇N0´sØÄa ‡¶Ÿiÿ«Ã÷¼Ê–¿ &~Q€ú]C¡×„|<:Ë~á{z䨭 "åœæ[lD_Þ•ê.ÆÅôV9,ç‚èúf¨Z8‡UÇm^3ÑËåpDŒ–ÒO>;vþ|Ñ6WÇý÷gm»RÁ„ÎQƒI€œî / [»æ°AôõŒd?q÷nÚO<›‹ âEˆ{ô÷on;‰í ÅHPyñ#¸ÌtŽù ¿Ìá ø‹R醌죋™ýÞâU³±HŒ¸“õPg7ˆQò¤ž{Çeæ’wG‹my£¢žaÔu¢#B^…Yd7 QØÛ¶ÃvÏœÐKˆ}úÙqXŽ×±ëÁËÕ(=¨:Ça|PØÐx¨¬NÂaie@—v=xƒ‡zÄ+Ü{kœ'0¶]»¢M]cL‚¶Ÿøìa—[×L¶r“R5š¤ÜÆyþû—{°vÁSÉ÷­hžŠ<‘ êƒm.ä$$«” 8OÀƒŒf^sÕ£ ´äƒo f žkkS´©…À”@v) ë.Þ`Boùáõ$¼†ê;…¨Oe{ 3ƒnÙx"e0hêóˆDg÷úwx€Ïj…ÛFñ;™Kü7}š—¸Û÷?”m_ÅV8Þt˜ëíÓHvéÓgxËK@ ¥$è“J¿rŽéä§|/<úÇå$BKÐh4­@ëñëh"^,Æ#ázjÅïy^JZtÚƒ›7}8Á6‚“ë¼€íË}¦äòç^„ÚØ‰öÔFò+X‰Èï­ S›\IbípŸq˜Åv?ŽA¯=PW%Æ ÿ;ç£< mG„ßÊ ;™åÚ‰($ºzëæ­Ö™ëH‚§è¨f’Ü$$ÉuE±,|Z¨o3ÒÒ) cBª˪ʋkƒžË<Ý×Fsš'"<òè³_wª  2 Ýr…z¾:o/áÄR‰f:¼W€FßGžõBþ×.¶ûJP…¨'E½yÝÑeÀ%h8b9#˜…b¤üš~”2U\ÊBDÓäÁNÜ”!êɪƒ½ ¥®%xS/²åßVù1Õ‚ÆÓ(õ´ñüt H•*HHàÔš;U 5¤e&ä…åm;†ÌÙ¼©)L§­L¿$oÙy=3>_»q*‰K/ÐèåÀ$@J¬¢@Q¥Õ–îæôú¼½þÈöµÁŒÁæŽZÃ>Áý+Ä)ksÓ!${K9`ê*}ƒúêgH28mJ¶*˜âlO&”`BÓ˨Ï54¼^ˆVò#YädÂNÈ5D¤™´Ýy‚Š‘ÜÃ÷»wôÒ>åΆ¬§…øù./‹9Û®½$µm&To²¤öê}aÇ?f›2ò®s&Ñ4•V‘¥†i:’Í^Üâx|ÅÓòÔ t[ªòYpÇR•(z¶:¯IŒòÐUœ÷gO“Š lîct—ŸÏö‚g>=ÕÍuîÔôÃÇ9ô9Ͱºª)b<Ž^Ð[ö^ÉÑsÚ™"ä”ke¢§d£'´-?ø9-Gß"‰yŸ.÷&WeÍ"ѼK–fY¢©¢oê¾æh|U .2Hô¹ç'deÀg–{FÒ.ªÜ}äžçåü0“€_…n±Ug š´Œ‰ çm‡°B,çÑêœ(ˆ…Z™¾“(«Þ[£ª´È‚ ‡ê½P™èmª- Ѱ‚{öœ}ZQ ‡ÛgïÈ"úVÙ»ôùfIíyg×Xô9›SYÒÜf~gyç6†ðÞýÍ꼈¹*MêŒ3ìG!ƒ²té`Ø£­"oôQmD‘›†ö¾1CYºa"û±(_“™žGž­Ù»-·ÛJnGä'¬çÙ㛞'A/ã'öY;t;‹¡2óôOQƒýSÜ OÊL.‡bÈÌÍÌcÌÆˆ¤5¦û£ ù6 »ÖAø<`9Ÿ4}^ðÒÐò8ƒ±´¬¦jGI”žkª> ògÏú̇ЫUKTé¿‚˜©ßG|Üqäà Nïkðm&÷NæWâ3!ëèí;¢ýàh7ç;nF„·›gcëG.£IzŽT‰^@ %±ÆÆÚ†:“©±ÑpÑœ7’·­ „÷zkUªÎ”C$)åòD×yö.ä$ñ¤NUTmÞá­Z É–\<ÏÛeÐjN®üÛÙOr”Jü¶ÏQ;ÑMºl"¬pöôˆkRFr6ì†Ýù…:}u]M¡ö¶“}™Œ€.RR#5ŒY‘@)yèÔõlZ }D=ÓñùöÈnƒYv»–±=ýÊB‹î¯ê齪[|/ƒCos<ê1¶uqß/Ö-gÉÑE»k›ZÊŽ½?¤,´$¼0æƒ úoñߺ%)ˆÀÏ»:b¿¼´§‘‘³ÍÈÿÍ$(ÿRH þ9[}®)×Ô q/H‡ ½€¨U&ƒÒ4©J¹#VÚc¥‘d:TÕk+†¶ª¶´ýá'¡P]⋦àŸí·Ž}÷]`¤ ù ÎΨ̃ZÆ]#K ñ<ìwº³õYn&©cˆ¢í»Wjév„|_~+Ù™.9^c4”ï®»°¸n<ÇBÜùi,ŒF4"ëüü$P¥*5 —-qÌŠ)­h,uõtGÎ)ˆkÛþ¼í¹:ŒkàOݢ쮱lJ™F'!zïò[ÛI¡#`óë¤sѵ€¬–ç]‡~Z bÞHkµHß³ Á¡ôùú|´«Kh¯ÕXu_øMønêÌÐ Œƒ›¿>~w¯ÄâwDá V¥ÆáÝÚÿ$ö†!ºmvÏ%ÑŸØ·¥ÌÜÔ Íý~àw£Ñà¶cÆF“Ø•FÎ¿É ÿ=§ûKЛ¾#Fâí¹<À»*²)|¿ò‰åýœƒ•UMGŒ' Z#ëܘž¤{Ðt‹S̾,ä'“†_‘•îOñü/G–Šhcª µQ‹i0q¹èüOßÕ ’㡝Ùeßžƒ/™ÇxÈ-ìÄ᛿qçŸÍÁû–ý³9zÈ 4XtL~Ù)µÁ¤å,WØvÈï°ßYe-0Ÿ—ݮ˂‚„DHNÕˆ•®QXiGãÐfHRê9ÅǸOƒéb”%ƒ6¯J—£ÝeÀÜCö€Ç`á,ÿ‘b[Ôjñ°È ͘"7¾›­•Õ…EÈ‚j£ åµÃ+ Ç>ÌI#+y¯³Pĵ4öädYº~¿˜§ˆÁ¸K†qu‚¶‹èÑE!JäÉÔQ´åœg.½ñ.šW‰§ù까dmjÁsG‡6Òf§™¢V%§rk7Y^´Ž¸öþoN=ç°Wi¬x¯¿)QÒ/cŠT¦€Üäê@pgm˜?÷Á¢SHœÆiÈ9‰½NÓÍâ4wi3r³¹Ì¬Ìã¡‹k¯&}@œ¦ÝŸÝ]Ó9ºŠÄ~Ý,êŸôïÿà[} ‰¾^Z* ¬‰2š*kŒžl‰þÓ3zèðô×íˆÊÔ'qYå`djbË$Ò¨øg%>kÿðÇ'ÐÄœÄ>; @÷ˆ¯ÜØ5ý…Åà«zQ!z§²'M¨v©KÓµ 2Ý%t$¾Æù‰štˆ‚Ãd‘M–ÄúÒ¯™<_BC%ä—gædÔ’ÝäˆJi!~ÇbÜÍýKO¢Ó¨š½sù£“‡ôÕZ©J$ÖÆµž=]“™iŸ’†GaÆž§ÿ ûu§‰)3÷Èßþ!}ù´ űEÑÅ1Å1…1å°ƒ‰UɹžÖ]Bfâ²^TxA\¦ ±k{ ([\'?Èd6Û¬Z¿·tĭį…'líéˆßÃæ1~Dãᣧ^mëÑrŽ·ùS×ÝÆ¯¡AŽh$zù ê‹^åzðmNŒÛkÿÅ)üÉæ@OÿS ¶?÷{ñIH˜ùKØUA¤åVÀ9,˜èá–ÀÅ]YS¶ ÆÇ¢¤pæ¯4róØÅö¼ëþTgþʳDÞÉ\3ÌÄo%³Ñ•O…<×õ [»¢“4šD‡¯t{[¡Æé´ŸŠh¬GƒÎ‘Õ«A2z¶ŠhÐçOM#è¹¼Ûm…³ ¥ÑuO£EFÄÞÉ/<#]Š Ž|[~@qÈ̬k/ÊKÕ)ÕjPîûìð¯'fÒ½($_¡W“¹‘I‚ä81>J/^¼K—¹KÏåí;qÎBGNl‰†ø€Ÿf½Ä36Ô;d,ƒåeí:Mñ`Œ±bw}SXAL\°bÛä“ìž>DƒóöØzÓ–„@_säÑå|Ûmr=´š˜"ÿ”ÍÉ#qU08[$…­4Ã-S“£*fBDØŸp1£(ïãÂ]yúO‰1(@4 sX„#ñpqP€G q5ù\+ hT²>ÖÎXòPÿð¶‘n4?@úo…üpÔÊbf¼VG»éêùæL r™:YId\X²ïâ³Ûï£WÑP4¡L_…­GLÄÃ0ûp<²úÛá#'ª8üòbÖÂÊ£šdMÄpœ -áPÍÁê½{ÊZÈH_»T§,w×¾1ŒË6ç)T#!ò%²u …È“ïÏ6„WH£Â%fÕj¨0«VGßûÙˆ7”£yÏ7“jþ‡ÍÒn! Ë¢W{nJKQk@ÉìÌIÓ\½…ú›Ò¯¬ ŠŠ -h,ÉÒé³Í‡ò–Ch 9p ÿ'O/âT ¶ì-¼sY·EÛ% 5ï­ÉÉËÊ"Nv2Ö•ÿ¼¶Lðc»}‰uì®|ÈÏÔh“ ÄøUÄ¢•€&:ˆú£WN-,ÐjAÏèÓ³ ›6îL!yÃŽ¼° `Ù¼ ƒyq&R‘C/  ß$ Ñ dí~Ò-â$ħƒ:_.FŽØa^xàÄ(ßè‘DxBVêRs‹ŸÈÙÅ¡—‘£-dÃô˜ âÐl’ Î<€DYùŽ_Ä&k” Ö¤á¡H züuÓ…ººÔØjN퓲=ØPQØsð”[ë4'<ÔÝüydŸqk¤A¨Í¥e˽ĸô¿?É èå U¤«ŠX…sªTD0ÓzŠøù ds&æäæÞ"Væët¤ˆÏÐòvA>j¢´œmÇËèœ=ÇØ™G‰ÙIödºŸ$äýº6±ú|È„ ¦@‘›”.—+9Œÿ1K)'’Ú!)WQPœ™ŸoDÑÆ.k£à`'ªìò‰è Ï’Ûƒ>ôüûÌŠu€ƒÇøD‰ÿ|õXpƒ¥º×[gw5ö¡Æ—EÇ7]ÑuB;ƒe¸ „EÅÑO’ÏÁÒ¦>†OòΔÿp#·àãø²Iùë`.±NïÁdùÚè‘‹ƒÌ}ä÷˜š >ä7…èR—[£Ö†sÛ½7D:>kHf߯aºs‰Dʹ€ ÑAÝ—p„¬»pÐ’ñÚY•ßDD Òðs‹ÐüküG‹Ä¬”p}÷ä‘Ipê ’}#Dh[\tÈT\xá\ÛøŠAýÇ\'ŠË¤)îûS Ëë‹öÜr®¶ãRõ1â|[]Þ~mþlˆñ`c>›|î5Ç™«çøU'êË*j›‚ £´Ü¡3×u»1\Kõn˜Wb‚84D¦^™«IÑ1$1JHE}1»Xxñ𞗚Ǘ&<õÝ;~¸Y—ÊÂ?_ü¶%…qAûßÓ…3Ù„­þßüˆøÓ‰G?±H@?¼»ÿrÎ>H¨å"ª(P@JFT¥B%'S ’Á}鞆ÿ{zK- ïúî¹h§ÇJqœÁ§v)0d ý+t—þO%¡»SnŽüñWMëêRâªÈUS½dù‰Å©Z´-é— Š$wW¹{åÑ&ëký9ë¾o—Ùô3êml®Û  ¨ÿ2ð½Ðendstream endobj 412 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 1041 >> stream xœE’mLSg€ßÛÛ;­$7QÚXE—˜Â"_ºð5B‡ Ñl¥4´”¶Pî¨eÈdÎñ:ÀŽŽ¸82‰AQ`aÎÍÁ¢¤a0ÀM Œ}d²à\vn÷‚Ùöçäœ?ç<Ï9‡B2 ¢(ŠNÏÕëþËâ…”°K"„Hq­?Ñÿ¦<mO;‚rò !]´V…•R¬”õýÓ¶ÎBõv( @RŠ²×œJwT¹–r3Ïi»¹ØÄD]$“È¥ÚLN‹Ñ`çr ¼Ùd3ðbQÉéF‹‰wsÚýfž¯z%:ÚårEl5QgyÊîHÎeáÍ\©Æä¬5•qvžË3ØLÜsÔ¨ç1Ýa«z›79¹\G™ÉiGmMÍtòµ†ãƲ=B¡(éQ!*Bi¨e L´Å#¥¨Œd(9ÑE4Kɨrêµ,‰•ÔKž gTþ7°2Ç„}Ý”îÏb#hre3Kþ#=Gäð =OŽÈÿ¢‰OxÆÂ²)'-´Êˆçá¬ÜßQץ ,°^òê \цۚÚ0ÅSüæ!¡q>–ÖaÝ:ÌÁ9ZϰGˆí¦–¥ Z]9ËÌ(tMy¥ÆÃÌ/)E‹ •Ô0 ³÷ÿî¦=”B¾È‚ö>ÑBP ŠÉY!Íô«]ÓÓ_÷tôªëÆ+jÌl890¬Q õ⌗E»Ï&Ñ$„¬¼Wõ>Ï4Ñ×S•sIPò¢Ž†ì|ù¯´JhiðÂÎ1§†B…O*¤Àû$yR›Sr¦W7§]ýx¤wWÐD{š„Wâ(¦¸/yåÖÂùëwÔ¦EE<„²Ø…«ÍµVwYÓI̼n¸Ü?ÉÍCþ½àfá0Xå{iJH$Èci8J¬¤”¸å?Ñ ùGòå«"X¥¸‚ÅIꚈt(¶ØZèÎÅLtáÌêÜU |Vo½­iŸ~­·º=“ˆ}daˆêw"]Sßz.ß'‰rK^¨êÛÀùE©`O¬#ÃtöçYÍvÌ„&…’ ´úç½Ñ+}ýš"Ÿ¨À:KMU%˜ÉL—n>¸ÑWgùD£ò§‹Ï2ì…Ö1jm ~ûY* /ݼæÃ_1O‰ê S“Îͬ5˜€DŠÈ„|Þ0,´B¤3…©™ ˆRdóÕÎ ËÓÔØ2|¹,Nù£Øì«©˜!{âÂö”¶—ŒÓ –7Ü{wÑÕsúSÛGÇ;ØÂäê³# ¢»&ö«ß‰ýÂþÈ2 š?ÖAöÃÑa—Gc8v!åR^‡µµ¦çD;߃û™»wG§æ­ÛjU­GȿŞ.M Þ-¾­ê-2]·ò…±¥Òס܆пZç%endstream endobj 413 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 1457 >> stream xœ%”{LSgÀo{¡½*"‚×Ç”Û:_SE_ÑÕÇ¢ÀøB­PÛ¥X âƒrZT0@WP ¼ AQqˆ ¢“qŠ (™q›aHÔùعæÓd…åKN¾óÏù~çür>å$¤³,0Øoä²”Ÿ.àgyO üåOË=© ß­õ¤‚B=׋f¸‚ .Nu3œï¸cÆDŒ€Ü(Z ÐJ“iã’tj¥J/1Oêãïï·@ºØÛÛ_ºN£Ð©#ä±Ò@¹^¥ÐÈõŽ$Fº]¡V蓤sW©ôú¸^^ƒa‘\¿H«S®™·@jPëUÒ`E¼B—¨ˆ”nÔÆê¥ArB:‚ºh$È´š¸½B' ÔF*t±Eõ]ºÌo¹¿·Ïâ%ÅR“©)ÔTjõ5‘r§<¨IÔ8G·”N}d Þ “h1­£Ûœ¾uú™Ïtå»ÁŽ“ìü.«ûh¾—³vBóÁâÈI8z RÒ9òûçýÎJlñ«áè .N)j5w™íÐ]ÆVc«Ë,{œxÔIÄ«ùÙìg!); jþyƒëmèñB€ãì˜;D£ßÁÞH­‚"`pjß_…ùdzOyJ,Ô°´jÞ\–~ÆYÀ¤BÚr]|Ž3™ ò¸æŽ{¹õp*vŸÉ°DUÀ6ÐÀVÍ>rü.Ø ßÛuÝ™?•AÓ|07Õ¸B||Ð'¡Óº£óÒ÷Äm_Xšb7Êdz¶§Cõ•‚ö×hyEã||ÊÂ@No㽺öö†.¨‚¦ kœÎ˜xb]qBU•ÅZyGÞ6—|Cæ©$š ëj¢ëÇAœ‚³ÿ&cVî Ý›Àá ò€Ý ‘•‰W’«2Zá.3|çÉÓkÕI˸‹ “Ü ò¸ÈxUôá]ÀŒÀ”ðÏ "†h¾†Ëæ€ÊR(Š€4“2;ì´1'³˜Q‹T$ʹNd.Ìë.ÊË=û ³ŒEËw&jɲö(Y[™4o¹WÏsMbÀЦ?›ÛG_©ùD;Ú½‹=7iDüÂö5¶¶@3<»m>úÈì)U[ œ%1çT§j{üÜøš!4ŠQ„ïŸ!Õyk[—𫲦ä2Äoˆ=T¨­€æþã†G}=»Ö/žµ×h4šNçdYô\UžÜê04míò=K½VuôvÜ~Tûÿìk†žÙø)˜WhyMóËðKLd‚⇄PsuwétvœcÊŽ”j4I†ØM·åÃ8 §;Nšˆðµñäö`N > ñx»Ç(mæMÛê¾£•r‚K?FÐNã½E£Æ"3m(~×Q)xþµÿÒXŽ/ؾ[!Râ&;¸-bI~§\Òbί…¦&þBŒV{äÀÂ÷A8%¯ßômé#Â Ü‡ÆÆ'pé÷î$4¡—…¬T—®ª¶^°UJbÌ\ý•žü`®ÞOð•§„Š‘D)´Æ™:Ç/q2‹1fÓRÎA±ÄÕ`åe…›wÎ*"ò|±}lÿ8n¬“ŸÕeŒ-×Å¥¿Èe> stream xœ%“yPSWÆ_x<”JQ_]yI©Zdqas£ X¤€q@ŠH (!1 ("kÅˈ€”‚ `HC¬)•PPjuªVÑ‘¢U;2n£¶÷9WÇýçÌ=ç;¿ïûæãöv‡Ãq‰Øì;ñXÁÎá°síXW2Ù[oü\ çà Wâ›È×uõ\f 8‘àdo™ë04}ï‚ÒQÒ§ÉáH÷ä‡ÈäYŠTIŠJ°0y‘À' ÀÏC°ÔÛ;@ð•T¬HM¦ "„ª±T¨²-i‚hYrªX•%X¸*E¥’zyíÝ»×S(UzÊ’5‹<{SU)‚Íb¥X‘) Beé*A¤P*L zNŒ™Tž¡+2‘X‘NĤå+|ýü¼}–.#šøŒ˜AÌ$f³ b*1˜NL¶©%ì‰â=§’ÃÚΤ‚ì¶÷¶ïe‹§°¬ÈÅÊÆj9,1J²CÈŸ¶bG6’וŠ,Mq^!ƒŸ¿Mp¡á•¼]çǤ‹¸ÚÒÇåЋµ%ÝNó­éÆÓ1)/<3È ¿o­øQ-”EÙ´K.Ý/\ˆì¢Ñ4´ðÙø³ëÑ·ðd#s[?< w©;K±v\hÎÓê[ÚZ3äLWï#¨ªgXé¿%íË/ŠøY%rM!”ä—À*»ê™)¶ª„Ô!Yõ kxÖI£“™Iö~Z'GS¥“Óh½Ó'ñ?—üÛ>endstream endobj 415 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 3588 >> stream xœ¥W{XSW¶?1prD”žŽÔzím}â£V«¶VTZµÑj‘Bï$„ð ` (°7†‡‘ð0€ÈCÄGµÖVJ¦ZQ{k§:j§ú;Ó~+Φ÷ÞC‚Gçùî÷åK²³÷Ùk­ßú­ßZáãGú-ùö–e*Ïòò8Ë4þ.œbéy´Öq1Ñwå4"`}æ´U5äRWäÂG.Æ—'wL†Ÿ'AéDˆt#ø<žDží'MH•ÅDE+:ºÌg»P­fŒy:!ãC¦ÜØ„ªÙø?ܾâÌU7Y­§³á°’¼íÐÉcº¤ –Ç TRîT3Ò±0{Ì)’³·Ëƒ&Þ3è9{Rˆ§ÁuîCLajîtìŠ'ÝŸ$÷ïÃ$Gâù´_à·wo |ûí@À¢«üXî ,3ÁœÇnkÍ|ˆÆ2zŒ/ÛÈqWZ/¢€?ô7p‡ Á‡§I£Øî$¤|*¤áÚ±!·¤Ò_œûì«‹çü/ð_·Ú'òæIÆnñ¼‰×a†}œã[¦ÑuÈ ¡$)#m ¯Fv™>»a×û³€üï}EE,ß!PIFóo“ºœb”Áî ½áïô†`_oŸußüxçÒ ùjïÊ@Æ–± ¶ZC›l>Çç ‡h˜LF¡äÎNd¨cî l†“£<¶BÎV2Ém{WÕe±Ì°³=¦¦ay‚}ey…l+/mgÿ:z‰<…yïØ ÷›x0Û ~÷ø „u4ð®_Àh#šsªÕ ¯=ÐzøÝù›ƒe2Öh³ƒyd†x4<3î÷¶¯ ÂžöWHÛõB„›xÍð¢ù×jh¤áÅçÇõ"„îçÄ5ɉqX.“¹yvp·ÛivÙ&¾eµÅ“Ö¡æXÅ®¬Ü<&¼Vˆ\ºV$-‰ž÷[úí ötöµ,´Š‰ õ_*™]Õ»+%ˆ!Y*»P BÒ¶òÚBc9Óû9ÚÇñjÅþ¿ur‰°BcDñêJÿŠõåèµÏÐ6t§0.UÃhãP¢FœgmÞœ5ý}¼ÌÂ|H±ðhXDêš9—¥HÅâEß‘¶èe©Ì;Km¤)Ÿp ‰ sÌÅ;¦K%(;q§ZVÀä‡G%ˆåƒÎÔ?¹þI [¦ÐV´žªÜHr¹¯J´8ÀÀ~Á³ûZrþ€ä‡SººfÄ€tˆºò¯•«‡_M¼jË.¾EÈù_ß<Š7v"Ÿ”ûq2w:y ª~u«êô®êH$B»ÓT[e‹óÕyéT¹º {Úm={ß,ÕåÇTT©;ÕpKS©­ãäJ“U«÷ÝûQ‘ªÕ¢zTÒ5*]s¾¾gâ2ƒ’ ÜÚé›k¯ài¿†V$Ô%ëš*iTªÑTi+45¨Q߉ b=ñžèsaÝÃïN˪X}|OÞ D55Z]Êe©Yï‹4R"EDJtjDA&¢6†·wœ‚qF 4£Ÿ7ï‰y[ŸE}¸€LE² ×2ÿV’ð Ï•œd“¥ÞÄ;¶†C»5¡bmEQyª¥Ú$(Á|2.O“R“Öä5 ”«X•¢ Iz?WÆ!·7·6ˆ`ÆÝh…e…å卿©Çwi“ërÙâüfîqðÄ¿M)ÌÑä¡|”—øNvjZL¢"2¯¬nJAk’Nލ„äT±°5´ÿFÿµ??&We—Ìû€o‰‚ÞŸV¦ÌÈ@Šl&5T¸3 QKWs«¯õwp?P“_°‡-,Рl­DŸÖ€(cÝþ¦kÿ…Bðš˜y »ÜŸ/À«­÷kžø `ŽÖÏÏ&C…íIઙ?s.vÇ®wgÁ¸S=†æ¬¥ažâ°7ü‰Þ²|å{ëM·ÿj2]>×·5ˆ±™rŽH7 ê:ßâ é:êÿŸ0'ΞÁu&ç¡×àe˜Ö{»a$Æ4uA’šMÞ±M‡|‘ï‰ô”fÖ] ÞÌjñú68 ·ÃýÂfL_u¤»ëÐáSÖ"²Î£EôìøkŸÕ©aùÿQŸß2ÚŒ&ÿÅ ¾æÃæ?¹g‚ÎÒG›:úЗÔïÞÄN ^c¦Üúa¹Ò¾'[³‘pIT‚tï»JêQ2&Ö§pEˆüQtSt}|[.ó×M磜 Š+`?ÍÊ¡c'PIÎ~Æ=SmH߯¬¦ðävz „_ ïôl Ûôû˜=—àÁ§{.XÎÐF¹A,–ËÅbƒÜh4Œ Äî´}Âàá~<ž´¯Û™2®’V4­[ä&^“ÜF^|øÜÒCƒÛsÚ­Ÿ™ÉúžìôÛ°|û˜¡få¿ô]åIŽò¼ó|ø Šè-¸€ƒc›= x®=‡Í$ö¯8›ÉuvïËwŒ‰ÅkÌôt™¼~ ñ뎉‡]o¿^bM²É’Áõ3¨8^…p6±ÿâ˜ñß!gl èbàN“àßœºGâî½94tïÐv !٠ߨ2á…cü4,ǓƬoØýç~2põräÅ·ˆà4]ÀáØpUL"“rHZ%á:Pln|Bn³©gë³ÔÊâ~Aœ|”à·m’"Gq,ø n9Ö­/Wlepš@ýXA¾ýŽXûÕd¡¸åêy~hèìL®'(©ø=™ h_ß0¦×ÚÝðëe»¼ù¯äÁì²&ž”¡Ü‚L%… “4ÄÃ]G|wdä„ïôPÊÙ˜øðû‡|PX.Ò«ÉÒŒhF‡?Ñ$kFäÑ‚´Ž‚`ˆ Û¥{“‰é1Üv¨V©IjAF¤­×ÔYŦÂb˜Êƒm0•?ÂÿÒÝqÍÑBi\´°UÚÞÝÜÚÎàÏžùmį]½Æë´lã[6Â}º¼uv Gòô ™!D¢NTÎÂoØ‹ÞÛº:EHÍâ’jEts™å.Xª¤àâXÉQ-p jv–äa¯!9ÌBß¡îâ+‡Íz»úV–j+Q1U•]+Å¡ø%\‚Kûl3”ÕVYYÄ÷zàFšÇ~â¸u´4j)úêá©úÊÍè êôŽ‹x9žê¿9¦!ÝÐP_w UeW0FãçÜdDýÜ#'É3sv³Øëíüw8ÑŽDéGšu…ÝLöV@ìηsÖþ _ÐPŠCá¥X…:-cg±º,—é”õ‹zÞmÙ¨{ 9Ñxž² 53gJ§2ÊÓêÛ€;%#7ÝÑ[^ânJ½4¢äÃteñÞR¤§•ªì‚‚Ì$ÆëCm‡¦ux4fÕÈ•ªLq¤&LɼþeFiž6y¨”;••¹UŸ²ÿz-(+7#)=õ;ueÜ·ŸùùóüÈ‚0é‘X™fÐ몌=ù̽ F•• ¾¢Q÷iYV)‹Çu‰Ìôȃ«Q)—–M(ÚÞxÈXuN4û‡¿ ¦<»€1MyÎ#kSeõàdcÓ>théL‘\‘™—’‹’¨´ê´Úƒ¿ŸÁWÂ#Û?‘ÅEÅÕG-/Ôhö0v»Y¢_?bGµ!À>$Ü?<ðœ1ëÛŸ2‰øÏÏò—;0óï¬eßòÎLº¢‹ã©e±ø¹3‚cc'ªÃÆäk|8ûh ]ºDÜnóÇ2ÁNÑÉ=UÉXr¿M&Þ•ùªƒ“ô?zWmÛ¸9pAæt\Äj4H£Ñ—#­¶œû®EETCJ½$Q‘³ð~0L‚—ÿòÓOgă˻˜^cû)®s^_1ˆIìì·Ã7®>µ±I¯oªÙ½gw1ÓÚ8€Juá¨(Bµ;2:ž•ȤùqùìN+@ñÔŒ«ºÜâW +ËIœR,09™'‡T©Ëx“‹‹ÙeAü£1Ÿendstream endobj 416 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 1918 >> stream xœ••yPùÇ{èi”…Ó9<º§¢%¬/ã*‚«‹ŠÇÀn«àŽ0À4=Ì r, ×<ä>äcÏx©UgÛ¬2Êz•k,£&{(˳)ýÍÖÏ?҈ͩêªî_ý~õ}ï}Ÿ÷ZB¸º‰DµxÉØg s¶Ä9ÇÅ9WjÁfgýOÝæ^¡!s‰Í[òæ®k!Wz‚‡<\ÏqÏðF=3P®R¿KH%a_˜NoæÔ‰I¼Ü7ÁO¾88xÅBù’€€`ùZŠS'Äiåq|’JÇ‹‹Ty¤.A­âÍrßUI<¯_¹h‘ÑhTÄiÒ:.qµßB¹QÍ'É•ªt—¡Ú-_¯Óòò-q•|<[Åø+L§Ñx'ÐíVqZ‚ ~¡Õé¹tÞgŠ7'¨Õ©_?Al%¶Û‰H"Šøøˆ%¢‰0âcb±žø€ØHD[wbáEТ%„+Á§%ó%Å.¿tIw¹-•þÝu­ë›Û§äJ²Š¼!ÓËÎS3©eÎzOg=è– ùÞ¢Rg  ˆÞ°KeÊdŽCcNf§VÛì)<?¹B>dcôô¤A6ÂÉÑá3ç…†d%ƒ÷ÆËÞ~Šã èo‚Dp v1ÞN”J£O0‰_<{çÿüùwß#oïÀAtXÄí‘GCW®] WøG„„±¢æ´äuÒ)ŠÅ=%“dÒ¦?®µõìÑš‰¼Vü0/4fW^>ÛM?/çÅ‘)å.ã:<èÙ‡ÈûìíÆ¥¿FK:#é•wDMôÊrçÕ7QKšïïƒöæôvw}[s`šæ4 ’Ó•:W‰˜t‘ÝbxýX!.“)w’º`%þ}¶’²EÅõyÝ dùµõ퇻I$Ç/*ò ,¿Ò´eYºÂòAhÖÑ'ͯ˜•_@³E †Që©sŠ¢Û[á8r,Þ¥ðÁžØû©òBô¿îª*íH5îË*d¹í‹Lá—?ýŠ*»@x༠ÿ„öÀŠhjÌ>³­· vo— ;P¤ãäçR¤ü‚Îó­å5l[vÎ1 ¾½vóagaoF?ÛÕh+o*«Jí±”ÕÐYs¸5«?Ò[œÇ&[Õ5: Þ]n¨TJdMº äBR‹®&?JcüvSAO7‹÷:óé¥ûh>~’™ˆ¤£cÏ^¼8pöOçÎÄ(•ŸÄ|Äöà§´_è٦ʣǭ¬µ©¾êÈ–Æ„îò›`ûÛo$(O„ê‚sˆî4¶qœÑÈqmÆÎζ¶N‡à9ôcÏñ0&§0Ü;É©/z@¾Dò¥¬³@lƒ(zæ5©ØgÊèô|Q€=§¬/Mé)OrâwvÎŽ¨±GêŒFvºëÒ$õÞ=z¦°ÉpbÄBª¶ ßÒ—f 4ãrVèeñ°,ÅðJúþë.KeÑfÙýcçΪIÚÉàYæë*nþ_GÄä‚ÚâŹºZJ!Jì«È®Ã̈̓QéCó‘|´±®ºPuEÍéÜÚíc˜È÷t XXúÅém›nn9$N¡Ê–Ú)‚}¢_£H áŽéiبz`íU P|af~nuÑÁ\ѵåx`7X¬Å>xžoN~AäPùU™ÖÎ[‚c€îu=†¯Çä&A1ÈMŠ®Úé~½M›Âq)š®´Ãý¶îÆ3·ÁV‹Ö4Ø\-¦9¦3Ó\Í:wÁÃÃáñAü›“;endstream endobj 417 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1ƒ0 Üó ÿ €*º ºthUµý@p”' aèïKBèÐá,ïN>Ëáz¹² Áá‹"Ë:ÐâÖ€#M–EÝ€¶ Ëgå…nÊ¿?ž`3Ùù]Í$Ÿ§¶Í«z¡Ó´x…O$ºªê;czA¬ÿ¤Mq6ç¾ Áì?”M%Ž›€kÄ17ÍMRËô{Æ;ŸR°A|C-Snendstream endobj 418 0 obj << /Filter /FlateDecode /Length 3539 >> stream xœí[Y·~äG à—žXÓá}IÇŽ±ÛF’‡Ñ^²°‡´£Õñâßž*²»HöTϬvWrŒû°6ºX,~Åy¹½\ üþ_.~÷} ËóíB,Ï/2}]ÿŽ/—:‚RÉeì£Snyt¶ÈCåÒÇ^E#—.ØÞµ<º\ü³;ZÕ©Bw½Zke»³•èu0RylQ¾R¨îxµ½¹ÐÝb(‚sÝ6±»¾BÚôNªîôÂH»gi:/cw:öUÝf ­6,NÓEc4-he³àO8]ðÆÅn“§ŽÂ­ÚGkãJDY³´¹ÄIT Ñ]‚ñ¦;Ï‚æ™TT†gytÒ ˜ÿçqBèØ=M+kTÃ9N&bJÁ²7ÇUŸÓ,’ ÑtYZi×=Á%]ïƒìÞ¬¡tæÃSªj„‰A=FxÔpú…ªåLX/z}Ÿ£”ÆvEc›E¤ÌKÚ×â_ã² Qè^¯þ}ô×.'¥O::¿9­Ö¸,ä 4å&â¾Í=`F;'y9Ì–Μí6ÇÕÈSd.j¡Cם֯HP…¯ð|m·Æ04ÖK"Ëk­c/œ\®¥ì£µ2³ž– ÂÆÑ…dèΟ¡¸èµó®³ÿDØ-ÉG 8½ÞÁ9¹ö—Ê5O2{Q‚`¨]gaØ-g™¶@7»…í}F¾{}ƒ=¼@FºKtßo«aÙÈ@ƒ#fÞìÈ[òðmã‘Ȳ·Æd9иÇ%bÁuWS6™ý‚LÆÖë6×¥.ŒÙ¸*ZB•U¨R fÃèd›•ì…ˆ÷ì.uCÔ)óµŒ}ÅôûWWÈÕÈJ5|wÁŒ¬œg›b¼F½+!u³ ¶YNÓ*Šv™3€9©|ü»ÖNöQ£†ëÐ7ít«¶} QÓáöµ9jG9-Ѧšb“Er ˜¿jvŃâݹ¯Ô÷ª¾i—Á^êaË»,f…íœÎúÔ”tÕ´¥P/¬ƒ`—¾KØ9ÅéwE1XzYØ¿mö ö€HBœ¦æãÒüÁSÏ7Û3Ÿ-pøxŸ¶t‰1Ìðg ÐRCH$ð¬z‚tBJÅ9¸Ô`3çG¯ú’q×wŒã^uETOÔçD}AÔ7L¿¢Îx—÷=ê#o} £$¡ê¼VB÷Që%½‚#9úãjm("æ'…VÆ£Ÿ¨WUÛÐ#'6Bú¶Ϩñ†¨S¢^uË|½"꘨wÜÔ×Ì"—IMtxümqôÛ9 Išk¢.&N¹!êQÛŠoâ¬ðóšë¦tÜ2Ÿß–ÏF'ås ¢Å«€‰ ˜Hý%cÌwŒY‹"ÊÊ=QŸõQß0ýNˆ:+\ÿžל¤ç ŹÎ9c–§ÅW{Ÿ—#7²û¹EŸ?í=¨¸_1Tùú)3ËcÇ?ž1mÌ|—ÕÌèpbôZSàºaDzNÔ-ÃÖÓöd•aî6Mûtµ–ÆyY…-îL/m ê!aõFŠ"íkÆï*yÈæÅOþNÔ· «[f¹¢û!7ù¯Ù€ì‘gŸ†;i>¤ÏÖ+ñð'¢Î¾oˆ*ÛòCrû¼„ÕÜKÜ|Æ´]”i.™Í_‚D9ˆžu;1ukÆø|H^Ozk¼^¦Œ /ÁÝLÄÚb®'$&sŸÛÑíÊðR²ô9R£¬¨ÑÙ\µ }<6«ð¡ [ØÊ†Z ­¥J¬¥Ñ=xHÃÍåàrfYMëN¦‡L­hjûP£UŒå¸Dv²Ô¨‰ UZÇ´ÄÝÃdóËC— ùûYY¹õb·pg»Í‹£¨ÑÙX‚i‘“¬öå:;^ѽ÷ãU#z‹éªŠnuzÏ}{!÷Ý:°K„«Þ&ß­‰ªŒôùnž¯ã’0O½äæi'.ªÞ4‰Þ\b@¤ bRBÂP‹`ÝVOÆ'PÕÈÝT…ßñzÚÞ_GQ+(* ziÀÇ‚* ._œ\T3–ùš¹ÎË?¶½ˆÓÀ›Ì„Íå4H3Žá„qÍ]¸åIÈZ¸#QKÛ[ÉÀ/ªd_%‘8 OèØ/({ƒ ŒV×ú)Qní;È‹ê•7jœïaé Çð¬Ò+1fy†ò„rùGWS 7‘'P#fC,4ÐsF æq¬4…´;{%5Ë´É€„];Ù%ÔXø˜AfÞÄ=½ÉsxéjŸ'Äô]þ¤«wæu½pa>ˤÀíqå$©Ô³h[ÖDœøkµã‰ÉªsåèŠ0¡›ÜÓWmªŽ ¢g=ƒbh¥ÆÓL5t8€ò#TŽ¢ã­^¿ôÍ)øˆ ¤m6ÚF ÆŽ| ‰À¡rsëÕSG¨Zª«â'³xNábÔÜ~± Ž¡GF ÜRÀ˜ï‰úš¨¿T»‰ÝÖ]-íÂMžßV]ÕÖ:ßÀìuÐ%™ú¼ ´l¬ZÃIÐ{ë&yèåÞ\½dò%»¿"ªŸÜEÚû)“¾—„aˆ|‰ä#ŸwxˆÉâdÓí.‚Ÿ?§#®†³§‚ûËT@A 0ÆëKaTïàpY÷>‚;åjÉ0É‹™ ±’;E˜7¾rÊÇhô>'‚tñu€åììz{žz‘³Ñhœ“ 4éÙéÀE÷  ÊE7H pbݥ‡ªÓé …ÓjpÝì霶M Ö¤íÏZT;éT ’ H^j©&ÕÁ1K½Ô0.µìU [È–Ú‚^U«–±­ÒõÀ£ß6ÿ¢&Kbг'õ\ð÷ ªzaU}ž`þo2`¬txîØý†¶LžJbbЮ–8Æ`ÔÔn‡xÝLÉ©:h Ä· G RÑúèíVHPzÂc&±å×k§):NŠ_ƒ…7…´ë’©$;L욬,©Hî&ûñZU-+óMbæpÖæ˜iJYøý| Sø¦a”ã-£Æ¢ä§Lj}‡(©²Q”7«ä-î¼vô6°Es%/=p$*Qñ¿â¤‹Uç¹â¹n¹vz1l ›æ}Um-è®UÕ{[µ±kUQnUuE=žUuåÌ{ 5r¶Wä°Ïª~¦^3J‘ßu V>Ê…¦èU~­£Ö’PèTÆÉ¸šP6Öæ ²¼QF¬ªÂa+…ñ‰«kRTÉ †*§S¹t†“ -°âuõd”¿‰pj{ámãà}/¼Ó$‚½ðÚÞÚ}O|Z ¦˜¼Tvøl7??Öû_'k|8¬;}Eã{n¿‹£ì+šòЈ†•Üc NïÿöîzÇ’ƒ6Éž+Þ!ºš~C€ø«‹Æ7õ™“l%¬B.É¥žÁìûIE2ƒJô†u’í À³†ß@Ø^˜8 F¦YW`sm\åÒ¯-°@"ôdÏ×>ƒå8^(çm’0Òº¨šËˆ:@óUþÙ…jž0lóÄø&aR±‹—Égˆ–Ë'Ä49©3t®fçMî wÞNÌ8qšM¸±"ç|Žycólµvø>÷‹œªvQ•ô®ò@'Ô^_Ë[t´ùu'iùq÷Ý{Å1Y…ÑZÈgEȲEê²Ü ðv˜<ÁãJcLš¾UŸ”¢LÔ»•¨Âq¹~–KîñÉFª\®ßç÷%é‡RØIæ=.ñçw»'¦Ä7uåòiz$ãð×#|@)¥«èÿê#wÊe? {×ç¶÷Ak”} $¤BmîóRb¤ |S²×Ϙ4üï°™÷î;O²g­ÿªÜf?DR°§úG9íÊŽÙùö» ˜6_“>áðË2õe%u¼ÉW?¸ (I'[Áì0•9WF·¹ÚÿamTksóãÛŒ'}œš k?Ö=gÏvAjĆåôç}¤æ bîý€ u‹æ¡÷Æ[Çšõû¼¶+ÅèçÌ^/#j°šÁñ¤ŽVC~ø9••ÕRŠõï7F‘1Æ·4?h*-4®Á»‘ Ó©loîg RwèY¿C0àÍþ‡fA×£çù-Ów‹ÿé³mendstream endobj 419 0 obj << /Filter /FlateDecode /Length 156 >> stream xœ31Ö3µT0P02Q02W06U05RH1ä2Ð30Š!Œ‘œËU¨`h`T025éæ`I'O.}O_…’¢ÒT.ýp J.} 2.}§gC.}—h ©±\únúÎÎn@n€¾·‚.PÆ9?§47¯¨ÅÎŽËÓEAmŽúßëñÿ¯ÿÿúñëßšYK¹\=¹6s+fendstream endobj 420 0 obj << /Filter /FlateDecode /Length 160 >> stream xœ]O1ƒ0 Üó ÿ €ªvA,tahUµý@p”' aèïKBèÐá,ïN>Ë~¸l#ÈGpø¢Ʋ´¸5 ÁH“eQ7 -ÆÂòÄYy!û›òï'Ø dv~W3Éç鼯ê=„NÓâRP<‘h«ªké±þ“J`4ÅÙ\º‚³ÿPR4•8n®!ÇÜ47I,Óïï|JÁñ@µSjendstream endobj 421 0 obj << /Filter /FlateDecode /Length 160 >> stream xœ]O1ƒ0 Üó ÿ €*± º0´ªÚ~ 8Ê€…0ô÷%!tèp–Îw'Ÿå0^G¶ä#8|QcYZÝ`¢Ù²¨Ðcay⢼ÃMù÷Çì2¿«…äóÒVyU!tšV¯‚â™DWU}gL/ˆõŸT“)Φí ÌþSIÑTâ¼ ¸…@sÓÜ$°L¿g¼ó);Ä@Siendstream endobj 422 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1Â0 Üó ÿ mA‚¡êR– | uœ*C(M~IS†³t¾;ù,ûá2° ïÁá“"Ë:ÐâÖ€#M–EÝ€¶ Ëgå…ì¯Ê¿Þžàk ³ñ›šI>ŽçC^Õ[¦Å+¤ x"ÑVU×Ó bý'•Àhг9u fÿ®¤h*±ß\C ޹in’ X¦ß3Þù”‚/ÄB‘Smendstream endobj 423 0 obj << /Filter /FlateDecode /Length 5197 >> stream xœí<Ër$7rw†?¢c}pÓf—ñ~hWŽÐce‡ÃÒzwèÕA«CÉáÌŠd†œW„ÃßîL$ª³ª›šV˜@¾‘•…ŸWb+åÿÅíÉ?ÿ)†Õõý‰X]Ÿü|"ÓÓUùwq»úòzH%WqˆN¹Õù³“êõÝ©òƒJ¯°Qi¹°ÎŒr®k¾‡f'ƒ‡é°9Ĩ±´ g¥[_æéœðë¿B«Æ›õëya< Ò1ƒ3CW5Dié"žc£t6¯]ÁP·¾ q:¬ ~}ƒÍ>†]nroâúYúvp19ŽÓ Vüs"…S€íuê!„޹C6 Pm½h­/ÈtdtWg° c‡ Âz{‡ËˆJD™é¢D2³¤›Æ÷™ˆˆcw7Y¥¢2™J㢠‘ l»¶nóªµÐŽ»?ýñüßO´@vBu~ "Ôá9µ}(‹€ý±ˆ`ýeÝ:lÿrZø*ÜÁäùöif<ì?fZD)M“åDý ´In„ ÑàŽ6Zº!8¹ÚHØšµ2ï좈®°n½»¤^:éãúe2ï}¡;PÊܦIwôªˆ·÷F­ÿûtchfôó2‹8@+Tf´"8dt¨í4—hÖw•@»ÜƒZ§5 #U^*nÆ.JjEP]gTB¨ã4¨jEVt+b’hc ô‰´u—$ L•ÔT‡ë%ëîTÂÜÑ¡ŽÐ®Bo+4‚zS¡¡B—Ú2#ZaY^gI!1a\Þçi¿ÁE³ÊþöéØNžŸVZ¢`#1C0³‚=2s÷´©þý„ð Öô–éÍ©µhÄ=ªbö°žÂr©Aîvw¾4¬¢4šé:;i¾$j•ÙýD6«"½“Z)`<ØÑQ­b¦Û=U«ÓF¦GÃi*Ø–è'3¦ðº2åω‰@–µdy(Í …ÑcwB:¢WÅT)UÙž‹Þãb.ÍÚMœ0¬cñ€Â‚)|ÙˆO_¦f áÔx#’¨yBÙ=Å Ü%ÙÌADZû-"W1€YÚÝ7kU7Ó áM²A ÜS¨°àhÞ&0§ñÔü¾Ïá…Ü8žõ7c[&`JÖ5‘~òÊ1c¿Ð¸‘оŵ€à'c™Ù$³óͲ‰ÔÄí@è±E%ìk^ÓêÞ•ïwÍÂ%©ŒàA2«ÑSzYüfÖB¾©ë­ •z`€R|¯BÁí´’ÄÍÒ’E_5O¶&Z~Ašß·‘I+¥—ƒÑ~ÉÙzëKôå¬YN4Ho´áhE,rZØenwÂècÀFåÌ_ahɈþ3ÑEƒ‹§\ÈxCïËîé:¨H­Ç.w1«PZŽLam3™ymÎKPü*ÄTZ)ÅPÛU~ƒ…ú?NÎÿñ‡õ¿@Œ¡Tr¸çYñÀ_þ¹5þ®6nZã¶Ú©W Ôž¾¯Pñr>TèM…† ]2h˜¶*tV![¡ÉÓ²jMè‹ ÝVèó ]p¾ÙVè¬B’ÎGhê¥'» ôc.+^Tè¶BwÚNžâþ¯*tÏ1ã¼BÍ÷¤(¶ÐìwµuÓojã û7Ï×ø› }Ë„H-€â"Ÿ÷ú CçÖv^¡×ºªÐýG™ã{ó%³ãçEywÃ~îØm>g6üê#oøó‹Gc&*pvšC®²å>äú§ÓMˆpD9 C¡BŸUH0PÁm-Tè3F»í‘â°4Ý€ÅÎà&|Æ@Ë‹mÄl‰†Ì.ÈV’Æ9<éã|Î.îþž™øžYK¥mÆõdm‡ÃxZ 惄®*t_!"Ì-ÖlqL‹œî*´ežþ²¸Ši„5BTïëJ§C®»Ê)pQðÕ±h¾c†4¦6©kŠÖDã§c'y ޖàÈRú–Æ]ÚÀLFl£¼¤GéÂ3‘Ó×·éDáh‡É¤œ*RVæƒK¹§k’M¢çÇ™è¶äY£ŒÓl ÄBɺ•WõØü_U(‘’åíz—Ç9©èiòÙØŽJÏ’ÓmÞUìϘµÝÐ KN³ oº|è~"6%XHè:cr>[7B³²÷-gL¬yˆ3DçSZ¶Ïœ”L3'‰.’¾î¸o=®ÚЏì”$”%í+’#ô”1uĸÏ¢ŽõššÑ¢)Øá ›;¢â{±2ôeꆱ M?c .l_Æwxg)é\_¶àÁyiNi7Іïòµ»âb€¹Ð xÍð  Zb»|ÉHíhö6Ë–1!mknqƒÖB8©ñÅLMJ½&šDô¹$­£°TwÔ* ´ —ò|IAÐî>³z×2«×ü{°VÞØýœ±åÓ¥z0FÕ×Ûû‘Ja6“ŸŸ»c^%ΧõŠŽç÷1 "½ê£íÃܧÚOb§¼}ÛX¸`¬QÔNÐY| †°þ7šRLø"ì$vÿ¾êÉ›•šP K ãc%ÔTbÝd‰Íe—w£Û˘æ/Zf™LKÌ趤(S2vdƒ\ïúäb¿»,×óoNP©œã2 ãJ—’ æ’»œØm¤–0jšy}W«&{:€%pÆÂnÛ[ŽÏšrpùÌM齩‰ÍIˆx|Zó]U%.ãøt9â£zX3eÿÓÀ–âÜKIÎv¶ääG}GXÓ,‹)ǯ:¾g(zàökIM2)dz„¿ÉÔ¤ìS“M¦¹“dSˆÏ+tT2³ÐŸkúÿdæ§Hfî¥&xD^s"V³éL©¬·{ÉÌ%߯*™9y¡6êêwŒ^6ò¶B» µW›?Íjò¬ÿø¢BÏWþªó¤Ó|ô†[\ÔôHkÛ¯ùÿ&ÕzæOŸ_eJkà„$ÇXº¼û2ç"Ÿà)ÏH¥E*4"’ZGrÏU/íP{Ÿûj8zØ~‹ .¢… ©¯–ÖÙ¦C ‘Vûé!ÐH<ùI•Mž®+¯…Ì&à¸“Ê %ÖZŠŒ@¯+f5-fEÔ6²dZ@ôS:æi›wWrSóeF섞*Ç%¢LˆÐì N3‰I/8ç´îÊÈ6`ñ’7€O­¿Èe ÖÂa{×gç3æIeÐóDpéb®¯CzŽÅÁû5u¸f)‘Ù=BØÞ?;š ý‰žƒñð«\)†R–é°ÆJ ‘µ“ø»&nt5{Ú¥š¥‰;ýC«uœ©Ó+©Ñè±ðf²@tI†dÇܦb_·dMg€¼³éœI:¥«@4ÆÉ½Êìý¼-[2Ô¥”IÉšÃE˜¾J—ˆ©º L‘b»úl§Ùײé]î'åè¯[vç!÷“RñÚìrÍà~†¥jN]†Ÿ*F™yÛ@½iIé¤f!ePz#‚Ro×{…ÖYès•Xʇº<Íâ+Õ`[V• £›/…©)/5“o[žÉÜl Õ¬ë,„BÉ®üõ¦õ&kzCÈMR;$OW ‘ñsËØ/)Ïèõ¨ä¢W*˜¨5Z©V²ö2iŒ!Ђòö¶a†LÓâEÄ&L'${Öëp®sœf—ï(°¾Ç˜WܪŒ¤ª³ã’>q´¯_3Xq{¹¸1#\Ù8©ÐûýùÉOò—2`4|«W ŸÇø°Ç¢aòyŒRfÐÖ¯ŒL™ôüyŒKd€È@¦Y@Ý× Ô1¯€ÕƒÕ&#þEÉ5.íÃñ5HV¨Õoh¦ß˜•pý÷-cÔBõ+&ŽzÙ^7riú¶£v\ŽäÉ Ìå°täcôåbÀOIñ3†Î@]íºÁOKÝ‹v+šMmÜ´Æ65ý£>Þ6P7°…ó[n½Ë¯ÿnÈn–3­-1Þè¯*ôºB“~sÉé› gŽË¾rPùè¼'q™ ÙÊtU 9TŽW y(IµØ‘¦;ã¡lÒbGv‡A‹Ù5Ÿ²9r»îã/)>h¢Aワ«ˆ¶×ï›ùø°Ä c¿­³ì*Ô^M\ò«lÐUðõkëÆÅ6çÓJÁ¡B„áÛãLÄFØÅÇ8Úz*º hV”=û-aó8tÓí=6Ýh\ ‡Y•gŠA+öžyÒ ™G?…ßçô²ñCî'ÜÍß4È·´Ùâ2õÙ*3Z/–/Õ2«C\V\·´ž†K´w¤×<íUÍ—}ÇPš¶"âÜhÎlGš/’’U‹^ÒöF+n4‡Ü.?f—f:äs‹ ”Ö  ³ ìÅ|,kÑYóÜ\âŒqJ“°L†@ aå"ÙÇ‚cghX0r†Mnªœ}Œ &Lî=£ê·o¦ÆÊ¢fªyÏ)G{±ü˜åœèö¸ ™á(Oï‘/™ú}¶¤<Ž,ýH3ÕÜ@£ùí°ŸeÄ›­™ ³£å2#ÇQZ9 {Ï⇨ 1fjôÆß3Æ©™¤;æé¼éÊu5î?ÚzQ—¿è.Yo—IÏ 9õþ‹ æ8äs>ä€ëëùCEkúŒ íô™ ~qE5i„·ƒÞYìmÂvÁ–ØmfÁ±̧D™FÍnO!ÄÞìÜ3»÷lª,G˜¥FÂ&¾¬Ž„œtÛqËXyçäŠ5T¬@s‹Ý·ÝK6?ìž‹Äk‰â4`z^¡:µDÈáЩìêès_f ÁpüðË£Yë–G³ÊÃyy„vP)Ÿ>Û×3nœÿ&§}ÇÊ7OíJÙc÷˜[>æ1Étì›38ª%ºxk<Òü4»ÍFEÔè0ÒK—Mù¾ð˜ˆ§e˜!–G³s³–‘;Œ6Å-<ó ÝíÞ3º6ÆÁÿ7ŒÝi寖OjqOS„+‘²Ç²£OÆ4Wsðà¼ð˜UBÝ!ï=ÅÒ‘œžíæŸ}”¸§9möXF…t1¨]á3¯8npøXs âöHºÄŠ9´ý3š‡ztDÔ›M?. 0:n4uQŒA›Ë½-<Ž{ŒÝûÈ`îl¶Ãóm´”¡¼p‰r´JøÙW/‘µ '_2•KŒ­9Ë_`©@+AÁ’Õà‚í^eïè=9W¹©—Æáõ×íÍ{±__üË| “ kufîO$+ê¾ßÃÉtð3_^?ϯá}ºÝIÅ|A_«–Ú–êòs,"˜–’à(1­ã w>Ñj0rõUYÂbd÷QP÷UzsK™`Ú”]æ»Ñrë~DßõþL÷³-u+iÒê÷ë·Ed®m¾ËÖïØR=I_RDÑyÛE–·‰eBâûÎüeR˜-,ʽq•O³tt1Ž… OýȳÉmO eCÐjâ§G™É«Y,Ssµb 5"K Ô±~1…_ü±°a¥Y½­9öömÉã¯&j#žUˆ»‰û@`XÄ|ô[Þ¹úN Aq—_î2!cÜ+qJ*÷¾E.«2Š'—fnÈm™YÇ•då¾~M:l/p («!µ• +ffÛ¶*,ZœC:ËCéÝZ{ :Þ 9½_«¿2 ÂÕÏ3i!À½aÚZ¿Ç|9lì œPË M[#âl§k¤‚n›/ÎK¸‘R°b|ÑR\“ê´¬™> ÎM±ùsÓ•hä[Ïlz: ª¹¤Öñ¾ŒÂ›`Ÿ5ÿÃÑÜ‚¿Äº)q\GI…—[VBžå:/fï‡û»*S´Ã芃šóÄÓO»m€ªEý©¿4w"ÎÈK)â»~Ñêe;­ÏLܶží²UjhÙ^!dèq«H^¶Š¿OuÓñû¼Sß2ú-øŒßäêúôÝÛ±ú×gô`WRD0vu{¢4È­ µåæäÉ|m^+;€Šô·bÛÁh« 3ÈW0¤ìO‚Çs¦ˆ€ð+c…L&cvµA–€ÿõ%ÓÚ›ÔŒŸµ.—ñ"uÑÁÈLéÜvUØœŒ×&@Z,…E鉫©s ª!ˆBøT0š,s:]f êš*Âç55I𠇂x*ÛéRµ¾…åE´Ê4e¥‰yìJ©˜nLl“—ΪT iBú¼M¾×2~ž>nø.÷ñR—‚ô¶ðº¬rùfhd…V—ÈE¥ T¾.kÁ;) ’BûæW …·§*BàÐû<öÔ-ŠvÞ‘…¿L|w>V^§”GX¶„R.Ý ‹zâøå¨ÆRÓzKÆxO¾è¦$v¯¦ý“ ¤‹–+Ÿåí8ï²´”ÆÉ5ÙZf­‡“GªF"ƒí’£àE°¥*€xh?G4:K²:<^ÇŒŽ½0V‡Qò~‹ ‚Ìäs˜s !›#JŸâ4;"ßµÝá 95‘r_PFŒ×•Ét·Ù„=šÚ‚ý.ÉZq2g1)à*›IÙ(ˆC> stream xœ­\ëÇqÿ~ɱHòaÖMú5ýPâ"e1b6öÅ–wÇ#%Þ-}Ñ ÿíùUwOwõlÏíQ:›³ÕUÕÕõžýu#F¹ôOþóüúäßÿüæêîDl®Nþz"㯛üÇùõæÅ ¤’›0«ìæìíIÚ*7ÒèQLfcý4N^mήO¾.¶bÔÞÉ0ìî·§bô!h;ì¶§Ú†Q=Ü@ # ÞpP"Èá=öÚ=\Ò>#„µ~ø ç­²X¨@Øt0Ã9¶M^+ÐØ?Ð> !a V †"ð†w•5B¼ðÃO[åF%dƒô¡ÞJ‚1ìÒMÆÎLé/g¿#ùÁå£Ãh ±9»€D>l%¬Ùyõ¦¬.˪½/«›²º*«ÿëà»íÀ÷I)S¿®Ë•»²º/«ÿeûM‡t=ÊO"»‘Ë.‹ç=&Î:(¯ RÑ–/ó¥5J­ Ì^ÚùÒÞFP z·¿Mša¡š»¸÷T{=*·9•r Ó$Ó®+ï±]I/…“ôX¦üX8¸¾ïˆ€V£“¾Ñýˆ^É©y|ÎÖј¤œ@gŒÖ:REãѤ“X á“g¥ ã’=hªŠÊ¯õhò¹3Ã>qèm Õ,¸‡}²3Á ÁŸ·VÓy+ÀMD+ð”1,„–ë]2O=A$òÙóY„ ´·=» À4ßO½ú}GweéjÕŠ2˜i^ª™ØÇân£áãôpûí)|‡ËÌéU×qÞ¸ŽxxéC4K:°†~¾Š¶Êdºà!ïáê¡^Ì-iŒ‡&Ù„$!$žsvò‡“äË!xçåæöK¸6fT!lì¤GG?]Ÿ n{öÃŒûËñ0ËÑ}?|]6ß©g[O•–£ÔjVS“$þ_³kÈ–. ÏwI«ë²º)«]ç×˲º+«|û´<+«?••ª€Û²ü·Å–'¸¡ÿ,Ox¶)+ßyöUÝ]†§ʲ-8ᩲÒcÓð¯u{Ï£ŸvÂTÏßópV¾íëmý¹oˤ-Z›˜¨DèuÁrñh`ûØá¿"üœÝz$—¼ºqšXõjßÑ«ªUOÏ{Z÷ŠñÛÎîÏõç_Õåw ×8ú©+àyõcÒ¯êò˜&wôê¢#ºûγ_w¢öŸÊJõPWzëÜYýõ¼C¸þzÑaáeõm÷æçÕ×ÝT#j¡pÂCÏeW°¿ìªªl’JfÚmª©!ËDèùÐqvWgÕQö„ôª¬¾)«ÿ)«×eõ›ÞÝÕ”ócuµÁwù÷¬öñ¤÷UGlï;"¿éШ¶_ï=§1ä!Ü9éF£6ÊZÄh»ÑȾFŸÿ@ˆ½Üüys“â­B °¯@³ñ¸-5þDÑ#,`OÞž|¿‘(ZLØ UrÚüû.N¨^Ò›OÀó;üûCÆ÷ÇßÃ÷ú$ÚBY) ˜ªDµ3éD,m)È×·¼3bTN£­ÐQD²QÀK }U€I;ùhÄxdËviÁ/ ü‰Ð7uPîÄ$GS¡<ŠÅ•8ô(×B) *@+ò!q]:ÜQ˜–Úr˜_Ý ,á/ô PaõDU9†¡G-Ÿ:;3Zk¼îQ`ŽÙz7ú ×nTô{JGßlýh¤PKÔÇ¡#îqÂ_Ò9 ŠÂΕ]qLÌþíÆ:4¿á:¶BÛQªiÍ•ßõz‘öi‡ì9¡ÕYʤ C>š÷µQë´&K½¶*=nÐM½XøèV£á“l µb–^æ(Ç:9MRÔN›¶¨€ŽH–X€ö°"l¸“Zj$'ӻǡ{¸óùfh¸P,—Úþè&{ôa„è<üçd ø!)X†5{lO¿Õ(Y&‹RQ8Ý£ÀÌù˜ØÉxÜ>æuàe˜Ð   §ˆ%–ªëÄêc X]C¨ pz‡>ýè+N§ÇP@XQþ°¬8ÚxnT1ž*Ë )ê'HOÓ&‹…§Šc(rn4ÃIaG8±NR{[­y£›Zäó©3Ž.ù²s¬£hR#p€ ¡èå‹ ð̵ƒBz~2Ž÷0Gê¿t\O€îàžX åhá2Vp?ýúñ¾< Â3ó¾<Ý"oô^€ˆùï.ï?mQ™ ßqý» }JÛ(j€ ¡æ0äw|»ÛR“EËÜ0Å‘•²vsªÝ™¢îÔpöþúòi/¼gÌVÒ8´öd¤&*M¤ýêv{ŠÂ%xå†÷»­"F”ÏŒü\RÆA¾Ö¼!‡xÌAŠ_ÂüŒQKE­ Bž,ˆ@¡é¢ô”&ßìÙ»‡Û»,Uoì3IU£Æ°g”*ÒJ”³Ï(Vd}±«úlbEŠêçw_Þjؤ<\^Ü\’€éïJ¨g0’UʹŸO¾Y¯0æùäkPi(ížO¾f²#ņ¤¶[j­ 9<á’°ŸI¸Æ#|Ûç”®¥€jŸQºpÒA„ç“îkdžäùýþæ— u@J£jððHÉeÒ†o^¼ü—ˆFV”~ á˯B#c'Žáûîg1xpG“ fzF5š ŒxÆØ2R£¥‘þáK§Á*л}e ªXŸÃvœppƆ8û’^®â¤Ö‚´SšpÀ* ¿hA£vp_oO êñø÷iè"Xí‡ý]œÂ2¨a÷¾N¨œ×­uZ£[¹'Ir¹MO|2›?yW'ÉS³^H(•(ï¨ùûÎÞ`‰D¾nB,i†åˆºójŠÃrÄ©7ÍLÑ;6’Tæ×>±yž(Ž ¥™†›z€«$!íp]ŠšávXº1yçrÍåömb6üPEÇGVn¶Q)ƒ²˜é!»qÆÏE3uTN÷‘F¨µÛNËÜÔ Á*¬€sö8£†ÖhzUIj€ìÜu¯ÐSõ>ÍwòsÞ?: òl¯†{ (ŲE¡Æö¨§ôbH Ë;Eí­P êÕå…„?ÈãEb‚J]“Á‘éNÜ$¥3($¿N®–®A¢$â>aóÞ,öÑS8Û¤jipsÇÑ}&Msží@XÜH“ÍDTÅq‘ &¨àÛªl Ø:éØMkêL«"s83 fTÉõL"DmmGM#@+e§`¾$›-áS¢Ï‘¸ªuÃ=”ðR³±×=;Í•À9À}Jv–ýûªiF6h¡=?ãOÛi¢YçÉ12¾÷u&í2M?PÍçݬaå Mò=yÊ–mký([Á­w!%( œÐ®1ô²‹áJ¤†ßÈ#•„.Oå* ÂKå§qž —=Uù£ó÷Ôùiæ+ÿ !8U>¼ù1šBÆ?6¬&š‰^2¿ü‚!yHNÁ0‰ì#꣹<":Ù‰)\Iî2·U_IøyвÌpÒc:Rû— Öˆ6}¨S­{ÆRa')Äu ÛY«ö75x­hB‰%ÿë>jÅ7/(bI-e”(dê—Ê2Ÿö»õ/ÓcÙÄ=W.è˜@³¨¥rÊÎR!â?ÚB± V¢½WÈP þÛ^lº[LŒÏ4Ò±'Jª#j¢NgBíØOÍl“M, åšËnêHn;wLþdBBÒ¤JŸ)7 ×f¡VÊÁÁ¡°þÆ}ÒïºmŽ¡$éK½M°È‹Ÿ¤‹k„æ˜ÑX¨bÑVè‘ 4Ô-SÏsªÂ¾H’@êUk]MâqÍÔî]zl?wEV*+%R"ø­—é1]xߪ6ÀÍÃx>cÍ J­+Ê70««ÀÚ23ýjÆO‹vâ^õ¸Š*z•$É*¹¥# W)—Ø­g±è˜¶ v¦5³M“Ï録¶1¤Ö*Kпc •dž›KGÏGôY:Í ûóÅH=õ‘ £üïR¤h#ÆÑ%…Ör•×*pÃátØ©£æÙÂÓ™‘fIŽb™_ç‚ꊘˆÃ~ýò_#öE£ßÅ.nsFEIy¨Uñ‹-Eñ2»âq+ñ¯´#Œ“óZO‰Ì5ß$ùgSvD,TW¶`¦Ø‚a¥<{‹Áuü.ë|%:~d?¥ææìWYgk!ÏO ³t\`Ü.üoa;Þj\ߦáÍ=ÕǬý×6«ÊU²–jl“QƒGoЏ9D–aö«lpDæ¥a¾ªq«åÜá‘dŒ àFšFrl§â/ž€“Ô}ªUæ$€U߬’^é†Þ•&ª1å p9377¿›†QL;XšÊßk¬ºµƒx‹”¶ZN$k%ÄÌ”weîRê<9;q±'SÉÇj«ìÏ+Xgeß^¦¶”úèõ´:&G+íÖœL´_Ç'ØøÆñ.a ÎQ´ K=å•¢ø2ýNÝ›þ'×<-Þ×ô>kÕcj!ø‹Ç&]`þŸ—£ðb“¶ñÜ#Ÿí?Î5kAüï1ä"iØŸ3s/×^zÄ«–Š[ÚÆXaaíµ )“ 2~?–Mä37¡¬+ëiô,žHä¶‹;ŒÅ­¯7Ì jè¿}Øóˆ—uL‡ié^³Ðvu©OY I…§®`· žS¨$’E$º;¸ß”þxŒhï«AdÚWuË´)fðàfá›3À}(ÐÜqIU‚çš¿Ÿñø‹ÞdÏÙ¯âÏYõжçTjsP\"YÊ‚.º§A 6âðÿt6Âendstream endobj 425 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1ƒ0 Üó ÿ €ZuA,tahUµý@p”' aèïKBèÐá,ïN>Ë~¸l#ÈGpø¢Ʋ´¸5 ÁH“eQ7 -ÆÂòÄYy!û›òï'Ø dv~W3Éçé|Ê«z¡Ó´x…O$ÚªêZc:A¬ÿ¤Mq6—® Áì?”M%Ž›€kÄ17ÍMRËô{Æ;ŸR°A|ARSkendstream endobj 426 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1ƒ0 Üó ÿ €*U• ]ZUm?eÀ‰Búû’:t8K结ϲ®Ûò¾(‚±¬-n H0ÒdYÔ h‹±°> stream xœì½YÓdIrf¦Ç"Dù núcì $HR„£$#Ð2HFê¡9Ó3X÷4g#ýzÅY<2¿êêÈW¦ÕCežïæ½qcñðp?îþ›é-Løçÿöý‡ö—{}üåï>¤¿üð›™ýèÿ~öýÇþõ¹"·ôq¿íQÆÇ¯ñA?Íç¢ñ–sý8Vë«|üúûÿæ‹¿þ2¿¥´Zþâ‡Ï|úî~úÅýôÕýôûûéÛûéwOýª—ò¶W}ºá¯àÿü™ç}ÿì—ü´GóSðéןÁ~u?ýþ~úûx\Üà/àÿqÁŸæÓïþå¿ùÌcþás?þ·_|¦á?<þ^/øÕýôןyàÓÛ|óôœz´âg÷ÓÛýôo¿üòÿùúýðUËos”ýñ«RßvNûã×??¯ŸJ<ã|úýg°?ýò«œÚ|Kù‹ÿóŽÕ¿¸Ÿþùýôç÷ÓÏ?3οúÌÌùÝgþú<…×á=r~Û½ç_ÿ«_ÿ7ÿæ‹ÿë¶ð«ûéw÷Óoî§?|æ~{?}ûÔè;cÿìñ1ß¿Ïûéí~Ú÷SºŸÊýô'Û†ÇqŸ¹÷}¾™Æ¥ŸùñkÍ•3 ÷Ùü4ÞmÌ_Åk}ueÞîW ýðiõQŸ»î»ûé!Z¼ýðã—Ÿ]£½¼•3ÊѽoxÉ~F«~ñ5øÜ±Í3>˜}­7ˆÇ¯jÚç±ÃÝà7Uþ‡wË÷=kªD¯þÙ~Zñß>­’§‘ñ{çØý­Í;«x뼺¸ôýñÌY; ³!­3XomŸùªýJ?ÚýtÐßr¢¤:Æø\/çôvääÑÞiPÓ:Kg½ky,ÓQ±.pE_ù!bÛxºq=W[ñÿÕc¾~®gØVÏ#®.Ämþ.ìoþÝcíK®ŽÔ†Æ}Q所+΃s ¼ÅbðKüÍÓâ¼Ëéԯõð~†í›ßŽçY{·M¿yZ|\q»¤-Ö˜Gˆ>5û¹ëŽ]wŽÞj!Á{œ)éô¸âIŒ<Íó¿Ó@³þæÑžÿS÷üVs!a¿ùùãú'ô‡'ñÉDýªäúVJH†6c²q–Ÿ©‚Îê‰û‰Ä×ès݇–~v¤wúÔ[½íò¼ó=-šž×Õ7X±mŸ/ÿ᳓ož=—G^ô4—Ê£ž{wüæ_~ýá_8JæúøwçÜò?}Èý´,“Ž$çnß8=ëk_ä»õÓ§›ñÙÓÍb÷LíÌʡӗôž¥üÄ’žGÑ’>ïøKš“¼n-AAG¹y÷û?ÁØ´·•dzªòƒF,Ýø'”èèö³ù>ÍÓô¾£Þß-€wr#SŸïñô÷-Ý!£œ,ýWb²|º´~4Y¬ܹò|ñ^–èß±û» äYü\õæ7Åû‡Ï*5OmÿÙýû/W²¶¸ùívVÎiæ±äÏ;£™ãhv_üã§…û´´¿}’¿~À?ûD@ðµß Ÿ_qa•}`ÌÇntïðË;|õ¾ó$c(4ê^ç 7>Q'ÞͶ+* é–FÉö#M—<}„ÒsΈjõ“XãÏÏæx7;tçÓÃVÉÏ^©ê+žöÜ9/è[÷gÿá1†ïçð(xô& î±Þkõïµ€«Bÿ^c‰Q?Ÿ±ó¾›¤±>¾õk¯þ¤Ñ>´¾÷®œëý]}¯}î“ßý¨e!bž—ÄÓv¥þ>2¶ôŸZÿíyÎXlÈÿß+=¯Ÿ§»ýZúvîo˜@ž‹¿Œ°ÖNç9ŽúÿÍÏŸÅß>mÔï§?.©©®OÎç&ë´é§6û§Eönž›ÌsïY>’ï¶Òx—÷«â1ÿnRR®WM}X8¶2F ôÐo'"ïgyÒÃ:dStÚ¯bô½ß¤žßõ«}ÿh½ÆÀe–ÏÁôß=zãwŸOÏòçNòë•ñ,%®ÛÇÙèÏ:·ú6×úÜ ÌƒÓuÿXÿ#¾ÞO^ M~âðú´~Bó ·É™‚A<ð¬DröîœÛ§› .>¯öÓ[.€ÎúXÞçʹpbcJ}w*|Vµž黇ݭæ';âŸ=hLì©ã‹ÿp:µ´“ ïßH›ÇÑeÒ=‹$kÞ¤9[ywdý‰íçnB~¿pî3ž§žÆ¶Œ+ƨöþðô-Ð6ÞÆè9µÞi÷…iNê=Øò7®g˜ÇÂ=»„•áo$QúóypŸî}NÇ×xó'xÏ~­å‹÷‰Æ„Cë“Äüåßhö\ è±ÿøTû[µsaó|Z–½b—:ß)èØì7t„Jáo ¯ý¬ØüîÑóÏkæçº“ÿÉîòùùøé9Îù©ý#þþÓö±hÙ½ôIÑü©=L=´×üÄö¥Î?iûúåÓæ÷P£Ê]ÿø6íY.ÅÆ±ß­¤g•óïõdØ,žíG þº5ÉÔß-§mã½:øØ¨Ø-;½3]=)ƒ?|ïŽ;o—!`#:êÊן?>=EG½¾×Ò‹µæ\&&ÝèúÝCr~r>×™éÙêúë‡>øËÓÂZÎqm=/¤_=ÝXrÿñn!üN?˜Šµ¾µ/þ‚M˜m¼_•Ï?“½0_‹šsO¦ò°¦ŸïÇßþ·Úû];žgžÕøcïãmÁ¬óý‡/Ê—_ÿmÜû?õ~%Ek¶çûÑ$÷lÓùOkö9•?μnj·\ÇÇ>Nì¬ÇLvÚéüÜ;ÌѸ>ó£}æc¯R^7þïÃÙPÂ%õ‰Ùòþù¿{òß\?ÈÏ>gÌü'÷Ó?ýŒçåñé_|æÓ_|æÓã~OšÏ=d|æÓ¾nǧñ™Oÿä~zòè<Àz?ýÅgnøçŸùëç°ðCÆ?ðÓã!é?çMx¤HåmÎ+˜ÿâ3wüãŸþÁÏû‡Þpýч„—4Úý^¢>Íçÿ\sý{ÿúgÍõܱú'Ÿq§>Üô¿ýŒëôïî§?»Ÿ>»´ÊgRŸ½ŸЇ«ùá~þî3Øoï§?»Ÿþøo=”F¿ùpÎ o»”GçYo©l½Ã\óñ+^?þöÛýñ×\l­õ±¿RË;A>ŽÆ\çǶʑM‚믾ý²l˜ ËÿñO?þoß|÷-Û‚ë¿Â36ÂÓšV ,.ýèÏþý¯~ÿ?üéÇÿûË ÛüQD¿µ¯ú+÷[©§Í»åe|îQñe…툠o¿ÿÑSK9‚=šóЖÏüßÀÞ£w<[iÙŸIS?}¶ö¿ÜÚÿñÛÿ UÔ7Ø<¿øý÷ßþú÷úñÏ}®_Ÿ&϶ÞίG;Ío0Ûžm÷ì¸ùH? ›<…œ-¿Gp‚0"à¼OûHó8Í+¾IªBN[Ê{d¾M6晴#ÔÞ#ãhZBÖÒßú{à̧b¤©gíŠè½IäÌc 8/YøÎ[ŽO‘l¤«û tf!MU–;ë ¦;ë(5äÛ½æ™ê®ut9!ÝÝu½C‘ZH꨻&ŒõDª»ë jòYîm YnOvîw““ûë dߦëQyãýˆ4=*Ÿ®4 'åóÌ*¤Ì@²ï«WÈýuþ ®8‡¶AdÀŸ@¤½m%€\…L?ú¨\ÈŸ}d€oÓýìü¶Œ´HÎBªz='Œ‘¢gÁ»å;'=+-O‚sbo¤%déYi¾Õ,dêYG[ëMH×k¥îYpÖ[$ùÎþQó$è˜eDª'Á¹Ô ,ž ¯gD“à,70{i% ù­šœ5ylØç…ÔH*Bò² #rVÓ2ýVõmU!sˆÈØBè"ÒV ÅHmBš_«¾ùá•ò‡§½Ñ[OÛ±‹ßª¼ #ƒM¤ùáɯUðH#ÉHÖ¯ÖÆ+$cºY~-. "±da‘ D.m_Èð$Ì-õø,.'" BŠÈlh&l~Dª0Î`Bø†D<+còµ¡?¡$:Èô³—ˆdlRc ‘à‡v!*@ ž~¤³ûyžA62ü«iM¤•@’‘’…t-¥ªÏÁq¾N¯Ð1ÏépÀc5‹¤vÿèÈdôúÙ|ßDáÕ¡PTýdlnux—0°ó’îÍH2âžSsâ E­CRü É×xc»d#žêç(Õª‰À1ªû®Z¾£¶ì!Ä]sÒ…4·9»ûjÌÚsìS÷ˆ##zú™-jóÑRêâ9Ú—û¯`™~zŽÙ§ÖˆTºó9Év_ãù×»„ vÄH62Ô«G­k]ˆ—y¯Úß¡õ@ ޳kºðÔêY“ÿ ~õž4ùÏÆ:w ¥ éê£Tt#E]ˆ½ŸÈ´ äNjD2™{W"mˆ; ‘ ;‘águI{ìÙ¾¦i²”â 2ÓˆdÝâ©qDÅ×ÙÅ=5Ž(YF,â \Œx‘4/s¸ÅtÍYÌzÓï…åmDÇÊ%PB2aa!# ÊRb `M9ÂG€÷ê’­put!½J[ÂRˆ“3…5ž)Fø êQêŒDŠzô(ÀPÆ–Oåš$ºòöVè‰vï}æÌ0âw8ósçh"ž¸©¤¯að¹F ¬èÆ|"óã& ú¾]Ê¢%?—/W¾¾éêÓ ´£Ê¬(@`‰yú½±à‚ž ÆÄ;†\c]ž.ä”KlÈèÂ*DjömŽYvz‡3(çhïyõiÄk.OÍ œbÍwï< Ù2ç»—eÚ±·eh ²S%~R¨Ý@øÞ)ta4}›Bc=ˆ¾zÁuÅ›ŒØÁÑÌ*¤õ]jšÀ—>J×Ðcη¯}´~Ô0o©ç‘œOñš¸5¤Z¿Æ#¸i¥~î8ù«=~nú©D'nD¢–Gy‹Zb¤«`$Šúƒ\u¿/B<ãÏÕÃ}ï˜I‡O Ì c ±éãoès êd|Ú¼fº“ù·*¤ùWg„¤v…¾cÆ'ˆFüŸ|›ÎÕ…ÿ{ ¤e!Ú':¼s³¤Á§’…dß§²3ø $!Ù¿*uü¯NíGgTgn^“=%;´Hù‰H²–Ûµ nÍ’ ÁÜšµ81€d#:ïâ/Ú±ä%DCÜV9оE©d«ž×f!<óR'/ a@  ðR߸òà$Í@ ïÓ<+;Ï-$å@2ï\c@âÙÙœâ²ë¶ÅÚC‡Êžð¶Ñº–­ •?¢RCd¹k’Ïf@8,øµ€ÉQ™Û;<¶ ¿UOžþˆxä&@õRSúÔ¹Ú¿&Òxúå¯G œ6Ó_+$~éy?y’i¸nù9<¦©9Â'W«ý(³s )#Â7(ÚhúäÂ&Kš!3„u? î0âA:—ó¹VGË>––óØÑÀƒ@[¢_nø‹ßó 쬃ø’Á þ«ÝSó(À\ÜuÏ ªË ÿǃunšÐç/¸£$sÜâù{®8üÅÓ ÿ[ *µDê£Y0Àù6.ÂgUÚpMÕ°d¡¡Ü$Àúß°×k˜ŽÐôÅö´ûâY³¢^‰Ä]gLß$iR#cÂI}Ôq:ñ2”÷5‡®Û NVÃùÂhðQÇ—‘¥Q8ˆE_µxÇNƒYi²<È´áL—@ÊE¸˸Ϣ ð ÝZ(:„”‹$^ÓbrUYjz 5—& ¨Fd»¿D6 v%:¹ä˜ÈµIxàü£%‚S=¯I±®ªì;ýèáñ¬*xxVÑ[à(_ç–_š¹jŠ¹Ò¡.zHô*Ã.ï'8¯h1—j’pôp¯!x5¤ pºåÓ㈷ÌWgÂA‹Í»Ò#;„”‹pKÄù@¯Pd2î0²^„s0­èуð’Šã/¡½Ò¯"įY†VÚѼÛEïÓl£Å¡ì=H¼h‡O Ã£Žo*1sÎâø5Û£Y¤Ê)á’­œfi;t™¢C1€{l¤mÛiÃݶMò˜†&m‡]4} å"°´ÝC[ÃTå}[¬ù"_«“©l]œ`€$6¯„œ?Ó'ĶÃh ¯ ¤\¤ðéaÚä”ÇÓ ì§gîD¬êñ€÷‚º|‘Âkf¨ÒgYl^3ló‚iÄ«§Z´ùè·^ ™JõAÂ`‡µ³ù«0ÆÁáH¹Hã¯Jˆ˜,%½ýu©72­CDòE&±*qç¹cÛδdEæ îì;íÔ»g¦áSi'ÿÈݳ ‘@Š{C'‚;lìx¨£t}÷äÏò>4Xa/}°Í°÷@&,¾ÔUܳ,@|¬8§åÓ ÈVPGhË£ $ §é ìmŒ8få”Î'îQ÷ErÓ§Ô]`†ò6%vÜ,ck;Z¢Wy¦B¼¤3 \H¡èfÙb”E?)i‘Ep_„}ÞCŸ>´†£©ß—Jšý}„ô…¸äm@´%Ä/“ :´7±€ð’úv¿sXz‰w‚¤ )á£sL5˜iø£r5„¹X<Ë(àÄÒÚ5$Ùœ[»Š0¢áð¤6â• ">¨%šìÒ-À¡lWí…ߪ¡Œœ=a ðy?ÉÕZ±Y Èà59þ45Üða×@8ÜížæáÐãà9¹BGÀ‰}oŒÓg;*ɺ7æj–›Èà¦Øê°ýÈ{VC¼¦U¬$Z«-t¹ƒŒ*ÄK3É ßjMaWx-x±çEØÂl¯.l· JÀ½3%kMw I޲ï$¼(/Ê‹òb ¼(/ÊÇåÅ@y1P^ ”åËòb ¼(/Ê‹òb ¼(/Ê‹òb ¼(/Ê‹bäÅ@á·åÅ@ù”ò7þÊ¶ŽªÄä'ž‡qÉ÷ùÚ¤UÁj`¾¿úVDÈ@ºßáÄ©ÈRm z Å»&V Vgݱ>€`×6î‹,ëi ¾¯§»@Ç©kÞöɃU×¸í ¤?îùRW³¨ ‚æ­p¤³þš·ÊS󞟚YVa6 žŸ:C¿‚+fv  ÁG¡X~'ëž[Íä˜:ÃhCAƒÖÀÀUÁ^[¡?´@°×I ÍÉO÷|zø `¿¨#NC@°¥ AÑö³(ø2½¯Á{pËÁÚ«£{·RyŸæ-Hç}êÓ±-Ö~0 üQ~ºÍfã YxMŠ¯Ø kGÖêÙçS Ý×§µ\Pøk>Õ«fnË hËH‡žÜYCóñlë9Âù×Ã@„S©‡ïÈ௒m@¸ªÀÎZp}\¾v-k „?Šã€N$\&Í ®zi$@ ×KöÀ–‰C ÊwÔGêr)š¹aõr €à¸Sa]€ãrIØœaq©hg „>é°ß¯«Èâ5áZhæ®!A¢Ž¥@°sÖg ¿ÊRaP˜Õp a—†o*ôýZVÈ)X`ð£2CÞr‚w8{çº×°K…ÖD¨Ø;G \z¥Æ>—U «<”".¢rg6†@¨°wÖ눂»ÂÉêkš:'¯˜‚k<<Ϙ:­©“óˆ©Ód’D6éâk:%iû8=tWŽÃ6J³ëº¤zˆæ#Ü䓮ФK 8(Ôët„≓>ëY”@` ¨Pï50Õ4îØLͯÔïØ,-ôÔîØ,êï5]!ØÄCÞ; ¦ÓkM9„W§ Hº×p=§7íòÈ•½c]wf ‹×Ì7_B«ÓFt!¬¹¼ÄZ4œxËÙ+E‘‚þ_T ñ}ðŠ;×;ÿ–œ)*/H±åñ”qÄ[ $.Y+f(N&È AÛ¹•In[ç¶SÖ¸ïܹ”³sÆ;wµþ,ƒÍ_Õ >ONDJ¼²¼î¦ï‘b}wš ¶ZÑðN×—„ÀŒ°“3u;Wߢ薭ñ™=Ï  zfS$¶v†’‹3&Är¶Q˵AÒÙA²‡hð±a^Åɵ1ÝŽ!ò>•ëÊÁy²§Œ0¦¤)ÒTyBÆCÎŽ¥¯¾Œ{ÄÖWF»m•²]F½mÕi À;£N´R$üØ@ŸœB‹:^•~•€¡ƒ[é+„ÓЈ÷Ç!#ré¡©ÂÄ€­¥ô»†Î½¥÷|8ÐxŸæCÎ_…ÅŽDÄþZÈÕ«7 ªD|‰qåú€pÉõÞ˜ZäAJŒïž’¹ÄøÂÖNÏÙVKšÃù«àŠÁ¢Àžå3mâ8Èe7ŸhE .0âøDËc’8ÇÁØ3õ{0®TÍ€ÄÁ†‰´{.–?)§çâÊ]["l%%]C¢M9@ÌIb[–t-‰2bå} ‰Ð[¬kC åHØ6‡?ïgCb'2®u”&˼ƒ¥`U!a•ËgÐÃ:Z)Qó®×:Zý¨kI”å@G¹]$_ëhç=ï *ÁÔÍG­}­£tþ)ÙDVØ{2Ím™aïÉŠ+’.²‰Œ°ùêäé×;•©6feìÐ(]×u‚ÁÍËz0@øÐ‹ÂøÉPÈùNùúÅDˆÈë…ìˆÊs_¿éDÊE&‘k\,<ääºÌ‹(@òEðJs\G#]{yÆÁÁî@ž|ö!æùðþžBóu TÒö \7hám¯ó'd"ùº»:oæðYËW}tŸ$‰Gp<Ãá $½sçqýãv.ç1ÃûÑ‚“®yårí8UñÎÏþñf$üã2zäÑÂWÖÄMÌãá׊.rì‘üØ’¤ÉÈš΃&30 LÊÔ<Ò¥h›’.Ý¢¹Žýê]üîܯÌŒ,ĦWÓ/24\_££aîÁ¯ƒÒƒš3¦YÐc7fNè%Ä´Ñ[óÃÁÊ €8ù€ËBIÑkLã.Z=k˵@æì…òÃé— €°‘¬³„Ø?Dr<»A¬ªƒðÆ-ŽK A+É-<èC'Ü.qt œ/› 'Å H„Òîá˜0Mê¨÷—1´¨b±s0Àâ qÊ ÚVnazBáÖêe¸e¿DH˜)â+€r‰f\2i~‰SRǃ¤p”L†*ñrœC\¯ç}N= \ ÖÁvKÜ1M/×8ÄžÇÅUÏDH^0Þâ›$bA2¥0×+•–tb ž8&+æ‡ d)dˆé•²„åZÃ{¿DÒΟÈò2«aó%Ñr1ídmm“°EØI’¶šÁtÐ\Sø¾A]Bì51Ñ4—­ÙôIqkLaÍ%lÇ ·Ž"¤^r,k‰C&´u qnY!rárÝÞ JıÙüUBt’­H3ç¤8¦Ï›LœÏ.f¶vªZœ%¨¤)c€Ëå 'Òøˆd“›‡äK1‰‘é% ¬@¸E— ™|  Ø¼’"¬!' ^ S ëDÌ&WlcÎû‰MÞ˜›Ú{==då!Á—Ë+_ç68Ï'мSúóRƒóðê=ø“DG»º#'ˆ¸ÏKѪË=úœ¥ã€´©S8ó·‹®|YèÅë#_a © 1§¿X½Ê%8ýeIÚå8{ ‚bùšG”Åä5——îØ Ž¹’ô 7<¶TްªÍC•bÞT’mrÚ7Rdh¨à©VwÕ©©•BÌ;àì áÑŠX žjIožÂr€(ÊùvWÿ5!v 4B*Õª„YÁ¶oŒ âP½Ö5IÓˆák: å$,Ñ!íF"u_ã÷jÖs`kt¨ɼD (ê€cî<•Sl2ˆ¯¢ðÂJväV£ˆ/ñôOAü@¸TžBÂԧ冄-÷i¹!a¢ä”oHØ––ÂÕŽµíkê lÛ6Š–QJö{!bn pPçhÚ'SZx7 è›,Àéú·S#¨ÓAHq v‚´WØ…ºÏ˜ªÏ’)ÎUÚAOCcã53"Ʀ„&Ç‘LNõ´çÌTÐË©Ió‡N‘åÒÙ#Ç1Eºâ»%%F9_ÂSÚzȶ³ÿæ*ÄKÈ1©@,íAÏKÆâ@ý$Rg>«€îÈîÅ“æöÁýD :r?yy™ ðÃ×$ËKHΫ_OmFD¦Ãô»o›l™ApÿBºãô¥;Ñþ‚¤£ ©~ôæ‘™BC7GÁ—ìÅ›,š‘Ô nã¸xh(ïÑhHŽ08P¾ˆô›Â»¤ +Ô8$f Xáˆl@È_ö½Mwr‰¤ö­ Ñ —D÷ë³KÈÔòŒH&âìíÞÇÉ9j¿- dÜ:9‡½²Ïˆ¼©D0ZéñÝÞLNÅÑò½I å6/¹ ÓºÉ9tÞy‡È«÷Œ ßçL7ï,µïà)ÁÉ‹Oòâ“|xñI^|’ŸäÅ'!ò⓼ø$/>É‹OBäÅ'yñI^|"/>É‹Oòâ“òâ“yñI^|’ŸäÅ'yñI^|’ŸäÅ'yñI^|’ŸäÅ'yñI^|’ŸäÅ'ù/‹O‚t%% ex BÉð®N€ ^´*žd‘º×(ÁKyºFˆ’1=NxJDÙ\œ9æ šþBŠ2ÐUYÆ~3€Þžò<a·8êÁéœ>ü3»5'S%’‹QuwK›0YU¿€\µ)N˜ô3¦‚jNÅC„i÷Ú”‡‚H \÷>NÚÕVɉ½â6ÙÉÊl7!2—òƒµ“vç$Òv,¾×¡Ìdu¢ìeCò˜°úIIIäŽX[ œnÝɉ‰t'\«pℌ“oŠD8+F•TR×)°iJaúÃÑ¥¸Q^9³ ˆ0Y˜‰0õØ0ká */ÑÂØH„™Æ"ŽU˜1‰0Eâ´Ó‚‡&òðDŽ zör ˆi&¦›]6"J8¤>d8 m¾D˜rkÚ­N„Ù³¦MD(à–½óD(u–+†D îáP³´£þ3˪.ãüg–]©¡é-ó5ÍIG¬b8UyÍŒÅÖe.¦/Ì×T-¤µc•ˆ??—g7f,³^昅½*óÙ6iè EݳkÌBÙ™3¦²°ÂeUî5ÌÅw”¿ìkh;*×8uûB„m{ë;ˆÌ)ð5Ũ;1urŠa"µèk–²ë§+¹›Sû§cÓœ/<­`MÉþÓùЦ* „wÓÿgŸÏ2”.8—XÙeL/[î5Ì*iê寳î#m-%ѱpZWö÷l ßAœš5ræažé|'f̵"û èWy»1+ºÂE Yö@˜–z^uXfZT ýÞoy*ùö´Mú ƒ²o*W%¿W}ß¾Iwá†k.k¼U%¢3/\%–®lÈZ¾(p•X­È®„0ïrÉE½…-ZÝ™Åʃ»Ä-^ö²ÄU`VÁD¦Y¼æj1˜:뻜ˆî«{b'1á-ñÄV&I–'ñ“’÷…(jr¡íΚ ,ü¼à. ÁT`¥®ÙùÞ¶éÑ»Ä ÈNÉz/Ö”RUÙ“]C˜Èɺ1î™YmÆúCÒa³¸··² Š©=mÂØ.¡‡Šüd@(vŽØÛ5OSTòÛÛõ×RÔ¢H®Z±«ð\,ë>¥à@ï- ŠœÖ{óL@ÖS,Ðv m³X©¹ÁE%‚¢½£ÆbŠ‚¢{»*j ·ÓV•¢`ÇnY^XJ&·½¢UޝYµ¥dÜPåT¡J¯½¦ŠH^ê.•ŸŠbX{¹ü” hmC<ÕÜÀJiyØK´Ws«`Ío…G™у‚P·Wуö}¹UGIñ•ýZ$ëÙ‚ºçVÝ»Æm ¬aUŠG{.—ž«1<“ç|–z«a±©ÒíwÝs¸ô\·u{ÏîÒsAÛØJó¹¨bÄsnV—ž‹JDQSe÷êE¢ÂŸ»ofW˜ âåžÉ&£vàÛ&‹­m@НÏÇ%Èï!õãQ7‡³šû¶ ¬A§y%£ý£òãM•GoùH Øi€ˆA°‡ ªÖÈ6¿‡ü>é"Ɉ¼€{DAÕ'Ø#¹ôhp°wŠªQDH”õ€öåÒ£5žÕ§‹?j÷áÒ£D¥Gû}V¿…b=yzséQëf‰’ªQ/HT ¶ÀêQR5j¯ížßWÞPix›°Naɱ²›ÎË@dÌÝM'_ø˜D¡Ù,-Üàbò›ãd0„øÍÛPõ+¢ÒbSýn:  åæ4Ü JcÜŠæA‡΃4‚Ý´?}l7²F펲Š}Ùº5Eˆè¾ö²ð¦º¢²6ZŠíÙPùÍöõ­"¤ ~+¹-w”PAÉØU¥ÎàÊr?¨0(/`{HáÜên®°ðn%77IŒ¨& „Òn˜z¹‹\XÛTSB%5œ`žêQqDöö]´Ã-æA(Ã/.Òm—2?¼ˆ•G™äµ8à²KÑ+ÌÓ±(4È*p/œÙvï]”0Ü(±]“{†zge 2Ëﬤðôµé%³WÝl1?)ûfõl{ùázó¾¦Ð"jl–¥Î8/Œ¨d8G¬ÓìÍd†ßgYÃÆŒJ¹@Xrš<¼ÓVñrXªG9#þj'†$ŒÕ€7¸©ø \zêƒ(K¸’½;¹:ñÊvbî$Û 7&UU™]ÁÎÜÉ‹i•h×Òª¡Œ$ÅšÑbZ®/h× !€ S±BÌ-i_Dº$SÁXþI—üZ¡Ñ¬íú£+¤ÞÚ"yÿ¨µ¿¶×Î –ÌÚ®6º¢nèÚb)Œ©¹–ÞãºE€p{¾®œµX}ŠNÉ%@œT ¾‹+‰î¨$µ–U¹ë`[Kþq#Kú@Ô§ZË+e—è`3¥X7Û7Hë°Á×ó¹¦t³mZÔRá<ÚÄÖtuàë¨]Óu¯w”0X³«òï¶5ò Mëm‡À]Sþ´qq‹Þ58M=#Ì_c•q7'I-ØA[]Cg>8Vk”<{y‰­á2¿;˜hkLÍØë¾_CVN–A÷ÅC‚;6û¹ÝY<]ý7HBà×tM:\÷ÃYn^ÊÂò–‘/juZŸœºR\–i@ü]ka^RÅê*93Sy‚íˆÇª« ü”óyuÚÓð5fâ>uÁ›ï¡P@Ò÷X!üQ7Åb¹X©CË¥Ù€øTì­Áì-k5yø9hYM1¬óVXMN& ^ðMv ^óˆLä5+&dÓÔŸ—±³dKj3üvyUºq‰ˆ¶°êr‡]|UzˆxÀ«8 @,)ªx p‚{ÖV;€TßY¡Pv”AÄ ÷We}úϹ¤ó·€k•ùw†§1¯"V(œëžØ¦u‘_Eúݦ L¸'ÂsïA|WÕg„¾ø€hw\EsÚ£yMóË[¦Èïo•iK Ì*ÃNÏ¿~“Å>¢Í}¹Š¸žÃY$†™£˜ÛÊbïÑ‘c9¢€ #*'4/Qn© (ÉÍ•h¹n%±æ-ï±0Wº’VÒáˆgcRÜ3˜ ÒºWà­²ÿADšðJò]±üLªŠÄ?ÉaÄ †´¬ ¬Dö“Bî2/qnÉB &—lù‰æå/Î=$`*3z{ææ¨‘¾1·gtT[ZæûÊëAB†±‹ˆ6ªAÄœ ìÀ~’iŠ“‹Ã\%Z‰¹ºF ÊÇ’0KD‹,éf2ŇÍp”$ß»1*0N@=<·oÕfæTì9ýjÒFD*ÞœC;I ÿÄŽY„èà0§´c bsÖ{g߆Ök~wkò½¯tì©ð"Zís¨ìµo¨ü7Qç˜÷>Rš¦ƒ.‰¨ÅCöb0ed‹˜£ù>D 0³’†,¿@|ãì_ ¹}—Údì3¢Ó$iS¸Ï€™@<õ{¿· ¤ÝÛRoó)û¨»ìÎ{FÒ}M#ò´¿CV4ðð”ÑäÅ?yñO^ü“ÿäÅ?yñO^ü“ÿäÅ?yñO^ü“ÿäËò⟼ø'/þÉ‹ò⟼ø'/þÉ‹ò⟼ø'/þÉ‹ò⟼ø'ß½ø'/þÉ‹ò⟼ø'B^ü“ÿä?‚ (ùc:ÿ~ûËû×LÇý÷~èÀíbJÅñ蔦ìƒOHWRÜñÜ+Ê­úŒ(·êsW*·ê£ÔÝψøyÜ×±«ý=2XÔˆ-ð çV-6@`Ƥþ)²üZ3¦™_aΘ‰Ñ8á™ÒæÙ¾‹S y»kßu î c ÖŠº+2@˜Vî&ŒÅ SwU«|\…KH­±P‹ïÓˆL´mÆâŽL´½…è¾ñÈ!$Šoó›é„â%z=;ïò ûuç-$tç­ô›³„î\D@¢oŠì@|ÈP">eå?̳ÞSF¥ ÈnpÁûé_5%¾yÇq¨¨F†5(f¤€ñ¼ÙËqVál—2ôÎ3Å׌È6'­y+›~U½‚¡f ‹@À²’ÎñtcîÙ›‡4fß„{Ö×4%É„i®„¡sʼvw`$êpÿ­˜·uøáËæqœ+Õ;f)üF|V¯ÛßÑ•zóSêyœs)WŠ)ز¶Âu>ñ¶©4ÒÜç{Ê.F|†Ežk#~uŸÍ™.ßG|•IâËO `ÜC?“ÍÄ“©MåÇ…Ú·Q=1fêŸÀìÄ«Þ"Nï£D1à¥Öå²²lÿÈ·l€Ï¤½¸l€]¿´ˆ$#^G]%³P~Àv¾Þ\9 ª-Ñ1ˆ¬¸ ¯ò¬G}„kÊèSÛʺ¦ [yPÁKBü ">z÷í]÷µ¶²Æ/ÛÑib2†*dXœ!çŽì0] #¶\é„„¢_h[ˆMNC  –À6›¡x„gò]ÈÃØÆ á Î1ý¬¯ $€x–k°}ßÍôÂ;Üb0RÝ-Vùt†éí´÷°(RUÜ=FŽun Û QdÓqZܲ@(E÷=áCƱܚ$ŽðEÆöŽ5XM‡‹p[VQ(ÄÆŽy+‡XA†4!ƒ-K„¥[U0®•·«D‰Û»$€ôæ|®k²Xl‹uOl®]U•[R sí’%ÕR,0æˆ/骯âù³dßÒ Ki!`§W•–ê{Ènø¨äB£úTµ—ÀÒ÷k‡ÇlB}ïK%D€tÛü“ ·DIÇå¼ð@<7vV½¯[Іž‚¢j5^[òˆeù¦ù•5nŠ%ëD ã誄ĭF\°(_-`‹9 äúGêRQžðˆzÃ2=a—ç&B$H~®í`ù’}½5Mõ€ÌZIJ¯ÍšA°ŠT¾NA;…†€Ì­bDõºŽX× «Ô+œÔIîÜ’ck„—JÕTÂoU‚µIßVUõ$» “Nµ@Âý¥q°¸„²r ªåí n6V(+å6FôñGu'øæ8±o9»\Q‰lîðñuö©æ¢ÊMA½ŽA ö³çª‚%ªÞQÙ¨¨ëG±2'+¶ˆ3ËNƒÊWf£°Ý-O'K¡•„‘,² Jl¥wþQ ÝÍsõ6°!/Âù‰( {27‡¥^Oa¶lªá),2O±ØE¸Rk½h§Q7ÌU\ª­Æ‰ÈȪ?Öì’®,îó|€…ãêŽI! eÍ̦)MK·^B„½ÖZh[L"óz¿Uü,òünû$â‘KH‹r9@8R^b zN~ f@QåIÔwË×9ÏÞëuÂ:‡­Æ´†î›…¸oàõÿÈRrö}‹›ƒrsñuëktLÕëd¶Â9}KØemÄS¸êd_ZHT ›­ÛA P‘AIÐQ¿ôŸàLT#«¾›DÀTïMRaïõ¨z¹ÅðaU@¹ê’䓆÷Üz š tì!Äë¥%×-ì–Õ@(úˆQh®ùØ£î Õ-Œzà ¨náŠ)ÚŠë^ŠxAüUø¤€p5ÜÊŠ@(DG² N šJI|g—ŽðUë „7ñßåEHÏFW=@­ÈËß¡1ÂY³]&ÍýjÓM / y@YˆIQ…Bʈ:| q©Œ(Ö„ý;¢ÞëÕˆ2Ýã])ºQ ϙ鲔dhFÑM-zVÁºõ·ƒþô(Õ¹EFb9Ï hYšar¢*¡Qd —Ô|´æı“Ÿ‰b£ÁÏêš3l=@8æ¼ïÍ»®ûÒC l†‚„Òi„2î#+ŠÛa{WØ2€°½+»¥ËŠ¢«–{}iV¹,c JµFoUº\Q¶—áj—ëG³ KÀ¶K£Ž²¬Ï]ÇY¿¢2¨6Û/Âuœb²ô80·F-½è¾´BWœÄ±G€jgبÀnŽÝpTɾ ‹E¸…ì‹ðöƒÀ(î!êòZŒádÔ„ì‹`¬v¿´B—ý¼õ~Mr,A;&@%`ÏX@°~daï,ÃzáÞ±¯ 9VÌ˸d½õ”.OsQO@ã|ì†@¼‹­ZÑ&2aE“KSß-ž&Ù¢,²<.R› 1׋°î}ñšS¦ªGùfpMYŸ>Å! ÈPhïR"|ñ¼á½>׎ÂBäv2,f<˜XëKE©ÇE ;P¸º^¦Q·Naq÷|å‚Y¶(‰Ý.5ŸU³K *‰…¶ød¬µ=.¢’ØQî’ߥšÝñô®¢â9ʃŒƒ!*}›8‡^½D™ ¬Á^Rh‹f£¨ø¸&; ׋°HvqA•·«• Xš'Å¥m`ïEÍór/aÍl¢\„5³Ké@’æÔ)3Tà)Æ*®{’N•h®å®ú)eÚÝKar(åîás]T{O—ˆwì¥]5âK »«Ž¼‡ÏÕX€x¯ZY˨¶`ƒSfúv¨ƒ(æ*P•K|ú *‹†^ëŠé¿ät¨u‡~¸ŠŠ²×»šÍ^›:¸öZ˜p€pN¶0Øh².)ž«±ÕËÐWE™ú %/Õ᪠%×@84T- pÚ6— Ť=!(ú¬KSÛº ý®iÜ®º½d`­òRÝÚÔfª'þ’ñå|ã’M¶öðVa‡÷«I‹Cx°_MȼAœùØ©ØüÑ…mñtAÄ@˜¹ÛkĨLIí Ò¥«iÄkj­ŒtßHªV7è`1Üô å¾O©÷yÀÐRÀ:bå.•G©£ßR˜Z㮥a÷쯚"¹„`}ÔqOúªxr}çXº@¬ç«ÞJ«3ÅÞ!jåAòwЬ«3ª¶n–!`‰  1©±˜\¥¨ÎvsDž©:¯ÊÓ±rÅ 6¸ä]‡ºÔyÏß*„s ê‚è™ì{ „y£ ¶(õu;e« ÏArŒ¿*õäÞ–ól…Ï›Q8uݪÒ?¹ tËêW/v«`‘ÛÛr]÷Œ|‘y#²DùZ·£²t€ucT¬è ûö”‘}Ï¿*qt¨œ‘E@v~Ü/+% *4_ ÝŽ’-³îþÉװϨüÒAÆS_ ™O}ÅþÜa|FÖ{`?õãÐSzê+!ùFe=G_ )O}%¤Æ‰ý"íÆe=G\–…¬~”áäE+yÑJ^´’­äE+yÑJ>¼h%/Z ‘­äE+yÑJ^´’/ZÉ‹VòáE+yÑJ^´’­äE+ù𢕼h%/ZI /ZÉ‹Vò¢•¼h%/ZÉ‹Vò¢•øÛ‹VòáE+yÑJ^´’ÿri%H\’>þòÃo>äÄü%þïgßüç_øg™Ï,Ç‘ötÁ׿pŠ“ü1ãxÉÄ>HÓ?~ýý‡/z._~ý·úužYøõ¿úðÅÚ_þ× MÇ™Á÷?ÿðÅчp-ëôÁWÏËú—_ø×®b†4¶“5}`T*H®ë¿ú™¥]'Ô÷2 YÐWTÇó¯î{ãFMäl[ë"ì9sgsÚÀwH2b§~uÎÍ'„¹ë„ˆQ%úRÜÓªª=#GQ †@˜½ñ iJ‡èDDΣÿô•)ba×âè-ÌN}5£Yº„quVg|‡”@ÔWe{`"ãðĘEQ–³$‡síi@̽8zÝ4b~FÎ’lWé,t›à»9…\ŸqÕ½ ׳ó@ ÓYps•¸è‘2™ ”Œ˜`R²òÕ¢™j 0ÉCu™ˆ8íT–ç™H „)+¯~Ь_£Q? È…i²s ¾¤+çæÌÇÉ:Œ2‚Äã|ó£ÄËoFLB¢¸b$’Œ˜’uŠbn’rÝ1}$9t ‡:3í e}:›<SU’“€Þ×@²s¹ÒpÂí$±$>ÓÏœ>ˆÞü¢‘K›y‰„ÇÂÙ‰8WV*~Ñà0³ã¢yáTñ´Ý@Š)øH3ÙÈ`‰t•›ƒ^}*¢ Êó`ÑäAòÌeD[<g/þQó}“£Äð7ó8€,@g¼¡mÈ ‹ê$)~pvº’QÖ?}Õ@’m&H•Z÷ǧ\éH°ª‰“=tÃ:(È™º•/‰7‡u蛄}X9 S#¥JOÅÌ4k$ûQ’y[Ú`¶Ó\ d¹íFºŸ-âÖȦ ߸ϗrIç0¨!/ï Á@ £öðQvä°Ë!#0Åê=ŸÑ“¦×/2 3—m—$Ò/#Z›@Ê’ý+Y+ß:EÆ×^¤iþ•óæfÊÊN¹Ì,üîÀ™µMeÊHÝÌį¹[gC¾g&u¾ö ¦„æÃ;"E|ŽC’MsÅ\;.Î|C¤¨æç )"±5'S²#N&ÐhÙì$¯@¸¥äòæK”‹Àp“÷sF¤Ï‘oØÞø(}€DàLzœ#‡îÐ9r›?]>ÉÆù5Å”ìNÒ›“7E$-×xFJvæ5ǯ’+¡¡TºF,äGÇb¤0á‚M+ú¯;ÉuZÑ]%ó€´˜7Åi)ÝùÞ)òY"<ßf·HîB¼{r­ŒH®‰œóœ^)#2ͧò®ÉÝŠ2Ú¨‡Ý@,6›åÇ5")>eUjñ¢‘|>íÉõ»¯ñB²dÈál„K+91*óø×O¥:†êt[v\ €ˆHËÚ‡R” Â-/E¢WàVr*‰‚ÜÛñ]•;¢®JPS…c’|N(ÃéêQ±dûìBÂìð¹+ŸX|Ôæz)+į½v@<¾µº^Ê´–8d#â‰ì´ö¬»âkäWDm 8H]b9Š©‚.L¨Ÿ¤Â@ÿí¾Dê 5⬺0R=¨?.!Þ› ØùïDeº/jŒÔ¤"Ä£ç\øýÚõ©¨!ž‘…Ž "]PTîÆ’Ýžæ[‡Hàí ×­I1VÅ­Â3ám•¥vd…ã–¹„X;¾ªódlc|𠽜[âU—Å™EÑ‹—L’Ñ­D@On$ö =9J•x'!â)K&EÁ!i (nr‹÷ë ²¸'âW‚ZëÊFn^Q²uu©¬¨ ‚,¢rWru#,è³Âj Ë ºÌ³NÖÀ³&m÷MŠ5Ÿ–ú&…K®Aµ’½ˆƒ*Bµcš'Y̸ӓìì¨åNO Äâ^O,(ËS`Ñ6—Ü>mþýz¤þTª*C|±b×>7R‚U¸=ÊŽÙcÓIRP+^Âe¨f@ö° Õ¬1Ñ!ªÉYH.è–=Ddiƒ\á`Í ›á4ž‘AR„3yF¶^ ê°pèBŒ¨Ã\‰‘{l“…r¤Iauªy•!#(a¢<îa|‡Œ`a»{^7ï …Ï†ïS4»G‘ªõ‚sß®W5ºÊ„þTa-C@"×+µÇ{U_Š#2aåÀQ¢É[ˆfjWÔ„ÓN„%?Œäˆz0 c«#m:—<Þ³Ç>ÓÍëâ\²^ôËdÀ‚¦xèËóHá}¦¥U7£ˆö–<㯆U.ñ!Ä#ºÊ­«×|Ùóúåla °ËëÀ¦ïÕ 0ÕÕ‹<ý,ÜÆ_Å 4 ÒGXîoâKš½Ú£´Ì_ù̃%=õ½Äwüý’`°z)p/O¦‹'Àb†ƒX½8pß‚‡DX¾­aÝeîX7qÏ@8KÛˆõ8åÊA±E¯YW|RV ¢¸–3z O?‹ðñQ5&;xâ©3åC5H/Ðéj-ÇPMžÀx¤fRÕ·(3™±šY±EJ¢B˜¥³Ïí±$knp]Õ8¦yų,¦ç’+P t¦¥Õê÷(¦I„ý^#†gE KÀö¸hŠEx¸¬>¨íyn6õJS3°xdFu‹ó}¶ æöË̃ þV¯ìÞ¢)„JQl*˜ÎWc±sBÕ‹‹þQ­Ô¨äEÁÁ‡Ö1ÁKº7Cœ:%JtUim!`ºÎ”@R„£Ԙ&ÝuPK¸}(XŠíÃ@Ð3Q"6CŠp&•lÓvLòtë„9œR3S€eo±µjÝ ]ä î…®“**ÜF/ˆDÒU9ˆˆ&@¼9t‘QPM·ßkøÞ—¬káDÄÓ±{#ÊQ^çZÛ„o\BRô,Ý"_9ù%ÀÛøw¬;ÙÒÔÅ¥dÝhï­«¼úÞÖ§ :1ÜítGÔ´5°vú÷®ªXOÿHñ±@ÒEXŠ{÷˜kçA}lÏ5TZ›B,››˜nM‘‹FZVåíylý@âÕÅ¡C½îxõìWO÷Õ]k~G)M ¼(î} Éá‰bá;Âç â4RŸ*ŒP}óc^i4!R.¢úæq¢õͽ':² H1²\ß<<Æ@Tß¼F7×åú榋sé*·î Vå¦GIö}‘>„x@«Lí†uQAx-;G‹µ( ¢òᬢ¢òáAàæTTT¾^ÇfžwSmÊÚô;Ö?ŸW”T™bQã>E‚¼2V`ÿÂÐnp"·8LÚ" ØÚL÷Q.ƒ0`Ÿhc‡:XÉ!;ȺÏR¹ø}Ê­ºq­]Lòƒ„'íã'vÏÍG]éQÅ¥iØ‚/Âi|C¢pßmºa'¨£]˜ t¨ ¾s¨¸„½>¢NJWE¡Ö·Ï†Ýñ‹­Gélå‘@¼#¢Š8ݧ­æ@°Ý=Ž(Œb•ѵŠZì­·Û¥¢Oµœ>ª ¼^ŠXÒ­1Èàƒ<„â¤ÁǪ ǪEÙ&×O'’îÓ‚ e å Á‰2ªth§›cšyïZ ú%õÞùžÊÒRlÁâì…©÷5E3o­Ü×®µ0|P¾¯©@שw‘Ö„¤‹à j{"ë¾%ÏIDÒA¯×yßrh¢Ôp v…ݶÚïhÊ:¤\„“´¶êEÜÁVMñë®Nà^Á­Š•ÚÀKB3 ÆÑV¯PH0€rèÍ­}:!öõV®Pˆ ‘ò@и2CM‚­+W…(¢LÉO/‘à„¥~‰T Ý5¸Z¹6†V"cEwès+9Ž–OH)Pþ•«’%=hy‡ö„x‚½e¢HÔ Â](&_|uµ,èÐ\PÙÅ/@ &’…äO(,ö@"6 7»\âÌz‘¨øù@®}!´ïškZÍ·¼Û¹ öEæ]v$÷HĨ=Ÿ¢È|Â!]û¸æû øƒßðû5¨}‚„Mµ™PoYÅÒ=ùHp™<ª¤;•{gÃnAÉê(RÅn1K b‹mOãGß–MüÚ¾­)¼+„@Ä‹¨+ Ŭ³u¯],N¬Í‰º"È‚“×l·yErØeÄ!9lF…v fÚ{c»äG5B×ÑpúýI‘3XK¤£ŒvdIÞŠ³@È’œù#ÁB¬õSLÅzo3Þñˆ&įlÒ$óm YÛQÕ‴ÚÑb0‹™Ê#Žn@HJ!RÅý”2d<è¢øº²è¤é>ˆTæÉ>š91µ‡âDiUEnÝ%f¬–¾D›]÷W$3öPŠ,Ÿè·•œ\=<£@8•Zp€~ÛÂI Ûľ·™æ{R娓Bü ð”G@ è J tŸˆ§ b@§˜[Eµ"A¶n3LX} œ5ø¡@8j¤ÀöCª,á 3È[¤™w^G $îC']ã\ƒËŽ¿Š£véå¯fÏ×H!¿&lª¤_~P3µëAæBžæåÍ!;²Dªl×Ùq–eÅ\ã’cö‹QòPà®Ï°†p¥@3PXÊEd~Â#,‹À—íƒsûò€TÇfh£BZnŽÌN­2 à · ÆG…”‹PH]ß=J Õ¡ 1heÅ]”@Q,Ñ6´ëuŒS¥Â£S¾«U3+µÛ6™ÊoÜ T5ô”Û<Ù¾7V"ýxPÂÜ¡BeÂ0¹}Y%!„G¾²z<¢”€0-Î5cÁѱçÀ3K¹Ï˜Á}<´K•¦J˜¬ŠS,¤ˆ³0xMAwd’…ûêLNQÑCÖMÅHÕu&âD%Ìm:ÑÁMIÓóc”b`VuO ºIs‰Z¤Ãðþ¡,\Ì¡aù<»^Ž8RØ™"ÈpÚ/³é¼‘#ls°ua·ìÎ8â µë™í¤Æ‚y¹M,°]Á—Qv Ji`hU²i¹Ž()rÈab˜¾ª’a\gl¾ÉÉd|ÒX“ôæ›é˜M¹® ÛikâIƒ+ä1½ã’µ¦¤ˆ0kË R„•´Í9áØ„=œ=ž‚Ò„©üö•¼‘žp›ÅÓ{V®!‹Ž^iÄQê%æ9ò€¾àDH–ñ FMçSò¯š,]EyKåë d;LpC4#îåIbb©¸ñä>…lT2ÃÒœÔÊGWÄÎ>jnöóŠŠót¶”ç$\pÈ”¢Ô]6.(ŸS€ÙÖ•”gˆÍ‰©*ÑÚ ½¤»Ì7r”Ù W‡Âñ¤Dk±Æà›ŠDk6[ÈÅôl6š$‰&~“')Ê#\wír䘳y+m=}‹8Žç ®=½tBâërö=œ²²øÙà„Õ—•10~Õ”t3¼»dû#¥!]‘Ì]”(¸+KdaÔk¢¤z’.Ây2‚îDg)žþ°Üg)mùa¹/IIÀ–û’œÙ²_ädÀ#Ÿ(ܽ|¯ñ°ÙU§7}ØìªÓ›>ŒvâVåqYI‘Áp<ŒvÊ`86;å!É=¸ºaë6;™;rØìdöÎ}†'Fvé"œî=t'z Ð;Pì?"÷`†§H8}”$÷zKNXh:]>¹—뇾íu’Û•‘Po¢Û¦ëó‘Ã,·K ²ã-Û«û4¹€øYMˆ¹ÍðP´¬sÃ}+šÊíšñZÕT~î[SÎÅÖÃÒº&S»„Ÿ6$†Zĺ’“…˜!Õ¦ß"B¾Ãóœ[ÄëÁ_Í•ß.½ÇnïÜ~q‘º•¶Ÿýí7L.y®ˆêxÂðõ°3Ò Ö8Å€iÀ™S/§“I¤^ÂÓ¯Öðóu*æ¹FšMp'˜õîaÝïRºs­adLÀn!s?r½ÒbeJ¬aŠì”¬5rˆ€ˆÂEö°÷GºÀQ A‹ÉÅG£àÒärýäCtÊ\Bko‡^"•ü>þêz†Lk¹„E<"®)YAˆÈT•Kdú‰‰Á@4Îã Ó±KRGX>Ï®Tº%"@ÖbÊÏy£IÙäKïÒäyf!^ôSau¹„¯pŠÏ›K¤¦ u ñ¦·¤4æ1ÿàÃ1›m¾îÃUÕd˜C(óÏšåLµyút„-Îq ëJÍ—³s\’B\˜A·´QåËìY:v± 0»1çH†$€æÈ—Þ$³rž*X¤•ù±e]Ò&§@¾lŸÍ£.KL³AsÎADP"õ"L˘SôÞ^š´9²)uçÇË×CÞ+ÅøÂFäÕâ¤HÁH]Ÿ˜¹|s0_ÌÃu8k G>ã¶u¸òšA_¦1µ9-‰…Ôƒ:Lûf5Ot9…¯iEû(GRžAà ‚Ej6s 'l‡“Íåt#Ô²˜0@Ú;ê8sß¡ ðšÿ¬üŽÉžSä7Õ3)¥j*b,D_E©O7›ÑwEˆ£•ŠRª§â~º®†Jú’t#0&iψ/q¼‡ßE§ís7㦀—â„Ê„B‰“F„U1¾jžI7„#³9ÁÅD˜G^Bª¯YHò5rÞ¤]oH‰Î@<šœi—ZãG)4úöˆ”6“ø¢´ÈÈâ¤àýG4å„MJc@dêëx ”¶Óš"Zhû¿R;7Ý$ˆLû")‘„±Äì4«ÓŠt"ùº’2_i¾qäüÌxC¥ò â7t­tÓ$ ÔK7™7¦,áëx (SÓœ 1g|Ã5np3w¤ßàºæ7 -pŒ®¯~agbjÿ2ž®Ûzìüe*êpZ'FpÒ¦´‚†ð@ݦz†+.Ÿ€'‡ó&±Ô™ŠZNJCCDþê´Ê_+ˆºzò®gSû=r£2u~s6Ÿt}ŒØ Ä‘0N¨Ä‹Ä£@ÜUNa“ÖÄr(*ßFÚqZN¨®ÜwÊ©À1‰Ka%Iiˆ€8KHšaTBPn`™¸å(âi¹EëâH(D.l!ÂÙ⦛maÆPË€X’JŽñìÝÊbÄ=¾EKÄ™ ‰äN=õ#}•I;ÍHf˜ì5…ÈŽHîÕ…t‡ˆ“„B¤Ýq®$¥úNaä» ÉÜÞúš÷îÉ:ƒˆ(wJ²iU…ð[€«NÀÍÓ„D|? ¹\Í"‹¡ $ÂûºŸæ¸Áüڌӌ\N‘€ˆqÓÓŒ ·È6f${v¬y—ƈ+{ØM ÄáýE\~ ®RºgDXا=|i†™™¦?Jg7 Å’Åá ©âgÉN}!_Wº spbùšä´JÿŒÈî˜f/‘è‚rˆ ¬À¸•4CcDŽÜß!-y4ˆ5@D±BÎÒß#Ô*Ÿæ±k—Ò¥ˆóiô{®4ÑÞïÎÝ蹆íÞ«¹õo_±Xä’`Ôµ5‹¦ü}ÇÒ’ÖîH±55•€9_¯XUA€1¢9Æ8öSÈ«>îN^—z3˜<qøe°}ˆpã_i@¦ù>¾7Çë{qW¥»1ýDªãþ­Wgà˜Y¦j"LCPbíT1 Èœaèêl1NÒ™¡µ@²S'X9rÂ<Åg`ìß\·ÅÚ^a~’¹c¹¥1…õÞ«(c*¬÷îã¢} Ö{÷qÑ®Dëý „bvµèQçÔs¼½Æ4¯»ûKß&âIP9@ˆp¬Ö ARVÈTðáÁÄ;ˆ÷cÖš×Ëä(²žU’4OÎ^k…J§ æjQw9õ,úî “´o昃ˆ0¾HžšÈ{…TðÙü&»9Èpf–*`ŽäAÉNjܕR-]]7’ñ¥½.K úÞ¥²Lû7½ÐA”t&þU©S©Œ|nÉ‘>(ĬÍ´ðû6tÃÁÀŸu—¤7Ϧ¡Š|QQ CN¡SGj¾œC[O*=6Ìn=€ª“Èg~®Äg5fMêN|Öâä;`:0½@ªÎ{6âè+`3ÏF0œÍ·]òfX#¢¬g;Vf’Óˆùnh}Ø{;qePÞ”úœÄnïåÄ•Å6ï½§Ò@Ý,|@ò~Îø·A`qÆ?§ô–}þ)³!D$òáì-ô#ñãÞ‘Ÿ3RU‰NL¼•ºb¿Ô”K׉¸Œp¦´[C&L"Q@*Üý"Ì×¢tläÊ5,ZÄY#S?^¼Íȉ÷ڊ侑x¯­Èƒ=UIggÛìÁ;ÝÊ1ÞàõðDQò†œx.·â2£GmÖ(WˆKtñà qŸv¤€øY]´qdís讃ü%Qv¤ûÎ7±9Þo:"Ó °p*`*¢l §2ŽdFTn•(í¢TSt«ø>Ρ;‹ Doq“Gö»±¢Ð‚ áÀ÷â,ÔMAð½¸äJSvy Qb§+•âÌ·Šã±éñ%Nl7¯0hUÂôÔºÐsGjH#•Ý”ÚZ‰m<¢7Ñ”€xŠUªæpá¨+\bi\Küvš:8uÔ˜J¢$\:QZJ¡š@4']#Š>ž³ßÍ[× ª&íxdL¯b}ÁóãjJ.s$êZ%æºÉŠx§ã‘F]™¼C.¶Qœ¾rEEm ìd-Óˆ¯¬GnuÇ‘ ÅÅØ+"P·Ô@,dJQŠÆ5-¥‹L7<%U«£ÁÍåФpŽGþõìet¡;;óäÚ12Îò_”zPéB ¦d©ôLù¹›˜RÖˆ3JîÈç³­ Î*ßFÁ_ô^í@ØÜÝnå8’xÆš¿w–»Ê –‰Iïqýé(úÇÞ#j—$F6¸¼<‡“3É &€|~¸ÀÜ—-â«þMƒSÌEË’Â Æå ˆ"Œnp“¹´ªïqÓMâGsÞáïTÁ: |i.?»•²qJ+úN¥!1NLhì_)Aã¼´ ”œl¼ÆÌŒ¨SÉLÉ[ˆÂTæM!0èÙK‹˜‹Öઙ¦üÐo§æ‰äÌÎß©Èç.BܺÅrôŠ„¦"À%c]Z9¥]õv±ä]›—<ƒ ¥Å×H+[ªgÕ›Zr”BâÂÆS‡*x]åvjÖÏ€Œ:¬˜Gðº¨Óh‰"±\èD¢N¬xap,F¡XQÇ€D¡XÅFÎìÛÊ”ÅÄÝúÉPàI¹å˜ÜÈkTÆå äˆ‰ê¹óÆk£æîäK…KuyçâÂ˃õ)\˜QÍWHTó%½™HTóUÎ?¸9];Vñ4Dü]^ ~ —1†+ÔïÞ刘yÝêÌ<^ðXõ¦y“#vn)5‘îkh'âJò]¤ ¸T£Î²Âæ¥ÈD_u×¶5ÿJÄ;¡º4LñtÂê9N ÄÌñzŽ+YÃ-ëõݤPq‰Üæ™]Ìd@Ílì¢>wñk”v*zw}›äÛDžÅ¨ü ¯ZãL3,êÔW¶“½Ì±a(K®— ¥7Š™ÏðDZ:\±@¢.9©$Q5=qi[%#"^chŸ7õʺs‹¬üÎq,A—YNó0/s9ÁÏ—‡¿Úõ‘KÑD»© PÐ~-!ž3…eΉ¸pyö¾)%–ªÍ6ø½=‹X¢–@óä9â²ÊYÞ, .ß ¶<š\StXV¦úy3\¬¬~^ ìR„  Ž,Mˆëc'>µÁéî釴Î\è=)™#Ëä®,û ‰”˜b€k8«3WÕFþP¼U-± ÁBÄÒ%)ƒ< ý$F#ØÊOR€h{IôÖ —ÍM;G›‘CdnÖñ—u<ån!bRƦd$bÞk†0µc«°&Ië©’ï+Uˆ9%ËÛu CÀä<ª·6Àòœ­Q>`yÊÖ[a`yÊÖŒ—U%é|v§kŒ¬,ùZãD6—"æMÆ2ED};Å×âûا D'©9•·ˆæt€—c˜¢CΠŠO‡0R<"[ÌœJNÄÍ‘ËeM´¾çØÇ[@Â^¸gD¶L"j±¢<ˆ¸…Ý^@\¬b(W%uù¨Ò€Lô Cú&«®èšn¡õ„0Ïï¼IrfWVßgdx(ˆ|6DÔÝÛîBõò(^ÉB7{öh>ä¥|‘ÆSâ;d½}¸Œ’Oø$Ï•q´t¦ò÷ºo¹€DVúŠó$Þ)×O‘ÈŽžèŸ * ûÙF\Ë¥+ßÉ3"Ä`º¶¾÷Éבvmޤˆ ÜÏV ¤%?!&ú¯‚ÅžÏ÷È6â‚/0ŽÌO‘aÓ§@\8ÆÞ÷ˆÂÝJH(»—3…×G`²Œ˜Ôfy¾覒§ ¹Kq±\›¬Àb +ÚÀŽEŸ?>Å¿M1´wrHQR˜l#†@@L4‡Èê!È’r™!À)a„økï"ö "¶lÊ ™ ÀEd–Îþ@Z¹ËBQ+‚k§úšˆå¡^‡ˆÅß’ïÈzÈx# 7dØ¡eÛEŒˆúx9ò`E‚ ÓˆåÔZ~ÑËïs8è# È2»ª|ªy]v!Ü1D‚€¸³_ô²wñ‹^f£¨D<×·CÇVT»ÃÖ;Œ¬Øi·¢l%6p½gdp2xï©€¸5£Ô^~ÏÈ D#‡DèÍȺ:ƒX|ÊÄ™œoù"¡¦tµ‡&SXWMųֶ'h?êÍÊkjîÍʫ€‰\­jxh^å]€+m$T¯éÞÜWõÒ&‰@Ùÿ½7é™mI²…æ÷Wœaæà¼oj‚x¢‘@‚Ê B ž _:—¦¨?ŸXÙŽ8…T ”Ê›÷[¹c7¾Ý}»›-[«ú<Ú«°t¶²x}[õõë[wj*pÆE¹%—¡ÝˆO£Y uÄ»¯¨½ak„Ó÷ cSlMXad n_X>sÍl#ÂÏÔžôój µë÷Õ ÑëuQÀ‰êÑÔ{ÅŽ‹¥nî4MÉÒzÃBõôâWŽ-±m²s#Ç2Õò‘'ªIƒª”»A n›.]p ±3ª ¼Qî äº`½ûêâ$ÉÍ)ë´S÷Xâ´•ÜÅ['ˆ7`]úA@Ü»êåš[³k†@}½÷ZCqx î{CNW­D•/öß}ªN߀5²pÿXˆÇîPA]Ká` wZÀW´´¥Üð±0ï;‡lÿZH#Ì0Ž€éÓX6§„äÖq‘j ^Ô:)v¡€“Å9ÊÊŒ¼Î€DHFl ÈMr ŽN›¬§€Âq!+íxK—ÚDH<Š4”“ñ›¥§(ôq]+”’•¡<…7Å«JÒ'+ŽÔ¸W/•BæÂ!K³ù ËAÃÝØ¢ï@<ƒ.ñ¾Zm%õ®±˜ Õ&ê̇ Ê'€x†-ù“Tr­æ±|^…8¼»®Û"òÚ@8êꎷ·µùoAÀ#ÂÑ€"ÕM©o„K©üSC‹ Ȱ|^)µµ í¹aU\€'¤-ÓhŸDÄ—!P"¾²¦H¨8j· îm·“BÉ–.³c§8 ÐUú-~-.àm-²jµn¦  æPËØ­½*Z‹Ê ˜-ŠL §R‹êÑÙIRxÈ02ò^xV8CÓa 'D臭ÕbEt ² b´+cU˜t2ˆ„{žœ'Š 7¡DW 2ÝÊ?3ªCõK­«ï>ô‚Áº -y’*Ó*COâ…Z1Ý~É8w¤ı½ U[Ð@âHOuU® žAN6ÐEr;1ÔÒ«'·žG m’Wj°}à^õã‘ÑFW31ÝÀ.>’S¤·€8h7 è?ÎàÔt‘ÆLQOzm<äø4bžBiªÀž=Fä鋤ÂÛYÔ+§7"æ(‘3 reOryÈqšÔÉ@hc9‹\–o'}׋§Äa²3‘Û„øRJÓC†+n‡¡"qÇŠt¶™ä¤"F2g•«§ÈY#«\µ]i郼(?Vé-äqW+--j"ß Ä¤Šª8Ézfiùõš¡‚ƒL.«fÔ‡_¹51¨*øÖfdö®¾y4¤Œ9‰Ï´y¯Š9Ce͹RQ¿ˆ8 _ýhi„ÈLT=jÔ•dYr1;‰6žTóÛ«RÝ2·•‰q)Ð92€ÊJkxY•RnndÆ}±³vóŠp ¯y€H€.êÚ€H€ÎªÊÈíK /Ü €ؾ’ ™t™$qWoÓy¡ÔDys¸2å¥8ßÃFÀov ŠŠ\ƒ‰Ì¤0`áþ¨þá$´k’Tí­@wlÙ;9#ªmÏšëJxwª„&`É䯥ýX3€¸mzQØ1³‚Î!ÅÈ+ÑS%%«[Lk#\…ïã÷Ò=£cMD:Š!={íwGÍ\†Äüî©?Ýã®! izD—‚Ð#dJºÏvQ̘øÖEC‡ åÈcØõOKexWz°t‰AÜÒ7wô×IZŽ„-CŽå9OÈN]Ë´“´‡~5LOÝ\9WØÉb/JuN h•t:À™ù†Ú¨F¸ÐmÑCp Ú‟ėt“¾hóͦO°É’š¾š“dIÃ)2ˆO@LŒݲ¤á'º”dI#bD²¤{"YÒýA¬+æˆqZšo9ÒZ5Åq¨èˆ?‡Cߨ^6—_Pqõô9$¾ÔÓuõ³Ó¨¥°ÓÝì2Ê\—‘­®Ù#* d]!îyCÁvHÑÎHo’«5g\)K§CìµC/!puÅm¤ ®{Á”n Sq§x­Ï5NÔzf(M¥{Zî^›Ñ¦(/øye p_E’2¾Á¤z…¤~¸úÛã}JB±§Ñ1±$!ìî6õb@*q=Ýš¯wxꛊ@¬xå¯Î– ±4¥ÒœNÖ@¤Ò\ò±§Uš#Ù „’ÚÍf– @R?»% ßí+æºÙN”GGtP+)}ÞBÝ4È–=}Þ°!ÒSþZüšÑq%‘6œØžíÄœ6ŒÛžeÅ îYS”n(X{Ž߂ʵ ]S 'ö¬Þ’‹ªêpèg{T-MìD(ÙŸªÛ׎Ïæ6u)°Ú{ìÌ€´ý.ç¼WH~{>Y ¾Ù‰Üwð+)wj‰Ç¥d_䡨bKrwì¥È TÊk"lÂQóR²©¤¶y]êç-þ¦zöC˜^ üB2Ý]É `¨ª{îXR â¾´dg-ö™ˆÜ#2Üb ^ˆ.)BPä=)¸ßØFPŸ€éîKŒ8¨Ç¯D¨à>k¬yMÚxF\ª,‚,ýrÓ¨b¹ÏÙ?òö%|Áx®X*”è3犵¤Z>“ˆ¸´Ür>¸ØÛ6eôªÏcfmÍ3¬çPö}†#Î'+·KÛB 7,`à/‘õ^újÙ¤ªlï"ráP[¡ö„ éaziõ·‚šßC˜^bÒÀã`$Âá¸rvàhX;{Ž‚û°N˜‰Pj}Ý|LàÜDä‘ÛuíQó1¯<#’½®=#Z>¥òÝðƒˆ§¼éQ?øø=Å%€ AÓh‚O¼ÞXýüÞ>Tè­Åca„Ývçâ~}«öÍ‚†¢Ùcߘs¶Mýá>«‚î¶[«çÇtãZ¾§§¨Î9'f‚]õ …œÅC'èªeé Òà•#¥„ýCiNä)œ¨²«8!ßÇ¢'ï…«Œó{ÝLMZ"œþîS8¡üNOå© Ýj"xÊ›…oÈN„Óè͸ºÍZê¿B¼,O$$¿®%’zꄽ#^"$r¢z`7 Î{³ÌEÈHrñ;âoe"A%BA‹FT˜DzL1‰„ãÕ;rCþ-$–<ä”@¼zÆŸÒVi ðG’_bk\?‘âcL=’ß|C\ÁúŽHÆ“ˆ­bÞò˜ñ·Iv9xGb¦D7¥ÿ aÐõ©üØ.3%¢Lv—>Ã'R}ššr;~;\BJÄâ?ÍCÙ‰ "Mnƒ+§Ö-ñK!°nbEð\Ï~Þ«WBUh­ýuù43Õ‰®«`L–qé'S lCD@ÉK™ `×ÓX„IÓ§¼fÛ©%8&Ôí‡XA‚¨²b6T”q®dkÕá‡HÒW•>éSܳ«/da¤ª1µÒ°ðî“XШŠš! E‰æŠÔ Ië]ˆ‡G9* \Óiòã«­uM©'bVK™®s²]Pj=%L@¢ÌÉŒšÒýÆÏGªK‘ÄÅR]æÜFÍ+ºµÆ\p¹¢~ˆZÇšfÇ‘HÔX¨P+$/w ¬Ê¦Ù“ˆUÙ¢´rEÒlÉ‹ˆ•ÿnwˤî¤ ‰X%óVŸ9tK6Ö@¨{ÑiÛ›N[wñ™õÏqÿŒ -ëp¶ µÿ©Ëž`)eyb€E¥*ʆtÖ²GiÑv!œ%ÙNóÈM¹ËÓ^¬5öU=Õñ¬FåbYß²Ù'+#þоÙFìQb1{ô^×oÔjÍÛ‘Dz¯[ð( Ä-xTù´2k½žc±[Ù»0ýÙâÿÖèOÑêø³×–âÿW‹å²ßÄÿùìƒ!À/lI¯nå9© cñ—¨2sáD…|x ü’î&Ý»mi°2ÇDá¤+ÄÞ'Ûüžk[b½x‰°™(ƒäd k»Ar†ŽÓ`7”­/#…÷uƒ‹v«ì¹½´)œ™‘  6ËXòBÃX±mÉ’ÌûÀ,iécôXÛÉ9c%..Ñ_ ø˜!Í/ćtË‹ó5‡,­m1|âêM*U76@šå’ì˰ÓÂðq_^ ÊñƒÊêÎãé—¼:¨éå>Åh²‹õT6ÕˆðÊ'-…æ–ÒIWp±)S6Š>"ºò •9}å(f³€™©|Œ,1˜ÜS’š@ü¦>Ã^nˆ)÷C &û¡OʰMÓlê&'}‡¦2-]~rËm'×PénGþ Ún‹R –ør‡”¿eÀl®$~[¨ ÍT]5‰‡ Òsád2–ÕÃz´ú˜V‹ aŠÚ!vÃVvLb¥ðº;míˆ1˜<5¼–rX[`0¹ Íf©¶&³&A¶q‘ö[Ú€ˆ'&·å)Wd‰¡ù'í·È$Sð qëVb(r¦èQ—B/4þÄ$*ìIs…Ú) {ÔÞáëaï;ÅÚ¦>¨»ˆ˜Ý†(^¼…A´Vù’ „]¤…LÿtL(“ Þ”ÿÉc†Ð+‡a±Ò­Ê5Ó}mÂnÓz̬U*È|_;¾—–+êoAK×8ñš ‘Z9#ªDãºç»Îž»‘$mì8 {+IÚ´+×’´»¥4.z[Ýäs9#†´-À#K©‡+<¡ ¾ Öþ‰e-5²»ûP»Ptì †•èØ±B1ï©) óÞ‘ù KÒ;§k)5ï÷ gj×­Ž[{:›¢Iwúz_©¦Aj=Gec4Ò ïš]0Í0Pñ{—<³îz·YõUÍOŠÁÑŸqÚf¥úí†û5æíTlWÈЯüÕ»R=ʬ§íåpÝ/M–O »Kßnê¨÷l’ÉŸ;^-¥Á$ ¹}; ß*ñú™>÷WÏ1ƒÊ„­>_Z¥Ê ö¿[ |Ÿs晋\ @‹ €jí¶þ%KÜûf4¹cÆÐª»¥a¦@§#äqÖíC65ÏÓä„àÆ±&ÜüçhàrqTâÄ#Ò“@øöFz†ËçN 6Ï> ÑÂ=OCŽ‘NÏG™zÅûŽþôëÕð—Ä „=i”|Õ“Ž_‡¡ë´DðÂÂR@jž—Jiް7)82TòMÿŒ•§á·ÅM’e– 0@d–’@8 ôé¨:ö34±úc!'¯Äy5¥‚†kòá,ÒÀݵdGÚk4(ÉÓ°7¶ñv)!!H¤ ©yvÙ¼ùð‹©0õ ÁÈ'òƒ.,žÚ·4j¹ûw¤æiØõk¸%=Hì΀ðsõÊþDÆÛµ8`j0ˆß‘:>‘(íæ· V«#¢äDÊM×î)Ÿç)Áˆx`<ÈJ‡ð@þ-J2Lþñõ«¡hkÚì¼!A¡r~¼Ñm €£c³ì!…3ðxêùD"ãø r‡xE}"±ßÁ\SÌXò÷ˆ+VÓ¸ŸHìn€Eœ^x@Z§Êo"‘’2âY+!/ö¯Yœi7Ô¨¤‰å!¤Îž– ¨)HÉcv#­æÅÉ<%ÕLÁ48bølQ9 A×ãqÓ´ ¹ƒé€Eg›)úÞÌcÈ ß=Û/¸ƒ!š du± Å("ò`ñf 9¥+ûö0ypE+?/M|G[¯O•÷¤¸Ði§‹4YòÒ®“j9¬ Ô“„DìdJ9÷èÉÏ]Ô_„©yObOoÌòù!Õ““Ì É_]y ‰§Ó|x~kˆ±îLÓÔ¶¥ë@È2Qì„,ÓÙ¢7MUMñŠn’­I2¯§¢©àjÏÔ2Vì™6BF<ºt„¯}ì|t­ß@ZyÌ6Õyæ1|¥c䣟fLÛÌk*j’u<ºrP{Xª§À5’43ÊEÏ.ÞO>:K¿{¦0‡U¨@Aï¾?iÚƒ¦> ~d'X"}w úpIÆCZ¾ñ¥™©×øjÌ­ù£—˜?æv¥Böû£ôÌüÁŠ ˆ^“[V Äßl½J`@\sàïç›sU]ÆìcÄOÚ»ïBŸ†Å&í:§ûº*×”T8dòœîË×JöåØª“YBrÒ tn\ZlP\‰ëÆ…®5a“é3Š5aÈ YÖ}gT SÆ¢yߦtzÆÌ†Ô·D²88MëJC=²Ójí© `H˜"u-]ÊP\Ö^KeYš6•5ŸD+ |’zb8!?_wÐ%å%HÀ™Á)A広9-²V÷>Å”»z݃§KUo‡²0‰6F‚Â#N`Ýï^üaðˆåS÷ÃàÙJ݃G(u? Ñëë~< ²Ô]“ʤÀ¤2±ñv蜓ÿ$À¯)PWȾM­Š^H„vp ¬d1.m¶ëÚÑKy" b(á P” ]ô±ºf^Y^6u%¡gÉ…ˆ'ç%¡Ÿ¢·m!5Iqz¦Ìå[ú HI*uH‘… þ]]!†0-uÄ£}k7Vg2|¶Tÿ¸iL¬ó±jkUçñkÙŒ¼€íQ‰ž¹¸‡š¿X᳤û;ÒQ¬3óÿ§j’og²'9rgèfNk”Õ™ ©ž¼ˆ;ƒJÆ®…`Ï,=„j†ÍY>m‹o›I¢uÖh,Û”Ô0¬&Ù{º*SR"Ö›k²X©Ú:‚%tµ ¨™šV«#ÖÔ`üñ…(œÖ «#Ò³ä^!î\– «c'u˜áÿ:’It•â™åªú»ŽHᙈ‡ü• ÷Mcp1õ ÁDÂÆÉT“9DÚºŽ¬[”ÃaÄêSž&k£ŠZuD1TQhªÊ†ô×Ãè¢o:9±\–S¬ž)+“m‰¸F©RÙè…„‚dÐÒPž¿öØYƒLÉï*8`óJËËP‘)gÚ³ÈÔz µBóÍ šc,õùAž¦NjÁ¯*Öbþ{U؈ùïV«ª=–h47\B\ Ó¤X{óº ÌrêÌöÈ€ŽÎ6î@JˆÚ½Ó'~ 0­ôy¡oDÎ}2à$ÐK¼…Fw Ų́Z6Óà ë:qéD×.º¶ØÔ,wÕôðX6÷ª¸¢¬Ô4Yö«‘ÍCá;u[Þ3 \ƒ³™èªyD‘‹·Óï$ªr€¸ÈexýÖf¼§Á2€Šœ`V#cÂ…M Š×²ðPmñ¥ˆê# >˜Ú5‰Ëú@µ…;Ù’–—î FSø¡ÆÕ‹kA;AÙ囿.8XPq +YX§¶øÞ œk p1Æì¾ß€J܈øíº–¬Ö¬°ï‰J6ö<…l¼á´ZV‚…l¢í‰z>UÌÕº£vÆ:+@´ŽEåo¹†æ²Ú ?¹UJj ÖÛN!‡® ⎽¶zE u7fóêÁ]ËFAŸꌖ°7…% jµ:(Mü–€eÄõ5«ˆdÙGD„åþŽo*ýùªWt*{ÀÝÈu @<‰mëL׈Káü^C=%§ü|×H›¡&U/*¤Ti­çc\þhP ®sEl­Vî£k^àîw–ßS„ÀPŠ;|ŒgÇã¥LZy¡¤—z Øe÷T .¼ Óñ zü⚢ë=G ~!j¹(7¾m—Z²0ʾ±@ü1VýsMË2ÚÐU!î×6­âνš½ËÉ’nÑÚDI·ØŽ@¢¤[…<µ„ÀëÀô¬ç<›iw:¿ù˜’Eèì%UŠX¦@|ˆrج P7b×ı^@U¤¯†/Jë—±HU&H¨¸ó—G À?Ýæ¢òÈv‰¾Ê>€h,n;C×$@w ù˜á«+ÃÄêM…n@šE”b~ÁK0ìüEÄzƵ¤\@C%åÚüi`[‘Á+Œ‡ØÐ¶:Òw¤h¶ )3 /ð5ö1–è^´—Ð¥¥“kty¢¾#t[¯%øu¡‹AD€Ïß^K”ˆOAã7@Õa@¬d0Êojú|C¼/aÔ¼GSÃ=€„ß‘¡†{qaˆènéEï€×ÝoˆÔM~æÉo4POZ ‰'õäAbRòƱyÓßÔ.:O=Ÿˆ?FDŠiAöBOÄùŒrD zÑrÛ{˜îÖJXÙƒ£1ÅdÚy p±HyCºÖxBªxT+¯N’\p­ˆÌ->ÖÌc¦[#æyõ¼x7¬ç1$Ò…ãš*FYË$µ.XgDJU­ú³ëb}N„tJ¯õ§p¶|ª©Iîä/HîŽãÓì»<¦1ô¸j!BjÝ(áä4¶´}†E'ˆZ7œ¯!fÒn/vÕÉtC¼äÍ +ž!Kn8D„¤¸a["$¼ ¡ˆS)m?Df ¶%eäæ!¤oͦ;²˜§Ëˆp|L4‰°¯Ï™=IÊä4œ@ö~cž!OwžìJd^’Á]IV?Èö÷ØKVÍž$CJæíuת Dæ; —ɔ˴y"ÍdÞ›Çðu®•OΩ 9îxpä3¥½a_Z7œ’ª¤#Ï<†DÝ]óÁUÆ?"tGd‘BÝãÁ±']ºµª)ÕñÆUœ´k"[À]q»M+%·¹MŽ"ÂY)¢£­tåj™½p^0½œÇ|Ør!õ9«^ˆDF8BaûW¤!'}ÅïZ“·ªÆRðÔûø“ãå­ùˆDzÌÚ$±vfŒîüY{é>(CiÖkÎKÀdDÄêX`¶@Øì³ÄôQ%„èÂj#ª¤­ñB«‹$gNÇ–D›áR@„å\sÄ"B{bVÅúRDåÍÊÙ¢5"rTþäi§Ê \w•¢u6KÜo ,Öž¹<)²îê_,½[%B{-û!´½aé±WâÅ¥íËŒx"aðB´ @J*ÎL¾Ï´/Àëoeòf†ˆp*];Ï[uÃ+G*ãx'^±Ú‚ro*ÒAœ;2D÷* ;‘ÿ+ Ö±—î$ëÍwF; ©¯ƒ”‰¨pÚÞ½½$bg…;5£íbœ—îÃO¥ý½rÒC Êé×õƒ;È—÷J§c>êü òPñÀ‘é{›^ÞIyìK£?d ¬ñ)EP"VuÔç*¸WÛcd ¬»y”÷šI´Âû;)‡-ÍÒ7% |WÚþÙ¾ú ¡æ³5£œà‰_Ô°üxÓy¸ÇUì¶¾•䄲nêAÉi)#’¯"ï=®Z¿=«Y¾"nO³|ÅÈǪêÈ7rM@$_Œ›{,å‘v [ôZêtÜ@,媳!åqCu6”«ÌûÑ͹ҕ¤ÚŽMA¬½ ÄOºDŠËOêxùztð¥¹H$Éæl;€,ëeUWCHýmK¶Ö›rð×òK%÷zB /Ô¢–!Äé‘ Ÿ•L;‘j‘&÷ëë?ÂNWê}D<ø,îÿ(FAÿ_La{}¥$Nå)ü©jc¤Ê¬“>«$c[09 áQºRã~Öã±PÏ»ÆÝŽt¸|ˆu(ûc´á?=f9ÿÌ®éo¹0°ZòP¡,ÒmÖÝb@"»æ‰at_7j„Ãv 7ÏnN=úhvøà9hIÊfÕ¬DŽO6 yôÜ`RBU©,‡£ nr[£8Mg¤±!ë (SgÓ¤®pÀŠët«6“C®3TÿÓ#ÉL€ˆ¢[k„™(=k„“ iÆkV8ìz©ÿ>Ñ#›ß,T9‚ÓyŠ•ì!¥ µCX-ëúóÖÄåF6Ðݾ k¦Ž>†Î;KtEÕN¼é< ‰ýlÖh›&$ãþ”¸^´~µZñì!|- "˜•ô}${ü-EÉ´«Ê2«¨kשYb¦è¶}«g '+•:Pû2ÊÎ ‡Ö£MzCç.MáÏ¥‡* \½nÈ©Àž&ýÕkâ%[uåâ/Ù#Î_T+\¥?«¡h·r+þn¥r-ШIøÈõ—.aÒÕãÕKO`ög[z¿ÕëR.oÍúÂ6Õ•(—´Œhþ($Æîüx“^†wîßyœ°·KÍiàu«[Û¯ôß}(Ÿ¦[‘õ‚‡ÁJ;È«XÖÚiy—b'ÂkÊ]Ò‰ð²tÇÒ!|—dßɬŸÏ݉P~†+m€prÚQ F‰ÌÇúJÖ˜Û#îOŒNÐþ’m#uswº?ŠÏ”íãþHíº=ÃÎÙ–‘ÈâÚw[[r‡YwxQ®t€_%¿NY+.—@l¼¼µÄDò·À9bGâ›”WÜQJNv’}Mu†q'_gZÀ3k;ázøŽMâ;^J8#…l3ÅEnl.åíVÖ•Âæ”3Íiáù*oT¦}ˆ…¿Oð*Ž)„@Üo–¬ œ›â„ø*,_ٱî°–'$\á%Ë]$œiM×3 ?V^[¶^ºX\Úùv¥€¢_„?öTdp­N`¹KyàJE°å¥è +´c½4fÉÕ˜ø—‘q`Ìùé„Öî¡â 6¶ngl¿¸ûfQ¬Bl°1æpI³†ð>¢v $ÿvKX éx±! ~·CÛ8 z&û93c¯óö«ec:>+Q#¼Â¾Á¢‚Ý4[ô†YR˜TSkÜg–Û=ßv«¬Þ4t6'ÛšåD$ÄÆêöÞb'ãÎXñ ö¹®Þd² Äî¸æšñ|ÃoãÑÍ­ÑŽ›Âš\+KÒ%¸VZTÀÊœæ 1mØsrCpû˜š¿ÒÆâÐK—€ÝSÍù&ô–B½¡*q\M†ƒç±*J¥ãO œãoT À>ž_ïkÓãòÐ"lrZ‡_ÕÞV½²¹7lOíwÄsIõp€!ªoPP;ëþO‘¹¸hC†..Ûk"þN&qÆNoŽã‚– §(òÞ†hÂÇ:IdrøÈQSþ$™äw*É»¿Ž)ÛêfÖƒø†É÷1ˆýÀõoì!3U¼øŽ(ùñލ„ˆYbÓÄ»A=ÈrÌm`£Ú†Äœí>}ès}"âlÞE§Ò•‚¼ŽŽWÚïÈ ‚{öÖ¸”Ùp.0b¿U˜™1.ªOc_!†?\z ǧ±«ÑÖ¢¥Žœ¼¶ìƒxòÚú`Õõ³ÚºÖˆ?É:áŠÉ”º$öcþX&ÿ{‚ÛÒˆ¬#'¸­"¿¬ÎUå¹Ü"‚­l<‘sàΈy²û4žòŽiÜYœ°ÏПZ†oUgñôv¦Xl©”µ]² Ä3š¤¨xþRõOWVéâéʵÇ@<;YŒ¬¦J>Mz‚œX®«ðFN W¡ó§Lc[®Ž$^“YEÀñ·sø<7¿¯-ê?n|ƒõiŽt%‡ûTl ÿ™'¾îÓçÙï ^W[Í*^\–êæ1¥‡ ‹í"ïBJwó„î±´c¹iWÿ­æ5;n.Ÿ–OìV¬–Ô67¶Ä帋†q&ãŨ€A.â:Cl–XVª<µ<Ç¢¡uÍkÂ6„x­äbÛ:ƒ. „óÄ J1V›¼YI„)ï 9R,Z›ë4b©ËhFz"¿@¼¹­×g¶3'ÖМDgø:`™ÝT ¥?›ÏúÈ@\=å Rëšš§™ÞDz,°òo¦x­sÄÒ³M1z§5ãe^ù·×™.†â Ssu> ¹eauÜŒrTnb\æ–òF§¦Þ6Cä¤Î nô&6é\îöÇ*Ë5uÕ¬éO¸Ò 'ª\çö¤û9–¹ùW|‹'7ŽŠ¨×T0Ä6‘cÕŽ¾®ìë´LöŸ$§≚ÖÀ±i}êô°±åKYÅ£›_Ödu~ˆa åó(ÔùÔbSÍC2`2\—jޱ7â15¬&±q‡Ozëj1&ÇÍ’ÅãøBqfÆPfѼ·º'fÄ8W¸7D “eHç€G“#Te)@ÊÂ(ÄtÛÌéÖŠšMÔZQ·„]xEmOÄY€¸LWûÉŽášÆ3GÍù±¸>?Õ’Œ%ˆ UWbžŒ 5?Ö‰ÈPóSE!'®7–¥é•²ó@Ý%þ<ú3æäåâ¹]"€·d™‚[‡M׃?Dø¢¶­¢gãhÙQçà\7â¿MlÏ’_ Ûˆ_Ã~@ü£®Þ—µÄ@®÷Ç׿‘åBÓ¬P>6°¨;[Ú@¨kö[Ú®¯Û‘yCü“sgÐb_•îÈÂ1TU­_‘AÀÂkklæÄ/ÅZ”kûëy\T´ot¬Ó}ñ(A(˜3õ Ùu Ûˆã,QÖzB¾3‚ÌOu9Îs§Å`8þ€ù‘ñl~}²lÖ'Ö¶Çn,@ëp œõðºV»¯ñ°BãÄ~øadeÜž=?jñó#Ø_OÐh?Þjük_SÏ9ë¸WÜÒ»×U'’q%rü¹n¸®æ»¹P¸Òç«ÁÎd„ãû†Œ:% 7dM[áBá6_"ÛRSíùF}ìÍ”a‘€0gåËðµ¢,¹žn}'4e÷BU‡¸‹»¡ßrm”ôHAá²ô®¼£!r£ˆÙªnÄ|'´! $½ø)¸iÇ-Ú-œ_­’1âT³ k ¦Ü®7¡,}(§á&¬Ã‚aÒƒì^]ï"×VW@œu­ih%‰ Îq¾·n „È•ã›1KÝÂ!ÎUÕË´ ¢f^³¥&:ŠôH}æ—ò£Ôº*!ŒˆjµØ‰mw[³Iòœym%Ìš€è~B‡ùZ ¤Ô;²¼Ô…ªQN dïw§†[Ílb á[RX‚ñËj¡ûÒóö$ÍÙ¤éÿëI\9™Ê¾Wš3N¼v…Á€˜ÉÓ•1v³6‘k©¯Ä:xj8Ò"ù. œPz»b(1[ª["«Æ$ „MÑJ’ïd)Õ™¢m Ìþ®ísm±ùSSL:h~S]„¥ÖbËåIGU¡$7ôñ.<„Z\íaM_ù5¡+?´ ÕåCöÈ´‰a©«”FºÒœ¤|’ßÃè¾zÒ&lkÄ^éØìK•É\®!C”ÖCžòº¹i…ND12(@QEs ª(‘ݨ¸–ª=ÒRà²4#ÁU‘;$ªÜš«Ø$~(­Ózò$°×ÛÒ¾jP6ªï ±MIûCAË,3tZ˜~€ÄÃñÛ“^(%RJs™u0µnÁ:R¤¾t·a&4Àì2>åâÄ ƒ)§(‡XRLq1 TØ!‚çÙ¡¢aø¬\WO·pb€¶á© j_Û8ÉJºšZF8D€¥› nÃR.Blž¯—E¯f¸ÖÑë‘o»"1Pâ­%2ÅMµ.†›‰P+ØÁD¨Á5#˜t]¦ýˆÐ](TçÁ,£d"7vü¹“â¶4)Ì“·eɹ“t:íçSIïÊ›b{~*ù³@¯À].`µ ºrYWfA&Bù¨v( RÎ*M’®kÄ[)jHa±{èëªqè.#Í÷·“²(ã£g¢#§Ù a¯Øº=ʉa!?4xw‰®³YQ ¹EwQó.!ÉØ“‰ÉN 9#œnvÈ!щÜ#/Î("ý¡Ú¢ØñøØŠD·‡c,¡ÇÈ‹ƒ]ÊåBÚ}a—Ø'f|ÉMSÒr&•ƒìDúTWvìS¢“œ"é½PÏÀê„Ç•Ö5E8k"ì&'ôpÒ:¡š.n·ºgO„j†g%_W‘¦ýÎcÇMØqŽQ‚ 5Rs»Î°²èØñWà–XñœiuÔë­3­Ž)FP—CõÐ- :6êp:;ÐPfu—8Ê(Ù‰ð»šz® Vs>¼ûgå¼'J%:@â¡XÂ%Ùx&±Ó 6ÏÄB žN†à‚Sá75kŒ!]Û™e©K(©§ú¸5Y攜NƒÆKêEu=#‰º@Ý]Ï—ÖooÒ¼×öæ@L$³ A/9zM‹ï%G/š¿K(¸%užݵ©í*ÍÐÓ*[§†‹4ŠC?dâ+55Hר$_s}õ½ƒþñÎc¨ñ[#"ƒ¢Ú¥£l"㕻磵|%™Àij#Åf?.NUgÏv—?*?{Ä\E½€Ü¬s 4ºtŒP= ^¯hÔöR”4JS»¿]}]{ºÍa[´Pf¶h¡~ÀÊŒ+Ul×ù}{T±H‹*¥µ‰P¢^Ëì·Ht÷üÛ¦·<­ GzïQ¨T¤ÀdŸ@(ÝGTˆ•Ö‡‚8kT0§Cd¼~T­@ˆÜ%¾¥…2¹+Ï\êñrW }û!p>`÷ ú®êe¶”Óã¼*©M-õ¬²Þº«¸09CÝ%\JñS³=N+3\躻v®HXH²ÄàWÖ÷P»<ãYÄÔ["¬ÈzŒ@Æâ:4ei)W¿²Þ³3$í](AÊÞ»J]9pŠå»¸XùìG>Ÿb‘=üç‰ðeÚÞ{>}G Ÿ;Šé Í?{ða³ˆ 'Yd±±Ä=#¬Å‚Ù@<‚-0‚ÜÊ .Ϋå#ØÌbõlU`_Þšùò X+›ÆŽ+ë‹'ÌC%@eòu²×ðþoÞØK”|¹ Ã"î?ì%jÞ˜I´(2,ZéÑ^"yId52¶ù?hRQó,´’Ø3*9%,Ck‹<„ƒ{ïèŸU%pÈpnµ·DXhá*f4›8%®TÉH€5GTIÚ}â´¸.éƒéÇÈÓÐ"ø‘BpgÆÈµÆKº‰ag v£ÜñÙQŒšÈS)ŠÜM¢þRE‡p@‰úK9çöûÔ_RcZ«ŠG¤·Ê’EËUËë®ãoÈÍópÞSh"+&gIzÐ!f¬O$çüDnœ»¼ÈmïÈ6Ù’j‰´,î ¤¿]KÈÈ:ÒyŠ;iöo:ê$¥äÿ¨˜ 1sq¾Z5aª¹'òëA2Ö‡4~õ¯‘üÕ۹ɔ8Ó;ó…¾OWmýêo>ȰâÕ?þ!KÆú~P ï™þv‘·ƒ–7{ÏA¼ä†zމ¦|9¶b}Ž äí ¨*x äý —¾dä9¨•u<|P"og »ê[:Q7¡‘]{?æAò 0Q~z<5PlÛ<èAò ˆIQÿ!zç ce˜ç Dò dÂØpyЃäAzz›MÙTŠ–½!yD´ØºyЃÜߺIŽrXð”ŒrïÿSœ,dÍ|AÛ·C¤oûöÿ#–ôšZÞ1òq”—³oGå÷í¨ÇD+Žú$¤ù(¥‹ÞŠÒÛQSž oGÍpQx;ʵToGeuÕÛQPÐz-"ߎ2⣔‘¢Í ¦ÛsXJãí(ÈE–ý~”‘£˜5hïGù<ê<­ê£Î{«ê(,söz?ÊÈÇQPñYçý(#GQüüý Ç Ôu~Ü»‘Ï£ÎÓ¦>êüë6Ó®Ý÷£Œà¨òãŸþø¿ÿ¨åþãÿùŸÿüñïÿöÇòß××j ¡½×WôoÿñªC~“ðºÖlë5Rÿöçÿã_þñïEòaßö—ÿ÷~ü7ÿá×ßÿú?ýí¿þÇÿ;ü§÷µ€ìÌTýè?û_þüßþå?ý‡ÿÃ_+cñó/ÿø³¼ȯmB<ûµÚüÿ¹Ôù׎åÆéùûŸÿêªH3\ìþ<¯ùßþŸýÛÿîgÄné~<"}®æ,ëjõÝþçÿ¿þÃ?ÿõßáNïù—?ÿþüË?üø÷¼âñ·?þ»?^ ¢×dŠWó?ÿùŸbòèœp×÷._zQ¯%­±ÓôoJn ˜3Ü)¿ùRü#óò(éö°\àáî^ˆ ˜ï'÷È„y="f Ê>ò7$ÓÔé^ó1Í6ìÌ1±1à×çZ¾Uô>Æ „XÒoa\1õ±q&bþwcÌRˆ‰0¥²1`GļòFp"¦fJ…ˆé›é"íRü+“>%\HÄTQ‰ÑŸ”Cí31Ù´2åDÄÕÊ(œŸ5kÌ}­t @z*8úB¦ÝVrgˆÔÔd“÷(ç V$€“Š“G½¯ŸPœ<ê|ý¤â$ó‰D¦¥,è'Òý«©ïa³ âÚˆÂ(×O”¦ïÁÿamÅb¾)‚,<: _k½JÄ“ˆUŽ%¼A¤¦èáyRúR‰ˆ¢|¬îYB–…HYYEÄ’ÚÑ ¢¢Ì` QQÖêvHãXðSÈöyªþ´ ·”S‰ŒTWe¯îÓ}o) Dnué>R1–lr"¡Ë"D"¡;9“ô‘‚±C3I)Û5“ô¤£ÞOï2\>P6¨w^=–«%â— ¸>ÅQÄØxž .åˆLC:$MR¨»ÔµZèÃnæš;T¬»=§ 7r)ú{ø7$ìq{*mÓ¡–d%q%{ˆlÿªüô!–W,7ບ {‰\tÙî3MtQ¯»(F¹”#b}Ý5Õ[Táä¼C¥D¬¦«ôMÎÔeFûµ02B=ôéB,•»Šf—¹œ%Ñ"–ŠŸWSz ÿRÔoÏ.D1·b¥ÔµÇT›)u½ü“gŹk KꚈUÀg÷3¤ ø$[’ˆûÛl¬-lãQFÏÁJØ_’âÖ‰C¤õùK€ï8n›­5˜ âÖÛm6hPÛè©%ÿСæq7†f›²Ð6ÐiÃ#ˆN$Th$ðÛñ¨€ê3§ øÐJZfîÃìFÍ0xÈ ji±ñ3‰wAÄÏ$±‡!5†¾|â³Tg1?zçÒ´Cíø’1‰LÃ"6"uÂC¢„Òü <úÝR·'âÉ£1“Öç#×ݼNª#º±BþsÔ>?Š5ÊJ()r…XÕ½yèÕ ‡¯æ¡'Õ."C¯üÑÕn]ï¦J;‘nz5Á€¡v "²[Õ X톋°Ÿ@½šíaëz´$SÀÔçìnWˆ9B´< †LRü+¦vûLåèU½”*Á‚^ ›K²^D<õÍ'ÒÔ\eƘ©M]9Å¡ŸÃ—ZÀPÝá'¹]u‰"GĦ}è³ôtÆ8ú•n뉲ýX-¬'ÊÒ7®„_ö‚ù(ì¹ôXÁå‚(Ñ6â†/¡%@š‘°÷ྫCÎÑ#ÙþP ¢C21a"v¶¹|µ}ܶU—UDlqvi:ÝǵT#J†‘0L¡ w‡le¦hÈ ÃJ­õq#yªiÄþ-╱™Œ¨§}Üžn(¬ù Ž) ½t mú<\Ù±÷Ž\_:ä9mu®¯®G8dÉôqãK/#vo9ÂÐ}lj¶ßÌÑÒmœø2ÚÙ¦Cpôq¿ÁŠH¸ßpÍÓéÔè33VM$ p˜îãÌ4À᪃ˆÝ‚@mÆcžÈDS?ÍH¸1cÕ‡¸¿Ò ˆˆ}Ѥ4ÑÇ ¯cè¹#¶Ú”®é”lU#C©B€Ò°<P~æß|›'ýÃDéІµýÛæmuz~&RØàJ¤ôÕY;\‰¸N¤ÀKítÛäâ±Iábý}§Ýi ¤^!¶ÚYLkö‘òÐëc'Ý1µiMˆí—¶ZÐßõH…! á9Åõéíõo†ÞïN“±¥½”~ãÚZ;B 8®ÝÔ—V쾦4-úÊØ/9p±‘WÔäÛ¥«Óo·Â'Ô;øä+æ[È/bÞ†¤qØœwB lÎH¢êPFö8Cþ@^+32"îµsªo­Xi´k‚¦$G«ØhF Of6T¢MÉ"Âñ8º}å;¸jŽ:îŒå)Íè« ‰!2@)N3ò~PqNw,—ûnç(þÔ³tHßBü g«sÏšÏ@Õ‡®_<ÃR·7Ÿ_þÞÇÉkiSÜÇ·˜j‹¶@Øcyb§—îgÄVHá1¡é‹ÝÇþˆªl ì;#ܯ¹œBÜOYBœK€£ x×:ÎÛ§G! ý7ðýäµ´I)ôúžië= ¹“Ä“v#lìÑ{ÝT œ{”I!нßÀÖ’3#“¿àû¤œ@°aÖÆa ±ƒTë©èTµ`‹o¶¤ØKƒ«×b~tÞÛÀ"¦·óªüí ·?$-´M†tVzo±¤ ϵe@8f›·‹Üó4ñÂi¹…,ø¤K!Ú½r ëë·!Aªbµ{åÖ®¤×Ðpã¶mQWL*•&¢]0Ž€«5¾tpÁ=Ç©âº÷µã@°šE¶|å¯8JjVVåvï5¶qCî×DüF™4!YHIœ¹œx£{i$•Í¬Ø (÷+ÅK E |¥%Äð•oÉpø!·á;ÖçÕ»SñÚkHÓ«£øÀ#tk ‡…SÙb»Å0DRòGˆs´»³ÙµîCµDÏ_aX·k ,UZ–Ó҇ϴImïí†PôØSF›k9‰’ºÁÜ€²‘ÞÁ.¥%#…hêQ¢k+Õ’•¿BDŠe-ù+|ˆÛ 1F[pÃÇær0)£‚¦O¤ÇT@ˆâÓÛˆHð"$vAC LuA7…¾¤ÎOädÿS¼®¥îÀƒÄ²éAfö?–ÿö¶Gö¿@zö¿‰qÎìgG¥U½ŸHU@"åç'°n6F ÇA–±ï¬ôÍ#Õ$ÿøGùñ_}(_ʯ/åË@ù2P„|(_Ê—òe |(_Ê—òe ù2P¾ ”/åË@ù2P¾ ”/ÅÈ—BäË@ù2P¾ ”/åË@ù2P¾ ”/åË@ù2P¾ ”/åË@ù7(ÿ+f^[-ýw½æ&–Îð~‹¥‹’˜5Ó/±p¯+¨O±Ü‡£]î,°G€/ÞÉÁPr;Â~1sÔxC·¿GÍÌEº÷Kuæ"Ý»,øîÜe±)f.ÒÃ`Dj“‡ä"Ý;EH¿Å]ÛKXHz<ž¦î>Kîл^ðÈÌ¡V6‘’;bLö´ö@£œN@Ø“ÇÊ``; T}mëiïzád0Fl»0¨#Ü(‡¼Z{9ŽЍbè¡‘× )ñ]qƒ²|OÑŸ Pè¤ö“‘ÆF_ÈŽáé˜Lí+#W7ÜgÌ­²'R2FÄ¹Eo¦çÂýV½™ž+÷«\Iíaº>äûB‚´;ä F$ƒcøÂÖv3ìÖõÚÉw½‡´Nâ8´àcFøH±#­ØÔ¯@/5â;tX[Xd ù»¿ð§òŸ%â‰KAdl“w\byvµL®õ:6€ £þöüë¸j­ùuÒ&ó…Ä8›´¬}ë-„‹›«I&pÜ·>™‰B«è2"[V˜¬µghVâ/ ¼‰–æ½´ø“ߥ±rlø“É–24 K,ƒ5g[–¤ ‘0…øGúdײ#U¥mþ YAG€À:€™ÉmK„=‘-àg Œ ²@ä÷WzFÑ•t¨ ú¬@¶‘Hâ5^©ÑTµhª+±y ëuærÃÂ0ò&@œ>¯Tü!™k¬út•ö:Sñ•ñ’̆š¹k'²×xñ ê“SnÏÜÔâ"HK¯¯ ¦ü‘¿* +ôµrƒ¿„I½ìŸFÚµì ×’2;„´Dpˈ– ÛÙ YÑ¡™* :-zKÉÜÒË…ÇŒÈb: Ä 3EñBÂKl*´úB’úÓ»ÞÃn™}Wò¢`Å69üU¤3kÇêH¤ù•/ëþL384²3«È€'Ïù›„Q“Jæ7Xfp„ÔØ%×ô…LK(kÄ»3oH(4Û•"ÐØè+H<Š¥ð¸šß 6‰þ‘²y´”88W îˆÊaŠø5D”[,ó$wDôå2í,4e5örÕ‘’„“ÃÓ¬xl•#”ùðuÄ®„-®Ê|;"Ä”ù0v”ñbâ ²%/$ÜlA¾Á&½dÖ̉eÕ»ÖŒU$x>fivâe±q*åCÄŒ“ŒÊx(;žrFçÕ á{3i5Ö¡?½’ò4¸³-Ï/™iž„Ešñ²è=QN®ŒLׯ®çáäÜ®2’¢¸¦fúQ“É%f‡E†F‘ÁgއìèÒ™œrÔ¤´-"™ÒßE—ÎtÍÜž)úN†"¥ç„cÎ]‰Δñ ˜É°ë=‚'SVÄ/$lΧ܈‰¸m4…½\Ul_LED[%&@|Èá»ë5ú(¶L*3ͱôòF…\Sˆs݇Ô^ZØÙ‚d‰ýsž^?Ò-Ì®@Í\Z²7ÙûZĸÁùä”Þv¼ªCg¦²b‚<ÊFñGç(RÚŒÖ2µ´ðž2~!¶žJ÷ð%à…äF´"þZˆZðBZÌ'¢ñËEá…DøŒÝv…ø–áRÅóD@ˆÚ+¶»à³Öð@ƒ˜ß†š¤E³ŒKuõW‘ø…‚P€¶À‚ÃÔé¢I±f–|…_HW±!áºi1’Ù-Ыyâ™|kÑ•€ßZ[€R³¢©,Ÿ8 ˜ÊÒJ£>¤ñ­wScâ$œgnî• šëÌái·DÐé° qÙD->s°Ak×™Ã} ÔwŽO,¨h¬_ ¸¯/nu|']—§:F"n –:pïß\…©>„Ÿ­ÓK7èfY‘•…—£ q¥Þ&)ƒˆë¢T Ð`¸éïÅ&¹ƒˆ§²}ôçÌBÑ3„ô,&å9‘¼[,xh÷„µîR›ˆËV›^œ"…Dº^\f¸—ÜD‰ÿ 5³íŠ¥cà!QhË‘(´eq;ß“\樂–TZeÅIÄß²[5ÄO¤‡P]ÜDùswOk1Ϫ‡ˆ«ÊLYO–6«‰c™±èGIÀ_Ä»5Ð @]µ4‘¨½>±+ÿåïØn&ûUÂMÀ…à14²<_î‰D\ž_ºo8‹ñ Ú‰TŸwê†Uy_–îv_¿&®ónwVÞ—£÷¿# ‹‚ø>…dÍ|[¢Ð¾ªCì§Ò¾©·î§ÒÞgu™}íº×uöu¨·î¨³¯SÝaGþ×BD¦¯ƒ‡µJ=\; ïUõ)Ä ;T"Í $¼q)~kšw–â7&/|‡¯%:烵ùð|¯F¬¡àÇÎjý¶4`ç"Ã'¶ª—¸¢aŸxXs¢h.O¢4'êøDšf°ZkWÑD³UÂNDCû~"SߟZ[õÿÈÖ¤‹iÊêGã7äj¤îXïá‘ð†pÜà¡ð ïÀíÀÐLøÓ]+,“÷Xú„½!Û}ëAŽúÖ'ð.{òe˜|&@¾ “/ÃÄÈ—aòÇ—aòe˜|&_†É—aòe˜|&_†É—aòãË0!òe˜|&_†É—aòe˜|&_†É—aòe˜|&_†‰/ÃäË0ù2L¾ “/ÃäË0ù2L¾ “5iE>#ÇäA¶ÎDHص{C¬Œ$ÄJ0œ¬ß7 ‘1…ðëò†X«‰ Û›šñì"à‡ômòÚ-4pÚ'âÙš“ÜÒ3BÚ‚FPÇÇ;ÇÆ­¼u|ÊýDòò#âó-„²>Þûaþ:oD„šB^¤˜ÔjZ³<¦ëCF„Ùþ+½XTEòŽ)wi…$ws-IF„Ä‚Xˆ6DxÞó3ÿ>¼‹6A€ÂN»'Ò¤ý´F •r(FH•†Tt¤P•šÝˆHUjþÌKTjeÏ U©­J‰IãÒÈRÆŠ3N‰gåßÈöØÆ¡ª ¶|=bå®é‡RŽê^qÃR‚X»P+®>cAÊN]#T׫ÏXS]7ô>¨–mŸY[0(›Å=wi‹!Õ™ÖL‹{–Ä1dÕâž•”‡ôZܳl¦ ÏÖ x#"‰¬*ä®@(1×VÞ²¶‘ІÛybIÌlÔ*Ç#²¥K7VãJ».Ó,o×Z |5½çôR$Ú×»–ßDØ º3¿D((‡DjÔ“ë+»-íÔ웺ÙÆPéÏÝÖfZ]ÎF(Ú!C"U¢‚Å?’¼33;pAŠÐ3C;Kì13´#Ñ·á¬JC¨Kb‰óç €Cïõ -n…Ŷá°îmˆ†)¥ÆÒ¡üÞðrŠÇÈ´Ù6ûk çþÖ¦úÿtêŠ{ûÌ1Ò¦”ôfŽ‘&ÒD*ýúš"˜@üúšØL¶<ù+6Œnü+ñ¤R3“5FÁÑ»p’YE‰V´ô¡>§çõÆüj—h¼ÊD.+Ã᣿öpþ5Ån»‚ÎF†E=ø´£i4Â}íèÊMD¼É"u ¹z[U¼À.­~#ÒR5‰‡_¨«6Äô8S†+êj"¦5ZGÌCu«_샺†¬ìŒV—ß %cýu’ß eeG „OÉ6"’•½?}È”T"Dçf NŧýÂW~rR–ß õsÇ ¤‡Æ®ŸJÞ!©ÃKí~füɹ^Ñb"ŠVBðwÝ@ØÏŽyI»êw?‚}P®3.EbýY4±^ˆ±—+U¦]ž@ŠB°Gî°;Þ}$•½Ê€T5áí4‘ÊŸ¬è‹e[tyG_,ŠCðÙ±,ßï‰Yª,ßï'@ذJ\ÚOPĨþô˜Ð礴˜µ›£²u<Â%Héyæ.õî2â%ÅÆ‡cŒÐØÁ2Û­ˆƒø(q©ïjÝ A@Ê‹—ã5c‘ÃD¿=©—buñ¢]©‹W³Vâ uqí8L +TÉ(`“lù`„¼V Uâç ü2âw„hÛ „>5b¶—{õ7uí]Šýj/¹öH»’t_¾ÝA+’V8Ð¥ßüˆ² òJ2B§‚¾ Êu)ÐÏ¥Î1g ÐWÂöeB÷ˆ6%`xm½ äÁ²ˆ"úz„cㄦ@z¢Ø:¢^Ç‚ýÚãáÛÅ÷^£à ­ê ¤qsÙ¯n®³l_A†«º¬ j"‹ï„B‰¯³Gz ˆì V>—rý°JˆçêÒßïáæÂâ±Dv†+CODN%z}ÿƈØ%9M”á‘)ÛÅ7׸´–П'Œ&vü-Ÿ‰áH÷ò[öÚUs_\…#Ë>š¾6 „sFO¤Û]ÃÝm]û8.€Æ7¯uÉ0ÒuñY¢ñ¶MÒêãb·Š‹' ñ*ºIÇ–5üg \|Mqƒ3rÌ׃0'Qæ®+Ûþxw‰v3’Jy—af”ï²#Lx§¨úÛØF®~D´Ïa…B"ì²3ÌLï"͈f/~JðÄñþW¨t_{’*o¤©%V(aK¬H ÝU寱†C\×.…ä”@äÃ3ãuÊb†ˆ[T3ôÏñµd1C×…q®BòDî D><'ºŽL÷%÷&õsZ6ùk€5zâgÛFOáx‡§ª @ꇅÔ¶y997Ø‘šÞ#J¤æýÈmœ]¿÷aW—Ó¢¦>vº;óJþc¦uG‘çÇ Nãµ/âH>ûÕ4FDáÏÛU>4N8XÞ¾õ’“)„@P(.ön?ÞŒÅn÷,sNt¹.Nˆ´ž­l”ÄMÑE778e@ø˜7\SnozÎë$ãíÒ],Ö.x³[ȺÈò-œXoÓ¢!‹n;¾vN2MÁ¨Ç'îê›H/9ŸeùÚ+®mãD ^ÈW†FvþìÉX†ˆûd“ðcˆ¤ñiͦy7§Ù0êæ\ÔŠLu±lºÐ,:MýnU) 7ŒZ€–𸺨À üŒR^õœuÉ¥6} ooWw'Üé4r¡or™13rF`°-uš+ÞZmÒÕ@ê”I£›«?T”¢Ý".Ðcÿx‹MåïlDucò,[–¸Yâs‹¨ÍiOy‹¶Lt°¼È4±m°„øí•nWаϼà×w‹Í[d†ŽkK„íU{´`!‘ ¤W5à±sâT®ƒˆèì3K¢ŽÌdˆè“zî¶+hXлì œâs§]AÃ^çÜaWа?"WPÐid‹ŒÉ o…Y£†åØ#‘¬X29 × „¯%M_Ϲz--ªÎ9i0«yNXïFæ9ÒæŸ-Èœ@tæ7x¦Ïn·çhtÎ,y;GÅ1@ÜÄGLåÇifq’i(^F _=Ö‘¯i ›‰³U‚Döìãó† ö­lö³Iõ~l‰Ï^öÈ üÙᾞ3G®14EÖ¨:&°=VÊÇ)8ðšáæ×´q>þ‘=LÓüùØËbúÕRèôË\uiD}Ö–‰s¾ñYZ·>×g© aöà54ž{ð‹Î ‹ø˜æÈ#¾ûµ,[ćÌýYaïmÓYáô§3 °Ëø±ùàìQ"„㫇/ÑÁ“×ÙÜwÈ7ÝÇÈ`åq[?SKÉÙ#¿uf·yìÆÎ´upÚÁ©ü•­”δ£nzѹÃñø‘;LXkŒ#w"~¬ñ±©;céMŽòMŽÈŸ¡åßÌ"Ü3ì?JôÅ¡ðø±<ÃîðYÌ{†ö®@<) UÕÍ_) å qÛtñË€èÃuºv¡¤3«µºr 3»ÃÚ¯ˆ ÒŽhžD¶Ï<40GðN÷øN^QkDUÝé6Vöˆjng?sÙ¸_4¹`T¨g— ÝN“{¸Ón ;æqc€k)@7ÓìT=Âíâ4í@¸.þÊ€¸ç4;ÂpH>­j ;",:ø£°,;õê;1‚6 ¤l!j,[Ôð©2#X7k Qºuêô»‹Í÷©Ãï.ìsމD<—T›P ¬(^À,Ñ^ZB|Þ¢y#ëáKc€xÜÙà.jbTèànfX•œ¢-Ï~Ei±™µ÷§0 À^»®ñ›*Š–ƒ‹îI±×û âÕqQ €úõvõÝH•½¿âh’ä.Àݾc%¨óâYƒ¶ïÔ톼À¾ÓiÑbßèÙ½¢Û Úv‘뻡•~ª>sñý*%V·H[tš£ + Ÿí@µê4G¦Ð@´ØgªKÌX4m-’·{ì îÝ>Ýw%Uû4uÙ+­}zb~Ä)²ëž3xd½pÛ ‰z*ûÑÖhÛCˆµº¸«f‰ˆTáê[T¨—lñÎTØ[¾Š@ªoG CV#ø~dšÄ€rvtƒKé7Zèê—6sEgs½7ÝŸ Éè˺a·>e¥€¸‹ºêžˆžÁþ@Üi-/ðŽT}ˆf°4·åÞ©HÀÔZv[\#ÿܨÛSþž1Ä{7dhj »±=c,¼#êYÒ4¾!Õ=+ü–ö”gú‚@FýDŽzÖoÀ›€É—\ò%—|É%_rÉ—\ò%—|É%_r‰‘/¹$/¹äK.ù’K¾ä’/¹äK.ù’K¾ä’/¹„À—\ò%—|É%_rÉ—\ò%—|É%_rÉ—\ò%—üø’K¾ä’/¹äK.ù’K¾äi—`oܼ¦µ×Ð?ÿùŸÙ“ËÉãOÝÁ}ß%µœn˜Ôàñ¥òö ë§µÆÜ ôH¿!÷~"çg'^ÌžˆLWˆDG©Ö{&É¢›³Ûê’•¹-»;ÓoHȞ܈Ä|Ì(‘>b’¬þ•¶'˜H‹3tw× Í¼Mi2VP:¡1[JjoÒ{˜ú­×ôqï{&èãÞ7C/I¤ "ñÍ+nöœou÷ËùøZ1íæì{­˜vsö½Œìñ\‹ ´… ûó1åßñ‰æLø&Tˆò¸BÜeo( æLv¯[=æ!|û)Å•¢ˆ@¨vC£+JÝÜΗPÜŽ7œ ƒa‚…åŠO,Ë =–ˆ´XöèOïG¹B¼¢-ÛM|­Äjj~èCbÅÕ·o²k±Šc¨JzáFÄë4®íxãP›tÑKqÐKDŠC—Ë´ª4(¯{ë”v)±î =úÐÅ QÑ[($î4_›œˆGnë`ŠáÊÌà—VÏm¿KtžÐ•Gý“eAýÝ?¡¥ÞW>‚‚£@ÜÆNPí+]Íd=¼ßçà·ªÇ*¹ê÷XDagKwäT†gð‹Àp)pnWT'"Î5ÃÁH|È‘Ñ ;ç…–Ú/˜£D¶=œì•ù}œjs~ ÈPŽ~ŒÈÀ_D Ý*z{¯ ¿dvëqú7×Ó-y3.XÆrôÈ%²3õgš£HÈñu¤0Õ©0N&ÿD/…ãÅgòÈö=+Aø8g9F"ù'ÒA™I‹¨WŸ½nˆ@(j>#¦‹æàlf±tÿ3€ð q{5ótA:•™\¨j—çǯ'- Ã’HÔNû“¬LÔN_+3m¥J$j­Ý’EæéÛá{;¾PX® ]íÊâÞÞäMäÉas>^É£À¢j95N+æìH—ÃGÆÓ»}oz4N÷Ú1Ýgš—óM8Ô\Q«hbã´¬D߉øÍô•Î73y²¾‰3¨ò¾‰÷Jž.;fVtù õ6Y ´êÙÐÛc?ÄŠ0 »F0"Ø»v5!Éýítùðt±íã€ÜÝ£%†H«@Ì…ÊíµÈÜ™nCÃàà}”´Šá d'­B”¬7%pLØÝöŽìýØên;©YCþk@ÌŽ’ =â~.A‚ÓK b×0‚šD˜Ey¼Î*ë%Ü¡çÅ —PpdÆT îÈ+|æîô)ñ¤ûp>ëärlÿhúâ‘»Üaœœ¹L=âÇF`‰wI. ºrXE€æÄü¹Ia:Zà!²œt©.àaKñ{r°ya ÂwÌ-lVÑg•ˆmŠnHŸƒÏÅ÷‚HÈ d ð7duõPeÚð“v“¥¸ì[tÃX‰„³#dÒàÓ6aǹ7æ»õÿ±÷.»²$YvØü~Å™©jÐE{?zBˆ5$dåD8 ÈbCÄMI¤z ÏWìõØî'›"ô‰Be¬ôãáánfnf{=À§²AòaP „ðå¹qz¾¤ƒÑó]Õ67A¢hñÞ–éyˆ¶P{A²^Þ[ÁšÅ;å k¶x%ȳÑ/݉˜´wICŸº¬ˆˆˆÃ»fŒô› ×ÐDc0q9ñªM$.§Öuᑵ½¨“w1‚Ð_Îi} '‘xyGn¡æ$›…È6l‰œõÎ?„où|G$^ò̘¢(„jƒúðÅ*­ÿøp)ùŒ3†E:‡®õ‹¹j,§*,ÒkBXæo†Eê’ “z¦°H¯åàú_‰ìD.£+õ7DŸpKDÌwþ¥É½‘‘i*ïdDc;b³‘µ©Ñõ,^]Ïb+írÛòq¼T"Ôs$‚|ßîÌ!$E\"j$´!.Ô? É—ˆI˜F5›&œ:×,k†—î$bÇtò 8 ²6îx† ‚nÕ½î@‚G%bB7Yà‘¶:aŽ«K—ÁHm5©¾gŽ«IõœÕÑ“U?”ã:̪WRæÃ?¾“=o8B%4õáÉ¿)òO«CH"@€mòìc,[óê—òjoòê7Ûþ,É«ß|2d…9|2³y`ºÔ­Ö‡m|Ézˆ¨^‘Ë.Ù²‘æ«sì­*Ýà˜e gaÁ¸Îe½9xH¶ÄÉÁÇ:­hã©ã%$ v¡"ȸ¦lñ¯«¤Î„{/ŽCÀ åÔH–¦ ån• ™{H^®)âÀ@¾†åUEQÂQù’:Mp=ún”V‡ –u<1C¸ê:yâÉF¹R@\Àm@ õLµ †Û]-h.‹C×VwLLÝšÌÁÝÙÝj%†¸#8[j¥²Ù®*œànúöÖ»ST\Ð퀜Tó v}§€…œóÈ÷à<¾ž’·‡¾õÔ¼=$ìU³{ uŸæÛ#9R$–[rUØ0Îð¨…a¶g¦ŠûY‘—®QÉ®¯gùQÔÊ›Á±î¥ŸŠ$vIè*·›3­ª+ôý›ÚÊÚØznŠ+kc(î­n¨µqÈb[JÀpb–Tà=wS~,™XÕ·”’¡ßÜ•ß5˜»~³·‘ùA¶e›UoôË °T­UF³ •u)ÖDJêÖJµ&²r' éšë¤.œ}LÄä¥[ Z'éì.RÞAÙ^RÙ^$ó%É’ë5ê¿„l³¥••Ưqò¡Œ7_+©µ”2°Õ’:@Î7Z­nîÒ¶šbËŠŠ›*ċ󤨲r;£Õ[VŒÙ$¥aLbù +[%Ëê-¶¦‘¿ÊÓ^üÉI}ÞÅãm!š:F¢C´V#û/ó;pRMÝÙaæÍÖ'd·¾pë#_±­š7ÂHsë3ÐSLmd¸Ç'b»ˆrrôB¸òý§®'¿(¿(¿(¿(¿(¿(¿(¿(¿(¿(¿(¿(¿(¿(¿(¿(¿(¿(?~1P~1P~1P¾~1P~1P~1P~üb |ýb üøÅ@ùúÅ@ùÅ@ùÅ@ùÅ@ùÅ@ùÅ@ùÅ@yÿÿ(°Fùú‡ÿéG-_ñþïßýþõ/~ûñÏþõçåô3†Ïcþí?ü¨<ä+&§>3ñêš_¿ýþãO}ö?ÿö¬p¢üôëßþ§ªûÏÿÍs¤Ï£Ëÿðïüé3®âØÏð|k÷þo'þË¿üíÇ¿z¶4Â÷Ó¿~ç|†J ÿú£NWõ¸÷9jmÛVu–¿>&1[Á³‰æ|Á½ÿóôˆþ˜^H•½æ51dpnù©ß‘~Ó^S¥½¹ÎEmò)ã;ÂÐ d-0[æ 0frB$¤´ïH§_\ñÅVžÊ”[cQò“,èÊ1뮇/KÑM¯D âJÐ`‘/¶Œ@X¢ÚÍþª®ðì&SÇ"ôÍw..v³½ªN·Â9fñ‚mÓÄ’ŸÉl ?áblk€ŸECÖáëæñWÝ ¿˜š|€£çïM¡@`tT¶VH»Êè±x¹5嘜"™ôñLs×@ŠÎ̽’­¹V :¤ÿESôÝ^Me™¬S›¿md·æ¦ˆó£9î,Zn|Vò~Œâ¾€¯—9íæÂ-û­É~ b-•¥_4Íl*“¦Pi{H# U”„ˆ³TdlÇ¥ YA­òh¸<ß±š“ïÕ1…MÚž¼€ÙוϾÀÑïñ.âºrZý UÇȲ¸ŸPWƒYW=Ô•÷̪N„!"„;ªëº‡Úxx]îRôW•¶bžòõr/^Üê˜EÙŸeá¥ç"ï£AÇî’|OeùtæXBd™NÊëÀÉŸQ •·çݸaÑ|õM+ÑY´/´bKÀ¸F8’Úhsª§R4t­#»ï«ì Ï ùÊÛ›Ï4!, ­}hÂ}ÍY{Ëۛ؋KRzQX̸ÇOî³²]õm`½ö`úÄ•+ér†À§õ°G®M·Éq] \»Ñ+>b¸„yv3›|/¼>H‘ öÒý[ì³pŸ2"„g Æä¦“÷‚ü·Ô´9=¸$¤ Q»_Swo¸•¯‘6âjætï‹/î‰Ñå5Ý=o­‡Êæ0ÐÞæk˜“~°æÕ7k­ 9(×â5n¿gn}s•©þâöýÕ·‘&„o•5§ Ìyù,¦ n ¨ÔÊ&\¾P/ž˜Ì†õ»ºì¬L68®›Ò„è'Ï—ã½å5M’6óT!¤n¬ÁÕßcN¿TÓ „ƒm Ug!k,¦zè±eŽo†B/ýåa1œ`uŒA„wÀ\ë{:?ª!0üãåÇ¿FÕO4ët n©<.þ óáV¿ú‰ÝCS'Á! D±‰«P«c6ŸâQ“W±4=iT#œàê˜ÉŒSµ9¹ºzú1gd1™!¬H¯®®~ŠoÉÇ@exÂÑ܉Dœïg'ó$¾ZaPÃQßoW™ÇO2BjåÀIÛ²¯ä§ð«ÛV$„“`VcÞb„Dµ¿S#VSßÿ zx<ÁÁPH¢ÖȼdVp¿|5z‰?)T!E%§ÿmRÓjØŒ¢GÅj Öj´¬¸áÜjÅFÇ"¢±½jDØÕo˺•¼á* LÞ(~4ué.;ÜgŬ ù!×ÍK4Ž@4Ô¡ÔOÀWíJ9:jÏÔ}USèˆÙ Kyè}^5T,ûf¯ÊŽ'ò$*¬å7IQˆ6›?©1è-[Íq+z¥e ×"¶E=©Xôr,¤Þ¢–R¸Ë÷¤¿¬Â½Â@4K(2cÔ© ˆú¯\™U¸ãÙ3ûÁ˜¿\[[ CbDá]ŸN’w}š!:+0˜äp„žJø‚V0MX´ÈJ ‡˜4ïâgš7FÒiÖí¼ädRõWJ.›~­‚ñl•Ê& ߯à @Ðg×ä|2ß qCú£ÊÞ7õF £ø4?zŠïHkF0¾MÇÌs3©%Â,&§›Ìs”Åd6ù<[YLG8ŒíQŠ"À˜×ˆkÚÅd“×)2à“ñ4q«†9ÿó %TYM0OWB•u S<Ĉ—ZBSq˜O#‡t?K&ѳŒ"¾J|–I€ReaqTšûêÆØH{*!¾ô‰óv¿æÖK<ƒ¸ævšØQg˜{1š‰ï>!L3ùgn®«Ÿ°@0hw9§Ï(äðsÐb»2&?ˆ³Ä†&ss‹,ÑŒà…Ñ»:âÜcÆ <© ¢ áîëÈŠp ¡éÕOjq9u»uy·ÌÈë( ÎÕè@˜g®ï\[YpšæOÅ2 Fsâ/dé!-­ÚvCZSIp˃ÂR¾Xóòw*¹!>8¼ŽÛ&suöø6òË».GEŽÒ8åmXazS"æªJ¯«z႘{sÞÀVÜçbVÓs˜¬EŽGõ@MèÙèT0Äà;U¦X|§ a6áÖ‚0Üš1#¨K]fng^¡!³Î¿$€È¨:<¶„Q8=†9MØ}æP4¡ƒN&£`è>L,Ö‘­¸†¤-eþâTþDd4ªûN©Å‚ØÔ"w§&0å¸UÌÊq¬(óºÂYŸ€ä,|ß%ùfJ)÷䀿²½:òðUÃCÅ`êy Í@ÜÏ’ÙPúc±à.¼KÓ=þ¾Èzæ½ÀóŸÞ¤ †‚«¶Êìáé]ÚIÂÅýåU¡àÍóØVf?«çPþ<©å>shE éWÍ‘Š"¤ëý•ñxrÃ[†àV ‡éMj•½žF­aro¤¡Ý¤žaJÙ9Ø éø+o¹MEÔd6|”W©‰ 2}TßÁʽ„>5)ÜBT½WÒm¢;XY"È€{ È|îÛ·°’?Û{~,ø)y#Ý ­-f¥Í·S?3)ORA/ëÞ)šGº7…,Ï2Æ“mA×=;¸Ô«½ç|½’4Ü#Ì(‘XõfR}¤Ú-˜‹÷¶µqH, zË)@„ãůlÓ‹…Ø2 ÐpG¥7ï¨LŠ?ˆÁím=›Dcj8;…<¸V³IpbÓc¢¨ïªþZq§ª•/Þê”Ô@ÐéëaIqžºµÿH# ÷ª"‰zu© ƒq/‚þ³ŒÄú¦×œ5×ÂGS‡çk•Ôñ^{¶tnõÚòg¾âà¹DjÄ“º@p³3ÙêAl½伸N/¤û#ž|†¨ &P‚SUúw¤¼NBú–9"˜¾õ\é[ûužM^Øsuèc¾~%ºNF ‚î5º¦”@yj>Oc/5/°±'’çÁ>]0äØÀá„ÈÀ‘ì.«9$ãùúÇ"Søˆ\²ü|}cZw[ ä’,Xu'¸0m|¨B0xvW‡x8&*è—\F_C#[Û¯ƒË¸4Æl›šc§ ”HÇ´q€‚}Ù €¹ØÌЫ³uWWT $ÞsÁmËx”õ¼NuF2,I¢l]š`B¢ìÌ3Oemgƒ=FPU]èLQ¶ O¶jßið)‚Û‡¼’gžx9ª`ÃSdß~Œ šý»¤h~Åi•Ø7­XDc æ"7 Á†`+9â4O] •Á„žÛH$KÏcÌqñ©@ÿ‘‚@Î&-{é«.…9YØ"¡Õë €@À«Wê4!BÓÇ(d–skyl >:_P@m‘5í@ æ ÇРçòÆ&x[/z<Òì=)‚¤¬ âkœ¼]_Ýw4*׎%6@’a!@h Pÿ ᦠ §>sï§~Ë ­RË æ7À]‚ÜÁŸ¡¸ÙÞ C ­õõ §R‹™‡€,¿ý2„" _Ô8 .XN¨EDbÇ@dï8äÑ•œf>Š… !Y›R£$S'-~0xÖ,ë!9Ærj1j=ƒJ [CÊÄ(Ðå1嘂3¦ÄÓÉ@w2-½ äH¤{ªÌÀJv†ùMH4¦„³»Ç€= =Òög<ðYݰьâët€ô©„ õ„~êꘑŠ*5Ç)-İãÂ`Rbè²ü«'æ˜ù«Y8­dV¡ˆ9x·þÕ’;ºâ¾¨æo¦!iœê¢` a›÷e-á u;: e!‡»ÃžSßž?LÉ»§¶,éìxÈ*©ÍSÿYEÚ¼îqw‰ó\§ /þÈ%ž0Y;8”xYÛõˆ´ÈM¢žÔ"á:t‹\*2¤dÔ|"Z3”ŒËL~Z$õ¾YƒãFyÁ—׺ÇóÅêh3}ÁK×W}Áؘ„èóèb¸ a¨îðâb'õ¨EoL‘)ªÞëHkÿ¥ PëÐ$åü}(Ó4d²á7é½5+$A€tJyKëXs’Þ$»½ ¢YéöÁÏÀíÌÂÇØÒ5obÏ£Kv| `)ùØ“®¸`.¦äLa/¶ØÒµ: ßm¡t\ Ó+5oƒåÕÅcáfµ$dÚW—ÃRv ê»ôáƒÜ[äÈ&æšÇÈ’\mâPkbsÝòÃÍEÒùíŠÔ D³ÙC£±ë¼npš¤…¿¬—×=>ƒö ®Ap¥Ì×p~¦œ®'³\ BϯùÙ‘¹Ì1øa´1XùM2';Óý‹cبYŸ#ƒávsnÚh¨»E_n®ÞàÊvü ÊGõ [éUrŠõŸ„Isƒž°kÛ %L "" %VžÙ†˳‡KéyJèî\JHˆ„†æ¡ šÿÁßBãj”µ&=`eP§KF0Oq/¶ÉblÙox¦ÈU$w­uXöM/wÓéC{¦+#؃\Ç"Z¢sˆ De6c8‘xg‡æYˆÉð çÉÎ:Š´×fÅ~[§x×&ìU´IYäCDNh~ó䯭"fäµOÞ/Ú©æÓ cŒ6vj¥7Ü®9ÔJï’9sç• Ôð«¹ÏŠÖ6ûkç78*À¹í»„èWUŽSå©TRݞظ^BT ¨›žzY ¨ç±2@¯£w%€^G;ëEE^G+ëEE^GY hrªz*ª”§К¼ŽrÑÖ†ÃÖ†QµéB´7ÖÐ^àÏ4ÎH–P“ãZZAEµ þÝÖ)ÀÏ/M¦¢†þÚ³ú×äFåNÕ ÛNwضkyaŒ¥%ÕÿiY4lòÛš‰ÀV¤Û–5Š‘›€~hì Ýû¢à(70–ü*ø’ˆSzc‘ß³$ôt:”–Û/vøž¸ìm{«kv½–ìÁ¡ð\›Yžòn;ßj㨊¤‚zxÀ™ Ϙ֓@oxÍ]#ü[{qà“&vA$€ëI˪ Ùå©! ne?–zÁjÀ@lŸ¾‚ÛU=¹ þ„ S:JR𬳦0e;ZMŒŸ¬ü½L ‘n߉Ôä–\9$ªmÑ .Š79+SÞ‹Ú`ž¬§šŸ‘ÅTmÉ ¹:sM®PÑ™‹wgX1O‰SÃZºY‚f\<~-н`5ÞšæH©w–•<%ù!–•ß-«¨2½‰Ïµ@¾¤3Ò¹e •¦ÁóâOHzÎæ|>|L ãìÞ Í~TVÚ÷`xŠ{ò9”söâ.6šON†µú&H]hs[œû¶Ë ޽˜½¢³Þ½0[€[zÏäV¹«ÙlÇÉU´³:+=šèÞ¬ó`öõrŠÚ$˜½, ²%=‚í'„ÌúÍN|’ÒÂÃDCä…(ˆšñ­2/¾’a€ð~†H#þè\_ —€è’ÉÂi—SA ˜mÑûânÞŒã}åI P [•¢î¡aòñTDÝ»‘‹ }ª³à" pƒm(É >ÕGËÔEç¤óƒmâŒäB:vgI&XÌç1xÁLú˜+|iø–ŸÔ^q!D\mò°ÛͪMpµû!"®v÷¶…u¸ä¾Ú'QµóŠwòœk§×õ¾VðÔÁ+Þ¶<]UVëÛ¤ƒE¿# õaÄÇ—ïmQF=¼Å[&$àÚãcÉõ“t|´í¥BøÑ¶í_Eôoaâ$UIè” ˆ^ÞÔCöHùþ ÈÔMš[ïñ’/l|yמ×"eˆd$íÐQ|§‚¨É]}'+R Ò/D9${´…lmW‹Õzç³ÚÞñ=Z×öfi¨H`LÎUøO*Mà‹¿½–Y]ýa]ËgúæO_¦JIçDÔú.3õeöTèeÐ…×Ñ.R¨lÖ""Éhl§K4ëE’3 hÆà3ÏZآNj‚ Ã×# ÚŒX^H-º µ0H³*HÞékZ§1Q¢›tI3R1Õè©o?@‘Q5˜%•ù—#ëôä¦s &¼BoâËȈ´ú­ç™‡\#¦æjá¢àKÀGM& ;ÏRÓÁÁ×Wu}7O#ÿçM‰Iàð®é&qàÒxžá‰âfUÈI“ ŸÇÈÄÚæx)í…?†Ï+àðòƵ›@T8¿¬V~CjžÅ6 -¯ÎHÏ«{¬B|#3/Ï^!+Ïcdç9¾À7ðbž¼i&Á;ie=T•ßýq7Ý,rñ›§7/$€mÐÖ¶Ø Ù¯?!rؽ€À~È3H"íE(zS^g†ÑѶÀ é¯3ÃüèäWÃéŒ×IÆ oJÎQkÒSÉ«" hùSãÇ×¥îpîë0sºÅæHk’#qÓŽ EA9óÇ#š/ ;¢íŽ<“Úœ‰›FpkОêî¼:z'­Žˆv÷užØmêÜî5"NŸnkZÁû“7Û’‚¢tŽ9@@¸-#OÜÉÞ/3/°“#ZÖë<ƒÜÅ< øý$#  Ü©­pâä™F@Ô5ÝÈ&#S`‹.YãMõ9‡·n€˜2êÓpë+y¦@¶¸¨>܃¯*—µ%ˆw¨€€2Úª½íV%ß¼iÏ»ÅFö£vW# ˆ¶!3'fhª+¯Åmüà÷ÊkjÑé,8À2˜S µs4 ¥ÈŲ$›ôY òÌcºhËr›W<ì&«‰fêó8Fö"=Z‡œÕT=ò¯ÀÝíËd“®t{š@è\:·0u½Ï 7ú bjó80y©ÚÁGúýC]&z[+`±·a\Ú1mu§Šyðãå…5iؽsdŠy_«‘6IÎ×È2‡”ÅÑfŽq“ûÂ!¸ê})ZÔ ~™Ýc² uÑ R·|~{à·¨7€’>·[ͤf¦+a”$¤«p£©lŸ+ ?g%Ý|Õl4U:fK?R!ñÓæehEä7©ÈìkæÃ•Lf¥a¼›%]‘§äLf»µ Éd,iQÿ ð§øÁðµÚš’‡Äæg·§|+¢®„FG¢è-!ã9yŒ¥?Æ’ögºC%ñ$E@¨þÙ¾©ŸvFõϱ=›ÜÄ:E½B T8Å7•þb¡“êÓúÉÆoºn—÷VDCê'ÍcªÀý¤Á ½ü€èu2$x9ë/ @{µ=>ŠZ•²3 øæëALœ­P¯ù7³°ÝoÍß\@Þ¿ÍŽtÙ PÒm#ÔÖ ÷îΪ:zh…ÖïòÃîR Øy½EÅ}÷n¿ý:kgý÷ ¾8¬Ý4ìKêËâq·KE^ª CÅ* Á£žRŸÔ‡–îÇ$OµaÎË/õ¤z—ü2-`EÝeå5“RO_³pÊÉk– FINrjCxª]üÞ§ªoöB ³ýÑôE!l•áá•Èuø6˳-ôµêTM>u¦1!£CB¥«·dc¹vpwé+£fghrΨ×Þ²žQ£¥ãkc­8äÈ6+äÜ0DÌvm”$½5Ï_·þCú¬–K’+Ó+í—„×rf• bˆ³íÛHÖP¸[z;BÂO› U×süe7ZΚ¼yDP¢c)5éº2ŒÌR&åïöô=)£·Ï.‰YƒéB&µ÷¶Ù•‡ôúôµ=”ôÛ˜uRn'ótÕ ¢§YY"=§ÆuÐï x¸¨ôóÕ¯¹ Ûî†Ã´;È>#Û¹´QHä@ì²+ýùHÎZèBaOX4Cq¡{¹JƒtX¬TYà_ÔØ_ºðåå½2äHﻼ ÎødIçÛ?öâ DÁ%2éí¡GVzÖ Âän¯0N*ˆã‚I?ìXþv,‡ù>¼q¶·d¢(…[ž?òù1 Ÿ#‰“á;Š ?èP{3nx ˜€ÀI€ž@+SXh F™<êL³¶·=/)Ì@” ;¼as4ÀñÄæ@ð]Ç{ÂQ<êBô«4D–¿ª²ßnSžˆÀæªáqÛŠ¢ZÖñVü%ñ®^Uˆ<€’³¼RÒPìr6d +RtGP¿¸åš-d'‚z‹“M6…¾ð^b‹¼ê\‰ôÈk™lb¼þJ6¡G^¦~­+>³àošäÍŒ‹±KžE!Qá·¯W†Ëµe_"Uv ±Xœ¦ d‚×åªCáp(¤Óµ±¸©0\àe·ˆÒÂ7G” –¼ õUE–£v ¹5`V©¨Ž,/—Ì@è¤:3yˆ<ŽÇèÓ‰@Å(vǿٛRÎÛm56¨{}ÄbóÙ6¹Bh½]¦ÎC{<(>§‘ý!%Õ_‘ 3köýA¾Ì¬ÞvŽð&œÆ{ÙÀy¶fPÎÏeu6räDá†ÕŒÙˆ ƒ;¸]ê7 © n—b¿³X ™æeÕ­»KÁ[ƒŒÆY3ykÈ‹¶fò–üïÀ÷й’š-“8˜ß¥:?Ó8`6×¢ÁiéÎ dé´ž'«;пt ":Dû:¦ëÌ#ÐþCŠLt %ô“³A ì 3õMÎ_ „‡4ÚËL¼áȰ†‚tŸíf6Ýâè?ب Ÿ=G™&¯äžñ;³øÙ­1t~Ýì-³é{DÏ×g“ÛlÏq§‘ˆnOhú0™l3kf‘ŒŠ§—¥·ˆSÅ%Okª Þ}ÎîlŲÎéPûsÅ5OÙˆçê,eF,/ÙºÕs4Os -ˆbcé ÓƒxÇ«Ù\ŽÎ™‘¦{ëŠ3ÒTîgS®ý;g&©sû§Ï,UrÌàwäcö^5‚†;¯/˜N/@ŽþŠ ¢ÁçSîîbZF Ê‘]TØÃïÉžÔN–ÜÁ—ÅHˆ1 “É3U:¤óI-ë8Ž|Â&·ë2fÞéË:(G1Ï¥õ 8@¯LŸœ!¢híIQØ\ö%ˆdhþòùJn:ÆéÑL#šŽ & s&k$‚«pìò¤œ/(‰Jü•¥ÖL†ŠS³ÑÃ4Ny#m`YZ‘Ýpó^âJF¨7ŒÜ×>9r­w\øDÔí‡ÚþzâÂö¡‚þè¸pŠ—g²~Ž¦Ñ…miÛyâÐ%ˆ~¨¨Â3ð‡Lh zæñ:Ç1~_†ÑidŽ3B¼kвMqHxøðïî†Ü[Ë?þt€(ƒ¸éu°3T\‘õèÍG“4®Ã5v™M…37ŽšÛö¤‡¥ jû4ý €Ï Uvp9?9ñÓ.§r!0·IÁGv9hÔ’ñM J%—…MI5,ˆè?“wÐ3v†”Ë&®w“Úh “Á (©üEòA D§)j26?² D÷_þà±òF†uU¡Õrg„sáShÎ6“/x ‹(L¼t5=/Ë5%3ý$N©|IžLH/Љ8EkÑ}/Ç„cõǦqÞô}Õ¦‰âûÂY6ÖȯZôIrÅÕôà8?s_%@œ¤m\.žÁ×Õy+_/ÇzÃ}YÒDŒsó<¢mHè4]Ôþ XúáÛwŸ¥óØTgØfÌ ”V+eBšEÑW5¾²6ûÔ<±NÃpâ8íØ|Œ=hÇJ±ØzóG]¿J‚¸@”±)‹Â/gMCÕ@tšž'V~ÇVȱcóÞ5O¬ ÷™„Ú-Ýâ·å¥ÍN @J՜ƣÚþ¦oÚ|‡Y:ÝœíÈÕW‰È‘Èu'•TbE°§¶j[EKEæ[ÕsÄg+Øš×à­ @CµÎ±›ß#û¼U]ÛòX 5·è÷XÙ/¬Ò‰h¾LVo™ŽtÅŸÅ‹° Xù²,:¯²p,{mÇ~GW‚šT¥é%€ßå›Gõk{ ™Ó¯vþ{áùõÈõ ZZêø„õxZ1 ô›Md#W¤ mQ/í&[Æ”†*@{Å´§ ÑìJîn Ì‘vµç ºlIµý~ä€Öü2l+Ýd1K[B´Ê‘5]* c®Wh5P‘ƒ€V¬ª¦ñPrDËšß™Óâ=—ê4Ñ¢‰±ì£¤­CÝle]yÃl+§Ê³99nÒ]êþÖ«§bÝeÌÁ1˜T„yzâ%D“2Ôb)Ìíq×mìóÿ)DûÖ¥v{ÁëR»µÝǺԞ;m©M8þäÈE5•ë•+ä诬Kuîy,{­$dûˆYOÊRå‰ãåTI/ÌXr5!jé©J½^w÷®;hOM/æÀÅКÑKx†]ëôE˜;åCûZES°€\!SßÌ8·2œ˜sx/-p¬a‹ÎîÙBÔ}Å‚ÇRruÜôn@G¨µ X^•ÑÝ™¼ªŠù_X©C£:F®ËÉëD—B›ˆbŠvð3—åôg*ÃŒGÚA*ú— 4åÖ!Úù(ìdÃy±i1‹ªX!WˆÆ¬Ù¨j¶-‰íÈ%GnZPÄ}rfèGÞ܆™ú•’”Ç.L ža+mظü˜»0ÜÜã«Ácœ¹91içÁ{n$U!;7’p7§³=ŽE­Ó²¸Ø¢jB4¬Æ»9»·2—ºûÌøÐ2ŽÆñ71õM6ýŽ=5~ÓôÇ#/õýµõ5ÞDX‡ßâä”cÍkš ÇîÞ”ëÀ³Uø´iy xóO{Eƾ!úþ<~l–ÁN›’‡†ézYlQb¾3õÛ˜UÈ\F0̤Ïrl~âñ.GÖz[уښŒ.¯žcWÍ(H²dÒÎgk@X^Ñb·¶I€Ö^eÆï²žm`œÆ–@®Œ'4ŠV@ÁÇæo CnêO§ó—/Ï`â?ÂõÂ{âSö+÷Ä5J¬ÜØ?Kþûµ'~e§á=q»•urOü𩯓{â‡fŽˆ y¼õ×}mÚcÆa–#ýõ29Je)æ=ÁL}W_ß%ݯ¤yT0Díœr¨Îˆ¶o.©‡áYÒ¶ŒùÛÎ@.`”ma]9pßCî{¬°2€nòeÄûcªšÁVöC”\ðEÛŸÐö£ª†„\Š’Fð`f¿É4F©ÿ5ÁBkøð„é‰ð.\WÉÉñƒ“ŒÊr…\ÝÇm&JOh}¦m¢Ü­|} J D ½ÉÐ4W”Ì+m³üžs%ùQ,ÜpÚriþÐ#¯ š|þŠÂì@ÔJ†Üöª‘ƒbò͇,ZÕ© (¨±š£ dÛßNÇ µ£6U—â3ýø\K J ýøú_€—]f]rBhÔ¥È~|"C ’YòOmb,úñ-7ǘ_úüõdõÐoçÕìô œÉâõÙ;Üáv=¦’@T3Y*8F¸ÁÝ^ .V"æ*±þQMý²d‚X’á„&ÐsØšØðxœA‚jDDšÜ »EÓ´híS\ „dóÚ½÷eÂU9š¦%Äþ¦uˆˆ~9É& ÓÈ¡ó°^Æ’b MšõÖñд8¨‡¥ú’¦Ñ Õ& ÁxrAVÍ"‚KÞ®€rüÒSl…»>þ™Ð?Ó{W‘°5³â8ó6%LA²ˆl³˜wvžb®¢jߪ@ðpÆucL!¬CW"1ªÓëëàôÁßQÏa1g>ÜH5î/òƒÃ°TDÊ%ÓÚùPùþ¨H<¼é<,pÈòG~úKþçÀöªAYÄ•æ@²èˆØš4‹…o«^ub>†·kMÏ-ï²­Ò‚S‰}%³jÙŽ¶ùµÊôKxͪ³.½•VŽ‹{Qem x’kæí´íÊÛi?Ú¤M)>ù1Ì…'Güvz(×dFy‚Ž·r¨Ø•Ïöá+“§ Kß–È" F³ĮۖZðß°Ëo—Á°Ö|ìfŒ7ãp`Aœ\Ÿw SrAÎdØ +‡—ûñÝò±%QXÈÆ‰½»T_Ú3{«~.°g¶sèU`v g98&)Ìq©_ðtn `P:9 lr z%rämófÛ?+¿œî@N"¸gk}™— ój½ÙéGƒë’üQ_ fî ÖƒY®¹/óåèöb0“nÛ¾…df‡#wO·Ç¡ÙAçÆ]ÎìmxŽ3Ià̯ ;p=?¥¤‡e¸ÆìÄÅzíîsù”à3®Ë¡_WÍ”t˜"7(Ëw‹®-·?2² $áï\Ò¶2>ÌÓÕüVªðWï‰ÜKvÍG¦˜UOKø¶¯j9"a;™H´Ûð7«¾3Ãä¡&ŸA«~Ç=¯ßè,oZ=+Ìá>_’Ó“Ö@ôÆ=¬«„e½Fz)[û@b ä&ñwµ5H S†ùbtÆw>¦úúîëõJï ~0àòí¤x!&x£ tâÏ÷Ù¡t+rÔÚ™ެ€™gŽÉ'žVkù³½Òzþ,åµ4[ù‚Øfþ,Á<$bª €ìD2ÑìiwéÉ‚ ÿ.Nò"kA¯Ïƒ ;‘àn·‡!|¿ÚzÍ~;Y=¸ºü÷–?‹;c­÷ìG‡í {]O‡˜WÈ„Õ(ìD*“)¬¹l)ݛȠ¥$¯÷0ÙúÉÁÜŠÖïK÷‚Æ4ŠgYÒÆD¸†u/\¢7ñ99V~ÐSöré–ðr/«{ÜTåÀu&÷BàÌ@Õ­¸7 ‡¼IËÓ’Y îD°úêw$Å—²€–´ÙËÁù{äÙ@ì+fwHÄ1Àùb:{ó29-ë; *ë¥Ó^Α3¿#Ç "Gf|èðäZ/õ…,šõþh +}tTt~TïçG®êãÌ›#s’v¿}ü¯§è$eäᕈöø;Uö›æ+»‘+éÓTÒÙ_H™¯¨¥…ó T/„œ‚7²Eæ“E2¾#K<<ÏdØ&¹›B†¸›~†Á ¹›ÅE÷Þ¨ y!UÜM/·ÃL§ëÌrÿÿ9‰£áÉ£ëëiÛ3tˆ(Ô¨€X"òñA²ó̬¤-É,þ‚k޹¼9NDx It ·¢#Ÿ+§*3ë6å,¤ËòZ*§Wvó…„·pS*:F§áÄòáá†)SÕ!E'¦$ˆNÌ 42 uUŒd¯e¶ô¦ óG1(ˆm­÷ê0¿KZÒ@Ä`)\ï"– T@Ä„¡õIÉH±™™Bµ¥ “Y¶V’z-¯ÐvQS_;-ÀÔÔ—54áæERrzöÝ£{ºÔ°g‚/®uø„U£CÈxÚá6Ötˆì3§;,˺hà²<¼U·ojüY‡'}~Òmêy·ñ¥:#@*1'=x~Òˆ $à›g@syÃ-üÜ öH&;,ßÈ"T“ʹ\[„d÷n#Å#)Ö²kíbpÌ‹XuMeü£†dóìµE]^å‡ÞÐ1òÜS )ý·´W«Ú S 8¨ø+v»½¦ežЫªÏ‡Éßýz©ÂÄåeÏܵÌó¥´J=C8 Vé"ä‘ÿTv¤ú£”3C4‚i¶Ã¢ÔŠ Þwz"¾$òM"û;n'QKŸWb¿3ã‘b”mo»y¤FñVO;V*BøqIŠbUI¸CvògÔ¾X v¦ŸtΗ3%º<¹?é^‰'õR~e O£]ÃG*³¤U|,þ´?‚S2MˆZãd^2péIZ—†e_ö¨±K§}—e‰) ÃP^­Ã¿M€ØŠôòrëˉoÌT'…[)/×îsáhŠËM“¸ðAíRBé'Y†8Ì8 ?Uˆ †Ý„—6øu®c¨ÞÑ çÖ+]Ö1€Û9\ªY¬P@ÝŇÒ.Ecéµ®±]*1 %msÌÓ–ŽÜ‡bë[˜ùÑ ÕÃ;62º¢!ÀŠCM‡ð ãh•”ïáhÅ{(èÔ+§ôÈ÷¶r 23S/ü}!T¦¿‚G—â½pF×’\*àÍ\>,V8 ÔUé)†t7"³ùÂÙw‚YfBŠ4‰kªÔT]Æ5ðLF+éi@¬ i"yÅ´^~)$éÎüQ.æËQ«QEžþünz½@ž©Sº¤—Þy 7é)ݧ^A¦t©P,óD ÉèÍFÚzwôšFÚšrôLT¿›s¸^lËç ¹X ³íûMòƹæÙÖ¤õ_i®ZFcÒ鈢@®&ÜͧÉò-ù¢ß–’ÚÞi bsÝ?ÜÉ1è7sÂÁ*¬æý„k„ÅÊ18š–´´‘?S?ÂDG0Ùë×KÄ ~„õqPHð'ŒÍ<Š)& 2¼uMTšË‡Ä@ÝmÅŸ,ã(¤ts_éÛ<3ãê_zï¹9GÏ­™ibÆ&† ´.‡,ˆîˆþÈÊro …€ #dµÏl¬£ ¬i#O.ÉKúŽ…Á$b¿ÿ˱·fìû¢´œ“êŸ @®ÕYÜo¹!†y#.ØÆ¤˜Â =ãÉÜš™Ïˆ¹ N~5·°3˜ÑÊ ãw—æ;ÅQ¤ÐÊ@Ž1‡¡“· 'yS@x#¦`Ŭ±@0*ÃæôsËÈ ;zf²¶óÃÉcÐàx‰0qP²€ÕÅDˆ9µ(Žžrð‹Ð½˜²‰¹×± ŒœÉ„úø‹(œ)”W3 ?ª‚H/ ~–LÚ]Lñ ÇuñuŠ+‰€$äh °1.Ã}C­QjÉKM€2…6[Œ‡ˆH˜ô:nQL³fà ¢TâÁ DN-#¥½H&=áVÂ?¥u(YDœ×q2õB¼ØÁÍV ²±ÊLÊxcÐÅkŒI¾3üVÏœ¸´(݃93@ü]ØËF2˜Xû“£ƒè_1“Aô0‘Q£’$3ŒfÔ¿4DcO×jd±'Æùqéi£<J€èá*°gdaiŠ<=²ø4) ¢vD- Lwô3Û !KÖ¹bȾf>ú¡ š¢ÞÓdþ’%µÉ Ø!¤’,ÍMúxÂjh$‚–´2ÎI30,r¶‡€ÁUÖisäüùˆš7’ãçe³žøy²êÃQÉñóœ Žå”ɨ:3‘S“òpjç#_Þÿq<ù»kc+žæe›¹éðÀ’1Fd7†dÜu—‚'óíaZõ„nÁÆhf‹§µäW)X>ܰž8õ «+«Ðø5ffF)+=œ·F"ðôÊš9hâ’íà„Žb×R†Å\«›4º…[˜Õ{:Š9Q½ÉQ̡ڊG 2‡”UŠ=éóU†b™>Oe`X )Y^½@”Ì|9[I‚ˆA@q6५*‚w‘M n-ŽÃäà⦿¢%j:½Aÿ”rë`”×­ß Í¢[–c7| úÈ´¦ó‰;½¶º;ì€ÑV¯yZŽS=ƒ¹éŸ K<!ã·f}ñ é2œõVzY± ?>¥Óªž}JW=ä=ÌìTÀ`ý)Ü™©20Tôãá4%  ý™eG«ùõáå¸ðl›»¢“ÃäP?âPs2’t5¨°„Yâ5³©šyÞ‡Ÿá°èLIÒwÃ…Qáhü«ŠVr\dØ;ú›hé9žÏCîp Ž”äÖá¨ÞR$ÈÅ»÷´8e_´¨ôgt»X2:óQ2ü&K#¼/JÏpÌTô#ópÕTô#s;Ãxs.xâÑe›Üâd6Æ ·¸d˜ýžL’39h-kP‡†ÒÀ³_ çâ _Ï)a×iLª›¼É©êÉ ôP€¿©zë¦=P N %7´?Qž sª²<sUÅU²˜ð°‰cš´f¼kø¸:ûöÒLõœŒ’eÕ0Ü_¡º¸þ°O”,,o“UìÚ0žu6/Ü5`N» Àú˜½0¸Ý ƒ[ÅI+ý6Üu•3/G÷¾mú1dÆÞM!’áöäæÁÀìW©¯‹>HiœÉÀi öÍjp6-ß6á’Üø…õòp0ü&ÇÊkfø Œ§ü¤# ì৬Àf‡/3£^àˬÎg›p5%1‡•³®¥B¸=ër'‰(á­ç¯ˆçpÖ@jÿïé¹S ð]Ÿf:<ºÏLºµÝ·%4Ãnßs8&–.6ƒ«7ý²•ú;öôVÕ@ÕÛ%)—>ÜÎÕž4rOœøcïýk÷F ðü#4óXÍèn‘ÖNâsÁ­¦Ù‚–?<ë,õ\jr†ßS/ÈÄ™³/Øó;s1³Uà—î_ÞØ;úq¶úlôCïÛomú7Áš½@Û¢õ‹Ø¡[Ós¯h&ö€ÏÐtØÄo#0©O¬£ÖÃn~å14©Ïù ÓL`[ïñB“ú:ä’ÀýÞO¼0P •äƒ#@ôÈ[}=rÀy¿æ!¸ÅͶà7¯%AgmÅC zsÍɉÒì#*@MŸ4Ä Ì<.õuù-9è ±+{õfB ¨9!¥áI@‚Cr´žìÕeëa[ó’⃚›@Jƒ'•ÁÄo(9ñ”ùö²ò~.ÅELw¡±2.¢å1_Љ©0.¢å å^~DSŒ<¯¿N•D‘p1ó¯@ººNK6$e¨IÆf;Ã4Ö4ËÛÆé®ð»®ÍM½ÞÛ±óŠ2%qȾ½]óÔW@èö"¦“suìô(yXOÇÅ…@b“½32áˆ|™pÒ¾91'…¹)ÝW§¹‹'RÕÅÝFÛ±ôHü¦ý4v#Þ= $^Ùm?݈}‘%l?ÝÈÈ'cÄ›ã@¾U³ç7àië2é[@r8ùPì|&ã¿H:ù'@2LþúK M' –¥Ùq½t0ùô½ñnð)+‘CâOkßûŠ::= EsƒsÙγNøFžK^—ü¦ž«*ô—Y_?‚Hyýô²LpÂêŒ@ÉÓ\Q²Z.éÐ{³Îö ¶‡vü{9¡¾,­·‘jÙ`âU*îÙoSÒÓ`%' §/½.g¬Á2ö™Zó0‘ñ2ë#Ø·1zç™1]@+ª»H†t1dð î´FðK}j®5U†Ð&¢eýÝ¢rçJöî¤rÜ¿ÂÛ"+àÅTLªLS¢áC6pÁ+ï 8¸'OLgÙ ¬¯÷nZ°ÜÝ﨎 Ä=Z2­4÷d‘>ï6+ÝÖõ^JìíAjs›«h Ôš?Oqù½ƒ 2\%7ô§v +EÞÀ¤¡n ÞÀ¤Ôë<û—Œ©zôÜKͳ‰[uÞû—ÐèœL¸×gè#¼ÉXíÐP$ñÁñÊV჈êl¥K›‘¹ÐeP›QTô+Ò‹í̳/Œ@ D¥Xf½Ô$“L Nªˆx¶½›ˆg»'Æ&1¤r;ãë‹4eÛd,©fTY-”‡²æê̤¢êF!É<ô8 «Í­NJ+òú ¡ÏM*‹§âPUË6Þ2QQî¡ã†-K‡½Ï‡®Ø‹GÓYÞלL‡Æi%!fl#h)±A€*åS K¬Õ@8º¼¸ˆâ$;©år!ô^jò¯¬ód…fD!$»Ð ”oÚ~“#0€à7ÍÌ”¯œŒÖ9²”!±óôú#hofFÈWëѬÎE\rº’šv*ðm—릌fX>Œ:Ô~ê°õHGx²Â¬€P ªd\¥>ÎIˆ"u„ËÅs–B 8 ŒÐ/ΨU¡•³Ì§ì“ydÔ¼ "sKÚä°#¼ß/×ò¸yjÍ‘žV¿YË“Œd.!Oõ“µ¼Nõ`ßYËÒ¥®¬åQ]ÝW–ò&UP)]ô¥ƒ¼Uuyîq)O °¬KÀï×)ë¦@J"ÛV—€›ôÖÝ{ª®}†ÒW ‘³ü¨£`Š~”ÕŠÉݨŠÕ‰»4Xm»æÆ=¨“‡òHCÒ|²¤;…è‡vNÕB=Á}o^GMº‚Q³¤Ø#Âv}2mŒ% $²qÅ%KÐT"Õjþv h5ÙB¢Ð•k¤TI<Ôè.wo©Ú“Ô7 êSZ—æ:•ï iáõZëL£ ½¿‹æ›¡²ßËÞ-5‰?ƒºëk á†$þz…Š@€DpO«7(‚u©gº LúR ×Ïèœ\”ãGL‹*è½ K{Æ R±xã7˜ ‹õÊ v"ªõ‹oñx/‚†ZÆ‹¶1äáðÐ6p Kòx½`QUwH]‚ˆM'Ô¼gIÓ‰fN BÒªÞó°Ò¸þó˜›ã…¸+á‘q РÃóÔ ÀÀ¤éz_ .ó2èÐyGúshÒÄÅ—«Ç´ ÖµR6+wÍ8hâK½@§|ïîÃÚ!ë.¬ItÓ瑃ˆw¢Ùímq2§¬ðÎÕ6¨Ñ›ˆÈIä"…™Š†ÎÅ`‚@DŒâH^ŽS&½—àâbòT£·`:½€‘^éÃYü”-| fOÉ,ïŒdOÑ>ëØ¶Ô|™Õ ]D¹ŸM0=ÉåaÆ#lq€ßα}Äd=î:"Ÿ,Ú[”¬÷ îJÛìñÉJËËÛJ›AĬ;JUÃ"h'ó ¾2ÛûäfÇ';‚A‡6›ÖD“îG´/J"|îvr~DÍ $_á”ûPW+ÝKæC6™6Ù Í6ŸƒM¶vó¤dsT +'M‰öÑͰI%|»M¢4ù¢ÑËGÊÄÅÂ[ „õÅ@LD¬`8Vi¥Aï¢@üѶWê1‡6|é•WäN?­‘f^à–¯¤(†;"ÑX_uÐüVŽlŒ®„Ù˜&:¢›>†dó–/X˜¤qãÊY~lËàOä,À Kã ÕhˆŠ²Z3BǺ“ÜÜ!+¹d!ípá1žÑNg‘ìLã%ç q·ÃMëV2s©¨›ÉM’“ÖÌ…Í´Ò /ÈSˆ8ò¥Ê´Î…‰Åú:‘±K“iÝCdÖX3[’ež¥,\°¨aÓ”6KÑ.a.øˆ®†å¿2½;œnŒùib¸ì”5̃\à'ÀÑ\lòý÷Äà“c€ž¶²…ˆr._¬±,LP0N s9BD¯\9R ’!kxÃ/ÜÑaÝþRrO "DbŠºô;û‹£¿„˜£ÒˆÓýK Ðæ»Îô¡0ïÔM—bà1ø UA¢VžVø+‡†a ø£Ù„Ö~¤Ð3\"ËÀ L³íÑ»$”´ôM¥O“é^7}É ê±f Mü³HÓüIÝFÝî¥×¢ƒŽb«kH;ÙÕe²ŸÔŸÔVˆ4ºÏ¹Ð5ºÔ¸È5ºÔ¸ÈO/­vCѲ„XãBžQ Ö¸¨Ó§õo t m©q‘}^owÑœ‡ÎÄóØ‚ÖÛúî\"UÇp7ùqI PÂ!'dBhþiÈšó–Oh‹ð’Jëgˆ}ù’s¹£Á¨ÙRº#g&¸[óY‘Ø‹l=O.Å_®ÚkèížOÈ!|Y|‡ « ‘œ‡\Ä—¿ø"_ñå\¾¦ìíÒÝ"°Ç[}‘ù²q_2>zãmŽüPŸ !j·óÈ[ßÌyq¤<-D¿U(¥ví{Év¨Ð.M†J3u/ú|û¹h¬D»¡{e‚ßZvy¯nË.ï6œ[ üRR'H—w»Œ®e—w“xáÍBY2à ¤HUÉt¸HÌålvÉJ©ºü«-Ãèê<“Å> q7°GήÚÙ xWDªr º5›2ä@$ ÛØs(Õ±-‹j jX²† d<ÊPmÀÐVí<´dξêQGá Õ%×¥èÏ@4DÍ]«è·–êÌ›%C’R§_ 2)Õó…%óú—羽ͯ–& $K3Ëuy ê•2qD:¶[uÚ”OÝʱ¹ß`FyöÈÖ¾]gΙ‰$Èh`»2æ&¤žÙÂKéµ…[—@6ü«Y !‘nB4¦_õÃÚÜò¯¬ž«6ZBÍó6§6Ðøˆ²¨é€¬ÈˆÒubW¾ÁÌ®jídµx²Ó/Sm¶¦|å‰åÉ_@ò(Õ òxÞ´JPX2—uby9×´/¨E!Í *Õ4Lvk:Tò’JMãÊ—dõVò®ZyÖ4¨²RM.Ëfå ˆ|Xb_Y5—«‘m³Êľ—~Ô%}š¦ÅÙr0ÀÏŒÓ @Ñ­±?—4hÊSH®Í¦0­—b~ ¸9%­wÒ€È`.ì%½ù”H³–Á•d@ \EæMþ¥Åe²0‘˜únÙôª{ó M÷ÆSã0¬ðo¥A—¡r±Š`S ù aÞs)i{Ð5“{![O8:-üÞˆsÅ,Ž åè Ü”ªßðxsथœŸt™@†ïƒLý€Yº5oäüá<[7­duòGà¡¡üs<¤”£šoʃlí¡À™o/Dá¢@âµÞ¶Æðo!"´üв•(/£‹Æ©oÈu¢|!÷,XFeýy.d‡£P D8ýê…ô׀ߑ3œˆx~¨gò6¿åß@mN;;ÏËÒO´|së÷çæo ñ­›7‡SœvKþêþ‰BÀ¸-ï©pÁ=+ùí]l´›ß‚˯3“§úRáVþ ª}ƒç»#rÜvˆS1îäÝ™¤ÿ\m »µhMd²÷Z3¾k©¯3ƒ \dn¤vò}f’Øzy{HbK~!}ÈAôíá{/Ÿ„® .®k¤by@@*/÷ubiý¦ „®¥‹åÉ.Ðjk˳r¸èU. ÚÖ‘gEZw¯ò­br>/]ƒ*:9¢“öfdmRNË52DBõ‰Y˜ ¢ê®FÀoo2¥ý »'æ@ÈÆåüɸZ&¢Ñî<†lÜñ:1(ã§ú ElÜå’q%"@2îæ¦ ’qÅŠ‚Ç”©BWf6=.ů@d«w/Ù·ILì[Ë„€gÞE4ºòÄé¹dwr¢å€H -xÓ>ñ!ß¾§óÅNdÏ™{)– Dö¼—ùS!¢uÛälSe0ŠíÀXÍé#³¨¡Ï}È@ÇxþŽä¡Ë-òNö¹1Ò ŒªÜN-®I‚»½Óhóóà¯fÔÁ“·w]€ƒKïÇÂ@¥@ŽÎǵ‹è•¤f¦ÓÜ¥*¿OOP¯&øÝ[† UÏLƒ†|Š9{j™Ð3îãÊv(¹ ^êû!EÐi`Y ¹Â´™•3Í!éÖ ÙƒÌ!å¦Öç±g¨ìßz¦nØ|.ò •Ï]§ýçã¨×s;ÚÎ|}eJ½ þúJËWö•ìçm½d cÓÂG5bóþÒq]–‰ýŸõbÏP 6öeƒõ#íβÕ…Ñ/áK h';­æäKr™•¸é;»¦<0ûîù£¸Pí.ŸØnó‘îØ¤3ä=þQÌBê,µA+Ø6¹²ghè†d whrÓ³rRsm¢XSt àEU«@ÐÖO}¬bP?^äÙ„5ôQ6†äž\ϪZ¸»‚_ÒÍxS­s²»ÒqÚ¬•žæñVÀ%³ê¥ñ [¼ÞŽ­SÁ7™ÈÔÇ'á¡‹6z‹?v)Ðø HÅ¢HÍãmù½S ¹´v•Op¨ßsa<&z‘¡üQÑ‚!þ¦QäîÐïJkܦËÝzp”Ýõe›*÷L\ج©Bh€—ëòŽýœC<¸‰®C‹%•%ÊÅ}QfòúSºÆ™•zG5{¹k"ÛÈÅ™|¨$ÕB§Å’2œ´¡ô,ËŸKbóbbëÕ]ˆ6›*)m Þß÷Rè¹¹óC„@ê»|â«K‰ZÝ—ÕªÍ]AÞå¢&Ò&媣}íBÑj#u²ÕFò?ÂZýL"K¨ƒ‡VVäätaT/«lì>’bó÷å>ñ­'¯åÊ#lTï(Øz~4ogí©nç¢ð´™MĽ͛A6ÆOœ¼¥-Íç zöIC˜ YŽ' aÒOi4oÛÙÝ?µÐhÔ–´Ç’jµ%Ù‘…ìZo7¥"ÔÉíƒkûó]ˆnñ…v$‹*ܽ^ýÃõƒÐ‰ëd¯ÑÓ ylúNôî¦=6?ʆyµútõ!:ú¿4ìW•†Ñóõ7HÆÉ9 ý£§ãùèôè.n#ÆbSdÿ=ê"M”¨ì‡X_ïyUPF·$("‰¿fCžIÀs\ÇÊ‘¼*ߌa|¤~`ˆÑ+³s{+ìÔ:¹!w‘q«<ÕFR‘I²il !S¹%ãIFè¡ââ¯TÝjŒœáòþÑèúòãi»²WÆßí’¬œòƦÓ,œ¼RùÕIÁŒd˜+cMìz‘yDµ-r»ôG˜µÔ®Ì#,Õ‰ 4â$‰‚1àIO`·¢ooÜx /IRT#kgÁð"Õ•Ç«®‡ÁcÛF$ýùo¨™´Îa}¦×z³‡÷/í€èÍ)ç·±¬X‰l"Œ‘+Ý×[ᓈzúÐ<¤j@\.Ü_•Çrq?–hžâMõKË zïÔÅLwG ¬~ '0ÁF1ÔÛN"9šïc„F.9+ÿ¼$àŠ’Á 2æP_•äÇÊq¬V:§ß)%^=n0WfrèÍÅ7\eôBŽ·ã&¢V¡.aìœÈË_.MMñø·ùtHþ¢WI¡ˆclë)B|1ðg __‹†¨íì@ð³·S¤#Ð ïC¦*¬Æ=þ\7º…ùàa¿àH9Mg’H¥i?bë2LøfX/—¡sw:)eðÊAc;ê—ž /û¢#wºq¼°:4¨“‘¼ ®ƒ½E"`S¥xð9fAGH¦HŸ‰¶BÉ×¢È+ë‹caÒ¡¯­œ®‘"DñŽÌCƒçH¢+T`à8^˜:-qì¡&,¶ýd\!Œ‡ŽCï©BžØC¼JÒÔ*¢l®{ÈpïØ^Vhß©¾(‘ÂH§.Å!JT˜ã¸ÖbQÑ eãFVýª Ú}ÁÞK·>w(ÊQ>wi Æ¢ Uñl®ýQÆà¦;õpF„cbz{Íœq€æÈTŒÃ€Dx˜)|r}÷ÌëeEh\[CZ8À ͹¦šÉÜåhÇE=Sèš JSƒ@I-ë;hŸtfò\Û["´Tˆ’™k:nƹ®Îûw ºaG¸fv~©¥Š”ÕÞÞ>r‘Ä WábÏÈo¥C7bùoq@O¤¾Ò¡®jJyćåq| D‘• œîáLÚP)ÀyÊ »d¢ëìrÞËH×ÉuE¨‡ôü&S’&6„È ¸ gOîJ¦£ß¡?G_ú—¦ `düÒw/ƒ^I£wà6SÇb/ç ?Žƒ9Ì_µˆ*ª÷,VHB?AgD¸1Ÿ–…³‡û{@¾<ªž–7¡· ᛨ0kì$ÐCï2Õ.uD†3Ì6«ƒ»Žhø³Ú™×YÐ(}¹Ë«2£Ô»½*3.¶3¾ Š£k„^•~KFt5Í*­Œxkܯô”ŒPl´‚êô…#ÍĬ^XG’v§}%>6&ðÎ T94ž2ð“ÞfŠÌ \‰pØpW[.GV8|*kfÐ’ÏDé¨mÐ,7½6ŒëB‘¤±¬1² ª%ýU¥mkõz%RÑaÛZ§op£qÍàÚzi=[½£‡öCDcGÅ;º&†ã{ÂDÒ;œ„Óc4Òàá]·°«zfµœéȃ.µ½`UÐÍTŽ4c¿é!kˆ^fÒ«A Èµ>mR«î°RUód}ו§l=îëž)3¹@Ô`ËÔ™­Ü„Ö¤– éá¨óà ³WG¦Å-“Ô%Æ„kÁ™[ñhÇBŒeÅ â4‘ XÂòñÝrÓu Î¾rønÞGݬüáòx_å:4;Ãoz^Àô–fß®3;˜'Úé6mÊo ½«Ü}«Îl§ú-e| L¶ÿpvÏl-Î>zƒXí½ÏNŸ^1VŽlŒ[7«åÐFòq÷ ƒo³£æ>C^½íƒ7%Ü$ù™ëÏ™ºó}ªÎkóÿ@` ßœo¼÷ú1‚VDxæ­×E¦í½i>ÛdÏ€O,*ÇF¥r8R¶<ò› öìÒÍêǽ©}9²®†²]r U–Êj]›i0ßäÃCž~f@~732€ˆ,#0¨óøÝ œx™7ïµtoìºòð º7žâ3ëË鸀wk5Ýœ©t8oÓ<¡UòGa$ý¦«ÌE^ÈѰàcÏÍ›óK¿ÁÓ÷=§îÎáoxÁ¼œ±2eþBäZüBdþBJþ†©ßÏ3®ðƒÿ†o@RKþH,yEï´Øœý ¿ó?CŒýˆO?k7c-âî|ÚÙ7C‘Îèlò Á!özV/`a‹™@>òϰö ÛHn:Ÿ!ˆFÆ|C°yCä¸Ý~ÞE@4Â0pˆ‚W"FÀu§™í‚n`gϺ[âEÑ‹Àuî8Â^æÑë; JAÙÑ!”0v5kCd Ä£vB€xLùˆ¨b$èYî& AéT=•a–ÚMˆ¥ƒ™z‹Š‚¥` ‚(Ó@ºGù¥C4â°€h|9˜îÑØA‹ Ço¡¡ïVŸ'÷¦E D=œo3 êšÍˆˆ‰jX ÊUº°N¢0¦ë†¾óD? º˜>·(Ò”|µóžJ÷¾é  OŽÑVDA‰È~Í<hå_°n(MÏ`Ö" à˜åðž¿è›P¡à©¨ç@ö3źD<5ºnN?Q¢©Q0N hfT›š¤9šœ%ÐtŸ-êiZÖÁžiV1!½:FKŸºØÃwÑ´'&¶¥K9ìÞÛ81[^<ÀK!RÎ,-*[’It˜ª£knóvc:–DGÏŸœò·MD7”Ä šk‚ÓÄʧаÑD_”3Ñ¢™¢s Ú kW¿ÛäèXø`dÛ ŒµºÐ6ëúP"Dsuúû5Ôey9}ðo;¢ºÑœ¿ãEäjÕ¥A~«Þ¤•#­(×¢R‡4£Ýí­ÜX¤îKD÷" šèŽrõ D rà65T¹ušÁ^Ÿ®e‡j Zz/=ªëM¸±õ¨l~ ÿJ@M‹R膌Õm£Ýɽ”YØ7ªâÚ…hÁ P׌Zµ› Õ1-úèþß‚| [3'/øØ;)‡ˆVñª Àéáà­iÒ$6mxÁÝ fŠÛ=#7ˆ°¶â ¢Ê[|FnrÁS²!ž”?buŽ'·R°]@ûCkò5\z ÃŸè˜ØÃèw¬Î‰Í3¼<±ã†þ{¶ÇÆQ4„xjK°°·•ö²Nn ¢És˰mxˆh l_'wÑËâUCÓ]êáoviêdó síÞ›¥xvרP-ȳŊ‡—.„ؘÅw9k2önq/®l{b{·âºP›ÂèÐwø}B6 u¡3ùêº3·¶ÑµhÄ¡ ·T¦3o¶“ë\¬CöH»&b‹ 9=cc]‡h{…{ïí^Û$r.Ò8¦Ó® è"@tÛ¯¹Ú¹@¿æj—,,‘Í]@ „\íܤ¾[dsûø¡¾±‰¨]\éébË2Ë&D§[ ´€§r¨½l"ªé–KIN½­¨ÜÊÂO)Y*‚n´ÈÖ8ªIŠ R9âš+íV%©‡-éHñ’D±«Í·H! bõ­˜pͬT‡@FYm¿UQy³TD\€J©G§=é_JøV¿5£.¸¥ó°tÞÌ®RCÔáºZ=’  cë‘ÅJ¥[%‡HÅS”C'ER®K'—b+×YC¥2y“N®Z®õÚ%y˜î–ª¼Ê,*ÁÐC§6-H@[VÉ;…^>?Y‡†î;j£V-@µØÆ­ùG¶w-k­Uí ¥SÍtQ^‡,8ýrU/ðäÑgwlŠ™¾É“ç½Ó\øq'ÇE¢e1`‡zf·ÏbgÐåƒ-ÜŒ©ªÕ$ÄÊ)ÝfrךÖáÍÉàû p²…èžOÉé† ràAö#ÝdÑ"ñHÄb*ô²²¨v™_=—ÄâI+ž[jñåß>ùBo¥ë°§ßµÒ5Uû&}•‘Ĉ%§™´¸ä6À2@šE}8è‘‹¨V¦Í\°Nÿ¢€A7‹c¡Ù¾^¾XŠÊ;Am‹Ä…—ãB¬Ú覃ÐÚü"—“Ì ³q‚HT¢'¼lܤ†]ØvÂ`&ާs¯‚?ÃÁÂÄI–ÇÛ ¬Mâ<é-Rê­nêäÈ0urRô>Xào’°ií^&2Y¨ðÑ7k WŒ-g±Fȉ ÓH6›œtEÁìû Æ)cÀ{sû½ëI@ñä–Ôno™d\¶·MM;˜jàÈÆ™ë&ÒgæÛÁä8 øÚ²Of®s.7‰\9á$ãË)¯ÐuØòÔüMW¶<––:O=Œz,(òåytU¾<ù¶½2xˆÈ—.“°[ž$İÉøÇc8oá@“Ó‘C×:ÃIÝݯÂi&re©dÝ¥a»dÝɘ\h÷Ñ Ð3æÚ3.Ü›‡x|¯ü¦¦G ©MŠ#> QÁXwW*›˜£¾UV6ÉBâžÔL±ÚP¼ÿÊÚu!׿¥ ª&§º(Œ¸º•NtêzDMø‹ü¿ÇÒSÓD-<Ât”Uó‰S¢îc>+ …áPfQËüébÖ ÕzRÝÕ)Ÿ¿jQbeSmÚBŸ?Y;7ˆÿ}þf52iØ—ÚüqË¥ÅÜŽtMÛ)@hó'–I+¬”«pw#°ïlš•6ÌFi‚8ª\ž[añfŠuÁMîÚ{rIÔ(àðmÔódnzžM>´}ùy¶N§ß¾õ<›Lôzv&–ã7J ¸ëýZEÉ’\-uכޕŒÒs‡Á¨Ú`#í*lIÕÛ•“juÃ貇ÍúÕ.{Ø©I[©j¸è5­T%Ðcº, X¯Ø+®ÅnS{Õ@à18S|ÉdÛǶŤšV¹Å°“¸Vm®ËKËÄÓlt›%®$œ§IAò!‰øptÄ}ÌCñ![ ÏçDò…À¾1âC¢ÄÈä&’ˆ¹Ô¶½ID‡ä?yR‹îd’H^.{Jö 7„lð8æÒí2èíwò°PR"›?GÉ ñ9(½_ÇÈüuP:”¿:¨ó½:ªü½ŽzR³|Ôwšêô+%äÛQÒ¯¾ŽJEëë(Ñð_G%1ÿ}YG qÔà=-yãgã=-ÏÀ’GçÎë¨ó¾ó.Å.f믣Œ|;ªŽç®6-Qßw•G5.;_G ùvTg]ðuTw¥ð}TTÛ¿àÛ1¡À;í}oG…ë¼ï)òí¨°oß¾PHU¾þáÇúQËWü£ÿûw¿ý‹ß~ü³ýY|Å4ø3kûí?ü¨<ä+šÓ§å}…Âk~–^¿ýþãûÓ_ÿögÔ‰nûÓÿó÷_ÿó¿ýù·?ÿ›ßþÇqüßÅÆ„æ|Ñôù¼ÖñGÿí¿ÿýÿÇþ÷_ÿ럱s°çŸþöãÏž¯Â:œÖç^ÿ¾éø3Öß§ÿéo¿ÿ“/ Ëœý™UD£]ŸùÁçïó;ÿ—ÿóÏ¿ýGýÆ(Ýo?qB#ˆt±£‹ýïÿöýÛÿüç¿‹ ½süéÿÛÿñÿõßáÿåo?þÕ (ðë3ûÏÿð£­îs³µM…˜‘ŸÄ–).7é.odƒŠ„¦EÁ0ô‰`ä‚*.p8~ ­àTçù¡7`‚Wyu^NY)ç;Ròй„ͦ3s‡asŒëz«n#—ïM º‡'…›º- ÜÔÅT%[‰8ü ü8õÅM—ëÅ^[«Î:t¹E—[µžެ/—K» Öên>‚¼¬²|¹ü8u­’ñµ *óbE¨ „—ΉÉ8©s6]¬²Ï[—Û"ÒuÚ¢Öà=Ÿ )뛹­Œ÷šNÌ¥lóâÄG^ Öz'Üaȇæp^?rƒ,Æ<ùãý§ðùk:¦ñ7 ´. úæÊ¦zD{lñÒ˜›×dáAˆÆz¼C>…èKÇ›_áe¨ÓtýÑb[ýÙû’Kº$©}>EîèÞ$g¶Hˆ  Þýµ€÷_pÝøª<W¨©J«/âFœ8£»¹Ù‘H":ˆ0l²‰ O#ýÀÒŠ]Ç"g!¿86u™÷ ]÷§ˆ~<ô,Ç:¦}‡<ä¹DɤL"úÞ±géDôuäÖ‡´a Dý¬!£0BQV«á´dé·‚´€1îPÓ<ºH¿„X>€ràk;Jõaì<”=1.ŽS¡þ9:‘ÆŸ®›}í 7z…~ß•\}ß:Õ׺§½:ôñº'£:85GVCÔ”¯yÎ`ù ÿZõñšGw-úxÍC· Úx¬£êÙ³‰èÓp èÓ(üa®"°¸J--„bû"¸wˆÉÎFDC¤t<Õ?µP©Ekì«\B(ÙbämG§AklGZBÿ¶é*N²!‘Ûq•üŸ[ë¢5¶Œ}[hïb†ÜÒh¡Ï»ãëì­¨IhøÞA„­ßK„“NhW]Å9f‘ë=(Õ  arý{Ã$-¤ê.¨?!Í>ÓÚ}Ól@mØÁb47ü´ªÿFj'Â1£pË(²alè*5×jvüTœ/>´J¤èo‡ãVUdƒ`t'Òõ7 áЗEÛ8˜ñAøOvŠø˜IwUt MŸ¶èe4?¦*FÛ¢ ðùî«¿A=.ÕÄ®(}ðÝù[ëøƒo £¥²×¶XO¤éo f”å°NÜ3Êr*6Òˆè½×ár°tDhøkœµqÖÂmSŠÿ†ÖYÎ\…2::äbm1¼“ª[ˆ©ó Di!¸Î7àç_•c~uM‘!ãÎç—£&>soœÚÈ,†C/~àä’òüUC5Tçyc‡ßá‹¶6¡UßNS¡pO€ÿê¼ëtê0Tòñe§XÕ@ðe§Ž¡¶)a…¤#—#ž’SËbHrØ9çò‡“MœÝø[B`€6<°˜ÓsÄЛ¾ sÌbSjå@° Oé[8`œÝM ÂKÄÏW¹ôϖχ4¡«‚ù @ƒ»_N çä— ’×$Âu`1A4`[0T!š¡"6û†7Ep’èDšn3ø½Çq'Îq´Ó Œæ±=5wìøh\28¢i™¶ ¤,"Ü „…öcj’m—ëËкa›ûªH¢1×6®¡j©gDlÊF÷|Ù4 ‡ufk!èi"9Œ·ªþÀ­ë­ª§é¥d „/U<¡·Ê—ê× È´8 ¡ªíT F+•€>'“ô#>¢º$K©h‘ªHø¬Z”êâ˜éË3oÕþª›¹2ØÊw*?(^d9ŒøÐš’jã®¶w7W­Üeô\;ªÎ"½¹¿U ØØ`(ÈrÐŒS.°ËD²…FQsÒhQöˆ–²²¸ßkÇߊâÇ@4’Å2`ø¢_ב¥)µŠFpsŽw‘;3V³¼uøíND©ÀÖeÀ‘FS1®ÚðúR¸¾Xœ­…{VãfÚ –€x6ž¤èAìÈáÓÄ×ôÉ-¾¦4ÓZø=a lª5i“´§Š9^¸FÅoUéÈ78K"ì·“"«ݬiBŠþgÿÝlè-´·«Îýà 멺J'¢:Üçr é„eƒ¯ª¸©MºCŒèyKW–¤,#Ø¥Uïk'ép†LצÿêY0¬ÄðËU%ìÆ0ÍÃÁ€,žôµ KpÙnùÓØÔì§q{U$*Ø&õ>GôÄ%¤°‘‹ó~àÅA߀RžcQ·I6¥û½÷ѳ4¿÷Þ\P ÉËZ#ÀíKÇ¡{•â·¹’GWÜBPפêÆÜWFçÔ'ßÁÜÑÕµ@üS¸Bnnmr39¢n!H÷)‰Äb““Pxó-ü¿“;C %‘‡£½…àÁÃX"~r×7¢rÅTÇÒ‰Ÿ:Þ L²>FôBžªÂÊðh tüÉv7Z˜Ft9 i•H‚-Û@'L¤àÎ#ß"8@üž`<Œè„=ü¸£q“Û±1MØÒ}L’¢Gtº‘H,èuB8 ¢™*¼*È©‚’–#úØL$–èQúñ‰40 q¤>–ÈÀrÆGt) Dz%©B:ûÎnî)“ûåèCjÁÉCh %‘†;¿É úÐ2‚¯µ¬sî¤h°e-ü@Ð`kk!œÜM t™j$öÃÑC¸Mˆ ULf“…mB"ÂßÅDw×§y¤ Înpð¬Fй—jÈZX¾òqr˜nÏã[k&¥´ãˆ/{AN{Ln§F|Z­"ÜN9¨oád{pšªÀÓ]n ,1ð§‘ÛHÈæ¡)ÊǤÔ%M“Lí1gusõËéaZµ~r5ðAøØ_hk†4Àf¬Lê_Žhl-<aM«Þ±¸øñ1=%wæÈ?ˆÁÃkLŸ|:KªáÓåˆ%½qÌ‘#¶¶°úu‡œñˆ—oÍHülwÐ2€ØœÇ‹ù®ã·;(ºlw8/ü¢1„¢>·ÁÄßspRCsÄ3-±Ñ€'t3@¹ù?ëM„ÛØ@0A÷êû6üŒªMÚ¤ÄæÀ¥Ó·I°…÷! NÕ6›¿å l›_ å¢Ó§ƒ„È6Òñ7Ã-Ó†¸Ë12'µL5Ðò•€ÈN3ÀÁðUêÍWè•õz:¤åá€ØQ3‚w¨Û‹µ<‡lç…`œÖ鱡Ïhó:ÖÏ%âߪìWÕÑxVLƵ¹6îGèjù·x< Å.ݦ|þUäFÜ ”4‰h6ª— }䦷 íÐ2ã/WØ h¢#è{ŧoHBųĆÀæ„U: 逌M¤V#?Õ¼ª×Å%±89‚†ŒÁQŠçêŠ ¤BxZƒ¨=:BNOý¼âÁƸÛݤ24 é¾n¤à>>ÝFó˜þš•ÑÓÐT ‡PnÜý5¹#¡kx¦‘…û8¾H´W %¯ªøŸBã»ÆŸñ{âß1xCßQ«K…îíUÈ•üP—ÔCObŒ3½S®ˆÖ±ZÙ‡õ<*.êÙ¨‘uÕØ¬ˆ(Pã\b0ŒS²!˜‡Û¬ªèf±w†Zh7]iìœÍ Öò¤-¯j¸j¹hÈŒ;U,0!‘ªí_9l®Ý½-*<‡+#Ú  z[YÐ=€hŸ^ ¡«Ù¤€²:B=¶ˆ£vÐ\F^‡Ò©U7‰óK$Ñõ3 +ÖXyä*L6†~îÉ¿ƒˆPÖiŒPÝÕxaEð­^—‚Àõ…_`xr! ¬M@™ìÄóþÉ?Á§›f£q€P)>yQÌl¡mìO‡b' Úe–Á~>§çŽ‚q2BYYk>I#ä˜û0ùÙ€ŠÇ«>k–ί2‹×Â4kHJk(H0Atºˆe`¸ƒt~ü‘'œ1¡r×q¥sôDì*ï¡1†§Ÿ*ô€üö6G†PìÖôÃZï1H{2ñ[%ŸÉ¨…_…=——‚¸Öd‰ ÁÀì;™éäÐ3÷#ƒï8B]Ó‹ÑGh§k·†#æyu?1ò#ÙýÄhýªínÐÊnÓK>2x@JÞ=©Ý|ä¼ ×µ¼1zqÛÞû—Êy¢­×Çy "÷ycôô–çŒÂsx¨çû‘´ì´…£A«sav'dúý¼FJöXÄG¨ÿ·¼-¦¨z²Ë1=H+ŸW@î\ ÷ãŽ×ãé¯Ç%ÒrŒ1úˆÖçDê뉉”cBÊ}=1‘ózb¬@e{ ”Èú³~Þ¨˜R’ÈýU~ÿ»/ÕäK5ùRM¾T“/ÕäK5ùý¥šøRM¾T“/ÕäK5ùRM¾T“/ÕäK5ùRM€|©&_ªÉ—jò¥š|©&_ªÉ—jò¥š|©&¿¾T“/Õä×—jò¥š|©&_ªÉ—j’È—jòùÿjòßC.æ3wö8ù¬´üÿCè$'mþéßßH}Mã¼]îÿ…ô°en‰Ü@VÎÑRó>ñvýΜ£ où)ú ='i#¹›Å(€ú'¯‰ïÜ“1‘£_B¢¿tÓ °¹ª<=¦dçç0²þäŸD×ígzΫÌýô3| $Þû8HܽWm{æ–±,`]E¯Ó=ÀÄÀî[ž­Dâ§öq LË}ßVËÅ–2îL ù×&¡ï\*Ñdæ§ù2r‡ÇÍFÿ¬Óêšæ0¤å;”¾«‡Ã…¤ä”Ÿ‰07ìÚéRòª˜ª{fáv¿µ¶£2 M~ {SÕ3ÿGdüVÆ{*"¤çÖrì@J^›úþY­k^upç Wä”>HÉý.VûÏt~=·Tu;ÃrRáWöYDo6:rÉ=lRòÌ!„ž9!¸4Ç;ZLÛç¤ÞóÔËxþ±è*\šãùR·Û'°Îð‚L)Œ&!q¯÷–CŸg»Þ-`ÈÄUÅ’Ê,zÃ<&%>€…9áÀ¼ˆ¨g7¦zÛ^”Âo9^¢å¾¤1wØ[îÙu¤ M.}ËV9fš¥'ó(ÀšÓqzÆph9ˆZcç2,9ŽÞ2F[‚ø›jÁß8©£c×ãÃ@èœÇ׿Åç±É GN! w–^+" ¸qîTmè5£.mrF©í´ÀüQ% ?^Ü&7ò ¢Øä ¼0Ò¹* Á /GѼÉUŠÒ7œ«>ÈÎð R'!èL%—#Etzù°Hù Oà…qû^zy.Ç^ÉÁ$üÖy!¥'d#²­Ó"áf ÑíIitì£?Èñ µ3ùÐîöˆíäÁ´›±Îˆ™áoò0¬¸Z{’éþv‡¿pÄÆF ýOìþ$7ü´eú 5oLêN åG`°=‹Î 2ü7âwÒØÃãO¶“~aåÔˆ¨ûw¸ÐåÝE'ר<(úÙŽuÙ#´ˆw8é°ù¢1ÜɺhG”Ý.ªVÅqëƒo : €hÇÚapðÙï_ÏCŠ·æÁýiÛ;ãÛܺ·½¼œ#…@si¯¢-žlÄh|âí€çÀ)«íyƒÇ·@ôÁo»ùƒª3È-k»8”ðAT¯çŠÁóq õ‰ù¤-k§‚Ï@GE H4èqv#±œR3©Ðð[Ó³6Ó !¬`å€IÍqD7¼ €¸µ®°e2àêí¿ÈR2wQw 5‘å8V^†éâAÞ‰ÄäÕh $æ5z«ÉÍPè+;uÃX(-®L-ÜfxŠ™p[¢ÉkvvZxa¯e*‚öXIáèƒäQ„Ì¢’IyåÃÚÈ`©rf(->Ût/Þþs£Ù0 é¹$0±“«Dœ»,ñÏ™™ËÂq6†zU=ÌðöiU=Lf'H!ßí´)“!ñ­öS¢2–HìÚ(ùëdT„xøId ÑÔ°¸€…äøHß³›PÉ4|Ͼ½á$ën„ò§–n»G¨¯Dí-¹$­›-ÈÅÛ„Æ9äPZ¿?òÌ´ÌE£Ázõ>šæ\#$ÜO"èp=ÃD˜ºïŠä‘h9  /¤³ÿÇi –´ÐƒUfñšóÇ@lóðM~«6ß©ýhÑ'×νˆSûšÆ›×&‘Zkî:ûè·Mé mÓª÷ô¤õŽÐã7y¢pUi9îÏ{¡â?“¹¡f4—f Z–OcW& ƒlá `âHgש+‰#Le¶'EB8G&GQ/âfÿêÝ£˜$˜7B¶I˜ ˜7Ö~ø$è‡4qÖÜMžË¦¨95°"!ü“•£Ñ˜Hüj9z Ö‚ÐÇf=ËÇõªË j˜Bܤî C”\ׯ&¿2=ÏÞÁ¾X¦7ý4ë­ØPT"\Õ“8´ØÞ%‰:—q°¶P{‹¶ˆºÚeÉFbˆtq\ Ï }þ{Ùù"pÀb`g„φÈvT 5‰W±£ w¸d5s5 Í\"b§ص ø~$3l4"fpq’ ·³Ç&Vö@ăǨ™Ï nª›ÜYàG DÜ:FÑFÍ )pð7¹FøS5¹uÜãbn Â8%ÿ}øo1ë*¶/áµ"ZiÅr>Âå&ÏïQÓTÆÜC \MCS, É(ŒÑÖ/b‘2r ¤%W1¦–Jžº+Ä$H$ FØÌˆÝ]œ „ãtÁÍj„7ˆÃ •)ñA+ß²™Ù°Y€ãM7©f@†S3‚Ìý D%~høí¨±|m¨±Úb¯ÉŒ˜…ÈFï¡zO@”à 5°Ìj¦a`˜~˜NJ{ z˜Þ88buèFŽÑȳáWt“îÛ…ˆ÷NÓ% %IÂè%Ûe} œ )É>îh ïîƒÊZ„ˆe݇Ýv‰X0b'±^Î7#œšÔñÅŽ†wS7ðw'xàã'8¢3@¦øà?õ;¤^„”f‰Ø@ "CW 7Q¹ÄØ|¥åÊ$0 /?ÝA8.VB®Þi¼xì|'³œ‰¸o bÁ2¤$CŸj¹<(XåèmKçþÐÕW53uò¬Q3¥ŒöÁ…—úÀÄiˆÆót@Tû0qj°¥çc1NNu¸¼ ÌÃT/Á˜3ÜÄŽ‘Øþ¢—Z ³VÜc¡4]Ä­}9[/&Éë4ãj±µ¸Xc¢kM;V¡HCˆ>Ã"a LÓî2‚ÃVÂXY š¶7àaÇ&*ù&Ï.uÍ7,Û†ܸiõ*ÞØÑÅ!DsÛfx¨NŸ±¢FKEœ=ª N:ýa§N{OFõ ïœ$õ¨ÂýCA”]m¾œ÷€¨2ˆu@ud¡j…™O=M-9Ÿq¢IéÓÈT‘Ôàb7¶ÇÐÑX¤qÖ Ö±\HEzkòX“â¬?`H¹†j&hQ‡ºˆ¸è¬òFÖœ5½€s¸ð†¿ííz‡ï£žL€æò8ˆ–xî4€èkßÍoKK4!üf·F æ¾aëÆ(“ÃÜ7j–ÛñÔ gÄ,·k•ˆËí;訮i. Œ„Á¢ª“c“4ˆ¨p´ð´\‡ƒP®¬™ˆF!­ßÞœëºU`Œõëßr8x»yDQÓ‡îÚ] uXj»,ÚQÙ‘¦¿at·v‡–‰h¸OnUfjë×MÚÊLÌ@ÝÀ¨çÄÚÚ&Û<\ŒðµT¡mE°ÈÒU‡‹6Í…`¶ëòøŒ‚S\ãÚ‚¨IEGï&FÙ*:J8x©–c*ˆyY [u‘J@I_aÍ©Pû€¨8›48 S?¾ØsºÓœQ¯ËÏicÔô6]¥&m$£ÃÔ&ãîZU ÊjÕ¢:jÕ ¢ŠèþÑÜ‹ûàj’¤‚¨vÆrÒS€ “©ˆn‹òd-몗.RÝm?jó,Þï—“KÏZýAæ[ ªÄUž•÷Œ‹éY~t•*×»:¶ AQcŽrÐ ô7ˆ³Ùº1ÎlD²¼½ê*UÒm®zVÒ“ã DµÖAƒÕUhŒOQ)ýd (ücUô;YˆJéƒmC@cq2ÆS›)s›16 s²-Z–×ÏÅù„&¨@(:ˆºu P1±¤ )RD ¼AmYp¿xj®ÝM‘|¡›€‘Ø\ÔR ›À’‚6J-‹ô+§jóÙ2D:Õ‹/FWéúíɵe!ÿ‚yÕ”/&ùkËÒ~Ɔ‰ðÇ7ÒoD£–åÿÕå# ‚¯„3˜p}#4¸ý@’dm)"°5f^ÈáJßl°Â˜§ähÌ¼Šµ ¥Áañ'‚Éã…tL/`ðìóB&'@ØÆ‡ÅÄoD瑼”N¾<“/Ïä×—gòå™|y&_žÉ—gòå™|y&_žÉ—gòå™|y&_žÉ—gòå™|y&_žÉ—gò噸òL¾<“/ÏäË3ùòL¾<“/ÏäË3ùòL¾<“/ÏäË3ùòLþúòLþ/<“9éeI%ÉL“ ~D\I¤A—{!ó(†E@iÿD&{è/Q ú²yÄ B ê~ Ðlªæû ƒÚdÔ³þ0ܲN±˜¿]’±ízªHAz@†ì!‘¼Œ@ ·Wè?È`—t²ÈzÉ'€~ïÞrñ¿A„yv#1îSó(ÓÉÐrÖR¥CïìË–zÌЄÖÝé Ü}òNß’éüù!UvöëoÇÀIAê]Pùºz^jÒå;vùZ£y¾kLÉ€Ä2Œ@ž}“ÄP/?F .>¦¬1~<9L…`ÌŒíkÌPOÌ&š}\òè2Gºáº/÷ŸA#ÑiÜ(¦þøá/5ÊDä7u/ùÔMó>€Pü|ä[ѺnÎÜd)6”¿U¡ü¿y(@ Áo®  Ù¹=ð륖5wB î#ZÌALjñ5hW¯âÞ-<°9®´éR&ñƒ`ížšƒÎ úî ¹NЇ¯Ü†U&>ÁÒà{W$耜DÐo×ò^²’ NÆ2Âg9ÞÂצçÍYDì~?Jÿ ÊÁ ¸ùj•ƒAñî²’à2yD&ò¶š ¯ôá c#A­X‰àSîὦ’sO/ge˽`z¾+KîËf™ós¢²‰™_r]Ñ@º÷õ[fçÑ[–.¯‰â·,M^5_´QkûTíäè—î@Ð÷O÷º] —³Ó={”B¯‚c*í¥è6Ì0˜„ ¤¸ŽŒ8–â%WÑê E0Ј8®%'Ž£(äU2 äê¢Î©uÛê¨Bbñ}TÙ+ÍñA´ ÜÞGŸ(×f—ú–_â‡c™>ã¼mÄñ‘£«d.º½nŸ˜Îï—ƒùᤄËÎQ ½Ù¢*›y¬ÑÏli°ÎYüÌÊ>N=J:¬sù?QæB€kÁGë®ï>ª2z¬ÜÏ?ð¶YØQAøjü©AW›Ç$þÄŽEWi|ÊÖòºNþ´Ÿ7@_{»}œ!ëmnâ4Æ€¯=€~¹¦l×jÚÖ,ÇŸ–IÝÚ&¤ËÂ}Y •€¾B'¡z±nÈP›»øÐO$<]ÔØæQZA@;®H»ëé4‚²6?¢‹ô¡TS 2oÜ¢_ÇÕiGn뺦 Ö±iã‰í6M/ªR ¤5#EuýB  hlˆ£@ÎaìzÆNc°?®ÜÍÖ±ïÖ©ä-µ]ÝjËåÑÐU]7¶U•oÌÀºkÕçd ò>M'Õ£ªü@4à¥ß©ÃÓ´Ÿ9RgXÇ–6§h¾>.A:åpiHu†S6gÕcBð {©MDó¬l¡€ð‘Yr „Ç¡Ã=õPù!½í"F ²—˜ž|ä=ó ›! ¢Ë\— árº96W~Ñ?=ˆ9—þŒoDÃå…t.ǶAûÊàý…hn~!¬~! 9ýRöOäp7z\jfÉ¿D^š&_¦É—iòeš|™&¿¾L“/ÓäË4ù2M¾L“/ÓäË4ù2M~}™&_¦É—iòeš|™&_¦É—iòešüú2M¾L“/ÓäË4ù2M¾L“/ÓäË4ù2M¾L“/ÓäË4ù2M¾L“ÿ'Ó$TMbéãÿ}´¿ÿ×{É¡Üñw>äy½Dhªôûâí£_³¨¶Ûº™/„yß7r(¡çᑜDr"fPýÙ¡èöBªž/§@Æ”€h%*]ü ŒßÒ³ëvÝY»£ˆ­^"š9?CÀϬÍd9ùÌ+ÎÕµ³3O5apëΞ3j>³6¥òz DS0ƒ=£¶â½™fˆÚ<)k°ßM ½¾'# ÇÈÕµ­‘óG5£³ÞÖ=qK6 Ú«š<§n쉻鉟‰»ë‰«WáH 4"ž¹¥±Ûª×eÙT¢uYxêÞtcÏ܇¬á÷±¼ø,ÝX37ƒ"¤IÌóÐLj¦êN¥53—°6²Is^f„ˆfá>$ï”S¬\¹¹TSÞ)§Ï¾Ù ZN–²GdxS ¿PcÔˆzרîlÖõö£6ûÕx‡Hû±‹ D½} IsÉù›¡Kàè6‹Ê\VV=•¹¦µ±ï‚ _³£BlÖ(G5uÌ t˚Ççs úTÒÿ¯Í^ Grÿh…§£õŠ`÷ "[¿å!²¼[›Ë‚T¹Ûµ#Œgz€ìÜëQéW_ÎÂ\±ñžë-ù›óñ‹ý;?£±Çç§³°Yœøé¬”&SUí,IP§[9(ëgÇ K$ §A%|ÈÇ"mjΠ^Ju‡§#ŠÙé@Fé^óC#5'ÅèÛ¦(bÊ<_Ê,¾¤ Aç7ÛarLeíÈà:þdû9ÑvÉÎ o“ üR¼v¬?E±#ùÿéÝçåDD! ÄB~[±ÿ%ÉsíÀ"7ýÖòÎéR—ÒßJ*ÄïZ*FbƒéX’üÒ1áå‹é|,Äé$€.Dß%¾Þûj;i2A½9ëîÍ·¼™BÜR’¾Õ‹Â>Ô@K%B~ÿôüO3lFšþŽý^ƒCa·M<´È€¡‰f÷£]šž}Û˜»<ÑÒ<%¬|×F›ÁÁ»I« Óˆ©,‹^7'B‘RÂÆTW1•…ä–0’ñó\DmQóœ+;œLÚ]šÔ†iMMú ¬™ÂئwZ ’´c¨4È ¦œ´ÊL쇊U¿ï•ýPÙ^¶®ì‡Šm½‚fÄçñ–-ÈIuÒHï.S +IN0c ÞD`Æ‘¿$KÁŒ©æ,E5\‹”´¾ØS†±‘ê* T¤ˆ¼úiLº! “ìŠ|ˆjró‹|ˆì³þ‚öb· Ö|u›"YÍê6ERµð}?xlá%B+ˆè„lpxLDh‹UM',ò\kÕtÂàùÐÎJlLiχ–8¹å¦+ÖmF`Z׆¹³R£¿­žm±–_TÃöPÉAlmû=+¦:x‰ÓW9ñ¤?tËvM´«Ì͆ÏXO¦Q=+e*ùHìDÐ:½™Q®=NhÉò„ETØO€kSg\ÌPôÒ¤×#s·eV{E¡J;Øùõ,ÓŠ=ä¢ÿ›™´…§ŸäWy»ÝdÒâøVs#©µpÕŸNd[XÖ a‚0m퀠­FO"-jÄÚÉ£e [ 7 ÃÈ4ë™”rØð‰GÛ`€ ó¾úƒ›ˆßŠüå°ü[‰`†7_K–k³äk¸ŒÍ’ô`nËK9„à‘gãäZ6pv×6P–¿©@dï6‡GT×8³(€DuØ,ª›wÉQÔÕÞ9ûÎm:i§¹R{8Ç}Ð*ïá“÷ž¾ ËcVXYøUsÑœ«&SŸÊåáA©ÁÛié>•=‰ù0Ï[Ã#ª3÷×–NY–˜ªÑê‹nd)wfø ε~~†ÃÝÖÏåA{òù/àÉ•EIÍ`T-ÑA¸˜<Î"øñάlŠïz³˜r‚ Ÿ_}×Y9îðOMÒcÑô4Ñh&øÁ‡ºËîiáÅl¶>YÌf áëÞ9õϬRƒi¹m’Ägƒ•rþ <ÊKËJ6–ÈusuÀ» W,«ø.W²-:—éJ6„AôÊ*?Dï¬dÛt/ÇÅ_Õòø“åý^x·µ@¶îŸX¾‰y¯Ç„¾ÄŽ´“¨!îä$sÙì|£e­4Ë€¸¢~Ó¯|xCµ6GÝÙC6 ÌÇÌBr&ûX.Ç^v;‹§¹Eèã¸ðs1t×G–¾Œ>³hRF$]‹çºì¿³æ—#‰²Ï–_ ÑÒÿ$€üp0çü ӳƢ tŸË³Æ&»Ç(¿Fðæ‘À.ü¸ó¸RoóÈÖçµÀFBrôU¼lìhúÊ¢ã™àƒ4ÏžéÇÒ=à¸ò~<ÚìÊYcMu5xwþ½\%7–¾ò³I–ë+÷=‘ÁUY ¹q"ý W áfÄ¥ïâ•|“ÑÞwõdÄâ¢Ò<Ýoò¤úîÞŒ,^j{ll¦»IZ@6·'c;Ý<) ˜Ò>;M–`]¼¤Ï¿Q~þ®×½‚NsrC³Éëê§æóvv#3^€Ä‘¸Ÿ4vçLyF>/Šý¤Ä&¦Vü“7ÆêsòlÁ2«²}Mä¼yãqîë‘c·ÔoN¹‰Ô×#džºß¬aÜ4ì·g¯x?3j2>ÈÈNkdæ3‡èÅXùÈvv ǃL¶>ýæ!AÈ(%™‘šl¤ý©?<øGDnÿ‰ Ï ‰Ìôßùê&_‚É—`bäK0ùL¾“/ÁäK0!ò%˜| &_‚ ‘/ÁäK0ùý%˜ùL~} &_‚É—`ò%˜ùL¾“/ÁäK0ùL„| &_‚É—`ò%˜| &_‚É—`ò%˜| &B¾“ÿŸ&5ùýß~ýÏ_µüŽÿ§ÿøÏÿýoþöë_ÿÇú™ŒC]úsÑßþë¯Ê?ùÇ¢)zó§ýþö÷_ÿT[ùç¿ý_TŸúÛ¿ÿõO­´þW¿â\³>Ÿ/ÿ—ÿÿËŒ?þ§ÏWûüÅ¿ýÛ¯ÿð«…t1Epùû±ä¦½ÿÄ`P4C„‡À…I=ö¯£‹ü§·– 6Á˜9”]~$~ZÊí}¥™ô 8I<ÚA;†Û? ÒsLÅ¡ N &™™°õQÞ*ŽÑÖ$ñhíMêÉÁú$íÅ@r ¤IRFj*íu¨rèn‰%™h±µšÂðàŽeº¿•›¶ÖÿGÝi‡0Ÿ®Zz’&µ$ó>¶\jR6jÇÖI ý¸U£ƒ;Μç­Gµç‘X’5«vì,uã|[;Ô¾ºÃHo1¬=§Z‚Y[f#¦Öž]aÝ­-Så ôƒXŽÈ¼=%—Š^{XŽHôŸÐ© ¶Ç‘Ž™yE[öH±h0}ûŽmÞQ²-“‹G·l_ä¼µÍ)ã­¶é%‰´«;So*jl°QþÐð‘Ä#½¶»„~Sžm‡HìKÀ-þM>Ã?È’@Ÿuà6¢SÔ»º&*áRH×meBC„N?U%›g)Û@Øê³ÛsÉ#x·Û¥^Wjâí&©SËæíFòÏ"ñ%{½Z†—aÞ7é-î)%Ìݤ·¸§è»YoqúK5ë-³ñm›Ùúqßa¸|×+Ù<'UAgO¹Â]%nšz†»J/.\’‹‹©‰XÉ_y¤w%ÇeiÕ!}¾æ¹BÑEåH M_³©=kÕï8—´kÑǬ̱mš¸ Ím ªü.Ò×™I@År^!´6‰hIå? —"5í½ÜÒô„1wéÒµxæŽS·ä4I·Ùq2_DÔÓãôm;L<òºWÚ¢‡Sâº1=êÄë’”’ œÉÖµŠéV—]×*¦VÁ^wRáUa¾ðˆ˜žfÄ"¦C·éRx]ê}ëâ º ²‰K{M¥ÃÖ-Òzí@0·,åÚ C€E„ãy1å„£w ©¥`ËáŠã*J¨å“ª¿Ñk Ê­¦¬4ë6½¦rd©ÕßT5ŽÙ,X[‘¡Ç))Ë^“'¿¯™3‹i×—°.”,çï—ön ¶F¨­|” ÄøŠÇ™N[-kÛ§4qtrŠ+/}â]ø/vëÅ,<¾õ’¼õtm-É[ûÙÐÉ}Ì%}ëiÖ:Ì%¢‹†4›»X‹4 SÓ¨ÙÜýå¬Xïo[t}©o)“2Ôœè¢aá0«6yhêADßijÜÍâþ7 „Oói|îa›‚xa|îaÂ_ X_â\OÉŠÐãÖ7™0-\C¹”åkÜvk¶ ʲÃ4GˆlÇ; S#ã©®x*Ü%᢭©hñ›X⇢ÎmÑRræ?C;©á¬3îø:éú·´è‡ÌÌÚ²ý°Àeü7 §{[ÿíQ‹Ó9BîKWi]pЦ…<^JaP|)KÆC_>nÆvhÐCg´{Gåz Þí¬.½ëT°Iüz=*÷!¤¿…L¿_Bù!Æ.búËJô)¸¿D0zDù©–/áþe%ú÷_V¢OØ1M.¬³Dƒ®îÛ»W^˜"Ñ„Ù6ÛO¡½Þ ˜š{õkSÍWüÈM‹I/þ¾MÇ™vý}›V›~§qJˆgªFlÂÀ„ù"gÖ C·åñ†r•Ÿ¤ÚjbúCVÒJÑø§+Ì'¸¿ ¤ ÑãÕÅ~¯èk ¼b³(9¼>€4I./KÕ·\HªY[õúX;7·­º©*×Ô–kD­éÞ¡ŽV+'x;ý”üXÈö¤0žˆ:~¹4 QðºÅz ?,|YøKæl‚éÝQñn.è45góÈÚ4ÒðÀ'ܦçokÀW6àHƒÛ4·M‘²ëçu5‰Yà¾ÖüñÊiøÓ<.Má— K|à"SP XŠkOˆÊÂE>ˆ@°©+²maèÃÐÃÖ5t•ÜöŽ|éÚäéˆÞ}s÷ £+ÞáUxa`´vÄhS’üÑ(s‰ŒO4Á²ôãæ©O‰ô£Q À¸,ìü6<õ©‰4…ù—hz¯{)[~cͱNñ‘I–®em2 ý8›µ)]ÿéô28%2"OH‹ˆ©¤ÓZ›´)xÜØ€À9l»8eòl󸺉91ßZXÖÁÑîp@j}ȵ0ºCǵÃôÙ…®MF‡|ö†½x`é9ˆèåãÃjT#Š;z”jzšzûö—¢õt&Áú”{§x0NO£Éèth›ÏŠ·êݽoÈŸ¯›½œMpkñôè½f{1>6°ê]vWduÊìa6ïöá·µ£ýt¼2>DÛîÖG¡@´Æ`•l.0ˆ@GoÖ‚YˬÂ;XüM÷ÌË£UšH„õ®îK/íÙ\žß†*Ð*øz§ KÀwi¬¯-íR·÷%“ð öÉC ›b]DªÊ¬90»œ>ëÈÃ|öÇúܽrÙªÞOÚŸÁWY[ NzM Uƒ£\x1k³ÕIÒ™0ý€á]Ž'±FÂI‹O!Xì’¬œâqˆW,ÞÎF/€eg±›_ pN*ñ䙿ÖÝmÓ¾±4·ŒÜ5f±ßU x8{dMúÔÁj{w#Ñ9ÃŽÛ/°è“~O¾‹GÂûØtùö+Nà~æ2Â?\Ûùx„Ÿ_aÐýö|XØ\W^a^Î[ôÜÝip¾ Ä{‡+úiF`LÚü$­ˆvQ üMØ­û§¸d†#»:~£c¸¶kBjàù‡Õ»æ£V4ìàï5‚¶8ö ÝÀC^×Ð4{díB »Ò‰^#·!µöõZd‰aq}þ“ç©*ë±··»Qj¿u8<¾}€éy‡–đа¿œ;¼ð@D´ª<ÇmW­ÉCß©Þ5WîÙÇöŽNÞ)#‹fÁ'XÇÝœ¶“Ä.d 7f üÉã§‚v£§EåXvN ·žj*«ºF"Mº`~æOY™ö˧¿‰¦YŽÀÅ$t€(Þ@L`#+¯à¯@N쪪¨ÁPñ'.v]|®l_ücuBbA3OlïöAºBpSýcº†s$žÏVà§J¶8cäy¬’¹0‚ï4N¶8c)#‹ ‰µ{ŒÜeUR4Fìò* §1¼Už2ÇèîÔðA ãl%§dÄ>Œ`Èš­ŽüÉsTeªzôë3[e:dt‘i€Äl>º9(yQì-G¬‘º¨rìö™mANÎèÃ3”|zF·é¤ìiÙÌœ¬t³a/Ùý@  äEízP\íd[…=b5È«b[0šƒ,XóâÎ-<•yÈÑF¶‹0)ºŠT¨‘š Ÿ¶–aÄñ ”Štw)rñ yž©$fº½'~!ÚaTèK|ç<D5DâÕëð hÀ5¡Òrä|T¥ú O*/Dk¤ļ~''%2ÒFq4ñATÎûÿGfÌ?ÒH@-‰õ]™õ÷7â±6_y˜9‰ôëŠí@†ˆB|ÄÙ:ÿ9/ºÑ qæAÄZ>™™M?’wåðºÖÈ‹ªþD¬E€½Ì"«ç¯ƒ†fÆ+RWû8Œ6$n>EaÖ0èc'2ÕrC§l êÁžSM‹F)bªÉÀ%á­çg')näis3ï»6™u+ï šä^žÛ¥oÓÓp/'3W{C°W·ãö€(½{>±øÃijˆå!§P«ˆQ¼K>r%uv9à>¥®ôÅ•w>d8î¼1Ù•Ë»ÏÒÄ®œùÈMìÊ™,fr”d #£““éGnâW6OŒ¥‰_Y=1]k’þ9óÎà/ÏÜ£ž.;D;€M9sÇ_Ähf°BH¹DKhÇyÚŸ{Ê+­;Ý8É€3V»Â"ÞszÑ2H°]yŒ™ äÄq³ˆ =ާ/¸ªÐD^PÚB°"Œƒ3¬í± à‚W¬é«0:ÜcņÊ"/7ŠCš‘N³ÎµEÜèa6×,æFWïйÑÅ‘Ò"nt¿Þÿq£yNn´³aàšåe˜‘uлµô¡{îHŠˆÐ½{å)"B÷Ü’”Ãù%ez1Ÿû3@.‡}Ï-Iœ"Á¹³¾½\†çé$ò‚ð|úûc¹ê”5zØnßr/ÜÁÓ÷ÂS“Éï…GìãfÉŒU"à5XÔâHyM#X/>¿ž]WOÛwmÞ¢6AÑCíø¢~!7| år?.d«0bç>ר@שׂ¡ñ=šT\d¤Uí^k¶6ç½>[ª?ô'OôÊFžSA@£´äæÉa𹀪X½,Ÿ ¨Ñ‹·qÔQ™ËÌ‹®Jaò˜…ÃYIÈ"P~œÍ:çl ý£”ÜñðÌ×nFL™ÅCm¦a%Ûµ B;ƒ5BŠúéŒHÍ£o¬,ÒÔÙ7ªF¡Ë`ÍÒÊcvW]SžÄ·*Ÿnžß+ÿôQ.¥o‰,(*ªF P6zŽ'jÚ“i$‚E=—“•{'çeF2¢Ì!—šWÌdlV–9èÃ(JTŸ9f²Y$× ½ Ð53ŠàDUÛ÷Aµé¾¯¸ê%÷q˜2vú•5u~’|Ú“ÐhWÕ|Ë/Ñ Ëãv@;ý(£ÀOwVü«íþŠˆ­MÄ1ŠuDå Cb$›¶]S-Ù6gu²à5‹¾qgAÜÊä'JpPèØ2\ˆš¿å¹[!Æ( ÔÜH‚Š,oþ ªßÖÌû.lT¢4Ó¡V¶¢žSóM—$À2p²’u¡¾ña«/ÛCOÖö šTÍÇnT ® êâùæñøÜB¢•cp§×291¸CkOrbTÖLNÛ};TÝfæ(GcO[r&ksuÌQ¶iñ‹)•¶¨ñÕ‡˜UP>*ƒÕ—·QN¬¨Â˜,›öŽˆ[ÕºñØž•ë¨mž™g@ ÷ð™ÁÙ‰@J&0ð0Ã[¤qõ,=[‹Â¬V \(%S'®àVV&ê¾g&i®JÃ33T([?îWS2}ûûOê_Eº¾?kš€8Ã\ªÝ¤«P¯‡Y,íy̘Xp\U<*¤»ùþ`Rí&ÄO9}g?hF¸È?óv)Ù¾ßP ¸þ'çš&®ô$×ð²¤R ñ¿h½œº@d ò\.Ãóöd*U¶R#ÁYДQÈDiJ-Yxœê9-ýx.&K“uóÈ%'õ~@¶·¼Å!€à«Ô<¼/º[‡ÈÄJ=¸æÞLÙç«r-h!„‰ˆF–2FÝ'¡!¼Ïа;»lÍLãîÒé°RÐ$Ù:3“õÕZO>_zÎð3çš 3™hrò·H…m¤39\H‹u”&iúP-ñoi>,Ë“üq™ÌL"â„"@³_ñgÙTU4>ŠÃfc4ž…@™¢TO.‡ò!FÈÀŒä†@%í^÷¥Cì@å–ÞÞu\!ÉÔè¥r¢ÙæH±ò&cç°¨?Sv¨ vódu+U©îð¼&²M%IH£šÝͳÖån;D|A¼H Q?¾ƒª‚7gŠ+±°››ÊàjOêiš½Ìº‡êæ¤Kzb øßMŨãòŒ ë@ëIº\*¡„æ‘zÛeJ(%‰²jݾ³L$ŽðC§ÜÒI¢Vb&™¤†ÏH&+eêIŠblÕ``@£žö"’ñE•« ^U›Ô™Á³ÐW¼žšD2rCjfoV‘>æI>b¹’ƺfÙ•›ÒXfÙIcdÙ±)¶Ä~­Œ˜Ò’¾Õ¨½Lu«œ-QëTÆBßK­S¹Ð"ž]eQUÈ„é68ºÇ¿E'E?Þ>D‚×W)M&2de]J ¢æV ‹U§5‰ˆIY`*¥‰Ê8w [º¾+Yê ~ö@ćlLS„pÛNZh¢Æ¢¶,àŽ‘#DÅ89„äD[e}CMTQ ¦] ç¼5ë‚yŠ×J]»`§âµRû.8­?äñ NŸIÁŒ¥PŸTöK>°|qÏv ?B~¦ §ÖøÀã÷£”a¨kMO~‹b /QÁ@ÐPÖ\]ºµó˜NÏ—|!ˆÒ“ˆºýÐÐä¯<—IÁÖÿ›Žì/JkBpÑLqéÿÍ©5)t¬é93xᘦO`“ã]b´˜Óâ$+=? ‚Å5ój ¶o"¢²qãLÛ¢T%qRg׋öß‹D¥ “uŒˆ{m¹¿aÁ·`ùK†3Ë0­,♌s"Öê$¥²Š?‹ªÃópÔ$@ql·ù"™3´FÕ'˜¢>Áä)=îêuÊÜeÐ>ÕU’QÎ< «1ÑWXK¯å7k뺋v ÐH©V]tõÍ¥I»è%š¶RQq]»Ì'.i÷¥r,JÅÆï—ºl”²ð#<µ6ƒ3üxjmÀA~„l£Ž†¿d±[TùÅO¥ .* ûï—hn”ì@f±_O»±dÄ»Y'(ãÜ¿_ú¼(ªDJ QŽ×1”]""hŸ.å_'<£\ 3xwrgYx¯[Ž.ê °t‡ £VŠêÊSs5 ¨xဢ€Š+Î#®¬ý21ˆù!WH@}òZ[9Kr¸×¢|%FÞ­…eÔVî.z¸R#Ï|ì"—åQ•Ž¢6¶±“Q‹Òƒ/qê¨Ã\–Ö`Hk—ØÕLënéÞ=ZÙ(Õ;¿9í¨æëº¨eÅŸÕ½U?L Á—rw V÷V%pYf¿Zqv‘&fs2 ŠÑ·› €¢mÑÌ«ÝäDQ¢\%“ Š¹j3[JŸ+SYÉJ­™ŠÞ’žƒ^º®b!Ù#©´°ìzT}RWߥDQÚu•Šmë‘€õ`£À”Âú–wjÍ´· U!'ÏêË&Kˆ6]ZËzWèÔ?5±xg£n–Êõ>ÄGm-Ö½TÄ‚\~@«æï¦Q“Êú[Òmú~TúR¹Þ;„¨æì:„DÅ0¥õ» F»öOÍ»ˆ¨EÆšQ¯Lm}» l‰ª=N[j[Á–bÚãh°)ˤf5µõUôLýÈÇKUÚí÷Ëo! ¹1›¦%Ãf¸æeÛà‚ðÇÚÁEã¨î~Hƒ¶g5Ë÷I_'"]‡£'΂þuó‰5©H)Š¿Œ´z\}Nu< *Ž–ÌËaÚ©²€ ˆDú“¨²YVð‘¯EM9}8;¼òð^H7Mƒ‡)™7ÂË ZxM¡f-~ L”¼‘É í…¬?ã'`·ŸDbÊ?P‚˜ÒËžä'™˜òB– ¶re§8Ú, ¸')LR`|þi¡´-V]M¬DR“14€€õT™uxǺ’[ä<+¨×U,i–Hµ¶ÛIŒ*)¦sø‰IO­sÛùB†õ)·XmZèN=p4åO‰´ß ˜Gæ„­¨¥êâfJ5¸eÏ#7ÒJÞ¹¸y㘆m5R¯htkÄIä(Èp;…™ö$aÎO,:!ëY_2LÃ&)Þl½nüÂ~­P(Q¨áíu GrÈMUÐO,~aó›^Ø-çF« )úy)…5lÕ@µ¬J²cÏÛv" 4R&kÊg-‹Z\ÄAS/)vDÏ›P³¬’PÚSÓ•]B,<û¤`Úðq>EÕ†b)»68 4[#IV}V @A¤µüòíàÞªÏ.HƒŸÛSóË©H’EÐ|Õg¥?7» ~Hmï±ýÄÃgÇ·^Ð’5+,J`sÙtLV%»øÄf@éâM[çòAe ,¥ ¶µDH™Þ#;….ZÖî’¼`ðºG^„&vœÐ*…Á×àXŒ¡\sæb/›H)ëŘ`ÕÕÙéEãôÔ|D, ø‘W}R¹ZÒKA¤D®ÚžÑƒöãä?‹øù–¤¸d°ú¥³¼šÊŠ¥›ªjJ_6U 4 ÛIér\ ÅI 3 ,}I\ š$™tQÎà†(l ›Aÿ¼9ï.¦{£pBŠ“!ÞXë62K2,9Êp¸zºŠF0àÏm#(õ)Ͳ„“êYÉÑÍ‚&eN¨wzXhbÅјœHMrnU•䀙4ujAS9)ëJáªYîKÖ•f.Òû«ºHQ½e]U3Ô¶ªf«ä·¤K^ü-É©é¥Á×mÍ‚³°8±ukƒŽËâ´–H‰ùH¿ÃoÙ–ç’ÁTJ šKM‘²b ‚Ƨ¾ªÊ€  .ðézî]ÄŽQѦ—tÖ@[pLÖ0öT¼e!jå´Â0k ;¿EøÁä¸FP¨Õ§»ú¯'HZ|=þæ#í>X¹ƒz¸fߦç¨T¬Ä­SÑ·FÉÓkU~z«ÂadŠnd¸Ò±A´«!¡È1rƒÓé ‚ªJ¾g?D±=FPl6¶ïÌŠF š; óQrÖÓš}¦F#:άÞR°âˆ6•l}T¤jøR)=è Z*î^BAòr[5&ÐçÊMRSEüÚVûn*l]dzu㢔íÇYã „eé%ß‹4€¹k¾—*wsÇR¸jš³Äeéê¬ÛD麺@£7ªùÔ«ªôÔ¢Vô,-Bõ°*=×»º9Õne×€ q§æn¥XOÔ©«Ö““R©’p‹Ü¿=+&RG¸)¤Þ»dÝ"Õ¯™¬ª*þävKñÇHþk¨Tô‰d¿-+hZÉþšå–0êS~CLAÀâYP`PW/Wò ÇS[¹’W¸þz´fÚƒ¾^TBÿ¡z/R6'Š[½ë)š aø¢‚÷Û½™* ¶ UõŠlÿHª)u]Tà~§_³t©l,Ä )Höó[2¥ìOý6’ýBH&E²ß@‘Â"ae­b‚A˜Ì@…(•Cþcª‹„5 tVJzÝ%{Ý%™×;_ná u2¡ÌŠ‹ëÂM‡2+ªý¸J!@zÆ@£‹Ü½.õ]Ùº‰´ƒ@n SñEvY—yîG&"åÙ2i=Ê‹ŒY“U­aÄ*“æ†FC*6lÜìçªßC},ÔþýEç$¼o¶¯;ñ€<ÃΠ’|¼@™xéô\%“ë3…P¹gÛ¿õ­P’ÿÑ‘0d½öè:œŠ‘'‘Ò]A AØ&-6ûoju5×rÜ-©È”>ºäC¼ä‘¦MÊ Ð^ "KúéÚµ¡°ó%ýòMŒ†qW' ¶ Ù×báô_é¶š¥ åD™3ÉR,Ò^l;öèQÙš,ˆêu²/ „+‡³ÕÍ”¶ Z úm9¥­îrI›©l —ÂdoùãÌ7=â]aÜe«nnw˜»AÙJ‚ú0„»d%‹¸  ÈÐn¡ˆÐâwÝ„ØÑ Ñw™‡ã¢›x¦v˜ | Ö‹¦ÇB Q×-)ÎyCˆÜÿ”n_Yhž|wQ¯™ûÃô(øøý~©Â…Õe÷¬‚”ݳ5âd]XPøÒC*ÂwâmS=ÁuPS ÉÜ‘Íæ…ЫJq¼08Ä ô¡oXY¶Ö‰èÃé´@ð´ÎÁ…ß"¦‘“Ù gÈ£NŽTÝK·µ!EÊq•ÿ¸¢"¯·N1„ L¨ñÄ"…bA7O¥Ã+6ÊÊ2§+ÂJ*ô--}—µQÚwÓ¬À+VÌÊrª0ÎÄè™.A¼t ¦ä#Û¦WÎüÓ…‹—ñPHGÚ³H(2-Û‚“ûHá!ì@!§5—]ú«ßAÖ¨F0óMWV†õ(_ÂÕï—yÉ—¨æepˆ½Qƒ N‡½Q»îlšA¸¥R+4í)[¥jIoT†ÿÖJã·j)XI×€a™ª¤—I(—Ú¡Vš”17&å²z@8ËR¡Õ\õ@*5z)Ъwª=Z5ûˆ÷¶².ð¢`€ýi™¹ ĵ%5feÛWnjÌjR(’ó^¦$†}/ß!çï"Ê,e „*³GÉÈðæKHJ7|ƒù®oaþ”¸èl©.ÄETêâ-žj÷‡%¿hþ¡äôK{ÿHPmÙ™~Û€ºÃ‘õ2,8J[ÔÑû‹Æôªr¿ºˆŒ« -qM:ËZøÝ.à‹;±cÕdz,…ïÒÁ³,…o"õ‘ Ù:æjœ)%ïc>Ç™RòÎjð3]†o^È™4%cŠ€öYÇì’3[Ç ”3ÉÐ^Ç,•3ü1“åpÕ"gu®•@ÔñX÷²ldÒΠ¬b ÌUŸqôá\9v†öPg+—~ÆÒ—SüíƒLn %õÎØ*Èïèd`å3¸g:æÝžÑdd`JÑ•ÓÆ±Ø×QqT ê8ܨÑ×ë<ÕåœN¢½ù}Q‡Öæ,@° ×Ù.åûcêÔéR¾?–<*cƒz¸~«±»]K˜hÔhº1C-p›z«DÖu±ÑiÚ]CŸ&ñßkIÑCá¦uõ¿/…¯wHG…Œ+uNÓè¸:M³ðõé4ŽŽk¶ÚQ'TÔ H|ûVÃzõ´Öª¤è¾™õðiÍ”;x%†©‹ßÿ6^5ðªä¶Y×D– ×J’GÒ/¨ÓHke]3ù…Yh(ÚÈ\W‰¢÷º‚ó’ûÀ'äªÀ<ø„ZB‹¢ QgA@ŠkíËS¤À{»û£´Þ… n¯ûpãr»§eI9üDŠn,FÒ•z6W[ L ½d|¯rY ùF&g…ÂB¢@ä@ÁrÞˆö)/„´7Rx”ºÝD«ËlÈ ¡ªéD®#/D3í¿@’iò<“~;AZü<ÙßùŒ&?Ê~á¶$?žlÒÌ„óÇ‹2IÙȱñª³ÄÇXý'rElôä%%@凮}¥êš§÷Üûaº5÷'™Æ·›ÝG„×LJ+ý½ ™-Ç„.ÒÄ)ƒÚ|ÂóØzX¨1þÜ'‡¨.ÒÜZ[¶„¶’ «f[ýמ³­ úRìQµ~í9K ,-•U´qskc¶šú©™sÜÔEž‘ Û")º‡n`ñ6O¦Ui}mbòõêõ•eë {’FKt×¾œ&²k&iR¯ú3I³¶<\@ÀKò›¾ç|LçÄu9Ö%6BN¿\º’îìÅ-͵Òb|8Ò±HVq¤½@O20{ΚÖ»zõÝ¢c{…¦peí9¹QãþEâ>TÆ¢n5ÄKLê÷¡^èáÚQSHëÞwl:¿H?$óãÝËÐE:ÑŒiÚ¹¶d´,x±×m€hOKsRÞs¿…ÙÆ¬ø#ÍÖ´–NU]$•>vv(X讚ôîücäꪥ­§êüºL3¼Ñ f¿¶d9‚‡üÛÕ.òÿÔÎS5}}+äthÝäN#G¥jbF€h&QÙðS°pBpr£^Uå ÑÍøÀ«â,ñ`»+ÏÏòø¹>]INù©±ˆ“f’Q<ÆãfhH-†Ñ &Ñ£òÞ@|ä¢ä}¥ ê"UÓËp€âoµ×™«¨&%Ï\*miyì$ gÝÊÙ*Æ=]K¿¥âÅÃc(ëa®‘«ºš’IÖçXÅ8Κè”Ãu’G’ãuX^àÐë ;š-Žê.†M"¹ªóÑ\pD—V 9,º @{þ8m"ê`g©>Ǿ@çhd¥´`芕th:YÜê£ÜI]‡;r Kq †ñÑtÍ-먓hº­K­ÎU­]ª#*çG}–†Ñ%ñ!Ç^fVuiR¼"˜g-XÄ^ð}§å°#>Àb/ÂA'-ßJµ²ÓRuøÁog¾PÒålQê*ySp·tÝxhw~åˆB‘… WÓéo¦îì‚»ëÒØ,Ê‹8Æp*Y"rv‰(Ã¥èZ [wf®ÎÌ®)Jˆ¢“”ÞceaÆöXÜgø¥E+c†¬í³îHÄ1—²þMÈqå£ÂžÚ*¹82"˜]哊õUUÜ-ÂjTE®\?¸ {Òù6ê6õÀäI”vn#`ÖgùçmÚ<­îðdS±Ð²0cÄ{›*M•g£h#ó#nõ©ºHÌõ5ýÏhO™k„¨Ñ\+Ó~M»©euã+O—@Ɔ&QM ×ÛËoÔàöŒŸoU垌ºc~jH<å¾—µ[, ®F¶†ň¨,®ºJÛ·ÝO¹ªPV—苃~{ãé–aûà)Š*d6#GÈÕULuD¹´2T\€\!•Û¥í²8¤U*õ롉E{/—ŠƒJÚQÞ3ɳ ¨E$»ŸJòK¤bƉzô®|Žjc·=å"#…æ;^JAóE|å£^2u¨1­é1Tcœ–ù=ƒQë1§•ùNï!‚TSV9€£¥|9[AàIð5!ú–ád)÷=éLÀÔuî9;|ôËÄ>-8çd†ïê5­v‰M¾æÍ\¬JÿÎuúa©0ë–ÌŪ°ý–ÌÅÒU ôœ‹%>Tœ‹¥ÿ_ 3H?XJ;’¼ü©ž¹Ø¥ŸÊãZÜã¤ÌD$”)*ašÈ9W7“Ðø¥•?M™Ð¼Ðü½U»r%z|eêVïy¥ÈQ*~OæÈU?x]Ò„¼ú¦H‡–“Í^ S‰ô!µG¡Z‰Ëá#ÍOµ’œyc¸¸I  X‰•…‚P@±Ÿ"±-åÕÐ3kb_©•LÏú!àM”™ˆÄë ‘snAœXÔ[ÑG8ÜV†J‹ZBü‹G·årì¼´].•/€hDn4C#Fƒè°P1tdÌ9á ®¥x~ðG¦ôhDþØ”1ª9ûihÕá^q$S¥Kbj&ãÎ¥²ZõFÊÙ©þ÷ZÔædÕ„0ÏHúN)ÞÜœÑ$ðC SK¬eÂŒš)@´FÉ|2„ƒLÎAârCºh²óµæùœƒEJJÞEÑ¢á9àn‰ sŸs1$-èb:…`’vV—ÕúÍl_0¦ çÔvrŸ¨"m&ô é«©ßܬ—º¨ZszØ¿…ÄÙ@DØ+L‰‡n”h›…U5­§ùzH"÷Í. …*1 =ÛCæ ôUIú|=YöeIvk™Y^•|l÷†œ…·DL,<¤†8×H„Â[É–7+¾Ä“UÀ’Õ—Ja@ð­F–@U `f~^•ÖèævÖÆ™c Ÿ+ém óÂ+½1CìlX%["ÍÕA º±TjS±Æ‡¨š²]ë `tkb yñ@üÃ$ø‡À›ØÊ•Žk!×–‘%DúÊsŠÉ‘ Šût×^ˆ1Ú±X¥ˆÇÙ GâœIe6±Í•Œ_M‡s»ùèD䀔ë"ÄŒ… Ø¼ Mä*&G‹aÈN$.ZÕ„Ç&]Â&^lcýKí]ŽÉÍÖG@ çµ²ÔJ>ËÍJÉ"nk¹ç‹i qC½¢U ]‰Ø,J˜Ã»ÑÓ«­k²(‰þTZlF0\vÉ7â>%U#Fuˆ:ŠëËùˆ¾S×¼»»çU‡€¤ºyƒ$4&Å •»u âÀªê<”)·<^rˆ;©*íáwZ „*¦Z‹=@t‘d OÝ®¹úd–xï¡ÈY“ ûÛTí‚Ï`eOìñÎp¯îjöDOŸ ¡+Y÷Ô ät#wÞ*UêtýR ÖÎsS³¯ôês]ŸÓé B¨{¬è­&B³æ«ºó°¼kwáϰ¼kÏ ˻÷äñÈ»Îw E(À>…xó‡ (¾¯Õ]“ã?¤îš•x¤f-…[Ö•"Б¦>¸D!†ë£º¾ó¤Â®®€¦+Dx["å.9ƒ½¬¬u![2Ô}]ëBvH»`4™P nYgC™ã’¥.ŒïôZòXÓkÍ’ßzmùäwöškÉ 7Ph!»Ö…µA®!@¨ÛiO"dRdyûßûH†9k— î\³ªfò´šjÎ@ðâ{Æ?ïK:+¢º½E@Ú¾µ¬JãêÒÓ*<å¢yj—¥q½ì-ÿ‰Ù¦g£Édch^k6R•WoY9yJ-ín’ð-Ç—ÊÇz»^Í'E';ÕqP"®÷ê¥u*ß A7"C@È”ÿQìÖûô1Ië̉•¯÷åíÊè `žet±ï6Fñ•ƯWkè¥X:t¾‘»º¹h0j¦‘‚Âì.‚ã> Û5™õ‚›úîî$Œ„¼‹àH xuG"vúÆ9ÕûUÐEòqò‹ò,²õî:‡ŸKîÖЋMB¹þÜŸ©š²Ï¦]Ó”ÇÅì 9g fŸY3å;a¿ð,Ü ÿ–…›ð‡˜:¶¡Ü_Âöæ@`þ0OVÔÊë!î׌àÛ˜xÉÒÒxû°ü4î¼ZVÔÊÈauÎÅ Jw5ª<ÖpÒÂA)\v^qµžÊb[6¤òÂz,z^…‡{*‹eÙ°s_¥¢Û¾‹·‹‹Yëð‹Piç’‰Céóª8„ï„«Âeâ``Vü†£EË«Ð)öʆ­ÃÞî€K¶æà¹øõã¸dô°óp¤ ä~J–YËúáÔ,³FPªŸ–!3€d]sš„£ÀÙ2Œ8ÓûŽÅ¬d7Ÿ++¦Ã²däUø©ó*³†§ÄÉSŽs–²; Æ%߉›Õʼnä™fQ\6¬Yfÿ‰Œl †ÒÃôÅ¥ÖFVÖœËÒâ®l #;k­œWÑyœ‡Ã–æ):¯)ï¢s ÕÅ›‰´,:7òTä“V<Êø‡fѹ‘ÿ·ÿN’H^T“Íÿ»TVø‹B,1Ø‘×aü…Ðì– þ¿ÈmýŠþB4“ Þà„¢—»XrKìãܺ‹S™["@R4gôŸ"‘xQ— H»h1D >°å@bF<ý@HßÅ |¡tucý {ävRtã&"ÎC»8«ÒE~ñ¢æ¤:m75‘ú"§£w mk!„ºÅ÷.ÖÎ’BS R7b½{’2š]ïíøú¦“,i8U½v±NѤEïº×DÓ¬*gÅ7cHh˜Ue?´=̪’¹e¨`¾šnP[UÞHËkP=8Gq ´ò{SqK\ï{Rq«ŠUxRq˽cÒ™Ž1ôc´Ík¦A(‰ý¤ØúÎü„##Ί¹!Y¦?‘ú]fÍÕtQà̹4”÷]Ó'BR8›I„ì)pËšW:Y"…+†â€ŒGâíiÀ5¾¦€X).D ÕÛ¤&ˆz—çº $®)›:é@ q;~,W·ìfª €w5aÕ”™ÖãÇ2«f«ò7†£)bí¥f_5ÑPbΪi¡ý÷ƒOò€ã-3$Á NÂgšªÇCüᚆ|áOÊjHòyE¡Ú ‚+}ÅMJprl7ìZAÊ•|`™ÉÖ*¦¬šî»™?xq„75;^Ìâ]4’ŒyE,fKVвznJVÂÞù!G/Ȧ˙%+ÚµU-AÕÞD( ³Ò Ÿþ÷‹ž¥’²›Gt)ÌÞÆêg¸}(ëxð½u&…Þ'ª'"7Ç{²¾‘‚óÙ"Ì‹ŸÅB4PIþ¢)ÆØ)b:P´’ ÿ¥´ÔMõK“±B₾÷«º`¹VõXfùU5Xép1•³,«Qq\›5Wuңݤqg5E“4ÕWXŽV%¨§¤-vOGÍ,UvèCK 6I«âtK“JæU&ú“¢ÊW¢X."´ìvËÕ4v' 7o¯¡Ì‹î»Aò*WÓèC.ꨭ´¢[”q"©cדf-OÔŒMÕûH|v‘ýó)Qž†fß©à¼(8…j#^5)øT$-—úø‰ä¡Z'>ò¤ Y4œyNVîØ;ŠY¹ã¢ª@\Š%µãÙ³KjǓǚµì7±Hú¢!G?VT†1o¸È8zŠÇ,K½–Íd¢å‚kûC(…ˆ…³5~VŠÃ  e“Š@ÐU²J%¨õ÷«’n¹DÔçK¨vc-Y–‹•¥§ Cìï8Ú#|íÇ¡ Žu9Õå»|; ¯®*µå%ª‚/îóÁgX)ÆL^&¦«8¬f È‹?ˆúzWåMÖV.2h_õ—‹Ô]Ôhjx‚X @¢Ø¤B£øSwñk¦„²ôãÑÅØzÒ˜DOC‰¿W)jTœ£D<ËUQ”®"WuÚ¦q6»?oÓîfÚè8*âXŸJtÑ1Õ~”2¢ökEÕ³vDÉ~%¢aOAÐWðªÒ+f´C0 ^¡ø à›§aÈà›W!‚­ßp5Ór‰èð"T)Ë x¾VU¡Å0™!óñòS@{eÕµ]¢Ëk’SC š9 OËHc 8ÿ¹¹ÌÛ–‡–ß©øS¹bt˜K,Õ·,d+]ðаèH%¤Ñt'ÎßS-¿ŠvMYQ¿xºaÕ}5‚ñù¸#ºŠ´§Á·ð¨ð—½Ü¥¯Ùz /5LV·.»íCàe ¿¡ÅÈê.5…2I´F¯i:2¸ìÖˇÀ‰›ŽtI:(×™”N@N/—rÛ1ædõÂ-Ääps ¿^KïôC®örÌÖ.˜8Û´ ëÝ€èqŽFqK;uמ#ï‰cêzÜE‹qië*–¯–ëG5v-Ý}d{ó¨bÈçΘÖÚEÞ¡"I=ö9;R$q¦=W"êÛŠ$fü…ºÏ¢4ˆ=ÂD<^‡£ Vëôko­DÕ“×d9.„J¶î3¹%¨éí#fôªéí³eZΣÀ ûÈšæ>Ûâ$ÖU±ÉÒ£½ªJCˆKâ$%M¾¸ˆCåE&_t· „ÿ<}I—ÔÅÈöbÁÍ_ô›¢ì‹©´àôQLÀ ©!Ý +.Æ»‘›˜ b‹='`ß»xƒb­EoàÿÍÞ»ôZ—dIóïWÜ™>üý¨ ñZ€„ È*@tS]ƒþù}–=Ö>7‰9u”Ê̸ûì§ûr÷å¶Ì~}Œ¶V±ŸD¨dm©óôD¨ûsò)¥~g•8Y÷§&±¾»|u&¹¡(´FÝ5C/´¿f£0‘®N±ˆÙî™Ñ•N9©‰H8Iµž¼±“Ü&—kÈýŠã&d›f"`:®å´h‚”D Ùx,øzkÑ,¢'O#žÒ¥á3ÄV€W—örØîv)r¼=ß„°€@‡íT›!Mq*—@AcnPrK6§ƒFVÑq_¸[DÏ4¸.ŒŽ+·®ÁÙ.®È î´&*•ÅQä:8÷p •3!zé6¸­¥ÖÒ‰ÛæJB¡3¬¡ä7(¦ð¡‚ ‰Œc¥ s¥’ÒëÑæôue}MÌ6±k¢Ù³Ûq8‡Þ¥­nQh8Mƒï†$¦'%Ûéiæ*ßhº”|€±O´‚q·JQ ¢Oõêa²Ýòc‘¬ðÜlF¥Qk˜*÷ž-»YÈþ " w²QŽÏ&‡voBöT¿šìymxŽS©?[š$Sjo{¡2js[!W‚J­:?MLŒ–‘o!øÚŒ`¬Ç´’‡=ÇGNB -+W>ydÇ}ÊOÅ®(@Ó2¶~ïTãŽkú½3—Œ+,# ¬Þ6±78ŽÀ¹ txí¤N²×,NÆOº‡z¯jdÃ>}ÁS•ìHåH’x¾©ãJãÍIݾt7Á¢a§x™i÷Yœ‡ „’Ä9S£=|ˆÛIô¹q µï|È*!¿^ÊÏ>Å™ay%k 8‰V›"Ï@°×wÓgµpK)};˜sSNº€ò;hÈâ]dôÇ €¬;ÙÍB â~r‚U¸ù=Ž3oS…Þ¡™ÍCÝCÈjû†™½DÑ£°þp‹èàÍ^ vsËfJJ`ØÑ0ˆ¯gÉÆTzHƒÛ«žckŠóŽã{@j}Û«>˜-¯úAËqH—óŽÓb{Ét¾·»G|Ö˜„&:‡úA£r覷n$Ö¡¡¶Þ›‘ÈP –¡d»ÎÁº³Pu_‰DÛéHÌCþ4#P·€<ªÙëfá… úrŒ,ªÔsfœ VWB!,k6šªC¿ëÄ$ú„jþFà›¥jƒÌ<ˆï¯fdPŸûo4鬭Ôû„ðÿÑ CÓæwºý*\ûtÂcÐâ¶­ÞýTOŸ$K#!±¾ipàÄJØ"Œk*è†Ä&Â\am#q¥áuEhi ywœÆbƒìG8=Ü<æÃãѯaÜ1ì<`S‰æ¿!oÛ è…r¦ ¯Øæ@ 1oƒ ´ÛðÔ!%.u胋xi¬<Æ~~ð*¿š^e¸aa‚!!Ž‘â  )5W a„Eˆœ ¢°ÉCÐÚÚÊ'çŠyØkpõÃ’~Áµ{>¹‘–_œ–#½žA´i%Ÿœvv#M‘Ä“ûAÑ\˜·œùó¬ Iâ'7â ä Î1µ}z>¸€§í p9op©2™N \…Ô9Qœ¾§åùßóQ’jò'Íþê×c»ó†˜cáâ0…Á@0qj"`ŸæÆøƒXŒ oš:'’«ßJmÓàÝù{D  õTzœukšÛï3ë.$@yÊ]Èò|›uhx¾Í~ÔÌ1'¬d^É!¹‘>쬙ǀª–r`€ªvº'; »=7l"5¯ îÚñÜ¡R×L9˹~ÒҀ̖®Ê­ˆÌfƒ©˜è‚º¶s¶Ý*©kÖ·Óª¢çî[à©ít`o¬È BÞ〕¶]ð4IõRrI¦ä®ö!o”Ð :`®q@dîìW®Œ¢£vêÄ —xm;·FZëÚ?ó?ïBæ Çqä1äRŽß-Ë‚9é% ‚]é% (ê}¹Ϋ» dzU&ÒzŒÏ¹œÔΙ§Y‹ôO7Gè¿ôÜ+ äɹ7Ðb3™¦@ÐþÌF‚Ö6sú­Ek§î‡’c]Òˆ¥n#¢É!%ŒA—UŽB:bÁ±UþKKæàậ‹5?Ì<d‹Ï{ó< ŽÌ«H³,XÀîW\°÷ñô+]¿X C-ó;òJ®HÉš’S­æ|W£dÏ­–@¦ÑùÐZûÍGà^}«o¦%@îiVÞéÏÝJHf€ ½ò(M´n?þ^ò„LÓ!â¯ÀŒI¶G°y†¹ÃÓ,ÉBG*&žÈDõÌÍ¢õ}Gæôvç¶ÄØÏÎA4hójHÝŒý\Òòš@Ê“<‡_•œ‘A¨¢ð«…vêêÍß'fÉ[o-OlÎ~u íæìWŸØœý’'fþ«×ìeÅïQ¹ ¦/í¾@å;Ìzum àÁ2 „3ã^sÜê˜MG©Å·^¥Ak¬žÛGš‘¢6‡}‘Tší“{n¨9ߊY/O^ìÞZV±;âì¢ÑM÷§d“¢zEñ™éÍ´œfûFŒÒ¦RVŠ: 1RYì)š•EÄYfºEùÓÌѹP¡ã,³8êÅB(‘.ë,ý¹þ5ŽT®‚’±kå'‡Ûõz`²BuHÎ^7/e6Siç(^ÒÖ woQà¤AG*•íZ/Çùì@”.a$B1•ÓדåKÕç æ_{6x"íÙg\T{ö”ò/uEVG°BLYpê FY&ò7ÿÖ >*<3ç/Áj5íð1F£¢Í ùËbÔc‰B¸É¨2n@Åâ;nJ•*êë4ØÔt Dùx)FQÞHÕÅûø‰X–r?õxR€4(”ÛÚÅ\3*VŸ=†(Û! üÞ륈ê…$D \Q“Û 9„¨hÔ¬iÒ22« ]m/Â#;Ž ÖmvÜmŽ*,¹BÔö‚IÄBÍ‘>ÂÊäŽÔc›ßËauåÊ\¨6Ê¢ht&2 K?WáŸê:‹Q‹:| ì—IN±Ó­e¡ÿIén -wÿ °Šßv£–öä."K„³ã.•+O×ác+d°Hw'Âáí/G©î§ØÈ|/‚Ö™C¨6Gi‰°D¸{Ú²¶J„›‚7'’jîÕvU5DPÿ9M‚AÏ$R9( ¿~Ð-‡‘ÓŠŠ¥FDu8 ¯lƒ@H«ÞÀ…YÇ  y, ºý1zjnòº´*©ò1j¢yI¨¹E[ˆžfªÂgáëMáuq𠬚(]³(LÜAËIÝT˜áËLo¨¹ -$hÏS©F@c-‡a õåb·7+à Ìšò·(œÝD\'ùìÙ²þ®êÌ5ëï•$gÍú;¤yÞ(ìëú‘¦BÓW¥ÏtÑ2³`çõx"TL³o7§nVÑh¹’q¾®®åëM•xS½{ƒK CËX¿ä ™úU—µ3k[Ž©¨Â±‰^›æ»¥ ˆËA5޾—ƒ2ß^ÇÎrÐCyô±³ÔrÚÎÙ¹¬4§ò´`¯ŠêT*\Û›k7«iO—©ûOU›7Y¤.uÆò <7ˆâY¾›é"ðÆÉb ^G .@jHY¦ÈÔ•`dÅ­æ‡Ãy‰( –”6ÿ¬©Ð­÷Û›¾¡ µd½£™2ÚÞˆ‚ek«Mwëh{šâ²çPWõô–ÞµEÆw·Œ¶g2»[F»½Õ\C|v¤lÁàÞgMÒDTsóR©0ÔÅFª îº=RéQ'ލ0²¢_µäÀú–bDØufмUö }?²Z_ºJ–û¦6] rHñ·kó¥Ì;LxŠ:{X]Œ÷R| ÏÃÖš.ίã©ÅçÂîQŸß$ÏB¡þª‚_Ã[jÖ‡T€µïÇæNíû•ú©}Oà¤ô½BÝ”H æo©ÍÔn'@Ë‘B£øw—F ÈD‘@×d!×%¶@¼GÀ?èš Îîf2€Šü×–;‡m¹:"/¹ÚßûÓ@jL S¥¿ÀÞÒ§ï©°¥¢kRNh’@èã[k}G$‰M¹g K> ‘«§)ô¿CŽ®­p¸¯<0¡tÝ7D‚¸iµ°D°ßlT0_xC(¼ùŽpÿ6Žüû,Šü¾!Ò4ù äá™üŽS<èSÕöÆ3yC®EH¥…ܘ=¿#ÝB›“ëš$á|CnŠl¸V§utY’Wp‚~ N±Ýªj§¤ 9Q©c´´b—ä×h)Þ–H³ŽÓd©¸O†èQíATR¤‘ºX 7¯ÝyW;/ÞDÍZùrê }käÅAµk×bYƒ¤0k¬‰xg· çJT¼ž:@“»¯AI³ö©A[K™µIbÛ£²n^Ÿo*kàæõ4T—X[P欲&n^W!‘JæUÖ¸b69Ù{ô¡<)‹¬‰­7ZЬ‰œ§¸yöÈHAº Ö;Å.9ûLv½…K“€€¬ý’ŠàÃ1ª»!«nPN\xÈ75ý¸Aز¤>äXL%šRƒámoÔÕ ~eÍCj½uúGj¼A ¸Çã¡Ô‚R¯o•vËGvèì¢ÞB¡­cú‚jÝž*uŠ;e²²ƒ¢¸UÕ|Y‚Qw!í·àÊL‘H§›²K(P¿Rvj °<¢¸Tö² \}¨€å¦¨pg]ê` ¬Vo “£œF‰Â󦜤‰ÙFX¤cIa >E!%…9íŒò mvÚµ  èA5’‹†R\yz—‹Ì—ÅG’l—%U5•I%L@å‹hÔV˜­ZŸ°~Ý–±QmaúOjH£zjA-’w !EOÙÞåªQ¤ÓÞåªQOØ–õ·'Ë [ªUs501úÆM› ²Y~{ð»µ›Ä œ¬6ƒ ÷qEšÔõü÷¨óv.d¾ÑðzOndëÀ Óe¸õÈM…À}z¬nt Ÿò¼ò¨Ì3À»Ýd¿ºÙ“Šì—¥Tý¦$ûQ½aIIö£zÃâ@£ôI¡4O¨.ˆlo’ìø®R„;¾ÁÈ~Zé|6GŠW³, ‘jiÜr)“u” Ì•SÛiòÔv¹rõ¨õVÉ&€Î¢N¾âŠ¹Ùœž«U2ð…¹•,¡•HDûÉŠÝÈ.ª7íþt•$ï9‡5r%í E9Éÿ"¡ßÂ$Á¤V®” ÎíîT6[ìLyëB¡çùÏòóž/Òç¼½|…ITp®â†RT׸ªŠ æjéè©0WÎñ Mþ¦?@µV.› ‡Ð@ÔÉ ̹¦g!Ò ‚a ˆ‚a)º÷ðB…ŒàÌÈ`ærQ M"u+\.3LRHµ¢1o"Àúõmã¬KI¿¹[^ˆ–=3É\àsfMy8qLU¢Wmvºµ\¦Ñ‚úÂüÜeå*ÜmBrÈ”`=¼DPð¼M^ G–Îßëòõ“Ö!._·ŸÕU}ç¾ièQØÚ¬a7œSm˜BâÌSý*è½Çe¢á»Ó–y$¡"ü±qæÅÖ*”dRÃéUˆðOÕ©‹ ‚ZØÔ7ß”)Ÿô ¢1Td˜#ÿ™`«C>S »Á‹97o†tãG¥árÓJeå(ªÃáªC9 +‘^úXA"Bž9›Ôì™ìÎ@hSj",}ÐlîÐ;ßKj΂¨ŠZ»_é´%j!ï˜=)F‘–RÚäzÔ2lS4¯wÂme\pm+s{mVEÌë*¢«m8hõèéø°¬StÁŠ·üái7"Ç$ƒŠ7NÂÕ R†´þ•VP”CéþkSr©xgçÒÈÒ™¹-²¬ºR0eúâÜæ\Å¥¶³ZId¶åÕ*fc„-Q¼{ˆäZdöµªQÒïÏjq)û„P«E«ÂÉkP#G8Öñ»–‹!‹×ærº4Š ¸rÑ%lƨÓÝSçäSWë˜jJM8šuéè¨ÉNi1Öé&;©S½j- ÌÕté›dÆíEJl ÂèÌE×rµ×’¹ªÎ×B­ol‡Üø&*t™ï êP5R„èÄ27ÏÿŠ´°è?„é¾Gð(Ê¥ Õ yëy7,)X­ç3À,» ÀaoEÓ@UDg4¾ Ÿ@¯¯³B#¸3z}±dµ•&Ì*›Æ&›/´í4 \zL'y/}~ß´¨Â° QÓïCim‚p:¼B4t°¹ß„°Â1¦›ˆuVf§%R„è•Ò&2\z¥~ÈWÌêu²ÎY(ˆ:}wBüXøZB¯Û¼¯uê1eý Ì0¥b¦U†™AîQti”[X=c¯Œ7ƒî£6[IúxTÕ. M¡¼¦xSÉ D³ŽJÊ~ +OA ÒcUªHŽ ÏL–ài˜¬ƒÑe¤á_%fj·ÆÊ¡teÙS PèRÑ CTô²Ñ󆹄Yvt€Õê% )[‘ªdÊß…?+šÄâjX¸¢E ðÃøõœ¯7]½@ª”ö4ë(SµÜe)›‚}v à$ÙСò›×>B€—õ«@®®M® ä0»Šš•ι´œnØíR ïÚ–ùJhr¤“=µž!p(sâKMài]±s%‘ªˆ‡JÊPN” ó¥È¨QÍHýz“[aQHíWZŽ#˜-oT‡-9Æ‚e¿03¯Ç§\£í’0~ W¢WÁúX zKcL–¼†Ù:F¸e: Ù!ÎJ éÄ.œ•­û#èjçwhû9ü£òW?’mù„ètcÏ”.ý2«â¨'¡³NÕƒú¡H*µ4U ­e¥‘#i·@dø=5z,óÄÏ”Xäò¶çQéDލ\*ˆwzwª±z$uÏЄq¥õùØzó+ÏXú˯s€D ù^E©!}×e®ÍCa»ÉŽ®‡pQçR¯_Rf9CóÁ¬®>T!…špFÂ)`-†€è½«àp­´UïRì^òûz!ë§NÌÙÀQÑdPõ!(' D‡¨- …œn½a³¢NW[Rú<Ò4[˳¢ÓK^INö±7¼‹»‹êq}æÃ¹ŽGzdp ?,}Óy>ªxSrÁPºM;ªä^6Æy!ô“6¥NÓ0;ÝÕ!IõôÈ*¾i·«£³D ÀÉä•*èœøPõä…Ǫ…¼µ<Õ:Þ¼>Õ:Þf6Ÿjo—ƒœjoÏOµŽws­Öñn&«u¼›æ½§ZÇ»iÒsªu¼=ó<Å:Þ.h;Ð?ç –Í¦NÅO +/¥Ð[Ô­¾!EªÑ¬ K†óOi«¦‚ø)ªì ^$â½ÍaÜMúް‘X&pߣù [ ²k`ØÌ,|C¨´¶«ÿ¶t}¾#e}G¸ïþH¥o¦K¾!š`½!ÜI~C·€¿!ßÿd1} ¢Ï‰Oò—HòI~Ï&ùæ¬Sƒ& cK•bºapJ2&¯È¸o)ˆÞuÅ›C8=xG(ˆÜ;î!½ë ¡èUM¡—#5­wuÐ(–Qâ‹«£¢Áý ²| û½ÍHäêÁ™?ù†l$MÇAÚz97»Y¤àjµ”@tEPƒk†Íªª–ša³Šâ^]óå@P-‚sªX´5ã¨ôüj5i6¢P# ÀÊœÖJ³Ž@N†»¢c4¡‘2L­–æ ‰'¯jYkuéëajˆc­XŠÕ’ºGJ(´¶©¯žG¥ãñ Ùê²Ûc@Ž«G›¸é5ã¨Æª@º84ìšA³W¬Ô ‘ëÂ@{g1BÍh×E‰­Ý£/¸|©ot$Yˆf¢…žµºC{ä¯\™jvÒy5©ûpÍ äæ¼…'ÕUË15'°:5˜zՊ⇢²h 6D1⽜P—*NžMÒ4¡ÙäGVëKƬ ¼·`9¸j ¢”a¶‡óZ/*&„üNfÒ©ÿ¢ñ˜ë/ j#³ëC©¬ÉóÓZm£9ì!Òt©¥ï´sž+þ8Ë×€ˆ?^-þsh_ðB¼1sèyDïx‰ÝZ­%q¤ù\«7o<ݯÕ5 GjÒè•®¡{¶ÉY"Œ§PW,?âvL¦ÃœâæÚÙà %Í_–jšv…Ž”·k«¹êªà>¶ú¶è §ŽØQýt ^u‰ÉÚZ®ºÔšm]bÉzuD-ùX ]Y§„»*• ( XoÖ:‹E*ÞhëŽR§0n5×'‚r˜6ryν£@V®™—~¥àqThݵ£úäÚ¬«Õyzð#>xª·#BxSúåЀç*¹hÛM鈸ڶ;çqµíÌBˆ¸šZrÎfTÓ™ðØDÔLT2\ƒè`­­YP"r+üxW⪃˜Öˆœ ¾]—îã¡}=”R?µ;ù,›ƒœS*" [¨ôªÌ·v‹‘E&Š^ÕG²ŠÞæ„~!½³vKzG¬-"΋™ü=œþ+&Ûp+’¿íüI9²¿§–Åìo+ F¶ìo‹mÞjö·•]#iHö·s+\±ƒ•x¯fçÎ^5ûÛœâÈaš“­   p“ý¹Ñ*ê¹€U5>ÝY’+¿Úsó¯j.2l’w$Þøý‘Ïí*%ÐÇQ…í¨™9Ö 2Ú“9>*lpæXôïÔ‡Œ”ôT)†3Ǫ¶=3Çš°4SÈ]DøPí¨ Áõ&‘SÇjb òIäúãuáVÐÕéÔ›»TóÆtœPiBè¡Í,M@o~hÇýèfŽ –T)ÈÕí]É–]7.9I‡†ÛP] gÛ¡ü¦R(U{@N•\År³D]ïajcÙ „ßfÖ ,ý pçC¸Ù‹'+jR(/ëpÚÃ*VõN›Y(¦šŸéSEH0+„Ü,/‚bÔª?³–¯ÔŠ€([Â+}8Ά eÁ–UV8ïÌGàT¢­éb7Õn…‚¡ßÍ¥¦ÔÚ~UŽ…0¢‹ÝŠîï)½d![(.ºô’q¸ØM’…û©½¤v~h9º°Ž…m÷,v#•$D"}-n)¦d–@âj7àíZ1V‚†^åüV»*—=w-0îëÆ?iæj™"ýOZ@¤¢&ÄèS³Ìss€8ÙÑ&7‚B»Sýhnépfù%}0BT 6"Ȇ*ŽOÖ XXôÐà¾Ùã ÔROŠ‹Z?¡aª ´è‹§úž‹^ͼÜÂeP»ÍßsU„6s‚¡çfåÍâvH£í„<Ä®êX ÍW½÷Eq’PŠÕ{_ îq8[}ø¨læÐ U%€jÛ»Mó€Pþ¶æÛ˜”ù5_d~¹S(2Évõ{!¬›éej`^t:í4¹û’—‹y«ü€^t9.ÐX´ã ±`àEjZ¯9‚±×A†XÁ~êC׬ýZdÑôšE% ©^ t‰qVE½Ûëñu£ GA[YÝ‘&mÐ_ÖçÛÔsì5»ÍfíT(;ëóí*9èâZb@ZEeŒ†4ÈA[gú…4ÊA·îøº¹UÕÍÕú6ݯ7sø{9ð*Z–TR]zÚ¾c2žBr»æ¯Ðtz΢è„Ä·LÂNƒ¼žUÊ›>W!î[žÔŒîYºµI¬íq¿eš †>¹šÁžÒ'ß xlóãOÚD–®¨ô(n{I¾8üo öQÝ…÷fY­²Iåv5ÛM¢hȽ«lusºÞ©©,rý–!,sì”G¯€‘Å‘ Æ­×µ©xÂö ö›fg}æè­³z^¸É‰ïöI‚þ;³Äj3wÐÍ‹‚.mÒØ_˜ÑJÁœ°©RyÕŸìò3»Â‘ÅÊ®pä8±ª{¯$$¶ ä1è+ç0Gî«;¼jôunå‘WÄÊ%Æ¡YÄZˆŽÌ"L1‚.%{r\_׋SoÌ«BgˆêÒ¼Ûún2É 3„Y@ Y¿‡ƒÀ‘WDô‘mêä;'(‡Þ{a€‘Ü#Âë>4¾h ༹08¬¦ ƒM¡ì$NµNÁaV6|>Fƒprh8T9 ?8-'bØî ÄפЬnÜÓÁÙùàF²*øÈ¨"(y ÂÍ-î¬GÖ·þÜßT~8Ã[˜1}F>¸ì.îô˜˜ÈòˆUHÌrÞžœˆ4j¥d[Ç$:ÌkÊüð<7€§é xZ¾€á —ÿ½·ŽÉ&úѺÙ 5öûÑ¡]9ùõAŽë£óW‰ä¯ÞÎýÈ©øLA]q¼ÔWëüíABPeó (Þœëý #o-në½däí 1CWʃŒ¼¤õv_æsÐ!‹ñí #o]‰U>yÂjV>(‘·ƒªØ0ÏAFÞ.7T#-I/3‘›/“þÏo=ÈsÐ!ßàí Dò Ð6÷ûA’…¤Þoô yСUùÛA’ÅN"^]ô ÏAgSÄà9(‘›o|½ßµÿÌ]ªö~ăÜß7’ì:‰d·øÅ’^o¶ÔxþFªV÷ƒêZïGÔµ¾ýkH«·÷#„¼Ôiûv÷ƒ#-ô¤¦ƒö÷Þ:Úñ{š”ø;hZôÿ9hqBôvÐòé9h³„æí !zÙ 0~#Q¯ß%oGUè¼%äÛQ‘zõû”À}y{Ÿ>êæD¤l : î¿m:ƒôæ ½\"•@¤ÊÔ,Ò–U¿š’¶DRˆE£(V¿Ò‰‘ƒ BæéHYpT5c×íljÙ)óY*©…^½ K „Zh›DÔ°;ôŠ€¨Õö¥wQ-ØJa ÎíCïÂlß–¸%¡ÍÎgXöÀ S¥Ã“R·,4:?…¡¶M›bÐb¡ÏöHàƒðdPj.•–*}‡«§2 è™öägGQC¯ÆÌ¬EÓ úâ»1B&kõWx¦ôõØU'®þ¾ ˆ¾/ÍÏûL_åE*¹k¬Ã3W—°/ª|Ñ礵Û€šÕCÔLí–ô- ¶!à<ˆFþ}&•] @š~Q9¥0k~Q,€š#™r@èUù{è´©ßM´C úPs3ègÁšúSÒüSñ½vŧ(Y“ˆZ%ïè«Lì.õ™Å‹†é@tÆ[]d·¨:„Ñ%jÇñQjµ›ÀÔ4)«bdÆDŸ€ª¢}f=P”Äct)f|-*qôس]´€ÈòdLÎõh$æL=Rr<Áô«È‡7ÄNºòæßUI§¡ LŒÂ¾]<ñƒBBþT¶"G‡~dK·¹È@äß@W—>øƒó È3—P”Àç/i Ò'ÿ: 8@˜©xÛ¡¥¨“4Æ«âIÁbï!À(#Ž<9ùEu$é{È6*ö4$fTѤCýQï³a;ˆjö;4#ùÙ@ä‘A²r¡IµXšÈh%<„H[†N€ºÔ‹{H\ªóÇœ…€gÃŽ[‡Ëý1Ò€T÷žÆÕ ,ãyÞŠþëôDj%¢Va•ÑC¸S¥"DNI•ë™qrø "”ÒÁeÕC#T'Ɔ.{Ja¸‡Ô¨ÂQå g8Ë÷ªÄüáEx z]ÛÛ}°$Hრµð9½s­ˆ|¡ ª¶_À΢XX@èì¡»j32T|±ÙÂHj­6#[¼Ö¾néeònW,.&À{èÀj*ƒït[%4Ÿ¢Ó‡ž¬ÝȰ5 ÄWoh:ÛZ,¡.5pš™fdE—J£&¦ô{¨ÛÊ/zYK†ñG¼ ƒl3õ'Ë>€È0¤:0K5€¶‰Ø!cV(öÚ!%'=D~åÀTL6ÈëlJë¤A:yìÂ2¯CNXOÀW(Û ´i 6È+xì¥ ý³A $3à‘²¡£Q5‘ÞKš»N(CEYnž÷è!³| ´J@Žãô2ì¡Í,ëGúÂ?Ò4ïòrñMæI7ƘáxÌÑ¡¦k‡&úÅ´ ]èÚ¡ãΕNŒŸ~Îtb„J^k91rsˆ™þ’=¤¯õÈ[qdº~+äûð‚9d!kºxlR¤‡·Ý.Q%D6­,ë¡ßͯÈ-Èß×.=Ž_. ;û®˜¬ôîc¤m3 Ü´Dp푆°»êÚÃjW~ý1ìrMÁ‚ÚæœJ†ãÅý¸T§@då»@ïꡚ>Aœ®Á H¼³€àÓt–Éýâî;Ü¡:)Í^û±Jº2w°Œ¿A&kÔDÐjúÈgD^£‡JýNíµ[˜D¾·}ôŒôÆí¡/É¥hÕP˜¬¤Âfr&qâvmWŒ-ûJüü …cèÒ4ÿí±¯Ô…`Cª‡Ä¿Lo'çܰà«c7ÒADk®›,!íÃLº_"r¢ñ› Ùà•@usœ`a÷p;Ðë›"ORr"z˜&Èù•NÎ=Œä:+zÝö¨§àÅöÉ }86Øs¹p¤©Ó·¤lôp~PŸ—ƒlM£×¥uìãñMŒÃ¦W]ÃHÅo\ J;mj÷d”ôaÒL¹ñyKšº’‡òBfeJ.Ø©+h‹½?åu°A6›‚HÚç+RAŒÕ(IrpyvAßlÚmCÖ,¡y¿kR¦O÷渚j´'iÈ=Ƚ~éœv#X]3ø8MuT§ MbñJ$f&A>>ù£xW5—cµ1„×åV]+~Íé%«zOðI œäè‘¿Âw©ÎùOÒ­{ìÊo±±Ìì“?Bœ¨Åf÷µðC•›í ¡ |ûM~¹bò@bRÌñ‘?êøUN¥(r܃’®XR™? ÚúÉ_œ§¹MÜÑ´­€ó݃¯wAÍ ¤ùž¿z8xõjÌñÜ{uÅÂTm0öwþ(¦ÁêW;¡æ’¿Š.µ5 %ê4Â0Ù{Ô(®€ê1ãïûs>%…Cr‘C†}o. ,Qq‰æˆúKáb3Š)´vKÄ”œÄ Ÿô£/Xö(äXù+\ÛûÁÄõ ç|G¼»:)† DÁ£Ðl+ÀΜFÁLµãDöÏ<$†Í¨ƒ‘û["Óá:K‚!1r¿#=¿ïŠ=HÍo‡™zË"nëæ·à$`ûçýö÷ò~£X#‰üõòõ/?l’›äÃ&ù°I>l’›äÃ&ù°I>l’›äÃ&ù°I>l’›äÃ&ù°I>l’›äÃ&ù°I>l’›äÃ&áŸ6‰›äÃ&ù°I>l’›äÃ&ù°I>l’›äï-›ä¿†ðË+”ó¿ëµB ©¤ˆGöÛ;RÝ«^ý":pÈøªW™ÇË·ÿ Z%$²=œ¬À{›¹êOd*Ý9é;ס`|¾#öÚ y¶·;$Rßî0FíXV¿§0jæ›gž”*˜Ãu¨9ÏïÈöüBb°jãiì$†´1ßn9Ö©!6=óÌ×zZ;Ôb{ˆVŸ$¾WƒVÚ3”&vB<üäÝ «Åd_Ï9Ø?sÇwRk¶·ÜÍ $ÖþÐ27çm7C+ä7{Ë Ö@b :ê+CpH­ïÄ'|VŽ›ÝŠ³¼“Ð ç6/‘’?ŠwPêï†ìð‚hßúð‹ð4Óƒ0™˜9=ˆ>ž¦#Èõp¯Øù«ˆápáé3ç8”œE¸ £Œšó ˆᥡЙüª½Mî ~•IYÍåÂÚCq‚Nµ=Ì?žÙ]Œìarsö ­c³Ì>ãÙsÈ3Ôkìù«ƒc¼Ý3ßè/az²rº½#œRü6°•ÖÃ\忝b6ˆR'šš‡#KËé;ÞÏ)Žššô‡ÕËÈ_Eb$ bÔÇ+3÷a+£•tÅä¯Ãz&×)§Ñ*´Bè·‡fKá-3›çæ%Q}6]*L²zøéxêHþW¸ðxêHex÷øÍsÍHÍ…_tò0r;¼È„IÐÈ…_¤ÝÂZHë[ì±öð#Ú¹êŒÑ*\Œú¤s(öStº‡÷ѳàÅ!9a×¢8\––™ðfÚF:Ž)þ¢­ñí¬Ì‡2¿Ú«õÄä;줴Õê?L¨Væ О-F6H^©/5ì‡Q ¡Kuã, `Ja„ÝÖø–æ“®•é’gÂÙkgº-~Ú ¼šxa³¸—R:»‡¯˜z)µ{¸‘ÍÌù ŽŒ'íœO}64·à–¦«_¶Ýñ¤].–ËaÃ62ããg˜·é)zaÿ¶/by»ˆè):ò-ærÃÚåpÆ“.¬=$ö|â hô0¿ÓªŒBÞ@|ÞΨÖsbß;cXϹN_¤‡ƒŸ¾±ò…µ›Šêœbu–IÇ0ôiÁ\éaK82›  º•vl²v˜$#ˆÂ½x‚Ù7cJËžÖI GE4ʈ÷ð‚TDc†6´#õæp¨ßò¼0³Tz›ô”0¼TÒjÀi ÃJ³A$h=ÓÛÜî«ÏVÅÈ{¸>éí¤¸mQž¼‡]éÍ4y#à7óæÕJõ™[‡ k3‚w\Wf¸çOêÓƒÓ¸Ïæv¤«(` ÇßÝ“„È?ÆŸ%wŽ#ápŒ Iך¯ –¶NÈÐ «ýWcw£pÉäçà’¼–ì…Ú7©%×ÒÚ[ Ä»6•·Ó*¸›";ÝØÅA$)#÷l0wâ=›ÎÞTzîÙ F€’ Ž0õbšSìD!֕ܧÞú q¯œ$<—+ÃìþÉ$\¹ÇsùÉaª\S“¼¹VîòJp‚>€Æ¿Ø´`zz8™,7§\‹Ô¹rMØ$VĨ—.²Êmž,QÆý…T¿¼VèÝ-в_ˆÙÆDƼœëw·˜I.çx .Ý$¢ˆ¿Èâ(Ç¡@â;…ÄîÍÍÖè§èRH¡—ܧÐn9C-OÛ¼x›”qšä¹ŒWsš÷t£4„ßyâ]tå*âÙ$ëÚõ+‘=ˆ(¯³‚!ПÀrœhOú¤‰Ð« ‘…cr+c“šN 7‘Fð_!€>Iˆp?ˆm$BO =‰1…ÅJ/C³ÓÍX&JÁ;DÆØï‰Å«&Ëm¸}˜*LC’Ñ€lO¢À7¢= ÿ趘2ÙWæmÏäç^®¤rÈ$`nÈæ _=¹!ð¢‘ò0 Ž/Šœ¯Ö]ð«äãœËÆ÷ìŽÜÂv¿Šg¼èq3\‘ªëDô­/†øã¨Y](÷)Éϰ=˜^¿áߣ!ÞO "3 MÄ*¦ _TžÍ”X!}ш]¥{ráxE§#=j€ˆKWŠ®R“UØ ¦óÅA«ZBZ2¯b–'ƒb fg)ÎäÌ¢×1죮eœYÂbÊ>íð5eCˆ©a›­h85H,x)IgÃë#¹a䯆·–/~À†öðƒ'‡çÉŒ®°¡"Ò(Ýáôµ«4áa³Õ ò!hŒd=‹ìWrÇ„ÀpÓÛg0}ñ ái–lDôËž$öJ^¼Ñtu¦ŒÃ?M”[š±Ï‘Në0a;¶Ñ²wÂË+ Çdb9Òë–p:mç ìæš‡i9€æ·ÕºÉ?‹äð§[É#E·ëfûa¯ÛIxÑ9HM!¶Ù‘8x­1Ÿ‚¥ßqçÖ!l÷øŽ;+õJn?³åpPÛfA**.×Záù·õ+®¬å›NÜðÔ ì“ÏÕ¼Sø‹„I/ư/n‰‰WÉ,“’1M™Äå’[[â6 ÚºèϰQ²û@D´¥­2µ@Ú*àq™‰Íؼ¾^,‚ƒÑ£¾1åÖ(^ÐV†‘üXtU õÀ/ìäòÈ"¢þ1.Ûuô$©#êW³˜-•$€Xí»è¨ DoxvØÛA´ïÈ=sõJ²óñYjw«‹Ÿ¥:þéq Ú8 xâ¦ç¢Õ3¿'Ëx TLëUtfÙÑG™®z1KˆØÎ«ë¼Î‚ˆSnÖb vz™˜¢¨Q³ÓŸUxˆ>ÃÒœ¸˜féŽ0RÝߊ<Ñ ¢BدòÌ»ðå”ä?« $–e(xˆâäD xíešñN»d ꜴR†e,—‚QƒÉkIŽô^ ÈÅ„®(®AH,ÝÝuo~ˆâña«ã•nVüf!waa~²&yO"* 푢˹Κ:5鮾r²Œ¿«¯œ,Ñï‡î¸ ¿c­`gí5¾ÜÉbûQôå²’>æÔ X÷M§ö÷Úp˜¹ës³|ˆêäYbD-€eè@V¯—AD‡´@T;ÎOOUüÖ‰«JòQwÿD>æó*©ž=î˜Ç%ú­Y*ýìq'‹äiï DEòs2öœ® w°­9›¢»QÝüĤHÑÕ±óB\ZòýQ3Yè²D`{€è‚-6t¯Î q¼!‚ G'VLXîY¿ã%ò€¸{dEþ:yË–X7oùQ–àKmŽ»ûwßìÄQ³ÝM÷œ•ý»3°½!˜Ñ9©ŽÁ{~äˆHc«K?ÈùÙuæ-yŒ«{Nä(ôžT8U÷üŽÜï@Ë[6Òó–FÏP;xÉaé ‰…Ã7`çCÞÕK>ä’¹Äȇ\ò!—|È%rɇ\ò!—|È%rɹäC.ùK€|È%@>ä’¹äC.ùKð÷‡\ò!—üøK>ä’rɇ\ò!—|È%rɇ\ò!—|È%r ¹äC.ùûK. 1$ŽÍ“ß`7mf–b;xk›ô ‘ê,жöÂÞfñ€â°%°ô†¶K"P”¹Üó{£Œ ªâÔçþ4U@(“£žDä º9Ï Sg¼ÝplÕ5Á7@K:¤Ð@ÿ8ZÁ‰OÞ0·C È7\I5¹%ßE¢BU¨å™AY¹Z‰ÓÜžwÌåw ·Æs Ù„”bžô»Þn¤ŸHX'€»9 äÈè"ݼc)K•|¡V–Òl” ‹ä¼ )Ba„ZSšR±ÖÔÊ3CÏ­HÎIH*K-Π@äͺV@ EYc õ¦ŽÅ\æzSйJ/›€@]¬jëäÅjcÿDÆóRÏk6{Qókt#˺`ù# Ž)¿¤m ŽÕjòŠ^ó"•z%Sv–èÇym dIíl5#ãRm䯚DÓú1R%µÖô«EEͦÑȾDÔG"Ÿ¨áv®èñµå>,§¨Þ²“,ÒlB-nèWÍ©]·¸Å®¡kÓšJ¯nr!½8(gWòWEwW1(ƒØ{~?î²…Pž¿¸L¡®7óG”퓼,òÈ·R¥Ï߯C’´.ôàƒ:9´Kd“éèE…À­_‘v²‚k<ÜhŽ‹Ž!OØóWé ÷(ýªRûÍùM xƒC+c J‰~UJ‰;_=$Rq^—U “X¹H‘ºcå¯BàxIr™“"‘Šö“éÖБTOŸÐGƒÖä4ñ¼9Ý.æ¡TÞ”:4ÜIc dÕ¡„8³×¼þ ­tNÚ€à,ÚkGš!eå°2¥éâ­5#Ð|½}µÀIüî“Li‹v7¿9¥-:8o‚¶µä3‚|½µEÇ4ȵÝ'퓺7WÃGï\ÇÑv‰‹^ÇÒI‹CÕÅ%»+ç¡@ÐJ¶˜s@ЫvÏY)ë]' h{äÇd #T^§Î·œ†U äÜò±mÁíez5²3‰ÜeÀu4 Éf(àN®±{€¨tªGƒ!ñØC™vEBºê›Ñä÷×#­ ÍúäÀ#o5+ô@ÅÔ0 Cö˜©üë'˜Ô¿=Llr §Í9MlüCª¸ä@];„‰58ŒÎî¶|Z‹¢o0¨{»ãì '(T’õ äò ­¬ U¶»ò¼’wõn5D­»óÍHÎõžˆå5t½G1=íob®É¡®A©Ó·1E®ä!3>—¨éP÷W%q>Ü´:ç{)hŽû.z„çÙë:÷qB<Õ¸ý€zsãŒìDômãˆé’/Ø.1¹xcFE€™ÍìÞÄm©»é¡;èŒd˽é*é{µ'ÞQzá[Aéí­Ý=9bÃÜ-Öc‚l ¸Ö1kéÊÏp1®6ö=ƒ„t¹f1ÛxãÝ ¹:Fñ&ó ;§C×}}æ)D_d ÀéëOÍQxy!› éHÊý.y­œã‹¯ÅÈLḋ¤úp+c÷.Š©Âüñ]z]~såzˆÚãb6-ÜÓ˜[¿‹y¦@N7‚àqM”»«°é\3\î¤7¬Üøz&©ž# %iBÔFçfç¼&àÞ)Ë“t– ÎcwkkçN®àGnÜ);“kÚ;Á `wñA¯6 ï”OɽJÀß`7:ê1Ñg‘½Ÿ™ÈÐÞ¯jƒèš§¢×%>ÊЮþ©È1°h|LïXr÷³±à\>惗vÓ@¸Ç cñ>É©<+ ÁÛ%9ñ7 M»£ê\äv—W(Яžçh³üÂಿ¸N:jwôyã±]o8*w¿ÈÛ§Ì*ͳ¹]öÒÕ¬¹ÛeNüò&­Í­ž†·@…;wò§”<g–H]y Â\“/!ÌF 8õUÛá—¯Ã1¢!ÿ÷æây\PÌ™hº~ëŠo'†f~¿Á,Ñ1ƒ˜uûþZ§ad–Š]z¢(EWbv~§`]¯ø7|1Þ]So¥þ} 곕Šó°Ðày+…×I º>œ9xw1W$0uÈdã%#íVÖæÁ>VˆlI›™ —–¥ í­ÌL ñ@F†ð²ÕH™‹¢u«`ÕÐd¤p‚<› ·°^»ò|¬¦Ï¥o l†–4Ðî’¸s.~ôidÖ>^Øçl†¢ôË>gÑ’·»ãõ1rà„ ÈuKÃÕuºyŸƒå1 ¿u¸Ã"¼è4Ü98ŸŠdæÇiül™Qw­ÐøÚóÆ³Œ¯·:âÙ¬]y¬ÐÚQŸÂ+m.g`Ï<E o6ìg³7;Œ Mf=ðÙ-}âu7UŸÓ£ÞŸ}X+ž…éÜc>ª†€°û¬Mó´µQmØ•ç¬Åa%†ýŠëGØñvi®7‡·9Ïêô‡^œ%çéá-Þ³*ãýpeB E¿Úú €QCš—ñexOüÌCàáô3·^„÷ßÏ\ŒÃ»ögN½ m÷½Ò«Ñ ›²tfœIg™9dGsdV7‡©èGÖt(ÔÑ ‰—rœAV9jwxƒã0P驜@@4.9‘jlDôNã¸û~ý5á ä 9: ·jç0¯ä¨Â)ªó°.*5cUSÂx|Tƒˆ¶<Ùálè¨Ú+ª‡8jž¾ôùLŸ9*, „aüô¡ÏgoéÕëúûéMŸï8tM‘†ù=GŽfQ—¤Ð›I‹‘7Ù¤-6ÃÆm7ÊÓ6}¶‡ÉFGbè7ÌùÀéæ‘óJ¢ªÄÑYÜÕ*pjudLj:ŸŠCQHE€”H¼¸±Q¯ÌŒá»©›ímš™uêâ ’R Ç÷¢¸_á–Õ£‹kÊS5ᙦ\,¢¯6ݯIbG%Õ@t3…Ñbz5x ·jQã*ê Ó¶TGö3¨¹ÝgêGœ¨Õ×™y©2ó–õi»µ#¡…w¤é–M;¥êzúzмÖÙ—“šÙÌÕ¹GwüŽX ŸKŽó;BBøb˜˜­>Ýo3qèö¼_#¬}Gª^èƒ߯€÷b â(nµ¼#'ï×ÈVø=ò¦Xò¡”|(%JɇRò¡”|(%B>”’¥äC)ùñ¡”|(%JɇRò¡”|(%JɇRò¡”üøPJ¾>”"J ¥äLJRò¡”|(%JɇRò¡”|(%JɇRò¡”|(%o)%¡Sb…ý+ÚcÓÿþíß¼‹œ ×ù›ž%=n—ô’tlbÍù@${×kº˜C¤‰v­â{­öŽÄ7ý] *¾{—ÊÓqÓ€`×tieÌ+ª²ÂÔ)–âyIm£ ‹Í#rܺç7éªè:D“„rõö2rV°1Þ°¢û‰d)–V,AÙ¬è³q3cýzéþnéIñ¦¾ñaèWg:¤´ó®Ùa§êW÷:4•oJ_\‰5Ç3KŠ9$KŠ9$Wé>ÙÙ""'…©,Çѵ¿ëE®ú‘ƒô”¼Å˹R<«Í±žZkOÞÔ òN† ʤí·QåMIMÓ“ómp¢þÚ}ÀÔTÄíÍøfÄÍ¥oºn,õ«‘êЯ;ƒÆ _©a÷MM©›±6”›!®ŸžÓÜ«_½ ÇŒ˜,ýJÓú˜&Ö¯7‘»˜p ýJ³±Ñ%÷×=£¼þ›v^Ìmø6º–£ýJßx,*‚¥*_ÌšÐ2n÷'ì5±@âß—òZ×)ƒ˜°¡Ó¤"à™:`ûOJ¹]çcH©?[ÅTQâ„]sÇñùqSZÓ3±)]´»<ÎÌEI»ÔAŒÉí‘z¢§»Gíoåt÷R´ëÚ,ãP})}­¤VU´RcLÑùÊ·»ýjìŠw»K¯Î¾xí9v–´¬îñë[“a)%kÄ¥{”Y‹u_ñÍB·*š‘3‰xA)Á”ºô’'Õ0½* ÀË-.œ°K¢uR§Àm)ŽÐ›y ì­hu5¨Rr±ƒÖyJzÆZ×Ò½l*Ý3Ä‚±H;T ® ^X1Û «LHô¦àh,Mg{Ó$Õ+K÷ä-H ;ÕË;°‡â5yÇê ûRŧ#¢¥3w·±™•kr^J¼:¬Û/É7U†cf¥ksfˆÎ+ö ™ø¥³Nl¿ÕL4P v;`ߦG°Æl ç›m¤0¨kœHsðŽûÐ…ê/ö¶®< Êu@¼Kq5hßMç€.uÈ¡²t-zªh—ë]’7Ò;ײ굪ÝG’èè¥qKÓimPâdS©¹9V¸÷û«Ú¼QÒ*¥÷”ØÂ®¬ŽAî+6n0ÛÔO‰c§Ð9™T»—RÉÊGÒ´QÖ¾ª…‹ ËšéÕù”ÈúÁØ ¬Oºèß×­€©BÉ=+g9hvQ-«ùGhÛÖã]¡ oŸGU:ÈèÖëJnc JñU:¢ÅF¼R|õJsÛ6Ps»pu}-Þj&…)ÁÜj&…IK@¹Zu»f‘Å@Ž”»Ÿ¬0%Ê3kÝ&?Lê}Gº™’Û#ÓÂKåÃiá-…òé´{“#B›ÞêlôÁ R…6 #«Ø‰ÌL™Ã ÙÍ+ÒêCÚç])þ¦×•ië. „öl0ˆöŸ”âîˆöŸ´ ðèµ{«àÑt÷vtß•õß|®.aÚØ”  ¼§/PPÞög—®K@´!5Š囦š±AR%f¯V¶B´<Ô?» £®Ò{îFÊWŸn¦ƒü 0ˆ„å’‘6l$ÜÜ=&‚6˜Âþ±´„ø~£YÄ.¢Y·] ˆõžco cÕ°Jìˆ5!34QsS6›xS¾–ÁSOšƒJà£9ÆMÒiA5Ó™¹ˆ>éd¢*8lê9séê#wõ¸\¸–ÛŽMˆ÷õ¸;ïëq³2ïÁrC3 'bϳð,ÙöoEì¹]è…Y]}\ïŽ¬ÎÆ2mû´xЙlˆÅ ¸ÛÈâ=ØE•ï™›Š\Ÿ-ïpsè›ÝŸk]k¼mQÓ $ K¦Cs:8î—:§_OÐÿ¿@SÍ¿/ÿöîxcÛ™;wÇe04íBÜ8æÑ”>6ëñL¹o¸åû‘V+ ™¦Kð`ü‘~-Á¨Ä—½¶/È@@¬t—é?Å#¸t±n±§P9’žÂFiìäFÐÆÍÑœqånŸåÛ—3¢P& Æyd´Žç‘mÐ:îÎG¶AË–¥hŽïdKœÅ´KòS4³Ü5ù)›-x·7~ ZðnÉO¹œÂà¨Á²G2s ƒÒÉÌ)¿³1Â:r|½YaE/96;¶„ßmïK¬Û6‘›¾ÿ¶Ã¬Y@aϤÑàrÃ'“l"¼˜?Œ£8æT?û]|§&Jn"§¹­Ý-G©þÆ€¢£TÏk]šY Çõ«™ì™f„–¢‹k»‰bý¸Áešx©ŒÙAnFxñmî\Ìqñcî\QÀ9×TòB—¡s]ØQF:}ÍDàWp«¨œer´ fÔ×w-,ÅôÒÛÂu¬½1àÒ˜Œ“r\g¸¡Z'ØàŒv¸Ðz£èÁå.“Í‹¼ˆî6Á’bö°[«‰lzÈZõïõCVÚ˜„©›þ•„ô@n"‡NpKœËóŇ6\1Û‚Åœxs\K *í©²Q,ÃôÀJ_‰p³k‰\bŠrxbÖE_²M¹­›Îå˜A_ihˆ¸‰•‘-ùTP¹a®}ªZ«—^•‘ÄkFŽQ‰^ŽDà:[{²c+=8kò÷[Å ¬ãÛéTøcá¤XW’c;ëJr,æøðD|çå¦m"©»õÍZ‘3Úúæ¾n€­¸ý5–Áű‰…_8=Šß8¥D…GÍn•Ý­¢ÑK)\%EmÜÔ çIÓ–YªˆŸŠË‡ð«ôc1’†Ç¥Ê§DªLÓ¾¹¯LÓ¾åjÜ®ƒG—‡q/np]~•ñoD׿j?¬;«¶õ¡w®CÂïÓ¬o2OÂ"T/¬3!6¢ýŠûtèøøðU#Ý=üIgàñÞû1óU´ù0>-ɵGÇê×m;ÎH³Ô•”}ÎŽêæÕɨw8fÁpjå_‡·;²l+6Ð7Í]§.sy»cº¹©Ä LbUY…>²Y§¯Ùc.’Kx‘žµ˜$£'Ú×–~aÔŒ3°Ç­Fð›$ùÇN7I’?+6ˆקí|‚9pbû î½êñƒãBxþú¼L‹T3€ üÍ,~¤-ü…Kb;zÉCb¹¦Ä뽺¥ší Ÿm5ç»0DÖ4¼ à©«Aÿ^Iè§ÙŒ–ýGžÎ+ßÌá3­O àuò ¸?UW–u+Sç›péöºŸº.òtÂuÚu]òÞîs•“ôuYÌnÖ. ãær8^»ø«ÒÕvO—;L&„Â9[ÃIì~Án»d±†Š|ýÙh¼[¾Š¶ªI·X(ÀHø¤Ùùà‘>4€ç>Y?LsòfÂúDöÆãê¶sÐÜõdaääJ7 Ì]™F6#LÎ Ä>è¢áÏ)ôíú•Å…{úÉj;ú©_ÇžIâIø²Ï¬¿C»¹5 È#–9¯Yµð5‹ÿè@ß=Y˜$Ž…sü1°ð£éF¢ŠÁ°¤× :¹U¶õjSöÍ.5 è³ xZmß›_“)·óFýê¢ð¼Ôü˜—í¥åÇdI+ž´©š²E®0Ó¸V¦ÃðB’ÈͪLlÄ–œ¢-f ZÙ.¸ j”™—“µÕ\´Ø]Ô¯X%Ú¢ÆÉ@¼ãVkVº"°¹ù£‰µŸYާ­=ë\iåÖjN¿óŠ­f@Wåk«+ë\Iÿm5˶T/Û\åM?¸VsªµKòkÖEª ·5×E.Œ[˲ÈÅ\H‹yQmq@ó ·zßè.ÜüÑÁi¦ßÉ_ÍEÖãÚ;_Í%[;?óœäzb¸˜Lm=«ïUÔÜž"ÏÖzÍÆÀÄMëÍ5r‹Ž•­ç|IÕÓ­,„†M^ë3ëØ™´m=K¯–Ž/dgûdcèÇ….‹9¢@Nþ ÍÃ|6 ÷ZÌòW‘ hfžqÍ5Z6Úöµá®¢Òò6Þ‹ÖÑ|ÇÓUX†Ö†…T¡ÞF–<&’% üÙ6R×A…îmÜ!@ú¯·Y4ç­Íœ™7¤¥ Kè¹ù«PmfhNdxNºèNØæü¹¾YE¬jý6w¶a#çgûܬbgÊ=»¾!«d»‘šUìFZ~<ÎÏÛê ùñŒL—¯%²Þ¾ž&vÿ’%"ɇHò!’|ˆ$"ɇHò!’|ˆ$"ɇHò!’øI>D "ɇHäC$!ò!’üøI>D"ɇHò!’|ˆ$"‰‘‘DȇHbäC$!ò!’|ˆ$" ‘äC$ùIþÏ ’@¨äëo~üϵ|Åôÿù·¯úËôï^ÃÎ×ë%½Zþ×/ý£ò¯ø²Úa¡™=¿~ùíÇ^ÓÜ?þòßD:n¿fÏ¿ü«€êÿÁPIY¯·šÿæ¿üøCÛ7þë¥½Žøç¿üø·Ì¹@žwa&<û YŒ—÷§¯—¯¸…äþo/dt&“ƒåŸÞZb¾ƒö™T€"‚71¨Õµvy!zC¤¾Ø§6²65ù¿!TMýŽ@7*ÛöÆÖà;°(H%2)õ÷†p7$&-÷¦Ìú;Ò¤“g•Ò-¿ž”Û‘Ål߀"Y5Å¥¶—ÔûÔþö^R[ìS›{q9ø ÑŸ¼]Å…GÉn¯™_L…­Ð;™ôÛBïêNš>Š·Ý¶º@ö5"U½¥ÓHç4uö¶f@xwšn"ÉžWdžŽf¬í,}H ømM´‘šQLØ0÷½'#àžÚTݹ²=¥Lü][Âp–  è4S×.¼’el÷¸RâëÊ~îq¤ÄוÜ2x„YË=–¤ø´Ò!SR|M@(ÅgIÄM'†¸³³…½Å-›…@ªN\©Ö›6?7¿7t÷(”ªêÎ=îNÑîGÂq÷£ U Òí°>hDÔlÂè@çQÛã‹Ñ?OµÃâ&¤Õ{ LÁo-ù!8©_Qµ´+C½N~t}ûN‡¨@ôñ;´P¶äûIF' ¢\;Ô¢kÞPß!¹Y¿Þ$3w£ªv Õ@‘§>Z›lâMn/d°‰§:ç~Í­bIˆ¾‘ÒP¦çn•R±Í©æM±òï¢G0Û`W‰7Y˜¼CµÉTÝJÑ… ©‚`¥Fö£KºCiöQ”QÆ0ª_ vÓ<ݵë¦v¿vmzï)R%œªž[«žbú[U‰È¥ë.Wb¸Ï•£§îb”ןž¥¾€­a=î]–^O×6ß.S¯§û…‰'nÚÝ„ŸØ:³»t¹¹å ˵æv^*f«˜¥°™ÆÖû4Â3k:ÞÖ•xi“‰Ô 9ÐWkÞ6YwëÄfþ¬«NÖ”PR¤·K®Åº“ÊhÁ'Ðy¸Mˆ®Ý9¤Öï ª u-1€IOÝ_¥ðу¯[t^­ôÛ:W"Í[ík#‘æ­**¹bîÛ®#qƪô ‘œi]⋬#9ëjeù%ûPOfÈ ¤èºT—úóÔžÉ:ôÞxT™×©`êÐ.S ”öÜ:% 9UXû²[§&ôÚÌåÂquí-hó¡R³Ù†+v7‘¢c&•š½ƒ¸¶«’;@²«5×–Æn5 a…"¦C*ÇœâýÌ@¨~ý…wá¨Ø~`Iþ‘ê^ëPo°˜~äb¦ï σ¸z1y­ÅŽž"ákMvô¿FŠ©«Kè1¸Z|ô¥!ïÕ›ÙC×j”).V9]‹Vˆ¡s~¯h8—¡ ÙZEÏ5äÖ”ôcl¦à^ý#²¾´Ÿ:ìW?Úz,kÉ.©ÏR*¦W7¹Éõå#¿æ r,µœp™O ˜C»b¬Ù¤àfÖ”mÌ½Š¹k2÷2öÇ@ Ç¢Â(ÄðÕœ"/I€£ØËçbQ~þQØ_’Ÿ„ßiÞBýœ8R…(ŒA#¤;´¾† '‚ɪ{c¢ð± ¤ B˜¥ „ãÁ ÕØN‚9T!$ð­~¥Þ_ªú¡ÓÌõ¾l °…ºž£¬Î5L zÁ}Ñ8Ày¿2åó k ôy8â}®ºA]§ANüñt è: rïá ÁÔê´h9¹WlÏ\Lè…ÆÌiÑ å$¶êvª¬+ÇÃé?lyYx'~µ¥g”½!&D·»`‡fº6äeQóvØÃaÑ!¤ókžâp×äQµ¯(4ê?¤º '˜_@‘'Šï¡'Êö·¬²ŸJ'’UÁn ¯…­ëö³*¨Åp/€át«8ÇÍ¥;™Ë‡ì”Ú~ dÏ‹…3|›8ëž—î|aöt®„¬Ùõ!æå*1¢V7B&óäf”Zàžk¾Œx£S¶ó*²L¯Äæ¡ÍÞ°ÁÌ”<˜^‘è7Ïá…‡©öÀToh¥3%Ø@²Õl¸¡{¸î`JÂ?~ð@pˆ¶V_÷†ÃÿK¯\šþ蕟Î87lý3ÅK¯±+„2ÿtGnPÃ*×à³P«ˆ†a;AµZ›ûrÌìÇþB`wÓ·}sdîQ PcÛ2ÃëÚ³ÓÅðxÛBÈÐ ¤#°Êê]~ª\æqÐ}í/÷B( :!Ü €K]"x]½hàˆPíºYíÆwÕŽ¯ÃÕpj廲ÿ4G›¹I ›=º@`ÔLòžërˆjS#h ˜pØÐï…~:›þAÛŠb+¾™„yàJ†xÍI”¹–îÇ>`piESÎ@èmx5Î5ÙÖM%—¦êú€¹¸öÄBbõ:Êõ ¬{"°È³ã ilþÕÕ`”¯^WÏUeÙÝiWìay„”4ŒT#eeL%uõye)ß: 4´´sL 4´¼ŽJ“„¬°¸T”œ2º*"6¼Í; ÅõDÐEŠWŠs>Ž›Ë†™X˜Œ(ETà%Wf‚>S¼v›SæXÅêì`œ/Ÿ“ Û0‰À¨÷ºh/8_{„`9¼\¯¹éa§Ëé@”tB)Î@Ôñg¡ ¯)n@`6|½ì™ŒÖ0fÕ€6˜:D½-Rš‡n®œ€ÁÈ`ÑV= ì*mbÕ.‚îp‰¨]Ðýæ²jc§ÿ¬ž‹Qµz®±èçúB:þÒê¶ÁšbÈ wí "1lêØCöÛÇ™ZØk,zóÎDðm’‡ +\¼úÏ>hùË)#B.m5sð€s°†ƒ!×줒†c ¾Õv=LMMŠý¶¸%>Æ-ìRà m¯c ‹€_V‘ù±³\áÓr.m•§:&W7íÎa&̙մ;³6¨uW©ÓÒ™+–)“°}fú`¢Ù$Ÿ8Üið"–=S.á/­ÓñïàA­ÐÞäš©MÖzÂËúL#•€Ú­Iá‡Ý üÿì½É®.˵×?O±šdCÛQ·#ȰlÀ€ ØdÇ0Ü$ª0x\È·áÇ÷š£˜™ÿ¹î ,$±Æþ«ÌŒŒŒbŒoÎÁ’ÙêH»«]e9°ÁQYz[—¥ä¼ùÁ,Š[jÝ*§»|2ÀŠ:àºÞÎê(îÏUõëéÚ,(k4X_\÷|§I¦§cËú‡uÊ5´ìÅ£gUOÖ¹x•åΡ,UD?PLz\?Ù;K2FeuuÏó娾®AcW…ìá9?áñ® vAYw=˽œQ ^ŸÜ¸Nãuƒ·ËŽkx¶ :ÑÔ7¦c¢½î…XaÇ›\‹ ëéñ¦îi7ÖÓã]Ý5±ž>©Ôf«»˜b(è9ºÂ!ĸ¬÷™ÇOïž1c=ýRÑÉi õÞ}\×ë½å¹Y¼É249TùVJ‚JvGêÉds¸ëëñkÚÉ#˜,^ßrÈø·²ò«™·™—a°ž}ZüÆŠ;  aöÖÝF8ÿVšÛh¼3dëëq&ZÑdÙâcêÕì7”?…â¯ê<éõha4tàu{:Ñ&üVò>kw}u8$\‡:ý'zêU¾Pð<Î âT©›^›ŸÙ=fÓ*«¯gr2Ü0ߊº¶8ñÁåz>ÄHÿ·rüU ž™ÑPpS¯¤¡L^Ϋ†€‹Y¼iKý¯q [øªîι^ÞÒò«T(½x½ Û !xJ×rm×%áBWé^?+–¶{ü|ixà+;[z‰?—ïÔÊÑ`»ÓÍœ<„o%ï¨Êò‰¡èŽª—ö«jëP`*¹^`‚‚ãj=Øá»ÊÍ©XÎãç/A Á‡š”š?nšX|³ÃÍq+(ñ‹Ã ïæ:Z;9ͪ4!´£`‡¢ÑVårz;ÍÄzè_;,ìýŠÇ)•Æá¶¯çŽ©O.B¡PÚ‡à<+vvâ¸w¬RQ}s(ïò¾þ£x'ýQœþz+5¿<î»¶s‚”J®¶XY.Rø(G[<²ýá%ÿE3Í'0£ •!Ò«~+.ýåÒÞÃ!ã£8…Ê•¹ˆcãG9º§eërÅ24li1Wù¦ÖA±RM{ÔýøÛIO› +¨À;h»òXñ%xÛ²²”CÓWŸVàv˲gXHoôŽù¸­¬Äu­ ËKOEìhÙϲ‚ž4 pB94IÞ|ÍhTJ¾f.Z+k·ƒ¦B- 8ÊÅWy…•ã8›“Ä¡Äð8Ü¡s[é0ƒz;x4›,¡À£Y ´wQn$‡‚ÖŸ»ÄƒsV(¥Zi°Éªîm¸­«ô°ÙôkmÛÐ-Ö®çA(èkÓ\ [Vø˜ªQD(ðÚ–l¤ôR7FMhX•O¶ÑÍçUÙÙHa'^yÈ%Å׉Öê°.û:q†çkeãs< …Öjù¦6ò¾`­žÓ­ÕJ4@³ÿ^ÍçB9¾nuÑ…EÃû­'n)2§/ÿ ·BÑc°æS®ÍØŸœ4žku¬TƲ²÷$a –v =T´.l—ÊðXn¬Å~i9´Ïû7s{#,ö×ÌéváCAÊNý6 ¯‚™_ôÂ){þý“¹±¡€|R"Çfl 7f 4X*0,Gaw ë0°pŽ•»jÐz·ªþÕ,[ 2UÛ&Š pÒú\®ÄÖCBWn¤04B-ôÎfR 2LNs´ØiG:f;g<Éß‚Rºdï¶óË“ù>dK4¾ãrò'våËrƒ(üh^KhŒ)\Ëîœ{d4V¯´ÄT+Pm:"5?sõ±ŒèõЍN í*Ís¬¬ñ ü@‰{Ê™ óÅ`褲.MI+é"~Ô,à€ží™Ú•|šžvÅ,ƒ9'_†’O=‡/ƒÉ§æ•·¬¿†+ÍWêTêɆŸ0Àຎ›CQ ´—‡¢¸©ŸDX ³ 2•ûǡlËž"rŒÏ&ŠÆƒPëØ±=ã%y§häɸœW…ç<1ÇÌ#„šSš£À]Ë9býä´‚3¬ˆòyVˉZ}vC4áËH çˆJ ÖœG":ÚGΠIå‰ôáÈ©/. e¤ óéõ5ƒFÃê%gГ·Ê³×Ñ”?kyó´Å ÑNN¡YÄ5¢˜ž×nŒì"®é9ôæ½Ð\[zrwä‰}Ba täQà¦LU˜x©Zc»úu®ÌíEžPÔ¡ji(¢¬:‚^ùäˆÙW³‚SÍaz' r´7—®p5W<»_Ï®EçX«Ö™ŸL$A¦z¡ ®çà/öвâZq§ç‚!­Õ5çCAG\=ûÕJd}¶("ÐI¡jQL è‘K¢—›Ýb!?¥,÷¹ƒÊ^=Õ/Žüµ—×aŠÐ¶z±ÁxR»w.O|{óTrПŠ:L-ù*ºÿìG\E÷=*ò¾F€vît,ûSO຅c„Aî»-ðlGDO½_(ìß/òlG¨\w V*Ä.ØÙûIÄ.äãmN1rà¤}ª`>¬T®¸7·»ˆ]°ËÚ›d_" G’A1áV ¦"…+”…:öE'nà.4¤G#$ã¦ÔÑö\#öÙyÇ®&üÆð)\4l„¢o׿öeþ!…ÅÞ‹7j'¯Ã®ºxk RÔ½-všA&Ñ7Âtÿjo·<;4½‚rRÙoLJl,w¡T´Á³þŠq'EÄ̾eSi(䝨Bêq$÷¤ÀbÇO(DÎT÷{&r¦|lÝKS­ ÌÜtÜ›GþllíUìßÖ‚Žj„Ò±€q.o÷ц•R /šÎ›Ì#èÏÔ_ü’2(- 4ÍÜo—u£ä¶Ç€,)Ea Á…œ9î;bªÌÂ?±æŠðtO3 I9K2pœt‘\‘ ¶‡ÞWü”ñ8wŠHPs‡†ðHŸëÇÜ–$(]ë+ Ùx;wH‚Ê­’˜eIQ –(b÷vî¨ÕË™Wx¦ñ8whÀ±p%´SÀÒèÔ%ÈTYЍa~†‚´§i±ÐŠYr%¬X¦†É£Z:9ÝÄPÐWw÷"aû“"™eá“ ÃdÆ$¨²üÉpV¼KÈl-½n‡t«Ä®…ºL¯ùÕœNƒu!CÇ¡ÈìFW]@4dH¬œÏ>è¸0çM(ÆË…ŸT–þùÊYf@=tnªXdmûÜÐ:lý˜%öžŸ³aIDoÚlø •{;ácÄA5Ñ4`~ì¢úɻߘ› ÅNâ>á“­²²ŸnÍ6îÖx_5‹ÃÀ‰¶ž Â0y¢ãhiKl˜… Ó%¯hÐ`v:Lqº”¸¸æªÌîVW”³a0·Í¸MwpÝiÆå¦F@nviѲëòÏéÔ´™wî,= É0èü@gàÂÔ‹¾¡*¼¶ºn×jOyŸú\»‹lÿ'-Æèá«f¹áBÆÃ$!œá]æçVߊNÅ Éb:|Hþgؤù¹Å_VêÀ)_2¸°Zr+ônÅþưmo :¡cŠak‹RØÁcŠŠ ºcs€VÜmÂRžÈU8ÎIe基 ËÔx=lëøõÅ«óáuÇóÔ ìðüñfȆrñ16C†‰žßÓ}³LÆÌC‘»}NcqmoŸK^g~ÂøOoõÅŸ;¼º}æ×€ÂÈà2ÏÏìAEãUÅT¾Nµ¬&¦òu÷³š˜Ê6VF,âˆO,{ëb*õ†a6{“#’Ñ¥8È!¸ùµ÷RÑÊš•s¨èMXvçLæà¸ûÃ‰Ž  yÊ^Xˆt yÊ%.2úˆTÓk†xÊÖ}&a^pl„eö©šUß îÝŒùÆwGT§ï7ô;âz@FÈìÿ}ž.šÄÞ™ÃÝl±{g÷ ÅnÛ#\¥ôŒ £nC‚6¼—ºUÅ¡%¡o²èf2µŒ›l/GsÃEFAAS"Ç T}Ìb©‰àƒ¾Yd' v ;¹ý²_؆ <Ê!(¼‹®egüTVØöOŠéþÑÙîwfùÇ`/–›þF_z¢ÙÄfAQ6{l]¨Lê£+5|«+õDõ ¯”ØZÝW)Îå7])Q] ðJ=ÁüÎîp; ”8ØÝWœÌ&( ÏÅZÛLÍàÌCE7ëd)Œ47cp¨¨ý­Â;jgxUÞQÛLô`(ð¦Ê8x´5Rrs (=¾&;Ú­±à&‚ºrJ 4Á¶ÚhBvl¨ð\¬«#O,À.¬‡òR*Ü‹OA±`ûJx@œn ê€öÐ? âÈP”©ÞKþ(›ö–ãJ°ŽOáêêØöeíŒG9…•D^JÕq?JÓÝZµ´ãp?•áã¶;̇²|ÜTöçʈòËIØP°Ö÷aCy)ÇÈCÆ^øb·àæÏ%ÃC›ŽhO©L£|RYF žÆÝû½ sKå˜Ât¸£Òܯ?Ê)†F»Ï)Æ£ÚìÂÃô•öÎÏIþÜ¡* àÙcE¥Ò‡õ9•D´ò9 ;—O‡þ½ð€ùÈ›˜œ†¾v“jt¸Ýv38²c°¤•|Í”Iíæk€Ô²µ Lj +qí޹eÖÐé/,t+_ÅÏ9yäÌv6סÒèßó‘:‘=¡" ØV‚ç„UPl«CÛ섺‡ù’†ÉO…aד9¬,ÃÏÊÝ <¼å‡·éñ ‡¤øS›Ž¾pQ @µéß ë¥Ð†K×áÎaiÓ›×=…†R]žBÉmšêÂ*&ê¦?.Ì£5_3; ¦bí#oè6ñmyCî+ŠPXW¿ã0†Wq¬8 ¡'¶ZÕ¾Y(½Ó[«FºÉPÿ­éæê˜vE¦ÜKþàåVº— ‰_ŒqHƒAØÂ¥ßN^+.•†§XØ´ÍMæ0" ›Æ1ÇcV†‚VÒ#€O{ï&inæ÷»×è¸ @+µPw¢'…ÛZýË4¹÷í›j:ûÎ_Ü™?éÇ,³Ýå·NØ"üpˆ‹ ËÇ<\än]'t´l]„„]àÌÝd|ù›Y5#lîÂó‰æVxÿfùûxî$’á¨oZ½‚‚ó5nþdÖð÷¾ › Þ0øû,ÆBfóO^,É£óh¶èfíCA۞ðÌuér6YÛ"05Ï®HNÂ× NÆ|àŠÜ‹̄™q‘åDÐÂ<=å~V5gr)÷³šÁ§Â¶…ò ã¦RºÕ³B4$YyÈ?¬ùâé¡…­‡ÇÍˈ¡ø+L`$WÚÄwɨ¥+\œÃÈĨ#)+’4+)Œðsï–H=çƒò6YÎ ?|–óA3/ ùZ‘zYðÉPNâ(›2Dú*fL<ªVÐÇ9œ”HÍ0%‡Ý0zÂ;ñ’æs1Y±#”(hd©Ì:åÖZä­¦_ó…„–A§tÕEŠË S8ñ"èeÎ)Q_dQ¼ÔPYœ´F¨LHb‘XûÔŸ4.F\M?ŽÕ‡"Ñfh-m”™zƒRsNÉ8Šôœ9§4FÂn&ü׉vE)SÙ=õäFºO÷â¬úÅÇ|ŽS‚†óFP—r*œ^´ú”Ð߈( 58ˆ«!^8áQºû¥!ÈEÉ{(fiŒKjH%–qF*YŠd1v©.p¹Šúq”ÇÃ,[5<§‘†ª=ŠááQ=»+dÀŠ2¨>²2j7ÔwQVñfm?€í#²þT–4û}pì6øûP£w/Î¥HìjðÁ)8r¾º¼î„PÔ’DõŽt0¿»ÃЉâ‘bð­'|]È6ÜBEEƒ‘·ÿhӈʾx½ÛroÓáh }´æ`Ák=™úÐϹzØwA9è’‚`|¯PõFb@Ï1?°½¹G 72"æªh œ|ÆÐ¡ EÕu­ã‹g×Ån|”Eä]»ñÁ1yÝ €ˆÒóOVɬ=´»~=hôTŽ‘s‚ÆÇM$ýÕ5rÉ&ÐsVH¸Àøh èIhŸ‡ûNHœÑ ØM›â-èâs™L]ü*ÖO”\’rĸ(Yýáˆq‘Óˆº¹hžòU«Cqõ‡%ÈE÷½ÂBÁ¦h)tÑ*\z{I´X) ÏåÑDïcî¬oÓj9Öß¡x +éRŽ^C¯Ò/Ó ®ænm[X ﹂üõȰ‚{)è4l.A!ñA :ø,Ê¡É^oH„‚…‰ø(R`IR×¢´P²Å¡ë(vÅjsU7#ñ¼eeÔ€¨,Õuð½<Mî&¯ç·[ºöP¢ª0Á飉xÃíq ?\fVuãléý@ñ\Xm^þ²²Ú¼ý¹”ùM"ÞÉ|&â¹ôöM"ž½g¡ð¨\Ú:*îŒX²Ÿ™•ÞIZ~„VVß®ÚÞe+dKºÀsT ì,º_!Ñ©f—|莡Q"u¾a0ðóƒÃVE›äO´ÝÁ³/e’(\³3/}ÕZrçÏ餭Ö,Ì6‘xSo0¶šw†Ò‡DO•e…œR—@‰I¢h‹û—.f%k‡J1§Teâú Ö1°¨x9@UÓ§>؆‰Û…ñn6U„BN©36Q€“œ^;¹£Hg Vu {¨w¸”\/âÚhç¡S |Ìò)m,½«=¯CÛänfŽ. –¶ºýÛÅÕé»Kpí ïÅQy†GUÓ†uÍ[KڰβlTÆøŠWî±RHXÆŸõ §]]kSUbg÷VKT’ÅuèY›·.á´³6oâi{ƒ+0»·Î¢°m[g]Þµ‡Âåq%¨UTŽO®xÔÝ=ŠÝê QXâ㡲‡Rñåî¶ËaÇÔ½Áe€É`÷¶¬*Ïî½Ü(&Lûv[/Cví_"ó.oé¢^».¯q`¡èC«pÔ×=Y)"£_W%¿¢÷üQvÒ5 àBf|5Š5ã–ÞÛ?¢¿…¢Ùû66ÿAea>ì,ˆjÒ8åÀTœÆ»ìž:Õ‹î÷—n7D&zZ§Ê±€"ÀøDTÈFgEãžiWÞ0¡@ÑââO{ùŽXQŒl+è§C§³´G(„¸žÎ sDH8„?}³?ž~á˜BQkïÌ$€àÆs¡”V(Œ²e»BáêôÆgÓ´‹ñtV³ŽtJÕ' ¦Qº§qI‘–+…ÒDqŒ E]©ât¡p\vÄâ¢áÒDÄ`hå;ÕÀC¡5ó´ÎkµœÅ<íUÜë(£8“qʃBà7 ~ŠnEe*Ö£ ÙÃ*šö…‚鲕ôT®ôAÑ»%En‡çª ŸÍÁ>²@Áõãõ­ÂÙ/OãNey—Øpàßš,§·O¹Ì¡è±Tä}½Ùy_¯xq›ù*¨ƒ¬E&i‡r¢Üãªï]^&¤>̶CQJºçËOý|zבòèKx)³[åÆ +Jgtõí›0RÎÅÂæ®mêÛOvv¼»£eÑxà݉¢îwñ<5”Káéx·>E#œÖ™ -OÏË<EJ+&^Åsœ’m#ÈõñìÇÓãPP¯*](ÍO &ùJö¡]¡Á’}fç½$~+”;ß<õ9êõÔŒ5\ugÌ]DŽŸ¾ˆ'w(ñ„F’§¨L+žâ¥RÙøU»ŸD †² %¡’3æ:ŒŒ=B[biú¼‡+O CšÆ¤†Ú±Xœ¡¨™Žéx‡Zi¸-ïXn{'@bÀÅo¯AÓ3eWéŽé Ñ,JwLÊ¢ðÓ¦²=dp§L_šIhD¬ãëÒ8\hѨµÆ"¡h &-‚1RKÂ&‚¾{ë¶{®y=*Û}Ð$—<ž-¹cs¢æ(œ‡`Ýqž7ßšŽÀ‘ȘtæžYBUf¨[àU0 ¦<7böMZ^ÌPpjñ=¤YLÆ•<ÑA}u}ÌåÕ­&¶Ç„‰á)ç@cRÅv4ü8Š[½Lr­¶ÂÆ\+mL瘜²U7¦|ËU[~cZÈOíjꊺ¤Í]ÇP4ÒÛŠáUÛŸcšÊO¶:¦²¸y«WVBA ¯Ú²`ܽö·a’<•¡)1סJÂxç±S§‰«æãØ»ÓçLF13w°ÓÀ ?}Ìf/ZMã>NÒÖíuˆƒå0l%ê@ëN H<„KÛaêž/\!Üaá*¨Vré| E½èíÊ÷]Ïa.oºÌ‡f›™O¡àVhÅOÖ»x7·š‹+[ÉgI†€ÑØ›-¹´ã€¤&ZÂŽ®Ö¼Š’ŽyÇÂ’=×Åš’Æ „‚79Á VÌ“]àXâêJyjý¯LåI§×ÿÊTž4÷׸v‚õ#¢@i×J——ÕËá=ÞœŠ¿Nɶ­>3Vûx¾¥Áª¬Öa3&kæ},#â×lŠ¥FD›væ¹)Ekïu°yõâ]¤:•_ÎíeãÁKA9Ùž[yu+¿ì%×9ÙÞò'_…[.s- ^:.¸X=—¶[åÅê†õÅ0nš.æÔeÑ1Ú,h •žÛü ’(ÚËZj¯ìt(WŠ޹]ºV¨é4 ûˆŽ›n׌m‡r xükèSÊÖR<7ÂÂÌRs)=t¿ºíc¹wÑ(þyZÒ‡oF êÚ„ÙF;F¬|Bî(¾Îúªz,<ÕÊ•¢Ý]Çh‡G&W•zž8½÷6à'ªV¶9°%i ‡¯PtýU\y)•Ï·Ø¢ gÎPa(è)†#JWµŽØ@l¡ Gm/WWiãÒ®çœeäæãàrLØÁÔUIJhûz±BÁq&&:ö³†-³kÏ vµÜÃs|ææÁÔ=>ýÔ‹íµ"e熞M3÷#g|ÂÈØÊÃIS‡²?H±ÈïZ¾¢•›alȹðªoì,)Ú¹pP79¡H¸úÜCÎÅuϪ}Î0KjãgÒP C¥ö|+ fÕvnåŸ+7XqyWîD.uKáÁÄÿP_£­Ü’ óKÚ‹#Ê¡ žeMAML‡‹eÄÙ—`Š!lþí=cõ kçÛbšä¦ãb9†püvý˜+¨Iz*ÖÔÄÓÄØ.'FÆàøØdǵÞŧf«—ØÕß¾‘ñÚâVÑwKw&x»)OsUfðfe >£6¼¹ÌÊN/I2¹¸’qyKdñÜ6ü¢I_/!à1Û ðë Ô±?"Û,°Ð¼B èñOnA‘Û¸Vp"NîûÅP¯~½ @¨”6‹£Î%‹„O­ïäžÞ[æ˜læ’.‘­-dw†¢[W&•FÌËÎÊß³õíÛ£Žƒ*5AZ²³Eyÿsòç°Jkà™4¢Ó&N:é—öÀó„è50Oi«¸$e.Jœ—¦€Pv*bLé!‰„s€©ü:£ðIa±Ú‡e…äz¥¢]É+lÀ]ð2{áwÂÍz·›Úeu£rÓ/qEª‰Ù’»æ„vÙlU³ÐV|ÈþÒ¿´¡†"§_a¥ß †ÍTH ë6ÂË Š\±ªnt2E‘ r… ˜É)Y- Eƾ‚êÇàž­«ZÐÒZšÙ€}<Ä…Þß`¬É†ÆG0Ø)),Ü6™2i¤ a§P`­Ãn¸®¦²¤ÈqÍ¢²A‘“!´° öœÌ“å’QZGžÒK*`6¹VÁëòaÊfÊMxÆÇh\¹×\I³–na½JeUbšC8€blEÞÓŠ¹z°ût1U¥7ù~˜Bâ|¶f7-˜A ôO2ããqe™¢P”Ô«,eðÁiaˆO(Ka]úuË–ÂÊ9r¡¢å›”ó4p‰5´Ðv{¨˜)²(Ó³J!š±§‚Ü[:w ð­ÝéÃÆ­ÖêzWÄÕŒ @éÜmD–öéèTcõªú2#ôU_fdx¡ÜaElKÛ‘¹ÓQ_vd‰ÁȼVÐ"ŒÑä tøÁûdpgã--?vò¼¸0/gÈxÍÈ´ =¯AWZfèa}§{‘èO¾€!÷lvwçiáÜ>ç>‚Ͷvož—M<|)þløRò¸œluÁa%ãס ¦A鎡àÓÙ?ô§¯ê ~æÏ¿BÁ¯ŒA]¡àw^Õ+üye¹bÉ­Ù5Ë ~§r3† yÍè±RbÉ»ç"@#ÿÉ!ê^»#W³0ï‚ç\(8T±PD-€ü+clHý×åP󬪰}ëΚÕôô™ª…Q3iÉ|\–€€¢-ãÈS%*â‰\­ üDkùU*6Ñú¯ü˜Ã¢%ƒx¨,Ñ2›?UG¢M§ªF´åçì¬Ó¶3!SU#Úñ£yrq6j9”üœ­ê'Ãh9½¸ožO•'ËH(¾£LbTšxr‹h]}¼r‹}R¹™vDш>3·¨¢}yÈ4Yè,Ja(6YF¢ߪä§<5£dŽüóàOG€¦ªJØÅH hü–Ѳ=ªÎÄÈØÄj5ª„¸¨òĘÙge`Þ9³òT`Ѷ\ÉjH½é «.Ķ#yê{«~­÷綦}³90n]ÎfbSb– nR0 n¥ÏÚ0¸)Kp1D« u#ÎmÅv¿v(p1aƒgGk¸…¼ëz“Œi\wx¼â[‹¨©+—8B1nU*gÖß fß±S]ïâ2öãf…¾þ­ó«ÅžéG"€òÍëc›|ý6Òv=ÂX¾Û+ ¬CâÀnôu“G\Šx+›>ÿ]œxp PŠL•‘&Ó±ÊHÓJ¬2k}%ö(Òä Q^i1&š ¡*,MÉjÞ2|wWK º¬MBooõÁÝ$ÂÀÖ Þ]³!жL4¸´ß…ÈOû¦Áñ‰e-EH#º%ò:Œ4,Þ­¼—Žaq™&"b:†%x¸ y!‚us©”ä…ÏÌæòT(MŸ¬@D;¾|‹E“ž [(ø® û¸+BÇ—„_ :a´§|ámùZ‰’Œú¸Ý e[II^¿”çy7žù“ŠnOA›‘ Ô›:†YãðgeEIž²‹g éCEÍKéÐ&£är8´Ùo¨ê«@¤®ÃPj¢U_‡±un\Q;ØSQKqvç(ÄïÖv?*(U+M!ΪïâäÑO ,»ø¤C×Ðà&¤¡…LŦ\:@•Ÿ17 ¢wú\#Ϊ¦Ýu£UØBAwV=Ô ž9žÐÕ£¦åt¨WÀIÇ#¤víy-§CkR›{W"Ø£ °%yuƒv,Í¡æÿ.™Vܹå…1•¦p[&ƒ³£ZA•Ç%/®ˆcò|„šUI@túY­,‚}„£Õl ¹¡úoå©ý[-+‹¢¢eúì5MÊôÙk'WÑ N?_eøb BL\­­*ýVò‰VBá.6¸Ø<˜H§°tLÕwÇG3Ø š@†ß£ˆòñâmWlwŒ›Pùʧøöþ…æÈU^3**Leóõ ¬Ü-Y±0°.®í‹Q¼„‘›æÐ¡Ðee!Žð*#P§E´"˜Éí¾V…¦V”V@ÀÉ:‡KŽ<Ô{å‘NEBñ·s×ndåÃP€Ò¸v¸ŠFò\hãA6„RßT‡¨×1ú›ú%=¶• R5à®KÂÙJ(ªc}™gf]:”Táº$|ÊŒc›ÅT65€*Aw;ó´¬KB_(*+u›ŽÓ•â.‡n!¨úέÄ/x±åV&1üØsy9·ÇÜ¡T±?x&ŽÑ$ÇÕo6/‚ ¢â7GdÅò£z Á$Ë5¶•fOÆc0‰ÙäQ¨f½'QÝæî7%”öN‰’8“dÕ0?¶­D®©ŠJþrƒ×¢’?a!!äÅÕ…®©/OÁŸ-TÌSð§Š&£8s§T©i/žÜ•%ö"&+þì)MVüÙ‚"-ýÙb”,•Á˜²\ícA€f¹/šÊËŠgK¡&ÕyÛ\“€*Ýì’ØŽÚ%±?*À¶˜ÖX o‘• ÐJ-|fuÔÅ9Ùx @ÆYúzØE(I5ßx£¬Z•¤,k•”$–¾Š5I Úül>ÐÕy2f–ôY,¢Ô&Uk⩃¢Âi‹‚`=õT:¨RY)rav )5áUx%F–T¤´xj¤‚/YÑ]„PÔDçÖ—géȹõåÝÕ·&q¦AÏ:©àîÍ0'ýÁÜÒÝ1™s P—ŠÚq?0¯k¡Hàa*š0},ëº2L e²M(:Jh¬M+èà»Kí`Ýг,ú>a .yGnRWMø»L悘ö ¤¦â‰ ‡¢ª±@|¾¦~e_&@Šíá4™áÆ‚|Š éÑ#à7OLµãƒ›ç[“ å¶… Áá— Å…1… x.Ñhú5YÂk4ý`Ïxb ÎßSjrˆ¾ô”š(4 ÜžÏHãUj’fñ*5ÉE¢Pø¦~ØÎª'˜ÌâMyóöÍ>ÀÞ(CŠzC¥¨<ȉœ«H›PåPc¢úž¡ ÙunCó–/LK/½„±Ña—ÜűC°8€WÌ’®3À4úƒ¹ÁŠ® ›àŽê:§áÁô'‹þT¼øè Ÿã)6©* ã)6Ù9U’G¡döHOæœKXK „l¦ÓI@¡MÇ„Ò_üLÔ>ôfAäþ8œêÆÚ&cÓÄ£˜M¬3£ÀwÐ=õûU™µ?U%U½5°¡þý´BZT¿Ÿe ƒ>ª+¦›‹Šç1Ýd–ú¸‰¤S=pÑ7CÕEhLáö4Z† ñSC²lŠîhÅÍ{Ú<1DµjdÐèÒÀ«ºÝF×_0`Õ÷µ"íÍ’ÈzïëZÑU°î}\Œ®ÞЪ·S’= µÓ´žË£Èð!ýVƒŽºyè;+ºVÂyBѰ ޽ËÀÄu½«z~qw_òy5TT©ä`øêI¡rÊAþujR÷úóP­×Π̃×s­¬À,ˆõò6E(½L¬ëY…¬^YÁµ Y½òa§šÒE¾ù.ÐmÓUlÕzÁ”¡ ™aZi/$3K^7b›}Öé|êé wÎ,EÍye(ª1ŒÚ¸`Fû\˜FíÇY}`Ô#ßûÌÚ¬Œn@­·¢ºoÖ*¿lãøLdž¡h@Qh%뙩peòþ^,Ø R·úù"8õ°Ç9|®Ó©€™ás›šªœÞ32ÁªÖýTªž®…‚;£;bŠí_à•ÿn Íõ˜(Ww;ÌC‰§sÀÒ]ažË^=#B®'ß3W„eNE­°„`€ÚuÕŠàÖݵSY˾g=C”»ß—™«)ôB`ã]a¾‹=Ÿ#¦"üuó¤‰ayëUB¹p¶І2¥“žÉ6¬HÄnÝUe ‡=cu“Á| õÕÎ WÉ»ïvÆ%„žÙÀPТëɆÆZ þÕÑ€ÂþÍ"m‡ÉuZÁ…¨9Å)•U †»£Â ªèƒ `é5oRT•à¹5ŠªdÕÒ"bw¹Ù Xø. )œüdТKŽxЍÞeçOæÒHÏtr¬¾Ä,^C äïâ8ˆ†xMÿ•ïÿ½4ýâPpÏeŒ=wÃTsç‡2Í\Äbõ¤7Ç#"+ûW¾ ~¼¹òÙl%gÓdvÀ½½Bù˜†;ºå´fù›õÒ¹‘aºÖålUþZ•¿{ùrjŠZuˆûåa÷®9óE6Ë&q(ÈFÜ\#å:Íæ× Â06¤sNmÃ:þBØ(7U< ‡ë½Y‰Á`=¹ÈÒ^g?~*ÎzMø+8+'áƒgêLwˆZ7.$ sz~+aDÀó}.HDˆÀó}šBÑÒu[ ’óýÅXÁ¾9ß'Á( ê[Úf²i{?ŒÛÆl„:-²„Rsé¼=½(ÕƒC{äJWo"«á£P˜ãä‚bQxÐŽ{×<".#%¢Þ1–9™,Ñ ¸ÐV>–´Š®B¯ ¬\Üìô§GÐeX`ÂfúÌh1 y™ké“•k,ÞÁˆØ¨¯étG.'…ÑÝñÓþñ¬èáîq ¢&Ž å:`}v&:ýÓH5+C‰¤“ˉLK-oõˆÿÁ&2×)#ù¤±@çÎN¤£V®w"|7 Um}¶üdŽY#ˆ¥‰O';²Z3W[q‰Ç}-Ñ"35Îk‰vPhþÍx˜_ ½è†Ýµ¡ â1lDb\†RsQ/‘i•9Ânê™GcÖeä‡âJ G¹:]òÁyldï´ð¯UðHìåB92^Ýóƒ)¦U(êë†"|}çJùdä§Ë^š[\Õ\ÈGl·Û6îr‰=WÿqYº‹Ãâ‡×60…ÆÊH$¶н86 ÁµžsêÁTqËmÅqÙš c0H"ꊽØA¦Ô[÷Þ%‰0¨*â¡zM欚KÙ‚Cû…Ô©Žir YU5€Ù•CuýVoôDÀuXàÏË{“÷r³7Ü ­¼vÐï×ëûeN¶‰gãA{UõÙx˜‹ç¸ÚB+>'—y¦ƒÇ¢•ÇÖYU„X½üd…‘ˆ«˜¬yáä•ÊÂÏiظ ¥¥‚±V/ .Ö@©U< <•¦Ö+XN3×Ã2ÚÞK_,«€Ðv³‚€YÙ^¿Yƒ·XYîÄŽ€¸†öKÏÈb#w(¸¼ÅàcC•?Æ`þP4·O»°ød×r§¯äúR¬¸˜™ úçRÜQ,>©"c¯vbÜØ5($ö’(¸váN›#ϯ#õèK É¡qt¤›¡á` oˆ7 ±ÌîdGÜ\$ÛC{V§)e7gK{’ïu[~Π#HÁÊdôôÐÿYÈ b˜âdÖ@+¯!hȳE9yÏ ÀXˆ“®…6¨4 ‡p‰ùát$…6nŽh‚'çW§‘0V ]_Ý|öÍ(1«VÌŒmá¢aK(‡ª7-Â:tze ARù4`A%Ñt‘lAAìaMÑP´FM+w¦_…d“Ç„SD61£d²îÊ‹cÎ|öcÂ.mç(ó6‘MšoV™pʳ‘FJöŠÖuïà™Ø9î4Am¯e^nË—G„q¨*º¥îú%€Œ I[x›´á˜¨¶L åU£.^¾¦’Û/a}"ÞFQ_(›‚lt¥Šo“&ÄRÅ·éiÍjâÛt{" Y‡æóÄß·¿>aðÂ}™˜Ÿ05 + T•UÒ&úát’À~³’[8ð¤}±®¢lbX´ Á_}…LJ{a1žÉþÂÊ•bTM”e®’˜4Ó1Wù=* ßÙLGN¡Û\¶c r°£ÿìÓFÖ.ÂbŸi›a±´M ¶Û‡{„¡û?&fš€Œ §±Äù¨…:M ÚønÏÑÃĽ¤Œ4zÔ=ØXL×Z1¬Ñ eT¢ï¶«Þb† à,¼ÇÚñœ,ø òÜ»›C~vî“= ÆeøšKFCÁa…iBNyn†Òä”ݨ¹¸Ñb°ʰ€þ²™M¶„Zk^0XB­5E}‘À•Jòe(hç-ûåUH¯kM£‚6=gÄVkÍÞÅ©A@\œièbaV'ÖÐÙÊÞ{M-£B±xp¢ÈaĹ©×ãÅ1}(¸V Ï„ÙiÕþ¬H…à™Z ÷[€Þ„¦.(X•‚pŠÕ«RÁUˆER ¤ÅÚ[²LE«wEöÝM¢a#cXìÐÇÀQSê¡ÙbÅ>Ñë¡»f†æ¶âêgÇ©bÅöŒi|\e¤?·ùëf”F9CkŠïRa¼ó ´üF¶‰ì_;H#ÿ$è°.å9„ªC¹*tbµhÌÊÞXÍñÑÈ3cc™)Éu= e徂2¥pÑåøÌ× ž¡ÉpŽ„z­²Ý›(V\¶*’j¤œg '†Yø®tÎ_ñy‹ bEç¦Ø#19ƒ·#Y‡“Süü‹ô]£ ¨¨ÊO>tíPxžßF¬a PDÿ&eƒ‹“eè„aýz!ÃCáuðcAį‹<²ŠÄ§› ºãz(ç‘x7I葊äWŸ¾Uq4… …ÖzÀì‘¶äwyÇ0™Dµ{¯ Œ)~"PÁ»º¸dÓ\2 l1.úfÛé Ðî—%Û—í~5±U$°„uÏ4)ûýºÁÓ~¿ñ+_²q\§ý~Óu˧ý~ëW~<žBo¾$žmÒݶF5ÁIw_X—âš|(7ß×§Çô¤xwš*M˜¢Q/Œ—¦Ñ–æL£ÁÍ›ÆKÑBOCŸèK(”YeÍwõM3©©O\ÆëšöA€÷¹•„>‘‚¾ÕúËgK4éäú`(ô©Ëû‚>ºf}މ“êš}'¸*ܸ-ø~=Ñ7F«{u i\aó5ó‰ô°‹Š'XØ…Í|"=̆âDŽ¥ç8Ae¡øbʸì+Ét˜›}% Që^êo_y%Y¶5ŒÔ¾˜Ì¼„×úáÉÑH}óÖd¤¾S%b.LÜPFêQ_ì:œÐXp³€ƒ=‰r Ê„â¯bÒ> ç&Ê:ø‡Y¹“n†îuÅnŽ_ZÛ¡À%mû{C9É œtÑ«=ŽË[aSÈ`lŸÕ­q(l3›Ÿ0ƒ–ŽPtËhÔµØË5S¥ÔÏr؈ €º–ÁZ)/hI_D’an7s?¡ø–’ ÇD¼±˜d¸yJ鬼£3‘€³ÒŒQ„ò…œ…žƒTéÈbÌdbâ°ÕOŠfd:t£´±+Í„’¯@Šgí<3t%E¼d4+h3ëúÒŽ¦(KÉKËœa(êkË^G¸Å×µqºqøX¿½JÍ4 ¸}vòËúfÔ†k¼xÍ|ÑVqP{%Æ•{Í‘êIŒ+lë{¶Ú¹¼Q “¯ÁåÞ7?˜[a22\w3pjÂuYÕ2ÓuQ© ÒK‚­Š_ '=V»BQú‹Ô…~(Ú™ÊE ONø¶R—Ö‡²U9Fêäòtïh&Å7;7"ë¥nYH«ˆ™®Ûy«I×mʼµ¤ë²BF÷Î-”®xšéºÄG¢Í\îjDèMCá±Bá—s¤œ9ά^i:!¶cæý…¼àŸM)š{Œ0mG±·«'|ÛŒs–bZd#n.ÒzØ4–÷ Ew¡8Ó‘ÔíÝ&Ü%1‰\N@QÜðÆu¼P4þiФ–éÞ» dc­Jߥ@¹=\>V^R &èDÂåÅîŽØåHÁæZÜ ÄÏô&«,óH¥læE.x@† qU¹Ìï¶ëA9ü9NV³€Þ¢¢Þ®lž¿žTÓ²O¦n!ä/rÞj˜Æ¡ìTº¢âj'e(<>ÝNÊ`x|ºcÕêidÐÕÿa4úÎãlJoÑU¡Áq÷,âà@µuq=GPm–6IBÀ}žh²¹_/|G»Á¹X\»N(ª#Ú[pAT’f ͱœA žôìFmž-&‰*qó*ö­>pÙe€Žà'O© FA­`Ó´Ä”¶ÖÆÊÒQ»“†¢¿([„,ý..g·Úeéç2¢=ÒE“Äwá~ÈÕÞàƒ„¹"Æ=ؘ¨Ït(èÎ[(`ð]ƒÅD(z…x;ë®EÌö®n”“^gå7ñuìwQª ç¨5ªpÕعBÊt¡£üÕ0=ú¢jŽÎËJ#Íí¡T):1BÌ “öQ¼ Í*mó¡€ør²tÔdÄ÷a]’~Á+R¯0Y]6˜Fº5D'{©9ØÑÆ^£\Ü3]Ý.¦¬P–Ûëd…útñÄI*SL3¸‰ƒ@d'׎Cr3•°Š‰ÞðP:Qãè³\ã&ú³*œ®»rܧru¹Ph1pº€\qáÞUêBÑ=º[Ylj½ÂN_žÜP‘®¨«®ÈÍÚ{]D‰;}ê:éúAüR#ï‡ƒŽ»ÜÐúæ#óní _Qêƒä9¥™M½‹óJ¤Eõ®‘25ü΂cÁê÷t¶‘{Ý#t–BBHWï"2‘i¡4RÕtN;˜VÈ"ó”¶KÈeÉb|i¡Pô{KÏR›ìÖ„¾E)HÔ£/Ýå¦Ú"'®ëCPd™qþØ&äjŽ\–2”Ê|Së\ÜAx}Mý¥—¨go$VM—<`ÍF…B½: »k£'X¹‰Ý‹Š›d>õ3•ž“™É«Pšî*|dµÙ7ê‡.A]-T¸ßj+È•.”¡w ÌZ«ŸËµŠ½˜Eþj{Ñ»0Wº©öZ®À‹Oñ×Cð¢MÎQ‘¿jk,j¶#›Q¶PÀÓ­JˆF¡×A2£®²*N×k@YR—Û|é¢.fg_š¨‹ŠôG­ZBÍø4[†€¿£âí™8<€B}@ÜNAõÈe åê%¬‚ œ^3uI²âüº$ÞÁV­_$Ü.ÒÌæ­q— ~PœG”9°R,ïØª6Ê¢818ª-+±Ÿ«¶½þ°?†¢£"U.Ô7R5)PL¹‰IÊçN\Û·u9:¢(3ؾÍÜ¡(å ¾zu îÁ–¥ÏO#d¶Í¬¸\I mS‡yФŽ~Fõi"H]ù0*T£³jK7a(ÏžìáHžñöÒ©ÙY—[€íŒ¡FMíK½ô‘¡=¬Ýu~ðùä m”÷æ!e–?µ‰uÅg5ß6^†§DùÑupº<*’ã ’œUËqº¡QØ Í^\(>Êp^**¯úè³ë Ø^ÝlTcñ4¦Ê°#ŸFF°ü­|¬3/±¦Ý%Ëç!4¡ÃQ4žØæ¦‘]Ô•ÇÖM°<Ê´L× …ª™MŽôBQaïÉ7pT¼â¦=Üä39Ù …çmÒ¼‚ΛBCH¼5+•‚n8QÐ"ÞÆ¾þŒ-:´Í¶gÐ’Ÿäç£ô [<Æ1}¦tÈÝÑ£C~Yú-, µ~¯Üݦº3DàîKã×£ [( àÎÒéÝnO:N7€;‹©w¸¤8¥MffÕOº”æmáÁÐ[R¾ˆèæ_¦o›JuºéÛöžNV}†äO3|;K´7÷•9Íìí,Ú®ü郥6*êSÚäC°ga÷6ØY%šü(¯ '8§5ò,þÞTý£›$tZ$Û“ÕPpC¢jM;QÏûë…I?ô_BÑZ5Ž&ÚžºÈÀ÷˜îˆ#Šnbåõ‘R\Vê× Ú~DEÕŽ¤pÜ4Šoü¡Xáo)— òa:à)\„™ÉDk(üyb½•%̼ÁóG¼!„(ù‹E …'œŒ8 üÅ¢÷|*EoR.UÀ}3KŽ6²ïÕ¹¨®Ûp[ßKÑÀh•¸#eõ)p[ô!î‡RÛ§2LÜWɈÛóTXiª<ð(ê¤_JQåTÎÍs!?Õ¹ªñ('Ï…•çâSIÉý#ïò8\ì¨(sñK¬è8Ž„ãc‡ßç¥LZŸ¶-ÅqRÇüT°÷ý¡:ïö¶30¨?o!®1|9;»¶RiQúP®¾\œ‚ì-”ùnsPÜz0|ÿP&Í‚;»HúX¹84} t}òÉÛ¤-)×7WÕ'«­…³í™âáA¹c%kBÞÕ0:íìjk§·q;"{¡rÜ£Àt¸³÷­Ø¥rÝ1 êŽë¡3kÛU~ˆ€r¶û¿¢v]hRÛOdäGW{*wÐ(ðÅ46„¹hç^}N +Ÿ0Pí§ÞúÅOÿ|ø‹Oñ3žºãt/YAU‘)æ‡åPPׯø0Ï;¼ &_¦…_ÖOÌVÛÛ™ª íË«IïËnã*“®º†c¸üÀ1¿>zÓÍ9ø–¥YëØn“sLåi·W:æÿ4#Ûa}ž§SÂÅ„“ G¾ñ–‹2›ë0¯à+Tìß¹ bøöàøÚns¼ÖOG},±ì~,ÃHÐZÉaë»^Y޵œ¦€–NYyà#¹ãRôx‹U#D­lPÄÂ3N5ÄâÓ|G!®ª'„²sQ«~„,bÝ‹ñó–CAz¨xÑ%–ÏÖdÄCË} Ó–žKa7ã%^í£¥°×ÊýH©\…g“  ‚k¼ã/±hÈxI®32[²¼ày*ånŽFK«ŽõK¤ŸŠ‘/±è9ú¹Zõ<úý'­rtåø×j}µØr E™$]”F*,¶Îµ(ËIÆŸb±GP‹:›ëluMË:'!hã±M^ê‚G±=éÒ%h[¬ÔW¬g%żzMãÒdR®>xä³ÓA×QÙ¹Ò§²õUû>•bG/’"NúTúB"@ÖšÂûn^_ï[84X.fƒ¡#€¹wÙPÇ{kXÙ±ÉÔ45¾ÞQÐõ¿CøgÐ0r“rÓct„Ña_°¶?b÷tBÑ}(ƒ(&ú˜#þÈIc’¨/ë¤ JÔ—u_†«¢K±u¯ÐŠ<Ó¥r€±«¢±Üö&º@AÝÔæ_ñÉÝØ¢žb+`Ëðm¥"›`ÑÏð~(è•vòÔÝÛ&ÅBêËÞ¿ÒP¾Ä> *øT€Ñ©ÖµÃ1 >iÏ/—gâTÿkáâ1Ë%Øe%Þ*XA:ÊÊb[Å5˜2aHŽ}VI ,ÑH¢eë­¼¡øËgV*KÔ$•T¯Üµ ²’¿|19~®=²áèüŽIÙJ4L˜4¾^¶Ð®…+Ô’O”¡€EÉ{ÝÔ(—ÅÐ)[J5SÀu¹ËgOow%S|S—t·Lñ­‹£e롬¹ÖRÑõ†çà[òw= Eɲ6É@S·BeiòÇ6¸®%ÿϰ0":b>"‚¬ªË¡Ðý\‰*+©\ÁËÒC Œ_Ùv6ЦTB/dÊ•k'§ Ü¡ÈñËø2Xkº:áVh³Gb@Üì‘§ÅÚÿQ«§‘R‘u½s’¤Ö÷>LÊëtݵ6­ø“ût¡8òÆiµq6Ó¨AçÕµ\Üï«Ë÷”±Aýíù&4ÇQIÅw ÎXp÷#’ ñGŠ: Å«[Áqºp#Ð÷zE¤/ÈÅ;#Ò"¾îãžxqVŠ’‹%Q궉~ v&§ä©@©•@8Ï| ˆ›Ç÷Ïöô<¬Êøheˆwf«Œ§ŸŠK­ö·¸Ž”¤j PÑNŽGb‹œà]«‡_‹]Ò¹~r,=Eoƶ!mÕ.Q(xHÝ–¡åMÒèÍdäÚ<57/‹Såà| èÇìî„‚3cîe=ÛÝ¿ò%ä‡?ÁIB¨Ü—î2B]ŠÇ¤’¦àýTn¢‰Ž6aìÍWnÁØ»×›ùæÂ»P\/™„ÜXc ‚»ºŽ]¾/Û¡˜]¹—[޳"›ñ¯öü97iòzä“ÛâüÉN6¨ôô…Ñ|‘\¯¦¶iË Þý´2AÍØãVyŠ:=åØ*#Q—»ºÍ"¦ïcš5ÿDù{ù =œÛ*äÐJþ\•mh9‘Ù*ÛÐ|glÕhh=.,Q©À¿V%ÚÌ_ËR\QïÀ?WZî¦od°÷ŒXÍÎJ Þld°AŒó¢Á þÁ“ì;¹ ØÐ3t¿U¡ç d« ƒ+F@™,*¡ðôfM²f‡§`ƒÊÍ/õ¿çHfÓ“ÒäÍ€ª =󊸲v¨`‡6ön@A©€ÑÜ3n×B™ŽnUÆðCbsq7ª}Ô| ê<Œ$:lÜÒPn~9*?ŒíxV*ÇùÆxÓõIøw®`Õ"fõc-•fQ$¨•¢!ÊVE‰™±ÃT2ÿ»NB-¾/­l®1@Aйý(Nå8$—Ês3°˜]T’qÏb幬ä$•þk| 9…ß*‚±f6-+™!|)ÿL‰œÇGò—ßZ'xƒ”+&•`_+O¥!+¦z×?Užw=Ÿý¢£ø“Þ~s*Oä÷eøý¥à¼þ…s´²|I*~I 6]´‡/IÅ/ °žù’Tò%>gÏK¬ø%q€x~äKRñKbѼ~üÜTôB÷ë%â—ÄDõ¾¿èQüECÞh;:uV¶®/ÊÖ÷÷KRÉ—RJ_/±â—,­?/IÅ/Ùb§U^ÃrN{`‰çÇJ‰W•¯ÿðÛÿý[-_ñýß¿ýýë¿þëoÿÕÿü=ŸøŠ¥Îï·üõßÿVù’¯ Y´¯ˆÜÍï!À_ÿíýÓ_þöçX9Û·ýéÿý‡¯ÿáßüýoþßþúßÿ¯þñ¶ŽŠÞó+¢àçûFÅ›þÕ¿ûý?ýã¿ü‡¯ÿåϱtq÷üÓßþ¼-¿èûj•»‡WãûTÿÿ|ÕûçØ÷ûžaýéo¿ÿ“o|øünªÑb×÷lèûýù¥ÿãÿùç¿þï:Ä`ûÜ#Œ@ì÷ó1u÷èçþ7û¿þÍþ󿈟zçøÓ?þþ·ÿãÿáë_ã+ÿõ_ûŸ~ 5ÿ»¾{´ÿü~ëðGÀ=[Üiýp­ÌÊßC §£áýû¾Ð †³Pø° á÷íüåûÌ~(:‡Â?1¶~þĬtŠ”È °7N œg³|>Œò¡p ¾Ê­wqñ¦åñ©`‹ W B)• ‡Eáê´ÀóÂòP8Ú;DI@á 6 ¤}I¡„&ËÌPª÷êjT8üÛÀaPhPŽhÉ•g+|²wR9J§ÀiD˜kk£ÂÁ_ôs¦¾UE¡ }.Ö¡t}0¶D¡pZ7íÔws*ƒr¥Âµ°°#o|²6{ؑϑ£:°¾B!î ”2©p%ÀxuPáÄ!ày­Qéú®¡ßlÀ]t4CŸÌŤõñ7{'ð~kPÙúd˜ ¡pAg³J5B/6SKP8Â¥è“y°ê“Ùò7­k#ð½WŸ<õ›½zìñ©L½«ë7k3·ƒ¡8¨>œE}LÑ,üÉçj²Güäã æfJ §2›Û‡PtÙWŒ è/,aAÑ5^XæA2Ö]X ƒÂYÓfUj(œæÜòl*U_UÙŽT›Û±# ÉGß…>>.(YM÷9Ѓ¾9ô&½dóV;žŽoî1CáÎM?÷¦¢³ÎúPÚµR]Ï€ƒ…àØ€Ž¶E…=Ê& J×› º³XYã›öeG ¦uW š¡pñ"F¬§SzàPtþ~×XôB™lo§ûÎbM(<Å£³¹…‘‡9Çû?ÁŒåÅk¾!FÑÅknþ´CÑÉéØn móçÀAç¸oö“Ç+}›Î(œñ*/)nµ}°g?ÅUï| Ö‚€3Á´!,ÒBQÒ+ÏÄ–ó°ÿ§b„¸Y;мžŠí€½‹K'¡ ìí{º->3ööÌú|PtÃÒê4Œ.¡ó)·Í žòîTÔtvFðÕOµ‚nhO7”Æ[f ÂÛ挳 Nü5(êß*b Pª^³Øww?Y'Ÿõ[ÎW(èñvw#$b&2íª.y7ßµñfÜÍ9úú`ô<†Zø¤ÞU;Û›±vqc/H_BQ§S޾«¸mÓâ8ŽÏÅ’M_$zA4Š.(Aâc±†¬€@éz òµ#fŸWŸÜØx–÷È6}¦Ptý˜O1yãRÍ&= ;Êà¤w|Ž÷ãí´P¸¯v4Ë%Âi@ÚÑ*×P·ˆ€†ÂŸ¼h#kÉv×yÊP8ÚÁ aP©zWc¿¸¼n§k£ÒõíUŸìÍI<»(ð=GwÖ2ô(z}ž/â9P©T½kóO¯/F_Œ§ÑTl¤£ë_3µ($®ðTò%”øS›]zÉée¥u4Ö™ÊQöèø©¦Þ.¦: ð ½Îôö ‡këF›¶~¬­mj¨¸¸v8¢DÆ©Vð˜ž^Ñ]\ÑQjƒÍõeCoÂÛGÔ縩àa:«n×ÅåúU=Ô6–¡ ‚CCU{ÜrUªƒ‚îm˜÷…r8Z‹&ß4Lô[Üö2RA=ü˜FmŒCE_4ØU Õ­êh»!Ìü-'kŒü½'køÑ¹–æ£ëô- –†×ýÑæðM~F tûÌÃ&<ü|‹æ‚[cˆRº¸q:¢^Œ®å\:$;®ÐZBéªPÜ£) tC‰kÐýЦ0)°Ÿ GC·e$Ü•ÝkÇkê®ì~ ,‚AGTÒÑ}:.[u·_eqû}Dý’ÐâªöME/AÞd¬ÞÝð¹èE?p,ŽŒzs×2&ÇÝü¾¸Bèzuß7{ù®ôG_C“´F ®N(Mkö¨¼³)¨ß• ½ürŒ®G,Åê˙߅¢n‚\ʱš;çE ]æ¡ÇjÙ÷ÍéVnÆ}ñÉÞT@¹ÇÙEwÝ4k[,ï ¡é%šÌ4ÒWú"r,Ö3†Ò8 Cvµ‚QES¼·Ç©\Rt.ze¯Z¯¦¸¨áÂñ©ˆaµ¦ÙNÝn£íðŒÖ­™Õ"ha,Vܦzl'­ShúÔÉfÃjåR0y­Á.:· ø·tô$U…™!”E¥èMºã«ÉºqŠp{TQ T ESØçïL£ÎW¼©x^J£ Ë_›QÙ>èª5…â]ñU9ƒ*Ë—¶..Vc‡CA×\†ïà ÈPð7 މJwC«ƒ7&ó«PØ+”æ3\‡2¥æ75Žz‹ýq’ЮJÑýP|”\󈚊z@r‹ c*£Sáè!ÎdŒÆba™k¡à%* ÝãDÆ”/ÊDÞTâ© wiãÔÆ $ªMŽT¢íG‰Ê-%ž•üÝØ“óˆý¥ã×x…!N.ÎÎöbA(1ÀE•ÍK¥óìlÏàCÁÙÙžŒO` ÇÜö¨Nb ¡4)˜ü(:†•‚/÷˜—_în`Òà …÷äd ²é¡ÂNjnᡌJ¥HÁ¶é˜[„(¸XË·ý¤ŸeÌ¥tœ°e²k\7œ°e*x(1kù¸Î`[^¦ÈNÖQ›VÇuàçâã–ñˆ·¥Å¥äïizº†ÂßS}¹XÑ JµÐ!õ›(}?pú®%¾{ªº[‹9óOÜõÓ¥¸öhësùdÅ€6„©Þx2®0ænó(¼‰B)PìŽö]✶’L–…sºôC´˜…wyÊ úÔ‡0x:‡"ãPp:‡ì*.Ë0g?”1¨èžÙÕ°­3šY\ƒ1}Uvãï^;ŽV¶ Ø@J[TüÁE?P95(_Uó“KüY|±×å ã 1Nà$`¯+èܺWÅ¢¥âòw/pͥΗ6\)èø»ÇΨ~|©Ôjç¼Ûªí‚nãÀÑqLH§f©‡î~êDƒ_“Š+ÁGÒ°Ôe·“Çèã‚“öÝ4”>Fx›ùU ͼÍ<T­²ÒÕ]Xé€d5+ ¯Q%›>™˜³ÕüªÂ§L³yu …£¢¸Ûâcêu?11kÁ!S¤J^&˜ t@¢©é£ˆá²Z¾¢L |(NÔ=¸U èûjÞcSÏŠÚdpGõp¼F(¯ñr’R9n§çÒœlmEùB(èlŠøPp‡¹èâæÇã­xþŒ¢èñULxHÁÅ,òëEÿ€´x»#tIŶíPâö)Í×r6vb%o¨ÉÍÀ!Îj%®n@Õ+Äó&¸‹Goa¿Q}MÌšŸr0fFp ›ÞÅidÐ#û±}_0'ÕaND–G*W¾+žmÀ[ò]ƒ{WÅäÚ¦Â5²P*>¹øL ¸D ËJ¼é¸6@ÑÙŒ#ƒs±ÞT¨ÞÃñkFÕøŠ Œ ’^ /é~°ú h¨@¼©ùÙ1ðcOîBxWñ#| ó4€hÕ»@×c»bC(89ÛUB™x—^¡Äè%”šïж-…8š¤Úv-,(9¸bPuW7 I¤#¸¹+ßUðûJ^Æò:C•ІÁDú¦}ïx`A9¾ó×@nVpÖçò7 ¯çt=ˆPbìiÿb8µG«Õ ®Ìƒjݬļ0”’ŸŒbÖüÉè…G0µõ›;8ÁÝÖoîŽ1òé\™ÃóðÀâ7G ¶€ÛhLÿä~Ù¡ Ï[B‰‘o0ÈO¾ ttÿä~x¯ ÙÍCˆ1jþbî7$Ý¿˜SÎà¨÷•Jž’°?˜Ip)±KP÷•_Óµ¿ïüàèéƒòƒc,Dyõ¿}ó¶îÞà.„æ;­ƒ=FW| 9üšâ[©‘T|õ“,ÄÎõ“}³shÎ.X_Õ–;‡¾Ù—¶éΡs8`ÿ'?9ÖÓQ`ýQ)ùÉøG ÀÏ‹chÕ]{*"Nº‡U¯û³Ž€Âõ¸?K%,¬G=¢ ‚ÆF©ä4‘þ…ô8J¥{Pý(þ·LuM±G©Ú¸}”âçƒÂ)>”°ÅŒ¨j1Û§’·ƒOì/IåßÿV¾þ»ßÉïäÇwòã;ùñüøN~|'?¾“ßÉïäÇwòõã;ùíÇwòã;ùñüøN~|'P~|'?¾“ßÉïäÇwòã;ùñüøN~|'T~|'V~|'?¾“ßÉïäÇwòã;ùñü—}'ÿ1@2ß³³ð“ïcâÿ%ž~ûyéïoeäC5þê¾­)ô{=BÖc·ßóë| ^â+ÏÇFŸ×ïúÕ>…éûjrs¦§ÓãQú¯ç%7„œ§RµTâqBו›`›~ÿYü¬´rnŸbù[È©ö'zô“wƒÆ#ÝÅå¡D×ßcnšBtî+‡Ï‹îR÷/¥ÿzF>ñûsg7”xìôSsøzõ·RrÈgeß< ØþV † %ºÃž»¤ƒõ½ó<à+¿•õ«[øà™£'åúy¸ËÖwÏa$!únîÆ¢h RrDϾëkYðÝysp`Ù×Ía$¸°ßÊñÍ<±]0‚C¤)ë|++G’ç}+Þ½ e/*5ß§|åXdâú|+=[vù¿•gðΚ¾r,¢u_%›ÞøEÁÅ<èEuYÁm1 ªŸÄ+9pŸl½ó¸O4ÎùŒÛ±Žù­ /GL®ìö9òŒOÜ´³ûÁ6±…ó­4÷=“û?}ªéÙHŸÅåÄêV7O8§05뤜IÅs0N|úدÉQüà±òVçl©/Ɇ‚v5rX27›ÑP½n(1âî£y <±2 ENݺ·„âï…o*Ù¦i«ëýæ, 3ûoåä,Œ{۽qŠûÅË“ÞWv#´-õ®åÕÉáÜ·`D2Ö‡¢Qw³ó1«Å›ŒZ]+ C‰#è^Jõd¹÷’3uî#vÖ–­Þr ¯ixÏ-˜˜©ÇzI(ÕÂÀ›–ïJMï{›ùU`´+ãµ(Ððšî®ôÖü<;Ò âEš{zËeR.HôV4WZƒ'¦Þ\Ö E¨×㥤ED¯ùäYܪ ¥æêÈÂkr`}C|w¿rÙ]MÍ%ѵرԞk \*Ňhç·’sá¾â·Rý8_ÜnîYŠ –ŠÐØÊ̓àâQ(zJ,µúVŽž %E¿…ê/.ÿÏnæ:U T{™ÏÂNWÉåÏÊËPtrvaÃ)fb{ɬÏðu}7ØûßJ®ÄìÊvRŠ›Ò®¼©JNµ`×nvX\è[É}£øé·²ÝH7|ßÊÊO†7Šº²!þ·2½$¶± ó­ 7ÿÿ½wÉ•eÉ–Ãúw»§ÇΑÿ?]ê©#¾šªÔüJ[f¶"öå#8& …ªm'3#ÂÃÿnŸÿïƒØ®È­D´Ðvf»ÍS¶M.C»¹fÞ1Œ|â탟™³O-¼™ÚTQ´ÛÚÎÉm\έÛÙ¹³øvÖk·Å¯ÎÜÆåºˆ·q#~äƒŒÜÆ²ý =·qƒÂÚŽÉžØt®Hy¶¡ã3Ùò´UÝŽ›Þ!W«íì÷OPƒ>ˆ-'C“Ñ{9\ ·›  ´u ¹„8ZòA|L dPgÇ\¥2r+=f¤»ÇÆ+¾Õ<ôê<HÙF¢rm“Ýp®Ðâ—‹GË‚Ïçfy…ék jŒ±yórÝ­C¶r•¡ƒ˜zª¶ÈÝ D#ò ÞÛ€epO}l…§ŽfHF‚ª¢Ë ­°FÂódøwñ»6¾S¸¸¶ñq,4‘šZ9¬lÕ:ol0ø)¯DЦÃsY©ãmšP‡ñ;yBÁ¨§Ñž#Š;YÈsxsï’]®Ð‰¬@º»3¦? ¸K'særc;gÀ’z˜ñFDZFøZû”/¸·èжè8>´-Á ¿lH­F¢JbOMH4°oîöÉÚÈPtü°óîx+ZÝà5šÔ0y;VKëð5œÄYñ»E•g¸UH9F°Ùƒr‘-Êf÷ÜÉeóUvOäp~¯ëþa$¾äEÐb~T b½è¤ºu%âà0;:)L‚ ÄWº‰•[[@Ä1Áº:®ÜLN«…Å$Ÿòô½"-‘¨DÝë¤U50´›ý<×ÒôË“¿<~Ÿì›Íƃd±ˆø¨Ÿ¤ÓfçÖ@â]µéCðzXÍ›=ÿ츥ݒâ÷ãÍ]Ó €´¤JDßÖšfB@î"R’_EK.O#IzíéêîÀåˆwƒ«dwÄ0Ym 4šM¤êwŽ1S01Ke b&ù€ž¼ñ”ù摇#óHxÌ© ÝHÔÛjâNS/Z“Ù8ényÌ~MT•ªÈ‰ à=I9QU…™<Ê/¶²!įè1ü ÄIˆËÑûë<ˆ_JâNÍã@Ôª:iãH®/ž²-Nrš¢J–é*ЩÙ@"†8 §¡¦_¶ÖÒ³ 8DÔ†è]ìØŠ×€ jEO©Ý1²“¹4m+Žc'¬(IÃ’AÄåÁz5xf@øS­n«q#ÉWë€¯Ë ¿È(Ñg¸³ D ¨q’ŽˆÁ'I¯[ˆèL“½âRD¯›‘ôˆzŸQ±+êÊgãu¼Ì!ƒ²ZT(Œ6 D¥2‡.UüB]4£®OlÄêüðd,";ÆLHN(ôid²“ É“& ºTä£ð÷7p¨®®¸¯Qóèi1( 7‹›QUòÁîÄ´4"sp›©:VgæéTðHã™’1$ª)µãŃOø$À‡ô¦Q8ÜâÉ8x¯UˆHÁë°Öä‘زñ6w’Š6áÈRºC2ˆngsýŒœ"JB\qksá‹t£•ÈR“)uv{ß?”HÓ·8^#II}.£5Q°#íp éèw¸l¢1}s¶©MËH”!—^@|Rä: S­IèR’ ½ã~L“A‹q­ÇZ$O‰ów‚’ˆÛ¤”Ô•=èáÀ DÃÖáqB®Ìç!©‹s¢ìðˆaYâìóx €îïrÌDæ–J´A¤ë[Oµ’Í|»žÊG~!¼DôÃÁÔüR;B²c'ÒÅôñ)ajÔ—ÊG Ð"*uéê2ÃÊÚ ð`Á  XvF¦a ™î9×<\„T"Jk^ëJì¾èÍ@ÔÉÄú`DaMO“6ù”è;!ÝÈ]“d¥p_óˆˆ>"Å X‹¸ßéKb°˜Ú—†ö¤Å…¦[˜ö×<Ó„€%šËt8‘E.È‘;)„é—ÈH­ ö"kÎzüùÜÄ™à†ÛB@,¸‰ŠD;iUêažÍâ£ÊÝ !¯)~Ù[*P ñ—«•€˜Éu"’5ý°'t@¢»™Å*¿ÖÙà§÷a sÂN’û¶J±—ˆ$8ÒKEº_7¿Œ}^]+¶,èïÃñnë/ÛÕïî Åî~ %¡ÃÄH7ª&€’Ò1þðråDç%¢Çd w 5ekQ)sØŒðfJÐ4¢0wýŸÄÅh6¼4…fŽ¡²ºèäÇp%EÏ`‚åy5Ÿ-á«Ã£"d~|óh7c·)©Œ7<´æ… 0zôažG³2k,Îq|ª½)ñˆ’TQŒÃJ:L<‚2êþ¨VóN.ƒ¨5à/žÜçWòRÅú¼Éc –ƒ’ç Duiªõ«åÔŽšžaYƒFóíÊ, ½è!`-(5¥@JÊNcwW(S+•:¼EÄêÚXz lÝ1U¿E=ñyRK6Ö Ä¶E×NζúNq-‰4@,®|Éi¢×nvÌXN§8*I÷Úšá˜DvS#¡4ŽÆÚ-„‚9jI÷æaÈ“7õe›k? \Q[æ DrDM¢æÉwÙçKQÍw9RQ½ù'ÇTˆ°ù.ÍÓ„P;ýîc]È»ù2µxÌ»Åù› Õ@ô^NãPÒM÷CTOéýáŽU¿J‘®éWw'tì“€êÄ¡\ˆ„ög³Ûê]-êp£ ³?žÃÎ¥?:{.¶"†Všþ¢"w4¨ü@Ô%ÝÆQ dNÇPüÁG¤¦}ÀÕ—ÔT¯[P ﯛP ïïVÅöI9ì b¶ÓSŠÏèh"`’Ðø÷”cÅ @Æ1rô!ËŠ£€tY6hqÔ½ç ‡¡oÉ‚²Hú•eÙi@Š,6×Ýñ’¶”$m'X%äª(ŠÆ2ØWL}‹ËÞÃ`è_HS ´Sœ2ª¾µ†‘»#±ô d¦ûÆî¿‘ÅA§k»¶ãüG†n‚à è7Òx¶ø¹¿¿Õj–ǜӚãƒô, #C­ñAf–…‘•eñyyŸ|É&/äK6ù’M¾d“/ÙäK6ù’M¾d“/ÙäK6ù’M¾d“/ÙäK6ù’M¾d“/ÙäK6!ò%›|É&_²É—lòó%›ù’Mþù%›|É&_²É—lò%›üóK6ù’M¾d“/ÙäŸ_²Éß OFUÆKÒM Ó$y§<&qú¥“K”ê éܦz!ãÏóyÈÄŽÓ YœöîL1G|ÿÉzH=´¯Y÷7¢Ãb"•68#¿U†¬rîo¤qnôAdúdÇ@§鄑H“sp¸iŠÒdÉhä—Ây¦i6ðƒ?q~+¼£ÚÍ’ßK7¿fRn#´E£m Ü¥Üyi´Bоt”Oÿqe Ô¨ôXйR g¹ù¥Mc¦˜®þS]~²„ÿÆöüRPuù\B*ùD©//©@h@ÕhÞ„}…ñBp¨ŠVí§t†H˜™Á 3¿–ccqùHøO ”‚Ž}Œã?Ë¥ëÖÕ…˜c4t[ ¼( Îtc [0íc|€Ø„ Ë;×êÎW³52î]˜Šå·¢ÀHu)aÒ§ÙÉá‰8Lùú2?L×¾ø;z¾ÏL¬L#ÑGí§U:ûmn·@g?¹¿Bg¿ËMÅ@ÂÙï_©ÈM±r H¼)zí Á„&‡®9…ÍgÒ~Â9‘ÝSQÜÕ(ÏeëA¨P7Ò ¨YîlÁµQUäÐð5œ¯‘h çræHÔü[òRÜF¢êxHЀ­¤ª#Né‹җ¢=ü*Õx·çÏC…DE0œ(Í;ÕŖì˜eªk9ôÀ‡¡¦ªùáÞ l8ý “¾’WéSdÐIÛÛpÉ&u¢U„¨:-&"…Kè¨F¬·ô|ˆNÇë2ò—˜~¤º:'ð,uñ4Ú¬ú<#°Y-›ëìç aZ „ç4ÎR¶‘ðÉ.7‹‡<(8±úÚךÅÃY]ÕãìàÕ¢:ºå“ý™=ìf$lVa»OÀ~³Ó]Ù–ßìÊßµáìr¯¹e8›­ì„Ó²mÈùVwÂSî4Ç MèêÞ˜½3›Îó /𦼠@0„YoÓÝÅ>W êä·ÌeÛðp²Ér„ ðÔíu¾Ì¶þ$ÖÆm»ÝváÍ“$ áÆÜŽëÖæI#Œ‹Y%v•sq¥Ý2ãîÙð6ù¿°Hö[)a0måÂ`+3>б1̘Õ%i} ä°(M`aꬡ tªEãguª‹'³a{ÈgXAA ä$Òé(­ øC×iMwí[Ãxs —öQ<ä€~Vén½ ÓøÑT\‹gÒð;U?µd“n#íBOÕ4Û$š ¹ ãñ±ò©¸™^´¢½Œ­9ÅâYlpýPU¶Þ7ª°ÔÇuX%mÇõPSi3›ædÆVP‘ ÐësŠÑ@4çÙ=*MÅ#˜òA¸ñ3M $ª± &„S“ôd¤Æïäôk’%1§6'‰:`ÂËáì`šHÔlì$ á@$Íi$ãªÄ'ãX‚_™H¢j19"ƒ¹Á-¯‘Èýƒ°FšMŒÇeµµç~ ˜e;QÈá+^Ù„‡rÖñÀƒú·s= nÝÏ•37m4‚!¨o,eT7¬.2 J"Qã°E`d´Á0Àdƒ¡Š¢@¾ð²'À³øLL„¹+Ÿ²²òèLHTÀ½=02LSÖˆÂÚÇí »’ñH×o³_…P¿Mt ˆ ´kp8ÕÚyV œ®Õ—R(š®P…“Ó¿NÎøm-‘6‰¨Þv¦‚ߦ~²Öö£SÇÒÙãœö¸ôvéMGK¢øyŽ'ØšAQœ\bu.?#‹b‰‘üH~«]ƃÜêµP#o1£H‰7s›Ç¸ÆmpÎZ"a$»› \o"EeL¦Md§è¹[dÆPãlêZïr÷ÖxL¹!º?¦ ÁLßâ¹ ødê†0ö"Z®5îƒP¦ÚÕÈ“]æÀ~…KabÆ:Yc<âht?UéDÅÊИen \Wº¦æu+©§{ãl¢ «Ð«Â¦Êt¡×‘A=êé•UX–ë@íº—íeB% ˆJ¢6õí ‡–\Öª žë¥Oe(Èañw‰”ü­j^¸ãµÌ¥þ ‡!«àõðîx˜ܰi 1*I®p¯ ˆjDa ˆ`½iB4-ƒéO3/ÍÙóÒ`R:“¼°“'DI‡U'Ò¤ê:¢oš„Ý´°Ø±»W¢UGæš-ݫеV´o/ ©Á «F"¸y—öÞÅK5§âÜË ³‰»fí×ubó†ú½Œk+Œä UÈÒ/ó´0Ý ²§àRm *Q îåÔ“4v£ŸÛÖ¶ü=Œ^Yq) Ý ­‡4yíîloìáC:ÿò¢#=ÌSëBÃÉ2sí2N]ÚçÏí¤WÝ«ô ‘ño¥’â4à©—ypM×Î+µ w_]×±@J}GÏ]Ñ>@,«ú©!@–u8"ÚFª¥ßáR3Âñô­¡ ;óš¯è,OÈ(½®e÷i*½£#Ô+ê ègCߪüóè+L[湋¤Ü'$ð..È"¶®)B¶¾µÀ–⟻ÔÚ‡i£w… 5ž‹ˆ®tt·Œ¥ SG!Þ¥°é1òþó3ÇA9ºŽ0¾«ò¡¦ XôªHj䕦Ùæ àYÅ1âD&¤jÀ ú9ý½Ø–0‡Ów8 á©ßi”wª!€ïTuj»î$[„8=ÓdNB„$ê[Ua£ÎI¸S™íÓ1à׊ÓÇJ@ Út¶sIÜ0õ¥­XPqïà$xuM™ º¶Æ‚ú”ðŠwÜ;ÞñèÊõåM¹ VVa.¨A×i‰Ó”Ì;Rz¯ÔC–À¢X÷BjˆÐœÛ5©óhˆÚj_Ê=ʸŽFÌ´ÖÛ¹åy›ËH|ÄÇá·sïfMg"Ü®†™I±W¬WÿÔœ»B…!K›F˜‘+fìmW9¿ÕM¨†D/snc´éª~(±}×rúþ®r/F‡ADQìûj¢ùÜÖ‘ëLÈÛ"éïÏ „¹æ^ÜF#X‡z¿"PáÉØÅyNûy…ßÊsC0yÚxc¸¦Ç…ª0Ó5ýÎ!*«DX¹êÔCX®tÁ8'0ô»]…“C]íz†í:€ÓÇAD€^F)ߪÉåòpX4z¬ã§=3š8ÇYƒ5¦L“ˆª:6ò~"üYU ƒµ,¹E1¥Ë,Ë‹’¾¥aµ0~›l~‹’áwq½)qdüÊ«¾E¡¤Û#﹜9—GïOö‘ÈÔ—44ló«ÎUévÔê¹Sá⦠ž;.î`îs»ÂÅͼÂpqËÏg¶?lJعUáâÚû Eáâ&’i€°C‘Z̳ä³.>Ôò€0\Üóƒs˜(ŠÍ>^êL~[apŽƒ3ëüœ®pñ¡ŽêH¨ éÇÃÅ-W:§*\\Û‡DÍ ²9þÎæâ/ÂØù;ÛiñBjßž÷³—^¨ÖhT}I/TZ= Kßú[XüÙOX<åJg;,Þ’÷³¿5 ;›‡O.=ªoéÀ~i‰–iögC“ˆ(8?" ÖrãV*ðRâ€è#úK$‰EžQ+–:õmÎåa’ÝZå¢n[º|VQ›?~&£7ƒÌ_™$IaÛÂw¸Ñ™nB:Nù ‹3Æí¸—ƒñRßâÌã` MßZºzW½v¸ó™ íMÿ‚#ÿ G¿¬>|_ g’#㼟q¹ù›€VFGÛ΄œZ \D¥rQï1S¢IB DÓ§£#þÂýØ7ÒYýŽi¶g4Ö¿ti8CËžcñïQÒÓ é<1¢žJ±I‘®N@ËœØY$ýDŽß©Ö@ÔúȲ0ÂÈßH Çq¡§7•Ń—úF´9UÏiÔ"¿‘£²xÆ1¾‘¥²x#µÿGÈËÓäK2ù’L¾$“/ÉäK2ò%™|I&_’É__’É—dò%™|I&_’É—dò%™|I&_’É—dò%™àåK2ù’L¾$“/ÉäK2ù’L¾$“/ÉäK2ù’L¾$“/ÉäK2ù’L¾$“ÿ€d/øßÏ"ôçÿû/ÏÇj _ññ&wzŽð¡«ò²ÃÇçÎÈ)µfÇÔÙ ½ù³½M²º\Wº\_i@T{°ï¸!ƒ›md} › šà ¯ûa½¢ú¤º $«wئÕåÊ2dU³ë²ó©Û/yp4 丹ÝJDÝë‘iµóQ‚Q Ù²§~Yö4Sök5»`¥É™îVº~X}òä¦ õÉ“›Ñd&c5{iø¥o"r)š‡®@5»éÍj$U»Ì»Ó¢QWé<]·~Wã0C4ì­ÊÇCFÕG4.¯¥ûÕNZŒ3aéTµQc‘õÊ‹ó@ ê•w¡9\Í>S¥ADó¬ÍN êqwlrÔæîUQÔ™n®¿€xäf(Ü<5¢ É U?ÍjËEó ªH§Ð´UW›Ã#E 3g1á…Ôlʉ™NÔ‰fkr̆Âç·9ŒHØÁµ¦-¢£x ž®­8fâéÚf­h-§k²5Å+&yQ>íÁ0 ׶Ƴ˜+V}Iãååá9Ï€9^Öæ¨ª#û ªšºš–ÒÚˆ&g—+M šœ]fýÔ6s |Ø7´™S`Ù¶é=¹òÑŠº„¡×Y^XÙ[µåÝ®ÂC1 ZfÐc€–`…\ W_š4=k¹RÍcÛöþMáy m‘2©*L˵9´—4'.K@´i¢õmúiÍU?£W'µéõ]M°rãû»¹,ü“u«¿è$»3ΰB¤¯aÉÅè’ý›ý̰Мõí¦†åiŸo›6,aeʦ%X+²Â³¯[á±bm|hk®ÉÒ³;–‹î)>ýмmËç•ûãîçÕýã „ŽgND‰M®ÜH û›þ´÷ M±Ñ啨gÄãËöP»]í%]ÜŸ—u#öF¶Ì½13ÒIÒ3óq’Ôg– ÷ì?‰ÍZI>3[V’¹ß¬$¯áÚ(ªf*r/éQÍ•ct:sbKо¯Åõ|ÈQ6=?±µ©ÞDCþ¬6Å™ Gµc7÷Ñr‡l§µ©b¨¥ý)öëèÏÚ½C&ûöôUÅF /åälÚŸU{—dÀåU{IÚs¬æRæ¾d¸Å²ÎNfÌÀQVw2Qæñ¡Å®hµW­¾%ç±½Sdkö±sïòè!Ž{)d‰PKÓ!.ž»!K–žãzïÒNìÃÞñØŽ1m†Må(YüX‹‰-a+¬kÉR6­‡cû:|…›{µôËÍ8Kxc«4yãÞãzؘ)ÄÔÎïîÍ»QúÐCüí”z‡½›1bdõÊØ›$Mˆúƒ¨™enbo’ Ü™üVº¹_ruG¦ÍLÒco拪$à P¨[YÉŠRÔÐà O ˆYQGÁ7'YQGÁ7—ãLp²"é¯\“ƒzÐÔ"%Hm}(Ÿ­VszGe’ŸÒ‡À–ŠôÒÊaÔ DÔPínbCQCu˜68H¶FZROä0bIuxD‡ˆz™Èí©Û, ±ŠX‰·#¨,ö@”BÍ^EÞñ•H”B+.ÍAr=§D{,|ÙÍ]ÈäþFk q3ãJ<ìY•z5LŸM©WÓŸÙ”z•ÆT„Z[I‹¤Œ"’º®^z»rΩT°£þiÎ ’ó×öp‹'ÇQ ')¢q½š<9£Ej™xL“š$›©Kž<¥r’ÂAX.ˆPËÃD¨µD˜ß¦ä£$ÐÙ‰0¿-™ˆ †v&º™ÚܘßvÜÕ¬®ü¶ki…8ÀÓ€'ž0"çÔÇ"ˆUˆø²+8qÈ©[ Dya«P—b°òïüTܵiêï1—ˆ™ùHìÁ"e/YÖ‘d8¶{‘Å BÄ÷Iر˜Cäˆ[¹¦ºmªÍâµ™š‚(Aõ•"“1ŸäEš‹Ï½ ¤Š2ºIâˆdC‘æÉvo·x3 ‰i>^&„PÝHdCÎmÒ<Ùøˆg4i~²ë™'IóŒPB¨£Ió‹ÙÅ"²¨æ«Zδ·‚!›{û½Yœ«¹KØ4#o‰x3AÁ•ú™'Û²¦"2Q™"ŠÌÈ $óœ£ yÆOmZÃ៞…Ãó ~†«pÚáÁ\Â(äÌjLS’Òi{J²"µõälõrv„ÛòKÈ䤌é´)¶‘B iº–Ü5¾'îÂ”à–š;§ww~×)Á=¯¥àzÊ)õB‚gJ÷b`¾+/Åâ‡[J÷¢~]qѬ ù†dƒÈ5¶‚’”¡NÝl Ü?Db²%w<;Cª²%wPzÒm/wHÎlÍéJ@¬¡Œ]mD<[w_ %õ}Df{9ù>9!qúѬbç9Õ~ŸÜ{ró[ ^˜ò]$`³k°.)Ùc‰dè*²TŠ›¿ÍI…Öˆè.)›Fÿäæ·îd°÷Ñ·8¼!|m#³22|#‘¥Þ<šÕÂyhÒ®ÂÉ‘ˆÒnnI@"å¾õ3¦Ü7Q¾A}ï" þâÔg~)BÄÛN]7·]ôÒzc”»eÝ“eÚKʺ§"ì ÷‚>ÈR„½È¢0ÂÞ¢I QzW¯dwæÓc$‚Æ»²µ?H¼=1÷DÄz·îH”—_ðV€½yìÃRï:ŠÇŸÌ!%ßÇõ>j¾9n›õÑòÍ&“ž*tRû©B¿lyCIñOœ¬Jäv#‘q?[HdÜ£]+·½;v$ Dhû,ú ãêæ(•x6WµÚ>MµÚäÂRÙ¢1_FP¦ M|Z` dÆï¼áÊw0½ÀrãwnÞ1Ýļ@âŬlµ±cXÙ8*ORûêyËÁ„ DEZ¡Ãù#o™yÓ}yÎ$z“µò–™¯ÒÍ> •bYê šØºyÇÜòì»h‚ºïšwLbßÍÖ•S°¾ógã߻뫂]ºéL„ºÆnR 5®¼òv¹ÙÚ7O”ˆ¶¼=¥ $.eý0Ht³GÁ½êæUÉwé§¹{¬ ã¢î±†4ñƒ`L·w¬Yuíˆ=üA¸ÜÏrˆàމzt@HôÄ×ê¥QŒl 4wb• @îúxºû¯ýf'›ˆ @¢[»:Šy!Ûýn"Çýnåì¹_QY¥dÝ7Rµô}É ài âÆÀtäQ†v¤ÄaÈÿ­É—^ò¥—|é%_zÉ—^ò¥—ùÒK¾ô’/½äŸ_zÉÏ—^bäK/ùÒK¾ô’/½äK/ùÒK¾ô_zÉ—^ò¥—|é%_zÉ—^ò×—^ò¥—øÒK¾ô’/½äK/ùÒK¾ô’/½ÄÈÿœô’04ùù/ý׿jùÁô?ÿ׿~þ·üõ¿þçú) Ø}ª×?þŸ¿*?òÓÃOúó;˜ê~Þõ?þõ׿ÍþŸþñÿþ…ýýù—üýÛÿéù |ˆõ)‹ü‡ÿû¯«ýà³ÿö™Û|>ñ¿ÿã¯ÿ󝿡òhýtÿú«Ã¹„{]cãø÷¿b“(–ŸNö_Zôpk=^É¿ÿõËÄåSà Ë49DžÛÛ-j+jåq”ÚJMy\§¶Pgª­0“ǽj_î:=žW[¹#á‹Åç:6’Ó~q µ¾Ý´¶Öc@xš´±6šo—®õÓÛÈk©Úèkž¨ú•8D»°k ¬·£Ø><‹~\Çö‘V:“íÍ3íǽloz“‡Ã/¾·ÜÞ̬Ú{ÉíÍ|, áûd3µ½eΘ~k{?X½ö–MêŸmï¦2·³ÛÞ•VoS'P[¡áW°È'] 7ÌÖ6!äÈéMÑ^dËêM„dŰ–v¦œÞ†:Z tî“ø*áá7Z•Üóô¢  뾫KÕ´îãKøL¢ìÜÇS³´£K„gÕ@èÜçSÔ=通>‚{ÊÓ;½·r €ð¼nk»D÷”‡÷Ö>N |—U廙ÀÔýÉEÎGÃ{V½ËúG€]ä|ž¼Ç¥qZz5á… y÷ »ç±|܃Sp…Téi'úqŽÜØôkD†>ÃéÊã@ DΖjCÜéc¹ÏÞ¯Ë=*{ÄôÃÜôÿOÇÌÝ嶦š@¢:Úws÷Cç·´æÜ†¼éÞ¹Á¢Ûe­& @ÂaÖ>¡®÷çe%º»åÒnt÷¦î%ºåÖpŸÐ‹ÁÖ¦@ª¬MyŒºÛÕ3t1ª6Žƒ~^.ªì™þórZÝm锣ËF†­»É”qù¤uc?žÊ| æÙ$¢jÓš~ؼ ᾊ™·NÜ€¨–`_‰¨ImOUøç‘Gîu¨‡?;¯f»Êëþ‡úÌ’Iîq ¨“?;FÄ]ýZçv[¨jbs{”ª]¿¼ýÆ+9›p¦íææé) U“ž¹\ϱ ·3EÝÙåÒ`vN¿_Ûà«+<·¿±zê²ØûÌ¡:\Ýàg÷«,œK™ú -ï½>ýòmüô{‰t67ÔÒ83˜Õ/û^•ˆ^\!EÝP—=î§âª‰Æà%õ9|ÝÓd‹uɈ„_õÔ·då8H8hsìÖ%•ˆ~˜;3ðÓæ-¯Ë  ,R 4÷V:Ði4÷֑˨Q\ÃüÃuåì8¦¦~@hî=5b®£‰Ô˜~¨sdX>ؤ֡9êz ¢®ql•~åd¤­3eXn²è:òzôJ |&SU—<ìGU°äa?jÞOæ`˾ TUß)œ uÏéb.'„¤Àµå‹Ý=«ŠyÐ&Òô­]ƒõ¨¨Å**2[½£UGÙÔ²3ÂÓ:Ú0NqH„†?Q^j ŒSt2š>ãuV㫊½AäÎ’ÅG†‚$¯î‡‰|ÃK4ÑO ^uôÑC}JîE Ý"Â6$bX¥1%—B:&Ç Ñ>†ƒ`ÐgÄ ËJ€Ä]ýw$èŵ”zHTÒáxt#QÊÃc&(fŽ‹B¢“ï7Ÿk°w‘ú‡_î[SSt5QoûÎçêÌìÒ9Eo_š~[§qènÒ@"æ°{Y ¥Ü%Ãß1.t+ТˊïˆôÚÑeEÌ&f«×H ÝûÉè²â†›÷oÄXÖ<û÷)Â4-¯O)ô€pÓsq½ÝˆŽŽ¹¸›á¦M{~"2°y÷. ㇽ„±¸ž¨O…M pwë[S±¸Ùb7ùöÓTÈý%«ê<2hõhw6zÐøŒö›ÐFƒ­ÞLš ­ˆœáüL¼«ê 5ºÔÈ5%4¨^Õ:“d÷@T•uÉÈÇH´k"ˆ Úê¹1ºâè´ËÍ.⊓†€D€bQŒ]ˆPÜÆ lfn ñ>‹76ÐÇû,:í$¢UËÐþ1èìK6ÇÅ=ÄXû—9eDÔõ6o¸8¢*¢P+UÜÅ€4¤j«ûZð>î¾–¼-$‰Ñã2ÁûL#Lð^.åCzû0ï»c„AŸ‚@q?Ĉj€Ðñ]tÆ’«õ­Áô›­oQ ƒ0ó6àÑ>  Z %]½à"“IêzËÊI¤­ûEpë H5€Š‚Ìv_©*ÞÁ3@6cßUI?€£áT¤Ç_ÿ ÿ¬ú‰à™3~‰hxt"ÍHWý1€9.ÂîÕ|©°[AÝÄÁ¬eôøÌô•&ã&‡Õ}*$elgÆXŒÒÜ^’A“»þÉôN€Ó ‘Ïf‰–³œm×N¹ €¨Ë1a1ÆãJKñ¶ÄÁ±¦32ä>N­W€`B3ÖpïK ˆ ‚‡š£µláFjV@²ÌG+Y1óõfA8Yý¸s>Ò_â…øÏ “T¸>HvÔ“2•Q¥Ä|€î!!k>Á3Öêb03¢òC ÑÀœr”ó*"ûU D¼bIà¿Kù;¡$H&“5ÕŸú×É)Q_ìÑÅáI ©@}Ê]%éB/Äg=b8úb&=@.5º(¾&8½æ"훼M¥^HuÃ~x}ñ­ˆœ'+ J}ý±~#>â2ÅÿZùËA‹:Óãh§\4ydkTT³Î‹ŒFàMX $x¥&µu,xJ}ñÞˆœ7.öæÏ¹¢Øm#ÁFMŽŒf êíüåù¦ó²Æ‹òȾ¦ÆŸg½8ƒä¾H…ø“äÉÌ1BòdNS:z@`œ‚æšGͱ\/d ûÅ”$ 6åÍß Æ¹^ÚàS02¯‘ Á®œö&+'³3!ö§nwpi†¨VEƒ»s=Fm=£K±Þ j¬Ù¨rNßðÁ{5´Gò„§1…q6c£‘ž8kÞ³Hà³ä=GncRurìÈ%6L¿'8ȱfw,þ«™ÃÇH3·xyÓªýè®°Èn¤1ë#<›H¦s -n¸¸¾òºÀ˜>ÍÈ2¥:¿%@MlˆInZv A?íËr‰ã3SãÀ0“ÜtlžIžSøñ0ÉÉ«»±a"¹°æ0‘Üfاi&¶#ÑPÁ}ßF¢üòC–¢lêÍ©‚h௚=O„4¤(9¿(’;ÏA‚’G!Û¹ª (Hý1 ¹ wKH…¤pÿÚý‚êP¯]îg@T<;¢²C½Øs3.D|ËŽ#@BÄ·vþ²TÏ+¢-ùÛÊù½6!·ô/Ç&}(2½á:X}žS‹Íc*(;oîpò[þ2| íÇÈÝ”®Ü)¥xöæ/KÇ‹½’܃í¼“SDÈY½‘{$]¹‘Ëíß–§;¼ ÚsH±cf%­6[§9Ðßj,=EºYU˜pòï’;ڔ͚x†}ñ"5°–„‡3(†õZ´¿M±^øé,Α r}ƒ6YkþCS!è—ÕªuóMd¨y²·3lj„·Ó5£Á1gÜMó#§8 ê[uжKëP}9죂ŠÈ³-*‘üÒÿ’ña ú·ž¢·JC¾üL—ë{ 5OŠzHãMƒ}W·u\5ÿ„2þ&­£{KGçVØëM]îA„¿‰Âê&0áÈ,šÂs>q齿*µ¢Âa  n]Çs@|FvdH`šõ¼vB°&NmH ÓáRdHàÆ „†žsÅ1äO¸!œ¶üªN_ƒ¬ C…‘ͪ™1ÅfÉ#*d(©A‡àe°£¨&Þ.¹ßÁâ&êþjê8ÖÏxٲ䇯‰“H(¾ó GÈÑMÕiFU!;¶~ÐK‰7ÄTD°•Þ-ò!7 4¸õA8šÈ‘‡H•”mxuˆuS;ßDMÎOåt*Õ:XOËÉcxñÑsÝ|ÎaÖH|%uˆPeêx_ÞÅN0臢M&¼¯ˆ@s©x馭µÂb/IæiêðŠOÜ[йËe¸òo1„Œb ¦‘tÛÀ,ÿ†O÷$‰„ä2 ‰$ÜÐî6‰D®Lw»|!|­t©1‰„Ùõ°¶9޳HH{Ži$ì¦à¢#²\j*3w8í®yvZN¸&^oB9“>zݽéZŦÞiý˜Ç)@h™tÕ‰“¨Ol#˜f$ÌwN’wú¢ÏÏIòNßôù9ÞºX¶ú;Ù}ô7‡Q’øýÒ¯ð ³.ÝV*OŽ…`ÓM*Aq 54›ÿ–t²–æPâGŒžæP/:Ù¥¥”éd“úÒ3Ü¡b¬Ž¿W¸CíëžlÄ®œ¬L$£UëÞyÕC÷Þ½]ï7Rjžæ¹BT'» q©ŠÀD4;‘üàæ%¶Ñ$ˆhv³Óü.O|‚B-DoeNz5ßáäBd7¹Œö$3'rÓÑË[Ú‹$Ž@L‰”%ÙòÑ+¨•t»"?Ù,pyÉi†&ÕÙUå·Í»[ÁÍP“.ÓP0à*€Jk >ÓJžÐâq*v(—gð—S^œœQ‹^jãpaã·ñí\yâi9»È   yœgá&èiú ­•`ǧî{wÖò•}ÝîôÄü•å·åk;½©xR¿¨J¾y \ó+„Tñï!…hK.„ê6ϲáT¨±u_úžáÄM$m.%€˜pM'ʳ³@j,ɼIÚðQ 4îMDz¸€£šÃ¡ÎˆøÖgêw-?­šu+eé)ÌÈ‚2$ºEl¦Rb$é)!/9D$Ùª…Úðìi“ ˆ„IµÊuØ<.¨]èµû¨oºìvõÍÝ‘EyÔ7Kv»Ç:¤J¯Î¾­]©[v»ÛÚÛüõmR½²Û]’[µÂVÜÍHĈn»K"M©Âû÷—R ÈL5Suw+P<…Y{jè›Lø@f §¢ßö¹(ÔVô3îF6yøcïU—:|Ÿé‹ ×&`%šZ¨Ê’¢ÉŠ3 ˜¡!‹^º{yYt½ZöÚ;»îͨӢCK;hØÂz5-£!{‹±:m¥¡Œc©—Ô¢Q=÷ØS‡ÀßJ kˆò¢ÜÓãHFú`CÊGƒm-N¡ö‹šnÚPFa¤ã6Dƒtp÷à©¡¾Tô™ú§¹7‹4p÷Qûóò‡Ì‘þí[s|(!£·O¯qè'Ão±-Kug•º%CPa†UrÚšC©ë6SÊu\Í#j > uˆBcM›õŽÆ“›qé"°¥$¥„ˆZÑäî.W!lF>]j_ôuKá–ZMó„oÏnµ{ÉѶu+þ _$ J¹¸´‚ß½e¶KŽüÍ2[²¸ÌÓ¾]®{!fb@s÷º‹ìÛ›ÿ¤w»w ¤Q Žk·Œ °¦ZVµÍVC³²ÙT‚iV¶òÒTw’TïÌ7°¤úd¾ÁlFœo F¿oæ¨[?E¥©àÅÈ Ñå ïÍWüèQšѰåröÄ8@Ç]EF=@ëÞôËjAGå™Í|pµu2fH(xÂÒû'›ò|Þò±:ûVݲwH!óg†D ío§'sM¥ýu„ŠÃ36]»HRy—îØ„qx0A%Å÷W†äÉ´€ÃTœö&¨d"HѲ£nUø+èKŠŒ(ÜŠx’FJ× gÞhQ‘½ÀáËvHA}ÜÜÈ/Ý/7·¨^ó¼ÜÜHe½YÌÜ›IJl QSVyÙÍEß°Eº€~»®‘!ŠîJÛº ¯á{†Õ¨¾ª…$°ƒ[šè5ÔÛ>Ä´M ñ¾– IËÓ›Páèog»nÉãg¹nÉpxO,‡ýõ ñnle‘D¼uÊH°÷J»ÃE6 wÿÒÜpìý'} ‹áêÑÊM[úèôQ±|¤³b´ýS9Ù $^×iémHËFPÚÕn8¹ÍHp{OÚÅÉ\yU:9H‚N/»8y5„ÿxS¶ ËgC‘[%¸ü~t´é]o×K ~$E[óÒ=Ò‚5ŒD‡jùAºp¦D!:!c°÷#  u¨ÓH”êUžRº‚B2!~9‡ŽÇwþÆ*Ê 9½_2ã Îù¥4ápì!Eoê¬#€)eˆ\{å‡ DŽ·—yГØ3UŠ¡2í®{ɰœ”/þ“î¬wQÌÒ»‘õ¼ÒÞ¢:Á"ö¥’ûÒÒ-ò›Ç—öJ¢óøÒ†b³ö—/mgêH_ÚNáLiL˽zH†lLk‘Ç—°Ò•I¦Û×"¢c¿ôËshšªî@da-«-(¡äÄ~i ý”̱/ƒP¦,AO“:, /¡V8‡f­É•!x„&«¨ð4ŽGhË:¯ú\6J±™LùäAíÚÞûlªCÛµ·.”/•26Û0S 4‡÷ÙÞ|Þé,³wÛãFÏ@f÷XIÇ»ê#‚'HëÙK:¼0qô3t*-ú $$¤ýñ’ÿ°àª^ñ^;w9Š·© UV®‘Íêi4Ígî݈ø‡)`Ý!!Z‘‚†8«ê1§xS;. ^ðæÖc° „¦tlçnlµRúHHÍ, [óèfqäÈæ1DpítƒT§‘î Q)ö¶Y[²÷™ msÆ8Õ67gu¡6ÙF¢„ç´…5N…â+KƒLàCø6•H(®çvÇ@®Ðñʦ|sã8Ô¿*Q®ËR!H´M«ˆÃÛ>TŠ«¦Ðfær(É .dˆÀŽ(Ï•FÓk³Ý-¯€D?ºf†pX+?¼Ý'-‰å×¶#ý’XÞŒ§ˆˆ×kVT T¬—ÖâÀ8ÓÓ}u Ö«{ìŤ•¹›=­iaòt?)Æ@üTE‚uO¢€P°>TÁ ˜©.TÛa 8©L B;cèá6MV‹ˆ=ˆÜDâ½íëNK®n ,©ÁÏ%cêÖ2§Œª+utG§yè˜CÆ]ë/bÐX`xLš’ÊŸ‘QM– Ó~øS]þÉÄ¢Ye¹à)ô\ Ptq úÆÏ'$þ9ÝßÇa±\Ï¿€ÐóÂë”Bû¥@ôÎÆâ˜z³Õz/N3,#ù$ÞÉã?¸ žw¸â Ú€$¿|¾ûÏã”±+qƒž¢z)…áF7€bÑ0(o¹y£FPmÃ×Ã1-²k,™5Ô/] Kf Ùo®TGµt ×ÁòÑ{ìÜ o’DÂPÇö%‘|C»•lÙ}ÉnÅ‹èty§øZCv+™ÕÐG:®¨sê]Ž+^!ᇖ+Û“2ÈÚEÓ^e¹r3óGFu6 ¤ÛX†Hc—,Ÿj$l1MmŽ$£°h¬Í­°mÕó•`ùè[¤.®ô¢G°Ò–ïŽÊ¹M¹ìL?{´ä«^V!ÄÉŽ>·áµ¶“6Z“£yC‘µ¨§Öáx>]_’¿]ÍÈ¡zéáÕ2r¨Ò½¼¾­ÊÜêµ*; }4¥¬äkØódiu!ªì•sÕ°~Ñg&ÿä:¡]ýeò¹^ñ~Ût¨²JË{ áêÖ¶çÍ•<æÕ¶ ¢r3{ٿد—ƒŽ(ÁüqܹÙi‰hacÛŸ¸3Òú‚ Ä,2 îõ•³vdÚÕ TÎZã®QØé3]ï*Tj®ž=Ž}ñz:u/ ‘%wè]¦¬ øîˆ]UÒ…ÿq: ]J?&iˆº #¦nòxu©A@Þ›óúÂA!\Üœ×7X|6zsÖš#­/ÞܨNW½TÒ™H¿Â•&•Hxxî´>*7»+¾ÀJ% 2 Ãèmø¬ H“·ž‚ L³†ÅjH;Œ6–Ž|ÈHäCeð™dý§(×ÃQtÙÿ?£ǡ§Gv•©!R±yreDè»!'Ì$GW^™H–Œ®TkNcDdNv™'ònö–Õc{%Wn9p+}Ö¾xS½_ ¶{ä#lõ,G D¦<#•]¹›Ü§ÿ¤Õ£I‹ˆë¤Õ£éˈô¤Õ£Âb"ô3|ÓòQ¡Ñ)LŸÌ‰ž/UdîhšNR¹t*ôxÉÕ{Ú„â.YÄÎL½]ƒ#鼎_]]¶§ 9ˆÜÔI@¯Ûžw«:Áx•t'UÆí*i{ªæ¶Ð"¢øÕyä{j³T¤ÁÒT~vˆ ÒeW ´=5Á³a¾†­>C®§^qÔ—XdÜFš²AãÒ?×ô Qe×rõ,*œåtPÛ×Ùú¡¼´nµ³‚{Y8'ƒ{eOé †¦±­½zÌKIŒaúÚšŠŒ¨aÚÏ^w[CÍ.E–@b Ûf!Ö/oW"|LoîêÔÎŒ£4C¾bU‚/ÆÃr„,Ó)ÚŽ0@èmrT„5ÇÏt·Õ>åmõ74,'YÆ]Cö< ¨•¥5§é-Eö×j1]N“{:Lµ] [qu›´ñ 7§^qŠWJuo“Kñ>êÇšUz¯#]<:çí“e !êN/xd’G—“žòW\êÇ›þò(îåzåÛöøé##=\"Ó˱ê¿Ó®â@¦~¹qì:v`‰¼ö¸VŽxµÈ…ßä¾Ó…ß>J‘ ¯o)6½¨S ‘:sã#;¦ˆ¡oT× ©R@T…gßA÷Ógx>¤ë3”€Q’u¡ ñ®:¥Ðþö˜=r®œ¾&@bŽ{¦(Mç2_6x„<‡•…x„EßZœ=ÉG àSþ±±î¹²Ê?–”ÉE@Gä x®ÍQØZ 1íIû‘z{Ås®^ŸÝµÎ9z}G%xÎÖë;âÀœ£…Õ1­èœIKÙã Á9Ž˜ÐnÜ!£ lÈj üßS¥ŽÖLðêÑÏ¿D'â‡ÁB@xûrqx®Ïë7åì`YŠ °eå}í=v¤ÒÂNúHÛµ®AGŠ0 z¡Ò‘¹ú–¢®}ÎÎnœY\Ïaήœá]‹Î.¼gîYrî¾6‰>,p`>KmäVµš³–nÙî@xËMíèÈj)ìšõÃC·ÜزŽëÚÏíÈíÈÒý‘@Dox)NâZvu$ö\ißp&É à©òŽgÇâïm ³·»Ì¥6ÉïÈg]‹`޼kÖ5+ýLõÁ×dÁ3™ÌY>¸ä7RæoD¾(ÿ ’l”¿sQ~ô€¥ùyµÿâ-Ž›ˆ£÷7“ÿ†ÐuˆxJNzoD4󞽓 LÞÈgÕâ¼å:þŽÜóæšùS¼ÎC€Ì¬búÒ|ª¡¾¤ŽPUõˆ•DWÔø¢oícäê[gº¸04ÑšàaТN1qËr#íbâÖlÇAý%%šºé»m1}wTwG_R—+Ó¬‡â‹Žfê[ê„ådg‡EÖ´©Âè䪾¥~Y²xŒÕ/ÏKaMRŽÑ}ݯ+±4ºØ¥o¹§f7üP—£«nDVuo2þtÕ“ÜôÞÄtrõ-uÌ‹kf§«ý­Ì£uH,ìÙ1K ^Ó¹ ÃX# ^X*o êse¹X{v±»©f‡º¹ÁDÝ窂Ù7Êm2äÕ#8‹=û4ˆP×[Å>=1`™[^ìÉC×Ï{~dÜ÷ˆjÅiÔ|ôé:pÌ’?žÚÏÐDzL‡¨ðÆÀ‘êþ?2Qâé™è±¦ìcvF)ÀÖ*ˆhþžœÝ¿I0í³`çÔµÚ"3L(›tÓäo ¤¸’!ˆéw#€&ÀSïÑ[1ݽDÔ×iJ D=Òuó±L"¦Öˆ:Ž«æ3쟂)z¼Ê``„ÆQ¼”ç1­>‹È䣪abÅ@éKõšZÌ@´/P(âKÍ–"ÔŒXGõ å;vµÀ‡ò«Q.…0’âcUt¤qѾA-¤û¦Æ««¨‹Ãî3WÆØÕ¬´X¥E]LõŒWr¡°áýÖ‘ -äD-Ã{ uIô2½ÉW·6Óå ‘nÓQŧMQ–¬Ø8ø Ý)áö.[˜ÎA~¤­·¦ödA€è°ÆöÖjëz„í­,kp‡=L°Tï¿ÔPXÎó%äÆKÔòá]—¶õ«ö˜Æ®KÆ:,ï,Tsñr÷ˆž‘š°âb±´[L¼‘.öù|va͹ 3Hž6×Ç~JÔÍ)·$AŸïM˜3æ³ ©!¼£ö5mîu•:yœÞu=%tØ6šÕqÈÀÎR´\øÖè3Œ«®SšPMhùžw±‰_ìt`'½0ª ·kðX©&ÔC(ÿ¡Ò'ôŸÜe‹^ÍÄØ›£šÐzlßQMhþµPvÞܦ,Rý]éé±16E’ØN çœ]ªÎÜ>™$´ѦߔoÙ[™Vuængåßë׆(” êŸ´g DG“þN@´ë1éÞTÓÔÈ¢ Ÿ%V{jU±Ïŵ¼­à(.=äR#\ËýÈ"E·¦á*v˜£‡ZÛUiI——bÛ«´rÕ€'€èK[JU»ûÇžø%¢ÈEþL]Ïf¶$F»äŽ<­„Q1öècj¸=5”ú]½7¼Õ8·Ý¯ps{R…ó‚è¿·W®@ÎÏK3†ùK<}­“ÝÃÒæÉcUÑléö¶'8ïBÔ3C+ˆg€D¿J㳕G#Žã—.äæ /}ÜOãxÛìG=aQÞÂ8jBÔð,œ5q4Ζ31ŽéÊ_E6ÑÛá– ÷] \Ý ¤KDZ@bÀ1Û4ŽÏb8N*ò [X$ÝRÚà¤.fúi-€Ó¼&D»ÃWÀ³óľs@Ôd,®='Oü¸ƒ dàƒzй’›êâ3¦u×)@b`‚ ÙjÛ[}<ÎÓQ>ŽóSX<èÈUÓ<ìÇä±, 'úëX–y,Ûu¡$I#Õ$¾9Bô¦,ÀÛ¢ë”x±‹¼+oÞËpÄ𡵚ý=yh-åÔMÖÃrÂC¬ ësï5°† nXx¨lª\elxçìáìR’5U«¼Lš)FUî`ÅÔsœé‡áÝÖ¡›Irj«œ¿Ê0'§rÏ þ&å ›’<Šºdg²Ìh+ˆNU.2%yU.2åH¹P”ñd%%bÒÆEÄ † ´‡‡ÛHF…÷‹ÞDcwǘštŒpÔ©É£hd™ÁgFe­â""j—¸ ­&B|xÚˆGÑä!cß›`™D¹Û'¸)áÅTWòW[ÄüŠáÃcþÊ¡QÉ+4:bò ;þ˜¼0‘WDÁMH»(°o¢3ˆª`o26jf€ö&c£næpïôIkÝ’Þål4L"ÁŠ·Ó ©&«O?8ƒÛÂVɤ%v:0c*‰„ñSÛ/‚TTÊvòZqß'ó£Ž|Ÿn^êÐ÷©—¼§Ã͉„ïS¯#åÅÓ›9ub‚Á¼J×[ ˆê×`–<°F"a#×§¯5hHä E×ú“Ì6Ztm‚FŒÓø[­ŠŠgxxÍ΢͗ú7‘ê€ÜDÂÊvaA΋lK±@.3âä‘5š»Ò!¬ÑÍEôÀ€ÁYI$|Ê’äBØ¢Mañ8²ó˜\ý¹‰„WÝØ¾x¤œÿ„I[O$ù¸®7“çeaöv„cÒÌÎBDË4 ä yp2ÙVs+‘(ÁÙÝIÏź>‡[Èä4¿Í‘O9> cñIÛÃ0žêç6kò€ÈN&kÔþy\„êï™É™çV`;+9ó3\íR@4B;™’¿ÿó˜Å¿ÆG’P¼ÉåƒiªQešÃXUS§M’,ÌW«Á62 ¡lO*¾`‡{óýÁªØ ›‚Y¬_‹Æ‰»òµé ‹YË< ‡­{RçQ8¦_·J‰/úC>dƒ±~¤Òô¼¤ÚFòÿ4ÎMåGšë‚ýŒþЃO£­t/ÉÝôñõ/s^¿çO§‹tÙ)!¡G<ƒ-åá äãrM¼<ƒ¾ÈÅdü¡%áW¬:{˜£Ó«_’ÛP—uÈ|ƒ3²ã‡kPX'Ï”öÌI{eÎ’5rrzåi6ÍŒ—‘së¼Ìž‰èL@õ3‡ÏݲÝIžÔ“¯{Ͳ[ËÌí) ¾TçaWí·pibÞ†u<ÜÀ‚ vOyUT“¶,Eº…Ñm{î|¹ÉÄ ,šË7 '¡ä 7ù‡i{™’/o_ªÊ <燗UÝA·DMêÙФ,ƒµøN=ZøH;t7ð‘î3…wä€ÁÅÜÂ;’­átîkEŠqï)¯¹]åçO~‡å7uwåa»þèî¢þúÒÝE;-uwƒ–ï-ew#-ß-»²|)»›ò|Ÿ)»›ò|_)»£õ.<é-»‹Mø>’`9Ÿ€‘½ªÉå¶a©÷º$Þ…þ12^–ùÌA[ý–JƨØ3¹ówÉæ¿{è’_ ûSXÑ3xW!3ûù«€‰¹ÿäGÂñÛ¹êÔgþDæ€&ƒ²áè39ð÷ð}›±H¸þ¯êeëU Äj,IŠT„žß ãð5,,¢a ’]Z™Ѭ2z—µRèÈi",tüòf‚è‰àTa-¿„a(%ª‘Màæ—âuïæ‰éUîÃ~ww»¹„¡¾~lÏTñ* bK¹en7å/AàÑð†þ–nê|‘αóK1 œ’Šf3ò1ò[a¡on\*ˆ‘ÒÀí‘‚fŠDºéi!Dú3ÿ$aN.!×¶. HÄPœóÒ3ÇPvnê™…Ü’ênEUÈN£p¶‹ð•3~#-¥Ýʲ¸¤Z¿€‘åž0Ȇi÷°\ ”c#`¦Œ¿#çþFÎS ™qï«¶ ”W)aÞß h¯B ¢ü’Ž2^…@äГô‘Ét-zà܂ՠ|¯ŽU+Uao„S¿]¼ ·ô,o„ÓN "ªUF¼Nx·ø-•QU»øÌíÈâ…”0X DTù.ìâ4½Ã üBH˜ Dî=ÔO†AÏ|eùD¼&Ù"÷Zcªp­DÌUá@¡ @~zê€È-ìÒt$¶tòð”˜Æ‰pðØM¡C€I¢<ÉNx^"_<Y±ýÕTiδ¿*M%¤À3«ë#-}µ†>"ß²ºX6^ýnQa:%^ka±lº±‘ÙæÑ,!#0ꜫ±IöÄÖfCÕ±ÛzpwUG§ÈmI”<öfuBh“D?C/“í^!|Õz4UÃ{­/"ß”XXaŽ_ÔZØeNSvuwEïÉ|²pÝÜo°9¨_h&˜Ö´%=ë¡8ÔӴÔl"˜¦‘hŸoJ6ÌF‡ØòmM÷ë D¸˜®_œñ­óŽ`š³°êÃÿÚ8´1ü»ÃЪ¾xïpY­âÆ«BT.®ÁŸ—gÕ/þÖñP°÷õ™N½MÍÁ*Ö ÃV1ü§;²ª?9ôm‹BS£ßZÒûM‡‘ôþáÁ¥lÑûÓ¾’ªæ—¬"†ˆNÄ.¾CêŒt¢,Üå¢þºt¦Ídi’g¤¯d¡U††c0µ*&R íóíŽë,ôQ¸,k<·½£J#kö+vêk¹QÒÄe²ØÚvF\WR²÷áý‘+ù­RîØ 0Èÿ’Éúr•Ä>ÖW:¬©Tì„ðo%D®À[áv8ùl[À™I}ñþ;ùlÚã,y~5™ÝÔSÈ/{e˜ŸµÃB\Wfè5•ù)a›¨ï+––qáðb›u_.àñ6×M§ðÅ·™Z8”@¼ÍÔˉ·¹¶KxSl½œ˜zCOÇ–Xky&7éUb,YÓïÖ/xi¼J=36pW!#[ÃÁ kë º=ûOˆÞöâöìò<Ìâû/Éäbÿ&ÛBQݤŸ•Ý<|¹Ó¡l¡3DÔ|èÎÿÈVÃÀ¿ý¼¤­½n bÕV&€8b€¢éi¸Ñ?Ns}!Ûë àj €Jjª}M'2‡ˆ¥0uÖ1gb¨“ãâõ1¶²EÆçƒ»6JY ê;Á?&sùÁ³. zIƒ;eø5E»ŒN Ðp*zÀëÏK‹ Ùx ú©×R…¨lF¡¼bØ‘’õаŒLÿèšÿdÊâ¢%RˆÎevßéïD/FÄÁ•Á‹Ðá‡Fd8*(¤úñ\ŵ¢K5QŒËzÎ~mµß¥MëÇ¡:=lBV×u$ûÏþ”S!óWAtž ,orFÄIôÝÝK˜!„„ªK%êì ßP¼ÊÊðF1 "ü“«ä0;ÐW¸~HC„@šùÞ·®+‡´ÆÙDË~OŒV *ðFÊ@Ü7 ¢q7·©.L6@Hä 3Œ@ø¸ªÄmv™Â?Õ,+[nËPžÐ[§l)¬°B[Цu*LABïÕ²»ªõ£Õ¼R×C¥T(eèðO.IÒÑ#¸Ûz<ÐX.Z·³uDŒŽçèFâ:ueLOpߎiÚì423¯52¦‰ñ$B¸‡OrŒÄ¤ÅV*-®Vws…™«Zó$¦^µøI­2-×%ZH7½ûIɨh(!Zîd¶¤³pïJó¼Üâ81š¯ãÞG"ñìe8äõnYètG¦Þš剛$y+Lu'|'»P¬;‰uG± L…b$.î%ÂxèÒöGAo·Óðflü%µc> ”L& w!.§d ó>bÁpK¢ dK£Ìl›×|@ eÐÉùu¹QèaÞnBùËgÓCíI¥<‹>p·8‹Xát¸‚¢þ¤ê¢Ç:ô‰k#á²u2Z^rüžžëTúez‹¤OU&õá÷œ`ǵ[üÞ6R䉥ÛáV6¾ì´Dn£Ï'¦r³ÓÁg¶ áÌ!¢Ûô™œþ“¡†a‰¶3b{Ë6m{uã0ù‡©ìÅw³­†÷Xx^Ƀ,’ÃòÊ6eŒK+3ÕîMÉ ÐüÜõq^¢ÜwÐ+]Îjþeª•"ãr gÄåeDø° "þe99.iÎÈ„=œr˜×euZÃM}o¸Ì]]NtJ-\ÌÐI·º ú´Ÿ]D†åòÔ=,é]ñ— ÎÇ$¢<â%WÆé\úEÑK„¾.#]¶}J“\ìö`íçGèòtHå"Ÿ(Òn§‘!Ï@?}T"kW?CmñœÞ%sÜ+ì õ†¥6†¥¡Þ0éùía†Í†5b3EjÆSdØ[,Àïïé8 κҔ Û˜®¤R-G|µ¾µd$ÙÝ®°M&ÛÈ¡o‰^“ÊŠœò`˜£˜¢ŒM< $z”Ì œdãFø;ËfrÙ±÷ÃHøªŠ—áÇi±HXÌÚ†3è€{ψbZLO³ö‰BïžH†Ç$ž¡× &ûÖ¡kfœÉ|&‡zév2¶š2ð-U`óà2È“[/¢­ŒIçÂލ®¦3\S“N¾5ŒV•“93V§ŽËϱU5ƒ´¨4utð¨Ü6~Íq Oì8¶‰Ã@VߊdrxÌ:uœ¬-øÐªŸ±H‚U­SÇ)&­SÇ#[¸ª( d‡O®SǹQ/]½>©Ç¨CÌ”‚¯^ßàb &½ÊÚV=¬}Õfgm°ÿUž®tŰVU<…°ÙQd¬HǨÓò'æ|HLìXHŒ{%»íè¡‹½PëÃa¶h oæÉºÝ~h« VßJSuX0ën;OâǵÇFËš;ëWè?ÿgõÏç=#ƒß–€z NK+¿–ÉIa>­q’2Ò0¨VOѧ ³ïŸüý²klòËvŒúe;+ý²}Ħ»çåÄÝÑ’ͺÕÈ¨Ý ‹ou8=Ø#”ÛI0/xUaq(…wx3°è.®N¾sh…¹z…Nº LÊÕ+tRbF [€DÀΖ@µvX¤«%Pã6êj©±ˆŠž'¼ûwÍç¦@–íJåEïÔñŸ¨xpyW×Ò¹á'x=v£_YúÇüö˜$ìÇíCH§U}ÏDy.™¢ ?òåýè;mñõ¾½`¿ò36Ü×xÜH·‚)ÿÎ_'øÌå‹þõÒÜ_Á×"$¨]6îÚ)ù™ÁÔõ$iH¨ù‘¨û³¹ëk4+3ë~#Ñ ¹=?³} (óÆEâF~$"†}DƒÉt: ¸6;dÑ„,yÅqã#¹ hœ Krì¸OÙ€D#R6ði7?®þ=ÿ=Ú`÷Ù_ÙY¸»9úʤÐsd„F¯è#‹.ø#帢Ф]È HÓ QiÒÿH¤´ë9oãžbAf~C¢CV~&,ô[ö÷‰ˆ¼H$K4ïõòÓüHÔÆÖ²6öLHqm4’aòÌA ¤ägÂw?ó•$'å¦IÈp©ã7²²L¹y…,ͱ4öAº“à Ÿˆ¡éë7buïƒxH$”›mÜÈÉ6nd»OO$ƒáùsM’EòïÅ”¿=•½Ÿ!rɗѬLH÷áªWàáÌñY^å-ñTïô(óBJþð$'¨õ_@÷.ÂbÛ fQýÔ×í‡-ý˜b}U‰Ôù I›©@©PꀵBêyŽ7e™JÕó A~â¿þFf>{žnc@‚ìxº_ûŠ® 0?„º'Û€–p@JþLð!O¦¶/±x·=L¬F–Ú\ÙN^{‹ëVògĆ{!ˆªfÌÅš3èfƒs]ì²Ý½A°6Ùe»å3l2ÇÒTHPÅvuç¿8΀èÊıdÁm rëxÿk‰eœ†‹±'±åÒ:Șyäµ5ˆŒ~žUþèÂáÛWþHpdWÏ¢‰;H•=WìAš]5•¦IÎäÂ? øÏh Уä^AÔê™óíuYcçÓÆ8ó—t{‚dS C8ÉDTíµ)†jîR ‘Xg"Áºœ-÷g8q=Öû(Œßâ ššœÚn wÒnŸm”)jî:F–ø¼þanµñÓN YÀD­3™¦`#AĈ# ²÷g: å£åÓµ Üæùk¿ügïüЕé“ߊ·Ð¯î=H2ïǃä¿vÙEz³ªçéqºµNÛƒDÖ>=¯Ü¢H÷\Än&¨‚->s£,š{÷ôl¨þóÎcw-JÔÄô@‚áÙ-ÞìÌfãQ Q€-›œŒu€¨?Ú›Œü\Õnq›[Næ7©ÌM$ì –øY;qÅâ%âJKÅfÏ“ ïq+ùÆE®öÀ^&%'·a‹%Þ…ƒ·æ„ÿˆŒ[WîÂRmÞíž»°P\<»°‡€.Íe8T=÷r£úÕ\h¿·ç±ÃŒ—[knYs‹ŠõO‡™œ•ŒÜFާ.¹ 8±×ÐíS@Œ#e¿v¬ãReåŽ5©E)ƒ $ÚaÉÕ±, ¸¬¸;ÉêÚáñd9.«%yO0œ-yOuÉD¨ü¨ß=ôkσŒ‘ $º‰„<óš $¤¾×ÔPìÒ×JÒJ$ä|wæs] æn6Î<”GÈl/Í*6uU73BâwMóBÁVž0PÊ.uÈ—®éí9aÐÁI{N.ÓìÚ1%È  ìäGö$¢ 3õþ¡Bë‰ÌN¥Ú\FBÉxFž=Q:{rÛS–OPÄõB'_(ßo"¼oW€~‘³ÿ’ÅWÓ ñ¢ S=DWÜ“­ÓHªj>]V?ÉÍ «^êÿ‰^ÖYbA ±z_I‘ ¶{±ÞFÔ½>MšhÒ¼ö‘\‘ûgЫî‘â&ªÂlàÖ}R bIUoÚC òI\ÊÁÁ Ù?éB ¡¦ÏC–ˆãø š½K»n½²c‡„©9S¶Î}Ž8ŽADMFï°›8èZ¡ Žäµæý·%y âô#׊‘eqåZáÝh³›€˜%Utñæ²…¥Ó’¾(g2¸qÌdRÅÅ‹kßPßÛlºT؉\·¹ÑùfjRv7ÔÓ?$pЍ^{‚«E{“í®Mæe0*Ëd,ٛȱ xgaRátÂ@$¯…-D­qù§/Ì)n:­$'ˆ\8›‹Ï8”!ØqruÑ—H"`Z_S9do4ÕÖâ:Ã~),e¶.%/‚r“×GNQÓ$aˆØ4ly°´háˆÈuSÝ\1ÅdE:ùL×Y¨ÎNÚÈ“G4È¥~­xHÌìädÊñvûI¶%Ôa.õc¥¹À”©g!JKÅ^úàzF'Z|Òk>hóAÖ²MÚõv~$Sì—õQPO› “v"Wˆ¹«°YÒûÃÜáÑû[Ì+„9“nGÉ%@tËvNó‰˜é´ÄNÛÊ¿Ã}ÎîºÔp bÑ\i9¥$‘#uõ›{äaTÕŒ!æ/]+çS›’& š¹mv:@ÔöÑµŠ»±–¨“V ü¸–Ô2?´eW„"ì—©€+ ½IÕ°S‰ž8G  Ë’¯Ìº@Ôƒ®ñS²4.S¿ªL U÷ÆÜØé:‡¥y’yŽž0;Úsõ„vV k|!fâs­ ?6³Ú)¹¬'»ÞÛèY{ª'‘WÖˆ§ºÒÞN߯S\Ä» 9±ìå'}ŠÛýåH^·÷%ƒèQ}© »Z"D¥n'µí™e'µ- H™HQW ãö²ä¥ .|ì]Ì ˆ"ë>é[Š<§·Ý…CßÖé(!O‘_âV¾M¨øêÛF0ô„•Vƒ’íHÆQ%R(Àí cC‰6ì—hŸCˆHX€ ;±]¢MU­~Û~{Ë% &‹[Ÿ‘+u1BôB»Ä¢¹2-CÇ-9ü—é¡[)l ä–'Z8En#áºY÷mëzd9$[ÁƒR𖯵iZ·:áe)YVã®%‰‘d¶»–v÷¶‚íªuE]ÏäC0(–¢]-ç„ì&ãê5-”êé1çu} «°šÜM—¶Œ¶"þÂõ3EX!(^=EXÜã"É_çÜ£îȹìñÝ * WÒm$ü%Wµb ¤æƒE…½ñNæ *%™œWÑÞ–—µ=й±‹ÖŽŠ«¤híÊeÖãâîr7MWkì€ÔÔá…÷á<ªÙCæ¦é „^°é¾/:×éSÄ­œÌÇw:Á.·Z ´„@TMÓG[(Ã9ršÆ Ùb8§NÙ9…²‘¶Â^¢CýXi¼f‘³°cެ™|Lx¡«T™¦\¯ç´¤©g/w_ v–>Xy«@f ?ÃSt7"‰Ckò€Ð؉O[^\@T`!V @jÇUd-lãÿŸ½wÙ¹&˶ƒúùª†Óë~qaa aW!–}|ÀÊlÜàñÙs\fÄW‚Ø*UU~#cÇÞ±b]çsŒ¨fåKïž?–„äSû8J`)-ìhb”ÉNꔋôšœ„-Þ@T.¸¦nìdÔèb™›æTF/vG)ßµ¾¼sÕf(ê1CôÙîëCÈ–ƒq jãÍPq š£wãÔ5ìŒreÞ·xÖfñ©õ)š–†jµ>4yãaѨžF¿ÞåD…5úmr/Á£'f]6Ô‘ÇõˆQ­ø9ãxr8%e·JËµŠŒãùâÐžè‘æŽºñ£ûèeIH ‚Þº3É u8†‹u<—*Ë%2šªà[RFux¶¥Sôh‰ï£¡—„’M¡Tÿ-õ T$G¾É-PÕ4–‡Ý,‡…Ò¹%ºšæ‘’:k°i€™ªêúª¥–yD´éJÀ‘\»™¡jÀgH¡)¦úFü òè¨Ú¿½‘Ï g:óŒÐ}ý×Åì>~ |ˆaÏ*Ûr5†iEOa!…Ó*WµBzõO„Oñ ,@~##¿ËȨ̀0"“áˆèß3Oþ†eÌÈ‚ŸgæÉ 9&>¢Š•káG#µYêpU &N‹H²/‘aSMÁñY?e%¤ÃxVP…Zý‰+!òÔêáæRGñ°¢.hJVp•­Õ×#i¯GÁ±=n¢»µþzÐÝÚ[ $È6%©„­ç9\%ï‚—¡> Èµcçí#ŽœÉa@À‘ëå%B·If·m®×— ]ăšæG‘®§.°Äì‚áæG0•næ3˜M·ò!̦Kªc6Ý~=Ùt'%ú̦»ù"Ȧ³ŸpçÄôK…¾ úD³£åìØ¤nÒ[~w%àÆaXvxò¤^`#1EÁ9óz¬/›]Pˆð’9é¶éØ[W©•tK·  ÇLS­Ã m^#`áΞ¿±ÛÎgœ5vÒ™"á‡!êàŒºïPr1˜¦ú9•Ýk¦¨Ûa¤{x{ d½)¬@ÐQæµ–-*A…õ®âêÖ¼3a@î%É ÙúÎÔ¥>A¥üd’u€Õ¸¦µP%Zˆµ8/åŸÖJO$‚(¬K søÔ okÝÔød>a¬”zÛ,V Ö²E>iOÌf‹‡¢~%ÈϾ1‚ - 9Iz"5c‰ë­Ú7FÅ!¨×VÅqôl‹|25n}ˆþ©œ”ÅëÛ) ¸™ý¶HéR°ÇÕ\ }ÃÜê¡Ô Äê¡\gƒ—nQ×I¶óÉÑ'õŸà»[Ô•ãÁ*:úFä@åZ¹2\LhL§A ·ŒªªZ=/F6g­žÜ—Š{R‰q¹¸§ÿž€Š}Ôµ·×ø‰ ö”B†ŠˆZ"E…F~"ÊÖ²A9Mµ6㢰Â|ôðYäõ„­xÀHHj¶\D$¦UzO“Ö;¨_O¡ANõIúý¤¶å|Ôï¥ö=[ê`S£E`š&KuP“ ³Zîß&•@³à,õÈ£c$Â’¸b9Î9UW,C9ŸÔ¹Y ¨óëu Åô³_es©ªŽò›f%q}¸_H+ŠZ"(&ë¹ÒLjFL'ª©û^_µ© 2¨Dæ&â7Hºh”[é ŽË‘6ŠmÎÄèË%´•õ‹@Pÿ9šÇÕ Og:ïQüÕX©=È€i1Š'µ|쥳¾@S ¦Þ–4ú£,OƒZ¸(æ“êîPñäØ^4·l(lFÊ»F”•ÈM$bDQ¶èmøàý€¤Å|ôîw2óѻܥD-¦nÂâu›0jåœú’ΤªB‡tõ™ªÚý°š|.·fW-ûÜnÍ®V)ÞöÍjéy'ˆ_(ýõ·,ÕÛŠÜÎGø¾«®})Üe¨«y"îÔ³ D³–úº~†(õ]ÝÓKg¾(êµlvš§ã,9”CD“@ç|"ø9y†ëªH\Û6#òÓÌ)¸ Óì*x†Üw8]D~¡Hüå2ôe9&4­H»º}Ûb?ÜÍMÜ–*Íó”×u N±:D›¬4ßÃÊÎMË˶Èpc8d`}ˆˆ¬Ÿ‚ ݽ¤õÜšêòs†h‹ßÞÞI2 üÛn¯nçö´‘|5ÅɃÏÌ…Ú@ž +Éí`Ér"–;M0`ÕËëÁ´ršGmÕ šçFé¬ÍGf¿²´!¨©êl•,‹?€ÅúùÉãc¥R¨©Ã:ÿYR`®`×¥ðBþÅT€¦åG‹¿ª€ý+çL:n‰¢¯ÎIE©¶ùÈó•«ß=.—/æÖ´^â~?u)à™„à6[/ч$8žõÒ–¼Åc½„Õø¦²x!q3œÚéÅŽªy|Ÿ0½ß•_Γ*œ°¬Þ­I®X×#ÅÇyÓãݱ¨ÇñP(õ¸éIÅ|lÐ.å*iQ¸¢ƒ, ²¨¦bšp TSq wmQMżÿk™¸Ò”º w®)ùˆ*ñ ¥”jÖÝ…vôkdÀ&¯°G‚%.™YÚÝ&1å@-³4èŒÚ.i|x)ÊÚŒ@«§X‘*LѨíbŽhX©QÛÅê3´F…µ×aŠ1ylr@×Fff2rƒöîL÷êUíax¾@®‹Gžà)ª¶rxòjÁ@ðèuØNï°b8ùžJ}¤êš–@Šù¡Vñ‚BIdË8<Å‚ÁK¯ÛMhºjÞ‘ø‚1YˆV5«:Üü`1_Ó¤p3¶²¨àZ¹®¦Iá–dV+¶Ô‹4þ "ÅÝ¥ûdÖåÒD6¤c+§þ)SÕXŸ¨édGfû—]ÒAJPÇH›.ɵ¡ìNx5RÛÊÌØkm»f÷ãpx¤îÓ²‰î¢A\ v$á¢\|㫳=™ëþƒŽ“|*s5.†YKÇB©Ýee l-)³ÅÔ«/W3¯Ïî˜Á¼d šBôú'E#Ñ€žš)zZЬˆB™Ð;“=!ö€†”ôïÔ*³Sú,í §Ä%{OÐ&é3“wÂŽ´ aâ.,K—ÄÐ^¶¦•ˆ~í¸Ò>K¯^1 Qµ°Pµðš^ÀÐ Ð-måϬmv­”?ÛžU®ƒá©yl µž ¶PŸ¡LݸKÎæ²%ÈáeV˜Òb´ôIì,ÀD]g-ý#cg\ÐDuͦȟ Ë`£ ‘¿æà‘ ˆ^UgæñÕ Ã^Ì"#ÍZ-†—R|áû ½NûaÁm̳`("‹´†¿ðÖ5vL¾z(“HáCå‚ÓÊr‚07þÅAvŸÑ»”?r z—òPD@z(«Ë„3zÛ0"*¦=okR.4/1\Ÿ±ÞŒãÜHÀ D}Ôx)¿~ÒM zΘׯ—fc8UcMsF)’~Ô%‹m6öØ¿^êpÉþÉŸFÚ¤¤Í¶ÑÏcnôÑ¡ ÏnL ö¯7„S½òª²!õ ʯ¼4/Ã?Ãuš‹ã ÎÎð!?ú”¦¤B’ ãª£is»4üÍ¡P=­}Yb²æ°#sapbM—k‡•:ºÛ´¼ìÖ+yÛJ4=z ¶m„ ÖîøÅ*7÷wì­¦é½áEÀézòp‘ßô·紤±*Ý”…¼é!·š5´á_¹Õicâ0½‡ëtÁß¹ÌP@uáwuî¹E¯ühÖ<çêΪÄ8ª4 2I&çlî’¦¥©¡äªÎ „ p—ùåR“ DÆ ‡ñl©¿ž£mçrH ¸í*ZÏÑHL Ù£Å@ô=Åwu®7Ê%¡2k%Ú@0Ä—õäÏ>ºquÞ]E ¢íÙ‹{U͈X?QTëˆÞ‚jhƒj]õ!ú¢÷»yª dèS$’v‹Ë{¤³¥»,štTŽý(÷žutco¤Î’´ä’ÀQA|r`Mÿ£|$xE_¶ß’Œä²„ÓY=ï+3 ɧ­5ü'ó²½ÁY•ƒy Ûƒ•Ž€z‘DР.Là¨],i¶bKÙŸ²Õ0ÃCw.=€ËãÏä‰1 æ©M^Ê'ŸÙõSt¡@6ïlz땞YÕ0Þ`ž)AØ5=' íêRº9ªsWtÆÑ3¤ †„¿á4{$ó·ÕgÌ|}h¨ç[4úPž+Øø¢È í^HSÿLbͨz“oŽ4³iÓH™?É_¢~.y«7¢sÛ2õçtÉ„¿É„§B µÿDtL[˃µ÷| #,®#5ŸB$ž^ó)Œ°fç…4šœ¾‘“OaDG°²4»½?ÅO$ 'K7ùa´S õ_õ-=+ Šõ¢…öß r:·-H4ýª?wK·ÿLGüð P©º¦ÛHt„ñ¨¤\•œP¢lNÀq'lºD3 û驾t»ì –d×4C‰1Rtü{—ªš*1²@”ugý¶¸òH°¦K ê­käQ4e4^rêšÜ-2º'”¥krΙºÄs—jbŠë¢bîºÆs—ø}å™»d^rîZb¾—œ»VUƒzªZej'3vÕ%šªEÑÄ´˜2 DÓšj¾œ†¤±ÈÍ%—䌲8­éòKÕÖ5^‹št½ÖÀ¥k–—IvÇ•k`W“š’x¤PQÓëèHl¢¦R¬ã]×äRßù÷lÞ 4]±–w Àñ.œÞrrW„˜}5÷*Eˆ7E⢗£.!­Ü´x¡¡lþiì›6®¹~¹g±|£˜Ùû¯¥kô.Õ%á±æH  š•–›¿Z]§Dt¤j„ØV¢X(µŸbë‰J¥Z¼ º`prO Fv{ø ô¦š]{`«K¨cŸ Z_u¼3öÒxìj£r÷@¡(Ì¢˜J©|ìšA‹B‘•@ÄRYz5õ§¼îj÷•rv +'àZ׌¤•sT4€€/^‡ÌEÅJÔQC=¿ŠÄh«Õæ>v¢S´ÄÈÙyžCM@Íl¥ÁL :'²Z;…D*™}9 µnÇ%1)TÌÆ •ï*#$õê]íüfu‡Ñ”¤ÊÜÚ§C¿òJ Äqi:Ön3®ˆB ÐܲDßNyÆocfë¶‹ 8­Ÿ°4Sìh-™~v_ÔËöEŠß4èê,ÄäߚĶ 놋;Þ¨pÉQçŸÐ: ®·¸›Ÿ‡ÃîÜKñ#=ƒÉ3 "©SUpÐ9?*"œB©…RAƒ?ßE‘têª'Q>ª¡Öûqòê©f‰êƒ\9°©úÿš«Â£Œû+»öTú\—ÁFˆO鿚…G™ÈÃxRQê/Š—œÅªá;wü®r«•IÆ­²-KµA.ìŠìeS™Vs—»ÎåHö=œ§k½;­–€E~u¨ÌM¼«B WË1i›õuÈên•éÉcºtKÚî6²Å´.dFùþzʰ0Í–‹öó«å¯ÛªìÊÑ—£zÉî$½K[×0­Ç¥­kf–^|×f"K_Uß)6ŠLÖ1A ©`r›VµõYI§ªÚó¯äÍWÊBDݪ¨[Uó=…ûDXªš•ëzÕÜÕ$µªcÀ–ž ØSE¼âÂÔ«’ݤK¸¶u']B~wYB :^×NºDkÜ(¸6ÈUý²ÈØMCs¯$© Žù¤«¦‚º¤«F&eT\›¥²Tõ|“¥Âãþ&©¨øÛ •£Úø’ •«Úø*^F×iâ$å°3Uˆ,Öåe5íI$B?®u=6ó B^Ý™î2RŒºüd:]Iba|ö»Êí°ÿÜÉÒp=;ËR˜ô/ì–o1©kvm)4²!¡sç*7P·™=*nZ}øÃC[³;ü]â¼¥”ì©î|qçø]Ë#|h„ßån>èZSÈb—±B€(QºçMÄ}n’i]yŒD†ŸôJ¤úIg¡RIi~ÒY¤TÒÝag•RIOFiKi3J•†ÊLF)œ):Þ'¤sЉVsHd¿¥Ý‚'‰P©ää—/)•\3ÙTTr+&”2ÝžñK<”õNäêC¤H¶äÄVU{‡N)Ç%¢ÝÐV•·8CÙqP¸Æ5²GgT×töƒÆ5Ê6ÁC\YVU§Ž+¼®»´OE×Hq^hv^©0:Œkn¾@šwC–¿§r±ÅZ½Àˆü.êÜr¦ dKùV/°ÖTÇån»ªFšºÍÕqÅé‰,…yõƒ+³©!Þ;®jûl‡(n¥*°0Ý¡R9D€kÍÌÒ÷$æá1ÛJD\ÉcÿàÁǬ#0ãÚ) ¤SAy.Ð]¯Þ7r6¥™5Hꢮzœ~ Àu Éɼ𠓆¿=,p—Èj79æ1hKMjuäÊX…¬…º=9ŒM9ìÙŒ@9º¹QÜEí¡kÈö-î>ŒÄLzÝm‰ñšÞu9‡BàšE&Qßjà‚†zøÍK"ò ã'/Ù—²äº¤Rµ\ƒUÂÝ.®Œý6ªŸ¯¼}mH\ ôã‡w3 õ†Ì2üm”g×S7®é¡é®§¦Ø7tß¹:#b<© ï§æºŸúñC¥zì|lÆ›C‡ÞÝ©v>®ößUZ!g¿ó´°H^"£['Ÿ‘çE-ýecw:"š|RGôäçòÈdƒÎWûá‚£bÕÑœ3—©2t~€@{e·—iM_ÙíelÓÍra„{¹ù) –Ø—ë™Y˜,Á(ø/Ø*ìiÊëË[ÿ*³žpgÐx—¡OX:¬¼†¾ÞuB_Šš=I¾-;’})DžE,’ÝщěÛÓS]³/ÅÊ´/ÅözÒìKq²7ý ¶·òUŽHá¿QòŒðS³Må]a 0¯8Ís¼Ü—Âää5˜Îðº$Ϧ°q£Òàâ¬2¸8’ºúsrtÊòâÜì’Bnq—¤ƒTªŒ¼®¡»øèÙ¤”T ë–þpáM¹#,aÚOÀa£¡(Ìt÷æÏkDŸ枟HÍ.ç¬Òr„é9ÂHÉä…ü»ê$eä_ýÖׄ1A,¡îJä9&ä§þÉO½îýè¦øNojKT9³õÑûþú ¡Š²u‘¨…¯‹„¼.ZJ|<y]ÒfXò"#¯‹ÜRÏEÙšÏE‘.ó}‘‘÷E‡D¦×EBž‹²t;/Jäu”‰Æû"#¯¯ ‡ÙÙøur³1«ÉŸ‹$/²ìôsуäE!ìH¿%_ô ÏEg¥M“.J$/ FIùñ›$/Š$$Ev|уÜlL&æE/$/Š­<Ú7/zç¢# ‡ç¢Dîßv•@‰äàˆ«*ªŸ(YsÈßÂæßÈ«Bñu Å_ÿ>BŸ%îu‰Wu–·¼®ê.xy]õØiùª¼3_EýÖ÷UVt}]5y8~]5}\~]%Æÿ몬x]ÙÉý¾JˆZ½±ÕBA%gí9z]…À~_%äÇUˆŒ¶÷UB~^užVÕUçÕªºª“±ùºª›ÃùºjPZìuÕ°ØØëªPøq+?®;Âß.äçUçiS]uþ¾M7-Ó^W ‰«Ê¯üí?üV˯øþïßüõ×?ÿËoÿô_ÖÏrœ±Ï2û—÷[å%¿b“Ô~a;õÙäü寿ýOúWÿðç¨Ù·ýéÿúg¿þ»ýÇ?üùþËó[\ýOâc±ÿü²_qš»Ÿõú/þí_ÿ×ÿôŸÿ³_ÿãŸ#ý|÷üÓ?üŸøX~Q°š?;™ˆ1ìÏVýÿá«þ«?© žþ§øëß}+ü?{š8î‹Ïç—þ÷ÿûŸÿòïõˆA2»?ž0’ŠŸ1Ü‘’ÖÏý/ÿáÿø×ÿñÏÿ$~êãOÿé¯ÿð¿ý§øñ•ÿâ/¿ý¿Eb¨ÿŠmúßÿø9yÀèöpÀ5lÚ`‡} sÉêâIì‚°'Öž H]!€éí à$ÿ ,CgÞnÄ?X¾Š¸ï â÷RàîœÊˆH’ïäcŠ]\P§ ð1Rô) ®l #ØL±D Bð¶D`WSÔ”lj‹äùÇMŸ²D`Qkdq@)j „daÓ§†ÔZCö¡…8É’^8õÒ¸XꈘdˆHç Òo¾HèÀ¸XZ¯ÉÚ뻚CüŒPúÐJ¥È5ˆ )P"€L„_~Ë¢Gò1Rô)K]BÞˆ$Ptˆi¤,µË8\°Ø%RXtVÖmÆ"õËÏa„“–²©ÓEïeý¼Ê.د%¨#|ÿ ÖÊä@„ èXDxÅm €DS™K2¥8ºØ»‹i)—HÓ';`—2ÄM1i)ñPÑèÝ¥<›ÂJ4vÖ}"À:ª…mÞSå›Y ަ’|œY°™?¢ù3ï%²‡ˆTH© ¤ë>“¿/˦2ÝE?èf¤"I]ðY"‘TÊüÐ!ZJ´•ã¼O÷“¨âÁ›05*ôuѵƒ¿ÞŒ”MDЧó°kw•…mÊêÐ3šÊÌ$=£ù QÙE@J¥Ä #u„ùB]üêà ·Ier°Ô0R²@,5 Å3 dÕ†:òÙDJê%<¦í@CS¹"SŸÒ¬Ý«Ú@Ü‘®ÑúRÖ0|Ö(ƒSUwµà¦l 6×è=uªGcOiò¬ jLŠiµ2Õ˜“¾')k:@/#è”ÍÞ‰!‰N‰O`šÐdÖ¦¹³®A] )9nèܼà’¸:ߘašëAC \€u¾«ža¥ÎwÑ3ÌÔù†µÝ’S¬œÏ`«Ô4çC ‹å60ÇicL`él&fÉf öØÕ6žd¨Ñ`l69¦!ì =Bïx»­¥Ð{UÓ˜ÍzñèI­¦~ýÕûu]q(Ñ–¯?ºq±Ô;µM€h $1ˆ&<ÆÊû´¶¨Åóh‰«ƒ¿Ë×WÈ?ÑJT9Ϊ=*BËÅÏ©)—\ pµ¯| X“ݶ=~`•ÿO8 `î¯v®?l²ê’öuYœ‚ªÕ©“ˆ^fAfºÏ¤u‡;Â" >K Í@Á]ì<> Ø TÕ‡UÃÀsëÏÂn•üö0€@‡MV| xÝÕµýa·RmÎf|+ÕŽFwqJªVï³_M³«ÝSIŸ¡;8ÙkÈ #ýÿl¡lXlÃÙ­Üü•m¬[¸ÆÛq†¿–›’&DG©"‹0Ág¦:Ð"_›‘S „:Yqé LK¹Ú¡Çé2ŽÁΫت"ŒQ°ô8Bï”)D}*µRÕ7µ?€ø×=‘wpáä‚‘^¼[›#²¤Zy?'¤›ÆE‘š% ¾ÀìmÅC5ïÆš ÄÆ7$顜hã €È`Š2=ôenõïY˜Ôè÷,nÐFznGEèXDdƒGâu-L¹Ý,ˆÉá~ µ¼@ÌnG‘ð$¢ÛpË:œ'‘&Dƒ‡NO=Ô;åC7( -‘ÄÚ( " ³pfÀÆÂ ìvU=ôFåDo@ ¬3BÝ=ªyøNÌK@ä«CoÀÚ¦z}4áꡈªæš•/ô˜ª:1Å†Žª\t&j€Èçg`JéÃ࡬[›ÔÿÇáWí­­ÍbÙä äwz(¿êÒÓ¬‡:¬=Í •Ćj\£ )»”CD7î$»§©ˆv@Ô:<œö´õ—󈚷š/ê€ØP­ð o;§ÊJˆæ.*ƒô±\_HLΡʫ®Ü7‡ÍJÿ ¾ùèkÛ–«/>úòA"DVðèY&&ˆ•³+½÷€¨çƾŸ²îB mÑ\{äº-˜:kc¥±"õC€¨1bC@Ͷ@¢ˆÚN…}¬b“Gjzˆ,Ûä‘ à˜916žçBšy@£Ïí¥-6útáKˆ¡Ñç²Ë#WýÔ%iëØ¡]ÄrêÒlonIZIÓgÈ ‘M;M{˜ !š«O­°ŠŽûTO²“ÒÁ­Ë¾Ïš«8u´²ÞDdj»õ©0ì™j·G<Ùð=Ì#Žš‡‡ñQ[6aå]«úiH3¢Qk±Íù-|ÅÅT²³7Dxä‡0d÷±u ®€}îO:}îQôÿDýUï×* Ä”’Â2RñsI•Rþ\s郈Újsï–í/!X¬œ $†^?Þé‡jê! +õ /žÞE˜Pp_ïáCŒuá¾7 ¤ã¾®…¨ëÝDj5²p­Fi¸¦*¶Ú°±Ùít72âÛ·÷ÞP”Dtc0!z'¹[ÈÀ%‚7µ—ßËF¥wªõ ‹ˆzñFÚ ƒ“£ˆpÞŽïûÔ߸òí¥Uú¤²ß±óé0u¥]øT˵ÏÐìÇZ*²ß½ï€Ð¯ÍeZèAjâá.Ø©EÊQÝk!:݃eŸ›ÍN¹,ÜHfÁ±ÚFðŠÓ}R8¨WËwFX¹GLžõ‹á à|ù!˜Óêê‘ Ô­41µÍšìµsç# ãØû\Ù6ø)Œ=Ä–c" ºšþ`˜Ö†œ”¶èÁqóÏG_íÁƒÓË]ýzúPHÇ7—üù]bÜ| ÜôR®4Þ8ÊUAÿ óm½dXØ3Ĩciš V‹ÅÓsP ÕE©˜¤@·=¿Šì΢(! ×x›8i5Ò#µê¯â!<èfÿD˜ZL=H”Œb‚¾Õ·RA oõåy5‚õø”O«!Ýûìs4aÏÃë.( ·éš„&Xª=˜£šÜ&½[Bb²$´um#±Ñ–êéFЛ4šGçâ<Ú$ºhózåñMi.ñi1ôÙfÝÈž"ê“SOïMÕ@ÐÄÍGÁ@&îcsÙI'ÄäᕟŠÍRŒõ:çäœ]w¾ÎÁ×Yw¾ÎÁ °®|X¹Ùo“¡ @ç]0 T—»ÜÈ“Ÿü¦…ê`WhÆ_µظ¯äù`q(WÇ”@b—Ô’Çð¬, ÅA©@0»”™ÍÎ V/› ý$vÅÃ&è"•èÞã6-[»² Å¯¼sx‘ÐĤXv½Éï@J~(Ækpòõà-þÚÙ«±hôàõÏj$Æspÿó’ÙÛ2…/ÎUÈÉ?ñ;š×‰É¨rÔ*ôüHÌ QÏ0ò‹â§œëÙp¢$ £,"?ÓF”NœüPô¼–ÎÐ@÷<‚+£­3%—zTrôùéÙˆ  dØø95ÂHn“&Ãå(>ÉOŀЕ:"Çžgu.ë'²²-Œxš(]í-½¤gS<ÈÓ?ˆÚ¶~"õÕDÊï?u_MA会‚ýÑ^€Ócf—$òï~+¿þë/ëäË:!òe|Y'_ÖÉ—uòe|Y'_ÖÉ—uòe|Y'_ÖÉ—uòe|Y'_ÖÉo_ÖÉ—uòe|Y'_Öɯ/ëÈ—ubäË:ù²N¾¬“ß¾¬“/ëÄÈ—uòe|Y'_ÖÉ—uò7¬“ÿ%„d>€ÿ]Ÿ±Ò'È'?—ýõ4ïÛ" @Õ9ÕÀ¼þ=öÜ¢Áã'²}#Œ^ËIè—û±Crâø.;[?HwvFC=yä·ãUÌ\>±™ÒïñŽGn^èÂ×CÌY/¹3ÍÐ7ïŒN8¶ª{ìЮùX±K y¨üîX B~ºé9¡¶Ü!Z½ÁmòœÐ‘Dé-ù% à‹Ð ~pr*&½¥¡  Ø­À&¢å¸Cu±·d²qãé)¼NTÎÁÏŽß½»êÌÚ‡Ú¸ÆP¿œÓ hÒ§²·žûÿYÓÞœ‡ó#¾&Í‹À‰oÛkzçq)š7ŽƒE¯ëçÒ$³‡öfÀ4WIwýäáÎÝËÀ€÷Z¡x-é²Þ2“ 6 î\”~›´èì¡I¯ßµ ‡ é¢Ù×chøŒ¹Ä` Ê¡ž Úõ¨V÷°qoCˆÎw.…=¼GZ~jãSϹ3ajâw‡~‡Ê4ã!ÜRv~*öûa±rr­Ä%åµ0¢mNy-Œ1¯ÖLy×ö/ýÇ~$,bžÍRôÚÊ…4F¬Ó¹)Cc0Ö ¤O"j íÿÂúÆTn#ÃÇ&n>ÃTÇ»„Î&Üyüåη.sy»\3Gã=6|€r#¸fFÆÛùðR„C‡€p!¹åe'¼‹VžRÐy–Ç ½—{ÍlK »©yˆ*ø94:1U› A¿˜7O^‡ýbžlvœßà •‡¾KV¼.:AfNÀ3ÍéÁù›†¿HÇÔJ‹W!xã3c£:W¤²§Ïš1òF`ו§h¼Üq3’Á{ø€9’Áhx¸‡9ÈÀP@À‹U„!Auš[ ćٙ#x]=üÒtüÕ<Ððˆ#Œ˜„O›Ž/p ïn¿ú;&¾Úoƪ 9uÄÅôCí¹oY›Ã´ïüý ÕLcLš˜÷°Âs„䲓õñû;H6|O*Ž4h· Ç[ϱ¤YÍ”¢hø}ÞÉlRÃPí¿ª™×鞉…`Òh¹•ߨPu‰E ÁÑL[˜4Š¢ !ÒÀxoCsù.Ýo!hôf`Eùûjm!ú ÄAۥߗ£KñWXo^#±9 ¥wm¹‡ ÄQÛÃNñd# ®™P ¹Ö ÄXËvøŸN#èäÕ4_£Ôžù†¸å¤±W×ê^“#¢w~¸Ã½šÁ€’OéeŸ ‰aÃpu—Éé³ä–èÀ!«WŠYœ#Š“ŽN4À¥Ywf®¦Ñ‰`&,ò-Œœ¦÷Ò¼¯‹cï"R2Y‚ÍB©^Á•P ;uä‹î2ñ»ÿ¢g¢Ñoc–HŒ˜rm‚ ØÃäGÞÎ%Ê!wÜÃÝD»‡ËHLX¢ÌLQá«zfÃëRÙøª ‹*?S—D6¾¼x2¿H‘w™Å‰ëµù’gGd¼P L|fÿž™Â8à„ï6oI†90߈/>>‡rqWH[Ʀ=ìzć+H"©‰ð»6uÒ4Ì‚~äU‹ïêC—‘¼“™õ‚xAo#ñ “´D"±ISå=Œ•N"wñS’®MzÊÊjƒpu:‰D‡ DÔŽÊ#…Ü¡„\\ãüE$ÎÑ»½²í Ÿª™Øç”öU'Óøè¡»dbbS.XJC“>ÙKf4‚ €ß³’ÑS9!†ãÖLæÁħDØ rB ^¸x%áàCÓô‰ _•n`'©µi6>ÔM i•íû£da”J¤ˆ³J8š‰5 d˜ž™5ÒÙ VIÖÌ8{‘=X#1©–yͱ$ü×ÔÄ"ŸÀ£ç,ظm#èé™ š úÛL¢Nã%Ë8!7NÒcg¬d¶$X8xã³›¶ÔÁ¡ëp° §é»\‘cÂO8á‰RÖ;{׬æSô¡ï²ip‹0½Œ$ë`:ÉÔé1ú{øô­$1Åæ,ñU:ëMJæWÁ¨ÉWŒH‚–Ä« Ä5 ¦bÁP×Ì…g¬@êác¨1=Ÿ2b5b!>ÞÃñau!º ö½dV&-Ä_¾8q’_ÎYÉÌÑ'¶vaˆ)jÓäñ9žüE4`ÏiiÂÆÈÃÄëëÉùapˆúÎäi4<PL!:žšv¶ˆ4ˆ:¿ˆÕa‹ûP­#ê£Ü$h/jÎC¾@¸ëª9c/Z‰h8„~žË‚+Ž!S‡{ÛY|-5¹âŒ;Ëì#´a¬ñÁ‡v­P0Û1 TWžwu˜48óxÙU‡ô`ÕcdÆŠÏï¹úS¤µËˆZøki»ŒŽ¢¡{è ¤êS‹·-Çü½Ë€T "|Ý­;+ѵ›«¹/9Í4þ »çÎãKI˜·€¨~ uJGu|hº<¯!0Do³!I„›Î@žÀÉѨÕÚTߨ1œ€¨Ü Ìõöê.7ÃD•w”4 ª>"W!ÿyð£fm"C·ÅJ DE‰‘O Àa÷¨Žc{WÍȨ ;DxpŽº»J@o?J|ñ¡â‡XÀ€¨Î8v"TC!íw}¨Iy£sÛòƒ†ŠFo?‘Éil+Ruþ p/"¡  ~ ‡Ëû _ïƒ8±b‡>"• ñ i\‰_Rø?ñ»n,}Œ†èÛD;’¿EÞ"'_nÉ—[òå–|¹%_nÉ—[òå–|¹%¿}¹%_nÉ—[òå–|¹%_nÉ—[òëË-ùrK¾Ü’/·äË-ùrK¾Ü’/·äË-ùrK¾Ü’/·äË-ùrK¾Ü"_nÉ—[òå–|¹%/e2"ÝÞì’Ùˆ‚tõnÃ/„ñx´ ¹ºúeá…HGˆDdpš~!ê„CÁqhÑ`k „²7æ/@³‘HFI ‰xDóôúB$% ÉüCZ_*¼¾R÷¹Ý˜ç0QBê>ÒPÒ F¤cñPÄbC7D$H”·gáŠV3ªÌ…Bæ¨-#H±_Ñs€lh#‰Â$Âúínÿ9ñ[Ä2ð[ÀÊ>@¤içWSVªäïí +õ÷¼"N‚ЇRë5*Åi€ü™²©DU—ŠJ-ν@ w'‡¼1$ßÊÎLU©î“²©žµtg†ùCakn#[*\þÉ” é®ÞŒ@á²öüÉT… °:Pllæo®›ù›‹ÄÆ”R²åÇü› ÅÆ÷í(6vO MIŸå}!¾gy´Q)e µº@¡Ñˆ¡,Pˆ³ùOþå_[h—jnê…\„P|S{r‡(ÜRÖnœB;ÏŒV»hœ›:0"¹}i˜^„rÜË/xŠŠ—í‹MÉnÝyË¥Ï LCè›k\ Ðã®Å½tSÕ4$ÄÛ5¥ö*ߨ)aŠŠ+Z $â­+×wÉÛG!xŠ*¿¬IbJ¥û/¸3ÔåÜšFuõ9LJ²÷f²û•:Ð#RÈð¤¨×SЦôdh½/µ'­&C!~4#oÍvÉ ¡5¿ðÅý@ªÑÂ) Šõ³ÁoS‹k‘µ^hákD/Ji²ug.É! ¯yiѨ¥iÞWÞºüåá¯ñ©!òÎøÅ‘lÕ7q™ €ÞLÚ”h[Ð~×]š\ º{qp×æË‘] f>Sù]ÀN€¦Ë4/%÷ûö럗’ûšy§Œ]úñRA9h1 ]³å.á½’Õ•ˆF¼<þÂóû6%«aѱ½„úô„Õàc8„ö¢ÒHVÚVhY’` WH“¿Äò<6›~àö$0eþ2Ž›/˜§ˆ–ìYôå×í7¨-L³h× ö¯Ç d( »MmcóËg÷—Eo¡™c— lX“èÝÈOpÌéõwð¬È"7˜ ‘õÔ•ôN@Ï9h3O>fÃ!>=ZFd«ç¯ÇÆeD²(ËÇš@䣅¡_ö“è»ÐTå"3˜åcNd¦a«±¼‹ n0kø›ºÜ`Vð.7˜åƒO ´àñήpƒ{õ÷ݧ˂çx2‰Ôt4ߺÞô& QGéšîvñ‘ua iQì2˜Ú¹ýc"ú1$‘dFÜ>0‚ѹ»‡c˜â>émN­{x‰aRJjçÆƒB8/©›Ì^‚}tŒÐäiûÙ™ÑË“hå„h2yÊe»ÉÜ%l¶4s¬»4/©ðBm ǦÄÓ¼ñ@~Àeoªãì䯒ɰÇY HrtÍäjqr¯Yi5®^<õ)Ó D®r÷9H("©³=SUÊȇɘæùZÙ™ÎñêUåÀrrB‰2‹@>[7i¹"¢ÉKyɰJÓnYXêÒé0\ãï+PAOO¶ùSôãÛ=Š&U* äOÍÙÌ–Â4NûpæFhÅ+Œ¤ùÜPÞˆc‘Êqg(© „ÑÆHa£ê:\ùI¬¥×rþWYè@˜b¼ÊT?F|H6 Ù&Ì‹æ¥)ÁK`/p‡.vºŠ Ì>‹Í½#›'Ô´ŒÌÜr¶X í–;2s˜8cÖ Æ‡L¶)§ö²K †–Y:sÅ/DudD]¶"ÙL[` ] (>àR*7Hå„`9„4ºÓ–›?…ËfØNú§Tz&×’?¥Ð÷6 -•ŒBö’ò0ÆdÖI±s GŸÚ2ÿìÊAܽdþi/Œ+ÃİídºíÊ01½Ì=äþ© Âét¬.¹›G ØŠê÷°6 µŒÀ«ºnåìï.òÿ´§é¥ìw8¡òOÚ)„1ªštÙŠå“î¢ Òã¸zë5!£óFöoѹ• †»dÜì kͽ¾«Ë ·Š;q—ÁUÒçy–6K½òÓªùŽÉGè6¬¼cÐHÝöWœÁ°‰Ö„5¶xÌ¥ï¨ô¯ï.)$~pºVßQdÙ®­ÔçÌÎàÂc‡}û‘}x÷´Ö·ìÃm´}û’}øð3ô%ûðá±*êæcû}ûHûð¢kzÚ‡kêëc]+Å©²_¢‰^zeÑt(GÁ@4žÛ•¸}‘._[,Øo£»ÈãÛ~Ûâ”nïWÜÛÇ#>J@/´ öõt£¿­ËæÛiÐÛ8¦º“©·U½P'e/+BQÊÌî­—/4jw®ÌÎÃv6W|gÔŒò*¹¨å3T”?°Twžl÷átù­ôC¡Ò‚ ÏÊg ¤PÝÚ¸tî~ y úrU“ ®Øé(ää Qƒ©ŽAõJ„‰ï[4¢p³Á,9Lã¾eª1Ì¢¸e°3-n¡é'ê4ù“KcLj:M} ÌÇ™5Ù·z; ®ë4»X`óIΕÿû0áäÈÍ•šúÝ;QrùNŽÈ±µæÎ ˆ’ßaÑŒ‡8ZeÎe² µ›L©ÆFã™°éfßùɉñÈ=.Ê2Ù‚çÈyzØzíÈ.òYΑßû9èÄ¿p*:gÐäzÚÒíRááìÆõô¾áå£YºÝØ`&¢ûŽôé½ÅÙä÷ÃÜ„÷Ù,Q‡# ï³å.%þgËÞ}Ú˜îÈ .Ð㦷(G^fsšµu¶ì¤§ª‹Bi7ÃÜ ›Ð’F7.4áž&ˆuõ‹½Ó9ëèó¶K#dÚkﬥŸÛ5Âßk­ƒ³¦~o×”rTuݹ³LUžEãöi Ü‘¯*Xù²¨‚w¿IfSð‰ä5SŠáLœG&9àèÜäa—pÁú‘W¼Žøò4 „“ÿ™šçЈ>“iY¸( ÁªgùP**„£þ f"gÊIœÁ," ›øsÆæê4]ÆfÉi&âaÑ'ërùUHs(þÀ¡mÌ ÑabÊ7è¨êp Ïò^>Ä ì ‘ÜA š–¤Å„%…‰Ÿ6^sº§wZÁ¾$mPIÌ'ïšl_¹ÓoDÞê)Úq:k7ßHÁôûMGƒr¸}!´Àû‰°÷»æÖ,“¿G^š&_ÖÉ—uòe|Y'_ÖÉ—uòe|Y'_ÖÉ—uòe|Y'_ÖÉ—uòëË:òe|Y'_ÖÉ—uòe|Y'_Ö ‘/ëäË:ù²Nþø²N¾¬“/ëäË:ù²N¾¬“/ëäË:ù²Nþ°NBí$$bû¯hȦÿýÿøÛ#•r!µýWþ,Èa]K¼Ç/÷­v(§þ7ÈÕ}$YÓ‘iüLêð¼Eé [½NÐ,çr¤‡–sZ¿ÒC{#÷¼•Ë£¯@4ïJEšÒ$‰f OôN}ê¼:ðK$-:¹%ÑÔcÆÊÆx+õµ’ÃgÿP_‹A· 襫¶h^EiL¤{xwý½=T}fV#WŸÑ,<ÁOi:´kj±8œæåŸú”æeÚ5¹Ýó›¢{âc;äL‚ˆfêy)ÙuŸ™ZŠXŒ)hÖ]úÔ\91ëS?æn {¿æw ZжßDŽ‘‹O ¯ú iªïm¯7Ë }Õ ÒÔ§47/)®]‘°°é­Ö»Joäj]ÕÿrÞݑРIvw "¥¤`¬»CÒt¹§=gǽÔî9ómlÄ^ò…‡¢ý@Ô¹öQÿ[^¿/5¼>€zÎÑå%”¨] Øbçq¨«Œ”ÕN§ÕUû£E-Fb5õ)m×BünñvmkÄŸ×v³”µ!¡Ääñ¢Jõ{ Jj¯XRb27øõ×ï„f7/¡Ê؃RaÒαO-’·ÜÓ&¯ß¶uë#ò¦½,r/áÌc•zÓ·r'ˆ.‘(} rr$ã©3—UèEú¶~ηÒglýñE:0%„ؤŽ*C:¤Ö šÃ!tJ`RGµ9úU–tT­K‡èÀÆÑC:Rlí>yBÚ8åNãŒD¹ÖápT…4©“UÕ Gk“`ëttçsdƒ4±YË…hq0€tÛA-æbK$Ÿ #¬PCHòN"ÏSº°:+i¨î<…i¨f„%‚õ—ˆ^¹ÜÔ¤½Vv/LJÐÖôûíó~ËñS·.ÍÙ øZÉò¯ðÐÎ;_ÇÚJ-^êÚæC¤o hàÿ`"¿EB«h€šm­~P#Q”C ³µºôÊ×[›Ïè‘‘ÂðÒ5]7~"3ƒý3bŠ{LöÏÚŒè“ý³¾#3«qdfS¯¼>‘™CùèúDf4ÀêtØ‘ Š=Œ¢;Ûn.NOçËx…ž @½n÷ìÔ,3(øµ (¬’9ºí’.µ%¡D³mf.Ð@t›#ýh{ÓDHúÑ­8p9.mšÍ¡®%×[õà´äz«Ž$á°@±ØÉ£l$º‹’vNïˆ\B¹u‰-^€(6éOÈÉ)´ÜÛð$5鉈"#ópJM%rÇc¡½BZ¤x*„tÒ@ðYšÆÍîC\!šRVWƒYr=¢Îl0'qJ]Y´p0÷ÜKÂA!oæ¿•røX©«ûìÑö݉¨á7ͨ"%ë=£öLè½yœo Ðn‡H`€¦@~$ AÐn«R÷(RKZüÄV]ïÓ&  þ-HÔ÷å!KUÌ®<7Kië¾ÝNù7ê?±TuëHCmõ< çRæ9xeˆ,ýÏÌ„ 9Cx-¾aó©@“­þX™‹êBükaFDÜ‚é¬ÌzŲæ£À1þÖìwȨ †ƒ’VeSrª:‰Ñ`"“t˜ëÆrî ƒ \h 羈ìK@Ð4Z"è÷#W&Kµû^Ý_.Àæ£dòoã!É+p𔤞À²4í¡x­Ý>“"Áý¬A8œ¦­ü¢ñÊÒÒºd˜Rš¾išSd_4“2QXÜgi‡¬K¶©J”ƒ"\Yú®ãd~°'‘­ß£ñ>¯ EûÀ8. #«d[šï¶zZâÓ &êA˜2“€I÷@D£°?|gD<âÒ äŠÐÙ•Ö0ƒ#‚æøÔ4û0ª2QtêLcœ^ ™W²*¸€>þ:à-Ð'YUŽDˬŠ*C¢uUq‹uͰª6$*æöÁý®A¢´Â_³«_DcÀ2{j ŠÉwÑü„`µÝݶ¤ýõØ -bõÈ#X¨¨4-Ê š}¯$³ÈŽh¯$³,îV¶ið Çíí^NrxˆcâBÓußü=Œ ÊQ‹\Ó`AU’ÌriKqªYÛñfNm#XsNò(¸ÖÀ¸J ]ÑÀ`ÝFh’•L­Þi“q¦9%½óUœiVX2ÉJÞD§9j 7iMø9;ùR @‚œf8ÊZ›/…/7ÃÞœ*q”Å»²}˜ùRöa·šMÓµ¾®IÁÒш諆ŒŒÂð¨l;ÌU¡ßw`Éß´•öýwåWÉÙèæÄ{ËFDãšÅ ],y!è<÷zdI›Á;¢ ã@´‹#jjRï`OYª»“{õ¡‹Ô§z!0,9« n½Q^‘Ü3Ëô“ÎBwÔ¢H)DOJ7§®â!p>ã'y3–ã:™ ‹ užI¶´[Ô é>»7Ô(%2é>(B!¾eXü EbÐx¨jÀÊ‚(º3É”;ü°J,I íBv"ðò|hÇQ±iÂØ‘dM‚•جµÞl»KŸÁVÜv ÓdxBêÅ­"'Éê¶[ä„·äJ¶-$e¯ Ž.^n^zÄÚ ä$KuøZjÔs‚¦f˜Ød:fŠ*¸èËØ²`É óxÁç¢'ΞÞv/fžG(pñSɹ×|*ú÷æ7³6%{χ"ß0ýE“^mRè}zñ\Œ„‘©†Ô’;f_ž;£µç€¦•LS§4z¿¦”nybâìŽø0­D[v™£&OŸ3v}˜Çd°‡¬–‘Ü1™^nÃc$™žuÌ™ƒn¬$Ó³*³jjÿ>=lÁчMµ}n‡ á…«žÙʸíTdìÿKD¤XU„Ë®ÖRZ7×óìMÏ×é*È@p‰é~Û&Âóܖ‰ðôØß‰°ÿ¾|Gsç+á:^Ä꿪D7>¶vU¤Öe¸gQÆòCæBýx%»¶#ו0úŽË.ÿ s9\™µî:lÕ‡,¬Â’pwöå,¼žŠ•ι\%Úâ¡«ª¥®…‡KRÝÅ>C®Ú¹…U<ծ¨ §kWM<ç<çNöÿaº«îdÿ«œ'?|üÒ›{hcñøw£NŽ´öø‚‰d_?‚lŠ>/üCá#7ñ$ûŸC«õ‡¾{˜ç ãr¿_?…¹¹_q}Ïåì `Ë÷YY6GSÁpVWÉ-\ÐÏɲ9Ʒ³Ýes…3Ôq-eì|Т׵” Eo5˜»#Ƭ;øTÏ*9ž¼kÒj¹p³P óÚMÚþ÷«ÁL÷Pf/ùõfÝ =G>Èõl}É.k¥øÏXZ)ž.ãà­To¾.N¤½êñ‚ØLu³M+yÂTY_‹µ¾ÿDS(k[Y®G¹“õÝ%±ÄWɆÈÆOΣâ冢Õ\ÚèšD]ò¨ýƒT÷IÕ3¶ÚÜ'/9³­ö|¬‚luäëd¶¼Õ¬.Qíd‹Ð¨ãøu²à²Å1îÁk¨'_'" ­^bhëÒ[+žöUÛÙZõjr™\h-‹¾Tˆö˜ÓÑéYhÊÃPk¹r]î“Z›YIæ@kËS×Å®¹H7€ã±©òÕ@®?¦ÅÁÊ@ìZOæú½|½¦6‰ÖìÏBàƒ<ˆ¡ŸÖGVöÒ<³õ™…Ù…= Ï,̦·vëË¥½…]¢ïWi/¾éX­A•½Íì> djCtr"ñܦåe…p3åú‘étD¢)̕۱BèOéq3¡ ¤ÆÊ"çJ¹†p!5E‡±_5ÎP}'KÀYÖÄáz€8Xä‡ '1k9sÛDvz=[ÕÔÍT!Ñ|¢â¼€™íP©~1]4ù ûiçÕÕ˜7Ká…¬ò´ƒ›·:Ǫ¯v Òžv Ð_í@D~…/d>í@`½ÚÁÈ?åïO¾$“/Éä/ÉäK2ù’L¾$“__’É—dò%™|I&_’É—dò%™|I&F¾$!_’É—dò%™|I&¿}I&_’ /ÉÈ—dò%™|I&_’É—dò%™|I&_’É—dò%™|I&_’‰N~ýãoÿá·Z~Åôÿ毿þù_~û§ÿò³2ý e¬Ï‹ûË¿û­ò’_á©° {Ãùë/ýíO­ýù/ÿþ·!ÚŸþô—ÿöƒŒ?ÿg¿…ÒÎútâüÿö·?}>×þé³R®øùíø  h=o¼ð¿>;ÝTnüW¿õÕ¸eŽãÊ  >©Ët /J]—VtÑiÜà ‰†¨Ò#jÅ,•Š ä²¥%XÌv©KZ‚2ó>Ê$×!‰¿7R¥»(>NñÆO¤åÅë‘ÁË)’)Ìâ–[3o ŒoÆM€`à}“7f(k±ƒlæ!"ÁÓÒÜr£„bV[›ƒ@6Éy…ü ŒE< ;Ø;[ˆ¾¼þ®µc’ }@( >«sV[ÖRÝçP6­nî÷‘ªQ5?cË£!ævDé&‘I`²kU?6Í ªçÖùà‘)ݧS’Ê{‡Ò¨¢&@.{ÔNwì³¥ˆ:›‘r‰p0l9 Ân½7z§ö-Ç¿PZeÝŸsÁ•ö*»c ÐЪŠp™Š­CŸbF>®Ou½m+x·-φÒôº-¶wU?²ÄØþœ‘øº-L¶?+}ˆ™ÄøºMoÚëèuwMÆ[†lÝF¢sÞø–juzvÇéP÷QGZŒhO tBw닽ªIfp€[…n/CÜöò84OnO2D¦÷®@8™æøØŽ¿™¦ÞsK7·ºÍ碎Zªÿnâáäusµø CcÃú€{vN½U–Ò8<ÊõhžÃ£\ÞÉ£R ]Ÿ’X]‘O߃¼žP=ÖÀ<¢aöÙÅ ½â<À;Íö*G‰À=h—òè)ï!»b5Ç­¸o !cPÆ­lMó[ånæH‘¼3§Ú=è+Ág•þ,Oˆyú‘hµÎ ~v dÓí®ÑYdÌòA4:Ëp·í›Ó|î¶}멆¦œ¾Ð |û’Ïîn&§Ìk„$©ÐÝ»žÈj¢»k‘,VÝAj¢QIxên]êÙÕã¹q«?IÍbùl«{Bùìâ•¢i5)Å=©-Î¥xŽoRˆÄ™¶·ÎøVê)®ïéT¿~F‹ã_³LvkÒr7e7²°)º†ù³PB×j…î ± ¨—>×|©]¥…Ä:Måòù›Œ Ì¡oe ¡Ý®k&í@îôúBJ¢k@zTá?H§¯Ã•[Øi´;Žè2R…¨—GÎTHÓCµ ]•”#{6W u!š"Ë¡½@p‹uͦìÿ­Î…â¬Cï¿pë—ŠúdÒ¼"Öq099Êí}:.¦Tÿ!Qe8¾„Æǯ®p(§ À¡êvh+S¸ „ÄÜ@à`b{¾”¡ .KYü@øªýé[ð@kƒ©Ï òg£%>Aø!ð/qáJ»ÄK_…k ˜úó ð­,Q ÂaW#EHÓ}xˆ„ǃ~½pO¾ùTpPÂ-zëyE½‡@`¢°¯–ùu? „ÓÂ’b~8Wp ]‡C~8, ¾[.›„1©ôÄ„úç!k.jóÃýd jôìÜ8‘ªï™Z¹6aá±·²_6}íËq¸­†¿6ud‡C¸@*E®>Dî^zŒ|.lèhŠã›Zx#4@ÑmÈK»¸ lydìâ.™ º¢,}¨±}—ç@ðÜKN“úÌž›üq TÝ„ËÂZòæZË º¤Â3ÅZÔCÞfmW¬é1¶˜Â_õÙ%Û­åÅ/ÌHŽýIk&;Ö|™l­–?gèç´üÉ^/«º‡Š»T«¸eRÁ45½.­%­i6ëS/Ü|–úý¿ÎHiF#Ïí×y•xò)VÀ‘e–³2dqÞšÓ£c2×ÈÐa±Ó¢sŒ`!˜^,Ö”EÖì:¬)‹¬it °™âM¯POæßÍÓv©ºµ&ãB”c¦OÓa5©Î†NäŠ.êZá®7L8CäöÅ  ÅÁk¸Ii¸Š<ˆz¢iÕ°#–em×°#–«¡#ÖÐnq ¹Ó s%=²>wÓøË;Îáµ¥ÖN.Ã$Æ@0-œ|GUë”üöÂnÍ^L—Ép_"ú¢™ k¦Šý'a?±Ü-ñ‡Ã…¬ë‚Í~Ó-¹ú³»™¦°7[Õ:mî´lÉ@x˜_¢<‡“šÒÎ=z ­M7×vuÒÖÒ¢íƒ4¾«^¹' ›©nÿÕ«ŒÞŠúh—N»ÙH­ÂRNt/œNd2×W»eM9A X¢»ÿ  Ñ„w·IÀF«ù|°š¦»–Ó]Ûì[m¨€i5žÍÃKO£!î8ˆèUÉ5!ø´chš›2¦„a38ù]#XvZÑ)l5jÄ¢©µÑ:<g"ØhT›p¬ÆSÃ0ÞguY@Äñ>«#¥ ÷×Ü*5m4ª’Î@ª-ƒõrßS‡g‚JzWø)êÙë¡ÿP5é4ÌÚÕ¦*«n¶juÙH hÕZuº ›£ê`䪋óxqØ.Ì^Åq²UeâU\šIÙÞÈý!šæ³âÈÎ’Dx[j°UM·ÅÑ•%‡ˆQF~Wc;‹Ç/yF¢õ)žßå²¹%‰0èÔmhRIÿqdªº^¬råBzu” „.¤>¯8ð¿Ft!Ã…Ù°nE/S­ÊeË…ÔGæUíaÍ5BßTû`,z7ÑI5ØÿÞÜÈÑ"üWYeÐü7h4ö R‰ŽV¯~N&_ÓHœú9ù <Ž…‰¬´ÑÝ7vÁ‰ÔFwÚ+ E -áwÂ!|oÕu C0ì!íÒ,—ëå”ÅFXìòü¯?ÛÜ›­p嬸AŽì|=AB;ߣAì|·\?ÈæûÚ;¿´ºN6‹Øñvr`€°z ;]™Â¦˜á‰@`¾»–¾©ê´@üí“opûÔ›wUo™wЀ}õ–@ø\>Ñ‚ç’ó 1$ ›Ý„]†Í;›°ó•.‡§¼D zL#ô¥¶V è˜ËqÔyåXoÛi sÓšzÀ;^TÓóåsżEÎÔ²ˆ‚!¹|Œ¤Ò7›C`²Ø#ëÀ3^öÛ`Fovø944ŸËýâšOGЦÜOÂú[ý‚µ°Ÿú*ÚË…ƒøIn—q xòi¯¸@Ð/¦X­„æCáhîÇZt~&SHì!ýÀP'‹Þ+i§e8ƒn¡@ìî>ó±x¢w÷¼„æîC3F rp&‚;3 çÒZ:ÓÈ?;>’CX>04_ („†Íý:F`öÞEÆ‚AÔ]«{E­Õg¶ó‹=kggxÙZ,æèÝöfP%Ï<Ì ÷îÒõ@0òº²À@Ð#»üoÄ«ÇV­ÁXl9^7C½9%´mw®M¿ãÞÄõÏioSËÆgê8T0?sæä6ü':qX¸5#˜¢›ƒà’ê—°é3Ú[ùý€'{óù> ðzý6è&ý³Ò̶y êÕ{×@ð ª ÁP}T3ÖN†W>ÈðŒM—Ò’#s/ŽÊЏKsþ,Õ•0tÜך“Eߟ7y½lfm©úÅpVìåxÀ‡óvܸøø¦›Ø,l#x“%ïÙ§äú²'G\ŸŸHŽ˜N™õb{¾I‹W ¾1 r{Ä¿›ôªRòm3ÝÛîUL%¸ÍÜÄá €çÆ ÞÜí•,B>A‡ÈÊö)Ϥ–EôÄ€oר ÚÅž†7‹:Z*ÞÐõÜtm©1YÐÿAJö"–•¢3øuÇÄ4ÒŽ¡ÄÁ´…ikþœh™³<¶eúˆÆöKùƒÌ׉ p«êø” $â´¼3¹vÔýÑ™2>N®F¶ ƉT"¥þDNöXýÚ¶AÞƒxÿh4{f—M$ÿŒ¿FöXýuS-;¬€šý•ĶëÏ?KvV!ëfg5r^w%"¿êòÿJŽù[& Ø%‘#G/çU}#ÕOÈ%éGª‘iäçéE‰HIýN–R¿?+b€ @®“æ=ÎìؤP©y*«l’fL¶¨úX?‘ìøu“Ý8rËTɢšY$š Š­¼¦^²Év^þÛè@õ—N‚ R¸¦ŽÀ&Ýíä% À¥e"CÚœæ÷JVx0ëNþàºc”“% è]]C‡ñ$ñ‘TrDôkF@oë>±€®rIÔbWsiéÒ f i‡d/d&¶¼±…1AÔ™¢<ò±Zÿr¹ë´"þåôBÔ¢h°æHÄ ª5WÃwö¬ÙŠè—v äâÆÎœÀ+4S³|&ëS)?“•?Ù_ ¢î¼dãCy†h¤È·tá dÏôî “YÐ`›ßpc6 ¥Mp ·)>Š4–Q·´$D€6Ò­ó•—ã]T#¦e>¼Î’›ŽF¿Ò–Ùò@нÊÈ7ø`sZ;¼ªÈòå%`å–š/f˜íÃÜIð§ýb“ Žõ2€Ê‹»ó½œïô.\rë‹Ì äŠð­†Y(è 6ö5@f¹Ò«€ N?I!—äsþ¹Yñc3r C”öc`UÑÞ  Ìêl@š˜ûKÁÝÈÜz8Ó×ÁÁy™ûZúš™ûͳX3sßBSìÆ*Í>½°Š@ݰ¥ǰŽ(s|“\ š!: ë+÷‚B\÷PóêpW$änVT¨uìEQt¡—i^Ý6˾‹w©,‹BŇ¼‹‚Ÿy³@Ϋp„þ½GÅ%ÍÈè,IÑëíƒÈ÷ ].ªXô¶;w©Qé¢Ó!WÅ0šÑ:(£‚F»¡ðîeÍèXÏ“qyV§¤&ÅÈPࣵ†Z(Òv¤>Ñ´ÔL x¢ O#¨ç™y¥„*’ü¢Hª%uÐNNb”5é$3ÈëŽÒ'­³ƒí:L ¥n.ª¢//ë°¼9¤G­–~ó€.ê¹ô›eþˆ:×à™8ªÀ4;Î atÿæP5À[þf¦åQ‚ÖŒ iØc†ºªmS«Ê¯2Êß4/F•k7Ó n¹Ô/}²ÒŽÖ¸(ÏËMñ`\¤fê%t·î8i¸ÓâåôæäŠBXϪèPG“ÁQ&j9 K¢Úñ¹y2øõh“¹Ó(ŠÔû“shmæÃÀüEÃïOv£hr Š4Õ2)BN’HhþB­§ÚBΦQ ª1=ë×\DJ+RTÎ:ýïO!ê*Ÿá2®º½S™ªÑ«¹¡˜ŒnF)¬:óäÖÓÕ²€{½¨§Ý‰ ¿‡|Úâþ.@ ídÌ6ªvµW^ÜÌ Ö÷A“–Üa¬ªÒãã.(ÏÙ(4fäy.Š'RŽÖ¯ü®®ÚãÜP/²îù þœbRf SÕšìe›ˆ:÷bÎP5@Tö–)‰@ GrÍEœd#tÕ¸Ù“ˆ¶žë¨v|{‡¸¤@!‘¾œÌ.”(5#±k‹Ú'ŸøXiV (Ôš¹={Röºíu2ƒúí¯ƒäÚ¬ÅßyF•ÿ-þ“%þOÜc¨Äÿd„Bê_g¿BPp:+|Kú¹'g`¦dZÁŸ¨Rô¡³Êd@*Ê(3 uš5ˆ^DP/sPƒ8Å3Î)Rƒ¸ÊWDÀ¯IÅAc*øH”~h[¤ÄV>sBÓÁŠkBÉêÉœAU 3­8’³èÍR¼Ðš§ˆoªbY“ˆ#ŸœVQœ!i4™Ì@ŽÔ?\0IQâð7QFý#¯Úk áGÁöMdIiDÏu;Ų2T Wª&¤ù²Ì–F6CR(NLpöC1ˆi¾,”#ßõ¹e’³4QKÿÊíT*¼¬ÌU)ÅÜD |23>rÉ*Œò¥_‹4_ærª×r€s)x¼ŠÔʬnƒZ‘&Žh¥±™§Ù²‹õ@Ú“Ó‹WJí![š=G<A©¢pRz”œ™ ¹‡éL )©]‡)ÏB ìŒe&¥ñ2’ûS©ñ2–”IÛG)ÒºÔI~N‹ Žäù(…ü2‚6-¿‰qG¨u(ûKÕ)z(Ç(ù]œQ†Ø›Kƒ–¡Zèlýè9z|mOO½|•±ë :@éù'ÔáºÏ0f<ª[ÁDÀÔýAD° }Aš+B0_öL™ùBˆx“ÍÖ}Ê f…8N #ûÑW3"õè˜J4±IF.¤àœ’ ÛŽV– Ÿ@¬ç›{¥(XÛÎm‹Õš9"‰t ©¤ð›é2¡½£ÑÓ%›Ò|^ZŒÈtiøÙ’¯Ÿ,T‹åÚã å·æfïßA?1=––‰õÎzî¨+Sk)S¸*ºxu\5´W½I¢ÒX®RÄ­ öc¢C v‚êmfP¸0kYÁ,¯JÀßM*ê#d±)ùD9Š1BÊô±+UÈ‘ô1©—Õþ¢·YòŠGWSR,9JV©€iGMHèÀÆÃ)n› iD ˜SZ¢Þ¼‚ X‰ˆÉ9µÊ—ä!’U.}÷‘çÃC¼’ât‚3ŽTC] ŒÕÂ?›ÄD§9'!Ý/®øœúå%¢×´FÊ®ª©–5?-»äOª‰zO¼°…*IõY’<*Õ=x1Žˆféu¥Zš³×fx¯d‚fQOð¥{LÙ"EÝ‘ˆ¤{Õ§cÃ6ˆ<œÜ©kÔ6±=ëo™à û6‰!ú¼ä‰¹ë­r¼âóK2²%–EÌŠ=—Äœ5¬—@Ô³â!’6Ûz环3ÞâØAÀ>ó-ĽŽÅνƒj·ÅÎMÂÆÂÄ$l‹—üyW^̬¼…NY@!›>×,¨‹í µ1å¿Û=®/ z<„Þã;“S¦€D!€IáíãÒ¡ÅÆn7SQAç/‡ˆè¡Tr¢Ī€‚‰¢cRMˆêÙ¨‹Ü ³¨ºĈ¸}ˆ ©JgGɔ֦r•ÀP8ˆJ`ʤ@ö1•rS§ˆÊ>ʦXø©.‰bM¡;_Š…Gzžw®…]ùø%}QåfňÊ2kçoÞ7K^ó¾OÉ ÿt½ËäÞ'ë]¤n¾OÖ»lþ`Ž”?XƒFÞ;ë].4±·…k£¬`keˆZ Þ ,UAÿˆ«™:»Å6M+*‘ Ç¾½ Òô{jÙD5þµ(«½GV3í߸˜ [$ E·¸”ç4ó«±0Ì(dú ¶0ÌvÏŠ­FI÷ÈIªð !, 3ëÇš®Q9`—ôÿ¶âqT¡ñ-IyCÉ|;žekWÿ^ýÒr-ÄZU÷×%[¾r‰š9ìâv"•ˆ›ú¢-d`UÒ×ùÔËÑLù5"*ƒ`}QQ·Î-ôeõ†Tþ×q;P ´I•ö,Uâ²ÈB]øe¦ù¦L™¥“EרӰ˜Hɂˋk–Â~Q”yu£‰íÕQñ4De‘S²þË'ߨÅô·¼1ŽÂÒ©kÔbä{Q}j¥$þ²òø¦¬Y íß«:×ή´†;àjTkšGA-FÐrõRÝ6]£"-ªµÐVð:4wXŽ~©äˆFâÆFˆêév¥À2w(êK%¢×Å“|ƒÊ±>­/ úÐà’²šçÞÐR"0²BzëU'¢¬€æµ}¸.ï<¢8›oÆ´ãÍ( ûSõfœ#ŒÊp¾‡¢zœýÝQ¾Í0 ½¬£m……1P¹Žõc™ÓÕí]ר_œÃ5oYÉ!Šæu 7BQWUz9 µ÷úŒÆµ@Tb{5E.ë$lê5‰V™5ÓôêÀ¨™VEØ$ Qaú=|åÓI€@®®ÑJE!"„Âåß]ª •Óï´B()¨mèt‘˜Ž×„Dcð %Ú09§5µ‰H´I[ G¿æp>›¿†f‰«€Wö›éœ0.ȹ#øâç'28¢gÚF0üü‘õÉ<ö©›Kâ 9ì6jÜG¯?8Åý*טÒ¸{IÎÆiÚ ¼Á~óB&gÛ"Ÿ¿E^L“¿a•ÓdD q3MÄ+#}H¨¡œÔƒè`H¤’·ƒ©ê…X3.H1yÃòBT‹dn2‹¨…<Í%ùhÕŸˆ%âêÿÝÚ›ôl–¤×aû²~D.³V:Æl@¬…w4[ð‚ŒêÕmÕÐên’ ~½ßç ϽoV)B/:¿S÷½cÌqm!­iæ1]Ô§qÞû2Ò)T#ï,ƒ½lŠÔÅÛ¶U Üó·!Á«ëuQµÐ„\²fg¥n.™ýR!ulØæ§›:6m€ÔEó|ž%ùpóã4‘êÊ0¦w)ò9ÒˆÈHI»½A×;-‘JJß©F®E" ¬Ÿ´k\ÁmÔÆbo7¸k¶©+‹NµqË«Ûn‡dG*8‚"é¢Ã¥¾ QÎiÌóÚlqÔ+ù¯^ŠÜ¡ pŽnlFÓ<‡¶ø@@]¬+?z!øn9%un† Õß~0Uó7 7[íÆF!íZ3•(8²%Ñ» `Úž9y D'Mî&@À‰mvâŽÈ'ÉwÈOD`}˜k0…WÓD(¾ò˜ÒICÖ‡itH ¦²>Œ¨AfÖ‡i4¸ ÂóÌcPº³cûôð~ùÃ4)\ú’=c ¯vL€†Ü ÔJgWšFÿÈà‡ËÄŠŽ´à—</uØé::ˆø¥Ròºú2Šø˜ùViJ´÷•ÇTQã/}ŠÎº6²]lØ1 ÊýÌCºˆújÎM:»íóGìêî.Âÿ12%èÀåö£†Ëo\;öt H¡P@AN×û@ª„j?(Ÿrº‘‰óÚѾ¢3„¤bMht®êÂ$ìfè¹ë‘Ã.…H¢Çÿ.„"zlñøzÑWSp2ªè‰ýÚ.´V†’EÖvaLA@¼ÒÊ?1@PWú¬Un|‡°féâôpñÍÕŒ€l¾4?Ò$ôQñ¯tï1:žÊ´‘ µe•fm÷µ³ „1­((v¶õµSMbEÓˆsê©ìü»É³R‚Õ?éVè‘z+ 8éte­…÷¿¯¼[*ÀCì¥æNÖbª0“ãæ_¨ÈJ"ƒ€*TÙ’°ß,Gެ©B•- ›M<« ÉB GgM¹‡ßr9 h#«B¹¨ý:ZÅBÂ@T$ ÷|B¼§6@.f¡øS(ºxS$¥i»¼ ÊíAÈ=¨Qd–¤Iبξˆ]%ÍEÖR0èƒ0¶„ENûÊFâPÙÒË3¾ [k¦È4M;Cˆz»‰UmîJ#H\›HŸëLsWÉÔï?aqÛúÉ”n×ãŠ@ð’ë¶ ¼¬\CÛ« †MéuÃ2Ž–Q œÈQ4ÌÙBk,cNÿ P¶ã*‰x¡k¶»+öv }N÷ÞE½´½]I ì ¿§å/ L¢›­F ò¿ Ñ·LwÚyø CÛKú /jt{M¿a’*±ß0y˜¡.·¡4;ÓaªÝ™×CÉ>óý» › Hä±;•Ø–Ñx¼/Û¥‚Qq¾B®·º¯CöïW™”ô}.|a£>ܶñùF³³:(ÓÃ$¶!ÎÌ l¬Î>% ` nSJËd<Œô¹@r»-лw³óÌSŽ Û–µAÎø§ÞïâИÅīӱdJ¨ ïfz¼ÔŒ%'Še—žYÍ–éspÑôÞs:*@T/F‡„á†ll/éÏçrxÕŸ°ò°ÿæŸò@¿$5·AˆÝ÷ÃCÄ&þ͆®ú0ñG£‹I(3JG tnI·fYÕ…WŠÞÕÅðÀ@lâßX@¯é6颭iµ °‚¸î`‚"«”凊eø l׎)C¬Û1@ð£)‰Àë(F¼ø$="n†è0Ãi‰ ¶®îPeO"ü)OŽ•ŽÎS­ÛÊ…^ŒÛ(}Ê#ËVA@ЖE' §SVÃåÄž©fk_XQé…„€¼áÛ8}0iìjcæÁ1¼™„Èh§åó¹Æî+BmsZE1 …€Z6DmÁpJmBäþ~¸M«˜ÂòNWO‘kÆ^Ÿ ¹ÔV…5ùöUïòÈØÇ¾ùœ—ÀÐKm„ÌïÂLWR@M:‡Á3ÞNë$|À¤Ìך2;ë6ï²»8Ãw4hš¦3S›VkÌïÙÙÄ#+ »º!¨ Q½ë/œŒâi°« @–òíØªN•£‘& s;ýz7 ÊÁ"›¶yJÓPJÜ÷ô+P¼nË>ä/­%Ÿ?ýjÐ+Pu¥qQ$E‘46éJˆt(xHÞîŒFø.#MΊ<›iøÛy[ô¾ ÇF'˜ÿDÍOåä7Ü!Õ¢Ê/|'·5iXêéSð+Žm0›–V¿òÚeXÚ2n«É°ÔcÈàSl¹ªÔ*ÃÍáWQË»a)ÒÈèOëYW‘ß]X¬:m}’/«ÃÑ– K—›ârɰtùâeʰtg8ÚEíÎp´.‹Úó)úôFÂŒí´ÌX8¸g X¡Weó¶ƒh î("öž¸9÷l¹*{sE”Êr:‘K~ÍÞì ö]•í6¥Th¨cýIöõ2ºx§çjs>-wð©†9DØMAþ²Ào³yk.x|oè%€‹n=ÃÉðC]Csâ€DÆBÆU ðíÕ˜›ZahNø.6WfC=âôC.kâøCj>gêÜ´ãä2ãN»à#ôî­Ý.AÚ(§\„·£,É@ø*½3ux±½q™ìºð߀œeFÝ*UìGPƨ23Cç¡Ï¹¸öˆ,¬L»ëwíQ ïâŠTDb¿oÑ U4h(8ôB/RKÑ÷½d!>ª#v/2CÁ1„Ø MoûŽPpt0.N%6úð#³/´dêÄ0¢T`q¥BÀ¡ˆQláP•™‹ïb\NTl ¤ò 3öÔ*s²Ù$)í{FÏ¢ü ÍÆO‹ò7¶_û¤¬%½v²…\C‰¢ÊÆ…¢ƒÏ0d>L¹9¨šÓNÁÂ[„(²u0+#ÔÜî‹,ßX}ž©’=|%@ô#ä‹…vC9´¢òÍi^Q$ ãv§¹GOŒB<­Â‰ãS‰¨ÜŒÂ?Ô|t»æMV€þ?„q¬Od³¼ÜÀRƒp#2+ù’¤’Ï)%oq91!}•ñx¿Ôd©yà– l‘/ÞRþmé¯R‚áÆ@ ª+&ÛOdKL`»ä@NGÅ®`ƒùX€–‘²ÝŒšoI¦ŸHÿdyDs…*.[ÚgMèUÐy⬨6’5¸šUµ'1\Á$\ïªËxÅy ÙOAGÔÜKwãšÛ(†ªYscäI@5w2 T«B ¤«YW'§ú¨®ÎK*•¬«l P5œKò—¬†sKþ’UJà@T]¸ªüP²Dûɲe'Z´±ºAu%—…5–ÖE[Í2zi°pOý©™‰Ÿòç¸D4«t륡vt%”¿\šEwÓtæ³Ü%A˜Tm¯‹NjáX,ú1>MLAW7žºŸè ë"r-#w¯ÚRB —ªý,EÇ}1B±êv‘\4‚U^»Hžª­·œ×´,¦‘$OÕŽ:›Æq±”¦ Ã6’§-¯»’o1 á7·E UøÍ=¿öp&ÓŒà›7köc„v©鬇m6¾Ø²D DãATÅVüñ6ãKb ±h”ÄUÞ@ZìÐz4ÇØlzL h«SXÇ›M¶u®Í~Ï›KÂ@áu¸VÒEo3í¢+ËÈ Ž¹ÙäÆ#Úb¦F½PЦÈ-#D'¦‡iiÖ^ÄhÄ÷‚Ñv¬ÇjAÃúÕf÷…Ã4N œ…”Ê2Ü,‰q¾T{Zòˆõ>ÃóEÙrûpaˆ–h¬TmÖ¤šìPxŒl!Zt(‹rt§0Ýèº-OÑ Š%MpËa7Õ,Yމª|³›‹'[± ®w#›â@¦ó&„<¦¶í, S›$޼–™9?ä÷?^oažd˜zò~D+DLâ٠窱ü¯ÊØêT|ž&\¯9¦EéNÓ_«7O‚5—.v·Œévà¹>''±a¡EšÆ8ƒ@<ןҾÚq^sý’Æš±€æ¹Û¦>%Äs}îõÆŽŠ³q×&Ïõ¹³ƒ]¾.æ|a›yd«Ž½Œc@Å¢“³%9/–J$÷ÕBŽõ©}yjo}j_¹–#}jʆcf¹×rP(z®zpñˆ×r6ÔÊýäZýñccKk9ƒ”Å@¼‚E*dAðÓ÷\ŒÂ¨cØD<¬ª6͇‹Ž@Ø{™ëÖWÅ”‘‹cPá;,cˆå³E€7<`ñ«Ä2ÜÛ}ËÀ& %ÃÁÄØ[ K}(÷Ã*¥c똹dȽ߬­ï¹ˆØ…Œ\h,B8ÂÝêX^5œ\•¹EY$„ׇšCòùíš9Õ#íFirM¨Ø¤7€ýá!¹?H…*ß7¬ê=åÿ%.>}Ê îWÁº8µdår/>å´8DBÔ_Ì»‹½fU¨‹¹›Ú}Âà‘Ûàp… ¦=׬-]ªñsæŠù¥kiÚ/u=ÌÑÁÌ+̹¨ˆÌ7kG:/@W%D/}Ñ/6ðU-±É¯E~zÈ åÆ@#°ü7š€Ki ±—0\B¸±µÖ Úðv©Ÿ…M…Ö×™:·qÅY²a¸ÄÀvë˹<±c‚éÀÕ]HÖ‡k†ZÙuø†¯ìŸ—¬`.÷ÇÔØPï»I³R‡6Hveó}eϺInˆÞ'U·@öeUèrRì0¡ ]–’@?Š_YÙûR}]VÕڦœ¤'²eW¢EKªhiB@³Æð ½Œ ãZv»Š½5¬&›TºðSÑ<šI.%WAþûááÁrŽ2h×p<ÔÛ Ú%=†Û‹>§ö Ñ!$1I•þ ö×Ë‘^‡_¸ÊÔܬÜò™Qý=\f Z”ê8%¿@ÔWM?Ó?¶Jñ&vqeP4Øm„Ù8oíj>º“é@U0 v¸]*åî»›S¸D3ïír*+w„ñöÌÍgÓâÍgÍZ“´Èõá¶ Âþ4=Ö§èG´rïYMÐÞ¹÷LâpPõ|òƒÙj ¹[.ã#o|oÙ%%¢`E/xƒ~9‡µóT¿œZør’ûZª bìÄ þ&g&]\$€’lX1ÏFÑyˆ­çýqé.QbßqQ£+\^È|ZK Íÿˆ#ʉ; ©Äô«Z.:'Y »‡äК4¡¨s!×""&ç¦] ß:lµt ň‰F¯.¾íƒæÈžFàÅèØÐEz§ï—y)láj2o‹bƒGÜ lþ}tä;ÁEžË<÷ÊV}£ßuæÈÐí \iýhò^ªÂ1 jThŠÒIÜ@Ÿ·jZJƒ[MÁÅJidQ…\÷Â1„hçBà¥W‡ß8i?õ&î6ÎÀ‹ÿ î¾dµÒÃl¢`à‹$ÇÈÍHH×–s¢¿t¨\aT' ¦ôÐx)7³·Ã̈¹s= $*fœ(Üžy@èê×̘ëò*lâÇÃ|gÐŒOt½~ÉÔO+d@lê7“÷ES¿K¯·/yú]â;2zf3Ùc}ÐP°%+Ŷ]¿Ä9«Žw‚𢍡þmf$ÚüPCͤ·™"êRbÀÕÞL8•Uçæ w ¬KÃÏÃ£Ñ >¶ÌáãØ{k_®·¤U ¢Z :Ú†e¤I|Su\Ž—9ÒhM-ÍÅ/Œ¨€aeMdÐ+e®\.§ZZZоÃèùJ™&ö™W"6ý˜\”_Î’H_DV"Knžª.³Ð+pl3e1¨°Ø@Ž?ç¬lÓfñçd.“œb^*(šÙL÷ž¶œ62æÙ‰ÐFÕz‘¢`å:gZGÚ¨žDh£z™V:䢺’Ï*—Ä›g,l ;º¨žä³rá'¬`["Sæ°'ÔòË¢˜ø£™ºš Šr§Ó†Ñ«»FÐÌ æµW"p¾´¦=4]€'®_!M”Ò%]r—Ú˜yd’+ÅaÌŠìxÕiÒ~ˆˆ¤â,Wçc¡°&“éfÀ%ÉÐ4NÍï`!¬5>¦]†À‹ue=¿¿ùšæ0r*ÿbõ— e×r¹ ôëöA‚²œtã‹ãܺŽ;cÎVà·\YôaỿCqá,ŒœG"h‡¶;<¦ðC],ìNªMÞz¸HDЖî+ŠÁªæäbŽHxS¯Dðßw>ÑbõO~'õÆÇÒ.pìãOõÕ;Ÿêv?Lñ“ö êÃóÛ2Çè´•÷b~†«Ïuä×=“ªX¬è„(­ß¹ÜúѨ‚ÞáÛ ÚÙþú„¹^Í#Ã¥Ü"f¢—èUéSï¤íÁé¼Ï[³ÑÝ¢N¯ɳœAOõ™j‰:é»®WŨ¬Û›È$ ‘?E0tW™â4¦ïbà.²êÂÞjpIá&oMz±&„Ë#áAˆÙ|ØÔ¿‹BÂÉÞ¢†sbQ×TÂÿ^ã(ƆÁ#ßaÐq¾Z>"ê2mœæ @4œ¡h%ýù´Aÿa :Ö²N.ê 9ob˜P!†éã–|*.aömÄÔ7•M Vëpg1 üe˜›ÖÄ®ëg ©"'¡ä1Pò·íÚ±.…ä “³m$0X´h+ÐkÊ€ÈJ DíÈ’9W„Æ?ä‰ÌÑG~ræpôœÒ.¥n˜àJ™ReÅJ™26z"—5z’÷—5Â¥<PHÔèÇÍðr¢F±ÀÈùõq≌æÎR¢Æèž /%jŒþ81Š›ã<€ ò`äÜsq}.bAî3#cc˜Š/W9ÑLd{8³È ŒL’q‰Û™9­\Šá0ó Ù)‹aQ$¢\©)CT‡CSHÒê—Â;L°|Cî3#Bf^Y¾ŒäøúµÌ}¹“å67ƒvõ*Ó©Yl´,¶FúãĹúãÌDƧ똹'6ò¯å$cäÁ+Åv¯$š…–~×Äçë&wU.å>‰’ ¹qó¨å{"v!@ÁÍ¥:ÔhÝH¹Þ‘FŽÝåLéMßòU3œoÚ|C ³²Èˆg‹¤ØÓˆ‡g@Óˆ‡,Å™b)¶´æáa ¶æé¢)¶´æÁ²Å´W”œƒT˜ ‰¤**T¡çõ¼í£¸ºó ²†kQ‘U†R\7"B'¨#õí1uÈD˜rÄ:’Ü–S§ë±Órê €ÎZõÔÞ£ ‹§õ$æ† ÔÒ!2äÛ[m6ïâ" ‘í¥Ç®öíÛ—>¸IÁkK`ð° zìÛ‚ÊåÚák»\§ÕFμ=¨è¬D._GÜñ4œ¢AÙK-nˆÎikü¥^/‘V`Gthù™­Ijîôqˆ=6íc¸ ½4¤‚©?sUq±mÆ´4&œé™&häbkº £´Kdñ#›6)É’,†k`kN[Œ.îN<(æKs‚@TÚ.ºýLÿ„ÒÑ›¹~pøRÓ{º‹1¢¾{ØÊ‘åïXÏåÜüRÓQÇr°"kj5çôÖV8ò†«­Å6N ¿Ä!¥…­‹î×Ñ|á»H]‚aÍH]Bù$¯Æ%YBQ»»º„"rÍ£ý#6|²ˆ¢·©D†è$\w¼%(‹+•¨$vñOÇvCBÆCÛΗGŠ=6i!Ì8éFP`‡ÝZÃR½Kkp%xz(WNÐnããèRÌaœ™¹~Ÿ¨`#í[7õ‘Ë`'uˆÚš„1Üâà ‰1<¤JÅ•¢£p,…Na¤k]¥+O¡fºtŒ˜×zÂ>õ!‰ZVKŽšF¨jbÃr[Ç`dG¼„•+ßCq©×îÊ­Ù KØ-¥—> ‡¦L bwSÜÂk–Ê9óâ–Ê9³é³Öžf®–Fö4s-]z¶å&¡0&ôlju‹FÝÅðÙ•N6œ´|†PÎ^Áqu³"ÂÒýD7éS¶¿LtF@JÏP»°FñëNЀŰÄ~²h>CÚ9Síéo¹Û\»ŸÓ( ¨Š½êÓ„2FM™6Éxðâ ‚}ø=»ú7÷/B ©„ô½(DkN~ ‡l›`oI’SaymÕ æ]Õð‡†n¦ÙÉŸ~Åš4|í²ù&qff_XU[$ºunµßBÒ°¼FÍlé*¾vÊOmÒMªwvá_ô¯Ÿ­§I·$ÈÍãÉ0éæ»ñ(/Œ¼QÔS0{QZðÕ†ý7_NM“î¦3ûÕKÙl•yYY=*¹è|ËPDþè✠Œü-koí0ŒÏ¡d–¼VéGdY^Ë<}fl  ÖQÆ5¤šÎÜZ²Q¡´4Òl"/«%ƒ´MS«[j ÓxtòµûÎÃAIm~ôÚ"ñ¹Ùßi1ŽõM›oz‘ÿrñ™ÆgPÞø]ˆ,ͧê•é°Ø_›H¿Œ@”Uì>›¤ôËy“ËÅ!“×Û²ž2c ‘ €›¢H"ÃrQ³B¤ÏÓ ‰’K†–Œ¥§ê.ý±yƒK5 ¤"×` ѧÜGS}ªÁ͇@c2 Ïš“!é<–"‘Ôs½1*í…Ž$¿‘懈³]”ø02@1R!Î|˜+08bÓ€A^hÒ°uÌE륓­q fªT}êêÔf†C„šÎ]µáäŪŜtVÖq2æ¤s/ +€Ð ¢¸ÝïÔö„ñ…¾BãÌcdB£ƒ?Â?CïBá a±¡2ÚWˆ^+p8u¨5ö(ýŠëŠÃ+ðˆ3Ù êQÄ' Ä÷C«´‘ÉŽ‘œ‚'µÙ"Dï½UZUeÖcd²À«Å[ m‘ù‰‚H”ë)*uóZË>ºà‹®é62ÍYŽ…Ž€™%DmC% ?Ü[T#be—/ª•îcͼ…6iƒœùð’AlΖÝÌ6K©ÌŒŒ¬ø­*Râx( à ¯Ì7‘žt\féGðÏ!â?;=uX1ÅÈÚ³¦Ëä½2ïQ_Q8S ƒŸ™H“-F/…li„H$ÔÍË%„$U9~‰‘@ôŒR¥ŽLž¼h^o#Ç/É1g·´…(HîÛ¦I38Í ˜æÖ …˜n„9“²ÆÎ’9ÓÌø,.k;žò¨ñ°T¦\¥Y’°ð2¦Wéˆé=*eüIŒÞU t>Üÿ{+çg1%åNßÂËùY”B„“–ó³¥8îKE‡R©òèr¸Ù¦iÓ0Û;"÷/ nPep¶™ÚœÑº/ÖD›Š!5 vF6‚×32ú}“Éfz¬MÂ09Sˆñ†<,lÐüT­Œ;årS­8ÌŽ‚‚ѧC@%‡ [6}­Mcá+VVÚϘ:˜iuëF\7zõËR¸Ý¸C/%ß Äqw0fmN h#›e@S t±ÒáŽg½Sh 4ÿ6³YùY(kÕpdz$/ä°øvÍÜcèlq™‚)Tøù)Go‘Š.€ú.‹sÙp T,ô¢»ð¨rQ €„[P@£ûª¶gÚ?ršÁdz.œ ÐxhÒU-–98š‚jÑ} 0G«ZjvªcÚ*Aù³õ"Dï²TGü‚wZ=ò±=¾p1y‚üp‚TD£$nᩜAÅW†£¤^ÎÅÙu ~†Î®ªÌ|ηëR™Ô—¬ÇŠ•°Zü&˜Êq—z” ãVLg ÕjŒ™\Xmtàô•˜m.Î9¥¢#œ=%ˆh®ôo^PDgæÈÒpÂډѤêR  ÌÎ댖ٴÎJELŸÃn0Ó´§2$|ÿØ[%Àݦ°¾V@é ƒOØc+ÝU‘×a¡­tÍAÛž™““]5|·ÕA+L»§@'Ô猘 …dÌ̧7POQQ °$ÎDÉ@à¨â}Îýºž‰à 6û-Á+î8µÈs‹7ZNÒÃ+.Õ<¯™Œô†Ë½†¨±6ùFøê©:• aŸ¯j&s‚ž9S=E¶°ûŸþã¿QŠS6ªoÍðÂSØû¸ýêeX¬wÆË«ž“‚ žÚö@ð jÖ…7d^„7Ԗ߉£—È=XÐÂsiZ“53¨Á ²@Pa(C w=#g&5\î‚2©az ²¡™D¯tí/Þñ†;Žiî;ýD#mb@YrÚйòÙNvwX#êâÊcÀ8™s¯,‡HÌP«%[Œv2ç¾rf޵6Sñ”¤æ¾$ø•HÍA;X†cä 7 ˜a¼Ãá… C‚¹¶òn@DÌ­?$Dw+yqQûy\*šâ Ò©[½hZ×r‡ÈyòÈ×è"í5#¸ØÀ¤þ©¢¤’¨¥ ®.A8óâ¨zÝäâ¶JPwΌФÏç”±TÕŽ]• PÛŸ@£ÕV^‰!!A¦\9 ›"\zJˆMžTL.ŠÎ‚¥Ùò4hœ[ÌAM4ÿ“L SD›œÆ‘nL!rÐ dÔ–K”ãš9órÝ N«/ÕÉÌ­!OY[n8¯žüY È®N’œíslñpGž¥Ú‘Ãæ øîœTÇ—©‡Æ,[¬`vŠ À%VµâræÍ7ÎÙzp’}).}¶’óÚ‹òá@JÎñQ˜Š'¶r± †ôÈu>´—àÃ÷Ìëõ¶¼äl?u|-7Z¼(ÑŠ]çbá‚Lñ“ϰ’)®É› s‚M¾óWä¿çpå"+8èU^ŒAæºbS™bv;ËL€²Yó~ŠC­Jôþ [8æ†yY'èøs_ßfà¹,¯y±÷Ä«iVÔÜÑÞ¢š`¾-<…â@åtÂU‚Êéââ|¥ÉŸ(\vwǼ`© Ä4µ5ôÃ{È(h‡·¥´ÐÝtIAr•S+n!ØP¿"ƒ¡š;z¹uåii`ˆЋ”Ph¹¤{\#O;%É«mòi]vQÌ¢ömÑÇ8$0=—'»d2êª×EÅ\ $¦‘¨Â)·Á2g¿9ø%é²÷R [²ž{AêÓk>T' 5ƒê<© ÒžI!2òr*jëeÂD x7WsmØE¢§êfIKÀ!zÊCP"¦m#cÑfÚrÒœŸŠ«m%Àª,,G7 ·4fÙbÇ™‹œ›Ž@¼¬ÝùÉçðhhÓ¨/„dW"íÌRmÚf¢Ï°¹˜QïMY]†°M³(y:Õa~Öï+‘’«þ(Çcç “dË9ðçaï ÁÈuϽ¨Ó9ðÒŽCè5È£ÌZÁ•ˆÔƒüSΑÓM>NhõL‡Òù2ê™—Å)‰@žirZì´ üõ\•ÅTíÚà÷öL½7 N£ŒèÞ€8ÅÍNlü êÃ_êÐX‘Š@¨OQŠ¡(Ízö‹Úvª=§*ÁZD4º<4ÀAæžî™‹¥ÈÏKŸ¡9?;6ƪ„²=‘ ¡lÖEùVàæ†[Í]‰CpÈxGîÒá´¶Ë `J®´µyõ4Þ'7;H(‘O"—&¤±ÝoªÀ FWMw@è=ŠØÀD‘Ùô1‚li56Bñ‚«g;@#[“ˆT&Uèu:Ê^[¬ .ûžÛ°]HMm/Sɾçæ- võ®â%óàZ“ST¤â¯Õô‚"5a:¶‘›î%¼Ò¢EXì=€n~ùo´ŤáØÀÆÍ•å2½_¹¼»_É÷­¹±È’ðß»û]–Cµ!”ÐxÙº–:8› ä¶|`ç¾?Í €ÐR æµ¸o†À®mþ'™>öö:çÁ$€yˆCÊÀ6˜JpQ9+÷÷É»w‡•Äøeä†H p°:öz‚¬‰Î0­M>ÙˆÚÒ¯MTrÄÔ‹°±ØIÏ€%ØIO“yß©IØ ow8f¨Òp¶q»j$9$œ7®¤”Àef{e8xyå>JQà–´WHä¶—9MmËÓãÊû!Å%FL ¡x'wŠ®~w˜#“n& Øàý웹Så("¹‰:•H3·¾mûä`÷LÚ­ðÏÁ/³oÚŽlWÒvH÷|X¿Bß™¤íØml%m§Ó8V¢6!°U\&Î͉Æ3—™x}Ëxfæ›`jPYž3¥ Ž+©ŒCö&6Ó  ÊrDG Gž<*íƒ %Tžb’u¾ŒU5^DA'ic0YZÅmçx3 ¦Œ¥®ä(ŠF™­¸¨8w6”Çkå…·ìž–Ëã @v þ&¾wš‹]JÄ#yÄU°ÃJñil™2¿ßüÑ‚‚Ê )Ò¼Õ*ç5õ ×’;[2/ÙÎd]¤»ÈBÈ–œé¸E~w'é¸U~wž\r„¨ÒÇ´OÎuêeV—ß]6‹ še ü4øXkÊîÎŽþ—mÊÒl/ }ÍEöcf8o¹Ýuþ¥¦ÃN€A‚† ݰÒ-ˆÒMÈHäãtsgâv* v¢ïkO²‘ìåM‹ž@šXäL1Ÿµ£JÆÛb3‚F)vý\l"ûΔYw–OÐÓáîÔíß.Έ0 Ü| ƒ±71Óˆ2˜ñè·ûtïzhèPrC0øôSˆÚ^†•Ó)³Á¡Œ¥gÐûQõº—.Æ—ÃÜó¤(F§Ý«|—íÄzó{?ê]»×"B’@KÐú)ZZ¦ÆÇæaÝUô¶ÓñëÀWµqªöƧü H´Š‚ù}9âöpêܬ–—½co{î²ñ‘ç—0AúZW"•Ä0AF yLžq$‡“`vùpsІ,~ÈDêp¶ĵ•§aáÌOyí&FÜÈÓÀé¬Ç•à“tÒO’þÂ@ä3|( >ž|‰~i)th(iZßÐJ-˜ò1ز><.õÒíªvê”Âí¿Ë$éPfK\ï1Q6NôLyÑW…¯{Øüt§€€Û-šäЪ1˜”Wž¬ÝšyÔÑ1S6T›´¹`lª0Òâˆl57'ºó“€€õ\3€~1?3kšæÑà”úZ[ÌÏm3­9TòÒÚd3tÏt†Ö´Ao½ò4M Øž§©$ÉÊT¾6ÝS<  þ¶´íÚ‹Ôßæ*‹lðsgž¥ž‡wÄRy,Át‡© 9mØ‚IÝÒíx“Sìaß?h'ÝÉ%ÊEZ²nä•Î9“°~½„Ìên£´X³i¤Dû¦˜Õ3`˜YíûbV/ÿÙÅÅöýÃÎ$èÚªû›"Ü tûþ¹SÚ½ „tö ‰ßä&%}|È syûíF¢íèyfŽñ‚ª.§7l€Íî7Üû>Ò‡uÓ]½û°Òg(Xò~/\B &½‹f—=Ï·48# ä¸ïÖþ2@¦ÿíåxDõ¿ 䎨þ·}ÕúÌú%W·P¼È…òà6Ci—»9ÜY½Fy݉öÛ „]ÞèÚ2 ûÇÑý/¤Aˆ+jº¼]”ƒT7ò+ †O<%éiò&mÖîé§X ¨MfZ"jþÓ¬:H…P²<XûxK‚f“‡ˆÍ;9à E‹Í;¹ÞWÖºÅ-»PÆØ¼“ è]ÆÅ‹+}µ‡yçÂýaÞ9$ÓñçB}HyüÁ :˾fû1u(ô?6öãÎdh„d|Ñv¼¯mû΋©¤}m÷4ñÀ¸Ã“ž°t–ê;ͯÍ>açˆéâfj 6…åt4ÄPêò¯%Vf@èMUþ'Wm¾vš,[÷N~Jï.íô@Û¥w%ÿªÐÙº—3}{ˆ…â͘=aiñÐïÜ–9¨Ñäò'ŸÞ@lrËiH¿c.’~úé~ .)wUÊKŠïÁ7ºdO;›ÜbÝ!}¢yt»i±:éÔïø€)!ML¹ à{œã-[æÐªžKúÄê¶}J@YšKžL¡Ñ0r^Ô'v{äÏ)}bº<Î)}bÖS¹V;€-ä’ÿ…»Qd)Ήy;˜,”ô\ÖO7jSBVó"€@.oî\¾ç¤Pt$' s0†6™€ìD  ¯= â9ÐI•*X3Ô‘ñXª‚¶õJ„2Õ+]Û—dªéþ?–dªËõi\”þÖôÖ‚ä¸|ÍòKÉ.J‹K¿ÜÆFËÚL“”Ó&çÀâéÖÝÊÃ>?iã[miüÍ}&˜eü=Б†Ä¹$P*õáò- ¥tO„·¼=ní$¢OÚå­a öЮtÚŽTXÒ{x”7zo~Ñ.á{O+ð.á{O+ð.á»I\@º´ç¬(2jt½S°®‚ÊaDíê%;ù€\úí•C ß @)ÝÓ\¾i)¨wrÄ0åmh/È1€1ŠKr“ðýNЪ} ÛÜUFš…+#}€ øRrOöÃì:ÆpÇá2< °Q[õ«M*w[dJHØ#h ‘qGp/ˆêP“ƒÇ8.ZŒž¢Rܤi·Ãо =Xüê‘™EuË^Ì,2ÅG¹+á'¡¢ÎI Õ*îˆÜ.H}94ãfÌ,Ôso†ŽŠQëjöéP¸-\2tÇ2ö°‘¸IÌíwQeì1szXeì1BdÆúä•ãÞ@ü.dìa;¡e Ç€LEÔ–Ùz\]CÑ¢&ñêg„`Ü0+ÑWÐÖb šT¢±¹S¸î D ›ìãÂ'EsÌBFæ £8Fæhs4 Yj"´RÙ™eTéip/)p©ˆz¯"ÿŽ´‚Ñj c…oU'ß®~‡éŒrÎéµÌDÊ›UMäW ¹Ù\ºÔ”oŽ…ÓÐ7Ç1¤‡±Ìá’Su;=}s/xäͱìv´kˆ®]é›#+¢¾eˆ) ìöù‰Ä0¼âuGšvÈËÂÁÁ Þ܆ÄC àBÛ²˜H8C©ÚÕb›;š(to“A?œ€4µ!D‘pr® £$ß ÆÑŹîJ¾æ}Ã)|Ý=3ŽbÈ@ºä´½ytGÈNV©ˆ\;4cŽS@öng;xqaÍ }ÊE— Ç+8b/EÁÊÚ.\®üº8+Dyž‹3îÛ+Ñz;„H¢ ×E›ÞpÝ"ó㈷“Æ\Ä«½" pP¢­îÂíKE‘«¾@¸álÄp S4ç%§¬3}wí!ÑÝ]jÒ,²‡%^ p 3EîÑ}ˆÞðE[ëpCS§XY(,VÌ­0USýžjÎq¸îÜzö×<´ô…W›BeWŠ*óÄlê¶w‹M˜–ꇚ ëѽÐß |ã8YÝg œ"Ñް¥;!uÒŠ µG›:·i]ÐŽKØ|(`t`‰áᆇäÒMDwÂöù–Ç}³ÅÑ–/mØÓˆ¶ù%„;A3ÂpüΗ¼½›ô?[®j0ˆÐ¦¼•=ŠÙkÈÇYÈ[–iÓ6@ŠN3ubÕ°´RÞ1H'ptš¢+y,´™CFç>ÔÅÑmÄ6Δl‰ocç-Yñ´®}_“€^YqÉâ5ý¡·$×°,™FèÂmÚØ–¶f(| ÛÓxz_E—j¦È&+|*9""ÂÃT§–ŽÖ{ÚðÛL·=mø­,®2Ù]ö®½§ ¿í•½g×µ¬%ÛœbácÉwkvËGö$ 4|Ð)üôåÞCþø½»eÌï…» ïg,V×n¿½-g«@ØËîÁj•à{¨Z=eŽt“÷Pµê¶ÛC'Q $W¨@(vá8¯À'’-ÓíO¾ûV1¹Ù}w[‘RßÙ}÷áÚЧ ä S&w—aþ‘aþRtqÕ*Šßªn¤1LâÐ.ð‰@¶ô†\²…"µ INÉ猒gO‹YÌëýÀëƒð³LŠÜ4“’Oz­7‰¡V¾Âw`’ô@м¹òóõúŽl²i–9›Q Z}CèéEä¸4•Ad_.p¥}Ž=¸š<–ÚR]¬ÁþYÙ¬;¼!P©Ã5´«•­äØzðl%é]¤O×ß®cÔnFðl€(¥'¤=–›—ª#”ô3§ž:[Væs6MlÕpm|©lY¹®D'd€¨ã¤Ã ·µEnb/ZßCD-Ã7\Ù®w£¦ÆŽ»>x¶¬—Ël vsÅ ˆÚÑkëѳeoD­&—³€¨Õ¤;µ‘ +@Ô".lž5­(¨ƒ•ˆÚ·5ôèÙ2Ñÿȼ{÷FÄð"KmÙÕ8 ÷<]âÂp¨|Å"Ò"Ò/Hö!¢ª*+Ç5«ñ0 rýv–c$b¸biaRá,£®EäÒy–j¬+b8‡ç\þ~û¨ š2¿©~¢¯uо–ÕÜ1–äײˆ3 üZÛC1R›V•€ }­ýëòkíì^$F®ƒ]Œ<ÛI³¬i·IDo™~Î@´útÌJD“èòy‹µ1­(•ÊFfçj=¨ ýª³ríâ°ÂÖtÛO4æ%¸åí´“˜»ôNdèG‹·¼­Éˆ9Pʼnmw}ðۢܘEñ~s=–!–@4õªºÝ\4©ªÛFþã4ùªC·Û=ùª“ÔÓm¥Š¦“@†®½ø†·Wu)|n±Îɶêp6 D“­zØlóòI€¦Ù0]h±¢ªùoÌ÷."SSfL¼°!Ðä¼Åʬž›êd #'ù{ÑÒK»È•ݹöÒëÙv(Ãi"°n˹c¢ëWç±PÑb%YÉX |7½±aØŽ"jñª^ïütÇjóXUAI;vÿŠ¥˜I@˺Þ™@ôL\Ñiç8Ò"V}Èn^æëDgE®ADºðDOP¹ØÊâ%±Æq$yÅâÿæÃG!4·.Fc¿‡š†ƒ)*pô# ÛÑz¿uÇëÞ‹î_±YHA Gâ£Ü XÒyÛàâ‰S¤Îï¡{:–ÊVÍ_½€Ö Ù×@…uïÃÌë¡Ü:äZSîE€N Å V Ç:2ýˆž[}v,‹­¨Ñ¢U¬Í ñb–+Hª½â£q¼$¼Ó芛ʼh¨år_-–å|{Ù’z’¦ù¥å$—÷o¢wüJ³M UÚ6ÿXÛì9u¬m]ÓØæÜ³c l³™•w6ƒ Åѳ%«œn}Ï¡®%aªHd×.s‹(A”{Ý cãï}oàâ¼É RhXê]‡ê$4±¢ŸQít+i‡tT@Ä„¶»'!º/ úï´B"GÏ­ï*DÄ\¥§D41«_ûH2ó !Mέù*‰³É´aåƒöóáUÒ“ /@‘¥2<1ÕÖCåŠlq&Þ¬ŠŠE½C@œŠpÚù¹¸(•&t)2¹)Ëòu¾íVøî,qYª³‘” ‹aGR*šôCc˜E¦H±ŸuÐK>@´/>Eã|Iúo#ï0XY"ï4:¨b ƒi`*ðA\xÐŒŒvo\¾ Ä,fà”Y¤'MÀÕ©Bô¾»ªñl~áe6\HšP¡¦RÈ–mƒùQR…»KÀ'3ˆ¤0•›/ãœNDÝU«ç2K£_é^¡IЧ/ß73ç1WE”A|]„¼ðÛIƒ 0xkT3£†êøÕtQ)¿Ô92ö)z4Ù• _ÊÒÚk&!­Ë¯$)ÅâÇAU<É1o¿’c„~%Ë\Ëm¯m‚ÒXò+9æã*/†*BìSTŠA ÖsÚ!FDOÛÃ4?ä,,S«»ÞLûÃd\ü´?ÌpˆõTÓàü-és™p3ås™{ÆÂÔÉ€[Ý¢z‰*v:‹…l¥†F¯eSŒDBÅ>¼ƒ›n<”Vnºñd´rÞs×?Iz÷0mþ*Ô·ïé—'ή4Bxñ•,ê&?¥å¯vu6^{»\]?æ’†\N§hRs[7ÞŒÂxª…—dNVuÂ.­–ÜàI]ʵøvÎ0%âM;TÑs.Kƒ.ÒmÊM;¾ nÚñRKt¶«Å,¿pÙjIo‡qY)—^åV2.ÈáèõF£Ó/½+Á*ì&èðtw’¢õ×2Lea†-™ ú „u™ úT„áYI%B”áz³Œþ\o–ñZô~“³Z¿Ó%cP²Mä9)0ªÓh¨Ûh4zSïÊz@{8©S°»žrT‘"êM¥= 'º°Àø-v¼¥+Æ%î|ZÚÙ„ žÏ˽ 0Ê+)¹¹êí¥áü[ŠÁ6›Âpä“’• >~MŠÊŠÂ믤ˆ~žÎ(‚^ Fy¶ /¸ ƒ t*|{Š©`§ØV>T[ágèGغ?“ÿ% D?‘™bê+söŠÖWÊ”¶7õ§”ÒUYaïþ¸‡þaöxÝB:DZ8'sY»H¦ 2œ&kª4áÁÙ· ¨$¢ìÔiâ÷c)'Ûázóy©9€Y¦ú¬#§Xjùãè)åd*Eu2Ø(ÊžçNë+)¯ÉÖ=Ø- ÿOË+—L8oy%™<õææ9Ú|ìƒGö…gË05yûlh`tª‚ ëðBe'«–ªü˜À u¨ª‹Ã©íþMê´,¸^Ny–AU¡Ž~Å ¹ê¸: °ý¥%»”­KæCm\ik72 œmïÒŒuD#‡wy< pM¶V“×ÌwA†g˜ï°N¼´ˆ ï"¹­°yÄ¥ŽÿDz) dÈVÁ+‚ò¸”š6¢1…¤9ª@ðy—å]hpqEchpñjL@2pžK#šYÞ¾É`kzé¨TèÃ|ÙßîÐ%tüvGVÐÅߎ*~8?SH—94‡ÌÀ ºs.Ó ZÓ#{”‰-) hoöôW$ÃΚmG (é[ã(YÂ[7\é ^Ú¾af«3H W:r{ѯ ÜȶÞÓýBkg+pFÞ±úÚãqRô=ô'Ÿ~ÇT¼Ài\︎ô'W!¨ƒÒ'ëRásî[¦…Tx¡ûƒVvém€ ¾45û@{o"(áÕ^ôX 9Då–V®0xßÃȤüšV}8Ç9Mîòº84:p W5«$6„Ký¨Fàon¤§Ô¦‘H ˆ\©²r*mô‹®¾™P•¤];~ ƒŽý{@F]Y¼”iQ÷§<U+¥C£ø[)Œ¢ÕüVŠžhw}PЄi[@àgà„ øXmºý­J‘0+jĦÍÛî@ØÕ¦Ì³Š€tå6äi˜!Š›tmaB„2#ˆæJpØ#"†›¤ôéçnJŒè6E1R/Ôt5’ž#£@†D?ùØ•q9£øOH8JͨùÌŒ”Êz&¤ç3+dbxô/Ged†ø™;1. bü†Ü ±ò™¹cçccTÙ†‡û¦ˆ«ân¦eD ÅhZRx#Ý¢®L{:‰››d†>€™OÍæH{Yý_sÈYúΑy çñÔ@®â~*‘,öMIWs?•H<7‘,ö‰Lw¬‰xé‰ü+±;7³äo¿¢V®~h›6û?|5IˆF¾ON›\qׯ~ŽÜ¿ºÏýpPñ™ž ˜àR£M&Wý‡‚–ôoéÚybć\bç߇$âC4‡$âCòå!‰øÃ”Ç!‰ä!2§|bD‡äd<¹RÉvy’ˆ/Ò#x˜\]Jå6^Ý zç>$~Íx—yH">$œñ.óDò­wybćo¤¼ÝK">$¶ Ï~’ˆÁê~ÞËø‰â]æ!‰ä![ýþ}ˆ‘¼]U—¬Y²ØÇQìW[l`¾Æé?z…­á!ïïû(m#Üå¾Âã x y?ŠÜçQ[&¦£î8-õF>ÓQzàÇQBÞŽŠê¼ßn^ÈÛQ“# ÇQÓc¢çQRÎçQ[BU¼ùƆªä›gj›‘÷£öýæuÔ~¼yu=±Õúã(#oGÅ<Óo•G y;*Fåy_FÞŽ ?ó¼/#ïG…‰ËÛA›i c¢VïöêWÿUå!è?}vyMõÃW÷ñ?}+{¯–õ|üþù*—¿û‡¯¿yMwJéç㟾Içk˜W?þîëoÂS¤Ÿëcû÷_j±]ö뱞tzûø~7ÏyMDõÃ×Ôvüû¯cË}¬ññ§ïq‘}ññ·qºèrÚëŸqÀªyëôŽ“Åãèÿˆ^Àóñûûˆo¿±óuÿÌ[[¯Óþýã~úñëԉߕW«|>þ>/÷»ûÿåWÿg¼¸×€ìñâ"æ­ÄëûÕo_/ë?žk|üù¯¿Îý‡ü×oó_?ä¿þÿúóþuÿןò_?>ŽÓͽ}Õà:¾ÆUº·û~Ë']»Íß~úú›ø¬=?þߟ×Dl~ùð‹íW+qñSÇ:ø¥÷ö*ãýg, ëuÃßDmz_ÔÅ·óê F‹ï+š2^WˆÎµ¿Ýæã«ÿÄ+—:~V*ê«ûø¿Ä)‚¶³xﱡxî=–å×ãz‹Žˆs¼æN¯Rñݸ¹×4®¾Ð×»„KãÇwÿyüðÛ¿Žý¼×˾‹ïï[úç8xý*Öñ¾i×~íÕ¬ü>ªµ-¾…U^ý`ýøÓ¯»…îácýĺQvýXþêÇf_ï÷.Û¿ÓÏôÝ–ëÚâÃÍuÅ݇󃄡–ßçÑåÕ³¿%O¾>þúq’oyø~M'_çϼ׾ŸàO_¯ðu,?ý…'~]ñu„kï?Æ ~Mûûëº1k‹ÙÕ[YyÖù?Þ_I/i½^Ø·oú[Â׿Txq¿ñóz¸¿â³&<ŽøþqD¾ ýñ¿áæ®×4ìã£Ýü–/üQ±^¥åBaŠmœµŸ÷öxƒßòàkáCEyŠÿò$ªž?[Åßß­â·¿ågz½¢Ã·xbŸþù.~ø:JÝ«`°]DÜW¿åzÖ3Ýg;Ïû||ž×ûsÀk÷>ßêä÷÷‹ùgöíã§»¾¡®„ Ük,ñá›W¼>äà× f_™¸þ)ó³â·r^ÝáÇÇè,ýÑN}áõ>ýïÂö,ð¿½Oý¥¶6Ü^³»òV—_ýÙ³Ãûñ®aÏ3ÿB3÷ã[1ŽwRjm_ê„@”xÂtñß [xU uñ¯¿ä¿¾Íý:ÿõÇìH¾ÿB7ó—ü×ÿóñþç×_ìr"õì~ -pDÔ=GÝzÖÚgÏ’ðx¿ýúK½Ð«€##öU³¾}ý3â2^÷Õð¸ðê¿ã!ãu‘ïžgùñî~¡‹ûö{^(Zþg£ò Ïòío¢Ü\Ñ8ÿž?|•¾gÝøé¾Ù/?î³æþãæ~ÿè‚þp÷Ñÿý³¦IG<ÊÍw÷Ÿý•ŠÖuÞ^ñó?+’¨ñÏ"ù¼)ÔÓØè¯Aã{§ö­:ØRæ…Vàªcìúüæq/ví×?¿½kõ=îüéÇGuz:ŒwÆóþHìuij]þéó6}åÛÏü¾ûy߇žòYaÿ9.;>×ótß>ûA´l'²Ø>+ºñÖ#ê W^OEt–óV~¸Ç1™”5Î[óê p#Ѽ—³´£öD]ÌóŽMð£ý…BùÍ}Kqp{5c=þ¯~š§ôG}yÞ<[í×Í­+¿dÊ¥)ìlG6h[ÿòXòîºü›Çí²˜×«@ÿö.¿Î~ë³aÙçÍj0ZúïÍê{Óø?¯áÁ¬æDâ³ÑŠïý1¦|ŽŽÿr^•»²¨”¹Ç~V€G§ËN5(¬¯vú½²~Ç=Ç«~5xùzéîªø›ŸUì×PtE÷ýú ¯6s|üJ ËÈï?kmY“¾ôÖáßøšìëEüñ ×ù¯ß忾ÿâ[,‰qÝoõñ€ß={ftu—Ÿ}c\_¢šÓaô‹ôãuýñ~Ö/W©Ï:³8󈮬ƒªÝ>=ð¬?~¹Œ³̾><à´¯Óóþ¬Dè¿õÁŸwW¾¡_h;Øj¾ uý…žúO{d5'âñjOq•Óþýè}Ãâ®óy7ê7ñŸ¾8\ûù"Xã>_1ú?~õÕß|…µ·ø ¯™è‡?}÷o\•ëp&ë@Ä(#Öå>¶¯õÿùÜÿöó…5å|žïï8ïzï ÷ã·ßÄá«•yÚìÃBû¿¿JãXCý㫲ý”ÿúKþëÛüׯó_|´ÞÿÒ”E½güó?ç¿þcþë¯ó_å»éÿ†~#ò¿Ýÿü| õÞ››ÿúuþëÿÍýýFÿÿúÓ¿xÜ=ЏWaÀ”¿E¤ç÷x[Âcÿ¶Öò\ø¬˜„~{àóUåþ)êÖ«Áã$),&^ýß?EÛŒžðm%?:ï#ÞŸø«òšr}yBóhbþ‰Ç¶:þ…媸ƫÙýÒr•º‚·ÉŠF’¿Ø¾MV°ñÚ³£f·öºø«Å}´øÈ?{´è_îËßú¡çêO=ÿÌ'ïûÕ~ÿ½_~ÅÔU/ï­{}6Ê–ö,ês.óYŽnèñÿñk¤M¾žê¹Tô¸ççº{‹FvÞÝ›>ÅŽº—c˸¹«¬5b•æn?Ýßþ¯Þ'…Ÿíà­zrÿG×Ãÿòå±Ý/ÞçŠn€Et}|[¤ÊÒþ¾¢õX³ûâú9~Ž*Ø}·|>×}¯a‚µè9þÝãßZe*Õ7³Çy[øÅð%Œ¼^ïø¹eñ-Æõ'îùí’ÏçcÁxMÄ´lŸ·uUœ=dz|¾¼ªó±ËÞ±ÚÿóyyÜàsXkŒ×?ÿ­«yC£ÿûsÜø˜2~>ØÕc½_'Ы7_©4Ÿ)”¹?ò/µÛ¿þBý»/”ß»ç‹ÿõ ÿõîþõ¹a¸ôÝï)Ñ¿6Ÿþk á¸p~iöå÷ôÇÇpø±0ÀaZèß–ƒüÃÃAã[qÇ@ûÑj=W}s·Z³ý¬r=OòÛ÷*‰‘Pd¿F5HíŽëѯQŒ‘ØýÅ‘Ò.‚Ó±ÞFJa2ÑgûbîÖHiaìýªÅ×ðjòúaq¥b&”½o‚ƒÕÜ+ÿJ…x½få¿ÇTsÍ©-‚‘j6ýÏï…glU•ªòýº©r‡€þâÄ*LèZÔ¼ù¼¦ñò¼Ìr^ÓëÕF-[åuME^שeóSé=Èn s‘Ýö½þqŸþ'<ÄlSePðß=î¼Gg±ÇõšÎüûÇ!ãqȹñ/½ÈhEJ®Ê¿Æ%á‹ÉuýýjA[gå§Ì·99ä×ûŽAÊþ)–]ûü髽¹÷ø8} …'æÅ (&gè¢'/ZËë ^oŸ*öé”Õ¸£·üû·sßø§ùªÅž/ŽÙa³ò3ëˆÇ—ýþq¾¿èÆËõyñÈbuüü «vvEëÕÌ¿:û?óå¼=ÍO:ói.>|ƒQª¯ójÈ1Iݯ/ÿjþÉod¢…ˆß½Æ¢ß=îãù.qíõú²óíVŸW¾WÊE%âùA~Äœ^5€vh( sUñ›{í–ï ÄO|.áz‘õõ¿F9jË+º¯wZW|êi®¹]@5Nz½wø:òGž«{ )W¡¯ñ;|ÏWÛ­=ÑgaÕñϲŠYûy{ezp¾:^Öp {µbß„„r¶®E±·_b°«ð‡„¯T_¾ôövW1Èxu KyÄwo§»+þžu•¥ùUÞg¶|‹AÀm1†Ã3¬8÷_x#­õ·îÙzáVN¹›É>Û×û©ß.×¹ÖuîwT®ñ~ȳµ}èo"Z§Ö_¬ØïeúÆ?ý¼¹}ÏFLÅéâö÷^è:>Ö•óü¿ùêÿúqÍ"endstream endobj 428 0 obj << /Filter /FlateDecode /Length 3613 >> stream xœÕ[Y·ò8ÈèÇž@ÓæÕ<Œ$€lÙ ËŽ¥Mü ûa´—ÖÙ‘G++ òãóÉ&«g8#‚Àˆ·L‹d_±'?ub òßó»ÅGOƒï®_/~ZH'ÇA»Îx)ë:%üôgwÙ}×mzpJzß½]ÈîKüïÇ…è¾X(áä u7Jg©º»…¶T¹]'êªÑÇ0ĉ¼öaìVr ÃhâÜóþÑå«õn¹ÒaκþþîrsíH_Ñ èô#©Ö`W=¤ÿj{Ý}sqñº{º¾¿Ùvß÷/înî»ÿt_\n..wß/iCÒFhxœnÙY¯eò¸[Ü^,qŠ%÷¼#ɯ»«¥¤Aoûí®K»¬7yŸîEœµ!ôï:Ü侫.ø-lT¼"ÇŒtBÀË;8¼¥D“SÊóNR¸w VÓý€“^M5ûN¤()gNôÑS…m#a'ÂM„Ÿˆ /ˆxžwžÎóvAº‘.µÓø·A˜?eÙná˜,EYOä¢S:²Ê;ºêmI=VŽ~ |8e«µIÇž2 t “<ýð»ëÖåó åe_4Îʃtãèä>8’ñUw¶[ô·ä1Ñp¿V¢¶T+,¯>„@ë¡ÒœÄÂ`GR¼ä‘ 7’å$\ôC´ð£1úÏ`”oIÄ„Ñð„·ÕûT<¯¢Þ™A˜y­=äÊ#3®¼òˆ¬\€‘†(²º{<ïx’g`<Ó²¶œ)p$‚P¬²#8ô`Âû¸òÈŒ+¯lË*˜ àÆÚ3”`µAå°ïášF8×´òˆ¬)ò…ŸRØ42º’ÂNpå‘W^yDÖ„UR¬p ½àAx×4¹¦•Gd•äÍ“öñJ€%ßž†OJvTl”ÕöIqTàР¨¨R5û| | ÿµ¡¿‰¤vA÷×o–T¿C@êìwW ¤ï/Q~5¼¡7KÖ•–*,  Bé[®(!yôX§=p(H³BèпN¬ÚKÈ¥ÑàÔýúž6Ϋ‘Ÿc›ß¿*ÒnÙü}bXè¯ê&Û]vRõI° Êô/+Ç%‘:ÜÊ0TfW\^ µEäñ÷¥„ÖXNÔ'…zX¨‹BÝê¦P¯TÝjÃøòáföDÒ‘™/®ª™+q=Ð ]ðÞ[²Ò¨:HTÝ VSZdLõ­Rš1i&wÎA= =è§Yi‚A°þvC[@Éð.qxœ€s\¦aíu¶‘g|Á<+ ¨…œùçåˆÂ hد£çy@Ea™B¤A»Œ„Ù9˜ æB¯óhô @ç¸GÞTlª*o«úú$N&ÍÜ›&fæLØŽt…öÄðó\s–¤#35[™Ng”µq÷¤Rjepo|ú'ñLHèeƒƒ¥÷üLGNÂ8ª:Ö<þ.“8@[oªnÖ¤^$$Tá~{½}ƒÿ2PyrMn .{"Û†àx]¼G÷¿þf‰ÔLú,©ôÀ˜ýÃû~ŠÐÃè%ÁqÉEGŠw4BˆÑæ;z9:ÎÁ}%™Á${ÊR©Yöo«N_2wÌ^„Á¥á ãôg3GøH¶­á„Kea5áqˆxRÇGPo0]ÎÚd0z²Z±ëyµ+÷è¹éÓΰæv:§XYY¶·ÕíßO¾#Їöys u±Ñ˸¹ˆÈz<³w¾Uñ’ùIw 8EÎO!>Ý ZÏ.óSŒeo!®t}[G)<½G°ôëëõ ‹ÛMöQ=P£”ô¼c»IÖBšÑ•òø6M#a›“ŠÔ†¼ªrD·²±øå„(–ùº`w°4lî¼q“y£nv5šs–·¼QñÇêšå¥µe«ºw«:¬ ☺@ÆPJ\€<ÙÅPÐÇ Žp”Šßˆ@rÒø¡ÔbxùÕ$×Pî+𨥒Øbî#1„ ’QPÑÆ™‹M c”“·Å³ÜT,“Kas8ÁÊÓ^}”ÔjzG@{Þ¤µ3b’¨ƒJ%¥FÇSP =( €•M—0ùÄ“B=.Ô}¡. õº1û}_ÉeÞ9\phœé9˜ÛÔ[pÕ_äp;;W#³b>‡ÐühÌX<÷ç˜õ…Ô³üsQóòg,˜þëGdƒeöp¡5ÑT©¦ÿ#ÚÑ>L›×Š>O¼ô6$O<£ðƒ¾©å˜ 9ª³û™ Óþ8𤻠ÄÃÌ£/Ó:Äî!þ‰É‘çµX‡3‚`~Ót*ÖOˆ›?N‚œ Ù'e't’.I(9‰Îœu[%Ù•”òZcØI‚u°¬”ž%¸í´™`§ Eÿ%ûmŒcçµgÛ‘O)9Â-_0tmúàYΰô‚ú 2„Ý2”xÑÆ°S~Qã”—±’¸S’r,XÖa!•Sv¶MJz£@,œÕ ¾ÓòlCºyF6sÁ\“«å*6¬xñe¡v ˆ\æªh¨b¹jšMÃHUÏV)R.òªþÝ”\áYã-Ùú(Üš=g|l^5rè®Nÿ–´[×¾õ]¬gü¥w¹b+È;­]Íá§Äð/á17 ä"7Sù‹?Kºˆ¿ü<4mÕz5mm)^5B²®x¹gÚcêa z¹ ŒX ðò¤:~ªº|ÄtTàòߘµð"iÁû>mìu ­¦k—±ò×M4vÊ•wlEqêÓ™¬•ÓªAöçqK Õ Uó’oä šÇjY分ÁÖL”lÌ*Æ×è¡j–âýí©ÌÅœæiCo«¶ïúä9®0þ–RS×¾·Bö´•âílѬihÖ4ìâØl¹¾>m!Õ°K•­³¾ZÈÐpþ{,ôiCŸ·ÿW:CUŸâwY¨Ú7œ´³¥­d¡ë†]6±"¸­ ù -ncUÛÕȚkf‚Cª­ò¦!ð縱¡×ªÃªõjGfŸpÚ><êeÍÙVÛ!N?ÄQaðÞ¥èËPÞ\麊þß0ß.þ ýŽ£Žendstream endobj 429 0 obj << /Filter /FlateDecode /Length 5546 >> stream xœÅ\[od·‘öQäiŸ7ýæîÀ}Ìû%X˜ØŽ;1[Ipv=º´”Ѩeµäùñ©"yÈ"OOÏx²ñ<˜ xÈ"ëöU±Øß/ØÀ ÿ¥ÿŸ½<ùèïÛý [lO¾?áᯋô¿³—‹_ŸÂ.øÂÞ³8½<‰Ÿò—r`^,ŒÓƒvbqúòä»å×»‡ÌÎbby±Z+$Z†^¦¸òË+hJg¹_np€<·ËW«5c^Ùå5öœ+]f3ðaì^(2Yí:…ãŽì…Þ+%—7ø™3š›8…uBÇ圴^.X Tp9÷ÝS¤Ø-÷äï/áï–IÉÕrsƒýÖ;ç |[¯—g0D;)ä,Ei°cry‰# cÒ§ 'D¿û>|çŒWËGzyÀmèõFqXºt_—õèè=tnµRË m8¼ú”3E‘Né PÏmŸóÈ÷¸+`Êe IYµÜݧƒÙ^¬þçôËáÝ`¤Á:=1 ä{ã}¢3ÍOC cÂAGNl¯qœ›u61s>Ò© Ðì—»(%,ˆFæÉå\èlç Á<¯È¸-Ýg¤;ŸŒ9´ò°[”æŒÄ­®%—ƒ¶v±æ|ðZó¸åÍ~ÂÍ`ÿ@0l‰Á9‹1c\â¤UƲìY%Ê»´kÏ^«µ¶(\nùÅn%ÌÊAç *9Ìœt@W<%'" ™i5°ÚÌçÅóû¢µ/R JªQÎV|³`@lrë~•>¬ “‡-øñ»YáÌ¢þ€„ÂF¼ldɇs¥ªLç¶«p›ñ;›×ibžô;ðƒ°æ–žòˆiçÒ ÀæQP6Á>éÊҌ 9µ(ׄÐ( °qkZi@ò4(bl @-yl2P˜s¢%—ãhŸ&¦¤KW«TÔ#bqYNnˆ"=Åùòv’ÆàÜ‚kÐ-Ñ)²âÏÈ2gd™[”\`Be¤†²`Ð3®åà”õÌG¹8%Ž&8¯´T’l1IdK„¼eÑ—]tWè"*×P˜æpÝInQëh1uy›ëbç©:##„¶­vTá„ ›ˆvñDÏŠ’£¤µ»-§´?¬º`Äa~>UÝ—¹u•[ÿ—[¹u‘[ûÎ_ÿº,ÍU×h8-ìHÀ†HÕyÞs×èx[—ýÏM^èynÝüRÞ$´Óà‘?Á=pŠk–¿Ç\+—_ÄŒ«‰ÕKŠVY½¦ ÆZ˜B´X Ln·~ÿ?ûßÜã¶ñÒ`ì¼,èáÔËßûs6ä­3=A÷ŒŒôü.oò:·^äÖEnÝtÆ]uøQZç…™ßdÈ5ù8b3ýCn±N«ì3øLaôM”œGvúxnéÜráÃÁ);Jë:ÇTöwPlo‘Béð ÇîÔºÞÏNYg§*VÁHðN˜‰?(T´¢1ó37Ã/z^ºë§â…äÈË/3?J™Sñu OøÐå[%EßerCáMÉ—ÖWmxT*ÆI$º)än.m?!1O~Š$à Òz¸ªŒ É™M¶K|^2Ç$ÛKš%cF#ñ*Š_—´õWÚR’›0¥ ^DÂqt(ܱ¹ví¯)žƒ€°0—È]Êf½nî[áX $î'±çF“!gxŽ:œ#@™Q»¢ÏNOþxK‚!vÇ×ý[×9¹…ÆâŸXü,fú=0\ÎiRMk•㎧áwXâ¶!†d¼‘ÀÔH{Ù4™ï>]–0±$Ef5ú*;'pi¶Ò<7+îzнšŸ¯Ž2)Mñˤà| hN Ž~-¡pQf*Ïí¿°ò|KæÀ¦Ó…[±ô1!¤,ÓXÊR¹]IÔKkcÆ,N2[wÎåñuç3åAŒ—PEpv×iu±ãL±ÄTA··å^‚‚CµÏXíãÛjì†Øéèb™ªÈ@T`Yަߩ«1˃ŸbJ6€áqcâ ”m(ôw¡Ê­©ŸÏ»"u3ÇÔѬ N)ÙOˆlì¢p”`qú¶øƒnˆÞ-D!´%3ö“J”êp0åëJ]Š#±ø±ÜÍ“Íã±½êU×)‹æf¾®Éö×Ršcvøš'µö.¼ ÝDÉóÎÐ Þ™XuÓ¾ÿ@EžÆšÈ+%—ÞÙÂ0Ø~в³™üÚ£7bøô÷dƛުt‡éà›RèR 0.É-õFõ…:3ÀôI뵫j{w7o%!Ù¨ÛîwsewÛ]?cÕ:˰”ŒÌ…¼ï! nn°Æ.æxUŸ¬rÔnÆj»q“(Ñ?ÖÁva£º¦ ¥{Åo·éªt9/éðµ£YˆpÛ‘B—œ`*a~×Ýx‘c³›¯)“«ë…›ÕøeMºCj .¾¢ï–yÁpŸñ†¢,ð5Æ›|£*üÓ ¬îCÒY^ª‚æÈj“òù÷Êó®,¶º~E6ê#fK¾’<àÜÌ¿–OË4†w-x6išÌeÒcEûÜ^eòò‚4s¼@tõž&0·ýÇ8 KEÖ|g´p¶KóäÑH°ƒû8+vé{ÿ§²—Ií;ü««†Úv\†"h¥J‰ÎuKz¸×}]%ßÏ#qꉨûKC¢p¬ð"Ï¢‰Ëä=û¶LþÙ*s¿1áK—Ц³-²ûЯA”x»Üí÷¢ö·zŽÍUÞçMk{ƒ×n¡cù‰‘±<ù<îÎpsð'A&÷üªÁ5Ö®SwÖ¦ª‘{:—R¢Ò‚|ƒ•”Ë¿õ¯eσܱQi}¨Zn”¶œ¼µ¨55kÅä¹PùO.w’þ†—30/w3sÜ@Ã…t!kå¼ÍÊ;óë!Í\KŒw.Â#‘ÒW¥÷¿rçºtþ¤0¬$éK&ûu'iV¤®¤Å~ÓwKµjª`ªyn`@5šœñ‚K Ü#¡w- Ö¿º”⪋$>+þ±WÏwß³çÖßsëÇNßXØ—wPÝÝÿ)ÓXê•^uιT=•kŽRåõÐùâ|þö[<·DnùÜ’ÇÖ™Ür>V¦ái\§¥]¹Ü2”’;Ù™Ðöîlþò¾Ž8\ÔôÎW³[6·Ô±»´•]’ºBypšòW~ìʽâMՙЮzÖ*WÓN¬U©ÆëÕùõj{U€•ÍkòDÊKXlñ ýç˜K÷ìУÒ‹—'€¤ 8ìÜssòí<îwl¡!¤°urZx| ƒ%xEBŠšÆÄ³ ˆÞ-ãñØv+Àˆá+5—¢èOc>‘3LŽ&±XpE¿_~çNO T¨ª‹ŠƒgÃÒ†§ßþIí˂鿦nô;΃Êñ²R°‹‹;`$OfÒàkºþ>­ÏUŒèÁËôÃi„,2M:'ȉ88E’%|AÝäíߑ޻x@øÐ%Œ0N2ˆùc œÆ`Ê8¨ÇLD›òË_”ôÛ¸9߆ִ0‹‰‰}áqJq &¸Íüöd!ˆµµy¡p»¨¼dØ#ËÁQˆòkH3§wí…ìJ¢`fE!ä‚fVVÇøHŽ‘.„¨Î1 :f!P,˜®÷¼ gÁ1Ïú²ðÿ‚LRSžÙ©öœóúÛtæLV³¸$lãæ^:É=iNÄüÓH.–N(Ð7!bèBE:Vÿ¾Jy:]MsO†\„ ‰Â¸ èrƒÌÐãò—>|ȉ?=rÎHÁ£¢Áx› sÂ,n+Å='_^Ó=Z®kÆT ž©I¿-#>JEñ1¨Žá«aPLmºa‡Qò·çÈfªZâ2t<Á{ü5»Å"I£…_¯%D"fÌ´ç<¾IËFî!Ë&½*­Ler^¯21I·ÃÓÿý4Ì4©îÊ9ÌíR°\w÷Ç“¯†Üendstream endobj 430 0 obj << /Filter /FlateDecode /Length 6611 >> stream xœµ\[od7rò¨ìH#ÀÝuÌû%ÉfaëØI6‰=³ñƒmI#É–ºeõhf”üù|EòÅÓ§¥ÖŽƒÅbè#²X,Ö½ŠýËB r!èåß³Û“O¿aq¹;‹Ë“_Ndúë¢üsv»øü5fH#qˆN¹Åë7'y©\83h»pÁ6¨ÅëÛ“ï—ÿ²’ƒQ¨åºŽîfFëÕiˆÂ/oÒ—`äò}ݯNµs~rù83.Åè~õ_¯ÿxFñS)‡h­\¼>Çöªlêè}]ÔÑu]Öm¯êèm]ÔÑùêÔ*5Ä —§õ£¨£aæ[Ù:’u¤fæé:ŠlT7•C©:ru¤ëÈ×Q¨£%fJ&ÆÅë=yý·ß/¿;‚\4ºª£äR"š!Ægˆfžü«™ù¦ž&’fGÞ'C˜!' " $3Xaô‡\ô:sßg`\)—ïV®à”Ñ  µ ËÝê+„°\-” q¹¾I3‚‹“1%!¥Zþ>[|Åpý–¦€Ò¶›±ÝÐnØ;nߌc¿\_ÞÓ,Â8œhrQ0p.,oi¬(Ú4ÅIðy³RŽ0‹Ë¼§‘&.?!Øfð í¸U?8°ežáƒ²`…´» ²ÍËÈ`½§&l‚VÚ”ñQ//¯òTʶ*ª‚®ŽÑ€ŒçìxÚ!‚¹dþ,÷gïPçVHoþÆ›åö¾lÒF¡c!¨°!,3¡]Ž©=aNÀŒ0ÝÅt@ˆ§´VþUš¸çŽþn#ð[ÜøC:Ï1nÔíï8í´~ÚCe5J põn¹½Í(iéø©?$ 1€ý–ë³t0-4ÑߣAš§ÚøAy_T­ñ]â3lq]y‹¸;oà-ðíÙ¬Ý'xÃc¬ép§86ëîk—¿j,»›=ò–î35#k´þ¼.âLž¡Ú>ÆþF€_ó7Ú1–¿š[_Ô@o ÇCw–+\W…ÔÌÊ}Õ$3£Û™o›:z;cþnêh;3ï‡e®f5Wƒ’QŽ(7†hü™eÌX£–gìÚ2…¬ö–{Ë$rXM*J/¿ä7Óøìò¡ÁÎ 6sœ;é?¼>ùæ$û)Pw>ÈÅýÎIw¾Ñ9!®… ,œŠð$"¹'°Ã¯a¿ž…5S‘ÃûŒJ²mÈpÕÈW¸3SƒƒçtÇÓL¢¯¸RÈÂ(ZYü Ê W2E´ˆKG⹫h©n®ZnÚ½l“ôÉ!@µsÍAª¾{!÷p-PVsrå˜êãÄ={5ã[|¨£‡ßñ¦Ž¾œÌ#D7³nåÀ¼é‚uÒkT‚ºƒU¯árüHPÛ–túz›ŒÀá³ê.32ÿjgM•NôŒ3J-×»<6puŸç|T8¢ñ8–I2ËÜæ2Õ}ÃDª€Ðг$ÑK(6°ÛkøªIßňš]~ž™Iâk²¯FFa`àΛds~O<Íæx¬3Ò1ºÌ*†è‹{‰Q‰U™LlYgFÓGÒÉCŠ¢Øè:ùŒNx`‡­1Ÿ“q©Lö»{iGFéÕ™$ày[lïÊ=á¶Ìô2¨YgJm²ó6ÊR¶±^“,òâŠC™Ü‡Õöæ)³£÷ÙÙÀ r6¸‹yÅ\LνÌ$Ÿ'D2Õ£»aMT'Ó…eè±àBóíASI×É1?2;±¾Ì§Î|OœC É­O,YØÒÀ\Ï,ù]ýM}]G»p7}h~ó$b¿ŸÙ¥D?U÷.¿y›-Uؼ͖ñÚÍä¹¹_~÷-.ë\f“}¸’R Yw|>³ÏffÇó:º®£÷3koطʾ¿É!~jßþ»Ž¾›¹éÆ:—ut5ݶ0NÁr½ÿhçä¯ M¥ûµƒRbš4™Ë…¶”pËlº'¿)ö×Së¦Tù\ÞÓÔQË{vêqäîãH7Ðr´ø¨Ñ“X‡å”v ñÞ#Œ üZ;7H·PRǶ¾ƒYÜ_,¾[lžÎÀp¤ü²< ,]mV “_ﻋ•‚HÃÂ.?,Þ÷¿åCSíÝJaRp)ï_>l¦'‰—h5¨'íq«ÓùöNâjaå,.Ôä“ÀÕ°w¹l;x\–ŒÞC¤Ét¢9úìM|u´yjÏ@3øÅLûúïAüf&îã7 m¿Yh/ãc ^±ç†/W&ælÄú¢¤Epzó¸ø÷7o/qh¤´=Bæ à(*¬6,þM¹ÍBò¡Ï„+&“ÇIpSg,£úó_M -q…ñQ &B8±@Bì½ÿiØ=7íÕ1Äq^Æ `+‡«ã†ÿ\yÒ‡LJ*RDæ–Oò‚QˆÄÑë3'xœÁ¬…O}€ø œ9ê‡ #ø˜¬öA¸¶P´]LYÒ(ýì¼WGÁGÂJL£ŽÐga´p`…µ”Òõ߆Üó_M íEÌ Këzføì–ÉæqÌ•Kwd)ð‚¥…:ry6…Äjaó œLÈÿv¼ ¥SÐ3pÿh‘V îiü·Òûˆ‰¯Žƒ&Ž…Fü aÆ[MóGS²gÉžŸ÷ê(XâHXùš¿Ý7*·'Àß R.L„×~qób=¢ "Ò""eòÝKE•%O)Ÿ¥ ð%þªÂRÒ8½¿ÔzÅjG ·“yÜögµ/Þ2¹aúžÍª_جQ µYõK›U9´áU¿°Y\9ömÖ˄ғ¥Ë‚N-Öv)#9‘8?Àœ§$×R ö-¤P´ÎƒþÒʨô ¯hÁíi>˜™e¥”ë—nV1™lVùÒѧ¨ü6küÒÍ*Ú€Í*_ÐzœÅ°ÿæ…C 7W’n wWŽÎ e´¼q]rëòaZƒªÙÕœ4³—ëU”#ÿ¬äp±ËIç4®¥ŒnW€™¤_AÅ8æc®~å¤ZJ¦G)M¥Áœ7´Óœ±¦ªˆ?œå›d½ûz¹qƒ75ëýj6ëó1¹ªÙz`ªmÕLqW¬=,ëIñT¨¡ˆ^Óí•|Ì×4I ¸Ë:',¡)£›a¨ŠtÕåXé>¢v*õ´!øô)›Ø'ÛK?Ï:grƒ}E2W80Ž{…­ríÔ cÄ“= ´7f?•R®ýFDZNH)姘‡f^#]o2"v¯Á¢„•¦ïÆ©]]n›áÆø¢jjá ͓ Œl-ï#lÙ?49»_Iïp"xmØ#f¦Û—:/ßt òq³TIÃ1¨ºËš‚®ùíŽÍ©&Í8€n¯Üª\î²éº¹ ©CžP ‚z€U‘êHYǃ• ÌœÖ:¨™€­ü$³µöÁgQ‚­†Ê›ÔEÒ IÜ¥R¥,Šê¼•ˆ¶Î ·ÏÙ© B¬í--Ýcˆ">ë\ÃJöu“ó´MvG}~äpzv²¶l——¹W4ñXàt`àÞ‘ô(!5Ÿšh^rHDƒõ>¯Rû͵‘$á(úJ »ÍËÔ³c„ÑÆä’yýòA0@HRÃ<­Ì`—_”î­è`4;”F©òù¾}¾¨R›-)xóä…ÑQ}q‹ÊºÍ9Ÿ¾9+G“vMa4ôÕfz ÀJ6”+ÊË“pieÚ D•ª¨2cWHb¥£Ÿ ¤0õXÔýo|¢.J"˜”ËÃSãì\hî¬ÁWsÆ{¡ÇÌâ$!®2xë#y)îÀÏ+k‰í©I?áiJÓ6¯è«‹Š;Î[æiУ\sÝeóBmŽ]ËGö„RwÎÊ’r½×x‘› ›µ=ÐPµË@ ÜÆÓä¥ßçϦ®kàn~kàz˜l^ôs3¸ åŒ3ÌD»Ž®q’µÙv¦‘uà8Ò}¿Á°ëdæà@4'kÑc³ïª²Ö·ð¤ÜE8Ð:5æ`ßð&s˜í›¹oY)¼+,ŸìËBì ëeÍÄÇ9c2¹-±¡·¾Ûä¼ó×¹ëÿó¹QÚ¤FéÞ_¡¯8o'•ÛÜè·‰‹{¿® À%ë~äßÏZœþ˜·™ÄÒðéëñßMƒz³µ ÞôAAvÁpyšílst"3^þ}¾’m?¦T›í kJVÅó—{ÍPìH÷мˆ³¬h¥Ö³p”ÿS±²! 7zŠ4`)(òñ"ÍäÏÈ¥‹NéªÚÆ%˜=8è‚üÔ˜ _#?éÐäðTkÍ¿aaS}ÙÃXpwˆÁf·{oŽïúçÄèšÅæÏKWahzDÎÓxŒÃýö4Û èûúx…‹81ĸi§ñ”!N%œ¥Úx<Ææ]¬Œ²³í”à9z¸Œ%Ußy3eôfÖ>R®FkqÚ-*¦O*_úQv}£8 ìYK>Wù>SNµ{,ÇUp‚MæoèÒg™Zšyönœ­úÀgVXL­©Mâ|ÆK0©)wœý;zéJŽ8ûð¹²BóӾõÃCSN§©g”†§§`M üÅžÒè±&ÔJ¼½””óŸîn¨KOU*ü=¼P(1ê¹Þ×0"à:Zô ù~)Sþ!PXïçcÑùgŒ¨»fB9Ù·ìÔ4Ó)aöÞ!&~ ËßT4Ø.ç4šj&Á&O"lçó}»|Vþ´9Iœdƒº®üñEsûlÕÏõ<Ãb2’]ªäÿP>ŽÝÓÊF¦#,¼%²b©åšúÐÓ¢k¦o >¹—MGXtëéÊùÛDƒX'„iËu¹x1]Nê+®ïu ó SìIZb<ƒBžàí(cTh‹è9Ê$Œ¥›Pæ'*d^AÈ ÔÞ>aÆFï.lÉN©òª{úyÉ›×Kê9÷|w Gµ'Ñ;7ìñrŠl\¶ê¥=ÜÈI®ÇzR‡ù4Uº¶Ì³èó15"à¾ø»S±ìÆa÷{óöËëc ¤t5,Ó¼OȽ7K\q ‘—?ÝìÜŠ±pj¨ôfÕ·=]šGbV7ãc?ó'U¾¡–_Qi8ûŠ×ºêkÕ¤ù=Yj3¾~Ë¥ ítr;‘ëÈÞ¤Œ²{N¥0uûИòm†ɨä1 ¨¹~ºk1S”tH{¼°ŸVåwÇj"ÓØoÏ]f¦î¬}eˆ|T€8n) ‹1éãC*ä©Ú³eUóQ¾zò C{µý‡:zdÕêÙß„NNU+p¸Îl5ø:?Ë,¦ðª‘˜‘mþí;ÉM ¦?œP®ñ|$™ëÀìI¦Ù¹raȳÙ9ô—>ý‚Ao}÷ß!‚éÜÄå .è”õ’µ©þÞꬠ¾æ$a5òǼ@bümÞG¯¥cj¦H.}b´c£ý—€‡MPnAä…¾”À нzáÙ8V‰/G‘ÜŠqÿõ¾¡£¼ä¾Êµÿ›„j×nðØ•/ :õÂËÿ´ }µ„áÝS¡þ,[&èf kŽ¢–è¶Ì•ÌO_2ÚÏìyÌØÐ§W>L”ÊäXÞoœŽhôÌ}Þ¸”G’ûÏK›Ž0º›œ7!Ì~R:mB©U%ÖðÞ¼Ê}Ýà §l®/„\®´p¡Êœï´¬ÍBä[žÏïú²H«—Œt4Y^›æ,eÇZÌ=Ÿmi`^Ý|«R¿mè }@ZžOİxŸš6 ©ô†Im~ðÊéÍHùòtÓXXx3¾w9ì`4y3JÁ×Ì›‘à7þ~‚ð µBæø\À'¢ß‚»®Í¼÷m"\Œ’%ƒ ¾”¿$UNr‘Bn*»Ù°¡T¾gK‘ËÁÎ'Ïó6I–”1v3R¶Nj™Ç’»‘šoTp_˜ |Þex‚òš«â–—2“ˆ›ü0•¦¨±H6«)¸Z¾N(S7$^tTDåŸOˆ»´¡ðV''Ñ(É(êàÆï^j™Ïë¡ÜÜyžê–^³ña耡ʅ5¨ø+Rþ:܉44ø[›„ÁG¡EÉÊá:„R®Û³PP¦ß;ã„¥ˆÓÛc²aº,XÑ+ÿþJÚNÝ<äê‚6©SŸc­·ô+ ³Œ’ …ê“~Ñh&pAÇœÌÌtø”î!$ðsZ¥>Å:´©ïhrÕ“01^Çv\H×M¸oø NÀld=Í÷‹Ó8WÓœâ—ɲZec÷ùaÊïIޝí;¶J9×Êí—SVØ·ìsæœñ?VM³Ì©j€÷Î5Yʆ ¶½(LMÿ|Ê~…[o¸š<Í [>•”M³í&juT„ô“3VQŸê¯¦Ô  T1¢•{RT€g)‚sÓ+?®¶þ…HiÊâcnL?À0#E”4T=üudHú)'Ýï˜Y¥ž~ÔÂ{j€˜ 5ï„Ì—¦SÛD¶Õ†¬ÇDÌÒ5DÏå–Y€ût0¸ù÷|,=uIfŠ!æ×´9p}½½,ŽlRÙ  „ãA­FÔ¥ñEƈ²Ð=F‰ÓÍÓŒ^¶(Ì%:’¼ƒ5» À>¦ŸçÌ,®Š)Zi¢€hŠ *â]Ûè“£[ ÏÛòÕ{nŽŠ&)Ed¦·8é.{ÚÓ7TŠnNߤËG0ü¬ºé ÷êF’š€xóPA²$Ä韢PéMû¢SKdp‚¯¬öÔ]«¾e§p‡¤9EGàK¯üAÆ;kêä"w’Âc®coFšîëteê!RÂZÙ£xÍÿc—á‹ò‹D.xû©YÑÃ_.qKÒÛžˆí—n òQÒ­ùL:mþõœ|RÎÝõ³í>³-×eF°¹À=ãýð•¤üá0[i;Á84Kåc®/Ñï1õB\5Ó.;ô»d­“Ëâ¡mî˜ÂJšƒï¦¼#9—AæO€ÍTnå¼2)]¾ÁºŽ(£±mzv/(ó¹©ÊÓ¡zmø˜ÉO…̧íjóÊ9{Ú9ÅQŸ‘Šm¦>=h<¨ƒzMÒÆ×lÁošßuXKöÎz™2 £èƒî,O"ìäqS¶Ãx´ÄÌõRØ/¡;桌ι+©/«4¯&”I6¬IÁÒD¦ŸÑdglü–Éá}ºmÀöÑqƒ|=q ¨?Ju€n†Eù}}Q ð&×àŒÆ|/nM’Lmç²S‚l£ú;²ÐJæ6öD È1s:²pnàtUÍSûl„Ø)†Ž5w´WQâúeGsÖ?*úžz›¦3)o2 ý(_…à“ŒA9%§Ú¶nò>iÂHyºO³Þ’v°º×[Lø ?öajçIœJCäfIÜA÷NܼÞÌÉgô$¤Úœ¹:Ä©CH„YÐ{ù|Ì:e}¬ùÍÉÿAAº2endstream endobj 431 0 obj << /Filter /FlateDecode /Length 5687 >> stream xœ½\[\·‘~ŸõØÇ†#݋鳼_vá]؉œda‰5X`íCK=ÉžéÏèbù×§Šä!‹§ëôôXJ¢Ùd±X¬ËWEžùy!¹ø¿òïË›³ÿ!†ÅÕý™X\ý|&Ó¯‹òÏË›Å70Bš°ˆCtÊ-.^å©r!eðì`ƒZ\Üœýmù_«µUjˆA/¯Wr"¹|S[/jë®¶6LßÇÚz¾¬Í—Ì”ç«Õÿ_üÏ™”C´V..¾;»ø·Ž vR£ù'æç7 COÖ.§\ ¤‘””ŒQ-Ö2ÚûÅÅ$ô†z[qÇì¶ýzËp{ÝöÊ‘¹d–{;!)±u?¬‚ÂÞËÚjs®˜_wµõž¡}][÷µ›ÿ‘DkÕ „Ž Êñ¤Q–rµ62Ê¥*ÿê4vmý`”“ÁÿË ol½#[kiô ÔRÔΡ¶Tmµ_CmùÊwké&PQ;‡Új}êÄ–?• GÚÖV¬­ä†–ó"†CÅyÉüÚTí-#à+*ÖQO›¸"#`=&¶¾\­¥õ’°<0j›„ñJD3Ä:r8E2­/³ž&«íœËšóCåc_[jë¾¶Šª¢öÙ¢|¦Xý7ŒÊm¹þŒ;p¬1”)ÈÚ]mík뱸µvÎBF{£Ú6Q4Š× Å™’ÌÈ©Á3µ¹ïþïI«ðY㑌É4ƒ2Œ€iKaÒî8]7D‡›½ž® ãŽäT#2­¶ˆ:nŠ!ظæ”×Wkü}Ãk&üék'ñvºÏþ$,Ó’L«Yè'!™=EfÇŽÌ8JP3 ÒaK? NC¸¾HH'ñgÆßZ³=‰ÖÇI³ E~=*8A$|¨]ŠÙ»yfúTÒÜ9j†4‹C¸ Ž’æ(ìGì}òù¨Ó¹#1L«NóImîÉGÂùãÈôÑCZ1¤9of¨úµNAZË3¤ƒb¬Ê‘»E“c8ÜN«ùºG‚œ;f©„l9ªœ·$á×Új(æºMùcí¼cÒL„R|ÂÀ§?Üá lSYÇø˜ßd95è<ñi)B:þëQ•m.È1}’áÚ5/‚ŠôÝ?ù¦Õü MêŽÊs—…1'"ŸÀLöÌÂd„–iqEÁ?ù4®™_ÛI¸ÇÍüÚöOÐ%x2òáZšiù òi"öL«Íhþ)œf¹ÐÅ%;ÔP>!àJÇ;\ø Aš°<ú(FsÜEFÒœôï¯î—ÅÈáȘ~´srGIDZڪ$­K›!#KµÕ†¿/Vk1(˜g<„1èh¬Q©ÈÕ<ü„Ø'#„s¢‰ °{(@·Q: ðA 2@;‚¿ƒ¦0ÒDÐìÕ>juð[˜á.‹Ì­OÏìJ›0ŽÞ¿YäL/Ä¥œ9‡ECÚ½5n$¡1êqU„®{³Ëý@e ô”*!×=¡þ¿¼m¤ÓŽ£”Æ.÷yË>([v¯¢2Yl…ô6Ó°ÁPöÞ%¹+`:`OZ`¸0BéåÇ|0Aør22H—)bß5²»vœÓÈ®-ÊëëJ‡pÍ’ ä£F¹ßVÝã*gok‹ÌÛ@wž€È• õøoyJ§kº$ÂrÒ(ØZ„%ñ´„°DÒ~¹ße™9!ìÕP\Ãjd@Ùhd†ä¿I#U†½Jæå—D…α;íƒOøÎPMð÷U>Úà"Þ(`Ó)ƒWi²R=Á¹tâ&@±Ô>_£D-‡ ƒzq‡ÝŒÏé®1®WѬwx`Åô“нÍ4€Þ„Ô žäÛ £>ØÁòê]ÓX¢ÜU|zqö׳|K„}‹»#WsýÑ”«9¥œÑB{ é^ÌC½øq¤ühj¼üÒÈý Cܸ-N¥À¥h%ôcͧª„§—=³Õh·ÕŠÝ€atÞê>k‹:ëP[O<åuS¾krû•rÞ$µ‰„’†*Èݪ…Ò¡[ðºéE]Û._¥µ™Üåµq½-±û7mÞŒÃØ‘Ñ»¬ZÛ2QjP¸ýÁà_»åÂ’ÖâÎæ¼1å{bÊ—Mz/¹\ BÊr”ø¨2ØH¿Ül[8ØòËîä` ƒ²~Rnø@Bû›¦xÃç´@‰Ã{¢/\ÓúZ¥„Üû¶‹ä9wû7ÕÀjŒÁà&!­Z#›x³?Q!ù“©$UþÕT’Ð ‡P@ŠQwZ–i ®$<6eé]Ö,ŒÝ'ÊÂ{£*s°ÞþÕHÎΚ‘·MV1ð_¶ó¨£)ZÙa¢á#H\OQN¥¨Ü9ÕûÜ­¡¹¡“$£™-÷ï(h@¤‡taŠZoÛ2¯š÷O– [󨂲“G,'0p°ûìð‚(5£õ4ÃÝ%ÅÂÅCómW‰r"žêEj¯'¶¤CñOoõˆJІ»ÖX¥¡š¹¿&CöãQe©ik­aeWö˜¦~DÏ ˲i°Á—wï3ÞÔ¨b•2: â‰÷ˆ§1·p“qjC´:»KØ»¢ ‘¸BªVIÝ4‚ª¸”ç¹VlÊ‚à.·È&n4ò…ŸÆÁÍíÛ<(£W¯Q4p!™Ž%Œ²v½^‚¸:èPBDÇÂàJZ€>{r:C¾Þøxp J1ïin‡¸gÜåCïv>çœ‚ÆæyË7oÑÞ¯´ï¯jëImq¥é6÷®-ò{Óp8‡LÙ0»4Ð9Ç!'ɆÕýß*hˆ[Z¦h®b*ÏLáÒØMK6Ù¿+ˆUhB8B‡p$;C›Ö@"dcŒœ¼bÔ§y¡êŸ Ñ—h¹A‡ajÙ—{pONcŠ›F¹ôeìÅJ͉õHçHÍÀ ‹m…Íï¿nPéYcôY=ö*†Ì´,±b—«1}úiɉŒ ÀU9?sYUD©Úå•ÂÛ<4 OÃíþjÝàÀu ŒX˜¢ †÷,àrôÖ;KÓÊHwÝÂlAëÞ“ò“ÌŸ’šÉÏ û8ahÁ‹ Á÷+k€etH$“$èj—!„Êû x¡Ë÷ #F0,‚-èNû •VÖJ„Aú8IÂ(xK‹õI©(`#y º‹3-=Z,ʸþ±˃`6Õ\›EP¼¼¿)$„K­¬±ˆ1ñ¿PHضŽNƒ—Tz«Þ7¾1ä°@Jò±:–`| ÁÊè¨{Àö²åiD)Ò{k+‘ªõÁqT>6÷£ä €ÁZêæ¤D×üµq5­ìÚTÜâá+AÔ—-óBÃÁi.ꉦœF@0lù÷)‰ÄeçÜ¿Ø Éßb7¤ªïóÊb&õ}¨(ŠqJš–o}š[k¼Þ{õä-ßÇAðùLÍt¦žSê§ £ÔO›#9L $C“ ðh91Kq‚K23x5'ŠF¡]µ]BÀvIÖ@^ £ Ø7€ø»ÚjUÄ=ƒ³î™÷¸õwìùˆgÂÖx–ó¹à0õ¼"‘79Jf}DUµsŒG‰ŒMFÛ,c'ëçoPƒõÍf®ˆÎ´TwOMËG[ˆ p]KªÖ¥ ‘ä=ñÄys¾ÏÛ×X%Í+ˆysæIÜ †¾dB‹™û\œÄ ®ı¿ÆfZ©ÂÐ0KLï×® :Wµ„Ð4§^+Ìé™T^vyëÑ+…Ñ5(DÉøAõ3Œ(6ú¦z¹² 5Ü£X²N)ŒÅÒŸœÜFýÛ ´:<†í³l¼q‘øóô)Æ€e?¢bÄœˆ‹N[ŒÁ‹ybnäò‘’ è#,'dBR¯Ú<¬a· ‡e¾Zû¯hl&ø’zÊ’SISÉVɼ$fÁKiâqgÀù;Μ&±¹‡¬%xz_Á×QI«ÝAta£+lUh@ &u¬ç‚Xãk‚O/3_¸Éî*5Ky°0Y+Ýô¢€\¢m( &5û_I7廌Æx—ÎTšO¹5%Ñ©!pzE¶¸ŸÌZ+ Yìõvé#gT²è‹¼IW޳FSPÄ]EõÝ&fÙ ýÐñ‚xt· Av—ãØ±èœŒ†–É*‰LÙÛåcïM ÙÇ3 OòÞ“šn=Fšuݤ¯±Íl:á&7`ØëE!1ëŒØ›µQT}>›ˆ’º—{Öpºê2{mY˜µ} GŸ\ènH>ñ±³ª}Þ¦šÇ$^wrKì"ùçèãeíNLjG<ðÁL’šMÆJ3٢ж‹Ëä¼ndtx-A-qs q&ž®+WÛf˜9Å4?êÙ«‹y{¨ÝéN¦âü³:oOÍ2ÝÙ€°DOï—÷ò‡Ž Ïf 6-À§k+LÖÇK{x­ÆÛıd¾T%‚’¶3 uš˜¦e‹ÑʾÕ4’b·ë¯Ri£@¿h„ßæ^H±NLQŽZ¤ÓCh0æ3$kºÚ÷äNk‚ðúGCTÎ3lì>½BçsWÝ­ÜŽ\ U-Ú8=ƒx·™œ«A¢\;¥c]Ãiµà2ñ3½4Òp¶NùhØ€éNzdÈë ¥Z‘¥M^A˜±ò“·GkÀƒ:È«¾î‘sEûôn ߤÂ2h²ˆmM°”Mh™^“Ä—\f7™åÔëæ•Ñ6ÿcžåcÿb_Ê Ñ’šÌTý(Áq;Ïð¢i8Q`"ˆmÏ}Øé0œELž¯,þ€eËjQ<Å÷Q}Ž—@çämV–h\~}>òï—ß´Tzƒ”ScM|D¼'myÌÐdݧ7ubþ –2¶yŒö¨‚¢¯û–uPdt ÒäJ‰H§éô{W®oŸ'ÖøK&3‹ÊXvÄéMm¨æc|†äÍd-<ÂîÒ„¼É˜yêL’,ùz‹åGQ :ˆ÷¢ïëKùµØ‹@·Éʰñ>;deAĉ•áPèhÿŒ æ•óFª§û…ï@Êë–Ú¶z}—¶E*§®Üç7r«€oE”:Á(>“ÏÆRÆeéÔ ÀwãûPõ ïCA(CД\òѰç!ÌÛ;ÕC‚ª×‹8QëËmœ»M¦Tcg*y¡Ênmðý+ÈùU9)9“ Å€sº³ÒÙ$-Ô®I1‚fptÂUPAKõ{R±j°––+«ö€†Ôô9Ö£t¬Þ¹ä“¡ø³¬«”¹‹8â\~i`ñ–˜7ñâ‡I¤¨Z¯fú*Â,Ìü­ïcìX99‘HŽ Îpê FÛd‘¸ƒ¤*ÿGäè&¹’q¶~0N£¨X&L\•ç‡ËLq´ƒ„ÉûQ:úq¾`0Hï(À2/>Òj¼,ÎÆ+ÓÑÃË€ŸíOÁq 2Â2Xg(+êÏÔjwãu]<õZ¤s`ÆÂéê0&¤5Ö l¹,VsÙ2‡¤ßF8sѺøê(ÐFHá;"úoºêÌèn’:V‹‚}á ›ø[I-³àEÈ‚ýÜK_rÅNqÌdžBnBoæGÐC ¶Ù_AJ½êüÕòùs x&ŽZHò?õÒó¿[ÚÆ bð`õŸRP»B£ÍôõÎ>÷â»ÊÉ#im3óu×f›gáå0B[I’`¤2z_hß[÷(QÆŒµÎ {ÐVÇnFŸf€;Byù]¾«a³zc(,àÑrÚyÓc"ŸTM¡x¬”ïReñÇSé<ùrâàšÊ‚Û&§ÕônòL¼bìóL‹ôò‹f3 ǰPÞß@÷~‚ž?ÏÕü Ï[qe¦2ùd%•…ÀfÆÀ£9&ǦÀ¿ÿâÀ7‰¿,‹‹›3à‘«=×gÏæ£@ ~Â÷1ëIÿr þ©l²«V9Ú€S®–”üÂ!!Kð-Ö 2‡’â|» 0D¤èž…s:{o GUï ();8UÓ¼+2ñ:¥FVÙX v"‹Ä Ô_8F•$y(ˆ—¢Ê åÔÉ,‚3ÓÍ#Rw'ó˜ÿã1<Ú .!@¨í0³˜õQ"Pù¯y‘ºœtãû3P4p^ø] žÃ÷ä¥^W¤7#=ß-¹+Ýã½q[½R¹ÌCÀÌ“!¦ëT›Å\F1£´ÁèÒóׄ %"Kìu¾1" ÜÌîRøWj„·…|°§$Æ­lÃ¥:d—síT¬õàεI÷Ö‹åk2s“Ü«w±R© d7yÃ>Ùm’¿ %çRVù´#'Ÿ¤A|ÍI5ƒ…kgé¦ `@’V“§Åñ»’²©èl·×©€ÝžI&]O?»ƒ„T ©|ÇÇm"¨…r…ˆ².#§²»7dðÓƒ)c¶õ »%GS-¦Ñ½bÝiÈj t®!gu€"[Ý€‡ ³&_\Œ ?œù+([šùXõm2:ãTyÒRÆ\’!?O˜–”¥”ö&ê_³áôm¬æ²æÐØ¥Ýøè¹ùK«bj½’5ãyCúÓþ©údHµû+2üIÚ¾ Ò·!ûŽbrO†Ü“!û6dG†1/C¨¯ Û×m( î!¹v@}VN\{uZtkÔù|Ìl!ñ÷+‘Ûiõõ7ÔUÑA/’æá§yãø*-Ý O€@À ÔDz¨Â+¼iîì´†–QQ†ã: Èlµº¦Ïî› ÁR$ ø¥Hv±VéñNšÀoêž©hi(ë¥|-¨|ŸÒÇÁÄHÊÒ!ŽE€Ù&ÉKœÊ—7fÅ}%mµõu61H,bú@*á|ÊÙQ=,ó6m'[Ø–À²pÂþ5Ùdø}(@”«VÎýëÙß²/endstream endobj 432 0 obj << /Filter /FlateDecode /Length 3240 >> stream xœí[ÝoG·ø#VÊË.ñÎõ÷ÇéÈ P¢$J¤$øá$ÂIËÚØ ¶—ØðËýíWÕÝÓU3Ûë5H.õT÷TU×ǯj–?&¢“ÿÊßåÙÞ?~‹ar|¹'&Ç{ìÉôtRþ,Ï&€C*9‰]tÊMžíå­r"µîDTlgƒšœí=žþ0›ÃáB¯§ç³¹ žÉé«®iâôH¼ŒÓ£ü܈0=ÅçAû¨§+Í•ïœ=Ö‰°uº~ÌQJc§—ìŒ}Ø(CÌW3åà8gý:ŸfÃô ²*!¢ñ j=á4é¼idö@®ŸáºBÇ,³T^9nŒbd'Ìôñž£)¢±FM‰i=¸…*û%¬:é­1ÜpgIä¢Oú».«áÜÁ q³å‹t*–OaÀñ–”~9{rð㞦 "øÔÁ!x¿]¦ç%;‚ÓGIdåà…LÓdÙ 5èÏ Ä÷e(":º‹QÎ MÜ6‚OÆé"ÔÅ8xÍ%½üŠ]à,{)œ7OÖh(Ôu®¥ï¼Õ“¹ß²VfŸ&³ÚÀ§Ìÿ³ƒ(í§ËÙ\Æ”ùn“á×YÚ½ rE±´>Ï:Dé,:&h€2²Háž‹XMà;hßq~µŸÊ" }-#!—³à/RBê[eÙ¬ö‘Ò› myÆžg4tïgs«øžœ>™þa^ЉBõüì&×½CoφŒÎH4o]f.¼æâÑì ¤"žL: ‡Ï$¤ð³Ã„ & .uùbfÑqCÆÀ¡<ˆÑŸò)ùwÁuÅœù¬ŒàÔ%Ô"„‰l™ÙCüWÃuih 'üVõ<£°¨92ÍΜÒDˆ HEY¥$ñ º^gSÐbtZð ]|÷m¿»^´ç,1”Z¤=æyzØ"³Úg®àà mÒ‘Û%E!¦¤aÔz!MÈ„9ØÁjN: ݦeÆ K¶ŒÚ@qs*è ¬t2rŠÚ)¬;ZC} ƒpXœçeÑÃ*áQ^5å´\vÒaƒBÈJçÕœ¦¤•ÞóŒNûšÑïð¡¥è—¿ÎLïWêA¥žTŠø²§-?†ìbMÍ.ûU÷Åe¶ž}=´.ñ„é3Kþ FàÅœ™ix÷x˜î¦IÃ(lMM¨Iã%Ò«YMg3— ±÷£Ñ5ÁÖ) À¹ZäÚÕ²ƒ ¡¦Í% ¼õ)Õj*3äºîCç‡$€˜I—ºþ3ª,b,ÈiWøU¯AŒÀ¡ ÀÌÝ*+”u‚]Ÿ…,É) x7øO©çäPåõýƒ‡>T#VIp§•œsƲ`ëˆXÁJä]QI9gÑQ^„R1—¨´ñ%V0ˆ$ÅÊÖ ¶qÐXÁrJMX Á”H§„ñ™ÅKL‹ÐÑö\ú÷)…lA€É}3‹¨w§† ïßgÈ~«pëRâ± áŒÒ D2ùr‘Þç,£&LF[ )¯,¹õTK bË!ˆ®Ø•—¹­QZŽ+7°žÏrþ¦ÕEµèU±à‚Å Xtð´¢Ê¿#ihSðŽ4k4·áà àþÁé Š„”¹Ô uO¼Ú,=×í­A^¶&*¾<èxéå qq(š6b¯äù5a‘%šÓÁuàqøm©îAI»‘_6 2ï0YŸ‡…ÏØÔ´÷½ä5Ÿ·=ï!Yk z2ÃqV."¢:Ï4VÉ7‹“C¯‰¦Æ²ŠæJÿ9³<šžºÅõf©I"±Ûa‚^¦¬ð»Þ¬‚ÃÞo(gvNš¬ ö^—¸³EÑSú(ºPUef^³TS=4L¿›Ý`dà 8fád[¨0ÇL ?‡}XÑ„ÉSkVzˆ\9’,¢@h{A·Œh†WUX ®à{©ÜÐÎèö xò\Ò*^¾pˆ»Ùä]ÝNÔ>`á¡Í3ðie: …+7R„œ,¤ÅR5­´Jâ«r^éJ¡ç²\¡à< ÷—…3¦’^“Q֣é£àRKè•«V;ÍÕ\z[ô:ÈMÇÉ™••(•Õ³ºšz²/_SfiÝê¼åì¥do¨RÀu©à€²4ï©u'&u—mÖh`PBJÆ1¯éi /wÂÚ—Ä ï¸¦,À.#â»ÂNÜ÷Yy1¦`;y`Ç?pŒ>å,êr˜\z†÷G­7˜ WS¼˜Aùà(äÄ!3Ëœ “õw®¶ÎXø‡ˆýüm‹·­Ù„ƒ¹«\Žª7#Ýu''0W/8,=Ìë`ŸÑW¢ßfÇŸU8–ã›y/ù°°}Ÿ° Þû¢R«J]TêÛJ]Uê¨R+µ®Ôicí¢‰‡yY¡þ‚Û^>`™÷üs¥! UÑä›þs„f¯&Už6”Z4ÖH=H0=ùs¥îWêQƒ‚<„W ÿ´wpw ÙW•õ+Zü¡.žWê°aå—µëwÐÓ%{Z?ÛœU º®Ôa¥Ž*uJ[ˆñ-žÔÅE¥V•º Fz|Î^XÓ¯˜õñ²!îicíæ/¼l(û¶R»Ìظ̓†ÅÉÎ]¥$mùW]œ·Ü•"å¸Ggc×D!ÿW©ïwÜBO}]©o›öÞä{t­™öÉž‡ ^5Öî½³Ð׋J×.Ïè©‘zóËme[ÈL [¼r.åGJ9§ êŸ-»µ{ ù¬™…7·¼n¸ kÜ#òæWÝâk"Û÷Øâü`NWïù‰ÝÜÀ·êo˜ È«†£ÐU³ËkŸ.|½k™.˜ñŒö?U¾ÿVJµ Ìrþ/¤%£oèyûÌ‘æmy~jèô¢­:Oq»nP‡äÇ¿ÝÐÔkÚ"«ü®Au•Ò• J’‘È ?ØD l5ßûÇ/ Ý[Øå²¡ñ9¾MÇN¨¦ø¦!~l\ñùÝâo|ñ1]4ÚÀ¦5};Ý_²Ÿê°î‚ÿgÁlhuÕOc÷óè#lý[Œlÿßh™ZÔ=¥ñ†´&ëáo½úóØoË6¦Óì›Hf^”OÿA¨OÜdéüƒ±ÅàòaÝä‰GùŸÔ8E瓦Ä;þ¿MžÒáôƒfÊ7DZ„á§¿áH?@;äy7TÆŽi¹Þþbôí(-²ï_mKnùv‡{We8¤éÿÊÖli´°,Dˆì”~£¥i6­¶«Õ¬v$¸Í½¦Ô×Qä`z³,¾l¬]¿ƒž.ÙS´Ô-Q™úk 7[üp`ˆÕÈÊy÷#࢞úlÀø­ïz ϳ[#pª¿Ÿ-ÿ¢ýïsÅåRj\nü߸üv¸ÜVª…ËiÍ7¨HF¢Ã&.§qÁÇå¦!V —›%v‹?†iJan<§ü±¡ÆŠ)D·5‚ ÛpÉöÑæm!Ã-F›dç3ҟ״ïùQG/=õ¥C ½}ÈgÞb£}¶ãƸâî.\ñ÷¯æÇMGù¼†|Rˆ|0¿0Aõ–jp¨T jðµF5¶´(gc žÐÙ´×°§>iY¤eßÐI½—EBC†–½Æ"•ˆ8pšÌUìBðɧÊÔÿðëÞÿ™> stream xœÍ<ÛŽ\Émïã|Dà àîÝíãº_Û@kcÞ8öàoZsÓ$3ÝÒÌHZ½øÛMÖ¬ÓÕÝ#­`û Úê*É"Y¼y»“\ü¯ü{qöË?ǰ¸y<‹›³·g2ýº(ÿ\Ü/þíVH%qŠN¹ÅùõYÞ*ÎLÚ.\°“ jq~ö×åížV[:#ÔòjµÖ:LF†åÓj-¦côjùV)Œ—ËÍS^¢^¾i¼ŒË\ì…N/o\°`lWÊOW/ÓÂHËïÚ€-n³1Jg—7.ü_üº‘P‹“Žqy¿ÂÃŒv«5œ $ ~É`]Á :µ¼+è67ùäà¢Y¾_)ã×j¥Ýò1a„_nò6!ýòvBh$SL*ÀžSq… /côò:ïsÒ"´6ý.‘¡`/ Ó*†èùpƒ µF-?VŒÝr×ÀUª¨ 0¥aÔ.L-o 3„Õ=_*·|òJØ:434U9:Â*½ü§âiõ?çÿy¦•œŒ‹ Qç— AÓjmœL,þ×z‰H¾@6ˆÊ Œoâcu6³B <%!ï¦D`Èsw‰÷NqŒ—lÁf[ ‹×ЀRÏa ëJÇZÊ)Z+39ð. ‰štÒGö,`Þ—ëR @ó[‰ç* H•+GÔ9‡QlÈû¯€oê ð§3œ ˆ9¨·L½C˜¬µíB6+ b žQG»6zßF´î‡% W«rDgA@ˆa=â:¡f¼:‰Œ-ÎF%@‡/Ñ™$«½åj³c«§d?bå¬ôvøZ*”q¿X+5‰ BFâ7«µ…‰T!%‚¼oÛh×FïÛˆÖ1’_·ÑU]´ÑÔF²¾©'ëgoV?e³Îh;î§“ä‹þÃÙùW]þ¡QùgbÑS›¼j£Ç6zbs ½ëø<Œ°¿Ðñ0³‡•#(ƒ½wtð]cÇ®nÚhÝF´î¶¶mtÕF›6jÔI0Iû‡\6Ó!éj”ž¢M)wñÝ@/WqG7%Ûä‹Ñ‚ÉÓPì”ñYúÿÖVýn Þ·mô@€¿&¾l|ýÈî®ýÌV~?¸Û^ë±j@?ƒbu”â ˜(.Êûe‰g°­üêsø?›úùðlâ¿ú²7oÕ¤À{ŒÂ.TÆë·ƒõôÚvÞ§‹.—mòš&G§ÜµÑÓ@©«´Àð·~ÏkSÿÑJRõǸÿïѦß|¡ÃOÒ˜eÂWñìQã«mÙ÷CM.×;f²­#ƒwYäW ø`#Äã)`±RZª·¥y üòN gL"´Rˆ²ƒô‚n])ܤ\¨qÔMŠ ¾D§¶„çÁ¥ Äô6Ef4OHýŸ!348 Q ˆ™£-ÿ—c|”…¡Ž|¶Ÿ…O4®¥ƒWFÃá›õ%l;/7àUÎ,騵ô<ŽE8†R1DÉÕÑ}ŽgAl‚‹2¿¡`ô ‹hsÆÇyÀo·¹XÕLEæÌ)¿})ëÐ6^bÆÃÿÕlTÊ0ý˜’jB€,æ¨|'ïø ã1M+X=‹›[ç1Ÿb„MÙ;¡³G!ƒ,©;é¢ã0Ò>L›øY:.Cp‡³M5¨/<š_‘ “ ¦] ð)Õ5¼ 9Yï}ÝHyÅÊrPÊÿ[Y Rãürss•Ñ7¸âŽ’zwŒI«šBúN±0¾± ]Ð4Špm÷¸Ô#y”²IǺ̿ƓYn0•—)5ÎQ=ä}^ô™)–ºÚ¤ëõAk[ïð[t?%'Ö¥¹”Ž“0væé• ŽT Ù€¬ýòèf·W}£\“Ï`âòmÎÔD0™_ù~Y‚ªdD€lóš2ç5y®v“õÀJŸ³®)ËØR`»ŒúÄ}"gÀÜ$z*'DÛ²× QEuwòм# /¶Ãºà»c¾A „Nv:ûÆÕéÑì2Óéè„¡åJóIžxNŠm@ÿ‚ãâŠ(9á½QJOyq4¦ËÓR;ßÓo“9Rª–ˆ$Z9&ÿToax¼£·ù¡V¢sU1ç–$Ý!>IdIš€¾!»• zc÷KI°smÆ[cfÓ錤>Õjˆ™*Â}ï" Ni%¤¨ÏŒeƒâàñ+æœ6²ÙXqdW”9*ll¾êâj@‘¹ŽdrÕ%?ªˆï¥ºĬìÞ¤¬Þ¸È™9[ ¶ˆãÊ̆-Ø0=GQ1:•+y™´ 9(ïéo™Þd“F\îÉw)ªnõ–¦ÇÎßUF3WvÛÅqÈL(ØŠÆgßñù†ïÜžTÚdÔ`+¯ƒWt N¶aâ|›ÕDb™³ÅPüŽÈdí½ÓÜO©.Zµœ£w ‚çô~‡½Bç B$À,€UÜïòÔe[ LôüAÚWºhÊìmš §ÖS5ÅÊ,M%4ÉUOÐ×Msçõ!W,½ ³£¿Ãê¿< *o^lÚ€ŠÍ ^#ïÔO£ú…ì9-ß_sÂwçÅpfTRÙÚô—™ÊYQMµ”Â÷½©g'—Gž!ÏyÄ¢V„úòüìOg¹­„ ¼€Åçö²è(QA€+^[îfÑy°vÅNúTèà,‚òÌ ÏÊÔ 4, ‹pœy`U4€™w\øÅý™ÔJMÒ¶™»³ïa•™¢^hPŽÉYXT'<ø”eÍA\ààê‚\DÁg\_&YÕ"ÕlXçŽ(¸ IÏ‹^&•”)3Z8x»ó%gy0|á@ªp¼ cÑø#Ñ#G4zÛ #¸±œÔ üƺ2s”-Ã+”ÚNÚªž-çM øaˆ@±s 6ê;<eWðâI_™¢ ×SñšöðY0“±±½o3û hÚfGî×á Ø…Tá¾G?R ©;Ÿ™ù/ÙÓó#…xOä?P°°!J ØtÛ[÷œRªèYßVÁyf<µ4ˆvN!X`/2ÖJXœ°[îi6àû´i¾ØÇ £’ƒ¡ ô«Ìܶ„ 9zsBåÆÀ4Kÿ–ââwý%–óŽš2]ð »åˆ.µ–u± ‹tsÄŠ08kT'Tµ½f§ò¼ƒo>æÙ’ì–yB—*úHzVØK|Ž ¨<Ø~XYûîxT±·®’à[ÉeßåL­õ &ýeàF“:¾X¦'ƒ8ÝVÅ3z©Z1ákÆÁ’€CñŽ3`z§…èãâ6jº”i{ªyª²sÒÓ6гƒA%w»¾²D²Ôóh§÷7<€onäæºý¢rЖî-Âd\3y¬ÑÆL¶â€Ç°ãÙœÔqç‚TÍ-LZuA×;÷”žÉ$‹þx#‘ܘ—LKÓltà\¢ÇÞ“Ÿ5íƒÏþuáAï`8YÖ˜§1’ÃZ]2 6é½y]=²ÚªSvò¤s ºª¾#6À5˺ߪ›ôö°ã°ÊÂìàýªbŠÍƒ®a&gäRŠj³×D˜Ë%;–…äŸè"D–«V Y÷ਠîa^LÂ!•ÇŸÿÂžwî­+ж_ìùvBÛÞŽŠñ£1ÍMmt=#©Çïžù,#ôj§O:<ÿß!†ô²¿(¥^Þ³ÑêùÔFøªYM¸5ÀPŒ}MUjM¿S? ïÝ«£ÛÁ¯¯±:õÇ/¨=ƨyÏÝ¿Œd†·Sæ2ú ¢øNâ>µ5QÑþzp7ÔE’5lÄüÐ&I“¬bDñûÁoel¬þù ROs¿Ì©g®ÓÏ\gŽ®›£„¾ü„¡:=Lo C_¸ùÑL,Ùp„Ö”e«îïß&[TÈ&Ù¢¾æ¹º÷fæÿ…ì™/.{9ºåÐÜwA{/By¾œ©¡@˜Ók“páG.M¶Nn:&fÒA)ûvðÖ‘¼½%»ËDŠš8ù³×VR;“€8š¤v¯ÀŦN:¢:€þYcüSCJju)oBéSž(í>Á´˜X._tZìz¬eËܺ’³SAIÞisÇðauø.ˆ€c4³šºàÀ—ĹƒÆOX`f‰à%0ÒXÓ—ñˆç5°TR»³k²Ó;–ëërÖ™©)Ã:/½§„¯™Ý°ÂzwÞ€C ÏšìØ'ËóŒy¢¤KtïJŠOOÞ¶nÌÃ1åÎr™‰‡@&3g™òBw÷™nÎé(¯ºØ›Ý^Ž*§ÈûOµl%rfñÚïz ©[áy­*…ìþÕ9ýΟ {l¶=¬@mc¾œO2+•…ïQJÂÝ×÷RbúR˜Ôk5˹¥i–}•ga篯òá:ê±ü2Xü>8B„ zVz¦ûàé¹ù•Q÷—Qä¼J–5>UÝ"‹/ž³Óõ¡¾½kø•Cb6ëûžaµÌÑŸ>˜ o×½'¼§þê@…½n9úYOnâö¨sü†¶üªy|kòÛoQ7EìÿÐÐ?G”-ôüš†=ÒæßŒ‚bI]q£Ä‹uMC ÒwƒÝƒ­ÛŽ;••!1ÎKäD™$9I'Å‘g9ŠÑ%ý…7CŽí%XÚâùÚ`j(ƒô»¥}jê]*‹µ¡Ôò󧳿R”'endstream endobj 434 0 obj << /Filter /FlateDecode /Length 5026 >> stream xœÝ\Is\Ç‘¾ÃçÜ15,ösí cäY–¥qXŽ1‰›8‡&‚°@”éƒûdÖ–Y¯ë5yb¤ P¬WKVn•_VU·“\ üWþž^ýæE «óÛ#±:?úîH¦¯«òçôjõ»h!XÅ):åV'orW¹rfÒvå‚lP«“«£oןËIˆ(Ôz×JÛVºl¥­tÛJ¬nc•šbÐëëTŒ\¿¥Ê߷ʳVúÐJ­´m¥]+²¾mÀ“A—×­tɺü÷É”ž¢quò§£“_»þ¦Ñ}ÝJoZ錭¾Ì§Ö²U>'"þЦ¹¬ï;jøO*^ ZÒZï˺¤ÎŸRñv6Ntêp±RNÑZù#«~‹-Ÿ%µ”ëì{-¶Òu+Ý´ø¼•¾xÂ(Èú'ò[ÿ|ù}5à㮕îZiÛJ—ÿB~K1 ÎpÓtþ ¸´nœIey*æ¾^ øJ<"÷õ¦•†#rßwxÂߨú¡•.lÝ þR ã'!×ÿ5àá«u+þv Šwƒù^'¾oÆŒ—Ç#£,Ηd7ÊÉ0Ÿ6­Ò3.Õ}¥Î¡Ù”ÚØo´4záÅÁp¡•"§¥r•ÖËjx°0ßJS+ÙVR­¤Iõã ‹` ÷‡!YnZI²¾mè_·J^jÎÝéêÜõc1Æ £,’Äè­D4(/=Xžm¨‹t‘ƒÎ†}ìX›f—¼TuáåÀ’Ï[i7øúvß΋*(`!ó™ùsb“HÔüæ…6ŠEY¢ çDZ` ëOŽOþºˆÙI©enòm‘.ÎK£±¥œƒæÀØh“¶Ž,²'ñˆ™æ³i9i%Ά+‰º_©éù@›î‘ä/ÙÖ˜hå&)”~0ûÔÔ¤šÈ ñ”Ú”>œÐiP%ï×S×2&u•b ¡õû­F'ÇôIiÞσÍc…YHQ’5§&«Õj£ ,Þ‡<ôÉñ€MŒ^)ØoĤ£ÖÒãÆ¥¤èˆ´ÐBÈhÁRa7Aëu`ªP¼ŒØ–)„s£Ì„__¿ÍcØ`¡& ÚG ^¡wØÑÆ(…A ¢v~nrG" åal¥þ¤J•Om{›êañ£)\m4F§e¡[±ˆu¨:î·wX†ŽÊò9¯wu²‰ÉI+ úŒ R¸#C4ÀJ1yi­ó8}f0®gw"ÈB·VÚuôÝ®Fß͸͌.šÌx`l{šF Rûn9×ß'ùkÐKW¾Þa‹ôɼ!tD7 …)HQpР L7FµPk=©êS¾]_³å½ÏÂE”€˜Ð˜ >óN‚²çÆÀªª6=ƒYb€©e©°¹IRHU›§ñÀ½‰ëówyM^±5¥X©ÒÙÇJY¢FB×L °1i´öÀiö=FpÖ‡D4À0 œÓLçbt:¬¿ÎßÁø‹r“_gBbGX„,c­@›dOÅPÁo{eÃŽ@+Ø'µÎÚ êíýzû?xðÁUW¸+©Þ!‡¿¥ û‹¶‹Âúº•îcþù”˜‹?Œ^"Öù>tVKÿ 3 .>m¥9†{ ¢zF!ä›AÃQçÏZé›AœúЉÌ ÂÞ¡'ë(CÒ$”|ÚJïZ馕()E1Þ¦ñê›GÈ—xO9"åëÁ(#$MrûÇpä†ã^.êXÏSBX·¿ž²vMÔo[%GÊí38ó¯[åÇëQê!LÞ¸»|ʸ©ìë>0ýnÀô­ô~Sw¹ò²Eì‡a-uþŽx@C çq4qýuFÇ Ç'õ&•¿)øíσõ’º”ì†õhê657ƒÒÈ ¶3F÷¢5îˆC‡á3!?BÀ®Ãàu9èMÁjÔÑ|#LÍûð%ù”«V¢]cäg˜"½°ø‡Ã8î‘)¶ÃBã›EëL@•rOäÞ˜Ãüð‰ø®|÷´|{Xjz0¶Œ}Xj¾}²™'ˆê‹c/ÿ‰Ê Kmù,U52Oêý/´e¥(s6Õ6÷KJ'}5àìnPGæµ·L,û%$ wtõ`§H2³®ÇÁ×0²)fh£qFš1J>–^MOHWÙIª”þÀP§ðÝ@€nJ¨)FÄG _RB¼na÷Údèÿ5B)ê¯w›Ë„àƒR +•âî¬zO^áy ÃÛ„;¬ÀMtÍÛÞ¤êh䪢…Ì$a¥-ËéR2> RH9HØEWh:†‘$Ãig¬EF‰°Øú<Ã¥(,¢%ΤL™i.ã©–Ã(+š­ >¯0÷œ¡R–-\Dhiœ÷ äDWaXÿíØZà`ãíya‘]bã¶0&V4¨€™úŒÖ·wy}QJÞ¯Ô†Ð7Îä'¸ø†ªS*0[V°Z§ÞbÜšÀݶAß “%#Z-¹,?Ðú²øœ—ž“yÖbÌÙÓ=0 n.½ÄoS"¤¤¢´š§¢ÊX—”½Yâ`Ç]nbàµ;–ÄHsDùuà”Ã`ë¤ Æ5ï·aÉŠË”¬°Îò4Èîx–GK àU9所ږ3Ü2{;à9»œÕQèBŸ!ÏbÊ­ü°h„-—ÇÆƒje5(¦9`@©…ð\ÍóÖDØ`8M|š< l™škÞMR,ðü.‰ÇÖ}‚l‡&¤ƒŠ ‰–›ÜËK»`6d÷H½‚n*eÚÆ™»HH7v ¬7y6—, KBö¹›œÊ RHN2V&çiÏDºÁdÞb2”HO÷/‚œ@’Õm+sD€›oÖßov”*<íìia“˜eAÛ2'ÄSÝCöó ö9 ý]—É«ºÄS€Í8x~//ß¡í÷f×úñ1î²h¢4]“¿“oÛî ®@mÂÂàÉ+ë6-ÆB–åË€€Ã†ÕÙõbÊ05–|•ìÁG£ZÎYtER9n.艌Ą  ÔN¡0Iì}µN{¸´ë^çÖ!Œ“÷ûšÝ’÷¸¦´çƒ§0†:{ŠNŒI-¸ç?=Nf Ð{¦µîg§IYÒ´åßf’Ñ}ls1ÔhëQ¿o_²yÇêøžµe¶ÁÉE[µ.øY^Ÿš|À)P“,½Ls#Ü!ô~7‰ZtÍuΖƒ+í¼nËÍ*AK˜¤gFˆ mTÛ»*Q_̧RF§s³}­iž/G>hmûÌ;?úh.ì"tÚ7Iå•ãûßår–=¯/›‰ñåLà-*’Í=P7ÚáYo°}M£÷{8Z‰+ 8 gº]†©A?DPaÅS»8s¿à¸¼•Æÿ™öƼ%íòŒvdݕŅ¢Iîš'Ü—Ž2g ã„°\µ˜^Ð8nelôt‡MÝßÔyBgGì kŽ”!p¿ÀŽ^º|ˆ%øêÞq±}µi3Ým‘óbøÈŒù]6ð ±Y `ãîä…ªW,ÜÓ%A!å½Üêæ€Gb£ëó«^çµ¼×cÞpCÔ½üÅ3.ìuÄ™A&täaq¸£/ßÉ.ÇÁÌ&yHXb `”ϱuÁJ—dílª…†6® ÏØV½M‡Ý0šv%†–Ñ›ÀHÍ g å1Î@ô~/Ö‡~L«Ñ5³Â`üÁ…úŠož|a׺kA¯9 œì¾ÑÎpÀÛŒt&L2Ýïéúl{¼†ls]\tÍ—ÏÒ|™Ù'{ âOø°‡‚•|ÿË;oJŸœç¹Qz|rî¦Ù]†wYfNèÅ:#ZçÁmÙâæB£mmÏÿzð^¾í¿š·Þ7tçþ÷CkH7ZoÄÝÎk¤Å›bËoVï60Eˆ’:–S‘ÞH¡ö¶[Á–æù‹ë«Ü@‹CùÇÔ"+ÞøâÆD¾÷åÉÑ_ŽòèDï|sà5ŠŒùõ5Šô dÁ®”ÿéÊ{‰µÒxB˜RÆÀ!6ãYú´o%@\[­r€fà,/{(..;£Ž»C®÷\•Yø±ÞRÂëCÖA ë“9}JƒÏvk…P[ÇJÛƒãÞÔ ¯mY‹œD,^ZàM›“žRâ…äM¯óW€e? œy¶÷Ý÷/·¹£‘¾ÇÃmu ¡w›w½œ¿ò@V»ó¹Ÿ™­©¥FѦyÅì F;C§ïKÌugÏå@AYÓŒÑ{_ï\EàôŽUŸRu»ö#Ò.ÜÄ¥Áýi±:] ¾?VW5â²Z>)VÇ—]ÿ‡Äêî)¡úˆÂ…Pø?£ê!A1aœûCu\COÕ#BõÇPÅpÍc‚¡Ìêöþ2Â~*Ú+îöçÁÖíð‚v7Ãâú¶ÝàVG–žü2I3¾º†ÃWÌžÈXu?cƒ·flsyßËX Z&k¯lÜ.Á+ð,÷šeÉ”ËW½³|‡ÎWU˜”¦ØñqG¬Æ65yÈ kó»xæ qËW‹»ÚÃýðD2yø¾MÇŸÅðø¿‡½Äì°i[üe›á®4…JŠ–XŠòæñðl ÷ =wé–…<}£z’U*ç$å²–ü­ãÚÉò4y`¨ìŽ®gÇ\î ºá ¢Õ€à¾:òØcä ” èêêHj¥`‡«—G/¡™ ŠÄ|9ÞA¸ªJØXš,ÿøAÌL`v|ñ”ž ü2-E‹ô€¢ê-–Hl;A¯"Š÷l×AÃÒ6ñùg7R“ÁZºóùI2€–[kØz+ÛdÄ{žñ­ÖäÊãÁp£ãJyu£Ë³û!ùÑ ò3‰ou4 ²×íŽ;Ý¡›†÷c÷t ^/õlÙÆ÷&bÁ¿òQÚ™T$OžÐ:Ö\Г.µjÆ’x¾oá®Â–ç–OBÏ´Ñ¥"–™LÝB:¶ÌÇq$Ýí=lKbÈ<_J2Èq“p©ÒbÿVÒè½åÔ8x.àÝ Ó!ð Ìû›Bì¢ßI³§¬ºÐ Ó2ä{9œýeNf©·á›)ü=—½7SôKôÞFÒçÿh•º.{>0˫Ѿø÷÷¾ia-Ÿþ˜ß´xÆÈ‘ÈâOêò/´•ίÁÀŒf¡ìa²óQqéS³ج×éì»ÜWï%D2½`R«%zÓóY+^÷Pßù-íùMëßXñÀ!ôTçõ`Ä‘ ˆÁh¥Ï[éO­ô²•¾œC‹Ù£“{¬Rýܬ²–>½GáG6I7èwƒÞ#MÿºÉÀ¬ÈŠb†Ÿï`©É{ÊùzÓ™ÿ]ñÐU•ËU™ð¶ýôýÂOÆÌ<Àèá= üÐJÔŽY8q„ltôË%ÄÍgdêô|^Gɇ6T£†ó·µKÆL\ü÷Vzø«Ìý¾…³ =9@X‚ šMhRi ûX¨AQ+;A0Ø"/Ó  UÄ ÜQÖ“ÉŒý„3¼\Š~qºÇp40@ð?v¨]ñ»€PÜ›|ÏßC0lËmG$5¿Â·ÍÉ\ ÔZÓÕßòú®Q†©!Xß×_òÿl Ь‘ÞÛÀ¨:¦ð8@°©ÓáÖêù\ï[®Ñôãc6 Ä(ìŒè»<*?ˆ ã®ëè&=éÀ{õåW/$DÀøëlš›~öŒ“„êJW"ƒP*Gîº+× ´Kq·w¢œo—ê36oj¢T”¾#ç¬,ŠÛñ€¼ú2‹:â=|Ì{£LOâkSä Ÿr""há-î-ï ïð^]>–Ô¢…x·IˆV9ß ±Ë"÷R§«Cbϲ,”êtÁ ¿*XN…Ä~Ø„-g‘¥ùâ:÷d‘ý¬ªµÇ4skI©t'I‡tÇçÊ7@hÛÒ"òå…SÒ«®_á¦ÆVúbL]=f·º®ÝòOƼ,‹P6[ æTÓŠAUL×kêmÀsïK>øDÄ Ï@´ÀÓ“Ž‚OqùZyü¡äÿþrô?Pmžendstream endobj 435 0 obj << /Filter /FlateDecode /Length 6117 >> stream xœí\[oGr~Ë×?‚ð‹×:ãéËôE°w×ÈëEVf^"ÁobDñP¼Èfìoߪê™^$z<:ìKuu]¾ª®î»}§v{ü¯üÿàýÎ7/cØ=¹ÚéwOv>ì¨ô×Ýò¿ƒ÷»ßíC eûÝØE§ÝîþñNîªv• ²a×…¡‚ÞÝ¿ójõížêú>Xµ:¯_›úuV¿në×Uý:e¿­­»Ìj[<¦ÿP<ª_…qòÌÑÙB ~Ô¯£q@µÚ¯?R—7õëŒuùiÿ_w´é¢êãîþŸwöûjõCˆ=(<£¨úãsúñûúã¥Ðû5ü}¾g¤Í×z-,ëŒVý5}^ ã\¥Å*ÕÅaPŸ±Xý¿c±RËêßoi _¯êç§9~=»Æ:ß^Ú›OÜóÿòÔž ³\±ùjÃô©øÎ~ʺQŒ‰Ñ`›÷Á?pèµcoMÜ]Ws€sí&I¡ßž‘ I>^SÃS³Á¤}­z¤!Lˆxy'43)E'yæ^7"RØÑNýA`öGéÏÇ÷õ¿N… ºÇŸý[ýñRR‰ßÕ¯ßׯ·Â|œñV¾[Ÿù®öÖVEÅ|Ù°·Ö%Æ•­cxƤñ‹þJ‚ÈÛ83z sxeqÄ( Þ ƒ“ÜÖÖ`}׫•ƒqú˜†”HãÑ×Z ^‰³#-Š\¨¡EÁ§QÖt½ÄEñ¥$ºh¾( â = 2ÜÉB’ǵ<Ì(q¿­?ò¯ÑI­«Œð¯ñ¯? ºu"X ú뱤Ëô„¤Êω }¢æ›—fÖ€ôœ Ù¦®¾ÚÛÿ¯:­£­î2¦C«¡sNù»†Nj«E—$¤¾”8ŸQàÌ}+ QIKÂ}ô‡ð0úeªµëTÿ²ÖÔ¤› `*Z‡;h}8©Ýçz·X,0³ïB¨ý¾BÍ1Iy¿*ÍÛy°yYCfRTÜÛe¬ÓØ&ƒPÇÚØ®h¹qlðJiÛ \Ö}×G°I/áC߃ ‚B¼ŠÜ6qjèÐqjrÖу auq¶‡?V‡<ŠÁi„ðœï#ƾó,+«ÑНFwÃÕÇôº¢_mSO£µfµ9Mtѵ®uÐ qu‘ÈŠQ¹au–ZÀw\mq~Ø. ü ›Î*a8\xß¾Ó¡ åˆG }ǎݶ&/6ô=ð ,Ò,tèÈÇy7çyX·ÊtMnšÆ?┟ånAàƒŸu °¸#hëT¯¢^eZuÔèì¡.#ak °Ææ†0@ð Þ¾‚×胰_¢Ö{œˆ°Þ®¶—¹Ÿ ®]\%˜3ø&õC=Hã–`wn•²Ãê¿q¯H€Ë*-ÎSOoô‰Ž¾÷°Iµñö™ œ |­ŒŽóʸιôsT pËA zØ^»ºÍËòA»G\ äæM 9f¡ÈäEU(îd…1ú8ï„Nn«2ÒÌ©ŒÕ™‘…¼–‘S™2¶.3ºz>È9#»2r"€å sPXÐBð/jû<ˆU6®þ´×"^Ë Ão¯†,›d>ø;|ƒ¡°gÂ&Vy¾c¼õE’I×çñ@,z48yJ Ü;ä1Ȳí}³.e·¤Ý…>P¾¦ÅæÇ ÷«w¹EP ãÇùÛAk¾ÚÜ †¦Áf\¹+¶"-e{’eÄ qh…²êb’0&`ð6°D=° ã7\lÐÀ½¢­°2"¡_ `#ñ_ ¢{Øe4´v¤³‡q¯Qsz°©«Ÿqù¡9Ù®3óq­ø³CNhè5Z²[âc7¨Eô3’:&I%¨œúãTçí¨]´n…0š"!ŠV(b¡N a¦ÞÀ‚Ÿ5®Â¶©§…\áò"ŠL—¤ßö°rïFé#¥L¢“Œ‚ÒS+mû$îذµ3‰ |£Ô!s:Ù¨DãtÒ”Р¸©9g³/1Fùfnf@˜ù˃ H‚Ýö¶ß¯^‘ º„À·¡Ì·4<MôàáGÿÄ$óU6RÊ/ >¤”{~—>¬ÍK{ < Á±yÀ‚T3á"ØV# œ‚ÍUºÚó¿U¡ù¶~}]¿¾å0Žw–\ÂUÞR`~rÝEAÇž†÷ô°á¶Η3v7ó¨„VÆÆ­ï—‚ˆrþà†LóK°NåÝmìÁ„GÍV8 =Œ§%)hE*´ }]Óš`H6=”Æ|$U4W ÓøÁ¦¼Ûèí¤ þfRœuu8’@QÀ¥:‡­÷€a.·­@ Õ)0T­Ã]v+(Ï@Y@r,XZ]üŠ‹qæVæFçŒ L7Éø)ªÁŸag8(KkO?ë§$÷›7qèq ¹Є;«mƒ_ [sy¾¢‘9„;Ì‘L–ØÀ„ɹ‡Ü^ç~±xÃÒø‘Gc[[°qÆÁ`ŒiÐl&s0~`h\ à ò¾š·3AˆO„ÄíûGˆ®¨o«3DØ`A@‚]ý(Á‘ñ 8Édüqç¯;ùXàòîåg‰­Y+g‰v°¶nבÍg‰ÕFq´w*›„F ›.Px¢W…æGÓ àÕ¡¥ø cG D=;ø)ù4+‚Ž^ö)”Gqk)ŸXÏ' 3 JM ŸC8¦GÍD4𔄱|mÁC¬:±’­›4Æ Ÿã!Jçàß³Än®§@às ª•Š~©17ÂÙº˜Ìñ¹îü)œ¿ÊÍAÐ߈€Åð,vnÐ#.iˆîŽÐ®LÆÏ§úŒ¿®à…[—Š£g‰…dжuŒ”Fr½“6‰¸6û’™å8ÂAGNö*s# r¾'áäÇÝœ#¨HæHêxÖ0¬þܵ¶`­xôÁ‚Ji°E&d®ì§Ígb}¥¦[g¡q€¦Û`'‹Uðn*C<&I»R­ÑFŸå¡1ȼ?–Ûöj˜dŸðWh¾º®f#aä,üØjàÅE €äÆ|³¦,ÜAõX‰KÒÜ4£1¶ßæ©}ôK™±74ßáƒ^¹e‘Sn X)›£‰AÍ’Îi#Ù×¹mäVîÍÅúoO"DÇ‘KÂiV+g;;Ô© ]§8­•†‚éúäX!ÇŠå#9ü<¹ái*ŠV㠫ׯYgž)c»Ó¦ï*_æQ ö™åìp3ÒN¬%ÇVYb¹ZPIfØìY#=–ÃÄÉå,²kã0>µ0m‹1}I9íhƒæËGœ1@kdà†u+¹2›0¬ ˜ó²ƒÝ\g"bïÑ•J[*8Ÿ8°Å¼Zûs£öÏ ÆœŠKj¶Lêa:Û„L(²k)IÌãƒJSœ”DR˜¥ú“ÙŽ£sгÐ,ÏÇ=Q²v0!ÀT?`B0»‹®|äÑC$,å²u„¸²ª%éL±Õ®â™œ›=ɱ,¦ÑŒÛ€ù‡›ä#êOΛ–ež0Ï9±§u¶ªa9P. zÒ@ _õØ_:Ù²°bÔ ^ÁDÃ"ìÔѤc#$‹† ì-Zƒi¨Ã 1,êX[K[ÒIÐ=ÄTž°Ép•ü8ày85Á^ã$÷†Å äš³bÇm[fÀ¯ )~vƒÁ^}#+ZÁþ=0žÏÆXøúRâƒrÒ£ÐBŒÚq£i°]óC4²ñ™ozɃq_Â#‰Ð nÃÒv•#MvªQÇ`Öí]nú|d12_Ìoc•ƒÝ<&ìÁWÊ™OfF›õÞdAÖN¸6mq’œ1ø§ªRÊJØ”giAhëç`zd—7Ìã³ÅsÙešcü¶-BláÁÿ1èuCȪÍ}Õ@…'â7ãìaÙγCËqM’f‹Þäð0!1 H,Ì\~^ÑÉI£äâH¤¯‹8€Âmh vŒœK§±\ -)–¼7•Å%n÷½}lSbmÏbÿVÇÄ”ƒ‰E»R¬V½Öqšþäé›— ‚&Êg¬Ç¥·êg&í¸Ñ!;L®VRA£7ðŒÔÅ糈€~»@4ZYnæBKìUDH/kåR2;úЦ<ë9lI]p3Ä}ÏØÌ²ªm¢ôsüTt[d ü¡lȺ6·7M ®L?ø0 eß åbTF—¨êŒ ÌJJ5}.ž„0ÔÓŽ)™YŠBôb 'çjÖàÊPêp·ÁC!6ËQ±‰,8lÃ@ìg[¬ÊN©RQDj ­~*_LwyPÞO0,wÝHºî6‡ÃaiÃ"üÉ£us$4…¥ó(ý~|=òXN ló M¾i’™erälPS`Gª ÿA ‹6=±«wÅâBÀŒY¯l-u§èØÙE°i¤ÍñSÞ4íµ›&­Òž¨7@vú¬ˆPðMŒKŒ* 0mÿ²F>¯ÒiÏã„%Ç?-•(æˆÔÜU1QÝ$шƒ óAŠÀ×1'Š‹×I ™Çžè= 9›Š@ØR{¼‡ýM]•¤é—¹1pù쉛2kÏ#X–ÙÅRYƒPzXý [ŒNì:O¦}äÂR ¨Ç†ƒ&K„16rã6”.É £Q Ù,òhðn 쨖jŒyHŒSëœ kA¡æ‡7¨{årRR9楘Vï¤Tû‚¶m˜(1h^^«4<¨Ìç—Öºk²¶¹",e30÷ŠÙŒ;j¥ÆžY>snš«óÖ<ÊE‚‘šz2–o¸’_•RÄèŸ:m?VNÊ… bý V­™°ûàÂÏJ›ï/LÈÍ bw÷e2 Œ°dBù!bÆfž+x+úd2úoI`Û{I‡/gþÆúQžzñÂiä8rï‹44µ¦‚úv¾V­ÝS@;È¢¶–¯Wœp,’M¢„6⯀pv7"fáY×K¬âL‹ÉŽWùJ/ZH×"éfãµðu$%­¾2^Ô…øÃÿeýb—ã©óFúðWªm$göB˜ä£0Þ¡ÐnÁB뢩pw’R\Ã&Ôˆ“â÷b–℟çÒºÄóOdG`tY»¦º`²QŸaD”ºÝÀÇKȘNm§G‚e% \˜Öºòy’ûªóÎø‹Š¤5ÇcèÍ€î®uŸŸ“ÂN¿›¾5óÙÇ5Ô¼y>öd yÚ áyΨ·Ëæ‘“OûOAÙh”_$JKž˜¦»&‘4õRè!í¸øˆcY¹—õëßë×ÙRꛘQÉ×±¿.‰Ð.K;J˜ç…eèí1iIo…K̢߈ÑÏi•¯êOoÕǯŸ¤éžÊ,Ó$ÆF£Fæ­úKNš(góém[À$”ýåÒÕ6à¡yja‚jÎF(_¬³hJgj]µñWeäX¯Y§ãäËü³óq†dŠŽrs,|f)ï|”åÚ¨—ù@wñZfaí£/çD­okbPÏv<ΖU;ñÔ=ÁO ‹¦EI›Z¬¤y*ëm9Í4éVKqÅ ¾ñËçesT˜äˆÇR¦%¿<.¤¹3X±ÝɼP1Õs,Q¢å”¿K…{}:­¼Ë—J‘BY|‹~Tòx¡©½¡1®q9=4—ó®“úQK%4!°9üƒQ.&î•~<ÊU‘_ýUAîü¥0~z!×}‘¤ðÈ‘!fÇn©˜iÁž06ðXè nZx¨ðNÆ}M÷¦Ù[hÆnê×{áKB`ô胡ùª6˜Óëûa¿;¾Ø;ˆ¿sQç7öd$Ó»T[ø÷µÈm#¢*º”þôú%=4øÔ0ƒ×Ã;‰•—ÿ­0õÃ1]>wñ=‘6ÎáÝøÌÝØ´µÑ¹S®I (ÑvyÏóµ˜`ÉO)NPÌX™#½Ü(¿Z+YzÞÎJÔ=ýª…U'HÙ$Í¡¸®Èí‹Ì£¡¶ÝR¤!½ w!ü6í±¤vâ{¤MìUHé9Nz »Oƒ“lÐØ¤$œŸÈÄ«ë89è&ëù¶~ þ Ȇ½¤HïʩͿD«q,Z-u"‹EÏ~øEŒ½‘¶…MóðÇ÷H­Æ/Ç8!†ueò^Vaqß[Ÿ/RXÓåyBhÛ\ü¤úæå%F•<ªžÜ°Ä#g§sÈ£m¾L‘Ê},bî8‹Ü(L²åþoûF˜–\æ‘]|ÈÅ’…j½<_ M"šUIÑU©õP%(½¦UN%xE™;íl©Òà©ï‚Á7Ô±:ÎÍÞ;¨å¹Œ‹1¥Tâñ4ÏïºÞ›–Ì_á8ÔÃo&ði>÷¹¼·ˆ¯¶”ïe Æ5Q+¯QJêâƒ1m­Ñó98nž®ÂP‘,¢üÞ…hLž:wøý8HòB‡‚ ¼•Æz<:­ÑÌäH¹)Ä’ø'øié& Y1Ìý§ª þ¶Ú/” çms[¼íÊ"~âÆŽ¶¿I‰“˜+HsÔJg&:~N òâb]î?Ý}ªì;oCŸ'Õ嬦Pï¥*zAéŸç‚äšÖ4_Êô¯VöÙXá«&7Ýqéý¤xl¡~û(·6ÑL³QíeÒY9cåð¼¼`zµ¼)/ÈÛ—n÷JwiÏÑaþ)u> stream xœÕ]K“¹q¾ÓþòÁ=¶¦\x–#¼ÒJ+…¥4 ë@o8š3ÍÇî¼43Ü%uðÙ?Û_(T¢ÕUœéYYÁØe•H‰|õçU߉UOòßçW/þåÁ¯ÞÞ¿øó d´X!\äJ ­:gWÂ+×±ºÛ®þ´º~Ñw^j`ZáÿZØÕÝÛÒ(ß¿2Zø®7+-;ïÌJØK®/ÞPÇ€€ê;‹ÎU7üÔÐM@Uºy¹b°Æ‡ÎŒCxßi—a Œw˜†lÂjKS04ÂÍ&‚W¶Ó¦žˆ:JOµ±cµp«@´Þ •è‚×:uT½«^ˆÕoðß·/úÕ¯¦Hеï#ÉÎ÷#â–­ò_çW«/_bÛ„GßNôZ®^¾y‘öS¬¤5qMF….8±zyõb}ŠÉœ¼¤Q1GaV//b›¤6ÑÛ užHDï°º>ÁÇÖܘ;‰éz"JnõÔúÕËÀм â +ºz!­]?®ñk¡Òžs¨ÔRAåÝfP¹¥‚ÊûÍ r ‡6w„Z*¨¼“ *·ÔP:ÔkZ.·o”_B0°öíÕú÷'à?pŽ^o7w÷7×'ß¼üÍ Ýáí¼Xßmïß_|Ø\Þqò ÍJÓ‹—ÿôj}{úý‰D‹wb½¹ü°]ý¬ê ¤½A¯¸kss“bílªBé΀e¬79—¦úïw'§ÄrÁ‡õÃ;üƒ¸DbæïÞß±z{»úÙêþÝæâýõÛÿ¾Ú|ŒÓI÷ÄQ:âb×›N9ì¦í”ö+g1u&y„-<+¤›^u¦ÀB¿aGUé$ÜY[ –Žªu F˜Òi¸³¶’,‡åæŽ$!û#LÎ Wh—ˆhÂÊÛ›ö剋(å!H›‡«íõCäI?±„j¤¡©O~}u{w£Î–ë›ø¤„_¿½` ýÙ²/”‚zN´{¬–töï/7çÛ×7E·.ÇÊVà;ˆ–I‰Ó~»‰¢¹þ.Γ_?nê{ƒX/ÕÖúìæj{„yÛž¸*~ý»›ëÇ ,&ÏFxN`b! °¼Œ ÀÉð —ÁÀöÍùBSѺ¦~ÇŽkC8KO©ýíDdï^Âö'z Úƒ¶`ê–—0xÚ@›ïy:l“ZejDÓ¯ðb4þÚAjƒ®Œ*»@1ƒ=@Fv€zš‘Å’;1idg.•ÖX—™0—Î0s eÔ0—¿Ì¯û`Öï·×—ŸFV%TàÕŽ6󰱤¹üñWK Ïj´¬å«\1×CJ1Á`z»úã\äµóјÒj°C–së!óц¦è°¹±ÎWHYËW™#ÕçúÔ¨Ûɜǵ•Ê´Ä®û¶ÒŠ [© H´â`+Iþ­Ƕ•JšNB¿?¯±,£ÑZ8h.u©îuÛ\6¬åÁDX X,™auÄd†¬isnV*·îNNé9x!Ö[åÜ™õo©Uô}o5ø)A€÷×÷èû¬¡Á7x0Õ&¬ß„Á®ûÎï¡mî|oÖo¢í}B’AîP®/YóÍ ="âQë·¬ùrlæã€êpÕ½Æð©·34]·¾‰ÑaÄun}±%$ªïEP·öØN‡•¡Ù#V†È jJ‹þ6Š´rúþD#ŒiÖïâ¤`]œ[Óœ wÈÒ\Ÿœj˜uïëO'jKXß|Z%úa¦/!˾‡ÀÈmTp™þ¾>¤Fê°†¸ò”°š›·ñt€õÛ¯ó …5à ••k’8ú6´ˆD@i§£äú¢Œœˆ‚ÖD4 ®;þüPîÀûÈ$^€_âÐ;°§™®ºgtÅb;˜0*¼œWÀM¾Vˆ\7<½-OãÛ«òô_ëññ䤵uÊ"ÌQÃ?ÅìÄìízÃfw‘e¼©–²¡ é¡oë­)ï‡n ø5ëô"n­Âb ¶wD÷¶BA".1Zƒœe>—#ƒ=¤1?ÒbI,e ù»Pc%MÚRàk­ý|Ú6©¾Q¾ñæõɩј¸WëoÓ´¥×"ó—ºó¨7@ ±š·óÒÐ. „¢qàšÁ€R=MªK$ü©ƒ:dA±Æy¢›2 ”³i—44”±™á陫"Ýf}·—’mîh­Ï´÷ÎFÙRd*€ú’GÌU1Éü”_ 6B®lÔ+˜OÝ9Ê “<ÎJ·@J_ÉìÎ| µØâ³¦µ»'pRážû´åMÒªfïíØ,9ðMj  )C|·ZIeùÜX·w-õÑT’²Œ~à­ÛSŽ,ûð9¢±Á ÄLËEë d=ã²:Rƒ}p5 9-ßBöß0€»Ôì°wY´…ç{qé$IÇ&?]:áPøõÏ£l¯¡`]&K©œw™×­Ôde†…·5ý©²vAÞĕõrX7žîËÓ¦<½/Oçåi`€^fÑoÊÓÛò4¾½*O]£ÇëòômyÚ–§óòôPž&âHð°ÌÖG÷HD6)æñ½® y3`5) 9‚& GØåAÄ_lm$Ó­\d—õÙˆ#rPè©õœ ˜ð4\ýÝŒ–¹òt$GËyAùÊIg¶¦ç’À^®4ôB'oVwqHoA±0hhcWq½5öÍëèAÓ‘ÛôÔ à\(O« ̈l!±Ïpå !R«÷.ï–éC,¡|Õšv\K…2áØ1?£xþ1€©³ƒÙ‹òÎpðækP÷N–É=TÂ]\ÃÞﺆÃå¨îÇæhæ1X'3Mšy22ŠŒŒ¯ƒ{×Åä7­H Ý^¬È×;J„¤õ}yº+O_•§O ™ÿyCK´tÍ]S X8ëV§i$Ú&-‚´Þtƒ?mVø¨#“5´¬ÄŒ"ú¼¶÷£FaL|ÛT9ÜÓá\ç€Ö›7ymÎð›sbKÛGÌ@Дú6.N;=ƒäö쑎•Õ^¡ÅÀ7¥%Þ+P‰)8Tj© 270¨ÜRAev`P¹¥‚Ê[Ï r ‡¶y„Z*(i}=¯¡eÉá{cKEØRŠhÉ¡–Ã黦cð^ðÓwÄŠB”½x²±NV×ÜèÈî,»æÆ»ÿ+öZÊí©°Kð X×—0èxö(E>¼º¹Ø^~±zõõæýÝ7¯¾ú”8n¿yu¶=QñÀÔ­?~3sÕ-@ÅÐQ€’A‘ Åc\+À(ÎŒw£žÍ A7ÕØ3$)j 1 Èniï=W8â!U§ä Ê[Çît͵p¦ dX­H¨Î)8†ÓBˆÁ<0‚b\½"qÕ|ñóp œiøÖÐlu oŽãôˆE €=ÔY «jÛŒ˜Š/~°}y°ôËDËýŒ‡JÙ`pœvá™ZFEYˆ^ÂBÝ`®iÀ}¬Ò÷ØhS€¡œ:éÔ>Ö€ŒÄÐõ¤ÀTÕ Ó',ýÓá凞GÌÂíã&P`ƒ…Ò@z°º²õ(]³ tËt÷ÙùJæXó 0ÝÛСua5xm¦`I¦WÒR.ÿð , mÆúåÙ ý0ËNUªf°Áb.ªõx‘BIW÷- ^ƒÅ¦X£ XƒVÓûÒI8ÎbtŽwÌ©Ð98tÎÄŸÀ Ž ë‘Ãæà8Óø³(§À8siKWófg9 ×À93Ë90®e•OÐèPÂÄ{¼fWÄ!Ð\ Î6°¦ °JÑý™Ð@: 7q³¼°‹ñÇl0 È¡\Y’GÑвç†Ðt)M—žP$ÒÛbÉÝzT4hEžDè¡EUB6 ×À™'0tía3gý0Ѐîhí!‡kà̘G: ÈÇtŒ¹,í„]Yã(¡½‚zòÚq}88ucè9P.÷Ä<¡IUcˆiÀù[ˆ­`Aˆ„+ ¯Æäc/xÙ‚,‘¡s—ê¢×”§_zÑ vºYGžy apèå)œ¯ÎY•²×pÒéV-¢ÕõÇå·úØTÇÁ¡zh2Ò•ÚÆ¾ë//o®/ßÁka°eVFæ…uÌ·ð~y"é]o¸Ú\nŸ6ï2]ýéøo ëS¦]ˆáèT"ÝGüãöH¤°ô›¦g%9ÑÒ™pÍ ¦iywùùú‡ë#Íšퟑ.ª×tu\º(o&qÙœ÷¤E_Ä(=pÂÀJ¥Fëñ'Dp%á³Vø¿Þüå㺫µ½|~È B4Àð¯u·Ý^?‰(±¶Ò//?,ÉÊ"˜M/œH=gRÁ Ç€Dz/(¡3— ž‡ká<œ ž‡q¦Ô£®s¼P{¹à€-¬,Ë à½lðÀfþKõXbï넯ÚO ÏÃÍ gé^µŸž‡c»7°O•ìûYá€ûXëd¯ØÏ /dXm¢á*Û§Pg…çáöqVÉ^‚ÝÍ /ä2á\ÌsÍe…6°Îe…2¬1¹óYá€M¶-ýæ²Â ,6—^ØÀ:“ž‡ã,¦]:¬?œo‡kà<œoãÌ¥éðt~–³p œ3³œã*œ®Mªù¬ðÀÖ™¬ð<\[fá©^˜¼½œðÀ‰œpéÉS½è¹—^È¥Á‰>¹^ÑËý¤è<Ü]r7–õ=ŒMZSA9›ž‡kàœK/œrø e[ä|®xàD®¸ôœË/|T®*¸£ÛÉÍ\qü5@•+–ð}/ÿ:¹bºRåâM cæŠUºnòŒ¹â2Â1sÅ…GÌ}®\qᘹâBŠçÊ—Y?[®xᨹâ.ÇÌ“ìyáŸ-W,áGWð™rÅðêÞ‰ãæŠ•P]äTqáˆgK—¥ŠÕì£c’¹T±CKöÓÐÿ)2 êNÄ´Ý»ßJÙ²1äÌö¼[Š3E¦Y*“CéègC®*¡!Æ'?Ì:2´ÌÊæ¢$‚ã0¢§è†®•ÈÜN-ã©™[º‡§ôÎ5]¨Å|¶ÓÆh·—£5A©è¤z'G[©VÈ^})/Ó¢KâU{­¨O¼¨¼ *·TPy TnáP¹G¨¡e*'KÔS’¥Úù˜,ÚÏ•ziŸ’+í­šÊ•Ò¯Î‚pO”*Ê7ÐæH”ÊX&Ò°LiLþNfJs0eƒ§”xó4ûÉÑ€g ¬<Íiöó¡ G¬Z"Ö4;W^á›ï%D¶°²L'€÷R¢ V„zó9Ñ€-¬39Ñ€+–_‘]ØÂ:s/v`ƒ³æîÅ.ÜÇ:{/v £@ïIlgïÅÎÃí㜽»°’˜òùè<Ü>ιôç<é ÞÎg?6°Î¥<6xj.å¹°u&å9Çy ê!þpêp2q®óp2qŒó“ÒK®ëÎÃ5pÎÌrŒk=‹ZϦ<6°Î¤<çáœ4w÷u`ëÜu×€ «³pŠäìu×y¸Î¹ë® VOÕ“Ül>s®s.Ÿ¹ð¬©àÎå3žµÃ¬Òs.Ÿ¹ðQùLðj j[ùÌ wó™TÃH û×Égš ~pGÎgjòVŒ|Æ|fá˜ùÌBŒ#æ3lj>W>³ŒpÌ|f!Åså3ˬŸ-Ÿ9ŒpÔ|æ@—cæ3ÌTòÙò™tÌX÷lùL*m%¬=n>SSI—ØŽ™Ð,,ñl Í2¡„æçÔsTŠ®¾Pà¡æù¿§H½Q’¶bŬx=.^«5‰F8Ç«ªí”Ë¥,S#‡5­?¥*R2Ö:Å©:Ün ZËž3›X±–|ÃÀ‹9}`ŠE“‚R;õ“Æ*QŸjpª$" ¡{ùùg ­ß)ÀH„Tö“ª­Që‡Va§X/w§¤e³„˜2T–xÕWµ «Â@÷krœOT"jÓ’8Ölü!=´ãuµêâRqƒvKN-,8 i®U÷ã¶Qd\ÑCãíÅôÊJ!¶ cï·X¬XÒP¨»é2tC ®®®šuªèž‘1;åä^¦’•Îl¢KbÖ,<ÂÇzcÐÔZˆvã-+7·IN±E²*-Æ¥ŽJƒ *,êRv-ú#@ô¬FççW¬Ì,u ìä^©ã¨<½ª˜+F-¹V9' ²©zq$¨P•ž®JÇfrµK§n•¸dU+«Êt­z’\‘ä©{*•k®Ù‘Ȧ6Qù’¢5Á]/xª^7Vpcë˜/£Éká%%ölŒ Ÿþ€ºWåé¶#f¶¦7¥¨],.‹VÇë(fQ%„7#UxÅGF¬· Eß¼.äIJžy¸j÷ˆèÛmÎMQyý^¹¼ˆê¦®;Ì!˧%ŸëTH*ç‘k½þ2&èù´>0œ•%ü±ÃÚªxº[Ê9ôÚO•rŽbUòwÌd±zç‘ äǦšœ"VÜ)ÔkÌSsJU§5Œ§" 7þ1*¶‘bÑœ/‹žýÏò$Ûõf}µ†ÿO…c5]êBŒÔ ÄÙ5“¬p¬RáXŒŽnërìŸS8–j€ÆŸ«‡dÃöU5cη7±ò+þåká¿`üz§z7}T/~¾aT&L1®Yuœòc!è$Û=±ïÚÞßq9~ˆuóáÁxÅâÂ*Ê|@iÄú»¡§¨,Oüý¸3ÜIÊ¢ ›ŒÈ¾”ßF3c{?T»Žʹ[}1:¼üèPZUˆ¡Ì~´åߟC‹wñ{ E°ù±a¾%îܰZ*UIýà²ÿbT©uiø\·t׿r¨/ÉÎzk•ª,ÕÕNý_Æm:ÞË¡ÏÍâøùRÐô‚1Ád•ÓFØq„½"§”£crƒ`VÃ*œJºIªSõ¯ÿ¥¸kWÂè‰î[dͶS —¦)RUå¼¹©Ù'§?^YõW¦£I%à±T“¶5iÇŠ³œ9¯KqY™‚Óú ûŒ·ŽïÆmV¶V ×£—Ñö—Ø×+v«ú&ºÜ¼©´èž±!‡T—âø¿k¸÷£ËÿC# ~×öFzÒ ØP”þz ¦˜4ŽŠ>ÓÒ‹´ $S79¶ÃÚ^PÎäŒPæ‘´ÍIQr&Ì,}è3Z¾L¿h~Ÿ¦ŒqT‹CüS‹ìþ4‡ Ä­aÿÇ“—²m~ °@õcËù³²ýíâÅIÏ;ßiX³¤çÓÝ¢/sTŹ“fÖ6¥­v‹ª½ŠõŸ}¨R—Ì)kF[Ð’>²› ñ?%túE£®çXùû‹òôr§êç.ñKqá/CÎ\ÉHt‡õê)}äLÑ7"èÅzu0>…iôEåPíãæ;ÚÍÝ)e¯U-VHÖ8Ax³_–¾1š.Ÿ]¹ÚSfû™‹ßv!ž°!ÔüÁ†c¦x³ûE¢¬ XøÀPO€9ÊŠDIRÒÁû:dˆÑΈmjÕ|ð+(Ð7VêGéåÆÏ§ìé›i 3«âIê&úëx"ˆñrRÇK©þ RÙ5©³þUžbÜâËq7GGdÇ%Nd?K÷©Œë³™ÆS:›P½Èºu8+PÉ?3”¾'Õk'Øw¸ÄñT(£Õ_˜ž½Ii5Ñ.¶›K&–?Q.ƒÎø|åÒÌÊ1+ÐÑI)v˜³.Â<¦Æ&˜n¦)þóä/GîC¥¸JA×ê龡™ŽÝ2eYz“§œoÚÚVMEv'ÝèmЕ°|ûŽ'E oèÝãEˆi¸Ë42Ežµ 5?bE' ô+_§³nP±ÖqM˜èäW£f¿ ¥ŠÿuÙÂïm#Ü‚|²CÃuÐWD*ü5ËYG /ä/± Ñš*}ýý.c®Œ{¦ÛÌ[þP?ccÇ$;)¶ýï+Qè¯äNX>ý‰ˆ©Òæc…ôù„³¦Ÿ¥±â‚µÃtë¶uó‡1ŸÉý€! ’¶6ôé8¬—.ÚñotâcU²™KÁ6÷–lݰϔ$­´ë*L¥Ë¦D‰Ío'KFÃÛ¾úHJ´Žt ˜d€æÚ„å’ꓱâÈïÃWå±ëìèìÃh ÇÜåg¥å²ǯ›P:>VÒëÝ/Hƒ·Í$)¹Ã–Î’è&¹mÇH×Lm®™p¾MÙˆâT_jÉ 2ù¡–"o1îô>å`ç}Ÿ)U¦‘ ÿ¡¨sendstream endobj 437 0 obj << /Filter /FlateDecode /Length 6527 >> stream xœí=Ms\ÇqwD×$·Ô+]²°¸Ïóý¡ ]eÇ´¥”¨’T|]ª%‚]¤xÉoO÷|ö¼Ýp.q©JÎöôôôô×ô›iþ<°‘ ÿKž\üú{ï†óÛ6œü|ÀïCúãäjøÝ1@på?z#Ìpüö åƒQ£ÔƒqzÔN ÇW?,~s¸ÔBŒÞ‰ÅÕ!óL,6¥u[Z«Òº(­“ÐrŠ/þ²(ÍÿÉåâ÷¥sUZëÏ_Ôæ·å÷³Òº+­¥µ)­›Òú[ãq¤Ò¨L/´®J무^ä!|qZ:W¥u×é{YZu’ÿ.-APÿõø¿8½Ö|8þæàøW?D¥}u©¿”Öu‡'–Ô_OKë%ÙƒÜêoAn}YZÇ¥ÕgWaÒµùûo>–Ö—¥õm‡í•¯JkSZ7¥õ·îÌÝÜ碡·£¾³£—ý9ïô­÷îÞJë·¥õMi•Ö«ÒzQå¾NwÝØÛÒz×Ù¼Ó[×¥u^Z?–Ö:ìºè°zÝ™ã²#49 Æýù,ЂI1r¥ef¹>d£ë&íb<\b¤‚ƒê-9jÔ‹×ØËcaK#„dÖÏFÁ•¸l€àJj„fL °i·‡KiÕè˜^¼ ÐÎE$ ä¦"LÄzF‰å ¨õhtOÁÄž#o¸•°di‹3 ¯€„ Œg픿ôY¡GÁ¬ËØÉ2Ö§‡]z„Žh›žõ³Ñ“‘_>Ï$2 Pžža¿dŒ{äô:Ä…ý€nǘæ:!œ8®%÷nÔ†Ðc>Nrv×{ “¼ «‘’ÛE˜CÁ¦ëL)“Ô Bý^¢ôý¡°@9—`_Ë8ODÇ„V ^iáì3ˆ óZ Ñ€N/°æ&2ÛNYÙ[ˆÀQ¸t³8Ëýbur( ¸céÀÌh «2v±:‡¥'`?<À"G½çFÃŒ@‡3»ÁqÞ1 ™J‚ cSZ`ò)vÝÚyiСÄÒÃ,ëØÖn±ŠË†%ò80.; ó ]:,äÚ E'Ù¬ë¸ÕeDgŠR ¼ó6Ìm8ì» 8ÞÕI@J`#.2?°åþn½sEè£[i%Ì#«Ðm›ïÛŽ}«–餴Ǝ}«–©bNNk¤š=6zTº*f0NY ê&rÆ-+œÁ…ãJAõÄ{¥¢ž¤-ÛæŸØ»Øé™ìrì~4fŽIëlâ¸uBGŽk»æ{<Ì+b#žÆ áFÏ+ºÕz"_» Y‚ -j|Ž›=àb×¥UãÕL6wÐcPò29cU\~u|ð§ƒdC§u|¸ÙYÃîYéhdí%¸0O†q°r:ÆÖ<ù (’]4Ü‘)÷LÓŸ§á MùÜ' ¶PÀâ•M¶Pi%Š-4ƧóZÄKÀ`ÅÌb‰V^Ðî5±”0oFçU2=Lû]¨£¬)®|4Œ¢“M’fÜÒ¡’B]`6¸P 3š£ñ d€g ¶b{Á¶¶„¨L&þ;4ŠÒ °*q^g¼JDjæeô7I½6ë:]UÖÑЋ-hàÄeµyÑœ½@÷¦@íôâM!æ¾¶Â<ߨÉU ”u±º®×ÇeµÀ#:4¯}朵­nã&qܺhþ—èoQú@BGðTY>Dwj˜2I„gEÜ}%Œ¡ìžÛŽ0 \PEFi»F¨³ “ZòK‰_JeFÔ·â}õžD°/¨‹~SÕ#m%`—RŽ[»M}à\×e]”¼e‰ft€P x Í8_îpk„¥¸ß£ˆrUÔ¡g$¸†á‰?XV¬©vÜ W‚«‘ëø×˃#ø]^湈Üáào°;M :V&¸*0fÚò1Ø9´3¯SAàñ´M„µ6Æ= ÂoÌÂ2™ŽëfKf¤H2­U7L’ÙY0¿ì°ÏÂâËbKO^ma—1ÀjK9–zöò¤oy!̵µ<ùŠnx:X{”©ÒVBƒü€[ÎX¡ÓÞÙ&¼‹R mL/‚8Ï©emÔLj9"]U\IÍ3P² nhñ§â·®°§’ìIÄ Ÿ7f¦¬CM]ƒÈùŽÌC§9³§2Zm&œzŠd7ÑOTÌä 8E?\Y?‰~ ¨DC\]Çꊄ7m´ÐÉɃâ·Nc¿á­šF£¥”i·9FÕŽ3NÔÿž KVC9g4ÉQfX"9ÏN€ìÑD%lÃÄ à ì9Ÿì8†y¦+‚Cg•\TI öXúí&Æ€š[KYÝ—ÅÕ]$nWd§ìœ…µpè.I‘·{sÚõ0ß9꯺‡h…i»š×¤¸¯JP÷©Ž!¼§ÊPv‡·Ç›ÞJÑÏ %†gHÿ̧2˜WðŽMV§h{–õ>|·w#û¼éÐÿª´Ž:“€‹™x›efTë$žÿZF•üÞÇç*äõù¨\ïôô¢ö}QZW¸:qÕ¤ªâUDÈ>öv îcuèõ«r‘žÔ±õv¯úúº3oOjn:˜{"ÜûþÝ“³ø‚[Œ³òP 2“~2£lûŒ>Wr|” í‰G½JðcßMŽâ—=ó»Ù9Kû)î²Ã¤7-ém,‘žOG³ÿÞÃË'øÐ‘øÓÇ4N ‰÷Ë~£4$J‘‘[RTu¸îzONêÍ•ÏKëõ#ô¿Ì»Îà©¡hѼ«ƒz‹ëË*'_ïÕÓº=«°Dýõ¤ókã:wÅžß ìWøý¿þúƒVpÿúƒŽ8íL÷ðБøO÷À¹*Îu§µŸ¬^¨Í§RÛ¸Ÿø¡Þdl]M52Õ[Ta¬‘íC¹^º¼éŒÝu³ t½ZKxZl½\Ժëg:ž½ë îIB÷Lö?úÍÞåõh¸©ƒ‰b+Üèʱñ1·y˜œkøîòsù:,¸q£„?Ã;5ùÏ›³áÏÃúsZ&zàFIü$ ¬5ƒÐ ±J ¯üJ5gGÇäìà-Žsb ÀÖûQØ Œ_æµÀH2èÝ„Ÿl R¡‘ ÄNh‘ÒzYFrgFË-¯I€7øpÀˆ-†ŸÂe‚]Ëą̀Q9z´ÿóìÆÄ¦o¿OY=rŒD„*ÜNZ,ÕÈq^ØV‹L‚m],Eìåx™ÅZX.ÞŒrá7Xbáñ꨽"÷r7\`è•£ÄÞ°±ÂÌ’•e]AÞ­)øÍ¿¥]©@©ƒå])@¹£JPRG¤7ž¥ŽËG²Z#‹ù ½0á%Çâæönµ>]ÝœÎâ]W˜8Å‘×Ë÷‡B ¶g±º¼?^…Ðxõ v ø Xг%ó‡ÿþ€/”“€Ö×Ü‘߂ՠð–š—à¯ÏÖw«ËáÝÙêòîÝpºº[}9|½>=»>ƒÿ­OÎÊþuÅt\~qü4€;2­îåëú2=wT4“cfÔø«´u0Û8gáN‚¦eXá0«ê·qÎÂu-f•&SÃOSôópG3|FŸ k 6ÈsÆvø¼Žòt j†U`š@Ä:Hç)§……¯X?Jï:Xçwð/éÔaÌ€¢ªmüóp„èÌ6†¿¢EÝA׃a»î°w'Á‰öŒg‚Uºèàœ…#8­fÔùÁ±áKƒmœ³p'0L¶,°FoãÛ³ÍGa-ì-0]Zà= “çjù¸Žò?®ó«ð7ïàœ…£|Tnäd~0\›mœ³p”Ü)°’ãÝJ»…s®+ö…e| 9+«ŽŸõ»¬Ý GØÀ= 3,Š:ômœ³p§³.œ¬ÐnÎwpÎÂí2…e» ”VŽœËγpûù,-DŸ@˜hh¸Á¯[°rZ¡1Ê°Šƒ«R=¤³p§Æ(ºKð) -ÜÒyÀ¼Ë,š§{p.ò½@ŠKÐ? ®Çá…:°|–£»ÕÝým =ᨠ£RÂ~‡·â=¬ï!\õ/¯Ü‚áèÕÑïzµ^™ÄHz´ƒJa4‰”™+=!Dùx¯1êçÁèal¸»PÏ‚Qrp(*b”σQâù#\š[ˆçÁˆj¯á-øc0Ò¨Y€ý€Ÿ˜‰ÒôçC‰7D¤]œ]^>JŽ rÍÂ1±b_¼¾¸<}½áz¸l)~½9=»9DÿµXÝ=‰j §²¿¾º^]Üœ=tŽ• {‡j¤a|”^UR ô®A¹|"Nð#£QhÁ˜>^h+‘ «Î5—R¬ú ÜD€Â`’Áçe?Jý鮣‡£8Ÿˆ2 ˜½Z¦”Ä“hdn„Є"¤IÌ/ ÏuL(íË/q˜[Yi04Ÿfb09Àœ‰î/All0æÎ@ÊáyFïËôìÂÿÔL‚x_M3=<Ê|Îéh;j\C›Óp×#Ñ¡—Æ Á)Å“ìr‚¡O¦é› •yQ¡rOU².êiiedãv¦]™7‘‚‡3Ù\Þ䛋õÙêfø%5ÊŒ(K£[פ~ðV½xJ… Âªá¯0ÿi¢ãû?>ÉQœLépö¥™ÛLUQ΀&„ïù6/Âå„úHm䢞‘+ù‘O’"v’7(óàà´¼72»–`™ìMœØ6Ѱ½–ÙA ¹ô.æƒIÅL‘ÏmË,’4™ƒsºÖMjʼn¹•ÌJ»ç(Ph’g1?i%óHÚ=i’0V4 ‹{²{PÞ†DÌdm2ÒYàÌ{0ðíKÛÌoô,’¼pŸæiÐß‘ÉÏç€'<'ÙˆšäÅ–µß œy ácXIMϘ Ò휔yΗæi¼œ"Ÿåù’Äs†z"÷æv¶V2?hÂû}™»à½iŸ00Á6îÍloËü¨„ÞB0¢Ì\ö'£‡Î;€Þi.áS8?ü¨¤„3±eú”¤døîôïô‘ ë:ØçKú„Ï–ô)Ÿ-é“1>_Ò§`|¶¤OÁø\I|®ï¦}´{rÚG‚¹e†?kÞGbØêÔß3ï£Âcóìy<°5IÒðñQšPhex{I48é‰öQË÷á…W“¦`â©‚æ0Ðt`÷ŸErA f} Â厅 t“÷y‰ÿjÌ{ÐäÔSIäbô˜  ©$Nò|úÄçÜpráC~:RÌç“ › ||ªÊÅsR‰¼äí>)Äwž\·zPY²ðפGº¹ÀyŠ zœª*¯o#‰/’¤Ó±Mz»NŠŽ5¯Ó0Àz!µàxí‹<ö›>+-Oñ©›ã©ØÚž":–,IÇð,®>a¿Ø终oßþÅéhR€ ½ý xºÅ<èŠð7NmãöRé!–Q² Ñ÷Ñä â<ÖåÙMŽhy ‚ämºfªGЉ,A³ ¥Îbµ=šIèñû²À25yüXv”š[J8Ä05½Tš«!…ǵ‘,¥ߪ㑠R§ ­JqQÍoèÛ÷Ûü˜Ó4/Ui™Œ7UÛê ñæƒ+ikìVý…íB\ãáR),`‰P<`ïL|•ߪî2JêÖ§²ŸJJ‘ÊÎ#e¤7j1çéý*Ä“ú7a²¦˜ÂízOÞÌ~¦gðûÍ‹Õâ#`ŽE+ì^bŽð l<èâ°xðË•QœbÙ0]*UË J˜L¬iJÂ’³~-à.~ý½ÐžÔƒ Uaâ ܉ âtÉhXÛOQg¤1¦gÅQ„Áƒ4ÑÎMXðåÖÞ–B;òú®ü_š²(Ó©àÐÊiø‹ý˜CËè.;[6€iESœT8 C~ËK9&dà.Q«ëÅF¤>µÀ4>ΦåChæüá4#pCóïZ¢Î´°M]7ìGf±zà'ÓYyûoD:§[ØÒŒ–™²SÜz³I¬µ3±èMÇ„P ±`†KÌÞ{ˆÑñ!9HºÝ[éµÚb4¿,»i;…÷¡,ÕíÄ®§Â#´ÂIìfVµOÿo"Ëc¾‰½ÞMjÐ qò}•­¶ë²Òéh9«Ð‰?µNVD&[p474¡¸!)xð ‘²X-%”¤¢8®«í¿KÔ».g•@\v[}À¨5µô퇸ûÀŒÅYz3…EXK@ûO‡´,P¡±6–X«4’b Ñ{ˆs|S)V³ EÀæ§"›fÎIX´krb‚KW¿Ì䶯ÀCöô«§ÀÁG¤ š”oÃnSâ°² ©PqOŠ…šCRЇ’pŒ[P¢„0ãm±¢UDqKsâY´åŸv•#a<‘Ñ¡IÑ¿¬×6{M˜s‰Å^<È„¡ÔC3&D"@›szZq5ÌÌ$Œ›T5Ù*¹L-×*VÌ2Ö[3­KÈcaÓjnǰ‚‚Èi‘&àBïèihE ­"–¥uùh¹®ní¢PÜZ¶£z Ö–B¤Å9+SÉng'Å ób'õúº%§èéxZ*ÌÙTxùyZµ¬Ôº&‘qÙ›:–'„]Oê–:3+ª\‰ã=!EÀŠã! #™ZM™Øéû–à «xõ¨tö¦©Ðk12tò¨¶XUzZ‚ù*DïLæ9ƒ×Jb;ÊìwPa«§¦%Š’l +šSp­Ó^N d®jI>¡Uk‹“¥E›¾Hõh½lõ’Ô$¤}Ö÷å¤,ЛX7QÙˆç »¯>ãîê^‚ynK¼WßöêwÕwzóõ­$(¹ÅùL þ„ÜáG°k‡,¼¬’ÚáwÇܳ·œªcƒÃjÛìÞÂ[ xAË)R/;ý£ ¢¶sÊËð€„ba”Æá #pGŠWrpéh6kLš\ý:D“œ¥dƒ€‹$¢ˆ`…XD>+’b¼›ÃˆIÐ(á¢xÇjo©Ö¦$¿«#ÉN’±l#5K—‚ 8<‡ó²å|P ×°;ú²DKüwâ/=ŽÀYÕ ]LAsY ‹BVdmü-ÈÛh¿AødtÀíHšð„Œ¼¬¤Ó9o åø\âßtxŽò–3>ym‹8Ôr‰5$p»àØ­#Ìš€\UÏ È‹„%$èxƒäò±H. ÄUCìt‡heV`Ùº"a ‡5"·²Dm¢, 'ÿíA H›HÚÝÝ Õ ‹·ø!q »ñ®ì]C+eWU)Ðüm5‹*‘´ŒèDRIÉ„I*é…(ªš(_÷i±rçõ”ÔÌ©‡m¢Õ”—÷S™I•§›‘© µàV„Ìͬ-f TÙ| ˜nFC‚.“ŽJJUËÝgÔä\&œ\=ØüSBÅ³Ž )ÜÄ®ÒS™(ƒÿzɰ„“¥Ñ"kɬv6ÆO0ÄfÅÎAwý^@Q…Ï#ôs ÿÉ“™Ýçvt†|-¨KZ…í2x\:AÎY8&à—䬆š†ó)'’¨ŒÄ ÷ü‡CRLà‰Å“¿œ|€ù_þQ¡•endstream endobj 438 0 obj << /Filter /FlateDecode /Length 5684 >> stream xœí\[o%¹q~‚üÁ >ñªÓ¼4/‹l;Þǰ 8#8kÃ8#i5ÊhÎÑJ£¹,Œýí©*²YEö‘ÆNó N6/ź~UìïO§QNø/ÿ½x{ò/ÿÃéõÃÉtz}òý‰¢_OóŸ‹·§??‡J«Ó8F§Ýéùw'éUuê㨣U§.ÌãôéùÛ“o‡›3|š¬7Ãns¦ƒݤ†|l§išÝ°»™F?¹‡+ì1Ãa¸È=œ ÃözÌÁZ§°KyŽ ÞÅá x3ÚÑ?ÜÀSŒfx—fŒ“Þâ{fšTœ‡íFûqš¼>¥>¸á4=,.¨´ ;šÓBMŒÖšá6Akxü÷b´ Z45L } Æ»¢5ÏV´Ûh‡oF^P€&FGm‡}Ú6ljؾÂI¢ž¢¨¼ŠÃ=Žëb°Êäq6nx/¦K».Úa›†öAÏr+ÄŒ÷iJ?¥õçot6~W"Èœ†€ã 3¬ouPAÁiŠÉa!0^œÑ©¤ ÷‡ Sá–ëJúýæç¿:1“õ쀻Î/—d‡[>¸]YÝ62þoé leð±·ìÙñ„b€½GÞÐzO§&ƒ»½Ì›UøÚaßÌ•*ù1¬v.ïN«‡ív™#ì4#Œû<3 Æ9=SjŒó¬Òn¿Ù3ªh4QXôjá*l>¦ÃŠ6ÒΡ­Â¤ü'ŠÊÍyÐŽ’oàm4cP:“Ã*ïu¡ãUšÏªPñ?ö"P.è…Çâ4W™–MW=Þ!¡,ÐYL÷&1­èu™1Óˆñ!½¦Õ,ØRò5Áã™ÏMœD7ÚâBì¨l8`Ç‚'y¸¢Jy±œˆ¯˜ ¬] lÒRó|(TÓ~—†pSößå6È7,v¢÷õ#ï…6šÕ’¾>?ùÝIÒçÔ[P§÷Ÿ«Ä 0ݬͩóftÚ£üæü¿—±?{<à³)z9Þ·x†°…iŠ6k먔¥h¾^bá‹xÁûÄD*l$Éhë±:|ýÇò E7 ¬Þ0÷%ÕaôÖ/B“fu­¬ÈÅ$wìJbH‡ü±Èšº÷,Žùpa)+¢Ü új¬dfŸžÅàjÁeO;ˆÆiÉ>È®Ó ©†¿k„fšc#ä4Eˆ+º8ÛOSp Ã/fg0aU¥§Ñf㫃~Zf@Wº9âY /Yè= Y•@‹IZYÃýÃúmÚ*}i­w™Ú“nØ'¬D øÝN‚ßõ :É„ÅT½ß(˜": “,­ëÒÚ•ÖUi½ë´øW°ÝKó'¥õÐyå¾´K뢴îJë¶´öQx°Õy»µx+gg–íÞuÙCÊF©‘I¼ÜTª5ËàCæ$8·äÎøz›}(ÂßH¦CWB5æsÏR•˜×Íð®FXª1zóÓã,£[é4â­Š“=äðâŽ2#f ?)¥v ’]Àv˜|ä_Ÿœÿó·Ã¿mÎf­VH"ý¤ÎÖÛÒÚ•Ö¶ùÄ&ó ¶J+s6ÏKë÷¥¥¹ã¦4ÿ±óÊMiÉùò²Íð¯åá?¼è­‚y+tøí‹åmÙ1>·£*­©ÏËgâxiø~E†åí¬#(üëûγ[ð»ŽlÝóÏo;??$×X Ä’ié¨ å²ó3+„ž ó€Ÿ6IphºZl'2×íKë¡Ã̧Ýóþ‘ùáåá¶´>ñÏ?åæo;<ý®´>”Ö¾´îKëMwÄó½ÞvÎñ‹‰™tï:Ͼ*-žä÷¥¥{Có|;gÆ¿^t&æ_/;Kø±´~Ñ=ù¥õegÑ79êòcŒÅ”I"•õÿ”›¿=jÌ>teai½´I,™çNL î-f3q»í0ÆuçÙ®´.Kë«Òú¦´~VZ¿.­—¥õui}ÁœÄÓÝuèÏ2øºCÿžÔ²7ÀŸJë›Ùn:$ßuæ`ÙÿÔšÈ ö;ÛŸ.YF_uÌýØ20Ñ™FêŒwßéÇë«Õ9Öœá1©_uZ=UßsÕzòÎoða_ô–p\îøeÃ/od³Hû^*ž*;ë4:à´€‡ï—ÃsmÇyøwrF@~ÀÀíÑ]‰Ý˜ÝÆ(Œˆý@xÌ þ¬NŽôµ^‹{Ñãã-øZ~¸»?–ü c_Þ¤9Á¹SÃú\¹Ï^<ÇHÃB44åPÃÃëù•à…).;û/Ž _‹à*Ç8œ|pÛ!„"ÇþBøpû8€û'ƒ˜öÉ‘x,NèþQÄQi`pø*<¥×¥X ?pz ;$„)…}d¶Q§iâˆÒ!ž^¥É!¯€©yj8?leàrË4’ØáJög‚9 Ò1M[Ã=Â<Mù ü¯fŠY/$±ÛímUŸÒfÀùɱÖÀ¨>O ð^PW ‡[šöUŽNÃÆX‘|—, h9“»øÈ0)-rå[H0²(¢–Àñ®yÛ ¡x(·"šIÁ½Ò^»*Ê¹Þ GŒ^½+Ö•1ˆ´¯&Òyd Bp{%2Iâ<£UCÀÇGUáã{:‚Ûb4âæÄ]Þ:hã .ü^ìk}‚—ñ)]‘aì€È|^?±mgý8ß²~e“pxM˜]ÀãØ BUÜräÉ’.€A#|‰š7aÛBzö×r–ˆúRp‘öeî‚öׇÏi<=&øã ¥€8îQ ‰æ¿âýt¹w+§ü”Æj3’ˆåä{Òò ¸Qª•ÊeKãUÖGlS¶¸ ÒU&[ˆ-ª ?ƒÞÝßeˆ8Áá`?8¢½”Ñ$EØžöÛH‘Ä ^p)Ìr:øHÈé=a_zÏCß|h Ë ™¦Að¡k€Ao†÷î8ßß²fð¨«U?aÅ}YZ¸ÄѯØÑÏÖV3®rAéÝ2UÈÊ@>_ .Ÿu°ÈÚõ$kÇL¾棱@‰Ò¶k!e»®èKùØ BŠÁhv˜S{—9dRã¬æE„ЉÅ€¶ª¥ÓÌqî$V€'ulu Œ:¨ë¢KŠ(4šh—ÄÛšˆŒˆ‹c>hø˜g±»Þ’Î Â:Ù‘‚0«Ê![Ú{-NZÿ~“)`7¥é|§ *cyó±Ž J°$\£VÂ÷Ú{ª]ßÔ=ûC‰8µfåtÿŽCŠ,Š“©mL0–óÈKq‰ ç/†¨¥‚”0;±€ÛäÝ‘ÚåÂÇès6•/‡¡¿ÆÈ%Lúô4˜_ìÐ}¥Ç ¤ÎäÇŸ“ç2Q©‘ŠxÄ2eA £àoÚéŒÚ%UˆÀ{seôe»ññ=, Üþ°<õõh²Þ$¯RvÇÚjÍ‹Ëïé ¤â@¢^‚Zp{XöY“ìSð@N!ú4£¢ÔÝÿR-‘Õˆ‡ÙSk°æ&×þü2¡=T–'¥X¸¨é HD37*F'WöÙ« ‘*4ªÕ`¤–qÆä¡Ê mþI†íwl÷FÏjÀ{“&ó‘5‡„Z7Ñ

zî©‹£šã"öœP¹;šªá”ŽHÚ®”}kD˜[–†!¤!˜åˆÈö w±@5’¡m®/Ÿúe¶ZC?$-Ø‘ªËVÄ2 ØES‚s…z"JŘ{ZÊä=ÖsŠ‚Û ^lTN˜:èïÒ4&ŸW‡#»ÉüÅd`ýx!i*=Ž#$üÜ…¯Ó`†0X`[šÙØ©4|-ƾd›™Í`ë; ›z²ÁüRmSŠéÊÆžíS¬î|'¤Y•óNÔ }Hñ 8SBCw~öåçQ«*ô Êg f/Ü›âK¸á6 Ž Äôº±MÆYÑ–·@ž¯ÒhV¦„¤íœYÏAoMOhš…ï(ˆSî€y™9ņ ¯ÂëÎ"¿$ŒÚÛ²#ÑÊÌøŠn+©¨&yÂS¦w 5I4GÐK·Ruµ.^k:Ð¥¾›c‚•¬Ìa&ü—nIwP%!Ý–ÌCx–t£1Mè§Çbqö¶  ?5±Æ„Z@i°âÓth0L„Ðóp’±Ù;¿æ‡Õ؇ä?i]TºÚ«"K\àœ¡3jêUTÍè¿A!JW"q¹ 1‹Š0-xꨯ ’ˆ‡eöò!uÆ,¯äøÖ&•’né‹&RDDð4©.^Þê;”’S¯_÷ÄG¬XÆ ÷:ëÙ©O\¥Td\¦ÜTˆ–0…e¼‰ä–Ý0%(Òȃ¿khˆnˆÊXsñ <Øa§„9c^%bMYˆ¦fÕV©½bk+½0JFF²­Y„”iá.•ñ<œCœd¿‚N·"pÛÚ¢ü½’ñ¼Z–vci;Œ;Iƪ 'Ëpœv]A#¾o#ñddëä~7{*܈ƒÐî—\ðyÒ`±.£«ÃG¼Epi‚FM²°(Ñ™òÑ4å„·{œ1 ˆLÛk3Ċ󺱀o&^ôF>Y/³Õ:H„™…áÖٹИúÎRbVÞC¥ƒj欳°’9ÿ §º¹á„;Œ%³þ”–8ÓˆqÓÈ[S[C6Dͳ u¤­È:Ƹ‡ÄL¤q~H\ÐÁ„`)LJÕÂ@‘…^êô{”j»7Êö‘`Œ•„R›-‚]ÀŒ‰™.ÏÙ2ÜœoBïÔ¡:Ûæ Ù¡-j€ûƒ”DÄk昌…q¼d-®…”î…^ç9)à6à&Ÿ~€¸û?N<8ú˜‡0xÓS™Ó·'Êh ±gyr{òzÙffB¡ÓòÀá½¶Ôg5‚·¶ÁÌ*€²0s-²&Fð_Ón ¨Ût!.GÛ[lOà´B”D÷²¿Kî£"‹™Éua²m=+f ñÊBÂ/™× }€‘ ì·<.¤›*Ê¥G‰Ò…5”AÔØÕDiîÂO®Mã°]Ýã0ãÇyšËÞ¹`–«8¹2ûuçWNu½ë&ðº©.pfžTÜ“…ax1D´ Ðó«´atꨗž"{¥ZÞŒt=l%Blx¤–Ä 4æTÛ¤§2ï;ÉÁÞ­‚Ë.u¬Ĉ?/ Ä="Xþ!Å"”[f³o(z;žÇ¼¯£.c¾ê.ÄB›’tÝ>¤ƒAlUÔä‚–äÕ‹DfsyýÒD6‘X»jªxSMœu(-á2¼ÅpWZ|³ïÝ\•ÖRò‹Í¸ ¬ýlz@üöMÒû®ÅÅÎðÎ3æ›0‚CŽ ¶®«+H}.¢Ñ~u«â¯¬ÖiģŠ“¦Ê‡ƒùÈV0ÖOS>tÙê'®bþ‘›ûìÑÓUU$ò$81!8§å«k¶g½ºÀë[)™ÒU[u#ܯ"¬îúãÚpLÑíÁmÅ„ ½=^­°Ïe>¸š#$»sN¯ªLÒ#ö*4ÁGyy5]d?å•èm {°ËÖÀšÆƒj JE–.c®]¦¨ü¶øŒéš&‡_8`ø1xøY $‹'<ä/gÆ^a]1[C%V¬«„V]ÊØ<{¯„ÂÔuÚËômh;¹J¶‹ðÀ0uR¹ûRÿgtžN Ñy¼rÌj\Es[%b%$[ºø;”Žù9T\²:̳‰ˆÆùg‘T”ëÃ4ÅîÝÔ;ð’ŽºË ÇÔé¾´¾X¡„ÅÚ¤k®ùäþ5×Ë…„UY^[-Q¾Sá(¹´Öõ¬úÊFJPˆêY ¦áì‚î-Å uâZTV b)¡–¦dO棑Ü4SPšÚf9æmÎÀôŽ‹‘_uüINXñ¢7âQ×CøâbÁ¤d^Kó)>£ñÔ÷+ðS(qC@RÉI¡EïÛ }ÒɇáM7üô¬p‡"‚Û¤¼‡,aœнí9³Ãrd•ðÚ:‰[ÔeG½’n´DåPˆ¬o&á7j•ì¥ÈKPf¥x¤) S5íVÜ–ÕUe ]î«Ð½»¹±KNíàK‡^‘ZuWJHÂ|À>HÙ¸¶úd™|¥òK| ƒM²ôv›©ãú¦  r§Y/Ö]À[«y²ºc³ý¬wÅÚ ëd>ãlYïYú¶B<¸Ï.âµÏ öšÚˆ¥Å_B¸ëØ)¾›þ¾c».ùB±¸!ÏèJïR³¸„|ÙÁönz‰¹¾â&Gè÷w^w~} ºáKÅe’ã—my±oÄ0½Å®ÝŽíéZK÷oÂrôZ"0ý9%ºLQëÔÂ)ׂgÎŒCwM‰¯ð²ð¦œæËNÇîg6~Óñ}ä—ÖÖ¯@=šƒ{õß–EëÒšJë‹Ò²Í¯8î7gjö*rŸL`&Äx#Þ™QÖŒ4rç¾¹Åvìzfwdfþt‰ªfNŸøˆbež×€'w°†v–z áY»k¼û¸éêªrİðÛßtÕÿg]ÅA^– ÷¥#ŸZÂë‰Ze»‹“çXc˜èýƒ—л¿i¼5·,ÿ¹Ïôe~‘pÇMÅÂ>ÑÄniGz–f]·ÜÔ­Jý,=:Ÿ'rsâyLožv¤çëkÖ•ú }mŸ°Çõµÿ+ôuxÂR´c?ËRÒv¨¬úX‹ 5x!Þ"Œ)_büÝÉÿ/Љïendstream endobj 439 0 obj << /BBox [ 2656.88 6207.5 2745.66 6289.18 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R274 171 0 R >> >> /Subtype /Form /Type /XObject /Length 99 >> stream xœË¹ Â@Ð|ªø Ç;Ç– 2Ž »‰¥ò÷N«eÇëYº¼eh3´$5­Wè’Ø¸à-Ô²9ÄÂKç ãøµr¥÷D¦Žà˜\žrÅ Äý¼ÉY¾7§-endstream endobj 440 0 obj << /BBox [ 3286.77 6207.5 3375.55 6336.17 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R274 171 0 R >> >> /Subtype /Form /Type /XObject /Length 99 >> stream xœË¹ ÂPÀ|«x¬÷ø{¸ @ÎÀΰû—0ÏLW«í ™H³'B+Y Z"\ì…ý…;Þ$Üvrëæt¤Éük,1 jƒçð“ÓJ,<ÿÁ7ºÐ2Û endstream endobj 441 0 obj << /BBox [ 2026.99 5550.76 2115.77 6211.27 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R274 171 0 R >> >> /Subtype /Form /Type /XObject /Length 99 >> stream xœË± A DÑÜULƒ×wc» ¸€‘q!qô/±%~½ºD­x}ícáÞ\•Qî,a¬\ ÇŽ+Þæì˜ÇäÑlAR0-º:‘JÖìÇnO»áÇã¼ÙÙ~4«!endstream endobj 442 0 obj << /BBox [ 2341.91 6070.19 2430.69 6211.27 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R274 171 0 R >> >> /Subtype /Form /Type /XObject /Length 101 >> stream xœË» Â@EÑ|ªxŒç³³o)3°3$ìþ%¶’»Ü‚ û)_ ³¡ÙQήpš) ÞF×$Ž7øˆéˆyHdK%Ñ¡V5ÛOjÖerÙ䉆×?ø.Wù/¡endstream endobj 443 0 obj << /BBox [ 2971.85 6156.34 3060.63 6211.22 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R274 171 0 R >> >> /Subtype /Form /Type /XObject /Length 98 >> stream xœË» Q ÀÜUl{¶ßó‡28‚Ÿ è_âè€|fY½&y‰«6G"¬’æ°RelvrÞ7œñeû>Ä5X†´h::¨Ñ¿MËi»–»l¸@qýÞä(_÷úendstream endobj 444 0 obj << /BBox [ 3601.74 6089.59 3690.52 6211.22 ] /Filter /FlateDecode /FormType 1 /Group 66 0 R /Matrix [ 1 0 0 1 0 0 ] /Resources << /ExtGState << /R274 171 0 R >> >> /Subtype /Form /Type /XObject /Length 99 >> stream xœ…Ë» Â@À|«x<ïÞÝ~\vàÀ"Ã&Cú—pä3ÓÒràù•·4Õb¸eÐ,U™ì‰Ï /QV»†ôÐÎ@èlŽrªWÀ,9Û…å”w(ÿé*7ù¼—/endstream endobj 445 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1Â0 Üó ÿ m…€¡êR– | uœ*C(M~IS†³t¾;ù,ûá2° ïÁá“"Ë:ÐâÖ€#M–EÝ€¶ Ëgå…ì¯Ê¿Þžàk ³ñ›šI>çc^Õ[¦Å+¤ x"ÑVU×Ó bý'•Àhг9u fÿ®¤h*±ß\C ޹in’ X¦ß3Þù”‚/ÄDkSpendstream endobj 446 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 351 >> stream xœcd`ab`dddsö Ž´±L~H3þaú!ËÜÝýsÒ>VY~'Yß#'e]æ°)ñuó0wó°,ü~Mè{‘à÷|þï9 ,ŒŒyÅMÎù•E™é% Éš †––æ: F– ޹©E™É‰y ¾‰%©¹‰%@NŽBp~rfjI¥‚†MFII•¾~yy¹^bn±^~Qº¦ŽByfI†BPjqjQYjŠ‚[~^‰‚_bnªØ­z`Ò9?· ´$µHÁ7?%µ(¯8373'±ˆQ‚±‹‰‘‘%þGßÏsßùN9ÍxlõOñÕÌ?'})ºº¡¿­¶³¨F®Ý9ÔÆ¤Û¿;z}ÀAïãa[ݺ9~³ûþæÑÌœÞÑ?iò¤%=òýoμÚ}¼{wú±ÐÓþ/ts|g=ýó£<_ñâŸö Ø~ËMgßɵ“{ç\ æe`×endstream endobj 447 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 1180 >> stream xœ“kPTeÇßÃYö—õx©C:°gO‰…Ž!jAš‚`^ÀUÒ5Ç"„UPwW–å&^Vó,ë-T c¼à•%1mU"ÆššœiÊpü M¦ŒåŸƒ/:¬EÓLŸšgæÌygžÿÿ}Îÿù†èÂÃ0úÔŒ”•¯ ½½ªF1jt˜jbÁó¤o`i¸‰ŒJ™c"‹3g›æÖ Y0êüÑáñcqç,…¶Ñ„eGѶTç¦rWÁº|·›;Iž–””8EžŸ$ϵÛ\¹99#Ço³ç¸µÃF9Ë™[`s—˱3óÝîMÉS§–––Æå؋✮u³'M‘K Üùò2[‘ÍUb˓ӷœ™c·É¡YãBÏT§}S±Ûæ’3œy6—ƒbpºŠs2òlk7Ú ±,²‚¤8’FÒÉ|’A2I„öÁDG–VFÇl ‹ ÛÖÏZغµFPUÏEuD€iÁ(P**ÀÉç6¶µ5Ÿð_Ë:³4aMší}Iñ†?æè¬íáVî˽g¦WX¹n_=ôÀP7k.'“8a`¦'€jƒ‚VXUV—ˆ‹–qo¬X™œ\Óô¹„¿qÔò]9Gu÷ÊûÏ~ 7®Ji.Üä,ð›qÔ_µ¡Óý`Z5ߢ!_/‰¿¤~GWÑ ›ÝRtöüÉæsgJ(^i_]ón?ð·Z³SÍŽNV¨u=PO¸_€Ì£Žk%û~ÛéŠþ£CpüxTf@}“™ Aq"5BµuõéÏnôu`¤O¢uÕT{ƒ N¿> a$«þˆ‘"Ý¥6Užª„bOˬHƒz(QgÍŠCÙ¼ÓªïõÕA·VuÐJÇÉB5¸}…>Å ¼Úæå°eð·åõUGaü1ØwxÏlWŽóޤĪo÷Ý… ZÝ…öžµI–qn¨R*|(lÕÀE=L»FFÓP,‘)Ãq/á–·Y÷¬žÎŠ}‰Šô¹“pÆ÷—››WqhŽX}ÈQað–¸zekÖ 8ºç+ÀgpBçŲ÷Z4jº=qŽž2x ŸU·`¾xô 4wlýæíK欮iàÖ¤é ”¥ù0ÓqÞ9$ GKÀ³¹ºvóvó†´h,ìD¾¶‹÷ë¼çwiÛ­Ý»é(o/¨mÑ}AÜ`p ò÷†ðÜ¡F‹(é³ýúkM×okÌÕ* ÔÖJ[ªvV@ŸÝZ|úÔ'Ú™Û™ø"±†FH“gZ{iߊ}Cëì ùkæ Eºp˜ÃÁ󉺴ª‡îh'<Ù­Í5Æ:„ö“xìјrrê oen,‘Š~z³a>Б°üõªb^ùCOSÿý¿\õµÂ9ðC \}jŒïpÈÞ¿ðukÓæY=ô?[8¡²AµìÇì#õ zúî>.h@!B2è#õF# 'Œ#½F?ƒî{]endstream endobj 448 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1ƒ0 Üó ÿ €ªvA,tahUµý@p”' aèïKBèÐá,ïN>Ë~¸l#ÈGpø¢Ʋ´¸5 ÁH“eQ7 -ÆÂòÄYy!û›òï'Ø dv~W3Éçé|Ê«z¡Ó´x…O$ÚªêZc:A¬ÿ¤Mq6—® Áì?”M%Ž›€kÄ17ÍMRËô{Æ;ŸR°A|AñSlendstream endobj 449 0 obj << /Filter /FlateDecode /Length 3560 >> stream xœíÛn\·ñ]ÉGi¥ÞÞi4ì\Ð Ð:êSÒ‡$ËJ¥]YZ9qòí^gxvvµ²d´n ?ˆà!‡3ùp.ë—û¢—û"þ+.ö>}üþéõžØ?Ý{¹'Ó×ýòçèbÿé!¬Jî‡>8åöŸïå­r_JßKã÷·½õjÿðbï‡îðÀ«ÞK¥ºåÁL…¡ÞvÌDoCÎv«2-}w§µ2Ønyuc©ŒñÝI\â{í]wIvÎãCØz¶†`ŒîÎâ´ôRÈâæ  Æ%Òë!èny J¡¨`‘†¢„WÒ&ć^H¡ºãÏ ¡ Öûî¡«8+TP¦{ðƒ Ý|‘iqrèNOÒ‹„…±Fut:‘èr®›ŸgNy kNé¥ãW+»£ˆ ×J;Jî|Wü)ŒÒ*³¦ðàÑÁL §˜î—åúÁ+Mo€€¥ÄWú.ºÂ]ä›ÂΨ8 øgÜÿ<üëžÖ¦WFƒpƒ(QÀ„/gqÚ =ø+x*à Ü”T˜rD6.óNš¡x]v¦Ï) ÅÕ™Boùû<)÷-\F °wQb€qÁ ,„#dp¸HÑl$i&e¬•™²9á?ˆ˜ÖѰzyž©Rª.K¬€*à‹¸SÁŽ0R,œóñÆ©å²Ðfˆ§‘¶(㻬32h `È"Y{P|PoT|%{emØ?üvïð“ºÏfV) [wÀá?dg˜¨]ãè?ŸÕÉE­ÈÂqtÊÌ]!˜ ;8¤£e]ÕÑI]›ã…s[êç+æó/Î8út+çdGdf€5æ}Áp쨎–„|†Ÿ®“3œœ3{†ñ7Ç †cȇ3†Ò»:üŽÙ¼¨£UÍæü‘aì 3÷ãAââ„uÏÞ³n »B)Œ:ù¬ëU\£eP°Ã Ù×UìN¶i`]\ò)ÎöC0ÅŠiìC_ÁøÀúbÙ… ­_Á@5'|€>šNSã—mŸPC2fˆoìÜ2ª}‘žÒ‡T0¶{Þƒû¼Af%Û Þ% Ét[lãý?ÈÖ]&ζÈì´z°-}r¤/rQ«©‹.2}œ¥oê‘° j£©¾AgI¢T°À'7zóçy©JŠIœòE/¨üÊ©º2€ÜuígHIöyÞ.¶LKmÒ;'“â«çŒ$T&¨ñï˜ó|?¥9q^î Èè~JH9á%Ý™œþHyqú!3àÃzÒ*;nB"9ºÿx‹ùbÓzÿtaAéaúÌ(êx’÷·qþ{³ÓælæùÄÖÅѱ„u3:èçÌ¿â!h‘ˆWF#†Nô˜1l縅Ò¼Çf¼<…ÛÉ)ÙW0ï·>[^"µ¿q¼ÚÝà×ÍÂ!Ï5nå3†«»>½ïxK¸'’€DŸ²ÆÑkÜü/ÏÈÑ)ìP é~´S×ÌŽ‡â1ƒ"å<‡Î}Ü}‘áYA·(xµÆgJõ/÷WÜü:ª`ðv™è2Oõ{]¾×‰ëºüÅ=,Â[Ñôu]æu~}ô.hºÓÛ4å’ °Qæ1¼!Áô¯wƒÓ¢MçùŽL>b®…ãì{AŒÆoÇšÿú»Òø‡ñÞï5þ-kü4¯ EãfÈ4Þ‹Ñô°eäÓ÷U€¹Ç=ލ½X×¢3†èƪœL±ßˆ IV/j8Yö M몗V ã¾W1“¬„läÏOe÷I0ÿNáU)¦uA[­®r¼$5Wó8])8¬î­Åäôeæ|¬`„ö ƒSB Ÿrˆ1 ÷2úˆ…öD Ì(ÙZ)mhÍ™[ë ºÙ…´¼oV,SæCZ ö%›§lE°Æç`ÔÇzŒl 7˜'Ù2LÕ˜œ«xAÄ×Ïó8FÜMò )ı¶Ð†Ä‹pGåN6Åø›–“%GMZ!ŸÜ4#‘æc➯´…&!Nšs2¶Â%‚G(ÉiK·ÚÉû¼)LmLÑ”$VNÑ ÆXQ²QZ÷~’jŠK½Ë9‹š3È’á‚TëÙ±ÌA’õ'ˆFÇB É}-óÒ¨tóªÛŽ Û;«F-ùèì,—ÝiÄTx­e0…$%„Š¢ÈÁ©!ßj´ˆ40ÝØà8ÍÔD\U{²ü l²®ÿy¢obØ”PëA‡@Ë ó{ÜG4hN{稓„—W™AƒP”—×9áRÕ¦ê6¢*„Û)(†Á¨ WxªQl#nË2 äwO¾! \uÂC‰aS±éiʉÚÁÙ.í,Ç¬È±ÒÆ¥“f û½v¦¸§íT.Á‚nær«Ã!9û Ì1ã…hõ¥zè2¹þL@_}Š[žÔÉo˜gy1<Ýu!>ðIpÄ=j_UbŽR¿ª£_ëhÅðùù˜}[G\%e5Ò"M²kaÝúãæ‚™ã^g+ w?\r=„×Û+y¸ù93"`¶—÷vÃQ¿ SÒÜ ˜@|­/„WÕú³±yêÍF(3¥áeR\Þ5>àþµQuÚu/µ£ 8©ËÞý]9yås¯åŸ¶òþqâ„U`òt˜8ß'õ°oêìXDÓÃWÝŠË·uòN¢j¿Ø*Ó$ü’HGü[D>éÿœ±%·8ŽÏ.Pæ )¼%8gð᪓¢ŽBa‘Ô’¯U0póLĵ¯#YG°YÚAv¦Î ÌzÃìôÈ_œ´HˆÞz®eF(g³:ôxÞ'u’ްöP›'”ur¨#ÃÜ„!wÂlö“QËKobÔÇ,T„ñÈNæ T|øJ³ìÕ†µ¸Ysœ{•m8.rò,ȈÙ“·qQlç¢gø8.Ší\D%áTÃ֗ƽrf»   ë¶ Ç:O°®¬ [ºeufWÖ9¢¯¬›U´éhüú=úS†Åøõ9áÜÖ3g Ñã>F 'Ê»@kGÞAÃЕ\Y÷ñÁáÏkOÛCôP“›­Ù´„°Ïï´„0Ô„r ˜‘dÏÓ²‡Àé6JH–éV"vòÁð籆°¬±TÓ’H\"ÇLÞT1)ºUû ¨ö÷Au;W7ðBS_÷}G'uÿ˜}Çåad €Ì¨L²­`¬Ví³ùë¯c'’ÊÑö“ M¦ì2¥slÐnÒ•TS0g$3ó!vIep¹ûxCzˆvLExpŠJÝ)°ÑÙ[’}¹mÚ&9 ܹ¨y(:ý¨äõ‚k[ƒG¬éá©5æ]ìV&Y©I³r\Úv©ŒŸwƒš\êë‘ʦe{UfCÓpMœrÊl°ÆÐ13 ½© ;L2ë¦Ç)[°ɵÇaÐ>ŸQi«f<¶H|¾¹|4›²’½ñ¢§;ö&ÕÌòufFĈ4ÓÆ(>Ë)ˆÇadµwaä‰Ñï«:Âu$lÙø7Q‡ê](Ê/ð†)­‹çq/7)‰6=†#Ë–Tò¢Ðê(´á™ØÖŒkÁnQÈ4{Ʀp7Hìq†á$íkÊ yEÌÆp4žzû- ÒÀتE›¿Öì'Åh™Ï¡é¤Úy:U³š½¬)˘Cµ:µþe­?Û “‰…ÑQÇÝš¤®UI(C×)‚dzÓ×RúÑr™a$PÈalü¦×²j!ÅË;sú“Œ&^Ϻ$Õ Ú}Â8 OhŠr½áiÓe[7IšÒÏšr·Ëɽ–‹Èý…¾¬-cùn©¤g±òƒ ŠAçß8»D\dž’Yµ¹øƒ‰!Qö¡ªé´æW|\R µÜЇõmÛ/é7æ;]X l«fÈ00T€^Ô„»6Ñæmb½Ê•å–4¸Rïuú"³Ã¦/ð™_þ®]ËiÛ’‹¸Y]oÏM…bÂü¶–P]uQóë’—„¨Dl’«Fè^͹U¸ÁĤåÔ EBBk¦¦¾œ‡rÆÿk•¿·ÆçIÂíœàC»kãZéM‹$)ù¼ÆJPnù-üi“lôwV'¥Ñ7—Hc-G#_~’SÎW©û¾«‰|@Wî—±b¼Ak_±'U¢¶ôµ$ôÜTNMºÔËŒÚ.õ¶ ¿L5†ˆ´áWQé[¦ñ?/âºsçLøŒQ1JpéBØ=Ó9ŽáÉÓ<È4g?ޏ¯_AX1[¸úÉguôQÝ=Œ{7ÔR˜¤ÔxaOFã '_3¹‹32W™WL²v_2W÷ŠÃ=`ÿ*S™·,_kH†‚LüHów;fiιô×cœüºN^1»_âÂßpøvÛȘlë«Þ bÿ·zæÆLâð[ÿÞù½E˜s†¯õÐφpV ÙÞs¦ˆØ§»AÌ™.¦FˆØ¿º¥„ÈïY7ñ·Ä1yü@5ɱþØò½üJBœOb 혛oTñuF­6ðŠ‘F÷ð@½Cñ`d†,UH½',_`ñFãqç ²p WplJ@·#7ô7rãU¨‘ 1\irdT·pÉ]$Ù0›ÃÿGÍeØ=‘~碋º½h±k%à}Ñå(ºŒõ¹Í5qßBÜ[,¹´-„´ä¢àÓ´ƒðü¨í¿ùµÿ;€ˆ¿ÿ6û3­úô’ꔋ÷ûÕáÞßáߥ™Ôendstream endobj 450 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1ƒ0 Üó ÿ €ª² ºthUµý@p”' aèïKBèÐá,ïN>Ëáz¹² Áá‹"Ë:ÐâÖ€#M–EÝ€¶ Ëgå…nÊ¿?ž`3Ùù]Í$Ÿ§öœWõB§iñ )(žHtUÕwÆô‚XÿI%0šâlÚ¾ Áì?”M%Ž›€kÄ17ÍMRËô{Æ;ŸR°A|CÌSoendstream endobj 451 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1ƒ0 Üó ÿ €*¡Jˆ….ZUm?eÀ‰Búû’:t8K结Ïr¸^®l#ÈGpø¢Ʋ´¸5 ÁH“eQ7 -ÆÂòÄYy!‡›òï'Ø dv~W3Éç©=çU½‡ÐiZ¼B Š']Uõ1½ ÖR Œ¦8›¶/h0û%ES‰ã&àqÌMs“TÀ2ýžñΧl_E¦Srendstream endobj 452 0 obj << /Filter /FlateDecode /Length 161 >> stream xœ]O1ƒ0 Üó ÿ €*U• ]ZUm?eÀ‰Búû’:t8K结ϲ®Ûò¾(‚±¬-n H0ÒdYÔ h‹±°> stream xœ31Ô35R0P0Bc3cs…C.=C Âɹ\… Æ`Ac# ÊRNž\úž¾ %E¥©\úá@u\ú@E\úNÎ †\ú.Ñ@3c¹ôÝôÝ€Ü}o] Œs~Nin^1P‹—§‹‚ÚBŸC™/â÷÷÷ÿòÿÑÿ'Wõÿ®¿uëÖaîl6.&.WO…@.Ó¶/Sendstream endobj 454 0 obj << /Filter /FlateDecode /Length 163 >> stream xœ]O1ƒ0 Üó ÿ €*A%ÄB—­ª¶Žƒ2àD! ý}I€ÎÒùîä³ì¯—+Ûò¾(‚±¬Ín H0ÐhY”h‹qgy⤼ýMù÷Ǭ2¿«‰äót.òªÜBè4Í^!Å#‰¶(ºÖ˜Në?i fw6M—QÕ5fÿ¡¤h*qÜ\B ޹in’ X¦ß3Þù”‚â G{S~endstream endobj 455 0 obj << /Filter /FlateDecode /Length 11829 >> stream xœí}]ÏGrÞýû+ruÞÀ<î¯é°6›]$vÅÀk_ÐEiCŠÅ]Ù¾ÈoO=UÕ3Õ§Ž(­m {1žyX3ÓS]]]ÝýTñ»K¸ÅKÀúççïžþúw£_Þ|ÿôÝSÌ9ÜFº”¾•Û/)ºÞö??¼¾üýå›§pË!]Þ<Åû-l—-Åx«í’úv£G\F‹·H^?}ù”oµ´Ò.?<ÅËoèÿ?<…˯~έŸÑ{jëæ=ôGºµC8xË1Ë{~†ž8––£i·0vá8Ê-å~÷ÄúìÐÅùûþA­[.åÒJ½…6ŸøÓBæ«§ðØÚ­m»p õV§v†ùj®¡–[>„+ÁƒT¶<ñBÇW§JÊÞòþA[N·TÒekôgÕ6þ ¡ã«wáÖé}y.±ÞRÈwOüq¡ã«§pMdıµÓwöõ‰ŸúÌí»§œb±MäíC©‰z¥ŠÔ)¼;#u¯ºwb¤î[aÚe¤邆8þ‹þñù»Ëß¾$€OÜn=o£^^’ ˆÌ%ný¶ÕxÙþ¤¿z÷ôûk Ï/rkùÖëµÐï˜HQ·t­ô»‡Ðn£_{xþÇ—¿yÂ3^àaqË·LÏÞ*ÍÒåA÷þ›×tK¦¦Ò-Ÿ½GW1öA ¾þ¯Wèb»…xý?Ï)П9__ÁOýå˧ß~â›k¥>£aúcš)ãÖÆÒc»Ô5ÅJ)²H5ÿ¹X)E¬T ä-He‡ÔD)Rä¬ÔH)²Hé)ófŸæÔoV6ÈåDôÄ5<¿„WNƒÜÛ6./¿ è– HîSÀò¬nï^2z1}^ßÈ.4ú:üÕ/Þ¼ùi/dQ}áõHóN{ùõ»o?¼†µ´º]ÿôœèW­e1‹ß’:r!kåçÎb¹•HªØ-‚,ë¸ÄTÈê·OÍb?ãÖóMáZ*i7ï‰×1[þ´ÐáÙváÃ>:MU¤îÃ> ww½}úʾøßý¬cv˜ò»z6êûFÓ™ ™¡~ñO NáQÂ.™i¢ŠõîiŒîæ·äÀù©‡ý¸Ìñ­%$š{Žx†ZO!C dîeÈã~RäøN/~÷TÙr;òö²Þ$ýúõ¤Csw©ŸaõöSFkúê=4Øâ ò•4Olj!?CÈhn wr%ïÂq#‰ZîžøãBæ{§ð¨¾c 7š®S¾{àʘoîaƒsØ?'’L¢?G¾•:må§…Ì7Oaü‰àE…K¦{c»{â ™oVášr½Ñ,ñÉ'~Bèøê-S ´ÌÄ[Îm‰!ë]bÓÿqÏ:tø—Û¾ñP˜¾ãF>©láé,õ3„>s“Ö£Ö=’ÚyÞ9ïcÖ°RwŽïÝX™»A~8ªEênX¼;#õHÿŠ,RëW=üû¢æLM¯ÙGÍ1VŽŽ3~É=qÚ¨´Ôºnˆ 3Íø=r4Ýë§xmwÑt&'S)*Ý2…rÃÍhš<E\E£é-S¸XÇOGÓ?® i>’¢5\_ gÚƒé)dbé]h¥§ ¥÷&ì¡´ ÙHzÚ#é)déû9„þíqôV­õôÿ;ŽŸ¤óFΞbáC ýIËOñ2覴.¬bûˆ”65ü—Ï/è*„QÚõ«gšFÙJº¾†Ù“îÇõs·ži¹¾ÿÂÛ‘Úõ ÷ûõkú{n#_?âg(4|%aZqƘp'¤·VÛõÕ[<»Ñ_yHoqà…¥ä+ ÷QK¼~o`¼†^i˜||NíF‹Ñë¿Ìgµë·­!P›ß¢#’_½¾ÿ¨}»þÃ"#Ð3vEÊfÆÔWþ\•¦o%•U¥ÿãý—üµuBÞ±~Gôáß-üFZUC éF$}ìðÚ1¬wR0…Ÿ[éãúÇãyßˤ1ºe{6Ÿ÷ç~F^¼ÿ¤xV-ÛÙßO#¡o“¿½_ùã;-Çlû> ïhÉHý/&’F*bpG÷ê™QH–ŸÕ¸¾±ø7‡VWq<º¤ZaZh\Mj-ÔÑ™Œù_ô+~}L-U+2M›U|{~AJ©I/ÄQC©Ò ÕvÒ2ß*ð®úAcÕ¼ž­?·ÞH" âÚŠªD_óÊ˜é ¶óe ï^$¬‘[%/Á޽Çï­îõ›yÈßÄ"ø¦œ ½:ÚqùÚÜù9Æ}Ò´0Ù¢çœc³Oûîø¶?á×òzòeÍd?2‚͸”Vn:på˜ñŒ«5™÷‚ŽÑU­[ a3ö¿>”úÍn^ïß|`‰A}G~‡;é5a,Íû+}#¹ŠÚo5FüOýBmz»ƒ´NlWÁ†Áñ ÍFôQq zü¾•¦®Í ¶¦F£«Wëßȃɲ¯¯¾5–õí1R>°/ Ë¢iCœíèý+íxŠd¾šFRe$Ѩ[_-͸ó>ÆL»jÿÔ^Ö!m | ¹³/Ì·ì£î{30?¿ÒnäC^Ú5o±£ë_&™GÛñü†uðb*AÆÔEfd,î=]–ÁòÏüt2{U0KZ‹µÓ#+0ôÐ+ùÌ÷⣶2f¬œ)¸¡Ñõ§g„Б¦öWFÀ £WÿtL o ˆR¶J__U2Ùl¼þâûÙêžÉœ¨ï²UŸ}€ñ“%íiwÓ8ƒãçÚy"m ƒæôʪ,PzèRµPÛêXÞÈï^ÙÅï¯{eÚ‹î£ùª3_%Œ¨¼‰ISßPXu?¥Q|Ÿ{êÓê_ýóáî}µ/ôû©0…•Ö¨búÐQxTaŒíëCµ2ØS-Ùª†ÕCo-Y“¶SÚòh8î÷¦IVäæµÓͧ¿ ÷Óâ#_hn"'ÑÀ¥5GB”Ÿi¸Ñj1`*ûbÿõíþëíþëýþëãþ‹â¨ùó×û¯wóáÁcþ´ÿzmš íŠ×ÿ{üüÿ=9:Àó×ëý×_‚_ìà«ý×Gƒí‚óèáÜóÕƒ¿ýúÜ}¿ÿ¢X =²;™ÿùôò?/ªÿaÿî¯(÷«GjþÅ•>ºåÃ'}`ßᅩOýãÍkkðóŸÔÉñëõ-šÇ|ºIºúó|¿ÿz»ÿú›ý×Úýð S5ûõƒ{Í‹¿{ œ¯ÜüñAc^²A¤Ž%K›Þèwû_ÿïý×/}½Ú‡óß=%Š)2¤Å\­8QN8¦­ÇŸóôä'öE¨)…|Ôø#ô!ƒ…âü”yÅ|¿^儈·•[/ëzu£öRd$ëÕASþýÁOʃ¼$5½cãù’-ÑiÑqJÜî}¦p¦‰ù–°»@H¥†É´2¹¼}J…fô™ÜpÎI3Ék2B ÛpW¢?+#)b³)ÇHþR€Ž­h:.Œ.hâÈ1ô[„[pN› ÔtW|´dã”ÌGŠ l0ÃÛ£Y@ê†ó=Bv¢4Ò)lˆtAŽÍ IØ Ò Îžs¨ ÊBê)©|<ÁH¥—úßÎ;:HÖ»(ðÄî_¥ß?y#õdÜUŠì´¤ÔSq}q™„•!yÓ/ÝH?29ª~°ß’ì¦H¥9HÆ ÒO‡ )¾vA¨=¶[—’z¨' I8’BêÙpuEI‚Ps`‰Æi'à¦UËiÜt<—”C „>“Eäx ´,È Ë>{“ìfFHÑ/ª¤X’? x@~A´ NÜ–Šîj›i-XÞÁn¡ õVpWó.ÒLƒ ê"ïÚ:!¤=y©&% ZBª)¸‹dåÁØ;ÀMÛ¦ºª``à&¨žÝ(Amb4XR4îÅ&©¦â¦B!@¿uÜ„s_þ¨A¶@w’X)i’NhXb@&s4ìœâ©øSXKVÔº õƤ 4´5|ÃcSÓÞnôµôý„l°n 8Õá»ŠŽ©FšI|WÒîní’ &Án=NÖ(ýÒ[ Ó.O&4)uþ*LÍÀ˜Æ1ÏvÖÀHÃè%dˆ&°ßÍ3daS¢ë9€€!EÏ$-Ã6 R<¿ðUǨ•¹šu@kõUvÆæ~c„&"˜Æ1§ñQ0#§_B*F#—Áx‰ri…†Ä^äD“Ù`™ 2Òá¨X;›H”ÐH•'§„#r YæOBø­¬A饔ù„T S„41üL‘ NŒéú])@õ5Uy¦ެ‚„)9Už]IÒfÒ¬ˆ*LŽ£ÍéÄñI £—_®AêÙº4™»†Œí"`Óž .€Í™/sƒ¸Ä%Ñššg+Ö ÊðF‰h™oÆ¢fé ìd àiš½ÁºÈ¤ŽIpH(í!í´,v°‰ …awõ"Î.ƒa‘pWÚ8Š,iJä ¤Š=ížÜ5ˆ« 20zÉÐ71ôLÚ)›DuÜš‚¶ã:ê‹ x› ºÈ^¢Fp¤ñ6…].f±É’øTš<ÀÀ·Á~‡ÄY2€ ©QC^]8€$o"Õpˆ”ÅÐd ×'Oõ„㘬‰ÃžMàwåÕM¸¬!H.=é<0e02༎i*Ò O'›¿@6jûúqŸµ6zš»vÔ†™œîêmËÞ`¾LZ®2{Ryèô½bµà¤2±9ç›ƒÃ’Ž²r¶"!\§FAH94‚)ИŽmS˜îj½Ë™mÞHO˜ÂHk“Ïh(q‘®mºÈÉiÓ@+%Òyh¹ho¤Ì&-iŒ™+i‡âO ïÀD‚8PeX95²ÍSؘ%ØÊUÆ?ÇšYn‚QÑüRACe»Y©E=mÍÅÕm“™•Fss8:ÖÖr2dHVg%å€nWÓì½JÊ!{àØ¼ ‚óVœÌeLW TXÅ÷Óß1I«XÒºôCí:$íßJº ™gÙ4V#Ö[Ó á66·Zã%9ì†-»¥€Ë¨¼WAºmUÞÜH1K`â[æn£ø"*‡¯ðJ`+iÞún"Ѫ[*¯Ë@XÁ"¦ÂˆÄpXÌ ÖB#½`ðniÞDj),%²Îj„ §èb `A±‘'™$†Ã¨K뤘Ìw%<:b¾kz!°¤ÀŠ´È]¤,)ÀË/ïY¢¸-Tu³¾xã»6irG„À"8.gaX$ʃI9™œe!—»,ª‘h§W â°d–ÇrB€#ÿÜÁöPQi'í`M$ŠÌ`žgè]2±ÒNÊö9ˆ™‹Þ%Ñ!Ö$MŸ,ÓÉ`» ìbp™%‚Ó%>#‡î&4u úð ,c͇Ú-ɵX\r‚ˆÉbZ‹¨SüV¶Cõ+M#ÅðžÐ®ñAŠaC:dâ·ã)%^º¶…‡i7í(dq‘f®æ JàãøìoH?‰LN2r@eA:Ï@x¤–ßd¤–@º!•1d·¢`«Š2º˜M •9BÚ|igã»t^ ‰vãY Œ",=² väá%2!:k–08ã]v¡¡aÃ]°±,HåUE¡à³ ™¾È»C¢1ÄPXU`nÜä9‰9B‚Ì’[c˜x—iÒy(Ôö*2…‰—„h0U˜{É2¼ïÄÊTÈÛUyé#óy™Ê~ƒ(j‰xÙSÙºöNì@È&.Î öÓª|…ÐŽR)º~*HÁL€]8ž©<£ò“›g#OQ Q {Bø~B6í‹”8”K`EÊÛv#pWêpÂáb)€ç*úQ0 d5“%ñŒ‚—* Üd–z@6(œÂ-Ø-Èr9xO `KÞCºa hS ÀÛKØùƒÃ&2†0ù éNtõš%æqMÇòE;òd@Î[û—âqžS2v©ª É!$•ž¢øž16š7A°Û„¢4žSëVy)ÛKüˆ»#M¬-RÆ–£•Ç0‹²bð<Ì´9^%dàïѰ¼ ¥°$¤ÀɇÓ‘ ÚÙD¦K$—±« ºìo¨K«Z-”‹9%Ǩc†1O…ËR¢™·I²‡Gˆ„rÅÊ~ÙÆ³…Ÿó9¤ 2hE1ˆîw7@žk~µ¤uà8ýFˆ,Àš2€(Ó¢o p—@zÍܯ®7€X‡Xç3€X¡ØP ¶l6÷a®eÈX€G•xÜ™k™°ÚÓ‘mü`÷`®ÅX€}ŒÄ @<•Ø™@üØ%€zMˆcµû^ˆ{6€xp °“7€Ìà©Â2›@& ðœd™¶ 3›xò3€Ìà)ôt–5€LÄà¹Ú2@f| pP`‰,Öñ¨Ñ‡$@±@Z FÃH$d’ ñ”8äÚ Ê̵„màÈÎû™k‰-À¤8Ä´×a¥šk‰c-À¡®ÒÒh •-ÀÑ´$à¶Çä a»$²·@Z<÷\@VÈ‹çžë ðRŲš1€¬w,ï?WM…•xíeYžY .–2yu Òj*º˜4€¬7-Ï=­àøÜ²ô5/Ží5/Ÿ +l {%+tðÞ^Gë¹ç6€x£À^§Å÷ÏÝ `?Â\ËŽ…dSÃqñÛskIJ{b´DÝ‚±@X|ÿÜÈ1€lõX ®CQ7Œ {JH‹ïŸ;SëhÔí-Șââ·ç>šòGÍÍ8È~Ââû箟ÂGͽÃÐíEȤx“Ò²i°øí¹jÙ/µ@\ü¶î¹ ¯a”nÜškÙÚµ6͵l ¬Q”î1׺ mÙ¦¶/Ú ÐÖ@jî‡@¶Ì »êˆ«±èÞ¼dûÞ¼Ão9°@X÷y”`9m°@\÷yda9Õ°@^…y6b°ŽH=`1€œÁX ®î_Or ‡=ÈKà>Œ,–À}<@Φ,W×­'\C0 ¬‘Ôieæ™Òâ¹'Íañþ“g!ÊY ®ãQév°^sÒõ@}ÖwOV ˜×a¯ãj+J>4ÓíuZME9Ž‹ûŸLI™Òq Ý'%ÓÂÚ´ÀIMî§Âj*Ê 5€L-÷?©ª6«Òâþ''Öa &³ÖB¾µ@\]·2x $_ ¬‘Ô¤ [ ,¡û—OÀ –ÄZ1ædŸ¬ã“u|²ŽOÖñÉ:>YÇ'ëødŸ¬ã“u|²ŽOÖñÉ:>YÇ'ëødŸ¬ãËÛ“u|²ŽOÖñÉ:>YÇ8YÇ'ëØxMœ¬ã“u¼'ëx'ëødÏë“ul€“u|²Ž8YÇ'ëøNÖñNÖñÉ:6ÀÉ:ž×'ëødÀÉ:žÀÉ:þ `ÛÑl§SônyÆ¥ord²óŒ 6R“iB.m'ÿ*ÏG&š›3{^‚M[}­AªªOÝ!IU·F+®Üju{¼(©ÃxFZ„ 9ô™Î3c»\9»â_3N:¤uê‚3N:†qÒ;èbYêÇyíÒ§Ïøç'» 26Õňu¾È8ùÐ7É”BÈ´ux…³™y)ãïvê"$é'èì–ñï5Ë2ÎØŽW±Ì‘„DmŽN£´ÒCÙ9Ó²©†u2&DÜç|MúSŠÙœÒ3öòí¤ŸñÏýÄn₌Ý}a{kè@K*=îšÑENEO{g’ñO=¢ð²k˜ †=Ežq!›ŽC …IrJ:£%B”Ì6ªŒ5T)ÅrñòMéÁ” ç±3l#DÏugdGHÔÞ”Ø/ãB”F‡„L£•ø‘€ª#AL¶Io–4ãÔBĨt=S#4ŠÍ(d/i‰s |Éa0/,»ðu%P&D£†Ò!J9–h›=Jž9!E-HcvBò¤KXψ•%ò'$Éqü\ð²V>Y×§.M(K |)KZøN–¿®RéêÉu!CH“S×¹ÖaDÞ,«!¦²uÁÄ+ja8隊¡”è²+ãHÈTº2#$+‹QoŒéI×w„$%éáYê*‘¨´]H¢š˜kMF¢]Žò¢_(k“_Œ3©®|b!ÇÞg{”aÁe¶KcF´=Ê0Ž]iÜ;Ã8ÂûKWÆ@D?“a{QÚ˜2ŒȺ] ÆØ•R²ò‹#¶±DDƲ2Ž…aŒc5!Ô(Ã@TN1ú©%fNŠ1a¤*Å8¶¡¼¸I1ŽMÙ“a @¡“aŒ¿2ÕÎ0Æ›´m2Œqp¡|beÃß Ÿx2Œ[™ÌàÉ0nY"ŸƒaÜôàû`·I»ÜÆ-¨õí cšqõÉ“aLž\< Ƶî”c%#–¹¼5übÙBkøÅUC˜ƒ`\ƒ[w‚1lp%ou~ù$ošBr0Œ©Y›rŽ•aLa—rŽ'øLíÎ0.“tµ3Œe‚~kƈ)ŸXÆ0(±áÉ0Ž t2Œq2ª”c!Çù8 Æ q'‘Q‚qD V‚±¡@+ÁÔó‰(Áø 9í;ŸÇÉžã¶©ñOŠqUîÏA1®šÝ±SŒ·i¥º—ÍÔ¥Í+Ib5Ýçü4q.ºeN­On°ìªõ+uã]òäU²7Oγ*{çI¹ß9ÆäÚ„p¸sŒû¤îOŽ1NKËÂ1“+µsŒ‘"N|rŒ1±HòÄäãˆTÌkrŒ‘³²)24÷iºÑI2ÆÔ'În’Œq2Z’ñuÊ(ɸӗJs”d܇¦øí$㎄8Vá$crW%wþ§}Q’1B„$ì[%ãiÆmL«4ãö_+˸aãžQ–qënO–q#¥ Ã~²Œ›¸µ·Ê2n=*Qu²Œ[2•M’qkCóX”dÜZ¿#ïóÉ1 J1„¬c EYÇÂ1²rŒñ&éÇI2FcDD9Æh®ŒÉ1†¨Òs•cŒo–ÇLŽq£)]8ú“c Õ‰chW<ä䣄Q>9Æx§ôÀä·ÑÔwL’qa2•d܃®v’qººÞIÆ0"1½I2FKÅ•M–q×UÛN2îiF“dÜS™Dd%ÃÊÅ^'ÉX?æíA2îY‰o;ɸ#cSîR’1>OYÅ0›.ÄdwZ))YÆøT‚q'ç-¤îI0îu2í•a|¸ƒÉ0î:‚qŸäÄI0†›‘ivŒ¡ e2+Á¸¤9J0îca >pg°IÓvv1!…n9ÙÅ»ŸÜÙÅÈ­ »ÞUˆž“]Œ|Te+»Y¬Ê7Vv1”ª-Tvñ˜»1;»x”p9ˆFGêáN-ÆZ¤.ÔâAscS¤ï“HD©ÅHFä˜t§ózEÉÆB-5*áqR‹÷ôÄZŒe¡ÅNj1åÃ'µx *‹—ƒéÅ)Œ<¼vjñhºÓ²S‹GÛŸ¬Ôb,Vjñhš±±S‹GóîÔb¤g xR‹a@úv¥c²ð¤#Í[訓Z[yÄÊ,†j„Ë®ÌbhOªÌâ=‘tg½ ÌbtÔB,F_ŠENbñà¡úrð59Ðz÷$ÙmÁ§6øôŸ"áÓ(|ª…KÇð)Ò:|ê‡Oñ)$>Íħ¢øtŸÒâÓ^|jŒOŸñ)6> ǧê¸tŸòó -ȧùô"Ÿ‚äÓ”|*“Owò)Q>mʧVùô+Ÿ¢åÓ¸|ª×}:˜Ë{”TæÏ\nšO_ó)n. ÎåɹL:Ÿlçòî2ö|JŸOûó©.}Ч>HCô©Š>Ñe<ú¤HŸ8é“+}¦ËÑôiœ>ÕÓ§ƒú”QŸVêRO}zêƒVŸæê2a}²¬O¨u9·>-×§îºì^Ÿì“„ﲈ}š±ËD¾ÏUöéÌ>ãÙ'EûÄi—[íÓ¯}жOãö©Þ>ܧŒ»¬rŸxî“Ó]þºKq÷iðRå}:½O¹÷iù.sß'÷û¾H€/$àj ør¾d/kàK¸ò¾„ƒ2 ¾ƒ/×àK:ø²®2„«áËKø®J…/dáj]ør¾dÆ}U WwÃWæðÅ;|_Ä ñÅD|Á_”Ä.ñÅM|_$ÅRñÅV\AW³åAY_úÅ—‡qd|‘_ˆÆ«ñm|Ñ_ÇÏñv|_¨Çóñò…ƒ|q!_€ÈÕ(òeŒÖ2G® ’/”t_JÉÕZòå˜\Å&_ÔÉ~òÅ¡òE¦|!*W«ÊU³ò¯|Q,_8Ë×rõ·|‰._ÆË—úòåÀîK†ù²bJùòd¾„™/sæK¡ùri¾¤š/»æK³ùòm®Â›«çËÄùRr®Üœ«H÷ h/lçjß¹êx®~Þ}=_‚ÏUéó…ü|±?WЗ ôe}éAWžÐ—0|PæÐ—Bôå}IE_vÑ—fôå}‰G_Ò—Š¼«%é‹Mú‚”¾hå}]K_úòAyLWAÓÙtu8}©N_ÎÓ—üôeA}éP_^Ô— õeJ})Ó»Z§®ª/˜êkªú²«®2«/Þê ¼º°¾L¬/%ëËÍú’´¾l­/mëËßú¹®Œ®/µû ¯/ÙëËúúÒ¿¾<°/!ìË ûRľ\±/iìË»ÊȾx²/°ìŠ0ûBÍwÅœQõùd]Ÿ¬ë“u}²®OÖõɺ>Y×'ëúd]Ÿ¬ë“u}²®ßž¬ë“u}²®OÖõɺ>Y×'ëúd]Ÿ¬ë“u}²®OÖõɺ>Y×'ëúd]Ÿ¬ë“u}²®OÖõɺ>Y×'ëúd]Ÿ¬ë“u}²®OÖõɺ>Y×'ëZ€“u}²®°®¥ê5yq[õzÃÆòåݾ¿­Àã-p¿Mî·Òýv»ß’wÛö~kÿÁö¿?"ðÇî¤ÁF¸ó ¤á=üш?>ñG,þÆÕøãwäs(ôàØÈ-¹Ó'@uw‚åN¸ü!˜;'óGiþ¸ÍÉùc;´çÿÜ¡?F|pÔè#ý‘¥?ÖôGŸþxÔ¡úcVëký‘®?öõGÃþøØ1ûcèGÕþ8Ûyûcñ»ss°îßý½?Ä÷ýž à žTà‰÷ä„•¼ðˆÜà Ž#áiŽiḞ®áŽóái!–4â)%žvâ©)޾â).h0ž*ãØ4žpãI9ž¸ãÉ=žäIBžHäÉFžäIKžØäÈOž õ€DåxVžŠåéZžÒåi_Žæ¸cŽ]æ hŽ£vÇaó$7ǃ»gÊy2£ÛyBž'í9^Ÿ§þyz §zš¡§"zº¢§4zÚ£§Fzú¤£Xz檦§szʧ§…:æ¨'—zª'©z"«'»zB¬'Ízb­'ß:‚®'ñ> úz2°' {R±'{r²ã/{г§A;¦´'S;¾µ§d{Úö=³Û“¿Ä=‰ÜÍ=ÝÖ=©ÝñÞ=5ÞÓç=ÅÞÓð=UßÓù=åߥøÔé>Á§)øTŸîà2"|Ò„Ë«ð©>=çpø4Ÿ âÓE\J‰O;yšâÓW|Š‹Ë‚qy2>•æ.ׯåâøt—Ñã“~\^Ð}âO-réG>EéA“OuòéP>eʧUùÔ+ŸžåS¸|š—Oóéb>¥Ì§ùÔ´ûô5Ÿâö( ÎeÊùd:Ÿpç“òî²ö|ZŸOýóé>…ЧúTDŸ®èSï³]^ä}ޤ˫t™—.7ÓeoºüNŸê²D}"©O6uù¨>eÕeµúÄW—ëh$Ùº<\—©ë“y}¯O ö‰Ã.·Ø§ûeŸÆìS}:´O™viÕ>õúAz¶ËàöIÞ>Ü'‹û„r—sîÓÒ]æºOn÷ ð>IÞ'ÒßeÚ»T|Ÿ®ï3ú}Ò¿« àKøò®/Rà øb®‚/™àË*øÒ ®:ƒ+àà‹<<(á‹Eø‚¾è„/Lá‹Wø¾†«“áKiør¾$‡/ÛáK{¸ò¾Dˆ­!²97µÏMísSûÜÔ>7µÏMísSûÜÔ>7µÏMísSûÜÔ>7µÏMísSûÜÔ>7µÏMísSûÜÔ>7µÏMí¿ìMmp¶ûôÝS¾Mdå48áÔñç‡×—¿¿|óD†S©Ó~xŠ—ßÐÿx —_a|q 0ˆ,Þa•€yý–^áDv a+.±„ÿéÚ+„Ž¢î=„&`…zãdæChFhmåÑj#.oH74]á?ýãów—¿}ùô׿£¸å‚(§Ò8ù%)ƒe#MþˆÞ„©Š¦ƒ—ïž~ÍáùEÁ °Œk1¿7þ½¹v­oáù_þæ Ï{óÃÕO°Sßå¡÷þ›×Ï/öµ·týìý»×|Ó/_R‡ÎsˆããîO*>{$ÓŠé$+“ý“BY¤¶Èû FJ‘EªcKÞ¾p"V*cO<ØgMd‘:NbTèø¾ŸÙq»’Qždc¡qOÅJ¦Ñ÷ü‚w)j¡‹Ä(BÅþMµÓõoÚ•Z|ôÊ{®Wv»nº;rz{ÿ \7$æw•¿‚@ÄLs µ˜Åt>{âR6-.B ¡­ð"ëÚ#„šaÙ°BHŠLËðWÀ 5žFŒÐ!> ²B !ì¸UÛ¦°B…ÏŒÐ!Þq¶B X¡ófŒÐ¬­—âÒ¦ !>q³mÚ+´ñTo„&°U,¬‡¯÷ªu“;`…$ÚMÀalvÓ-°m«ëÞ+ÔØý¡ !ZÞŒšöëE¤!²2 X!”u+4;¡t'sß)­`e0B%sާ™¹&`…°m°MÀ¡ K4#å!Þ‚±B X¡mÀ u>0BX„(ieäÚˆ`ÓØZ÷,B¼Åf…°B¹/îë¬PÝ–±{‹VF®­ÅËØ#„s€e(í€*¼z4B°B(a_·‹Ðê—`ºÓÒX¡Êµ¡ŒÐ¬P¯Ëœr‹ Y!¬ÐàÍC#4#Ô%Ý„oX„˜L`…X„Ò2vçõ"Â[@VFEˆ yX!¡¶ØÒ¼^DxÇÊ(`…hšÍ‹Ð¬PK6zݯÈÉ!³‹P_æï°Br¬m„&°q$+¤€ÂïEh‹PÇÆRÀ áXÈj`¬ÐH«ËÙE(/Só,BãN8„ •m›`â7+¤À"4WqV}n#ø0B1ðªõÚ+ãÒ-°­ó÷X¡׆c #4+4¸¾€š€Âö§ †`JëëvÀ aßΪ`¡¸¶i!>Ô±B X¡¶úÝX„ú2óÀ!ôÛ_é¤x§jW¨¨– }6¼i(ð:ç¿?‡[¤eXׯù'ùå|}óÇç8”±n×ÏðÁ±Æ~¥u'Ž©rO×þ_hÙÊÛÌõúß Mú¥]ßCâà‹”|ýÂkŒë·­ÆÔõ-^=HÛõýG‘ W_¿äçÒªòúþƒ utm2ªc\¿:aÚƒµ)¾«øNKSš›Ãåå¤ê_<Ó¢0P‹I£ó×Çý×Wþöëröý³¾témÔºÈ!ΗÊ{õ­”s}%_FÏY”÷ §ÔÇõý3Üzè×Lç~}èì›C oÄ*ð´EO¦Û¦ÕüéõnFm×WlV½ÐŠP­ æ’–VJ_†Fvã@‹7Ê—ów¿þÒúÂwÜÐÑG£nÆ7Urç¤MB[½fóQ‚m¤35GŒ´"©0W´³Æ&Mí e¶®&Rнlƒ,w ^$Š jÛ./È ¦µÏß°”há{½áw!Æ„{RÀÉ`¹¼(8Ý+¸ãšËóË?È®ÁoŸþ–dªôendstream endobj 456 0 obj << /Filter /FlateDecode /Length 6715 >> stream xœÍ=Ûn%¹qïr>Bp l+^õ6ïäclï$ÄŽí ìÀèÑ‘4³+éh%Ímüí©â­ŠÝì3šÙ‰ÌÃ<ì"Y¬{©ާQOø/ÿv}ôÅŸ‚?¾¼?šŽ/~8ñ×ãüßÙõñ¯ŸÁ!Åqƒ•öøÙÅQúT»0ʠűõf4^?»>úóðÍÉ)Ÿ&íÔpsr*5þæ‡{ìÖÓ4;¼:™Få…uv8;‘v´Â‹áEk…nãïN„á ZÂ+Ô°Àö&“`©´VÃç𙕣 nx‰Ý~š„ÃBÓcMï}šCäWuí<¯ÒZ?\ÄÊÍ_%:Óz´ÐaاÞ˜Ӓ½ º\‡ *&ئ™BžæN‹“´ R#Z`ô4©PL~ØQïK„ІfîâØàµPipÂçóº§W,~ßÁɰ¡û|Žv²Ãþ¢´Ãû—¹Nݸ8 ¡\³ë»“ÿyöoGÒM£œPÕ³ÐÐëc`SÖ sšÃyi–+"ÀoÃ|Ö>• {~y8œìjý&lÚÒdÁÑ>bÑM“·‹V´ã®”T×´ Ã=:k‹h ÍéÛEdJ#âñÃÀ«ÁôŽWªÑ(0¯O¤…tsÆDÝ.!dºé­H€¾WZQ‚eà+ñC‚ÝvjÊ=öª¨´è¤IŽ övãgÑø4À鹺N´ÒáÚ dš öùGHÀ©vœ$™ñbµ?í[ŒÚ¢2öž¼WiŸÿ|rj¤ƒ—åìã*Kë¶¶®jk_[=©ýMm]×Ömm‘ðß×ÖëÚ:¯­]Y—þFÍ'…Aß|NwµsîÐç\Šá+j>©¿ßu¾yÑùõegõÝ/™ ’º)ýïGÏþ©Aý›ŠÒ— äbëEmÆ‚æ“Jé<^t~}ÙG}÷=4>ZUwpjw°È¤Áems4–æçtl$²H‘xúª¶~^[o:‡Ú;Êóηlâj竘ó@Z̳d¸z6JU´ª?ÿgm}ÝÛý–P-ðZzâÈ ÐvL$DÕ ~@¡>‚´ZÈÊ΃ Å$ C Ü|]hÚO` ©G¥Áˆ™(ÍŸl ˜Þ1hæ¢ú÷®o„. yÄ|¾ÍiTÑ>úØ‹.wîæäŽhm‡ÍíÏwiž¬˜††²3Z7~L™]¨ -þø`úØB¹bÔmÞûDÖæaÐ^&Bðà~F ×Î7:meu·d%ãDáGÂ9ù4Ù¥‘¯wŸ€Á ¹¥Ã‘QV Äöˆ•ÞÄíW dŠ 'ü-³–Šõ€HIˆÒ£Ž>Žv@ ¿C“w BNëßG6S:À “²ýíZÓê >t2îVÆ Lì˜ll¿…¥ƒ(  #&ÏèAÇsœ op0Ø.Ò·4` 7î:7aÎ"¸ ¬\ðu%éyI±y¿xt°<Œ1<ÊÁÁvxrA€æóŧt‚YµAˆ«÷Ám’ùÂE‡9"õÄ…<3ŠaÝ,¸‘ÈAœäœvtÛp}%aËXèfë¼ZºâËH;¸| ÊÅ.’ ª“›šÃëËDq`Œj™än *7Y?ïrW;7` ~’2Û§°lkŠ+Œqƒt.Iì“#¹ä½ÂTSÂ’ aK­=†¸<É×È® ÇiÅ2Ƀ4%Ô`‘§ÀQ¬ÇÁ×óÀïò6 Þ— 8ÅA'›d¼€!0xw^SŽh„ì†Óçs$ @šI))°TE¶IópÔÀ“·ñ˜ j‰¡#hÂ@“E!¬CMéœõ'Pa_‚5ÎÞK':I‘å)È È>Ä9pU8OÝãžMù»auQ¼áhÔ×sdV­@:ÝìN(¦È1¨%®ÁËËȨÁKØçU–.ÚåLXè_žtC+°V(a±Æˆ4.ѤÐ9ß §kD7V;°i Ñ€làf³+€–Tz2+ZfÔT¾Ö€ùÍs`uc¾!óyËÛXÉ ò­t(3ž‘ѵoã¨uÊÆ9g6Ëå$M³3ÑtäÏc2 IîÕÙ›Y@oP³‘'–óÛYã4YðÜúg`Ï[®MÛn’´Ü¬‰.F€/– Yí[~-d ×Kɇ³•ñÍlïÖ¬ˆ‹hÖEá?`¬êkìi..ä÷ÑõlÇ´µGó±ly)…&ví~×`™~n§ÑÔZ8Y/lšeÕΨ»W1ªQ<“$÷3"Nk˜ñÑx´•{p7ĨÞuº5×Ù³ÄÚƒÅZ¸ýŒðÀåÙkÆ‹ÜËäîù̽AWà«a¹1+Û¤Exø¯î–¬cÆs UL¨Úªyx_6²§šÒäÒUOuœwØ®ûžóc˜„€è©ß%h®%Æ}ê„iC#‘°ec#mÄXô‹´ÆhŠëŠÑk–@%Ïó&-Mðý±øÆÜ7éœÐ=Z¸n“….×HpúOèÌLv2¹¹´Ë =†bJœî]êt ‹—yülŽ&(PÙf»<ÉE ­=µÍWѲº¥>cAªB·®†#Ý2\G zU¢ ;Y“è½äÔº±–ðç.µWR1EWW¼H\g™•×ãìsja Ï-ÄU öàÂèÿ€rœ³²ŠdZFei€¯%ã呸(&ªô0)|*€•õ*C&ð“ICtûe€3@±‰¸ñve¶U#vă®öWÍ!]uŒª‘`p|©L/ï–¼Wš6«FîÓÖA†Že¬„í–¼©Á€DŸÎiɽº¦"i»¸ §0ÑD«DÜ8@Äóg',Á#ò¼ N l«ñCu#ûFñX9¬Î:µtÉR«Ý_°.Oik+`?š&yTm H5B¡y_Ùk¦N˜]E)¢‰”à×¥lT †#4…þc,x›q÷¦`æ´  P˜ÜäÊJƒþ:Éà|8²aš>܈ ˜}Wg9Ñ#a`uS‘=Û\>4°Š¨èáÅ-IÑ1<ÚÄ"Vu÷;vÄ‹ «Idn¥oŠeâ6¬£h<+4žõÝ.é!k¸X©"‹~“f\×5ÆÍgÕ¥¨Ó5Þýo¿"^l=/?˜I$Tjµg:äe‘Õc‚JÃ]·™sL»Tj«µ~Z6å’jô(úU¥ax„µ»5gàqjYC3óÉdëÙF"`iÍI¿èTP Uí±Jê¤)ª¤ù¾ók¯"§W]C¿^tZçÙ¨@æ¾oÞËV{n•"1 «Ì[ö_¹‹žìnaEr;¤@«,²‡”(j‡ïV¾s‡ºã‡h/mñ˜ñ+ëû˃ü.p­»{ÕmòÔbòA0ìÊ9§I˜0“‘íp³î¨Ð–¹Î)6“×3Ëi'M3¼l}«PÞªZƒ9é÷ Ó(„/Òäè“m®îÙ †ýhTX†jc^š‡R¨6aGv©JýJ7 eZJ3÷äÿE÷¶ƒþ~ô¯oP²Äã)3¹1ˆîõV49ÃUVtjÙ¥u"­/M´#kó!é\ä:FÕtXh0îò,ËŽ\î¹ÏÌ ¹*ó©<óª«îЭ+V?ì Ù ôµ–ŸÐÕKô}”‘wË(ƒ©¨…ÓËå€9~s4ÿË‘ àV€ FÊ_‰(ÙLí¹:úFaÝó±ž]À Ò!‘Ò˜Íú;­;õw y5J^—+𾎪@H%·?»±üÌñ°ë“zŒ·L‡âæÕd£N’ÇݽJ "î †¿lã ?€‚)¢ î·ö° WÔMN‚`á¨Ë=ѲQ–hF-}‹–g,„ð‚$íââÉê~ 4•~Û©¨¾é´¨Nú®/ú! ì*SŽÆư¤Ú¸¤‘ôJˆ¬oTÔ7K†ãFM#Å-©ªQ³¨âoìÿ3ýšÉ:LÕËÉËÜF ÒEäú6ð‚”ª¤SÙ]â#Ÿt*¼ß9+rÒ» ®e… 3ƒÆƒ#9QFoÔ€ž—±mé +]á‡ñšD;€Æ—à¼(ú7©‡0¥6FÉu“U¥^#>fT› tÔÜ9¨MÔ”ë%Õ§òÞŒ‹3@ðäá%‹YhšÇ\ó’kaë€ÓVZÐb5à l±âLÅZ'"Åüà6^²tÌž[b·yp0"&¡m¼Ý*(Ü3‹ª Œâ‡*´•7\ŸïÓIø¦†c#2\•œ0êö[ƒw´ìEÂÄù1´ÅYHNæFµ¾UÃv³€—pù›“Z6s~`aÔÆžªy"–Üecç\§‹¦~§Òë¦ì. Âk z5R~[©Øõâß%$øI—š)ÝU—n'/%I+Jæ’U+úf1‹|$Â#¯M)Æ„‹»x^cÈ1aãSIÃMàÛgf‘FpïÒèt™—ˆ9þBë{ÿÈàýH&ÏPîÒJ¬È…py%ÛÞ P‹‚«™eÖßðº¡û9Ã]wÂÈ£øÝò2´ÆÇäî·½@ÄÆrxà>Aæ~ÏI%Š_PSpÞlmeô$›ô QÜäRF0=¦ª¡xIS/®0†•Páb'Kéž“‰¶}÷ên&ïlh®|0—¾üÖî˜ ±°m•C}kÁŠü$yæ›/ÊFR¶M8pµ5íÏ4·/XÚ0*0‘kÌÖïÊ vãæÿòJC€1C¥£;ÖT–6©|€$uç¢h´ÑÏd8š“&Áü³h22Û˜Ý* [¢l¡‘%0rB…ò­æ¬jvŸêU¬¨£báY¹ÀE:%ÑÜ¢*>Ešp­?Ø„Ýxu%®ŸýjåÞ,npæÝž¢em½^Ýà¤{›sç!õ±k„ã=t. ÒϹ¶Îjëûί½ œ_ÕÖÏ;—C ÝM¦ëqo;áDºcHaÌU`qûVæ}‰wÛH¤»±‚Ä-}Þ^iI,qUÃg'Ÿ­³ã#šü‚¼µÇüX[‹-µ§sÍÖÜ[ž‘@¨%â |ö]w…A1gLP1„ x"ê!¬=_ÜGÅséÊ%Ýð‘FFÀC¿¿îxïº.ýú¼0§5-âËT̺RcľU]¿ š¹è ýº¶èv/ÅÞÑN¾$”<=ÈPt¡–PwG¿©_P§èa¾·é×¾+†’Šmb¥_w°}ÞiÝtØðóNßokkßùö¦ùî ä±Óú’6¨)>¶Yn :*YyïW½e'ÙuH Zß‘×emÅ«=6NJt#›$ѬªG,Mgm¾Ÿ²tMºvNÔij§èâ^QÓ?I5 $÷/z’û¼¶Þv‘óéùL.¹«½÷þÜõ´óÉuG–õ˜ù0Ÿý0®‘ݦøÉúÍm2Ø<ûW­ýz!‹·ßŸ(­ÇsÒÔe•÷4?Zä<!d3ô¬ðÏ…q͆ÿP·ùµõmmÑKÿ}w˜xÑe¨µÕÑ{µãËoôñL•–­­Óβó+õFýuâß×Î'9‰ÕÅ ‹e1á“T‚’ kî]Œýd¿vÏk$j&­Ü©ßº šnc¡hÙ=£hX˜´|&Dz¤Íšå#7ËÛü¹ö„E}ûÎw“-—ß^‹ºœM‹À+v£œnž—MÀtÏ«•we­@áQK¶Ë¶úYÑ_/Д.ù%‡þÃòzù¬Â"}\–—iÈ6Ù§FOïü¸V žwða°ýŠ‰Ñ˜©~þK 7ô¦ÁG4XmÊ=‹ÒÆÈ”@C±Ô^z  f;Uî²oG”î´ó%ØÑïÝT”«÷ÁCªq¬lËÑX~Q—§²LóÅã€7Ä7ªº. Ï5Ï|u13UɱÐ;£Ì`°].”pTÒo?{‹ðá¡Û+ïW1ìnl# îÓ`h¯Ÿ©2 ëCšë¬ev›47*–ÌÍy"dÒ951¸Éa³pÚÆk«ûTÁ ððçžIÝ„¼ƒS55Âß/MÚ¨³²vâµBfr˜ÝÙο¦ÝÚáûÍÛNFÅ;ªË & 3ý°wóÚf=¹E[î}dùõ¢bÁ¢~!ž]+Ñžb¥Þè ²/÷VïY¬÷?úæ°¾H×å%ª™K^MÉ¢Óu—»âµ'¬Oñ <ß”=dÆx9À¢Ú…œy²¦zOuÑ+i;2ž"Kˆ"\‡ Kh3£xxfmoÑã_dd‘=;v<ž»Îb˜+ÂI#ï ÷ÐY÷ @*þ#â»î€!ÐïºS?ï|C¡€ýAØ=÷f«NÇY¼Ù„…ý…1õCÕ^Ót#H?FõëßuÖþª³âþ»u [£æÛîæ¼©6…F«WuWKª–wFrc®îXY½£sNöTáMecLD9¦ü¹÷fÅ|—j·À ‚Yk=«MÞ×Â馲MèÅã-ó‘ã–dº.ïFðìM4^gk–­³\²$KØ|ú7Ô¤pÍ"€÷Q ?åaY¬jTžÃãob.ote„]T”3íÅÌ·nVC`îÓûÖ%bQø<Ö¡*{âŒúø„Mϵ?ëù²½7ïi O Tþì[/õ羘¨?Sàæ‹ÚzJA1Œ®Ò'¼ì`ïÑï{~¢ŒÑÜxß›ïpʈA$5ôŠ¡¬´(îö²ÓG¡.Ê-=úÉî 1}ß™†ÒUí»Žõ£_ö’S¿Ì!ò]Y­-JÐÑ©õÃY#è}|ßÓ{W´CüŠÛA·4¿°ÚK1ûÝA(½g>»o’ö.WÀËθ‡.LõwŠQÅÙ½a¦·ÜBX$ýöøá¹RÛi™ ÕÖ…êzùÐe ±¥:"¢É}§Õ³]ý}Œ¹Û£µ¥Íݲ1Ù^Ýç^éçÞÓ®túD”t>tâ½¼$-æÿ¡Í]ˆÒt¿â“™ß,ÿýÌcš¼>û„{\‘¾•Þ1%D&IE¢À¹KwSýœ/J»uUáîÇÝѽ«®Ü˜;Ÿ°×¨{dÒ#‰«Ÿ¬Íºþšy¡f힤'L&£šöê¦V­ù£ËýhÕÚë+üM•ïä±¢“½¡Ë­T,·Z>¬&« g~ï1-?½:U®Ä¾I0d”± ÀðÅ+›7"S+',oÿרg*¹Uñ¢kYÜÆã¹,(z௟ ÷¾*¿ˆ^u}[Û0ö#"ržûØm=+¬ ½íë>³.ßèL>×;ÚS“É(nJ?òc¨é&ÀßK"õ !.ž;ìË…ØZ lþé… ‘/5,‡YðÏ×¾¬”JpFýD¾lÀZz0B’?z2?Ñ—…Ma‡Ç}Ùîf—/Ž'â0ZÛ‹ø„ £¢9•aôùéõA™ºì?ý/Ÿ>•endstream endobj 457 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 2062 >> stream xœ­•{PSwÇï%{UŠÕivqª7éÔ®ø(`uëc\ºõD©õE‘ðyH¢J€$÷$$ä!„B€" ºÁ:ÒJ}¬c×Öή³ítÖuëêÔb»ú»x™½Èêîv¶ýkç7ó›dæœß9ßóø\‹ŒÀpç'§¤.^6ñëmæUœ™ÁÌá›ñtíØÆ¨9ØôÕ¿žƒmÞ²gÎ?6¢yÙ={JýLä›ôÓÑÁ—1Ž–'+”zuN¶L#Š“Ì-^±bÙ"Ñ[‰‰+Dïʥꉸ@”"ÖȤr±†û“/Ú¡äH5zQÜ*™F£\™ ÕjãÅòÂx…:;iþ"‘6G#¥J ¥êbi–h¢@#Ú"–KEÏ’v'+äÊ"T-JQdIÕ†Í(P(Õ…­X’%Í–åäå§cØVl¶Û¥a;±÷±ÕØl-¶{Û€mÄ6c)Ørl&W,;‚}…Kñ˲ˆ.^&ïqäæÈóQTÔ(_ο@$šÉ\ÆÃ8 Ȉ¼øhe…yÌtQ0xìKvêªÌ-’Šå³SF©ˆWå!ÛIHU›(Ñb A;Ðb>ýzØd´õ Qñõð'¿¿åܰ‡b ?mÃì740K½xÛ5ÔqÇ$£\z%þ[gñøyì+ìÌ{ Žð{ÐLŠÝÏfö‚¤]Ó§0 CÎX? ø»úNrïþFÝ"ëÜ  ²Ì[•{û2tR 9mlzýÒË$6àÃÈs‡”쯓¹n%:Äg”a ¿½ƒf£™KÌÝúA®D!üŒ`§½<.ÿë™x ©¯×në>qÈÏ>Ndù,Süªì,߇raijÀKÆ '“~8ˆˆ<´<8}÷s‡ƒ69((ªP6«¯ 8ÚOæùsÞÉ}WvLH#œÏ’ÿ³ìg¹ :`Ú&K™\ÂļÌrÞz =ÅCW˜•‚ÚRi¤é’2Êd,?f2æuï·’åI¶o_ü¨¸£Rxµ4`lÒוŠ ›îLXšýŽž2yÀbÒ@ÓZ!;—Ð[è»Åâ´S5®ú–·_Ù£òaøÂ•m¥µBy§Ì!v¦Õ®õÀe2è}R3/ÏJYªìt5.¨öqù=áZMÑ ?þMQ <´‰ Q¡=ßÚù–ÿÚ²_°ÓÙŸÝ_øE†79*Á\A›ËŒTÑú´œtn^2Z”¡üAIK½Àæþád?Ü€³›,läDïØ ÚºëCŠ+¸¿ßÕŒöŒœëç!zOÐXÒu¨ÈÛ7®|4ôË;…í'N»›ÁÆí¼Ýd¤Ëp˜,u•×Õ;Ýu%þ<‰L§RPò¶lw×ïõ)ËS»²ÝÇ„rdB¨ÜÒîâôÒL ì&“ïïB|ñpðs4›ƒ$M&Xª¨|{q;ô€ òëÁ@˜ßÏ]wî‹iµÜµDdÿ.˜·.<Ôéëx…­>7'†üñ!>mÝ®yÂ笟èÑq¿B²!·Œ¢âÿ'üþŸ¼GV.ø´^4ç ü7 ̸  oÔéôš¢]s ±¡™b§°Û¹™vþ€z¿…VhÓÐ<ùà6sÍ)!Ã'bǦõrlå!ò hkU Tƒõæï\΋çNÙƒ|àÚ4ý«àZ®´’*ÂC» p„Ôju4Å.d©(¦”øQÙ­Ü ý·ìâ“«£ì‚Z½û”ÎÒëMe/²l€^ÚG?³Ý G]òùLøD’ªßÛ(ø‰HÁDŠyºåpÓ˜(ˆ»ž®â}ÏܸÚê{.Y9U‡h=A¹P8É¥ºQÍQ­‚6² ãå±ì2ÆgrÓ6Ž{í!häXV`ÞÍ}"Õ°ÿ9÷ÎB­É­·•{Ó¹2š+i;o<{ÂÓ[åáÉ:áéá©úw´ÁVéÎCéãÿˆµ—€Ùñ¬6JbúbQÒxŸíHµÙ ³\ɘÃ^fu-R8=^>›á"‚SÃÓ¨©‘zEô¯;::\ý†ý¬Èžùendstream endobj 458 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 345 >> stream xœcd`ab`dddwö 641M~H3þaú!ËÜÝýãúO/VY~'Y¿TY—9lZ|Ý<ÌÝ<,k¿úž+ø=‹ÿ{º3#c~i‹s~AeQfzF‰‚F²¦‚¡¥¥¹Ž‚‘¥‚cnjQfrbž‚obIFjnb “£œŸœ™ZR© a“QRR`¥¯_^^®—˜[¬—_”n§©£PžY’¡”ZœZT–š¢à–ŸW¢à—˜›ªq«„rÎÏ-(-I-RðÍOI-Êc``` b`0f`bdd úÑÁ÷ãS÷‚ï{N|Ÿq˜ñÉÝï‡o1ÿû~K´·ç»ÁÁžï Ýó$_i?û­"÷»ä¯×Ýï¿«³}wý}‘õ>Û|¢ßg|gÿ.¼Ã904Ùþ·„<_õìNÓ¿gÍ^8›íwÂTö\×¹å¸X*óy8Oàá¹>™‡—‡}¬endstream endobj 459 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 2297 >> stream xœ­UyPgﱡ»5811BƒÝÓñØh ·]‰’]• ¢Ãæfp ÌÌ›áafÎ<0j-èšDcŽŠeÈjLe7Ùݦ²1ÇדÝÚ™hv«’ÿ¶º««»ê{¯ßï÷Þï÷TØ4J¡P0‰ëÖ/^2õ¶L¾G!GM“çÑ€·> $…Ï£n_ñôææpDl4¾ßõõCˆA3®…îpNã³ ­]{J?j ŽÂ!xcp´ÿØñöa8 GÍ}™ƒ[a-dA¬+HÕ½´µD œ2PR \R «Whù+ù/ü¾±-Ç€“Ø 69ªRŒ“T¥e¥ÎWŠ^±Y+vŠø Ü“ÑxµÛV s:{ëk°%€Ž\+¡(ð´ZýÙ(‰7³YÖŠD¡XËt9¯¬\ƒî Úbv…Í=,¢TV)Ï qþí…®Ðè1dá¯üýƒ&'Øë…ªÚÕ ç´]åí>_k߀¾û2m\Iµ`G4ƒçþB úg`úà=èºõ£=De`6Hè4¹(êC´ÑOËïžà'û´>æR]E–8Ù{«Ønç§$Ç0|R0‡†ý­­å€ˆœŸ°Ÿ»ó–‹~Ônd–Wä=,,EnfŸ³ñ¸€î`н ½Ð†£9?£”ï®ðÉË<Š}£È-Ç ¿ò}o¿êðÂOà_A²!„ä ­Ùæ)ðµ °+Ó¨§º À1§w||m Ñ‘š„¬…œƒ¦šÖ|39-r[4Øj«J5zupù%a¸~¸»¤Õ\jªÐªf=rêB/éèê ?’®ûÃ(í#ý[fø²,ØQÔQÚÝåm“Þ~ödž¾pþ”‘|ƒÂQäŠhnÞ ÖJ»µÌ*W¦hR€KNC±(áã#§Ù•Ù%*ä² ±[î¸L£Ò@,¿Ûk'9wKCV©±Aw#îÖc7Ý-^íåcUÿkF}y]•æ4ƒÓÕ°Õ”‘ŽÉ÷™Ÿ½âjÐ+®Þ’p¢Õ}P”ûÙƒuîK‚OÂqŒìöª¬|+’˜6[œåŽLö\ ÑXà)þ¦3°&iûp2)“»?ß…gƒ¦M?ímW±(Ú.±««O‰ò÷lÒäežôÈj1åìÐn®SnôÂ¥–ÿ>Ñ782xÎÃpš{gK)é’™ ±­÷¢Ä [W§åÙhïw‚÷Ÿ)Ÿã™˜Yz/ž…#¿yÑè¶QDu6ÔÔWW‘ù·‰ËgH.ý¾~ôŒèe^¾ -þÄÓ?@Žvär7wѬ_[G@F’%Ÿ³£o,Þ2ÃÉÎÿë>Bݤn;®@ Bî y’ï5ué Œ&½¦»¤{ÐÓë°€“ åo6WÊ{~6ã“$ã8ûC9“j]‡E9ŒU6‘q¢ö+Ð[Òè@?y1ã¢p?Ó†–4@KÀCªi²ù¬Ž20p“› –![Ë 9oÍv¾…¡`Îl6Úö‹²ÄB³¹½¾É)H‚æì6¼$hÆ­3„ŠFÈŽ¡Ñ>äáÁ±» Pwa¼©éÝ×;:§ «ØV(€Á`‡½¥vrÅÅf»€‰bÂeû‹> stream xœuW TSW~1<©Ö>•.ï¥*TZw­Z±RTcA"›Ê*;²¯’üeQA@‘*aÂêB-:¶SkÇV«SëXµÕv¦Îÿè¥sæ»YÎ;ç䜛{ÿûýßòG˜Œa$‰ÌÑe•×bã§â+ñÕ1âkR Þ?}0ìjú󪕯1ë7¬xmu©ÌÜÌ¥`nÒüªEù‹xi"î{'0R‰$,*Ã1<"!rWÐÎh…Ÿ­Âañâ…vŠ7çÎ]¬x'4 r—Ÿ*LᢊÞªŠ¦_B›ÃývD'(l–팎ŽX2gN\\ÜlUhÔìðÈ å¶vЏ]Ñ;nQ‘±þŠ5áaÑŠ ªÐÅh±³Gߎá¡1Ñ‘ —pÿ€È0†a„wÂÂ#öDFEÇ8«v¬Oðsñ Úå2oþ‚Eo¾Å0Ó˜Œ+3ƒÙĸ1›wf ãÁÌdV1ŽŒ=£dV3³™w™5Œó³–™Ï¸03™™ÂX2/2“˜—Œ1aö2÷%s$McfŽ©ó/iô†‰I¤ÉÓ-¦E2N–&YòØ-f/›šµš=»`ìGã|ÆÝ6·5?6ÞV̵øipºA\W.S†çsiµI@Ÿ$H‹ G¾›š•šÈܽêxÍ^`c•²óù%y}Ð :¸œÛm>Ý €]^4€S§â)Ù ™nj¯”5i¿‡h…»pvtÑL9ùA áÐŒ4˜’T™Å°=õg½Ç¢¢™T¼„9çX3¼ tHÊâ5?ÊÈ;™¦Jùà!ðÈE'¥üSm1èé¶…ðÙè¦N¨’£ÉƒöÁ³å±+yRö¿–È-Ä{Ùm¢¹^Ò€–XŒ–RqÃ8œ2çd9Y=HÈ‹dÂ}k\‰Nw~Äq<9Jœ92~üêpXtøþ½èԹɬ%î»’ã….9q2âXÇ $÷x.¤A,;’ûß`9‰§¹²ÊNý5`?p!DX½ÕÙË·®%‚žmòûÙÍ£@Ò–uˆœÌëm?ÃÖ:gzok;bI,ÛᬾÞã§õñøŸ-!R<ÇA(ËS8ðž0]ÂQÁçlðe%Û­/„vð¨D“ý®ù|^òŠ [w Š^Ø7» 2àŠrÉ3Úáq=RŒgÈ,Üì%ÓØ¿?Ã*—õ»CªåDö$éòÍ®ó×ù>¯­òÕA ‘›`°Š§<ÉÖ#¹ðm7p¿‚á&_ê¹uùò}•}<>”—ßaJ”É_ã¾­¿7zø•rt:hjoÌ„fÿEÑrƒx× Á™h%Ůቚd;s *ú ìÔR`•h:r—Ð=›´ÈPh”@Ó(äÄ”îò ­MÙˆÖzÉuœŒ‹ŒÕÅcwÏñOd¶yI=ǧ1¶¾±¾ª¥noi®–/̯ÚØOZ‚– Ár²„,ó%RœÿÉg½ µÂØöOÞ ;~ª«!]X—G±¬b‰Ù7°°ÏKÕÐÑPv=?Ç…nü…O¥FùD ¯åžëc„‘G=òÎ º`CÈŸÁçþð_ÆWf~Cl–zµtp3W­ëÒ}@u6ˆ¹³ ª±¾±¢ã!Q€?q_C½„Úì}+\rÆp°¦•¿ÐTÜ ZÖâ§g¿@eaxU·x§pD-Ö¤Ö$C,u”h#.çe£Ñ„çFåª!‚ªM)»¢= mеpu”Y{h©!J®ÕäA%+¶æÉñÔÈÃý‡öV‚å (8v O‹ÿœš7r‡‚LÓé1µ$=|û+3ÚÆ@¢ä±©‘ì¹bÄ¡ÈÈS“áU¿©h‹|»îý’m‡…v6„#–íÐîÃUŽ.¿)Gü»œ˜Üæbvnó£í'¨Í_½˜×VsF¨Ñw7A7´ÄÔû³b3þšÓ¸àyP©¿“ŠÌUÃÉJõ½µý‚Ï™‡•t‡ CÌÉ”'Ö¸·>;VÙ êìäl!ÚsCÄzckÔÈ„Ó&ùç ‹r@3ØO,òýX ü™úé†ÐäÒFGM2úË8™øD†’›‡ªµ… ¦”R«ù´”œdj°ªÆ˜º†ú*}‡ûÛÓãCLx›åÊ¿'ÿÇVïo=Z?¹¦ÇPz™¯ñé×R ¹ËC2ì-I*É,¥ÎX^P˜—w&¾=«Jɾ÷Ë+-ñ ±Bí‘òü’|Í>u6¤³ U•'Žª‰kòVª¢’ãyUµws{ÏEk݃ʒ„ØèÔ]T¾»*âWĦûB »ñ#'Ü‚Ž÷>kJ¹¼FÇoªñ`s‘á`¨½†î߇uÛ}`ƒp‰øqõÈÂÆà@wò*™ªên¾8G ¼5<÷ÐAB£[*Å,q„ëUÕyïö÷W5vœÕŸê≠1RâbÁ¯)ó':µ‚J4çÕF86ˆVTÏC”å×?à£!¨‡p?¿'‹!ûL{d…(=ü%ZûùÈuBǦv-7z郴?§¦½xýs9]rhA!‘°=²ÑÒÄ»~©+ñ?üë­âM\£ºp}[]}KC|ßaþPAË©!`¿‚µÁ›cçÍÏra·:M“¨¦ƒrWe¼Ej±èz·—ˈwÜ@'«qüX“…åæfúÃææhvÒ||ž¹Ãü¥@Èendstream endobj 461 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 5400 >> stream xœx X×¶uµ-]% b xc7qг¨‰ŠQ£q‚‚‚ˆLÒLM# ÍÜÝ»›yFæfjep€¢â§“h¸×隘Ą$š„˜¸‹rߺ1¹~ߟ—¼÷¾âcúúœ:{ÕÚk­]"fð F$Y8¯su2ý2Wø›Hxe0F Dÿ˨¾D‹1Œõ²%c˜µëÝÆ¼}H2Ì ,Å`9¸ñûK#ð¿†c¡5ú cÄ"Qp¸Ê94,Fè t˜ä3Ùa–“Ó¼i³–û*}¼CÖy+|ƒ½•ô¹Ã¦PŸ@_eŒÃ¤…JeØ‚™3£¢¢fxGÌUø/ž<Í!*Pààêá«Øç»ÇÁ%4Dé°Þ;Ø×ÁtÖ¦oΡÁa‘J_…úÐ=¾Š†aæ, u _®p‰P®ŒÜå½;ÆgÝþ=ë}7øùnÚ$~ý¹óæ;-˜4yÚô޳fÏa˜±Ìf#3žy‡qe&2›˜×˜ÍÌf2³•ÙÆ¸1˘íŒ3ãμÍÌ`<˜åÌLÆ…qdV0+™UÌjæuæ f-³ŽYÏðÌ(Æ–±cì™ÑÌߘ—˜!ÌËŒcÍ gF06ÌHæuŠ43˜IÙ‹ˆž R 껉/ ž5øŽÅ‹ ’×$eì.¶»ú’ËKw‡ÌòñÐéCY:ZfZ~þò+[« «~ëÃäÃNûyxçˆ#Zl\l #§<7ò~3ßÄ÷ð¿Ž3ªxÔç¶®¶ß j«¾|0¢‡QXQ.jìs ÓûÖð©…êœXˆMJJ, êÿÒ.~¯2`±–“³GÓk¡š R{AsÚrœ‘uÕí7Bäê3õ™ÈÛa䙚‘ SCª}ð&—ÊÙ“ég¡NB:yƒx’ݸ€ÌÆÍ¸àGƒ/åAîÁ8HNÔÊ’È4b¹tûÜÄcŒáOäãk(‘jY66"¸°ŠýUUÅ ‡ƒ<7x{¯SJMD}Ê’±@Ô.}¡î ‡jí5sõQì2u~ -m6-­Ãˆí¥¹ÓÒ¾5•véÿ/Mr2ý*eçqøšÌ}Ä:¦´ÈPóˆ}œ4w °¹ª )ÒY˜ù¼°œ†M%Dj.l²Ê 8VŠÏ`Ü1>ð&†¥j >NªÑ$&¦¤Ê+åY5»½ËÜ‚ŠÂ«"eò£ÉgÏ$RWîÏK2Dƒ/7g"œþé¹x©:O›ÜAHŠ‘‘‘ì~H,ÈÒéÊJ¥Yù…™™­»ÛÕÊ<ë¶[7NFÕFVÉøg/- Ï\V¹ÃU­_"“=_¡—êâ!…¿²Ë(ÍÆÑþµ1âürÑÏÝ8ºNŒw„ñ¼‘LÀµ¡¬v^Àø)jNñ©gÉàûÊ®M×oJ/†na—ÊýýWÂç¥Rl`„ñªfd¯= -Ë“kxPҢϿ-­2J<Õ›`'½6g:ET^Å`^ùæ(æípNš1W²‹5ô¹a^#Eõ)¡’ý‹;Ñ®VR²µþG4|“Ÿ½}óüÔ¶sRü„%SM„º êÙ22……ÕàÝPu"ù8œ…ΪúËœM´n¾'å[Ù‰Ù2ÄçkÌ$¤M«ªFEå‚FŒ+yu–& ²¹šf(“Ù]¥Ö  üÍGc;¨!ûáD‚vY±9ê\È…}ö!|ïÛUœi*hÒqF6Xí ¹.To^U¯9tÂ!A«IS#™vT!Ÿß¶ïqßpüŒ¯¼r"£Â´ÖG­½à ÁgàŽmÚ¢0ˆ‚MŠ:y&I·›ˆµTÓ!ݾƨ¯¡gÜ©ÞMýLNÏ4°â0”kóæª }@ šThH¤Ý4,L)¦¢—a_{Ì\ÛNZ›½W¬£ú£)Džül—ŸžP Å‘Ÿ‘׋Gì~$G22TÅ`_ éyyœÉF QMèo¨ÿºÉ0¢íŽG¾Ù}«ÎÖÆC…ßÜâ3½ëýÛ€3\+¾'»º™UÁv8U"Eg–:G£ç ¿#ÅÛ›·Ü{¼bMmYyM›W¦VV[s*·¸3}çÈö²65ÌÍ’Ô•AΑÛÀ‡szvCJïñŽž/ªã?ä§, öÚáÕxúý‹áô,êÇÌ%*òÅx #ùä\ÐÑK盜cccõÌ!ê0Z¸/•Öç0k “(̱‰ªÔ”5 í}—œ«ÑC¶=äT™Vøjµ¦Çê {V|zMûÆ o¾uU¤bûWÍ> A¢ž‰z8 ‡µ‡Ô\‘®>yô™—Vk w'Ú•EÒÍR@•aJ­új]}<ᘹMX†È:n  *úµRÍEµ¶#ÔÚnñ”Šºü’ÆöÎ’£ÐÇ#j¼kvé7BìHÙî´[¾<Á·:¢å` ŒKÜ̽N‡žûÝ Ÿô±\±# ,BFýòÆs¿3·FOkò4oCÉ'‚’RõS1®ÆŸù”"È=WÓ”Dîš(­7¥Ù靖„2MAÎîgí#·†¬X ö Ø_¨Öçèêr ‘«‰.Q*bbÃ<[ýÎþ½íÒùj)®Þ̯)=q]oodƒÔrˆ{±wÚ5ài JKHKšHí¬„ª:lþ²Þì²^=bÆðûÒ¢B^USSqÈxuùÑ9däT2œðO'Ó &;Œ/ç¨@­Íþ©ÒeS¸pkœÎâ$t¼×q¹ðœFA-ç—¥`ìQUœ`šï‹û<û¤¼NkÑãw9»œ~ƒöEtõ?–3²P²’!9†(û¯Û¥p9©\«{²ðwßJeÃÀcµ¾óú"]'5þbí¹ªL€Sµ2á0«Ó¡öW&3.#±ìó!37#—RaÔ VQ›¯ÅÂaŸ›oji®8öKÉ¿$*2ߢN’Õ[Ú‚ô)–¦~.‘ÖLkl.Põï‘BWšˆ8Î §nö’ä7¥eæEÕp“rL°GKSöPØ'÷Üí»À§:¼5¼Mó%wGVWÍñÊÃ'Œ]Ð Ç÷ÕúÄ‚¢¹ô+|j–g×"º±µÃd“áßOÄ¡Ÿl«(¥[ÃâÔߥXxú"ß̦dÆÞÁ(,1ŠQÕ7‡O*PC4Ýj¡¼Nr É:àú›ÿ¼BVXÔÿ,'>“òÔ¾rÊdB[ §z)]·hh=Àât›ØÕ a²þ·(ÓwRVTãbs2V` o1|e&Tü¸é‹-§#ÈíO¯¡Zu>«ÊNÍJLP'§iefk€ì®U6îíÔ^†VNWÍçâŒG%up.{ëÈHš‰GÖ÷ͪ7ebe¯¸/¿âq™‰sÉzÍ$óˆñBGŠ×ÓËg£—”|C>á'Ùç˜jw÷ñc·¨I>‘Í™(³Â¡õZÜìùBŒ]æ$0-Ø/;O¶ååiÒò¤ i4‘)¸ð²CmiYÍQß#›9m|MJØ…ò»äŸ°›B“Gv×âÚk÷j1¹VTÜn¼ Æá¸‚7Ä´÷Í­îv†6ÆÉŽ7w@¤ª“R!†SeGW•äU”4ú„xªü¤a•{Šè„0ÍyÅB·ÿ’2&4€Z`‰mŒ[¬ìàÞþÎG¢õO]·OD¿ëW-ÝÔôL£ÒIú¬èZJ¾’ÌâÜB®7Ÿ O*kõ‡ªeÅåµÅuÀ=‚)J%L™%k'‹ùÃh­]àºØsúä¥çKòÞ(ÄÌ< „ 1®é[Í¿¨{%žÊq5ûŸÚV‚{€|Ú¬ð¨éáµ¢%2#î"³”v\®Öód¼é(×sShD®ømû“ú Ý»PB[­sÀV§e“Ùü.å"ªƒåJEHH•ÂØP]Ý`‡èüåX'ºxï_c”0—³<ü:¼ËÝ8wýäÏNX$MÖ¥˜C%¦¶°åŸ¬ÕÆ'Kãbò ÷nÛ_¾íôt“:Ìsœ°¨Ã#3UvÒ«4‡†ÞÎÕ Ôiý`#·Â}‰ãbBÅÕ)Iª¥²gŽ»Q2²…‚¤‚l® [ZXšwäìÕ w:©&YÝÿþÙÅ}ïF4Èv÷Ì_]l)¤EÔbÄüG˜g÷-ésâe%Q$’ªY:•×ãêvõ§)MÓ%P:ÓÆJ÷È%Íé8ˆú_ÐÍftö°c¡¸Ñ¤ÉP£Ë-ÏÈÓ×B)Ð-ÊCŠˆ“9ì›B‚¤Ã[EÏnŠéÌ áì²à]½A× 'À =ù ~`¯Ÿ{LlK"sâá }‚:^n~JÕºcfë¯xJ{!B犌Å-vO>m?öA:Í4àùQo×m bmA2 pѱq‰Iœ™}¢VSG‹qíŸQ³:à¶óÅ´ïñ‹¯üM_¶6æ7 ÈÜ9}ù|éžåFÚßþ?úð ÿ êö¶ì«ö)Ø Îàîº5þ=Ÿæyð&¸oÛ·“³¹¯}"ùß½¢ ãÉh£à].Â+ÝâztãåXóë§V-‘FÊ%×Ò?‚lªî·ž{p$»Ä¤U{wÊäìèx­ö`¢”é÷ã‰E¿Ä$c‘u(z‚« hû@t¦©Ôˆ ÄOñ ª¢œôŒÜ,iCë©ê³À=2Ã_ëB=eaÞÞÑ{`#¬l‰èÖA–&¸òúÂfcdiؾÐxÇÎD1¾üã†/¿ñy¸ÇúÈp;ÛhL5šÆ»[_æ¦d¤j¨'ÊvDïLò¦N꙳³,!SC-“‹‡„ý2óé™áÁWÇišþƒ{VikÓ£ñzó/Öž-?Ñy©Ê¡=­ÊŸó‹ /ª«++m¸àÓåH}öUbChŠû…T£ÕʯQü}Ú¡ÍÌo‰…Ô¦ì ÝÀá0r‡w…½ñõÑ'µ „»aDæÛ’NXwXZæKу^îÑ~rW¯ä00¥üÈ áê=ÑEã×4hÅ ÏGGô2Ô…ê2.T²—øÒ>ͨʹPXuŠhÙ~ÓÛŽQdE<™®ö*ðlï€w ¥,àk•ÈuÝ2ï^×gMë=~ ëN…Ÿ³µ¹Kw>¸LY -Üß?:öþí«[–/ß°ÓG)5Äò§«?> ÷¹¯§ŸŸ5ÖiÉLß"ÿêé¡è¬´K~œÍs¼7®wýêO+‘AîÉ¿y¼ô”¢DªÊ )Ïæ&ìâqÄ¢ª§ƒ`½ýÂ7=¾¾ìêƒ[§/Ô,5¤å«O «&ðW÷`0ÿ™`ð‘,5\ñ.W¥[Æ‘pi†$Gs,Žs51å¡¡Ñ1Á«ºv=Bž^v*µù«‰åû³;a†éUÏWS‘‚ÿìc8ÞUÚÆþm>–Bt5ú(˜rË)8]q´á̱Âføj]2Šü2ÖÃ.nÙÞqÓ¥&! u!#ú1º /›•>8""8¤ZÑÐPm¨—’®0Ðÿ)B‚«#ž«ÿ8¢¯øÅµ\ôÃ51~MyÈT§§d‡ìÁ•€Î€3p#¾Š£32ôéÉe§f$ÇÏr&¢-R©ã ç\"9O†¼G¬{–Q—ÈÎÌÈ–ýggŒ§[ŸÁ þƒ¶¡x#ZÏø8¸M’:U¯ÎN–]Ú†ƒÈ+@ŽN<Èò7:- R9µ>)¯àÑ-“žú¶úG@K“ê‘-¸„KÛ“Î%<LÞrNK¥Ñ_Ë¥f%gçß¿C> oÙVÖd(.ÖQ˜¾ÁuWEùX+Æ£¸Ž¿Jj×±Ø=ÇíÿåEúFDÇbÑC±ŒópCý0à^_¿Uýl€·"¼¦¹¤.€·`µ~JÛÂŽ¹wbÛáïÐ]×úUìÀ=ŽÈÉG|08E£EÂ{ðNnÀÝ‚óå_\/m„£p-ªlFþfXkaÌKظH–&(Ì=EEº…æÜ¨Ûb¼Øgôz…tÝš™t†§ƒ`\s5\•»øÒ “Á}AýçŒ\áï=%³ÛSÖœZg sãEEˆ"ø€÷ìäpxÏ£^©Í¿á3—Ud¥à|Cò³*%dwkÒ=T:dð¼rË— Y––Ý%–/3Ìÿø°§éendstream endobj 462 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 341 >> stream xœcd`ab`dddsö ¶±L~H3þaú!ËÜÝýSà§«,¿“ƒ,ƒ_’¬Ë6=¾næn–µß÷ }ÏüžÎÿ=E€™‘1¯¸É9¿ ²(3=£DA#YSÁÐÒÒ\GÁÈÀÀRÁ17µ(391OÁ7±$#57±ÈÉQÎOÎL-©TаÉ())°Ò×///×KÌ-ÖË/J·ÓÔQ(Ï,ÉPJ-N-*KMQpËÏ+QðKÌMU;UL:çç”–¤)øæ§¤å1000103012²øÿèàû©Û½àû®3ß'c|óâûýGÌ?ܾßý¸üòå~sÜÿ­,÷{Â_¯ßÏ7aûôû<ëk¶?A?D¿Oþ®Îþã|°³—Ÿïo~y¾Ò…?æ|]8w!Ûï¤éìg¸îqËq±˜Ïçá<6‘‡çÞ ^tŸ|endstream endobj 463 0 obj << /Filter /FlateDecode /Subtype /Type1C /Length 3423 >> stream xœ½W TSg¾¿1¯Ö¢Óg,ií½hµ­­ÅÚvt´µ‹¸ÔÖ¥°® *v ! d'Ë?{B $,$ ;ˆXPk]žB­Öμ¾¶vÆYZûfÆyg:ó…^æ÷˜¥súÞëygΜœ““=ÿï·ý—EÄÌ"X,'qgòöõÑGk#³"KfEa ˜ôo¬Gf?B,ØôÊ#ÄŽ]²G6×r‹…ùl˜Ó¶dáŸ@Ã?@š(s!Áf± JÔ‰…EAv_ÿı•ñkÖ¯_·*þÙgžYÿj~† ûXzAüÎt!?#?]ˆŸäÅ'ËÎJâŸx‘/mX½Z,'¤ç—$ ²^Z¹*^œ-äÇïÉ(É”fßZX Œß•žŸ?5lÂÔ}ba~‘H˜!ˆßYx¦söÓ6ç9]¤nîú¹wæíš÷Õ}Êùkî_t[lNìYàb'¬F‡Ã‘-AÚ‡ærS7¦ëŠŒJ“Ê2Rk×9)´Zß0tNÀÈüGm^¨«Vš t‘E^õPmuT7Ú\ƒ6“HÍ©a^V”dâUT€¤Òª©¦» ×Ú ÐmèøóÔ(A±×èl1Ò&HSÁ.žT “LÎä.JB·›™Û³c#)C‘ÇC˜öËÞ16ŠëáêËÁ•¤Ä«òµÛƒïÖPö€÷œÕë꨻áh±Õ´žmjòù::ÎÙãÎÉ}ë3OšJ+rŒ:(%Sz û¿êF;«+@_n2”ê(UVᆠ5æ–¦êvW]w¦±á=òýc½[¦åSê‘ÌætÈX’!ã—'ž¬ð€Ã¶  ~ènÒ_¥Î+Õ›ôtl¤~zdľ…jo±#u¨•ûÕîW3ÿÂ<È,eK¸¸é.ZŒ¡h)Åx/—ŸÖôioðëÞ1ºã½Ñö Oufmã—1œÔ­tîC¹é@ÆNDቧÃ,ôÔöÄõ‰¹f›ÙvÒ£õÊóGnîðï“Åä3"¦œyök†ž¸{ïܽèŽaÿyp’Õ*Ðjµ 1Po1o2ñ‡·¹võôØx zþ§¤oþâWýyïÌúeÙzèh>£×lóÙ¨Ø ?„'XaÖÄÁÏØhË-®š¯Ó¾¤L Œ„ŸFW9èm”‚²Q Š{Ía–ÑÌéÌÛA«ÒƒQ_Nl溻\XKäלÿßÎpRŠ˜¶›ü¨kqžB î€Z¡^ž¯ª,)ö)Í áC)»ee—PÊ!~èè÷ÖS"Ó?-y^TòÞ€#ì¢b#딡󟠄ñw00«/¢/±#K"n¦òÀHªeM õµÝCÇú™¹i Éß/oÓRê!¼M”¼À°Š“œÃ2jØý`&[„¥.ÍðfÒ¹£Ÿ¡§{QìP«ªY\O7 í©î|g–É:ÿÍ;m] ~‹îb‰/£Dó¯±Q]d9÷gœZR2úö¶¹l¼Ú ¬H)’ê©ãÌ3³×N‹¢ kæß9›™^0UJ@ÎK ¾ô»wÑŠ*O*L†rU¸%=“ª”ZzÝÖ{£=<ã«5cî+ìH0òwê?ìd]y•mÑÑ»·O6t®ñƒ“_ÒîF›‚ä;Ù©«2™jE5ؽf{µr6·üæUc±@™¯-£3 TG!\yS|z ;Ð’ò¥uÉ:à4Ô÷¹jëºkÞ‹ g‡Ü¤/ÓSÒí’·s€,ÓúC33N¼43ã³ãÎ+¿Cy£ì%rM•ÚB€X#.+’–”)D@&ÂÚ!ßhïéªf[Õ÷0•¹_u Ž‘ñw$gú:‚ƒÍTKò%Mþ ú®N‡?ìFËày¡¨FVZT*PY*]jÊSÁ!8žÃ°åÙå¹»ßr‡áJà g«³…îEÏØªzÀÁ‹¢(“h ´Ô_#,ê9MdÞ4Ð62Jf…HSˆc3Xܵ{.{nóÔŠKÉã—O©О´NAÿ÷ìW½ã%l´dÃG¾Ñ64Ë×C7_=söWÜë´•`2¨0õ•Lzmü[úQú}SSE—JÕ…Zj3¨:šzáþÊ‘–ÿ7ýKo+:Öö.*‡Y¹)9 Š È«91l7÷áãeÕiŽrÿÞì hÖˆø²#–§ ¡ÎOú§cŸÿ9;Rú;®¾FíÐA%ZõžMD‡ÜŸ×3>¼×u>1::0Õ0h¬=žViÌÀ)PТw¾¡á#½Û˜EÌbf³â¹sï}ñó‘í¡b¿Y5Âß,|sÔáò/žcrD Esor?9ó¼NÇ¢÷±Ik?ŽHÂ(áãÚKpÝÄ{± ï†Ï§w^²ô‹œÀ\ÝV_ÕPÕéNlßFð˲T†<Бb¿¢®¡o­ÃõÇ÷e§«´T1¿|3˜DÌœ¸mQ1}•³e’P¼FÉVž¶¨<…A¹¿«ÊÇ”=EEÒÌ´QßÝSèI´(µg4@^ E–` b0cêo‡šX&/«¤Åáh‘ ˜X†bVƼÜ×ÝÐÚF—®S¨çfó só”|¼b‹í6«9ä…d½ÜS&É/ÏH}Š8ݼ^K•¡ƒ˜¯N±sË7|±k{é¶ëýݸ?¸Öêq0jèoÉèË©~{1*$¯Ùü? É:½œb–OÞšjzžç{ØjsRÕƒ`©ê÷ Öœuuzúƒ=mmÿ°"œœ&Èäã¨5àË…Ñ ³"¦/ßù%;Ò€¾à¢˜æ_ ?'oÿhœYH1õß9Oöþ}8tCŸµz¡ßÐý÷á-b>z ÕÌFÎç¾¤Ü “F/§åIªL("—ÿ^øYøžêh]0ªlW^aB¥w°Î6`ñÑØ%C„‹ÙÝûÈ7ž3†|Óå“÷§z+[¥=ܤp´ìå´ÌƒIêTܳÑûÓœ<ô -¾Ü…>›`sýŠÚÒ’’RAa£Îjhl¢ª'…ÜïÄ®ÖÑè ÓÞn°Ôœ¼çnô´AàÿØ>%Z±ôÍOÐIdé5OSª.ñ_9¸ aºY‘÷"ˆë<á´ºÍNË œÊätj¶k›4´©Ò ‹6I/X«m–€•jE[Ábqƒ™çU”š¬ç”b§è$»™×âÀ¤W€‰¬Àái³ƒÕJY-`q7 GCAô¢¶4.¹z9óÐdgœš¯U®²œ#ÃuÑ¿ GàœuFá¼a$Êk†;ZΑƒØ,7 ­3øaªPcT}ÊÐ㬻×f4þ¿†åŒÆ0©Ÿé‡Oyw(ØpSu]5—¾•ŒÂÕeÅ‚ow0«¨~b)¾XÑEƸ}c¾^›ÇìÆÃ.]KéAhØùc-ˆŽ]$QxLVÚmj7B!ÈÔr­’™”Åí‰Ô,õW̼|8»Á¥ ³à¸!k&ËšýneŔʫGÌ¡…¾>ù_–2³®xµ`w9=Q-"P."Xhb³ÑåȃÜzyc~vn^NQP:ÙÞ¦60?á6Jë ³ò³ó…µ¥­]m'Û¨XQC$у„f‡I÷Ì Ï»q5/f]pþÜcþ|‚øo×tµendstream endobj 464 0 obj << /Filter /FlateDecode /Length 160 >> stream xœ]O1ƒ0 Üó ÿ €*± º0´ªÚ~ 8Ê€…0ô÷%!tèp–Îw'Ÿå0^G¶ä#8|QcYZÝ`¢Ù²¨Ðcay⢼ÃMù÷Çì2¿«…äóÒ^òª>Bè4­^!Å3‰®ªúΘ^ë?©&SœMÛ4˜ý§’¢©Äyp 8榹I*`™~ÏxçS vˆ/BSmendstream endobj 465 0 obj << /Type /XRef /Length 365 /Filter /FlateDecode /DecodeParms << /Columns 5 /Predictor 12 >> /W [ 1 3 1 ] /Info 3 0 R /Root 2 0 R /Size 466 /ID [<6deda6c40034dd66080a92682b20be17>] >> stream xœcb&F~0ù‰ $À8JŽ’XÈÿ BµË€l¶tPjyQ9šZFI|©EõÝ2`:¹7šNFI|éÄ`(\Ü5šNFI|éÄ®TûH-¥ÙÙ R° DrŸ‘œ[ÁìK Rø;ˆTü"Ù&IF“• ¶Hˆ”‰‘gA$k'ˆdù6ÓD2Þ§ÃW`5`0[xØFi)m Ö{,ûD29ƒÙ`2IÄLFƒMÛ 6ç7È=" ¶X=ØU‘`—ëÕü“`sÄu@¤dXå°;sÀ>ⓜ ’?D²ßD¸“OÌ~Šp”ËŠNÙ.ĉ“Ì߀MKÛõDJÍk©ç· …À®«”Ї$8¬$O‚][6-A2…‚Ù"{w‚Ãð;B„ßòŒÈä©ßv[+ÈUœ™`ñÍ ’ l>¯2ØÅ`»ÖƒCRª¼da Pè endstream endobj startxref 540157 %%EOF vcdExtra/inst/doc/extdata/0000755000176200001440000000000013163461153015211 5ustar liggesusersvcdExtra/inst/doc/extdata/tv.dat0000644000176200001440000002541213163461153016340 0ustar liggesusers1 1 1 1 6 2 1 1 1 18 3 1 1 1 6 4 1 1 1 2 5 1 1 1 11 1 2 1 1 6 2 2 1 1 29 3 2 1 1 25 4 2 1 1 17 5 2 1 1 29 1 3 1 1 10 2 3 1 1 10 3 3 1 1 12 4 3 1 1 8 5 3 1 1 7 1 4 1 1 20 2 4 1 1 24 3 4 1 1 26 4 4 1 1 14 5 4 1 1 40 1 5 1 1 20 2 5 1 1 15 3 5 1 1 28 4 5 1 1 11 5 5 1 1 13 1 6 1 1 26 2 6 1 1 91 3 6 1 1 64 4 6 1 1 2 5 6 1 1 30 1 7 1 1 27 2 7 1 1 32 3 7 1 1 24 4 7 1 1 10 5 7 1 1 21 1 8 1 1 36 2 8 1 1 63 3 8 1 1 59 4 8 1 1 54 5 8 1 1 77 1 9 1 1 50 2 9 1 1 36 3 9 1 1 26 4 9 1 1 13 5 9 1 1 38 1 10 1 1 68 2 10 1 1 18 3 10 1 1 35 4 10 1 1 3 5 10 1 1 34 1 11 1 1 34 2 11 1 1 9 3 11 1 1 53 4 11 1 1 14 5 11 1 1 23 1 1 2 1 8 2 1 2 1 11 3 1 2 1 14 4 1 2 1 6 5 1 2 1 10 1 2 2 1 38 2 2 2 1 9 3 2 2 1 15 4 2 2 1 11 5 2 2 1 8 1 3 2 1 22 2 3 2 1 15 3 3 2 1 20 4 3 2 1 9 5 3 2 1 8 1 4 2 1 38 2 4 2 1 28 3 4 2 1 32 4 4 2 1 24 5 4 2 1 20 1 5 2 1 27 2 5 2 1 7 3 5 2 1 15 4 5 2 1 9 5 5 2 1 15 1 6 2 1 54 2 6 2 1 5 3 6 2 1 5 4 6 2 1 5 5 6 2 1 8 1 7 2 1 26 2 7 2 1 7 3 7 2 1 1 4 7 2 1 7 5 7 2 1 16 1 8 2 1 39 2 8 2 1 10 3 8 2 1 19 4 8 2 1 21 5 8 2 1 11 1 9 2 1 19 2 9 2 1 8 3 9 2 1 9 4 9 2 1 13 5 9 2 1 12 1 10 2 1 23 2 10 2 1 18 3 10 2 1 15 4 10 2 1 11 5 10 2 1 15 1 11 2 1 10 2 11 2 1 7 3 11 2 1 5 4 11 2 1 12 5 11 2 1 10 1 1 3 1 0 2 1 3 1 15 3 1 3 1 6 4 1 3 1 17 5 1 3 1 9 1 2 3 1 31 2 2 3 1 43 3 2 3 1 7 4 2 3 1 67 5 2 3 1 5 1 3 3 1 12 2 3 3 1 26 3 3 3 1 6 4 3 3 1 32 5 3 3 1 8 1 4 3 1 30 2 4 3 1 10 3 4 3 1 47 4 4 3 1 27 5 4 3 1 34 1 5 3 1 13 2 5 3 1 9 3 5 3 1 19 4 5 3 1 42 5 5 3 1 15 1 6 3 1 14 2 6 3 1 67 3 6 3 1 12 4 6 3 1 97 5 6 3 1 14 1 7 3 1 19 2 7 3 1 23 3 7 3 1 25 4 7 3 1 44 5 7 3 1 13 1 8 3 1 10 2 8 3 1 36 3 8 3 1 42 4 8 3 1 80 5 8 3 1 22 1 9 3 1 14 2 9 3 1 29 3 9 3 1 17 4 9 3 1 26 5 9 3 1 13 1 10 3 1 11 2 10 3 1 25 3 10 3 1 27 4 10 3 1 24 5 10 3 1 10 1 11 3 1 10 2 11 3 1 17 3 11 3 1 18 4 11 3 1 12 5 11 3 1 14 1 1 4 1 11 2 1 4 1 6 3 1 4 1 5 4 1 4 1 14 5 1 4 1 8 1 2 4 1 10 2 2 4 1 8 3 2 4 1 12 4 2 4 1 10 5 2 4 1 10 1 3 4 1 11 2 3 4 1 6 3 3 4 1 10 4 3 4 1 7 5 3 4 1 3 1 4 4 1 29 2 4 4 1 7 3 4 4 1 66 4 4 4 1 23 5 4 4 1 10 1 5 4 1 9 2 5 4 1 8 3 5 4 1 19 4 5 4 1 9 5 5 4 1 9 1 6 4 1 6 2 6 4 1 10 3 6 4 1 5 4 6 4 1 6 5 6 4 1 5 1 7 4 1 9 2 7 4 1 10 3 7 4 1 19 4 7 4 1 2 5 7 4 1 8 1 8 4 1 75 2 8 4 1 41 3 8 4 1 47 4 8 4 1 31 5 8 4 1 65 1 9 4 1 0 2 9 4 1 0 3 9 4 1 0 4 9 4 1 0 5 9 4 1 0 1 10 4 1 0 2 10 4 1 0 3 10 4 1 0 4 10 4 1 0 5 10 4 1 0 1 11 4 1 0 2 11 4 1 0 3 11 4 1 0 4 11 4 1 0 5 11 4 1 0 1 1 5 1 65 2 1 5 1 46 3 1 5 1 45 4 1 5 1 52 5 1 5 1 40 1 2 5 1 62 2 2 5 1 65 3 2 5 1 65 4 2 5 1 31 5 2 5 1 52 1 3 5 1 42 2 3 5 1 56 3 3 5 1 40 4 3 5 1 58 5 3 5 1 37 1 4 5 1 122 2 4 5 1 100 3 4 5 1 104 4 4 5 1 63 5 4 5 1 53 1 5 5 1 66 2 5 5 1 59 3 5 5 1 51 4 5 5 1 71 5 5 5 1 51 1 6 5 1 73 2 6 5 1 57 3 6 5 1 71 4 6 5 1 58 5 6 5 1 44 1 7 5 1 58 2 7 5 1 71 3 7 5 1 67 4 7 5 1 66 5 7 5 1 51 1 8 5 1 109 2 8 5 1 111 3 8 5 1 130 4 8 5 1 135 5 8 5 1 90 1 9 5 1 77 2 9 5 1 112 3 9 5 1 100 4 9 5 1 85 5 9 5 1 111 1 10 5 1 113 2 10 5 1 127 3 10 5 1 109 4 10 5 1 93 5 10 5 1 107 1 11 5 1 75 2 11 5 1 87 3 11 5 1 92 4 11 5 1 59 5 11 5 1 68 1 1 1 2 18 2 1 1 2 49 3 1 1 2 30 4 1 1 2 25 5 1 1 2 18 1 2 1 2 16 2 2 1 2 84 3 2 1 2 106 4 2 1 2 17 5 2 1 2 69 1 3 1 2 7 2 3 1 2 19 3 3 1 2 16 4 3 1 2 13 5 3 1 2 10 1 4 1 2 73 2 4 1 2 59 3 4 1 2 53 4 4 1 2 26 5 4 1 2 44 1 5 1 2 41 2 5 1 2 18 3 5 1 2 15 4 5 1 2 19 5 5 1 2 15 1 6 1 2 25 2 6 1 2 94 3 6 1 2 80 4 6 1 2 7 5 6 1 2 50 1 7 1 2 29 2 7 1 2 31 3 7 1 2 19 4 7 1 2 14 5 7 1 2 5 1 8 1 2 65 2 8 1 2 128 3 8 1 2 116 4 8 1 2 99 5 8 1 2 80 1 9 1 2 29 2 9 1 2 14 3 9 1 2 22 4 9 1 2 16 5 9 1 2 32 1 10 1 2 65 2 10 1 2 14 3 10 1 2 31 4 10 1 2 6 5 10 1 2 44 1 11 1 2 23 2 11 1 2 8 3 11 1 2 15 4 11 1 2 9 5 11 1 2 33 1 1 2 2 13 2 1 2 2 24 3 1 2 2 22 4 1 2 2 7 5 1 2 2 20 1 2 2 2 61 2 2 2 2 20 3 2 2 2 49 4 2 2 2 21 5 2 2 2 23 1 3 2 2 16 2 3 2 2 13 3 3 2 2 7 4 3 2 2 8 5 3 2 2 26 1 4 2 2 76 2 4 2 2 83 3 4 2 2 152 4 4 2 2 86 5 4 2 2 74 1 5 2 2 22 2 5 2 2 26 3 5 2 2 28 4 5 2 2 11 5 5 2 2 11 1 6 2 2 50 2 6 2 2 11 3 6 2 2 6 4 6 2 2 7 5 6 2 2 5 1 7 2 2 27 2 7 2 2 6 3 7 2 2 6 4 7 2 2 8 5 7 2 2 45 1 8 2 2 63 2 8 2 2 23 3 8 2 2 41 4 8 2 2 56 5 8 2 2 22 1 9 2 2 18 2 9 2 2 18 3 9 2 2 10 4 9 2 2 39 5 9 2 2 24 1 10 2 2 10 2 10 2 2 13 3 10 2 2 5 4 10 2 2 19 5 10 2 2 14 1 11 2 2 14 2 11 2 2 10 3 11 2 2 2 4 11 2 2 8 5 11 2 2 9 1 1 3 2 26 2 1 3 2 26 3 1 3 2 21 4 1 3 2 25 5 1 3 2 27 1 2 3 2 71 2 2 3 2 77 3 2 3 2 4 4 2 3 2 73 5 2 3 2 9 1 3 3 2 22 2 3 3 2 22 3 3 3 2 4 4 3 3 2 14 5 3 3 2 10 1 4 3 2 110 2 4 3 2 76 3 4 3 2 65 4 4 3 2 41 5 4 3 2 110 1 5 3 2 35 2 5 3 2 16 3 5 3 2 17 4 5 3 2 13 5 5 3 2 34 1 6 3 2 20 2 6 3 2 149 3 6 3 2 24 4 6 3 2 102 5 6 3 2 26 1 7 3 2 14 2 7 3 2 11 3 7 3 2 25 4 7 3 2 21 5 7 3 2 20 1 8 3 2 12 2 8 3 2 77 3 8 3 2 112 4 8 3 2 94 5 8 3 2 102 1 9 3 2 23 2 9 3 2 32 3 9 3 2 11 4 9 3 2 10 5 9 3 2 17 1 10 3 2 44 2 10 3 2 12 3 10 3 2 12 4 10 3 2 12 5 10 3 2 7 1 11 3 2 5 2 11 3 2 12 3 11 3 2 5 4 11 3 2 6 5 11 3 2 5 1 1 4 2 22 2 1 4 2 31 3 1 4 2 13 4 1 4 2 23 5 1 4 2 14 1 2 4 2 15 2 2 4 2 13 3 2 4 2 14 4 2 4 2 41 5 2 4 2 15 1 3 4 2 13 2 3 4 2 17 3 3 4 2 9 4 3 4 2 12 5 3 4 2 8 1 4 4 2 70 2 4 4 2 37 3 4 4 2 145 4 4 4 2 51 5 4 4 2 30 1 5 4 2 26 2 5 4 2 23 3 5 4 2 20 4 5 4 2 12 5 5 4 2 11 1 6 4 2 7 2 6 4 2 5 3 6 4 2 18 4 6 4 2 3 5 6 4 2 9 1 7 4 2 7 2 7 4 2 16 3 7 4 2 19 4 7 4 2 12 5 7 4 2 13 1 8 4 2 180 2 8 4 2 142 3 8 4 2 127 4 8 4 2 156 5 8 4 2 269 1 9 4 2 0 2 9 4 2 0 3 9 4 2 0 4 9 4 2 0 5 9 4 2 0 1 10 4 2 0 2 10 4 2 0 3 10 4 2 0 4 10 4 2 0 5 10 4 2 0 1 11 4 2 0 2 11 4 2 0 3 11 4 2 0 4 11 4 2 0 5 11 4 2 0 1 1 5 2 61 2 1 5 2 70 3 1 5 2 44 4 1 5 2 45 5 1 5 2 51 1 2 5 2 34 2 2 5 2 57 3 2 5 2 55 4 2 5 2 35 5 2 5 2 28 1 3 5 2 37 2 3 5 2 58 3 3 5 2 28 4 3 5 2 49 5 3 5 2 38 1 4 5 2 255 2 4 5 2 177 3 4 5 2 141 4 4 5 2 85 5 4 5 2 104 1 5 5 2 51 2 5 5 2 67 3 5 5 2 63 4 5 5 2 43 5 5 5 2 41 1 6 5 2 70 2 6 5 2 26 3 6 5 2 57 4 6 5 2 23 5 6 5 2 44 1 7 5 2 30 2 7 5 2 39 3 7 5 2 37 4 7 5 2 51 5 7 5 2 41 1 8 5 2 114 2 8 5 2 102 3 8 5 2 104 4 8 5 2 152 5 8 5 2 86 1 9 5 2 72 2 9 5 2 54 3 9 5 2 62 4 9 5 2 54 5 9 5 2 49 1 10 5 2 44 2 10 5 2 57 3 10 5 2 41 4 10 5 2 35 5 10 5 2 52 1 11 5 2 78 2 11 5 2 27 3 11 5 2 11 4 11 5 2 35 5 11 5 2 49 1 1 1 3 146 2 1 1 3 244 3 1 1 3 233 4 1 1 3 174 5 1 1 3 294 1 2 1 3 151 2 2 1 3 181 3 2 1 3 161 4 2 1 3 183 5 2 1 3 281 1 3 1 3 156 2 3 1 3 231 3 3 1 3 194 4 3 1 3 197 5 3 1 3 305 1 4 1 3 83 2 4 1 3 205 3 4 1 3 156 4 4 1 3 181 5 4 1 3 239 1 5 1 3 325 2 5 1 3 385 3 5 1 3 339 4 5 1 3 187 5 5 1 3 278 1 6 1 3 350 2 6 1 3 283 3 6 1 3 264 4 6 1 3 198 5 6 1 3 246 1 7 1 3 386 2 7 1 3 345 3 7 1 3 279 4 7 1 3 211 5 7 1 3 245 1 8 1 3 340 2 8 1 3 192 3 8 1 3 140 4 8 1 3 86 5 8 1 3 138 1 9 1 3 352 2 9 1 3 329 3 9 1 3 237 4 9 1 3 110 5 9 1 3 246 1 10 1 3 280 2 10 1 3 351 3 10 1 3 228 4 10 1 3 122 5 10 1 3 232 1 11 1 3 278 2 11 1 3 364 3 11 1 3 203 4 11 1 3 117 5 11 1 3 233 1 1 2 3 337 2 1 2 3 173 3 1 2 3 158 4 1 2 3 196 5 1 2 3 130 1 2 2 3 293 2 2 2 3 180 3 2 2 3 126 4 2 2 3 185 5 2 2 3 144 1 3 2 3 304 2 3 2 3 184 3 3 2 3 207 4 3 2 3 195 5 3 2 3 154 1 4 2 3 233 2 4 2 3 109 3 4 2 3 59 4 4 2 3 104 5 4 2 3 81 1 5 2 3 311 2 5 2 3 218 3 5 2 3 98 4 5 2 3 106 5 5 2 3 129 1 6 2 3 251 2 6 2 3 235 3 6 2 3 103 4 6 2 3 116 5 6 2 3 153 1 7 2 3 241 2 7 2 3 256 3 7 2 3 122 4 7 2 3 116 5 7 2 3 136 1 8 2 3 164 2 8 2 3 250 3 8 2 3 86 4 8 2 3 47 5 8 2 3 126 1 9 2 3 252 2 9 2 3 274 3 9 2 3 109 4 9 2 3 102 5 9 2 3 138 1 10 2 3 265 2 10 2 3 263 3 10 2 3 105 4 10 2 3 84 5 10 2 3 136 1 11 2 3 272 2 11 2 3 261 3 11 2 3 110 4 11 2 3 84 5 11 2 3 152 1 1 3 3 263 2 1 3 3 315 3 1 3 3 134 4 1 3 3 515 5 1 3 3 195 1 2 3 3 219 2 2 3 3 254 3 2 3 3 146 4 2 3 3 463 5 2 3 3 220 1 3 3 3 236 2 3 3 3 280 3 3 3 3 166 4 3 3 3 472 5 3 3 3 248 1 4 3 3 140 2 4 3 3 241 3 4 3 3 66 4 4 3 3 477 5 4 3 3 160 1 5 3 3 226 2 5 3 3 370 3 5 3 3 194 4 5 3 3 590 5 5 3 3 172 1 6 3 3 235 2 6 3 3 214 3 6 3 3 230 4 6 3 3 473 5 6 3 3 164 1 7 3 3 239 2 7 3 3 195 3 7 3 3 264 4 7 3 3 446 5 7 3 3 169 1 8 3 3 246 2 8 3 3 111 3 8 3 3 143 4 8 3 3 349 5 8 3 3 85 1 9 3 3 279 2 9 3 3 188 3 9 3 3 274 4 9 3 3 649 5 9 3 3 183 1 10 3 3 263 2 10 3 3 190 3 10 3 3 289 4 10 3 3 705 5 10 3 3 198 1 11 3 3 283 2 11 3 3 210 3 11 3 3 306 4 11 3 3 747 5 11 3 3 204 1 1 4 3 222 2 1 4 3 130 3 1 4 3 316 4 1 4 3 146 5 1 4 3 131 1 2 4 3 233 2 2 4 3 141 3 2 4 3 327 4 2 4 3 135 5 2 4 3 132 1 3 4 3 245 2 3 4 3 165 3 3 4 3 364 4 3 4 3 146 5 3 4 3 146 1 4 4 3 179 2 4 4 3 154 3 4 4 3 178 4 4 4 3 101 5 4 4 3 126 1 5 4 3 210 2 5 4 3 145 3 5 4 3 163 4 5 4 3 135 5 5 4 3 242 1 6 4 3 229 2 6 4 3 146 3 6 4 3 165 4 6 4 3 148 5 6 4 3 269 1 7 4 3 241 2 7 4 3 158 3 7 4 3 157 4 7 4 3 160 5 7 4 3 291 1 8 4 3 0 2 8 4 3 0 3 8 4 3 0 4 8 4 3 0 5 8 4 3 0 1 9 4 3 0 2 9 4 3 0 3 9 4 3 0 4 9 4 3 0 5 9 4 3 0 1 10 4 3 0 2 10 4 3 0 3 10 4 3 0 4 10 4 3 0 5 10 4 3 0 1 11 4 3 0 2 11 4 3 0 3 11 4 3 0 4 11 4 3 0 5 11 4 3 0 1 1 5 3 766 2 1 5 3 704 3 1 5 3 738 4 1 5 3 573 5 1 5 3 549 1 2 5 3 783 2 2 5 3 744 3 2 5 3 726 4 2 5 3 581 5 2 5 3 546 1 3 5 3 901 2 3 5 3 817 3 3 5 3 797 4 3 5 3 642 5 3 5 3 587 1 4 5 3 616 2 4 5 3 632 3 4 5 3 641 4 4 5 3 580 5 4 5 3 511 1 5 5 3 662 2 5 5 3 637 3 5 5 3 756 4 5 5 3 637 5 5 5 3 592 1 6 5 3 615 2 6 5 3 645 3 6 5 3 723 4 6 5 3 634 5 6 5 3 579 1 7 5 3 636 2 7 5 3 726 3 7 5 3 737 4 7 5 3 651 5 7 5 3 574 1 8 5 3 509 2 8 5 3 609 3 8 5 3 599 4 8 5 3 443 5 8 5 3 490 1 9 5 3 636 2 9 5 3 675 3 9 5 3 743 4 9 5 3 575 5 9 5 3 674 1 10 5 3 557 2 10 5 3 587 3 10 5 3 673 4 10 5 3 541 5 10 5 3 604 1 11 5 3 510 2 11 5 3 540 3 11 5 3 642 4 11 5 3 498 5 11 5 3 591 vcdExtra/inst/doc/vcd-tutorial.R0000644000176200001440000006547113163514377016343 0ustar liggesusers### R code from vignette source 'vcd-tutorial.Rnw' ################################################### ### code chunk number 1: preliminaries ################################################### set.seed(1071) #library(vcd) library(vcdExtra) library(ggplot2) #data(Titanic) data(HairEyeColor) data(PreSex) data(Arthritis) art <- xtabs(~Treatment + Improved, data = Arthritis) if(!file.exists("fig")) dir.create("fig") ################################################### ### code chunk number 2: case-form ################################################### names(Arthritis) # show the variables str(Arthritis) # show the structure head(Arthritis,5) # first 5 observations, same as Arthritis[1:5,] ################################################### ### code chunk number 3: frequency-form ################################################### # Agresti (2002), table 3.11, p. 106 GSS <- data.frame( expand.grid(sex=c("female", "male"), party=c("dem", "indep", "rep")), count=c(279,165,73,47,225,191)) GSS names(GSS) str(GSS) sum(GSS$count) ################################################### ### code chunk number 4: table-form1 ################################################### str(HairEyeColor) # show the structure sum(HairEyeColor) # number of cases sapply(dimnames(HairEyeColor), length) # table dimension sizes ################################################### ### code chunk number 5: table-form2 ################################################### ## A 4 x 4 table Agresti (2002, Table 2.8, p. 57) Job Satisfaction JobSat <- matrix(c(1,2,1,0, 3,3,6,1, 10,10,14,9, 6,7,12,11), 4, 4) dimnames(JobSat) = list(income=c("< 15k", "15-25k", "25-40k", "> 40k"), satisfaction=c("VeryD", "LittleD", "ModerateS", "VeryS")) JobSat ################################################### ### code chunk number 6: table-form3 ################################################### JobSat <- as.table(JobSat) str(JobSat) ################################################### ### code chunk number 7: relevel (eval = FALSE) ################################################### ## dimnames(JobSat)$income<-c(7.5,20,32.5,60) ## dimnames(JobSat)$satisfaction<-1:4 ################################################### ### code chunk number 8: reorder1 ################################################### HairEyeColor <- HairEyeColor[, c(1,3,4,2), ] str(HairEyeColor) ################################################### ### code chunk number 9: reorder2 (eval = FALSE) ################################################### ## Arthritis <- read.csv("arthritis.txt",header=TRUE) ## Arthritis$Improved <- ordered(Arthritis$Improved, levels=c("None", "Some", "Marked")) ################################################### ### code chunk number 10: Arthritis ################################################### mosaic(art, gp = shading_max, split_vertical = TRUE, main="Arthritis: [Treatment] [Improved]") ################################################### ### code chunk number 11: reorder3 ################################################### UCB <- aperm(UCBAdmissions, c(2, 1, 3)) dimnames(UCB)[[2]] <- c("Yes", "No") names(dimnames(UCB)) <- c("Sex", "Admit?", "Department") ftable(UCB) ################################################### ### code chunk number 12: structable ################################################### structable(HairEyeColor) # show the table: default structable(Hair+Sex ~ Eye, HairEyeColor) # specify col ~ row variables ################################################### ### code chunk number 13: structable1 (eval = FALSE) ################################################### ## HSE < - structable(Hair+Sex ~ Eye, HairEyeColor) # save structable object ## mosaic(HSE) # plot it ################################################### ### code chunk number 14: setup ################################################### n=500 A <- factor(sample(c("a1","a2"), n, rep=TRUE)) B <- factor(sample(c("b1","b2"), n, rep=TRUE)) C <- factor(sample(c("c1","c2"), n, rep=TRUE)) mydata <- data.frame(A,B,C) ################################################### ### code chunk number 15: table-ex1 ################################################### # 2-Way Frequency Table attach(mydata) mytable <- table(A,B) # A will be rows, B will be columns mytable # print table margin.table(mytable, 1) # A frequencies (summed over B) margin.table(mytable, 2) # B frequencies (summed over A) prop.table(mytable) # cell percentages prop.table(mytable, 1) # row percentages prop.table(mytable, 2) # column percentages ################################################### ### code chunk number 16: table-ex2 ################################################### # 3-Way Frequency Table mytable <- table(A, B, C) ftable(mytable) ################################################### ### code chunk number 17: xtabs-ex1 ################################################### # 3-Way Frequency Table mytable <- xtabs(~A+B+C, data=mydata) ftable(mytable) # print table summary(mytable) # chi-square test of indepedence ################################################### ### code chunk number 18: xtabs-ex2 ################################################### (GSStab <- xtabs(count ~ sex + party, data=GSS)) summary(GSStab) ################################################### ### code chunk number 19: dayton1 ################################################### str(DaytonSurvey) head(DaytonSurvey) ################################################### ### code chunk number 20: dayton2 ################################################### # data in frequency form # collapse over sex and race Dayton.ACM.df <- aggregate(Freq ~ cigarette+alcohol+marijuana, data=DaytonSurvey, FUN=sum) Dayton.ACM.df ################################################### ### code chunk number 21: dayton3 ################################################### # in table form Dayton.tab <- xtabs(Freq~cigarette+alcohol+marijuana+sex+race, data=DaytonSurvey) structable(cigarette+alcohol+marijuana ~ sex+race, data=Dayton.tab) ################################################### ### code chunk number 22: dayton4 ################################################### # collapse over sex and race Dayton.ACM.tab <- apply(Dayton.tab, MARGIN=1:3, FUN=sum) Dayton.ACM.tab <- margin.table(Dayton.tab, 1:3) # same result structable(cigarette+alcohol ~ marijuana, data=Dayton.ACM.tab) ################################################### ### code chunk number 23: dayton5 (eval = FALSE) ################################################### ## Dayton.ACM.df <- ddply(DaytonSurvey, .(cigarette, alcohol, marijuana), ## plyr::summarise, Freq=sum(Freq)) ################################################### ### code chunk number 24: collapse1 ################################################### # create some sample data in frequency form sex <- c("Male", "Female") age <- c("10-19", "20-29", "30-39", "40-49", "50-59", "60-69") education <- c("low", 'med', 'high') data <- expand.grid(sex=sex, age=age, education=education) counts <- rpois(36, 100) # random Possion cell frequencies data <- cbind(data, counts) # make it into a 3-way table t1 <- xtabs(counts ~ sex + age + education, data=data) structable(t1) ################################################### ### code chunk number 25: collapse2 ################################################### # collapse age to 3 levels, education to 2 levels t2 <- collapse.table(t1, age=c("10-29", "10-29", "30-49", "30-49", "50-69", "50-69"), education=c(":"ê{0΋e^$YU ¾R3}õ=R _ï¸ö´§f(ØÓžž]©¶ tá^*ÊTu`ñBQXφŸ ©Ûô§lrþ=©yVÏ<Ï“³òâ)eVÀî$MJÔ“q^Le!§Zé,™Ty¡F׊þÇ–ÿè¥ô€þƒ=qê>êµ@ñQ¿!:h€2th¡ÑŠÖÇê]÷̶õÐFÿlª4Yóm]'||sµWÛŽwñå|h»ö›¨Î>;0®>ûø‹ùCŗ¢âÙê#1 l€`÷ ,L/\V§ààF¬¾½ú¤|x»ëÑx=7þ;÷n=z¿Â 5Öú¨nžšï=(Á2ÌUä±ÉÛ°©Ì¹òsöºâøèrѶ~Ú:މg³…²!.±áòŠÉºt-×xsöûäÌ×V ›Z£bä‚éš/›Nì¦ýÝ–ÇE6ï‹Máa8¾ØF}mttýh$ÈòzÿHŠj‘¤õÏhvž¦ u»<ÿXwr]!fµ›spòcòfåpüp‡qú:ĺ9†<˜. Çæ7æ;#LŽò‹Š+CP)¬ŒÉOùŒùAùõa1ÆôPº1_09ªÆ¸ZË Ø}uÃîë‚Ié¶Õ UgÜ>åü§|Åd`î¹}€å³ëÜž¦j–ÛÏTÞ©uª™B;y˜æz¸\èÞͧó]Ûøc‘­•póaÛ.¶påI•pÓ1põ£í˜Ûüäbᛳ¦üµá¹ÚkòE1Ù&ŸÐáj¿í2a®aø6»¡ ä¥ìÆt¸æÖüÆð¹˜D„>KÎ6Úx¤ö/Yè#y8“º¥÷—¹êÕI%ÛêëðøÐA¾©›Ûo¸†éÄôqöp<6?(¿|nö8›\x1ÿ1²ÅÜç¦s›»Æß¥>\lälåôq}€Ê—/_8{]â å8û¨}ç»G0y—ZqÑIárys© ×}î“?šÝèëº)GS™èŽÏe’VóK—'ãÂ|× aÞçôëqL¦ßïB‡bgnÞŸù¾‚‰ë‚57¬ç"B>ªí!E;$ŽCAÛ:» vÇàElâk óÝŒŽ«ö êè8Ðîž!ãÈ5,˜»Ð°ý5Õk#"N:OÏ™Øh ‘ñ[×Ì)|&pkËmݾ404þÈYO=q¿–»bÓ~³ö5žY_‹ï!Xë Ü7XGÔ> ‰±ËÅî_*&?f |–ëÝ#Η]ˆŠ¡A›ÿøË’¥,Ū%÷õñ{¢ŽÞ§yq©»ù…”gº›—•ùÂ(“e¥;ù\ªcºþOeV.*1¬ )¿­.Z÷$?¿÷ă"¿8ÒFÝuÁ+õqssý›¸~ ¦I•Í %"V×ë? bpÖÀ.vcdExtra/data/Hoyt.rda0000644000176200001440000000115313163461153014361 0ustar liggesusers‹u•]kQ†O’¶‰% ^xDDDĦ*ø±öËRS´µ*f³YóµÉ³E/‚þ/ýý)½ôÒ_PÜí>o`Zؾ3gfÞ3sÎÌIsë R<(c²&›Ë„_(.dÃcJ—"¥êŒÉ] å?Ö›ýYã2xhÅ8B/÷À³Ð x ¬ao¢_+à°£sšä}zí?£—ÁupŸ¼k©¼Å3ûKžøÝe}üYJÚ•·¥º°¿eýfŠ?ÀÏ!þ>ëàì€wÀUðqV¬?ÿ—òûn³Ï5ô&™÷6ºIÙë©|eOÞGèºÕÑn'íŠW¼š%ë)¤ìïãk+•¿îAçj]À¯8õ‘êR| ì’ç3tõúsº£úcSû“Ÿîëx[ùŸÀsA>Cì^9Æ*ë¯ñ¯þT^ò›UÉ¿=ìmø~êk±_¿â5WOÁÞ)|ð¿dßö)ý=þc]vpľêGÝŸæä¼/ð×¹m ;èšõɨùe®æý?ï?øWÐuïš—xõ¿ò’¿Þ¯BŠ_}®þ{lyÌ÷×\«?4GâßùSok®;‡p9Zç‹Þß¼‰žã¤k!tØc7|(²Wñ-‘=¿é{žÛsQ—ZNß÷=´\Íï .6‚¾û…àœìuÿ«wÝ®'š…ê ×Ç9ÏZfYBEŠ„U kÖ%<„&+ê]{¾ÍÒŽ;>×%/R¯ê,Å…vp<ÕSÓžŒ‹ Ç9>²ƒ?Ññ¶Üo!œ¥yÏžŠ7£s ìÎy>gÑ÷&c£^vcdExtra/data/Vietnam.rda0000644000176200001440000000055013163461153015041 0ustar liggesusers‹ Í“¿JAÆ7{{Æ ‘ XØYH K;5Dl H¬äÈmP8sqÁ€•…­•o ­oâÓäü67#fñO“Ââ·óíÝîÇÌÜÜQ£½´!„r«©$–’Õ bùø\g½ðB¯fß!."nGfŠœ-SeÌÅz ãªZ¼<-5±LÄ¡sÞïÄaÊÇ?Mºa'K Ôø—6¸©xÊÁw˜±Ÿ-KæT‰OõúM£Ó3Ú¨VÒg- «+µo¨0ò؈;XNL¤Žþìü’í?ÛOuHñhì°Øe±ÇÃÒpšñí|®ÞÆiþ.öô:e°Ö@Ï:€° ¶A@wWÁ-y½€ Ü%ð nÀ=Ý!x"ßGp ÞÀ&€W$}‚øàŽ=þtÍcÏ­ðR=$)¯ )…ê³Æ4]:N“\ÕÙͦ-mþyž¿ÿм ³°Þ5¸B ê†äžvcdExtra/data/TV.rda0000644000176200001440000000107613163461153013773 0ustar liggesusers‹ ]’_ˆLQÇϽ;³3»ÚºQ‹ÁE’fÃe8Wð¦Ã¸÷ð²Änx¤s’?‚¿v‚Ì]'¶Ã:˜Ç܇è]°&3~ ßÑ>ñìÎhöáõÂRôW¨èÚ)°ý¶ÃGÝg úÔm=Îrâ8wAöœA¼;àÄ˯Âc†cÚ2,„à_Î<âKØ›€:Í|†~‚£ð-l¶=›x{à4üÒºféþ¿É›¨ûlÔz¹!'….o×:$µ~ñŽØGζ[¯XÎý þØ;u¤–×ðÅöËœ%>'þÔþrF“Ç{C< oÑ5{¿nx^ëzï™{¡5ŒØ½ä.›èsöNLà­%®Ô{½aës÷ê;’ºXãLÇ»mß‚¼ ó„Ø'û=üïµ6KÒy®R¦´a‚ÕãÒÒ¤UúÊÁ ¿™¤¹žæË|ó²j¥Ø·MG)¿ FÖÕA±7Òþ@½7/©•ìÈÓýÅMÌ_ÉD:;7Ò³cþœ~.–Ÿ‹åçbù¹(?™ÍD dÏ„K<-5ìÔ¢|gCv滲 î¸^%µQÕ¶Ï]R_(~©¨N‚¡­ÕÚf}Y!Í¿¥cvcdExtra/data/ShakeWords.RData0000644000176200001440000000104213163461153015732 0ustar liggesusers‹uÔOs‹QÇñÛ')ÿâTEEUDªÿ´Ú¤"­H•¤$ª4$e¦$$ÅŒUf쬬¬,x^‚à%XXZÚÙ‰ï“{®wX|òë¹çœÛL2yr©b

â’QɰU›ýɘä dÄÚ3ý¸µ×gÝköBÿ¹ÏÞ [ç!k¾ßšZs½V¶2jõMꟅ²žpÝ5÷I•‡[çä˜ÃrýymC ºQ}fmõ4ê/£fÓ{¬9-^Úíö{ûúòc·i®ÿóT­¸nt­Á>Õ/Ïo؇{ƒ‹vcdExtra/data/Accident.RData0000644000176200001440000000107413163461153015377 0ustar liggesusers‹íU»nÔ@=k{½¬! …G”9ò.á$Š hÅJ¢- 4cO¯-¼(!$Z*| ièéhé¨%@–3±Ç° „‡»(Å™;wæ>fî¹ö´fêÖ‚@ƒ.G9‡ T)wÌx^Ûaè{©—)‡(çÓy!¡„V¥b諳ˆ»"èr¶{£:éªîØÓÙ´\sìZ®Ô»ž+'{J)úIçø@貸]YË+Qì‹Xø*ý¢ë%QÌÙúÿéE ´Åý% Zo ðÆl;¯z¥.¯H6LueZÚ”Õý*‘êfu¨íýM÷ûH1²JW§ì«·„¾3s¾í­zÈTk."}+óÂÝ$n»á?±§Új[þ—üå'e6DÇÍ™1æäü/8AƉö†÷‰'ÔÏPÿDy‰Ræý 컌6‘ŒjãÏùO½ Œ1úèm>5â)qYW)oÒÿå[`×2ûi?°“ºqh~†™÷ìgàȱ4—ôÃiÙWô{FÙây\Jæ)aÌ<üaîqíÂCžå%0ñØ#oÕ Æˆ‹ÄQú1N/ˆsÌ=ÁïyÝË4«ï1Ì/bˆkÖxfw8„ô]63ȸ3´c*¯Ò¹ñšxL“iš·›Ð툟ž1wI±cÆ¢»$Š«Näç;K"䓤v±¸3»G÷&U|IžöˆC¯×[û Ý–ï&îäbL—ŒòõoüZòvcdExtra/data/Bartlett.rda0000644000176200001440000000032713163461153015221 0ustar liggesusers‹ r‰0âŠàb```b`bfb “… H020pi§Ä¢’œÔ’f>ß!¹BPi˜xÔ°BÛCå³ Tó™S2s/ˆ v£:¨,/17µ(% U+U ’guÌÉ,K…rX\RSÐ0ûå—C™lÁE™yéh X|òÁb`Ó‚3ò‹JМÀ µ¬‹&µÌa ÉÌ…±Ù|RóÒK2€¬èæ$ç$ÃÌa„ÙX’˜”“ QÍð!PÏû—vcdExtra/data/HairEyePlace.RData0000644000176200001440000000052513163461153016160 0ustar liggesusers‹ r‰0âŠàb```b`f`d`b2Y˜€# 'æñHÌ,r­L ÈILN*ãŠi8”$0€€Ck„.-€ÐP¾3„Ž…Ò¡PÚJç)@è&Ý“¡+ t4Ôœt(]yjĨ9PÚ fÔ¼©P÷-Q‚òw@è¨x*T]í]åç?€Ð\ ôR9½ö „ž•¯€Ò½PÌ9¡OÓö îQƒÒPÄ0@J3§dæ)^$Ĭ  (AUÆT–—˜›Z ”†ª«‡È³8唦BÙ¬>™é%P›ojJfi.L™KbQ6T++LÌ Ë0×¥¦àÑ5ß ˜`¦0A93K2òR‹‹a.vLJ-JIMÍCó +Ô`ÝÌ0{© f“Â9¬Ð”Çð†ä‚aµvcdExtra/data/Hauser79.RData0000644000176200001440000000050413163461153015271 0ustar liggesusers‹ r‰0âŠàb```b`âfd`b2Y˜€#s1‡Gbiqj‘¹%³0TÀÌ ¤%¡j˜ b ,@ÌJ}1·°å¤–¥æYP Q–Ð?_Û'Îf-€3}òá*Ü‹rÑÌeMÎI,† ·,-1¹$¿Èú‡æetÌ„3cÁ,X0+&9é?µ<É 3 «¿ø@þr˜&Á ºè„nŠ»0@è¶Pu :ÿ„ný¡= ê:>@õ;@è.¨ºY:ª®j^Ϩ>¨}+Aè\¨|%T¾,BWCù3u ôˆ:ôhÍKÌM…E+3,”‚óó`!á–X’‘Z³¢ÔB48‹òËõ`¦€’SøÿÿÿsË•’X’¨—VÔ àÖÃt¾cvcdExtra/data/Donner.RData0000644000176200001440000000273413163461153015116 0ustar liggesusers‹¥WÍoE_ÇI'vì|@‹¨ÐJ RHH• "MS×i­:ÐÒ{¯¼Þ1³» éî\ø¨¸rçî¿ÎH¥¥¥¥õ•PáÍÎÇÎLÖn–žgç÷Þ¼yoÞì^^¿úÖôÕiÇqÆœ±2PÇÇà/çŒ;­“ ÀÔqò 0›€±ã5 B"õt*h4ö)oÓ5.hJ£‰g¤Â:fÑ´À'5*X6‡ÑôfÛ&Ÿ™ ïc?„§J¢£«ãÀJŸŸmµÅóÔ [Wˆ¿+%ÏQ´Cɽ€C¼ƒéž˜¶škqÔl+Å…­˜‚†º×‘f/EíÔÐeŒ[–·M…ÒYÂ.jF Êy$‹é¸xΉÍ½ô¶ØŒ¼ØÔ—€N ^Iàoˆd½§mÓ±4#d&¾ô‚ÀrB×Iaÿy1ž¼²ß:-t2=E Eak^è;#Æ¢ð}QØp…6n½tBÄ‘#³óжvJðžsxñ3½Ë"FW‹±$Ö2’ö²ÀOl‘?«Ã+VnÈsÙ{”Žœµ&K–…gù—¥;Ëþ0?Gùø4ûkØqûÇá'`Lž€ ÜE>–g‹='’y)9ú¬8¹È"ý§c¹ ÌÆ³~¶¬­s˜Í,Ÿ²ÖeÙÏâãÙÏ£öÃþeèɳ³x­?øý›}öûŸüæÿŽýÁàÌÇðÔÜY~”Œ|~TîÖ4¼?øéî©L¹Û7{‰ž‡§?7äyý”1çüþàæ»õd¼Wçó[ß¹|þ}Å\ÿ7Ç|õþˆøåú üv]‹OΟyŸ:ßÞHÖ=>®ïðó_fÚ»ó‘kÈýúµ1X^Opÿåž±þþÁšßÃótw©’‰ÿùIÖ¾<½>dž¸Ÿé\Ž{Äóõªðëž÷ïÜÈôçq‡ùouÙ“&ë—ç¯6#Ùx’idßçÑu`ÇìQù¶×œ¤A™²êby÷O¤ý¬ëùò}$öḋøSùÆtßÛ‡— ¡¨…QÔÒ §[(BË»lYö ”,ë>\“‘®ñi¹”¼;-¹g[ˆJ›óÜBtQ¹ç—§(8ÕD3Çf$FÚí¬€ê(¢^³cY¨[¥¶¥:ŽÔK–Ä^—HSåÕ˜îaºä®µõ•SÅu qš^Í®qa¿0Âø^ÔFîZÜõ¤wë:rë„HOæ Ά YK6( š^èž]N­sÎ9L¨Êü¼‰j°`p<ä®2’djÚ`5ÉŽm´†Q [´†¼tǨ'Mb Ô±o£Û¨ª÷ã{Õf–Ô>ÍpP/—YÉÚ2W_ñ|ßC]% 3°çG*ãEöùcj•£ k Šd2J€FÜý,KдU‘pSÜõRqþ•`6—Âf,a¸k2¥ó’ÇŠ£ã{{Å^•rܪôLòª$€V€T9K¼Fö=•[%ÍҨ׌Ä/B]ÚÁYû¾‰|ŸPvVjqGµŠMèBz [UiySÊΙ©¡´IHgØ >IÏïLRàIÓ&?Å–Ü•–Œ«œ‚zÕ.X°v€SV½Íé¥m¯x2«zaŒ×b¥v^}ý~Êc†§s`«¨§Š»õë³ÚÚô U»—”‡ðIÙkò&åÖÀ U$¯†´ì¦(ÖdY¢IηT‘pÒ/NZ†ÏX =±OXí6à°éý /"¢£3ãŸaŸÀæY/k–™Y[-`N¡¬ W‘òjž3¶Û¤‹B7Éœ¹äCd-R5[`x\­ˆmnˆ{2ÉS äï£VzŽmŸ##M®ØèÂu`†Sl‡­¤G¤—Ê\£çE׸ïÅaKáJ#Bü 2o¦Ê6wH ³NƒwUº\» U¹Š}UŠåOZŠ'þ®$Wß:¡J|VÃÙ1åßNOþ8·6rêvcdExtra/data/Burt.RData0000644000176200001440000000052513163461153014601 0ustar liggesusers‹ÍT½NÃ0¾Ú ÐH H,Œb`ê@„ b†‰®§Æ¥†×ågãAyž Á ¹ ›@È€À’}w¾ßÏ?7Œ÷‚q 8ëã†õ˜Y:àA7JðMÃû†®ºSèÛMÖròvÓªyEŠ{!ç† m‰äUò|(bÚ ºvø‰sò¯¢Nq¢Se¸å7çЄã—ís,sû§ÉÕL“p–<Öçäõ3¼”ÿe /£ÊÏQ©ô0^&±h‘ýZ‹Ñ»@)éJG³4ÿ¤Ía#‡õ QXR°ÇÁ«MIïÚ¹£²ÛªßweŠßd÷)ÏvƒßZIwºzvÞhß±)‚Ý nñFP7ðéüO>º‰wôdô´/~×êH£^(Q5 %îœØ]óNû?ìÙ,Y–½|qAŒûSe\àýJ—o<ÏûèÅvcdExtra/data/Fungicide.rda0000644000176200001440000000034113163461153015331 0ustar liggesusers‹ r‰0âŠàb```b`âfd`b2Y˜€#s‚°[i^zfrfJ*P P@ÀA„  tP„Ö€ò9 4”…Ò:PZMšÞé¶ãnLD>1Ï#Ÿ! |,!†{’[KÄVQ`—-"®•Lã54À§ù1ÈŒ!1 -—»ãn|À?&åÒuŽŒN„µ©V“Þlç2•Rô]9ß¹½fZmî]+:UÖ©ùïý—ª“YÞË´A6ªÝˆ÷Íaø÷½b4?ÿ§;`•P£mªl·Á&˜+.®ƒgp†`«P‡@wp–À#x­7½:«`õ6xª¿ˆâW¢zö5Zv!Œ«˜ÕÊ¥Á®7¿ø‰û<ã­Äà5Èè»V¥Sõ°äyþáÌ>¼¥W :vcdExtra/data/Titanicp.rda0000644000176200001440000001574213163461153015222 0ustar liggesusers‹íÝM¯$W}Çñ¶Çv RÐHaÁ$ËB BÝó|Ž˜šñ#˜óˆíÄG6X3ÆIVÜ]^AVÙDY’7Yä°É>R)Ê*+ÈŒoýÏ÷ßÿ¾ÕUÝ·ïƒÍ-©9U]¿SçœúTuÝAþÁË?öŸÿñçW«Õã«þ÷Ñÿ<ñøÃÿylõÄês˧ôî‡wúîÛ¬V7þèáòSË?\­žüÊ'Û\®?ןëÏõçñëÏõçúsý9àçÆõçúsý¹þ\®?ןëÏõçúóûýY{7ùÔ{÷>º÷Þƒ‡s7?Y{òí ÷àC™õ?}GfÃýwÌîO¾ýÞݲw?æOî¾ýáÏî?œûíŠ7òD~Ú/þÇÌ:û?µýiÛÍ}wÚòÒÜ¥ûœVN}N;·Óö]ÒÛ²–¶ÍÒöZš9ÕSǘË_zÜ©>ZR¹·SKÇÊT?NÍŸ6v–Œ×mm1u.suªÏܺ%ãbÉøš{+¸dÌlk«%}9uÞKìÚÖ÷Sãl*{×ëmÛ>ÛÆÖܸÙå¼·ß¶vÞg\/©ó¶õKÎqéXß–=w/KÆÒ’q³¤®K޽´}1Vwé'›5×SÇ_zýî:Æ·£¹ú-iË©mv¹Ö箣]®­]®émç;Õ·»Ôa×:âXKöߥ—ôß\;M]Kë¿ÍÁ]ë=wÝ/©Û’1{–þ;ËØ;ÄøØ6F–85U÷%}¶Äí%û-=¯}Æê®}3çèiËÛÚö¼úü´:,½vçê;wÏ9«­KmXR‡©ëoɽd[{-5r®¾û˜wÖ1³´oæêxè:í3nÝKÆÙܘ=D;ízœ³º½´M§®«]®û³´Ï¾†.ØîÑ«ÊÇ·:y;ùøøvò‰wÞ½'o4Ÿ~ðóû½ûѽñµæ Ùvû›Ìm­¿­õ¶ÉuZM 7ÕÃsw²Ó²¦´]RN }Îmjû%wŽmç¼m›]Úxiþ®çpÚ¾ÛúhÉ•;5&¶Õm[Lm?×ïÛ4™¯sm·í\¶µç¶6ÞÖ'sã®çút®—fïÒ»Œõ]ÛlI[OåÌo—¶ŸK®#»Ý\½–Ž÷¹ãí²¼ä³Ë9ÎÕyÛ8˜Ê›‹Kúaj ìÛ/K¶_Ò¦‡Z?×§µÿ¾ÇXêë.ÞlëËmÙKúlnÜ-é·©º,'sãf×kdiv“KúlΜ}mÚ§®s}¿¤½·ÿÒ>\²ßÜ5²íœ–´å.ÇÞöýYÍ[r].iË¥õ˜ÛvÊŽ)—žÃ®}9µÍœ¥gÉÝõ:ÜvíÏí»d /¹öö¹^æú—1³Ä¿¹ñµ´¯æÎuª]–ÔgI»Ì¹¾´¿§®±%ÇÚv˜k×]Çý¶zï2¶—{ÛÚ6–{®ý·ímm°tìÎë]ÆÁÒñv–ö9kßÍy±4sßsߥ_N©ï©o2Ÿúɽ÷ï¾wOÞk~2¿à=æê Þc/®>™nýÏŸÿÓ0üÍ? '‹«á[cÙÆòõ±üÞñIùÒX~g,¿ÿñIùÚ¸ìÇíëX>7–?:)ñþ÷þ¥¯ÿöXŽîHy¼~\9Î+cù¢©§”a,oû½0.¿:.¿lÎóÏÌqe»lêûòøýëæûWå8ãú7ÇeÙ~<^ß^r¤ýä|¥ƒ©·Ë?Ëo¬Ö÷“òu“#ßÿp,_“r<þó&O¶—v1õîÇ•ý¤?l;ÉñeœÜË·Æíß0õü¾É—ídÿ—ÆåoOÔûÖjý{Y–v“v¾)ùr<³ßë¶4çQ¤^æ8wÖÏ£½l»Ê¸—þ•ãʸ•ós}¿~<©÷+Çëßg³Ÿ§wÆóëâ5³,ûËø—zÚq%×»´«Ô[ÚEƱœ—OÆŸ½^M{s]™úÉuû¼9/©ç8®†ïŒË¯˜ídý‘Ôk,e\Éu)yR¾dŽc¾ïý!ãå5óý[o——Íywú÷2îä¸2~nOìoÏOÎKœL«õãH}¾i¾?2ç+ýk]—ó4׌ë¾ÿ[ãùK{woÌyIý_3ëåz–ñ$õ¶×ËØ^׃ä¿`Ö¿úñúq¥d?9_qDê)ãKêù]É7çm]|ÅœíoiiWiOù^Úç óý×M}¥Ÿß×[ªY–ñÜÇ×XÊõ&ëe<˸”óǾkJ›cÇ´û—ÖÛ£ç~÷x}ÿaüþ{«õvvýŽÉ“íÞúxÜoÜδK¯—ôëKëëûñä:tëÛõõÒ¯Rɱξ%ëM)ç#ý'ý-ÇÛ£·“½}Ûì/×2çqË,K;ÙûÔ fYÚIܾmö“úÉvŬ—ãI½†q9¬¯¿õ›qùæX~Íç¶ÙOrìó¡´‹Œ7q®?Ÿe4çuÇ,¿²^Þúï_¾ó«_ýãßÛçŸ —¥~âžô¯Œ³;'Ûo<Ǿj¾—v³Ër^ÏŒå-SJ½ì󃲿ô›\’3Œ¥Œ{©‡äK½½)¥Ÿä¼%ç³^úIÆ™ô§Œc9žô׋æ{óüÕïËö>}Û”Gc)ý(שäJYM)ûÛûÁ¸|ë?ÇßS²½ô¿8'íú=sœ/¥xfŸÃÌóXï§7Ìvƒ©¿´“Œs{]Úça©¯}n~ëŽJÞ¸ìÌvý9ÁÔOÆ¥ÔSês{uêv÷)i{q/ýiÆ}wåé±”q+õq6oü^ÚIê%®÷çðÕ©õéy·Ö÷ïÇ—~—q-í!ßËs€Üßå÷dÞ—eÈï(¹žÞ0ÛË2îä¹T~O®ÆRÚçθ^ÚÓ8ÝÇ“ŒɱýfŸ¯ìó¡´c0ßËù¦õõýú+¼q=ÊùËù<#ç;îgï#ö~7Ö·;/íaßWH}¤~ÒN2®ž7Ëö:ìÏáÇëËö÷Žì/íÑ7.KÿI½¤½å|¤/še9¾ŒOë¯}NøšÙOÆ=þj,o®ßß3}qüÞ>ß¾u¼žgó¥}¥亓ëãh,¥ýdYÚKÚ]úAÎWŽ;žWïwoÖ?gŽcï’+û Ñlj´‡ÜVcÙŸ'ÇòÙ±”÷=r]Ù÷.ò}4ßÛñÖï[ãöÒþ’#ã[ÚQÖK?Û÷âo2ëíï$ù^Ž+õ¤<9έÿZ­•=_êmïÿ¶d½ÔKÚI\f?ë“­—½oÊ~ÒoÏ®ï¿ñ¾Çz.Ç“úX7³)û8:^ßßþ·¿çå{é'gJû\cž{ýíóåÑXÞëc×Ëñä|弤ݬ{v\ÚûëÌsìðý>)å> í%ýÌövœ÷÷­ÇëÛK;½j–ó¸y/¼‘#ÎýúGÿþ‹_|ýûýÅ:,×§}’v4Ï!¼¿ÖëoÞ×ÈóÆ­û‹÷{ì—_îí/¾¶‰ãJ¾×¶Gÿýs¼¶ÿ†oöý¦\ÿÝ÷áôõö}Ëx>“ïä^ò{Nž/¿¸~~ÃËcÞXÚ÷“×ïs§¯ïýc®ÿz¿>±ÿ\û™ãʸ•ö“û¿\æ~ÑŸ7å¾ñUS/yNŸäú1¿Ç7\¼3®—q,ý-ý85^­ý÷‡9§Ò¾¶ßM»ÉùÙ¿Glô‡׎³Ö˾ßÊl'íc®ßó´ï™íïmsÝ÷ïÍï«v—çú©ö_ì{ûwɹyRöq"¿#e{é׉óÜÈï×÷ú~½þö¾`ïCß²Û¯ŸÇXß÷l~}y²~·O‚Åó¹í'¯Ki_©¹¿õöœzaûýÎöúo<ŸHþ‹¦4Ï ‹Ïk÷ÛgòïUâ­}Εû ½ESo{}™çmÆûx¼þs\6ãkò|&Þll÷¢­Ïñéç/ÏÃéëÏZöû–}_(ça~oìÿÆÇëûY?ìsÙ §¯÷§<×.½îo/Ûnö8uaž\W2ÞÇöë×x#~-ì·þ{Vî‹öýª½H{Ê}K®ÞÎÇëõ|sY=¦\]êÃÆó‚ÔßÞ—n™õÃXÚ÷‚Ñ”Gci¯Æïeü<=–ÏŒ¥<ÿf\–÷›²Ÿ·?—Ÿ”æ÷Žu«ÿN”¼‰çµç6û¾oâylÒóü³q™ßý}½Ü÷–Žwy®“zËsÜͱ´íøÌXš÷oÖ—y_ôÚéÛõã÷ßßëÛõý¶Ÿ×\;/ ëÇýÓeí´«3½=Íó„-ûx•ýnŽ¥}T·ÇÞoßïåùܾ‡zÈxë}5–²ßĿۙÌ7¿ßìõÆòÉq7þ>óæÉ†ý82þì¿Ë‘zгGRß±”öýÁ˜#.T³~¼ÿÜúÍÇÇ¿þÂäy­×ê9hãûñ¹À>7n<ÇÉu(ý ç%ÏEÒoãßKú~2~/¿O¿jöÛ臱>æwËÆ¿·“ö>2Ëö½Ó7Ì÷sÏMòž×þ;,ûj‘~¹?ÊööwÛõïûsÒøï46Žgœž;^ßìÕøïe¡¿'™(ív2 «Ó×Ûíöv=Þ¡ów†s>þÒ~Zºý¾íµt¿©í†‰í¦öÛ5gé8>k»ì;ÎúsÚD9—?¬¶o·ë÷sÓy_ׇªï¡Îoν³¶Ã¡®Ã‹šÕÃÄú}9ôtÕûe®]myhÇöSûïêò\}–n75í{Ùuº(§ö.êºÜu<jº*ÏAWíþ·ëö‡ê÷óoÕŽvÿ}§}ŸO‡‰R¶³ßÛ¼©ü¹úM­·9SùK»t:Ôýeßœ}·»èißûí§eZ³ïtÕ¼¬ç£‹þ~QÓ¾¿§û÷§¹ìóû»ÒÔtèçþCç\õi˜(w®ZûœÕ‹óλ¬ýv®ÚýíPÓRo–þÞ\:×{Â˾?OM‡n?™Îë}оSÿw›g,—Ö{˜(¯êõzÖ7rU¦óþ}4õýe¿o<¯ëmÒ¼O»¨ûÚeµë¹kÿÞq˜)ÏkºjÞNå_Tξ>«óõü:Ìl7˜å]§}¿«Wg†¹òæz¹ëq/ëùÿ²ß·þ¾ÿî–ißûðU›æÞ“§¯ÞØÿ¢Ïžu»]§aª\ø¾ë²§a,/ûïKŸGìö‡šæÞ»N¹:UŸO«s“çyÆqæÞ‹^ßÇÎkzâä¿™}ò_º~ò§wß¿÷`uòŸ¿~Jþó×¼ýÞÝÆ¥§üüþGï~tïqùƃ{'³wÿJþsÚO>x÷/| ܽÿö_›¨ÏÝÿÙß~SÅ=ù•qÅcNf¼Ì™‰2“d&ËL‘™*3mœyÜõ9×ç|Ÿ }.ö¹ÔçrŸ+}®ö¹žá{†ï¾gøžá{†ï¾gøžá{†ï¡g„žzFè¡g„žzFè¡g„ž{Fì±gÄž{Fì±gÄž{Fì©g¤ž‘zFê©g¤ž‘zFê©g¤ž‘{Fî¹gäž‘{Fî¹gäž‘{Fî¥g”žQzFé¥g”žQzFé¥g”žQ{FíµgÔžQ{FíµgÔžQ{Fí­g´žÑzFë­g´žÑzFë­g4ɸᎎ˜uÌzf³‘ÙÄlf¶0[™%Í‘æHs¤9ÒiŽ4Gš#Í‘æHó¤yÒóvcdExtra/data/CyclingDeaths.RData0000644000176200001440000000130313163461153016401 0ustar liggesusers‹Õ ˆ a€ñÿάË]N’$I’¤í¬»µïÇJg“$é’.IÖ}8ZK{›¯¤K’$I’$I’$I’$I’$]’$I’$I’$ízÖZÍŽ™µS¿ifÞ¹™yæíºÓ=ñ–žqÄ•ˆ8.›Q‡ŸˆD¥™u뢽ٹ éþLapHÄÿûdq[YˆÌýŒ¢H| &cH£ëÃnÄ œÇ5Ü׈¿ÂGü™71q,Är¬Á&ìÀ~ÃY\Ám< ¼Ç7‘Ž(Æab°XŠ  €½8‚Ó¸„›x€gx‹/"FÞÅÜ1µÖ÷!œ)%.Èã iâë÷vcdExtra/data/PhdPubs.RData0000644000176200001440000000503113163461153015227 0ustar liggesusers‹í\]o\5½wo²›Ý$KÒ´M+  - UˆWÄ+â‰×@T4€Ò"^ùü]¬é<{:cû†Š´FîõµÇóyfìE»ùâÓ/?ÜûrOD:é×ÿuýz8íÖÿLd*óõs÷óïÎ?ÿù«§"ýñúugÝ÷×ãßdÛ¶mÛ¶íå´É¶oû¶ÿïz·íÛþ?èý¶¿Ô>Ýöúðö—ÐwGöٺϵ§š‹u_j_­{ú\i?Sâì°c¼K°ÆtÞ>¡5æeŸLW£-ñŽø1¯’¾5Þ­*éíg^þ”°ØêW¯µä€G?ƾRîz2jòÆê×ÏÖüˆtŒôˆxG¸/í¯Éo©Ѿ’î5ŒEv°ÞZ‹-Ïš]5ÝÆøí-ùªe¿Ç¯æƒÚ¸EÖ¹ô^‹G-ž^µ½ž¬#úVy->háÙ‚iqækxªé4ïLñg¶³ÅÞ’­¸+éݺ¯5æl£çÓ’-5;Çú&âéñ³Þêß1{<{jºÕü^‹×X _SQ+Ùñ¬ÙþopÕâ“_°}Þ{4ŽÖZôó¾ûŠý=Kzzt½‘Ó“NQkµ+ÒztM„ë’,ðóøG1æ}]°î5+kB<ìÞÞ¡eº(Ç#}™ÆãÙÓÓÙÖï‡ÖÓ¯¥y±)åyMfGcèjÇmOs×±1ʳ.kõ#Ò¡&Ÿ}ÈXöünåÛ¼/圥­5Æóðæ¼©5[?¼|Šä”êyÄÇ®{kÁ ÇÂò(å··×¶(_KqàæÕ¾(Þ^ý€OGÌ[œ±¥+ù~ÌyéÚ²‡×½|kåñàgT #z´è ­5ÈòêADoeFØôüS«mœ[×É+³^Ýe==/o,ÿÖZíÅÅÖ+¯`mLî¶à±VÏÐ#¢µë­÷NgIŽ·ÏÓ³Ež‡ÉÒy`i<~5Ù¥¼h[ljÑÇžQÞ=0j5¬XÞÞÈúp‰øµ6®‰-þa¥{¨Ç«&':Ó¢»ËõxÖZëùãÕB»ŽÏ”8¿§ºŽïã;ÓƒlžíS³ÇîÅ÷«;3î‰Æîí =wO»w"›úáûÌV6‡¾¤Ooö[l³o¬îžýv?ëçD6m‡þ¶w$/²‰ÇÖv}Îþ´´ƒøºYžƒ£ƒå3•mìiŽ¿—oýÂò¦Î|Gól»çàÈá¸y¿@½³­žÖoͱ ¶sqÅþ-ñä\á|âXpܼ=‘ýµƒç{þ™µwWçÓsnø@7ød¦O1Ï™îå6—Œ¥9ùú$¾§ª×L×S[ê{Ú¿œø<ýœ+{fMt rl=ÆüGºÿXßß4ï3#¿×¹Aòo_w•¿µs8!® £'×!Ûñ»ZøaßøXŒ·$c(­ßSÿÚ8ÚfÏ=è]Ós)ù÷¼ö¼˜è0÷¹ú1Ѭ$Çü ÿž‘u ùwËhðùR×Õ†¥ÑSdKûJ‡¶Ò'xØ{“õÅ®á1“kø94HΛ‘;QZøÿ®ÑwOmM{nëü@ûà'¬}‘£S£|¸ŒÅÁðC“Ú›’c(ÊöÏ$c9½”ͦñMãË•Î-Õ6‹à¾AŽ*¯7$×1~CŒEiê¯C#¶"Ÿ û ]‡ŸIÆ?j$ò~}E2vÌ“ïÀ´­É¨ŸÀÍR2®vtýDe®Ô~‘ó ü€©¥Ú€\HÎiþÿ¬¡Cž¿®¶ Ù;δ#µ)õw$cAtn¥:Ãðb ? ’ë&bhóVÔ)g%ãîDyK®û²y-ŒáàÛÞQðŽ²§2÷%׃4Gljï}Ýó`ÝI®—¶uê Ø ù¨+ɘɵ^tïCµ<%×èèuCm?Œ3Ø”ž¯Jư{@:¦õSCÓËfýFœñw.Nd3ŸÉã-¥&v¬»ä‹…ÒÙÏ<à•öÜ’œ'’ñóŠîÅ9ý,nqNóÈÍ4¾#9àÔ¥}å—è?’\+,_ÔÍD÷¾‘ÿšú«“œ7;Æ~ÌãlKöÝ“œ;sÉ™)ßÁÄdÐõ´÷³4_rmÆ9°oÔu‘sœuSr¾“^À™½³¥˜| 4·%ß?£+õõMÉç¾Å<ÎIè#FOä êà‘Ž“ŸÞÖ5äóL×Ne3'{z®$Ÿ©öþoÏÀÔJÆb²ñcÉ9žôÿDí‡-Àø¾ä3á¡äšsO2^QkfÊûcãä˜áœ=–\ÿ`r¹zKǰµ 1@~Áÿ+ÉwÈ´vWòÝx9U™¸#ÝÓuÔdÄ÷»Gjë±ò°÷•·4f8á Ä;g.ùî§ç •ºwu|¬ý®‘Kcp¤¾Á}ðäù¾? ;üpvyñTpÉ{>9;»zöøë'ëùçï;ß\\ž=¹Ð·Ý˳««Ççú:ýþñù#/~úîü§«‹§Ï êË‹žýxÅb¿~röb1¹w~öììÁ7Wkˆ|~õã/ ir_÷ëúŸ?ÿüýrýø#õ¿^Ò3WvcdExtra/data/Abortion.rda0000644000176200001440000000033113163461153015210 0ustar liggesusers‹ r‰0âŠàb```b`bfb “… H020piǤü¢’Ìü<f>ß!5’ t¢„Ny¡ƒ@è¤U—°Â@5Ÿ9%3Hñ‚˜`gÀ1ª3€ÊòsS‹RÂPµPµ y6·ÔÜÄœT(ÅÂFVÁä“cyd¢É1G‚ †Húå£YÍ µ¬ƒæìàÔ (“-¸$±¤´Ê.-(†X<"Üþ¡›˜œ“X 3‘j1kIbØÑÿ@v¯™vcdExtra/data/Mental.rda0000644000176200001440000000046413163461153014662 0ustar liggesusers‹ ­’¿NÃ0ÆÛ-m$¤J‰``ŠDù#@B- a`¡’Õ¸R$§'‚•à=yІÏíÝO ?çû.çËçûÅ4]¤B)ä(RAj‰%b2Æs˜Ûukœê¯Õ>žG»Œ2BEèˆAİO¿ g?¬k &”¢É)‹)‹3ç,.X\F5Kg.))¸WûÂz[ðwWfÙÖjMÎêÿ÷¡KÙQcšÚ×/Ö9Öyé Ò£¼Fæµì«7Sn'u×ùË€‚¼¯àdàŠ<ŽPœ/Á¸àä”w å?C0'?¯am*Ë× (¨šÚuXѸuzæí{TaìëÏŒ«„ä–®ë~¢3à iaZ“­<¶Ðà›_Y»6 vcdExtra/data/Mammograms.RData0000644000176200001440000000030513163461153015761 0ustar liggesusers‹ r‰0âŠàb```b`âcd`b2Y˜€#s1—obnn~zQbn1PPDÀÁ‘ $ 4*pPÒ PZBÛ@U—ÀaŒÖ€Ò:02§dæ)^°'€rŒ¢„¨$/17äa¨:¨Z<›cRqj^ ”Çî›™—™›˜Ó웟’Z”X’ SœZ–Z”J#P\É u"Ø\&˜IA©‰@½F¨\C ó ÂdÇ×ÅvcdExtra/data/Caesar.rda0000644000176200001440000000037313163461153014637 0ustar liggesusers‹¥P½ Â0¾¦õ¥àâà "núZÑÅ¥ˆ(èk„`MÅtñí|$Ÿ@¼Ö«‚.¤ßÏ}9®YÍ·£ê¶ ˜máAê0üXn±8ãBó+€Ý@ÕªÉãÞŒÊüO¿oæ¼1¡›C/§Û„›;ü*¯G8 ,Éu3fü}g„zâ'’‡ø#Zƨâg¡±Õ¤¼KwÒ‡Zß.¢34Ôˆ”ãGJPž‘gï’Q)e~ôOÓX³@;¦72+ uA,#E†³’úD¼6U±ÜË(–&«´ ¹Râ€ô™Ÿ„\gó-Z£ó}(Þix¾€Ò¥?UvcdExtra/data/Heart.rda0000644000176200001440000000036713163461153014507 0ustar liggesusers‹ r‰0âŠàb```b`bfb “… H020piVÔÄ¢f> ‡Ç!ÿ8´ï€ÐU ôÑ0 •_± B{1@ù":*¿L B@å« Åræ”Ì\ Å b‚ÝÁÌhÊ8€ÊòsS‹ÒÂPµPµ yv—ÌâÔÄâT(—Å/?/M ‹obLžÍ-5€ ÍKÍ-Èò8Â32KRóa|v§œR5ô Î‚™$€ê0—Í=5/%µ¦Ã?9¹´Èø‡nTrNb1Ì(F¨µ¬%‰I`—þaèwÎvcdExtra/data/Alligator.RData0000644000176200001440000000061713163461153015605 0ustar liggesusers‹íVÍNÃ0 N“f•†&ñ;ðHHp!»š.]«e Jˆ8ñH<áJ²Å%†´¡ÞX$×±ëøïk’>Ýή“YB¡„^F„23©yD†.,ÝHY, VšveÜð±á»Eýë™âž)ê—:H±²2³‰‹¶ÓÞ ¥ÂIÃ{(S•.8zXJØ@(?kÈ2¥çkžJ¨Ðs/ƒt‡*ÙþÑc«ügþlÛiã:MÛN‹HÄ0ÞÏ­%CËÓ19¶–³ÝaL¸„ŸmÅ+ƒ‰<{ÂpgŸu§è:˜p×éø¥°G×~žUŽåZèÁRu.´†Z¼ÖÅqû‰ vCgƒ»€á#ø>Oá3ø¾€/á+ø¾oá;ø~€a¬†Ÿàgø~…5ðüÀŸð¬…¿á¨ƒú ¾ñÃðßðßðßðßðßðßðßðßðßðßðß”øoøoE€ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿÖðßðßðßðßðßðßðßðßðßðßðßðßðßðßðßzþþþþþþþþþþþþþþþÛÀÃÃÃÃÃÃÃÃÃÃÃÃÃÃ+ü7ü7ü7ü7ü7ü7ü7ü7ü7ü7ü7ü7ü7ü7ü·q€ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿ†ÿVçvœßúüîÒÚxæ²a®ôø /?Óà0{úì>?Êíéñ\Éë¾á¼äŠ>“gˆ2S™:6¼?eŸÞ¿{ªyWÎÍ gä¬ ŸÓ³nqø^Üå²Uy Ž»²4œ¹:ÂÏtR˜ÏlÍú¢ðy土¿ºCØ>žÇ6a^<›f÷ü|•åsâyI‡ï×sD¹Ò)œ£»½*>K͸\H…sSÞhÖ娾{wX¥yUï)ëuî-cŸæ‡YsÇåz«pæž·x,G”òK]às£ï ˜I÷2Ï„:Dó¦VF­ˆçO(Ú†9ð<íæ×3K께ÜñLT®«k"wÒ;lIÈŸqu`‡xÆ4w]ƒCžG‚;žä’¼ó~W‡kþFÅó§TÒÊõ§æXY¬\PÊ u™æÙ;XÙTæÜûEï •‹òÁ35öD½¦,QvËIïz;rFïE§3é;ª.WïéÌ•óÊUe¿öu‘:D¯'ß•!Ú?ä©ÎL;‰ú@ݪœÑ.¤~Ó>¡ŽUõˆB· y¡.TöxgèóhGQßÑ“žAzí.!¼Ÿ•qš¥5Áï¹~!W|ÇR¨åµò_i´‹ÑgÊeßMÔ¡Ê}õ †W{€:‰\÷üW/ê|ÕY ¢°¨ãÔÏ:zEÙ¯žölÔ¾¡Ý@¢¾SȺdz[®«¿Ô?Ú‹Ø ”ÓÞËô‰²ÜwGõѲ(ô‘võ§zuaä;¥:Å÷õœî×ühwì磺Fß·Õ÷ªy(‰¼C|ïÓù&û‹víEÚE´ƒ¨£Õ÷Ú3´·jP—²×¨´3ª|×ÒÁ÷D{‡Ö®+÷µwiçT¶©‹´CjGQ— ˆ|wSÿøµ: û’öFýí3Úïä²öí‹Ú¿Š¢°gãGú¿ûw¾Håt)ÈòHã<Ñ$4Í#ÍòDaiž'ZäÿýKãŠìÔl…þ^äMÉT%—ƒ²ã“Ë¡™\r9°&·îÑé ·M©ZwYÑðÜ)’ˑٚärximr9¬zjrYœ-]ï6*­ÈLNÞ`*~°Iu®,›Ë–%Ÿ¢ ¾ ýׂo &g ¦> .Fï¿ ïeè›C¿èu§Èƒ[àóðÿFÏÑàm`'v°~èaø_ÃïÃ®éØµ > Ÿ…{±3rÈ#7ƒ›?þÄžèü³|ÄúK0æÅoO€Ï‚ïãGì_Bþ¶ð‰ð‹"OÜüÝð÷ ö¢ïÁ·“#ÁN—|UøLJøìpè—öìƒÂçZÁ§EžŸ/\ þ$‚Ç…Õ‚ÃDŽC~¸9qêA~;¸ùð·È3àK‚áÒž?ûÂ_Fþ}ã—ð]ðœ›ðã ç+Œ_ÜûH¥È#À‹ÀàRú­¿þüýž< þM0:\Ú£¬Å_Qâ+º¾~ø/‘ÇÎ>v><ñ›N/›èÿ ø$øKÚ9G±—Eç\ÇœËxˆöW„O|Uøy$áý¨ð½7‚È“çW‚sÀvp—ôOîßDþŽð©JáSCAòNêÛà"O× ŸÆOéf°öu ç>MM“wÓoŠiïý!~{\äiôL“WÓ¯ãêŒ`ö²oÁg°?‡B Í¯S°‹õ£ ç5zþ]ìbýØ×@â8îuÙC½‘¤žÉp?ço>O>Ì“wúÈgÁuè¿|}Ñ+ôøëÀåÈÉáG@ê‘p9hï‹Eðv_Ù—õHdšˆ—ƒŒ‹´‚a°|ü~’Ÿ£—âGÎCtþ«öØÙðØ£.ˆq/ĸ—cö²Ÿ±œ`ü{àO@îõøsðÜ«qò{‚º.ÁýØ_W¬y’|ž¤NH~‡öC‚©‰ y;åwƒäÃôlÁÌx:#{Ì›}HøõKøŒÔc÷NºQÚsø!_+òüTÅœôùLö'MæE>ÅüuÌSϺ“‘Ù}¬P…˜ƾ|KI|™ù?§i sJÓV|q+ckÏÄÇöC{ãg悼 ý×`Ûzülì5Ö<Æš>ãécl[ŽÏMÜ>ÀÞîàÙè>}½DIÌý7°¾ýi´C¿M¦pŸê™9ôËq….ª·ÃÏ”±efÎA‡Ì[« †ÿÞOÛ¹øm(üYÌ6z=Jÿ æ3çÖi?@Oã÷š®aì0ü`.`óRüºc/^gƨ¹1ÉêÖ –cC=²1ª?ªñ…ªGsFgàÃjڌΙs<òJ|:öK•œ“üM<5(É3f¿ÏW+µ;Æá³ZUˆ)óA…)bôy*›ëðÙ,Ö^L»ÝO«ï8ü6zƨB<˜x‹Þ7ê)†°oÆwŸÖt-º•#[Ëžt³¶)n¤©YS€uL¬¼Æ~-ÄWåªða¦Ñs k=.R’·Mÿ9ªpží˜+è· 4z ¡m ýæ¡O³*ä þN›Í×feàRú_›wŒym®hTrîcL vަÏ=ô߯Z3˜çëøåü7õ­mø¹Œ9¸ãl±öX;œd}äôS)*eg1›?ŒÜ>)F¥üTÊgnÊõ¡=.*æ_Kn»i¿‹¸èà)èE½ä¢b{èÜG»n9cÝIÖ3ÇǧŸÜv—²Õ}&zf~ýó˜*nNuÜ{âôå!U<–k–Š™>×z/°G\tºgåtÎÂï§Ûî=wûïT{êŽ]g|ÚxL«Bîî£ÉÓÙs§ÝνvÛ`cÐêx@Î@š~ûTñÜUÌ×ÎsáôS©xwúä úàÝR…x³º¤Ý¿Ÿñ{Û}jà7‡U-ÞNû;ªÞî~¦r¥·ÿ<µ‹¼]Þ¶Ö®V=žÖ›}ÞR_Öê‘ޙͽœK…!Žž™V S3•ß¡ÿ8qÜ|dÞ;ÿƒ‚ w+vcdExtra/data/DaytonSurvey.RData0000644000176200001440000000056313163461153016343 0ustar liggesusers‹ÅTKNÃ05NÔ *$.Ð%bQ‰²`ƒ ‹Š%B°) $FÁ%§'m‰Ä‚£ô(…#p‚‡Œ#ÅJÕð)X²ßÌØÏ3϶|ÞôZƒ!„ÚVÝR¦MÕ°¦zSõ>$±]Œå„%„XÛ*æ(ÜTØÁuôÿ°T«ÃÙ„ñHY[¸*‹Z—,B“ž ƒÒð8DšQì3/RYsC©ÎNÿÞÏʦi m–^TOŽ™ÖL¿¢ùUËYVÞ²r¿È¯”ã Yœ¡gçö/jú©æïhjLý Ö’"ö™¬£‰´3MnrF²æî~9~/£ûðšã½›ãƯÐFþá>É1è”ù׳Ð'Õ­˜ß7öÝA<ªÎ[ð‘çö;¤T‡ùùŒ dúóq0Øô‚[,.y¸'|Áõ|2¸Ã0`Eì M[‚W<·É„M)¦]4{e4;¯4MßÜaëb襢üç §c3vcdExtra/data/Detergent.rda0000644000176200001440000000044113163461153015356 0ustar liggesusers‹]RMKÃ@dÓj#–‚þ=(ˆJ‹mm´~a<•µ™Ö@³‘MŠ×þòâ&}Q²Í›÷‘Ù%σèÜ|"rÉŽY¦ô\órˆzƒ¬W¬ "Ñ5ÂipAÕL·{¼´øü|~ ~gÛ¦~oåÇÀ+`þü 8„>¿¥FÎ: ˆ“ÔÀq©W÷ð¿„=4Q%SÎ}‚|ÙÒ÷FÉê µ˜f?–->Ê/«Ò}È,󠯥ŠÏ¢& ‘u‡—lY n‡'›ô¯·Ô±5o ÃV[x^9ýf-‹fHípþ–³ógš—¬Y-j¿û.Í?0ÏMoÅy¹áÎî´X˼îä`¤V!?×¼OÓîò ÓZkvcdExtra/data/Dyke.rda0000644000176200001440000000043013163461153014327 0ustar liggesusers‹ r‰0âŠàb```b`bfb “… H020p‚8.•Ù© Ì|@¶‚C(8TL€Ð t”Ö‚Ê{Bi4šJ«@i(­¥PÅí? © ‚Ú“Ü¡3!Ò ´#TÜJG6 ºSjŽ”¶†ÒF8ÜaŽ*Ž6Ì)™¹@ŠˆY@AˆQ”s•ç%榥„¡z jÁA힟Ÿcäç¡É3ùåCYÌ‘ Ch$‰âfV¨ƒÁ:X¡‚œÞyùå9©)é©Pö ÔĔ̼t˜¦ /ækŸÔä’Ò"!Í~©åʼn© ïýC·.9'±f#ÔY¬%‰I9©Õ ÿ‹æØD³vcdExtra/data/Geissler.RData0000644000176200001440000000110113163461153015431 0ustar liggesusers‹½VMKQ=3™ÄLê*hiãmŠ.?JĶVp#"’ª]¤í¨hpFÁ?к±tW馺(ÖÒºéÆUE܈¸Ô;»\ˆãç Õ¡1Múàä¼yïÜûnn&s¦çi¸>áÙé"óC‚•ì®E #¦é€§ØÜ$òˆ~¸’KÈ.àIJð¦€ïä$ÿP\Â\"ßìwèÈjb›hf›_p¡àÁe‚ï ®\-øàG‚kpedëœlåÏtÞLå»mž›Æ»KWŸJ—lß¹~iü—û:Ûù³–7Óù2•ç¶ñ7s«OW—j?ÙºóúïsÔòGiå+°ž÷ù8/¥½tÒÚ~pÿÑN”ÃòºïÆ7HƒsÀvý4|Ê^Ж7ϸ5ê>Ìvñç&°?¬žñ] µ¨8 â´¢#Úâ6cLϺ[€ÙY`rèÛàwj›¨=áñϨe-U¬-´ÃšÉÍ @ k. Z6ä£7Êù·íaÙÀÃ.êë»ìP»ágô1ßcjÞŬ§è=ÛñŽöFoô†?Å}Båû„J/Í¡^ ólö@þÍ9=Zi`=­=„TAü!öD¯^ëÖ÷G-.¼õÊûŠw,2ªœÂ4skQyŸ6lÁpTÙŠÑìy‡®;³½ŽE ;›½x™ˆÔ é<È!WõøT]€yÈüÙH$>“ÎLœžW×q vcdExtra/data/Vote1980.RData0000644000176200001440000000052213163461153015121 0ustar liggesusers‹½“½NÃ@ ÇË%i£"U0° 1#¡Š–ÏL Ø:@–SzJi".§°ò ¬¼ÏÑà46J*P"ýÎ_dǾs&×ÓQ8 @€ ¥¸8 ¡‹¶sŸY=_¸;ÕK´[h÷V1›álF«.?Ñ…NrTýUVªö.KžçV“ïÕN;Ô‹•säw¾XE63¨–ùë\¹KHÂ#|" 6Œ«Ê%UPeÎňÅ1‹§,ÎXœ×é\N÷{£¢Qð?ØV£|yþ•2Vö&Z=©ô/]@óºzä‘Cäù@Þ‘Gä™!Èv52P?>Ù>U¹Kùöi¿Cûùr}ªRµÐ WÁ¡j883@h(íÞ¡Õ |(­¥m ´)”–a@JPZJs@i(í 5?ÈÕ¨9†hæ ë·€ÒÞhî–@sŸšzG¨½P—¢sJf.âbVP(b`å@åy‰¹©Å@)a¨¨Z<«¹a$HÌaòËÇ6À+mf‰_Ú¿´9š4ŠÓY¡îëa… ¦Zš"s 8f–È dŽ9ñ݆äœÄb˜ pÁ’ĤœTˆj†yḬvcdExtra/data/Depends.RData0000644000176200001440000000031613163461153015245 0ustar liggesusers‹ r‰0âŠàb```b`f`d`b2Y˜€# 'fwI-HÍK)f``ærùô-Vg ý¨k-Pe9kåf±#+±0ƒ´³ ÑLhF3§dæ)^°Øhi t^bn*Ðj&a¨°:ˆ<£ŒacÁÆ0† Œa c˜Áæ0†Œa e0ÀY†p–œe g™ ¹™ê`°C1ƒáºúäœÄbtõ¬%‰I9©Õ ÿa´£vcdExtra/data/HospVisits.RData0000644000176200001440000000033013163461153015772 0ustar liggesusers‹ r‰0âŠàb```b`fbd`b2Y˜€# 'æòÈ/.Ë,Î,)f``æŠp:¸60€€ƒ„V‚ÒPZ JAi(­‚ªÍ&æ”Ì\ Å v Á(J8€JòsSNa†ª«ƒÈ³¥¦—æ$ÁÜî™—V”ZXššWaõK-K-BÓÅl¤k “74Ð5´„‹h£9€j;X?L° óVÍéo{ÿ-,»¶a›ì¦O=”/~v0~ëø| ¯«¨ëƨ›þµR‡<;À²[éçnÆWÊvÑú~ÝŒU>Ju—´î‘öÃKn¥|¯·—Û}˜¼0vTÚâ–Ö·xù• ò׫¼WÛÅ'LL´¿_Û…‰ŸŽ _Âö·0¶µ…L™~Œÿj×ÏQ=v6•±y×Ð+é¹’¦*é:7•Æœ€î‘N#à>i½^r¼ôF‘«' Ë;(¾ë>ú²*ÛÚ¯¼á‘®¦jAÜöps¦Zúmˤ¬ãêfoqÔ.¨›ƒ£éb´èõxx¾¸ÂȈ3iMa'¬°/÷°“©Ÿ|¿8¹Õu[\„i¯ ÕKÍaËÅÃoÂŽê{ÔI,L?õÃoÒ[Ök§­T†·Mý…‘¹û8/„¡ ŽK˜€G‘5ÐqÊ&íoX’¶1h.t¼Âê¨&¦å\žÛDnµp±Po«ÑWë¸Gjatû½Fë/ŸüæÈ²tÒÃ-hš [˜%Iœæ ZÊyÕ÷[êx]ýt¸5‰ÛÒ*L7ò7¨;„íª~i¯˜…µ7êPñëS~ñ÷kǸ>yÕó‘·f7?ž3Úb¶9î"8)Ý‹)n1¨vf ;½„†[iGía0†™Î‘Qp à“ v| Ì‚ð$x|LkÎÎÏ0DþWXÿeI¹«àq–“¼s¼Ÿ¦þ0ÿIÚq•÷3¬/y£ gy½ÆôY~@svxŠ´EòžW¨kŠv§ÎO2£œs´ï¯Ræ ÆfŒ¶JÙSî,eÊÎÏaÚù e=U&;Gy“”%÷Ï‚S¼ŸaŒÏSž\¢=¢ó4ãs•éëý¶Ì#Ì—ú²Þƒš³‹t…>‰žË´õr™Ž‹eñ’ÿ˜¾YWµÙyÚ<Í´´É½”#ýä‹ô{Š×Ë ÓîGÿY扬}´Aµ‰Èî¤}ÊWyÖ¾>Â2Ó”1@ÖRçeê”xm£O§h“È꡽çiË4}=L[„;ç2âJãdϤþc~:Dn;óh—”µ(w'cy }ÝÇ:²“—§mÂQ¶ãý”gió}k1h§ ¢Gv-‡XFÉx€q¼‡: e1P;’1¶’î§½g¨ï8랤‘%»ÜÒ—&iÛh™Mò\vD¥ý[4g7t+u¡ì·³ì,ó”éˆå¿0–²+;®gÙNâ»ô·6ú*yiÓ}—2xν$ÖiúÖŸн—¨ç ý²å3‘+ýá;”?D»6kó}sœ~µñz„±ËR^?ëô±Î^ÒGù»´ùþT`¾´å‡—¾²vd{ÜOÙèçÆéåœf,Å'5Fe×ünÆ{Ÿ6?^îÓœä!æÝǸ SÏÝŒ…¤O0Ò¾Ò×gø|’¶ódÄÿí<‡£¬¬£.ªùTv ¯Åò$ªïqm¨Öö¨KC·2µŠ_\Ùaúkœ~Ét-þ9áö¡ÐÏT¯tn¡ *ë§ÃM§~ÝÞëCª› Q仕ñZ‰G•ç%£üìðk{/âèk«›ßuÂu¸5ugl«˜žTG=–wg툳NÆþGEesÔ‹0zÂ6¿aœ”õô+ û݆m5ö$)«ñ¬Çÿ£š¼PÃt¡uWkW=†¡×ŒVïøzÉ©ÊIÛRe_vnfçÆuž±½éŲÍübÆ-FQa­b\‹N]?Qó뇲gîgˆ;;65·nPgÛçnê½û•Ä¢0î?Ê[2 ?’’ÔÙË Kj!]í «E§6û‹cyõ­´†íùl*ÖºG: *åyÉOJo­äWö$í‹"O÷)ï÷Ì«­ýÊWÓ/’îGµÐEÿ]§;~÷“»sŒ·ˆýeÐj0’³Ä ‡¤ú„×{­²L˜¼ y.vç®ÿŸ\*+տНQ3æpÚVRS#jT›ãö”*QHY*{Ù¸]˜L—n­TΚÿJxÁΕ¾©Ú˜Î³­¢g©RMã'Ç‹ùLÚR"G ÅBªô:Y6V°¡aBÉL gÓJJÓ1˜Rœ(”ŒË—–Ýc£¥L«¥£óZ¬‚*¦sÊ'KÖ Tq¢ìmÖ8‘³ò9¯Ø®NSͰ"kWÄvy!¢YÅW½ãšÀKÀË4ç—nnoÑœ_™‘_‘_¬iÕœ_—‘_¨‘ó€òK4rþPΞîל³…êÌéQÍ9xLsÎ¥ŽiΙB9—(ç#åܨœ%ý²æœñþ:øø¦æœ½ý6ø)øø9øøø#øø'üi+ÀËÁ*°¬¯kÁ-¶ë°]oÁ°ô‚½à @€ô@¼ Cêˆ>N÷‚3à,x<`¿þi0¾ ¾àƒþð4ø3øø;[ÿv¿À^17gãf; ÄÙ@œ€øˆ«;»À0wƒ!pdÀûÀð~€Øˆ«ñq»ŒÇÁ,øx|üüüüü žì4`£ñ/ vþÈü‡øšè&º•‰˜šè&âi6ØjÂN³ôÄÒ’Á`Ÿ™•Þ Oó$@üÌ‹à€æ§À[ÍÏJ/ל_ä’_l’³°ðmîLªœþ<øø‘æœ'üü ü±] ÞÞÐõÃà48 Cÿ@ßÒžâw#€­ÚÞ€ã3à»à÷mgüv¡?™#à1ç¥=÷?I©öOvcdExtra/data/WorkerSat.RData0000644000176200001440000000040113163461153015577 0ustar liggesusers‹µR=oÂ@ uîÂW¤JHüfVVÔ­ÊV71¡jàZ“–•?Èÿé”Ô!ñS‘º0<Ûwö³ŸO÷²Ú,¢Ml€±†FL!ŒÄbÇïÄk,ì¤ÎŠ?¼uaºè4êçôMùQ¢ñ5ÛÜÚWLÛ0ÌœK=R/ÉñèsŽSbR^‹IáX¢òi*Ç;×SLå‹ÉÝIÅìÞ²]Sfý²Ï7·ý½æƒÎϱ`*˜ –‚½àYpñ_ù€{ÒWµí0#-Y}kªù*õ‘éÓÿ5ìNsíZK2g1UUýx;)!J±Àù–…Ò.Rþ[¤YœvcdExtra/data/Mobility.rda0000644000176200001440000000042613163461153015230 0ustar liggesusers‹ r‰0âŠàb```b`bfb “… H020pißü¤ÌœÌ’Jf> _Òaš84x@袺*î¡Û@Õ%@èüºõ„ö€ªëøÕï¡» êf@èh¨ºv¨y= ú ö­ü¡s¡ò•Pù²] åÏÔÐs êÐüÉœ’™ ¤xÁAÁÀÀ Á¨AT’—˜›Z T" U'U –-ðËÏóM̃ñ}òQù¡@NibB…Ïâ–X”K3QüÁ õØ"&¨ _p~^¼rriAbIf~TTÐ-±$#µE‚áºyÉ9‰Å0ó¡³–$&å¤BT3ü†æÚ(hvcdExtra/data/Vietnam.RData0000644000176200001440000000051113163461153015263 0ustar liggesusers‹Í“±NÃ0†;ª`A l ¨#Puca@eBV¹JHivTÊ‚@bb`eâ `åMxš†ßå\KeÊÀðå~'¾³ýûrÒéí%½D!…‘ 2–xD"MÄÅÓK*2=BµÜ7ÄÄé”z‘õR9ÆBJ#J-Ôêô+¿íÒP§Ä£øØéj^£ŸjëÓfź_äjæÙnIÄ€šëU®MþÊùgcg¿,Ùñ˜¼8ôâÈ‹ÎOŽò9ß’‹-ÖÞ7{›w°¶@æ*€e°öA¹›àžk½ƒ<€5ðîÀç=‚1xåº/à|‚]0ØôâsØ|øïÈ7Ÿ·BYû†½%mX/²WyfgÍÜ5t=Ç™äBº=0¨¬Ø4ùMÛ¯êŽ'Ý9˲üb'ßhPôb2vcdExtra/data/Yamaguchi87.RData0000644000176200001440000000100113163461153015741 0ustar liggesusers‹ 픿/AÇÇîqB$Z…JD!h¨8¿ƒsÄÙäD‚ÉYçâîöŒõ+‘“h$‚F” …F¢T*ý J…R¡rfó¾³ÉmVBr•¸äÝç½ï¾ùÎìÌf&ûãí¡xˆ1¦1­ªŒiºLšü+“Q)£jšgxr-±”êì`L¯sžKVKŽ¢Íi×IgÁí;­h[ËÓæº™^•Y-F8jÀÈEÇT±Ü\7rn±ÜŽA.2ß`"ÍW•­;Ù"OØ–Ù§ç輡ù„îŸúÄ™ÃÙb­PªÃÒ•ÓïϧÔá·O¥¿}/qŽ]ÕŒ˜›" ŽðÏþäXs á³7æüÂëÄ•ÛÀ“âÁ qóš¸}GìV}ôÅQOˆCûÄÓzâôó[âÕq~ùgâžòǸ øä[‰»XÏ<Öoà}æ0®ЗÃDs5üз ÝBþ•xˆõµ¢ïà û¿cÌùNì…o ýpý"GL¾@Ç|ÓxÞ‰þ.° 4à›Â¸ èQèͨPOÂou”÷õ@Ç¡O Vû7ŸÖ™„n]m,ÄskgyÆT·v¢³²H˹½d T}ÖZÖ[LÝ,Â\ñV k£E™:Šæ¬³P(Üó퇸Í[…Âèûÿü¹.·ÓvcdExtra/data/Cancer.rda0000644000176200001440000000033213163461153014627 0ustar liggesusers‹]PM‚0 - I4&þžP1œ¸ÈŃ—* !#“³¿œØIgMº××öe·ì~Hî øà9…+Ÿ`·&Œ.¨^B[ÊâôôkiÁ˜3>`Â+c™Nxt9°Í¶UÓlløñ÷-&šÂVôÔÚ3wÇ\Û_e¨\\¾õ°è¯ ”M­P.Dg¡¨° Eç®7rC¹Ä¶Ó5ª…š¥üf'Ñžm”Ž”k¬Äÿ…2öa\îzIìÝ./‡ŸRLl¿”™Fš¤vcdExtra/R/0000755000176200001440000000000013163461153012236 5ustar liggesusersvcdExtra/R/logLik.loglm.R0000644000176200001440000000127013163461153014713 0ustar liggesusers# logLik method for loglm objects, to allow use of AIC() and BIC() # with MASS::loglm, giving comparable results to the use of these # functions with glm(..., family=poisson) models. # allow for non-integer frequencies # allow for zero frequencies, with a zero= argument logLik.loglm <- function(object, ..., zero=1E-10) { fr <- if(!is.null(object$frequencies)) unclass(object$frequencies) else { unclass(update(object, keep.frequencies = TRUE)$frequencies) } df <- prod(dim(fr)) - object$df if (any(fr==0)) { fr <- as.vector(fr) fr[fr==0] <- zero } structure(sum((log(fr) - 1) * fr - lgamma(fr + 1)) - object$deviance/2, df = df, class = "logLik") } vcdExtra/R/logseries.R0000644000176200001440000001147213163461153014362 0ustar liggesusers## Original from gmlss.dist ## I think this is working correctly 01/03/10 #LG <- function (mu.link = "logit") #{ # mstats <- checklink("mu.link", "LG", substitute(mu.link),c("logit", "probit", "cloglog", "cauchit", "log", "own")) # structure( # list(family = c("LG", "Logarithmic"), # parameters = list(mu = TRUE), # the mean # nopar = 1, # type = "Discrete", # mu.link = as.character(substitute(mu.link)), # mu.linkfun = mstats$linkfun, # mu.linkinv = mstats$linkinv, # mu.dr = mstats$mu.eta, # dldm = function(y,mu) (y/mu)+1/((1-mu)*log(1-mu)), # d2ldm2 = function(y,mu) # { # dldm <- (y/mu)+1/((1-mu)*log(1-mu)) # d2ldm2 <- -dldm^2 # d2ldm2 # }, # G.dev.incr = function(y,mu,...) -2*dLG(x = y, mu = mu, log = TRUE), # rqres = expression(rqres(pfun="pLG", type="Discrete", ymin=1, y=y, mu=mu)), # mu.initial =expression({mu <- 0.9 } ), # mu.valid = function(mu) all(mu > 0 & mu < 1), # y.valid = function(y) all(y > 0) # ), # class = c("gamlss.family","family")) #} #----------------------------------------------------------------------------------------- dlogseries<-function(x, prob = 0.5, log = FALSE) { if (any(prob <= 0) | any(prob >= 1) ) stop(paste("prob must be greater than 0 and less than 1", "\n", "")) if (any(x <= 0) ) stop(paste("x must be >0", "\n", "")) logfy <- x*log(prob)-log(x)-log(-log(1-prob)) if(log == FALSE) fy <- exp(logfy) else fy <- logfy fy } #---------------------------------------------------------------------------------------- plogseries <- function(q, prob = 0.5, lower.tail = TRUE, log.p = FALSE) { if (any(prob <= 0) | any(prob >= 1) ) stop(paste("prob must be greater than 0 and less than 1", "\n", "")) if (any(q <= 0) ) stop(paste("q must be >0", "\n", "")) ly <- length(q) FFF <- rep(0,ly) nmu <- rep(prob, length = ly) j <- seq(along=q) for (i in j) { y.y <- q[i] mm <- nmu[i] allval <- seq(1,y.y) pdfall <- dlogseries(allval, prob = mm, log = FALSE) FFF[i] <- sum(pdfall) } cdf <- FFF cdf <- if(lower.tail==TRUE) cdf else 1-cdf cdf <- if(log.p==FALSE) cdf else log(cdf) cdf } #---------------------------------------------------------------------------------------- qlogseries <- function(p, prob=0.5, lower.tail = TRUE, log.p = FALSE, max.value = 10000) { if (any(prob <= 0) | any(prob >= 1) ) stop(paste("prob must be greater than 0 and less than 1", "\n", "")) if (any(p < 0) | any(p > 1.0001)) stop(paste("p must be between 0 and 1", "\n", "")) if (log.p==TRUE) p <- exp(p) else p <- p if (lower.tail==TRUE) p <- p else p <- 1-p ly <- length(p) QQQ <- rep(0,ly) nmu <- rep(prob, length = ly) for (i in seq(along=p)) { cumpro <- 0 if (p[i]+0.000000001 >= 1) QQQ[i] <- Inf else { for (j in seq(from = 1, to = max.value)) { cumpro <- plogseries(j, prob = nmu[i], log.p = FALSE) QQQ[i] <- j if (p[i] <= cumpro ) break } } } QQQ } #---------------------------------------------------------------------------------------- rlogseries <- function(n, prob = 0.5) { if (any(prob <= 0) | any(prob >= 1) ) stop(paste("prob must be greater than 0 and less than 1", "\n", "")) if (any(n <= 0)) stop(paste("n must be a positive integer", "\n", "")) n <- ceiling(n) p <- runif(n) r <- qlogseries(p, prob=prob) r } #---------------------------------------------------------------------------------------- vcdExtra/R/summarise-old.R0000644000176200001440000000571213163461153015147 0ustar liggesusers # summarise a glm object or glmlist summarise <- function(...) { .Deprecated("LRstats") LRstats(...) } # summarise <- function(object, ...) { # UseMethod("summarise") # } # # stat.summarise <- function(deviance, df, onames, n) { # p <- pchisq(deviance, df, lower.tail=FALSE) # aic <- deviance - 2*df # if (missing(n)) { # result <- data.frame(aic, deviance, df, p) # names(result) <- c("AIC", "LR Chisq", "Df", "Pr(>Chisq)") # } # else { # bic <- deviance - log(n)*df # result <- data.frame(aic, bic, deviance, df, p) # names(result) <- c("AIC", "BIC", "LR Chisq", "Df", "Pr(>Chisq)") # } # # rownames(result) <- onames # attr(result, "heading") <- "Model Summary:" # class(result) <- c("anova", "data.frame") # result # } # # # summarise.glm <-function(object, ..., test=NULL){ # dotargs <- list(...) # is.glm <- unlist(lapply(dotargs, function(x) inherits(x, "glm"))) # dotargs <- dotargs[is.glm] # if (length(dotargs)) # return(summarise.glmlist(c(list(object), dotargs), test = test)) # # oname <- as.character(sys.call())[2] # result <- stat.summarise(object$deviance, object$df.residual, oname, sum(fitted(object))) # result # } # # summarise.glmlist <-function(object, ..., test=NULL, sortby=NULL){ # nmodels <- length(object) # if (nmodels == 1) # return(summarise.glm(object[[1]], test = test)) # if (is.null(names(object))) { # oname <- as.character(sys.call())[-1] # oname <- oname[1:length(object)] # } # else oname <- names(object) # # resdf <- as.numeric(lapply(object, function(x) x$df.residual)) # resdev <- as.numeric(lapply(object, function(x) x$deviance)) # n <- as.numeric(lapply(object, function(x) sum(fitted(x)))) # result <- stat.summarise(resdev, resdf, oname, n) # if (!is.null(sortby)) { # result <- result[order(result[,sortby], decreasing=TRUE),] # } # result # } # # # summarise.loglm <-function(object, ...){ # dotargs <- list(...) # is.loglm <- unlist(lapply(dotargs, function(x) inherits(x, "loglm"))) # dotargs <- dotargs[is.loglm] # if (length(dotargs)) # return(summarise.loglmlist(c(list(object), dotargs))) # # oname <- as.character(sys.call())[2] # result <- stat.summarise(object$deviance, object$df, oname, sum(fitted(object))) # result # } # # summarise.loglmlist <-function(object, ..., sortby=NULL){ # nmodels <- length(object) # if (nmodels == 1) # return(summarise.loglm(object[[1]])) # if (is.null(names(object))) { # oname <- as.character(sys.call())[-1] # oname <- oname[1:length(object)] # } # else oname <- names(object) # # resdf <- as.numeric(lapply(object, function(x) x$df)) # resdev <- as.numeric(lapply(object, function(x) x$deviance)) # n <- as.numeric(lapply(object, function(x) sum(fitted(x)))) # result <- stat.summarise(resdev, resdf, oname, n) # if (!is.null(sortby)) { # result <- result[order(result[,sortby], decreasing=TRUE),] # } # result # } # vcdExtra/R/seq_loglm.R0000644000176200001440000000571313163461153014351 0ustar liggesusers#' Sequential loglinear models for an n-way table #' This function takes an n-way contingency table and fits a series of sequential #' models to the 1-, 2-, ... n-way marginal tables, corresponding to a variety of #' types of loglinear models. #' @param x a contingency table in array form, with optional category labels specified in the dimnames(x) attribute, #' or else a data.frame in frequency form, with the frequency variable names "Freq". #' @param type type of sequential model to fit #' @param marginals which marginals to fit? #' @param vorder order of variables #' @param k indices of conditioning variable(s) for "joint", "conditional" or order for "markov" #' @param prefix #' @param fitted keep fitted values? seq_loglm <- function( x, type = c("joint", "conditional", "mutual", "markov", "saturated"), marginals = 1:nf, # which marginals to fit? vorder = 1:nf, # order of variables in the sequential models k = NULL, # conditioning variable(s) for "joint", "conditional" or order for "markov" prefix = 'model', fitted = TRUE, # keep fitted values? ... ) { if (inherits(x, "data.frame") && "Freq" %in% colnames(x)) { x <- xtabs(Freq ~ ., data=x) } if (!inherits(x, c("table", "array"))) stop("not an xtabs, table, array or data.frame with a 'Freq' variable") nf <- length(dim(x)) x <- aperm(x, vorder) factors <- names(dimnames(x)) indices <- 1:nf type = match.arg(type) # models <- as.list(rep(NULL, length(marginals))) models <- list() for (i in marginals) { mtab <- margin.table(x, 1:i) if (i==1) { # KLUDGE: use loglin, but try to make it look like a loglm object mod <- loglin(mtab, margin=NULL, print=FALSE) mod$model.string = paste("=", factors[1]) mod$margin <- list(factors[1]) # mod$margin <- names(dimnames(mtab)) # names(mod$margin) <- factors[1] if (fitted) { fit <- mtab fit[] <- (sum(mtab) / length(mtab)) mod$fitted <- fit } mod$nobs <- length(mtab) mod$frequencies <- mtab mod$deviance <- mod$lrt class(mod) <- c("loglin", "loglm") } else { expected <- switch(type, 'conditional' = conditional(i, mtab, with=if(is.null(k)) i else k), 'joint' = joint(i, mtab, with=if(is.null(k)) i else k), 'mutual' = mutual(i, mtab), 'markov' = markov(i, mtab, order=if(is.null(k)) 1 else k), 'saturated' = saturated(i, mtab) ) form <- loglin2formula(expected) # mod <- loglm(formula=form, data=mtab, fitted=TRUE) mod <- eval(bquote(MASS::loglm(.(form), data=mtab, fitted=fitted))) mod$model.string <- loglin2string(expected, brackets=if (iChisq)") rownames(rval) <- as.character(sapply(match.call(), deparse)[-1L])[1:nmodels] rval[,1] <- -2 * ll + 2 * par rval[,2] <- -2 * ll + log(ns) * par rval[,3] <- -2 * (ll - saturated) rval[,4] <- df rval[,5] <- pchisq(rval[,3], df, lower.tail = FALSE) if (!is.null(sortby)) { rval <- rval[order(rval[,sortby], decreasing=TRUE),] } ## return structure(as.data.frame(rval), heading = "Likelihood summary table:", class = c("anova", "data.frame")) } vcdExtra/R/mosaic3d.R0000644000176200001440000001634213163461153014071 0ustar liggesusers##################################### ## Produce a 3D mosaic plot using rgl ##################################### # TODO: provide formula interface # TODO: handle zero margins (causes display to be erased in shapelist3d) # DONE: handle zero cells # DONE: generalize the calculation of residuals # DONE: allow display of type=c("observed", "expected") # DONE: if ndim>3, provide for labels at max or min # DONE: make object oriented and provide a loglm method # mosaic3d: provide observed array of counts and either residuals, expected frequencies, # or a loglin set of margins to fit mosaic3d <- function(x, ...) { UseMethod("mosaic3d") } mosaic3d.loglm <- function (x, type = c("observed", "expected"), residuals_type = c("pearson", "deviance"), # gp = shading_hcl, gp_args = list(), ...) { residuals_type <- match.arg(tolower(residuals_type), c("pearson", "deviance")) if (is.null(x$fitted)) x <- update(x, fitted = TRUE) expected <- fitted(x) residuals <- residuals(x, type = "pearson") observed <- residuals * sqrt(expected) + expected if (residuals_type == "deviance") residuals <- residuals(x, type = "deviance") # gp <- if (inherits(gp, "grapcon_generator")) # do.call("gp", c(list(observed, residuals, expected, x$df), # as.list(gp_args))) # else gp mosaic3d.default(observed, residuals = residuals, expected = expected, type = type, residuals_type = residuals_type, # gp = gp, ...) } mosaic3d.default <- function(x, expected=NULL, residuals=NULL, type = c("observed", "expected"), residuals_type = NULL, shape=rgl::cube3d(alpha=alpha), alpha=0.5, spacing=0.1, split_dir=1:3, shading=shading_basic, interpolate=c(2,4), zero_size=.05, label_edge, labeling_args=list(), newpage=TRUE, box=FALSE, ...) { if (!requireNamespace("rgl")) stop("rgl is required") type <- match.arg(type) if (is.null(residuals)) { residuals_type <- if (is.null(residuals_type)) "pearson" else match.arg(tolower(residuals_type), c("pearson", "deviance", "ft")) } ## convert structable object if (is.structable(x)) { x <- as.table(x) } ## table characteristics levels <- dim(x) ndim <- length(levels) dn <- dimnames(x) if (is.null(dn)) dn <- dimnames(x) <- lapply(levels, seq) vnames <- names(dimnames(x)) if (is.null(vnames)) vnames <- names(dn) <- names(dimnames(x)) <- LETTERS[1:ndim] ## replace NAs by 0 if (any(nas <- is.na(x))) x[nas] <- 0 ## model fitting: ## calculate expected if needed if ((is.null(expected) && is.null(residuals)) || !is.numeric(expected)) { if (inherits(expected, "formula")) { fm <- loglm(expected, x, fitted = TRUE) expected <- fitted(fm) df <- fm$df } else { if (is.null(expected)) expected <- as.list(1:ndim) fm <- loglin(x, expected, fit = TRUE, print = FALSE) expected <- fm$fit df <- fm$df } } ## compute residuals if (is.null(residuals)) residuals <- switch(residuals_type, pearson = (x - expected) / sqrt(ifelse(expected > 0, expected, 1)), deviance = { tmp <- 2 * (x * log(ifelse(x == 0, 1, x / ifelse(expected > 0, expected, 1))) - (x - expected)) tmp <- sqrt(pmax(tmp, 0)) ifelse(x > expected, tmp, -tmp) }, ft = sqrt(x) + sqrt(x + 1) - sqrt(4 * expected + 1) ) ## replace NAs by 0 if (any(nas <- is.na(residuals))) residuals[nas] <- 0 # switch observed and expected if required observed <- if (type == "observed") x else expected expected <- if (type == "observed") expected else x # replicate arguments to number of dimensions spacing <- rep(spacing, length=ndim) split_dir <- rep(split_dir, length=ndim) if(missing(label_edge)) label_edge <- rep( c('-', '+'), each=3, length=ndim) zeros <- observed <= .Machine$double.eps shapelist <- shape # sanity check if (!inherits(shapelist, "shape3d")) stop("shape must be a shape3d object") if (newpage) rgl::open3d() for (k in 1:ndim) { marg <- margin.table(observed, k:1) if (k==1) { shapelist <- split3d(shapelist, marg, split_dir[k], space=spacing[k]) label3d(shapelist, split_dir[k], dn[[k]], vnames[k], edge=label_edge[k], ...) } else { marg <- matrix(marg, nrow=levels[k]) shapelist <- split3d(shapelist, marg, split_dir[k], space=spacing[k]) names(shapelist) <- apply(as.matrix(expand.grid(dn[1:k])), 1, paste, collapse=":") L <- length(shapelist) label_cells <- if (label_edge[k]=='-') 1:levels[k] else (L-levels[k]+1):L label3d(shapelist[label_cells], split_dir[k], dn[[k]], vnames[k], edge=label_edge[k], ...) } } # assign colors # TODO: allow alpha to control transparency of side walls col <- shading(residuals, interpolate=interpolate) # display, but exclude the zero cells rgl::shapelist3d(shapelist[!as.vector(zeros)], col=col[!as.vector(zeros)], ...) # plot markers for zero cells if (any(zeros)) { ctrs <- t(sapply(shapelist, center3d)) rgl::spheres3d(ctrs[as.vector(zeros),], radius=zero_size) } # invisible(structable(observed)) invisible(shapelist) } # basic shading_Friendly, adapting the simple code used in mosaicplot() shading_basic <- function(residuals, interpolate=TRUE) { if (is.logical(interpolate)) interpolate <- c(2, 4) else if (any(interpolate <= 0) || length(interpolate) > 5) stop("invalid 'interpolate' specification") shade <- sort(interpolate) breaks <- c(-Inf, -rev(shade), 0, shade, Inf) colors <- c(hsv(0, s = seq.int(1, to = 0, length.out = length(shade) + 1)), hsv(4/6, s = seq.int(0, to = 1, length.out = length(shade) + 1))) colors[as.numeric(cut(residuals, breaks))] } # provide labels for 3D objects below/above their extent along a given dimension # FIXME: kludge for interline gap between level labels and variable name # TODO: how to pass & extract labeling_args, e.g., labeling_args=list(at='min', fontsize=10) label3d <- function(objlist, dim, text, varname, offset=.05, adj, edge="-", gap=.1, labeling_args, ...) { if(missing(adj)) { if (dim < 3) adj <- ifelse(edge == '-', c(0.5, 1), c(0.5, 0)) else adj <- ifelse(edge == '-', c(1, 0.5), c(0, 0.5)) } ranges <- lapply(objlist, range3d) loc <- t(sapply(ranges, colMeans)) # positions of labels on dimension dim min <- t(sapply(ranges, function(x) x[1,])) # other dimensions at min values max <- t(sapply(ranges, function(x) x[2,])) # other dimensions at max values xyz <- if (edge == '-') (min - offset) else (max + offset) xyz[,dim] <- loc[,dim] if(!missing(varname)) { loclab <- colMeans(loc) # NB: doesn't take space into acct xyzlab <- if (edge == '-') min[1,] - offset - gap else max[1,] + offset + gap xyzlab[dim] <- loclab[dim] xyz <- rbind(xyz, xyzlab) text <- c(text, varname) } result <- c(labels = rgl::texts3d(xyz, texts=text, adj=adj, ...)) invisible(result) } vcdExtra/R/blogits.R0000644000176200001440000000116113163461153014023 0ustar liggesusers# calculate bivariate logits and OR blogits <- function(Y, add, colnames, row.vars, rev=FALSE) { if (ncol(Y) != 4) stop("Y must have 4 columns") if (missing(add)) add <- if (any(Y==0)) 0.5 else 0 Y <- Y + add if (rev) Y <- Y[,4:1] L <- matrix(0, nrow(Y), 3) L[,1] <- log( (Y[,1] + Y[,2]) / (Y[,3] + Y[,4]) ) L[,2] <- log( (Y[,1] + Y[,3]) / (Y[,2] + Y[,4]) ) L[,3] <- log( (Y[,1] * Y[,4]) / ((Y[,2] * Y[,3])) ) cn <- c("logit1", "logit2", "logOR") colnames(L) <- if(missing(colnames)) cn else c(colnames, cn[-(1:length(colnames))]) if(!missing(row.vars)) L <- cbind(L, row.vars) L } vcdExtra/R/HLtest.R0000644000176200001440000000401213163461153013561 0ustar liggesusers# Functions for Hosmer Lemeshow test # original function downloaded from # http://sas-and-r.blogspot.com/2010/09/example-87-hosmer-and-lemeshow-goodness.html # # see also: MKmisc::gof.test for more general versions HLtest <- HosmerLemeshow <- function(model, g=10) { if (!inherits(model, "glm")) stop("requires a binomial family glm") if (!family(model)$family == 'binomial') stop("requires a binomial family glm") y <- model$y yhat <- model$fitted.values cutyhat = cut(yhat, breaks = quantile(yhat, probs=seq(0, 1, 1/g)), include.lowest=TRUE) obs = xtabs(cbind(1 - y, y) ~ cutyhat) exp = xtabs(cbind(1 - yhat, yhat) ~ cutyhat) chi = (obs - exp)/sqrt(exp) # browser() table <- data.frame(cut=dimnames(obs)$cutyhat, total= as.numeric(apply(obs, 1, sum)), obs=as.numeric(as.character(obs[,1])), exp=as.numeric(as.character(exp[,1])), chi=as.numeric(as.character(chi[,1])) ) rownames(table) <- 1:g chisq = sum(chi^2) p = 1 - pchisq(chisq, g - 2) result <- list(table=table, chisq=chisq, df=g-2, p.value=p, groups=g, call=model$call) class(result) <- "HLtest" return(result) } print.HLtest <- function(x, ...) { heading <- "Hosmer and Lemeshow Goodness-of-Fit Test" df <- data.frame("ChiSquare"=x$chisq, df=x$df, "P_value"= x$p.value) cat(heading,"\n\n") cat("Call:\n") print(x$call) print(df, row.names=FALSE) invisible(x) } # Q: how to print **s next to larg chisq components? summary.HLtest <- function(object, ...) { heading <- "Partition for Hosmer and Lemeshow Goodness-of-Fit Test" cat(heading,"\n\n") print(object$table) print(object) } ## Q: how to display any large chi residuals on the bars?? rootogram.HLtest <- function(x, ...) { rootogram(as.numeric(x$table$obs), as.numeric(x$table$exp), xlab="Fitted value group", names=1:x$groups, ...) } plot.HLtest <- function(x, ...) { rootogram.HLtest(x, ...) } vcdExtra/R/glmlist.R0000644000176200001440000000462513163461153014043 0ustar liggesusers# glmlist - make a glmlist object containing a list of fitted glm objects with their names # borrowing code from Hmisc::llist glmlist <- function(...) { args <- list(...); lname <- names(args) name <- vname <- as.character(sys.call())[-1] for (i in 1:length(args)) { vname[i] <- if (length(lname) && lname[i] != "") lname[i] else name[i] } names(args) <- vname[1:length(args)] is.glm <- unlist(lapply(args, function(x) inherits(x, "glm"))) if (!all(is.glm)) { warning("Objects ", paste(vname[!is.glm], collapse=', '), " removed because they are not glm objects") args <- args[is.glm] } class(args) <- "glmlist" return(args); } # loglmlist - do the same for loglm objects loglmlist <- function(...) { args <- list(...); lname <- names(args) name <- vname <- as.character(sys.call())[-1] for (i in 1:length(args)) { vname[i] <- if (length(lname) && lname[i] != "") lname[i] else name[i] } names(args) <- vname[1:length(args)] is.loglm <- unlist(lapply(args, function(x) inherits(x, "loglm"))) if (!all(is.loglm)) { warning("Objects ", paste(vname[!is.loglm], collapse=', '), " removed because they are not loglm objects") args <- args[is.loglm] } class(args) <- "loglmlist" return(args); } # generic version: named list nlist <- function(...) { args <- list(...); lname <- names(args) name <- vname <- as.character(sys.call())[-1] for (i in 1:length(args)) { vname[i] <- if (length(lname) && lname[i] != "") lname[i] else name[i] } names(args) <- vname[1:length(args)] return(args); } # coeficient method for a glmlist (from John Fox, r-help, 10-28-2014) coef.glmlist <- function(object, result=c("list", "matrix", "data.frame"), ...){ result <- match.arg(result) coefs <- lapply(object, coef) if (result == "list") return(coefs) coef.names <- unique(unlist(lapply(coefs, names))) n.mods <- length(object) coef.matrix <- matrix(NA, length(coef.names), n.mods) rownames(coef.matrix) <- coef.names colnames(coef.matrix) <- names(object) for (i in 1:n.mods){ coef <- coef(object[[i]]) coef.matrix[names(coef), i] <- coef } if (result == "matrix") return(coef.matrix) as.data.frame(coef.matrix) } vcdExtra/R/Kway.R0000644000176200001440000000171113163461153013274 0ustar liggesusers# Generate and fit all 1-way, 2-way, ... k-way terms in a glm Kway <- function(formula, family=poisson, data, ..., order=nt, prefix="kway") { if (is.character(family)) family <- get(family, mode = "function", envir = parent.frame()) if (is.function(family)) family <- family() if (is.null(family$family)) { print(family) stop("'family' not recognized") } if (missing(data)) data <- environment(formula) models <- list() mod <- glm(formula, family=family, data, ...) mod$call$formula <- formula terms <- terms(formula) tl <- attr(terms, "term.labels") nt <- length(tl) models[[1]] <- mod for(i in 2:order) { models[[i]] <- update(mod, substitute(.~.^p, list(p = i))) } # null model mod0 <- update(mod, .~1) models <- c(list(mod0), models) names(models) <- paste(prefix, 0:order, sep = ".") class(models) <- "glmlist" models } vcdExtra/R/Summarise.R0000644000176200001440000000732313163461153014333 0ustar liggesusers# fixed buglet when deviance() returns a null # fixed bug: residual df calculated incorrectly # but this now depends on objects having a df.residual component # TRUE for lm, glm, polr, negbin objects # made generic, adding a glmlist method Summarise <- function(object, ...) { UseMethod("Summarise") } Summarise.glmlist <- function(object, ..., saturated = NULL, sortby=NULL) { ns <- sapply(object, function(x) length(x$residuals)) if (any(ns != ns[1L])) stop("models were not all fitted to the same size of dataset") nmodels <- length(object) if (nmodels == 1) return(Summarise.default(object[[1L]], saturated=saturated)) rval <- lapply(object, Summarise.default, saturated=saturated) rval <- do.call(rbind, rval) if (!is.null(sortby)) { rval <- rval[order(rval[,sortby], decreasing=TRUE),] } rval } # could just do Summarise.loglmlist <- Summarise.glmlist Summarise.loglmlist <- function(object, ..., saturated = NULL, sortby=NULL) { ns <- sapply(object, function(x) length(x$residuals)) if (any(ns != ns[1L])) stop("models were not all fitted to the same size of dataset") nmodels <- length(object) if (nmodels == 1) return(Summarise.default(object[[1L]], saturated=saturated)) rval <- lapply(object, Summarise.default, saturated=saturated) rval <- do.call(rbind, rval) if (!is.null(sortby)) { rval <- rval[order(rval[,sortby], decreasing=TRUE),] } rval } Summarise.default <- function(object, ..., saturated = NULL, sortby=NULL) { ## interface methods for logLik() and nobs() ## - use S4 methods if loaded ## - use residuals() if nobs() is not available logLik0 <- if("stats4" %in% loadedNamespaces()) stats4::logLik else logLik nobs0 <- function(x, ...) { nobs1 <- if("stats4" %in% loadedNamespaces()) stats4::nobs else nobs nobs2 <- function(x, ...) NROW(residuals(x, ...)) rval <- try(nobs1(x, ...), silent = TRUE) if(inherits(rval, "try-error") | is.null(rval)) rval <- nobs2(x, ...) return(rval) } dof <- function(x) { if (inherits(x, "loglm")) { rval <- x$df } else { rval <- try(x$df.residual, silent=TRUE) } if (inherits(rval, "try-error") || is.null(rval)) stop(paste("Can't determine residual df for a", class(x), "object")) rval } ## collect all objects objects <- list(object, ...) nmodels <- length(objects) ## check sample sizes ns <- sapply(objects, nobs0) if(any(ns != ns[1L])) stop("models were not all fitted to the same size of dataset") ## extract log-likelihood and df (number of parameters) ll <- lapply(objects, logLik0) par <- as.numeric(sapply(ll, function(x) attr(x, "df"))) df <- as.numeric(sapply(objects, function(x) dof(x))) ll <- sapply(ll, as.numeric) ## compute saturated reference value (use 0 if deviance is not available) if(is.null(saturated)) { dev <- try(sapply(objects, deviance), silent = TRUE) if(inherits(dev, "try-error") || any(sapply(dev, is.null))) { saturated <- 0 } else { saturated <- ll + dev/2 } } ## setup ANOVA-style matrix rval <- matrix(rep(NA, 5 * nmodels), ncol = 5) colnames(rval) <- c("AIC", "BIC", "LR Chisq", "Df", "Pr(>Chisq)") rownames(rval) <- as.character(sapply(match.call(), deparse)[-1L])[1:nmodels] rval[,1] <- -2 * ll + 2 * par rval[,2] <- -2 * ll + log(ns) * par rval[,3] <- -2 * (ll - saturated) rval[,4] <- df rval[,5] <- pchisq(rval[,3], df, lower.tail = FALSE) if (!is.null(sortby)) { rval <- rval[order(rval[,sortby], decreasing=TRUE),] } ## return structure(as.data.frame(rval), heading = "Likelihood summary table:", class = c("anova", "data.frame")) } vcdExtra/R/cutfac.R0000644000176200001440000000056413163461153013633 0ustar liggesusers# Cut a variable to a factor cutfac <- function(x, breaks = NULL, q=10) { if(is.null(breaks)) breaks <- unique(quantile(x, 0:q/q)) x <- cut(x, breaks, include.lowest = TRUE, right = FALSE) levels(x) <- paste(breaks[-length(breaks)], ifelse(diff(breaks) > 1, c(paste("-", breaks[-c(1, length(breaks))] - 1, sep = ""), "+"), ""), sep = "") return(x) } vcdExtra/R/seq_mosaic.R0000644000176200001440000000415713163461153014513 0ustar liggesusers#' Sequential Mosaics and Strucplots for an N-way Table #' This function takes an n-way contingency table and plots mosaics for series of sequential #' models to the 1-, 2-, ... n-way marginal tables, corresponding to a variety of #' types of loglinear models. #' @param x a contingency table in array form, with optional category labels specified in the dimnames(x) attribute, #' or else a data.frame in frequency form, with the frequency variable names "Freq". #' @param panel panel function #' @param type type of sequential model to fit #' @param plots which marginals to plot? #' @param vorder order of variables #' @param k indices of conditioning variable(s) for "joint", "conditional" or order for "markov" #' @export seq_mosaic <- function( x, panel = mosaic, type = c("joint", "conditional", "mutual", "markov", "saturated"), plots = 1:nf, # which plots to produce? vorder = 1:nf, # order of variables in the sequential plots k = NULL, # conditioning variable(s) for "joint", "conditional" or order for "markov" ... ) { if (inherits(x, "data.frame") && "Freq" %in% colnames(x)) { x <- xtabs(Freq ~ ., data=x) } if (!inherits(x, c("table", "array"))) stop("not an xtabs, table, array or data.frame with a 'Freq' variable") nf <- length(dim(x)) x <- aperm(x, vorder) factors <- names(dimnames(x)) indices <- 1:nf type = match.arg(type) for (i in plots) { mtab <- margin.table(x, 1:i) df <- NULL if (i==1) { expected <- mtab expected[] <- sum(mtab) / length(mtab) df <- length(mtab)-1 model.string = paste("=", factors[1]) } else { expected <- switch(type, 'conditional' = conditional(i, mtab, with=if(is.null(k)) i else k), 'joint' = joint(i, mtab, with=if(is.null(k)) i else k), 'mutual' = mutual(i, mtab), 'markov' = markov(i, mtab, order=if(is.null(k)) 1 else k), 'saturated' = saturated(i, mtab) ) model.string <- loglin2string(expected, brackets=if (i ncol(coords))) stop("dim must be valid dimensions of the coordinates") labs <- pctlab(obj) if (xlab == "_auto_") xlab <- labs[dim[1]] if (ylab == "_auto_") ylab <- labs[dim[2]] if(isTRUE(rev.axes[1])) coords[, dim[1]] <- -coords[, dim[1]] if(isTRUE(rev.axes[2])) coords[, dim[2]] <- -coords[, dim[2]] plot(coords[, dim], type='n', asp=1, xlab=xlab, ylab=ylab, ...) points(coords[,dim], pch=rep(pch, nlev), col=rep(col, nlev), cex=cex) text(coords[,dim], labels=coords$level, col=rep(col, nlev), pos=pos, cex=cex, xpd=TRUE) if (is.logical(lines)) lines <- if(lines) 1:nfac else NULL if(length(lines)) multilines(coords[, dim], group=coords$factor, which=lines, col=col, lwd=lwd) abline(h = 0, v = 0, lty = "longdash", col="gray") if (legend) { factors <- coords$factor factors <- factors[!duplicated(factors)] legend(legend.pos, legend=factors, title="Factor", title.col="black", col=col, text.col=col, pch=pch, bg=rgb(.95, .95, .95, .3), cex=cex) } invisible(coords) } pctlab <- function(obj, prefix="Dimension ", decimals=1) { values <- obj$sv^2 if (obj$lambda == "JCA"){ pct <- rep_len(NA, length(values)) } else { if (obj$lambda == "adjusted") { values <- obj$inertia.e pct <- round(100 * values, decimals) } else { pct <- round(100 * values / sum(values), decimals) } } pctval <- ifelse(is.na(pct), NULL, paste0(" (", pct, "%)")) paste0(prefix, 1:length(values), pctval) } vcdExtra/R/split3d.R0000644000176200001440000000363513163461153013752 0ustar liggesusers# split a 3D object along dimension dim, according to the proportions or # frequencies specified in vector p split3d <- function(obj, ...) { UseMethod("split3d") } split3d.shape3d <- function(obj, p, dim, space=.10, ...) { range <-range3d(obj) min <- range[1,] p <- p/sum(p) # assure proportions uspace <- space/(length(p)-1) # unit space between objects scales <- p * (1-space) shifts <- c(0, cumsum(p)[-length(p)])*diff(range[,dim]) result <- list() for (i in seq_along(p)) { xscale <- yscale <- zscale <- 1 xshift <- yshift <- zshift <- 0 if (dim == 1 || tolower(dim)=='x') { xscale <- scales[i] xshift <- shifts[i] + min[1]*(1-xscale) + (uspace * (i-1)) } else if (dim == 2|| tolower(dim)=='y') { yscale <- scales[i] yshift <- shifts[i] + min[2]*(1-yscale) + (uspace * (i-1)) } else if (dim == 3|| tolower(dim)=='y') { zscale <- scales[i] zshift <- shifts[i] + min[3]*(1-zscale) + (uspace * (i-1)) } result[[i]] <- rgl::translate3d(rgl::scale3d(obj, xscale, yscale, zscale), xshift, yshift, zshift) } result } # split a list of 3D objects, according to the proportions specified in # the columns of p. split3d.list <- function(obj, p, dim, space=.10, ...) { nl <- length(obj) if (!is.matrix(p) || ncol(p) != nl) stop(gettextf("p must be a matrix with %i columns", nl)) sl <- list() for (i in seq_along(obj)) { sl <- c(sl, split3d(obj[[i]], p[,i], dim=dim, space=space)) } sl } #range3d <- function(obj, ...) { # UseMethod("range3d") #} range3d <- function(obj) { if (!"vb" %in% names(obj)) stop("Not a mesh3d or shape3d object") x <- with(obj, range(vb[1,]/vb[4,])) y <- with(obj, range(vb[2,]/vb[4,])) z <- with(obj, range(vb[3,]/vb[4,])) result <- cbind(x,y,z) rownames(result)<- c('min', 'max') result } center3d <- function(obj) { range <-range3d(obj) colMeans(range) } vcdExtra/R/Crossings.R0000644000176200001440000000230213163461153014330 0ustar liggesusers# crossings model (Goodman, 1972) # Ref: #Goodman, L. (1972). Some multiplicative models for the analysis of cross-classified data. #In: Proceedings of the Sixth Berkeley Symposium on Mathematical Statistics and Probability, #Berkeley, CA: University of California Press, pp. 649-696. crossings <- function(i, j, n) { npar <- n - 1 result <- list() for(c in 1:npar) { overi <- c >= i overj <- c >= j result[[c]] <- (overi & !overj) + (overj & !overi) } result <- matrix(unlist(result), length(i), npar) colnames(result) <- paste('C', 1:npar, sep='') result } Crossings <- function(...) { dots <- list(...) if (length(dots) != 2) stop("Crossings() is defined for only two factors") if (length(dots[[1]]) != length(dots[[2]])) stop("arguments to Crossings() must all have same length") dots <- lapply(dots, as.factor) n <- nlevels(dots[[1]]) if (nlevels(dots[[2]]) != n) stop("arguments to Crossings() must all have same number of levels") result <- crossings(as.numeric(dots[[1]]), as.numeric(dots[[2]]), n) rownames(result) <- do.call("paste", c(dots, sep = "")) result } vcdExtra/R/vcdExtra-deprecated.R0000644000176200001440000000014713163461153016241 0ustar liggesusers#summarise <- function (...) { # .Deprecated("summarise", package="vcdExtra") # LRstats(...) #} # vcdExtra/R/update.xtabs.R0000644000176200001440000000126413163461153014766 0ustar liggesusersupdate.xtabs <- function (object, formula., ..., evaluate = TRUE) { if (is.null(call<-attr(object, "call"))) stop("need an object with call component") extras <- match.call(expand.dots = FALSE)$... if (!missing(formula.)) call$formula <- update.formula(call$formula, formula.) if (length(extras)) { existing <- !is.na(match(names(extras), names(call))) for (a in names(extras)[existing]) call[[a]] <- extras[[a]] if (any(!existing)) { call <- c(as.list(call), extras[!existing]) call <- as.call(call) } } if (evaluate) eval(call, parent.frame()) else call } vcdExtra/R/collapse.table.R0000644000176200001440000000170213163461153015251 0ustar liggesusers# collapse a contingency table or ftable by re-assigning levels of table variables # revised to accept an array also collapse.table <- function(table, ...) { nargs <- length(args <- list(...)) if (!nargs) return(table) if (inherits(table, "ftable")) table <- as.table(table) if (inherits(table, "array")) table <- as.table(table) if (inherits(table, "table")) { tvars <- names(dimnames(table)) table <- as.data.frame.table(table) freq <- table[,"Freq"] } else stop("Argument must be a table, array or ftable object") names <- names(args) for (i in 1:nargs) { vals <- args[[i]] nm <- names[[i]] if(any(nm==tvars)) levels(table[[nm]]) <- vals else warning(nm, " is not among the table variables.") } # term <- paste(tvars, collapse = '+') # form <- as.formula(paste("freq ~", term)) # cat("term: ", term, "\n") xtabs(as.formula(paste("freq ~", paste(tvars, collapse = '+'))), data=table) } vcdExtra/R/print.Kappa.R0000644000176200001440000000103213163461153014544 0ustar liggesusers# Print method for Kappa: Add a column showing z values ## DONE: now set digits ## DONE: now include CI print.Kappa <- function (x, digits=max(getOption("digits") - 3, 3), CI=FALSE, level=0.95, ...) { tab <- rbind(x$Unweighted, x$Weighted) z <- tab[,1] / tab[,2] tab <- cbind(tab, z) if (CI) { q <- qnorm((1 + level)/2) lower <- tab[,1] - q * tab[,2] upper <- tab[,1] + q * tab[,2] tab <- cbind(tab, lower, upper) } rownames(tab) <- names(x)[1:2] print(tab, digits=digits, ...) invisible(x) } vcdExtra/R/expand.dft.R0000644000176200001440000000214513163461153014416 0ustar liggesusers# Author: Marc Schwarz # Ref: http://tolstoy.newcastle.edu.au/R/e6/help/09/01/1873.html expand.dft <- function(x, var.names = NULL, freq = "Freq", ...) { # allow: a table object, or a data frame in frequency form if(inherits(x, "table")) x <- as.data.frame.table(x, responseName = freq) freq.col <- which(colnames(x) == freq) if (length(freq.col) == 0) stop(paste(sQuote("freq"), "not found in column names")) DF <- sapply(1:nrow(x), function(i) x[rep(i, each = x[i, freq.col]), ], simplify = FALSE) DF <- do.call("rbind", DF)[, -freq.col, drop=FALSE] for (i in 1:ncol(DF)) { DF[[i]] <- type.convert(as.character(DF[[i]]), ...) } rownames(DF) <- NULL if (!is.null(var.names)) { if (length(var.names) < dim(DF)[2]) { stop(paste("Too few", sQuote("var.names"), "given.")) } else if (length(var.names) > dim(DF)[2]) { stop(paste("Too many", sQuote("var.names"), "given.")) } else { names(DF) <- var.names } } DF } # make this a synonym expand.table <- expand.dft vcdExtra/R/CMHtest.R0000644000176200001440000002251613163461153013676 0ustar liggesusers# Cochran-Mantel-Haenszel tests for ordinal factors in contingency tables # The code below follows Stokes, Davis & Koch, (2000). # "Categorical Data Analysis using the SAS System", 2nd Ed., # pp 74--75, 92--101, 124--129. # Ref: Landis, R. J., Heyman, E. R., and Koch, G. G. (1978), # Average Partial Association in Three-way Contingency Tables: # A Review and Discussion of Alternative Tests, # International Statistical Review, 46, 237-254. # See: https://onlinecourses.science.psu.edu/stat504/book/export/html/90 # http://support.sas.com/documentation/cdl/en/statug/63033/HTML/default/viewer.htm#statug_freq_a0000000648.htm # DONE: this should be the main function, handling 2-way & higher-way tables # With strata, use apply() or recursion over strata # DONE: With strata, calculate overall CMH tests controlling for strata # FIXED: rmeans and cmeans tests were labeled incorrectly CMHtest <- function(x, ...) UseMethod("CMHtest") CMHtest.formula <- function(formula, data = NULL, subset = NULL, na.action = NULL, ...) { m <- match.call(expand.dots = FALSE) edata <- eval(m$data, parent.frame()) fstr <- strsplit(paste(deparse(formula), collapse = ""), "~") vars <- strsplit(strsplit(gsub(" ", "", fstr[[1]][2]), "\\|")[[1]], "\\+") varnames <- vars[[1]] condnames <- if (length(vars) > 1) vars[[2]] else NULL dep <- gsub(" ", "", fstr[[1]][1]) if (!dep %in% c("","Freq")) { if (all(varnames == ".")) { varnames <- if (is.data.frame(data)) colnames(data) else names(dimnames(as.table(data))) varnames <- varnames[-which(varnames %in% dep)] } varnames <- c(varnames, dep) } if (inherits(edata, "ftable") || inherits(edata, "table") || length(dim(edata)) > 2) { condind <- NULL dat <- as.table(data) if(all(varnames != ".")) { ind <- match(varnames, names(dimnames(dat))) if (any(is.na(ind))) stop(paste("Can't find", paste(varnames[is.na(ind)], collapse=" / "), "in", deparse(substitute(data)))) if (!is.null(condnames)) { condind <- match(condnames, names(dimnames(dat))) if (any(is.na(condind))) stop(paste("Can't find", paste(condnames[is.na(condind)], collapse=" / "), "in", deparse(substitute(data)))) ind <- c(condind, ind) } dat <- margin.table(dat, ind) } CMHtest.default(dat, strata = if (is.null(condind)) NULL else match(condnames, names(dimnames(dat))), ...) } else { m <- m[c(1, match(c("formula", "data", "subset", "na.action"), names(m), 0))] m[[1]] <- as.name("xtabs") m$formula <- formula(paste(if("Freq" %in% colnames(data)) "Freq", "~", paste(c(varnames, condnames), collapse = "+"))) tab <- eval(m, parent.frame()) CMHtest.default(tab, ...) } } CMHtest.default <- function(x, strata = NULL, rscores=1:R, cscores=1:C, types=c("cor", "rmeans", "cmeans", "general"), overall=FALSE, details=overall, ...) { snames <- function(x, strata) { sn <- dimnames(x)[strata] dn <- names(sn) apply(expand.grid(sn), 1, function(x) paste(dn, x, sep=":", collapse = "|")) } ## check dimensions L <- length(d <- dim(x)) if(any(d < 2L)) stop("All table dimensions must be 2 or greater") if(L > 2L & is.null(strata)) strata <- 3L:L if(is.character(strata)) strata <- which(names(dimnames(x)) == strata) if(L - length(strata) != 2L) stop("All but 2 dimensions must be specified as strata.") ## rearrange table to put primary dimensions first x <- aperm(x, c(setdiff(1:L, strata), strata)) d <- dim(x) R <- d[1] C <- d[2] # handle strata if (!is.null(strata)) { sn <- snames(x, strata) res <- c(apply(x, strata, CMHtest2, rscores=rscores, cscores=cscores, types=types,details=details, ...)) # DONE: fix names if there are 2+ strata names(res) <- sn for (i in seq_along(res)) res[[i]]$stratum <- sn[i] # DONE: Calculate generalized CMH, controlling for strata if (overall) { if (!details) warning("Overall CMH tests not calculated because details=FALSE") else { resall <- CMHtest3(res, types=types) res$ALL <- resall } } return(res) } else CMHtest2(x, rscores=rscores, cscores=cscores, types=types,details=details, ...) } # handle two-way case, for a given stratum # DONE: now allow rscores/cscores == 'midrank' for midrank scores # DONE: allow rscores/cscores=NULL for unordered factors, where ordinal # scores don't make sense # DONE: modified to return all A matrices as a list # DONE: cmh() moved outside CMHtest2 <- function(x, stratum=NULL, rscores=1:R, cscores=1:C, types=c("cor", "rmeans", "cmeans", "general"), details=FALSE, ...) { # left kronecker product lkronecker <- function(x, y, make.dimnames=TRUE, ...) kronecker(y, x, make.dimnames=make.dimnames, ...) # midrank scores (modified ridits) based on row/column totals midrank <- function (n) { cs <- cumsum(n) (2*cs - n +1) / (2*(cs[length(cs)]+1)) } L <- length(d <- dim(x)) R <- d[1] C <- d[2] if (is.character(rscores) && rscores=="midrank") rscores <- midrank(rowSums(x)) if (is.character(cscores) && cscores=="midrank") cscores <- midrank(colSums(x)) nt <- sum(x) pr <- rowSums(x) / nt pc <- colSums(x) / nt m <- as.vector(nt * outer(pr,pc)) # expected values under independence n <- as.vector(x) # cell frequencies V1 <- (diag(pr) - pr %*% t(pr)) V2 <- (diag(pc) - pc %*% t(pc)) V <- (nt^2/(nt-1)) * lkronecker(V1, V2, make.dimnames=TRUE) if (length(types)==1 && types=="ALL") types <- c("general", "rmeans", "cmeans", "cor" ) types <- match.arg(types, several.ok=TRUE) # handle is.null(rscores) etc here if (is.null(rscores)) types <- setdiff(types, c("cmeans", "cor")) if (is.null(cscores)) types <- setdiff(types, c("rmeans", "cor")) table <- NULL Amats <- list() if("cor" %in% types) { A <- lkronecker( t(rscores), t(cscores) ) df <- 1 table <- rbind(table, cmh(n, m, A, V, df)) Amats$cor <- A } if("rmeans" %in% types) { A <- lkronecker( cbind(diag(R-1), rep(0, R-1)), t(cscores)) df <- R-1 table <- rbind(table, cmh(n, m, A, V, df)) Amats$rmeans <- A } if("cmeans" %in% types) { A <- lkronecker( t(rscores), cbind(diag(C-1), rep(0, C-1))) df <- C-1 table <- rbind(table, cmh(n, m, A, V, df)) Amats$cmeans <- A } if ("general" %in% types) { A <- lkronecker( cbind(diag(R-1), rep(0, R-1)), cbind(diag(C-1), rep(0, C-1))) df <- (R-1)*(C-1) table <- rbind(table, cmh(n, m, A, V, df)) Amats$general <- A } colnames(table) <- c("Chisq", "Df", "Prob") rownames(table) <- types xnames <- names(dimnames(x)) result <- list(table=table, names=xnames, rscores=rscores, cscores=cscores, stratum=stratum ) if (details) result <- c(result, list(A=Amats, V=V, n=n, m=m)) class(result) <- "CMHtest" result } # do overall test, from a computed CMHtest list CMHtest3 <- function(object, types=c("cor", "rmeans", "cmeans", "general")) { nstrat <- length(object) # number of strata # extract components, each a list of nstrat terms n.list <- lapply(object, function(s) s$n) m.list <- lapply(object, function(s) s$m) V.list <- lapply(object, function(s) s$V) A.list <- lapply(object, function(s) s$A) nt <- sapply(lapply(object, function(s) s$n), sum) Df <- object[[1]]$table[,"Df"] if (length(types)==1 && types=="ALL") types <- c("general", "rmeans", "cmeans", "cor" ) types <- match.arg(types, several.ok=TRUE) table <- list() for (type in types) { AVA <- 0 Anm <- 0 for (k in 1:nstrat) { A <- A.list[[k]][[type]] V <- V.list[[k]] n <- n.list[[k]] m <- m.list[[k]] AVA <- AVA + A %*% V %*% t(A) Anm <- Anm + A %*% (n-m) } Q <- t(Anm) %*% solve(AVA) %*% Anm df <- Df[type] pvalue <- pchisq(Q, df, lower.tail=FALSE) table <- rbind(table, c(Q, df, pvalue)) } rownames(table) <- types colnames(table) <- c("Chisq", "Df", "Prob") xnames <- object[[1]]$names result=list(table=table, names=xnames, stratum="ALL") class(result) <- "CMHtest" result } # basic CMH calculation cmh <- function(n, m,A, V, df) { AVA <- A %*% V %*% t(A) Q <- t(n-m) %*% t(A) %*% solve(AVA) %*% A %*% (n-m) pvalue <- pchisq(Q, df, lower.tail=FALSE) c(Q, df, pvalue) } # DONE: incorporate stratum name in the heading # TODO: handle the printing of pvalues better print.CMHtest <- function(x, digits = max(getOption("digits") - 2, 3), ...) { heading <- "Cochran-Mantel-Haenszel Statistics" if (!is.null(x$names)) heading <- paste(heading, "for", paste(x$names, collapse=" by ")) if (!is.null(x$stratum)) heading <- paste(heading, ifelse(x$stratum=="ALL", "\n\tOverall tests, controlling for all strata", paste("\n\tin stratum", x$stratum))) # TODO: determine score types (integer, midrank) for heading df <- x$table types <- rownames(df) labels <- list(cor="Nonzero correlation", rmeans="Row mean scores differ", cmeans="Col mean scores differ", general="General association") labels <- unlist(labels[types]) # select the labels for the types df <- data.frame("AltHypothesis"=as.character(labels), df, stringsAsFactors=FALSE) cat(heading,"\n\n") print(df, digits=digits, ...) cat("\n") invisible(x) } vcdExtra/R/mosaic.glmlist.R0000644000176200001440000001137613163461153015316 0ustar liggesusers#' Mosaic Displays for a glmlist Object #' @param x a glmlist object #' @param selection the index or name of one glm in \code{x} #' @param panel panel function #' @param type a character string indicating whether the \code{"observed"} or the \code{"expected"} values of the table should be visualized #' @param legend show a legend in the mosaic displays? #' @param main either a logical, or a vector of character strings used for plotting the main title. If main is a logical and TRUE, the name of the selected glm object is used #' @param ask should the function display a menu of models, when one is not specified in \code{selection}? #' @param graphics use a graphic menu when \code{ask=TRUE}? #' @param rows,cols when \code{ask=FALSE}, the number of rows and columns in which to plot the mosaics #' @param newpage start a new page? (only applies to \code{ask=FALSE}) #' @param ... other arguments passed to \code{\link{mosaic.glm}} #' @export mosaic.glmlist <- function(x, selection, panel=mosaic, type=c("observed", "expected"), legend=ask | !missing(selection), main=NULL, ask=TRUE, graphics=TRUE, rows, cols, newpage=TRUE, ...) { # calls <- sapply(x, mod.call) # get model calls as strings models <- names(x) if (!is.null(main)) { if (is.logical(main) && main) main <- models } else main <- rep(main, length(x)) type=match.arg(type) if (!missing(selection)){ if (is.character(selection)) selection <- gsub(" ", "", selection) return(panel(x[[selection]], type=type, main=main[selection], legend=legend, ...)) } # perhaps make these model labels more explicit for the menu if (ask & interactive()){ repeat { selection <- menu(models, graphics=graphics, title="Select Model to Plot") if (selection == 0) break else panel(x[[selection]], type=type, main=main[selection], legend=legend, ...) } } else { nmodels <- length(x) mfrow <- mfrow(nmodels) if (missing(rows) || missing(cols)){ rows <- mfrow[1] cols <- mfrow[2] } if (newpage) grid.newpage() lay <- grid.layout(nrow=rows, ncol = cols) pushViewport(viewport(layout = lay, y = 0, just = "bottom")) for (i in 1:rows) { for (j in 1:cols){ if ((sel <-(i-1)*cols + j) > nmodels) break pushViewport(viewport(layout.pos.row=i, layout.pos.col=j)) panel(x[[sel]], type=type, main=main[sel], newpage=FALSE, legend=legend, ...) popViewport() } } } } mosaic.loglmlist <- function(x, selection, panel=mosaic, type=c("observed", "expected"), legend=ask | !missing(selection), main=NULL, ask=TRUE, graphics=TRUE, rows, cols, newpage=TRUE, ...) { models <- names(x) strings <- as.vector(sapply(x, function(x) x$model.string)) if (!is.null(main)) { if (is.logical(main) && main) main <- ifelse(as.vector(sapply(strings, is.null)), models, strings) } else main <- rep(main, length(x)) type=match.arg(type) if (!missing(selection)){ if (is.character(selection)) selection <- gsub(" ", "", selection) return(panel(x[[selection]], type=type, main=main[selection], legend=legend, ...)) } # perhaps make these model labels more explicit for the menu if (ask & interactive()){ repeat { selection <- menu(models, graphics=graphics, title="Select Model to Plot") if (selection == 0) break else panel(x[[selection]], type=type, main=main[selection], legend=legend, ...) } } else { nmodels <- length(x) mfrow <- mfrow(nmodels) if (missing(rows) || missing(cols)){ rows <- mfrow[1] cols <- mfrow[2] } if (newpage) grid.newpage() lay <- grid.layout(nrow=rows, ncol = cols) pushViewport(viewport(layout = lay, y = 0, just = "bottom")) for (i in 1:rows) { for (j in 1:cols){ if ((sel <-(i-1)*cols + j) > nmodels) break pushViewport(viewport(layout.pos.row=i, layout.pos.col=j)) panel(x[[sel]], type=type, main=main[sel], newpage=FALSE, legend=legend, ...) popViewport() } } } } # from effects::utilities.R mfrow <- function(n, max.plots=0){ # number of rows and columns for array of n plots if (max.plots != 0 & n > max.plots) stop(paste("number of plots =",n," exceeds maximum =", max.plots)) rows <- round(sqrt(n)) cols <- ceiling(n/rows) c(rows, cols) } # from plot.lm: get model call as a string # TODO: should use abbreviate() mod.call <- function(x) { cal <- x$call if (!is.na(m.f <- match("formula", names(cal)))) { cal <- cal[c(1, m.f)] names(cal)[2L] <- "" } cc <- deparse(cal, 80) nc <- nchar(cc[1L], "c") abbr <- length(cc) > 1 || nc > 75 cap <- if (abbr) paste(substr(cc[1L], 1L, min(75L, nc)), "...") else cc[1L] cap } vcdExtra/R/loglin-utilities.R0000644000176200001440000001375013163461153015664 0ustar liggesusers#' Loglinear Model Utilities #' These functions generate lists of terms to specify a loglinear model #' in a form compatible with loglin and provide for conversion to an #' equivalent loglm specification. They allow for a more conceptual #' way to specify such models. #' models of joint independence, of some factors wrt one or more other factors #' @param nf number of factors for which to generate model #' @param table a contingency table used for factor names, typically the output from \code{\link[base]{table}} #' @param factors names of factors used in the model when \code{table} is not specified #' @param with indices of the factors against which others are considered jointly independent #' @export joint <- function(nf, table=NULL, factors=1:nf, with=nf) { if (!is.null(table)) factors <- names(dimnames(table)) if (nf == 1) return (list(term1=factors[1])) if (nf == 2) return (list(term1=factors[1], term2=factors[2])) others <- setdiff(1:nf, with) result <- list(term1=factors[others], term2=factors[with]) result } #' models of conditional independence of some factors wrt one or more other factors #' @param nf number of factors for which to generate model #' @param table a contingency table used for factor names, typically the output from \code{\link[base]{table}} #' @param factors names of factors used in the model when \code{table} is not specified #' @param with indices of the factors against which others are considered conditionally independent #' @export conditional <- function(nf, table=NULL, factors=1:nf, with=nf) { if (!is.null(table)) factors <- names(dimnames(table)) if (nf == 1) return (list(term1=factors[1])) if (nf == 2) return (list(term1=factors[1], term2=factors[2])) main <- setdiff(1:nf, with) others <- matrix(factors[with], length(with), length(main)) result <- rbind(factors[main], others) result <- as.list(as.data.frame(result, stringsAsFactors=FALSE)) names(result) <- paste('term', 1:length(result), sep='') result } #' models of mutual independence of all factors #' @param nf number of factors for which to generate model #' @param table a contingency table used for factor names, typically the output from \code{\link[base]{table}} #' @param factors names of factors used in the model when \code{table} is not specified #' @export mutual <- function(nf, table=NULL, factors=1:nf) { if (!is.null(table)) factors <- names(dimnames(table)) result <- sapply(factors[1:nf], list) names(result) <- paste('term', 1:length(result), sep='') result } #' saturated model: highest-order interaction #' @param nf number of factors for which to generate model #' @param table a contingency table used for factor names, typically the output from \code{\link[base]{table}} #' @param factors names of factors used in the model when \code{table} is not specified #' @export saturated <- function(nf, table=NULL, factors=1:nf) { if (!is.null(table)) factors <- names(dimnames(table)) list(term1=factors[1:nf]) } # models of conditional independence, given one pair of variables ## Not needed: handled by condit, with length(with)>1 #condit2 <- function(nf, factors=1:nf, with=1:2) { # if (nf == 1) return (list(term1=factors[1])) # if (nf == 2) return (list(term1=factors[1], term2=factors[2])) # others <- setdiff(1:nf, with) # result <- rbind(factors[with], cbind(factors[others], factors[others])) # result <- as.list(as.data.frame(result, stringsAsFactors=FALSE)) # names(result) <- paste('term', 1:length(result), sep='') # result #} #' markov models of a given order #' @param nf number of factors for which to generate model #' @param table a contingency table used for factor names, typically the output from \code{\link[base]{table}} #' @param factors names of factors used in the model when \code{table} is not specified #' @param order order of the markov chain #' @export markov <- function(nf, factors=1:nf, order=1) { if (nf == 1) return (list(term1=factors[1])) if (nf == 2) return (list(term1=factors[1], term2=factors[2])) if (length(factors) < order+2) { warning(paste('Not enough factors for order', order, 'Markov chain; using order=1')) order <-1 result <- rbind(factors[1:(nf-1)], factors[2:nf]) } else { if (nf <= order+1) result <- factors[1:nf] else { result <- NULL for (i in 1:(order+1)) result <- rbind(result, factors[i:(nf-order+i-1)]) } } result <- as.list(as.data.frame(result, stringsAsFactors=FALSE)) names(result) <- paste('term', 1:length(result), sep='') result } #' convert a loglin model to a model formula for loglm #' @param x a list of terms in a loglinear model, such as returned by \code{joint}, \code{conditional}, \dots #' @param env environment in which to evaluate the formula #' @source Code from Henrique Dallazuanna, , R-help 7-4-2013 loglin2formula <- function(x, env = parent.frame()) { terms <- lapply(x, paste, collapse = ":") formula(sprintf(" ~ %s", do.call(paste, c(terms, sep = "+"))), env=env) } #' convert a loglin model to a string, using bracket notation for the high-order terms #' @param x a list of terms in a loglinear model, such as returned by \code{joint}, \code{conditional}, \dots #' @param brackets characters to use to surround model terms. Either a single character string containing two characters #' or a character vector of length two. #' @param sep characters used to separate factor names within a term #' @param collapse characters used to separate terms #' @param abbrev loglin2string <- function(x, brackets = c('[', ']'), sep=',', collapse=' ', abbrev) { if (length(brackets)==1 && (nchar(brackets)>1)) brackets <- unlist(strsplit(brackets, "")) terms <- lapply(x, paste, collapse=sep) terms <- paste(brackets[1], terms, brackets[2], sep='') paste(terms, collapse= ' ') } vcdExtra/R/GKgamma.R0000644000176200001440000000320013163461153013660 0ustar liggesusers# Calculate Goodman-Kruskal Gamma # Original from: Laura Thompson, # https://home.comcast.net/~lthompson221/Splusdiscrete2.pdf GKgamma<-function(x, level=0.95) { # x is a matrix of counts. You can use output of crosstabs or xtabs in R. # Confidence interval calculation and output from Greg Rodd # Check for using S-PLUS and output is from crosstabs (needs >= S-PLUS 6.0) if(is.null(version$language) && inherits(x, "crosstabs")) { oldClass(x)<-NULL; attr(x, "marginals")<-NULL} ## TODO: add tests for matrix or table n <- nrow(x) m <- ncol(x) pi.c<-pi.d<-matrix(0, nrow=n, ncol=m) row.x<-row(x) col.x<-col(x) for(i in 1:(n)){ for(j in 1:(m)){ pi.c[i, j]<-sum(x[row.xi & col.x>j]) pi.d[i, j]<-sum(x[row.xj]) + sum(x[row.x>i & col.x 1, maxTitle=NULL) { # make sure requested packages are available and loaded pkgs <- .packages() for (i in seq_along(package)) { if (! package[i] %in% pkgs) if (require(package[i], character.only=TRUE, quietly=TRUE)) cat(paste("Loading package:", package[i], "\n")) else stop(paste("Package", package[i], "is not available")) } dsitems <- data(package=package)$results wanted <- if (incPackage) c('Package', 'Item','Title') else c('Item','Title') ds <- as.data.frame(dsitems[,wanted], stringsAsFactors=FALSE) # fix items with " (...)" in names, e.g., "BJsales.lead (BJsales)" in datasets ds$Item <- gsub(" .*", "", ds$Item) getDim <- function(x) { if (is.null(dim(get(x)))) length(get(x)) else paste(dim(get(x)), collapse='x') } getClass <- function(x) { cl <- class(get(x)) if (length(cl)>1 && !allClass) cl[length(cl)] else cl } ds$dim <- unlist(lapply(ds$Item, getDim )) ds$class <- unlist(lapply(ds$Item, getClass )) if (!is.null(maxTitle)) ds$Title <- substr(ds$Title, 1, maxTitle) if (incPackage) ds[c('Package', 'Item','class','dim','Title')] else ds[c('Item','class','dim','Title')] } vcdExtra/R/zero.test.R0000644000176200001440000000261013163461153014315 0ustar liggesusers# Score test for zero inflation in Poisson data #https://stats.stackexchange.com/questions/118322/how-to-test-for-zero-inflation-in-a-dataset # References: # Broek, Jan van den. 1995. ?A Score Test for Zero Inflation in a Poisson Distribution.? Biometrics 51 (2): 738?43. doi:10.2307/2532959. # Yang, Zhao, James W. Hardin, and Cheryl L. Addy. 2010. ?Score Tests for Zero-Inflation in Overdispersed Count Data.? Communications in Statistics - Theory and Methods 39 (11): 2008?30. doi:10.1080/03610920902948228 # Van den Broek, J. (1995). A Score Test for Zero Inflation in a Poisson Distribution. Biometrics, 51(2), 738-743. doi:10.2307/2532959 zero.test <- function(x) { if(is.table(x)) { # expand to vector of values if(length(dim(x)) > 1) stop ("x must be a 1-way table") x <- rep(as.numeric(names(x)), unname(c(x))) } lambda <- mean(x) p0_tilde <- exp(-lambda) n0 <- sum(1*(!(x >0))) n <- length(x) numerator <- (n0 - n*p0_tilde)^2 denominator <- n*p0_tilde*(1-p0_tilde) - n*lambda*(p0_tilde^2) stat <- numerator/denominator pvalue <- pchisq(stat,df=1, ncp=0, lower.tail=FALSE) result <- list(statistic=stat, df=1, prob=pvalue) cat(paste("Score test for zero inflation\n\n", "\tChi-square =", round(stat,5), "\n", "\tdf = 1\n", "\tpvalue:", format.pval(pvalue), "\n")) invisible(result) } vcdExtra/R/modFit.R0000644000176200001440000000164713163461153013613 0ustar liggesusers## ## One-line summary of model fit for a glm/loglm object ## `modFit` <- function(x, ...) UseMethod("modFit") modFit.glm <- function(x, stats="chisq", digits=2, ...) { if (!inherits(x,"glm")) stop("modFit requires a glm object") result <- NULL if ("chisq" %in% stats) result <- paste("G^2(",x$df.residual,")=", formatC(x$deviance,digits=digits,format="f"),sep="") if ("aic" %in% stats) result <- paste(result, " AIC=", formatC(x$aic,digits=digits,format="f"),sep="") result } modFit.loglm <- function(x, stats="chisq", digits=2, ...) { if (!inherits(x,"loglm")) stop("modFit requires a loglm object") result <- NULL if ("chisq" %in% stats) result <- paste("G^2(",x$df,")=", formatC(x$deviance,digits=digits,format="f"),sep="") if ("aic" %in% stats) { aic<-x$deviance-x$df*2 result <- paste(result, " AIC=", formatC(aic,digits=digits,format="f"),sep="") } result } vcdExtra/vignettes/0000755000176200001440000000000013163514377014054 5ustar liggesusersvcdExtra/vignettes/vcd-tutorial.Rnw0000644000176200001440000023435213163461153017163 0ustar liggesusers% !Rnw weave = Sweave %\VignetteEngine{Sweave} %\VignetteIndexEntry{Tutorial: Working with categorical data with R and the vcd package} %\VignetteDepends{vcd,gmodels,ca} %\VignetteKeywords{contingency tables, mosaic plots, sieve plots, categorical data, independence, conditional independence, R} %\VignettePackage{vcdExtra} \documentclass[10pt,twoside]{article} \usepackage{Sweave} \usepackage{bm} \usepackage[toc]{multitoc} % for table of contents % from Z.cls \usepackage[authoryear,round,longnamesfirst]{natbib} \bibpunct{(}{)}{;}{a}{}{,} \bibliographystyle{jss} \usepackage{hyperref} \usepackage{color} %% colors \definecolor{Red}{rgb}{0.7,0,0} \definecolor{Blue}{rgb}{0,0,0.8} \hypersetup{% hyperindex = {true}, colorlinks = {true}, % linktocpage = {true}, plainpages = {false}, linkcolor = {Blue}, citecolor = {Blue}, urlcolor = {Red}, pdfstartview = {Fit}, pdfpagemode = {UseOutlines}, pdfview = {XYZ null null null} } %\AtBeginDocument{ % \hypersetup{% % pdfauthor = {Michael Friendly}, % pdftitle = {Tutorial: Working with categorical data with R and the vcd package}, % pdfkeywords = {contingency tables, mosaic plots, sieve plots, categorical data, independence, conditional independence, R} % } %} % math stuff \newcommand*{\given}{\ensuremath{\, | \,}} \renewcommand*{\vec}[1]{\ensuremath{\bm{#1}}} \newcommand{\mat}[1]{\ensuremath{\bm{#1}}} \newcommand{\trans}{\ensuremath{^\mathsf{T}}} \newcommand{\diag}[1]{\ensuremath{\mathrm{diag} (#1)}} \def\binom#1#2{{#1 \choose #2}}% \newcommand{\implies}{ \ensuremath{\mapsto} } \newenvironment{equation*}{\displaymath}{\enddisplaymath}% \newcommand{\tabref}[1]{Table~\ref{#1}} \newcommand{\figref}[1]{Figure~\ref{#1}} \newcommand{\secref}[1]{Section~\ref{#1}} \newcommand{\loglin}{loglinear } %\usepackage{thumbpdf} % page dimensions \addtolength{\hoffset}{-1.5cm} \addtolength{\textwidth}{3cm} \addtolength{\voffset}{-1cm} \addtolength{\textheight}{2cm} % Vignette examples \newcommand*{\Example}{\fbox{\textbf{\emph{Example}}:} } % R stuff \newcommand{\var}[1]{\textit{\texttt{#1}}} \newcommand{\data}[1]{\texttt{#1}} \newcommand{\class}[1]{\textsf{"#1"}} %% \code without `-' ligatures \def\nohyphenation{\hyphenchar\font=-1 \aftergroup\restorehyphenation} \def\restorehyphenation{\hyphenchar\font=`-} {\catcode`\-=\active% \global\def\code{\bgroup% \catcode`\-=\active \let-\codedash% \Rd@code}} \def\codedash{-\discretionary{}{}{}} \def\Rd@code#1{\texttt{\nohyphenation#1}\egroup} \newcommand{\codefun}[1]{\code{#1()}} \let\proglang=\textsf \newcommand{\pkg}[1]{{\normalfont\fontseries{b}\selectfont #1}} \newcommand{\Rpackage}[1]{{\textsf{#1}}} %% almost as usual \author{Michael Friendly\\York University, Toronto} \title{Working with categorical data with \proglang{R} and the \pkg{vcd} and \pkg{vcdExtra} packages} \date{\footnotesize{Using \Rpackage{vcdExtra} version \Sexpr{packageDescription("vcdExtra")[["Version"]]} and \Rpackage{vcd} version \Sexpr{packageDescription("vcd")[["Version"]]}; Date: \Sexpr{Sys.Date()}}} %% for pretty printing and a nice hypersummary also set: %\Plainauthor{Michael Friendly} %% comma-separated %\Shorttitle{vcd tutorial} %% a short title (if necessary) %\Plaintitle{Tutorial: Working with categorical data with R and the vcd package} %\SweaveOpts{engine=R,eps=TRUE,height=6,width=7,results=hide,fig=FALSE,echo=TRUE} \SweaveOpts{engine=R,height=6,width=7,results=hide,fig=FALSE,echo=TRUE} \SweaveOpts{prefix.string=fig/vcd-tut,eps=FALSE} \SweaveOpts{keep.source=TRUE} %\SweaveOpts{concordance=TRUE} \setkeys{Gin}{width=0.7\textwidth} <>= set.seed(1071) #library(vcd) library(vcdExtra) library(ggplot2) #data(Titanic) data(HairEyeColor) data(PreSex) data(Arthritis) art <- xtabs(~Treatment + Improved, data = Arthritis) if(!file.exists("fig")) dir.create("fig") @ %% end of declarations %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \begin{document} \SweaveOpts{concordance=TRUE} \maketitle %% an abstract and keywords \begin{abstract} This tutorial describes the creation and manipulation of frequency and contingency tables from categorical variables, along with tests of independence, measures of association, and methods for graphically displaying results. The framework is provided by the \proglang{R} package \pkg{vcd}, but other packages are used to help with various tasks. The \pkg{vcdExtra} package extends the graphical and statistical methods provided by \pkg{vcd}. This package is now the main support package for the book \emph{Discrete Data Analysis with R: Visualizing and Modeling Techniques for Categorical and Count Data} \citep{FriendlyMeyer:2016:DDAR}. The web page for the book, \href{http://ddar.datavis.ca}{ddar.datavis.ca}, gives further details. \end{abstract} %\keywords{contingency tables, mosaic plots, sieve plots, %categorical data, independence, conditional independence, generalized linear models, %\proglang{R}} %\Plainkeywords{contingency tables, mosaic plots, % sieve plots, categorical data, independence, % conditional independence, generalized linear models, R} {\small % \sloppy % \begin{multicols}{2} \tableofcontents % \end{multicols} } \section[Introduction]{Introduction}\label{sec:intro} %% Note: If there is markup in \(sub)section, then it has to be escape as above. This tutorial, part of the \pkg{vcdExtra} package, describes how to work with categorical data in the context of fitting statistical models in \proglang{R} and visualizing the results using the \pkg{vcd} and \pkg{vcdExtra} packages. It focuses first on methods and tools for creating and manipulating \proglang{R} data objects which represent frequency and contingency tables involving categorical variables. Further sections describe some simple methods for calculating tests of independence and measures of association amomg categorial variables, and also methods for graphically displaying results. There is much more to the analysis of categorical data than is described here, where the emphasis is on cross-tabulated tables of frequencies (``contingency tables''), statistical tests, associated \loglin\ models, and visualization of \emph{how} variables are related. A more general treatment of graphical methods for categorical data is contained in the book, \emph{Discrete Data Analysis with R: Visualizing and Modeling Techniques for Categorical and Count Data} \citep{FriendlyMeyer:2016:DDAR}. An earlier book using SAS is \emph{Visualizing Categorical Data} \citep{vcd:Friendly:2000}, for which \pkg{vcd} is a partial \proglang{R} companion, covering topics not otherwise available in \proglang{R}. On the other hand, the implementation of graphical methods in \pkg{vcd} is more general in many respects than what I provided in \proglang{SAS}. Statistical models for categorical data in \proglang{R} have been extended considerably with the \pkg{gnm} package for generalized \emph{nonlinear} models. The \pkg{vcdExtra} package extends \pkg{vcd} methods to models fit using \codefun{glm} and \codefun{gnm}. A more complete theoretical description of these statistical methods is provided in Agresti's \citeyearpar{vcd:Agresti:2002,Agresti:2013} \emph{Categorical Data Analysis}. For this, see the \proglang{Splus/R} companion by Laura Thompson, \url{http://www.stat.ufl.edu/~aa/cda/Thompson_manual.pdf} and Agresti's support web page, \url{http://www.stat.ufl.edu/~aa/cda/cda.html}. \section[Creating frequency tables]{Creating and manipulating frequency tables}\label{sec:creating} \proglang{R} provides many methods for creating frequency and contingency tables. Several are described below. In the examples below, we use some real examples and some anonymous ones, where the variables \code{A}, \code{B}, and \code{C} represent categorical variables, and \code{X} represents an arbitrary \proglang{R} data object. The first thing you need to know is that categorical data can be represented in three different forms in \proglang{R}, and it is sometimes necessary to convert from one form to another, for carrying out statistical tests, fitting models or visualizing the results. Once a data object exists in \proglang{R}, you can examine its complete structure with the \codefun{str} function, or view the names of its components with the \codefun{names} function. \begin{description} \item[case form] a data frame containing individual observations, with one or more factors, used as the classifying variables. In case form, there may also be numeric covariates. The total number of observations is \code{nrow(X)}, and the number of variables is \code{ncol(X)}. \Example The \data{Arthritis} data is available in case form in the \pkg{vcd} package. There are two explanatory factors: \code{Treatment} and \code{Sex}. \code{Age} is a numeric covariate, and \code{Improved} is the response--- an ordered factor, with levels \code{\Sexpr{paste(levels(Arthritis$Improved),collapse=' < ')}}. Excluding \code{Age}, we would have a $2 \times 2 \times 3$ contingency table for \code{Treatment}, \code{Sex} and \code{Improved}. %\code{"None" < "Some" < "Marked"}. <>= names(Arthritis) # show the variables str(Arthritis) # show the structure head(Arthritis,5) # first 5 observations, same as Arthritis[1:5,] @ \item[frequency form] a data frame containing one or more factors, and a frequency variable, often called \code{Freq} or \code{count}. The total number of observations is \verb|sum(X$Freq)|, \code{sum(X[,"Freq"])} or some equivalent form. The number of cells in the table is \code{nrow(X)}. \Example For small frequency tables, it is often convenient to enter them in frequency form using \codefun{expand.grid} for the factors and \codefun{c} to list the counts in a vector. The example below, from \cite{vcd:Agresti:2002} gives results for the 1991 General Social Survey, with respondents classified by sex and party identification. <>= # Agresti (2002), table 3.11, p. 106 GSS <- data.frame( expand.grid(sex=c("female", "male"), party=c("dem", "indep", "rep")), count=c(279,165,73,47,225,191)) GSS names(GSS) str(GSS) sum(GSS$count) @ \item[table form] a matrix, array or table object, whose elements are the frequencies in an $n$-way table. The variable names (factors) and their levels are given by \code{dimnames(X)}. The total number of observations is \code{sum(X)}. The number of dimensions of the table is \code{length(dimnames(X))}, and the table sizes are given by \code{sapply(dimnames(X), length)}. \Example The \data{HairEyeColor} is stored in table form in \pkg{vcd}. <>= str(HairEyeColor) # show the structure sum(HairEyeColor) # number of cases sapply(dimnames(HairEyeColor), length) # table dimension sizes @ \Example Enter frequencies in a matrix, and assign \code{dimnames}, giving the variable names and category labels. Note that, by default, \codefun{matrix} uses the elements supplied by \emph{columns} in the result, unless you specify \code{byrow=TRUE}. <>= ## A 4 x 4 table Agresti (2002, Table 2.8, p. 57) Job Satisfaction JobSat <- matrix(c(1,2,1,0, 3,3,6,1, 10,10,14,9, 6,7,12,11), 4, 4) dimnames(JobSat) = list(income=c("< 15k", "15-25k", "25-40k", "> 40k"), satisfaction=c("VeryD", "LittleD", "ModerateS", "VeryS")) JobSat @ \data{JobSat} is a matrix, not an object of \code{class("table")}, and some functions are happier with tables than matrices. You can coerce it to a table with \codefun{as.table}, <>= JobSat <- as.table(JobSat) str(JobSat) @ \end{description} \subsection[Ordered factors]{Ordered factors and reordered tables}\label{sec:ordered-factors} In table form, the values of the table factors are ordered by their position in the table. Thus in the \data{JobSat} data, both \code{income} and \code{satisfaction} represent ordered factors, and the \emph{positions} of the values in the rows and columns reflects their ordered nature. Yet, for analysis, there are time when you need \emph{numeric} values for the levels of ordered factors in a table, e.g., to treat a factor as a quantitative variable. In such cases, you can simply re-assign the \code{dimnames} attribute of the table variables. For example, here, we assign numeric values to \code{income} as the middle of their ranges, and treat \code{satisfaction} as equally spaced with integer scores. <>= dimnames(JobSat)$income<-c(7.5,20,32.5,60) dimnames(JobSat)$satisfaction<-1:4 @ For the \data{HairEyeColor} data, hair color and eye color are ordered arbitrarily. For visualizing the data using mosaic plots and other methods described below, it turns out to be more useful to assure that both hair color and eye color are ordered from dark to light. Hair colors are actually ordered this way already, and it is easiest to re-order eye colors by indexing. Again \codefun{str} is your friend. <>= HairEyeColor <- HairEyeColor[, c(1,3,4,2), ] str(HairEyeColor) @ This is also the order for both hair color and eye color shown in the result of a correspondence analysis (\figref{fig:ca-haireye}) below. With data in case form or frequency form, when you have ordered factors represented with character values, you must ensure that they are treated as ordered in \proglang{R}.% \footnote{In \proglang{SAS}, many procedures offer the option \code{order = data | internal | formatted} to allow character values to be ordered according to (a) their order in the data set, (b) sorted internal value, or (c) sorted formatted representation provided by a \proglang{SAS} format. } Imagine that the \data{Arthritis} data was read from a text file. By default the \code{Improved} will be ordered alphabetically: \code{Marked}, \code{None}, \code{Some}--- not what we want. In this case, the function \codefun{ordered} (and others) can be useful. <>= Arthritis <- read.csv("arthritis.txt",header=TRUE) Arthritis$Improved <- ordered(Arthritis$Improved, levels=c("None", "Some", "Marked")) @ With this order of \code{Improved}, the response in this data, a mosaic display of \code{Treatment} and \code{Improved} (\figref{fig:arthritis})shows a clearly interpretable pattern. <>= mosaic(art, gp = shading_max, split_vertical = TRUE, main="Arthritis: [Treatment] [Improved]") @ %\setkeys{Gin}{width=0.7\textwidth} \begin{figure}[htb] \begin{center} %<>= %mosaic(art, gp = shading_max, split_vertical = TRUE, main="Arthritis: [Treatment] [Improved]") %@ \includegraphics[width=0.7\textwidth]{fig/vcd-tut-Arthritis} \caption{Mosaic plot for the \data{Arthritis} data, showing the marginal model of independence for Treatment and Improved. Age, a covariate, and Sex are ignored here.} \label{fig:arthritis} \end{center} \end{figure} Finally, there are situations where, particularly for display purposes, you want to re-order the \emph{dimensions} of an $n$-way table, or change the labels for the variables or levels. This is easy when the data are in table form: \codefun{aperm} permutes the dimensions, and assigning to \code{names} and \code{dimnames} changes variable names and level labels respectively. We will use the following version of \data{UCBAdmissions} in \secref{sec:mantel} below.% \footnote{ Changing \code{Admit} to \code{Admit?} might be useful for display purposes, but is dangerous--- because it is then difficult to use that variable name in a model formula. See \secref{sec:tips} for options \code{labeling\_args} and \code{set\_labels} to change variable and level names for displays in the \code{strucplot} framework. } <>= UCB <- aperm(UCBAdmissions, c(2, 1, 3)) dimnames(UCB)[[2]] <- c("Yes", "No") names(dimnames(UCB)) <- c("Sex", "Admit?", "Department") ftable(UCB) @ %There is one subtle ``gotcha'' here: \codefun{aperm} returns an object of class \class{"array"}, %whereas \data{UCBAdmissions} is of class \class{"table"}, so methods defined for \code{table} %objects will not work on the permuted array. %The solution is to reassign the \code{class} of the result of \codefun{aperm}. % %<>= %class(UCBAdmissions) %class(UCB) %str(as.data.frame(UCBAdmissions)) # OK %str(as.data.frame(UCB)) # wrong % %class(UCB) <- "table" %str(as.data.frame(UCB)) # now OK %@ % \subsection[structable()]{\codefun{structable}}\label{sec:structable} For 3-way and larger tables the \codefun{structable} function in \pkg{vcd} provides a convenient and flexible tabular display. The variables assigned to the rows and columns of a two-way display can be specified by a model formula. <>= structable(HairEyeColor) # show the table: default structable(Hair+Sex ~ Eye, HairEyeColor) # specify col ~ row variables @ It also returns an object of class \code{"structable"} which may be plotted with \codefun{mosaic} (not shown here). <>= HSE < - structable(Hair+Sex ~ Eye, HairEyeColor) # save structable object mosaic(HSE) # plot it @ \subsection[table() and friends]{\codefun{table} and friends}\label{sec:table} You can generate frequency tables from factor variables using the \codefun{table} function, tables of proportions using the \codefun{prop.table} function, and marginal frequencies using \codefun{margin.table}. <>= n=500 A <- factor(sample(c("a1","a2"), n, rep=TRUE)) B <- factor(sample(c("b1","b2"), n, rep=TRUE)) C <- factor(sample(c("c1","c2"), n, rep=TRUE)) mydata <- data.frame(A,B,C) @ <>= # 2-Way Frequency Table attach(mydata) mytable <- table(A,B) # A will be rows, B will be columns mytable # print table margin.table(mytable, 1) # A frequencies (summed over B) margin.table(mytable, 2) # B frequencies (summed over A) prop.table(mytable) # cell percentages prop.table(mytable, 1) # row percentages prop.table(mytable, 2) # column percentages @ \codefun{table} can also generate multidimensional tables based on 3 or more categorical variables. In this case, use the \codefun{ftable} or \codefun{structable} function to print the results more attractively. <>= # 3-Way Frequency Table mytable <- table(A, B, C) ftable(mytable) @ \codefun{table} ignores missing values by default. To include \code{NA} as a category in counts, include the table option \code{exclude=NULL} if the variable is a vector. If the variable is a factor you have to create a new factor using \code{newfactor <- factor(oldfactor, exclude=NULL)}. \subsection[xtabs()]{\codefun{xtabs}}\label{sec:xtabs} The \codefun{xtabs} function allows you to create crosstabulations of data using formula style input. This typically works with case-form data supplied in a data frame or a matrix. The result is a contingency table in array format, whose dimensions are determined by the terms on the right side of the formula. <>= # 3-Way Frequency Table mytable <- xtabs(~A+B+C, data=mydata) ftable(mytable) # print table summary(mytable) # chi-square test of indepedence @ If a variable is included on the left side of the formula, it is assumed to be a vector of frequencies (useful if the data have already been tabulated in frequency form). <>= (GSStab <- xtabs(count ~ sex + party, data=GSS)) summary(GSStab) @ \subsection[Collapsing over factors]{Collapsing over table factors: \codefun{aggregate}, \codefun{margin.table} and \codefun{apply}} It sometimes happens that we have a data set with more variables or factors than we want to analyse, or else, having done some initial analyses, we decide that certain factors are not important, and so should be excluded from graphic displays by collapsing (summing) over them. For example, mosaic plots and fourfold displays are often simpler to construct from versions of the data collapsed over the factors which are not shown in the plots. The appropriate tools to use again depend on the form in which the data are represented--- a case-form data frame, a frequency-form data frame (\codefun{aggregate}), or a table-form array or table object (\codefun{margin.table} or \codefun{apply}). When the data are in frequency form, and we want to produce another frequency data frame, \codefun{aggregate} is a handy tool, using the argument \code{FUN=sum} to sum the frequency variable over the factors \emph{not} mentioned in the formula. \Example The data frame \data{DaytonSurvey} in the \pkg{vcdExtra} package represents a $2^5$ table giving the frequencies of reported use (``ever used?'') of alcohol, cigarettes and marijuana in a sample of high school seniors, also classified by sex and race. <>= str(DaytonSurvey) head(DaytonSurvey) @ To focus on the associations among the substances, we want to collapse over sex and race. The right-hand side of the formula used in the call to \codefun{aggregate} gives the factors to be retained in the new frequency data frame, \code{Dayton.ACM.df}. <>= # data in frequency form # collapse over sex and race Dayton.ACM.df <- aggregate(Freq ~ cigarette+alcohol+marijuana, data=DaytonSurvey, FUN=sum) Dayton.ACM.df @ When the data are in table form, and we want to produce another table, \codefun{apply} with \code{FUN=sum} can be used in a similar way to sum the table over dimensions not mentioned in the \code{MARGIN} argument. \codefun{margin.table} is just a wrapper for \codefun{apply} using the \codefun{sum} function. \Example To illustrate, we first convert the \data{DaytonSurvey} to a 5-way table using \codefun{xtabs}, giving \code{Dayton.tab}. <>== # in table form Dayton.tab <- xtabs(Freq~cigarette+alcohol+marijuana+sex+race, data=DaytonSurvey) structable(cigarette+alcohol+marijuana ~ sex+race, data=Dayton.tab) @ Then, use \codefun{apply} on \code{Dayton.tab} to give the 3-way table \code{Dayton.ACM.tab} summed over sex and race. The elements in this new table are the column sums for \code{Dayton.tab} shown by \codefun{structable} just above. <>== # collapse over sex and race Dayton.ACM.tab <- apply(Dayton.tab, MARGIN=1:3, FUN=sum) Dayton.ACM.tab <- margin.table(Dayton.tab, 1:3) # same result structable(cigarette+alcohol ~ marijuana, data=Dayton.ACM.tab) @ Many of these operations can be performed using the \verb|**ply()| functions in the \pkg{plyr} package. For example, with the data in a frequency form data frame, use \codefun{ddply} to collapse over unmentioned factors, and \codefun{plyr::summarise}% \footnote{ Ugh. This \pkg{plyr} function clashes with a function of the same name in \pkg{vcdExtra}. In this document I will use the explicit double-colon notation to keep them separate. } as the function to be applied to each piece. <>== Dayton.ACM.df <- ddply(DaytonSurvey, .(cigarette, alcohol, marijuana), plyr::summarise, Freq=sum(Freq)) @ \subsection[Collapsing levels]{Collapsing table levels: \codefun{collapse.table}} A related problem arises when we have a table or array and for some purpose we want to reduce the number of levels of some factors by summing subsets of the frequencies. For example, we may have initially coded Age in 10-year intervals, and decide that, either for analysis or display purposes, we want to reduce Age to 20-year intervals. The \codefun{collapse.table} function in \pkg{vcdExtra} was designed for this purpose. \Example Create a 3-way table, and collapse Age from 10-year to 20-year intervals. First, we generate a $2 \times 6 \times 3$ table of random counts from a Poisson distribution with mean of 100. <>= # create some sample data in frequency form sex <- c("Male", "Female") age <- c("10-19", "20-29", "30-39", "40-49", "50-59", "60-69") education <- c("low", 'med', 'high') data <- expand.grid(sex=sex, age=age, education=education) counts <- rpois(36, 100) # random Possion cell frequencies data <- cbind(data, counts) # make it into a 3-way table t1 <- xtabs(counts ~ sex + age + education, data=data) structable(t1) @ Now collapse \code{age} to 20-year intervals, and \code{education} to 2 levels. In the arguments, levels of \code{age} and \code{education} given the same label are summed in the resulting smaller table. <>= # collapse age to 3 levels, education to 2 levels t2 <- collapse.table(t1, age=c("10-29", "10-29", "30-49", "30-49", "50-69", "50-69"), education=c(">= as.data.frame(GSStab) @ \Example Convert the \code{Arthritis} data in case form to a 3-way table of \code{Treatment} $\times$ \code{Sex} $\times$ \code{Improved}. Note the use of \codefun{with} to avoid having to use \code{Arthritis\$Treatment} etc. within the call to \codefun{table}.% \footnote{ \codefun{table} does not allow a \code{data} argument to provide an environment in which the table variables are to be found. In the examples in \secref{sec:table} I used \code{attach(mydata)} for this purpose, but \codefun{attach} leaves the variables in the global environment, while \codefun{with} just evaluates the \codefun{table} expression in a temporary environment of the data. } <>= Art.tab <-with(Arthritis, table(Treatment, Sex, Improved)) str(Art.tab) ftable(Art.tab) @ There may also be times that you will need an equivalent case form \code{data.frame} with factors representing the table variables rather than the frequency table. For example, the \codefun{mca} function in package \pkg{MASS} only operates on data in this format. Marc Schwartz provided code for \codefun{expand.dft} on the Rhelp mailing list for converting a table back into a case form \code{data.frame}. This function is included in \pkg{vcdExtra}. \Example Convert the \data{Arthritis} data in table form (\code{Art.tab}) back to a \code{data.frame} in case form, with factors \code{Treatment}, \code{Sex} and \code{Improved}. <>= Art.df <- expand.dft(Art.tab) str(Art.df) @ \subsection{A complex example}\label{sec:complex} If you've followed so far, you're ready for a more complicated example. The data file, \code{tv.dat} represents a 4-way table of size $5 \times 11 \times 5 \times 3$ where the table variables (unnamed in the file) are read as \code{V1} -- \code{V4}, and the cell frequency is read as \code{V5}. The file, stored in the \code{doc/extdata} directory of \pkg{vcdExtra}, can be read as follows: <>= tv.data<-read.table(system.file("doc","extdata","tv.dat",package="vcdExtra")) head(tv.data,5) @ For a local file, just use \codefun{read.table} in this form: <>= tv.data<-read.table("C:/R/data/tv.dat") @ The data \code{tv.dat} came from the initial implementation of mosaic displays in \proglang{R} by Jay Emerson. In turn, they came from the initial development of mosaic displays \citep{vcd:Hartigan+Kleiner:1984} that illustrated the method with data on a large sample of TV viewers whose behavior had been recorded for the Neilson ratings. This data set contains sample television audience data from Neilsen Media Research for the week starting November 6, 1995. \begin{flushleft} The table variables are:\\ ~~~\code{V1}-- values 1:5 correspond to the days Monday--Friday;\\ ~~~\code{V2}-- values 1:11 correspond to the quarter hour times 8:00PM through 10:30PM;\\ ~~~\code{V3}-- values 1:5 correspond to ABC, CBS, NBC, Fox, and non-network choices;\\ ~~~\code{V4}-- values 1:3 correspond to transition states: turn the television Off, Switch channels, or Persist in viewing the current channel. \end{flushleft} We are interested just the cell frequencies, and rely on the facts that the (a) the table is complete--- there are no missing cells, so \code{nrow(tv.data)}=\Sexpr{nrow(tv.data)}; (b) the observations are ordered so that \code{V1} varies most rapidly and \code{V4} most slowly. From this, we can just extract the frequency column and reshape it into an array. <>= TV <- array(tv.data[,5], dim=c(5,11,5,3)) dimnames(TV) <- list(c("Monday","Tuesday","Wednesday","Thursday","Friday"), c("8:00","8:15","8:30","8:45","9:00","9:15","9:30", "9:45","10:00","10:15","10:30"), c("ABC","CBS","NBC","Fox","Other"), c("Off","Switch","Persist")) names(dimnames(TV))<-c("Day", "Time", "Network", "State") @ More generally (even if there are missing cells), we can use \codefun{xtabs} (or \codefun{plyr::daply}) to do the cross-tabulation, using \code{V5} as the frequency variable. Here's how to do this same operation with \codefun{xtabs}: <>= TV <- xtabs(V5 ~ ., data=tv.data) dimnames(TV) <- list(Day=c("Monday","Tuesday","Wednesday","Thursday","Friday"), Time=c("8:00","8:15","8:30","8:45","9:00","9:15","9:30", "9:45","10:00","10:15","10:30"), Network=c("ABC","CBS","NBC","Fox","Other"), State=c("Off","Switch","Persist")) @ But this 4-way table is too large and awkward to work with. Among the networks, Fox and Other occur infrequently. We can also cut it down to a 3-way table by considering only viewers who persist with the current station.% \footnote{This relies on the fact that that indexing an array drops dimensions of length 1 by default, using the argument \code{drop=TRUE}; the result is coerced to the lowest possible dimension. } <>= TV <- TV[,,1:3,] # keep only ABC, CBS, NBC TV <- TV[,,,3] # keep only Persist -- now a 3 way table structable(TV) @ Finally, for some purposes, we might want to collapse the 11 times into a smaller number. Here, we use \codefun{as.data.frame.table} to convert the table back to a data frame, \codefun{levels} to re-assign the values of \code{Time}, and finally, \codefun{xtabs} to give a new, collapsed frequency table. <>= TV.df <- as.data.frame.table(TV) levels(TV.df$Time) <- c(rep("8:00-8:59",4),rep("9:00-9:59",4), rep("10:00-10:44",3)) TV2 <- xtabs(Freq ~ Day + Time + Network, TV.df) structable(Day ~ Time+Network,TV2) @ Whew! See \figref{fig:TV-mosaic} for a mosaic plot of the \code{TV2} data. \section{Tests of Independence} \subsection{CrossTable} OK, now we're ready to do some analyses. For tabular displays, the \codefun{CrossTable} function in the \pkg{gmodels} package produces cross-tabulations modeled after \code{PROC FREQ} in \proglang{SAS} or \code{CROSSTABS} in \proglang{SPSS}. It has a wealth of options for the quantities that can be shown in each cell. <>= # 2-Way Cross Tabulation library(gmodels) CrossTable(GSStab,prop.t=FALSE,prop.r=FALSE,prop.c=FALSE) @ There are options to report percentages (row, column, cell), specify decimal places, produce Chi-square, Fisher, and McNemar tests of independence, report expected and residual values (pearson, standardized, adjusted standardized), include missing values as valid, annotate with row and column titles, and format as \proglang{SAS} or \proglang{SPSS} style output! See \code{help(CrossTable)} for details. \subsection{Chi-square test} For 2-way tables you can use \codefun{chisq.test} to test independence of the row and column variable. By default, the $p$-value is calculated from the asymptotic chi-squared distribution of the test statistic. Optionally, the $p$-value can be derived via Monte Carlo simulation. <>= (HairEye <- margin.table(HairEyeColor, c(1, 2))) chisq.test(HairEye) @ \subsection{Fisher Exact Test}\label{sec:Fisher} \code{fisher.test(X)} provides an exact test of independence. \code{X} must be a two-way contingency table in table form. Another form, \code{fisher.test(X, Y)} takes two categorical vectors of the same length. For tables larger than $2 \times 2$ the method can be computationally intensive (or can fail) if the frequencies are not small. <>= fisher.test(GSStab) @ But this does not work because \data{HairEye} data has $n$=592 total frequency. An exact test is unnecessary in this case. <>= fisher.test(HairEye) @ %# <>= %# #cat(try(fisher.test(HairEye))) %# @ \begin{Soutput} Error in fisher.test(HairEye) : FEXACT error 6. LDKEY is too small for this problem. Try increasing the size of the workspace. \end{Soutput} \subsection[Mantel-Haenszel test]{Mantel-Haenszel test and conditional association}\label{sec:mantel} Use the \code{mantelhaen.test(X)} function to perform a Cochran-Mantel-Haenszel $\chi^2$ chi test of the null hypothesis that two nominal variables are \emph{conditionally independent}, $A \perp B \given C$, in each stratum, assuming that there is no three-way interaction. \code{X} is a 3 dimensional contingency table, where the last dimension refers to the strata. The \data{UCBAdmissions} serves as an example of a $2 \times 2 \times 6$ table, with \code{Dept} as the stratifying variable. <>= ## UC Berkeley Student Admissions mantelhaen.test(UCBAdmissions) @ The results show no evidence for association between admission and gender when adjusted for department. However, we can easily see that the assumption of equal association across the strata (no 3-way association) is probably violated. For $2 \times 2 \times k$ tables, this can be examimed from the odds ratios for each $2 \times 2$ table (\codefun{oddsratio}), and tested by using \verb|woolf_test()| in \pkg{vcd}. %<>= %oddsRatio <- function(x) (x[1,1]*x[2,2])/(x[1,2]*x[2,1]) %apply(UCBAdmissions, 3, oddsRatio) % %woolf_test(UCBAdmissions) %@ <>= oddsratio(UCBAdmissions, log=FALSE) lor <- oddsratio(UCBAdmissions) # capture log odds ratios summary(lor) woolf_test(UCBAdmissions) @ We can visualize the odds ratios of Admission for each department with fourfold displays using \codefun{fourfold}. The cell frequencies $n_{ij}$ of each $2 \times 2$ table are shown as a quarter circle whose radius is proportional to $\sqrt{n_{ij}}$, so that its area is proportional to the cell frequency. Confidence rings for the odds ratio allow a visual test of the null of no association; the rings for adjacent quadrants overlap \emph{iff} the observed counts are consistent with the null hypothesis. In the extended version (the default), brighter colors are used where the odds ratio is significantly different from 1. The following lines produce \figref{fig:fourfold1}.% \footnote{The color values \code{col[3:4]} were modified from their default values to show a greater contrast between significant and insignifcant associations here.} <>= col <- c("#99CCFF", "#6699CC", "#F9AFAF", "#6666A0", "#FF0000", "#000080") fourfold(UCB,mfrow=c(2,3), color=col) @ %\setkeys{Gin}{width=0.8\textwidth} \begin{figure}[htb] \begin{center} %<>= %col <- c("#99CCFF", "#6699CC", "#F9AFAF", "#6666A0", "#FF0000", "#000080") %fourfold(UCB,mfrow=c(2,3), color=col) %@ \includegraphics[width=0.8\textwidth,trim=80 50 80 50]{fig/vcd-tut-fourfold1} \caption{Fourfold display for the \data{UCBAdmissions} data. Where the odds ratio differs significantly from 1.0, the confidence bands do not overlap, and the circle quadrants are shaded more intensely.} \label{fig:fourfold1} \end{center} \end{figure} Another \pkg{vcd} function, \codefun{cotabplot}, provides a more general approach to visualizing conditional associations in contingency tables, similar to trellis-like plots produced by \codefun{coplot} and lattice graphics. The \code{panel} argument supplies a function used to render each conditional subtable. The following gives a display (not shown) similar to \figref{fig:fourfold1}. <>= cotabplot(UCB, panel = cotab_fourfold) @ When we want to view the conditional probabilities of a response variable (e.g., \code{Admit}) in relation to several factors, an alternative visualization is a \codefun{doubledecker} plot. This plot is a specialized version of a mosaic plot, which highlights the levels of a response variable (plotted vertically) in relation to the factors (shown horizontally). The following call produces \figref{fig:doubledecker}, where we use indexing on the first factor (\code{Admit}) to make \code{Admitted} the highlighted level. In this plot, the association between \code{Admit} and \code{Gender} is shown where the heights of the highlighted conditional probabilities do not align. The excess of females admitted in Dept A stands out here. <>= doubledecker(Admit ~ Dept + Gender, data=UCBAdmissions[2:1,,]) @ \begin{figure}[htb] \begin{center} \includegraphics[width=0.9\textwidth]{fig/vcd-tut-doubledecker} \caption{Doubledecker display for the \data{UCBAdmissions} data. The heights of the highlighted bars show the conditional probabilities of \texttt{Admit}, given \texttt{Dept} and \texttt{Gender}.} \label{fig:doubledecker} \end{center} \end{figure} Finally, the there is a \codefun{plot} method for \code{oddsratio} objects. By default, it shows the 95\% confidence interval for the log odds ratio. \figref{fig:oddsratio} is produced by: <>= plot(lor, xlab="Department", ylab="Log Odds Ratio (Admit | Gender)") @ \setkeys{Gin}{width=0.5\textwidth} \begin{figure}[htb] \begin{center} <>= plot(lor, xlab="Department", ylab="Log Odds Ratio (Admit | Gender)") @ \caption{Log odds ratio plot for the \data{UCBAdmissions} data.} \label{fig:oddsratio} \end{center} \end{figure} \subsection[CMH tests: ordinal factors]{Cochran-Mantel-Haenszel tests for ordinal factors}\label{sec:CMH} The standard $\chi^2$ tests for association in a two-way table treat both table factors as nominal (unordered) categories. When one or both factors of a two-way table are quantitative or ordinal, more powerful tests of association may be obtaianed by taking ordinality into account, using row and or column scores to test for linear trends or differences in row or column means. More general versions of the CMH tests (Landis etal., 1978) are provided by assigning numeric scores to the row and/or column variables. For example, with two ordinal factors (assumed to be equally spaced), assigning integer scores, \code{1:R} and \code{1:C} tests the linear $\times$ linear component of association. This is statistically equivalent to the Pearson correlation between the integer-scored table variables, with $\chi^2 = (n-1) r^2$, with only 1 $df$ rather than $(R-1)\times(C-1)$ for the test of general association. When only one table variable is ordinal, these general CMH tests are analogous to an ANOVA, testing whether the row mean scores or column mean scores are equal, again consuming fewer $df$ than the test of general association. The \codefun{CMHtest} function in \pkg{vcdExtra} now calculates these various CMH tests for two possibly ordered factors, optionally stratified other factor(s). \Example Recall the $4 \times 4$ table, \code{JobSat} introduced in \secref{sec:creating}, <>= JobSat @ Treating the \code{satisfaction} levels as equally spaced, but using midpoints of the \code{income} categories as row scores gives the following results: <>= CMHtest(JobSat, rscores=c(7.5,20,32.5,60)) @ Note that with the relatively small cell frequencies, the test for general give no evidence for association. However, the the \code{cor} test for linear x linear association on 1 df is nearly significant. The \pkg{coin} contains the functions \verb|cmh_test()| and \verb|lbl_test()| for CMH tests of general association and linear x linear association respectively. \subsection{Measures of Association} There are a variety of statistical measures of \emph{strength} of association for contingency tables--- similar in spirit to $r$ or $r^2$ for continuous variables. With a large sample size, even a small degree of association can show a significant $\chi^2$, as in the example below for the \data{GSS} data. The \codefun{assocstats} function in \pkg{vcd} calculates the $\phi$ contingency coefficient, and Cramer's V for an $r \times c$ table. The input must be in table form, a two-way $r \times c$ table. It won't work with \data{GSS} in frequency form, but by now you should know how to convert. <>= assocstats(GSStab) @ For tables with ordinal variables, like \data{JobSat}, some people prefer the Goodman-Kruskal $\gamma$ statistic (\citet[\S 2.4.3]{vcd:Agresti:2002}) based on a comparison of concordant and discordant pairs of observations in the case-form equivalent of a two-way table. <>= GKgamma(JobSat) @ A web article by Richard Darlington, \url{http://www.psych.cornell.edu/Darlington/crosstab/TABLE0.HTM} gives further description of these and other measures of association. \subsection{Measures of Agreement} The \codefun{Kappa} function in the \pkg{vcd} package calculates Cohen's $\kappa$ and weighted $\kappa$ for a square two-way table with the same row and column categories \citep{Cohen:60}.% \footnote{ Don't confuse this with \codefun{kappa} in base \proglang{R} that computes something entirely different (the condition number of a matrix). } Normal-theory $z$-tests are obtained by dividing $\kappa$ by its asymptotic standard error (ASE). A \codefun{confint} method for \code{Kappa} objects provides confidence intervals. <>= (K <- Kappa(SexualFun)) confint(K) @ A visualization of agreement, both unweighted and weighted for degree of departure from exact agreement is provided by the \codefun{agreementplot} function. \figref{fig:agreesex} shows the agreementplot for the \data{SexualFun} data, produced as shown below. The Bangdiwala measures represent the proportion of the shaded areas of the diagonal rectangles, using weights $w_1$ for exact agreement, and $w_2$ for partial agreement one step from the main diagonal. <>= agree <- agreementplot(SexualFun, main="Is sex fun?") unlist(agree) @ %\setkeys{Gin}{width=0.5\textwidth} \begin{figure}[htb] \begin{center} %<>= %agree <- agreementplot(SexualFun, main="Is sex fun?") %agree %@ \includegraphics[width=0.4\textwidth,trim=50 25 50 25]{fig/vcd-tut-agreesex} \caption{Agreement plot for the \data{SexualFun} data.} \label{fig:agreesex} \end{center} \end{figure} In other examples, the agreement plot can help to show \emph{sources} of disagreement. For example, when the shaded boxes are above or below the diagonal (red) line, a lack of exact agreement can be attributed in part to different frequency of use of categories by the two raters-- lack of \emph{marginal homogeneity}. \subsection{Correspondence analysis} Use the \pkg{ca} package for correspondence analysis for visually exploring relationships between rows and columns in contingency tables. For an $r \times c$ table, the method provides a breakdown of the Pearson $\chi^2$ for association in up to $M = \min(r-1, c-1)$ dimensions, and finds scores for the row ($x_{im}$) and column ($y_{jm}$) categories such that the observations have the maximum possible correlations.% \footnote{ Related methods are the non-parametric CMH tests using assumed row/column scores (\secref{sec:CMH}), the analogous \codefun{glm} model-based methods (\secref{sec:CMH}), and the more general RC models which can be fit using \codefun{gnm}. Correspondence analysis differs in that it is a primarily descriptive/exploratory method (no significance tests), but is directly tied to informative graphic displays of the row/column categories. } Here, we carry out a simple correspondence analysis of the \data{HairEye} data. The printed results show that nearly 99\% of the association between hair color and eye color can be accounted for in 2 dimensions, of which the first dimension accounts for 90\%. <>= library(ca) ca(HairEye) @ The resulting \code{ca} object can be plotted just by running the \codefun{plot} method on the \code{ca} object, giving the result in \figref{fig:ca-haireye}. \codefun{plot.ca} does not allow labels for dimensions; these can be added with \codefun{title}. It can be seen that most of the association is accounted for by the ordering of both hair color and eye color along Dimension 1, a dark to light dimension. <>= plot(ca(HairEye), main="Hair Color and Eye Color") title(xlab="Dim 1 (89.4%)", ylab="Dim 2 (9.5%)") @ \setkeys{Gin}{width=0.7\textwidth} \begin{figure}[htb] \begin{center} <>= plot(ca(HairEye), main="Hair Color and Eye Color") title(xlab="Dim 1 (89.4%)", ylab="Dim 2 (9.5%)") @ \caption{Correspondence analysis plot for the \data{HairEye} data.} \label{fig:ca-haireye} \end{center} \end{figure} \section{Loglinear Models}\label{sec:loglin} You can use the \codefun{loglm} function in the \pkg{MASS} package to fit log-linear models. Equivalent models can also be fit (from a different perspective) as generalized linear models with the \codefun{glm} function using the \code{family='poisson'} argument, and the \pkg{gnm} package provides a wider range of generalized \emph{nonlinear} models, particularly for testing structured associations. The visualization methods for these models were originally developed for models fit using \codefun{loglm}, so this approach is emphasized here. Some extensions of these methods for models fit using \codefun{glm} and \codefun{gnm} are contained in the \pkg{vcdExtra} package and illustrated in \secref{sec:glm}. Assume we have a 3-way contingency table based on variables A, B, and C. The possible different forms of \loglin\ models for a 3-way table are shown in \tabref{tab:loglin-3way}. The \textbf{Model formula} column shows how to express each model for \codefun{loglm} in \proglang{R}.% \footnote{ For \codefun{glm}, or \codefun{gnm}, with the data in the form of a frequency data.frame, the same model is specified in the form \code{glm(Freq} $\sim$ \code{..., family="poisson")}, where \texttt{Freq} is the name of the cell frequency variable and \texttt{...} specifies the \textbf{Model formula}. } In the \textbf{Interpretation} column, the symbol ``$\perp$'' is to be read as ``is independent of,'' and ``$\given$'' means ``conditional on,'' or ``adjusting for,'' or just ``given''. \begin{table}[htb] \caption{Log-linear Models for Three-Way Tables}\label{tab:loglin-3way} \begin{center} \begin{tabular}{llll} \hline \textbf{Model} & \textbf{Model formula} & \textbf{Symbol}& \textbf{Interpretation} \\ \hline\hline Mutual independence & \verb|~A + B + C| & $[A][B][C]$ & $A \perp B \perp C$ \\ Joint independence & \verb|~A*B + C| & $[AB][C]$ & $(A \: B) \perp C$ \\ Conditional independence & \verb|~(A+B)*C| & $[AC][BC]$ & $(A \perp B) \given C$ \\ All two-way associations & \verb|~A*B + A*C + B*C| & $[AB][AC][BC]$ & homogeneous association \\ Saturated model & \verb|~A*B*C| & $[ABC]$ & 3-way association \\ \hline \end{tabular} \end{center} \end{table} For example, the formula \verb|~A + B + C| specifies the model of \emph{mutual independence} with no associations among the three factors. In standard notation for the expected frequencies $m_{ijk}$, this corresponds to \begin{equation*} \log ( m_{ijk} ) = \mu + \lambda_i^A + \lambda_j^B + \lambda_k^C \equiv \texttt{A + B + C} \end{equation*} The parameters $\lambda_i^A , \lambda_j^B$ and $\lambda_k^C$ pertain to the differences among the one-way marginal frequencies for the factors A, B and C. Similarly, the model of \emph{joint independence}, $(A \: B) \perp C$, allows an association between A and B, but specifies that C is independent of both of these and their combinations, \begin{equation*} \log ( m_{ijk} ) = \mu + \lambda_i^A + \lambda_j^B + \lambda_k^C + \lambda_{ij}^{AB} \equiv \texttt{A * B + C} \end{equation*} where the parameters $\lambda_{ij}^{AB}$ pertain to the overall association between A and B (collapsing over C). In the literature or text books, you will often find these models expressed in shorthand symbolic notation, using brackets, \texttt{[ ]} to enclose the \emph{high-order terms} in the model. Thus, the joint independence model can be denoted \texttt{[AB][C]}, as shown in the \textbf{Symbol} column in \tabref{tab:loglin-3way}. Models of \emph{conditional independence} allow (and fit) two of the three possible two-way associations. There are three such models, depending on which variable is conditioned upon. For a given conditional independence model, e.g., \texttt{[AB][AC]}, the given variable is the one common to all terms, so this example has the interpretation $(B \perp C) \given A$. \subsection[Fitting with loglm()]{Fitting with \codefun{loglm}}\label{sec:loglm} For example, we can fit the model of mutual independence among hair color, eye color and sex in \data{HairEyeColor} as <>= library(MASS) ## Independence model of hair and eye color and sex. hec.1 <- loglm(~Hair+Eye+Sex, data=HairEyeColor) hec.1 @ Similarly, the models of conditional independence and joint independence are specified as <>= ## Conditional independence hec.2 <- loglm(~(Hair + Eye) * Sex, data=HairEyeColor) hec.2 @ <>= ## Joint independence model. hec.3 <- loglm(~Hair*Eye + Sex, data=HairEyeColor) hec.3 @ Note that printing the model gives a brief summary of the goodness of fit. A set of models can be compared using the \codefun{anova} function. <>= anova(hec.1, hec.2, hec.3) @ %Martin Theus and Stephan Lauer have written an excellent article on Visualizing %Loglinear Models, using mosaic plots. There is also great tutorial example by %Kevin Quinn on analyzing loglinear models via glm. \subsection[Fitting with glm() and gnm()]{Fitting with \codefun{glm} and \codefun{gnm}}\label{sec:glm} The \codefun{glm} approach, and extensions of this in the \pkg{gnm} package allows a much wider class of models for frequency data to be fit than can be handled by \codefun{loglm}. Of particular importance are models for ordinal factors and for square tables, where we can test more structured hypotheses about the patterns of association than are provided in the tests of general assosiation under \codefun{loglm}. These are similar in spirit to the non-parametric CMH tests described in \secref{sec:CMH}. \Example The data \code{Mental} in the \pkg{vcdExtra} package gives a two-way table in frequency form classifying young people by their mental health status and parents' socioeconomic status (SES), where both of these variables are ordered factors. <>= str(Mental) xtabs(Freq ~ mental+ses, data=Mental) # display the frequency table @ Simple ways of handling ordinal variables involve assigning scores to the table categories, and the simplest cases are to use integer scores, either for the row variable (``column effects'' model), the column variable (``row effects'' model), or both (``uniform association'' model). <>= indep <- glm(Freq ~ mental + ses, family = poisson, data = Mental) # independence model @ To fit more parsimonious models than general association, we can define numeric scores for the row and column categories <>= # Use integer scores for rows/cols Cscore <- as.numeric(Mental$ses) Rscore <- as.numeric(Mental$mental) @ Then, the row effects model, the column effects model, and the uniform association model can be fit as follows: <>= # column effects model (ses) coleff <- glm(Freq ~ mental + ses + Rscore:ses, family = poisson, data = Mental) # row effects model (mental) roweff <- glm(Freq ~ mental + ses + mental:Cscore, family = poisson, data = Mental) # linear x linear association linlin <- glm(Freq ~ mental + ses + Rscore:Cscore, family = poisson, data = Mental) @ The \codefun{Summarize} in \pkg{vcdExtra} provides a nice, compact summary of the fit statistics for a set of models, collected into a \class{glmlist} object. Smaller is better for AIC and BIC. <>= # compare models using AIC, BIC, etc vcdExtra::LRstats(glmlist(indep, roweff, coleff, linlin)) @ For specific model comparisons, we can also carry out tests of \emph{nested} models with \codefun{anova} when those models are listed from smallest to largest. Here, there are two separate paths from the most restrictive (independence) model through the model of uniform association, to those that allow only one of row effects or column effects. <>= anova(indep, linlin, coleff, test="Chisq") anova(indep, linlin, roweff, test="Chisq") @ The model of linear by linear association seems best on all accounts. For comparison, one might try the CMH tests on these data: <>= CMHtest(xtabs(Freq~ses+mental, data=Mental)) @ \subsection{Non-linear terms} The strength of the \pkg{gnm} package is that it handles a wide variety of models that handle non-linear terms, where the parameters enter the model beyond a simple linear function. The simplest example is the Goodman RC(1) model, which allows a multiplicative term to account for the association of the table variables. In the notation of generalized linear models with a log link, this can be expressed as \begin{equation*} \log \mu_{ij} = \alpha_i + \beta_j + \gamma_{i} \delta_{j} \end{equation*} where the row-multiplicative effect parameters $\gamma_i$ and corresponding column parameters $\delta_j$ are estimated from the data.% \footnote{ This is similar in spirit to a correspondence analysis with a single dimension, but as a statistical model. } Similarly, the RC(2) model adds two multiplicative terms to the independence model, \begin{equation*} \log \mu_{ij} = \alpha_i + \beta_j + \gamma_{i1} \delta_{j1} + \gamma_{i2} \delta_{j2} \end{equation*} In the \pkg{gnm} package, these models may be fit using the \codefun{Mult} to specify the multiplicative term, and \codefun{instances} to specify several such terms. \Example For the \code{Mental} data, we fit the RC(1) and RC(2) models, and compare these with the independence model. <>= RC1 <- gnm(Freq ~ mental + ses + Mult(mental,ses), data=Mental, family=poisson, , verbose=FALSE) RC2 <- gnm(Freq ~ mental+ses + instances(Mult(mental,ses),2), data=Mental, family=poisson, verbose=FALSE) anova(indep, RC1, RC2, test="Chisq") @ \section{Mosaic plots}\label{sec:mosaic} Mosaic plots provide an ideal method both for visualizing contingency tables and for visualizing the fit--- or more importantly--- lack of fit of a \loglin\ model. For a two-way table, \codefun{mosaic} fits a model of independence, $[A][B]$ or \verb|~A+B| as an \proglang{R} formula. For $n$-way tables, \codefun{mosaic} can fit any \loglin\ model, and can also be used to plot a model fit with \codefun{loglm}. See \citet{vcd:Friendly:1994,vcd:Friendly:1999} for the statistical ideas behind these uses of mosaic displays in connection with \loglin\ models. The essential idea is to recursively sub-divide a unit square into rectangular ``tiles'' for the cells of the table, such that the are area of each tile is proportional to the cell frequency. For a given \loglin\ model, the tiles can then be shaded in various ways to reflect the residuals (lack of fit) for a given model. The pattern of residuals can then be used to suggest a better model or understand \emph{where} a given model fits or does not fit. \codefun{mosaic} provides a wide range of options for the directions of splitting, the specification of shading, labeling, spacing, legend and many other details. It is actually implemented as a special case of a more general class of displays for $n$-way tables called \code{strucplot}, including sieve diagrams, association plots, double-decker plots as well as mosaic plots. For details, see \code{help(strucplot)} and the ``See also'' links, and also \citet{vcd:Meyer+Zeileis+Hornik:2006b}, which is available as an \proglang{R} vignette via \code{vignette("strucplot", package="vcd")}. \figref{fig:arthritis}, showing the association between \code{Treatment} and \code{Improved} was produced with the following call to \codefun{mosaic}. <>= mosaic(art, gp = shading_max, split_vertical = TRUE, main="Arthritis: [Treatment] [Improved]") @ Note that the residuals for the independence model were not large (as shown in the legend), yet the association between \code{Treatment} and \code{Improved} is highly significant. <>= summary(art) @ In contrast, one of the other shading schemes, from \citet{vcd:Friendly:1994} (use: \verb|gp = shading_Friendly|), uses fixed cutoffs of $\pm 2, \pm 4$, to shade cells which are \emph{individually} significant at approximately $\alpha = 0.05$ and $\alpha = 0.001$ levels, respectively. The right panel below uses \verb|gp = shading_Friendly|. \setkeys{Gin}{width=0.5\textwidth} <>= mosaic(art, gp = shading_max, split_vertical = TRUE, main="Arthritis: gp = shading_max") @ <>= mosaic(art, gp = shading_Friendly, split_vertical = TRUE, main="Arthritis: gp = shading_Friendly") @ \subsection[Mosaics for loglinear models]{Mosaics for \loglin\ models}\label{sec:mosaic-llm} When you have fit a \loglin\ model using \codefun{loglm}, and saved the result (as a \code{loglm} object) the simplest way to display the results is to use the \codefun{plot} method for the \code{loglm} object. Calling \code{mosaic(loglm.object)} has the same result. In \secref{sec:loglm} above, we fit several different models to the \data{HairEyeColor} data. We can produce mosaic displays of each just by plotting them: <>= # mosaic plots, using plot.loglm() method plot(hec.1, main="model: [Hair][Eye][Sex]") plot(hec.2, main="model: [HairSex][EyeSex]") plot(hec.3, main="model: [HairEye][Sex]") @ \setkeys{Gin}{width=0.32\textwidth} <>= plot(hec.1, main="model: [Hair][Eye][Sex]") @ <>= plot(hec.2, main="model: [HairSex][EyeSex]") @ <>= plot(hec.3, main="model: [HairSex][EyeSex]") @ Alternatively, you can supply the model formula to \codefun{mosaic} with the \code{expected} argument. This is passed to \codefun{loglm}, which fits the model, and returns residuals used for shading in the plot. For example, here we examine the \data{TV2} constructed in \secref{sec:complex} above. The goal is to see how Network choice depends on (varies with) Day and Time. To do this: \begin{itemize} \item We fit a model of joint independence of \code{Network} on the combinations of \code{Day} and \code{Time}, with the model formula \verb|~Day:Time + Network|. \item To make the display more easily read, we place \code{Day} and \code{Time} on the vertical axis and \code{Network} on the horizontal, \item The \code{Time} values overlap on the right vertical axis, so we use \codefun{level} to abbreviate them. \codefun{mosaic} also supports a more sophisticated set of labeling functions. Instead of changing the data table, we could have used \verb|labeling_args = list(abbreviate = c(Time = 2))| for a similar effect. \end{itemize} The following call to \codefun{mosaic} produces \figref{fig:TV-mosaic}: <>= dimnames(TV2)$Time <- c("8", "9", "10") # re-level for mosaic display mosaic(~ Day + Network + Time, data=TV2, expected=~Day:Time + Network, legend=FALSE, gp=shading_Friendly) @ \setkeys{Gin}{width=0.75\textwidth} \begin{figure}[htb] \begin{center} <>= dimnames(TV2)$Time <- c("8", "9", "10") # re-level for mosaic display mosaic(~ Day + Network + Time, data=TV2, expected=~Day:Time + Network, legend=FALSE, gp=shading_Friendly) @ \caption{Mosaic plot for the \data{TV} data showing model of joint independence, \texttt{Day:Time + Network} .} \label{fig:TV-mosaic} \end{center} \end{figure} From this, it is easy to read from the display how network choice varies with day and time. For example, CBS dominates in all time slots on Monday; ABC and NBC dominate on Tuesday, particularly in the later time slots; Thursday is an NBC day, while on Friday, ABC gets the greatest share. In interpreting this mosaic and other plots, it is important to understand that associations included in the model---here, that between day and time---are \emph{not} shown in the shading of the cells, because they have been fitted (taken into account) in the \loglin\ model. For comparison, you might want to try fitting the model of homogeneous association. This allows all pairs of factors to be associated, but asserts that each pairwise association is the same across the levels of the remaining factor. The resulting plot displays the contributions to a 3-way association, but is not shown here. <>= mosaic(~ Day + Network + Time, data=TV2, expected=~Day:Time + Day:Network + Time:Network, legend=FALSE, gp=shading_Friendly) @ \subsection[Mosaics for glm() and gnm() models]{Mosaics for \codefun{glm} and \codefun{gnm} models}\label{sec:mosglm} The \pkg{vcdExtra} package provides an additional method, \codefun{mosaic.glm} for models fit with \codefun{glm} and \codefun{gnm}.% \footnote{ Models fit with \codefun{gnm} are of \code{class = c("gnm", "glm", "lm")}, so all \code{*.glm} methods apply, unless overridden in the \pkg{gnm} package. } These are not restricted to the Poisson family, but only apply to cases where the response variable is non-negative. \Example Here, we plot the independence and the linear-by-linear association model for the Mental health data from \secref{sec:glm}. These examples illustrate some of the options for labeling (variable names and residuals printed in cells). Note that the \code{formula} supplied to \codefun{mosaic} for \class{glm} objects refers to the order of factors displayed in the plot, not the model. <>= long.labels <- list(set_varnames = c(mental="Mental Health Status", ses="Parent SES")) mosaic(indep, ~ses+mental, residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, main="Mental health data: Independence") mosaic(linlin, ~ses+mental, residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, suppress=1, gp=shading_Friendly, main="Mental health data: Linear x Linear") @ \setkeys{Gin}{width=0.49\textwidth} <>= long.labels <- list(set_varnames = c(mental="Mental Health Status", ses="Parent SES")) mosaic(indep, ~ses+mental, residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, main="Mental health data: Independence") @ <>= long.labels <- list(set_varnames = c(mental="Mental Health Status", ses="Parent SES")) mosaic(linlin, ~ses+mental, residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, suppress=1, gp=shading_Friendly, main="Mental health data: Linear x Linear") @ The \pkg{gnm} package also fits a wide variety of models with nonlinear terms or terms for structured associations of table variables. In the following, we fit the RC(1) model \begin{equation*} \log ( m_{ij} ) = \mu + \lambda_i^A + \lambda_j^B + \phi \mu_i \nu_j \end{equation*} This is similar to the linear by linear model, except that the row effect parameters ($\mu_i$) and column parameters ($\nu_j$) are estimated from the data rather than given assigned equally-spaced values. The multiplicative terms are specified by the \codefun{Mult}. <>= Mental$mental <- C(Mental$mental, treatment) Mental$ses <- C(Mental$ses, treatment) RC1model <- gnm(Freq ~ mental + ses + Mult(mental, ses), family = poisson, data = Mental) mosaic(RC1model, residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, suppress=1, gp=shading_Friendly, main="Mental health data: RC(1) model") @ Other forms of nonlinear terms are provided for the inverse of a predictor (\codefun{Inv}) and the exponential of a predictor (\codefun{Exp}). You should read \code{vignette("gnmOverview", package="gnm")} for further details. \subsection{Mosaic tips and techniques}\label{sec:tips} The \pkg{vcd} package implements an extremely general collection of graphical methods for $n$-way frequency tables within the strucplot framework, which includes mosaic plots (\codefun{mosaic}), as well as association plots (\codefun{assoc}), sieve diagrams (\codefun{sieve}), as well as tabular displays (\codefun{structable}). The graphical methods in \pkg{vcd} support a wide of options that control almost all of the details of the plots, but it is often difficult to determine what arguments you need to supply to achieve a given effect from the \code{help()}. As a first step, you should read the \code{vignette("strucplot")} in \pkg{vcd} to understand the overall structure of these plot methods. The notes below describe a few useful things that may not be obvious, or can be done in different ways. \subsubsection[Changing labels]{Changing the labels for variables and levels} With data in contingency table form or as a frequency data frame, it often happens that the variable names and/or the level values of the factors, while suitable for analysis, are less than adequate when used in mosaic plots and other strucplot displays. For example, we might prefer that a variable named \code{ses} appear as \code{"Socioeconomic Status"}, or a factor with levels \code{c("M", "F")} be labeled using \code{c("Male", "Female")} in a plot. Or, sometimes we start with a factor whose levels are fully spelled out (e.g., \code{c("strongly disagree", "disagree", "neutral", "agree", "strongly agree")}), only to find that the level labels overlap in graphic displays. The structplot framework in \pkg{vcd} provides an extremely large variety of functions and options for controlling almost all details of text labels in mosaics and other plots. See \code{help(labelings)} for an overview. For example, in \secref{sec:ordered-factors} we showed how to rearrange the dimensions of the \code{UCBAdmissions} table, change the names of the table variables, and relabel the levels of one of the table variables. The code below changes the actual table for plotting purposes, but we pointed out that these changes can create other problems in analysis. <>= UCB <- aperm(UCBAdmissions, c(2, 1, 3)) names(dimnames(UCB)) <- c("Sex", "Admit?", "Department") dimnames(UCB)[[2]] <- c("Yes", "No") @ The same effects can be achieved \emph{without} modifying the data using the \verb|set_varnames| and \verb|set_labels| options in \codefun{mosaic} as follows: <>= vnames <- list(set_varnames = c(Admit="Admission", Gender="Sex", Dept="Department")) lnames <- list(Admit = c("Yes", "No"), Gender = c("Males", "Females"), Dept = LETTERS[1:6]) mosaic(UCBAdmissions, labeling_args=vnames, set_labels=lnames) @ In some cases, it may be sufficient to abbreviate (or clip, or rotate) level names to avoid overlap. For example, the statements below produce another version of \figref{fig:TV-mosaic} with days of the week abbreviated to their first three letters. Section 4 in the \code{vignette("strucplot")} provides many other examples. <>= dimnames(TV2)$Time <- c("8", "9", "10") # re-level for mosaic display mosaic(~ Day + Network + Time, data=TV2, expected=~Day:Time + Network, legend=FALSE, gp=shading_Friendly, labeling_args=list(abbreviate=c(Day=3)) ) @ %\subsubsection{Fitting complex models with glm() and gnm()} \section[Continuous predictors]{Continuous predictors}\label{sec:contin} When continuous predictors are available---and potentially important--- in explaining a categorical outcome, models for that outcome include: logistic regression (binary response), the proportional odds model (ordered polytomous response), multinomial (generalized) logistic regression. Many of these are special cases of the generalized linear model using the \code{"poisson"} or \code{"binomial"} family and their relatives. \subsection{Spine and conditional density plots}\label{sec:spine} I don't go into fitting such models here, but I would be remiss not to illustrate some visualizations in \pkg{vcd} that are helpful here. The first of these is the spine plot or spinogram \citep{vcd:Hummel:1996} (produced with \codefun{spine}). These are special cases of mosaic plots with specific spacing and shading to show how a categorical response varies with a continuous or categorical predictor. They are also a generalization of stacked bar plots where not the heights but the \emph{widths} of the bars corresponds to the relative frequencies of \code{x}. The heights of the bars then correspond to the conditional relative frequencies of {y} in every \code{x} group. \Example For the \data{Arthritis} data, we can see how \code{Improved} varies with \code{Age} as follows. \codefun{spine} takes a formula of the form \verb|y ~ x| with a single dependent factor and a single explanatory variable \code{x} (a numeric variable or a factor). The range of a numeric variable\code{x} is divided into intervals based on the \code{breaks} argument, and stacked bars are drawn to show the distribution of \code{y} as \code{x} varies. As shown below, the discrete table that is visualized is returned by the function. <>= (spine(Improved ~ Age, data = Arthritis, breaks = 3)) (spine(Improved ~ Age, data = Arthritis, breaks = "Scott")) @ \setkeys{Gin}{width=0.49\textwidth} <>= (spine(Improved ~ Age, data = Arthritis, breaks = 3)) @ <>= (spine(Improved ~ Age, data = Arthritis, breaks = "Scott")) @ The conditional density plot \citep{vcd:Hofmann+Theus} is a further generalization. This visualization technique is similar to spinograms, but uses a smoothing approach rather than discretizing the explanatory variable. As well, it uses the original \code{x} axis and not a distorted one. \setkeys{Gin}{width=0.6\textwidth} \begin{figure}[htb] \begin{center} <>= cdplot(Improved ~ Age, data = Arthritis) with(Arthritis, rug(jitter(Age), col="white", quiet=TRUE)) @ \caption{Conditional density plot for the \data{Arthritis} data showing the variation of Improved with Age.} \label{fig:cd-plot} \end{center} \end{figure} In such plots, it is useful to also see the distribution of the observations across the horizontal axis, e.g., with a \codefun{rug} plot. \figref{fig:cd-plot} uses \codefun{cdplot} from the \pkg{graphics} package rather than \verb|cd_plot()| from \pkg{vcd}, and is produced with <>= cdplot(Improved ~ Age, data = Arthritis) with(Arthritis, rug(jitter(Age), col="white", quiet=TRUE)) @ From \figref{fig:cd-plot} it can be easily seen that the proportion of patients reporting Some or Marked improvement increases with Age, but there are some peculiar bumps in the distribution. These may be real or artifactual, but they would be hard to see with most other visualization methods. When we switch from non-parametric data exploration to parametric statistical models, such effects are easily missed. \subsection[Model-based plots]{Model-based plots: effect plots and \pkg{ggplot2} plots}\label{sec:modelplots} The nonparametric conditional density plot uses smoothing methods to convey the distributions of the response variable, but displays that are simpler to interpret can often be obtained by plotting the predicted response from a parametric model. For complex \codefun{glm} models with interaction effects, the \pkg{effects} package provides the most useful displays, plotting the predicted values for a given term, averaging over other predictors not included in that term. I don't illustrate this here, but see \citet{effects:1,effects:2} and \code{help(package="effects")}. Here I just briefly illustrate the capabilities of the \pkg{ggplot2} package for model-smoothed plots of categorical responses in \codefun{glm} models. \Example The \data{Donner} data frame in \pkg{vcdExtra} gives details on the survival of 90 members of the Donner party, a group of people who attempted to migrate to California in 1846. They were trapped by an early blizzard on the eastern side of the Sierra Nevada mountains, and before they could be rescued, nearly half of the party had died. What factors affected who lived and who died? <>= data(Donner, package="vcdExtra") str(Donner) @ A potential model of interest is the logistic regression model for $Pr(survived)$, allowing separate fits for males and females as a function of \code{age}. The key to this is the \verb|stat_smooth()| function, using \code{method = "glm", method.args = list(family = binomial)}. The \verb|formula = y ~ x| specifies a linear fit on the logit scale (\figref{fig:donner3}, left) <>= # separate linear fits on age for M/F ggplot(Donner, aes(age, survived, color = sex)) + geom_point(position = position_jitter(height = 0.02, width = 0)) + stat_smooth(method = "glm", method.args = list(family = binomial), formula = y ~ x, alpha = 0.2, size=2, aes(fill = sex)) @ Alternatively, we can allow a quadratic relation with \code{age} by specifying \verb|formula = y ~ poly(x,2)| (\figref{fig:donner3}, right). <>= # separate quadratics ggplot(Donner, aes(age, survived, color = sex)) + geom_point(position = position_jitter(height = 0.02, width = 0)) + stat_smooth(method = "glm", method.args = list(family = binomial), formula = y ~ poly(x,2), alpha = 0.2, size=2, aes(fill = sex)) @ \setkeys{Gin}{width=0.49\textwidth} \begin{figure}[htb] \begin{center} <>= ggplot(Donner, aes(age, survived, color = sex)) + geom_point(position = position_jitter(height = 0.02, width = 0)) + stat_smooth(method = "glm", method.args = list(family = binomial), formula = y ~ x, alpha = 0.2, size=2, aes(fill = sex)) @ <>= # separate quadratics ggplot(Donner, aes(age, survived, color = sex)) + geom_point(position = position_jitter(height = 0.02, width = 0)) + stat_smooth(method = "glm", method.args = list(family = binomial), formula = y ~ poly(x,2), alpha = 0.2, size=2, aes(fill = sex)) @ \caption{Logistic regression plots for the \data{Donner} data showing survival vs. age, by sex. Left: linear logistic model; right: quadratic model} \label{fig:donner3} \end{center} \end{figure} These plots very nicely show (a) the fitted $Pr(survived)$ for males and females; (b) confidence bands around the smoothed model fits and (c) the individual observations by jittered points at 0 and 1 for those who died and survided, respectively. \bibliography{vcd,vcdExtra} \end{document} vcdExtra/vignettes/vcd.bib0000644000176200001440000006353013163461153015306 0ustar liggesusers%% general graphics & original methods @Article{vcd:Cohen:1980, author = {A. Cohen}, title = {On the Graphical Display of the Significant Components in a Two-Way Contingency Table}, journal = {Communications in Statistics---Theory and Methods}, year = {1980}, volume = {A9}, pages = {1025--1041} } @InProceedings{vcd:Hartigan+Kleiner:1981, author = {J. A. Hartigan and B. Kleiner}, title = {Mosaics for Contingency Tables}, booktitle = {Computer Science and Statistics: Proceedings of the 13th Symposium on the Interface}, pages = {268--273}, year = {1981}, editor = {W. F. Eddy}, address = {New York}, publisher = {Springer-Verlag} } @Article{vcd:Hartigan+Kleiner:1984, author = {J. A. Hartigan and B. Kleiner}, title = {A Mosaic of Television Ratings}, journal = {The American Statistician}, year = {1984}, volume = {38}, pages = {32--35} } @TechReport{vcd:Young:1996, author = {Forrest W. Young}, title = {{\pkg{ViSta}}: The Visual Statistics System}, institution = {UNC L.~L.~Thurstone Psychometric Laboratory Research Memorandum}, year = 1996, number = {94--1(c)} } @Book{vcd:Cleveland:1993, author = {William S. Cleveland}, title = {Visualizing Data}, publisher = {Hobart Press}, year = 1993, address = {Summit, New Jersey} } @Article{vcd:Becker+Cleveland+Shyu:1996, author = {Richard A. Becker and William S. Cleveland and Ming-Jen Shyu}, title = {The Visual Design and Control of Trellis Display}, journal = {Journal of Computational and Graphical Statistics}, year = {1996}, volume = {5}, pages = {123--155} } @InProceedings{vcd:Riedwyl+Schuepbach:1994, author = {H. Riedwyl and M. Sch{\"u}pbach}, title = {Parquet Diagram to Plot Contingency Tables}, booktitle = {Softstat '93: Advances in Statistical Software}, pages = {293--299}, year = 1994, editor = {F. Faulbaum}, address = {New York}, publisher = {Gustav Fischer} } %% color @InProceedings{vcd:Ihaka:2003, author = {Ross Ihaka}, title = {Colour for Presentation Graphics}, booktitle = {Proceedings of the 3rd International Workshop on Distributed Statistical Computing, Vienna, Austria}, editor = {Kurt Hornik and Friedrich Leisch and Achim Zeileis}, year = {2003}, url = {http://www.ci.tuwien.ac.at/Conferences/DSC-2003/Proceedings/}, note = {{ISSN 1609-395X}}, } @Article{vcd:Lumley:2006, author = {Thomas Lumley}, title = {Color Coding and Color Blindness in Statistical Graphics}, journal = {ASA Statistical Computing \& Graphics Newsletter}, year = {2006}, volume = {17}, number = {2}, pages = {4--7} } @Book{vcd:Munsell:1905, author = {Albert H. Munsell}, title = {A Color Notation}, publisher = {Munsell Color Company}, year = {1905}, address = {Boston, Massachusetts} } @Article{vcd:Harrower+Brewer:2003, author = {Mark A. Harrower and Cynthia A. Brewer}, title = {\pkg{ColorBrewer.org}: An Online Tool for Selecting Color Schemes for Maps}, journal = {The Cartographic Journal}, year = {2003}, volume = {40}, pages = {27--37} } @InProceedings{vcd:Brewer:1999, author = {Cynthia A. Brewer}, title = {Color Use Guidelines for Data Representation}, booktitle = {Proceedings of the Section on Statistical Graphics, American Statistical Association}, address = {Alexandria, VA}, year = {1999}, pages = {55--60} } @Article{vcd:Cleveland+McGill:1983, author = {William S. Cleveland and Robert McGill}, title = {A Color-caused Optical Illusion on a Statistical Graph}, journal = {The American Statistician}, year = {1983}, volume = {37}, pages = {101--105} } @Book{vcd:CIE:2004, author = {{Commission Internationale de l'\'Eclairage}}, title = {Colorimetry}, edition = {3rd}, publisher = {Publication CIE 15:2004}, address = {Vienna, Austria}, year = {2004}, note = {{ISBN} 3-901-90633-9} } @InProceedings{vcd:Moretti+Lyons:2002, author = {Giovanni Moretti and Paul Lyons}, title = {Tools for the Selection of Colour Palettes}, booktitle = {Proceedings of the New Zealand Symposium On Computer-Human Interaction (SIGCHI 2002)}, address = {University of Waikato, New Zealand}, month = {July}, year = {2002} } @Article{vcd:MacAdam:1942, author = {D. L. MacAdam}, title = {Visual Sensitivities to Color Differences in Daylight}, journal = {Journal of the Optical Society of America}, year = {1942}, volume = {32}, number = {5}, pages = {247--274}, } @Book{vcd:Wyszecki+Stiles:2000, author = {G\"unter Wyszecki and W. S. Stiles}, title = {Color Science}, edition = {2nd}, publisher = {Wiley}, year = {2000}, note = {{ISBN} 0-471-39918-3} } @Misc{vcd:Poynton:2000, author = {Charles Poynton}, title = {Frequently-Asked Questions About Color}, year = {2000}, howpublished = {URL \url{http://www.poynton.com/ColorFAQ.html}}, note = {Accessed 2006-09-14}, } @Misc{vcd:Wiki+HSV:2006, author = {Wikipedia}, title = {{HSV} Color Space --- {W}ikipedia{,} The Free Encyclopedia}, year = {2006}, howpublished = {URL \url{http://en.wikipedia.org/w/index.php?title=HSV_color_space&oldid=74735552}}, note = {Accessed 2006-09-14}, } @Misc{vcd:Wiki+LUV:2006, author = {Wikipedia}, title = {{Lab} Color Space --- {W}ikipedia{,} The Free Encyclopedia}, year = {2006}, howpublished = {URL \url{http://en.wikipedia.org/w/index.php?title=Lab_color_space&oldid=72611029}}, note = {Accessed 2006-09-14}, } @Article{vcd:Smith:1978, author = {Alvy Ray Smith}, title = {Color Gamut Transform Pairs}, journal = {Computer Graphics}, pages = {12--19}, year = {1978}, volume = {12}, number = {3}, note = {ACM SIGGRAPH 78 Conference Proceedings}, } %% url = {http://www.alvyray.com/}, @Article{vcd:Meier+Spalter+Karelitz:2004, author = {Barbara J. Meier and Anne Morgan Spalter and David B. Karelitz}, title = {Interactive Color Palette Tools}, journal = {{IEEE} Computer Graphics and Applications}, volume = {24}, number = {3}, year = {2004}, pages = {64--72}, } %% url = {http://graphics.cs.brown.edu/research/color/} @InCollection{vcd:Mollon:1995, author = {J. Mollon}, editor = {T. Lamb and J. Bourriau}, booktitle = {Colour: Art and Science}, title = {Seeing Color}, publisher = {Cambridge Univesity Press}, year = 1995 } %% Friendly publications @Article{vcd:Friendly:1994, author = {Michael Friendly}, title = {Mosaic Displays for Multi-Way Contingency Tables}, journal = {Journal of the American Statistical Association}, year = {1994}, volume = {89}, pages = {190--200} } @Article{vcd:Friendly:1999, author = {Michael Friendly}, title = {Extending Mosaic Displays: Marginal, Conditional, and Partial Views of Categorical Data}, journal = {Journal of Computational and Graphical Statistics}, year = {1999}, volume = {8}, number = {3}, pages = {373--395} } @Book{vcd:Friendly:2000, author = {Michael Friendly}, title = {Visualizing Categorical Data}, publisher = {\textsf{SAS} Insitute}, year = {2000}, address = {Carey, NC}, URL = {http://www.math.yorku.ca/SCS/vcd/} } @Book{FriendlyMeyer:2016:DDAR, title = {Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data}, year = {2016}, author = {Friendly, Michael and Meyer, David}, publisher = {Chapman \& Hall/CRC}, address = {Boca Raton, FL}, isbn = {978-1-4987-2583-5}, } %% Augsburg publications @Article{vcd:Theus+Lauer:1999, author = {Martin Theus and Stephan R. W. Lauer}, title = {Visualizing Loglinear Models}, journal = {Journal of Computational and Graphical Statistics}, year = 1999, volume = 8, number = 3, pages = {396--412} } @Article{vcd:Hofmann:2003, author = {Heike Hofmann}, title = {Constructing and Reading Mosaicplots}, journal = {Computational Statistics \& Data Analysis}, year = {2003}, volume = {43}, pages = {565--580} } @Article{vcd:Hofmann:2001, author = {Heike Hofmann}, title = {Generalized Odds Ratios for Visual Modelling}, journal = {Journal of Computational and Graphical Statistics}, year = {2001}, volume = {10}, pages = {1--13} } @Article{vcd:Theus:2003, author = {Martin Theus}, title = {Interactive Data Visualization Using \pkg{Mondrian}}, journal = {Journal of Statistical Software}, volume = 7, number = 11, pages = {1--9}, year = 2003, url = {http://www.jstatsoft.org/v07/i11/}, } @Unpublished{vcd:Hofmann+Theus, author = {Heike Hofmann and Martin Theus}, title = {Interactive Graphics for Visualizing Conditional Distributions}, note = {Unpublished Manuscript}, year = {2005} } @Article{vcd:Hummel:1996, author = {J. Hummel}, title = {Linked Bar Charts: Analysing Categorical Data Graphically}, journal = {Computational Statistics}, year = 1996, volume = 11, pages = {23--33} } @Article{vcd:Unwin+Hawkins+Hofmann:1996, author = {Antony R. Unwin and G. Hawkins and Heike Hofmann and B. Siegl}, title = {Interactive Graphics for Data Sets with Missing Values -- \pkg{MANET}}, journal = {Journal of Computational and Graphical Statistics}, year = 1996, pages = {113--122}, volume = 4, number = 6 } @Manual{vcd:Urbanek+Wichtrey:2006, title = {\pkg{iplots}: Interactive Graphics for \textsf{R}}, author = {Simon Urbanek and Tobias Wichtrey}, year = {2006}, note = {\textsf{R} package version 1.0-3}, url = {http://www.rosuda.org/iPlots/} } %% Software @Manual{vcd:R:2006, title = {\textsf{R}: {A} Language and Environment for Statistical Computing}, author = {{\textsf{R} Development Core Team}}, organization = {\textsf{R} Foundation for Statistical Computing}, address = {Vienna, Austria}, year = {2006}, note = {{ISBN} 3-900051-00-3}, url = {http://www.R-project.org/} } @Article{vcd:Murrell:2002, author = {Paul Murrell}, title = {The \pkg{grid} Graphics Package}, journal = {\proglang{R} News}, year = 2002, volume = 2, number = 2, pages = {14--19}, month = {June}, url = {http://CRAN.R-project.org/doc/Rnews/} } @Book{vcd:Murrell:2006, author = {Paul Murrell}, title = {\textsf{R} Graphics}, publisher = {Chapmann \& Hall/CRC}, address = {Boca Raton, Florida}, year = {2006}, } @Book{vcd:Venables+Ripley:2002, author = {William N. Venables and Brian D. Ripley}, title = {Modern Applied Statistics with \textsf{S}}, edition = {4th}, publisher = {Springer-Verlag}, address = {New York}, year = {2002}, note = {{ISBN} 0-387-95457-0}, url = {http://www.stats.ox.ac.uk/pub/MASS4/} } @Manual{vcd:Ihaka:2006, title = {\pkg{colorspace}: Colorspace Manipulation}, author = {Ross Ihaka}, year = {2006}, note = {\textsf{R} package version 0.95} } @Manual{vcd:Meyer+Zeileis+Hornik:2006, title = {\pkg{vcd}: Visualizing Categorical Data}, author = {David Meyer and Achim Zeileis and Kurt Hornik}, year = {2006}, note = {\textsf{R} package version 1.0-6} } @article{vcd:Ligges+Maechler:2003, title = {\pkg{scatterplot3d} -- An {R} Package for Visualizing Multivariate Data}, author = {Uwe Ligges and Martin M{\"a}chler}, journal = {Journal of Statistical Software}, year = 2003, pages = {1--20}, number = 11, volume = 8, url = {http://www.jstatsoft.org/v08/i11/} } @Manual{vcd:SAS:2005, title = {\proglang{SAS/STAT} Version 9}, author = {\proglang{SAS} Institute Inc.}, year = {2005}, address = {Cary, NC} } @Manual{vcd:SPLUS:2005, title = {\proglang{S-PLUS} 7}, author = {{Insightful Inc.}}, year = {2005}, address = {Seattle, WA} } %% data @Article{vcd:Azzalini+Bowman:1990, author = {A. Azzalini and A. W. Bowman}, title = {A Look at Some Data on the {O}ld {F}aithful Geyser}, journal = {Applied Statistics}, year = {1990}, volume = {39}, pages = {357--365}, } @Article{vcd:Obel:1975, author = {E.B. Obel}, title = {A Comparative Study of Patients with Cancer of the Ovary Who Have Survived More or Less Than 10 Years}, journal = {Acta Obstetricia et Gynecologica Scandinavica}, year = 1975, volume = 55, pages = {429--439} } @InCollection{vcd:Koch+Edwards:1988, author = {G. Koch and S. Edwards}, title = {Clinical Efficiency Trials with Categorical Data}, booktitle = {Biopharmaceutical Statistics for Drug Development}, editor = {K. E. Peace}, publisher = {Marcel Dekker}, address = {New York}, year = {1988}, pages = {403--451} } @TechReport{vcd:Knorr-Held:1999, author = {Leonhard Knorr-Held}, title = {Dynamic Rating of Sports Teams}, institution = {SFB 386 ``Statistical Analysis of Discrete Structures''}, year = {1999}, type = {Discussion Paper}, number = {98}, url = {http://www.stat.uni-muenchen.de/sfb386/} } @Article{vcd:Snee:1974, author = {R. D. Snee}, title = {Graphical Display of Two-Way Contingency Tables}, journal = {The American Statistician}, year = 1974, volume = 28, pages = {9--12} } @Article{vcd:Bickel+Hammel+O'Connell:1975, author = {P. J. Bickel and E. A. Hammel and J. W. O'Connell}, title = {Sex Bias in Graduate Admissions: Data from {B}erkeley}, journal = {Science}, year = 1975, volume = 187, pages = {398--403} } @Book{vcd:Gilbert:1981, author = {G. N. Gilbert}, title = {Modelling Society: An Introduction to Loglinear Analysis for Social Researchers}, publisher = {Allen and Unwin}, year = 1981, address = {London} } @Book{vcd:Thornes+Collard:1979, author = {B. Thornes and J. Collard}, title = {Who Divorces?}, publisher = {Routledge \& Kegan}, year = 1979, address = {London} } @Article{vcd:Dawson:1995, author = {Robert J. MacG Dawson}, title = {The ``Unusual Episode'' Data Revisited}, journal = {Journal of Statistics Education}, year = 1995, volume = 3, url = {http://www.amstat.org/publications/jse/v3n3/datasets.dawson.html} } @Article{vcd:Haberman:1974, author = {S. J. Haberman}, title = {Log-linear Models for Frequency Tables with Ordered Classifications}, journal = {Biometrics}, year = 1974, volume = 30, pages = {689--700} } @Article{vcd:Wing:1962, author = {J. K. Wing}, title = {Institutionalism in Mental Hospitals}, journal = {British Journal of Social Clinical Psychology}, year = 1962, volume = 1, pages = {38--51} } @Book{vcd:Andersen:1991, author = {E. B. Andersen}, title = {The Statistical Analysis of Categorical Data}, publisher = {Springer-Verlag}, year = {1991}, address = {Berlin}, edition = {2nd} } @Article{vcd:Haberman:1973, author = {S. J. Haberman}, title = {The Analysis of Residuals in Cross-classified Tables}, journal = {Biometrics}, year = {1973}, volume = {29}, pages = {205--220} } @Book{vcd:Everitt+Hothorn:2006, author = {Brian S. Everitt and Torsten Hothorn}, title = {A Handbook of Statistical Analyses Using \textsf{R}}, publisher = {Chapman \& Hall/CRC}, address = {Boca Raton, Florida}, year = {2006} } @Article{vcd:Salib+Hillier:1997, author = {Emad Salib and Valerie Hillier}, title = {A Case-Control Study of Smoking and {A}lzheimer's Disease}, journal = {International Journal of Geriatric Psychiatry}, year = {1997}, volume = {12}, pages = {295--300} } %% inference @Book{vcd:Agresti:2002, author = {Alan Agresti}, title = {Categorical Data Analysis}, publisher = {John Wiley \& Sons}, year = {2002}, address = {Hoboken, New Jersey}, edition = {2nd} } @Book{vcd:Mazanec+Strasser:2000, author = {Josef A. Mazanec and Helmut Strasser}, title = {A Nonparametric Approach to Perceptions-based Market Segmentation: Foundations}, publisher = {Springer-Verlag}, year = {2000}, address = {Berlin} } @Article{vcd:Strasser+Weber:1999, author = {Helmut Strasser and Christian Weber}, title = {On the Asymptotic Theory of Permutation Statistics}, journal = {Mathematical Methods of Statistics}, volume = {8}, pages = {220--250}, year = {1999} } @Book{vcd:Pesarin:2001, author = {Fortunato Pesarin}, title = {Multivariate Permutation Tests}, year = {2001}, publisher = {John Wiley \& Sons}, address = {Chichester} } @Article{vcd:Ernst:2004, author = {Michael D. Ernst}, title = {Permutation Methods: A Basis for Exact Inference}, journal = {Statistical Science}, volume = {19}, year = {2004}, pages = {676--685} } @Article{vcd:Patefield:1981, author = {W. M. Patefield}, title = {An Efficient Method of Generating $R \times C$ Tables with Given Row and Column Totals}, note = {{A}lgorithm AS 159}, journal = {Applied Statistics}, volume = {30}, year = {1981}, pages = {91--97} } %% own @InProceedings{vcd:Meyer+Zeileis+Hornik:2003, author = {David Meyer and Achim Zeileis and Kurt Hornik}, title = {Visualizing Independence Using Extended Association Plots}, booktitle = {Proceedings of the 3rd International Workshop on Distributed Statistical Computing, Vienna, Austria}, editor = {Kurt Hornik and Friedrich Leisch and Achim Zeileis}, year = {2003}, url = {http://www.ci.tuwien.ac.at/Conferences/DSC-2003/Proceedings/}, note = {{ISSN 1609-395X}}, } @TechReport{vcd:Zeileis+Meyer+Hornik:2005, author = {Achim Zeileis and David Meyer and Kurt Hornik}, title = {Residual-based Shadings for Visualizing (Conditional) Independence}, institution = {Department of Statistics and Mathematics, Wirtschaftsuniversit\"at Wien, Research Report Series}, year = {2005}, type = {Report}, number = {20}, month = {August}, url = {http://epub.wu-wien.ac.at/dyn/openURL?id=oai:epub.wu-wien.ac.at:epub-wu-01_871} } @Article{vcd:Zeileis+Meyer+Hornik:2007, author = {Achim Zeileis and David Meyer and Kurt Hornik}, title = {Residual-based Shadings for Visualizing (Conditional) Independence}, journal = {Journal of Computational and Graphical Statistics}, year = {2007}, volume = {16}, number = {3}, pages = {507--525}, doi = {10.1198/106186007X237856}, url = {http://statmath.wu-wien.ac.at/~zeileis/papers/Zeileis+Meyer+Hornik-2007.pdf} } @TechReport{vcd:Meyer+Zeileis+Hornik:2005a, author = {David Meyer and Achim Zeileis and Kurt Hornik}, title = {The Strucplot Framework: Visualizing Multi-Way Contingency Tables with \pkg{vcd}}, institution = {Department of Statistics and Mathematics, Wirtschaftsuniversit\"at Wien, Research Report Series}, year = {2005}, type = {Report}, number = {22}, month = {November}, url = {http://epub.wu-wien.ac.at/dyn/openURL?id=oai:epub.wu-wien.ac.at:epub-wu-01_8a1} } @Article{vcd:Meyer+Zeileis+Hornik:2006b, author = {David Meyer and Achim Zeileis and Kurt Hornik}, title = {The Strucplot Framework: Visualizing Multi-way Contingency Tables with \pkg{vcd}}, year = {2006}, journal = {Journal of Statistical Software}, volume = {17}, number = {3}, pages = {1--48}, url = {http://www.jstatsoft.org/v17/i03/} } @InCollection{vcd:Meyer+Zeileis+Hornik:2006a, author = {David Meyer and Achim Zeileis and Kurt Hornik}, title = {Visualizing Contingency Tables}, editor = {Chun-Houh Chen and Wolfang H\"ardle and Antony Unwin}, booktitle = {Handbook of Data Visualization}, series = {Springer Handbooks of Computational Statistics}, year = {2006}, publisher = {Springer-Verlag}, address = {New York}, note = {{ISBN} 3-540-33036-4, to appear} } @Article{vcd:Hothorn+Hornik+VanDeWiel:2006, author = {Torsten Hothorn and Kurt Hornik and Mark A. van de Wiel and Achim Zeileis}, title = {A {L}ego System for Conditional Inference}, journal = {The American Statistician}, year = {2006}, volume = {60}, number = {3}, pages = {257--263}, doi = {10.1198/000313006X118430} } @TechReport{vcd:Zeileis+Hornik:2006, author = {Achim Zeileis and Kurt Hornik}, title = {Choosing Color Palettes for Statistical Graphics}, institution = {Department of Statistics and Mathematics, Wirtschaftsuniversit\"at Wien, Research Report Series}, year = {2006}, type = {Report}, number = {41}, month = {October}, url = {http://epub.wu-wien.ac.at/} } %% bad color examples @Article{vcd:Gneiting+Sevcikova+Percival:2006, author = {Tilmann Gneiting and Hana \v{S}ev\v{c}\'ikov\'a and Donald B. Percival and Martin Schlather and Yindeng Jiang}, title = {Fast and Exact Simulation of Large Gaussian Lattice Systems in {$\mathbb{R}^2$}: Exploring the Limits}, year = {2006}, journal = {Journal of Computational and Graphical Statistics}, volume = {15}, number = {3}, pages = {483--501}, note = {Figures~1--4} } @Article{vcd:Yang+Buckley+Dudoit:2002, author = {Yee Hwa Yang and Michael J. Buckley and Sandrine Dudoit and Terence P. Speed}, title = {Comparison of Methods for Image Analysis on {cDNA} Microarray Data}, year = {2002}, journal = {Journal of Computational and Graphical Statistics}, volume = {11}, number = {1}, pages = {108--136}, note = {Figure~4a} } @Article{vcd:Kneib:2006, author = {Thomas Kneib}, title = {Mixed Model-based Inference in Geoadditive Hazard Regression for Interval-censored Survival Times}, year = {2006}, journal = {Computational Statistics \& Data Analysis}, volume = {51}, pages = {777--792}, note = {Figure~5 (left)} } @Article{vcd:Friendly:2002, author = {Michael Friendly}, title = {A Brief History of the Mosaic Display}, year = {2002}, journal = {Journal of Computational and Graphical Statistics}, volume = {11}, number = {1}, pages = {89--107}, note = {Figure~11 (left, middle)} } @Article{vcd:Celeux+Hurn+Robert:2000, author = {Gilles Celeux and Merrilee Hurn and Christian P. Robert}, title = {Computational and Inferential Difficulties with Mixture Posterior Distributions}, year = {2000}, journal = {Journal of the American Statistical Association}, volume = {95}, number = {451}, pages = {957--970}, note = {Figure~3} } %% pointers from Hadley @article{cleveland:1987, Author = {Cleveland, William and McGill, Robert}, Journal = {Journal of the Royal Statistical Society A}, Number = {3}, Pages = {192-229}, Title = {Graphical Perception: The Visual Decoding of Quantitative Information on Graphical Displays of Data}, Volume = {150}, Year = {1987}} @article{cleveland:1984, Author = {Cleveland, William S. and McGill, M. E.}, Journal = {Journal of the American Statistical Association}, Number = 387, Pages = {531-554}, Title = {Graphical Perception: Theory, Experimentation and Application to the Development of Graphical Methods}, Volume = 79, Year = 1984} @article{huang:1997, Author = {Huang, Chisheng and McDonald, John Alan and Stuetzle, Werner}, Journal = {Journal of Computational and Graphical Statistics}, Pages = {383--396}, Title = {Variable resolution bivariate plots}, Volume = {6}, Year = {1997}} @article{carr:1987, Author = {Carr, D. B. and Littlefield, R. J. and Nicholson, W. L. and Littlefield, J. S.}, Journal = {Journal of the American Statistical Association}, Number = {398}, Pages = {424-436}, Title = {Scatterplot Matrix Techniques for Large N}, Volume = {82}, Year = {1987}} @book{cleveland:1994, Author = {Cleveland, William}, Publisher = {Hobart Press}, Title = {The Elements of Graphing Data}, Year = {1994}} @book{chambers:1983, Author = {Chambers, John and Cleveland, William and Kleiner, Beat and Tukey, Paul}, Publisher = {Wadsworth}, Title = {Graphical methods for data analysis}, Year = {1983}} @book{bertin:1983, Address = {Madison, WI}, Author = {Bertin, Jacques}, Publisher = {University of Wisconsin Press}, Title = {Semiology of Graphics}, Year = {1983}} @book{wilkinson:2006, Author = {Wilkinson, Leland}, Publisher = {Springer-Verlag}, Series = {Statistics and Computing}, Title = {The Grammar of Graphics}, Year = {2005}} vcdExtra/vignettes/vcdExtra.bib0000644000176200001440000000113713163461153016305 0ustar liggesusers@ARTICLE{Cohen:60, author = {J. Cohen}, title = {A coefficient of agreement for nominal scales}, journal = {Educational and Psychological Measurement}, year = {1960}, volume = {20}, pages = {37--46}, owner = {Michael}, timestamp = {2009.01.21} } @BOOK{Agresti:2013, title = {Categorical Data Analysis}, publisher = {Wiley-Interscience [John Wiley \& Sons]}, year = {2013}, author = {Agresti, Alan}, series = {Wiley Series in Probability and Statistics}, address = {New York}, edition = {Third}, isbn = {978-0-470-46363-5}, lccn = {QA278.A353 2013} } vcdExtra/README.md0000644000176200001440000000515113163476374013331 0ustar liggesusers[![CRAN_Status_Badge](http://www.r-pkg.org/badges/version/vcdExtra)](https://cran.r-project.org/package=vcdExtra) [![](http://cranlogs.r-pkg.org/badges/grand-total/vcdExtra)](https://cran.r-project.org/package=vcdExtra) [![Rdoc](http://www.rdocumentation.org/badges/version/vcdExtra)](http://www.rdocumentation.org/packages/vcdExtra) # vcdExtra ## Extensions and additions to vcd: Visualizing Categorical Data Version 0.7-1 This package provides additional data sets, documentation, and a few functions designed to extend the [vcd](https://CRAN.R-project.org/package=vcd) package for *Visualizing Categorical Data* and the [gnm](https://CRAN.R-project.org/package=gnm) package for *Generalized Nonlinear Models*. In particular, vcdExtra extends mosaic, assoc and sieve plots from vcd to handle `glm()` and `gnm()` models and adds a 3D version in `mosaic3d()`. `vcdExtra` is now a support package for the book [*Discrete Data Analysis with R*](http://ddar.datavis.ca) by Michael Friendly and David Meyer. The main purpose of this package is to serve as a sandbox for introducing extensions of mosaic plots and related graphical methods that apply to loglinear models fitted using `glm()` and related, generalized nonlinear models fitted with `gnm()` in the gnm package. A related purpose is to fill in some holes in the analysis of categorical data in R, not provided in base R, vcd, or other commonly used packages. * The method `mosaic.glm()` extends the `mosaic.loglm()` method in the vcd package to this wider class of models. This method also works for the generalized nonlinear models fit with the gnm package, including models for square tables and models with multiplicative associations. * `mosaic3d()` introduces a 3D generalization of mosaic displays using the [rgl](https://CRAN.R-project.org/package=rgl) package. * A new class, `glmlist`, is introduced for working with collections of glm objects, e.g., `Kway()` for fitting all K-way models from a basic marginal model, and `LRstats()` for brief statistical summaries of goodnes-of-fit for a collection of models. * For square tables with ordered factors, `Crossings()` supplements the specification of terms in model formulas using `Symm()`, `Diag()`, `Topo(),` etc. in the [gnm](https://CRAN.R-project.org/package=gnm) package. * In addition, there are many new data sets, a tutorial vignette, _Working with categorical data with R and the vcd package_, `vignette("vcd-tutorial", package = "vcdExtra")`, and a few useful utility functions for manipulating categorical data sets and working with models for categorical data. vcdExtra/MD50000644000176200001440000002062613163540741012354 0ustar liggesusersbae9b25b53b2f1d151845f3a66b4eff8 *DESCRIPTION 54bd83d63a6c74ec0543e1704b88a5fc *NAMESPACE 14d2e4e7edfa4b464cf43efd95306651 *NEWS 779bcf552908cea9c4d8aee4763b5d94 *R/CMHtest.R 7bf2895634e1b0faca9385bd81ccee37 *R/Crossings.R 1963eb3fb347cf99287b4794b4fd57dc *R/GKgamma.R a7a0bf2f6825917c8e31f99c7b25d265 *R/HLtest.R 10217e3a0582ec4cddba2bfc130a8fad *R/Kway.R c8079ae2c1058550fa5b5fda9624dc8b *R/LRstats.R 8e444f0aaaea055b108a478f8ddf2def *R/Summarise.R 650530a623dec352a740a6626baddac7 *R/blogits.R d3838c45cc40201160f85f0eb732d3c5 *R/collapse.table.R b0483049482397dff2bb9ca2ef123725 *R/cutfac.R 64863f026cfeb85a89bc3644ecf955dc *R/datasets.R e037051f710da39f3e511a8448bdf6b3 *R/expand.dft.R c1cf46896f5087dccb67fe9e27c7d200 *R/glmlist.R ea1051fba28687e9b147a8912baf9802 *R/logLik.loglm.R 564524d08cf38c5eea4926a1c6e05fb3 *R/loglin-utilities.R 67c413ec19c991bf083e0459eb3ff71f *R/logseries.R e66f1348beffe4281c1eb9b80036e586 *R/mcaplot.R 1384d1ce36ba17ce25b00603306a258c *R/modFit.R 014cf5c8db1d36327db64bc35c5a7ed3 *R/mosaic.glm.R 5ce7705b0b6911d86ed0ee6547fc7218 *R/mosaic.glmlist.R de9b6b1b1b831283a118c2e5e9e72932 *R/mosaic3d.R 1d5324d4c8b0e2cb572aa9e61ddb7148 *R/print.Kappa.R 1925c7203e978e7b778c25a5b2afa0c1 *R/seq_loglm.R 1987e35844f89995f5f071e3c219e848 *R/seq_mosaic.R ebdbee3b568920673218c4610b858562 *R/split3d.R 559c21fac93caf6fe54e3e48a6265c17 *R/summarise-old.R e782855ba6eaecbc19b51e1c4b8d69ba *R/update.xtabs.R c2c481edc3058a26cd2eb178510decaa *R/vcdExtra-deprecated.R 5fe4e1529aa78269afa34eac8595e84c *R/zero.test.R caa13f95d94cd4a29dc72407b061be69 *README.md 047d5c5155a0a078b531cfaa5f1d54f7 *build/vignette.rds 6d2a9f9d45fe43a9627455addb4561bc *data/Abortion.rda ba2b62141b5cbaf77c2b06698553e983 *data/Accident.RData d295c350b177b842ccee2e5b164a58f4 *data/AirCrash.RData d775f487470c49d64119693325aa910c *data/Alligator.RData 3bbe75f0bcf3a92f29625be117d1c296 *data/Bartlett.rda 5fc3e4c48e1397773e90780afecbb6ca *data/Burt.RData 0d56582a4b5af7562b3116810d53f961 *data/Caesar.rda b302d59a1a8d1f53d639375b7caf2919 *data/Cancer.rda 3813cf7ec7f60ce7fa8e521d6ba63ba2 *data/Cormorants.RData 3b9f36583cfb6342dcc4d4dcd3998004 *data/CyclingDeaths.RData 1e23f322f74ea63465f67781fb6b2a4d *data/DaytonSurvey.RData eabd8206f271bdb144bb53c0df6a2f92 *data/Depends.RData fd6545bf5e14066aab8b349b62b11fbe *data/Detergent.rda 008242aa4f2235a9c015b7423429ded1 *data/Donner.RData c66f21ad5716aee11c585e033e95fd7f *data/Draft1970.RData 1daeff3b42be51c00f8ad0b77fd83e21 *data/Draft1970table.RData 2d37442d326c44677d93a931a277e212 *data/Dyke.rda 7f3834a840e7060e63a12dfaf43307a6 *data/Fungicide.rda f5aa7f75d2342c90fd71ef90961bdbde *data/GSS.rda f1d55ca4db128ee7bf40f27e7e33daed *data/Geissler.RData 0562850df808df96f4d889575e340c18 *data/Gilby.rda 8bee58957caa50dfed1e572ff2f06ce6 *data/Glass.RData 59e31226e1a72831b243a8b1d9ac2309 *data/HairEyePlace.RData 462a843912c7fa9d65e80216c1b9bf21 *data/Hauser79.RData 7286137b66891022c7040ee165106cd6 *data/Heart.rda 962147756d43e545ce4195c0a008b3c2 *data/Heckman.rda 395c246ca156be6ff2f3c4ddd85e6e02 *data/HospVisits.RData 0be197a140533b84d2aa4e9cbdf53847 *data/Hoyt.rda 3ecbf5b367c5a5dc92de5ca67399a42d *data/ICU.RData 6760c1bde6b7c2f02ee7092bb1afde50 *data/JobSat.rda a9598a5882e39e718e8a011872a99bf0 *data/Mammograms.RData c6c47e0e0aaa247eb5deb17bc8dabc1a *data/Mental.rda 8e4b0f4f2fafb81451d25679de372f12 *data/Mice.RData 5792d4777f84482b21e8924c56f293bc *data/Mobility.rda 245b886dcb1ae47cb60a9a77da485b01 *data/PhdPubs.RData dbe73c96e2d34286a5f527db6e197ddc *data/ShakeWords.RData 6d0ea682819c64ba8bf93637a3252d31 *data/TV.rda 2f9862582e0cea8b0920f1f587829d80 *data/Titanicp.rda 4cd539822e2217dedcc6878c9d571831 *data/Toxaemia.RData a80d9998ac4582bc3bd52e4c426e3c2a *data/Vietnam.RData 8c762b9361d60b4b2fa6ea53e7a47a6a *data/Vietnam.rda b7392e8ea7f5a6d921090b4eb70aaa3a *data/Vote1980.RData 2d049ddb7fcd57fcdcbd381d07bd4e01 *data/WorkerSat.RData 67a9eda069524ea4e9156ba5f79eb05b *data/Yamaguchi87.RData 45bc99442529ca940cb36f44333030f8 *demo/00Index b4a763fc0987f60fd6102f1026d483c1 *demo/Wong2-3.R 55e972f403ae9de662d7e9b6891bd2c7 *demo/Wong3-1.R 688251f0404f6dc0b98773ba16415c1e *demo/housing.R 02eba55e051f4afcbdd923d80dd5f7b0 *demo/mental-glm.R f5c6bad4251ba4fc2879ac10f7755bd9 *demo/mosaic3d-demo.R ba4d09ac1c0d1f3708df0885d8b66dfb *demo/mosaic3d-hec.R bcb29a6d18ecaefa75284a3982aed586 *demo/occStatus.R b8e990080da869639effe9c5be96990f *demo/ucb-glm.R be799fa72e29b45c9c5117ad1396ab88 *demo/vision-quasi.R ee93f55378d740b1ab23b8a298739aaa *demo/yaish-unidiff.R ffd20b4a85ca58f1dc3d8d1c3f4efd92 *demo/yamaguchi-xie.R 3321d04d1eea258c261c59f337c21d1a *inst/doc/extdata/tv.dat 63f6dd54596430cac8d0a13f6b3b443c *inst/doc/vcd-tutorial.R 01cd08ecaba3adaa2d0f247148239fd5 *inst/doc/vcd-tutorial.Rnw 02bc6e94f98f98830a2879f9c06e298e *inst/doc/vcd-tutorial.pdf d87499ead4bdcce6563da42731b65914 *man/Abortion.Rd 72060b65be9ae4869974146b79384a7b *man/Accident.Rd 070efb8d763bb42ecd90deb786742143 *man/AirCrash.Rd f5f3416a87eadefaee0e88f37642724c *man/Alligator.Rd e8aa86902e20ea5a9dd4b739bf6fca6c *man/Bartlett.Rd 5380b712bb57acc195ae82e3fb1fa71e *man/Burt.Rd 27f5225eab21c53a6257aeba94fdafe7 *man/CMHtest.Rd d8ae910df0c07de129ec3b69b3acf481 *man/Caesar.Rd b16ff46a4fdb113c298fa978aec0a4bd *man/Cancer.Rd 86a18a592e7a7ec24b1f53caa89268bc *man/Cormorants.Rd 8491383b7e4ee8a89e64fda529931a7f *man/Crossings.Rd 85c2f9ea62584feba1130a9b8d11ec15 *man/CyclingDeaths.Rd 54dfe1dda33d5483207a39102853a817 *man/DaytonSurvey.Rd 3b12bd35839a927d6a9a33588ee14ce8 *man/Depends.Rd 67adef7a8e34cfd04a95d6dd9e68c9a4 *man/Detergent.Rd 398fb5f555db1f51f0a12127a661cdb7 *man/Donner.Rd 16a2111b4172cf3760ccba1c0bff7fc3 *man/Draft1970.Rd 6fb4be32f341d520ed96a419211a27c2 *man/Draft1970table.Rd 650d7169a8a2b1416ad223862148b157 *man/Dyke.Rd a913825f7fe4e3bcfc7a63c567c4bc3a *man/Fungicide.Rd 065181216f1521a5f18b3a5f45efc22b *man/GKgamma.Rd 30e2d9db679aa94d2c49a5a676ca89e4 *man/GSS.Rd b1437e6a9cea52525fc1f43ec00ac844 *man/Geissler.Rd acea91c7604486003eeee95a8906bb5a *man/Gilby.Rd 386e653640e44e7b5cb55f53c83aba57 *man/Glass.Rd 1ee0e737248ef8617605dc042e55ccce *man/HLtest.Rd e7b3ee0e86a80bd6ae303c8180c44100 *man/HairEyePlace.Rd 53988fccc0bf853b9bb522076381c83d *man/Hauser79.Rd 25cc2198699598097f1aa330b443d7de *man/Heart.Rd 991025420ed138c76aa5d5d88a0da9dc *man/Heckman.Rd 9d625171fb376059b71b99d1f94ee07b *man/HospVisits.Rd ef7a954e4beefdb7cf84f8fe1f84deac *man/Hoyt.Rd 275b18dbbd78838684086e492b916abe *man/ICU.Rd e3240cc9b40bdfea2d1113c0f970ef5c *man/JobSat.Rd b6c19374f7876f389ada9678dfeb34d4 *man/Kway.Rd 9c79c35060171b0ea5d2d7960b35edd0 *man/LRstats.Rd b43fcc6e93a991acd79038b9cc20396b *man/Mammograms.Rd 00778fdd583d325df5651b084909cd50 *man/Mental.Rd 2f67f04ad9d467d577843413db969c91 *man/Mice.Rd 65bb70e078e10f773017651a7e5a13b0 *man/Mobility.Rd 56f7462356a9f2a984ed8628deee9ee8 *man/PhdPubs.Rd 779137e44c6beead4a4eba3f502eb004 *man/ShakeWords.Rd 19ff4c40b4f445d48021fa0dcf0c59a1 *man/Summarise.Rd 5e1054cc83204f1399834dc51ee52e0e *man/TV.Rd dd4c8019dbb90747a6ee87037341b909 *man/Titanicp.Rd a62b058972e3b3c914ebe81bb096928e *man/Toxaemia.Rd ba46552a51ef7fd2e3acb6bc2c12a0cd *man/Vietnam.Rd fe59d7a7284aef51cba27c21df4cfa29 *man/Vote1980.Rd 8bccb08096f91083e1431926c8c8fef9 *man/WorkerSat.Rd e2d1c5edd827bcd18b394325e3eb7466 *man/Yamaguchi87.Rd 6286e9d054cf682373b0880b5f93a238 *man/blogits.Rd dc603b906841137948794e90dddd80b9 *man/collapse.table.Rd d64c532ae465bcf918eee40c427b3550 *man/cutfac.Rd cdca7088a63932032492eb4bc17938f1 *man/datasets.Rd 72e270c4160957d1d79da5e8331b0ce8 *man/expand.dft.Rd 578f9088e7c46eb22a338b1008d06794 *man/glmlist.Rd c79c3f89cd4a8342acb6efbbfd95f936 *man/logLik.loglm.Rd 72447d319ac7398e41dd9848f63602ed *man/loglin-utilities.Rd 44f02f21d3b8f896179f9030f161decc *man/logseries.Rd 98d1b49c2bf0b482d4cf1ab1f836276e *man/mcaplot.Rd d1668c07d59be03d219b588834149f84 *man/modFit.Rd 3ba8cea437e17389940b106d6aefdcb0 *man/mosaic.glm.Rd 7d7b3267571cd0fa2d0fe03fb0c6d869 *man/mosaic.glmlist.Rd 53f3f37172fd350e4a1ccc3287b7ceb8 *man/mosaic3d.Rd 05453d62ff0d832c737ea9cb1062767f *man/print.Kappa.Rd cbc06b6d710e4ecaa9068cb2d296dd21 *man/seq_loglm.Rd c9bc86df1d05dee6cb577694863aeeb2 *man/seq_mosaic.Rd cec84e260d0b884b8c27073e224a998f *man/split3d.Rd 684b8e8a4e00d39ac7ab3587bc3d833d *man/update.xtabs.Rd 86a2288b811a6468dfa0b3ea39bca4ef *man/vcdExtra-deprecated.Rd 9f5f37196a2cfe66f08b0c5984cc337a *man/vcdExtra-package.Rd 1b2bbebd32281250c18ba60d4b3d6dfc *man/zero.test.Rd 01cd08ecaba3adaa2d0f247148239fd5 *vignettes/vcd-tutorial.Rnw 2f31459bee0ca5db8893f77b1ae65c7a *vignettes/vcd.bib d20caf9dc57fb78d792790e721f6d365 *vignettes/vcdExtra.bib vcdExtra/build/0000755000176200001440000000000013163514377013143 5ustar liggesusersvcdExtra/build/vignette.rds0000644000176200001440000000047713163514377015512 0ustar liggesusers‹}PËNÃ0tmhy¨¢ŽþúQÕ U¢×ÅÞ¦V];JL£Þørʦu€„%[ë™ñŽwVcÆXÌ’4bqBe2¥cH{B;b)µõ^È{÷æl¥@ÏrÓüãÒsüÅV[e Þ(·áDÐ\‚ƒ3šs0’» rjËK[(ð_ÇR®þ¦ÿ#azbϪīRù2+vV¢®ý5ÀXÿè80°Ã:°ÉæX¢‘-üxd^r+¬q41qà^5vW;[ƒ¼ÔÖuØe­p=h¦Ô=WFžÜ©3zìŽÌ¤rÊÒþÁGùyªè rñ„‡ÆVÝ$=ÍÐkÒ…Ò]£ÁR¹¯Kò<_|ÉüêßKuTÙfÖ%{ÝÿNÇ‘V¿ÐP‡ñÛ8fëŠÞ·ÿþ6êò¬vcdExtra/DESCRIPTION0000644000176200001440000000327413163540741013552 0ustar liggesusersPackage: vcdExtra Type: Package Title: 'vcd' Extensions and Additions Version: 0.7-1 Date: 2017-09-28 Authors@R: c(person(given = "Michael", family = "Friendly", role=c("aut", "cre"), email="friendly@yorku.ca"), person(given = "Heather", family = "Turner", role="ctb"), person(given = "Achim", family = "Zeileis", role="ctb"), person(given = "Duncan", family = "Murdoch", role="ctb"), person(given = "David", family = "Firth", role="ctb") ) Author: Michael Friendly [aut, cre], Heather Turner [ctb], Achim Zeileis [ctb], Duncan Murdoch [ctb], David Firth [ctb] Maintainer: Michael Friendly Depends: R (>= 2.10), vcd, gnm (>= 1.0.3), grid Suggests: gmodels, Fahrmeir, effects, VGAM, plyr, lmtest, nnet, ggplot2, Sleuth2, car, lattice, stats4, rgl, AER Imports: MASS, grDevices, stats, utils, ca Description: Provides additional data sets, methods and documentation to complement the 'vcd' package for Visualizing Categorical Data and the 'gnm' package for Generalized Nonlinear Models. In particular, 'vcdExtra' extends mosaic, assoc and sieve plots from 'vcd' to handle 'glm()' and 'gnm()' models and adds a 3D version in 'mosaic3d'. Additionally, methods are provided for comparing and visualizing lists of 'glm' and 'loglm' objects. This package is now a support package for the book, "Discrete Data Analysis with R" by Michael Friendly and David Meyer. License: GPL (>= 2) URL: https://CRAN.R-project.org/package=vcdExtra BugReports: https://github.com/friendly/vcdExtra LazyLoad: yes LazyData: yes NeedsCompilation: no Packaged: 2017-09-29 19:03:29 UTC; Friendly Repository: CRAN Date/Publication: 2017-09-29 21:57:53 UTC vcdExtra/man/0000755000176200001440000000000013163461153012610 5ustar liggesusersvcdExtra/man/Burt.Rd0000644000176200001440000000372713163461153014024 0ustar liggesusers\name{Burt} \alias{Burt} \docType{data} \title{ Burt (1950) Data on Hair, Eyes, Head and Stature } \description{ Cyril Burt (1950) gave these data, on a sample of 100 people from Liverpool, to illustrate the application of a method of factor analysis (later called multiple correspondence analysis) applied to categorical data. He presented these data initially in the form that has come to be called a "Burt table", giving the univariate and bivariate frequencies for an n-way frequency table. } \usage{data("Burt")} \format{ A frequency data frame (representing a 3 x 3 x 2 x 2 frequency table) with 36 observations on the following 5 variables. \describe{ \item{\code{Hair}}{hair color, a factor with levels \code{Fair} \code{Red} \code{Dark}} \item{\code{Eyes}}{eye color, a factor with levels \code{Light} \code{Mixed} \code{Dark}} \item{\code{Head}}{head shape, a factor with levels \code{Narrow} \code{Wide}} \item{\code{Stature}}{height, a factor with levels \code{Tall} \code{Short}} \item{\code{Freq}}{a numeric vector} } } \details{ Burt says: "In all, 217 individuals were examined, about two-thirds of them males. But, partly to simplify the calculations and partly because the later observations were rather more trustworthy, I shall here restrict my analysis to the data obtained from the last hundred males in the series." \code{Head} and \code{Stature} reflect a binary coding where people are classified according to whether they are below or above the average for the population. } \source{ Burt, C. (1950). The factorial analysis of qualitative data, \emph{British Journal of Statistical Psychology}, 3(3), 166-185. Table IX. } %\references{ %% ~~ possibly secondary sources and usages ~~ %} \examples{ data(Burt) mosaic(Freq ~ Hair + Eyes + Head + Stature, data=Burt, shade=TRUE) #or burt.tab <- xtabs(Freq ~ Hair + Eyes + Head + Stature, data=Burt) mosaic(burt.tab, shade=TRUE) } \keyword{datasets} vcdExtra/man/expand.dft.Rd0000644000176200001440000000445713163461153015144 0ustar liggesusers\name{expand.dft} \alias{expand.dft} \alias{expand.table} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Expand a frequency table to case form} \description{ Converts a frequency table, given either as a table object or a data frame in frequency form to a data frame representing individual observations in the table. } \usage{ expand.dft(x, var.names = NULL, freq = "Freq", ...) expand.table(x, var.names = NULL, freq = "Freq", ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{A table object, or a data frame in frequency form containing factors and one numeric variable representing the cell frequency for that combination of factors. } \item{var.names}{A list of variable names for the factors, if you wish to override those already in the table} \item{freq}{The name of the frequency variable in the table} \item{\dots}{Other arguments passed down to \code{type.convert}. In particular, pay attention to \code{na.strings} (default: \code{na.strings=NA} if there are missing cells) and \code{as.is} (default: \code{as.is=FALSE}, converting character vectors to factors).} } \details{ \code{expand.table} is a synonym for \code{expand.dft}. } \value{ A data frame containing the factors in the table and as many observations as are represented by the total of the \code{freq} variable. } \references{ Originally posted on R-Help, Jan 20, 2009, http://tolstoy.newcastle.edu.au/R/e6/help/09/01/1873.html Friendly, M. and Meyer, D. (2016). \emph{Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data}. Boca Raton, FL: Chapman & Hall/CRC. \url{http://ddar.datavis.ca}. } \author{ Mark Schwarz } %\note{ ~~further notes~~ % ~Make other sections like Warning with \section{Warning }{....} ~ %} \seealso{ \code{\link[utils]{type.convert}}, \code{\link[gnm]{expandCategorical}}} \examples{ library(vcd) art <- xtabs(~Treatment + Improved, data = Arthritis) art artdf <- expand.dft(art) str(artdf) # 1D case (tab <- table(sample(head(letters), 20, replace=TRUE))) expand.table(tab, var.names="letter") } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{manip} \keyword{array} % __ONLY ONE__ keyword per line vcdExtra/man/HospVisits.Rd0000644000176200001440000000261613163461153015217 0ustar liggesusers\name{HospVisits} \alias{HospVisits} \docType{data} \title{ Hospital Visits Data } \description{ Length of stay in hospital for 132 schizophrenic patients, classified by visiting patterns, originally from Wing (1962). } \usage{data("HospVisits")} \format{ A 3 by 3 frequency table, with format: table [1:3, 1:3] 43 6 9 16 11 18 3 10 16 - attr(*, "dimnames")=List of 2 ..$ visit: chr [1:3] "Regular" "Infrequent" "Never" ..$ stay : chr [1:3] "2-9" "10-19" "20+" } \details{ Both table variables can be considered ordinal. The variable \code{visit} refers to visiting patterns recorded hospital. The category labels are abbreviations of those given by Goodman (1983); e.g., \code{"Regular"} is short for \dQuote{received visitors regularly or patient went home}. The variable \code{stay} refers to length of stay in hospital, in year groups. } \source{ Goodman, L. A. (1983) The analysis of dependence in cross-classifications having ordered categories, using log-linear models for frequencies and log-linear models for odds. \emph{Biometrics}, 39, 149-160. } \references{ Wing, J. K. (1962). Institutionalism in Mental Hospitals, \emph{British Journal of Social and Clinical Psychology}, 1 (1), 38-51. } \examples{ data(HospVisits) mosaic(HospVisits, gp=shading_Friendly) library(ca) ca(HospVisits) # surprisingly 1D ! plot(ca(HospVisits)) } \keyword{datasets} vcdExtra/man/Summarise.Rd0000644000176200001440000000630713163461153015052 0ustar liggesusers\name{Summarise} \alias{Summarise} \alias{Summarise.glmlist} \alias{Summarise.loglmlist} \alias{Summarise.default} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Brief Summary of Model Fit for glm and loglm Models } \description{ For \code{glm} objects, the \code{print} and \code{summary} methods give too much information if all one wants to see is a brief summary of model goodness of fit, and there is no easy way to display a compact comparison of model goodness of fit for a collection of models fit to the same data. All \code{loglm} models have equivalent glm forms, but the \code{print} and \code{summary} methods give quite different results. \code{Summarise} provides a brief summary for one or more models fit to the same dataset for which \code{logLik} and \code{nobs} methods exist (e.g., \code{glm} and \code{loglm} models). %This implementation is experimental, and is subject to change. } \usage{ Summarise(object, ...) \method{Summarise}{glmlist}(object, ..., saturated = NULL, sortby = NULL) \method{Summarise}{loglmlist}(object, ..., saturated = NULL, sortby = NULL) \method{Summarise}{default}(object, ..., saturated = NULL, sortby = NULL) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{object}{ a fitted model object for which there exists a logLik method to extract the corresponding log-likelihood} \item{\dots}{ optionally more fitted model objects } \item{saturated}{ saturated model log likelihood reference value (use 0 if deviance is not available) } \item{sortby}{ either a numeric or character string specifying the column in the result by which the rows are sorted (in decreasing order)} } \details{ The function relies on residual degrees of freedom for the LR chisq test being available in the model object. This is true for objects inheriting from \code{lm}, \code{glm}, \code{loglm}, \code{polr} and \code{negbin}. } \value{ A data frame (also of class \code{anova}) with columns \code{c("AIC", "BIC", "LR Chisq", "Df", "Pr(>Chisq)")}. Row names are taken from the names of the model object(s). } %\references{ %% ~put references to the literature/web site here ~ %} \author{ Achim Zeileis } %\note{ %% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[stats]{logLik}}, \code{\link[stats]{glm}}, \code{\link[MASS]{loglm}}, \code{\link{logLik.loglm}}, \code{\link{modFit}} } \examples{ data(Mental) indep <- glm(Freq ~ mental+ses, family = poisson, data = Mental) Summarise(indep) Cscore <- as.numeric(Mental$ses) Rscore <- as.numeric(Mental$mental) coleff <- glm(Freq ~ mental + ses + Rscore:ses, family = poisson, data = Mental) roweff <- glm(Freq ~ mental + ses + mental:Cscore, family = poisson, data = Mental) linlin <- glm(Freq ~ mental + ses + Rscore:Cscore, family = poisson, data = Mental) # compare models Summarise(indep, coleff, roweff, linlin) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{models} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/ShakeWords.Rd0000644000176200001440000000317213163461153015154 0ustar liggesusers\name{ShakeWords} \alias{ShakeWords} \docType{data} \title{ Shakespeare's Word Type Frequencies } \description{ This data set, from Efron and Thisted (1976), gives the number of distinct words types (\code{Freq}) of words that appeared exactly once, twice, etc. up to 100 times (\code{count}) in the complete works of Shakespeare. In these works, Shakespeare used 31,534 distinct words (types), comprising 884,647 words in total. Efron & Thisted used this data to ask the question, "How many words did Shakespeare know?" Put another way, suppose another new corpus of works Shakespeare were discovered, also with 884,647 words. How many new word types would appear? The answer to the main question involves contemplating an infinite number of such new corpora. } \usage{data(ShakeWords)} \format{ A data frame with 100 observations on the following 2 variables. \describe{ \item{\code{count}}{the number of times a word type appeared in Shakespeare's written works} \item{\code{Freq}}{the number of different words (types) appearing with this count.} } } \details{ In addition to the words that appear \code{1:100} times, there are 846 words that appear more than 100 times, not listed in this data set. } \source{ Bradley Efron and Ronald Thisted (1976). Estimating the Number of Unsen Species: How Many Words Did Shakespeare Know? \emph{Biometrika}, Vol. 63, No. 3, pp. 435-447, %\url{http://www.jstor.org/stable/2335721} } %\references{ %% ~~ possibly secondary sources and usages ~~ %} \examples{ data(ShakeWords) ## maybe str(ShakeWords) ; plot(ShakeWords) ... } \keyword{datasets} vcdExtra/man/GSS.Rd0000644000176200001440000000174513163461153013542 0ustar liggesusers\name{GSS} \alias{GSS} \docType{data} \title{General Social Survey-- Sex and Party affiliation} \description{ Data from the General Social Survey, 1991, on the relation between sex and party affiliation. } \usage{data(GSS)} \format{ A data frame in frequency form with 6 observations on the following 3 variables. \describe{ \item{\code{sex}}{a factor with levels \code{female} \code{male}} \item{\code{party}}{a factor with levels \code{dem} \code{indep} \code{rep}} \item{\code{count}}{a numeric vector} } } %\details{ % ~~ If necessary, more details than the __description__ above ~~ %} \source{ Agresti, A. Categorical Data Analysis John Wiley & Sons, 2002, Table 3.11, p. 106. } %\references{ % ~~ possibly secondary sources and usages ~~ %} \examples{ data(GSS) ## maybe str(GSS) ; plot(GSS) ... (GSStab <- xtabs(count ~ sex + party, data=GSS)) mod.glm <- glm(count ~ sex + party, family = poisson, data = GSS) } \keyword{datasets} vcdExtra/man/ICU.Rd0000644000176200001440000001200113163461153013511 0ustar liggesusers\name{ICU} \alias{ICU} \docType{data} \title{ ICU data set } \description{ The ICU data set consists of a sample of 200 subjects who were part of a much larger study on survival of patients following admission to an adult intensive care unit (ICU), derived from Hosmer, Lemeshow and Sturdivant (2013) and Friendly (2000). The major goal of this study was to develop a logistic regression model to predict the probability of survival to hospital discharge of these patients and to study the risk factors associated with ICU mortality. The clinical details of the study are described in Lemeshow, Teres, Avrunin, and Pastides (1988). This data set is often used to illustrate model selection methods for logistic regression. } \usage{data(ICU)} \format{ A data frame with 200 observations on the following 22 variables. \describe{ % \item{\code{id}}{Patient id code, a numeric vector} \item{\code{died}}{Died before discharge?, a factor with levels \code{No} \code{Yes}} \item{\code{age}}{Patient age, a numeric vector} \item{\code{sex}}{Patient sex, a factor with levels \code{Female} \code{Male}} \item{\code{race}}{Patient race, a factor with levels \code{Black} \code{Other} \code{White}. Also represented here as \code{white}.} \item{\code{service}}{Service at ICU Admission, a factor with levels \code{Medical} \code{Surgical}} \item{\code{cancer}}{Cancer part of present problem?, a factor with levels \code{No} \code{Yes}} \item{\code{renal}}{History of chronic renal failure?, a factor with levels \code{No} \code{Yes}} \item{\code{infect}}{Infection probable at ICU admission?, a factor with levels \code{No} \code{Yes}} \item{\code{cpr}}{Patient received CPR prior to ICU admission?, a factor with levels \code{No} \code{Yes}} \item{\code{systolic}}{Systolic blood pressure at admission (mm Hg), a numeric vector} \item{\code{hrtrate}}{Heart rate at ICU Admission (beats/min), a numeric vector} \item{\code{previcu}}{Previous admission to an ICU within 6 Months?, a factor with levels \code{No} \code{Yes}} \item{\code{admit}}{Type of admission, a factor with levels \code{Elective} \code{Emergency}} \item{\code{fracture}}{Admission with a long bone, multiple, neck, single area, or hip fracture? a factor with levels \code{No} \code{Yes}} \item{\code{po2}}{PO2 from inital blood gases, a factor with levels \code{>60} \code{<=60}} \item{\code{ph}}{pH from inital blood gases, a factor with levels \code{>=7.25} \code{<7.25}} \item{\code{pco}}{PCO2 from inital blood gases, a factor with levels \code{<=45} \code{>45}} \item{\code{bic}}{Bicarbonate (HCO3) level from inital blood gases, a factor with levels \code{>=18} \code{<18}} \item{\code{creatin}}{Creatinine, from inital blood gases, a factor with levels \code{<=2} \code{>2}} \item{\code{coma}}{Level of unconsciousness at admission to ICU, a factor with levels \code{None} \code{Stupor} \code{Coma}} \item{\code{white}}{a recoding of \code{race}, a factor with levels \code{White} \code{Non-white}} \item{\code{uncons}}{a recoding of \code{coma} a factor with levels \code{No} \code{Yes}} } } \details{ Patient ID numbers are the rownames of the data frame. Note that the last two variables \code{white} and \code{uncons} are a recoding of respectively \code{race} and \code{coma} to binary variables. } \source{ M. Friendly (2000), \emph{Visualizing Categorical Data}, Appendix B.4. SAS Institute, Cary, NC. Hosmer, D. W. Jr., Lemeshow, S. and Sturdivant, R. X. (2013) \emph{Applied Logistic Regression}, NY: Wiley, Third Edition. } \references{ Lemeshow, S., Teres, D., Avrunin, J. S., Pastides, H. (1988). Predicting the Outcome of Intensive Care Unit Patients. \emph{Journal of the American Statistical Association}, 83, 348-356. } \examples{ data(ICU) # remove redundant variables (race, coma) ICU1 <- ICU[,-c(4,20)] # fit full model icu.full <- glm(died ~ ., data=ICU1, family=binomial) summary(icu.full) # simpler model (found from a "best" subsets procedure) icu.mod1 <- glm(died ~ age + sex + cancer + systolic + admit + uncons, data=ICU1, family=binomial) summary(icu.mod1) # even simpler model icu.mod2 <- glm(died ~ age + cancer + admit + uncons, data=ICU1, family=binomial) summary(icu.mod2) anova(icu.mod2, icu.mod1, icu.full, test="Chisq") ## Reproduce Fig 6.12 from VCD icu.fit <- data.frame(ICU, prob=predict(icu.mod2, type="response")) # combine categorical risk factors to a single string risks <- ICU[, c("cancer", "admit", "uncons")] risks[,1] <- ifelse(risks[,1]=="Yes", "Cancer", "") risks[,2] <- ifelse(risks[,2]=="Emergency", "Emerg", "") risks[,3] <- ifelse(risks[,3]=="Yes", "Uncons", "") risks <- apply(risks, 1, paste, collapse="") risks[risks==""] <- "(none)" icu.fit$risks <- risks library(ggplot2) ggplot(icu.fit, aes(x=age, y=prob, color=risks)) + geom_point(size=2) + geom_line(size=1.25, alpha=0.5) + theme_bw() + ylab("Probability of death") } \keyword{datasets} vcdExtra/man/LRstats.Rd0000644000176200001440000000625713163461153014505 0ustar liggesusers\name{LRstats} \alias{LRstats} \alias{LRstats.glmlist} \alias{LRstats.loglmlist} \alias{LRstats.default} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Brief Summary of Model Fit for glm and loglm Models } \description{ For \code{glm} objects, the \code{print} and \code{summary} methods give too much information if all one wants to see is a brief summary of model goodness of fit, and there is no easy way to display a compact comparison of model goodness of fit for a collection of models fit to the same data. All \code{loglm} models have equivalent glm forms, but the \code{print} and \code{summary} methods give quite different results. \code{LRstats} provides a brief summary for one or more models fit to the same dataset for which \code{logLik} and \code{nobs} methods exist (e.g., \code{glm} and \code{loglm} models). %This implementation is experimental, and is subject to change. } \usage{ LRstats(object, ...) \method{LRstats}{glmlist}(object, ..., saturated = NULL, sortby = NULL) \method{LRstats}{loglmlist}(object, ..., saturated = NULL, sortby = NULL) \method{LRstats}{default}(object, ..., saturated = NULL, sortby = NULL) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{object}{ a fitted model object for which there exists a logLik method to extract the corresponding log-likelihood} \item{\dots}{ optionally more fitted model objects } \item{saturated}{ saturated model log likelihood reference value (use 0 if deviance is not available) } \item{sortby}{ either a numeric or character string specifying the column in the result by which the rows are sorted (in decreasing order)} } \details{ The function relies on residual degrees of freedom for the LR chisq test being available in the model object. This is true for objects inheriting from \code{lm}, \code{glm}, \code{loglm}, \code{polr} and \code{negbin}. } \value{ A data frame (also of class \code{anova}) with columns \code{c("AIC", "BIC", "LR Chisq", "Df", "Pr(>Chisq)")}. Row names are taken from the names of the model object(s). } %\references{ %% ~put references to the literature/web site here ~ %} \author{ Achim Zeileis } %\note{ %% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[stats]{logLik}}, \code{\link[stats]{glm}}, \code{\link[MASS]{loglm}}, \code{\link{logLik.loglm}}, \code{\link{modFit}} } \examples{ data(Mental) indep <- glm(Freq ~ mental+ses, family = poisson, data = Mental) LRstats(indep) Cscore <- as.numeric(Mental$ses) Rscore <- as.numeric(Mental$mental) coleff <- glm(Freq ~ mental + ses + Rscore:ses, family = poisson, data = Mental) roweff <- glm(Freq ~ mental + ses + mental:Cscore, family = poisson, data = Mental) linlin <- glm(Freq ~ mental + ses + Rscore:Cscore, family = poisson, data = Mental) # compare models LRstats(indep, coleff, roweff, linlin) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{models} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Titanicp.Rd0000644000176200001440000000404013163461153014650 0ustar liggesusers\name{Titanicp} \alias{Titanicp} \docType{data} \title{ Passengers on the Titanic } \description{ Data on passengers on the RMS Titanic, excluding the Crew and some individual identifier variables. } \usage{data(Titanicp)} \format{ A data frame with 1309 observations on the following 6 variables. \describe{ \item{\code{pclass}}{a factor with levels \code{1st} \code{2nd} \code{3rd}} \item{\code{survived}}{a factor with levels \code{died} \code{survived}} \item{\code{sex}}{a factor with levels \code{female} \code{male}} \item{\code{age}}{passenger age in years (or fractions of a year, for children), a numeric vector; age is missing for 263 of the passengers} \item{\code{sibsp}}{number of siblings or spouses aboard, integer: \code{0:8}} \item{\code{parch}}{number of parents or children aboard, integer: \code{0:6}} } } \details{ There are a number of related versions of the Titanic data, in various formats. This version was derived from \code{ptitanic} in the \pkg{rpart.plot} package, modifying it to remove the \code{Class 'labelled'} attributes for some variables (inherited from Frank Harrell's \code{titanic3} version) which caused problems with some applications, notably \code{ggplot2}. Other versions: \code{\link[datasets]{Titanic}} is the 4-way frequency table of all 2201 people aboard the Titanic, including passengers and crew. } \source{ The original R source for this dataset was compiled by Frank Harrell and Robert Dawson: \url{http://biostat.mc.vanderbilt.edu/twiki/pub/Main/DataSets/titanic.html}, described in more detail in \url{http://biostat.mc.vanderbilt.edu/twiki/pub/Main/DataSets/titanic3info.txt} For this version of the Titanic data, passenger details were deleted, survived was cast as a factor, and the name changed to \code{Titanicp} to minimize confusion with other versions. } %\references{ %%% ~~ possibly secondary sources and usages ~~ %} \examples{ data(Titanicp) ## maybe str(Titanicp) ; plot(Titanicp) ... } \keyword{datasets} vcdExtra/man/modFit.Rd0000644000176200001440000000373313163461153014327 0ustar liggesusers\name{modFit} \Rdversion{1.1} \alias{modFit} \alias{modFit.loglm} \alias{modFit.glm} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Brief Summary of Model Fit for a glm or loglm Object } \description{ Formats a brief summary of model fit for a \code{glm} or \code{loglm} object, showing the likelihood ratio Chisq (df) value and or AIC. Useful for inclusion in a plot title or annotation. } \usage{ modFit(x, ...) \method{modFit}{glm}(x, stats="chisq", digits=2, ...) \method{modFit}{loglm}(x, stats="chisq", digits=2, ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{ A \code{glm} or \code{loglm} object } \item{\dots}{ Arguments passed down } \item{stats}{ One or more of \code{chisq} or \code{aic}, determining the statistics displayed. } \item{digits}{ Number of digits after the decimal point in displayed statistics. } } %\details{ %%% ~~ If necessary, more details than the description above ~~ %} \value{ A character string containing the formatted values of the chosen statistics. %% ~Describe the value returned %% If it is a LIST, use %% \item{comp1 }{Description of 'comp1'} %% \item{comp2 }{Description of 'comp2'} %% ... } %\references{ %%% ~put references to the literature/web site here ~ %} \author{ Michael Friendly } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link{Summarise}} (soon to be deprecated), \code{\link{LRstats}} } \examples{ data(Mental) require(MASS) (Mental.tab <- xtabs(Freq ~ ses+mental, data=Mental)) (Mental.mod <- loglm(~ses+mental, Mental.tab)) Mental.mod modFit(Mental.mod) # use to label mosaic() mosaic(Mental.mod, main=paste("Independence model,", modFit(Mental.mod))) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{utilities} \keyword{models}% __ONLY ONE__ keyword per line vcdExtra/man/Yamaguchi87.Rd0000644000176200001440000000756513163461153015202 0ustar liggesusers\name{Yamaguchi87} \alias{Yamaguchi87} \docType{data} \title{ Occupational Mobility in Three Countries } \description{ Yamaguchi (1987) presented this three-way frequency table, cross-classifying occupational categories of sons and fathers in the United States, United Kingdom and Japan. This data set has become a classic for models comparing two-way mobility tables across layers corresponding to countries, groups or time (e.g., Goodman and Hout, 1998; Xie, 1992). The US data were derived from the 1973 OCG-II survey; those for the UK from the 1972 Oxford Social Mobility Survey; those for Japan came from the 1975 Social Stratification and Mobility survey. They pertain to men aged 20-64. } \usage{data(Yamaguchi87)} \format{ A frequency data frame with 75 observations on the following 4 variables. The total sample size is 28887. \describe{ \item{\code{Son}}{a factor with levels \code{UpNM} \code{LoNM} \code{UpM} \code{LoM} \code{Farm}} \item{\code{Father}}{a factor with levels \code{UpNM} \code{LoNM} \code{UpM} \code{LoM} \code{Farm}} \item{\code{Country}}{a factor with levels \code{US} \code{UK} \code{Japan}} \item{\code{Freq}}{a numeric vector} } } \details{ Five status categories -- upper and lower nonmanuals (\code{UpNM}, \code{LoNM}), upper and lower manuals (\code{UpM}, \code{LoM}), and \code{Farm}) are used for both fathers' occupations and sons' occupations. Upper nonmanuals are professionals, managers, and officials; lower nonmanuals are proprietors, sales workers, and clerical workers; upper manuals are skilled workers; lower manuals are semi-skilled and unskilled nonfarm workers; and farm workers are farmers and farm laborers. Some of the models from Xie (1992), Table 1, are fit in \code{demo(yamaguchi-xie)}. } \source{ Yamaguchi, K. (1987). Models for comparing mobility tables: toward parsimony and substance, \emph{American Sociological Review}, vol. 52 (Aug.), 482-494, Table 1 } \references{ Goodman, L. A. and Hout, M. (1998). Statistical Methods and Graphical Displays for Analyzing How the Association Between Two Qualitative Variables Differs Among Countries, Among Groups, Or Over Time: A Modified Regression-Type Approach. \emph{Sociological Methodology}, 28 (1), 175-230. Xie, Yu (1992). The log-multiplicative layer effect model for comparing mobility tables. \emph{American Sociological Review}, 57 (June), 380-395. } \examples{ data(Yamaguchi87) # reproduce Table 1 structable(~ Father + Son + Country, Yamaguchi87) # create table form Yama.tab <- xtabs(Freq ~ Son + Father + Country, data=Yamaguchi87) # define mosaic labeling_args for convenient reuse in 3-way displays largs <- list(rot_labels=c(right=0), offset_varnames = c(right = 0.6), offset_labels = c(right = 0.2), set_varnames = c(Son="Son's status", Father="Father's status") ) ################################### # Fit some models & display mosaics # Mutual independence yama.indep <- glm(Freq ~ Son + Father + Country, data=Yamaguchi87, family=poisson) anova(yama.indep) mosaic(yama.indep, ~Son+Father, main="[S][F] ignoring country") mosaic(yama.indep, ~Country + Son + Father, condvars="Country", labeling_args=largs, main='[S][F][C] Mutual independence') # no association between S and F given country ('perfect mobility') # asserts same associations for all countries yama.noRC <- glm(Freq ~ (Son + Father) * Country, data=Yamaguchi87, family=poisson) anova(yama.noRC) mosaic(yama.noRC, ~~Country + Son + Father, condvars="Country", labeling_args=largs, main="[SC][FC] No [SF] (perfect mobility)") # ignore diagonal cells yama.quasi <- update(yama.noRC, ~ . + Diag(Son,Father):Country) anova(yama.quasi) mosaic(yama.quasi, ~Son+Father, main="Quasi [S][F]") ## see also: # demo(yamaguchi-xie) ## } \keyword{datasets} vcdExtra/man/Crossings.Rd0000644000176200001440000000366613163461153015064 0ustar liggesusers\name{Crossings} \alias{Crossings} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Crossings Interaction of Factors } \description{ Given two ordered factors in a square, n x n frequency table, \code{Crossings} creates an n-1 column matrix corresponding to different degrees of difficulty in crossing from one level to the next, as described by Goodman (1972). } \usage{ Crossings(...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{\dots}{ Two factors } } %\details{ %% ~~ If necessary, more details than the description above ~~ %} \value{ For two factors of \code{n} levels, returns a binary indicator matrix of \code{n*n} rows and \code{n-1} columns. } \references{ Goodman, L. (1972). Some multiplicative models for the analysis of cross-classified data. In: \emph{Proceedings of the Sixth Berkeley Symposium on Mathematical Statistics and Probability}, Berkeley, CA: University of California Press, pp. 649-696. } \author{ Michael Friendly and Heather Turner } %\note{ %% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[stats]{glm}}, \code{\link[gnm]{gnm}} for model fitting functions for frequency tables \code{\link[gnm]{Diag}}, \code{\link[gnm]{Mult}}, \code{\link[gnm]{Symm}}, \code{\link[gnm]{Topo}} for similar extensions to terms in model formulas. } \examples{ data(Hauser79) # display table structable(~Father+Son, data=Hauser79) hauser.indep <- gnm(Freq ~ Father + Son, data=Hauser79, family=poisson) hauser.CR <- update(hauser.indep, ~ . + Crossings(Father,Son)) LRstats(hauser.CR) hauser.CRdiag <- update(hauser.indep, ~ . + Crossings(Father,Son) + Diag(Father,Son)) LRstats(hauser.CRdiag) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{models} \keyword{manip}% __ONLY ONE__ keyword per line vcdExtra/man/Vote1980.Rd0000644000176200001440000000303013163461153014332 0ustar liggesusers\name{Vote1980} \alias{Vote1980} \docType{data} \title{ Race and Politics in the 1980 Presidential Vote } \description{ Data from the 1982 General Social Survey on votes in the 1980 U.S. presidential election in relation to race and political conservatism. } \usage{data(Vote1980)} \format{ A frequency data frame representing a 2 x 7 x 2 table, with 28 observations on the following 4 variables. \describe{ \item{\code{race}}{a factor with levels \code{NonWhite} \code{White}} \item{\code{conservatism}}{ a factor with levels \code{1} \code{2} \code{3} \code{4} \code{5} \code{6} \code{7}, \code{1}=most liberal, \code{7}=most conservative} \item{\code{votefor}}{a factor with levels \code{Carter} \code{Reagan}; \code{Carter} represents Jimmy Carter or other.} \item{\code{Freq}}{a numeric vector} } } \details{ The data contains a number of sampling zeros in the frequencies of NonWhites voting for Ronald Reagan. } \source{ Clogg, C. & Shockey, J. W. (1988). In Nesselroade, J. R. & Cattell, R. B. (ed.) Multivariate Analysis of Discrete Data, \emph{Handbook of Multivariate Experimental Psychology}, New York: Plenum Press. } \references{ Agresti, A. (1990) \emph{Categorical Data Analysis}, Table 4.12 New York: Wiley-Interscience. Friendly, M. (2000) \emph{Visualizing Categorical Data}, Example 7.5 Cary, NC: SAS Institute. } \examples{ data(Vote1980) fourfold(xtabs(Freq ~ race + votefor + conservatism, data=Vote1980), mfrow=c(2,4)) } \keyword{datasets} vcdExtra/man/loglin-utilities.Rd0000644000176200001440000001615413163461153016403 0ustar liggesusers\name{loglin-utilities} \alias{loglin-utilities} \alias{conditional} \alias{joint} \alias{loglin2formula} \alias{loglin2string} \alias{markov} \alias{mutual} \alias{saturated} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Loglinear Model Utilities } \description{ These functions generate lists of terms to specify a loglinear model in a form compatible with \code{\link[stats]{loglin}} and also provide for conversion to an equivalent \code{\link[MASS]{loglm}} specification or a shorthand character string representation. They allow for a more conceptual way to specify such models by a function for their type, as opposed to just an uninterpreted list of model terms and also allow easy specification of marginal models for a given contingency table. They are intended to be used as tools in higher-level modeling and graphics functions, but can also be used directly. } \usage{ conditional(nf, table = NULL, factors = 1:nf, with = nf) joint(nf, table = NULL, factors = 1:nf, with = nf) markov(nf, factors = 1:nf, order = 1) mutual(nf, table = NULL, factors = 1:nf) saturated(nf, table = NULL, factors = 1:nf) loglin2formula(x, env = parent.frame()) loglin2string(x, brackets = c("[", "]"), sep = ",", collapse = " ", abbrev) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{nf}{ number of factors for which to generate the model } \item{table}{ a contingency table used only for factor names in the model, typically the output from \code{\link[base]{table}} and possibly permuted with \code{aperm} } \item{factors}{ names of factors used in the model formula when \code{table} is not specified } \item{with}{ For \code{joint} and \code{conditional} models, \code{with} gives the indices of the factors against which all others are considered jointly or conditionally independent } \item{order}{ For \code{markov}, this gives the order of the Markov chain model for the factors. An \code{order=1} Markov chain allows associations among sequential pairs of factors, e.g., \code{[A,B], [B,C], [C,D]} \dots. An \code{order=2} Markov chain allows associations among sequential triples. } \item{x}{ For the \code{loglin2*} functions, a list of terms in a loglinear model, such as returned by \code{conditional}, \code{joint}, \dots } \item{env}{ For \code{loglin2formula}, environment in which to evaluate the formula } \item{brackets}{ For \code{loglin2string}, characters to use to surround model terms. Either a single character string containing two characters (e.g., \code{'[]'} or a character vector of length two. } \item{sep}{ For \code{loglin2string}, the separator character string used for factor names within a given model term } \item{collapse}{ For \code{loglin2string}, the character string used between terms in the the model string } \item{abbrev}{ For \code{loglin2string}, whether and how to abbreviate the terms in the string representation. This has not yet been implemented. } } \details{ The main model specification functions, \code{conditional}, \code{joint}, \code{markov}, \dots, \code{saturated}, return a list of vectors indicating the marginal totals to be fit, via the \code{margin} argument to \code{\link[stats]{loglin}}. Each element of this list corresponds to a high-order term in a hierarchical loglinear model, where, e.g., a term like \code{c("A", "B")} is equivalent to the \code{\link[MASS]{loglm}} term \code{"A:B"} and hence automatically includes all low-order terms. Note that these can be used to supply the \code{expected} argument for the default \code{\link[vcd]{mosaic}} function, when the data is supplied as a contingency table. The table below shows some typical results in terms of the standard shorthand notation for loglinear models, with factors A, B, C, \dots, where brackets are used to delimit the high-order terms in the loglinear model. \tabular{llll}{ \strong{function} \tab \strong{3-way} \tab \strong{4-way} \tab \strong{5-way} \cr \code{mutual} \tab [A] [B] [C] \tab [A] [B] [C] [D] \tab [A] [B] [C] [D] [E] \cr \code{joint} \tab [AB] [C] \tab [ABC] [D] \tab [ABCE] [E] \cr \code{joint (with=1)} \tab [A] [BC] \tab [A] [BCD] \tab [A] [BCDE] \cr \code{conditional} \tab [AC] [BC] \tab [AD] [BD] [CD] \tab [AE] [BE] [CE] [DE] \cr \code{condit (with=1)} \tab [AB] [AC] \tab [AB] [AC] [AD] \tab [AB] [AC] [AD] [AE] \cr \code{markov (order=1)} \tab [AB] [BC] \tab [AB] [BC] [CD] \tab [AB] [BC] [CD] [DE] \cr \code{markov (order=2)} \tab [A] [B] [C] \tab [ABC] [BCD] \tab [ABC] [BCD] [CDE] \cr \code{saturated} \tab [ABC] \tab [ABCD] \tab [ABCDE] \cr } \code{loglin2formula} converts the output of one of these to a model formula suitable as the \code{formula} for of \code{\link[MASS]{loglm}}. \code{loglin2string} converts the output of one of these to a string describing the loglinear model in the shorthand bracket notation, e.g., \code{"[A,B] [A,C]"}. } \value{ For the main model specification functions, \code{conditional}, \code{joint}, \code{markov}, \dots, the result is a list of vectors (terms), where the elements in each vector are the names of the factors. The elements of the list are given names \code{term1, term2, \dots}. } \references{ These functions were inspired by the original SAS implementation of mosaic displays, described in the \emph{User's Guide}, \url{http://www.datavis.ca/mosaics/mosaics.pdf} } \author{ Michael Friendly } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[stats]{loglin}}, \code{\link[MASS]{loglm}} } \examples{ joint(3, table=HairEyeColor) # as a formula or string loglin2formula(joint(3, table=HairEyeColor)) loglin2string(joint(3, table=HairEyeColor)) joint(2, HairEyeColor) # marginal model for [Hair] [Eye] # other possibilities joint(4, factors=letters, with=1) joint(5, factors=LETTERS) joint(5, factors=LETTERS, with=4:5) conditional(4) conditional(4, with=3:4) # use in mosaic displays or other strucplots mosaic(HairEyeColor, expected=joint(3)) mosaic(HairEyeColor, expected=conditional(3)) # use with MASS::loglm cond3 <- loglin2formula(conditional(3, table=HairEyeColor)) cond3 <- loglin2formula(conditional(3)) # same, with factors 1,2,3 require(MASS) loglm(cond3, data=HairEyeColor) saturated(3, HairEyeColor) loglin2formula(saturated(3, HairEyeColor)) loglin2string(saturated(3, HairEyeColor)) loglin2string(saturated(3, HairEyeColor), brackets='{}', sep=', ') } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{models} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Bartlett.Rd0000644000176200001440000000275313163461153014667 0ustar liggesusers\name{Bartlett} \Rdversion{1.1} \alias{Bartlett} \docType{data} \title{Bartlett data on plum root cuttings} \description{In an experiment to investigate the effect of cutting length (two levels) and planting time (two levels) on the survival of plum root cuttings, 240 cuttings were planted for each of the 2 x 2 combinations of these factors, and their survival was later recorded. Bartlett (1935) used these data to illustrate a method for testing for no three-way interaction in a contingency table.} \usage{ data(Bartlett) } \format{ A 3-dimensional array resulting from cross-tabulating 3 variables for 960 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{Alive}\tab \code{"Alive", "Dead"}\cr 2\tab \code{Time}\tab \code{"Now", "Spring"}\cr 3\tab \code{Length}\tab \code{"Long", "Short"}\cr } } %\details { } \source{ % \cite{Hand-etal:94 [p.15 #19]} Hand, D. and Daly, F. and Lunn, A. D.and McConway, K. J. and Ostrowski, E. (1994). \emph{A Handbook of Small Data Sets}. London: Chapman & Hall, p. 15, # 19. } \references{ % \cite{Bartlett:35} Bartlett, M. S. (1935). Contingency Table Interactions \emph{Journal of the Royal Statistical Society}, Supplement, 1935, 2, 248-252. } %\seealso { } \examples{ data(Bartlett) fourfold(Bartlett, mfrow=c(1,2)) mosaic(Bartlett, shade=TRUE) pairs(Bartlett, gp=shading_Friendly) } \keyword{datasets} vcdExtra/man/glmlist.Rd0000644000176200001440000000663213163461153014561 0ustar liggesusers\name{glmlist} \Rdversion{1.1} \alias{glmlist} \alias{loglmlist} \alias{coef.glmlist} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Create a Model List Object } \description{ \code{glmlist} creates a \code{glmlist} object containing a list of fitted \code{glm} objects with their names. \code{loglmlist} does the same for \code{loglm} objects. The intention is to provide object classes to facilitate model comparison, extraction, summary and plotting of model components, etc., perhaps using \code{\link[base]{lapply}} or similar. There exists a \code{\link[stats]{anova.glm}} method for \code{glmlist} objects. Here, a \code{coef} method is also defined, collecting the coefficients from all models in a single object of type determined by \code{result}. } \usage{ glmlist(...) loglmlist(...) \method{coef}{glmlist}(object, result=c("list", "matrix", "data.frame"), ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{\dots}{ One or more model objects, as appropriate to the function, optionally assigned names as in \code{list}. } \item{object}{a \code{glmlist} object} \item{result}{type of the result to be returned} } \details{ The arguments to \code{glmlist} or \code{loglmlist} are of the form \code{value} or \code{name=value}. Any objects which do not inherit the appropriate class \code{glm} or \code{loglm} are excluded, with a warning. In the \code{coef} method, coefficients from the different models are matched by name in the list of unique names across all models. } \value{ An object of class \code{glmlist} \code{loglmlist}, just like a \code{list}, except that each model is given a \code{name} attribute. %% If it is a LIST, use %% \item{comp1 }{Description of 'comp1'} %% \item{comp2 }{Description of 'comp2'} %% ... } %\references{ %%% ~put references to the literature/web site here ~ %} \author{ Michael Friendly; \code{coef} method by John Fox } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ The function \code{\link[Hmisc]{llist}} in package \code{Hmisc} is similar, but perplexingly more general. The function \code{\link[stats]{anova.glm}} also handles \code{glmlist objects} \code{\link{LRstats}} gives LR statistics and tests for a \code{glmlist} object. } \examples{ data(Mental) indep <- glm(Freq ~ mental+ses, family = poisson, data = Mental) Cscore <- as.numeric(Mental$ses) Rscore <- as.numeric(Mental$mental) coleff <- glm(Freq ~ mental + ses + Rscore:ses, family = poisson, data = Mental) roweff <- glm(Freq ~ mental + ses + mental:Cscore, family = poisson, data = Mental) linlin <- glm(Freq ~ mental + ses + Rscore:Cscore, family = poisson, data = Mental) # use object names mods <- glmlist(indep, coleff, roweff, linlin) names(mods) # assign new names mods <- glmlist(Indep=indep, Col=coleff, Row=roweff, LinxLin=linlin) names(mods) LRstats(mods) coef(mods, result='data.frame') #extract model components unlist(lapply(mods, deviance)) res <- lapply(mods, residuals) boxplot(as.data.frame(res), main="Residuals from various models") } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{utilities} \keyword{models}% __ONLY ONE__ keyword per line vcdExtra/man/AirCrash.Rd0000644000176200001440000000414013163461153014572 0ustar liggesusers\name{AirCrash} \alias{AirCrash} \docType{data} \title{ Air Crash Data } \description{ Data on all fatal commercial airplane crashes from 1993--2015. Excludes small planes (less than 6 passengers) and non-commercial (cargo, military, private) aircraft. } \usage{data("AirCrash")} \format{ A data frame with 439 observations on the following 5 variables. \describe{ \item{\code{Phase}}{phase of the flight, a factor with levels \code{en route} \code{landing} \code{standing} \code{take-off} \code{unknown}} \item{\code{Cause}}{a factor with levels \code{criminal} \code{human error} \code{mechanical} \code{unknown} \code{weather}} \item{\code{date}}{date of crash, a Date} \item{\code{Fatalities}}{number of fatalities, a numeric vector} \item{\code{Year}}{year, a numeric vector} } } \details{ \code{Phase} of the flight was cleaned by combining related variants, spelling, etc. } \source{ Originally from David McCandless, \url{http://www.informationisbeautiful.net/visualizations/plane-truth-every-single-commercial-plane-crash-visualized/}, with the data at \url{https://docs.google.com/spreadsheet/ccc?key=0AjOUPqcIwvnjdEx2akx5ZjJXSk9oM1E3dWpqZFJ6Nmc&usp=drive_web#gid=1}, downloaded April 14, 2015. } \references{ Rick Wicklin, \url{http://blogs.sas.com/content/iml/2015/03/30/visualizing-airline-crashes.html} } \examples{ data(AirCrash) aircrash.tab <- xtabs(~Phase + Cause, data=AirCrash) mosaic(aircrash.tab, shade=TRUE) # fix label overlap mosaic(aircrash.tab, shade=TRUE, labeling_args=list(rot_labels=c(30, 30, 30, 30))) # reorder by Phase phase.ord <- rev(c(3,4,1,2,5)) mosaic(aircrash.tab[phase.ord,], shade=TRUE, labeling_args=list(rot_labels=c(30, 30, 30, 30)), offset_varnames=0.5) # reorder by frequency phase.ord <- order(rowSums(aircrash.tab), decreasing=TRUE) cause.ord <- order(colSums(aircrash.tab), decreasing=TRUE) mosaic(aircrash.tab[phase.ord,cause.ord], shade=TRUE, labeling_args=list(rot_labels=c(30, 30, 30, 30))) library(ca) aircrash.ca <- ca(aircrash.tab) plot(aircrash.ca) } \keyword{datasets} vcdExtra/man/DaytonSurvey.Rd0000644000176200001440000000462413163461153015561 0ustar liggesusers\name{DaytonSurvey} \alias{DaytonSurvey} \docType{data} \title{ Dayton Student Survey on Substance Use } \description{ This data, from Agresti (2002), Table 9.1, gives the result of a 1992 survey in Dayton Ohio of 2276 high school seniors on whether they had ever used alcohol, cigarettes and marijuana. } \usage{data(DaytonSurvey)} \format{ A frequency data frame with 32 observations on the following 6 variables. \describe{ \item{\code{cigarette}}{a factor with levels \code{Yes} \code{No}} \item{\code{alcohol}}{a factor with levels \code{Yes} \code{No}} \item{\code{marijuana}}{a factor with levels \code{Yes} \code{No}} \item{\code{sex}}{a factor with levels \code{female} \code{male}} \item{\code{race}}{a factor with levels \code{white} \code{other}} \item{\code{Freq}}{a numeric vector} } } \details{ Agresti uses the letters G (\code{sex}), R (\code{race}), A (\code{alcohol}), C (\code{cigarette}), M (\code{marijuana}) to refer to the table variables, and this usage is followed in the examples below. Background variables include \code{sex} and \code{race} of the respondent (GR), typically treated as explanatory, so that any model for the full table should include the term \code{sex:race}. Models for the reduced table, collapsed over \code{sex} and \code{race} are not entirely unreasonable, but don't permit the estimation of the effects of these variables on the responses. The full 5-way table contains a number of cells with counts of 0 or 1, as well as many cells with large counts, and even the ACM table collapsed over GR has some small cell counts. Consequently, residuals for these models in mosaic displays are best represented as standardized (adjusted) residuals. } \source{ Agresti, A. (2002). \emph{Categorical Data Analysis}, 2nd Ed., New York: Wiley-Interscience, Table 9.1, p. 362. } \references{ Thompson, L. (2009). \emph{R (and S-PLUS) Manual to Accompany Agresti's Categorical Data}, \url{http://www.stat.ufl.edu/~aa/cda/Thompson_manual.pdf} } \examples{ data(DaytonSurvey) mod.GR <- glm(Freq ~ . + sex*race, data=DaytonSurvey, family=poisson) # mutual independence + GR mod.homog.assoc <- glm(Freq ~ .^2, data=DaytonSurvey, family=poisson) # homogeneous association # collapse over sex and race Dayton.ACM <- aggregate(Freq ~ cigarette+alcohol+marijuana, data=DaytonSurvey, FUN=sum) } \keyword{datasets} vcdExtra/man/CyclingDeaths.Rd0000644000176200001440000000311313163461153015616 0ustar liggesusers\name{CyclingDeaths} \alias{CyclingDeaths} \docType{data} \title{ London Cycling Deaths } \description{ A data frame containing the number of deaths of cyclists in London from 2005 through 2012 in each fortnightly period. Aberdein & Spiegelhalter (2013) discuss these data in relation to the observation that six cyclists died in London between Nov. 5 and Nov. 13, 2013. } \usage{data(CyclingDeaths)} \format{ A data frame with 208 observations on the following 2 variables. \describe{ \item{\code{date}}{a Date} \item{\code{deaths}}{number of deaths, a numeric vector} } } %\details{ %%% ~~ If necessary, more details than the __description__ above ~~ %} \source{ \url{http://data.gov.uk/dataset/road-accidents-safety-data}, STATS 19 data, 2005-2012, using the files \code{Casualty0512.csv} and \code{Accidents0512.csv} } \references{ Aberdein, Jody and Spiegelhalter, David (2013). Have London's roads become more dangerous for cyclists? \emph{Significance}, 10(6), 46--48. } \examples{ data(CyclingDeaths) plot(deaths ~ date, data=CyclingDeaths, type="h", lwd=3, ylab="Number of deaths", axes=FALSE) axis(1, at=seq(as.Date('2005-01-01'), by='years', length.out=9), labels=2005:2013) axis(2, at=0:3) # make a one-way frequency table CyclingDeaths.tab <- table(CyclingDeaths$deaths) gf <- goodfit(CyclingDeaths.tab) gf summary(gf) rootogram(gf, xlab="Number of Deaths") distplot(CyclingDeaths.tab) # prob of 6 or more deaths in one fortnight lambda <- gf$par$lambda ppois(5, lambda, lower.tail=FALSE) } \keyword{datasets} vcdExtra/man/datasets.Rd0000644000176200001440000000424213163461153014711 0ustar liggesusers\name{datasets} \alias{datasets} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Information on Data Sets in Packages } \description{ The \code{\link[utils]{data}} function is used both to load data sets from packages, and give a display of the names and titles of data sets in one or more packages, however it does not return a result that can be easily used to get additional information about the nature of data sets in packages. The \code{datasets()} function is designed to produce a more useful summary display of data sets in one or more packages. It extracts the \code{class} and dimension information (\code{dim} or code{length}) of each item, and formats these to provide additional descriptors. } \usage{ datasets(package, allClass=FALSE, incPackage=length(package) > 1, maxTitle=NULL) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{package}{ a character vector giving the package(s) to look in } \item{allClass}{ include all classes of the item (\code{TRUE}) or just the last class (\code{FALSE})? } \item{incPackage}{ include the package name in result? } \item{maxTitle}{ maximum length of data set Title } } \details{ The requested packages must be installed, and are silently loaded in order to extract \code{class} and size information. } \value{ A \code{data.frame} whose rows correspond to data sets found in \code{package}. The columns (for a single package) are: \item{Item}{data set name} \item{class}{class} \item{dim}{an abbreviation of the dimensions of the data set} \item{Title}{data set title} } %\references{ %%% ~put references to the literature/web site here ~ %} \author{ Michael Friendly, with R-help from Curt Seeliger } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[utils]{data}}, } \examples{ datasets("vcdExtra") datasets(c("vcd", "vcdExtra")) datasets("datasets", maxTitle=50) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{package} \keyword{data}% __ONLY ONE__ keyword per line vcdExtra/man/Accident.Rd0000644000176200001440000000711313163464336014621 0ustar liggesusers\name{Accident} \alias{Accident} \docType{data} \title{ Traffic Accident Victims in France in 1958 } \description{ Bertin (1983) used these data to illustrate the cross-classification of data by numerous variables, each of which could have various types and could be assigned to various visual attributes. For modeling and visualization purposes, the data can be treated as a 4-way table using loglinear models and mosaic displays, or as a frequency-weighted data frame using a binomial response for \code{result} (\code{"Died"} vs. \code{"Injured"}) and plots of predicted probabilities. } \usage{data(Accident)} \format{ A data frame in frequency form (comprising a 5 x 2 x 4 x 2 table) with 80 observations on the following 5 variables. \describe{ \item{\code{age}}{an ordered factor with levels \code{0-9} < \code{10-19} < \code{20-29} < \code{30-49} < \code{50+}} \item{\code{result}}{a factor with levels \code{Died} \code{Injured}} \item{\code{mode}}{mode of transportation, a factor with levels \code{4-Wheeled} \code{Bicycle} \code{Motorcycle} \code{Pedestrian}} \item{\code{gender}}{a factor with levels \code{Female} \code{Male}} \item{\code{Freq}}{a numeric vector} } } \details{ \code{age} is an ordered factor, but arguably, \code{mode} should be treated as ordered, with levels \code{Pedestrian} < \code{Bicycle} < \code{Motorcycle} < \code{4-Wheeled} as Bertin does. This affects the parameterization in models, so we don't do this directly in the data frame. } \source{ Bertin (1983), p. 30; original data from the Ministere des Travaux Publics } \references{ Bertin, J. (1983), \emph{Semiology of Graphics}, University of Wisconsin Press. } \examples{ # examples data(Accident) head(Accident) # for graphs, reorder mode Accident$mode <- ordered(Accident$mode, levels=levels(Accident$mode)[c(4,2,3,1)]) # Bertin's table accident_tab <- xtabs(Freq ~ gender+mode+age+result, data=Accident) structable(mode+gender ~ age+result, data=accident_tab) ## Loglinear models ## ---------------- # mutual independence acc.mod0 <- glm(Freq ~ age+result+mode+gender, data=Accident, family=poisson) LRstats(acc.mod0) mosaic(acc.mod0, ~mode+age+gender+result) # result as a response acc.mod1 <- glm(Freq ~ age*mode*gender + result, data=Accident, family=poisson) LRstats(acc.mod1) mosaic(acc.mod1, ~mode+age+gender+result, labeling_args = list(abbreviate = c(gender=1, result=4))) # allow two-way association of result with each explanatory variable acc.mod2 <- glm(Freq ~ age*mode*gender + result*(age+mode+gender), data=Accident, family=poisson) LRstats(acc.mod2) mosaic(acc.mod2, ~mode+age+gender+result, labeling_args = list(abbreviate = c(gender=1, result=4))) acc.mods <- glmlist(acc.mod0, acc.mod1, acc.mod2) LRstats(acc.mods) ## Binomial (logistic regression) models for result ## ------------------------------------------------ library(car) # for Anova() acc.bin1 <- glm(result=='Died' ~ age+mode+gender, weights=Freq, data=Accident, family=binomial) Anova(acc.bin1) acc.bin2 <- glm(result=='Died' ~ (age+mode+gender)^2, weights=Freq, data=Accident, family=binomial) Anova(acc.bin2) acc.bin3 <- glm(result=='Died' ~ (age+mode+gender)^3, weights=Freq, data=Accident, family=binomial) Anova(acc.bin3) # compare models anova(acc.bin1, acc.bin2, acc.bin3, test="Chisq") # visualize probability of death with effect plots \dontrun{ library(effects) plot(allEffects(acc.bin1), ylab='Pr (Died)') plot(allEffects(acc.bin2), ylab='Pr (Died)') } #} \keyword{datasets} vcdExtra/man/Gilby.Rd0000644000176200001440000000342713163461153014153 0ustar liggesusers\name{Gilby} \Rdversion{1.1} \alias{Gilby} \docType{data} \title{Clothing and Intelligence Rating of Children} \description{Schoolboys were classified according to their clothing and to their teachers rating of "dullness" (lack of intelligence), in a 5 x 7 table originally from Gilby (1911). Anscombe (1981) presents a slightly collapsed 4 x 6 table, used here, where the last two categories of clothing were pooled as were the first two categories of dullness due to small counts. Both \code{Dullnes} and \code{Clothing} are ordered categories, so models and methods that examine their association in terms of ordinal categories are profitable. } \usage{ data(Gilby) } \format{ A 2-dimensional array resulting from cross-tabulating 2 variables for 1725 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{Dullness}\tab \code{"Ment. defective", "Slow", "Slow Intell", "Fairly Intell", "Capable", "V.Able"}\cr 2\tab \code{Clothing}\tab \code{"V.Well clad", "Well clad", "Passable", "Insufficient"}\cr } } %\details{ } \source{ Anscombe, F. J. (1981). \emph{Computing in Statistical Science Through APL}. New York: Springer-Verlag, p. 302 } \references{ % \cite{Gilby & Pearson 1911, from Anscombe 1981, p 302} Gilby, W. H. (1911). On the significance of the teacher's appreciation of general intelligence. \emph{Biometrika}, 8, 93-108 (esp. p. 94). [Quoted by Kendall (1943,..., 1953) Table 13.1, p 320.] } %\seealso{ } \examples{ data(Gilby) mosaic(Gilby, shade=TRUE) # correspondence analysis to see relations among categories if(require(ca)){ ca(Gilby) plot(ca(Gilby)) title(xlab="Dimension 1", ylab="Dimension 2") } } \keyword{datasets} vcdExtra/man/Cancer.Rd0000644000176200001440000000211613163461153014272 0ustar liggesusers\name{Cancer} \Rdversion{1.1} \alias{Cancer} \docType{data} \title{Survival of Breast Cancer Patients} \description{Three year survival of 474 breast cancer patients according to nuclear grade and diagnostic center.} \usage{ data(Cancer) } \format{ A 3-dimensional array resulting from cross-tabulating 3 variables for 474 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{Survival}\tab \code{"Died", "Surv"}\cr 2\tab \code{Grade}\tab \code{"Malignant", "Benign"}\cr 3\tab \code{Center}\tab \code{"Boston", "Glamorgan"}\cr } } %\details { } \source{ % \cite{Lindsey:95 [p38]} % \cite{Whittaker:90} Lindsey, J. K. (1995). Analysis of Frequency and Count Data Oxford, UK: Oxford University Press. p. 38, Table 2.5. Whittaker, J. (1990) Graphical Models in Applied Multivariate Statistics New York: John Wiley and Sons, p. 220. } %\references{ % \cite{Morrison etal} %} %\seealso { } \examples{ data(Cancer) # example goes here } \keyword{datasets} vcdExtra/man/collapse.table.Rd0000644000176200001440000000720413163461153015772 0ustar liggesusers\name{collapse.table} \alias{collapse.table} %- Also NEED an '\alias' for EACH other topic documented here. \title{Collapse Levels of a Table} \description{ Collapse (or re-label) variables in a a contingency table, array or \code{ftable} object by re-assigning levels of the table variables. } \usage{ collapse.table(table, ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{table}{A \code{\link[base]{table}}, \code{\link[base]{array}} or \code{\link[stats]{ftable}} object} \item{\dots}{ A collection of one or more assignments of factors of the table to a list of levels } } \details{ Each of the \code{\dots} arguments must be of the form \code{variable = levels}, where \code{variable} is the name of one of the table dimensions, and \code{levels} is a character or numeric vector of length equal to the corresponding dimension of the table. } \value{ A \code{xtabs} and \code{table} object, representing the original table with one or more of its factors collapsed or rearranged into other levels. } %\references{ ~put references to the literature/web site here ~ } \author{Michael Friendly} %\note{ ~~further notes~~ % % ~Make other sections like Warning with \section{Warning }{....} ~ %} \seealso{ \code{\link{expand.dft}} expands a frequency data frame to case form. \code{\link[base]{margin.table}} "collapses" a table in a different way, by summing over table dimensions. } \examples{ # create some sample data in table form sex <- c("Male", "Female") age <- letters[1:6] education <- c("low", 'med', 'high') data <- expand.grid(sex=sex, age=age, education=education) counts <- rpois(36, 100) data <- cbind(data, counts) t1 <- xtabs(counts ~ sex + age + education, data=data) structable(t1) ## age a b c d e f ## sex education ## Male low 119 101 109 85 99 93 ## med 94 98 103 108 84 84 ## high 81 88 96 110 100 92 ## Female low 107 104 95 86 103 96 ## med 104 98 94 95 110 106 ## high 93 85 90 109 99 86 # collapse age to 3 levels t2 <- collapse.table(t1, age=c("A", "A", "B", "B", "C", "C")) structable(t2) ## age A B C ## sex education ## Male low 220 194 192 ## med 192 211 168 ## high 169 206 192 ## Female low 211 181 199 ## med 202 189 216 ## high 178 199 185 # collapse age to 3 levels and pool education: "low" and "med" to "low" t3 <- collapse.table(t1, age=c("A", "A", "B", "B", "C", "C"), education=c("low", "low", "high")) structable(t3) ## age A B C ## sex education ## Male low 412 405 360 ## high 169 206 192 ## Female low 413 370 415 ## high 178 199 185 # change labels for levels of education to 1:3 t4 <- collapse.table(t1, education=1:3) structable(t4) structable(t4) ## age a b c d e f ## sex education ## Male 1 119 101 109 85 99 93 ## 2 94 98 103 108 84 84 ## 3 81 88 96 110 100 92 ## Female 1 107 104 95 86 103 96 ## 2 104 98 94 95 110 106 ## 3 93 85 90 109 99 86 } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{manip} \keyword{attribute}% __ONLY ONE__ keyword per line vcdExtra/man/Draft1970table.Rd0000644000176200001440000000527113163461153015475 0ustar liggesusers\name{Draft1970table} \alias{Draft1970table} \docType{data} \title{ USA 1970 Draft Lottery Table } \description{ This data set gives the results of the 1970 US draft lottery, in the form of a frequency table. The rows are months of the year, Jan--Dec and columns give the number of days in that month which fall into each of three draft risk categories High, Medium, and Low, corresponding to the chances of being called to serve in the US army. } \usage{data(Draft1970table)} \format{ The format is: 'table' int [1:12, 1:3] 9 7 5 8 9 11 12 13 10 9 ... - attr(*, "dimnames")=List of 2 ..$ Month: chr [1:12] "Jan" "Feb" "Mar" "Apr" ... ..$ Risk : chr [1:3] "High" "Med" "Low" } \details{ The lottery numbers are divided into three categories of risk of being called for the draft -- High, Medium, and Low -- each representing roughly one third of the days in a year. Those birthdays having the highest risk have lottery numbers 1-122, medium risk have numbers 123-244, and the lowest risk category contains lottery numbers 245-366. } \source{ This data is available in several forms, but the table version was obtained from \url{http://sas.uwaterloo.ca/~rwoldfor/software/eikosograms/data/draft-70} } \references{ Fienberg, S. E. (1971), "Randomization and Social Affairs: The 1970 Draft Lottery," \emph{Science}, 171, 255-261. Starr, N. (1997). Nonrandom Risk: The 1970 Draft Lottery, \emph{Journal of Statistics Education}, v.5, n.2 \url{http://www.amstat.org/publications/jse/v5n2/datasets.starr.html} } \seealso{\code{\link{Draft1970}} } \examples{ data(Draft1970table) chisq.test(Draft1970table) # plot.table -> graphics:::mosaicplot plot(Draft1970table, shade=TRUE) mosaic(Draft1970table, gp=shading_Friendly) # correspondence analysis if(require(ca)) { ca(Draft1970table) plot(ca(Draft1970table)) } # convert to a frequency data frame with ordered factors Draft1970df <- as.data.frame(Draft1970table) Draft1970df <- within(Draft1970df, { Month <- ordered(Month) Risk <- ordered(Risk, levels=rev(levels(Risk))) }) str(Draft1970df) # similar model, as a Poisson GLM indep <- glm(Freq ~ Month + Risk, family = poisson, data = Draft1970df) mosaic(indep, residuals_type="rstandard", gp=shading_Friendly) # numeric scores for tests of ordinal factors Cscore <- as.numeric(Draft1970df$Risk) Rscore <- as.numeric(Draft1970df$Month) # linear x linear association between Month and Risk linlin <- glm(Freq ~ Month + Risk + Rscore:Cscore, family = poisson, data = Draft1970df) # compare models anova(indep, linlin, test="Chisq") mosaic(linlin, residuals_type="rstandard", gp=shading_Friendly) } \keyword{datasets} vcdExtra/man/print.Kappa.Rd0000644000176200001440000000235113163461153015267 0ustar liggesusers\name{print.Kappa} \alias{print.Kappa} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Print Kappa } \description{ This is a replacement for the \code{print.Kappa} method in \code{vcd}, adding display of \code{z} values to the \code{vcd} version and optional confidence intervals. } \usage{ \method{print}{Kappa}(x, digits=max(getOption("digits") - 3, 3), CI=FALSE, level=0.95, ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{ A Kappa object} \item{digits}{number of digits to print} \item{CI}{Include confidence intervals in the display?} \item{level}{confidence level} \item{\dots}{ Other arguments } } %\details{ % ~~ If necessary, more details than the description above ~~ %} \value{ Returns the Kappa object, invisibly. } %\references{ ~put references to the literature/web site here ~ } \author{ Michael Friendly} \seealso{ \code{\link[vcd]{confint.Kappa}} } \examples{ data("SexualFun") Kappa(SexualFun) print(Kappa(SexualFun), CI=TRUE) # stratified 3-way table apply(MSPatients, 3, Kappa) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{htest} \keyword{category} vcdExtra/man/Depends.Rd0000644000176200001440000000314113163461153014460 0ustar liggesusers\name{Depends} \alias{Depends} \docType{data} \title{ Dependencies of R Packages } \description{ This one-way table gives the type-token distribution of the number of dependencies declared in 4983 packages listed on CRAN on January 17, 2014. } \usage{data(Depends)} \format{ The format is: 'table' int [1:15(1d)] 986 1347 993 685 375 298 155 65 32 19 ... - attr(*, "dimnames")=List of 1 ..$ Depends: chr [1:15] "0" "1" "2" "3" ... } %\details{ %%% ~~ If necessary, more details than the __description__ above ~~ %} \source{ Using code from \url{http://blog.revolutionanalytics.com/2013/12/a-look-at-the-distribution-of-r-package-dependencies.html} } %\references{ %%% ~~ possibly secondary sources and usages ~~ %} \examples{ data(Depends) plot(Depends, xlab="Number of Dependencies", ylab="Number of R Packages", lwd=8) \dontrun{ # The code below, from Joseph Rickert, downloads and tabulates the data p <- as.data.frame(available.packages(),stringsAsFactors=FALSE) names(p) pkgs <- data.frame(p[,c(1,4)]) # Pick out Package names and Depends row.names(pkgs) <- NULL # Get rid of row names pkgs <- pkgs[complete.cases(pkgs[,2]),] # Remove NAs pkgs$Depends2 <-strsplit(pkgs$Depends,",") # split list of Depends pkgs$numDepends <- as.numeric(lapply(pkgs$Depends2,length)) # Count number of dependencies in list zeros <- c(rep(0,dim(p)[1] - dim(pkgs)[1])) # Account for packages with no dependencies Deps <- as.vector(c(zeros,pkgs$numDepends)) # Set up to tablate Depends <- table(Deps) } } \keyword{datasets} vcdExtra/man/Hoyt.Rd0000644000176200001440000000652013163461153014025 0ustar liggesusers\name{Hoyt} \Rdversion{1.1} \alias{Hoyt} \docType{data} \title{Minnesota High School Graduates} \description{Minnesota high school graduates of June 1930 were classified with respect to (a) \code{Rank} by thirds in their graduating class, (b) post-high school \code{Status} in April 1939 (4 levels), (c) \code{Sex}, (d) father's \code{Occupation}al status (7 levels, from 1=High to 7=Low). The data were first presented by Hoyt et al. (1959) and have been analyzed by Fienberg(1980), Plackett(1974) and others. } \usage{ data(Hoyt) } \format{ A 4-dimensional array resulting from cross-tabulating 4 variables for 13968 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{Status}\tab \code{"College", "School", "Job", "Other"}\cr 2\tab \code{Rank}\tab \code{"Low", "Middle", "High"}\cr 3\tab \code{Occupation}\tab \code{"1", "2", "3", "4", "5", "6", "7"}\cr 4\tab \code{Sex}\tab \code{"Male", "Female"}\cr } } \details{Post high-school \code{Status} is natural to consider as the response. \code{Rank} and father's \code{Occupation} are ordinal variables.} \source{ % \cite{Hoyt-etal:59} % \cite{Fienberg:80 [pp.91-92]} % \cite{Plackett:74} % \cite{minn38{MASS}} Fienberg, S. E. (1980). \emph{The Analysis of Cross-Classified Categorical Data}. Cambridge, MA: MIT Press, p. 91-92. R. L. Plackett, (1974). \emph{The Analysis of Categorical Data}. London: Griffin. } \references{ Hoyt, C. J., Krishnaiah, P. R. and Torrance, E. P. (1959) Analysis of complex contingency tables, \emph{Journal of Experimental Education} 27, 187-194. } \seealso{ \code{\link[MASS]{minn38}} provides the same data as a data frame. } \examples{ data(Hoyt) # display the table structable(Status+Sex ~ Rank+Occupation, data=Hoyt) # mosaic for independence model plot(Hoyt, shade=TRUE) # examine all pairwise mosaics pairs(Hoyt, shade=TRUE) # collapse Status to College vs. Non-College Hoyt1 <- collapse.table(Hoyt, Status=c("College", rep("Non-College",3))) plot(Hoyt1, shade=TRUE) ################################################# # fitting models with loglm, plotting with mosaic ################################################# # fit baseline log-linear model for Status as response require(MASS) hoyt.mod0 <- loglm(~ Status + (Sex*Rank*Occupation), data=Hoyt1) hoyt.mod0 mosaic(hoyt.mod0, gp=shading_Friendly, main="Baseline model: Status + (Sex*Rank*Occ)") # add one-way association of Status with factors hoyt.mod1 <- loglm(~ Status * (Sex + Rank + Occupation) + (Sex*Rank*Occupation), data=Hoyt1) hoyt.mod1 mosaic(hoyt.mod1, gp=shading_Friendly, main="Status * (Sex + Rank + Occ)") # can we drop any terms? drop1(hoyt.mod1, test="Chisq") # assess model fit anova(hoyt.mod0, hoyt.mod1) # what terms to add? add1(hoyt.mod1, ~.^2, test="Chisq") # add interaction of Sex:Occupation on Status hoyt.mod2 <- update(hoyt.mod1, ~.+Status:Sex:Occupation) mosaic(hoyt.mod2, gp=shading_Friendly, main="Adding Status:Sex:Occupation") # compare model fits anova(hoyt.mod0, hoyt.mod1, hoyt.mod2) # Alternatively, try stepwise analysis, heading toward the saturated model steps <- step(hoyt.mod0, direction="forward", scope=~Status*Sex*Rank*Occupation) # display anova steps$anova } \keyword{datasets} vcdExtra/man/CMHtest.Rd0000644000176200001440000002046013163461153014410 0ustar liggesusers\name{CMHtest} \alias{CMHtest} \alias{CMHtest.formula} \alias{CMHtest.default} \alias{Cochran Mantel Haenszel test} \alias{print.CMHtest} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Generalized Cochran-Mantel-Haenszel Tests } \description{ Provides generalized Cochran-Mantel-Haenszel tests of association of two possibly ordered factors, optionally stratified other factor(s). With strata, \code{CMHtest} calculates these tests for each level of the statifying variables and also provides overall tests controlling for the strata. For ordinal factors, more powerful tests than the test for general association (independence) are obtained by assigning scores to the row and columm categories. } \usage{ CMHtest(x, ...) \method{CMHtest}{formula}(formula, data = NULL, subset = NULL, na.action = NULL, ...) \method{CMHtest}{default}(x, strata = NULL, rscores = 1:R, cscores = 1:C, types = c("cor", "rmeans", "cmeans", "general"), overall=FALSE, details=overall, ...) \method{print}{CMHtest}(x, digits = max(getOption("digits") - 2, 3), ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{ A 2+ way contingency table in array form, or a class \code{"table"} object with optional category labels specified in the dimnames(x) attribute. } \item{formula}{a formula specifying the variables used to create a contingency table from \code{data}. This should be a one-sided formula when \code{data} is in array form, and a two-sided formula with a response \code{Freq} if \code{data} is a data frame with a cell frequency variable. For convenience, conditioning formulas can be specified indicating strata. } \item{data}{either a data frame, or an object of class \code{"table"} or \code{"ftable"}. } \item{subset}{an optional vector specifying a subset of observations to be used. } \item{na.action}{a function which indicates what should happen when the data contain \code{NA}s. Ignored if \code{data} is a contingency table } \item{strata}{ For a 3- or higher-way table, the names or numbers of the factors to be treated as strata. By default, the first 2 factors are treated as the main table variables, and all others considered stratifying factors. } \item{rscores}{ Row scores. Either a set of numbers (typically integers, \code{1:R}) or the string \code{"midrank"} for standardized midrank scores, or \code{NULL} to exclude tests that depend on row scores. } \item{cscores}{ Column scores. Same as for row scores. } \item{types}{ Types of CMH tests to compute: Any one or more of \code{c("cor", "cmeans", "rmeans", "general")}, or \code{"ALL"} for all of these. } \item{overall}{ logical. Whether to calculate overall tests, controlling for the stratifying factors. } \item{details}{ logical. Whether to include computational details in the result } \item{\dots}{ Other arguments passed to default method. } \item{digits}{ Digits to print. } } \details{ The standard \eqn{\chi^2} tests for association in a two-way table treat both table factors as nominal (unordered) categories. When one or both factors of a two-way table are quantitative or ordinal, more powerful tests of association may be obtained by taking ordinality into account using row and or column scores to test for linear trends or differences in row or column means. The CMH analysis for a two-way table produces generalized Cochran-Mantel-Haenszel statistics (Landis etal., 1978). These include the CMH \bold{correlation} statistic (\code{"cor"}), treating both factors as ordered. For a given statum, with equally spaced row and column scores, this CMH statistic reduces to \eqn{(n-1) r^2}, where \eqn{r} is the Pearson correlation between X and Y. With \code{"midrank"} scores, this CMH statistic is analogous to \eqn{(n-1) r_S^2}, using the Spearman rank correlation. The \bold{ANOVA} (row mean scores and column mean scores) statistics, treat the columns and rows respectively as ordinal, and are sensitive to mean shifts over columns or rows. These are transforms of the \eqn{F} statistics from one-way ANOVAs with equally spaced scores and to Kruskal-Wallis tests with \code{"midrank"} scores. The CMH \bold{general} association statistic treat both factors as unordered, and give a test closely related to the Pearson \eqn{\chi^2} test. When there is more than one stratum, the overall general CMH statistic gives a stratum-adjusted Pearson \eqn{\chi^2}, equivalent to what is calculated by \code{\link[stats]{mantelhaen.test}}. For a 3+ way table, one table of CMH tests is produced for each combination of the factors identified as \code{strata}. If \code{overall=TRUE}, an additional table is calculated for the same two primary variables, controlling for (pooling over) the \code{strata} variables. These overall tests implicitly assume no interactions between the primary variables and the strata and they will have low power in the presence of interactions. } \value{ An object of class \code{"CMHtest"} , a list with the following 4 components: \item{table}{A matrix containing the test statistics, with columns \code{Chisq}, \code{Df} and \code{Prob} } \item{names}{The names of the table row and column variables} \item{rscore}{Row scores} \item{cscore}{Column scores} If \code{details==TRUE}, additional components are included. If there are strata, the result is a list of \code{"CMHtest"} objects. If \code{overall=TRUE} another component, labeled \code{ALL} is appended to the list. } \references{ Stokes, M. E. & Davis, C. S. & Koch, G., (2000). \emph{Categorical Data Analysis using the SAS System}, 2nd Ed., Cary, NC: SAS Institute, pp 74-75, 92-101, 124-129. Details of the computation are given at: \url{http://support.sas.com/documentation/cdl/en/statug/63033/HTML/default/viewer.htm#statug_freq_a0000000648.htm } Cochran, W. G. (1954), Some Methods for Strengthening the Common \eqn{\chi^2} Tests, \emph{Biometrics}, 10, 417-451. Landis, R. J., Heyman, E. R., and Koch, G. G. (1978). Average Partial Association in Three-way Contingency Tables: A Review and Discussion of Alternative Tests, \emph{International Statistical Review}, \bold{46}, 237-254. Mantel, N. (1963), Chi-square Tests with One Degree of Freedom: Extensions of the Mantel-Haenszel Procedure," \emph{Journal of the American Statistical Association}, 58, 690-700. } \author{ Michael Friendly } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[coin]{cmh_test}} provides the CMH test of general association; \code{\link[coin]{lbl_test}} provides the CMH correlation test of linear by linear association. \code{\link[stats]{mantelhaen.test}} provides the overall general Cochran-Mantel-Haenszel chi-squared test of the null that two nominal variables are conditionally independent in each stratum, assuming that there is no three-way interaction } \examples{ data(JobSat, package="vcdExtra") CMHtest(JobSat) CMHtest(JobSat, rscores="midrank", cscores="midrank") # formula interface CMHtest(~ ., data=JobSat) # A 3-way table (both factors ordinal) data(MSPatients, package="vcd") CMHtest(MSPatients) # also calculate overall tests, controlling for Patient CMHtest(MSPatients, overall=TRUE) # compare with mantelhaen.test mantelhaen.test(MSPatients) # formula interface CMHtest(~ ., data=MSPatients, overall=TRUE) # using a frequency data.frame CMHtest(xtabs(Freq~ses+mental, data=Mental)) # or, more simply CMHtest(Freq~ses+mental, data=Mental) # conditioning formulae CMHtest(Freq~right+left|gender, data=VisualAcuity) CMHtest(Freq ~ attitude+memory|education+age, data=Punishment) # Stokes etal, Table 5.1, p 92: two unordered factors parties <- matrix( c(221, 160, 360, 140, 200, 291, 160, 311, 208, 106, 316, 97), nrow=3, ncol=4, byrow=TRUE) dimnames(parties) <- list(party=c("Dem", "Indep", "Rep"), neighborhood=c("Bayside", "Highland", "Longview", "Sheffield")) CMHtest(parties, rscores=NULL, cscores=NULL) # compare with Pearson chisquare chisq.test(parties) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{htest} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Kway.Rd0000644000176200001440000001037713163461153014022 0ustar liggesusers\name{Kway} \Rdversion{1.1} \alias{Kway} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Fit All K-way Models in a GLM } \description{ Generate and fit all 0-way, 1-way, 2-way, ... k-way terms in a glm. This function is designed mainly for hierarchical loglinear models (or \code{glm}s in the poission family), where it is desired to find the highest-order terms necessary to achieve a satisfactory fit. Using \code{\link[stats]{anova}} on the resulting \code{\link{glmlist}} object will then give sequential tests of the pooled contributions of all terms of degree \eqn{k+1} over and above those of degree \eqn{k}. This function is also intended as an example of a generating function for \code{\link{glmlist}} objects, to facilitate model comparison, extraction, summary and plotting of model components, etc., perhaps using \code{lapply} or similar. } \usage{ Kway(formula, family=poisson, data, ..., order = nt, prefix = "kway") } %- maybe also 'usage' for other objects documented here. \arguments{ \item{formula}{ a two-sided formula for the 1-way effects in the model. The LHS should be the response, and the RHS should be the first-order terms connected by \code{+} signs. } \item{family}{ a description of the error distribution and link function to be used in the model. This can be a character string naming a family function, a family function or the result of a call to a family function. (See \code{\link[stats]{family}} for details of family functions.) } \item{data}{ an optional data frame, list or environment (or object coercible by \code{\link[base]{as.data.frame}} to a data frame) containing the variables in the model. If not found in data, the variables are taken from \code{environment(formula)}, typically the environment from which \code{glm} is called. } \item{\dots}{ Other arguments passed to \code{glm} } \item{order}{ Highest order interaction of the models generated. Defaults to the number of terms in the model formula. } \item{prefix}{ Prefix used to label the models fit in the \code{glmlist} object. } } \details{ With \code{y} as the response in the \code{formula}, the 0-way (null) model is \code{y ~ 1}. The 1-way ("main effects") model is that specified in the \code{formula} argument. The k-way model is generated using the formula \code{. ~ .^k}. With the default \code{order = nt}, the final model is the saturated model. As presently written, the function requires a two-sided formula with an explicit response on the LHS. For frequency data in table form (e.g., produced by \code{xtabs}) you the \code{data} argument is coerced to a data.frame, so you should supply the \code{formula} in the form \code{Freq ~ } \dots. } \value{ An object of class \code{glmlist}, of length \code{order+1} containing the 0-way, 1-way, ... models up to degree \code{order}. } %\references{ %%% ~put references to the literature/web site here ~ %} \author{ Michael Friendly and Heather Turner } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link{glmlist}}, \code{\link{Summarise}} (soon to be deprecated), \code{\link{LRstats}} } \examples{ ## artificial data factors <- expand.grid(A=factor(1:3), B=factor(1:2), C=factor(1:3), D=factor(1:2)) Freq <- rpois(nrow(factors), lambda=40) df <- cbind(factors, Freq) mods3 <- Kway(Freq ~ A + B + C, data=df, family=poisson) LRstats(mods3) mods4 <- Kway(Freq ~ A + B + C + D, data=df, family=poisson) LRstats(mods4) # JobSatisfaction data data(JobSatisfaction, package="vcd") modSat <- Kway(Freq ~ management+supervisor+own, data=JobSatisfaction, family=poisson, prefix="JobSat") LRstats(modSat) anova(modSat, test="Chisq") # Rochdale data: very sparse, in table form data(Rochdale, package="vcd") \dontrun{ modRoch <- Kway(Freq~EconActive + Age + HusbandEmployed + Child + Education + HusbandEducation + Asian + HouseholdWorking, data=Rochdale, family=poisson) LRstats(modRoch) } } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{models} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/JobSat.Rd0000644000176200001440000000217313163461153014264 0ustar liggesusers\name{JobSat} \Rdversion{1.1} \alias{JobSat} \docType{data} \title{Cross-classification of job satisfaction by income} \description{ This data set is a contingency table of job satisfaction by income for a small sample of black males from the 1996 General Social Survey, as used by Agresti (2002) for an example. } \usage{data(JobSat)} \format{ A 4 x 4 contingency table of \code{income} by \code{satisfaction}, with the following structure: \preformatted{ table [1:4, 1:4] 1 2 1 0 3 3 6 1 10 10 ... - attr(*, "dimnames")=List of 2 ..$ income : chr [1:4] "< 15k" "15-25k" "25-40k" "> 40k" ..$ satisfaction: chr [1:4] "VeryD" "LittleD" "ModerateS" "VeryS" } } \details{ Both \code{income} and \code{satisfaction} are ordinal variables, and are so ordered in the table. Measures of association, visualizations, and models should take ordinality into account. } \source{ Agresti, A. Categorical Data Analysis John Wiley & Sons, 2002, Table 2.8, p. 57. } %\references{ % ~~ possibly secondary sources and usages ~~ %} \examples{ data(JobSat) assocstats(JobSat) GKgamma(JobSat) } \keyword{datasets} vcdExtra/man/Fungicide.Rd0000644000176200001440000000353113163461153014776 0ustar liggesusers\name{Fungicide} \alias{Fungicide} \docType{data} \title{ Carcinogenic Effects of a Fungicide } \description{ Data from Gart (1971) on the carcinogenic effects of a certain fungicide in two strains of mice. Of interest is how the association between \code{group} (Control, Treated) and \code{outcome} (Tumor, No Tumor) varies with \code{sex} and \code{strain} of the mice. Breslow (1976) used this data to illustrate the application of linear models to log odds ratios. } \usage{data(Fungicide)} \format{ The data comprise a set of four 2 x 2 tables classifying 403 mice, either Control or Treated and whether or not a tumor was later observed. The four groups represent the combinations of sex and strain of mice. The format is: num [1:2, 1:2, 1:2, 1:2] 5 4 74 12 3 2 84 14 10 4 ... - attr(*, "dimnames")=List of 4 ..$ group : chr [1:2] "Control" "Treated" ..$ outcome: chr [1:2] "Tumor" "NoTumor" ..$ sex : chr [1:2] "M" "F" ..$ strain : chr [1:2] "1" "2" } \details{ All tables have some small cells, so a continuity correction is recommended. } \source{ Gart, J. J. (1971). The comparison of proportions: a review of significance tests, confidence intervals and adjustments for stratification. \emph{International Statistical Review}, 39, 148-169. } \references{ Brewlow, N. (1976), Regression analysis of the log odds ratio: A method for retrospective studies, \emph{Biometrics}, 32(3), 409-416. } \examples{ data(Fungicide) # loddsratio was moved to vcd; requires vcd_1.3-3+ \dontrun{ if (require(vcd)) { fung.lor <- loddsratio(Fungicide, correct=TRUE) fung.lor confint(fung.lor) } } # visualize odds ratios in fourfold plots cotabplot(Fungicide, panel=cotab_fourfold) # -- fourfold() requires vcd >= 1.2-10 fourfold(Fungicide, p_adjust_method="none") } \keyword{datasets} vcdExtra/man/Geissler.Rd0000644000176200001440000000473513163461153014665 0ustar liggesusers\name{Geissler} \alias{Geissler} \docType{data} \title{ Geissler's Data on the Human Sex Ratio } \description{ Geissler (1889) published data on the distributions of boys and girls in families in Saxony, collected for the period 1876-1885. The \code{Geissler} data tabulates the family composition of 991,958 families by the number of boys and girls listed in the table supplied by Edwards (1958, Table 1). } \usage{data(Geissler)} \format{ A data frame with 90 observations on the following 4 variables. The rows represent the non-NA entries in Edwards' table. \describe{ \item{\code{boys}}{number of boys in the family, \code{0:12}} \item{\code{girls}}{number of girls in the family, \code{0:12}} \item{\code{size}}{family size: \code{boys+girls}} \item{\code{Freq}}{number of families with this sex composition} } } \details{ The data on family composition was available because, on the birth of a child, the parents had to state the sex of all their children on the birth certificate. These family records are not necessarily independent, because a given family may have had several children during this 10 year period, included as multiple records. } \source{ Edwards, A. W. F. (1958). An Analysis Of Geissler's Data On The Human Sex Ratio. \emph{Annals of Human Genetics}, 23, 6-15. } \references{ Friendly, M. and Meyer, D. (2016). \emph{Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data}. Boca Raton, FL: Chapman & Hall/CRC. \url{http://ddar.datavis.ca}. Geissler, A. (1889). \emph{Beitrage zur Frage des Geschlechts verhaltnisses der Geborenen} Z. K. Sachsischen Statistischen Bureaus, 35, n.p. Lindsey, J. K. & Altham, P. M. E. (1998). Analysis of the human sex ratio by using overdispersion models. \emph{Journal of the Royal Statistical Society: Series C (Applied Statistics)}, 47, 149-157. } \seealso{ \code{\link[vcd]{Saxony}}, containing the data for families of size 12. } \examples{ data(Geissler) ## maybe str(Geissler) ; plot(Geissler) ... # reproduce Saxony data, families of size 12 Saxony12<-subset(Geissler, size==12, select=c(boys, Freq)) rownames(Saxony12)<-NULL # make a 1-way table xtabs(Freq~boys, Saxony12) # extract data for other family sizes Saxony11<-subset(Geissler, size==11, select=c(boys, Freq)) rownames(Saxony11)<-NULL Saxony10<-subset(Geissler, size==10, select=c(boys, Freq)) rownames(Saxony10)<-NULL } \keyword{datasets} vcdExtra/man/logLik.loglm.Rd0000644000176200001440000000617013163461153015435 0ustar liggesusers\name{logLik.loglm} \alias{logLik.loglm} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Log-Likelihood of a loglm Object } \description{ Calculates the log-likelihood value of the \code{loglm} model represented by \code{object} evaluated at the estimated coefficients. It allows the use of \code{\link[stats]{AIC}} and \code{\link[stats]{BIC}}, which require that a \code{logLik} method exists to extract the corresponding log-likelihood for the model. } \usage{ \method{logLik}{loglm}(object, ..., zero=1E-10) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{object}{ A \code{loglm} object } \item{\dots}{ For compatibility with the S3 generic; not used here } \item{zero}{value used to replace zero frequencies in calculating the log-likelihood} } \details{ If cell frequencies have not been stored with the \code{loglm} object (via the argument \code{keep.frequencies = TRUE}), they are obtained using \code{update}. This function calculates the log-likelihood in a way that allows for non-integer frequencies, such as the case where 0.5 has been added to all cell frequencies to allow for sampling zeros. If the frequencies still contain zero values, those are replaced by the value of \code{start}. For integer frequencies, it gives the same result as the corresponding model fit using \code{\link[stats]{glm}}, whereas \code{\link[stats]{glm}} returns \code{-Inf} if there are any non-integer frequencies. } \value{ Returns an object of class \code{logLik}. This is a number with one attribute, \code{"df"} (degrees of freedom), giving the number of (estimated) parameters in the model. } %\references{ %%% ~put references to the literature/web site here ~ %} \author{ Achim Zeileis } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[MASS]{loglm}}, \code{\link[stats]{AIC}}, \code{\link[stats]{BIC}}, } \examples{ data(Titanic, package="datasets") require(MASS) titanic.mod1 <- loglm(~ (Class * Age * Sex) + Survived, data=Titanic) titanic.mod2 <- loglm(~ (Class * Age * Sex) + Survived*(Class + Age + Sex), data=Titanic) titanic.mod3 <- loglm(~ (Class * Age * Sex) + Survived*(Class + Age * Sex), data=Titanic) logLik(titanic.mod1) AIC(titanic.mod1, titanic.mod2, titanic.mod3) BIC(titanic.mod1, titanic.mod2, titanic.mod3) # compare with models fit using glm() titanic <- as.data.frame(Titanic) titanic.glm1 <- glm(Freq ~ (Class * Age * Sex) + Survived, data=titanic, family=poisson) titanic.glm2 <- glm(Freq ~ (Class * Age * Sex) + Survived*(Class + Age + Sex), data=titanic, family=poisson) titanic.glm3 <- glm(Freq ~ (Class * Age * Sex) + Survived*(Class + Age * Sex), data=titanic, family=poisson) logLik(titanic.glm1) AIC(titanic.glm1, titanic.glm2, titanic.glm3) BIC(titanic.glm1, titanic.glm2, titanic.glm3) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{models} \keyword{htest}% __ONLY ONE__ keyword per line vcdExtra/man/Heckman.Rd0000644000176200001440000000505313163461153014450 0ustar liggesusers\name{Heckman} \Rdversion{1.1} \alias{Heckman} \docType{data} \title{Labour Force Participation of Married Women 1967-1971} \description{1583 married women were surveyed over the years 1967-1971, recording whether or not they were employed in the labor force. The data, originally from Heckman & Willis (1977) provide an example of modeling longitudinal categorical data, e.g., with markov chain models for dependence over time. } \usage{ data(Heckman) } \format{ A 5-dimensional array resulting from cross-tabulating 5 variables for 1583 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{e1971}\tab \code{"71Yes", "No"}\cr 2\tab \code{e1970}\tab \code{"70Yes", "No"}\cr 3\tab \code{e1969}\tab \code{"69Yes", "No"}\cr 4\tab \code{e1968}\tab \code{"68Yes", "No"}\cr 5\tab \code{e1967}\tab \code{"67Yes", "No"}\cr } } \details{ Lindsey (1993) fits an initial set of logistic regression models examining the dependence of employment in 1971 (\code{e1971}) on successive subsets of the previous years, \code{e1970}, \code{e1969}, \dots \code{e1967}. Alternatively, one can examine markov chain models of first-order (dependence on previous year), second-order (dependence on previous two years), etc. } \source{ % \cite{Lindsey:93 [p. 185]} Lindsey, J. K. (1993). \emph{Models for Repeated Measurements} Oxford, UK: Oxford University Press, p. 185. } \references{ % \cite{HeckmanWillis:77} Heckman, J.J. & Willis, R.J. (1977). "A beta-logistic model for the analysis of sequential labor force participation by married women." \emph{Journal of Political Economy}, 85: 27-58 } %\seealso{ } \examples{ data(Heckman) # independence model mosaic(Heckman, shade=TRUE) # same, as a loglm() require(MASS) (heckman.mod0 <- loglm(~ e1971+e1970+e1969+e1968+e1967, data=Heckman)) mosaic(heckman.mod0, main="Independence model") # first-order markov chain: bad fit (heckman.mod1 <- loglm(~ e1971*e1970 + e1970*e1969 +e1969*e1968 + e1968*e1967, data=Heckman)) mosaic(heckman.mod1, main="1st order markov chain model") # second-order markov chain: bad fit (heckman.mod2 <- loglm(~ e1971*e1970*e1969 + e1970*e1969*e1968 +e1969*e1968*e1967, data=Heckman)) mosaic(heckman.mod2, main="2nd order markov chain model") # third-order markov chain: fits OK (heckman.mod3 <- loglm(~ e1971*e1970*e1969*e1968 + e1970*e1969*e1968*e1967, data=Heckman)) mosaic(heckman.mod2, main="3rd order markov chain model") } \keyword{datasets} vcdExtra/man/blogits.Rd0000644000176200001440000000531713163461153014550 0ustar liggesusers\name{blogits} \alias{blogits} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Bivariate Logits and Log Odds Ratio } \description{ This function calculates the log odds and log odds ratio for two binary responses classified by one or more stratifying variables. It is useful for plotting the results of bivariate logistic regression models, such as those fit using \code{\link[VGAM]{vglm}} in the \pkg{VGAM}. } \usage{ blogits(Y, add, colnames, row.vars, rev=FALSE) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{Y}{ A four-column matrix or data frame whose columns correspond to the 2 x 2 combinations of two binary responses. } \item{add}{ Constant added to all cells to allow for zero frequencies. The default is 0.5 if \code{any(Y)==0} and 0 otherwise. } \item{colnames}{ Names for the columns of the results. The default is \code{c("logit1", "logit2", "logOR")}. If less than three names are supplied, the remaining ones are filled in from the default. } \item{row.vars}{ A data frame or matrix giving the factor levels of one or more factors corresponding to the rows of \code{Y} } \item{rev}{A logical, indicating whether the order of the columns in \code{Y} should be reversed.} } \details{ For two binary variables with levels 0,1 the logits are calculated assuming the columns in \code{Y} are given in the order 11, 10, 01, 00, so the logits give the log odds of the 1 response compared to 0. If this is not the case, either use \code{rev=TRUE} or supply \code{Y[,4:1]} as the first argument. } \value{ A data frame with \code{nrow(Y)} rows and \code{3 + ncol(row.vars)} columns } \references{ Friendly, M. and Meyer, D. (2016). \emph{Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data}. Boca Raton, FL: Chapman & Hall/CRC. \url{http://ddar.datavis.ca}. } \author{ Michael Friendly } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[VGAM]{vglm}} } \examples{ data(Toxaemia) tox.tab <- xtabs(Freq~class + smoke + hyper + urea, Toxaemia) # reshape to 4-column matrix toxaemia <- t(matrix(aperm(tox.tab), 4, 15)) colnames(toxaemia) <- c("hu", "hU", "Hu", "HU") rowlabs <- expand.grid(smoke=c("0", "1-19", "20+"), class=factor(1:5)) toxaemia <- cbind(toxaemia, rowlabs) # logits for H and U logitsTox <- blogits(toxaemia[,4:1], add=0.5, colnames=c("logitH", "logitW"), row.vars=rowlabs) logitsTox } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{manip} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Caesar.Rd0000644000176200001440000000474213163461153014304 0ustar liggesusers\name{Caesar} \alias{Caesar} \docType{data} \title{Risk Factors for Infection in Caesarian Births} \description{Data from infection from birth by Caesarian section, classified by \code{Risk} (two levels), whether \code{Antibiotics} were used (two levels) and whether the Caesarian section was \code{Planned} or not. The outcome is \code{Infection} (three levels).} \usage{ data(Caesar) } \format{ A 4-dimensional array resulting from cross-tabulating 4 variables for 251 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{Infection}\tab \code{"Type 1", "Type 2", "None"}\cr 2\tab \code{Risk}\tab \code{"Yes", "No"} (presence of risk factors)\cr 3\tab \code{Antibiotics}\tab \code{"Yes", "No"} (were antibiotics given?)\cr 4\tab \code{Planned}\tab \code{"Yes", "No"} (was the C section planned?)\cr } } \details{ \code{Infection} is regarded as the response variable here. There are quite a few 0 cells here, particularly when \code{Risk} is absent and the Caesarian section was unplanned. Should these be treated as structural or sampling zeros? } \source{ % \cite{Fahrmeir:94} Fahrmeir, L. & Tutz, G. (1994). Multivariate Statistical Modelling Based on Generalized Linear Models New York: Springer Verlag, Table 1.1. } %\references{ %} \seealso{\code{\link[Fahrmeir]{caesar}} for the same data recorded as a frequency data frame with other variables.} \examples{ data(Caesar) #display table; note that there are quite a few 0 cells structable(Caesar) require(MASS) # baseline model, Infection as response Caesar.mod0 <- loglm(~Infection + (Risk*Antibiotics*Planned), data=Caesar) # NB: Pearson chisq cannot be computed due to the 0 cells Caesar.mod0 mosaic(Caesar.mod0, main="Baseline model") # Illustrate handling structural zeros zeros <- 0+ (Caesar >0) zeros[1,,1,1] <- 1 structable(zeros) # fit model excluding possible structural zeros Caesar.mod0s <- loglm(~Infection + (Risk*Antibiotics*Planned), data=Caesar, start=zeros) Caesar.mod0s anova(Caesar.mod0, Caesar.mod0s, test="Chisq") mosaic (Caesar.mod0s) # what terms to add? add1(Caesar.mod0, ~.^2, test="Chisq") # add Association of Infection:Antibiotics Caesar.mod1 <- update(Caesar.mod0, ~.+Infection:Antibiotics) anova(Caesar.mod0, Caesar.mod1, test="Chisq") mosaic(Caesar.mod1, gp=shading_Friendly, main="Adding Infection:Antibiotics") } \keyword{datasets} vcdExtra/man/Mammograms.Rd0000644000176200001440000000220713163461153015200 0ustar liggesusers\name{Mammograms} \alias{Mammograms} \docType{data} \title{ Mammogram Ratings } \description{ Kundel & Polansky (2003) give (possibly contrived) data on a set of 110 mammograms rated by two readers. } \usage{data(Mammograms)} \format{ A frequency table in matrix form. The format is: num [1:4, 1:4] 34 6 2 0 10 8 5 1 2 8 ... - attr(*, "dimnames")=List of 2 ..$ Reader2: chr [1:4] "Absent" "Minimal" "Moderate" "Severe" ..$ Reader1: chr [1:4] "Absent" "Minimal" "Moderate" "Severe" } %\details{ %%% ~~ If necessary, more details than the __description__ above ~~ %} \source{ Kundel, H. L. & Polansky, M. (2003), "Measurement of Observer Agreement", \emph{Radiology}, \bold{228}, 303-308, Table A1 } %\references{ %%% ~~ possibly secondary sources and usages ~~ %} \examples{ data(Mammograms) B <- agreementplot(Mammograms, main="Mammogram ratings") # agreement measures B Kappa(Mammograms) ## other displays mosaic(Mammograms, shade=TRUE) sieve(Mammograms, pop = FALSE, shade = TRUE) labeling_cells(text = Mammograms, gp_text = gpar(fontface = 2, cex=1.75))(as.table(Mammograms)) } \keyword{datasets} vcdExtra/man/PhdPubs.Rd0000644000176200001440000000315413163461153014447 0ustar liggesusers\name{PhdPubs} \alias{PhdPubs} \docType{data} \title{ Publications of PhD Candidates } \description{ A data set giving the number of publications by doctoral candidates in biochemistry in relation to various predictors, originally from Long (1997). There is a large number of zero counts. Is there evidence for a separate group of non-publishers? } \usage{data(PhdPubs)} \format{ A data frame with 915 observations on the following 6 variables. \describe{ \item{\code{articles}}{number of articles published in the final three years of PhD studies} \item{\code{female}}{dummy variable for gender, coded \code{1} for female} \item{\code{married}}{dummy variable for marital status, coded \code{1} for married} \item{\code{kid5}}{number of young children, age 5 and under} \item{\code{phdprestige}}{prestige of the PhD department} \item{\code{mentor}}{number of publications by the mentor in the preceeding three years} } } %\details{ %% ~~ If necessary, more details than the __description__ above ~~ %} \source{ Long, J. S. (1997) \emph{Regression Models for Categorical and Limited Dependent Variables}, Sage. } %\references{ %% ~~ possibly secondary sources and usages ~~ %} \examples{ data(PhdPubs) # very uninformative hist(PhdPubs$articles, breaks=0:19, col="pink", xlim=c(0,20), xlab="Number of Articles") library(vcd) rootogram(goodfit(PhdPubs$articles), xlab="Number of Articles") # compare with negative binomial rootogram(goodfit(PhdPubs$articles, type="nbinomial"), xlab="Number of Articles", main="Negative binomial") } \keyword{datasets} vcdExtra/man/mcaplot.Rd0000644000176200001440000000715313163461153014544 0ustar liggesusers\name{mcaplot} \alias{mcaplot} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Simple and enhanced plot of MCA solutions } \description{ This function is intended as an alternative to \code{\link[ca]{plot.mjca}} for plotting multiple correspondence analysis solutions. It provides more flexibility for labeling factor levels and connecting them with lines. It does not support some features of \code{plot.mjca} (centroids, supplementary points, arrows, etc.) } \usage{ mcaplot(obj, map = "symmetric", dim = 1:2, col = c("blue", "red", "brown", "black", "green3", "purple"), pch = 15:20, cex = 1.2, pos = 3, lines = TRUE, lwd = 2, legend = FALSE, legend.pos = "topright", xlab = "_auto_", ylab = "_auto_", rev.axes = c(FALSE, FALSE), ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{obj}{An \code{"mjca"} object} \item{map}{ Character string specifying the map type, i.e., the scaling applied to coordinates for different types of MCA representations. Allowed options include: \code{"symmetric"} (default), \code{"rowprincipal"}, \code{"colprincipal"}, \code{"symbiplot"}, \code{"rowgab"}, \code{"colgab"}, \code{"rowgreen"}, \code{"colgreen"}. See \code{\link[ca]{mjca}} for details. } \item{dim}{Dimensions to plot, an integer vector of length 2} \item{col}{Vector of colors, one for each factor in the MCA} \item{pch}{Vector of point symbols for the category levels, one for each factor} \item{cex}{Character size for points and level labels} \item{pos}{Position of level labels relative to the category points; either a single number or a vector of length equal to the number of category points.} \item{lines}{A logical or an integer vector indicating which factors are to be joined with lines using \code{\link{multilines}}} \item{lwd}{Line width(s) for the lines} \item{legend}{Logical; draw a legend for the factor names?} \item{legend.pos}{Position of the legend in the plot, as in \code{\link[graphics]{legend}}} \item{xlab,ylab}{Labels for horizontal and vertical axes. The default, \code{"_auto_"} means that the function auto-generates a label of the form \code{"Dimension X (xx.x \%)"} } \item{rev.axes}{A logical vector of length 2, where TRUE reverses the direction of the corresponding axis} \item{\dots}{Arguments passed down to \code{plot}} } %\details{ %% ~~ If necessary, more details than the description above ~~ %} \value{ Returns the coordinates of the category points invisibly %% ~Describe the value returned %% If it is a LIST, use %% \item{comp1 }{Description of 'comp1'} %% \item{comp2 }{Description of 'comp2'} %% ... } %\references{ %% ~put references to the literature/web site here ~ %} \author{ Michael Friendly } %\note{ %% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[ca]{mjca}}, \code{\link[ca]{plot.mjca}} \code{\link[ca]{cacoord}} returns CA and MCA coordinates, \code{\link[ca]{multilines}} draw multiple lines according to a factor, } \examples{ require(ca) data(Titanic) titanic.mca <- mjca(Titanic) mcaplot(titanic.mca, legend=TRUE, legend.pos="topleft") data(HairEyeColor) haireye.mca <- mjca(HairEyeColor) mcaplot(haireye.mca, legend=TRUE, cex.lab=1.3) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{hplot} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/WorkerSat.Rd0000644000176200001440000000230313163461153015016 0ustar liggesusers\name{WorkerSat} \alias{WorkerSat} \docType{data} \title{ Worker Satisfaction Data } \description{ Blue collar workers job satisfaction from large scale investigation in Denmark in 1968 (Andersen, 1991). } \usage{data("WorkerSat")} \format{ A frequency data frame with 8 observations on the following 4 variables, representing the 2 x 2 x 2 classification of 715 cases. \describe{ \item{\code{Manage}}{Quality of management, an ordered factor with levels \code{bad} < \code{good}} \item{\code{Super}}{Supervisor satisfaction, an ordered factor with levels \code{low} < \code{high}} \item{\code{Worker}}{Worker job satisfaction, an ordered factor with levels \code{low} < \code{high}} \item{\code{Freq}}{a numeric vector} } } %\details{ %%% ~~ If necessary, more details than the __description__ above ~~ %} \source{ \url{https://onlinecourses.science.psu.edu/stat504/node/131} } \references{ Andersen, E. B. (1991) Statistical Analysis of Categorical Data, 2nd Ed., Springer-Verlag. } \examples{ data(WorkerSat) worker.tab <- xtabs(Freq ~ Worker + Super + Manage, data=WorkerSat) fourfold(worker.tab) mosaic(worker.tab, shade=TRUE) } \keyword{datasets} vcdExtra/man/split3d.Rd0000644000176200001440000000605413163461153014466 0ustar liggesusers\name{split3d} \Rdversion{1.1} \alias{split3d} \alias{split3d.shape3d} \alias{split3d.list} \alias{range3d} \alias{center3d} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Subdivide a 3D Object } \description{ Subdivides a \code{shape3d} object or a list of \code{shape3d} objects into objects of the same shape along a given dimension according to the proportions or frequencies specified in vector(s). \code{split3d} is the basic workhorse used in \code{\link{mosaic3d}}, but may be useful in other contexts. \code{range3d} and \code{center3d} are utility functions, also useful in other contexts. } \usage{ split3d(obj, ...) \method{split3d}{shape3d}(obj, p, dim, space = 0.1, ...) \method{split3d}{list}(obj, p, dim, space = 0.1, ...) range3d(obj) center3d(obj) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{obj}{ A \code{shape3d} object, or a list composed of them } \item{\dots}{ Other arguments for split3d methods } \item{p}{ For a single \code{shade3d} object, a vector of proportions (or a vector of non-negative numbers which will be normed to proportions) indicating the number of subdivisions and their scaling along dimension \code{dim}. For a list of \code{shade3d} objects, a matrix whose columns indicate the subdivisions of each object. } \item{dim}{ The dimension along which the object is to be subdivided. Either an integer: 1, 2, or 3, or a character: "x", "y", or "z". } \item{space}{ The total space used to separate the copies of the object along dimension \code{dim}. The unit inter-object space is therefore \code{space/(length(p)-1)}. } } \details{ The resulting list of \code{shape3d} objects is actually composed of \emph{copies} of the input object(s), scaled according to the proportions in \code{p} and then translated to make their range along the splitting dimension equal to that of the input object(s). } \value{ \code{split3d} returns a list of \code{shape3d} objects. \code{range3d} returns a 2 x 3 matrix, whose first row contains the minima on dimensions x, y, z, and whose second row contains the maxima. \code{center3d} returns a numeric vector containing the means of the minima and maxima on dimensions x, y, z. } \author{ Duncan Murdoch, with refinements by Michael Friendly } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link{mosaic3d}} \code{\link[rgl]{shapelist3d}} for the plotting of lists of \code{shape3d} objects. } \examples{ if (require(rgl)) { open3d() cube <- cube3d(alpha=0.4) sl1 <- split3d(cube, c(.2, .3, .5), 1) col <- c("#FF000080", "#E5E5E580", "#0000FF80") shapelist3d(sl1, col=col) open3d() p <- matrix(c(.6, .4, .5, .5, .2, .8), nrow=2) sl2 <- split3d(sl1, p, 2) shapelist3d(sl2, col=col) } } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{dplot} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Heart.Rd0000644000176200001440000000146513163461153014150 0ustar liggesusers\name{Heart} \Rdversion{1.1} \alias{Heart} \docType{data} \title{Sex, Occupation and Heart Disease} \description{Classification of individuals by gender, occupational category and occurrence of heart disease} \usage{ data(Heart) } \format{ A 3-dimensional array resulting from cross-tabulating 3 variables for 21522 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{Disease}\tab \code{"Disease", "None"}\cr 2\tab \code{Gender}\tab \code{"Male", "Female"}\cr 3\tab \code{Occup}\tab \code{"Unempl", "WhiteCol", "BlueCol"}\cr } } %\details{ } \source{ % \cite{Karger, 1980} Karger, (1980). } %\references{ %} %\seealso{ } \examples{ data(Heart) # example goes here } \keyword{datasets} vcdExtra/man/Cormorants.Rd0000644000176200001440000000601013163461153015223 0ustar liggesusers\name{Cormorants} \alias{Cormorants} \docType{data} \title{ Advertising Behavior by Males Cormorants } \description{ Male double-crested cormorants use advertising behavior to attract females for breeding. In this study by Meagan Mc Rae (2015), cormorants were observed two or three times a week at six stations in a tree-nesting colony for an entire season, April 10, 2014-July 10, 2014. The number of advertising birds was counted and these observations were classified by characteristics of the trees and nests. The goal is to determine how this behavior varies temporally over the season and spatially, as well as with characteristics of nesting sites. } \usage{data("Cormorants")} \format{ A data frame with 343 observations on the following 8 variables. \describe{ \item{\code{category}}{Time of season, divided into 3 categories based on breeding chronology, an ordered factor with levels \code{Pre} < \code{Incubation} < \code{Chicks Present}} \item{\code{week}}{Week of the season} \item{\code{station}}{Station of observations on two different peninsulas in a park, a factor with levels \code{B1} \code{B2} \code{C1} \code{C2} \code{C3} \code{C4}} \item{\code{nest}}{Type of nest, an ordered factor with levels \code{no} < \code{partial} < \code{full}} \item{\code{height}}{Relative height of bird in the tree, an ordered factor with levels \code{low} < \code{mid} < \code{high}} \item{\code{density}}{Number of other nests in the tree, an ordered factor with levels \code{zero} < \code{few} < \code{moderate} < \code{high}} \item{\code{tree_health}}{Health of the tree the bird is advertising in, a factor with levels \code{dead} \code{healthy}} \item{\code{count}}{Number of birds advertising, a numeric vector} } } \details{ Observations were made on only 2 days in weeks 3 and 4, but 3 days in all other weeks. One should use log(days) as an offset, so that the response measures rate. \code{Cormorants$days <- ifelse(Cormorants$week \%in\% 3:4, 2, 3)} } \source{ Mc Rae, M. (2015). Spatial, Habitat and Frequency Changes in Double-crested Cormorant Advertising Display in a Tree-nesting Colony. Unpublished MA project, Environmental Studies, York University. } %\references{ %%% ~~ possibly secondary sources and usages ~~ %} \examples{ data(Cormorants) str(Cormorants) library(ggplot2) ggplot(Cormorants, aes(count)) + geom_histogram(binwidth=0.5) + labs(x="Number of birds advertising") # Quick look at the data, on the log scale, for plots of `count ~ week`, # stratified by something else. library(ggplot2) ggplot(Cormorants, aes(week, count, color=height)) + geom_jitter() + stat_smooth(method="loess", size=2) + scale_y_log10(breaks=c(1,2,5,10)) + geom_vline(xintercept=c(4.5, 9.5)) # ### models using week fit1 <-glm(count ~ week + station + nest + height + density + tree_health, data=Cormorants, family = poisson) library(car) Anova(fit1) # plot fitted effects library(effects) plot(allEffects(fit1)) } \keyword{datasets} vcdExtra/man/seq_loglm.Rd0000644000176200001440000000737213163461153015072 0ustar liggesusers\name{seq_loglm} \alias{seq_loglm} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Sequential Loglinear Models for an N-way Table } \description{ This function takes an n-way contingency table and fits a series of sequential models to the 1-, 2-, ... n-way marginal tables, corresponding to a variety of types of loglinear models. } \usage{ seq_loglm(x, type = c("joint", "conditional", "mutual", "markov", "saturated"), marginals = 1:nf, vorder = 1:nf, k = NULL, prefix = "model", fitted = TRUE, ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{ a contingency table in array form, with optional category labels specified in the dimnames(x) attribute, or else a data.frame in frequency form, with the frequency variable named \code{"Freq"}. } \item{type}{ type of sequential model to fit, a character string. One of \code{"joint"}, \code{"conditional"}, \code{"mutual"}, \code{"markov"}, or \code{"saturated"}. } \item{marginals}{ which marginal sub-tables to fit? A vector of a (sub)set of the integers, \code{1:nf} where \code{nf} is the number of factors in the full n-way table. } \item{vorder}{ order of variables, a permutation of the integers \code{1:nf}, used to reorder the variables in the original table for the purpose of fitting sequential marginal models. } \item{k}{ conditioning variable(s) for \code{type} = \code{"joint"}, \code{"conditional"} or Markov chain order for \code{type} = \code{"markov"} } \item{prefix}{ prefix used to give names to the sequential models } \item{fitted}{ argument passed to \code{loglm} to store the fitted values in the model objects } \item{\dots}{ other arguments, passed down } } \details{ Sequential marginal models for an n-way tables begin with the model of equal-probability for the one-way margin (equivalent to a \code{\link[stats]{chisq.test}}) and add successive variables one at a time in the order specified by \code{vorder}. All model types give the same result for the two-way margin, namely the test of independence for the first two factors. Sequential models of \emph{joint independence} (\code{type="joint"}) have a particularly simple interpretation, because they decompose the likelihood ratio test for the model of mutual independence in the full n-way table, and hence account for "total" association in terms of portions attributable to the conditional probabilities of each new variable, given all prior variables. } \value{ An object of class \code{"loglmlist"}, each of which is a class \code{"loglm"} object %% If it is a LIST, use %% \item{comp1 }{Description of 'comp1'} %% \item{comp2 }{Description of 'comp2'} %% ... } \references{ These functions were inspired by the original SAS implementation of mosaic displays, described in the \emph{User's Guide}, \url{http://www.datavis.ca/mosaics/mosaics.pdf} } \author{ Michael Friendly } \note{ One-way marginal tables are a bit of a problem here, because they cannot be fit directly using \code{\link[MASS]{loglm}}. The present version uses \code{\link[stats]{loglin}}, and repairs the result to look like a \code{loglm} object (sort of). } %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link{loglin-utilities}} for descriptions of sequential models, \code{\link{conditional}}, \code{\link{joint}}, \code{\link{mutual}}, \dots \code{\link{loglmlist}}, } \examples{ data(Titanic, package="datasets") # variables are in the order Class, Sex, Age, Survived tt <- seq_loglm(Titanic) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{models} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Vietnam.Rd0000644000176200001440000000330313163461153014501 0ustar liggesusers\name{Vietnam} \alias{Vietnam} \docType{data} \title{ Student Opinion about the Vietnam War } \description{ A survey of student opinion on the Vietnam War was taken at the University of North Carolina at Chapel Hill in May 1967 and published in the student newspaper. Students were asked to fill in ballot papers stating which policy out of A,B,C or D they supported. Responses were cross-classified by gender/year. The response categories were: \describe{ \item{\code{A}}{Defeat North Vietnam by widespread bombing and land invasion} \item{\code{B}}{Maintain the present policy} \item{\code{C}}{De-escalate military activity, stop bombing and begin negotiations} \item{\code{D}}{Withdraw military forces Immediately} } } \usage{data(Vietnam)} \format{ A frequency data frame with 40 observations representing a 2 x 5 x 4 contingency table on the following 4 variables. \describe{ \item{\code{sex}}{a factor with levels \code{Female} \code{Male}} \item{\code{year}}{year of study, an ordered factor with levels \code{Freshmen}, \code{Sophomore}, \code{Junior}, \code{Senior}, \code{Grad student}} \item{\code{response}}{a factor with levels \code{A} \code{B} \code{C} \code{D}} \item{\code{Freq}}{cell frequency, a numeric vector} } } \details{ For some analyses, it is useful to treat \code{year} as numeric, and possibly assign grad students a value \code{year=7}. } \source{ Aitken, M. etal, 1989, \emph{Statistical Modelling in GLIM} } \references{ Friendly, M. (2000), \emph{Visualizing Categorical Data}, SAS Institute, Cary, NC, Example 7.9. } \examples{ data(Vietnam) ## maybe str(Vietnam) ; plot(Vietnam) ... } \keyword{datasets} vcdExtra/man/Hauser79.Rd0000644000176200001440000000737613163461153014523 0ustar liggesusers\name{Hauser79} \alias{Hauser79} \docType{data} \title{ Hauser (1979) Data on Social Mobility } \description{ Hauser (1979) presented this two-way frequency table, cross-classifying occupational categories of sons and fathers in the United States. } \usage{data(Hauser79)} \format{ A frequency data frame with 25 observations on the following 3 variables, representing the cross-classification of 19912 individuals by father's occupation and son's first occupation. \describe{ \item{\code{Son}}{a factor with levels \code{UpNM} \code{LoNM} \code{UpM} \code{LoM} \code{Farm}} \item{\code{Father}}{a factor with levels \code{UpNM} \code{LoNM} \code{UpM} \code{LoM} \code{Farm}} \item{\code{Freq}}{a numeric vector} } } %\details{ %%% ~~ If necessary, more details than the __description__ above ~~ %} \source{ R.M. Hauser (1979), Some exploratory methods for modeling mobility tables and other cross-classified data. In: K.F. Schuessler (Ed.), \emph{Sociological Methodology}, 1980, Jossey-Bass, San Francisco, pp. 413-458. } \references{ Powers, D.A. and Xie, Y. (2008). \emph{Statistical Methods for Categorical Data Analysis}, Bingley, UK: Emerald. } \examples{ data(Hauser79) str(Hauser79) # display table structable(~Father+Son, data=Hauser79) #Examples from Powers & Xie, Table 4.15 # independence model mosaic(Freq ~ Father + Son, data=Hauser79, shade=TRUE) hauser.indep <- gnm(Freq ~ Father + Son, data=Hauser79, family=poisson) mosaic(hauser.indep, ~Father+Son, main="Independence model", gp=shading_Friendly) hauser.quasi <- update(hauser.indep, ~ . + Diag(Father,Son)) mosaic(hauser.quasi, ~Father+Son, main="Quasi-independence model", gp=shading_Friendly) hauser.qsymm <- update(hauser.indep, ~ . + Diag(Father,Son) + Symm(Father,Son)) mosaic(hauser.qsymm, ~Father+Son, main="Quasi-symmetry model", gp=shading_Friendly) #mosaic(hauser.qsymm, ~Father+Son, main="Quasi-symmetry model") # numeric scores for row/column effects Sscore <- as.numeric(Hauser79$Son) Fscore <- as.numeric(Hauser79$Father) # row effects model hauser.roweff <- update(hauser.indep, ~ . + Father*Sscore) LRstats(hauser.roweff) # uniform association hauser.UA <- update(hauser.indep, ~ . + Fscore*Sscore) LRstats(hauser.UA) # uniform association, omitting diagonals hauser.UAdiag <- update(hauser.indep, ~ . + Fscore*Sscore + Diag(Father,Son)) LRstats(hauser.UAdiag) # Levels for Hauser 5-level model levels <- matrix(c( 2, 4, 5, 5, 5, 3, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 4, 5, 5, 5, 4, 1 ), 5, 5, byrow=TRUE) hauser.topo <- update(hauser.indep, ~ . + Topo(Father, Son, spec=levels)) mosaic(hauser.topo, ~Father+Son, main="Topological model", gp=shading_Friendly) hauser.RC <- update(hauser.indep, ~ . + Mult(Father, Son), verbose=FALSE) mosaic(hauser.RC, ~Father+Son, main="RC model", gp=shading_Friendly) LRstats(hauser.RC) # crossings models hauser.CR <- update(hauser.indep, ~ . + Crossings(Father,Son)) mosaic(hauser.topo, ~Father+Son, main="Crossings model", gp=shading_Friendly) LRstats(hauser.CR) hauser.CRdiag <- update(hauser.indep, ~ . + Crossings(Father,Son) + Diag(Father,Son)) LRstats(hauser.CRdiag) # compare model fit statistics modlist <- glmlist(hauser.indep, hauser.roweff, hauser.UA, hauser.UAdiag, hauser.quasi, hauser.qsymm, hauser.topo, hauser.RC, hauser.CR, hauser.CRdiag) sumry <- LRstats(modlist) sumry[order(sumry$AIC, decreasing=TRUE),] # or, more simply LRstats(modlist, sortby="AIC") mods <- substring(rownames(sumry),8) with(sumry, {plot(Df, AIC, cex=1.3, pch=19, xlab='Degrees of freedom', ylab='AIC') text(Df, AIC, mods, adj=c(0.5,-.5), col='red', xpd=TRUE) }) } \keyword{datasets} vcdExtra/man/Alligator.Rd0000644000176200001440000000410513163461153015015 0ustar liggesusers\name{Alligator} \alias{Alligator} \docType{data} \title{ Alligator Food Choice } \description{ The Alligator data, from Agresti (2002), comes from a study of the primary food choices of alligators in four Florida lakes. Researchers classified the stomach contents of 219 captured alligators into five categories: Fish (the most common primary food choice), Invertebrate (snails, insects, crayfish, etc.), Reptile (turtles, alligators), Bird, and Other (amphibians, plants, household pets, stones, and other debris). } \usage{data(Alligator)} \format{ A frequency data frame with 80 observations on the following 5 variables. \describe{ \item{\code{lake}}{a factor with levels \code{George} \code{Hancock} \code{Oklawaha} \code{Trafford}} \item{\code{sex}}{a factor with levels \code{female} \code{male}} \item{\code{size}}{alligator size, a factor with levels \code{large} (>2.3m) \code{small} (<=2.3m)} \item{\code{food}}{primary food choice, a factor with levels \code{bird} \code{fish} \code{invert} \code{other} \code{reptile}} \item{\code{count}}{cell frequency, a numeric vector} } } \details{ The table contains a fair number of 0 counts. \code{food} is the response variable. \code{fish} is the most frequent choice, and often taken as a baseline category in multinomial response models. } \source{ Agresti, A. (2002). \emph{Categorical Data Analysis}, New York: Wiley, 2nd Ed., Table 7.1 } %\references{ %%% ~~ possibly secondary sources and usages ~~ %} \examples{ data(Alligator) # change from frequency data.frame to table allitable <- xtabs(count~lake+sex+size+food, data=Alligator) # Agresti's Table 7.1 structable(food~lake+sex+size, allitable) plot(allitable, shade=TRUE) # mutual independence model mosaic(~food+lake+size, allitable, shade=TRUE) # food jointly independent of lake and size mosaic(~food+lake+size, allitable, shade=TRUE, expected=~lake:size+food) if (require(nnet)) { # multinomial logit model mod1 <- multinom(food ~ lake+size+sex, data=Alligator, weights=count) } } \keyword{datasets} vcdExtra/man/cutfac.Rd0000644000176200001440000000606013163461153014346 0ustar liggesusers\name{cutfac} \alias{cutfac} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Cut a Numeric Variable to a Factor } \description{ \code{cutfac} acts like \code{\link[base]{cut}}, dividing the range of \code{x} into intervals and coding the values in \code{x} according in which interval they fall. However, it gives nicer labels for the factor levels and by default chooses convenient breaks among the values based on deciles. It is particularly useful for plots in which one wants to make a numeric variable discrete for the purpose of getting boxplots, spinograms or mosaic plots. } \usage{ cutfac(x, breaks = NULL, q = 10) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{a numeric vector which is to be converted to a factor by cutting } \item{breaks}{ either a numeric vector of two or more unique cut points or a single number (greater than or equal to 2) giving the number of intervals into which \code{x} is to be cut. } \item{q}{ the number of quantile groups used to define \code{breaks}, if that has not been specified. } } \details{ By default, \code{\link[base]{cut}} chooses breaks by equal lengths of the range of \code{x}, whereas \code{cutfac} uses \code{\link[stats]{quantile}} to choose breaks of roughly equal count. } \value{ A \code{\link[base]{factor}} corresponding to \code{x} is returned %% If it is a LIST, use %% \item{comp1 }{Description of 'comp1'} %% \item{comp2 }{Description of 'comp2'} %% ... } \references{ Friendly, M. and Meyer, D. (2016). \emph{Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data}. Boca Raton, FL: Chapman & Hall/CRC. \url{http://ddar.datavis.ca}. } \author{ Achim Zeileis } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[base]{cut}}, \code{\link[stats]{quantile}} } \examples{ if (require(AER)) { data("NMES1988", package="AER") nmes <- NMES1988[, c(1, 6:8, 13, 15, 18)] plot(log(visits+1) ~ cutfac(chronic), data = nmes, ylab = "Physician office visits (log scale)", xlab = "Number of chronic conditions", main = "chronic") plot(log(visits+1) ~ cutfac(hospital, c(0:2, 8)), data = nmes, ylab = "Physician office visits (log scale)", xlab = "Number of hospital stays", main = "hospital") } %\donttest{ %# countreg not yet on CRAN %if (require(countreg)) { %data("CrabSatellites", package = "countreg") % %# jittered scatterplot %plot(jitter(satellites) ~ width, data=CrabSatellites, % ylab="Number of satellites (jittered)", xlab="Carapace width", % cex.lab=1.25) %with(CrabSatellites, lines(lowess(width, satellites), col="red", lwd=2)) % %# boxplot, using deciles %plot(satellites ~ cutfac(width), data=CrabSatellites, % ylab="Number of satellites", xlab="Carapace width (deciles)") %} } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{manip} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/HairEyePlace.Rd0000644000176200001440000000332313163461153015373 0ustar liggesusers\name{HairEyePlace} \alias{HairEyePlace} \docType{data} \title{ Hair Color and Eye Color in Caithness and Aberdeen } \description{ A three-way frequency table crossing eye color and hair color in two places, Caithness and Aberdeen, Scotland. These data were of interest to Fisher (1940) and others because there are mixtures of people of Nordic, Celtic and Anglo-Saxon origin. One or both tables have been widely analyzed in conjunction with RC and canonical correlation models for categorical data, e.g., Becker and Clogg (1989). } \usage{data(HairEyePlace)} \format{ The format is: num [1:4, 1:5, 1:2] 326 688 343 98 38 116 84 48 241 584 ... - attr(*, "dimnames")=List of 3 ..$ Eye : chr [1:4] "Blue" "Light" "Medium" "Dark" ..$ Hair : chr [1:5] "Fair" "Red" "Medium" "Dark" ... ..$ Place: chr [1:2] "Caithness" "Aberdeen" } \details{ The hair and eye colors are ordered as in the original source, suggesting that they form ordered categories. } \source{ This data was taken from the \code{colors} data in \pkg{logmult}. } \references{ Becker, M. P., and Clogg, C. C. (1989). Analysis of Sets of Two-Way Contingency Tables Using Association Models. \emph{Journal of the American Statistical Association}, 84(405), 142-151. Fisher, R.A. (1940) The precision of discriminant functions. \emph{Annals of Eugenics}, 10, 422-429. } \examples{ data(HairEyePlace) # separate mosaics mosaic(HairEyePlace[,,1], shade=TRUE, main="Caithness") mosaic(HairEyePlace[,,2], shade=TRUE, main="Aberdeen") # condition on Place mosaic(~Hair + Eye |Place, data=HairEyePlace, shade=TRUE, legend=FALSE) cotabplot(~Hair+Eye|Place, data=HairEyePlace, shade=TRUE, legend=FALSE) } \keyword{datasets} vcdExtra/man/mosaic.glm.Rd0000644000176200001440000002006613163461153015134 0ustar liggesusers\name{mosaic.glm} \alias{mosaic.glm} \alias{sieve.glm} \alias{assoc.glm} %- Also NEED an '\alias' for EACH other topic documented here. \title{Mosaic plots for fitted generalized linear and generalized nonlinear models } \description{ Procduces mosaic plots (and other plots in the \code{\link[vcd]{strucplot}} framework) for a log-linear model fitted with \code{\link[stats]{glm}} or for a generalized nonlinear model fitted with \code{\link[gnm]{gnm}}. These methods extend the range of strucplot visualizations well beyond the models that can be fit with \code{\link[MASS]{loglm}}. They are intended for models for counts using the Poisson family (or quasi-poisson), but should be sensible as long as (a) the response variable is non-negative and (b) the predictors visualized in the \code{strucplot} are discrete factors. } \usage{ \method{mosaic}{glm}(x, formula = NULL, panel = mosaic, type = c("observed", "expected"), residuals = NULL, residuals_type = c("pearson", "deviance", "rstandard"), gp = shading_hcl, gp_args = list(), ...) \method{sieve}{glm}(x, ...) \method{assoc}{glm}(x, ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{ A \code{glm} or \code{gnm} object. The response variable, typically a cell frequency, should be non-negative. } \item{formula}{ A one-sided formula with the indexing factors of the plot separated by '+', determining the order in which the variables are used in the mosaic. A formula must be provided unless \code{x$data} inherits from class \code{"table"} -- in which case the indexing factors of this table are used, or the factors in \code{x$data} (or model.frame(x) if \code{x$data} is an environment) exactly cross-classify the data -- in which case this set of cross-classifying factors are used. } \item{panel}{Panel function used to draw the plot for visualizing the observed values, residuals and expected values. Currently, one of \code{"mosaic"}, \code{"assoc"}, or \code{"sieve"} in \code{vcd}.} \item{type}{A character string indicating whether the \code{"observed"} or the \code{"expected"} values of the table should be visualized by the area of the tiles or bars.} \item{residuals}{ An optional array or vector of residuals corresponding to the cells in the data, for example, as calculated by \code{residuals.glm(x)}, \code{residuals.gnm(x)}.} \item{residuals_type}{If the \code{residuals} argument is \code{NULL}, residuals are calculated internally and used in the display. In this case, \code{residual_type} can be \code{"pearson"}, \code{"deviance"} or \code{"rstandard"}. Otherwise (when \code{residuals} is supplied), \code{residuals_type} is used as a label for the legend in the plot. } \item{gp}{Object of class \code{"gpar"}, shading function or a corresponding generating function (see \code{\link[vcd]{strucplot}} Details and \code{\link[vcd]{shadings}}). Ignored if shade = FALSE.} \item{gp_args}{A list of arguments for the shading-generating function, if specified.} \item{\dots}{ Other arguments passed to the \code{panel} function e.g., \code{\link[vcd]{mosaic}} } } \details{ For both poisson family generalized linear models and loglinear models, standardized residuals provided by \code{rstandard} (sometimes called adjusted residuals) are often preferred because they have constant unit asymptotic variance. The \code{sieve} and \code{assoc} methods are simple convenience interfaces to this plot method, setting the panel argument accordingly. } %\note{ %In the current version, the \code{glm} or \code{gnm} object \emph{must} have been fit using %the \code{data} argument to supply a data.frame or table, rather than with variables %in the global environment. %} \value{ The \code{structable} visualized by \code{\link[vcd]{strucplot}} is returned invisibly. } %\references{ ~put references to the literature/web site here ~ } \author{Heather Turner, Michael Friendly, with help from Achim Zeileis} %\note{ %} \seealso{ \code{\link[stats]{glm}}, \code{\link[gnm]{gnm}}, \code{\link[vcd]{plot.loglm}}, \code{\link[vcd]{mosaic}} } \examples{ GSStab <- xtabs(count ~ sex + party, data=GSS) # using the data in table form mod.glm1 <- glm(Freq ~ sex + party, family = poisson, data = GSStab) res <- residuals(mod.glm1) std <- rstandard(mod.glm1) # For mosaic.default(), need to re-shape residuals to conform to data stdtab <- array(std, dim=dim(GSStab), dimnames=dimnames(GSStab)) mosaic(GSStab, gp=shading_Friendly, residuals=stdtab, residuals_type="Std\nresiduals", labeling = labeling_residuals) # Using externally calculated residuals with the glm() object mosaic.glm(mod.glm1, residuals=std, labeling = labeling_residuals, shade=TRUE) # Using residuals_type mosaic.glm(mod.glm1, residuals_type="rstandard", labeling = labeling_residuals, shade=TRUE) ## Ordinal factors and structured associations data(Mental) xtabs(Freq ~ mental+ses, data=Mental) long.labels <- list(set_varnames = c(mental="Mental Health Status", ses="Parent SES")) # fit independence model # Residual deviance: 47.418 on 15 degrees of freedom indep <- glm(Freq ~ mental+ses, family = poisson, data = Mental) long.labels <- list(set_varnames = c(mental="Mental Health Status", ses="Parent SES")) mosaic(indep,residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals) # or, show as a sieve diagram mosaic(indep, labeling_args = long.labels, panel=sieve, gp=shading_Friendly) # fit linear x linear (uniform) association. Use integer scores for rows/cols Cscore <- as.numeric(Mental$ses) Rscore <- as.numeric(Mental$mental) linlin <- glm(Freq ~ mental + ses + Rscore:Cscore, family = poisson, data = Mental) mosaic(linlin,residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, suppress=1, gp=shading_Friendly, main="Lin x Lin model") ## Goodman Row-Column association model fits even better (deviance 3.57, df 8) if (require(gnm)) { Mental$mental <- C(Mental$mental, treatment) Mental$ses <- C(Mental$ses, treatment) RC1model <- gnm(Freq ~ ses + mental + Mult(ses, mental), family = poisson, data = Mental) mosaic(RC1model,residuals_type="rstandard", labeling_args = long.labels, labeling=labeling_residuals, suppress=1, gp=shading_Friendly, main="RC1 model") } ############# UCB Admissions data, fit using glm() structable(Dept ~ Admit+Gender,UCBAdmissions) berkeley <- as.data.frame(UCBAdmissions) berk.glm1 <- glm(Freq ~ Dept * (Gender+Admit), data=berkeley, family="poisson") summary(berk.glm1) mosaic(berk.glm1, gp=shading_Friendly, labeling=labeling_residuals, formula=~Admit+Dept+Gender) # the same, displaying studentized residuals; note use of formula to reorder factors in the mosaic mosaic(berk.glm1, residuals_type="rstandard", labeling=labeling_residuals, shade=TRUE, formula=~Admit+Dept+Gender, main="Model: [DeptGender][DeptAdmit]") ## all two-way model berk.glm2 <- glm(Freq ~ (Dept + Gender + Admit)^2, data=berkeley, family="poisson") summary(berk.glm2) mosaic.glm(berk.glm2, residuals_type="rstandard", labeling = labeling_residuals, shade=TRUE, formula=~Admit+Dept+Gender, main="Model: [DeptGender][DeptAdmit][AdmitGender]") anova(berk.glm1, berk.glm2, test="Chisq") # Add 1 df term for association of [GenderAdmit] only in Dept A berkeley <- within(berkeley, dept1AG <- (Dept=='A')*(Gender=='Female')*(Admit=='Admitted')) berkeley[1:6,] berk.glm3 <- glm(Freq ~ Dept * (Gender+Admit) + dept1AG, data=berkeley, family="poisson") summary(berk.glm3) mosaic.glm(berk.glm3, residuals_type="rstandard", labeling = labeling_residuals, shade=TRUE, formula=~Admit+Dept+Gender, main="Model: [DeptGender][DeptAdmit] + DeptA*[GA]") anova(berk.glm1, berk.glm3, test="Chisq") } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{hplot} \keyword{models} \keyword{multivariate} % __ONLY ONE__ keyword per line vcdExtra/man/vcdExtra-package.Rd0000644000176200001440000001274313163461153016257 0ustar liggesusers\name{vcdExtra-package} \alias{vcdExtra-package} \alias{vcdExtra} \docType{package} \title{ Extensions and additions to vcd: Visualizing Categorical Data } \description{ This package provides additional data sets, documentation, and a few functions designed to extend the \code{vcd} package for Visualizing Categorical Data and the \code{gnm} package for Generalized Nonlinear Models. In particular, vcdExtra extends mosaic, assoc and sieve plots from vcd to handle glm() and gnm() models and adds a 3D version in \code{\link{mosaic3d}}. This package is now a support package for the book, \emph{Discrete Data Analysis with R} by Michael Friendly and David Meyer, Chapman & Hall/CRC, 2016, \url{https://www.crcpress.com/Discrete-Data-Analysis-with-R-Visualization-and-Modeling-Techniques-for/Friendly-Meyer/9781498725835} with a number of additional data sets, and functions. The web site for the book is \url{http://ddar.datavis.ca}. } \details{ \tabular{ll}{ Package: \tab vcdExtra\cr Type: \tab Package\cr Version: \tab 0.7-1\cr Date: \tab 2017-09-28\cr License: \tab GPL version 2 or newer\cr LazyLoad: \tab yes\cr } The main purpose of this package is to serve as a sandbox for introducing extensions of mosaic plots and related graphical methods that apply to loglinear models fitted using \code{glm()} and related, generalized nonlinear models fitted with \code{gnm()} in the \code{\link[gnm]{gnm-package}} package. A related purpose is to fill in some holes in the analysis of categorical data in R, not provided in base R, the \pkg{vcd}, or other commonly used packages. The method \code{\link{mosaic.glm}} extends the \code{\link[vcd]{mosaic.loglm}} method in the \pkg{vcd} package to this wider class of models. This method also works for the generalized nonlinear models fit with the \code{\link[gnm]{gnm-package}} package, including models for square tables and models with multiplicative associations. \code{\link{mosaic3d}} introduces a 3D generalization of mosaic displays using the \pkg{rgl} package. In addition, there are several new data sets, a tutorial vignette, \describe{ \item{vcd-tutorial}{Working with categorical data with R and the vcd package, \code{vignette("vcd-tutorial", package = "vcdExtra") }} } and a few functions for manipulating categorical data sets and working with models for categorical data. A new class, \code{\link{glmlist}}, is introduced for working with collections of \code{glm} objects, e.g., \code{\link{Kway}} for fitting all K-way models from a basic marginal model, and \code{\link{LRstats}} for brief statistical summaries of goodnes-of-fit for a collection of models. For square tables with ordered factors, \code{\link{Crossings}} supplements the specification of terms in model formulas using \code{\link[gnm]{Symm}}, \code{\link[gnm]{Diag}}, \code{\link[gnm]{Topo}}, etc. in the \code{\link[gnm]{gnm-package}}. Some of these extensions may be migrated into vcd or gnm. A collection of demos is included to illustrate fitting and visualizing a wide variety of models: \describe{ \item{mental-glm}{Mental health data: mosaics for glm() and gnm() models} \item{occStatus}{Occupational status data: Compare mosaic using expected= to mosaic.glm} \item{ucb-glm}{UCBAdmissions data: Conditional independence via loglm() and glm()} \item{vision-quasi}{VisualAcuity data: Quasi- and Symmetry models} \item{yaish-unidiff}{Yaish data: Unidiff model for 3-way table} \item{Wong2-3}{Political views and support for women to work (U, R, C, R+C and RC(1) models)} \item{Wong3-1}{Political views, support for women to work and national welfare spending (3-way, marginal, and conditional independence models)} \item{housing}{Visualize glm(), multinom() and polr() models from \code{example(housing, package="MASS")}} } Use \code{ demo(package="vcdExtra")} for a complete current list. The \pkg{vcdExtra} package now contains a large number of data sets illustrating various forms of categorical data analysis and related visualizations, from simple to advanced. Use \code{data(package="vcdExtra")} for a complete list, or \code{datasets(package="vcdExtra")} for an annotated one showing the \code{class} and \code{dim} for each data set. } \author{ Michael Friendly Maintainer: Michael Friendly } \references{ Friendly, M. \emph{Visualizing Categorical Data}, Cary NC: SAS Insitute, 2000. Web materials: \url{http://www.datavis.ca/books/vcd/}. Friendly, M. and Meyer, D. (2016). \emph{Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data}. Boca Raton, FL: Chapman & Hall/CRC. \url{http://ddar.datavis.ca}. Meyer, D.; Zeileis, A. & Hornik, K. The Strucplot Framework: Visualizing Multi-way Contingency Tables with vcd \emph{Journal of Statistical Software}, 2006, \bold{17}, 1-48. Available in R via \code{vignette("strucplot", package = "vcd")} Turner, H. and Firth, D. \emph{Generalized nonlinear models in R: An overview of the gnm package}, 2007, \url{http://eprints.ncrm.ac.uk/472/}. Available in R via \code{vignette("gnmOverview", package = "gnm")}. } \keyword{ package } \seealso{ \code{\link[gnm]{gnm-package}}, for an extended range of models for contingency tables \code{\link[vcd]{mosaic}} for details on mosaic displays within the strucplot framework. %~~ Optional links to other man pages, e.g. ~~ %~~ \code{\link[:-package]{}} ~~ } \examples{ example(mosaic.glm) demo("mental-glm") } vcdExtra/man/Abortion.Rd0000644000176200001440000000263713163461153014664 0ustar liggesusers\name{Abortion} \Rdversion{1.1} \alias{Abortion} \docType{data} \title{Abortion Opinion Data} \description{Opinions about abortion classified by gender and SES} \usage{ data(Abortion) } \format{ A 3-dimensional array resulting from cross-tabulating 3 variables for 1100 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{Sex}\tab \code{"Female", "Male"}\cr 2\tab \code{Status}\tab \code{"Lo", "Hi"}\cr 3\tab \code{Support_Abortion}\tab \code{"Yes", "No"}\cr } } \details{ The combinations of \code{Sex} and \code{Status} represent four independent samples, having fixed \code{Sex}-\code{Status} marginal totals. Thus the \code{Sex:Status} association must be included in any loglinear model. \code{Support_Abortion} is a natural response variable. } \source{ % \cite{Christensen:90 [p. 92]} Christensen, R. (1990). \emph{Log-Linear Models}, New York, NY: Springer-Verlag, p. 92, Example 3.5.2. Christensen, R. (1997). \emph{Log-Linear Models and Logistic Regression}, New York, NY: Springer, p. 100, Example 3.5.2. } %\references{ %} %\seealso { } \examples{ data(Abortion) # example goes here ftable(Abortion) mosaic(Abortion, shade=TRUE) # stratified by Sex fourfold(aperm(Abortion, 3:1)) # stratified by Status fourfold(aperm(Abortion, c(3,1,2))) } \keyword{datasets} vcdExtra/man/HLtest.Rd0000644000176200001440000000707113163461153014307 0ustar liggesusers\name{HLtest} \alias{HosmerLemeshow} \alias{HLtest} \alias{plot.HLtest} \alias{print.HLtest} \alias{rootogram.HLtest} \alias{summary.HLtest} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Hosmer-Lemeshow Goodness of Fit Test } \description{ The \code{HLtest} function computes the classical Hosmer-Lemeshow (1980) goodness of fit test for a binomial \code{glm} object in logistic regression The general idea is to assesses whether or not the observed event rates match expected event rates in subgroups of the model population. The Hosmer-Lemeshow test specifically identifies subgroups as the deciles of fitted event values, or other quantiles as determined by the \code{g} argument. Given these subgroups, a simple chisquare test on \code{g-2} df is used. In addition to \code{print} and \code{summary} methods, a \code{plot} method is supplied to visualize the discrepancies between observed and fitted frequencies. } \usage{ HosmerLemeshow(model, g = 10) HLtest(model, g = 10) \method{print}{HLtest}(x, ...) \method{summary}{HLtest}(object, ...) \method{plot}{HLtest}(x, ...) \method{rootogram}{HLtest}(x, ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{model}{ A \code{glm} model object in the \code{binomial} family } \item{g}{ Number of groups used to partition the fitted values for the GOF test. } \item{x, object}{ A \code{HLtest} object } \item{\dots}{ Other arguments passed down to methods } } %\details{ %%% ~~ If necessary, more details than the description above ~~ %} \value{ A class \code{HLtest} object with the following components: %% If it is a LIST, use \item{table}{A data.frame describing the results of partitioning the data into \code{g} groups with the following columns: \code{cut}, \code{total}, \code{obs}, \code{exp}, \code{chi}} \item{chisq}{The chisquared statistics} \item{df}{Degrees of freedom} \item{p.value}{p value} \item{groups}{Number of groups} \item{call}{\code{model} call} %% ... } \references{ Hosmer, David W., Lemeshow, Stanley (1980). A goodness-of-fit test for multiple logistic regression model. \emph{Communications in Statistics, Series A}, 9, 1043-1069. Hosmer, David W., Lemeshow, Stanley (2000). \emph{Applied Logistic Regression}, New York: Wiley, ISBN 0-471-61553-6 Lemeshow, S. and Hosmer, D.W. (1982). A review of goodness of fit statistics for use in the development of logistic regression models. \emph{American Journal of Epidemiology}, 115(1), 92-106. } \author{ Michael Friendly } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[vcd]{rootogram}}, ~~~ } \examples{ data(birthwt, package="MASS") # how to do this without attach? attach(birthwt) race = factor(race, labels = c("white", "black", "other")) ptd = factor(ptl > 0) ftv = factor(ftv) levels(ftv)[-(1:2)] = "2+" bwt <- data.frame(low = factor(low), age, lwt, race, smoke = (smoke > 0), ptd, ht = (ht > 0), ui = (ui > 0), ftv) detach(birthwt) options(contrasts = c("contr.treatment", "contr.poly")) BWmod <- glm(low ~ ., family=binomial, data=bwt) (hlt <- HLtest(BWmod)) str(hlt) summary(hlt) plot(hlt) # basic model BWmod0 <- glm(low ~ age, family=binomial, data=bwt) (hlt0 <- HLtest(BWmod0)) str(hlt0) summary(hlt0) plot(hlt0) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{htest} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Donner.Rd0000644000176200001440000000743513163461153014335 0ustar liggesusers\name{Donner} \alias{Donner} \docType{data} \title{ Survival in the Donner Party } \description{ This data frame contains information on the members of the Donner Party, a group of people who attempted to migrate to California in 1846. They were trapped by an early blizzard on the eastern side of the Sierra Nevada mountains, and before they could be rescued, nearly half of the party had died. What factors affected who lived and who died? } \usage{data(Donner)} \format{ A data frame with 90 observations on the following 5 variables. \describe{ \item{\code{family}}{family name, a factor with 10 levels } \item{\code{age}}{age of person, a numeric vector} \item{\code{sex}}{a factor with levels \code{Female} \code{Male}} \item{\code{survived}}{a numeric vector, 0 or 1} \item{\code{death}}{date of death for those who died before rescue, a POSIXct} } } \details{ This data frame uses the person's name as row labels. \code{family} reflects a recoding of the last names of individuals to reduce the number of factor levels. The main families in the Donner party were: Donner, Graves, Breen and Reed. The families of Murphy, Foster and Pike are grouped as \code{'MurFosPik'}, those of Fosdick and Wolfinger are coded as \code{'FosdWolf'}, and all others as \code{'Other'}. } \source{ D. K. Grayson, 1990, "Donner party deaths: A demographic assessment", \emph{J. Anthropological Research}, \bold{46}, 223-242. Johnson, K. (1996). \emph{Unfortunate Emigrants: Narratives of the Donner Party}. Logan, UT: Utah State University Press. Additions, and dates of death from \url{http://user.xmission.com/~octa/DonnerParty/Roster.htm}. } \references{ Ramsey, F.L. and Schafer, D.W. (2002). \emph{The Statistical Sleuth: A Course in Methods of Data Analysis}, (2nd ed), Duxbury. Friendly, M. and Meyer, D. (2016). \emph{Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data}. Boca Raton, FL: Chapman & Hall/CRC. \url{http://ddar.datavis.ca}. } \seealso{ \code{\link[alr3]{donner}} in \pkg{alr3}, \code{\link[Sleuth2]{case2001}} in \pkg{Sleuth2}(adults only) provide similar data sets. } \examples{ # conditional density plots op <- par(mfrow=c(1,2), cex.lab=1.5) cdplot(factor(survived) ~ age, subset=sex=='Male', data=Donner, main="Donner party: Males", ylevels=2:1, ylab="Survived", yaxlabels=c("yes", "no")) with(Donner, rug(jitter(age[sex=="Male"]), col="white", quiet=TRUE)) cdplot(factor(survived) ~ age, subset=sex=='Female', data=Donner, main="Donner party: Females", ylevels=2:1, ylab="Survived", yaxlabels=c("yes", "no")) with(Donner, rug(jitter(age[sex=="Female"]), col="white", quiet=TRUE)) par(op) # fit some models (mod1 <- glm(survived ~ age + sex, data=Donner, family=binomial)) (mod2 <- glm(survived ~ age * sex, data=Donner, family=binomial)) anova(mod2, test="Chisq") (mod3 <- glm(survived ~ poly(age,2) * sex, data=Donner, family=binomial)) anova(mod3, test="Chisq") LRstats(glmlist(mod1, mod2, mod3)) # plot fitted probabilities from mod2 and mod3 # idea from: http://www.ling.upenn.edu/~joseff/rstudy/summer2010_ggplot2_intro.html library(ggplot2) # separate linear fits on age for M/F ggplot(Donner, aes(age, survived, color = sex)) + geom_point(position = position_jitter(height = 0.02, width = 0)) + stat_smooth(method = "glm", method.args = list(family = binomial), formula = y ~ x, alpha = 0.2, size=2, aes(fill = sex)) # separate quadratics ggplot(Donner, aes(age, survived, color = sex)) + geom_point(position = position_jitter(height = 0.02, width = 0)) + stat_smooth(method = "glm", method.args = list(family = binomial), formula = y ~ poly(x,2), alpha = 0.2, size=2, aes(fill = sex)) } \keyword{datasets} vcdExtra/man/GKgamma.Rd0000644000176200001440000000431013163461153014401 0ustar liggesusers\name{GKgamma} \alias{GKgamma} \alias{print.GKgamma} %- Also NEED an '\alias' for EACH other topic documented here. \title{Calculate Goodman-Kruskal Gamma for ordered tables} \description{ The Goodman-Kruskal \eqn{\gamma}{gamma} statistic is a measure of association for ordinal factors in a two-way table proposed by Goodman and Kruskal (1954). } \usage{ GKgamma(x, level = 0.95) %\method{print}{GKgamma}{x, digits = 3, ...} } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{A two-way frequency table, in matrix or table form. The rows and columns are considered to be ordinal factors} \item{level}{Confidence level for a significance test of \eqn{\gamma \ne =}{gamma !=0}} % \item{digits}{Number of digits printed by the print method} % \item{...}{Other arguments} } %\details{ % ~~ If necessary, more details than the description above ~~ %} \value{ Returns an object of class \code{"GKgamma"} with 6 components, as follows % If it is a LIST, use %\describe{ \item{gamma}{The gamma statistic} \item{C}{Total number of concordant pairs in the table} \item{D}{Total number of disconcordant pairs in the table} \item{sigma}{Standard error of gamma } \item{CIlevel}{Confidence level} \item{CI}{Confidence interval} % } } \references{ Agresti, A. \emph{Categorical Data Analysis}. John Wiley & Sons, 2002, pp. 57--59. Goodman, L. A., & Kruskal, W. H. (1954). Measures of association for cross classifications. \emph{Journal of the American Statistical Association}, 49, 732-764. Goodman, L. A., & Kruskal, W. H. (1963). Measures of association for cross classifications III: Approximate sampling theory. \emph{Journal of the American Statistical Association}, 58, 310-364. } \author{Michael Friendly; original version by Laura Thompson} %\note{ ~~further notes~~ % % ~Make other sections like Warning with \section{Warning }{....} ~ %} \seealso{\code{\link[vcd]{assocstats}}, \link[vcd]{Kappa}} \examples{ data(JobSat) GKgamma(JobSat) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{htest} \keyword{category} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Glass.Rd0000644000176200001440000000416413163461153014155 0ustar liggesusers\name{Glass} \alias{Glass} \docType{data} \title{ British Social Mobility from Glass(1954) } \description{ Glass(1954) gave this 5 x 5 table on the occupations of 3500 British fathers and their sons. } \usage{data("Glass")} \format{ A frequency data frame with 25 observations on the following 3 variables representing a 5 x 5 table with 3500 cases. \describe{ \item{\code{father}}{a factor with levels \code{Managerial} \code{Professional} \code{Skilled} \code{Supervisory} \code{Unskilled}} \item{\code{son}}{a factor with levels \code{Managerial} \code{Professional} \code{Skilled} \code{Supervisory} \code{Unskilled}} \item{\code{Freq}}{a numeric vector} } } \details{ The occupational categories in order of status are: (1) Professional \& High Administrative (2) Managerial, Executive \& High Supervisory (3) Low Inspectional \& Supervisory (4) Routine Nonmanual \& Skilled Manual (5) Semi- \& Unskilled Manual However, to make the point that factors are ordered alphabetically by default, Friendly \& Meyer (2016) introduce this data set in the form given here. } \source{ Glass, D. V. (1954), \emph{Social Mobility in Britain}. The Free Press. } \references{ Bishop, Y. M. M. and Fienberg, S. E. and Holland, P. W. (1975). \emph{Discrete Multivariate Analysis: Theory and Practice}, MIT Press. Friendly, M. and Meyer, D. (2016). \emph{Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data}. Boca Raton, FL: Chapman & Hall/CRC. \url{http://ddar.datavis.ca}. } \examples{ data(Glass) glass.tab <- xtabs(Freq ~ father + son, data=Glass) largs <- list(set_varnames=list(father="Father's Occupation", son="Son's Occupation"), abbreviate=10) gargs <- list(interpolate=c(1,2,4,8)) mosaic(glass.tab, shade=TRUE, labeling_args=largs, gp_args=gargs, main="Alphabetic order", legend=FALSE, rot_labels=c(20,90,0,70)) # reorder by status ord <- c(2, 1, 4, 3, 5) mosaic(glass.tab[ord, ord], shade=TRUE, labeling_args=largs, gp_args=gargs, main="Effect order", legend=FALSE, rot_labels=c(20,90,0,70)) } \keyword{datasets} vcdExtra/man/Toxaemia.Rd0000644000176200001440000000456213163461153014655 0ustar liggesusers\name{Toxaemia} \alias{Toxaemia} \docType{data} \title{ Toxaemia Symptoms in Pregnancy } \description{ Brown et al (1983) gave these data on two signs of toxaemia, an abnormal condition during pregnancy characterized by high blood pressure (hypertension) and high levels of protein in the urine. If untreated, both the mother and baby are at risk of complications or death. The data frame \code{Toxaemia} represents 13384 expectant mothers in Bradford, England in their first pregnancy, who were also classified according to social class and the number of cigarettes smoked per day. } \usage{data(Toxaemia)} \format{ A data frame in frequency form representing a 5 x 3 x 2 x 2 contingency table, with 60 observations on the following 5 variables. \describe{ \item{\code{class}}{Social class of mother, a factor with levels \code{1} \code{2} \code{3} \code{4} \code{5}} \item{\code{smoke}}{Cigarettes smoked per day during pregnancy, a factor with levels \code{0} \code{1-19} \code{20+}} \item{\code{hyper}}{Hypertension level, a factor with levels \code{Low} \code{High}} \item{\code{urea}}{Protein urea level, a factor with levels \code{Low} \code{High}} \item{\code{Freq}}{frequency in each cell, a numeric vector} } } %\details{ %%% ~~ If necessary, more details than the __description__ above ~~ %} \source{ Brown, P. J., Stone, J. and Ord-Smith, C. (1983), Toxaemic signs during pregnancy. \emph{JRSS, Series C, Applied Statistics}, 32, 69-72 } \references{ Friendly, M. (2000), \emph{Visualizing Categorical Data}, SAS Institute, Cary, NC, Example 7.15. Friendly, M. and Meyer, D. (2016). \emph{Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data}. Boca Raton, FL: Chapman & Hall/CRC. \url{http://ddar.datavis.ca}. Example 10.10. } \examples{ data(Toxaemia) tox.tab <- xtabs(Freq~class+smoke+hyper+urea,Toxaemia) ftable(tox.tab, row.vars=1) # symptoms by smoking mosaic(~smoke+hyper+urea, data=tox.tab, shade=TRUE) # symptoms by social class mosaic(~class+hyper+urea, data=tox.tab, shade=TRUE) # predictors mosaic(~smoke+class, data=tox.tab, shade=TRUE) # responses mosaic(~hyper+urea, data=tox.tab, shade=TRUE) # log odds ratios for urea and hypertension, by class and smoke \dontrun{ LOR <-loddsratio(aperm(tox.tab)) LOR } } \keyword{datasets} vcdExtra/man/mosaic.glmlist.Rd0000644000176200001440000001270513163461153016031 0ustar liggesusers\name{mosaic.glmlist} \alias{mosaic.glmlist} \alias{mosaic.loglmlist} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Mosaic Displays for \code{glmlist} and \code{logllmlist} Objects } \description{ This function provides a convenient interface for viewing mosaic displays associated with a collection of glm models for freqency tables that have been stored in a \code{glmlist} or \code{loglmlist} object. You can plot either selected models individually, or mosaics for all models in an array of viewports. } \usage{ \method{mosaic}{glmlist}(x, selection, panel=mosaic, type=c("observed", "expected"), legend=ask | !missing(selection), main=NULL, ask=TRUE, graphics=TRUE, rows, cols, newpage=TRUE, ...) \method{mosaic}{loglmlist}(x, selection, panel=mosaic, type=c("observed", "expected"), legend=ask | !missing(selection), main=NULL, ask=TRUE, graphics=TRUE, rows, cols, newpage=TRUE, ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{ a \code{glmlist} or \code{loglmlist} object } \item{selection}{ the index or name of one \code{glm} or \code{loglm} object in \code{x}. If no selection is specified, a menu of models is presented or all models are plotted. } \item{panel}{ a \code{\link[vcd]{strucplot}} panel function, typicially \code{\link[vcd]{mosaic}} or \code{\link[vcd]{sieve}} } \item{type}{ a character string indicating whether the \code{"observed"} or the \code{"expected"} values of the table should be visualized } \item{legend}{ logical: show a legend for residuals in the mosaic display(s)? The default behavior is to include a legend when only a single plot is shown, i.e., if \code{ask} is \code{TRUE} or a \code{selection} has been specified. } \item{main}{ either a logical, or a vector of character strings used for plotting the main title. If main is a logical and \code{TRUE}, the name of the selected glm object is used. } \item{ask}{ logical: should the function display a menu of models, when one is not specified in \code{selection}? If \code{selection} is not supplied and \code{ask} is \code{TRUE} (the default), a menu of model names is presented; if \code{ask} is \code{FALSE}, mosaics for all models are plotted in an array. } \item{graphics}{ logical: use a graphic dialog box when \code{ask=TRUE}? } \item{rows,cols}{ when \code{ask=FALSE}, the number of rows and columns in which to plot the mosaics. } \item{newpage}{ start a new page? (only applies to \code{ask=FALSE}) } \item{\dots}{ other arguments passed to \code{\link{mosaic.glm}} and ultimately to \code{\link[vcd]{mosaic}}. } } \details{ Most details of the plots produced can be controlled via \dots arguments as shown in some of the examples below. In particular, with \code{panel=sieve} you need to also pass \code{gp=shading_Friendly} to get a color version. } \value{ Returns the result of \code{\link{mosaic.glm}}. %% ~Describe the value returned %% If it is a LIST, use %% \item{comp1 }{Description of 'comp1'} %% \item{comp2 }{Description of 'comp2'} %% ... } \references{ David Meyer, Achim Zeileis, and Kurt Hornik (2006). The Strucplot Framework: Visualizing Multi-Way Contingency Tables with vcd. \emph{Journal of Statistical Software}, 17(3), 1-48. \url{http://www.jstatsoft.org/v17/i03/}, available as \code{vignette("strucplot", package="vcd")}. } \author{ Michael Friendly } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link{glmlist}}, \code{\link{loglmlist}}, \code{\link{Kway}} \code{\link{mosaic.glm}}, \code{\link[vcd]{mosaic}}, \code{\link[vcd]{strucplot}}, for the many parameters that control the details of mosaic plots. } \examples{ data(JobSatisfaction, package="vcd") # view all pairwise mosaics pairs(xtabs(Freq~management+supervisor+own, data=JobSatisfaction), shade=TRUE, diag_panel=pairs_diagonal_mosaic) modSat <- Kway(Freq ~ management+supervisor+own, data=JobSatisfaction, family=poisson, prefix="JobSat") names(modSat) \dontrun{ mosaic(modSat) # uses menu, if interactive() } mosaic(modSat, "JobSat.1") # model label mosaic(modSat, 2) # model index # supply a formula to determine the order of variables in the mosaic mosaic(modSat, 2, formula=~own+supervisor+management) mosaic(modSat, ask=FALSE) # uses viewports # use a different panel function, label the observed valued in the cells mosaic(modSat, 1, main=TRUE, panel=sieve, gp=shading_Friendly, labeling=labeling_values) data(Mental) indep <- glm(Freq ~ mental+ses, family = poisson, data = Mental) Cscore <- as.numeric(Mental$ses) Rscore <- as.numeric(Mental$mental) coleff <- glm(Freq ~ mental + ses + Rscore:ses, family = poisson, data = Mental) roweff <- glm(Freq ~ mental + ses + mental:Cscore, family = poisson, data = Mental) linlin <- glm(Freq ~ mental + ses + Rscore:Cscore, family = poisson, data = Mental) # assign names for the plot labels modMental <- glmlist(Indep=indep, ColEff=coleff, RowEff=roweff, `Lin x Lin`=linlin) mosaic(modMental, ask=FALSE, margins=c(3,1,1,2), labeling_args=list(abbreviate_labs=5)) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{hplot} \keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Mobility.Rd0000644000176200001440000000216213163461153014670 0ustar liggesusers\name{Mobility} \Rdversion{1.1} \alias{Mobility} \docType{data} \title{Social Mobility data} \description{Data on social mobility, recording the occupational category of fathers and their sons. } \usage{ data(Mobility) } \format{ A 2-dimensional array resulting from cross-tabulating 2 variables for 19912 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{Son's_Occupation}\tab \code{"UpNonMan", "LoNonMan", "UpManual", "LoManual", "Farm"}\cr 2\tab \code{Father's_Occupation}\tab \code{"UpNonMan", "LoNonMan", "UpManual", "LoManual", "Farm"}\cr } } %\details{ } \source{ Falguerolles, A. de and Mathieu, J. R. (1988). \emph{Proceedings of COMPSTAT 88}, Copenhagen, Denmark, Springer-Verlag. % \cite{FeathermanHauser:78} Featherman, D. L. and Hauser, R. M. Occupations and social mobility in the United States. \emph{Sociological Microjournal}, 12, Fiche 62. Copenhagen: Sociological Institute. } %\references{ %} %\seealso{ } \examples{ data(Mobility) # example goes here } \keyword{datasets} vcdExtra/man/Detergent.Rd0000644000176200001440000000327013163461153015022 0ustar liggesusers\name{Detergent} \Rdversion{1.1} \alias{Detergent} \docType{data} \title{Detergent preference data} \description{Cross-classification of a sample of 1008 consumers according to (a) the softness of the laundry water used, (b) previous use of detergent Brand M, (c) the termperature of laundry water used and (d) expressed preference for Brand X or Brand M in a blind trial.} \usage{ data(Detergent) } \format{ A 4-dimensional array resulting from cross-tabulating 4 variables for 1008 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{Temperature}\tab \code{"High", "Low"}\cr 2\tab \code{M_User}\tab \code{"Yes", "No"}\cr 3\tab \code{Preference}\tab \code{"Brand X", "Brand M"}\cr 4\tab \code{Water_softness}\tab \code{"Soft", "Medium", "Hard"}\cr } } %\details{ } \source{ % \cite{Fienberg:80 [p. 71]} Fienberg, S. E. (1980). \emph{The Analysis of Cross-Classified Categorical Data} Cambridge, MA: MIT Press, p. 71. } \references{ % \cite{RiesSmith:63} Ries, P. N. & Smith, H. (1963). The use of chi-square for preference testing in multidimensional problems. \emph{Chemical Engineering Progress}, 59, 39-43. } %\seealso{ } \examples{ data(Detergent) # example goes here mosaic(Detergent, shade=TRUE) require(MASS) (det.mod0 <- loglm(~ Preference + Temperature + M_User + Water_softness, data=Detergent)) # examine addition of two-way terms add1(det.mod0, ~ .^2, test="Chisq") # model for Preference as a response (det.mod1 <- loglm(~ Preference + (Temperature * M_User * Water_softness), data=Detergent)) mosaic(det.mod0) } \keyword{datasets} vcdExtra/man/mosaic3d.Rd0000644000176200001440000001654413163461153014613 0ustar liggesusers\name{mosaic3d} \Rdversion{1.1} \alias{mosaic3d} \alias{mosaic3d.default} \alias{mosaic3d.loglm} %- Also NEED an '\alias' for EACH other topic documented here. \title{ 3D Mosaic Plots } \description{ Produces a 3D mosaic plot for a contingency table (or a \code{link[MASS]{loglm}} model) using the \code{\link[rgl]{rgl-package}}. Generalizing the 2D mosaic plot, this begins with a given 3D shape (a unit cube), and successively sub-divides it along the X, Y, Z dimensions according to the table margins, generating a nested set of 3D tiles. The volume of the resulting tiles is therefore proportional to the frequency represented in the table cells. Residuals from a given loglinear model are then used to color or shade each of the tiles. This is a developing implementation. The arguments and details are subject to change. } \usage{ mosaic3d(x, ...) \method{mosaic3d}{loglm}(x, type = c("observed", "expected"), residuals_type = c("pearson", "deviance"), ...) \method{mosaic3d}{default}(x, expected = NULL, residuals = NULL, type = c("observed", "expected"), residuals_type = NULL, shape = rgl::cube3d(alpha = alpha), alpha = 0.5, spacing = 0.1, split_dir = 1:3, shading = shading_basic, interpolate=c(2,4), zero_size=.05, label_edge, labeling_args = list(), newpage = TRUE, box=FALSE, ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{ A \code{link[MASS]{loglm}} model object. Alternatively, a multidimensional \code{array} or \code{table} or \code{\link[vcd]{structable}} of frequencies in a contingency table. In the present implementation, the dimensions are taken in sequential order. Use \code{link[base]{aperm}} or \code{\link[vcd]{structable}} to change this. } \item{expected}{ optionally, for contingency tables, an array of expected frequencies of the same dimension as \code{x}, or alternatively the corresponding loglinear model specification as used by \code{link[stats]{loglin}} or \code{link[MASS]{loglm}} (see \code{\link[vcd]{structable}} for details).} \item{residuals}{ optionally, an array of residuals of the same dimension as \code{x} (see details). } \item{type}{ a character string indicating whether the \code{"observed"} or the \code{"expected"} frequencies in the table should be visualized by the volume of the 3D tiles. } \item{residuals_type}{ a character string indicating the type of residuals to be computed when none are supplied. If residuals is \code{NULL}, \code{residuals_type} must be one of \code{"pearson"} (default; giving components of Pearson's chi-squared), \code{"deviance"} (giving components of the likelihood ratio chi-squared), or \code{"FT"} for the Freeman-Tukey residuals. The value of this argument can be abbreviated. } \item{shape}{ The initial 3D shape on which the mosaic is based. Typically this is a call to an rgl function, and must produce a \code{shape3d} object. The default is a "unit cube" on (-1, +1), with transparency specified by \code{alpha}. } \item{alpha}{ Specifies the transparency of the 3D tiles used to compose the 3D mosaic. } \item{spacing}{ A number or vector giving the total amount of space used to separate the 3D tiles along each of the dimensions of the table. The values specified are re-cycled to the number of table dimensions. } \item{split_dir}{ A numeric vector composed of the integers \code{1:3} or a character vector composed of \code{c("x", "y", "z")}, where \code{split_dir[i]} specifies the axis along which the tiles should be split for dimension \code{i} of the table. The values specified are re-cycled to the number of table dimensions. } \item{shading}{ A function, taking an array or vector of residuals for the given model, returning a vector of colors. At present, only the default \code{shading=shading_basic} is provided. This is roughly equivalent to the use of the \code{shade} argument in \code{\link[graphics]{mosaicplot}} or to the use of \code{gp=shading_Friendly} in \code{\link[vcd]{mosaic}}. } \item{interpolate}{a vector of interpolation values for the \code{shading} function. } \item{zero_size}{ The radius of a small sphere used to mark zero cells in the display. } \item{label_edge}{ A character vector composed of \code{c("-", "+")} indicating whether the labels for a given table dimension are to be written at the minima (\code{"-"}) or maxima (\code{"+"}) of the \emph{other} dimensions in the plot. The default is \code{rep( c('-', '+'), each=3, length=ndim)}, meaning that the first three table variables are labeled at the minima, and successive ones at the maxima. } \item{labeling_args}{ This argument is intended to be used to specify details of the rendering of labels for the table dimensions, but at present has no effect. } \item{newpage}{ logical indicating whether a new page should be created for the plot or not. } \item{box}{ logical indicating whether a bounding box should be drawn around the plot. } \item{\dots}{ Other arguments passed down to \code{mosaic.default} or 3D functions. } } \details{ Friendly (1995), Friendly [Sect. 4.5](2000) and Theus and Lauer (1999) have all used the idea of 3D mosaic displays to explain various aspects of loglinear models (the iterative proportional fitting algorithm, the structure of various models for 3-way and n-way tables, etc.), but no implementation of 3D mosaics was previously available. For the default method, residuals, used to color and shade the 3D tiles, can be passed explicitly, or, more typically, are computed as needed from observed and expected frequencies. In this case, the expected frequencies are optionally computed for a specified loglinear model given by the \code{expected} argument. For the loglm method, residuals and observed frequencies are calculated from the model object. } \value{ Invisibly, the list of \code{shape3d} objects used to draw the 3D mosaic, with names corresponding to the concatenation of the level labels, separated by ":". } \references{ Friendly, M. (1995). Conceptual and Visual Models for Categorical Data, \emph{The American Statistician}, \bold{49}, 153-160. Friendly, M. \emph{Visualizing Categorical Data}, Cary NC: SAS Insitute, 2000. Web materials: \url{http://www.datavis.ca/books/vcd/}. Theus, M. & Lauer, S. R. W. (1999) Visualizing Loglinear Models. \emph{Journal of Computational and Graphical Statistics}, \bold{8}, 396-412. } \author{ Michael Friendly, with the help of Duncan Murdoch and Achim Zeileis } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[vcd]{strucplot}}, \code{\link[vcd]{mosaic}}, \code{\link[graphics]{mosaicplot}} \code{\link[stats]{loglin}}, \code{\link[MASS]{loglm}} for details on fitting loglinear models } \examples{ # 2 x 2 x 2 mosaic3d(Bartlett, box=TRUE) # compare with expected frequencies under model of mutual independence mosaic3d(Bartlett, type="expected", box=TRUE) # 2 x 2 x 3 mosaic3d(Heart, box=TRUE) \dontrun{ # 2 x 2 x 2 x 3 # illustrates a 4D table mosaic3d(Detergent) # compare 2D and 3D mosaics demo("mosaic-hec") } } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{hplot } %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/logseries.Rd0000644000176200001440000000670513163461153015103 0ustar liggesusers\name{logseries} \alias{Logseries} \alias{dlogseries} \alias{plogseries} \alias{qlogseries} \alias{rlogseries} %- Also NEED an '\alias' for EACH other topic documented here. \title{ The Logarithmic Series Distribution } \description{ The logarithmic series distribution is a long-tailed distribution introduced by Fisher etal. (1943) in connection with data on the abundance of individuals classified by species. These functions provide the density, distribution function, quantile function and random generation for the logarithmic series distribution with parameter \code{prob}. } \usage{ dlogseries(x, prob = 0.5, log = FALSE) plogseries(q, prob = 0.5, lower.tail = TRUE, log.p = FALSE) qlogseries(p, prob = 0.5, lower.tail = TRUE, log.p = FALSE, max.value = 10000) rlogseries(n, prob = 0.5) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x, q}{ vector of quantiles representing the number of events. } \item{prob}{ parameter for the distribution, \code{0 < prob < 1} } \item{log, log.p}{ ogical; if TRUE, probabilities \code{p} are given as \code{log(p)} } \item{lower.tail}{ logical; if TRUE (default), probabilities are \eqn{P[X \le x]}{P[X <= x]}, otherwise, \eqn{P[X > x]}{P[X > x]}. } \item{p}{ vector of probabilities } \item{max.value}{ maximum value returned by \code{qlogseries} } \item{n}{ number of observations for \code{rlogseries} } } \details{ The logarithmic series distribution with \code{prob} = \eqn{p} has density \deqn{ p ( x ) = \alpha p^x / x } for \eqn{x = 1, 2, \dots}, where \eqn{\alpha= -1 / \log(1 - p)} and \eqn{0 < p <1}. Note that counts \code{x==2} cannot occur. } \value{ \code{dlogseries} gives the density, \code{plogseries} gives the distribution function, \code{qlogseries} gives the quantile function, and \code{rlogseries} generates random deviates. %% ~Describe the value returned %% If it is a LIST, use %% \item{comp1 }{Description of 'comp1'} %% \item{comp2 }{Description of 'comp2'} %% ... } \references{ \url{http://en.wikipedia.org/wiki/Logarithmic_distribution} Fisher, R. A. and Corbet, A. S. and Williams, C. B. (1943). The relation between the number of species and the number of individuals \emph{Journal of Animal Ecology}, 12, 42-58. } \author{ Michael Friendly, using original code modified from the \code{gmlss.dist} package by Mikis Stasinopoulos. } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[stats]{Distributions}}, ~~~ } \examples{ XL <-expand.grid(x=1:5, p=c(0.33, 0.66, 0.99)) lgs.df <- data.frame(XL, prob=dlogseries(XL[,"x"], XL[,"p"])) lgs.df$p = factor(lgs.df$p) str(lgs.df) require(lattice) mycol <- palette()[2:4] xyplot( prob ~ x, data=lgs.df, groups=p, xlab=list('Number of events (k)', cex=1.25), ylab=list('Probability', cex=1.25), type='b', pch=15:17, lwd=2, cex=1.25, col=mycol, key = list( title = 'p', points = list(pch=15:17, col=mycol, cex=1.25), lines = list(lwd=2, col=mycol), text = list(levels(lgs.df$p)), x=0.9, y=0.98, corner=c(x=1, y=1) ) ) # random numbers hist(rlogseries(200, prob=.4), xlab='x') hist(rlogseries(200, prob=.8), xlab='x') } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{distribution} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Mental.Rd0000644000176200001440000000317413163461153014324 0ustar liggesusers\name{Mental} \Rdversion{1.1} \alias{Mental} \docType{data} \title{ Mental impariment and parents SES} \description{ A 6 x 4 contingency table representing the cross-classification of mental health status (\code{mental}) of 1660 young New York residents by their parents' socioeconomic status (\code{ses}). } \usage{data(Mental)} \format{ A data frame frequency table with 24 observations on the following 3 variables. \describe{ \item{\code{ses}}{an ordered factor with levels \code{1} < \code{2} < \code{3} < \code{4} < \code{5} < \code{6}} \item{\code{mental}}{an ordered factor with levels \code{Well} < \code{Mild} < \code{Moderate} < \code{Impaired}} \item{\code{Freq}}{cell frequency: a numeric vector} } } \details{ Both \code{ses} and \code{mental} can be treated as ordered factors or integer scores. For \code{ses}, 1="High" and 6="Low". } \source{ Haberman, S. J. \emph{The Analysis of Qualitative Data: New Developments}, Academic Press, 1979, Vol. II, p. 375. Srole, L.; Langner, T. S.; Michael, S. T.; Kirkpatrick, P.; Opler, M. K. & Rennie, T. A. C. \emph{Mental Health in the Metropolis: The Midtown Manhattan Study}, NYU Press, 1978, p. 289 } \references{ Friendly, M. \emph{Visualizing Categorical Data}, Cary, NC: SAS Institute, 2000, Appendix B.7. } \examples{ data(Mental) str(Mental) (Mental.tab <- xtabs(Freq ~ ses+mental, data=Mental)) # mosaic and sieve plots mosaic(Mental.tab, gp=shading_Friendly) sieve(Mental.tab, gp=shading_Friendly) library(ca) plot(ca(Mental.tab), main="Mental impairment & SES") title(xlab="Dim 1", ylab="Dim 2") } \keyword{datasets} vcdExtra/man/Mice.Rd0000644000176200001440000000303413163461153013754 0ustar liggesusers\name{Mice} \alias{Mice} \docType{data} \title{ Mice Depletion Data } \description{ Data from Kastenbaum and Lamphiear (1959). The table gives the number of depletions (deaths) in 657 litters of mice, classified by litter size and treatment. This data set has become a classic in the analysis of contingency tables, yet unfortunately little information on the details of the experiment has been published. } \usage{data("Mice")} \format{ A freqency data frame with 30 observations on the following 4 variables, representing a 5 x 2 x 3 contingency table. \describe{ \item{\code{litter}}{litter size, a numeric vector} \item{\code{treatment}}{treatment, a factor with levels \code{A} \code{B}} \item{\code{deaths}}{number of depletions, a factor with levels \code{0} \code{1} \code{2+}} \item{\code{Freq}}{cell frequency, a numeric vector} } } %\details{ %%% ~~ If necessary, more details than the __description__ above ~~ %} \source{ Goodman, L. A. (1983) The analysis of dependence in cross-classifications having ordered categories, using log-linear models for frequencies and log-linear models for odds. \emph{Biometrics}, 39, 149-160. } \references{ Kastenbaum, M. A. & Lamphiear, D. E. (1959) Calculation of chi-square to calculate the no three-factor interaction hypothesis. \emph{Biometrics}, 15, 107-115. } \examples{ data(Mice) # make a table ftable(mice.tab <- xtabs(Freq ~ litter + treatment + deaths, data=Mice)) library(vcd) mosaic(mice.tab, shade=TRUE) } \keyword{datasets} vcdExtra/man/vcdExtra-deprecated.Rd0000644000176200001440000000126013163461153016754 0ustar liggesusers\name{vcdExtra-deprecated} \alias{vcdExtra-deprecated} \alias{summarise} %\alias{summarise.glm} %\alias{summarise.glmlist} %\alias{summarise.loglm} %\alias{summarise.loglmlist} % \title{Deprecated Functions in vcdExtra Package} % \description{ These functions are provided for compatibility with older versions of the \pkg{vcdExtra} package only. They are replaced by \code{\link{LRstats}}. } % \usage{ summarise(...) %summarise.glm(...) %summarise.glmlist(...) %summarise.loglm(...) %summarise.loglmlist(...) } % \arguments{ \item{\dots}{pass arguments down.} } % \details{ \code{summarise.*} have been replaced by \code{\link{LRstats}} functions. } vcdExtra/man/zero.test.Rd0000644000176200001440000000471313163461153015041 0ustar liggesusers\name{zero.test} \alias{zero.test} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Score test for zero inflation in Poisson data } \description{ Carries out a simple score test (van den Broek, 1995) for excess zeros in an otherwise Poisson distribution of counts. It gives a \eqn{\chi^2_1} statistic on one degree of freedom. } \usage{ zero.test(x) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{ A vector of non-negative counts, or a one-way frequency table of such counts. } } \details{ The test first calculates the rate estimate from the mean, \eqn{\hat{\lambda} = \bar{x}}. The number of observed zeros, \eqn{n_0} is then compared with the expected number, \eqn{n \hat{p_0}}, where \eqn{\hat{p}_0=\exp[-\hat{\lambda}]}. Then the test statistic is calculated by the formula: \deqn{\frac{(n_0 - n\hat{p}_0)^2}{n\hat{p}_0(1-\hat{p}_0) - n\bar{x}\hat{p}_0^2}} This test statistic has a \eqn{\chi^2_1} distribution. } \value{ Returns invisibly a list of three elements: \item{\code{statistic}}{Description of 'comp1'} \item{\code{df}}{Description of 'comp2'} \item{\code{pvalue}}{Upper tail p-value} } \references{ The original R code came from a Stackexchange question, \url{https://stats.stackexchange.com/questions/118322/how-to-test-for-zero-inflation-in-a-dataset} Van den Broek, J. (1995). A Score Test for Zero Inflation in a Poisson Distribution. \emph{Biometrics}, \bold{51}(2), 738-743. \code{doi:10.2307/2532959} Yang, Zhao, James W. Hardin, and Cheryl L. Addy (2010). Score Tests for Zero-Inflation in Overdispersed Count Data. \emph{Communications in Statistics - Theory and Methods} \bold{39} (11) 2008-2030. \code{doi:10.1080/03610920902948228} } \author{ Michael Friendly } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ %\seealso{ %%% ~~objects to See Also as \code{\link{help}}, ~~~ %} \examples{ # synthetic tests zero.test(rpois(100, 1)) zero.test(rpois(100, 5)) # add some extra zeros zero.test(c(rep(0, 20), rpois(100, 5))) # Articles by Phd candidates data(PhdPubs, package="vcdExtra") zero.test(PhdPubs$articles) phd.tab <- table(PhdPubs$articles) zero.test(phd.tab) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{htest}% use one of RShowDoc("KEYWORDS") %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/seq_mosaic.Rd0000644000176200001440000000647613163461153015237 0ustar liggesusers\name{seq_mosaic} \alias{seq_mosaic} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Sequential Mosaics and Strucplots for an N-way Table } \description{ This function takes an n-way contingency table and plots mosaics for series of sequential models to the 1-, 2-, ... n-way marginal tables, corresponding to a variety of types of loglinear models. } \usage{ seq_mosaic(x, panel = mosaic, type = c("joint", "conditional", "mutual", "markov", "saturated"), plots = 1:nf, vorder = 1:nf, k = NULL, ...) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{x}{ a contingency table in array form, with optional category labels specified in the dimnames(x) attribute, or else a data.frame in frequency form, with the frequency variable named \code{"Freq"}. } \item{panel}{ a \code{\link[vcd]{strucplot}} panel function, typicially \code{\link[vcd]{mosaic}} or \code{\link[vcd]{sieve}. NOT yet implemented.} } \item{type}{ type of sequential model to fit, a character string. One of \code{"joint"}, \code{"conditional"}, \code{"mutual"}, \code{"markov"}, or \code{"saturated"}. } \item{plots}{ which marginal sub-tables to plot? A vector of a (sub)set of the integers, \code{1:nf} where \code{nf} is the number of factors in the full n-way table. } \item{vorder}{ order of variables, a permutation of the integers \code{1:nf}, used to reorder the variables in the original table for the purpose of fitting sequential marginal models. } \item{k}{ conditioning variable(s) for \code{type} = \code{"joint"}, \code{"conditional"} or Markov chain order for \code{type} = \code{"markov"} } \item{\dots}{ other arguments passed to \code{\link[vcd]{mosaic}}. } } \details{ This function produces similar plots to the use of \code{\link{mosaic.loglmlist}}, called with the result of \code{\link{seq_loglm}}. } \value{ None. Used for its side-effect of producing plots %% If it is a LIST, use %% \item{comp1 }{Description of 'comp1'} %% \item{comp2 }{Description of 'comp2'} %% ... } \references{ These functions were inspired by the original SAS implementation of mosaic displays, described in the \emph{User's Guide}, \url{http://www.datavis.ca/mosaics/mosaics.pdf} } \author{ Michael Friendly } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link{loglin-utilities}} for descriptions of sequential models, \code{\link{conditional}}, \code{\link{joint}}, \code{\link{mutual}}, \dots \code{\link{loglmlist}}, \code{\link{mosaic.loglmlist}}, \code{\link{seq_loglm}} \code{\link{mosaic.glm}}, \code{\link[vcd]{mosaic}}, \code{\link[vcd]{strucplot}}, for the many parameters that control the details of mosaic plots. } \examples{ data(Titanic, package="datasets") seq_mosaic(Titanic) # models of joint independence, Survived last seq_mosaic(Titanic, type="condit") seq_mosaic(Titanic, type="mutual") # other panel functions and options: presently BUGGED \dontrun{ seq_mosaic(Titanic, type="mutual", panel=sieve, gp=shading_Friendly, labeling=labeling_values) } } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{hplots} \keyword{ ~kwd2 }% __ONLY ONE__ keyword per line vcdExtra/man/Draft1970.Rd0000644000176200001440000000446613163461153014472 0ustar liggesusers\name{Draft1970} \alias{Draft1970} \docType{data} \title{ USA 1970 Draft Lottery Data } \description{ This data set gives the results of the 1970 US draft lottery, in the form of a data frame. } \usage{data(Draft1970)} \format{ A data frame with 366 observations on the following 3 variables. \describe{ \item{\code{Day}}{day of the year, 1:366} \item{\code{Rank}}{draft priority rank of people born on that day} \item{\code{Month}}{an ordered factor with levels \code{Jan} < \code{Feb} \dots < \code{Dec}} } } \details{ The draft lottery was used to determine the order in which elligible men would be called to the Selective Service draft. The days of the year (including February 29) were represented by the numbers 1 through 366 written on slips of paper. The slips were placed in separate plastic capsules that were mixed in a shoebox and then dumped into a deep glass jar. Capsules were drawn from the jar one at a time. The first number drawn was 258 (September 14), so all registrants with that birthday were assigned lottery number \code{Rank} 1. The second number drawn corresponded to April 24, and so forth. All men of draft age (born 1944 to 1950) who shared a birthdate would be called to serve at once. The first 195 birthdates drawn were later called to serve in the order they were drawn; the last of these was September 24. } \source{ Starr, N. (1997). Nonrandom Risk: The 1970 Draft Lottery, \emph{Journal of Statistics Education}, v.5, n.2 \url{http://www.amstat.org/publications/jse/v5n2/datasets.starr.html} } \references{ Fienberg, S. E. (1971), "Randomization and Social Affairs: The 1970 Draft Lottery," \emph{Science}, 171, 255-261. \url{http://en.wikipedia.org/wiki/Draft_lottery_(1969)} } \seealso{\code{\link{Draft1970table}} } \examples{ data(Draft1970) # scatterplot plot(Rank ~ Day, data=Draft1970) with(Draft1970, lines(lowess(Day, Rank), col="red", lwd=2)) abline(lm(Rank ~ Day, data=Draft1970), col="blue") # boxplots plot(Rank ~ Month, data=Draft1970, col="bisque") lm(Rank ~ Month, data=Draft1970) anova(lm(Rank ~ Month, data=Draft1970)) # make the table version Draft1970$Risk <- cut(Draft1970$Rank, breaks=3, labels=c("High", "Med", "Low")) with(Draft1970, table(Month, Risk)) } \keyword{datasets} vcdExtra/man/Dyke.Rd0000644000176200001440000000504413163461153013776 0ustar liggesusers\name{Dyke} \Rdversion{1.1} \alias{Dyke} \docType{data} \title{Sources of Knowledge of Cancer} \description{Observational data on a sample of 1729 individuals, cross-classified in a 2^5 table according to their sources of information (read newspapers, listen to the radio, do 'solid' reading, attend lectures) and whether they have good or poor knowledge regarding cancer. Knowledge of cancer is often treated as the response.} \usage{ data(Dyke) } \format{ A 5-dimensional array resulting from cross-tabulating 5 variables for 1729 observations. The variable names and their levels are: \tabular{rll}{ No \tab Name \tab Levels \cr 1\tab \code{Knowledge}\tab \code{"Good", "Poor"}\cr 2\tab \code{Reading}\tab \code{"No", "Yes"}\cr 3\tab \code{Radio}\tab \code{"No", "Yes"}\cr 4\tab \code{Lectures}\tab \code{"No", "Yes"}\cr 5\tab \code{Newspaper}\tab \code{"No", "Yes"}\cr } } %\details{ } \source{ % \cite{Fienberg:80 [Table 5-6]} Fienberg, S. E. (1980). \emph{The Analysis of Cross-Classified Categorical Data} Cambridge, MA: MIT Press, p. 85, Table 5-6. } \references{ Dyke, G. V. and Patterson, H. D. (1952). Analysis of factorial arrangements when the data are proportions. \emph{Biometrics}, 8, 1-12. Lindsey, J. K. (1993). \emph{Models for Repeated Measurements} Oxford, UK: Oxford University Press, p. 57. } %\seealso{ } \examples{ data(Dyke) # independence model mosaic(Dyke, shade=TRUE) # null model, Knowledge as response, independent of others require(MASS) dyke.mod0 <- loglm(~ Knowledge + (Reading * Radio * Lectures * Newspaper), data=Dyke) dyke.mod0 mosaic(dyke.mod0) # view as doubledecker plot Dyke <- Dyke[2:1,,,,] # make Good the highlighted value of Knowledge doubledecker(Knowledge ~ ., data=Dyke) # better version, with some options doubledecker(Knowledge ~ Lectures + Reading + Newspaper + Radio, data=Dyke, margins = c(1,6, length(dim(Dyke)) + 1, 1), fill_boxes=list(rep(c("white", gray(.90)),4)) ) # separate (conditional) plots for those who attend lectures and those who do not doubledecker(Knowledge ~ Reading + Newspaper + Radio, data=Dyke[,,,1,], main="Do not attend lectures", margins = c(1,6, length(dim(Dyke)) + 1, 1), fill_boxes=list(rep(c("white", gray(.90)),3)) ) doubledecker(Knowledge ~ Reading + Newspaper + Radio, data=Dyke[,,,2,], main="Attend lectures", margins = c(1,6, length(dim(Dyke)) + 1, 1), fill_boxes=list(rep(c("white", gray(.90)),3)) ) drop1(dyke.mod0, test="Chisq") } \keyword{datasets} vcdExtra/man/TV.Rd0000644000176200001440000000501513163461153013431 0ustar liggesusers\name{TV} \Rdversion{1.1} \alias{TV} \title{TV Viewing Data} \description{ This data set \code{TV} comprises a 5 x 11 x 3 contingency table based on audience viewing data from Neilsen Media Research for the week starting November 6, 1995. } \usage{data(TV)} \format{ A 5 x 11 x 3 array of cell frequencies with the following structure: \preformatted{ int [1:5, 1:11, 1:3] 146 244 233 174 294 151 181 161 183 281 ... - attr(*, "dimnames")=List of 3 ..$ Day : chr [1:5] "Monday" "Tuesday" "Wednesday" "Thursday" ... ..$ Time : chr [1:11] "8:00" "8:15" "8:30" "8:45" ... ..$ Network: chr [1:3] "ABC" "CBS" "NBC" } } \details{ The original data, \code{tv.dat}, contains two additional networks: "Fox" and "Other", with small frequencies. These levels were removed in the current version. There is also a fourth factor, transition State transition (turn the television Off, Switch channels, or Persist in viewing the current channel). The \code{TV} data here includes only the Persist observations. } \source{ The original data, \code{tv.dat}, came from the initial implementation of mosaic displays in R by Jay Emerson (1998). Similar data had been used by Hartigan and Kleiner (1984) as an illustration. } \references{ Friendly, M. and Meyer, D. (2016). \emph{Discrete Data Analysis with R: Visualization and Modeling Techniques for Categorical and Count Data}. Boca Raton, FL: Chapman & Hall/CRC. \url{http://ddar.datavis.ca}. Emerson, John W. Mosaic Displays in S-PLUS: A General Implementation and a Case Study. \emph{Statistical Graphics and Computing Newsletter}, 1998, 9(1), 17--23, \url{http://www.stat.yale.edu/~jay/R/mosaic/v91.pdf} Hartigan, J. A. & Kleiner, B. A Mosaic of Television Ratings. \emph{The American Statistician}, 1984, 38, 32-35. } \examples{ data(TV) structable(TV) doubledecker(TV) # reduce number of levels of Time TV.df <- as.data.frame.table(TV) levels(TV.df$Time) <- rep(c("8:00-8:59", "9:00-9:59", "10:00-10:44"), c(4, 4, 3)) TV2 <- xtabs(Freq ~ Day + Time + Network, TV.df) # re-label for mosaic display levels(TV.df$Time) <- c("8", "9", "10") # fit mode of joint independence, showing association of Network with Day*Time mosaic(~ Day + Network + Time, data = TV.df, expected = ~ Day:Time + Network, legend = FALSE) # with doubledecker arrangement mosaic(~ Day + Network + Time, data = TV.df, expected = ~ Day:Time + Network, split = c(TRUE, TRUE, FALSE), spacing = spacing_highlighting, legend = FALSE) } \keyword{datasets} vcdExtra/man/update.xtabs.Rd0000644000176200001440000000322013163461153015476 0ustar liggesusers\name{update.xtabs} \alias{update.xtabs} %- Also NEED an '\alias' for EACH other topic documented here. \title{ Update method for a \code{xtabs} object } \description{ Provides an \code{update} method for \code{"xtabs"} objects, typically by removing terms from the formula to collapse over them. } \usage{ \method{update}{xtabs}(object, formula., ..., evaluate = TRUE) } %- maybe also 'usage' for other objects documented here. \arguments{ \item{object}{An existing \code{"xtabs"} object} \item{formula.}{Changes to the formula ? see \code{\link[stats]{update.formula}} for details} \item{\dots}{Additional arguments to the call, or arguments with changed values. } \item{evaluate}{If \code{TRUE}, evaluate the new call else return the call} } %\details{ %%% ~~ If necessary, more details than the description above ~~ %} \value{ If \code{evaluate == TRUE}, the new \code{"xtabs"} object, otherwise the updated call } %\references{ %%% ~put references to the literature/web site here ~ %} \author{ Michael Friendly } %\note{ %%% ~~further notes~~ %} %% ~Make other sections like Warning with \section{Warning }{....} ~ \seealso{ \code{\link[stats]{update.formula}} for details on updates to model formulae \code{\link[base]{margin.table}} does something similar, \code{\link{collapse.table}} collapses category levels } \examples{ vietnam.tab <- xtabs(Freq ~ sex + year + response, data=Vietnam) update(vietnam.tab, formula = ~ . -year) } % Add one or more standard keywords, see file 'KEYWORDS' in the % R documentation directory. \keyword{models} %\keyword{ ~kwd2 }% __ONLY ONE__ keyword per line