Pathogen-specific correlations.

Pathogen prevalence level by predictors in the final model of Table 2. Spearman's rank-order correlations were used for all variables other than annual precipitation. Two correlations are given for mean annual precipitation (“Yrly Precip”): the polynomial model uses the full sample, but without significance values, which are probably unreliable for the individual pathogen scores. The linear model is limited to the 78% of the sample where pathogens increase with precipitation (i.e., up to 2000 mm). Other variables were transformed as indicated in the text, except that frost was not dichotomized. Sample sizes are 180 for log temperature, log precipitation in dryest month, and number of frost months; sample sizes for the polynomial and linear correlations with mean annual precipitation are 185 and 143 respectively.Pathogen-specific correlations.