Predictors of the number of males in primate groups.

Notes: The table summarizes statistical tests of the predictions in which the number of males is the dependent variable, and statistical models include the number of females on its own (i) or in combination with two other predictor variables that indicate aspects of female spatial distribution (ii) or in combination with one variable reflecting temporal distribution (iii to v). Predictions follow numbering given in the Introduction. For λ, we give a 95% two-tailed credible interval from the MCMC sample (n = 10,000). R2 is the average value across this same sample (although the range of variation was relatively narrow). For the number of females and other variables, we give the predicted effect, the mean coefficient from the MCMC sample, and the proportion of samples that supported the prediction. If a predictor has no influence on the number of males, we expect only about 50% of the samples to give regression coefficients in the predicted direction.