Comparison of models of the relationship between annual population growth and human-caused mortality for wolves in North America.

1Expanded model descriptions:(i) General additive model (GAM) that allowed regional differences, fit in the ‘mgcv’ package of R with cross-validation used to determine the optimum amount of smoothing. GAM models allow curvilinear functions if the data support curvature.(ii) General linear model (normal errors with log link) with no regional effect on slope and intercept.(iii) General linear model (normal errors with log link) that allowed regional differences in the slope and intercept.(iv) Constant total mortality (no effect of human offtake on total mortality), with regional differences.(v) Constant total mortality (no effect of human offtake on total mortality).2Number of parameters in the model (non-integer values are expected for general additive models).3The coefficient of determination (R2) adjusted for degrees of freedom.4Akaike model weight.