Generalised linear mixed models (binomial logit) describing the Scandinavian wolves’ selection of roe deer (1) as prey over moose (0) as a function of local roe deer density (Roe deer), local moose density (Moose), winter day number (Day) and snow depth (Snow).

Wolf territory is included as a random factor and the number of parameters in the model is indicated by K, AIC corrected for small sample size AICc, differences between models ΔAIC, the model weights AICcWt, and the log likelihood value for each model LL.