Model selection of 3 <i>a priori</i> hypotheses proposed to explain lynx occurrence patterns across 3 landscapes differing in the amount of suitable landscape-level land cover in 2 regions (Chapleau and Mississagi) within an area of 25 km² for each lynx track and pseudo-absence.

Asterisk (*) indicates significant coefficients at p<0.05, †indicates significance at p<0.1. aMeff.r is the residual of habitat regressed against mesh size (km2; see text), a measure of functional connectivity, within each lynx- and pseudo-absence area. bHabitat is the proportion of suitable habitat within each lynx- and pseudo-absence area based on the habitat suitability model. Model selection of 3 a priori hypotheses proposed to explain lynx occurrence patterns across 3 landscapes differing in the amount of suitable landscape-level land cover in 2 regions (Chapleau and Mississagi) within an area of 25 km2 for each lynx track and pseudo-absence.