Fatty acids
On the basis of fatty acid composition, piscivorous lake trout were divided along a resource use axis into four groups (1-4; Fig. A2), containing 14, 16, 21, and 28 individuals, respectively (Figs. 2 and A2; Table 1). Average dissimilarity was 14.61 (SIMPER analysis); whereas, the most discriminating 26 fatty acids, explaining together ~89% of the separation among groups, were: 22:6n-3 (12.5 %), 18:1n-9 (10.8 %), 16:1n-7 (6.8 %), 20:5n-3 (5.0 %), 20:4n-6 (3.9 %), 18:2n-6 (3.8 %), 22:4n-3 (3.7 %), 16:0 (3.5%), 20:4n-3 (3.3%), 18:1n7 (3.3%), 20:2n-6 (3.1%), 14:0 (2.8%), 20:1n-9 (2.7%), 22:5n-6 (2.7%), 20:3n-3 (2.3%), 22:2n-6 (2.1%), 18:0 (2.0%), 18:3n-3 (1.9%), 18:4n-3 (1.8%), 22:4n-6 (1.7%), 20:1n-7 (1.5%), 22:5n-3 (1.4%), 21:5n-3 (1.3%), 22:1n-11 (1.2%), 20:0 (1.2%), 16:4n-3 (1.2%), and 16:2n-4 (1.1%) (Table 1). The first two axes of the fatty acids PCA explained 65.2 % of the variation in diet and the four groups were supported by PERMANOVA (F3,76 = 23.9, P < 0.01) and pairwise comparisons between all pairs (all P < 0.01; Bonferroni corrected). Spatial differences in fatty acids composition were found among arms (F4,76 = 3.2, P < 0.01). Pairwise comparisons identified differences between Smith and McVicar arms (P = 0.02; Bonferroni corrected; Fig. A3). Interaction between fatty acids groups and arms was not significant (p > 0.05). Finally, depth of capture did not differ among fatty acid groups (p > 0.05). For all groups, most lake trout were caught between 0 and 20 m (Fig. A4).