Figure legends:
Figure 1 : Map of Ånnsjön and its position in Sweden including the locations where minnows where caught in the lake (L1, L2, L3) and the surrounding tributaries (S1, S2, S3). © Landmäteriet.
Figure 2: Position of the 35 digitized landmarks used in geometric morphometrics. Homologous landmarks (red dots): 1-4 : most posterior, dorsal, anterior, and ventral point of orbit; 5: insertion of pectoral fin; 6-8: most posterior, dorsal, and ventral margin of opercular (principal opercular bone); 9: tip of the snout; 12: dorsal transition head to body; 16: anterior insertion of dorsal fin; 17: posterior insertion of dorsal fin; 21: dorsal insertion of caudal fin; 22: ventral insertion of caudal fin; 26: posterior insertion of anal fin, 27: anterior insertion of anal fin; 30: anterior insertion of pelvic fin; 33: ventral transition head to body. Semi landmarks (blue dots) were based on proportional distances between homologous structures.
Figure 3: Results of geometric morphometrics. a) Shape differences between minnows caught in the lake (blue line) and the streams (green line). Shape-change outlines of Discriminant Function Analyses are magnified threefold. b) Ordination of shape based on Canonical Variate Analyses of minnows caught in the six different locations with confidence ellipses (probability 0.9) drawn for lake and stream habitats respectively.
Figure 4: Average proportions of gut content of minnows caught in the lake (L1, L2, L3) and the streams (S1, S2, S3).
Figure 5 : Relationship between proportion of zooplankton in the minnow guts and morphological distances (canonical variate 1), including results of Spearman’s rank correlation.