Electronic Supplementary Material ESM (for publication as supporting information on the Proc R Soc London B website)
Phenotypic evolution o dispersal enhancing traits in insular voles
Anders Forsman1 Juha Merilä, Torbjörn Ebenhard
1 Author for correspondence E-mail:
Supporting Results
Physical condition. To evaluate the possibility that differences in body size among populations may reflect a phenotypic response to spatial differences in food quality or quantity, we compared the relationship between body mass and overall structural body size between wild-caught and laboratory-reared individuals. If wild-caught voles are food stressed, one would expect them to have a lower body mass for a given body size, compared with laboratory-reared individuals given high quality food ad lib. Data for males and females were pooled, and all females with embryos were excluded from the analysis. We tested for a difference in least-squares linear log-log regression of body mass on our measure of overall body size (PRIN1) between wild-caught and laboratory-reared individuals using analysis of covariance. We performed separate tests for mainland and insular voles. We also tested for differences in size-adjusted body mass among populations using data from wild-caught individuals.
To evaluate the possibility that differences in body size among populations reflect phenotypic plasticity in response to food quality or quantity, we compared the relationship between body mass and structural body size, under the assumption that food-stressed voles would have a lower mass for a given metric size. Wild-caught voles from the mainland were heavier for a given body size than were mainland voles that had been raised in the laboratory (ANCOVA on log-transformed data with PRIN1 as the covariate: laboratory vs. wildcaught, F 1,296 = 7.08, P < 0.01; PRIN1, F 1,296 = 413.80, P < 0.0001; interaction, F 1,295 = 0.11, P = 0.74). This superior physical condition of wild-caught voles argues against the interpretation that mainland voles were smaller than insular voles because they were food stressed.
Among insular voles, the relationship between body mass and structural size did not differ between wild-caught and laboratory-raised individuals (lab. vs. wild, F 1,366 = 0.68, P = 0.41; PRIN1, F 1,366 = 594.42, P < 0.0001; interaction, F 1,365 = 3.71, P = 0.055). Size-corrected body mass of wild-caught voles varied among insular populations (F 5,165 = 3.86, P < 0.01), but these differences in average physical condition (measured as least-squares means from the log-log regression of body mass on PRIN1) were not associated with population mean structural body size (r = - 0.20, P = 0.69, n = 6 localities). This argues against the interpretation that the variation in body size among populations reflects non-genetic responses to food availability.
1