1254394701

106

inter-individual variation in metabolic performance due to sex and body size without removing the effect of organ mass on metabolic rates. BMR and Msum variations were not significantly affected by scx and Msum was not related to sizc. However, the effect of body size on BMR approached significance (p = 0.06). We therefore chose to use a conservative approach and extracted size residuals of BMR to consider this weak effect. We testcd for the effccts of sex and body size on organ masses with the same approach and extracted residuals for fiirther analysis when the effects were significant. This was found to be the case for muscles, cardiopulmonary organs, excretory organs, brain and skin but not for digestive organs. The following step was to test for relationships between residual BMR or whole Msum and residuals of organ mass (except for digestive organs). We did this using data for the whole year and then for each period independently. We used a model selection approach based on second order Aikaike information criteria (AICc) to highlight the best model explaining variations in metabolic performance. Standardized coefficients (P) were used to order organs according to their relative influence on the dependent variable.

In GLM analyses, significant differences between periods were examined using Tukey tests. For all analyses, model residuals were tested for normality using one sample Kolmogorov-Smimov test. Data are presented as least sąuare means ± s.e.m. Model selection were performed in R using the MuMIn package (Barton, 2013).

4.5 Results

4.5.1 Seasonal variation in body mass and metabolic performance

Size-independent body mass (size: Fj.72-8.5, p < 0.01) varied over the year (period: F₃.t2=7.3, p<0 .001). Body mass increased by 7.0% between fali and midwinter (Tukey: p< 0.0001), decreased by 4.0% (Tukey: p < 0.05) by the end of winter and remained constant until summer (figurę 4.2). Size-independent body mass was positively related to time of capture (Fł,72=5.5, p < 0.05) and was also dependent on sexes (F|_i72 = 35.7, p< 0.0001) with males being on average 7.8% heavier than females (males: 11.6 ± 0.1 g; females: 10.8 ± 0.1 g,) and this effect did not change between the periods (period*sex: p = 0.4). Size-independent lean dry mass (size: F 1,73 — 11.5, p < 0.01) was not affected by time of capture (p = 0.2) or the interaction term