103
organs was calculated by adding fresh biopsy mass es to the remaining firesh mass of the organs and recalculating lean dry mass according to the proportion of water and lipids contained in the tissues. We did not record the mass of the reproductive organs neither the lean dry mass of the spleen, pancreas and proventriculus because lean dry samples were too smali to be weighed accurately. Technical difficulties associated with manipulating very smali dried organs also led to the loss of some samples. In total, we obtained data for all organs in 64 birds. Each bird was sexed by inspection of reproductive organs during dissections.
4.4.4 Statistical analysis
4.4.4.1 Seasonal yariation and sex effects on body mass and metabolic performance
Body mass varies with individual body size. To study changes in body mass throughout the year we therefore had to consider this effect. We used a measure of structural body size as covariate to analyse change in size-independent body mass. The structural body size (thereafter cal led “size”) was calculated from a principal component analysis on morphological data (length of head plus beak, wing and tarsus) (Rising & Somers, 1989). Then, we used a generał linear model (GLM) testing for the effects of “period” (fali, midwinter, end of winter and summer), “sex” (małe and female), the interaction term “sex*period”, the variable “relative time of capture” (time sińce sunrise / day length, hereafter “time of capture”) to consider the effects of daily fattening (Mandin & Vezina, 2012) and “size” as covariate to investigate how size-independent body mass varied through time. We did the same analysis to study the yariation in lean dry body mass.
Metabolic parameters are influenced by the amount of metabolically active tissues which forms most of body mass. Therefore, studies on metabolic ratę yariation should consider the body mass effect. However, Black-capped chickadees are fattening on a daily basis during winter (Mandin & Vezina, 2012) and most of the body mass yariation during that period is thought to result from the accumulation of fat tissues known to have very Iow metabolic activity (Chaplin, 1974; Scott & Evans, 1992). Statistically controlling metabolic ratę for body mass can lead to underestimations when body mass contains a large proportion of fat tissue (Petit et al., 2010). We therefore first studied the yariation of whole BMR and whole Msum throughout the