Data analysis
A measure we called “fuel stores” was extracted from the variables fat
score and muscle score for each species using principle component
analysis. Both fat and muscle scores were included, because migrating
passerines depend on both fat (90-95%) and protein (5-10%) for energy
during endurance flight (Jenni & Jenni-Eiermann 1998). To test if there
is an effect of fuel stores on IgY levels a general linear model was
constructed for each species separately. Sex was added as a covariate to
the model as immune function can differ between sexes (Arriero et al.
2015). Julian day (date of capture) was added as a covariate as early
migrating individuals may differ in immune function from late migrating
individuals (Hegemann et al. 2022). Model selection was performed by a
stepwise elimination of non-significant variables by using the drop.1
function in R. To test if there were differences in the level of
immunoglobulins between the four species, a one-way ANOVA along with the
Tukey test was performed. The microbial killing ability data was not
normally distributed. Therefore, Spearman’s correlations were performed
to test whether microbial killing ability is correlated with fuel stores
and Julian day. A Kruskal-Wallis test along with a pairwise Wilcoxon
test was performed to test for sex and species differences in microbial
killing ability.