jane12023-sup-0001-FigS1-S2-TableS1-S5.docWord document342K

Fig. S1. Prevalence of inbreeding in the captive population.

Fig. S2. Information about relatedness among individuals in the captive population.

Table S1. Basic statistics describing the pedigree.

Table S2. (a) Test of the significance of age and year of sampling on each of the traits under study. To test these effects, we used a mixed model including bird identity as a random effect to account for repeated measurements. ASReml gives an F value equivalent to a Wald statistic, but for a sample size as large as this data set, ASReml do not provide estimates of the degree of freedom for the denominator. In order to be conservative, it was hence calculated as the number of observations minus all estimated effects (intercept, levels for identity, age and year, both treated as factors). The F statistics was then compared to an F distribution. The issue of significance testing of fixed effects is still under debate, but please note that our results are the same if we run similar models in r using lmer and then functions to assess significance such as pbkrtest or LMERConvenienceFunctions. Both age of birds and year of sampling were significant for all traits. (b) Size of estimated fixed effects and their standard errors. Age 1 and Year 2001 are the references for other estimates (ASReml uses the first occurrence in the data as the reference point, hence reference is 2001 and not 1999), and were given a 0 value.

Table S3. (a) Sequential test of significance of random effects. Significance of effects was assessed with a likelihood ratio test (See 'Material and methods' for details). Estimates of these random effects are given in the main result section. (b) Estimates of fixed effects and their standard errors in the univariate animal models. Age 1 and Year 2001 are the references for other estimates, hence the 0 value.

Table S4. Estimates and significance of maternal effects.

Table S5 (a) Covariances and correlations for random effects (additive genetic variance, permanent environment, birth year and residual) in the multivariate models. Permanent environment and residual covariances are not estimable for between sexes analyses (see material and methods for details). Significance test of additive genetic covariances is also reported. (b) Estimates of fixed affects. Age 1 and Year 2001 are the references for other estimates, hence the 0 value.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.