Final models for effects of parents' ages and interactions between them on breeding traits.

P-values for terms comprised in significant interactions are not shown because they are meaningless. For random factors we show variances (and SD) for slopes (S) and intercepts (I). Piecewise regression on significant quadratic effects showed that all initial increases and final decreases with age were significant (P<0.05). Effects of ID may have been nonsignificant because individuals were observed with few repetitions; e.g., 2.87±2.37 repetitions for the 1,860 fathers and 3.00±2.58 repetitions for the 1,789 mothers analysed for fixed effects of models 3 and 4.