Offspring emergence and starvation tolerance
During the 50-day pupal observation period, 96% of 355 pupae from the control, 93% of 187 pupae from the mating delay, and 91% of 232 pupae from the nutritional stress group emerged successfully. All emerged offspring subjected to starvation were dead by 15 days post emergence. Although the wet weight of pupae was similar between females and males (Fig. 4a), across all three treatments there was strong evidence that female offspring survived on average longer than males, but not by more than a day (Tables S6 19 – 21), with ω ≤ 0.01 for models not including sex (Tables S6 16 - 18, Fig. 4b).
There was strong evidence for an additional effect of maternal age on offspring starvation tolerance, in addition to offspring wet weight, for the nutritional stress treatment (models without age ω = 0.000). Offspring from young mothers in the nutritional stress treatment were particularly vulnerable to starvation (Fig. 4b), with evidence for a cubic effect of maternal age in this treatment (cubic effect including random intercept ω = 0.705). While there was also evidence for maternal age effects on offspring starvation tolerance in the other two treatments, there was not sufficient information in the data to quantify the relationship between maternal age and offspring starvation tolerance for the control or mating delay treatments (lowest AIC model: control – quadratic effect and no random effects ω = 0.298; mating delay – quadratic effect with no random effects ω = 0.212) (Tables S6 16-17).
For the mating delay treatment, singular fits were obtained for models assuming a quadratic or cubic effect of mother age and including a random intercept and slope, therefore there was not sufficient data to assess for this treatment whether differences among individual mothers carried through to emerged offspring starvation tolerance. There was also insufficient information in the data to provide evidence for among-mother variation in the control treatment, with the difference in AIC weights between models with and without random effects <0.15 (Tables S6 16 - 18). However, there was evidence for an effect of individual mother on starvation tolerance in the nutritional stress treatment (all models without random effects ω ≤ 0.04) (Tables S6 16 - 18), but not for variation among individual mothers in the effect of maternal age on offspring starvation tolerance (ω ≤0.005).
The results of the above analyses for the probability of abortion, offspring wet weight and offspring starvation tolerance were unaffected by removing from the analysis mothers in the nutritional stress treatment that had died before the end of the experiment. The analyses can be found in S10 File and compared with results in S6 File.