Environmental factors can have profound effects on the strength and direction of selection and recent studies suggest that such environment-dependent selection can be sex-specific. Sexual selection theory predicts that male fitness is more condition dependent compared to female fitness, suggesting that male fitness is more sensitive to environmental stress. However, our knowledge about the effect of environmental factors on sex-specific reproductive performance and on sex differences in the opportunity for selection is still very limited. In the present study, we investigated the sex-specific effects of diet quality (yeast deprivation and flour type) in the red flour beetle Tribolium castaneum. Specifically, we manipulated yeast supplementation in wheat and whole-wheat flour in competition assays allowing to test for sex-specific effects of food quality (i) on reproductive success and (ii) on the opportunity for selection. Our data show that yeast deprivation in wheat flour had significantly negative effects on body mass and reproductive success of both sexes, while high quality flour (whole-wheat flour) was able to buffer the negative impact to a large extent. Importantly, our data suggest no sex-specific effect of dietary stress on reproductive success because the magnitude of the negative effect of yeast deprivation was similar for males and females. Moreover, our study demonstrates that low food quality inflated the opportunity for selection and did not differ between sexes neither under benign nor stressful dietary conditions. We discuss the implications of our findings for the adaptation to stressful environments.

Stock cultures and experimental cultures were kept in incubators at 30°C (±1°C) at 50% humidity (±5%) without light. Experimental procedures were performed at room temperature. All flour had been frozen for sterilization prior to experiments at -30°C for at least 24h. The Ga1 (wildtype line) and Rd (‘reindeer’ mutant line) used in this experiment had been originally supplied by the U.S. Department of Agriculture and kept in the laboratory for over a year in standard culturing conditions. Rd is a dominant mutation that affects antenna morphology and was used as a paternity marker. All stock beetles were kept in a flour mixture that consisted of organic wheat flour (type 405, Alnatura, Darmstadt, Germany) and 5 % dry baker’s yeast. In this study we manipulated the diet in two ways: First, we manipulated yeast availability in standard, organic wheat flour (type 405, Alnatura, Darmstadt, Germany), with either control (5% yeast) or no yeast. Secondly, we manipulated yeast availability in organic whole-wheat flour (Alnatura, Darmstadt, Germany), with either control (5% yeast), low yeast availability (1% yeast) or no yeast. The low yeast availability treatment was with 1% yeast similar to a previous study (Godwin et al., 2017) but can still be considered arbitrary and results might be sensitive to the chosen range.

Impact of food quality on reproductive performance

In the first assay, we manipulated the diet quality (i.e. yeast availability) of all focal individuals to investigate the effect of dietary stress on reproductive success of males and females in a setup allowing for intra-sexual competition.

We set up base cultures of 80 individuals of Ga1 (wildtype) and Rd adults in each of the treatments. Stock adults were left to mate and oviposit for three days in 60 g of flour according to treatment, afterwards the adults were discarded and the eggs were left to hatch. 31 days after the base cultures had been set up, we collected and sexed pupae and established sex-separated groups of 40 individuals. 8 days after the last pupae had been sexed, all pupae had emerged and were fully adult. We then set up the mating groups for the experiment.

Groups contained the focal individual (female or male) and a competitor (Rd), as well as two mating partners. All mating group beetles were synchronized in age (±3 days) and competitors as well as mating partners were always raised on the control diet. All wildtype mating partners were marked with a dot of Revell emaille paint, to differentiate them from the focal. We allowed the treatment groups to mate undisturbed for three days in empty arenas (plastic Petri dishes, diameter 3.3 cm). All mating trial arenas were scratched at the bottom to increase traction for the beetles and contained no flour to prevent oviposition. After the mating trial, we froze the males at -30°C for preservation to take phenotypic measurements (see below), while we separated the females to oviposit for 2 weeks in the control environment. In addition, to investigate if females recovered from the treatment when returned to the control diet (i.e. 5% yeast) to lay eggs, we transferred all yeast-deprived females from the whole-wheat flour treatment after one week to a new egg-laying vial (see Appendix Figure A1). After two weeks, females were removed and frozen at -30°C until further processing. Once all offspring reached the adult stage after 50 days, they were frozen and scored for genotype (wildtype or Rd; with Rd phenotype clearly discernible from the wildtype due to shorter and thicker antennae).

All frozen individuals from the mating groups were weighed to the nearest 0.01mg on a Sartorius R200D balance (Göttingen, Germany) within two months after the main experiment. Body mass was used as a proxy for condition (Rowe and Houle, 1996). The repeatability of body mass, estimated from measuring twice a subset of the samples was high (Intra-class correlation coefficient, ICC = 0.937, P < 0.001, N = 96).

Impact of food quality on the opportunity for selection

In the second assay, we manipulated the diet quality (i.e. yeast availability) of all individuals (i.e. focals, competitors and mating partners) during development. Hence we mimicked population level dietary manipulation, aiming to measure the opportunity for selection in males and females separately. After the mating trials, all females were given the opportunity to lay eggs on their respective rearing diet (control or yeast deprived). Apart from this, the experimental design followed the procedures described above and was performed in parallel with the first assay.