(a) Conceptual diagram: predicted responses of temperature-dependent selection on and intergenerational plasticity of embryo metabolic rates (MR) at cool and hot offspring temperatures. We hypothesise that hot offspring temperatures will select for lower metabolic rates, while selection at the cool offspring temperature will be relatively relaxed (positive but no significant correlation between fitness and metabolic rate shown). If intergenerational plasticity aligns with selection when environments across generations match, then similar trends in the direction and strength of selection should be observed. We therefore predict that parents in the warm environment (pink) will produce offspring with lower metabolic rates (smaller curved arrows) compared to parents from the cool temperature (blue), and that this will be correlated with development time and growth rates, with fitness benefits for offspring. (b) Experimental design: parents were held under one of four treatment combinations: 24°C or 30°C and low or high feeding frequency, then bred to produce offspring that were reared at either 24°C or 30°C. Embryo size (diameter, area, mass) and yolk area were measured at 1–4 h post fertilization (hpf), and metabolic rates (measured as rate of oxygen consumption) measured at three stages: 25% of embryonic development (MRembryo), 1–4 h post-hatching (hph; MRhatch), and 1 week post-hatching (1 wph; MRlarva). Offspring were then monitored for survival up to two weeks post-hatching.

Offspring phenotypes in response to parent and offspring treatments. Responses of (a) embryo mass, (b) metabolic rate (MRembryo, MRhatch, MRlarva), (c) development time and (d) growth rate, measured across combinations of parent temperature (24°C or 30°C), feeding frequency (low; ‘L’ or high; ‘H’) and offspring temperature (24°C or 30°C). The first number in a treatment description refers to parent temperature and the second refers to offspring temperature. Note that metabolic rates shown in panel (b) do not account for the significant effect of embryo mass.

The relationship between offspring metabolic rates (log10MRembryo, log10MRhatch, log10MRlarva) and log10 embryo mass in response to parent and offspring treatments. Coloured data points and lines reflect parent temperature (24°C or 30°C), feeding frequency (low; ‘L’ or high; ‘H’), and offspring temperature (24°C or 30°C). Grey bars are standard error. The first number in a treatment description refers to parent temperature and the second refers to offspring temperature.

Correlation plots for offspring phenotypes. Pairwise correlations between offspring traits: metabolic rates, embryo development time and larval growth rate, across combinations of parent and offspring treatments. Coloured plots represent significant correlations between traits.

Selection surface plots. Selection on embryo mass (μg) and metabolic rate (MRembryo; μlO2h−1) across combinations of parent (P) temperature (24°C or 30°C) and feeding frequency (high; H or low; L) and offspring (O) temperature (24°C or 30°C) environments.

Acknowledgments
This image is the copyrighted work of the attributed author or publisher, and ZFIN has permission only to display this image to its users. Additional permissions should be obtained from the applicable author or publisher of the image. Full text @ Phil. Trans. Roy. Soc. Lond., Series B