Assessing the relative importance of environmental effects, carry-over effects and species differences in thermal stress resistance: a comparison of Drosophilids across field and laboratory generations

There is increasing interest in comparing species of related organisms for their susceptibility to thermal extremes in order to evaluate potential vulnerability to climate change. Comparisons are typically undertaken on individuals collected from the field with or without a period of acclimation. However, this approach does not allow the potential contributions of environmental and carry-over effects across generations to be separated from inherent species differences in susceptibility. To assess the importance of these different sources of variation, we here considered heat and cold resistance in Drosophilid species from tropical and temperate sites in the field and across two laboratory generations. Resistance in field-collected individuals tended to be lower when compared with F1 and F2 laboratory generations, and species differences in field flies were only weakly correlated to differences established under controlled rearing conditions, unlike in F1-F2 comparisons. This reflected large environmental effects on resistance associated with different sites and conditions experienced within sites. For the 8 h cold recovery assay there was no strong evidence of carry-over effects, whereas for the heat knockdown and 2 h cold recovery assays there was some evidence for such effects. However, for heat these were species specific in direction. Variance components for inherent species differences were substantial for resistance to heat and 8 h cold stress, but small for 2 h cold stress, though this may be a reflection of the species being considered in the comparisons. These findings highlight that inherent differences among species are difficult to characterise accurately without controlling for environmental sources of variation and carry-over effects. Moreover, they also emphasise the complex nature of carry-over effects that vary depending on the nature of stress traits and the species being evaluated.