Scaling Across Environments: Temperature and nutrition independently shape the genetics of size plasticity and morphological scaling

Across animals, variation in adult body size is accompanied by coordinated variation in the size of individual morphological traits. However, the same morphological trait can scale differently with body size depending on what drives the size variation. In Drosophila melanogaster, for example, wing size scales differently with body size when size varies because of developmental nutrition versus developmental temperature. Whether the genetic basis of size plasticity and scaling is shared across different environmental regulators of size remains unclear, but is central to predicting how selection acts on the developmental mechanisms that regulate trait size, plasticity and morphological scaling. Using ~200 isogenic D. melanogaster lineages, we measured wing and leg size across nutritional and thermal treatments. For each lineage, we estimated nutritional and thermal plasticity for both traits, as well as the wing-leg individual-level scaling relationship, or ILSR, generated by each environmental source of size variation. We found extensive genetic variation in both thermal and nutritional plasticity for wings and legs, and in the slope of the ILSR between them. However, a lineages thermal plasticity was genetically uncorrelated with its nutritional plasticity for either trait, and we detected no genetic correlation between the slopes of thermal and nutritional wing-leg ILSRs. We also found no genetic correlation in the slope of nutritional wing-leg ILSRs across temperatures. Thus, the slope of a lineages nutritional ILSR at 17{degrees}C was not predictive of its slope at 25{degrees}C of 28{degrees}C. Nevertheless, the overall pattern of nutritional ILSRs was conserved across temperatures. These results suggest that the genetic architecture of size plasticity and scaling depends on the environmental source of size variation. Consequently, the evolutionary response of scaling to selection in heterogeneous environments may not be predictable from genetic variation measured in any single environment.