Classical adaptational and genetic engineering approaches offer complementary insights to understanding biological variation: the former elucidates the origins, magnitude and ecological context of natural variation, while the latter establishes which genes can underlie natural variation. Studies of the stress or heat shock response in Drosophila illustrate this point. At the cellular level, heat shock proteins (Hsps) function as molecular chaperones, minimizing aggregation of peptides in non-native conformations. To understand the adaptive significance of Hsps, we have characterized thermal stress that Drosophila experience in nature, which can be substantial. We used these findings to design ecologically relevant experiments with engineered Drosophila strains generated by unequal site-specific homologous recombination; these strains differ in hsp70 copy number but share sites of transgene integration. hsp70 copy number markedly affects Hsp70 levels in intact Drosophila, and strains with extra hsp70 copies exhibit corresponding differences in inducible thermotolerance and reactivation of a key enzyme after thermal stress. Elevated Hsp70 levels, however, are not without penalty; these levels retard growth and increase mortality. Transgenic variation in hsp70 copy number has counterparts in nature: isofemale lines from nature vary significantly in Hsp70 expression, and this variation is also correlated with both inducible thermotolerance and mortality in the absence of stress.