All 36 possible crosses among 6 homozygous lines of Drosophila melunoguster were tested for their phenotypic response in developmental time and dry weight at eclosion to variation in temperature and yeast concentration. This method was chosen because it allows one to produce the same heterozygous offspring repeatedly for testing under more conditions than could be handled at once. We estimated the effects of yeast concentration and temperature and their interaction on both the phenotypic and the environmental components of variation and covariation of the two traits. Development was slower at low temperatures and yeast concentrations, and dry weight and viability were lower at higher temperatures and lower yeast levels. Interactions of the two factors with the crosses and with each other indicated that there were genetic differences in plasticity and that the sensitivity of a trait to one factor depended on the level of the other. The covariation of the two traits was generally weak within an environment. Across environments, its sign depended on the factor that changed between the environments: positive for temperature, nega- tive for yeast concentration. These findings can be explained in terms of an established growth model for Drosophilu larvae. We conclude that for plastic traits with moderate or low heritability, the relationship between the phenotypic and genetic covariance matrices may be a complex function of the environmental factors that affect the traits. Some implications for the prediction of the evolution in fluctuating environments are outlined.