Environmental variation (V-E) in a quantitative trait - variation in phenotype that cannot be explained by genetic variation or identifiable genetic differences - can be regarded as being under some degree of genetic control. Such variation may be either between repeated expressions of the same trait within individuals (e. g. for bilateral traits), in the phenotype of different individuals, where variation within families may differ, or in both components. We consider alternative models for defining the distribution of phenotypes to include a component due to heterogeneity of V-E. We review evidence for the presence of genetic variation in V-E and estimates of its magnitude. Typically the heritability of V-E is under 10%, but its genetic coefficient of variation is typically 20% or more. We consider experimental designs appropriate for estimating genetic variance in V-E and review alternative methods of estimation. We consider the effects of stabilizing and directional selection on V-E and review both the forces that might be maintaining levels of V-E and heritability found in populations. We also evaluate the opportunities for reducing V-E in breeding programmes. Although empirical and theoretical studies have increased our understanding of genetic control of environmental variance, many issues remain unresolved.