Genetic variation for ecologically important traits determines the potential for evolutionary changes and should be measured directly. Such measurements of genetic variation based on quantitative genetic theory rely on assumptions of environmental constancy. These assumptions are not likely to hold in nature. Instead, natural environments are structured, and systematic variation in environmental conditions is an important determinant of phenotypic variation. Here we provide an introduction to quantitative genetics using a reaction norms approach, because we believe that this provides us with a good framework for combining ecology and genetics. We subsequently review the literature on genetic variation for clutch size of birds, and we show that, in spite of the inherent limitations of the methods employed, there is strong evidence that clutch size has a heritable component in natural populations of several species. However, the number of studies on the amount of genetic variation for clutch size in different species and across a range of environmental conditions is still far too small to study patterns in the relationship between heritable variation and properties of species and/or their environments. Furthermore, the role of both correlations and interactions with the environment in these estimates requires much more attention. Above all else, however, we need more information on the structure and magnitude of the environmental variation present in these studies. Future work should focus on how to obtain such data, and how to subsequently incorporate them into the proposed reaction norm framework. This requires the search for, and measurement of, relevant ecological variables. Also a more detailed investigation of the within-individual variation and the use of animal model methodology may prove to be valuable. Such additional data are essential for interpreting the amounts of genetic variation present for clutch size as a model system in the general problem of better understanding the maintenance of genetic variation in heterogeneous environments and the estimation of evolutionary potential.