Abstract
The biometrical genetics of copper tolerance has been investigated in two Californian populations of Mimulus guttatus by crosses to a nontolerant British population. A simple biometrical model involving only additive and dominance effects is not sufficient. When the first order interactions are included, the model is shown to fit the data. Interactions between the dominance effects of different loci, and between dominance and additive effects, are the most important. These interactions can be explained either by a threshold model, or by postulating dominance modification.