One of the major aims in physiological genetics is the tracing of the reaction chains which link the presence of a gene with a morphological phenotype. The general method of approach is ideally that of working back through decreasing levels of phenotypic complexity-morphological, anatomical, physiological, and eventually to the biochemical steps governed by single genes. This approach has been valuable in showing how genic interference in a basic reaction of development may, by upsetting the balance between subsequent morphogenic processes, give rise to very complex phenotypes. In practice, however, the analysis often stops at some level in the development of the phenotype that may be far removed from the primary developmental action of the gene. Usually, it is not possible to find the physiological difference that causes an anatomical change, probably because vf the lack of knowledge of specific chemical activators in differentiation. This situation is particularly evident in the flowering plants, where mutant phenotypes seldom receive more than a general morphological or, at best, a histological description. The studies reported in this paper represent an attempt to provide a more detailed explanation of the action of a gene in controlling morphogenesis.