The occurrence of intra-sterile inter-fertile classes has been demonstrated within red clover families derived from the mating of pairs of self-sterile individuals. As inNicotiana and other genera, self-and cross-incompatibility in red clover are controlled by an extensive series of at least seven sterility allelomorphs which render pollen bearing any one factor non-functional on a plant containing the same gene in its constitution, by reason of the inhibited growth of the pollen-tube. The existence of a non-inhibitory self-fertility allelomorph has also been shown.

It has been possible to recognize particular members of the gene complex controlling number of lint hairs on the cotton seed by their effect on the glabrous lintless mutation. The latter was used as an analyser to separate the genes controlling lint development into two series, those hypostatic to it, and those epistatic to it and acting as modifiers. The basic lint percentage effect of a single epistatic gene in the heterozygous phase is about 6–10% on the lintless genotype. The epistatic genes show a reductive interaction with other genes working towards the same effect, so that in the fully linted genotype their effect is less than one-fifth of their basic value.

The epistatic genes are shown to have a clearly defined varietal and specific distribution, not associated with the ginning capacity of the strains in which they occur, showing that in different taxonomic groups a typical quantitative character may be controlled by quite distinct complexes of genes.

Seven additional types of anthocyanin in cotton are described and shown to be controlled by members of an extensive allelomorph system of which fourteen members are now known in the cultivated diploid Asiatic species,Gossypium arboreum andherbaceum. All seven of the new types were found in China, six in field material, and the seventh as an anomaly in experimental cultures. The information relating to the genetical behaviour of this and similar series, whose manifold expressions do not conform to a single simple seriation, is reviewed from the standpoint of deciding between the alternative interpretations of close linkage and multiple allelomorphism with pleiotropy. Although the cotton series, in which a new ‘pattern’ arrangement appeared under controlled conditions, might have been expected to be a particularly suitable subject for this purpose, it was not possible to reach a definite conclusion on the experimental evidence.