RELATION OF DIMORPHISM TO MUTATION. 35 



The two kinds of leaves borne by the two kinds of branches of the 

 cacao tree, for example, are probably much more different than any 

 mutations that have ever been reported. Changes of expression in 

 dimorphic specializations are as great as or greater than in those that 

 give rise to distinct mutative varieties or sports. Dimorpliism not 

 only covers at least an equal range of variation, but affects the same 

 kinds of characters as mutative variations. This is shown very defi- 

 nitely in the Upland type of cotton, where cases of okra cottons with 

 dimorphic leaves arise as mutative variations from broad-leaved vari- 

 eties. The dimorpliic condition, at least in such cases, has to be 

 looked upon as a direct product of mutative variation. 



Another form of mutation, more common in Upland cotton than 

 mutations to narrow leaves, is the shortening of the internodes of the 

 fruiting branches, as in the so-called "cluster" varieties. This vari- 

 ation also has relation to dimorpliism. The shortening of the inter- 

 nodes of the branches, which characterizes the "cluster" varieties, 

 affects only the frvuting form of branches. The vegetative limbs of 

 "cluster" cottons grow quite as long as those of other varieties. The 

 expression of the cluster character is accompanied by the expression 

 of the other characters of the fruiting, branches, like the peculiarities 

 that come into expression in only one sex of an animal though capable 

 of transmission tlirough the other sex. Breeders consider that special 

 egg-laying or milk-producing qualities are transmitted by male ani- 

 mals as well as by females. 



A further analogy between mutations and dimorphic changes of 

 expression of characters may be found in the fact of coherence. 

 Dimorphic branches do not differ in one character alone. One form 

 of branch differs from another in all of its parts. A whole group of 

 characters clings together, as it were, in expression. In a similar way 

 a mutative change usually involves a large group of characters. The 

 fact of coherence is of practical importance in relation to selection, 

 for it enables mutative variations to be much more easily recognized 

 than if each detail of structure or color were free to change inde- 

 pendently. In dimorphism, as well as in ISIendelism and mutation, 

 there seems to be a tendency to contrasted expression instead of to 

 blended or graded expression. In other words, these phenomena 

 may be said to be free from the law of regression enunciated by 

 Galton. Contrasted characters not only maintain themselves in 

 expression, but the contrast gains reenforcement by combination 

 with other alternative characters. 



When hybrids are made between different species, such as the 

 Upland and Egyptian cottons, it becomes evident that some char- 

 acters have much more freedom of combination than others. For 



221 



