HABIT OF VARIETIES FROM THE SAME SOURCE. 29 



Celebes, No. 21813 had buff-eyed seeds; in 21814 they were marbled 

 like Whippoorwill; in 21815, black-eyed; in 21816, buff; and in 21817, 

 black. In all cases the plants proved to be perfectly prostrate with 

 large, very pale leaves, and so late that they did not even bloom at 

 Arlington ; indeed, the plants were indistinguishable from each other. 



In a similar series from Mount Selinda, Ehodesia, the seeds were 

 as follows: No. 22929, buff; 22930, blue; 22931, speckled, resembling 

 New Era; 22932, black; 22933, speckled, resembling Taylor. All 

 these proved to be very procumbent, forming flat masses of herbage 

 and being practically indistinguishable. 



Where several kinds of seeds are mixed together, the resultant 

 plants are generally very similar. Thus in No. 11076, from Abyssinia, 

 three kinds of seeds — marbled, speckled, and marbled and speckled — 

 were mixed. These all bred true in the greenhouse. In the field the 

 plants formed very viny procumbent masses that were indistinguish- 

 able from each other, though quite different from other varieties. 



A number of similar instances can be cited, so that it would ap- 

 pear to be generaly true that varieties from the same source are 

 very similar in habit. A case of unusual interest is the cowpea that 

 now occurs wild or half wild in southern Louisiana. In this variety 

 the seeds are mostly buff, but mixed in are black, pink, maroon, and 

 marbled. Five forms with different-colored seeds were selected out 

 of Steckler's "Wild Louisiana " cowpea, and all bred true. They are 

 all rather small seeded and of very similar habit, all so late that only 

 a few pods mature at Arlington. Chinese varieties of cowpeas with 

 few exceptions are very susceptible to rust. When coming from the 

 same locality, the different-seeded varieties are often much alike. 



In seeking an explanation of this phenomenon, three hypotheses 

 may be suggested: (1) That the varieties were originally different in 

 habit as well as seeds but under the influence of environment have 

 become alike; (2) that the different-colored seeds are mere color 

 sports from one original variety — hence very similar in habit; (3) 

 that more or less continuous crossing has resulted in a blending of 

 habit characters to one type, the seed colors remaining in accordance 

 with Mendelian principles. 



The strongest probability lies in the third hypothesis. Indeed, 

 in the lot of seeds from South Africa occurred a good many seeds that 

 looked as if they might be heterozygote, which proved to be the case 

 in No. 22958. Heterozygote seeds are frequent also in cowpeas from 

 China. 



Again, in the Abyssinian variety, No. 11076, the marbled and 

 speckled form, 11076B, is undoubtedly a hybrid between 11076 and 

 11076 A, just as Groit is a cross of New Era and Whippoorwill, the 

 type of coloration being the same in both hybrids. 



229 



