American Breeders Magazine 4(1): 68-69 (1913)
Gainesville, Florida.

When one has made a wide cross between strains of ornamental plants (for instance, cannas), it is not difficult to pick out the required types from the first or second generation, by simple inspection. For these strains of plants being grown only for show, good looks are the most important requirement. But when one is working with field crops, fruitfulness and vigorous growth come first; and appearances are only considered secondarily, or not at all. In scientific breeding, after the second generation has been grown, and all possible eliminations have been made from it by inspection, the remaining strains may be tested for vigor and productivity by sowing about a hundred seeds of each, and crowding them among another crop, so that only a small percentage will survive. Those families whose survivors yield the greatest crop, in spite of the crowding, will show, by the survival of the strongest, that they are in this respect well adapted as forage plants for field culture. (For, even with corn, the greatest crop on a given acreage can only be had by crowding the stalks more closely than would be indicated by the conditions for the best individual yield. And the same is the case with most other field or forage crops.) Such an automatic selection of the many undetermined characters which combine to produce a good yield, will often be a quicker and cheaper method than trusting to the eye alone as a means of selection and elimination.

The Velvet and Lyon beans are grown in Florida as forage crops. The fields have to be kept clear of weeds at first; but when the vines have spread between the rows so that cultivation ceases, their growth is usually sufficient to stifle the weeds. Hence the first recommendation for a new bean of this type is a superior ability to choke down weeds. From the second generation of the cross between the Velvet and Lyon beans, grown in hills 8 feet apart, 82 individuals were selected, and their seeds sown in an elimination field, for the third generation. Along with them were sown, for comparison, pure lines of the Velvet and Lyon beans. The elimination field was set out with lines of sorghum 8 feet apart, the sorghum being sown thickly as if for silage. When the sorghum had sprouted, 100 seeds (in most cases) of each of these second-generation plants were dibbled by hand in the rows of sorghum, 2 inches deep and 4 feet apart. When the young plants came up, they had to struggle with the sorghum; and, out of 6093 seeds, only 740 produced plants that bore pods. The amount of the crop produced by each family under these circumstances was a valuable guide to the properties of that strain; for, although most (or all) of the parents were heterozygous in some points, yet, as is usual in self-fertilized plants, the majority of their offspring resembled the parent in somatic characters. The further growth of strains which headed the list, or were low down in it, justified the placing of some confidence in the results of the elimination. The more factors in which the original parents differ, the more useful will probably be the elimination test.