The above advice rests on the principle that improvement is made by means of selection. This' is the Darwinian principle. Selection, however, rests on variation. Why variations (or differences) arise, nobody really knows, although nearly everybody has an opinion. It is known, however, that variations accompany changes in soil, climate, methods of growing, and other changed conditions. Variation may also be induced or started off by crossing one plant with another, and such differences are likely to be marked. Some variations appear without any apparent reason, and they may be more or less stable from the first ; they are " sports," or, as we now say, mutations (following the terminology of DeVries). These marked so-called "sudden" variations may reproduce remarkably true from seed. The recent evolution discussions have tended to divide variations into these two classes,—the small individual variations that do not reproduce or "come true" (and are therefore presumed to be of no permanent effect in the evolution of the type), and the variations, usually wider, that do "come true." We do not know, however, what are the ultimate origins or what the physiological differences. Divested of technical questions and controversial phases, the practical difference between mutations and other variations is one of definition,—the mutations come true, the others do not. The mutation theory controverts the older doctrine that variations may be augmented by selection until the differences become morphologically great, and until they also become "fixed" or able to reproduce themselves,—that is, that species originate by means of selection ; but the mutation theory does not controvert the importance— but rather empha sizes it—of selection as an agent in the improvement of agricultural plants. Even if a mutation (or hereditable variation) appears, it may still be greatly improved in its minor features by careful selection.
The mathematical law of chance or probabilities applies to hybrids as well as to other numerical com binations. If a plant with three given characters, for example, were to be crossed with a plant of three contrasting characters, the law of probability would predict about how many of the offspring would have one combination of characters and how many would have another combination. The law might not be exemplified in any one plant, but it would very likely be apparent in the average of a number of plants ; and the greater the number, the more regular the results, due to the subordination of exceptions. Mendel found that this law applies to characters that are united in crossing ; if the law applies, it means that the characters or marks have an identity or individuality of their own, that they are carried over entire rather than as blends. In order to explain the application of the mathematical law of chance to hybridization, therefore, we suppose that characters are units and that they are represented directly in the germ-cell ; and hereby arises the theory of the "purity of the germ-cell." That is to say, the mathematical law requires a biological hypothesis to explain why or how it works with animals and plants. Very many experiments have shown that the characters of parents reappear in offspring approxi mately in the given mathematical proportions ; on the other hand, other experiments show a different or contradictory result. Some hybrids also are blends. A very complex body of speculation has been built
up around the so-called Mendellian law, as there has been about other pronouncements in times present and past ; how much of it is truth time only can tell. The Mendellian discussion has challenged our notions of hybridization and heredity and has modified the methods of experiment ; but there is no indi cation that the Mendel law will enable us to produce new plants with certainty, as some its early adherents predicted.
Plant-breeding societies.
This Editorial is written from the viewpoint of the farmer : the professional plant-breeder will take care of himself. The farmer needs help in this particular effort, as he needs it in other ways. The organ ization of breeding societies is one of the best means of spreading and unifying the work. A number of these societies are now in existence, indicating the interest in the subject and the grip that it has on practical men. Associations for plant-breeding are as necessary as societies for animal-breeding. As an illustration of the kind of effort that these organizations stand for, citations may be made from the literature of the Ohio Plant Breeders' Association : " The purposes of this association shall be to encour age the improvement of plants and to provide an official record for breeders who are giving special attention to this work." The rules for the registry of seed corn are as follows : "Section I.—Eligibility.
" In order that a strain of corn may be eligible to registry with the Ohio Plant Breeders' Association, it is necessary that it trace directly and exclusively to remnants of ears that have ranked not lower than fourth in point of yield of grain, protein, starch or fat in a duplicate ear-row test of not less than twenty-five ears ; and that each year's breeding or testing work shall have been conducted and recorded in accordance with the requirements of the Association.
"Section IL—Ohio Pedigreed Corn.
"Any corn which is the product of a cross between two ear remnants, one as sire and the other as dam, each of which has been selected as per Section I, shall be entitled to the name Ohio Pedigreed. The records shall show whether the cross was made by artificial or natural pollination.
"Section III.—Ohio Standard Corn.
"Eight or more registered ears, as per Section II, or ear remnants, as per Section I, may be merged by shelling and mixing together the grain from all, before planting. If this merged corn, or corn descended exclusively from it, shall, on the average, excel in yield of grain, protein, starch or fat per acre, each of three other varieties (including the one from which it has descended and a standard variety which shall be supplied by the council upon request), when tested upon not less than tenth-acre plots for three consecutive years, the owner of it shall be entitled to a certificate under the seal of the Association, setting forth the record numbers under which the work upon this corn has been recorded, together with a statement that it has filled the requirements of the association and is entitled to the name Ohio Standard. A fee of $10 shall be required for this certificate and copies of same shall be issued at 25 cents each to accompany any corn that traces directly and exclusively to this merging.