4. The superior vigor of this grain is transmissible in different degrees to its progeny.
5. By repeated careful selection the superiority is accumulated.
6. The improvement, which is at first rapid, gradually, after a long series of years, is diminished in amount, and eventually so far arrested that practically a limit to improvement in the desired quality is reached.
7. By still continuing to select, the improvement is maintained, and practically a fixed type is the result.
Thin Seeding with Selection.—Let us discuss what is possible by a combination of thin seeding with selection. In order to do this, we must look at the present modes of cultivating the cereals. Confining ourselves for the moment to wheat alone, we know that from two to five bushels per acre are sown. The bushel of ordinary wheat contains 700,000 grains and more, and, taking two bushels per acre as the quantity sown, we have about 1,500,000 grains per acre. Major Hallett has counted at harvest the number of ears upon a quarter of an acre of wheat (drilled 20th of November with one and a half bushel of seed per acre, and which proved an exceptionally heavy crop of fifty-six bushels per acre), and the number of ears found was 934,120 per acre, or not so many ears as the grains sown. Here it is evident, from the number of grains sown, that either the natural powers of tillering could not have been exercised, or that the greater part of the seed must have been sown uselessly. Doubtless some of the grains did produce more than one ear, but this only makes the case still worse for the remainder. Not only was the number of ears below that of the grains sown, but each ear was but the stunted survivor of a struggle for existence. A high authority has said that, if a square yard of thickly-sown wheat be counted in spring, and the supposed number of ears then recorded, it would be found that ninety per cent of them would be found missing at harvest. Beyond all question, in thickly-sown wheat, very many of what appear as stems in the spring die away before harvest, and have thus grown not only uselessly, but in the struggle for existence have starved and stunted those which ultimately came to ears.
In ordinary English crops the number of ears produced per acre being taken as about 1,000,000, and the crop as 34 bushels, we have, at 700,000 grains per bushel, 23,800,000 grains per acre, or an average per ear of only 23 to 24 grains; and, if more than 1,000,000 ears per acre be claimed, it must be at the expense of their contents. Five imperial pints (= 6·1 American measure) of wheat per acre planted in September, 12 inches X 12 inches, gave 1,001,880 ears per acre, or 67,760 ears in excess of those produced on the other side of the hedge from l2 bushel, or more than thirteen times the seed. Again, 6·1 pints (American measure) of wheat planted 12 inches X 12 inches, October 17th, gave 958,320 ears per acre; and planted similarly, October 4th,