The average yield of lint cotton in the southern states is only about 200 pounds per acre. This figure proves conclusively that the bulk of the crop is raised under poor conditions, for under moderately good conditions, the yields are from 400 to 600 pounds. I have observed for years that cottons produced under poor conditions as to soil, fertility and culture, are invariably of weaker and shorter staple than the same varieties produced under more favorable circumstances. This difference will not run less than one-sixteenth inch in length, and the strength, though not determined, is markedly inferior. So I think it is safe to say that the general adoption of a proper rotation with cover crops will increase the average length of the cotton fiber one-sixteenth inch, besides greatly increasing its strength. The general employment of these improved methods will not cause an increase in the crop, for a proper rotation requires that not more than one-half the area of our farms shall be in cotton, whereas now, at least three-fourths and often a much larger proportion is so employed.

A number of our southern experiment stations carry on cotton variety tests each year and among other data from these tests is set down the length of staple and the number of bolls per pound of the different varieties. The farmers of our section pay no attention whatever to the length of fibre of short. staple varieties of cotton, as they are unable to realize any more from a variety producing one and one-sixteenth inches staple than from one producing only five-eighths inch.

The whole system of purchasing short staple cotton from the farmer is wrong and uneconomic. If a fair difference in price were made by the buyer between each length of cotton, paying the man with one inch cotton more than the one with threefourths inch staple, it would quickly improve the staple of the cotton crop. Many of the most productive varieties will produce at least one inch staple under normal conditions and farmers would quickly abandon the shorter staple sorts if it were made to their interest to do so.

The farmer is, however, beginning to appreciate the import

FIGURE 2.-SHORT STAPLE, NON-PEDIGREED, LENGTH 1⁄2 TO %. ance of size of boll in the economy of his operations. There are a number of short staple varieties of high production which run from 50 to 60 bolls to the pound, the average of all varieties being around 75, while in many varieties it requires at least 100 bolls to weigh one pound. The picker can harvest a much greater quantity per day of a large bolled than of a small bolled cotton, thus a large bolled crop is more quickly gathered, the grade is higher and the labor is available much earlier in the season for other necessary farm tasks. It is also a fact that most of the larger bolled cottons have longer staple than the small bolled sorts and the general adoption of large bolled cottons will mean an improvement in the length and grade of the crop, and will thus benefit the spinner.

I will add right here that the small bolled cotton in the South is largely caused by the way the farmer estimates his crop. They see a field of cotton growing and they think, well, that is

a fine field, it looks fine. Across the road may be another field that doesn't look as if it will produce as much. I find invariably the farmer considers that the small bolled cotton will produce more than the large bolled, although the contrary may be true. We judge by the weight and not by the staple.

We now come to the most effective means of improving the quality of the cotton crop-plant breeding. The plant breeder has done wonders in increasing the yield and good qualities of many of our most important crops. His work has meant mil

FIGURE 3.-WEBBER COTTON, PEDIGREED.

This cotton came originally from a short staple variety. Length 1 5/16. Uniform, strong, and averages in South Carolina greater production than short staples.

lions of dollars in additional profit to the great corn and wheat states. The beet sugar industry is absolutely dependent upon his labors. Many of our finest fruits and vegetables have been created or greatly improved by him within a generation. Our splendid Sea Island cotton, the finest in the world, is the result of years of patient labor by a few intelligent breeders.

Until a few years ago, however, little work worthy of the name of plant breeding was done upon upland cotton and we are only beginning to see the results of the pioneer work of this kind upon our great staple crop. We have in the south hundreds or possibly thousands of so-called varieties of cotton,many of which are indistinguishable from each other and nearly all of which are composed of a great number of different types. of plants.

Dr. H. J. WEBBER was the first man I know of who realized that the most prolific field for the improvement of upland cotton lay in the segregation of the many types, the testing of each separately and the propagation of those which proved themselves of highest quality and yielding capacity. Dr. WEBBER'S work, beginning about 1900, was done under the auspices of the United States Department of Agriculture with which he was then connected, and continued until he severed his connection with government work, about eight years later. In that brief period he produced two excellent varieties of upland staple cotton by selection and pedigree breeding from short staple sorts. These cottons, the Columbia and Keenan, are now planted over wide areas in many parts of the south and have been frequently proved to be fully as productive as the average of the short varieties.

My own plant breeding work at Hartsville dates from my acquaintance with Dr. WEBBER and my observation of his work. Beginning in a small way about ten years ago I have rapidly expanded the plant breeding work at Hartsville until now it requires several trained experts and over twenty acres in breeding plots to carry on the work. Cotton, corn, oats, rye, peas, beans, sorghum and melons are all receiving our attention. We work strictly by the plant to row method and any variety which we put out can be traced back through our records to the original plant which produced it and often through as many as nine generations of individuals. Such work is known as pedigree breeding and it is (just as in animal breeding) the only sure method of producing varieties of desired characters. and keeping them true to type.

When I began breeding cotton I observed that the cotton plant presented a great diversity of characters and my selections were made with a view of combining the greatest number of those which were desirable and propagating them in a single uniform variety. The most important of these characters are yielding capacity, length, strength and uniformity of lint, and size of boll. It is no trouble to find plenty of one and