Yields obtained by this procedure have exceeded those of the parental check by more than 50 percent.

Cytoplasmic male sterility has been discovered in pearl millet. This will provide an alternative procedure for the production of hybrid seed.

As work with the various forage species progresses, growing emphasis may be placed on procedures that permit wider utilization of hybrid vigor.

THE MAIN SPECIES of cotton grown in the United States are sea-island and Egyptian (Gossypium barbadense) and upland (G. hirsutum).

Sea-island is outstanding in length and quality of fiber, but its production has been limited to South Carolina, Georgia, Florida, and the islands near South Carolina. Wilt became a problem in parts of this area about 1895.

Erwin F. Smith, of the Department of Agriculture, determined the cause of the disease. Resistant varieties were developed. Before resistant types of the upland could be developed, the production of cotton had to be abandoned in many areas in the Southeast.

The boll weevil, first found in Texas in 1892, moved eastward and did so much damage that the late-maturing, long-staple types could not be grown profitably. Early efforts of plant breeders were devoted to the development of productive early types, which tended to escape serious damage. Fiber qualities, particularly length, were sacrificed for earliness.

After productive early varieties were available, greater attention was directed toward increasing the length of the fibers.

Varieties of Egyptian cotton were introduced into the irrigated sections of Arizona and southern California in 1903, but as a group they were poorly adapted. Selection, however, was continued, and improvement in earliness and productivity was gradual.

In 1908 two types differing from normal were found and increased. One of them was later named Yuma. A selection from Yuma was later named "Pima," which continues to be important in the Salt River Valley of Arizona.

Cotton often is cross-pollinated. The extent varies with varieties and the number of pollinating insects. This chance hybridization has provided the variability that permitted the isolation of the many different varieties.

Controlled hybridization also has been used to produce new, variable populations, in which further selection could be practiced.

The breeding methods most commonly used have been mass selection or the selection of individual plants. When pollination is not controlled, neither leads to a high degree of uniformity. Thus variability was a factor that permitted the rapid change in varieties when wilt and the boll weevil became serious.

The commercial utilization of first-generation hybrids of cotton has been suggested by many workers. Four methods have been proposed: Inter-planting of the desired parents, with cross-pollination to be effected by bees; controlled hand pollinations; the use of genetic or cytoplasmic sterility; and induced male sterility induced by chemicals (gametocides).

For various reasons, only the use of gametocides has been tried on a commercial scale. Further developmental work is required before this method of producing hybrid seed can be considered entirely satisfactory.

Soybeans have been grown in China for several thousand years, but only 50 thousand acres were grown in the United States in 1907. Then the crop was used largely for forage.

The possibility of utilizing soybeans as an oilseed crop was visualized about 1920. Mills were built, and by 1929 there had developed a considerable demand for high-oil, yellow-seeded varieties.

Some of the leading varieties in 1929 were Illini, Dunfield, Mukden, and Scioto, which were direct introductions from the Orient or selections from such introductions.

It became apparent that various combinations of the desired attributes could not be obtained readily by a direct selection. Between 1930 and 1940, hybridization (followed by selection) or hybridization and backcrossing (followed by selection) became the standard breeding procedures.

Most of the varieties released since 1940 were developed following controlled hydridization.

Breeding of soybeans has been effective in isolating types superior in yielding ability, resistance to lodging and shattering, adaptation to suit various requirements as to maturity, and resistance to disease.

Yielding ability seems to be conditioned by relatively few genes, because transgressive segregation types that exceed the normal range of the two parents for yield has been observed in the progeny of many crosses. Resistance to lodging and shattering are required for satisfactory harvesting by combines.

Few of the original introductions were adapted to the Northern States. Varieties having a suitable daylength response are now available. Evidence of the importance of this development is that in Minnesota soybeans were grown on 97 thousand acres in 1938 and on 3.1 million acres in 1958.

The percentage and quality of oil are under genetic control but are subject to some environmental modification. Oil content has been increased through increased yields and a higher oil content of seed. Many of the newer varieties are superior in oil content to their parental strains. Much soybean oil is used in the manufacture of shortenings, margarine, and paints.

The protein in soybean varieties ranges from 35 to 50 percent. Oilmeal, the residue after the oil is extracted, is an important protein concentrate for feeding livestock. The protein concentrate also has many industrial uses, including the making of plastics and adhesives.

Increased emphasis has been devoted in the last few years to the breeding of new varieties with a higher percentage of protein.

In the early years of production in the United States, soybeans were relatively free of diseases and insect pests. Several diseases and the root knot nematode have become more bothersome in recent years. This development has been so new that the genetic basis for resistance is still undetermined in several instances.

The production of varieties that have adequate resistance to all of the major diseases remains a task for the future.

EFFORTS to improve ornamental plants have varied with time and the method of propagation.

At first probably no more was involved than the selection of the more attractive types that grew wild and their transfer to cultivated conditions.

Further progress depended on the type of propagation used. The rose, one of the most ancient of cultivated flowers, is an example. Several wild species are involved. Most of them are confined to the North Temperate Zone. During the 16th century, propagation was largely by cuttings, and only a few varieties were cultivated. Some 21 double varieties were cultivated in Europe by 1800.

Now the total number of varieties of rose probably would be close to 20 thousand. This big increase has come about by hybridization and the evaluation of the resultant seedlings.

The development of large numbers of new varieties in other forms also depends on the evaluation of seedling progenies. The development of new types under vegetative propagation is limited to the recognition and maintenance of chance bud sports.

The breeding of flowers and ornamentals has been done largely by amateurs, with some participation by private seed firms and public agencies. Some of the work undoubtedly has suffered from a lack of understanding of genetics, yet the most modern techniques sometimes have been used. I give several examples.

Stock, Mathiola incana, has both single- and double-flowered forms. Doubleness, the desired type, is conditioned by recessiveness for a single gene. The double-flowered forms are sterile and must be propagated from heterozygous singles, which yield only 25 percent of double-flowered plants.

In one such strain, a mutation occurred that kills all of the pollen grains carrying the mutant gene. When the pollen lethal is linked with the gene for single flowers, approximately 50 percent of the plants produce double-flowered plants. Still later a trisomic type, called "slender," was discovered. Trisomics are 2^N + 1 types, in which one chromosome of the complement is present in triplicate. The extra chromosome involved in "slender" is the same one that carries the gene for doubleness. Because of the unequal fertility of gametes carrying the extra chromosome, the selfed progeny of the slender type yields approximately 90 percent of double-flower plants. Furthermore, the single-flowered plants can be recognized in the seedling stage and discarded. Thus, although the desired double type cannot be propagated directly by seed, genetic techniques have made it possible to increase the frequency of doubleness from 25 to approximately 100 percent.

Colchicine is an effective agent for doubling the chromosomes in many plant forms. Such doubled chromosome forms, called tetraploids, have larger flowers and usually a sturdier form of growth.

Colchicine treatment has been used to produce tetraploid forms of many of the annual flowers, such as snapdragons, marigolds, and petunias. Tetraploid forms in vegetatively propagated types such as forsythia and lily, have also been produced.

Crosses between tetraploids and diploids yield triploid progeny, which are sterile. Where seed propagation is required, triploid forms must be propagated anew each generation. Where vegetative propagation is possible, desirable triploids may be maintained indefinitely. The tiger lily and certain varieties of canvas and day lilies are triploids which are propagated vegetatively.

In plants propagated by seed, species hybridization has played an important role in improvement. Roses, dahlias, gladiolus, amaryllis are common flowers that involve extensive species hybridization in their ancestry.

The limit to the usefulness of species hybridization is determined by the percentage of seed set and the viability of such seed. Viable seedlings often can be obtained from otherwise sterile crosses by use of embryo culture. In this method, young embryos from developing seeds are dissected out and transferred to an appropriate nutrient media. Embryo culture has been used extensively in the breeding of iris.

G. F. SPRAGUE is leader in charge of Corn and Sorghum Investigations, Cereal Crops Research Branch, Crops Research Division, Agricultural Research Service, Beltsville, Md.