The bee breeder should know what inbreeding will do, to his stocks. If he starts an inbred line by backcrossing for two generations and then makes brother-sister matings, he should expect that each line will become more and more uniform as inbreeding progresses. Most noticeable, however, for the first few generations will be the quality of the brood.

If queen B, a daughter of A, is mated to several drones (sons of A), the brood viability of queen B will average 75 percent. A daughter queen C is then mated to sons of queen A and will have brood viability that will average either 75 or 50 percent. If it is 50 percent, the line has been reduced to two lethal alleles and the brood of queens D, E, and F will also be 50 percent if mated as shown in the diagram. If brood of queen C is 75 percent viable, then that of D may also be 75 percent but somewhere not far from E or F in the diagram the viability will probably drop to 50 percent and all future generations will remain at that level. However, by selection it is possible to keep viability at 75 percent but the breeder would be reducing the effectiveness of inbreeding slightly by selection for heterozygosity of lethal alleles and other genes linked to these. It is probably advisable to select the matings that produce 50 percent viable brood in the C or D generation and thus quickly reduce all inbred lines to two lethal alleles and consequently have 50-percent viability in all inbred lines. If this is done, an analysis of the lethal alleles in all inbred lines is more readily accomplished. By test crossing to identify the lethal alleles in each line, the breeder can then predict which crosses will give high brood viability in hybrids and which crosses will give intermediate or low viability.

In one season of inbreeding it is possible to get as far as producing a number of sister queens of the D generation and get these mated to their brothers ( drones produced by their mother queen C). These queens will be wintered, and the following year the breeder can make test crosses while continuing to inbreed the lines by brother-sister matings. It is advisable to test the inbred lines at the E generation of queens (50-percent inbred). One generation of brother-sister mating should be made each year after the first season. This insures continuing the inbred lines until they are selected in hybrid combinations.

The bee breeder can accomplish very little by selection while inbreeding. He can surely select queens and drones in each generation for color and general appearance. In a sense he can Progeny-test each generation by measuring such qualities in the workers as tongue length, wing length, color, or temper but the economic value of bees is measured by the total productivity of the entire colony.

Thus very little selection can be made by testing inbreds as inbreds mated to drones of the same line. Since viability of brood in two-allele lines is only 50 percent, colonies headed by inbred queens lacking in vigor do not develop sufficient populations to accurately evaluate such economically important characteristics as honey production, swarming tendencies. and Wintering qualities. This selection in inbreds as inbreds is supplemental rather than substitutable for selection between inbred lines when in crosses with other inbred lines.

The division of bee culture undertook bee breeding in earnest in 1937 when a program for development of disease-resistant strains was initiated. Earlier work had been limited largely to control of mating, introduction of races, and studies of the characteristics of races.

The program was begun in cooperation with the State experiment stations of Iowa, Texas, Wisconsin, and Wyoming. Work in Iowa had shown that some stocks were more resistant than others. Each year colonies headed by daughters of highly resistant queens of the previous year were tested. The queens were mated at isolated mating stations. In the three principal lines carried until 1945, one showed definite increase in resistance, another less increase in resistance, and a third line apparently little or no increase. The lines became stabilized at characteristic levels considerably below complete resistance. Selection of negative colonies (showing no disease after inoculation) was more effective than selection of recovery colonies (recovered from disease produced in inoculation).

In the resistance work, artificial inseminations were first used in 1943 on a small scale and were so successful that natural matings were discontinued in 1946. Resistance increased immediately, indicating that the slow progress with use of natural matings might have been due to mismating. Soon entire test groups developed no disease. Hybrids produced by crossing the resistant lines were resistant and also more productive than the inbreds.

These encouraging results with disease-resistant hybrids stimulated expansion of this method of breeding to include other economically important characteristics. Inbred lines of bees were developed from a number of queens that produced outstanding colonies of bees. These queens were selected because their colonies not only produced large crops of honey but possessed other desired qualities such as gentleness, vigor, high egg production, or nonswarming tendencies. After the lines were inbred, they were crossed and tested in various hybrid combinations. Artificially inseminated queens were tested for honey production for the first time in 1943. As expected, differences between various hybrid combinations were apparent early in the breeding and testing program.

It soon became obvious that testing of hybrids under a wide range of climatic and environmental conditions was desirable. The Department of Agriculture, division of bee culture, therefore entered into an agreement with the nonprofit Honey Bee Improvement Cooperative Association. Through this agency a large number of hybrid queens of various types were distributed and tested by beekeepers throughout the United States. The queens are produced on Kelleys Island in Lake Erie. As the island is isolated, one can control matings. All hybrid queens are allowed to mate naturally to the drones produced by other unrelated hybrid queens. Thus the test colonies are headed by single-hybrid queens and the workers in these colonies are double hybrids. Having only one such isolated mating station, the various types of single-hybrid queens must all be mated to drones of a single type of hybrid. Some of the queens are tested at the various bee culture laboratories of the Department of Agriculture. The Department also is continuing to test other hybrids that are artificially inseminated.

Beekeepers who have obtained test queens have been favorably impressed by the superiority of certain double hybrids, which have produced as much as 50 percent more honey than comparable commercial lines. The hybrids also have shown greater uniformity, more brood per colony, and brood of higher viability than the commercial lines.

WE HAVE SEEN that the problems involved in breeding bees are too great for the individual breeder to go far in improving his own stock. He can do little more than avoid the mating of closely related individuals, select for high brood quality, and outcross to unrelated strains whenever low brood viability becomes too frequent.

Breeding and testing have shown that hybridization can produce superior bees. The best solution is thus a hybrid-breeding program such as only State or Federal research organizations, widely supported cooperative organizations, or large commercial firms can conduct. The ultimate objective of such programs is a number of four-way hybrids adapted to different regions or systems of management. The research agencies can then supply the foundation stock for the production of large numbers of these hybrids.

OTTO MACKENSEN is in charge of the bee breeding work of the division of bee culture, Bureau of Entomology and Plant Quarantine. He has specialized in bee breeding and artificial insemination research since 1935, when he joined the staff of the Southern States Bee Culture Laboratory, which is maintained in cooperation with the Louisiana State University at Baton Rouge. Dr. Mackensen, a native of Texas, holds degrees from Texas Agricultural and Mechanical College and Texas University.

WILLIAM C. ROBERTS, an apiculturist in the Bureau of Entomology and Plant Quarantine, has been associated with the North Central States Bee Culture Laboratory and the University of Wisconsin since 1943. Dr. Roberts, a native of Louisiana, is a graduate of the Universities of Louisiana and Wisconsin. Between 1935 and 1943 he worked at the Southern States Bee Culture Laboratory at Baton Rouge, and was an instructor in apiculture at Louisiana State University. His present work in bee breeding includes supervision of the Kelleys Island, Ohio, hybrid queen production project.