by William C. Roberts and Otto Mackensen
U.S.D.A., Agr. Res. Adm., Bureau of Entomology and Plant Quarantine*
(*In cooperation with the Wisconsin Agricultural Experiment Station and Louisiana State University.)

IV. Inbred and Hybrid Bees

MANY beekeepers have produced excellent queen daughters of imported stock or daughters of some selected queen that they obtained from other beekeepers. The new stock queens usually mate with drones of the beekeeper's own stock. If the sex alleles of the two stocks are different, the eggs from most matings are high in hatchability.

The new stock may also be genetically different from the stock that the beekeeper already has. Thus the queen daughters of the imported queen produce hybrid progeny after mating with the drones of the beekeeper's own stock. These may be racial hybrids or hybrids between strains of bees. Such queens often produce excellent colonies with high-quality brood. The beekeeper then praises the imported stock apparently without realizing that his own drones contribute half to the genetic make-up of the workers in these superior colonies.

The story is often quite different after a few generations of back-crossing to the original drones or to drones produced by the hybrid daughter queens. Segregation and recombinations occur, and in the third or fourth year the queens and their progenies are frequently more variable than the original stocks. By this time the beekeeper concludes that the original imported stock was not so good after all or had "run out"; so he gets some other new stock for breeding purposes.

By this procedure the colony superiority may again be restored to a high level. It also falls again after a few years. To keep the quality of his bees up, the beekeeper continues to get new or different breeding stock every few years.

These results are best explained as due to hybrid vigor and the crossing of stocks having different sex alleles. The first-generation colonies were superior because they were hybrids - that is, they were headed by cross-mated queens whose worker progeny were hybrids between
the imported and local stocks. By mathematical analysis of the action of genes in future generations of segregation and recombination after outcrossing, it can be shown that by selection alone the beekeeper is powerless to keep this stock at the high level of production and uniformity exhibited by the first-cross progeny.

Most of the complicated physiological characteristics, such as egg production and vigor are dependent for their maximum expression on the harmonious interaction of many genes and their alleles. Since these genes are located on many chromosomes and there is linkage of genes on the same chromosomes, the likelihood that certain factors conducive to vigor will be linked with injurious factors reduces the chance of obtaining maximum vigor by selection within inbred lines. Moreover, in breeding within a line the sex-allele problem again arises and selection is considerably weakened by a high proportion of low viability matings.

In plant and animal breeding it has been found that certain hybrids are superior to the lines from which they are produced. The superiority
of hybrid corn is unquestioned. Animal breeders have also obtained superior hybrids. Because a large proportion of the progeny is required for breeding purposes to maintain the lines, animal breeders have not used hybrids for extensive production. However, in hogs a modification of the controlled-hybrid program is often used. This is the crisscrossing of three lines, using hybrid sows and purebred boars in each generation. The three-breed crossed pigs are more productive than those of the pure breed. Hybridizing inbred lines of chickens has become more popular and profitable within recent years.

Hybrid vigor is generally thought to be a characteristic of dominance. Dominant genes tend to have more favorable effects than their recessive alleles, and in a hybrid the dominant genes find expression at more loci than do the recessive alleles. If one inbred line is dominant (AA) at one locus and recessive at another (bb) and another inbred line has the reverse (aa and EB) characteristics, the hybrid will have one dominant gene at each locus (AaBb). If both loci contribute to vigor, then the hybrid will be superior to either inbred line. It is also thought that certain genes contribute more to vigor in the heterozygous than in the homozygous condition.

An ideal breeding system is the production and maintenance of purebred but unrelated lines and the crossing of these lines for the production of superior hybrids. Not all - in fact, only a very few - of the inbred lines will give outstanding hybrids when crossed. The breeder, however, can perpetuate those lines that nick favorably for the repeated production of superior hybrids.

Plant breeders can take advantage of hybrid vigor because of the large number of seeds produced in each generation and the relatively low cost of producing seed for the production of hybrids. Animal breeders, on the other hand, are not in this favorable position. The expense of producing and maintaining highly inbred lines together with the small number of offspring makes this practice unprofitable, except with chickens and to some extent hogs.

Honey bees more nearly approach the condition existent in plant breeding. Less than 1 percent of the daughter queens and drones of a queen is needed to perpetuate the line. Whereas only 4 to 10 queens are needed in each generation to insure the continuance of an inbred line, 3 or 4 outcrossed queens can be used to produce 10,000 hybrid queens.

The low cost of producing and maintaining inbred lines of bees, together with the large number of offspring that can be produced from a single breeding individual, suggests that a bee-breeding program based on hybrid vigor is practical. The effect of the sex alleles on egg hatchability further indicates that other breeding plans, such as line breeding, are slow and expensive because too many poor-viability matings occur.

To produce hybrid bees the bee breeder may cross different races, strains, or inbred lines of bees. Unless these bees are homozygous for the desired characters, the resulting hybrids will be variable. Furthermore, other crosses of the same races or strains will differ from each other. A sure method of having uniform hybrids is to have strains homozygous for the desired characters.

The fastest method of increasing homozygosity in a line of stock is to inbreed. Inbreeding is the mating of closely related individuals, such as parent-offspring, brother-sister, or cousins. Related individuals are likely to have many of the same inherited qualities, and mating of these individuals tends to fix these qualities in a homozygous, or pure, condition. Inbreeding is used chiefly for the one purpose of producing homozygosity, or genetic likeness of offspring.

Inbreeding within a population tends to separate the population into many distinct families. Each family becomes uniform within itself but distinctly different from other families. Selection between families of such inbred lines can then be made with more accuracy than selection between individuals. This is especially true for characteristics of low heritability, such as egg production and vigor.

The inbred individuals become lower in average merit than noninbreds. They are not nearly so vigorous or productive. Inbreeding is the severest test of the heredity worth of the individual that can be made, for it causes fixation of both good and bad traits.

Because of the mating habits of bees, it is necessary that all inbreeding matings be made by artificial insemination. After highly inbred lines are obtained it is necessary to control matings to insure keeping the stocks pure. Artificial mating is the only safe method known at present with bees.

If the breeder makes very close matings, such as mother-son, brother-sister, or backcrossing to a queen, he can accomplish very little by individual selection. It is true that he can accomplish a great deal when selecting for color or other visible characteristics that are highly heritable. However, most of these characters have little economic importance. Characteristics of greater importance, such as brood production, vigor, or resistance to disease, are not visibly detectable in individual queens or drones. Selection for these traits is more effective after the lines are inbred.

To determine traits in inbred lines, crosses are made to a line or lines known to possess certain inherited qualities. Thus the test of inbred lines for characteristics of low heritability is to cross all inbred lines to certain "tester" lines and compare the unknown inbred lines with each other. Selection then is between inbred lines.

The mathematical consequences of the various systems of inbreeding have been worked out by several authors. Kalmus and Smith (1948), Crow and Roberts (1950), and Polhemus, Lush, and Rothenbuhler (1950) have reported on the various systems of inbreeding possible with honey bees. The formulas of Crow and Roberts can be used to determine the inbreeding and relationship coefficients of stocks of bees.

Before beginning to produce inbred lines the bee breeder chooses the stocks that he will use. He may desire to cross two or more of the selected noninbred stocks before beginning to inbreed.

He should obtain a few queens each from a large number of sources and test all queens for one season under uniform conditions in a single yard if possible. In this way he can get first-hand knowledge of the potential breeding stocks. He may select some colonies for immediate inbreeding and perhaps allow the others to cross freely for use as future breeding stocks.

It has been shown in both plant and animal breeding that the best hybrids usually come from crosses of unrelated stocks of different origin. The breeder should therefore select a number of different types of bees. He should select some black, some yellow, and some queens intermediate in color rather than all of one color type. He should also select some long, slender and some short, stubby queens. He should obtain some gentle bees and perhaps even one or two lines with a lot of evil temper!

The superior hybrids are genetic heterozygotes. They are the result of crosses between stocks that are unlike genetically. Differences in type, color, or temper denote genetic differences for these characteristics and probably also for other characteristics that the breeder cannot readily see or measure.

There is some evidence that gentleness in bees is dominant over viciousness. In one of our experiments the hybrids of a cross between a vicious and a gentle line were gentle rather than vicious. The vicious line may bring into the hybrid the few genes that mean the difference between an average and a superior hybrid bee. By no means are we suggesting that we need viciousness in bees. This is merely an illustration of how genetic diversity might be obtained.

Once the stocks are chosen, close inbreeding is begun to fix the lines. The breeder should know which matings to make to obtain the desired inbreeding in the shortest time with the least expense. Figure 1 shows the percentage of inbreeding obtained in successive generations by the various systems of inbreeding possible in honey bees. The percentage of inbreeding is the percentage of heterozygous loci of the original selected individuals that become homozygous by inbreeding. Inbreeding has no effect on genes already homozygous in the lines; so we are concerned only with those genes that are not homozygous. Since the bee breeder cannot tell which genes were originally heterozygous and what effect each gene has, he can only measure the increase in purity of the stocks by the percentage of inbreeding.

With present techniques the two systems of mating that increase inbreeding fastest, backcrosses to a male and a mother-son, are not advisable economically. Loss of breeding individuals, and consequently of inbred lines, is high when these systems are followed.

The next most rapid method of increasing inbreeding is by brother-sister matings. This system is recommended, because at present it is the most practical of all systems shown. For the first two generations the systems of brother-sister and aunt-nephew matings and back-crossing to a female are identical in percentage of inbreeding. Because the aunt-nephew system depends upon the use of only one drone for each mating, this system is not so dependable for insuring the survival of an inbred line. Using only one drone to inseminate a queen often results in a poorly inseminated queen, who may turn into a drone layer before the next generation is produced.

Figure 1. Percentages of inbreeding and of heterozygosis (assuming the initial value to be 50 per cent) in successive generations of various systems of inbreeding in honey bees. (From Crow and Roberts, 1950.)

Figure 2. Arrow diagram of a recommended system of mating for inbreeding in bees. After two generations of back-crossing to a selected queen, the line is continued by brother-sister matings.

Because drones mature more slowly than queens, backcrossing to a female produces inbreeding to 37.5 percent faster than brother-sister mating in time consumed per generation for the first two generations. Thus backcrossing to the original queen for the first two generations followed by brother-sister matings in all future generations is recommended for the production of inbred lines in the shortest time with the greatest chance of success in maintaining the inbred line's. Multiple drone matings may be made in all generations to insure good inseminations.

The bee breeder should know what inbreeding will do to his stocks. If he starts an inbred line by back-crossing and then makes brother-sister matings as illustrated in Figure 2, he may expect that each line will become 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 that are sons of A, the egg hatchability of queen B will average 75 percent. A daughter of queen B, queen C, mated to sons of queen A will have egg hatchability that will average either 75 or 50 percent. If her egg hatchability is 50 percent, the line has been reduced to two sex alleles, and queens D, E, and F will also have eggs of 50-percent hatchability if mated as shown in the figure. If queen C has egg hatchability of 75 percent, then D may also be 75 percent but somewhere not far from E or F egg hatchability will probably drop to 50 percent and all future generations will remain at that level.

However, by selection it is possible to keep egg hatchability at 75 per cent, but the breeder would be reducing the effectiveness of inbreeding slightly by selection against certain alleles. It is advisable to select the 50-percent egg-hatchability matings in the C or D generation and thus quickly reduce all inbred lines to two sex alleles, so that they will also have 50-percent egg hatchability. If this is done, an analysis of the sex alleles is more readily accomplished. By test crossing to identify the sex alleles in each line the breeder can then predict which crosses will give high egg hatchability in hybrids and which crosses will give intermediate or low egg hatchability.

Other than improving egg hatchability, the bee breeder can accomplish very little by selection while inbreeding. He can select queens and drones in each generation for color and general appearance, but selection for most other characteristics will be very inefficient.

In one season of inbreeding it is possible to produce a number of sister queens of the D generation and mate them 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-cross the inbred lines at the E generation of queens (50-per-cent inbred). Brother-sister mating should be made each year after the first to insure continuing the inbred lines until they are seleoted in hybrid combinations.

Since egg hatchability in inbred lines with two sex alleles is only 50 percent, the inbreds do not develop colonies with sufficient populations to permit accurate evaluation of such economically important characteristics as honey production, swarming tendencies, and wintering. However, such qualities as tongue length, wing length, and temper may be measured in inbred workers. It is also possible to select for such characters in the queen as size and number of ovarian tubules. This selection in inbreds as inbreds is supplemental to rather than a substitute for selection between inbred lines when in crosses with other inbred lines.

All crosses should be made artificially and the colonies tested under environmental conditions as nearly uniform as possible. The various hybrid crosses will differ, but the individuals of a given hybrid will be uniform as a group.

After it has been found that the colonies having (line 3 x line 4) hybrid queens mated to drones produced by (line 1 x line 2) hybrid queens are superior in production, then the breeding stock to produce this four-way combination can be distributed for commercial production. A commercial queen producer then needs only two artificially mated breeder queens from the breeding source to produce several thousand hybrid queens.

One of these breeder queens is used to produce daughter queens that will head the 50 or more drone-producing coloxies necessary to supply ample drones. An inbred queen of line 1, artificially mated to drones from inbred line 2, is the mother of the hybrid daughters needed to furnish drones. Since drones are produced parthenogenetically, those drone-producing queens can be naturally mated to the drones of the breeder's own stock.

An inbred queen of line 3 artificially mated to drones from inbred line 4 is used as the breeder queen. Her daughters will be hybrids (3x4). These virgins are naturally mated to drones produced by the (1x2) hybrid queens. The honey producer receives (3x4) hybrid queens mated to (1x2) drones, and the worker offspring in his colonies will be four-way hybrids - (1x2) x (3x4).

The test queens produced at Kelleys Island and distributed by the Honey Bee Improvement Cooperative Association are produced in this manner. In the last article of this series we will describe the methods and practices used in large-scale production of four-way hybrid bees at Kelleys Island.

Literature Cited

Crow, James F., and William C. Roberts. 1950. Inbreeding and Homozygosis in Bees. Genetics 35: 612-621.

Kalmus, H., and C. A. B. Smith. 1948. Production of Pure Lines in Bees. Jour. Genetics 49: 153-158.

Polhemus, Martin S., Jay L. Lush, and Walter C. Rothenbuhler. 1950. Mating Systems in Honey Bees. Jour. Heredity 41: 151-155.