UGA Bee Breeding Project

Fig 1. Each colony in the breeding program is composed of five deep Langstroth combs. Each is provided a food-storage super above a queen excluder and syrup feeder as needed. Colony fronts are marked with identifying numbers and distinctive geometric shapes to encourage homing by workers and queens.

Overview. The UGA Bee Lab maintains a closed population breeding program designed to select simultaneously for (1) reduced colony Varroa levels, (2) high brood production and (3) viability, (4) hygienic behavior, (5) high honey production, and (6) gentleness. We are using conventional propagation and mating methods (maternal selection, drone saturation, and open mating) – techniques easily adaptable to any beekeeping operation. We gratefully acknowledge the Georgia Beekeepers Association for a startup grant in 2002 which enabled us to purchase initial queens and 50 nucleus colonies for housing queens and making selections (Fig. 1) . Additionally, Mr. Bob Binnie of Rabun County made a generous donation of bees. Efforts are ongoing to maintain funding for this program with, potentially, a productive life of 20 years.

Fig 2. Spotty brood.

Genetics and Breeding Theory behind the Program. We are approximating the closed population breeding model developed by Robert E. Page, Jr., Harry Laidlaw, and co-workers. The closed population method permits a breeder to select for targeted traits in a delimited queen population ‘closed’ from uncontrolled introgression of new genes. Thus, selection can proceed rapidly. However, the breeder must also minimize the inbreeding that occurs simultaneously at the sex locus. In honey bees this inbreeding has the undesirable consequence of poor brood viability, expressed as a brood pattern that is “spotty” (Fig. 2). Hence the closed population model seeks to progressively improve a stock on some pre-selected criteria while at the same time minimizing the effects of homozygosity at the sex locus. In the so-called mass-selection / random mating approach, superior queens in the closed population are selected to produce all the queens for the next generation. Although the expected rate of sex allele loss is high with such intense maternal-side selection, this disadvantage is partially offset by inseminating daughters with a homogenized mixture of semen representing the whole closed population. Since we aim to work without instrumental insemination, we are achieving this population-wide representation of drones via drone saturation with drones produced by every queen in the population. We acknowledge that our open-mating scheme compromises the ‘closed’ aspect of the theoretic ideal. However this liability is offset by reduced risk of inbreeding at sex loci, elimination of queen performance problems associated with II, and greater adoption by industry.

Large numbers of breeder colonies in a closed population program help guarantee a long life to the project. Based on the theoretic work of Laidlaw and Page, we conservatively estimate that our closed population has >90% probability of maintaining at least 85% brood viability for 20 generations.

A dedicated apiary of fifty nucleus colonies was set up in 2002 to house queens, perform selections, and propagate naturally-mated daughters. Each nucleus colony is composed of five deep Langstroth combs, a food-storage super above a queen excluder, and a syrup feeder. Queens, purchased or donated, were chosen with the intention of maximizing genetic variation and sex allele number in the incipient closed population. Some queens were used that had already been selected for suppressed mite reproduction or hygienic behavior. Sources of queens included: Robert Binnie, Rabun Co., GA; Glenn Apiaries, Fallbrook, CA; Heitkams’ Honey Bees, Orland, CA; Jesse McCurdy, Perry, GA; Dann Purvis, Blairsville, GA; Rossman Apiaries, Moultrie, GA; Shumans Apiaries, Baxley, GA; B. Weaver Apiaries, Navasota, TX; Carl Webb, Clarkesville, GA; and Wilbanks Apiaries, Claxton, GA.

Fig. 3. Sticky sheet for monitoring Varroa levels.	Fig. 4. Measuring brood area.

Colony Varroa levels are determined with standard sticky sheets modified to fit the nucleus hives; 24-hour natural mite drop can linearly predict colony mite populations (Fig. 3). Brood area is determined by measuring sealed brood to the nearest cm² with a marked grid (Fig. 4). Brood solidness (inheritance discussed above) is determined by placing a grid that delimits 100 cells over a section of randomly-chosen sealed brood and subtracting empty cells to determine the percentage brood solidness. Hygienic behavior is appraised in the field by freezing a section of brood with liquid nitrogen and determining the percentage of freeze-killed brood removed after 24 hours (Fig. 5). Seasonal honey production is predicted by 7-day colony net weight gains (Fig. 6). Defense behavior is measured by dragging a leather patch across the tops of exposed combs for a total of 60 seconds and counting the number of stings received by the patch (Fig. 7). Measurements of all characters is replicated at least twice in the season, confirming queen identity and survival each time.

Fig. 5. Liquid N2 hygienic test.	Fig. 6. Measuring weight gain.	Fig. 7. Defense test.

Measures are first transformed to the same unit of standard deviations to make them comparable. Then the characters are weighted in the following manner to produce a Selection Index (SI): brood area (0.1), brood solidness (0.3), colony Varroa levels (-0.2), hygienic behavior (0.2), honey production (0.1), and defense behavior (-0.1).

SI = (cm² brood x 0.1) + (% solidness x 0.3) - (Varroa x 0.2) + (% hygienic x 0.2) + (weight gain x 0.1) - (no. stings x 0.1)

The index is generated and used to select the top 20% queens (highest index score). These procedures were followed in 2003 and we now have selected queens from which we will rear daughters for the 2004 season. All nucs in the breeding population will be requeened with daughters from these selected queens. Ripe queen cells are placed on comb surfaces under a wire emergence cage; upon emergence the virgin queen is marked and after she has successfully mated her wings are clipped. Drone saturation in the mating area is encouraged by providing each colony a comb of drone cells derived from drone-sized foundation. Thus, the compromise inherent in the closed population model is approximated – maternally-selected daughters are out-crossed with all (including non-selected) queens in the population.

And the Future? Once we begin realizing measurable improvements in our selected characters we plan to solicit funds to subsidize commercial scale production of the stock for use by beekeepers. Modest numbers of production queens may be available for sale at the Bee Lab on a pick-up basis.