CHAPTER THREE - INTERHIVE VARIABILITY

Since the 1960's, resistance of insects to a number of insecticides has been documented. Resistance of houseflies to DDT was the first in a long and continuing series of pest species which have become resistant to chemical controls. Only recently has this phenomenon been exploited to select for pesticide resistant beneficial insects and mites. It might be expected that honey bees could also be selected for resistance to pesticides. This study concentrated on examining various populations of honey bees for resistance to synthetic pyrethroid insecticides in all three races of bees commonly sold in North America.

Background and Objectives

The development of resistance in insect populations is influenced by a number of factors. These include the generation time of the species, the amount of insecticide pressure on the population, the mobility of the species, the amount of outbreeding with populations not resistant and whether or not the females mate with more than one male (Georghiou 1980). In a population of leafminers in a greenhouse environment for example, it can be seen how each of these factors favors the development of resistance. A population of honey bees presents a very different picture, however.

Genetic changes in honey bees colonies only occur when the queen is replaced (Collins 1980). This may happen under several conditions, but, it most commonly occurs when the colony swarms or when the old queen is superseded. In managed colonies, the intercession of the beekeeper in replacing the queen must also be considered. The most frequently this might be expected to occur would be once every year or two. This is a very long time considering the development of insecticide resistance.

It is possible that colonies exposed to pesticides will be more likely to supersede their queen. This does provide the opportunity for the selection of a larva that is surviving at a time when the colony is challenged by an insecticide. Neither swarming or beekeeper replacement of queens is likely to provide any selection for insecticide resistance. Only strong colonies are likely to swarm and it may be assumed that colonies exposed to substantial levels of insecticide would not be strong enough to swarm. Beekeepers do not have a commercial source of pesticide resistant queens available. Queens are not likely to come from colonies which have been selected for any trait other than honey production. Beekeepers raising their own queens may be able to do some indirect selection for pesticide resistance by choosing colonies that perform well in areas of pesticide use.

The mating behavior of queen honey bees is also not well suited for the development of insecticide resistance. The virgin queen leaves the hive at a few days of age and mates with 5 to 10 drones. From this mating she will store all the sperm necessary for egg laying for her entire life. The drones she mates with are likely to be from colonies other than her own, and may come from colonies several miles away. Because drone bees do not contribute to the welfare of a colony, weaker colonies are likely to produce fewer drones than strong colonies. Therefore colonies challenged by insecticides are not likely to produce as many drones as other colonies.

The development of resistance by selection also requires the assumption that some level of natural resistance exists in the population (Graves 1965). In honey bees a number of traits have been selected for at one time or another (Rothenbuhler 1980). Some of these programs have been quite successful, such as breeding for increased pollen collection (Boelter 1984). Other honey bee queen breeders claim to have strains selected for resistance to various honey bee diseases, honey producing ability, overwintering ability, a preferred color or resistance to some pesticides (Tucker 1980). Most of these traits have been selected for from within one race; others from the offspring of selected crosses. Differential resistance to carbaryl was shown between lines of Apis mellifera ligustica Spin, and A. m. scutellata Lepeltier by Danka (1986). The relative impact of race and previous selection pressure on the two lines was not addressed by Danka.

In the United States, the honey bee genetic pool has been closed since 1912 when Congress closed the United States to the importation of live bees. This was done to prevent the introduction of a mite pest which was having devastating effects on the honey bee population of England at the time (Phillips 1925, Rennie 1921). The border was not closed to Canada. Canada had adopted similar laws, but had excluded New Zealand as well as the U.S. and therefore genetic material could come to the U.S. from New Zealand via Canada. Since 1912, there have been some very limited importations under permit from the USDA as well.

At the time the border was closed in 1912, there were four races of bees known to be present in the United States (Culliney 1983, Severson 1985). These were Apis mellifera ligustica Spin or the Italian, by far the most popular and common; A. m. mellifera L. or the German or Black bee, the first race imported by settlers but an aggressive and unpopular race; A. m. caucasica Gorb. or the Caucasian, a gentle grey bee noted for its ability to overwinter well but not popular for other reasons; and A. m. carnica Pollmann or the Carniolan, also a gentle grey bee but noted for its tendency to swarm. None of these races were genetically pure, with the possible exception of the Italian. The lack of artificial insemination at the time and the sheer number of italian bees, made keeping pure lines of the other races nearly impossible.

Over time A. m. mellifera became nearly impossible to find in a distinguishable form, although its influence on some strains is still detectable. Renewed interest in the grey races (A. m. carnica and A. m. causica) resulted in some breeders selecting strains that showed morphological and behavioral characteristics very close to those of these races as seen in their native areas (Carlisle 1955). Instrumental insemination greatly aided this work because it allowed the use of single drone controlled matings. Today relative pure lines of A. m. ligustica, A. m. laucasica and A. m. carnica are available in the U.S.

A long term selection program for honey bee lines was started in England in 1912 when the native strains were decimated by the Isle of Wright Disease, now believed to be a combination of infection with Nosema apis Zander and infestation with Acarapis woodii Renie (Bullamore 1922). This program was carried out by Brother Adam at Buckfast Abbey. Brother Adam has traveled the world looking for different genetic lines of bees that might contribute to his breeding program and he has imported a number of these races and strains to Buckfast Abbey. His intensive selection and crossing have resulted in a world famous line of bees now known as Buckfast. This line was imported under permit into the United States as eggs which were reared into virgin queens and inseminated with semen also imported under special permit from the USDA. The Buckfast bee is now sold by a single licensed breeder in the U.S. (Sugden 1983). These queens are mated naturally to wild drones, however, so that their progeny are only 50% Buckfast genotype.

The objectives of this study were: 1) To look for resistance to one or more synthetic pyrethroid insecticides in genetic lines of bees representing all three common races found in North America and a line representing a long term selection program from England. 2) To determine if any resistance found was related to race of the colony, possible previous selection pressure or simply individual variation between colonies.

Materials and Methods

In the summer of the year previous to the study, colonies of bees were established in a common site near West Lafayette. This was accomplished by removing frames of brood from established colonies and introducing marked queens representing the different lines selected for the study. All of these colonies were established in new hives so that previous pesticide exposure would not be a factor. The old frames removed from other colonies and used to establish the new hives were replaced with new frames as soon as possible.

The queens selected were as follows: two caucasian lines, one each from two breeders; two carniolian lines, one each from two breeders; two buckfast queens from the sole breeder in the U.S.; two Italians from one breeder and two Italians from hives in a high pesticide use area of southern Indiana which had not been requeened for at least seven years. These ten lines allowed comparisons between races, between queens from the same breeder, and between bees which could be reasonably assumed to have been under selection pressure and those not under such pressure within the same race or from other races.

The seven treatments consisted of a control of 50% sucrose and carbaryl as Sevin 50W, permethrin as Ambush 2EC and fluvalinate as Spur 2.4 EC in 50% sucrose at 1 and 10 PPM each. The bees were handled and mortality measured as previously described. All bees were held in the dark in a single environmental chamber at 25 degrees C. The study was conducted as a randomized complete block design and replicated four times. Each treatment was applied to a single test unit (cup of 25 bees) except for one fourth of the treatments which were duplicated to provide an estimate of between cup variance. The duplicated treatments comprised a fifth complete set of insecticide-bee source combinations which were randomly assigned to a replication.

Results and Discussion

Significant differences were found between the different sources of bees examined. These differences were related to the race of the colony. There was surprising little variation between the different sources of the same race. No significant differences in susceptability were found in the hives which were assumed to have been exposed to insecticides and those which were assumed to not have been exposed.

Data Handling.

The data were analyzed using the ANOVA, GLM and MEANS procedures of SAS. Mortality is presented as the mean percent of bees dead at each of six observations made over five days. Means were tested for significant differences using Duncan's Multiple Range test with the error rate set at 5%. Two data points of the 350 combinations of 10 hives and 7 treatments for five replications were missing due to the loss of one or more observations used to calculate the mean percentage of bees dead per observation. The significance of differences did not change if the fifth replication was dropped.

Hive to Hive Variation.

There was considerable variation between the hives as to their susceptibility to the products tested (Figure 14). Differences in overall mortality as measured by the mean percentage of bees dead per observation (Table 4) were found to be significant at the 5% level. The ANOVA indicated that variance between duplicate cups within a hive/treatment/replication combination was not a significant component of the total variance (<1%). Analysis by treatment showed that differences between hives were significant for only two treatments, carbaryl and permethrin at 10 PPM (Table 5).

Analysis of Racial Differences.

The influence of race on the hive differences was assessed by grouping the hives into Italian, Caucasian, Carniolan and Buckfast. While this latter group is not a true race, it does represent a different genetic lineage than the other hives and does not fit well into any of the other races, although each of these races was used at some point in the development of the Buckfast bee.

Figure 15 shows the mean percent mortality by race. Each race is represented by two hives except Italian which was represented by four hives. The Italian and Caucasian races were significantly different from the Carniolan and Buckfast races. Further analysis of the four hives grouped as Italian was conducted by splitting the group into Indiana and California strains. Since the Indiana strain was assumed to have been exposed to the materials in question, while the California strain was selected from breeding stock in Canada in an area of no pesticide use, differences representing selection pressure should have been evident between these two groups. The means and standard deviations of the Indiana and California strains were 13.4%(6.1) and 15.9%(7.4), respectively. The difference was not significant the 5% level.

The racial differences were also examined by treatment. Figure 16 shows the mean percent mortality by treatment. The treatments were: Trt. 1 - carbaryl 1PPM, Trt. 2 - carbaryl 10PPM, Trt. 3 - permethrin, Trt. 4 - permethrin 10PPM, Trt. 5 - fluvalinate 1PPM, Trt. 6 - fluvalinate 10PPM, and Trt. 7 - the control of 50% sucrose only. The Italian race was significantly more sensitive to carbaryl than any other race. The Caucasian race was significantly more sensitive to permethrin than the other races and while not significant, showed an interesting sensitivity to fluvalinate at 10 PPM.

Discussion.

Racial differences in susceptibility to the treatments examined were noted. It was especially interesting to note that the Italian race, the most popular in the United States, was the most susceptible to carbaryl and more susceptible to permethrin than the Carniolan and Buckfast races. The similarity between the Buckfast race and the Carniolan is not surprising since the Carniolan was a major line used by Brother Adam in developing the Buckfast bee. It was also very interesting to note the small amount of variation between the different sources of a race compared to racial differences.

Since the so-named Indiana and California strains of the Italian race were not positively known to have represented lines selected under insecticide pressure and a lack of such pressure respectively, the lack of significance of the difference should not be taken to suggest that selection for increased tolerance to carbaryl or the synthetic pyrethroid insecticides would not be effective. The racial differences would suggest that racial hybrids should be considered in any breeding program whose aim was increased resistance to insecticides.