DETERMINING LOW LEVELS OF AFRICANIZATION IN UNMANAGED HONEY BEE COLONIES USING THREE DIAGNOSTIC TECHNIQUES by Katherine Darger A thesis submitted to the Faculty of the University of Delaware in partial fulfillment of the requirements for the degree of Master of Science in Entomology Spring 2013 Copyright 2013 Katherine Darger All Rights Reserved ABSTRACT â€œKiller beesâ€ arrived in the United States in the year 1990. Questions have arisen regarding low levels of Africanization in regions bordering the locations with established, Africanized bees. Honey bees were collected and examined using three methods of testing to determine levels of Africanization. With morphometrics, mitochondrial DNA, and nuclear DNA tested with the use of microsatellites we found that the known Africanized bees collected by the Florida Department of Agriculture did not exhibit Africanization other than in the preliminary, morphometric test performed by the Department of Agriculture. Fast Africanized Bee Identification System (FABIS) is the field test for preliminary identification of suspected Africanized honey bees (Rinderer 1986). FABIS measures forewing length, fresh weight, dry weight, and femur length. This process takes approximately twenty minutes per colony (Sylvester and Rinderer 2009). Once the colony is identified with FABIS, the sample is analyzed with USDA-ID. Universal System for Detecting Africanization (USDA-ID) (Sanford 2006) is the test necessary to officially declare a case of Africanization in the Unites States. This process is laborious due to the obligation to accuracy. Twenty-five mounted parts of each specimen are needed to determine whether a colony is Africanized. This requires thorough training to insure accuracy, as it includes miniscule measurements, like the number of hamuli on the hindwing, the length and width of the basitarsus, and the distance between wax mirrors (Rinderer et al. 1993). Two newly established morphometric methods of analysis are less expensive and faster than the well-established methods. The Automatic Bee Identification System (ABIS) uses a comparison of plotted wing-vein junctions with a digital image of the forewing of the specimen. The process takes two minutes per sample and the accuracy rate is estimated to be 98.05% among bee species and 94% among honey bee subspecies (Francoy et al. 2008). Geometric morphometric analysis is used to delineate between honey bee subspecies. The test takes five minutes (Francoy et al. 2006) and has a 99.2% estimated accuracy rate looking at the wing venation angles (Francoy et al. 2008, Francoy et al. 2009). Due to its precision and short preparation time, this test was used on the samples in this study. USDA-ARS Carl Hayden Honey Bee Research Center and the Florida Department of Agriculture and Consumer Services; Bureau of Plant and Apiary Inspection, respectively. The samples from the Florida Department of Agriculture were collected in Alabama, Georgia, and Florida but were grouped together for this study. The samples from Jerry Hayes, formerly of the Florida Department of Agriculture, were grouped together to increase the samples size. Known Africanized samples from South America as well as African samples from Pretoria and Kenya were provided by Dr. Walter S. Sheppardâ€™s personal collection from Washington State University. Forty known Africanized honey bee samples were sent from Belize by Brenna Traver. Altogether, there are 82 samples of known African/Africanized honey bees. Morphometrics For our morphometric diagnoses, ten right forewings per colony from an assortment of colonies among our sampling collection were removed (Table 5). Our samples and the control samples were diagnosed in Brazil by Dr. David De Jong using the geometric morphometric technique of comparing wing venation (Francoy et al. 2006, Francoy et al. 2008, Francoy et al. 2009). The results are displayed in Principle Component Analysis figures to exhibit the clustering of populations according to morphometric restrictions. Total DNA Extraction A hind leg from two bees per sample, totaling in 188 unmanaged bees and 81 known Africanized bees, were cut into 4 to 6 pieces and placed in 150 μl of 10% Results The collection contained limited samples from certain states due to the time constraints of this project affecting the number of collecting seasons. Morphometrics Through the geometric wing analysis performed by Dr. David De Jong, several patterns emerged. The unmanaged and managed samples clustered together, separately from the representative type of each subspecies (A. m. liguctica, A. m. mellifera, A.m. carnica and A. m. caucasica) (Figure 2), showing that the bees in the US are a distinct hybrid. The known Africanized samples separated from each other, however the samples from Florida, Georgia, and Alabama all clustered together but outside the range of Africanized bees (Figure 3). The Africanized bees from Arizona clustered separately from the known Africanized bees from Florida as well as the bees from Africa and Brazil (Figure 3). Based on morphological characters, the bees from Florida may not be fully Africanized. Mitochondrial Testing The unmanaged and managed honey bee colonies were tested and found to have European mitotypes (Table 7). The Africanized samples from Florida and Belize contained European mitotypes. The Africanized samples from Arizona had a variety of mitotypes from the African lineage: A26, A26c, A1e predominantly, and A1. The Africanized samples from Belize had four European mitotypes that were C1 while the rest of the mitotypes were mostly A1 and A1e. The samples from Africa and Brazil had no European mitotypes. They had mitotypes of A4, A47, A26a, and A26c (Table 1). T Tests were performed in Microsoft Excel to test for significant differences between the Africanized populations according to mitotypes. The p values for each comparison were well below the significance level of 0.05, suggesting that the Africanized samples from Belize, Arizona, Brazil, and Africa had significantly more mtDNA of African origin compared to the mitotypes of Africanized samples collected from Florida, Alabama, and Georgia. In the comparison of Africanized bees from Florida versus Arizona significant differences were found (p value less than 0.0001,60 degrees of freedom). The comparison of the bees from Florida versus Brazil and Africa was also significant (p value less than 0.0001, 31 degrees of freedom). The p value for the comparison of Florida versus Belize was 0.001 (96 degrees of freedom). Even the bees from Arizona, Brazil, Africa, and Belize were significantly different from each other based on mitotype. The p value for Arizona versus Brazilian and African samples was 0.000991 (53 degrees of freedom), while the comparison of Arizona versus Belize was 0.0075038 (118 degrees of freedom). The bees from Brazil and Africa were different from the bees from Belize (p value = 0.0000509, 89 degrees of freedom). Microsatellites The average number of alleles in each population tested ranged from 5.23 Â±2.12 in Africanized bees from Florida to 7.85 Â± 3.25 in unmanaged populations (Table 2). Allelic richness was the lowest in the Africanized bees from Florida while the Africanized bees from Arizona had the highest allelic richness measured (Table 2). The expected heterozygosity was very similar with the largest range being between 0.79Â±0.1 (in samples from Africa and Brazil) and 0.62Â±0.2 (from the Africanized in Florida/Alabama/Georgia and Managed populations) (Table 2). The STRUCTURE output consistently provided two populations: Unmanaged and managed clustering with the Africanized samples from Florida/Alabama/Georgia versus Africanized from Arizona, Africanized from Belize, and samples from Africa and Brazil (Figure 4). STRUCTURE is a program that uses multi-locus data to infer population structure, assign individuals to populations, and is often used to study hybrid areas. FST stands for fixation index and shows population differentiation due to structure variation as measured by Single Nucleotide Polymorphisms (SNPs) or microsatellites. According to the FST results (Table 3), there was a significant difference between the managed and unmanaged populations. There was also significance between the African and managed samples. Highly significant FST values were noted between Africanized bees from Arizona and managed bees and also between Africanized bees from Arizona and the unmanaged bees. Importantly, the Africanized samples were not significantly different from each other. Through the use of Genepop, Hardy Weinberg equilibrium was tested using Fis, or inbreeding coefficient, estimates to compare each locus for every population. In each instance, using the Fisherâ€™s Method, the results were highly significant, thus the populations are not in equilibrium. Allele frequency data, which is the proportion of alleles compared to the number of genes, can be found in Table 6 in the Appendix. Discussion Currently, Africanized bees are officially diagnosed through morphometric methods only. The genetics of a honey bee are not often utilized in official determinations of Africanization. Should we combine morphometric and molecular data to identify the level of Africanization? Determining the process of Africanization as either incomplete or absolute can be done with the use of mitochondrial and nuclear DNA, as was performed in this study, and will serve as a good tool for tracking the introgression of Africanized genes into commercial and unmanaged honey bee populations in North America. Morphometrically, the data shows that there is an American bee, an amalgam that is distinctive from the originating, European subspecies (Figure 2). The GWV morphometric study further distinguishes between Africanized bees from Arizona and those from Africa and Brazil. Interestingly, the samples from Florida/Alabama/Georgia which were originally diagnosed as Africanized by the USDA-ID technique were found to have European ancestry using mtDNA markers, European morphology based on GWV and grouped with managed and unmanaged honey bee colonies in a population structure analysis using microsatellite markers. A panmictic population is one in which all forms of recombination are possible due to a lack of restriction caused by genetics and behavior, and all fertile individuals are potential partners. We hypothesized that the samples would separate into distinct populations delineated by lines of latitude. The samples tested in our study are not in Hardy Weinberg equilibrium, as shown through the use of the Fisherâ€™s test in Genepop, which is indicative that they are not separate and distinct populations as would be expected based on geographic collecting locations. Essentially, our study indicates that bees on the east coast are not in distinct populations as would be surmised based on collection latitude. Northern bees were not differentiated from southern bees. Studies in Europe showed the M lineage, consisting of western and northern European bees, is genetically much more similar to A, or the African lineage than to C (eastern Europe) or O lineages (near East and central Asia) (Whitfield et al. 2006). Our data shows that two populations could be discerned based on microsatellite allele frequency. The Africanized samples from Arizona and Belize were genetically similar to the African bees from Africa and Brazil based on allele frequency data on the tested microsatellites. Furthermore, the unmanaged, managed, and Africanized bees from Florida/Alabama/Georgia shared enough alleles to be considered one population. These low levels of introgression can be seen in the Africanized samples collected from Florida/Alabama/Georgia (STRUCTURE output Figure 4) from the known African and Africanized populations from Arizona, Belize and Brazil. However, the Africanized samples from Florida/Alabama/Georgia still grouped with unmanaged and managed samples collected along the east coast. This could be due to the very recent Africanization of Florida (2010) or could be due to an incomplete Africanization of the entire east coast as supplanted by the migratory queen rearing and caged bee trade. High linkage disequilibrium in A. m. ligustica and A. m. mellifera accounts for the potential for high genetic variation and recombination (Whitfield et al. 2006). The opposite is true for A. m. scutellata and A. m, intermissa (Whitfield et al. 2006), accounting for infiltration of A. m. scutellata into European genetics and not the other way around. The mtDNA data supplants this conclusion, as the samples of Africanized bees from Florida/Alabama/Georgia did not exhibit Africanized mitotypes in the majority of the samples, rather they had European maternal influences, which were also found in the managed and unmanaged samples. According to Whitfield et al. (2006), New World bees in Brazil were highly Africanized, yet individuals had alleles from the M lineage. Our data shows bees from Florida had low levels of Africanization. Mid to late infiltration time showed substantially, but not exclusively, Africanized genes (Whitfield et al. 2006) which is substantiated by our data; the Africanized bees from Arizona, one of the earliest states declared Africanized, were determined to be Africanized based on morphometrics, mitochondrial DNA, and when tested with microsatellites. During the process of Africanization in Latin America, bees have historically had African nuclear and mitochondrial markers become the majority after the first five to ten years following the invasion (Schneider et al. 2004). This increase of African nuclear and mitochondrial genes over time is supported by our findings: the samples from Florida/Alabama/Georgia have European mtDNA and low levels of Africanized alleles while samples from Arizona consist of mostly African mitotypes and are accompanied by African nuclear alleles. The onset of the invasion into the southeastern U.S. as opposed to the desert southwest is supported by the observed differences in genetic composition between the two sets of samples and is due to the fact that Arizona was one of the earliest states colonized by Africanized bees due to the proximity to Central America. It stands to reason that this amount of time allowed for more complete introgression to occur, making Arizona further along in the Africanization process. All of the managed and unmanaged honey bees had European mitotypes (Table 7). In Texas, the first state to have Africanized bees, it was suggested that the rate of Africanization was enhanced by the decimation of European feral and managed colonies by the Varroa destructor mite (Pinto et al. 2005). The Varroa mite created a selection pressure as Africanized bees survive infestation. It was also suggested that Texas may be the northernmost range of Africanization possible (Pinto et al. 2005), while the most recent map of Africanization in the United States shows that as being a false prediction. This map, however, was constructed using USDA-ID as the diagnostic tool. If molecular tools were used diagnostically, as in our study, Florida would not be included. It is possible that the diagnostic techniques (USDA-ID) for determining Africanized swarms may be sensitive only to phenotype that is not yet linked to genotype and could find a diagnosis in bees with few genetically African indicators. Genetics should also not be used alone diagnostically. Through the use of mtDNA testing along with microsatellite data it was shown that the forefront of the Africanized bee movement is caused by Africanized drones mating with European queens (Pinto et al. 2005). The resulting hybrid drones were tested and found to be at a competitive disadvantage (Hall 1991) because of reduced flight speed, eventually leading to direct competition between European and Africanized queens (Pinto et al. 2005). Our data shows that the Africanized bees from Florida/Alabama/Georgia were morphometrically Africanized only using USDA-ID not the GWV technique, and not necessarily genetically similar enough to the African bee to be considered an Africanized bee. The Africanized bees from Florida/Alabama/Georgia were different from the other Africanized samples from other locations using each of the three testing methods. Using GWV morphometric techniques, the Africanized samples from Florida/Alabama/Georgia clustered together as being identical to the managed and unmanaged samples from the north, central, and southern aspects of the eastern seaboard of the United States and not within the same populations as the Africanized bees from Arizona, Brazil, or Africa. No clinal gradient in the degree of Africanization was observed in the unmanaged samples collected along the eastern seaboard using microsatellite markers. Using mitochondrial testing, the mitotypes found in the Africanized bees from Florida/Alabama/Georgia proved to be of European ancestry which differed from the mostly pure African ancestry of the African and Africanized samples collected from Arizona, Belize, and Brazil. The difference in mitotypes found in the samples from Arizona, Belize and Africa warrant further research into possible routes of entry of Africanized colonies. Using microsatellite analysis to reconstruct actual population dynamics, the Africanized bees from Florida/Alabama/Georgia clustered into a population with the unmanaged and managed bees from the east coast while the populations from Arizona, Belize, Brazil, and Africa clustered together into another separate population. The lack of definitiveness in the three diagnostic tools leads us to the conclusion that sensitivity of genetic markers may not be the most useful factor in determining desirable stock for managing honey bees. Perhaps the best way to evaluate a hive is by phenotypic traits, such as aggressive tendencies, and other undesirable traits such as swarming and absconding, and not by genotypic traits. Containing a blend of markers denotes a level of Africanization but perhaps negative behavior should be the first line of diagnosis. Excessive stinging and swarming indicates a need for further tools to be used, but even without the negative diagnoses these traits should be discouraged. In conclusion, I hypothesized that unmanaged honey bee colonies have blends of genes from European and African influences and low levels of Africanization occur throughout the United States, and can be detected using microsatellite markers. This blend of genes was made evident by comparing morphometric and mitochondrial test results. The results from the USDA-ID technique conflicted with the results of the geometric wing venation, which depicted the Africanized samples from Florida/Alabama/Georgia as non-Africanized. Future Plans For future analysis of our data, more software using Bayesian analysis will further indicate population structure and degree of admixture. Bayesian Analysis of Population Structure (BAPS) is one program that shows population structure and information on introgression. Before processing the samples, we will prepare the data by forming â€œgroup-wise clusteringâ€ of the individuals from each population (Corander et al. 2009). All of the data will be pre-processed through the BAPS program prior to analysis of genetic information (Corander et al. 2009). Format is the key to successful analysis. In the future, the reasoning for the difference in sensitivity of the tests should be explored. Perhaps the diagnoses do not necessitate genotype, as morphologically and behaviorally the colonies are being confirmed to be Africanized. The USDA-ID test is performed after a colony is deemed overly aggressive in the field. With those two delineations, a need for genotypic examination may be unnecessary. Three of the tools used: geometric wing analysis, mtDNA testing, and microsatellites showed the Africanized bees from Florida, already confirmed Africanized by USDA-ID, to be incompletely or not at all Africanized. The probable Africanized drone front is a possibility in the Africanized counties in Florida, leading to the incomplete genetic introgression. The possibility of exchanging the morphometric program with a purely molecular diagnostic program is appealing in that it is more precise, less expensive, and less time consuming, but the sensitivity may under-estimate the aggressiveness of the hive due to incomplete introgression. REFERENCES Alpatov, V. V. 1929. Biometrical studies on variation and the races of bees. Quarterly Review of Biology 4: l-58. Calderone, NW. 2012. Insect Polllinated Crops, Insect Pollinators and US Agriculture: Trend Analysis of Aggregate Data for the Period of 1992-2009. PLoS ONE 7(5): e37235. Doi:10.1371/journal.pone.0037235 Clarke, K. E., Oldroyd, B. P., Javier, J., Quezada-EuÃ¡n, G., & TE Rinderer. 2001. Origin of honeybees (Apis mellifera L.) from the Yucatan peninsula inferred from mitochondrial DNA analysis. Molecular ecology, 10(6), 1347â€“55. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11412359 Clarke, K. E., Rinderer, T. E., Franck, P., Quezada-EuÃ¡n, J. G., & B.P. Oldroyd. 2002. The Africanization of honeybees (Apis mellifera L.) of the Yucatan: a study of a massive hybridization event across time. Evolution; international journal of organic evolution, 56(7), 1462â€“74. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12206246 Corander, J., P. Marttinen, J. SirÃ©n, & J. Tang. 2009. BAPS: Bayesian analysis of population structure manual. v. 5.3 Last updated 10/28/2009 Ã…bo Akademi University, Finland. http://web.abo.fi/fak/mnf/mate/jc/software/BAPS5manual.pdf Delaney, D.A. 2008. Genetic characterization of U.S. honey bee populations. Doctor of Philosophy thesis at Washington State University. Delaney, D.A. , M.D. Meixner, N.M. Schiff, & W.S. Sheppard. 2009. Genetic characterization of commercial honey bee (Hymenoptera: Apidae) populations in the United States by using mitochondrial and microsatellite markers. Genetics 102(4): 666-673. Estoup A., L. Garnery, M. Solignac, & J.M. Cornuet. 1995. Microsatellite variation in honeybee (Apis mellifera L.) populations: Hierarchical genetic structure and test of the infinite allele and stepwise mutation models. Genetics 140: 679- 695. Francoy T.M., P.R. Prado, L.S. GonÃ§alves, L.D. Costa, & D. De Jong. 2006. Morphometric differences in a single wing cell can discriminate Apis mellifera racial types. Apidologie 37: 91-97. Francoy, T.M., D. Wittman, V. Steinhage, M. Drauschke, S. MÃ¼ller, D.R. Cunha, A.M. Nascimento, V.L.C. Figueiredo, Z.L.P. Simoes, D. De Jong, M.C. Arias, & L.S. GonÃ§alves. 2009. Morphometric and genetic changes in a population of Apis mellifera after 34 years of Africanization. Genetics and Molecular Research 8(2): 709-717. Francoy T.M., D. Wittmann, M. Drauschke, S. M_ller, V. Steinhage, M.A. Bezerra Laure, D. De Jong, & L.S. GonÃ§alves. 2008. Identification of Africanized honey bees through wing morphometrics: two fast and efficient procedures. Apidologie 39: 488-494. Frank P., L. Garnery, A. Louiseau, B.P. Oldroyd, H.R. Hepburn, M. Solignac, & J.M. Cornuet. 2001. Genetic diversity of the honeybee in Africa: microsatellite and mitochondrial data. Heredity 86: 420-430. Garnery L, JM Cornuet, & M Solignac. 1992. Evolutionary history of the honey bee Apis mellifera inferred from mitochondrial DNA analysis. Molecular Ecology 1:145â€“154 Garnery, L., M. Solignac, G. Celebrano, & J.M. Cornuet. 1993. A simple test using restricted PCR-amplified mitochondrial DNA to study the genetic structure of Apis mellifera L. Experientia 49: 116-1021. Hall, H. G., & DR Smith. 1991. Distinguishing African and European honeybee matrilines using amplified mitochondrial DNA. PNAS, 88(10), 4548â€“52. Hall, H.G. & K. Muralidharan. 1989. Evidence from mitochondrial DNA that African honey bees spread as continuous maternal lineages. Nature 339: 211- 213. Statistics Canada. 2010. Honey highlights. Date modified 12-23-2010, date accessed 3-6-2011. http://www.statcan.gc.ca/pub/23-221-x/2010000/part-partie1- eng.htm Kalinowski, S.T. 2004. Counting alleles with rarefaction: private alleles and hierarchical sampling designs. Conservation Genetics 5: 539-543. Kraus, F. B., Franck, P., & R Vandame. 2007. Asymmetric introgression of African genes in honeybee populations (Apis mellifera L.) in Central Mexico. Heredity, 99(2), 233â€“40. doi:10.1038/sj.hdy.6800988 McGregor, S.E. 1976. Insect pollination of cultivated crop plant. US Government Printing Office, Washington, DC. Needham, G.R., R.E. Page, M. Delfinado-Baker, and C.E. Bowman. 1988. Africanized honey bees and bee mites. Ellis Horwood Limited. Neilsen,D.L., P.R. Egbert, G.J. Hunt, E. Guzman-Novoa, S.A. Kinnee, & R.E. Page. 1999. Identification of Africanized honey bees incorporating morphometrics and improved Polymerase Chain Reaction mitotyping procedure. Annals of the Entomological Society of America 92(2): 167-174. Pinto, M. A., WL Rubink, JC Patton, RN Coulson, & JS Johnston. 2005. Africanization in the United States: replacement of feral European honeybees (Apis mellifera L.) by an African hybrid swarm. Genetics, 170(4), 1653â€“65. doi:10.1534/genetics.104.035030 Pritchard, J. K., Stephens, M., & Donnelly, P. 2000. Inference of population structure using multilocus genotype data. Genetics, 155(2), 945â€“59. Rinderer, T.E. 1986. Bee genetics and breeding. Academic Press. Rinderer, T.E., H.A. Sylvester, M.A. Brown, J.D. Villa, D. Pesante, & A.M. Collins. 1986. Field and simplified techniques for identifying Africanized and European honey bees. Apidologie 17: 33-48. Rinderer, T.E., S.M. Buco, W.L. Rubink, H.V. Daly, J.A. Stelzer, R.M. Riggio, & F.C. Baptista. 1993. Morphometric identification of Africanized and European honey bees using large reference populations. Apidologie 24: 569-585. Ruttner, F. 1988. Biogeography and taxonomy of honeybees. Springer-Verlag Berlin Heidelberg publishing, Germany. Ruttner, F., L. Tassencourt, & J. Louveaux. 1978. Biometrical â€“ statistical analysis of the geographic variability of Apis mellifera L. Apidologie 9: 363-381. Sanford, MT. 2006. The Africanized Honey Bee in the Americas: A Biological Revolution with Human Cultural Implications. American Bee Journal. In 5 parts: March-July. Schiff, N.M., W.S. Sheppard, G.M. Loper, & H. Shimanuki. 1994. Genetic diversity of feral honey-bee (Hymenoptera, Apidae) populations in the southern United States. Ann. Entomol. Soc. Am. 87: 842-848. Schneider, S. and G. DeGrandi-Hoffman. 2002. The influence of paternity on virgin queen success in hybrid colonies of European and African honeybees. Animal Behaviour 65 (5): 883-892. Schneider S., T. Deeby, D. Gilley, and G. Degrandi-Hoffman. 2004. Seasonal nest usurpation of European colonies by African swarms in Arizona, USA. Insectes Sociaux 51: 359-364. Shaibi, T., H.M.G. Lattorff, R.F.A. Moritz. 2008. A microsatellite DNA toolkit for studying population structure in Apis mellifera. Molecular Ecology Resources 8: 1034-1036. Sheppard, W.S. 1989. A history of the introduction of honey bee races into the United States, part I. American Bee Journal 129(9): 617-619. Sheppard, W.S. 1997. Subspecies of Apis mellifera pp. 519-533 in: R.A. Morse & K Flottum (Eds.), Honey Bee Pests, Predators and Diseases, A.I. Root Co., Medina, OH, USA. Sheppard, W. S, & M.D. Meixner. 2003. Apis mellifera pomonella , a new honey bee subspecies from Central Asia. Apidologie. 34: 367â€“375. doi:10.1051/apidomondia Vergara C., A. Dietz, & A. Perez. 1989. Usurpation of managed honey bee colonies in migratory swarms in Tabasco, Mexico. American Bee Journal 129: 824- 825. Vergara C., A. Dietz, & A. Perez de Leon. 1993. Female parasitism of European honey bees by Africanized honey bee swarms in Mexico. Journal of Apicultural Research 32(1): 34-40. Walsh P.S., D.A. Metzger, & R. Higuchi. 1991. Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. Biotechniques 10 (4): 506-513. Whitfield, C. W., Behura, S. K., Berlocher, S. H., Clark, A. G., Johnston, J. S., Sheppard, W. S., Smith, D. R., et al. 2006. Thrice out of Africa: ancient and recent expansions of the honey bee, Apis mellifera. Science (New York, N.Y.), 314(5799), 642â€“5. doi:10.1126/science.1132772 Wilson, M.L. 1988. The impact of a tropical-evolved honey bee in temperate climates of North America. Pp. xx-xx. In: G. R. Needham (Ed.). Africanized honey bees and bee mites. Ellis Horwood Limited publishing co. Winston, ML. 1992. Killer bees: the Africanized honey bee in the Americas. Harvard press. Cambridge, MA. 162 pp.