Hybrid Status of Honey Bee Populations Near the Historic Origin of Africanization in Brazil

Original article

Hybrid status of honey bee populations near the historic origin of Africanization in Brazil

WS Sheppard AEE Soares D DeJong H Shimanuki

1 USDA-ARS, Bee Research Laboratory, Bldg 476, BARC-E, Beltsville, MD 20705, USA; 2 Univ de Sao Paulo, Faculdade de Medicina de Ribeirao Preto, Dpto de Genetica, Ribeirao Preto, 14 049, SP, Brazil

(Received 29 November 1991; accepted 2 January 1992)

Summary — Africanized honey bee populations are genetically heterogeneous across their extensive new world range. Over 35 years have elapsed since the introduction of A m scutellata to southeastern Brazil and we hypothesized that populations from this region should have achieved the high- est degree of genetic equilibrium following the perturbation of introduction. We report here the results of a population genetic study of honey bees sampled near the origin of neotropical Africaniza- tion combining analyses of morphological, allozyme and mtDNA characters. Data from this study support previously reported allozyme frequency estimates and support the expectation that populations from this region are comparatively stable in genetic composition; and further, that significant polymorphism of European origin persists in the Africanized population of the region. Morphological and mtDNA data from these neotropical populations reveal the strong influence of the African race, A m scutellata. Apparent discordance among data sets from the several analytical methods reflects variation in selection and population size on the inheritance or persistence of such characters and indicates the importance of multiple character analysis.

Africanized honey bee / Brazil / population genetics / morphometry / mt DNA / enzyme polymorphism

INTRODUCTION 1988). Despite the varied African, Europe- an and Middle Eastern origins of intro- Over two dozen subspecies of the honey duced subspecies, it is generally held that bee, Apis mellifera, occur throughout its new world feral and managed honey bee old world range (Ruttner, 1988). A subset populations were largely or wholly of Eu- of these races has been introduced to the ropean derivation prior to the mid 1950s new world over the past 4 centuries (Kerr, (Michener, 1975; Taylor, 1985; Rinderer, 1967; Morse et al, 1973; Sheppard, 1986). At that time the South and Central 1988), although the effective sampling of African honey bee race, Apis mellifera races for some of the introductions was scutellata Lepeletier was introduced to quite limited (Aeppler, 1922; Sheppard, tropical Brazil through both the inadver- tent release of scutellata queens and de- range. For example, Africanized honey liberate distribution of African x European bee populations near their climatic limits in racial hybrids to improve Brazilian bee- the temperate zone (Sheppard et al, 1991) keeping (Kerr, personal communication, in or in recently colonized neotropical regions Spivak et al, 1991). Prior to the release of (Del Lama et al, 1990) containing a sub- A m scutellata, neotropical A mellifera pop- stantial pre-existing European population ulations were poorly adapted to tropical (Rinderer et al, 1991) can exhibit signifi- ecological and climatological conditions cant amounts of both African and Europe- and existed primarily in managed apiaries. an allozyme or mtDNA characters and hy- The density of European-derived feral col- brid morphologies. Established neotropical onies during that period was apparently Africanized populations show a lesser in- quite low (Michener, 1975; Taylor, 1988). fluence of European races based on simi- However, the success of the African honey lar characters or nuclear DNA markers bee introduction can be measured in the (Lobo et al, 1989; Hall, 1990; Moritz and extensive and rapid range expansion that Meusel, 1991) or virtually none at all (Hall followed. Within 20 years the descendent and Muraldiharan, 1989; Smith et al, "Africanized" honey bees reached their 1989). When considered in the context of southern limits in Argentina (Kerr et al, the current geographical range of bees 1982), while their northward expansion with African-derived genetic characteris- continues in the southern US more than tics, the heterogeneous nature of new 35 years after the initial Brazilian introduc- world Africanized honey bee populations is tion. not an unexpected result, although the A number of hypotheses have been phenomenon of Africanization and the proposed regarding the mechanisms un- "Africanized" population has sometimes derlying "Africanization". These range from been more narrowly defined (Hall, 1990). suggestions that genetic contributions by As the geographic range of the Africanized both races produce more or less "hybrid" honey bee has expanded, the most recent- populations, with the relative advantage of ly occupied regions undergo a transition to either race being dependent upon a num- Africanization over a period of at least sev- ber of ecological or behavioral factors eral years (Boreham and Roubik, 1987; (Kerr and Bueno, 1970; Michener, 1975; Taylor, 1988; Rinderer et al, 1991). Given Rinderer et al, 1985; Rinderer, 1986) to that the number of generations following the notion that expanding neotropical pop- contact between African- and European- ulations are maintained as a relatively derived populations differs throughout the pure African gene pool through natural se- extant range of Africanization, one might lection, the effects of pre-mating isolating therefore predict the region of initial con- mechanisms or hybrid inviability (Taylor, tact in southeastern Brazil to have 1985; Fletcher, 1991). Consequently, the achieved the greatest equilibrium. Recent genetic composition of Africanized honey analysis of honey bees from this region, in bee populations, vis-à-vis the relative con- contrast to an earlier report (Smith et al, tribution of progenitor European and Afri- 1989), found evidence of prior hybridiza- can races has been a matter of considera- tion between the 2 groups, based on the ble research and discussion. Despite presence of a presumptive European differences, one generalization possible mtDNA marker, although the population from the body of research is that African- was morphometrically indistinguishable ized honey bee populations appear hetero- from African A m scutellata (Moritz and geneous throughout their geographical Meusel, 1991). A large-scale allozyme study also revealed that populations of this ta in Rio Claro, Brazil (Kerr, 1967) and at least region express significant levels of Europe- 60 km from each other (fig 1). The locations and an-derived introgression, although morpho- number of colonies sampled were as follows: Ri- beirao Preto, SP (13), Luis Antonio, SP (14), are "African" metrically they quite (Lobo et Santa Rosa do Viterbo, SP (33), and Ibitiuva, In this we the rela- al, 1989). paper report SP (66). All colonies originated as collected tive genetic contribution of African- and Eu- swarms maintained without subsequent queen ropean-derived subspecies to Brazilian management practices, and therefore are con- honey bee populations near the original sidered representative of local feral populations. of adult workers were collected alive site of introduction of A m scutellata, based Samples and then frozen in liquid nitrogen for later on a combined analysis of mtDNA, mor- mtDNA and allozyme analyses or stored in phological and allozyme character data. EtOH for morphometric study.

MATERIALS AND METHODS Allozyme analysis Samples of honey bees were collected from 126 colonies at 4 locations within the state of Sao Ten individuals per colony were characterized Paulo, Brazil. These locations were all within for 3 enzyme loci, Mdh-1, Pgm-1 and Hex-1, 200 km of the initial release site of A m scutella- known to be polymorphic in honey bees. Gene frequencies at these loci have been widely used Purified mitochondrial DNA was prepared in studies of Africanization and racial hybridiza- from fresh honey bee flight muscle as follows: tion in the honey bee (Cornuet, 1982; Sylvester, 8 g of degastered adults were homogenized on 1982; Badino et al, 1983; Sheppard and ice in a modified Erway grinder in 2-g aliquots McPheron, 1986; Spivak et al, 1988; Lobo et al, with 40 ml of mitochondrial isolation medium 1989; Smith et al, 1989; Del Lama et al, 1990). (MIM) buffer (220 mM mannitol, 70 mM sucrose, Homogenates of individual bees were run on 2 mM HEPES, 1 mM EDTA, 1 mM EGTA; pH standard starch gel conditions and visualized 7.5). Mitochondria were isolated through a se- using appropriate histochemical techniques ries of differential centrifugations and then lysed (Mdh-1 and Pgm-1; Sheppard and Berlocher, with 1% SDS for 5 min at 25 °C. Protein was 1984; Sheppard and McPheron, 1986; Hex-1: precipitated with cesium chloride (vol x 0.22 g/ Del Lama et al, 1988). Allozymes were scored ml) on ice for 15 min and pelleted at 12 000 g. using relative mobilities and gels were photo- The density of the supernatant was adjusted graphed for documentation. Analysis of allo- with additional cesium chloride to 1.63 g/ml and zyme data was performed with the Biosys-1 pro- the mixture was placed into a Beckman quick- gram of Swofford and Selander (1981). seal tube. Ethidium bromide was added (1 mg/ tube) and the sealed tubes were centrifuged in a Beckman 80 Ti rotor at 50 k rpm for 18-24 h. The resulting band of purified closed circular Mitochondrial analysis mtDNA was removed and repurified with an additional CsCl density gradient centrifugation (Beckman SW65 rotor) at 40 k rpm for 18-24 h. The mtDNA band was removed, treat- Total nucleic acids were extracted from 2 indi- repurified ed with butanol to remove ethidium bromide and the method viduals/colony, generally following then and of Sheppard and McPheron (1991). Restriction drop-dialyzed (Marusyk Sergeant, to remove the salt. Radioactive was were made on of the resus- 1980) probe digests aliquots from this bee mtDNA total nucleic acids to condi- produced purified honey pended according standard nick translation conditions tions specified by the supplier (Bethesda Re- using (Mani- search Laboratories, Bethesda, MD). All atis et al, 1982). samples were digested with Eco-R1 and Bcl-1, Prehybridization and hybridization of nitrocel- enzymes that have been used to provide diag- lulose filters occurred at 50 °C with standard so- nostic identification of African mtDNA (Smith, lutions containing salmon sperm DNA and 25% 1988; Hall and Muralidharan, 1989; Sheppard et formamide (Maniatis et al, 1982). Hybridized fil- al, 1991; Rinderer et al, 1991). Samples were ters were washed, dried and placed on X-ray run on 1% agarose gels overnight at 24 V and film for 6-24 h for autoradiography. Fragment nucleic acids transferred to nitrocellulose filters sizes were estimated from the autoradiographs using standard techniques (Southern, 1975). using a digitizer and Bioscan software. scutellata (Nunamaker and Wilson, 1981; Sylvester, 1982; Smith et al, 1989; Lobo et al, 1989) was 0.82, similar to previous estimates from large-scale studies of Sao Paulo honey bee populations (table II) and identical to that reported from Rio Claro, the initial site of African honey bee release (Lobo et al, 1989). The mean frequency for the Hk83 allele, reported only from African- derived populations (Del Lama et al, 1988; Spivak et al, 1988) and variation at the Pgm locus were remarkably similar to previous studies of populations from this region (Lobo et al, 1989; Del Lama et al, 1990) and indicates the degree of genetic equilibrium of this population some 35 yr after Africanization. Contingency table analysis of heterogeneity among populations, performed with the HETXSQ option of BIOSYS, revealed non-significant het- Morphometric analysis erogeneity for all loci among the populations (P = 0.83, 18 df). Inclusion of published data from Sao Paulo State Ten bees from each sample were dissected, (table II) stained and slide mounted as described by Daly gave similar results for the MDH locus (P = and Balling (1978). For each bee, 25 characters 0.99, 16 df) and supports the existence of were measured using a microscope slide projec- extensive homogeneity among these 9 tor and Houston Instruments digitizer. The re- Brazilian populations. sulting data were analyzed with a discriminant Mitochondrial of the colonies analysis program in widespread use for African- analysis ized honey bee identification (Daly and Balling, produced the restriction fragment patterns 1978). Samples were scored as "Africanized" or "typical" of A m scutellata (Smith, 1988). "European" based on at least 98% probability of An additional Eco-R1 polymorphism report- group membership. A m scutellata is not readily ed from Argentine and Mexican African- discriminated from neotropical Africanized hon- ized honey bees (Rinderer et al, 1991; ey bees by this program, although it can detect and known to occur Africanized populations containing genetic con- Sheppard et al, 1991) tributions from any of several European races in the honey bees of Spain (Sheppard, un- (Sheppard et al, 1991). Samples that were out- published observations) was not found. side of the 98% probability criteria were scored Recently Smith et al (1991) have reported as "non-discriminated". that some colonies of honey bees collected in southern Spain, considered to be the RESULTS AND DISCUSSION race A m iberica based on geography, share a mitochondrial Eco-R1 polymorphism with A m scutellata and many Afri- Results of the allozyme survey are sum- canized populations. However, they were marized in table I. The mean frequency of able to differentiate these Spanish honey the Mdh100 allele, commonly believed to bees from A m scutellata by screening with be fixed (freq = 1.0) or nearly so in A m an additional enzyme, Hinf-1. Although we did not examine our Brazilian samples with pressed by A m scutellata might be at a Hinf-1, analysis of over 30% of them with a selective advantage. 1.4 kb honey bee mtDNA probe cloned Extranuclear markers with specific pat- from the A+T-rich region revealed a scutel- terns of inheritance, while also assumed to rather than A m infermissa or lata, iberica, be selectively neutral, may be eliminated origin for the mtDNA in our samples (SSP- from populations through their association 1 digestions of honey bee total nucleic with specific selected genotypes. Thus, in acid extractions with this probed fragment the case of maternally inherited European produce fragment patterns diagnostic for mtDNA, the genetic composition of colo- several African m subspecies, including A nies produced by European virgin queens scutellata; Sheppard, unpublished obser- mated with Africanized drones approaches vations). 50% African-derived genes as a maximum. Morphometric analysis of the colonies In the neotropics, these colonies may be at determined all samples to be Africanized such a competitive disadvantage com- with a probability of at least 0.98. The pared to colonies headed by African overall discriminant score for the collection queens (which approach a maximum of was 4.18 (SD = 1.04). In 125 out of 126 100% African-derived genes) that selection cases, colony discriminant scores were could be a very strong barrier to the entry over 2.16, the score that defines the 0.99 of European mtDNA into the feral popula- level of probability for determining African- tion. However, such hybrid colonies would ized honey bees (S Buco, personal com- still provide an early source for entry of Eu- munication). ropean genes into the feral population, their drones As a group the 4 populations of honey through mating by and, thus, be in bees were quite "Africanized" based on may important contributing Europe- morphology and all individuals exhibited an-derived allozyme genes to Africanized "African" mtDNA. However, allozyme data populations such as found in this study. In locations with reduced selection for from these populations are in accord with tropical previous studies indicating that Brazilian genotypes, such as the temperate regions of leads to the dis- Africanized honey bees are descended Argentina, hybridization from both African and European progeni- semination of mtDNA into the "opposite" tors (Lobo et al, 1989; Del Lama et al, morphologies on either side of the hybrid 1990; Moritz and Meusel, 1991). Such zone (Sheppard et al, 1991; Sheppard, un- seemingly incongruous conclusions in published observations). population genetics can perhaps be best The validity of using African and Euro- explained by differences in inheritance and pean mtDNA frequencies to assess the ge- selective parameters influencing the vari- netic composition of Africanized honey bee ous genetic markers (Sene et al, 1988). populations rests on prior knowledge, or at Thus, while allozymes are presumed to be least reasonable estimates, of relative pop- largely neutral, with respect to selection ulation sizes (Page, 1989). While the popu- (Berlocher, 1984), morphology may be lation of European colonies in neotropical more closely linked to physiological and Brazil was apparently small, relative to the behavioral traits of tropical and temperate quickly-established feral AHB population, adapted African and European races, re- over 1 million European-derived colonies, spectively. In the neotropics then, traits ad- part of an extensive beekeeping industry, vantageous to tropical survival and ex- was present in the Yucatan of Mexico prior to the arrival of AHB to that region. Rinder- Les échantillons d’abeilles ont été préle- er et al (1991) reported that neotropical Af- vés dans 126 colonies en 4 endroits de ricanized populations collected from this l’état de Sâo Paulo, Brésil. On a caractéri- region several years after Africanization sé 10 individus par colonie à l’aide de 3 express high levels of European mtDNA. locus d’enzyme, Mdh-1, Pgm-1 et Hex, Thus, both natural selection and relative connus pour leur polymorphisme chez l’abeille. Les population size may play roles in the ulti- analyses morphométriques colonie ont été faites mate genetic composition of Africanized de 10 individus par honey bee populations. en suivant la procédure de Daly et Balling (1978). L’ADNmt de toutes les colonies a Based on the data herein on presented été caractérisé par les digestions par les feral bees near the honey sampled geo- enzymes de restriction EcoR1 et Bcl-1 des graphic origin of Africanization in the new extractions totales d’acides nucléiques et world, and genetic investigations from simi- par électrophorèse sur gel d’agarose. Les lar sites in Sao Paulo State, Brazil (Lobo et fragments de restriction ont été visualisés al, 1989; Del Lama et al, 1990; Moritz and par hybridation avec une sonde «nick- Meusel, 1991), it is evident that Africanized translatée» faite à partir d’ADNmt purifié honey bees from these long-established d’abeilles, suivie d’autoradiographie. populations express substantial mtDNA En tant que groupe, les 4 populations and morphological similarities to the tropi- d’abeilles sont, d’après leur morphologie, cal race, A m scutellata, while still retaining très «africanisées» et toutes possèdent un evidence of prior hybridization. The com- ADMmt «africain». Pourtant, les données positional flux or stability of these popula- des allozymes provenant de ces populations as they become established in tem- tions concordent avec les précédentes étu- abeilles brési- perate North America will certainly be a des qui indiquaient que les liennes africanisées descendent à la fois matter of great interest to population ge- neticists and apiculturists alike. de parents africains et européens (Del Lama et al, 1990; Lobo et al, 1989; Moritz et Meusel, 1991). L’homogénéité des esti- Résumé — Statut hybride des popula- mations de la fréquence des allozymes les étudiées ici et tions d’abeilles près du lieu d’origine de parmi populations précé- demment indiquent que les populations de l’africanisation au Brésil. Les populations cette région ont une composition généti- d’abeilles africanisées sont génétiquement que relativement stable. Les données de la dans toute leur aire de répar- hétérogènes et de l’ADNmt de ces popula- tition dans le Nouveau Monde. Plus de 35 morphologie tions néotropicales montrent une forte in- ans se sont écoulés l’introduction depuis fluence de la race africaine, A m scutellata. d’A m scutellata dans le sud-est du Brésil La discordance apparente entre les en- et nous émettons l’hypothèse que les po- sembles de données issues des diverses pulations de cette région ont atteint le plus méthodes analytiques reflète la variabilité haut degré d’équilibre génétique après la des influences de la sélection et de la taille perturbation due à l’introduction. Nous don- de la population sur la transmission ou la nons ici le résultat d’une étude de généti- persistance de tels caractères et montre que des populations d’abeilles échantillon- l’importance d’une analyse par des métho- nées près de la source de l’africanisation des variées. néotropicale, réalisée en combinant les analyses des caractères morphologiques abeille africanisée / Brésil / génétique avec ceux des allozymes et de l’ADN mito- des populations / morphométrie / chondrial (ADNmt). ADNmt / polymorphisme enzymatique Zusammenfassung — Hybridzustand heren Untersuchungen überein, welche von Honigbienenpopulationen nahe gezeigt haben, daß die brasilianischen afri- dem historischen Ursprung der Afrika- kanisierten Bienen Abkömmlinge sowohl nisierung in Brasilien. Die Populationen von afrikanischen wie von europäischen von afrikanisierten Honigbienen sind hete- Vorfahren sind (Del Lama et al, 1990; rogen über ihrem gesamten weiten Ver- Lobo et al, 1989; Moritz and Meusel, breitungsgebiet in der Neuen Welt. Seit 1991). Die Homogenität der Schätzungen der Einführung von. A m scutellata nach von Allozymhäufigkeiten bei den hier und Südostbrasilien sind jetzt über 35 Jahre früher untersuchten Populationen weist verstrichen und wir stellen die Hypothese darauf hin, daß die Populationen aus auf, daß die Populationen aus dieser dieser Region in ihrem genetischen Region nach den Störungen der Einfüh- Aufbau relativ stabil sind. Die morphologi- rungszeit jetzt den höchsten Grad eines schen und mtDNA-Daten dieser neotropi- genetischen Gleichgewichtes erreicht schen Populationen zeigen den starken haben sollten. Wir berichten hier über die Einfluß der afrikanischen Rasse A m scu- Resultate einer populationsgenetischen tellata. Die offensichtlichen Unstimmigkei- Studie an Bienenproben, die nahe dem Ur- ten zwischen den Ergebnissen dieser ver- sprung neotropischer Afrikanisierung ge- schiedenartigen Untersuchungsmethoden sammelt wurden. Dabei wurden Analysen geben unterschiedliche Einflüsse der Se- von morphologischen, Allozym- und lektion und der Populationsgröße auf die mtDNA-Merkmalen kombiniert. An vier Vererbung und den Bestand dieser Merk- Orten des Staates Sao Paulo, Brasilien, male wieder und unterstreichen die Bedeu- wurden Proben von 126 Bienenvölkern ge- tung einer Analyse nach unterschiedlichen sammelt. Zehn Individuen je Volk wurden Methoden. nach den drei, bei Honigbienen polymor- phen Enzymloci Mdh-1, Pgm-1 und Hex afrikanisierte Honigbiene / Brasilien / charakterisiert. Die morphometrische Ana- Populationsgenetik / Morphometrie / lyse von 10 Bienen je Volk wurde nach mtDNA / enzymatischer Polymorphis- dem Verfahren von Daly und Balling mus (1978) durchgeführt. Die mitochondriale DNA von allen Völkern wurde durch Ab- bauprodukte von Extrakten der Gesamtnu- ACKNOWLEDGMENTS kleinsäuren mit dem ECOR1- und BCL-1- Restriktionsenzym und Agarosegel- We thank LW Dick, HR Yoo and CM Arias for Elektrophorese charakterisiert. Die Restrik- technical assistance in analyzing samples. CM tionsfragmente wurden durch Hybridisie- Arias and D Zarlenga collaborated in producing rung mit einer nick-translatierten Sonde the honey bee mtDNA probe for differentiating hergestellt aus gereinigter Honigbienen- African races (manuscript in preparation). Bee- mtDNA, und nachfolgender Autoradiogra- keepers M Taveres and AC Meda kindly allowed us to sample their colonies. A portion of the col- sichtbar Als waren phie gemacht. Gruppe lecting expenses and travel for one of the au- die vier nach ihrer Bienenpopulationen thors (WSS) was provided by a joint BID-USP Morphologie und der bei allen Individuen grant. We thank BA McPheron, SH Berlocher, N gezeigten "afrikanischen" mtDNA vollkom- Schiff, S Passoa and GJ Steck for valuable dis- men ’afrikanisiert’. Aber die Allozymdaten cussion and comments on earlier drafts of the von diesen Populationen stimmten mit frü- manuscript. REFERENCES Kerr WE, Del Rio SL, Barrionuevo MD (1982) The southern limits of the distribution of the Africanized honey bee in South America. Am Aeppler CW (1922) About that quarantine. Am Bee J 122, 196-198 Bee J 62, 259 Lobo JA, Del Lama MA, Mestriner MA (1989) Badino G, Celebrano G, Manino A (1983) Popu- Population differentiation and racial admix- lation structure and Mdh-1 locus variation in ture in the Africanized honey bee (Apis mellif- mellifera Apis ligustica. J Hered 74, 443-446 era L). Evolution 43, 794-802 Berlocher SH (1984) Insect molecular systemat- Maniatis T, Frisch EF, Sambrook J (1982) Mo- ics. Annu Rev Entomol 29, 403-433 lecular Cloning. Cold Spring Harbor, NY, Boreham MM, Roubik DW (1987) Population 545 pp change and control of Africanized honey Marusyk R, Sergeant A (1980) A simple method bees (Hymenoptera: Apidae) in the Panama for dialysis of small-volume samples. Anal Bi- Canal area. Bull Entomol Soc Am 33, 34-39 ochem 105, 403-404 Cornuet JM (1982) The MDH polymorphism in Michener CD (1975) The Brazilian bee problem. some west Mediterranean honeybee popula- Annu Rev Entomol 20, 399-416 tions. Proc 9th Congr IUSSI, Boulder, CO, Moritz RFA, Meusel MS Mitochondrial USA (1991) gene frequencies in Africanized honeybees Daly H, Balling SS (1978) Identification of Afri- (Apis mellifera L): theoretical model and em- canized honeybees in the western hemi- pirical evidence. J Evol Biol (in press) discriminant J Kans Ento- sphere by analysis. Morse RA, DM, Ambrose JT, Conner mol Soc 857-869 Burgett 51, WE, Fell RD (1973) Early introductions of Af- Del Lama MA, Figueiredo RA, Scares AEE, Del rican bees into Europe and the new world. Lama SN (1988) Hexokinase polymorphism Bee World 54, 57-60 in Apis mellifera and its use for Africanized Nunamaker RA, Wilson WT bee (1981) Comparison honey identification. Rev Bras Genet 11, of MDH in the African hon- 287-297 allozyme patterns ey bee (Apis mellifera adansonii L) and the Del Lama MA, Lobo JA, Scares AEE, Del Lama Africanized populations of Brazil. J Kans En- SN (1990) Genetic differentiation estimated tomol Soc 54, 704-710 by isozymic analysis of Africanized honey Page RE Jr (1989) Neotropical African bees. bee populations from Brazil and from Central Nature (Lond) 338, 181-182 America. Apidologie 21, 271-280 Rinderer TE (1986) Africanized bees: the Afri- Fletcher DJC of (1991) Interdependence genet- canization process and potential range in the ics and ecology in a solution to the African United States. Bull Entomol Soc Am 32, 222- bee problem. In: The African Honey Bee 227 (Spivak M, Fletcher DJC, Breed MD, eds), Rinderer Hellmich RL Danka Collins Westview, Boulder, CO, 77-99 TE, III, RG, AM (1985) Male reproductive parasitism: a Hall HG (1990) Parental analysis of introgres- factor in the Africanization of European hon- sive between African and Euro- hybridization ey bee populations. Science 228, 1119-1121 pean nuclear DNA RFLPs. honeybees using Rinderer Stelzer Buco Genetics 125, 611-621 TE, JA, Oldroyd BP, SM, Rubink WL (1991) Hybridization between Eu- Hall Muralidharan K Evidence from HG, (1989) ropean and Africanized honey bees in the mitochondrial DNA that African honey bees neotropical Yucatan peninsula. Science 253, spread as continuous maternal lineages. Na- 309-311 ture (Lond) 339, 211-213 Ruttner F (1988) Biogeography and Taxonomy Kerr WE (1967) The history of the introduction of Honeybees. Springer, Heidelberg of African bees to Brazil. S Afr Bee J 2, 3-5 Sene FM, Pereira MAOR, Vilela CR (1988) Con- Kerr WE, Bueno D (1970) Natural crossing be- trasting patterns of differentiation inferred tween Apis mellifera adansonii and Apis mel- from traditional genetic markers in the pro- lifera ligustica. Evolution 24, 145-148 cess of speciation. Pacif Sci 42, 81-88 Sheppard WS (1988) A comparative study of genomes in Spanish honeybees (Apis mellif- enzyme polymorphism in US and European era iberica). J Hered 82, 96-100 bee honey populations (Hymenoptera: Api- Southern EM (1975) Detection of specific se- dae). Ann Entomol Soc Am 81, 886-889 quences among DNA fragments separated Sheppard WS, Berlocher SH (1984) Enzyme by gel electrophoresis. J Mol Biol 98, 503- polymorphism in Apis mellifera from Norway. 517 J Apic Res 23, 64-69 Spivak M, Ranker T, Taylor O Jr, Taylor W, Da- Sheppard WS, McPheron BA (1986) Genetic vis L (1988) Discrimination of Africanized variation in honey bees from an area of ra- honey bees using behavior, cell size, mor- cial hybridization in western Czechoslovakia. phometrics, and a newly discovered isozyme Apidologie 17, 21-32 polymorphism. In: Africanized Honey Bees and Bee Mites RE, Del- McPheron BA Ribosomal (Needham GR, Page Sheppard WS, (1991) finado-Baker Bowman Ellis Hor- DNA in In: in M, CE, eds) diversity Apis. Diversity Apis wood, Chichester, 313-324 (Smith DR, ed) Westview Press, Boulder, CO, 89-102 Spivak M, Fletcher DJC, Breed MD (1991) In: The "African" Honey Bee (Spivak M, Fletcher Sheppard WS, Rinderer TE, Mazzoli J, Stelzer DJC, Breed MD, eds) Westview, Boulder, Shimanuki H Gene flow between JA, (1991) CO, p3 African- and European-derived honey bee Swofford Selander RB BIOSYS-1 : a in Nature 349, DL, (1981) populations Argentina. (Lond) FORTRAN for 782-784 program the comprehensive analysis of electrophoretic data in population Smith DR (1988) Mitochondrial DNA polymor- genetics and systematics. J Hered 72, 281- phisms in five Old World subspecies of hon- 283 bees and in new world In: Afri- ey hybrids. HA identifica- canized Bees and Bee Mites Sylvester (1982) Electrophoretic Honey tion of Africanized bees. J Res GR, RE, Delfinado-Baker honey Apic (Needham Page 21, 93-97 M, Bowman CE, eds) Ellis Horwood, Chi- chester, 303-312 Taylor OR Jr (1985) African bees: potential im- pact in the United States. Bull Entomol Soc Smith DR, Brown WM, Taylor OR (1989) Neo- Am 31, 15-24 Africanized have African tropical honeybees OR Jr and economic im- mitochondrial DNA. Nature 321, 674- Taylor (1988) Ecology (Lond) of African and Africanized bees. 679 pact honey In: Africanized Honey Bees and Bee Mites Smith DR, Palopoli MF, Taylor BR, Garnery L, (Needham GR, Page RE, Delfinado-Baker M, Cornuet JM, Solignac M, Brown WM (1991) Bowman CE, eds) Ellis Horwood, Chichester, Geographical overlap of two mitochondrial 29-41