Heredity 74 (1995) 315—325 Received 5 July 1994 Genetical Society of Great Britain spired, conical shells reduceBURROWS, the M. areas T. AND HUGHES, exposed R. N. 1989. Natural to foraging solar of the radiation and thus helpdogwhelk, prevent Nuce!la lapillus thermal (L.); the weather stress and whether (Vermeij, 1993). This shapeCHANG, H. W. isAND characteristicEMLEN, J. M. 1993. Seasonal variation of of Racial admixture of Apis meiifera in littorinid snails from uppermicrohabitat intertidal distribution of theareas polymorphic in land the snail densitytropics (Vermeij,also induce 1978). marked Differences differences in in population shell form Tasmania, Australia: similarities and (Boulding & Hay, 1993) COOKE,as do A. H. adaptations 1915. The geological distribution between of Purpura differences with natural hybrid zones in populations living amongCROTHERS, boulders i. H. 1985. or Dogwhelks: on rock an introduction faces to the Europe inter-related,it is very difficult vary with to separateoverall size them, and withand shelleach wear,may interact with the colour of the shell to give a popula- BENJAMIN P. OLDROYDt*, JEAN-MARIE CORNUETtt, DARRYL ROWEI-, THOMAS E. FORBES, E. AND HANLEY, s. 1838. A History of British Mollusca RINDERER & ROSS H. CROZIERI- environmental factors couldand explain Their Shells, vol. the 3, pp. unusual 380—387. J. Van result Voorst, 1-School of Genetics and Human Variation, La Trobe University, Bundoora, Vic 3083, Australia, 1-Laboratoire de from Porthtowan (R= 9.0, %L= 43.9). Nevertheless, Neurobiologie Comparáe des Invertébrés, INRA-CNRS URA 1190, La Guyonnerie, B.P. 23, 91440 Bures-sur- Yvette, this study shows that thermalHEDRICK, i'.ecology w. 1986. Genetic polymorphism is one in factorheterogeneous France and § Honey-Bee Breeding, Genetics and Physiology Research Laboratory, USDA-ARS, 1157 Ben Hur Road, controlling the distribution of colour morphs in Baton Rouge, LA 70820, U.S.A. 1990. Genotypic-specific habitat selection: a specific habitat selection. newThis model may and its application. promote Heredity, genetic 65, 145—149. Feraland domesticated honey bees were collected across the island state of Tasmania, Australia, polymorphism (Hedrick, 1990).HUGHES, R. DifferentialN. 1972. Annual production behaviour of two Nova Scotian and typed for malate dehydrogenase and a mitochondrial DNA polymorphism. They were also of the various morphs maypopulations help ofto Nucella explain lapillus. Oecologia, the exten-8, 356—370. compared morphometrically with reference specimens of Apis mellifera ligustica and A. m. HUGHES, R. N., BURROWS, M. T. AND ROGERS, S. E. 1992. Onto- genetic changes in foraging behaviour of the dogwhelk mellifera from Europe. These measures were correlated with temperature and elevation. In warmer Nucella lapillus. J. Exp. Mar. Biol. Ecol., 155, 199—212. coastal regions, the two subspecies readily hybridize and most samples showed evidence of considerable hybridization. In cooler mountain regions, there is much less hybridization, with the A. m. mellifera subspecies characteristics strongly predominating. There is no evidence for cyto- prey, Nucella lapillus. J. Anim. Ecol., 48, 65—78. nuclear incompatibilities between these subspecies or for dines caused by direct selection. We thank Matthew HutchingsJONES, J. s. and1982. Genetic Chris differences Hayes in individual behaviour for hypothesize that A. m. ligustica and F1 hybrids have lower fitness than A. m. mellifera in cooler assistance in the field in Cornwall,associated with shell and polymorphism James in the Mallet snail Cepaea regions, and that there may be assortative mating in cooler regions only. The significance of these results for the understanding of honey bee hybrid zones in Europe is discussed. Keywords: Apismellifera, chine, hybrid zone,malate dehydrogenase, mitochondrial DNA, JONES, J. S. AND PROBERT, R. F. 1980. Habitat selection maintains morphometrics. W. D. AND NEWBURY, S. E 1984. The adaptations of characters differentiating the two subspecies. Equally the rough winkle Littorina rudis, to dessication and to dis- Introduction fit alleles may diffuse across the hybrid zone. The Understanding how two adjoining subspecies maintain width of the dine will depend on how recently the Factors Affecting Variation in Natural their identity is of considerable interest because the contact was made and the rate of dispersal of individ- phenomenon may represent an early stage of specia- uals (Endler, 1977). BERRY, R. I. AND CROTHERS, J. H. 1974. Visible variation in the tion (Barton & Hewitt, 1985). Two situations have to If the fitness in F1 hybrids of the two subspecies dogwhelk, Nucella lapillus. VERMEIJ,J. Zoo!. G. J. 1978. Lond., Biogeography 174, and Adaptation: 123—148. Patterns of be considered: (i) when adjoining subspecies have differs from that of parents, a hybrid zone, or tension BOULDING, E. G. AND HAY, T. K. 1993. Quantitative genetics of evolved parapatrically, and (ii) when adjoining sub- zone, may become virtually permanent, impeding gene shell form of an intertidal snail: constraintsA Natural History on of short-term Shells. Princeton species have diverged allopatrically but have become flow between adjoining subspecies for many genera- reunited by secondary contact. The first case is the tions (Moore, 1977; Moore & Buchanan, 1985). The normal consequence of genetic drift and selection width of such a zone depends not only on the rate of resulting in genetically distinct populations. The second dispersal, but also on the relative fitness of F1s (Hewitt, case is much more interesting as there is the potential 1988). Following Bazykin (1969) the model of Barton for competition between two expanding subspecies & Hewitt (1985) requires hybrid unfitness (from the which are already genetically distinct. In some cases breakdown of coadapted gene complexes or abnormal- this may lead to the displacement of one subspecies by ities in chromosomal pairing in F1s) but no environ- the other, the merging of the two (Moore, 1977) or, if mental gradient for the maintenance of hybrid zones. speciation has proceeded far enough, a hybrid zone In contrast, Moore (1977) and Moore & Buchanan may develop characterized by coincident dines for the (1985) show that permanent hybrid zones can result from heterosis in F1s, coupled with an environmental *Correspondence gradient in which F1s are fitter than either parent within 315 316 B. P. OLDROYD ETAL. the zone of intermediate ecology between the native generally identifiable by malate dehydrogenase (Mdh) ranges of the parental subspecies. allele frequencies (Cornuet, 1983) and mt-DNA The western honey bee Apis mellifera is native RFLPs (Garnery, 1992). Studies which simultaneously to Europe, Africa and western Asia. This broad examine nuclear and cytoplasmic markers (cytonuclear geographical distribution has resulted in the evolution systems) provide the opportunity to obtain insight into of three distinct evolutionary branches: western Euro- the nature of the hybridization process (Arnold, 1993). pean (M), northern Mediterranean (C) and African (A). As climes correlated with environmental temperature These branches were initially discovered using have been reported for morphometric characters morphometrics (Ruttner et a!., 1978; Ruttner, 1988) (Ruttner, 1988; Daly et al., 1991) and Mdh allele and have been largely confirmed by mitochondrial frequencies (Nielson eta!., 1994) we looked for signifi- DNA studies (Smith, 1991; Garnery, 1992; Gamery et cant correlations between genetic markers and envi- a!., 1992). As defined by morphometrics, these ronmental temperature. By these means, we attempt to branches represent one to eight subspecies including determine if the observed dines are environmentally A. m. mellfera (branch M) and A. m. ligustica (branch induced or are the result of a contact zone. In the light C). In northern Italy, these two subspecies have formed of these analyses, the coexistence of the two subspecies a contact zone which has been well characterized by in Tasmania and in contact zones in Europe are dis- allozyme and morphometric studies (Comuet, 1983; cussed. Badino et a!., 1983, 1984, 1988a,b; Comparini & Biasiolo, 1991; Nazi, 1992; Meixner eta!., 1993). This contact zone provides an ideal system for studying a Materials and methods secondary contact between two formerly allopatric subspecies. Because of the high level of genetic diver- Field collections gence between mellifera and ligustica, there are Collections were made from individual colonies across morphological, mitochondrial and allozyme markers Tasmania (Fig. 1). Bees were caught outside colonies in available to discriminate between them (reviewed by sweep nets or by aspiration, or if the colony was in a Oldroyd et a!., 1992). Secondly, these two subspecies moveable frame hive, by scraping bees off honey have been exported to many places around the world combs. Bees were killed immediately by freezing in during the past century, providing opportunities to liquid nitrogen. For each colony, about half the bees study artificial contact zones in new environments. were preserved in liquid nitrogen and half in ethanol. Tasmania (Australia) is such a new environment. In Except for a collection from commercial colonies 1831, a first Apis mellifera colony was introduced from originating from Mole Creek, all bees were feral. Feral Great Britain, presumably of the subspecies mellifera. colonies were defined as those that met at least one of By 1860, honey bees were apparently feral and wide- the following