Variation was assessed among mitochondrial DNAs (mtDNAs) from geographically dispersed individuals of two species of white-eyes (Zosterops lutea and Z. lateralis) in Australia. The survey revealed high levels of intraspecific divergence. The mtDNA of Z. lutea was paraphyletic; eastern samples were more closely related to eastern Z. lateralis than to western Z. lutea. The mtDNA of Z. lateralis showed a major phylogenetic break within the distribution of one subspecies, Z. I. halmaturina, and was relatively uniform across four subspecies distributed along the east coast of Australia. The discordance between species boundaries and the distribution of mtDNA haplotypes is probably due to historical hybridization. The poor correspondence between subspecies boundaries and mtDNA discontinuities in Z. lateralis implies that patterns of morphological variation may reflect local processes more than evolutionary history. In a sample of the morphologically divergent island race, Z. I. chlorocephala, mtDNA haplotypes were identical to those found on the adjacent mainland. However, the frequencies of mtDNA variants differed considerably between the two places. These data suggest that the island race is recently derived from the mainland, but that current gene flow is rare. Received 1 July 1991, accepted 23 February 1992.

Department of Zoology, University of Queensland, Brisbane Q 4072, Australia STUDIES OF geographic variation in birds have contributed significantly to the development of evolutionary theory, particularly concerning speciation (Mayr 1963) and adaptation (James 1970). The evolutionary significance of geo- graphic variation depends upon the extent to which that variation has a genetic basis and reflects the joint effects of selection and history. In the past, this has been difficult to assess in birds because studies of genetic variation, re- lying upon the identification of polymorphic enzyme loci, have usually reported low levels of intraspecific variation (Barrowclough 1983, Zink and Remsen 1986). Recently, analysis of mtDNA has revealed a more sensitive source of genetic variation, for reasons that are now well documented (Avise 1986, Moritz et al. 1987), and is becoming popular in studies of geo- graphic genetic variation. To date, a small number of studies have as- sessed intraspecific variation of mtDNA in widespread passerines, with contrasting re- sults. A continentwide survey of the Red-winged Blackbird (Agelaius phoeniceus), which shows considerable geographic variation in morphol- ogy, revealed a large number of mtDNA hap-  Present address: Marine Science Institute, Uni- versity of California, Santa Barbara, California 93106, USA. lotypes characterized by low levels of sequence divergence and widespread geographic distri- bution (Ball et al. 1988). A similar widespread distribution of mtDNA clones was documented for the Song Sparrow (Melospiza melodia), a spe- cies that exhibits relatively little morphological variation in the regions surveyed (Zink 1991). In contrast, mtDNA variation in some species is characterized by distinct geographic struc- turing (e.g. Branta canadensis, Shields and Wil- son 1987, Van Wagner and Baker 1990; Amrno- dramus rnaritirnus, Avise and Nelson 1989; Pornatostornus temporalis, Edwards and Wilson 1990; Colaptes auratus, Moore et al. 1991; and Passerella iliaca, Zink 1991). In some of these cases (e.g. Ammodramus maritimus), the geo- graphic pattern of mtDNA variation is discor- dant with that of morphological variation. mtDNA also has provided diagnostic markers between avian sibling species, although genetic distance between congeners is usually small (summarized in Avise and Zink 1988). Phylo- genetic analysis of mtDNA and allozymes from congeneric sparrows (Ammodramus, Zink and Avise 1990) revealed affinities other than those predicted from morphological comparisons. It is apparent, therefore, that observed mor- phological variation, which has so often been used to interpret evolutionary relationships in birds, may not always reflect evolutionary his- tory. In this paper, we report on mtDNA vari- 4 "Z-.'-... 8 ;:.:::::::::::.-.':::::.':.:.:.:.:..-.:.:..-.:..-.:.-.-.- - - - &'' dd '.:: Zosterops lateral is  Zosterops lutea Fig. 1. Map of Australia showing distribution of Zosterops lutea (1, Z. l. balstoni; 2, Z. l. lutea) and Z. lateralis (3, Z. l. ramsayi; 4, Z. l. chlorocephala; 5, Z. l. familiaris; 6, Z. l. lateralis; 7, Z. l. halmaturina; 8, Z. l. gouldi). Stippled areas indicate subspecies boundaries where appropriate. Lowercase letters are positioned to indicate sampling localities, and each represents mtDNA clone type of a single individual. ation within two species of white-eyes (Zoster- opidae) that exist, for the most part, allopatrically in Australia--the Silvereye (Zosterops lateralis) and the Yellow White-eye (Z. lutea). The species and subspecies of the Indo-Australian Zosterops have been determined by plumage color and morphometrics (Mees 1969). It is now recog- nized that difficulties arise because of the im- mense color variation within some species and because of potentially convergent similarities between some species (Moreau 1957). Zosterops lateralis, the Silvereye, is a very mo- bile and wide-ranging species found through- out most of eastern, southern and southwestern Australia, where six subspecies are recognized (Fig. 1; Blakers et al. 1984). Winter flocks on the southern and central east coast comprise both local residents and migrants (Lane 1966) from more southerly (including Tasmania) popula- tions (Kikkawa 1968). Of special concern to the present work is Z. l. chlorocephala, a race that maintains high-density populations on small wooded cays of the southern Great Barrier Reef. Mees (1969) concluded that the race is of recent origin and, yet, it is morphologically distinct from the mainland race and is the most abun- dant land bird on some of the cays, such as Heron Island (Kikkawa 1970). Similarities in plumage color of island birds and those on the adjacent mainland coast suggest that chloroceph- ala was derived from a nearby mainland pop- ulation (Mees 1969). Zosterops lutea, the Yellow White-eye, is en- demic to the north coastal parts of Australia and arguably comprises two subspecies (Fig. 1; Mees 1961, Blakers et al. 1984). There is an outlying population on the east coast of the continent, within the range of the Silvereye (Lavery and Grimes 1974) and, in the west, the range of the Yellow White-eye overlaps narrowly with that of the Silvereye. A third species, Z. citrinella, occurs only on wooded islands off the far north- eastern tip of the continent (Blakers et al. 1984) and was not considered in this study. The work presented here is part of a broader study of genetic variation in the Heron Island population of Z. I. chlorocephala, which is also the subject of a detailed study in evolutionary ecology (e.g. Kikkawa et al. 1986, Catterall et al. 1989). Here we focus on macrogeographic patterns of variation in the Silvereye, using samples of the Yellow White-eye as outgroups. The specific questions addressed are: (1) is the mtDNA variation geographically structured; (2) are the evolutionary lineages revealed by mtDNA analysis concordant with species and subspecies boundaries; and (3) does mtDNA provide information on the origin and diver- gence of the Heron Island race? MATERIALS AND METHODS For restriction-enzyme site mapping of mtDNA, Z. lateralis specimens from Brisbane (n = 2) and Heron Island (n = 1) were collected during the breeding season of 1989. Dissected heart tissues were either snap frozen and stored at -80øC (Heron Island spec- imen) or processed immediately (Brisbane speci- mens). For macrogeographic analysis, additional sam- ples of heart from Z. lateralis (n = 21) and Z. lutea (n = 6) from several localities around Australia (Fig. 1) were kindly donated by R. Schodde (CSIRO, Can- berra) and L. Christidis (Museum of Victoria). Sub- specific identifications of all specimens were based on morphology and collection location. Voucher specimens are listed in the Appendix. The mtDNA was purified by ultracentrifugation in a cesium chloride gradient, followed by dialysis, as described in Dowling et al. (1990). For restriction- enzyme site mapping, the Brisbane and Heron Island mtDNAs were digested to completion with 16 five- or six-base-pair-recognizing restriction endonucleases (listed in Fig. 2) in both single- and double-digestion combinations. After digestion, mtDNA fragments were end-labelled with 32p-nucleotides and their sizes de- termined by electrophoresis through agarose and polyacrylamide gels, followed by autoradiography. Fragment sizes were calculated against AvaI/BglII- digested lambda-phage DNA as a size standard. The positions of all restriction sites for 15 of the enzymes were mapped relative to each other by dou- ble digestion (Dowling et al. 1990). Sites for SpeI could not be mapped because of the presence of several small fragments that had no internal sites for any of the other mapped enzymes. The construction of a restriction-enzyme site map tests the interpretation of fragment changes as site gains or losses, and al- lowed for an alignment of the Silvereye mtDNA mol- ecule with that of previously characterized and se- quenced mtDNAs. It also allowed for an assessment of the distribution of restriction sites, particularly polymorphic sites, within the mitochondrial genome. If the sites were strongly clustered, estimates of se- quence divergence may be biased because of variation in evolutionary rates among genes (Brown 1985). Sequence divergences between mtDNAs were es- timated from comparisons of restriction-enzyme sites using the methods of Nei and Tajima (1983) in a pro- gram described in Nei et al. (1985). This matrix of divergences was used to construct a dendrogram by UPGMA clustering (Sneath and Sokal 1973). Restric- tion-enzyme site character data (presence/absence) were analyzed using both Dollo-parsimony and Wag- ner-parsimony criteria (DeBry and Slade 1985, Swof- ford and Olsen 1990) using PAUP (version 3.0, sup- plied by D. L. Swofford, Illinois Natural History Survey, Champaign). The stability of the resultant parsimony cladograms was evaluated by bootstrap- ping (Felsenstein 1985) using PAUP. The mtDNA variation among an additional 15 in- dividuals from Heron Island and an additional 15 individuals from the adjacent mainland (Brisbane) was nondestructively assayed for three endonucleases (EcoRV, AvaI and HindIII) that in combination diag- nose the known east-coast variants of Z. lateralis mtDNA. This was done by Southern hybridization analysis using purified 32p-labelled Silvereye mtDNA as a probe against total genomic DNA extracted from a small volume of blood. Hybridization procedures were based on those described by Church and Gilbert (1984) and were performed at 65øC with a final high- stringency wash in 0.2 x SSC at 65øC. Composite re- sults from these three enzymes were used to designate mtDNA haplotypes to the 15 Heron Island and the 15 Brisbane individuals. The two populations were treated as demes and the amount of among-deme ge- netic variation (Gsr) was estimated as described in Takahata and Palumbi (1985), using a program pro- vided by Stephen Palumbi (Univ. Hawaii, Honolulu). The observed Gsr was tested against a null hypothesis of random genetic variation across demes, using the bootstrap procedure described by Palumbi and Wil- son (1990). RESULTS A restriction-enzyme site map for Silvereye mtDNA is presented in Figure 2. The total size of the mtDNA molecule of this species is esti- mated to be approximately 17.5 kilobases, in the upper-middle of the size range (16.2 to 18.1 kb) so far reported for birds (Shields and Helm- Bychowski 1988). There were no cases of mtDNA length variation or heteroplasmy. The map for the Heron Island individual was identical to that of one of the Brisbane individuals (clone h; see below). Three restriction-enzyme sites that appear to be widely conserved across the animal kingdom (see Cart et al. 1987) are in- dicated on the map. These sites were used to align and orientate the restriction-enzyme site map of the Silvereye to the mtDNA sequence of the chicken (Gallus gallus; Desjardins and Mo- rais 1990), allowing for the identification of par- ticular regions such as the rRNA genes and the control region. The sites assayed in Z. lateralis appear to be fairly evenly distributed along the length of the molecule, although the subset of polymorphic sites tend to be less uniformly dis- tributed (Fig. 2). No polymorphic sites were clone rRNA genes 125 16S Cont rol rgon D D HH A L  I I I t I I I I I PP E b D HH HA L ill I I PP 14 E D D H HH A L C I I I I I I I I I I PP E d D D HH A L I , l I I I I I I PP H E D X H A R B A L I i i i I A pp P E D X H A R B A L f I I I I I 1 I I I I I I I A PP P H E D X HH A R B A L g I iiI I 1 I I I I I I I A PP p e D X H A R A L I I I I I A PP P E E HH R A E H R A k I I I I Fig. 2. The mtDNA restriction-enzyme site maps for Zosterops lateralis (clones a-h) and Z. lutea (clones i- k). Sites enclosed by boxes represent sites conserved across animal kingdom and used to align map to chicken mitochondrial genome (top of figure). Other sites that were monomorphic among all Zosterops sampled are circled. Enzymes designated as follows: A = AvaI; B = BamHI; D = DraI; E = EcoRI; R = EcoRV; H = HindIII; C = HpaI; M = MluI; P = PvuII; S = SacII; L =SalI; X = XbaI; g = BglII; t = BstEII; n = NcoI. SpeI fragments were not mapped. Enzymes designated by lowercase letters not used to screen individuals in phylogeographic analysis. 1 2 3 4 5 6 M 7 8 9 10 11 12 13 14 15 --4.3 Fig. 3. Autoradiograph of a 1.0% agarose gel following electrophoretic separation of fragments of Zosterops mtDNA digested with SpeI_ Fragment sizes (kb) of size standard (AvaI/BglII digested lambda phage DNA, lane M) shown at right. Three distinct mtDNA clones can be seen: Z. lutea (Western Australia), lanes 1-3; Z. lateralis (Tasmania and Victoria), lanes 4-6; and Z. lateralis (South Australia and Western Australia), lanes 7- 15. mapped within the ribosomal genes; two (for AvaI and SaII) were mapped within the control region. Among the 30 individuals analyzed for 13 restriction enzymes, 62 different restriction-en- zyme sites were observed, with an average of 39 sites (representing 234 nucleotides or 1.3% of the mtDNA genome) scored per individual. Of the 62 sites observed, 28 were polymorphic; 16 varied among the 24 individuals of Z. later- alis, while 22 varied among the six individuals of Z. lutea. Of the 13 enzymes used to screen all 30 individuals, 10 enzymes produced two or more fragment profiles, as illustrated by SpeI (Fig. 3). Fragment sizes for these polymorphic enzymes are available from the senior author upon request. The sizes of fragments (in base pairs) produced by the remaining three (mono- morphic) enzymes are as follows: HpaI (12,000, 3,500, and 1,700), MluI (15,400 and 1,800), and SacIl (12,300, 3,200, and 1,700). Sizes of frag- ments greater than about 8,000 bp are approx- imate only. Eleven different mtDNA clones were repre- sented among the 30 individuals (Table 1). Within Z. lateralis, eight different clones were observed, the most common (clone e) being rep- resented by eight individuals spanning eastern Australia (subspecies ramsayi, familiaris, halma- turina and lateralis; Fig. 1). The next most com- mon clones, d and a, were localized to southern (halmaturina, n = 5) and to southwestern (gouldi, n = 4) Australia, respectively. Clone h, found only in subspecies familiaris (n = 1) on the main- land, was also identified on Heron Island (chlo- rocephala, n = 1). The remaining four clones were each represented by a single individual from various sites around the southern part of TABLE 1. Restriction-endonuclease descriptions and distributions of mtDNA haplotypes among white-eyes Zosterops lateralis and Z. lutea. Uppercase letters in descriptions, from left to right, correspond to restriction- fragment profiles produced by digestion with endonucleases AvaI, BamHI, DraI, EcoRI, EcoRV, HindIII, HpaI, MluI, PvuII, SacII, SaII, SpeI, and XbaI. Restriction profiles designated by adjacent letters of alphabet differ by only a single restriction-enzyme site; nonadjacent letters denote a difference of two or more restriction-enzyme sites (following Avise and Nelson 1989). By combining data from all enzymes, each individual was assigned a composite mtDNA haplotype (designated by a lowercase letter). mtDNA Subspecies No. haplotype mtDNA description designation  birds a CADAABAACAAAA gouldi 4 b CADAAFAACAAAA gouldi 1 c CADAAHAACAAAA gouldi 1 d CADAAEAACAAAA halmaturina 5 e EBCABCAADAAEB ramsayi, familiaris, 9 halmaturina, lateralis f EBCABDAADAAEB lateralis 1 g EBCABBAADAAEB familiaris 1 h CBCABCAADAAEB familiaris, chlorocephala 2 i EACABCAADAAFB lutea 3 j AAACBBAAAABCA balstoni 1 k AAACBAAAAABCA balstoni 2 Total 30 Subspecies represented by haplotypes a-h belong to Z. lateralis, and i-k to Z. lutea. the continent. Within Z. lutea, three clones were recognized: clone i in the east (lutea, n = 3), and clones j (balstoni, n = 1) and k (balstoni, n = 2) in the west (Fig. 1). Clearly, the distribution of mtDNA haplotypes in these two species bears a strong relationship to geography. Estimates of percent sequence divergence be- tween clones ranged from 0.17% to 4.91% (Table 2). The mean percent sequence divergence among the 11 clones was 2.24%. The mean among Z. lateralis clones was 1.48% and among those of Z. lutea was 3.12%; for both species, this di- vergence was largely explained by the differ- ences between eastern and western mtDNA clones. The UPGMA dendrogram of these se- quence-divergence data clearly separates east- ern and western clones of Z. lateralis, at a level of approximately 2.3% divergence (Fig. 4). The east-west dichotomy within Z. lateralis occurred well within the range of Z. l. halmaturina rather than between Z. l. halmaturina and Z. l. gouldi as might be expected. Even more striking is the divergence of approximately 3.9% between eastern and western clones of Z. lutea, and also between western Z. lutea and all clones of Z. lateralis (Fig. 4). For phylogenetic analyses, using the branch- and-bound algorithm of PAUP, presence or ab- TABLE 2. Estimates of percent sequence divergence (lower left) between mtDNA types of Zosterops lateralis and Z. lutea. Estimates calculated from numbers of shared restriction-enzyme sites, presented in upper right. Number of restriction-enzyme sites used in each comparison on diagonal. mtDNA mtDNA type type a b c d e f g h i j k a 46 46 46 46 42 42 43 42 42 38 37 b 0.35 48 46 46 42 42 43 42 42 38 37 c 0.18 0.53 47 46 42 42 43 42 42 38 37 d 0.18 0.53 0.36 47 42 42 43 42 42 38 37 e 2.23 2.57 2.40 2.40 50 50 50 47 49 36 35 f 2.40 2.74 2.57 2.57 0.17 51 50 48 49 36 35 g 2.01 2.35 2.18 2.18 0.17 0.33 51 48 49 37 36 h 1.97 2.23 2.05 2.05 0.69 0.51 0.51 48 47 36 35 i 2.23 2.57 2.40 2.40 0.34 0.50 0.50 0.69 50 36 35 j 2.82 3.18 3.00 3.00 4.44 4.62 4.16 4.09 4.44 44 43 k 3.08 3.45 3.26 3.26 4.74 4.91 4.44 4.37 4.74 0.19 43 mtDNA clone b c distribution h 4.0 $.0 Z.O 1,0 0.8 0.6 0.4 Percent sequence divergence 0.2 0.0 El Fig. 4. UPGMA dendrogram based on sequence-divergence estimates. Distributions of Z. lateralis mtDNA clones are solid black, while those for Z. lutea are gray. sence of restriction sites were treated as char- acter states. Initial analyses, using all three clones of Z. lutea as outgroups, precluded monophyly of the ingroup because of the unexpected affin- ity of eastern Z. lutea clones with eastern Z. lateralis clones. These analyses were rejected and, subsequently, only the two western clones of Z. lutea (clones j and k) were designated as out- groups. The consensus Dollo-parsimony clado- gram, generated by 100 bootstrap replications with a branch-and-bound search, revealed that the eastern Australian samples of Z. lateralis (clones e-h) and Z. lutea (clone i) form a strongly supported clade (Fig. 5). The western samples of Z. lateralis (clones a-d) that formed a tight group in the UPGMA analysis (Fig. 4) are not united by any derived character states (cf. Fig. 2) and, therefore, are unresolved in the phy- logenetic analysis. The mtDNA clone (clone h) found in the is- land race (Z. I. chlorocephala) and in one sample from Brisbane (Z. I. familiaris) was the sister group to other east-coast clones in the phylogenetic analysis (Fig. 5) and was also the most distinc- tive of these (Fig. 4). To further investigate the distribution of mtDNA variants between the Heron Island population and those on the mainland (specifically Brisbane), larger samples (each of n = 15) were assayed with AvaI, HindlII, and EcoRV by Southern hybridization. The composite haplotypes (Table 3) showed a major difference in frequencies of mtDNA clones be- tween the two locations. In nine of the samples from Brisbane, digestion with EcoRV revealed a new fragment pattern, differing from others by a site gain. The sequence divergence be- tween the two populations, corrected for within population variation, is 1.5%. The correspond- ing values within populations were 0.18% for Heron Island and 0.48% for Brisbane. The difference in frequencies of mtDNA clones between the two populations is highly significant (heterogeneity X 2 = 22.8, P < 0.005). The Gsr value, obtained by treating the two pop- ulations as demes, is 0.30. The maximum value obtained in 100 randomizations was only 0.08, indicating that the value of 0.30 represents sig- nificant among-population differentiation (cf. Palumbi and Wilson 1990). DISCUSSION Phylogeography of rntDNA variation in Z. later- alis.--The clearest result within Z. lateralis is the separation of an east-coast group from other samples, suggesting a long-term separation of eastern and western populations. The sequence divergence of 2.3% between these two phylo- geographic groups is among the highest values reported within a species of bird and is well 1 oo I 441 221 71-'? 1 O0  j I k Fig. 5. Relationships between mtDNA clones as described by a Dollo 50%-majorit'y-rule consensus tree, generated by 100 bootstrap replications of branch- and-bound search by PAUP program. Tree length was 38 steps, and consistency index was 0.737. Percentages of bootstrap replications that supported each node are given. within the range of previously reported inter- specific divergences (e.g. Avise and Zink 1988, Shields and Helm-Bychowski 1988, Avise and Nelson 1989, Tegelstrom et al. 1990, Van Wag- ner and Baker 1990, Zink and Dittmann 1991, Zink et al. 1991). In comparison, a continent- wide survey of mtDNA diversity in the Red- winged Blackbird revealed very little diversity or geographic structuring, with a maximum ob- served sequence divergence of only 0.8% (Ball et al. 1988). Even in the Seaside Sparrow, for which two geographically defined clusters of mtDNA clones were observed, maximum se- quence divergence was estimated to be only 1.0% (Avise and Nelson 1989). The geographic location of the break between the two major assemblages of mtDNA from Z. lateralis occurs to the east of the "Eyrean" and "Nullarbor" biogeographical barriers, discon- tinuities defined by species-level comparisons in other Australian passerines (Keast 1961, Kik- kawa and Pearse 1969, Cracraft 1986, Ford 1987). However, the mtDNA break does correspond to the "Mallee" barrier, recognized as possibly the most severe southern barrier for semiarid bird species during the peak of the last glacia- tion, and as a region that now has numerous TABLE 3. Occurrence of composite mtDNA haplo- types among Brisbane (n = 15) and Heron Island (n = 15) populations of Zosterops lateralis. Uppercase letters in descriptions, from left to right, corre- spond to restriction-fragment profiles produced by digestion with endonucleases EcoRV, AvaI, and HindIII. Letter designations for each endonuclease follow those in Table 1, except that haplotype C for EcoRV represents a new haplotype not observed previously. Heron Island Brisbane mtDNA type n Percent n Percent f BCC 14 93 1 7 BEC 1 7 5 33 h CEC 0 0 9 60 contact zones (Ford 1987). It would be of inter- est to survey mtDNA from other widespread species to evaluate the congruence between in- traspecific and interspecific biogeographic breaks. Comparison with patterns of morphological vari- ation.--The relationships and evolutionary his- tory implied by the mtDNA analysis are incon- gruent with species and subspecies boundaries. Most strikingly, mtDNA from Z. lutea is para- phyletic with respect to Z. lateralis, since the eastern mtDNA clones of Z. lutea are more close- ly related to those of eastern Z. lateralis than they are to western clones of Z. lutea. On mor- phological grounds, M ees (1961 ) treated Z. lu tea as an independent species, rather than as part of a wide-ranging superspecies, and suggested that it probably evolved in northwestern Aus- tralia, derived from rain forest inhabitants of the Lesser Sundas, Indonesia. If so, the species would have extended its range towards the eastern portion of the continent, where populations may now be isolated from the west by the "Carpen- tarJan" geomorphological barrier commonly recognized by zoogeographers (Macdonald 1969, Ford 1987; barrier "B" in fig. 9 of Cracraft 1986). The presence of the Carpentarian barrier was also reflected in the geographic pattern of length and site variation in mtDNA of Grey- crowned Babblers (Pomatostomus temporalis; Ed- wards and Wilson 1990). Previous studies reporting discordance be- tween mtDNA haplotypes and species classifi- cations have attributed this to either the sorting of ancestral polymorphisms (Neigel and Avise 1985; e.g. Avise et al. 1990), or to introgressive hybridization (Ferris et al. 1983, Tegelstrom 1986). The former hypothesis seems an unlikely explanation in the present case because of the large sequence divergences involved. Intro- gressive hybridization seems more plausible; even a low level of hybridization may be suf- ficient to establish a neutral mtDNA clone with- in a foreign population (Takahata and Slatkin 1984). On the east coast, Z. lutea has established a population within the range of Z. lateralis (Fig. 1; Lavery and Grimes 1974), providing the op- portunity for hybridization between the two species. However, the two species are now eco- logically and morphologically distinct, and no hybrids have been reported. It would be of in- terest to sample individuals of Z. lutea from cur- rent areas of overlap (e.g. the west and north- east coasts; Fig. 1) to determine whether the presumed hybridization was an isolated inci- dent. If hybridization was recent or is continu- ing, it should be evident in nuclear DNA poly- morphisms (Degnan in prep.). The distribution of mtDNA variation is also discordant with the currently recognized sub- species boundaries (Fig. 1). The existence of an mtDNA clone (clone e) common to the entire east coast and extending across the range of four subspecies is consistent with high levels of his- torical gene flow. This gene flow probably con- tinues, since migrating individuals from the southern part of the continent are known to mix with resident individuals further north, at least during the nonbreeding season (Lane 1966, Kikkawa 1968). It seems plausible that not all individuals complete the return journey before settling to breed, particularly in view of the occasional record of movement of banded birds in a continued northerly direction at the end of the migratory season (Lane 1966). A partic- ular anomaly exists in the southern part of mainland Australia, where the recognized boundary of the two subspecies, gouldi in the west and halmaturina in the east (Mees 1969), lies much further to the west of the continent than does the mtDNA discontinuity. The race gouldi is conspicuous in lacking the grey back that is characteristic of all other silvereye races, and Mees (1969) considered it a well-marked subspecies that "has clearly lived in isolation for a long time." However, both Mees (1969) and Ford (1987) recognized that, currently, a narrow zone of intergradation exists between these two southern subspecies. Lack of correlation between the distributions of Z. lateralis subspecies and the patterns of mtDNA variation indicate that morphology alone may not provide a clear picture of the evolutionary history of this species. Not only are historical barriers and morphological boundaries discordant, but also the magnitude of mtDNA divergence and morphological dif- ference are unlinked. The greatest difference in mtDNA occurs between individuals of Z. I. goul- di (and western Z. I. halmaturina) and eastern Z. 1. halmaturina, subspecies that differ in plumage color and only slightly in morphometrics (Mees 1961). Conversely, the mtDNA difference be- tween the much larger island race, Z. I. chloro- cephala, and mainland birds is relatively small. External morphology, so often used to define taxonomic groups and to make evolutionary in- ferences in birds, can show substantial variation in the context of very little or no genetic dif- ferentiation. A similar situation was reported for the Seaside Sparrow (Avise and Nelson 1989), for which mtDNA data suggested that the sub- specific taxonomy did not adequately reflect the evolutionary genetic relationships of the pop- ulations sampled. As suggested by Zink and Avise (1990) in reference to similar observations in sparrows (Ammodramus), morphological vari- ation, to the extent it is heritable, may be more a reflection of local selection pressures than evolutionary history. However, remember that mtDNA represents only a single molecular lineage. In Z. lateralis, mtDNA phylogeographies are consistent with type I of Avise et al. (1987), although these au- thors proposed that type IV may prove to be the most common in birds that are highly mo- bile. If this observed spatial separation of mtDNA clones, with distinct phylogenetic dis- continuities, is due to long-term, extrinsic bar- riers to gene flow, then analysis of RFLPs for nuclear DNA (Degnan in prep.) should reveal congruent patterns of geographic variation. Origin and divergence of Heron Island popula- tion.--The mtDNA of the Heron Island popu- lation (subspecies chlorocephala) is closely relat- ed to that of the adjacent mainland populations, including Z. I. familiaris from Brisbane. Identical mtDNA haplotypes were found in the two pop- ulations, strongly suggesting that the Heron Is- land population may have derived from the nearby eastern Australian mainland (although it is also possible that the immediate ancestors were from other island races to the east; J. Kik- kawa pers. comm.). However, the significant frequency differences of composite haplotypes derived from three restriction-enzyme frag- merit patterns suggest that the level of gene flow from the mainland to the island is pres- ently very low. This is consistent with the ob- servation that occasional winter immigrants of the mainland race disappear by the onset of the breeding season on Heron Island, so that no interbreeding with island birds is known (Kik- kawa 1970). The low mtDNA differentiation between the two populations is of particular interest in view of the substantial morphological differences that exist between them. A low sequence divergence indicates a short separation time so that mor- phological differentiation must have occurred rapidly, presumably as the result of strong se- lection. In addition, the continued existence of at least two founding mtDNA lineages in the island population also suggests a relatively short separation time (or occasional gene flow), since the random extinction of mtDNA lineages can be very rapid under the kind of demographic conditions that may prevail in an island pop- ulation (Avise et al. 1984). ACKNOWLEDGMENTS We greatly appreciate the help of Les Christidis and Richard Schodde in providing samples. We thank Da- vid Swofford and Steven Palumbi for their PAUP and GsT software, respectively. Leo Joseph, Jiro Kikkawa and Scott Edwards provided many helpful comments on the manuscript. Sandie Degnan was supported by a University of Queensland Biological Sciences Post- graduate Research Award. This research was sup- ported by grants from the Australian Research Coun- cil to Jiro Kikkawa and to Craig Moritz. LITERATURE CITED AVISE, J. C. 1986. Mitiochondrial DNA and the evo- lutionary genetics of higher animals. Phil. Trans. R. Soc. Lond. B312:325-342. AVISE, J. C., C. D. ANKNEY, AND W. S. NELSON. 1990. Mitochondrial gene trees and the evolutionary relationship of Mallards and Black Ducks. Evo- lution 44:1109-1119. AVISE, J. C., J. ARNOLD, R. M. BALL, E. BIRMINGHAM, T. LAMB, J. E. NEIGEL, C. A. REEB, AND N. C. SAUNDERS. 1987. Intraspecificphylogeography: The mitochondrial DNA bridge between popu- lation genetics and systematics. Annu. Rev. Ecol. Syst. 18:489-522. AVISE, J. C., J. E. NEIGEL, AND J. ARNOLD. 1984. De- mographic influences on mitochondrial DNA lineage survivorship in animal populations. J. Mol. Evol. 20:99-105. AVISE, J. C., AND W. S. NELSON. 1989. Molecular ge- netic relationships of the extinct Dusky Seaside Sparrow. Science (Washington D.C.) 243:646-648. AVISE, J. C., AND R. M. ZINK. 1988. Molecular genetic divergence between avian sibling species: King and Clapper rails, Long-billed and Short-billed dowitchers, Boat-tailed and Great-tailed grackles, and Tufted and Black-crested titmice. Auk 105: 516-528. BALL, R. M., JR., S. FREEMAN, F. C. JAMES, E. BIRMING- HAM, AND J. C. AVISE. 1988. Phylogeographic population structure of Red-winged Blackbirds assessed by mitochondrial DNA. Proc. Natl. Acad. Sci. USA 85:1558-1562. BARROWCLOUGH, G.F. 1983. Biochemical studies of microevolutionary processes. Pages 223-261 in Perspectives in ornithology (A. FI. Brush and G. A. Clark, Jr., Eds.). Cambridge Univ. Press, New York. BLAKERS, M., S. J. J. F. DAVIES, AND P. N. REILLY. 1984. The atlas of Australian birds (RAOU). Melbourne Univ. Press, Carlton. BROWN, W.M. 1985. The mitochondrial genome of animals. Pages 95-130 in Molecular evolutionary genetics (R. J. MacIntyre, Ed.). Plenum, New York. CARR, S. M., A. J. BROTHERS, AND A. C. WILSON. 1987. Evolutionary inferences from restriction maps of mitochondrial DNA from nine taxa of Xenopus frogs. Evolution 41:176-188. CATTERALL, C. P., J. KIKKAWA, AND C. GRAY. 1989. Inter-related age-dependent patterns of ecology and behaviour in a population of Silvereyes (Aves: Zosteropidae). J. Anim. Ecol. 58:557-570. CHURCH, G. M., AND W. GILBERT. 1984. Genomic se- quencing. Proc. Natl. Acad. Sci. USA 81:1991- 1995. CRACRAFT, J. 1986. Origin and evolution of conti- nental biotas: Speciation and historical congru- ence within the Australian avifauna. Evolution 40:977-996. DEBRY, R. W., AND N. A. SLADE. 1985. Cladistic anal- ysis of restriction endonuclease cleavage maps within a maximum-likelihood framework. Syst. Zool. 34:21-34. DESJARDINS, P., AND R. MORAIS. 1990. Sequence and gene organization of the chicken mitochondrial genome. A novel gene order in higher verte- brates. J. Mol. Biol. 212:599-634. DOWLING, T. E., C. MORITZ, AND J. D. PALMER. 1990. Nucleic acids. II: Restriction site analysis. Pages 250-317 in Molecular systematics (D. M. Hillis and C. Moritz, Eds.). Sinauer, Sunderland, Mas- sachusetts. EDWARDS, S. V., AND A. C. WILSON. 1990. Phyloge- netically informative length polymorphism and sequence variability in mitochondrial DNA of Australian songbirds (Pomatostomus). Genetics 126: 695-711. FELSENSTEIN, J. 1985. Confidence limits on phylog- enies: An approach using the bootstrap. Evolu- tion 39:783-792. FERRIS, S. D., R. D. SAGE, C.-M. HUANG, J. T. NIELSEN, U. Rirr, aND A. C. WILSON. 1983. Flow of mi- tochondrial DNA across a species boundary. Proc. Natl. Acad. Sci. USA 80:2290-2294. FORD, J. 1987. Hybrid zones in Australian birds. Emu 87:158-178. JAMES, F.C. 1970. Geographic size variation in birds and its relationship to climate. Ecology 51:365- 390. KEAST, A. 1961. Bird speciation on the Australian continent. Bull. Mus. Comp. Zool., Harvard 123: 305-495. KW, KWA, J. 1968. Social hierarchy in winter flocks of the Grey-breasted Silvereye Zosterops lateralis (Latham). Jap. J. Ecol. 18:235-246. KIKKWA, J. 1970. Birds recorded at Heron Island. Sunbird 1:34-47. KIKKWA, J., aND K. PRSE. 1969. Geographical dis- tribution of land birds in Australia--A numerical analysis. Aust. J. Zool. 17:821-840. KIKKAWA, l., I. N.M. SMITH, R. PRYS-JONES, P. FISK, AND C. CATTERALL. 1986. Determinants of social dominance and inheritance of agonistic behavior in an island population of Silvereyes, Zosterops lateralis. Behav. Ecol. Sociobiol. 19:165-169. LANE, S.G. 1966. The cooperative Silvereye project. Aust. Bird Bander 4:73-74. LAVERY, H. J., AND R. J. GRIMES. 1974. The Yellow Silvereye in north-east Queensland. Sunbird 5:42- 43. MACDONALD, J.D. 1969. Notes on the taxonomy of Neositta. Emu 69:169-174. MAYR, E. 1963. Systematics and the origin of species. Columbia Univ. Press, New York. MEES, G. F. 1961. A systematic review of the Indo- Australian Zosteropidae (Part II). Zool. Verh. (Leiden) 50:1-168. MEES, G. F. 1969. A systematic review of the Indo- Australian Zosteropidae (Part III). Zool. Verh. (Leiden) 102:1-390. MOORE, W. S., J. H. GRAHAM, AND J. T. PRICE. 1991. Mitochondrial DNA variation in the Northern Flicker (Colaptes auratus, Aves). MoL Biol. EvoL 8:327-344. MOREAU, R. E. 1957. Variation in the western Zos- teropidae. Bull. Brit. Mus. Nat. Hist., Zool. 4:311- 433. MORITZ, C., T. E. DOWLING, AND W. M. BROWN. 1987. Evolution of animal mitochondrial DNA: Rele- vance for population biology and systematics. Annu. Rev. Ecol. Syst. 18:269-292. NEI, M., J. C. STEPHENS, AND N. SAITOU. 1985. Meth- ods for computing the standard errors of branch- ing points in an evolutionary tree and their ap- plication to molecular data from humans and apes. MoL Biol. EvoL 2:66-85. NEI, M., AND F. TAJIMA. 1983. Maximum likelihood estimation of the number of nucleotide substi- tutions from restriction site data. Genetics 105: 207-217. NmGEL, J.E., aND J. C. AVISE. 1985. Phylogenetic relationships of mitochondrial DNA under var- ious demographic models of speciation. Pages 515-534 in Evolutionary processes and theory (E. Nevo and S. Karlin, Eds.). Academic Press, New York. PALUMBI, S. R., AND A. C. WILSON. 1990. Mitochon- drial DNA diversity in the sea urchins Strongy- locentrotus purpuratus and S. droebachiensis. Evo- lution 44:403-415. SHIELDS, G. F., AND K. M. HELM-BYCHOWSKI. 1988. Mitochondrial DNA of birds. Curt. Ornithol. 4:273-295. SHIELDS, G. F., AND A. C. WILSON. 1987. Subspecies of the Canada Goose (Branta canadensis) have dis- tinct mitochondrial DNAs. Evolution 41:662-666. SNTH, P. H. A., AND R. R. SOKL. 1973. Numerical taxonomy. Freeman, San Francisco. SWOFFORD, D. L., AND G. J. OLSEN. 1990. Phylogeny reconstruction. Pages 411-501 in Molecular sys- tematics (D. M. Hillis and C. Moritz, Eds.). Sin- auer, Sunderland, Massachusetts. TAKAHATA, N., AND S. R. PALUMBL 1985. Extranu- clear differentiation and gene flow in the finite island model. Genetics 109:441-457. TAKAHATA, N.,ANDM. SLATKIN. 1984. Mitochondrial gene flow. Proc. Natl. Acad. Sci. USA 81:1764- 1767. TEGELSTROM, H. 1986. Transfer of mitochondrial DNA from the northern red-backed vole (Clethrionomys rutilus) to the bank vole (C. glareolus). J. Mol. Evol. 23:218-227. TEGELSTROM, H., H. P. GELTER, aND M. JAAROLA. 1990. Variation in Pied Flycatcher (Ficedula hypoleuca) mitochondrial DNA. Auk 107:730-736. VAN WAGNER, C. E., AND A. J. BAKER. 1990. Associ- ation between mitochondrial DNA and morpho- logical evolution in Canada Geese. J. Mol. Evol. 31:373-382. ZINK, R. M. 1991. The geography of mitochondrial DNA variation in two sympatric sparrows. Evo- lution 45:329-339. ZINK, R. M., AND J. C. AVISE. 1990. Patterns of mi- tochondrial DNA and allozyme evolution in the avian genus Ammodramus. Syst. Zool. 39:148-161. ZINK, R. M., AND D. L. DITTMANN. 1991. Evolution of Brown Towhees: Mitochondrial DNA evi- dence. Condor 93:98-105. ZINK, R. M., D. L. DITTMANN, AND W. L. ROOTES. 1991. Mitochondrial DNA variation and the phylogeny of Zonotrichia. Auk 108:578-584. ZINK, R. M., AND J. V. REMSEN, JR. 1986. Evolutionary processes and patterns of geographic variation in birds. Curt. OrnithoL 4:1-69. APPENDIX. All specimens used in analysis, Registered specimens housed as voucher skins in the CSIRO collection, Canberra, Australia, except for B17363 which is in Museum of Victoria, Australia. Unregistered specimens designated by MV series field numbers also housed in Museum of Victoria. For remaining three specimens, skins were not retained. Field Collection mtDNA number Species Subspecies locality Museum no. clone C926 lateralis ramsayi Helenvale, QLD 39879 e F105 lateralis ramsayi Clarke Range, QLD 41408 e F267 lateralis ramsayi Atherton, QLD 41564 e H293 lateralis chlorocephala Heron Island, QLD Unregistered h C401 lateralis familiaris Agnes Water, QLD 39353 e C452 lateralis familiaris Kroombit Tops, QLD 39404 e B37 lateralis familiaris Brisbane, QLD Unregistered h B38 lateralis familiaris Brisbane, QLD Unregistered e S3078 lateralis familiaris Gunghalin, ACT 38457 g S3079 lateralis familiaris Gunghalin, ACT 38458 e MV068 lateralis familaris Gramplans, VIC B17363 e B724 lateralis lateralis Upper Blessington, TAS 38901 e B725 lateralis lateralis Upper Blessington, TAS 38902 f D357 lateralis halmaturina Sinclairs Gap, SA 40332 d D381 lateralis halmaturina Sinclairs Gap, SA 40356 d D388 lateralis halmaturina Port Pirie, SA 40363 d 42487 lateralis halmaturina Kangaroo Island, SA 42487 d 42568 lateralis halmaturina Kangaroo Island, SA 42568 d MV191 lateralis gouldi Esperance, WA Unregistered a MV192 lateralis gouldi Esperance, WA Unregistered b MV240 lateralis gouldi Albany, WA Unregistered a MV241 lateralis gould Albany, WA Unregistered a MV265 lateralis gouldi Pemberton, WA Unregistered a MV305 lateralis gouldi Hoffmans Hill, WA Unregistered c F353 lutea lutea Normanton, QLD 41649 F354 lutea lutea Normanton, QLD 41650 i F362 lutea lutea Kurumba, QLD 41658 D058 lutea balstoni Point Torment, WA 39181 j D059 lutea balstoni Point Torment, WA 39182 k D060 lutea balstoni Point Torment, WA 39183 k