Printer Friendly

Patterns of mtDNA genetic divergence.

Abstract. - We surveyed mitochondrial DNA haplotype divergence within and between populations of six species of North American chickadees (Parus, Subgenus Poecile) with the following results. (1) Genotype diversities (range 0.3 to 0.7) and low nucleotide diversities (range 3 to 27 x [10.sup.-4]) within populations were typical of known vertebrates. (2) The two widespread, northern species (atricapillus and hudsonicus) exhibit little mtDNA genetic differentiation throughout their previously glaciated continental distributions, most likely because of recent, postglacial range expansions. (3) Newfoundland populations of atricapillus and maritime province (Newfoundland plus Nova Scotia) populations of hudsonicus have distinct mtDNA haplotypes which differ from continental haplotypes by single restriction site changes. (4) Haplotypes of the southeastern U.S. species P. carolinensis divide into eastern and western sets which have diverged by three percent. This heretofore unrecognized, divided population structure may correspond to the Tombigbee River/Mobile Bay disjunction known in some other vertebrate taxa. (5) Allopatric populations of the southwestern species sclateri and gambeli exhibit divergences of one and three percent respectively. (6) Prevailing interspecific divergence distances of three to seven percent suggest speciation early in the Pleistocene rather than during late (e.g., Wisconsin) glaciations. (7) Phylogenetic analyses suggest that North American taxa include two clades, hudsonicus-rufescens-sclateri versus carolinensis-atricapillus-gambeli and that carolinensis and atricapillus are not sister species.

Key words. - Biogeography, birds, chickadees, genetic divergence, haplotype diversity, mtDNA, Parus, phylogeny, phylogeography, Quaternary, speciation.

The geographical diversity of mtDNA haplotypes in populations reflects patterns of historical fragmentation, changes in population size and distribution, and taxon-specific dispersal characteristics (Avise, 1989). Phylogenetic analysis of relationships among mtDNA haplotype clones further allows extrapolation of population processes to patterns of vicariant bio-geographical history (Avise et al., 1987; Avise, 1989, 1992; Zink, 1991; Zink and Dittmann, 1991). For example, geographically concordant branches of intraspecific mtDNA phylogenies reflect Pleistocene separations of regional populations of estuarine fishes, oysters, horseshoe crabs, and terrapins in the coastal southeastern United States (Avise, 1992). Differences in their mtDNA population structures relate in part to life-history patterns and dispersal characteristics. Inland are found even deeper, concordant phylogeographic disjunctions between mtDNA clones of freshwater fishes across the Apalachicola river drainage in southwestern Georgia (Bermingham and Avise, 1986).

Widespread species of North American birds tend to have large effective population sizes with high levels of dispersal and potential gene flow (Rockwell and Barrowclough, 1987). The few studies to date provide mixed results with respect to concordance between mtDNA genetic and morphometric population structure and also with respect to geographic structure of mtDNA haplotypes. MtDNA analyses of the fox sparrow, Passerella iliaca, revealed more population structure and greater consistency with patterns of geographical variation in morphology than did allozyme analyses (Zink, 1991). The widespread, abundant red-winged blackbird (Agelaius phoeniceus), however, exhibited little geographic structure of mtDNA haplotypes throughout its continental distribution, apparently due to high gene flow (Ball et al., 1987). Flickers (Colaptes) exhibited more geographical structure in haplotypes than blackbirds, but this structure did not match patterns of geographic variation in morphology or allozymes (Moore et al., 1991). In contrast, morphometric subspecies of Canada geese (Branta canadensis) are well defined by differences in their mtDNA genomes (Shields and Wilson, 1987a; Wagner and Baker, 1990). In such waterfowl with strong female philopatry, maternal lineage markers may reveal geographic structuring whereas nuclear genes affected by male dispersal do not (Wagner and Baker, 1990).

In this paper we establish a broad perspective of mtDNA variation and divergence in one set of six congeneric passerine bird species, the chickadees (Parus, subgenus Poecile) of North America. Few taxa of birds are as well known in their behavior, ecology, and population biology as are members of the genus Parus (Perrins, 1979; Gill and Ficken, 1989). These small passerine birds are common to abundant in deciduous and coniferous habitats of Eurasia and North America. For example, estimates of local densities of 10 to 25 pairs of atricapillus and gambeli per square kilometer (Hill and Lein, 1989; McCallum, 1990) suggest that populations exceed half a million females per 100,000 square kilometers of suitable habitat. Chickadees are permanent residents, although some species undertake seasonal altitudinal movements or disperse widely in poor seed crop years (McCallum, 1990). Reduced gene flow compared to migratory blackbirds, for example, should foster regional structuring of divergent clones. Also of interest are published scenarios of avian speciation in the late Pleistocene which invoke particularly the Wisconsin and Illinoian glaciations 100,000 and 250,000 years ago respectively (Brewer, 1963; Selander, 1965; Mengel, 1970; Hubbard, 1973; Dixon, 1978). In his model of chickadee speciation, for example, Brewer (1963) invokes the Illinoian glaciation as the key to the separation of Parus atricapillus and P. carolinensis, which have long been presumed to be sister species and perhaps conspecific (American Ornithologists' Union, 1983; Braun and Robbins, 1986; Robbins et al., 1986).

Materials and Methods

We obtained mtDNA samples of six North American species of chickadees including two or more recognized subspecies for each (Table 1, Figs. 1, 5). Voucher specimens are preserved in the collections of the Academy of Natural Sciences of Philadelphia. We did not include Parus cinctus, a Eurasian species found sparingly in the high arctic of northwestern North America. We surveyed restriction site variability within and between localities across the continental distributions of three species, P. atricapillus, P. carolinensis, and P. hudsonicus. To evaluate population variability, we used species specific sets of 11 enzymes which reliably cut two or more restriction sites. The number of restriction sites and the proportions of total mtDNA base pair composition sampled in each species were: atricapillus - 46/1.6%; carolinensis - 56/2.0%; hudsonicus - 43/1.5%. Protocols for the preservation, extraction, and restriction fragment analysis of mtDNA are summarized elsewhere (Gill and Slikas, 1992).

[TABULAR DATA OMITTED]

Our most comprehensive sample concerned the most widespread species, atricapillus, of which we sampled six subspecies from 10 states or provinces. Atricapillus and carolinensis hybridize in a zone of contact from New Jersey to Kansas (Rising, 1968; Robbins et al., 1986; Braun and Robbins, 1986), but we found no introgression of carolinensis mtDNA into populations of atricapillus or vice versa, except in the narrow hybrid zone itself. An analysis of this hybrid zone in Pennsylvania will be presented elsewhere (Gill et al., unpubl. data). We did not survey as fully the extent of geographic variation and genotypic diversity in three remaining species, P. sclateri, P. gambeli, P. rufescens. Instead we examined four individuals from two different populations each of sclateri, and five individuals from each of two rufescens and gambeli populations. The following number of restriction fragments/sites were analyzed in each population: (P. sclateri eidos - 32; P. s. rayi - 37; P. g. gambeli - 38; P. g. baileyi - 42; P. rufescens - 40).

For the standardized comparison of mtDNA divergence among taxa, we used a larger set of 15 restriction enzymes: Apa I; Ava I; Bam HI; Bgl I; EcoR-V; Hae I; Hinc II; Hind III; Hpa I; Nde I; Pst I; Pvu II; Sal I; Sst I; Sst II. All but two of these enzymes recognize 6-base pair sites; Ava I has the recognition sequence 5'-C PyCGPuG-3', Hinc II has the recognition sequence 5'GTPy PuAC-3'. These analyses were based on the prevailing haplotypes for each taxon and the following number of restriction sites: atricapillus - 48; carolinensis (EAST) - 53; carolinensis (WEST) - 50; gambeli baileyi - 58; gambeli gambeli - 52; hudsonicus - 48; rufescens - 53; sclateri eidos - 57; sclateri rayi - 60.

Relative mtDNA divergence was assessed in terms of distances (d) calculated from matrices of shared restriction fragments (Upholt, 1977; Nei and Li, 1979; Nei, 1987). These distances estimate the proportion of all base pairs in the mtDNA genome which have undergone substitution since two taxa or haplotypes had a common ancestor. The critical assumption of homology of comigrating fragments requires comparison of equal-sized mtDNA genomes and low probability of equal-sized fragments being cut by different restriction sites. We converted proportions of shared fragment matrices to estimates of nucleotide divergence following Nei's (1987) equation 5.55, weighted for products of 6-base pair and 5-base pair restriction enzymes. Unless noted otherwise, distances presented in the results below are based on matrices of shared restriction fragments.

The use of shared restriction fragments rather than shared restriction sites tends to underestimate divergence distances particularly distances [is greater than or equal to] 5%) and also is subject to greater sampling error (Nei and Li, 1979; Nei, 1987). Therefore, we also calculated divergence distances based on shared restriction sites which we inferred from the additive relationships of fragment lengths. The fragment profiles generated by Ava I and Hinc II were not amenable to this inference. For the remaining 13 enzymes, we inferred the fewest possible sites responsible for the combined fragment profiles of compared taxa, the spatial relations of sites to one another on the mtDNA genome, and their presence or absence in each taxon. Partials, or bands resulting from incomplete digests, aided this definition of sites. The inferred sites also were used to define phylogenetic relationships among taxa by parsimony analysis.

We converted matrices of inferred shared sites to estimates of nucleotide divergence (d) following Nei and Li (1979). Because our data set of inferred sites is based on 6-base restriction enzymes, Nei and Tajima's (1983) maximum likelihood estimate of d produced essentially the same result to three decimal places.

Results

MtDNA Genome Size. - The mtDNA genome sizes of these chickadees were approximately 16,600 base pairs based on direct comparisons of linearized mtDNA with comigrating size standards and from the sums of moderate size fragments. This genome size may be typical of most passerine birds (Shields and Helm-Bychowski, 1988). We detected no heteroplasmy nor any length variations due to insertions or deletions.

Variation in Parus atricapillus Table 2; Figs. 1, 2). - The five continental subspecies shared one widespread mtDNA haplotype (CCCCCCCCCCC), which defined 72% of 64 individuals across the broad continental distribution of this species from Alaska to Nova Scotia. Samples from Alaska, Washington, Utah, Missouri, and Nova Scotia contained this haplotype exclusively. Samples from Pennsylvania, Ontario, and Alberta contained the widespread haplotype plus two or more variant haplotypes. The six total variant haplotypes differed from the prevailing haplotype by one or two site changes. Four of the six variants shared a Sma I site loss which constituted most (78%) of the remaining continental sample (N = 18) and which dominated the Alberta sample.

[TABULAR DATA OMITTED]

The endemic Newfoundland subspecies P. a. bartletti lacked the continental haplotype CCCCCCCCCCC. Haplotypes of all 18 individuals examined of this subspecies were distinguished by a unique (Pvu II) site gain. In addition, the haplotypes of three individuals (17%) lacked a prevailing Bgl II restriction site.

The (unweighted) average nucleotide distance (p) between pairs of atricapillus haplotypes was 0.53 [+ or -] 0.23% (N = 36).

Variation in Parus carolinensis (Table 3; Figs. 1, 3, 4). - Prior to this survey of mtDNA geographic variation, we had no indication of an east-west division of carolinensis populations based on morphological or vocal characters. The poorly defined subspecies of P. carolinensis include a northern one (extimus), a southern one (carolinensis), and two (agilis, atricapilloides) with limited but largely unresolved distributions in the southwestern part of the species' range. In contrast to atricapillus haplotypes, carolinensis haplotypes split into two divergent geographical sets: (1) "eastern" haplotypes from Pennsylvania, New Jersey, eastern Kentucky, North Carolina, Georgia and Alabama; and (2) "western" haplotypes from Texas, Mississippi, Missouri, Louisiana, and western Kentucky. The eastern haplotypes were represented in two morphological subspecies, P. c. carolinensis and P. c. extimus. The western haplotypes were represented in four morphological subspecies, P. c. agilis, P. c. atricapilloides, P. c. carolinensis and also P. c. extimus (one specimen from western Kentucky). The two most widespread subspecies, carolinensis and extimus, thus include both eastern and western haplotypes.

[TABULAR DATA OMITTED]

Our samples revealed six variants each of the principal eastern and western haplotypes. Single site differences defined four eastern and all western variants; two site differences defined the remaining two eastern variants. Variant haplotypes constituted 50% of the eastern samples (N = 28), 33% of the western samples (N = 24), and 0 to 60% of (small) local subsamples. The 21 possible paired comparisons of eastern haplotypes and western haplotypes respectively suggest average nucleotide divergences within each set of 0.45 [+ or -] 0.21% (EAST) and 0.40 [+ or -] 0.13% (WEST).

The fragment profiles of the eastern haplotypes differ markedly from those of the western haplotypes. Eight of 11 enzymes produced diagnostic profiles, 5 of which reflect single site changes. Multiple or independent site gains relative to a putative ancestral haplotype were responsible for the others. Divergence based on these selected informative enzymes is estimated to be 3.7% including a correction of 0.4% for haplotype divergence within populations. A parsimony network of all 14 haplotypes connects eastern haplotypes and western haplotypes via haplotype #12, a Louisiana variant which lacks an Nde I site present in other western haplotypes, and haplotype #4, a North Carolina variant which lacks an Ava I site present in other eastern haplotypes. One Hind Ill site gain featured by eastern haplotypes #6 and #7 was defined by the parsimony network as convergent with the same site present in all western haplotypes. A different Hind III site change distinguished all eastern haplotypes from all western haplotypes.

Using two diagnostic enzymes (Hind III, Pvu II), we screened small samples of carolinensis populations on a coarse transect between the western slopes of the Appalachian Mountains and the Mississippi River to locate the region of contact of these two clones (Fig. 1). These small samples preclude accurate definition of the contact zone but suggest that the two divergent haplotypes come into contact in eastern Mississippi, possibly in western Alabama (Fig. 4), and also due north in southwestern Kentucky (Trigg County). Our sample from eastern Scott County in central Mississippi comprised four western and one eastern haplotype. The three individuals collected 185 km east northeast in western Bibb County, Alabama, were all eastern haplotypes. Thus, the contact zone appears to be centered between these two localities near the Mississippi-Alabama state line and the valley of the Tombigbee River system which drains south into Mobile Bay. Almost due north in southwestern Kentucky, we collected one eastern and one western haplotype at the same locality in Trigg County. Samples from opposite sides of the Mississippi River in Madison Parish, Louisiana and western Mississippi (western Issaquena County, 36 km north of Vicksburg) contained only western haplotypes. Samples from eastern Alabama (eastern Talladega County, 12 km sw Talladega), and eastern Kentucky (Rowan County) contained only eastern haplotypes.

Variation in Parus hudsonicus (Table 4; Figs. 5, 6). - The results for hudsonicus, a northern boreal forest species, paralleled the results for atricapillus. A single widespread hudsonicus haplotype (CCCCCCCCCCC) occurs across the continent from Alaska to Ontario. The weakly marked subspecies P. h. columbianus of coastal Alaska did not differ from the nominate subspecies which occupies central Alaska. The principal continental haplotype CCCCCCCCCCC, however, was absent from the maritime provinces of Canada. There we found a unique haplotype (CBCCCCCCCCB) in all individuals of the Newfoundland subspecies, P. h. rabbittsi, and in all individuals in our sample from Nova Scotia. The two restriction site differences that distinguished this haplotype were present separately in other locations; see below. The Nova Scotia population is assigned to the eastern continental subspecies P. h. littoralis. However, the maritime province haplotype was not present in our sample of four littoralis from northern New York state.

[TABULAR DATA OMITTED]

Six variant haplotypes each differed by one restriction site loss from the widespread haplotype CCCCCCCCCCC. The maritime province haplotype differed from this haplotype by two site differences, one of which (Ava I) was present also as a variant in Ontario and one of which (Xho I) was present in both Ontario and New York. Variant haplotypes constituted 3 to 8% of the continental North American sample and were most evident in the samples from New York (3/4), Ontario (2/4), and Alberta (3/5). The average nucleotide divergence between pairs of haplotypes was 0.46 [+ or -] 0.19%, N = 28.

Variation in Other Species. - Within population variation was limited to two variant haplotypes in P. gambeli baileyi and one variant in P. gambeli gambeli. All three variant haplotypes reflected single site losses. Each of the two P. sclateri populations was homogeneous. Divergences between the subspecies of gambeli and sclateri were 3.1 % (adjusted for within population variation of 0.2%) and 1.4% respectively. The difference reflects a total of 16 site changes between the two populations of gambeli versus six site changes between the two sclateri populations. The two subspecies of P. rufescens featured the same invariant mtDNA haplotype.

Genetic Diversity (Table 5). - Across localities on a continental scale, genotype (haplotype) diversity within species ranged from 0.46 to 0.71, the values partly reflecting sample heterogeneity. The highest values were slightly lower than the value reported for the migratory red-winged blackbird Ball et al., 1987). Genotype diversity in Newfoundland atricapillus was low (0.29). Also, continental atricapillus (0.46, excluding Newfoundland) appear to have slightly lower genotype diversity than in comparable samples of carolinensis or hudsonicus. Genotype diversity may be higher in some local populations (atricapillus - Alberta; hudsonicus - Alberta/New York/Ontario; carolinensis - North Carolina) than others, but larger and better controlled local samples are needed to verify this.

[TABULAR DATA OMITTED]

Estimates of nucleotide diversity varied from 3 x [10.sup.-4] to 27 x [10.sup.-4] excluding combinations of eastern and western carolinensis haplotypes. The upper value is close to that reported for the widespread red-winged blackbird, Agelaius phoeniceus (Ball et al., 1987). These low values reflect the absence of divergent haplotype polymorphisms.

Divergences among Taxa (Table 6). - Comparisons of the nine chickadee taxa were based on shared proportions of a total of 178 restriction fragments. The sites responsible for these fragments sampled approximately six percent of the chickadee mtDNA genomes. Estimates of divergence distances among chickadee taxa ranged from 0.01 (subspecies of sclateri) to 0.07 (gambeli gambeli versus rufescens) with a mean of 0.046 [+ or -] 0.012 (SD). As expected, distances for the conspecific comparisons were lower than those based on the enzymes selected for surveys of population and geographic variation. Divergence distances between eastern and western P. carolinensis (0.024) and between P. g. gambeli and P. g. baileyi (0.026) ranked slightly below those for most pairs of recognized species, e.g., 0.04 to 0.06.

[TABULAR DATA OMITTED]

We also compared taxa with respect to the proportions of shared inferred sites (N = 111). These data excluded Ava I and Hinc II sites, which we could not infer from fragment patterns. The site distances (0.052 [+ or -] 0.013) averaged slightly higher than the fragment distances.

Detailed phylogenetic analyses of relationships among both New World and Old World taxa of Parus will be presented elsewhere. Here we summarize the relationships among the six North American species of chickadees based first on a parsimony analysis of shared inferred sites rooted with Parus bicolor (subgenus Baeolophus) and Parus major (subgenus Parus) as the designated outgroups (see Appendix). The one most parsimonious tree weakly defined two principal clades of chickadees: (1) hudsonicus-rufescens-sclateri and (2) carolinensis-gambeli-atricapillus (Fig. 7). The basal nodes of the hudsonicus clade and carolinensis clade were supported by only two and three changes respectively. All subspecies or conspecific populations paired appropriately, including eastern and western carolinensis. The definition of gambeli and atricapillus as sister species within the carolinensis clade was supported by six synapomorphies. Carolinensis and atricapillus did not align as sister species in this or any other mtDNA analysis which included gambeli or the Old World species montanus.

FITCH analysis (PHYLIP 3.4) of divergence distances based on the same inferred sites and rooted by major as the designated outgroup produced a best tree minimized sum of squares) with the same topology as the parsimony analysis. FITCH analysis of fragment distances agreed with the site distance analysis in the definition of eastern and western carolinensis as sister taxa, and in the placement of atricapillus as the sister species of gambeli, but not in the positions of carolinensis which was placed basally to all other chickadees, or hudsonicus which was placed in the atricapillus-gambeli clade.

DISCUSSION

The largely allopatric distributions of New World chickadees correspond to those of other superspecies of North American birds presumed to have speciated during late Pleistocene (Selander, 1965; Mengel, (1970). In his model of chickadee speciation, Brewer (1963) invokes the Illinoian glaciation as the key to the separation of atricapillus and carolinensis. Our mtDNA divergence distances of 4% (fragments) to 5% (sites) between these two species suggest a much earlier separation, roughly two million years ago based on the current estimate of the rate of mtDNA divergence of 2% per one million years (Shields and Wilson, 1987b; Shields and Helm-Bychowski, 1988; Hillis and Moritz, 1990). More generally, the narrow range of restriction fragment divergence distances among North American chickadee species (3 to 7%) suggests that speciation in this group took place primarily in the first half of the Pleistocene, if not earlier. Corrections for haplotype divergence within populations (0.1 to 0.4%) would not change this projection.

Our mtDNA-based projections of early rather than late Quaternary speciation in chickadees are in accord with similar conclusions for a variety of congeneric vertebrates, including desert tortoises (Xerobates), grasshopper mice (Onychomys), and brown towhees (Pipilo) of the arid habitats of western North America (Lamb et al., 1989; Riddle and Honeycutt, 1990; Zink and Dittmann, 1991), and freshwater fishes of the southeastern U.S. (Bermingham and Avise, 1986; Avise, 1992). Even though the mtDNA divergence rate of 2% per million years is weakly founded, carries a large potential error, and varies among taxa in relation to population history, frequency of bottlenecks, etc. (Hillis and Moritz, 1990), this often invoked rate of divergence would have to be an order of magnitude faster to bring our results into accord with late Pleistocene glaciations. There is no agreement, however, whether rates are potentially faster, slower, or taxonomically variable (Lewin, 1991; Avise, 1992). What emerges nevertheless is an apparent conflict between classical hypotheses about the timing of Quaternary speciation events based on plausible biogeographic scenarios and more recent hypotheses based on mtDNA comparisons. Either our assumptions about the approximate rates of mtDNA divergence are wrong, or else some North American species are much older than we once presumed. Both perspectives probably need adjustment.

Not only are carolinensis and atricapillus apparently much older taxa than Brewer (1963) surmised, they probably are not sister species. Distance and character parsimony analyses both define gambeli and atricapillus as sister species. Furthermore, at least two Eurasian species, P. montanus and P. palustris, constitute the sister group to the atricapillus-gambeli species complex, phylogenetically distancing carolinensis even further from atricapillus (Gill and Slikas, unpubl. data). If these phylogenetic inferences are correct, they challenge the worth of debates about the species status of carolinensis versus atricapillus based on the extent of hybridization and gene flow in their extended but narrow zone of secondary contact (Rising, 1968; Braun and Robbins, 1986; Robbins et al., 1986). Our analysis of the phylogenetic relationships among North American chickadees supports the growing awareness that hybrid zones may involve taxa which are less closely related than previously supposed and whose similarity reflects retention of ancestral features (McKitrick and Zink, 1988).

More interesting perhaps than the relationship between carolinensis and atricapillus are the two, previously unrecognized, apparently monophyletic, genetically diagnosable, geographical components of the species "carolinensis." The case for species recognition of the western "Cajun Chickadee" will be presented elsewhere (Gill, unpubl. data). Eastern and western carolinensis appear to have been separated at least one million years. Divergence between eastern and western carolinensis mtDNA haplotypes is half that characterizing currently recognized species of chickadees and the same as between two well-marked, geographically isolated subspecies of gambeli. To the best of our knowledge, eastern and western carolinensis have not diverged in their external morphology and vocal repertoires. Our preliminary transect through the southeastern United States suggests that the geographical disjunction between eastern carolinensis and western carolinensis corresponds to old disjunctions between other pairs of vertebrate taxa of the southeastern United States (Avise, 1992). Many of those disjunctions center near the Apalachicola River in northwest Florida and southeast Alabama, but a variety of sister taxa of freshwater fish (e.g., Fundulus nottii x F. escambiae; Wiley and Mayden, 1985), water snakes (e.g., Nerodia rhombifera x N. taxispilota; Lawson, 1987), and perhaps rabbits (e.g., Sylvilagus palustris x S. aquaticus; Chapman and Willner, 198 1; Chapman and Feldhame, 1981) have vicariant disjunctions which correspond to the Mobile Bay/Tombigbee River disjunction we project for carolinensis chickadees. In addition there is a growing assortment of cases like the chickadees which were not easily recognized by classical taxonomists but which involve disjunctions between mtDNA clades east and west of the Tombigbee River (e.g., Lepomis gulosus) (Bermingham and Avise, 1986; Avise et al., 1987; Avise, 1992).

Previous protein (distance) comparisons of chickadees revealed few differences among species and suggested that the relationships among chickadees be viewed conservatively as an unresolved polytomy (Gill et al., 1989). MtDNA analyses do not confirm the speculation that sclateri and gambeli may be terminal sister species. Hudsonicus and rufescens, however, cluster together with sclateri. These three species now occupy geographically peripheral spruce/fir forest habitats. In particular, rufescens is restricted to the coastal forests of the Pacific northwest coast whereas sclateri is restricted to the montane fir forests of central and western Mexico (and barely southeastern Arizona) south of about 32 degrees north latitude. The three species in this clade also have vocal repertoires which differ substantially from the other North American chickadees in both the nasal quality of their "chickadee" calls and in lacking the whistled song for territorial advertisement (Hailman, 1989; Ficken and Nocedal, 1992). In addition they retain certain repertoire elements which occur in most members of the genus but which are absent from the repertoires of carolinensis-atricapillus-gambeli. Further phylogenetic analysis is needed to determine whether hudsonicus, rufescens, and sclateri position reliably as primitive members of the subgenus Poecile.

The limits to estimating mtDNA genome divergence from restriction fragments versus inferred restriction sites remain poorly defined. Both are subject to a variety of potential errors of interpretation and inference (Nei and Li, 1979; Swofford and Olsen, 1990). The use and comparison of fragment distances versus site distances must include recognition (a) that the theoretical and empirical relationships between the two are not linear and (b) that the intrinsic variance of fragment-based estimates is greater than that of site-based distances (Nei and Li, 1979; Nei, 1987). Also, shared site matrices are more amenable to parsimony analysis of phylogenetic relationships than are shared fragment matrices (Swofford and Olsen, 1990; Zink and Avise, 1990). Double-digest mapping of a minimum of 111 restriction sites across the nine chickadee taxa, however, would be a formidable task of questionable return. Our procedures of site inference push practical limits of the procedure but not, we believe, to the point of distortion. We do not infer sites by simple direct comparison of two divergent taxa. Rather we infer most sites initially from the comparisons of pairs of the most closely related taxa in the matrix in a conventional way including the use of partial digests. We then invoke these sites to explain the fragment profiles of more divergent combinations of the same taxa. We invoke new or unique sites only when we cannot interpret differences between profiles in the simplest possible way consistent with comparisons of the least divergent profiles. Use of the fewest possible inferred sites probably underestimates divergence among species.

Genetic diversity within populations potentially increases with time as new variants accumulate and diverge but potentially decreases as a result of losses due to stochastic extinction of lineages, skewed distribution of lifetime reproductive success among females, and other demographic factors especially in small populations (Avise et al., 1987). Our estimates of genotype diversity for chickadee populations broadly defined fall within the range of reports for vertebrates with a variety of population characteristics (Densmore et al., 1989; but see Tegelstrom, 1987) on haplotype variability in Swedish populations of Parus major based on surveys with 4-base site restriction enzymes). Nucleotide diversity within the North American chickadee populations was low because polymorphic haplotypes typically differed from each other by only one or two restriction site changes. Limiting the potential for divergence of alternative haplotypes in parid populations are their well known fluctuations in relation to seed crop abundance and associated overwinter survivorship (Perrins, 1979; McCallum, 1990). Regularly reduced population sizes combined with negligible lifetime reproductive success by most female tits (Dhondt, 1989; McCleery and Perrins, 1989) should promote mtDNA haplotype homogeneity within populations. Regular pruning of divergent maternal lineages will also tend to synchronize patterns of haplotype divergence with population divergence.

We found no simple correspondence between the geographical patterns of mtDNA genetic divergence and the geographical patterns of morphological divergence defined by classical subspecies. Like migratory blackbirds (Ball et al., 1987), resident North American chickadees exhibit little geographic structuring of alternative clones. Most notable were the two widespread northern species (atricapillus and hudsonicus) which occupy the once-glaciated northern regions of North America and which comprise nine and five morphological subspecies respectively (American Ornithologists' Union, 1957; though see Phillips, 1986). Low genotype diversities, especially in the western populations, and the widespread continental distribution of a single haplotype seem best explained by recent expansion from a smaller refugium during the past 10,000 years, i.e., since the retreat of the Wisconsin ice shield. Variable eastern populations of atricapillus and hudsonicus may be older than the genetically homogeneous Alaskan populations. Highton and Webster (1976) invoked a similar explanation for the low levels of allozyme polymorphism in the salamander, Plethodon cinereus, which expanded its range northward in the last 10,000 years as the Wisconsin ice shield retreated. Salamanders restricted to the older unglaciated regions of the southeastern United States have substantially higher levels of allozyme polymorphism within populations. Historically recent colonization provides an alternative explanation for lack of differentiation among populations which once were ascribed to current homogenizing gene flow (Slatkin, 1987). The only populations (and subspecies) of atricapillus and hudsonicus with fixed, albeit minor, haplotype distinctions were the populations in extreme eastern Canada - Newfoundland in the case of atricapillus, and Newfoundland plus adjacent Nova Scotia in the case of hudsonicus. Coastal land refugia existed in this region during the last glacial maximum (Pielou, 1991).

Beyond the need for modern reevaluation of chickadee subspecies (Phillips, 1986) lie issues of concordance between genetic versus morphological patterns of geographical population structure in North American birds. Except for one isolated subspecies, Agelaius phoeniceus of central Mexico, the subspecies of blackbirds surveyed by Ball et al. (1987) were not genetically diagnosable. Similarly, the woodpecker Colaptes auratus exhibits different patterns of geographic variation in morphological, allozyme and mtDNA traits (Moore et al., 1991). In this study the morphological subspecies of three species of chickadees, viz. atricapillus, carolinensis, and hudsonicus, were not diagnosed by their mtDNA haplotypes. With the notable exception of the Newfoundland populations of atricapillus and hudsonicus, we found little correspondence between mtDNA variation and subspecies assignments. Certain subspecies, e.g., P. carolinensis carolinensis and P. hudsonicus littoralis, apparently embrace more than one geographically structured mtDNA haplotype. In the southwestern United States, however, the geographically isolated subspecies of sclateri and gambeli feature divergent haplotypes. The divergence between a Pacific coastal and an interior subspecies of gambeli (3%) parallels mtDNA divergence (4%) between coastal and interior subspecies of the titmouse P. inornatus (Gill and Slikas, 1992) and also morphologically distinct populations of the sparrow Passerella iliaca (Zink, 1991). Thus, molecular data reaffirm previously recognized biogeographic distinctions of sister taxa of birds in this part of the continent (Miller, 1956; Johnson, 1978). The few cases now available suggest therefore that genetic population structure may match current taxonomic distinctions among some historically resident, isolated populations, but not among bird populations which expanded into the northern, postglacial forests less than 10,000 years ago. Such recently expanded populations continue to share common mtDNA genotypes but have diverged in size or coloration as a result of recent genetic or ecophenotypic change.

Literature Cited

American Ornithologists' Union (A.O.U.). 1957. Check-list of North American Birds, 5th ed. The Lord Baltimore Press, Inc., Baltimore, MD USA. _____. 1983. Checklist of North American Birds, 6th ed. Allen Press, Lawrence, KS USA. Avise, J. C. 1989. Gene trees and organismal histories: A phylogenetic approach to population biology. Evolution 43:1192-1208. _____. 1992. Molecular population structure and the biogeographic history of a regional fauna: mtDNA analyses of marine, coastal, and freshwater species in the southeastern United States. Oikos 63:62-76. Avise, J. C., J. Arnold, R. M. Ball, E. Bermingham, T. Lamb, J. E. Neigel, C. A. Reeb, and N. C. Saunders. 1987. Intraspecific phylogeography: The mitochondrial DNA bridge between population genetics and systematcs. Annu. Rev. Ecol. Syst. 18:489-522. Ball, R. M., S. Freeman, F. C. James, and E. Bermingham. 1987. Phylogeographic population structure of Red-winged Blackbirds assessed by mitochondrial DNA. Proc. Natl. Acad. Sci. USA 85: 1558-1562. Bermingham, E., and J. C. Avise. 1986. Molecular zoogeography of freshwater fishes in the southeastern United States. Genetics 113:939-965. Braun, M. J., and M. B. Robbins. 1986. Extensive protein similarity of the hybridizing chickadees Parus atricapillus and P. carolinensis. Auk 103:667-675. Brewer, R. 1963. Ecological and reproductive relationships of Black-capped and Carolina chickadees. Auk 80:9-47. Chapman, J. A., and G. A. Feldhamer. 1981. Sylvilagus aquaticus. In Mammalian Species, No. 151. Published by The American Society of Mammalogists. Chapman, J. A., and G. R. Willner. 1981. Sylvilagus palustris. In Mammalian Species, No. 153. Published by The American Society of Mammalogists, Stillwater, OK USA. Densmore, L. D., J. W. Wright, and W. M. Brown. 1989. Mitochondrial-DNA analyses and the origin and relative age of parthenogenetic lizards (genus Cnemidophorus). II. C. neomexicanus and the C. tesselatus complex. Evolution 43:943-957. Dhondt, A. 1989. Blue Tit, pp. 15-33. In I. Newton (ed.), Lifetime Reproduction in Birds. Academic Press, London, UK. Dixon, K. 1978. A distributional history of the Black-crested Titmouse. Am. Midl. Nat. 100:29-42. Ficken, M. S., and J. Nocedal. 1992. Mexican Chickadee. In A. Poole, P. Stettenheim, and F. Gill (eds.), The Birds of North America, No. 8. The Academy of Natural Sciences, Philadelphia, PA; The American Ornithologists' Union, Washington, DC USA. Gill, F. B., and M. S. Ficken. 1989. Comparative biology and evolution of titmice: The centennial symposium of the Wilson Ornithological Society. Wilson Bull. 101:180-181. Gill, F. B., D. Funk, and B. Silveren. 1989. Protein relationships among titmice (Parus). Wilson Bull. 101:182-197. Gill, F. B., and B. Slikas. 1992. Patterns of mtDNA genetic divergence in North American crested titmice. Condor 94:20-28. Hailman, J. P. 1989. The organization of major vocalizations in the Paridae. Wilson Bull. 101:305-343. Highton, R., and T. P. Webster. 1976. Geographic protein variation and divergence in populations of the salamander Plethodon cinereus. Evolution 30: 33-45. Hill, B. G., and M. R. Lein. 1989. Territory overlap and habitat use of sympatric chickadees. Auk 106: 259-268. Hillis, D. M., and C. Moritz. 1990. An overview of applications of molecular systematics, pp. 502-515. In D. M. Hillis and C. Moritz (eds.), Molecular Systematics. Sinauer Associates, Sunderland, MA USA. Hubbard, J. P. 1973. Avian evolution in the arid-lands of North America. Living Bird 12:155-196. Johnson, N. K. 1978. Patterns of avian geography and speciation in the intermountain region. Great Basin Nat. Mem. 2:137-159. Lamb, T., J. C. Avise, and J. W. Gibbons. 1989. Phylogeographic patterns in mitochondrial DNA of the desert tortoise (Xerobates agassizi), and evolutionary relationships among the North American gopher tortoises. Evolution 43:76-87. Lawson, R. 1987. Molecular studies of thamnophine snakes: I. The phylogeny of the genus Nerodia. J. Herpetol. 21:140-157. Lewin, R. 1991. The biochemical route to human origins. Mosaic 22:46-55. McCallum, A. 1990. Variable cone crops, migration, and dynamics of a population of mountain chickadees (Parus gambeli), pp. 103-116. In J-D. Lebreton and R. E. McCleen (eds.), Population Biology of Passerine Birds, An Integrated Approach. Springer-Verlag, N.Y., USA. McCleery, R. H., and C. M. Perrins. 1989. Great Tit, pp. 35-53. In I. Newton (ed.), Lifetime Reproduction in Birds. Academic Press, London, UK. McKitrick, M. C., and R. M. Zink. 1988. Species concepts in ornithology. Condor 90:1-14. Mengel, R. M. 1970, The North American central plains as an isolating agent in bird speciation, pp. 279-340. In W. Dart, Jr. and J. K. Jones, Jr. (eds.), Pleistocene and recent environments of the central Great Plains. Univ. Kans. Dept. Geol. Spec. Publ. No. 3. Miller, A. H. 1956. Ecological factors that accelerate formation of races and species of terrestrial vertebrates. Evolution 10:262-277. Moore, W. S., J. H. Graham, and J. T. Price. 1991. Mitochondrial DNA variation in the Northern Flicker. Mol. Biol. Evol. 8:327-344. Nei, M. 1987. Molecular Evolutionary Genetics. Columbia University Press, N.Y., USA. Nei, M., and W.-H. Li. 1979. Mathematical model for studying genetic divergence in terms of restriction endonucleases. Proc. Natl. Acad. Sci. USA 76: 5269-5273. Nei, M., and F. Tajima. 1983. Maximum likelihood estimation of the number of nucleotide subsitutions from restriction sites data. Genetics 105:207-217. Perrins, C. 1979. British Tits. Collins, London, UK. Phillips, A. R. 1986. The Known Birds of North and Middle America. Part 1. Hirundinidae to Mimidae; Certhiidae. A. R. Philips, Denver, CO USA. Pielou, E. C. 1991. After the Ice Age. The University of Chicago Press, Chicago, IL USA. Riddle, B. R., and R. L. Honeycutt. 1990. Historical biogeography in North American arid regions: An approach using mitochondrial-DNA phylogeny in grasshopper mice (Genus Onychomys). Evolution 44:1-15. Rising, J. D. 1968. A multivariate assessment of interbreeding between the chickadees Parus atricapillus and P. carolinensis. Syst. Zool. 17:160-169. Robbins, M. B., M. J. Braun, and E. A. Tobey. 1986. Morphological and vocal variation across a contact zone between the chickadees Parus atricapillus and P. carolinensis. Auk 103:655-666. Rockwell, R. F., and G. F. Barrowclough. 1987. Gene flow and the genetic structure of populations, pp. 223-255. In F. Cooke and P. A. Buckley (eds.), Avian Genetics. Academic Press, N.Y., USA. Selander, R. K. 1965. Avian speciation in the Quaternary, pp. 527-542. In H. E. Wright, Jr., and D. G. Frey (eds.), The Quaternary of the United States. Princeton University Press, Princeton, NJ USA. Shields, G. F., and K. M. Helm-Bychowski. 1988. Mitochondrial DNA of birds. Curr. Ornithol. 8:273-295. Shields, G. F., and A. Wilson. 1987a. Subspecies of Canada Goose (Branta canadensis) have distinct mitochondrial DNA's. Evolution 41:662-666. _____. 1987b. Calibration of mitochondrial DNA evolution in geese. J. Mol. Evol. 24:212-217. Slatkin, M. 1987. Gene flow and the geographic structure of natural populations. Science 236:787-792. Swofford, D. L., and G. J. Olsen. 1990. Phylogeny reconstruction, pp. 411-501. In D. M. Hillis and C. Moritz (eds.), Molecular Systematics. Sinauer Associates, Inc., Sunderland, MA USA. Tegelstrom, H. 1987. Genetic variability in mitochondrial DNA in a regional population of the Great Tit (Parus major). Biochem. Genet. 25:95-110. Upholt, W. B. 1977. Estimation of DNA sequence divergence from comparison of restriction endo-nuclease digests. Nucleic Acids Res. 4:1257-1265. Wagner, C. E. V., and A. J. Baker. 1990. Association between mitochondrial DNA and morphological evolution in Canada Geese. J. Mol. Evol. 31:373-382. Wiley, E. O., and R. L. Mayden. 1985. Species and speciation in phylogenetic systematics with examples from the North American fish fauna. Ann. Mo. Bot. Gard. 72:596-635. Zink, R. M. 1991. The geography of mitochondrial DNA variation in two sympatric sparrows. Evolution 45:329-339. Zink, R. M., and J. C. Avise. 1990. Patterns of mitochondrial 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 evidence. Condor 93:98-105.
COPYRIGHT 1993 Society for the Study of Evolution
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Speciation in North American Chickadees, part 1
Author:Gill, Frank B.; Mostrom, Alison M.; Mack, Andrew L.
Publication:Evolution
Date:Feb 1, 1993
Words:6642
Previous Article:Examination of population structure in red-cockaded woodpeckers using DNA profiles.
Next Article:Responses and correlated responses to artificial selection on thorax length in Drosophila melanogaster.
Topics:

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |