Postglacial range fluctuation, genetic subdivision and speciation in the western North American spotted frog complex, Rana pretiosa.
At the borders of a range, there is continual flux between the rate of establishment of new populations and the rate of extinction of old populations (Hoffman and Blows 1994). If establishment exceeds extinction, the range will expand into available, unoccupied habitats. At the population level, the potential for expansion via the net emigration of individuals can be generated if there are growing populations where the intrinsic rate of increase, r, is positive (Gilpin 1987; Pulliam 1988; Gilpin and Hanksi 1991). In an expanding frontier, populations are thus the products of successive founding events as new regions are colonized and should predictably tend toward loss and partitioning of genetic diversity (Hewitt 1993a) in a pattern of progressive isolation-by-distance (Wright 1943). A continuing flux of gene flow and expansion will keep populations out of genetic equilibrium (Slatkin 1993), and the repeated founding of new populations during expansion will tend to reduce heterozygosity (Nagylaki 1976). Range expansion can happen quickly under favorable conditions. The explosive population growth and range expansion of introduced exotics produces populations with demonstrably reduced genetic variance (Easteal 1988).
A range will continue to expand until the rate of founding new populations fails to exceed the rate at which they go extinct. When the rate of extinction begins to surpass the rate of population establishment, the range will start to contract. Particularly in short-lived species with low vagility, those populations on the edge of a range are likely to be net population "sinks" (Pulliam 1988; Gilpin and Hanski 1991), and their persistence is dependent upon immigration from neighboring areas. Those populations that cannot be rescued because of insufficient immigration will be eliminated permanently. The range will fragment as it shrinks, leading to differentiated, isolated, and relict populations (Hewitt 1993a). Shrinking gene pools and repeated bottlenecking will increase interpopulational variation and the populations should settle into genetic equilibrium once gene flow is curtailed (Slatkin 1993). There may also be reductions in within-population heterozygosity. Therefore, both the expanding and contracting fronts of a range should exhibit reduced levels of genetic variability and heterozygosity compared to the range center, although with differing distributions of that variation.
In this paper, we examine the predictable genetic effects of postglacial recolonization in a frog that has shifted its range northward in postglacial times. The western North American spotted frog, Rana pretiosa, has expanded into previously glaciated regions of British Columbia, Washington State, Idaho, and Montana, all the way north to extreme southwestern Yukon (Turner and Dumas 1972; Nussbaum et al. 1983; Stebbins 1985), but its range is highly fragmented throughout its southern extent [ILLUSTRATION FOR FIGURE 1 OMITTED]. Particularly in Nevada and Utah, there are disjunct populations isolated on mountain-top refuges or in springs surrounded by dry desert inimical to frogs (W. M. Tanner 1931; W. W. Tanner 1978; Linsdale 1940; Banta 1986; Hovingh 1993). Geological evidence puts the postglacial isolation of these springs and rivers at about 6000-8000 years ago (Brues 1932; Banta 1986; Thompson 1988) and the opening of northern habitats for spotted frogs even more recently than that (Pielou 1991). The postglacial range expansion of R. pretiosa should have resulted in demonstrable isolation-by-distance among the recently colonizing northern populations whereas the genetic fragmentation of relict southern populations should be discernable as genetic equilibrium and inbreeding. We assayed for these consequences of the postglacial range shift in R. pretiosa by examining patterns of allozyme and morphological variation between populations from all parts of the frog's extensive range.
MATERIALS AND METHODS
Allozyme variation was examined in 264 specimens of R. pretiosa from 26 populations (Table 1, [ILLUSTRATION FOR FIGURE 1 OMITTED]). For comparison, we also used samples of the related species R. cascadae (two populations), R. aurora, and R. muscosa (Case 1978: Green 1986). Preserved specimens are deposited at the Canadian Museum of Nature, Ottawa. Samples of heart, liver, spleen, kidney, and body wall muscle of each frog were used for horizontal starch gel electrophoresis to resolve 25 specific enzyme systems as described in Green (1986), Green and Borkin (1993) and Murphy et al. (1990). We stained for Aconitate Hydratase (ACOH, E.C. 22.214.171.124), Alcohol Dehydrogenase (ADH, E.C. 126.96.36.199), Aspartate Aminotransferase (AAT, E.C. 188.8.131.52), Carbonic Anhydrase (CA, E.C. 184.108.40.206), Creatine Kinase (CK, E.C. 220.127.116.11), Fructose-bisphosphate Aldolase (FBA, E.C. 18.104.22.168), Fumarase (FUMH, E.C. 22.214.171.124), General Proteins (GP), Glucose Dehydrogenase (GCDH, E.C. 126.96.36.199), Glucose-6-phosphate Dehydrogenase (GPI, E.C. 188.8.131.52), Glyceraldehyde-3-phosphate Dehydrogenase (GAPDH, E.C. 184.108.40.206), Glycerol-3-phosphate Dehydrogenase (G3PDH, E.C. 220.127.116.11), Glutamate Dehydrogenase (GTDH, E.C. 18.104.22.168), L-Iditol Dehydrogenase (IDDH, E.C. 22.214.171.124), Isocitrate Dehydrogenase (IDH, E.C. 126.96.36.199), L-Lactate Dehydrogenase (LDH, E.C. 188.8.131.52), Malate Dydrogenase (MDH, E.C. 184.108.40.206), Malate Dehydrogenase [NADP+] (MDHP, E.C. 220.127.116.11), Dipeptidase (PEP, E.C. 18.104.22.168), Tripeptide Aminopeptidase (PEP-B, E.C. 22.214.171.124), Phosophoglucomutase (PGM, E.C. 126.96.36.199), Phosophogluconate Dehydrogenase (PGDH, E.C. 188.8.131.52), Superoxide Dismutase (SOD, E.C. 184.108.40.206), Triose-Phosphate Isomerase (TPI, E.C. 220.127.116.11), and Xanthine Dehydrogenase (XDH, E.C. 18.104.22.168).
TABLE 1. Samples of spotted frogs, Rana pretiosa complex, and related species used for allozyme studies and for multiple discriminant function analyses (MDA) of morphometric characters. For MDAs, two alternative group assignments of Rana pretiosa and R. cascadae populations (MDA I and MDA II) were used in separate analyses (see text).
Morphology Allo- MDA zymes Groupings Locality n n I II
Rana pretiosa complex
British Columbia and Alaska 5 -
Stikine River, AK 7 - Terrace, BC 7 - Prince George, BC 8 - Manning Park, BC 16 - Champion Lake, BC - 23 J B Fernie, BC - 31 J B McLeod Lake, BC - 1 J B Moose River, BC - 9 J B Storm Creek Headwaters, BC - 4 J B Valemount, BC - 22 J B
Little Campbell River, Surrey, BC - 8 H A Seattle, King Co., WA - 2 H A Puget Sound, WA - 1 H A
Rainy Pass, Skagit Co., WA 4 - Ellensburg, Kittitas Co., WA 10 2 G B Fort Walla Walla, Walla Walla Co., WA - 8 G B North Yakima, Yakima Co., WA - 3 G B
North and Central Idaho
Craig Mountain, Nez Perce Co., ID 5 - Muldoon Creek, Blaine Co., ID - 7 F B McCall, Boise Co., ID - 7 F B State 99 Farm Pond, Latah Co., ID - 4 F B Moscow Mountain Pond, Latah Co., ID - 3 F B Poor Man Gulch, Latah Co., ID - 3 F B
Montana Rocky Mountains
Glacier National Park, Glacier Co., MT 6 - Red Rock Lake, Beaverhead Co., MT 5 - Skalkaho Pass (Mud Lake), Granite Co., MT 18 22 E B Missoula, Missoula Co., MT - 18 E B
Yellowstone and Teton Mountains
Crazy Mountains, Park Co., MT - 8 D B Teton National Park, Teton Co., WY 16 - Yellowstone Lake, Yellowstone National Park, Yellowstone Co., WY 15 - Canyon Creek, Yellowstone National Park, Yellowstone Co., WY - 6 D B
San Pitch River, San Pete Co., UT 1 - Provo River, Wasatch Co., UT 20 5 C B Wasatch Mountains, Wasatch Co., UT - 2 C B Jackson Lake, Uinta Co., UT - 3 C B
Utah Western Desert
Snake Valley, Tooele Co., UT(*) 10 - Tule Valley, Millard Co., UT(*) 10 17 A D Ibapah, Tooele Co., UT - 4 A D
Mary's River, Elko Co., NV 15 - Owyhee River, Elko Co., NV 10 - Reese River, Nye Co., NV 15 - Hurry Back Creek, Owyhee Co., ID 12 9 B E Anthony Lake, Baker Co., OR 20 3 B E Maggie Canyon, Elko Co., NV - 13 B E Lower Maggie Creek, Elko Co., NV - 3 B E Humboldt River Valley, Elko Co., NV - 1 B E Lower Annie Creek, Elko Co., NV - 17 B E
Fall City Mills, Shasta Co., CA - 1 K F
Paulina Lake, Deschutes Co., OR 5 - K F Crane Prairie, Lane Co., OR 6 3 K F Gold Lake Bog, Lane Co., OR 10 6 K F Conboy Lake, Klickitat Co., WA(**) 6 - Trout Lake, Klickitat Co., WA(**) 1 1 K F Klamath (Wood River), Klamath Co., OR 1 - Fort Klamath, Klamath Co., OR - 2 K F Little Deschutes River, Deschutes Co., OR - 6 K F
Total 264 288
Mackenzie Pass, Lane Co., OR 5 - Gold Lake, Lane Co., OR 11 - Mount Rainier, Pierce Co., WA - 10 L G Feather River, Plumas Co., CA - 10 L G Deer Flat, Shasta Co., CA - 1 L G Squaw Creek Valley, Shasta Co., CA - 1 L G unnamed Lake near Cliff Lake, Sisson Co., CA - 2 L G
Total 16 24
Rana aurora Loon Lake, Douglas Co., OR 4 -
Rana muscosa San Bernardino Mtns, San Bernardino Co., CA 2 -
* Snake Valley and Tule Valley, Utah, samples were combined for allozyme analysis.
** Conboy Lake and Trout Lake, Washington, samples were combined for allozyme analysis.
Allele frequencies tabulated from all scoreable loci were used to calculate indices of genetic distance and similarity between samples (Nei 1972, 1978; Rogers 1972). UPGMA (Sokal and Sneath 1963) and Fitch-Margoliash (Fitch and Margoliash 1967) phenograms were produced from Nei's genetic distance, [D.sub.N], and Rogers' genetic distance, [D.sub.R]. Lessa (1990) has noted the drawbacks of hierarchical clustering techniques such as UPGMA, proposing instead the use of multidimensional scaling (MDS) analyses to detect patterns that cannot easily be discerned with dendrograms. MDS plots, in general, are consistent with phenograms but can clarify the placement of intermediates. Accordingly, we subjected the matrices of [D.sub.N] and [D.sub.R] to MDS in two dimensions using SYSTAT 5.02 software.
To examine the relationship between genetic diversification of populations and their geographic relationships, a measure of genetic differentiation can be regressed against map distance between populations. [D.sub.N] has been used before in comparison with map distance (Green 1983; Slatkin 1993) but, as pointed out by Slatkin (1993), [D.sub.N] is dependent upon mutation rates which may vary between loci. By contrast, Wright's (1943) [F.sub.ST] does not depend upon mutation rate so long as mutation rates are small and Slatkin (1993) defined another measure, [Mathematical Expression Omitted], derived from [F.sub.ST] and calculable from allozyme data, to indicate the amount of gene flow (and thus, inversely, the degree of isolation) between two populations. Slatkin's [Mathematical Expression Omitted], rather than [F.sub.ST], can be regressed against map distance for all pairwise comparisons of populations. A significant negative slope of [Mathematical Expression Omitted] against map distance indicates isolation-by-distance.
To test for differences in conformity to isolation-by-distance in the expanding and receding parts of the range of R. pretiosa, populations were sorted into northern and southern groups. Interspecific comparisons (see Results below) were not considered. The northern subset of 16 populations consisted of those from Washington State, northern Idaho, Montana, Wyoming, British Columbia and Alaska. Southern populations were those from the Great Basin, the Utah Western Desert, and the Wasatch Mountains. Regressions of genetic distance and isolation ([D.sub.N] and [Mathematical Expression Omitted]) against map distances measured from 1:6,000,000 scale maps of northwestern North America (conic projection) were computed separately for all northern populations and all southern populations. Like [D.sub.N], [Mathematical Expression Omitted] values were computed using the allele frequency results from all loci. The significance of the regression coefficients, r, was tested by a randomization test because the variables could not be assumed to be independent. The distribution of r values obtained after 1000 regressions randomizing the dependent variable ([D.sub.N] or [Mathematical Expression Omitted]) was tested against the actual regression coefficient by deriving a two-tailed value of p.
For morphometric analysis, 288 fluid-preserved museum specimens of R. pretiosa from 38 collecting localities and 24 specimens of Rana cascadae from six localities were examined [TABULAR DATA FOR TABLE 2 OMITTED] (Table 1). A number of the specimens used for allozyme study were also used for morphometric analysis, but many were too small to be of use. The sex of each frog and 20 length measurements (Table 2) were recorded. All measurements were made by one person (JK) and were taken to the nearest 0.1 mm using dial calipers. Raw morphometric data were log-transformed and compared to snout-vent length to ensure isometry of the measurements. Data were tested for normality using normal probability plots, residual plots, and serial correlation plots. Principal components analysis (PCA) and multiple discriminant function analysis (MDA) were performed using log-transformed data in variance-correlation matrices using SYSTAT. Sexes were treated separately as well as combined. PCA was used as an exploratory device to determine the minimum number of informative variables (factor plots; Wilkinson et al. 1992) and to obtain preliminary specimen groupings for MDA. Groupings (Table 1) were subsequently refined based on geographic proximity and allozyme results. MDA was used to compute discriminant scores and posterior probabilities for each individual, standardizing data within each group. Pair-wise, between-sample Mahalanobis distances produced by MDA were used to compute MDS plots in two dimensions to discern logical subdivisions.
Thirty-one enzyme loci were resolved. Five of these, CK-2, GDTH, GPI-2, IDH-2, and TPI, were monomorphic in all samples and species. Among samples of R. pretiosa, a further six loci were monomorphic: CA-2, GP-1, LDH-2, MDH-2, PEP, and PGDH (Table 3). Of the 20 remaining polymorphic [TABULAR DATA FOR TABLE 3 OMITTED] [TABULAR DATA FOR TABLE 4 OMITTED] loci, six of them, amounting to 19.4% of all loci examined and 28.6% of loci polymorphic within R. pretiosa, provided evidence of differentiation at the species level [ILLUSTRATION FOR FIGURE 2 OMITTED].
Fixed differences in ADH, FBA, IDDH, LDH-1, MDH-1, and MDHP (Table 3) completely differentiated the samples from western Oregon (Gold Lake, Paulina Lake, Crane Prairie, and Klamath Lake) and southwestern Washington State (Conboy Lake and Trout Lake) from all other samples. No other divisions among samples were so clearly differentiated. In addition, two other alleles, AAT-[2.sup.B] and [FUMH.sup.B], were unique to the five R. pretiosa samples from western Oregon and southwestern Washington ([FUMH.sup.B] also occurred in Gold Lake R. cascadae). Analyses of genetic distance (not shown) showed unambiguous distinctions between these two major groups of populations. That they constitute two species within the R. pretiosa complex can be illustrated by plotting [D.sub.N] against map distance (Table 4) for all pairwise population comparisons (Green 1983; D.A. Good, pers. comm.). A lower cluster of points [ILLUSTRATION FOR FIGURE 3A OMITTED] encompasses relatively small genetic distances between samples, as would be expected for variation within species, whereas an upper cluster of large genetic distances irrespective of map distance characterizes differences between species. Both a UPGMA phenogram [ILLUSTRATION FOR FIGURE 4A OMITTED] and a two-dimensional MDS plot [ILLUSTRATION FOR FIGURE 6A OMITTED] based on [D.sub.N] clearly showed the western Oregon and southwest Washington State samples to be distinct from all others. In view of this clear differentiation, these western populations will hereafter be referred to as species A whereas all other samples will be referred to as species B in the R. pretiosa complex. The two species were separated by an average [D.sub.N] = 0.278 [+ or -] 0.047 SD, a value comparable to other interspecies distances within the same species group of frogs ([ILLUSTRATION FOR FIGURE 4B OMITTED]; Green 1986).
Within species B, the Utah isolates from the Western Desert at Snake and Tule Valleys and from the Wasatch Range at Provo River constituted the most distinct subdivision, clustering with other populations at the level of [D.sub.N] = 0.154 [ILLUSTRATION FOR FIGURE 4A OMITTED]. The Great Basin populations of Nevada, southwest Idaho, and eastern Oregon clustered at [D.sub.N] = 0.109. All other populations from the Rocky Mountains, British Columbia and northern Washington State formed the largest cluster, with a dichotomy at [D.sub.N] = 0.061 between British Columbia Coast Range populations (including northern Washington State and the Alaska panhandle) and Rocky Mountain populations. The single frog from San Pitch River, in the Wasatch range of Utah, clustered amongst the Rocky Mountain samples and not with the nearby sample from Provo Riven A Fitch-Margoliash tree of [D.sub.R] for species B populations [ILLUSTRATION FOR FIGURE 5 OMITTED] yielded a near identical result except for the clustering of the Provo River and Snake Valley samples and the position of the San Pitch River sample between the Great Basin samples and the northern samples from the Rocky Mountains and the British Columbia Coast Range. The MDS plots based on [D.sub.N] and [D.sub.R] [ILLUSTRATION FOR FIGURE 6 OMITTED] showed the clusterings of population groups within species B irrespective of the hierarchical arrangement imposed by a dendrogram. Rocky Mountain and Northern populations formed a cluster largely distinguishable from the Great Basin samples whereas Snake Valley and Provo River were more distinct.
Southern populations of species B possessed various unique alleles (Table 3). The Snake Valley and Provo River populations alone were fixed for ACOH-[1.sup.C]. The Provo River sample was also fixed for the unique alleles AAT-[1.sup.B] and AAT-[2.sup.C], whereas CK-[1.sup.B] was unique to the Owyhee River population and [FBA.sup.B] was unique to populations from Mary's River and Reese Riven In contrast, northern populations of species B, such as those from Stikine River, Terrace, Prince George, and Manning Park had no unique alleles, although clines in relative frequencies of G3PDH and LDH-1 alleles were evident [ILLUSTRATION FOR FIGURE 2 OMITTED]. Whereas all other populations were fixed for [G3PDH.sup.A] (Table 3), [G3PDH.sup.B] was at 0.250 relative frequency in the Manning Park sample, rising to fixation in the farthest north in Stikine River. LDH-[1.sup.B] similarly increased in frequency from south to north, beginning at a lower latitude, and reaching fixation in British Columbia and Alaska populations.
The highest average heterozygosity ([Mathematical Expression Omitted]) values in populations of species B were found among those nearer the center of the range such as at Ellensburg ([Mathematical Expression Omitted]), Champion Lake ([Mathematical Expression Omitted]), Glacier ([Mathematical Expression Omitted]) and Craig Mountain ([Mathematical Expression Omitted]). Values of [Mathematical Expression Omitted] were lower, on average, in populations toward the northern and southern extremes of the range [ILLUSTRATION FOR FIGURE 7 OMITTED], though with greater variance in the south: [Mathematical Expression Omitted] SD for the three northernmost populations of Stikine River, Terrace, and Prince George; [Mathematical Expression Omitted] SD for the seven southermost populations in Nevada, Utah and southern Idaho.
With the discrimination of R. pretiosa species A from species B, further regressions of [D.sub.N] and [Mathematical Expression Omitted] against map distance were computed for the widespread species B only [ILLUSTRATION FOR FIGURES 3B, 8 OMITTED]. Regressions of these comparisons should have significant positive or negative slopes if there is isolation-by-distance (Slatkin 1993). Using [D.sub.N] against map distance [ILLUSTRATION FOR FIGURE 3B OMITTED], r = 0.570 for all comparisons among northern populations (in the randomization test, [Mathematical Expression Omitted] SD for 1000 randomizations; P [less than] 0.001). The low values of [D.sub.N] indicated a relative genetic homogeneity among the northern populations, despite the significant correlation with map distance. The relationship between [D.sub.N] and map distance for southern populations was not significant (r = 0.248; [Mathematical Expression Omitted] SD for 1000 randomizations; P = 0.196). Regressions of [Mathematical Expression Omitted] against map distance [ILLUSTRATION FOR FIGURE 8 OMITTED] clearly indicated isolation-by-distance among northern populations. There was a significant negative correlation (r = -0.508) of [Mathematical Expression Omitted] and map distance for comparisons between northern populations ([Mathematical Expression Omitted]; P [less than] 0.001). For southern populations, the correlation was not significant (r = 0.018; [Mathematical Expression Omitted]; P = 0.751). The higher values of [Mathematical Expression Omitted] among northern populations indicated genetic disequilibrium and out-crossing that was lacking among southern populations, whose low [Mathematical Expression Omitted] values indicated lack of gene flow.
In contrast with the genetic diversity, detectable patterns of morphological variation in these frogs were weak. Factor plots from PCAs (not shown) indicated that all measured variables were orthogonal to some degree in either or both of the first or second principal components and that the removal of any one variable might detract from the total discriminating power of the MDAs. The presence of juveniles in the complete data set was problematic because most of the outliers in the PCA-generated distribution were juveniles. To try and reduce problems caused by the potentially less reliable measurements obtained from small individuals, and to minimize the introduction of allometry, two restricted subsets of the data were created, one containing only specimens [greater than] 35 mm SVL, based on the SVL of the smallest gravid R. pretiosa female reported in the literature by Turner and Dumas (1972), and the other containing only specimens [greater than] 50 mm SVL, an even more conservative estimate of minimal adult size which obviated any possible influence from allometry.
Multidimensional scaling of Mahalanobis distances between samples (Table 5) revealed the extent of variation present. An MDS plot of all populations and using all specimens (not shown) exhibited considerable scatter, especially among populations considered to be species B. Within this scatter, species A was best discernable from R. cascadae along Dimension 1. Restricting the analysis only to individuals [greater than] 50 mm SVL improved the resolution and R. cascadae, species A, and species B each formed reasonably distinct clusters [ILLUSTRATION FOR FIGURE 9 OMITTED]. However, and because no conclusive groupings were discernable from the preliminary PCA plots, test groupings of samples for MDAs (Table 1) were based on the geography of the collection sites and the results from the allozyme study. Grouping I subdivided species B into subsets of samples whereas Grouping II lumped all species B samples together. Frogs from Puget Sound and Surrey, British Columbia, representing as near as possible the region of the type locality of R. pretiosa, were kept as their own group.
An MDA of the complete data set produced no clear patterns of group separation, with the exception of R. cascadae, whose group assignments by posterior probabilities were statistically correct for all specimens. Despite the observed statistical distinctiveness of all preassigned groups, graphic analysis did not permit a cleanly delineated resolution of the data. A discriminant score (DS) plot of DS2 against DS1, with presumptive groups indicated by morphospace polygons [ILLUSTRATION FOR FIGURE 10A OMITTED], revealed only poor separation of R. cascadae along both the size axis (DS1) and the shape axis (DS2), with considerable overlap for the polygons of the other groups. Reduction of the MDA results to 50% concentration ellipsoids [ILLUSTRATION FOR FIGURE 10B OMITTED] removed the scatter within the data and improved the graphic characterization of presumptive groups. R. cascadae was differentiated without any overlap, as were the specimens from Puget Sound and Surrey, although considerable overlap of ellipsoids for species A and species B remained [ILLUSTRATION FOR FIGURE 10B OMITTED]. The removal of juveniles (all specimens [less than] 50 mm SVL) and resultant recalculation of discriminant scores improved the separation for both group-by-group comparisons (Grouping I in Table 1; [ILLUSTRATION FOR FIGURE 10C OMITTED]) and species-by-species analysis (Grouping II in Table 1; [ILLUSTRATION FOR FIGURE 10D OMITTED]). In both cases, the graphic representations support the perception of morphometric distinctiveness for species A, species B, R. cascadae, and the small series of specimens from the Surrey/Puget Sound region. Groupings were resolved no better or worse when sexes were separated, indicating the absence of significant sexual dimorphisms among the characters examined in this species complex.
The results from R. pretiosa species B satisfy the predicted effects of postglacial range expansion and contraction at the population level. Heterozygosity and numbers of alleles are markedly reduced among northern populations of R. pretiosa [TABULAR DATA FOR TABLE 5 OMITTED] species B, the same phenomenon of genetic uniformity that has been repeatedly observed in northern populations of other organisms affected by glaciation (Sage and Wolff 1986; Mercure et al. 1993; Gill et al. 1993; Zink and Dittmann 1993; Highton and Webster 1976; Cwynar and MacDonald 1987; Critchfield 1984) in North America. The widespread loss of alleles and increased homozygosity of northern populations is certainly not restricted to large mammals (Selander and Kaufman 1973; Sage and Wolff 1986) but is a general phenomenon to be expected from repeated founder events at an advancing edge of a range in any species (Nei et al. 1975). At the same time, unique losses of alleles, genetic equilibrium, and interpopulational diversity resulting from progressive isolation and bottlenecking, as demonstrated in southern populations of R. pretiosa species B, are characteristic of relict populations at a receding edge of a range (Slatkin 1993).
The entire western North American terrestrial biota exhibits effects from range shifts, not just postglacially but repeatedly during glacial advances and retreats over the 1.7 million years that comprise the Quaternary epoch (Pielou 1991; Dansgaard et al. 1993). These shifts are held responsible for the conditions leading to episodic bouts of differentiation and speciation. As with the speciation of chickadees Parus atricapillus and P. carolinensis in western North America (Brewer 1963; Gill et al. 1993), or subspeciation of grasshoppers Chorthippus parallelus ssp. in Europe (Hewitt 1993a), the sequential Illinoian and Wisconsin glacial periods may have promoted the differing levels of divergence seen among the spotted frog populations of the present day. The earlier isolation of proto-R. pretiosa complex populations resulted in the two morphologically cryptic species, A and B. Their differentiation contrasts with the levels of postglacial within-species divergence over the past 8000-10,000 years ([Mathematical Expression Omitted] SD for species A and [Mathematical Expression Omitted] SD for Species B).
Considering the existence of related coastal and interior forms of many western North American small vertebrates (Brewer 1963; Hagmeier and Stults 1964; Good 1989; Good and Wake 1992; Gill et al. 1993), the morphological and genetic differentiation previously noted within R. pretiosa (Thompson 1913; Green 1986) might have been expected to follow a similar pattern. Our morphometric analyses indicate that the species within the R. pretiosa complex are almost indistinguishable. Because previous morphological evidence of differentiation was also inconclusive (Dunlap 1955; Turner 1959a,b 1962), a taxonomic subdivision into R. p. pretiosa, occupying most of the species' range, and R. p. luteiventris, occurring in the Great Basin region (Thompson 1913) was not acknowledged (Stebbins 1985; Turner and Dumas 1972; Nussbaum et al. 1983). This taxonomic distinction is supported only by the clustering of Great Basin populations (cf. R. p. luteiventris) within species B. Complexes of similar-looking, or cryptic, frog species have been previously documented within North temperate Rana, including the R. pipiens complex of leopard frogs (Pace 1974; Hillis 1988) and the large Eurasian R. temporaria group of over 20 outwardly similar "brown frogs" (Mensi et al. 1992; Green and Borkin 1993). Previous morphological studies (Dunlap 1955; Turner 1962; Dumas 1966) therefore should not be faulted for failure to identify clear historical subdivisions within R. pretiosa at the species level. Licht (1975) noted life-history differences between R. pretiosa from Surrey, British Columbia, and from Yellowstone, Wyoming, but attributed them to altitudinal differences in habitat. The dividing line between the ranges of species A and species B is not fully delineated and the nomenclatural identity of these frogs is problematical because the type locality of R. pretiosa Baird and Girard (1853), identified only as "Puget Sound," lies geographically between their known ranges.
The paucity of morphological differentiation accompanying post-Pleistocene speciation of many frogs is in contrast with observations on small mammals, particularly rodents (Martin 1993), or birds such as the song sparrow Melospiza melodia (Zink and Dittmann 1993). Fossil mammalian teeth, which are the most durable part of the skeleton and can be found well preserved, offer evidence of significant historical change in morphology. By comparison, neither frog teeth nor the rest of a frog's skeleton readily fossilize and Tertiary North American frog fossils are comparatively rare (Chantell 1972). Other cryptic or poorly differentiated species of amphibians occur in western North America, such as salamanders in the genera Dicamptodon and Rhyacotriton (Good 1989; Good and Wake 1992). Thus some evolutionary change, including speciation, may go undetected for lack of morphological evidence. In the case of song sparrows, by contrast, detection of morphological races is not accompanied by concomitant patterns of genetic variation (Zink and Dittmann 1993).
The postglacial isolation-by-distance of the northern populations of species B is recognizably a case of primary divergence. At the leading edge of an expanding range, a novel allele that confers a heterozygote disadvantage may nevertheless achieve fixation if restricted to a founder population (Hewitt 1993). Successive founder populations containing this allele would present a barrier to the expansion of populations carrying the competing allele. Thus G3PD[H.sup.B] is fixed in spotted frogs from Stikine River, Alaska [ILLUSTRATION FOR FIGURE 2 OMITTED] and is present only in frogs from the next three closest populations but in diminishing frequency. But there is no evidence of heterozygote deficiency at that locus in these populations. Hewitt (1993) argued that hybrid zones can be established by this type of primary divergence in an expanding range, in contrast with a secondary intergradation model (Endler 1977). But, as Endler (1977) pointed out, distinguishing primary from secondary contact may be virtually impossible on present day evidence alone. Repeated bottlenecking could also provoke subdivision between populations as much as homogeneity within populations (Hewitt 1993) since novel alleles, regardless of their effects upon fitness, may more easily achieve high frequency in repeatedly restricted populations through stochastic sampling error.
The piece-meal contraction of ranges caused by local extinctions and isolation can be viewed as an engine for enhancing [Beta]-diversity - the proliferation of taxa (Thorpe 1984). Taking into consideration the prospective long-term isolation of southern populations of the R. pretiosa complex, it may be reasonable to expect that the subdivisions within species B may actually represent a conglomerate of cryptic, newly independent species. Hewitt (1993) has proposed that the Quaternary glaciations have promoted speciation by periodically pushing the ranges of species toward the south and, in so doing, fragmenting them into refugia. In fact, the present epoch is but a brief interglacial interlude because the normal climatic regime for the Quaternary has been glacial (Dansgaard et al. 1993; Thomson 1993). Therefore, the more pertinent phenomenon to consider is not what glaciation has done to the interglacial biota of North America but the other way around. The ranges of the R. pretiosa complex and other north temperate species will inevitably shift south again at the onset of the next glacial epoch. The fragmentation at the southern edge of the range during the present interglacial range shift to the north may lead to full speciation when the range is once again pushed towards the south.
We thank C. Peterson, D. Darda, J. Kezer, D. Good, F. Schueler, J. Lindell, L. Roberge, M. Hayes, M. Jennings, J. Applegarth, F. Cassirer, the staff of the Champlain, New York, Post Office, and, especially, P. Hovingh for help in providing and procuring specimens. R. Nussbaum and K. McAllister provided information on profitable sites for collecting. We also thank J. Vidum, J. Rosado, S. Orchard, F. Cook, J. Bogart and R. Heyer for loans of preserved specimens. This research has been supported by a research grant from NSERC Canada to DMG.
Baird, S. F., and C. Girard. 1853. Communication. Proceedings of the Academy of Natural Sciences Philadelphia 6:378-379.
Banta, B. H. 1965. A distributional checklist of the recent amphibians inhabiting the state of Nevada. Biological Society of Nevada. Occasional Papers 7:1-14.
Bennett, K. D. 1986. The rate of spread and population increase of forest trees during the postglacial. Philosophical Transactions of the Royal Society, London, B 314:523-531.
Brewer, R. 1963. Ecological and reproductive relationships of Black-capped and Carolina chickadees. Auk 80:9-47.
Brues, C. T. 1932. Further studies on the fauna of North American hot springs. Proceedings of the American Academy of Arts and Sciences 67:185-303.
Case, S. M. 1978. Biochemical systematics of members of the genus Rana native to western North America. Systematic Zoology 27:299-311.
Chantell, C. J. 1970. Upper Pliocene frogs from Idaho. Copeia 1970:654-664.
Critchfield, W. B. 1984. Impact of the Pleistocene on the genetic structure of North American conifers. Pp. 70-118 in R. M. Lanner, ed. Proceedings of the 8th North American Biophysics Workshop. Utah State Univ., Logan.
Cwynar, L. C., and G. M. MacDonald. 1987. Geographical variation in lodgepole pine in relation to population history. American Naturalist 129:463-469.
Dansgaard, W. S., S. J. Johnson, H. B. Clausen, D. Dahl-Jensen, N. S. Gundestrup, C. U. Hammer, C. S. Hvidberg, J. P. Stefenson, A.E. Sveinbjornsdottir, J. Jouzel, and G. Bond. 1993. Evidence for general instability of past climate from a 250-kyr icecore. Nature 364:218-220.
Dumas, P. C. 1966. Studies on the Rana species complex in the Pacific Northwest. Copeia 1966:60-74.
Dunlap, D. G. 1955. Inter- and intraspecific variation in Oregon frogs of the genus Rana. American Midland Naturalist 54:314-331.
Easteal, S. 1988. Range expansion and its genetic consequences in populations of the Giant Toad Bufo marinus. Evolutionary Biology 23:49-84.
Endler, J. A. 1977. Geographic variation, speciation, and clines. Princeton Univ. Press, Princeton, NJ.
Fitch, W. M., and E. Margoliash. 1967. The construction of phylogenetic trees. Science 155:279-284.
Gill, F. B., A. M. Mostrom, and A. L. Mack. 1993. Speciation in North American chickadees: I. Patterns of mtDNA genetic divergence. Evolution 47:195-212.
Gilpin, M. E. 1987. Spatial structure and population viability. Pp. 125-139 in M. E. Soule, ed. Viable populations for conservation. Cambridge University Press, Cambridge.
Gilpin, M. E., and I. Hanski, eds. 1991. Metapopulation dynamics: Empirical and theoretical investigations. Biological Journal of the Linnaean Society 42:1-323.
Good, D. A. 1989. Hybridization and cryptic species in Dicamptodon (Caudata, Dicamptodontidae). Evolution 43:728-744.
Good, D. A., and D. B. Wake. 1992. Geographic variation and speciation in the torrent salamanders of the genus Rhyacotriton (Caudata: Rhyacotritonidae). University of California Publications in Zoology 126:1-91.
Green, D. M. 1983. Allozyme variation through a clinal hybrid zone between the toads Bufo americanus and B. hemiophrys in southeastern Manitoba. Herpetologica 39:28-40.
-----. 1986. Systematics and evolution of western North American frogs allied to Rana aurora and Rana boylii: Electrophoretic evidence. Systematic Zoology 35:283-296.
Green, D. M., and L. J. Borkin. 1993. Evolutionary relationships of eastern palearctic brown frogs, genus Rana: paraphyly of the 24-chromosome species group and the significance of chromosome number change. Zoological Journal of the Linnean Society 109:1-25.
Hagmeier, E. M., and C. D. Stults. 1964. A numerical analysis of the distributional patterns of North American mammals. Systematic Zoology 13:125-155.
Hewitt, G. W. 1993a. Postglacial distribution and species substructure: lessons from pollen, insects and hybrid zones. Pp. 97-123. in D. R. Lees and D. Edwards, eds. Evolutionary patterns and processes. Academic Press, San Diego, CA.
-----. 1993b. After the ice: Parallelus meets Erythropus in the Pyrenees. Pp. 140-164. in Hybrid zones and the evolutionary process. R. G. Harrison, ed. Oxford University Press, Oxford.
Highton, R., and T. P. Webster. 1976. Geographic protein variation and divergence in populations of the salamander Plethodon cinereus. Evolution 30:33-45.
Hillis, D. M. 1988. Systematics of the Rana pipiens complex: Puzzle and paradigm. Annual Review of Ecology and Systematics 19:39-63.
Hoffman, A. A., and M. W. Blows. 1994. Species borders: Ecological and evolutionary perspectives. Trends in Ecology and Evolution 9:223-227.
Hovingh, P. 1993. Aquatic habitats, life history observations and zoogeographic considerations of the spotted frog Rana pretiosa in Tule Valley, Utah. Great Basin Naturalist 53:168-179.
Lessa, E. 1990. Multidimensional analysis of geographic genetic structure. Systematic Zoology 39:242-252.
Licht, L. E. 1975. Comparative life history features of the western spotted frog, Rana pretiosa, from low- and high-level populations. Canadian Journal of Zoology 53:1254-1257.
Linsdale, J. 1940. Amphibians and reptiles in Nevada. Proceeding of the American Academy of Arts and Sciences 73:197-257.
Martin, R. A. 1993. Patterns of variation and speciation in Quaternary rodents. Pp. 226-280 in Morphological change in Quaternary mammals of North America. R. A. Martin and A. D. Barnovsky, eds. Cambridge University Press, Cambridge.
Mensi, P., A. Lattes, B. Macario, S. Salvidio, C. Giacoma, and E. Balletto. 1992. Taxonomy and evolution of European brown frogs. Zoological Journal of the Linnaean Society 104:293-311.
Mercure, A, K. Ralls, K. P. Koepfli, and R. K. Wayne. 1993. Genetic subdivision among small canids: Mitochondrial DNA differentiation of swift, kit, and arctic foxes. Evolution 47:1313-1328.
Murphy, R. W., J. W. Sites Jr., D. G. Buth, and C. H. Haufler. 1990. Proteins I: Isozyme electrophoresis. Pp. 45-126. in D. M. Hillis and C. Moritz, eds. Molecular systematics. Sinauer Associates, Sunderland, MA.
Nagylaki, T. 1976. The decay of genetic variability in geographically structured populations. II. Theoretical Population Biology 10:70-82.
Nei, M. 1972. Genetic distance between populations. American Naturalist 106:283-292.
-----. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89:583-590
Nei, M., T. Maruyama, and R. Chakraborty. 1975. The bottleneck effect and genetic variability in populations. Evolution 29:1-10.
Nevo, E., A. Beiles, and R. Ben Shlomo. 1984. The evolutionary significance of genetic diversity: ecological, demographic and life history correlates. Pp. 13-213 in G. S. Mani, ed. Lecture notes in Biomathematics. Vol. 53. Springer, Berlin.
Nussbaum, R. A., E. D. Brodie Jr., and R. M. Storm. 1983. Amphibians and reptiles of the Pacific Northwest. University of Idaho Press, Moscow.
Pace, A. E. 1974. Systematic and biological studies of the leopard frogs (Rana pipiens complex) of the United States. Miscellaneous Publications of the Museum of Zoology, University of Michigan 148:1-140.
Pielou, E. C. 1991. After the ice age: The return of life to glaciated North America. University of Chicago Press, Chicago.
Pulliam, H. R. 1988. Sources, sinks and population regulation. American Naturalist 132:652-661.
Rogers, J. S. 1972. Measures of genetic similarity and genetic distance. Studies in Genetics VII, University of Texas Publications 7213:145-153.
Sage, R. D., and J. O. Wolff. 1986. Pleistocene glaciations, fluctuating ranges, and low genetic variability in a large mammal (Ovis dalli). Evolution 40:1092-1095.
Sattler, P. W., and S. I. Guttman. 1976. An electrophoretic analysis of Thamnophis sirtalis from western Ohio. Copeia 1976:352-356.
Selander, R. K., and D. W. Kaufman. 1973. Genic variability and strategies of adaptations in animals. Proceedings of the National Academy of Sciences, USA 70:1875-1977.
Slatkin, M. 1993. Isolation by distance in equilibrium and non-equilibrium populations. Evolution 47:264-279.
Sokal, R. R., and P. H. A. Sneath. 1963. Principles of numerical taxonomy. W. H. Freeman, San Francisco, CA.
Stebbins, R. C. 1985. A field guide to western reptiles and amphibians. Houghton-Mifflin Co., Boston, MA.
Tanner, W. M. 1931. A syntoptical study of Utah amphibians. Utah Academy of Sciences 8:159-198.
Tanner, W. W. 1978. Zoogeography of reptiles and amphibians in the Intermontane region. Great Basin Naturalist, Memoirs 2:43-53.
Thompson, H. B. 1913. Description of a new subspecies of Rana pretiosa from Nevada' Proceedings of the Biological Society of Washington 26:53-56.
Thompson, R. S. 1988. Western North America. Pp. 415-458 in B. Huntley and T. Webb, eds. Vegetation History. Kluwer, Dordrecht, Netherlands.
Thomson, K. S. 1993. Northern exposures. American Scientist 81:522-525.
Thorpe, R. S. 1984. Primary and secondary transition zones in speciation and population differentiation: A phylogenetic analysis of range expansion and contraction. Evolution 38:233-243.
Turner, F. B. 1959a. Variation in skeletal proportions of Rana p. pretiosa in Yellowstone Park, Wyoming. Copeia 1961:63-68.
----. 1959b. Pigmentation in the spotted frog, Rana p. pretiosa, in Yellowstone Park, Wyoming. American Midland Naturalist 61:162-176.
-----. 1962. An analysis of geographic variation and distribution of Rana pretiosa. American Philosophical Society Yearbook 1962:325-328.
Turner, F. B., and P. C. Dumas. 1972. Rana pretiosa. Catalogue of American Amphibians and Reptiles 119:1-4.
Wilkinson, L., M. Hill, and E. Varg. 1992. SYSTAT: Statistics, Version 5.2 Edition. Systat, Inc. Evanston, IL.
Wright, S. 1943. Isolation by distance. Genetics 28:139-156.
Zink, R. M., and D. L. Dittmann. 1993. Gene flow, refugia, and evolution of geographic variation in the song sparrow (Melospiza melodia). Evolution 47:717-729.
Museum specimens examined for morphometric analysis. Abbreviations are as follows: USNM (United States National Museum of Natural History [Smithsonian Institution]), NMC (Canadian Museum of Nature), CAS (California Academy of Sciences), UMMZ (University of Michigan Museum of Zoology), RBCM (Royal British Columbia Museum), JPB (James P. Bogart, uncatalogued), DMG (David M. Green, uncatalogued).
Rana pretiosa. British Columbia: Fernie (NMC 981811-23]); McLeod Lake (NMC 8665[1-16, 18-32]); Moose Lake (USNM 48642); Moose River, North Fork (USMN 48627, 48639-40, 48643, 48648-52); Storm Creek Headquarters (RBCM 1855, 1857-59); Surrey (JPB 13251, NMC 258851, 25853, 25854, 25855[1-2], 25856, 258862); Valemount (NMC 17278[1-22]). California: Shasta Co., Fall City Mills (USNM 38806). Idaho: Blaine Co., Muldoon Creek (UMMZ 183905-11); Boise Co., McCall (CAS 45781, 45784, 45788-89, 45793, 45795, 45797); Latah Co. Farm Pond, State Route 99 (UMMZ 133421[a-f]; Latah Co., Moscow Mountain Pond (UMMZ) 133735[a-c]); Latah Co., Poor Man Gulch (UMMZ 133736[a-d]); Owyhee Co., Hurry Back Creek (NMC 30803[1-9]). Montana: Park Co., Crazy Mountains (USNM 60311-13, 60421-25); Granite Co., Mud Lake (NMC 30799[1-17], 30810[1-5]); Missoula Co., Missoula (USNM 205130-47). Nevada: Elko Co., Maggie Canyon (UMMZ 42991, 42996, 43003, 43007, 43011, 43013, 43016-17, 43021, 43024, 43029, 43031. 43035); Elko Co., Lower Maggie Creek (UMMZ 43015, 43027-28); Elko Co., Lower Humboldt River Valley (UMMZ 42992); Elko Co., Lower Annie Creek (UMMZ 42993-94, 42998-99, 43006. 43008-09, 43012, 43014, 43019-20, 43022-23, 43034, 43036-37); Nye Co., Reese River (DMG 4489, 4491, 4493-94); Oregon: Baker Co., Anthony Lake (NMC 30827[3-5]); Deschutes Co., Little Deschutes River (UMMZ 3447-49, 125812[a-c]); Klamath Co., Fort Klamath (CAS 44370-71); Lane Co., Crane Prairie (DMG 4174-4176); Lane Co., Gold Lake Bog (DMG 16886, 1735, 1742-43, 1765, 1769). Utah: Millard Co., Tule Valley (DMG 3670, 3675. 3678, 3680, 3682-83, 3697-3702, 3712-13, 3777-80), Tooele Co., Ibapah (UMMZ 91727[a-d]); Wasatch Co., Provo River (DMG 3978, 3981, 3988-89, 4075); Wasatch Co., Wasatch Mountains (CAS 38574, 38588). Washington: King Co., Seattle (USNM 35638-39); Kittitas Co., Ellensburg (UMMZ 134188[a-b]); Klickitat Co., Trout Lake (USNM 61473); Puget Sound (USNM 310765); Walla Walla Co., Fort Walla Walla (USNM 14498, 310770-74, 310779, 310783); Yakima Co., North Yakima (USNM 45795-96, 46067). Wyoming: Teton Co., Jackson Lake (USNM 48139); Uinta Co., Jackson Lake (USNM 30001-03); Yellowstone Co., Yellowstone National Park, Canyon Creek (USNM 15872, 17617-21).
Rana cascadae. Washington: Pierce Co., Mount Ranier (CAS 30098, 30101, 30106-07, 30118, 30120-21, 30123, 30125, 30127). California: Plumas Co., North Fork Feather River (UMMZ 77904[1-9, 11]); Shasta Co., Deer Flat (USNM 45878); Shasta Co., Squaw Creek Valley (USNM 45877); Sisson Co., Lake near Cliff Lake, Mt. Shasta (USNM 38832-33).
|Printer friendly Cite/link Email Feedback|
|Author:||Green, David M.; Sharbel, Timothy F.; Kearsley, Jennifer; Kaiser, Hinrich|
|Date:||Feb 1, 1996|
|Previous Article:||Resource-associated population subdivision in a symbiotic coral-reef shrimp.|
|Next Article:||The evolution of oviparity with egg guarding and viviparity in lizards and snakes: a phylogenetic analysis.|