Niche relationships of clonal and sexual fish in a heterogenous landscape.
The literature concerning unisexual vertebrates has expanded rapidly over the last few decades in response to the growing number of discoveries of unisexual taxa (Dawley 1989) and their usefulness in examining basic principles of ecology and evolutionary biology (Vrijenhoek 1994). The unisexual fish, with their peculiar reproductive mechanisms, have been of particular interest. Most presently known unisexual fish employ one of two modes of sperm-dependent reproduction: gynogenetic systems, where the entire genomic constitution of the mother is inherited intact and sperm is required only to initiate cleavage (Dawley 1989); and hybrid-ogenetic systems, where only the ancestral maternal lineage is clonal and the paternal complement is replaced each generation by sperm from a related sexual species (Schultz 1969, 1977, Vrijenhoek 1984a, b).
From an ecological standpoint, the mere existence of sperm-dependent unisexuals is problematic, since interspecific hybrid origination from sexual ancestors appears to be common to unisexual vertebrates (Dawley 1989). Unlike truly parthenogenetic reptiles capable of colonizing unexploited habitats, most unisexual fish are forced to coexist with one or both parental species, with whom they share considerable genetic and ecological similarity and over whom they enjoy a twofold reproductive advantage due to the cost of sex. Thus a key problem unisexuals face is to coexist with and not outcompete the sexuals, lest their sperm source vanish. The fact that stable unisexual-bisexual coexistence occurs in natural populations implies that these organisms can be successful despite the genetic and ecological challenges they face (Vrijenhoek 1989).
The hypothesized ecological adaptations possessed by unisexual vertebrates that permit their coexistence with sexual forms generally take one of two rather disparate views. The first is that unisexuals are broadly adapted generalists, which may compete extensively with sexuals, but which possess adaptations that allow them to exploit a broad range of habitats and conditions that may serve as a refugium from intense competition. Numerous investigators have noted the somewhat peculiar and occasionally widespread distributions achieved by various unisexual vertebrates (Maslin 1968, 1971, Wright and Lowe 1968, Ghiselin 1974, Uzzell and Darevsky 1975, Cuellar 1977, Lynch 1984). One of the better known explanations for the geographic distribution and ecology of parthenogenetic Cnemidophorus lizards was made by Wright and Lowe (1968), who described them as ecological "weeds." The authors based this conclusion on characteristics of the habitat of Cnemidophorus such as marginal, disclimax, ecotonal, or perpetually disturbed - characteristics resembling those frequently associated with many opportunistic weedy plants (Baker 1965, Gray 1973). Similarly, Maslin (1968, 1971) concluded that the hybrid nature of unisexuals means they will not be particularly well adapted for any single environment but moderately adapted to many, achieving their greatest success in habitats where competition with other species is at its lowest. Moore (1984) noted that in essence, the feature common to each of these interpretations is the suggestion that unisexuals are broadly adapted, and their geographic distributions and local success are determined by competitive interactions with their sexual progenitors. Hence, according to the generalist model, unisexuals would be characterized as possessing a broad fundamental niche capable of exploiting a wide range of habitats and conditions, but competitive interactions would frequently confine their success to underutilized resources and areas to which their sexual hosts are poorly adapted.
Development of the hypotheses that comprise this generalist model occurred prior to the discovery of clonal diversity within many asexual populations. Contrary to early expectations, the application of modern tissue grafting and electrophoretic techniques revealed the vast majority of unisexual populations contain a high level of genetic diversity (reviewed extensively by Moore 1984). Particularly among unisexual fish, clonal diversity frequently takes the form of several clonal lineages coexisting in close proximity to one another (Vrijenhoek 1978, 1979, Schenck and Vrijenhoek 1986). While the above generalist model provides a sufficient explanation for unisexual-bisexual coexistence and is generally consistent with the geographic distributions of many unisexual taxa, it is a more complex task to account for the ability of genetically similar clones to coexist (Vrijenhoek 1994).
Identification of ecological differences (e.g., habitat utilization, trophic morphology, etc.) among coexisting clones in the Poeciliopsis monacha-lucida complex led Vrijenhoek (1979, 1984a) to develop the frozen niche variation model of unisexual-bisexual coexistence. The model employs Roughgarden's (1972) division of niche breadth into within-phenotype and between-phenotype components. The within-phenotype component is a function of resource utilization by any single phenotype, and the between-phenotype component is variation in resource use due to different phenotypes in the population. Vrijenhoek's model suggests that where there exists the opportunity for polyphyletic hybrid origins (i.e., where both sexual species are or have historically been sympatric), variation in the ancestral sexual populations for the between-phenotype component of niche breadth can be "frozen" in the form of genetically and ecologically distinct clones. In a heterogeneous environment, interclonal selection should favor clones with ecologies sufficiently distinct to avoid competition with one another and their sexual progenitors (e.g., Schenck and Vrijenhoek 1989). Such processes, therefore, should lead to highly endemic, locally adapted clones, each specialized to utilize a subcomponent of the available ecological niche space (Vrijenhoek 1979). Since the success of any individual clone is probably limited by its relatively restrictive niche (Bell 1982, Vrijenhoek 1984b), the numerical success of the entire unisexual population depends on the clonal diversity (Vrijenhoek 1979, 1994). Unisexual-bisexual as well as clonal coexistence under this set of predictions is maintained through clones with narrow fundamental niches, and the persistence of asexuality relies very strongly on the opportunity for recurrent hybridization, due to the genetic and ecological inflexibility of such clones and the high probability of extinction (Weeks 1993a).
Our objective was to evaluate the applicability of the generalist vs. frozen-niche variation models of unisexual-bisexual coexistence to a gynogenetic complex of cyprinid fish inhabiting a north-temperate landscape strongly influenced by beaver (Castor canadensis) activity. We examined spatial, physiological, and morphological niche relationships among the members of the complex across a landscape with beaver ponds in various stages of succession. We first used DNA fingerprinting to assess clonal variation in the gynogenetic complex. We then examined the distribution of sexual and gynogenetic forms of the fish across a variety of spatial scales, including: (1) among drainages, (2) among environment types created during beaver pond succession, and (3) within a drainage across physical (littoral vs. pelagic vs. benthic) and chemical (oxygenated vs. unoxygenated) gradients. We also used a transplant experiment to directly assess the tolerance of the sexuals vs. gynogens to severe oxygen stress. In the final portion of the study we examined morphological niche relationships between the hybrid gynogens and their sexual progenitors and assessed whether the presence or absence of gynogens influences trophic morphology of the sexual populations, which might suggest that morphologically mediated competitive interactions are important in structuring the sexual/gynogenetic complex.
THE HYBRID COMPLEX
Northern redbelly dace (P. eos) and finescale dace (P. neogaeus) are found over much of north-central and northeastern North America (Scott and Crossman 1973, Stasiak 1980a, b, Goddard et al. 1989). Both species are commonly found alone and in sympatry, and hybrid gynogens may or may not be present in both cases (Goddard et al. 1989). As in most other clonally reproducing fish (e.g., Schultz 1969, 1977, Echelle et al. 1989, Vasil'ev et al. 1989), asexual reproduction in Phoxinus hybrids occurs via gynogenesis. Diploid clonal females produce unreduced ova containing exact copies of the maternal P. eos-neogaeus genomic complement (Goddard et al. 1989). Sperm from one of the parental species is required only to stimulate development of the embryo, but syngamy is not necessary. The resulting offspring are diploid fish genetically identical to one another and the mother (Dawley et al. 1987, Goddard et al. 1989).
Gynogenesis in Phoxinus hybrids is not 100% efficient, however, and syngamy of the sperm and ova pronuclei occasionally does take place. In such cases, the offspring produced may be a triploid individual carrying the clonal P. eos-neogaeus genome plus a unique set of chromosomes from the sperm, or a somatic mosaic comprised of varying proportions of both clonal diploid and triploid cells (Dawley et al. 1987, Goddard et al. 1989, Goddard and Dawley 1990, Doeringsfeld 1996). Either parental species may serve as the sexual "host" for Phoxinus gynogens, so triploids may carry an additional eos or neogaeus genome. Presently however, only mosaics possessing the extra eos genome in their triploid cells have been identified (Goddard et al. 1989).
The study was conducted on the 294-[km.sup.2] Kabetogama Peninsula in Voyageurs National Park of northern Minnesota, USA [ILLUSTRATION FOR FIGURE 1 OMITTED]. Since the early part of this century, increased beaver activity on the peninsula has transformed the [greater than]300-km existing stream channel (Naiman et al. 1988). Modification of the aquatic ecosystem by beaver has resulted in a spatial and temporal mosaic of habitat complexity associated with age, stage of succession, and local environment of various beaver ponds (Johnston and Naiman 1987, Naiman et al. 1988). Spatial complexity exists both within and among ponds in the form of deep upland vs. shallow lowland ponds, littoral vs. open water zones, and upper aerobic vs. lower anaerobic regions of ponds (Johnston and Naiman 1987). Temporal heterogeneity involves daily and seasonal fluctuations in temperature and oxygen levels, and long-term changes in pond morphology associated with the building and collapse of beaver dams (Naiman et al. 1988).
Five drainages along the southern portion of the Kabetogama peninsula, spanning [approximately]18 km of shoreline distance, were sampled during the study, including Locator, Sucker Creek, Lost Ponds, Shoepack, and East Shoepack [ILLUSTRATION FOR FIGURE 1 OMITTED]. The Lost Ponds drainage was particularly heterogeneous, consisting of a series of four distinct reaches of modified stream channel produced by spatial variation in valley morphology and beaver ponds in various stages of succession. The reaches included (1) a steep-sided upland pond with a maximum depth of 2.3-2.4 m, (2) a shallow (0.8 m) beaver meadow dominated by sedges and cattail (Typha spp.), (3) a moderately deep (1.8 m) area of open water upstream from a collapsed beaver dam that had been partially rebuilt, and (4) a newly created lowland pond of moderate (1.8 m) depth. Because of its high environmental heterogeneity, the Lost Ponds drainage was intensively sampled to establish spatial variation in gynogen and sexual progenitor abundance across physical and chemical gradients within a drainage.
MATERIALS AND METHODS
Fingerprint assessment of clonal variation
Phoxinus eos-neogaeus gynogens from three drainages - Sucker Creek, Lost Ponds, and Shoepack - were assayed for clonal identity and variation using multi-locus, "in gel" DNA fingerprinting. DNA fingerprinting using oligonucleotide probes (Schafer et al. 1988, Tautz 1989) to detect hypervariable variable-number tandem repeat (VNTR) loci has proven an ideal tool for assaying genetic variation in a number of clonal and semiclonal fishes (Turner et al. 1989, 1990). Gynogens to be fingerprinted were identified using pharyngeal tooth counts and intestinal morphology, following Goddard et al. (1989). As previously described in detail (Elder and Schlosser 1995), 464 gynogens from three different habitats (littoral, upper pelagic, lower pelagic) and the four environments (upland pond, collapsed pond/meadow, collapsed pond/dam, and new lowland pond) were fingerprinted from the Lost Ponds drainage. The 15 base pairs (bp) oligonucleotide [(CAC).sub.5] probe (Schafer et al. 1988), which produced hypervariable, multi-locus, VNTR DNA fingerprints, was used for the survey (see Elder and Schlosser 1995). Forty-eight additional gynogens, 24 each from the Sucker Creek and Shoepack drainages, were also fingerprinted, with 12 gynogens collected from two sites in each drainage.
Spatial variation in abundance of gynogens and sexual progenitors
Spatial variation among drainages and among environments associated with beaver pond succession. - Abundance and relative frequency of gynogens and sexual progenitors were determined at 15 different sites located among the five different drainages on the Kabetogama Peninsula, with a minimum of two sites sampled in each drainage. The 15 sites included four different types of environments commonly associated with beaver pond succession; three streams, seven beaver ponds, two collapsed beaver ponds without rebuilt dams, and three collapsed beaver ponds with partially rebuilt dams. Each site was sampled once per year between mid-May and early August in each of three years (1993-1995). Twelve unbaited Gee's minnow traps (40 x 19 cm, 2.5 cm funnel opening, 0.5 cm mesh) were run for 3 d at each site. Traps in pond and collapsed-pond environments were restricted to littoral regions, since previous research indicated that most fish would be captured there (He and Lodge 1990). A random subsample of up to 250-300 dace was collected from each site and categorized as P. eos, P. neogaeus, and hybrid gynogens using pharyngeal tooth counts and intestinal morphology. All dace in the sample were retained if [less than]250 dace were captured. Intestinal and pharyngeal tooth characteristics used to identify hybrids were not size-dependent and were not included in the subsequent analysis of size-related morphological relationships between hybrids and sexual progenitors. In addition, independent measures of biotype identity indicated a high level of accuracy in the ability to segregate hybrid and sexual forms using intestinal and pharyngeal tooth characteristics. Out of 464 individuals identified in the field as hybrid gynogens based on intestinal and pharyngeal tooth characteristics, DNA fingerprinting revealed only four that may have been misidentified, and these may have actually been somatic mosaics (Elder and Schlosser 1995). based on the total number of dace captured and the frequency of gynogens and sexual progenitors in the subsample, a catch-per-unit-effort (CPUE; number captured per trap per day) was calculated for each species/biotype, location, and year.
Spatial variation along physical/chemical gradients within the Lost Ponds drainage. - Abundance and relative frequency of gynogens and sexual progenitors were measured at each of the four sites in the Lost Ponds drainage. Each site was sampled twice per year, spring (mid to late May) and summer (early to mid-August), in each of three years (1993-1995). Twenty-four unbaited minnow traps were set in three different habitats at each of the four sites. Of these, 12 traps were positioned in the littoral zone immediately surrounding and including the beaver dam, 6 traps were suspended near the deepest point of a given site, [approximately]0.25-0.5 m below the water surface in the open-water pelagic zone, and 6 more [approximately]0.25-0.5 m above the bottom in the benthic zone. Traps were run for 3 d during each sampling period. A random subsample of up to 250-300 dace was collected from each zone during each 3-d sampling period, and frequency and catch-per-unit-effort of the gynogens and sexual progenitors were calculated according to the previously described procedure. Concurrent with the fish sampling, we measured vertical profiles of dissolved oxygen at the deepest point in each of the four sites, using a YSI Model 54A DO meter, to associate oxygen concentrations with trap locations. Preliminary sampling indicated that oxygen concentrations exhibited minimal diel fluctuations at the sites (I. Schlosser, unpublished data) so most vertical profiles were measured during time of day 0900-1200.
Survival of gynogens and their sexual progenitors under oxygen stress
To directly assess the ability of gynogens and their sexual progenitors to survive under reduced oxygen levels, we conducted two transplant experiments similar to that described by Rahel (1994). The experiments were conducted in the upland pond of the Lost Ponds drainage from 27 July to 3 August 1994, a period of decreasing oxygen concentration in the benthic environment of the pond. Two sets of experiments were conducted, one from 27 July to 29 July, when oxygen concentration at 2.25 m was extremely low (0.4-0.5 mg/L), and one from 30 July to 3 August, when oxygen concentration was negligible (0.0-0.1 mg/L). Water temperature at 2.25 m during this period was relatively stable, ranging from 15.5 [degrees] to 17.0 [degrees] C. Adult dace were collected daily with minnow traps from the littoral region. Individual dace were placed in a minnow trap in which the openings had been blocked with rubber stoppers. Dace were allowed to acclimate to the trap in the littoral zone for [approximately]24 h prior to the experiment. Twenty traps were then suspended at 2.25 m depth and retrieved every 10 min until mortality occurred. Mortality was indicated by lack of responsiveness by fish to stimuli and flaring of gill covers. Fish were then individually preserved and returned to the laboratory, where they were identified as gynogens or sexual progenitors based on pharyngeal tooth count and intestine morphology. A total of 116 fish were evaluated in the two experiments; sample sizes ranged from 10 to 31 fish for a given species/biotype and oxygen condition. Size of fish (mean length [+ or -] 1 SE) was similar in the two experiments but varied between taxa, with P. eos being the smallest (51.9 [+ or -] 0.6 mm), P. neogaeus the largest (59.1 [+ or -] 0.5 mm), and hybrids intermediate (58.4 [+ or -] 1.1 mm).
Morphological niche relationships between hybrid gynogens and sexual progenitors
Fish used in the analysis of morphological relationships were collected in 1994 during the normal spring sampling of the four sites of the Lost Ponds drainage and an additional upland pond site in the East Shoepack drainage. The upland pond site in the East Shoepack drainage was included because of its similarity in basin morphology to the upland pond in the Lost Ponds drainage, but with much lower gynogen frequency. Each of the five sites was sampled for three consecutive days using a total of 24 unbaited Gee's minnow traps as previously described. A random subsample of fish was fixed in 10% formalin. After several days, fish were transferred to a solution of 20% Caro-Safe (Carolina Biological Supply, Burlington, North Carolina) for preservation. Specimens were returned to the laboratory and identified as P. eos, P. neogaeus, and hybrid gynogens based on intestinal and pharyngeal tooth characteristics.
We focused on growth-related morphometric characteristics associated with head and body morphology, since numerous previous studies indicate that these attributes are potentially important in controlling resource partitioning among fishes (Werner 1977, Schluter 1994). Eleven external morphometric characters were measured for each of 1132 Phoxinus specimens initially included in the study; measurements followed Hubbs and Lagler (1958) unless otherwise noted: standard length; total body mass to the nearest 0.01 g; predorsal length; maximum body depth; maximum body width; caudal peduncle depth; head length; interorbital width; left pectoral fin position; horizontal distance from tip of the snout to a vertical line ascending from the insertion of the pectoral fin; upper jaw length; with the mouth closed, the distance from the anterior margin of the premaxilla to the posterior margin of the junction of the upper and lower jaws, and mouth width; with the mouth closed, the lateral distance between the corners of the mouth. All distance measures were recorded to the nearest 0.1 mm. Upper jaw length, mouth width, and interorbital width were measured with an ocular micrometer at 10x magnification. Remaining measurements were taken with a hand-held dial caliper. Prior to data analysis, all measurements were [log.sub.10] transformed to remove any positive correlations between variance and measurement size (Sokal and Rohlf 1981).
Principal component analysis (PCA) of the 11-character correlation matrix was used to summarize variation in multivariate character space within and among the gynogens and their sexual progenitors. The correlation matrix of log-transformed data was employed throughout the study because of the heterogeneity of measurements included (Neff and Marcus 1980), and because it has been shown to separate size from shape more effectively than the covariance matrix (Somers 1986) in lieu of more in-depth matrix manipulations (Bookstein et al. 1985, Rohlf and Bookstein 1987). Morphometric relationships among Phoxinus species/biotypes were examined by constructing simple bivariate scatter plots of the component scores for each fish on the first two principal component axes. A 95% bivariate density ellipse (Johnson and Wichern 1992) was produced from the component scores for each taxon. The component scores from which the ellipses were calculated are linear combinations of all characters, and therefore represent variation in multivariate character space (Neff and Smith 1979).
In addition, discriminant function analysis (DFA) was used to determine if differences in ecologically relevant morphological characters occurred between populations in the upland pond of the Lost Ponds and East Shoepack drainages in response to disproportionate frequencies of hybrid gynogens in the two drainages. One-half of the fish from the Lost Ponds site was used as a "training set" to develop a discriminant function for classifying fish as P. eos, P. neogaeus, or hybrids. We then evaluated the frequency and patterns of misclassification of the remaining Lost Ponds fish and fish from the East Shoepack site to assess whether morphological changes occurred in the sexual progenitors in response to the change in hybrid gynogen frequency.
Assessment of clonal variation in the gynogens
As previously reported (Elder and Schlosser 1995), all 464 individual gynogens fingerprinted from the Lost Ponds drainage were genetically identical, indicating that a single gynogenetic clone occupied the entire drainage. In addition, all 48 gynogens sampled from the Sucker Creek and Shoepack drainages identically matched the Lost Ponds fingerprint pattern ([ILLUSTRATION FOR FIGURE 2 OMITTED] for a representative comparison). For each sample of 24, a variant clone present at a frequency of 0.12 or greater would have been detected at an [Alpha] [less than or equal to] 0.05, using binomial probabilities of a fingerprint matching the common, Lost Ponds pattern, or not. If both the Sucker Creek and Shoepack samples are pooled, then a variant clone present at a frequency of 0.06 would be expected to be detected at [Alpha] [less than or equal to] 0.05. based on our previous results in the Lost Ponds drainage (Elder and Schlosser 1995) and these more recent results, we conclude that only a single clone of the Phoxinus eos-neogaeus gynogen is present in all three of the Lost Ponds, Sucker Creek, and Shoepack drainages.
Spatial variation in abundance of gynogens and sexual progenitors
Spatial variation among drainages. - Associated with clonal uniformity in the gynogens, strong variation occurred in the abundance (CPUE) of P. eos and frequency of hybrid gynogens among drainages along the Kabetogama Peninsula ([ILLUSTRATION FOR FIGURE 3 OMITTED], left column). Repeated-measures ANOVAs on log-transformed CPUE data revealed no significant (P [greater than] 0.05) year x drainage interaction for each taxon, so data were pooled across years. P. eos exhibited significant ([F.sub.4,40] = 7.67, P [less than] 0.01) variation in CPUE among drainages, being less abundant (Tukey hsd multiple comparisons, P [less than] 0.05) in Sucker Creek, Lost Ponds, and Shoepack than in Locator or East Shoepack [ILLUSTRATION FOR FIGURE 3 OMITTED]. P. neogaeus exhibited no significant differences ([F.sub.4,40] = 0.23, P [greater than] 0.20) in CPUE between drainages, and was at lower CPUE than P. eos in all drainages [ILLUSTRATION FOR FIGURE 3 OMITTED]. Hybrid gynogens exhibited no statistically significant ([F.sub.4,40] = 1.59, 0.10 [less than] P [less than] 0.20) difference in CPUE between drainages, but exhibited a trend of increased CPUE in Sucker Creek, Lost Ponds, and Shoepack relative to Locator and East Shoepack [ILLUSTRATION FOR FIGURE 3 OMITTED]. Associated with the decreased CPUE of P. eos and the tendency for increased CPUE of gynogens in Sucker Creek, Lost Ponds, and Shoepack, the frequency of hybrid gynogens was significantly different (one-way ANOVA with arcsine transformation; [F.sub.4,40] = 7.67, P [less than] 0.001) among drainages and was significantly higher (Tukey hsd multiple comparisons, P [less than] 0.05) in Sucker Creek, Lost Ponds, and Shoepack than in Locator or East Shoepack [ILLUSTRATION FOR FIGURE 3 OMITTED].
Spatial variation among environments associated with beaver pond succession. - Strong spatial variation also occurred in the CPUE of P. eos and frequency of hybrid gynogens among environments associated with beaver pond succession ([ILLUSTRATION FOR FIGURE 3 OMITTED], right column). Repeated-measures ANOVAs on log-transformed CPUE revealed no significant (P [greater than] 0.05) year x environment interaction for each taxon, so data were pooled across years. P. eos exhibited significant variation ([F.sub.3,41] = 3.38, P [less than] 0.05) in CPUE among environment types, being less abundant in collapsed pond or stream environments than active beaver ponds [ILLUSTRATION FOR FIGURE 3 OMITTED]. Phoxinus [TABULAR DATA FOR TABLE 1 OMITTED] neogaeus exhibited no significant difference ([F.sub.3,41] = 1.66, P [greater than] 0.15) in CPUE between environments and was at lower abundance than P. eos in all environments [ILLUSTRATION FOR FIGURE 3 OMITTED]. Hybrid gynogens exhibited no statistically significant difference ([F.sub.3,41] = 1.85, 0.10 [less than] P [less than] 0.20) in CPUE between environments, but exhibited a trend of increased CPUE in collapsed pond or stream environments relative to active beaver ponds [ILLUSTRATION FOR FIGURE 3 OMITTED]. Associated with the decreased abundance of P. eos and the tendency for increased abundance of hybrid gynogens in collapsed pond and stream environments [ILLUSTRATION FOR FIGURE 3 OMITTED], the frequency of hybrid gynogens was significantly different (one-way ANOVA with arcsine transformation; [F.sub.3,41] = 5.65, P [less than] 0.01) among environment types and was higher in collapsed pond and stream environments than in active beaver ponds [ILLUSTRATION FOR FIGURE 3 OMITTED].
Spatial variation of gynogen frequency in environment type classified by drainage. - Gynogen frequencies were also examined for different successional type classified according to drainage (Table 1). Collapsed ponds with and without dams were combined because of limited sample sizes. Strong variation appeared to occur between drainages in types of environment present, suggesting that differences in gynogen frequencies among drainages [ILLUSTRATION FOR FIGURE 3 OMITTED] may be partially due to differences in kinds of environments sampled. Differences in gynogen frequencies also occurred, however, between drainages independently of environment type, especially in ponds (Table 1). Lastly, in all drainages where both ponds and collapsed pond or stream environments were present (Lost Ponds and Shoepack), gynogen frequencies tended to be higher in collapsed ponds or streams relative to ponds (Table 1).
Spatial variation along physical-chemical gradients within the Lost Ponds drainage. - Oxygen concentration exhibited considerable spatial variation at the four sites in the Lost Ponds drainage, associated with changes in water depth, surface area, and riparian vegetation. All four environments exhibited dramatic vertical variation in oxygen concentration, especially during summer [ILLUSTRATION FOR FIGURE 4 OMITTED]. Benthic environments were virtually anoxic in all locations, although the depth at which anoxic conditions occurred varied between years and between sites. Variation between the environments in degree of oxygen depletion in the upper portion of the water column was even more interesting. Oxygen concentrations at the surface (0-0.5 m depth) of the deep upland pond were always [greater than]4-6 mg/L [ILLUSTRATION FOR FIGURE 4 OMITTED]. In contrast, oxygen concentrations in the shallow meadow of the collapsed pond were consistently [less than]2.0 mg/L in summer and in some years (1993) were [less than]1.0 mg/L throughout the entire water column [ILLUSTRATION FOR FIGURE 4 OMITTED].
TABLE 2. Catch-per-unit-effort (CPUE, no. fish per trap per day) and relative capture frequency (in parentheses) of P. eos (EOS), P. neogaeus (NEO), gynogenetic hybrids (HYB), and all dace in four different environments (upland pond, new lowland pond, collapsed pond/dam, and collapsed pond/beaver meadow) and three different habitats (littoral, pelagic, and benthic) in the Lost Ponds drainage. CPUE is based on samples collected during two different seasons (spring and summer) in each of three different years (1993-1995). Total number of trap days (TD) for each environment and habitat is also indicated. Species/biotype All EOS NEO HYB dace TD Environment Upland pond 21.4 4.0 6.1 31.5 488 (0.68) (0.13) (0.19) New lowland 3.9 0.9 1.8 6.6 431 (0.59) (0.14) (0.27) Collapsed pond/dam 6.3 1.4 3.4 11.1 472 (0.57) (0.13) (0.30) Collapsed pond/ 4.5 1.0 5.4 10.9 453 meadow (0.41) (0.09) (0.50) Habitat Littoral 17.1 3.1 6.4 26.6 895 (0.64) (0.12) (0.24) Pelagic 3.8 1.3 3.7 8.8 465 (0.43) (0.15) (0.42) Benthic 0.3 0.1 0.6 1.0 484 (0.30) (0.10) (0.60)
In conjunction with the physical-chemical heterogeneity observed in the Lost Ponds drainage, gynogens and sexual progenitors exhibited dramatic variation in CPUE and relative capture frequency in different environments (Tables 2 and 3). CPUE for P. eos, P. neogaeus, and hybrid gynogens were all largest in the oxygenated (Table 3), littoral environment of the upland pond (Table 2), suggesting this was the preferred habitat for all species/biotypes in the complex. Two primary changes in fish abundance occurred proceeding from the preferred habitat to more marginal environments on the landscape, as represented by the poorly oxygenated, pelagic or benthic environments in the collapsed pond/meadow. First, absolute abundance of P. eos, as indicated by CPUE, decreased, and its associated relative frequency in the complex decreased from 0.60 to 0.70 in the upland pond to 0.30-0.40 in the collapsed pond/meadow. Second, hybrid gynogens became the predominant ([greater than]50%) member of the complex, especially in the poorly oxygenated meadow area of the collapsed pond (Tables 2 and 3).
TABLE 3. Catch-per-unit-effort (CPUE, no. fish per trap per day) and relative capture frequency (in parentheses) of P. eos (EOS), P. neogaeus (NEO), gynogenetic hybrids (HYB), and all dace under different oxygen concentrations in the Lost Ponds drainage. CPUE is based on samples collected during two different seasons (spring and summer) in each of three different years (1993-1995). Total number of trap days (TD) on which the CPUE is based is also indicated. Species/biotype Oxygen (mg/L) EOS NEO HYB All dace TD 7-8 73.54 7.41 12.41 93.36 48 (0.79) (0.08) (0.13) 6-7 31.92 6.60 11.38 49.90 200 (0.64) (0.13) (0.23) 5-6 4.50 1.94 0.19 6.63 36 (0.68) (0.29) (0.03) 4-5 9.33 3.61 4.17 17.11 162 (0.55) (0.21) (0.24) 3-4 9.43 1.73 4.28 15.44 377 (0.61) (0.11) (0.28) 2-3 3.88 0.84 5.11 9.83 162 (0.39) (0.09) (0.52) 1-2 3.80 0.77 4.89 9.46 309 (0.40) (0.08) (0.52) 0-1 0.43 0.09 0.47 0.99 550 (0.43) (0.09) (0.47)
Survival of hybrid gynogens and their sexual progenitors under oxygen stress
In conjunction with the increased relative use of poorly oxygenated environments by hybrid gynogens compared to their sexual progenitors, the hybrid gynogens also exhibited increased survival time under extreme oxygen stress [ILLUSTRATION FOR FIGURE 5 OMITTED]. Significant differences in survival time of gynogens and sexual progenitors occurred under both negligible (0-0.1 mg/L; Kruskal-Wallis one-way ANOVA; H = 17.95, P [less than] 0.001) and very low (0.4-0.5 mg/L; H = 13.86, P = 0.001) oxygen concentrations. Gynogens exhibited the longest survival times in both experiments [ILLUSTRATION FOR FIGURE 5 OMITTED], while P. eos had the shortest survival time and P. neogaeus was intermediate. Furthermore, the gynogens and sexual progenitors exhibited approximately double the survival time in the experiment with very low vs. negligible oxygen levels. The latter observation suggests that changes in oxygen concentration as small as 0.30.4 mg/L can influence the ability of these fish to survive and forage for short periods of time in the organically rich but anoxic benthic environments typically found in beaver ponds.
Morphological niche relationships between gynogens and their sexual progenitors within the Lost Ponds drainage
In light of the extreme clonal uniformity of the gynogens in the Lost Ponds drainage [ILLUSTRATION FOR FIGURE 2 OMITTED], we assessed the morphological relationship between the gynogens and sexual progenitors. PCA of the 11 morphometric variables, when all P. eos, P. neogaeus, and hybrid gynogens from the Lost Ponds were pooled, produced two components summarizing 94% of the variance in the data [ILLUSTRATION FOR FIGURE 6 OMITTED]. All characters loaded positively and nearly equal in magnitude on the first principal component (PC1) suggestive of a component summarizing general size variation (Bookstein et al. 1985). The large amount of variance explained by PC1 reflected the broad range of fish sizes in the data. PC2, which is statistically independent of PC1, was loaded most heavily on two mouth characters, upper jaw length and mouth width [ILLUSTRATION FOR FIGURE 6 OMITTED]. The negative slope of the relationship between PC1 and PC2 [ILLUSTRATION FOR FIGURE 6 OMITTED] strongly suggests that meaningful variation in the data occurred along both axes. Furthermore, the 95% bivariate constant density ellipses for P. eos, P. neogaeus, and the hybrid gynogens indicated that the single gynogenetic clone was morphologically intermediate between the sexual progenitors. The clone filled most of the niche space related to trophic morphology between the longer and wider mouthed P. neogaeus and the shorter and narrower mouthed P. eos [ILLUSTRATION FOR FIGURE 6 OMITTED]. Separate PCA of the upland, collapsed (meadow and dam areas were combined because of small sample sizes), and new lowland pond samples produced patterns similar to those from the pooled sample [ILLUSTRATION FOR FIGURE 6 OMITTED]. This suggests that the morphological relationships between the clone and the sexual progenitors were fairly robust in the Lost Ponds drainage, even though large changes in habitat structure and moderate changes in gynogen frequency, from [approximately]20 to 50%, occurred along the drainage.
Morphological response of the sexual progenitors to disproportionate gynogen abundance in different drainages
Because of the large amount of variation in trophic morphology generated by the gynogenetic clone in the Lost Ponds drainage [ILLUSTRATION FOR FIGURE 6 OMITTED], and because of the apparent preference of all members of the complex for littoral/oxygenated environments in upland ponds (Tables 2 and 3) and the spatial variation in gynogen frequency between drainages [ILLUSTRATION FOR FIGURE 3 OMITTED], we assessed the morphological responsiveness of the sexual progenitors to large changes in gynogen frequency within upland ponds of different drainages. Specifically, we used DFA to examine whether a trophic release occurred in the morphology of the sexual progenitors in the East Shoepack drainage in response to the near absence of hybrid gynogens.
TABLE 4. Loadings for the 11 morphological variables on the first and second canonical variates of the discriminant function analysis (DFA). The DFA was constructed using one-half (n = 140) of the Lost Ponds fish randomly selected as the "training set." Morphological variable CV1 CV2 Upper jaw length -0.759 -0.195 Mouth width -0.582 -0.349 Interorbital width -0.456 -0.324 Head length -0.368 -0.332 Body depth -0.354 -0.309 Left pectoral fin position -0.354 0.028 Total body mass -0.344 -0.250 Standard length -0.329 -0.228 Predorsal length -0.328 -0.334 Caudal peduncle depth -0.280 0.014 Body width -0.276 -0.464
Hybrid gynogens were present in the East Shoepack upland pond but in significantly lower densities than in the upland pond of the Lost Ponds drainage. In the random sample of fish collected for this morphological analysis, hybrids comprised 6.3% and 36.2% of the total dace community in these two ponds, respectively.
Discriminant function analysis, using one-half (n = 140) of the fish randomly selected from the Lost Ponds site as a "training set," resulted in a first canonical variate most heavily loaded by mouth and head characteristics (Table 4). As in the PCA, upper jaw length and mouth width were the primary variables segregating the gynogens and sexual progenitors in the DFA (Table 4), with head width and head length being of secondary importance. Classification of the remaining fish from the Lost Ponds site revealed little misclassification of gynogens and sexual progenitors when the actual classifications, based on pharyngeal teeth and intestinal characteristics, were compared to the classifications based on the DFA (Table 5). In contrast, at the East Shoepack site a large number (31) and high percentage (12) of fish that were actually P. eos were misclassified as hybrids (Table 5). Chi-square analysis indicated that the classifications for P. neogaeus ([[Chi].sup.2] = 2.16) and the hybrids ([[Chi].sup.2] = 1.79) were not significantly different (P [greater than] 0.05) between the two sites, while the classification for P. eos differed significantly ([[Chi].sup.2] = 192.54, P [less than] 0.001) between the sites. These results suggest P. eos shifted their mouth and head morphology toward those more characteristic of hybrid gynogens when hybrids were at reduced abundance in the East Shoepack pond.
TABLE 5. Actual and predicted classifications of P. eos (EOS), P. neogaeus (NEO), and hybrid gynogens (HYB) at the Lost Ponds and East Shoepack sites. Actual classifications were based on pharyngeal teeth and intestinal characteristics. Predicted classification was based on the discriminant function analysis. Numbers in parentheses are the percentages of the total classified for each biotype/species. Predicted Site Actual EOS NEO HYB Lost Ponds EOS 65 (98) 0 1 (2) NEO 0 24 (96) 1 (4) HYB 2 (4) 0 47 (96) East Shoepack EOS 225 (88) 0 31 (12) NEO 0 52 (100) 0 HYB 0 2 (10) 18 (90)
The shift in P. eos mouth and head morphology was further illustrated by a comparison of first canonical variate scores for the hybrid gynogens and sexual progenitors at each of the sites [ILLUSTRATION FOR FIGURE 7 OMITTED]. P. eos exhibited a significant (Mann-Whitney U, P [less than] 0.0001) change of scores in East Shoepack relative to Lost Ponds, with a tendency for increased upper jaw length and mouth width in the P. eos of East Shoepack. Since all of the characters contributing to variation on the first canonical variate are size-related (Table 4), one could argue that a change in body size of P. eos between the sites was causing the observed shift in covariate scores. A comparison of mean standard length of P. eos at the two sites (East Shoepack [ES] = 44.1 mm, Lost Ponds [LP] = 45.7) revealed, however, no significant difference (Mann-Whitney U, P [greater than] 0.05).
To further investigate this trend, we examined one additional character associated with differences in trophic morphology in Phoxinus. Das and Nelson (1989) noted mouth angle as an important characteristic distinguishing the two parental species; P. eos have relatively upturned mouths, while those of P. neogaeus are less obliquely positioned. Mouth angles in fish from the Lost Ponds and East Shoepack sites showed trends similar to upper jaw length and mouth width, with P. eos exhibiting a shift toward the hybrid form in East Shoepack [ILLUSTRATION FOR FIGURE 8 OMITTED]. P. eos from the East Shoepack population had mouths significantly more horizontal (Mann-Whitney U, P [less than] 0.001) than Lost Ponds P. eos (means: ES = 52.1 [degrees], LP = 58.3 [degrees]). While mouth angles appeared to decrease slightly with increasing fish size (M. Doeringsfeld, personal observation), the lack of difference in the length of P. eos from the two ponds suggests that the more vertical mouth angle in Lost Ponds P. eos was not an effect of smaller fish size. P. neogaeus from East Shoepack had slightly more vertical mouths than those from Lost Ponds (means: ES = 42.1 [degrees], LP = 40.8 [degrees]), although this difference was nonsignificant (Mann-Whitney U, P [greater than] 0.05) and is well within the probable range of measurement error. Mouth angles in the hybrids from the two ponds did not differ significantly (Mann-Whitney U, P [greater than] 0.05; means: ES = 47.6 [degrees], LP = 49.5 [degrees]).
Spatial and physiological niche relationships in the P. eos-neogaeus complex across a heterogenous landscape
Examination of spatial variation in Phoxinus abundances within ponds suggested that sexual progenitors and clonal gynogens did not partition local spatial resources based on microhabitat preferences. Absolute abundances for P. eos, P. neogaeus, and hybrid gynogens were all highest in the oxygenated littoral zone of the upland pond. However, increased frequencies of gynogens in pelagic and benthic zones, along with their greater survival times under oxygen stress, indicate that the gynogenetic clone is more general in its use of marginally suitable habitats and is physiologically more tolerant to anoxic conditions than its sexual progenitors, especially P. eos.
The increased likelihood of the gynogenetic clone to use marginally suitable habitats has been previously predicted by simulation modeling of early clonal "spinoffs" from sexual populations (Weeks 1993a). Weeks (1993a) used a Monte Carlo simulation to predict that the first successful clonal "colonists" of sexual populations are likely to be those occupying marginal habitats. Our data on the Phoxinus complex appear to be in general agreement with Weeks' (1993a) prediction.
Broad tolerances to abiotic conditions have been documented in other unisexual taxa as well. For example, the gynogenetic frog Rana esculenta apparently has greater tolerances to hypoxia when compared with its sexual progenitors (Tunner and Nopp 1979, Semlitsch 1993). Similarly, Bulger and Schultz (1979) identified thermal tolerances superior to either parental species in one of the most widely distributed hybridogenetic clones of Poeciliopsis monacha-lucida (Vrijenhoek 1979). A broad physiological tolerance to such conditions may allow Phoxinus gynogens to utilize partially collapsed ponds and hypoxic zones within ponds as refugia where the parental species, especially P. eos, are less abundant.
Analyses of the distribution of the Phoxinus eos-neogaeus gynogenetic clone at the landscape level indicate that the generality in its use of marginal habitats persists at larger spatial and longer temporal scales. Our results indicate that the spatial distribution of this particular clone includes at least three drainages, spanning [approximately]10-12 km, along the Kabetogama Peninsula. In addition, the increased frequency of the clone relative to sexual forms in collapsed ponds/beaver meadows, to the point of being nearly 50-60% of the assemblage in these areas ([ILLUSTRATION FOR FIGURE 3 OMITTED], Tables 2 and 3), indicates that the clone has extended the generality in its habitat distribution to environments associated with long-term succession in this beaver-dominated landscape (Johnston and Naiman 1987, Naiman et al. 1988). The extended geographical distribution of this single P. eos-neogaeus clone and its very high frequency of occurrence in some regions of the landscape are fundamentally different from the pattern in monoclonal populations of Poeciliopsis. Poeciliopsis clones apparently partition food and spatial resources and rarely achieve proportions higher than 10%, even when only females of the sexual species are considered (Vrijenhoek 1979, 1984a).
Morphological niche breadth and relationships among clonal gynogens and their sexual progenitors
The analysis of trophic morphology of the clonal gynogens and their sexual progenitors revealed a morphological intermediacy of the gynogens, with the gynogens filling most of the morphological space between the sexual progenitors [ILLUSTRATION FOR FIGURES 6-8 OMITTED]. At least three related interpretations can be proposed as to why the single clone could generate sufficient morphological variability to fill the niche space between the sexual forms. First, novel interactions between incompatible gene complexes in the eos and neogaeus genome might produce unusually high levels of developmental flexibility (cf. Wetherington et al. 1989). Secondly, phenotypic responses to fluctuations in environmental conditions might be fundamentally different for sexual vs. clonal genotypes (Parker 1979). Environment-genotype interactions can produce significant effects on growth rate among clonal genotypes raised under different laboratory conditions of diet and temperature (Schultz and Fielding 1989, Wetherington et al. 1989). Whether similar responses in developmental stability occur within a single genotype over environmental gradients remains to be conclusively demonstrated. Phenotypic plasticity is well recognized as a trait common to many opportunistic (weedy) plant species (Baker 1965, Gray 1973) and it might be reasonable to assume that this characteristic could extend to developmental processes in parthenogenetic and gynogenetic vertebrates of hybrid origin as well (Weeks and Quattro 1991, Weeks 1993b). Thirdly, the triploids and diploid-triploid somatic mosaics potentially found in the P. eos-neogaeus gynogenetic complex would undoubtedly increase the morphological variability present in the clone (Doeringsfeld 1996).
The morphological relationships identified in this study among P. eos, P. neogaeus, and their hybrids are generally consistent with previous descriptions of foraging behavior and trophic morphology in the two sexual species (Becker 1983, Cochran et al. 1988, Litvak and Hansell 1990a, b). The relatively small, upturned mouth, elongated digestive tract, and slender pharyngeal teeth of P. eos are indicative of its tendency to graze on filamentous algae, various plant materials, and organic detritus. P. neogaeus, with its larger, more horizontal mouth, robust pharyngeal teeth, and simple s-shaped gut frequently consumes more invertebrate prey. Despite this obvious trophic differentiation, Litvak and Hansell (1990a, b) identified significant dietary overlap and evidence that competition among these two species may be intense. To date no studies have been conducted to directly compare the foraging behavior of Phoxinus hybrids with their parental species. Given their strict morphological intermediacy, extensive morphological variability, and common use of littoral environments, it seems likely that Phoxinus gynogens experience similar competitive interactions with their sexual progenitors. Behavioral factors in addition to morphology are, however, likely to play a role in establishing the ultimate nature and intensity of these interactions (Weeks et al. 1992).
Results from the comparison of the two upland ponds with disproportionate frequencies of hybrid gynogens suggest that the ecological and morphological generalism by the gynogens might translate into significant competition among the gynogens and at least one of the sexual progenitors. Significant changes in the trophic morphology of P. eos occurred in response to the presence/absence of dense populations of the gynogens [ILLUSTRATION FOR FIGURES 7 AND 8 OMITTED]. While the phenomenon of ecological character displacement (Brown and Wilson 1956) has become excessively difficult to document among natural populations of potentially competing species due to the increasingly stringent criteria upon which evidence is evaluated (Grant 1994, Robinson and Wilson 1994, Schluter 1994), it remains a possible explanation for the morphological differences observed in the populations of P. eos. A fundamental limitation of the present study was, however, the lack of replication of sites with clonal presence or absence. Hence, one is uncertain as to whether the shift in P. eos morphology was due to environmental differences between the ponds or was truly a reflection of a release of the sexual species from competition with the hybrid. Further tests of the competitive release hypothesis need to be conducted across a number of different ponds and drainages with disproportionate gynogen abundances.
The role of heterosis vs. natural selection for a generalist clone
There is some disagreement among investigators studying unisexual vertebrates as to whether characteristics such as phenotypic plasticity and broad ecological tolerances in unisexual populations are an immediate result of exceptionally high heterozygosity (i.e., heterosis) or the evolved products of interclonal selection. Unisexual organisms possess some of the highest levels of heterozygosity known among all vertebrates (Moore 1984). based on this knowledge one could argue, as several investigators have done, that heterosis confers a spontaneous fitness advantage to unisexuals (Schultz 1971, 1977, Cole 1975, White 1978). Such a conclusion is plagued by the fact that studies of naturally occurring unisexuals with such characteristics are generally unable to account for the evolutionary history of the unisexual population, and as such cannot distinguish apparent heterotic fitness from traits evolved through interclonal selection (Vrijenhoek 1994). Wetherington et al. (1987) demonstrated under laboratory conditions that artificially synthesized Poeciliopsis hemiclones are frequently inferior to their sexual progenitors for traits affecting Darwinian fitness, providing the strongest argument against spontaneous heterosis as a significant contributor to unisexual fitness.
Given that polyphyletic origins are possible and indeed probable for most unisexual populations (Moore 1984), and that newly generated clones have the potential to vary considerably for ecologically relevant traits (Wetherington et al. 1989), it would be a mistake to assume that selective forces do not play a significant role in the structuring of unisexual populations. The factors determining whether selection favors a clonal lineage with specialized adaptations vs. ecological generalism appear to be a function of the long-term stability of environmental conditions. Lynch (1984) provides a plausible explanation of this relationship:
[E]ach surviving clone must have had a tolerance (positive fitness) to the full range of environmental conditions to which it has been exposed since its incipience. Relatively specialized clones will surely arise (perhaps frequently), but they will only survive as long as the narrow niche to which they are adapted remains available. Therefore, in the long term, clonal selection will promote the evolution of highly generalized (or general-purpose) genotypes, which are characterized by both broad tolerance ranges and low fitness variance for relevant physical, chemical, and biotic gradients.
Hence, the clonal specialist and generalist hypotheses are not necessarily mutually exclusive, though we would hesitate in taking the view insinuated by Lynch (1984) that selection will always promote broadly adapted clones. As Vrijenhoek (1984a) clearly points out, one can imagine that selection can favor a generalist strategy for one suite of traits, while producing a specialized strategy in yet another suite, each pertaining to a different aspect of the gynogen's ecology. One might also imagine that selection may have sequentially favored both strategies during different periods of the gynogen's history. Thus, while the data presented for the Phoxinus clone of this study seems to support the generalist hypothesis, it by no means falsities the clonal specialization hypothesis. If temporal stability in physical-chemical environments and available resources persists for an extended period of time, then a clone that is highly specialized to utilize those resources will clearly be at an advantage over a generalist.
The single Phoxinus eos-neogaeus clone inhabiting the drainages along the Kabetogama Peninsula in Voyageurs National Park is not highly specialized, either in utilization of spatial resources or in body morphology, and is more common in the marginal and stressful environments associated with beaver pond succession. Apparently, the temporal and spatial unpredictability of environmental conditions in these north temperate landscapes, including extreme temporal variation in temperature and oxygen and spatial variation in habitat structure due to beaver pond succession, does not permit the establishment of narrowly adapted clones. Selection under these conditions has instead favored a physiologically flexible, "general-purpose" clonal genotype that occupies a broad ecological niche and may make establishment of additional clones in the same ecosystem difficult.
We thank the staff of Voyageurs National Park, especially Larry Kallemeyn, for logistical support during our field work. Barry Williams, Joe Totman, and Donovan Verrill assisted with field activities. David Reznick, Robert Vrijenhoek, and Steve Weeks made numerous suggestions on an earlier version of the manuscript that greatly improved the quality of the paper. Support for this work was provided by a National Science Foundation Experimental Program to Stimulate Competitive Research grant (STIAA-RII-EPSCoR) to the state of North Dakota.
Baker, H. G. 1965. Characteristics and modes of origin of weeds. Pages 137-172 in H. G. Baker and G. L. Stebbins, editors. Genetics of colonizing species. Academic Press, New York, New York, USA.
Becker, G. C. 1983. Fishes of Wisconsin. University of Wisconsin Press, Madison, Wisconsin, USA.
Bell, G. 1982. The masterpiece of nature: The evolution and genetics of sexuality. University of California Press, Berkeley, California, USA.
Bookstein, E L., B. Chernoff, R. L. Elder, J. M. Humphries, Jr., G. R. Smith, and R. E. Strauss. 1985. Morphometrics in evolutionary biology. Academy of Natural Sciences of Philadelphia Special Publication 15, Philadelphia, Pennsylvania, USA.
Brown, W. L., and E. O. Wilson. 1956. Character displacement. Systematic Zoology 5:49-64.
Bulger, A. J., and R. J. Schultz. 1979. Heterosis and clonal variation in thermal tolerance in unisexual fishes. Evolution 33:848-859.
Cochran, P. A., D. M. Lodge, J. R. Hodgson, and P. G. Knapik. 1988. Diets of synoptic finescale dace, Phoxinus neogaeus, and northern redbelly dace, Phoxinus eos: a reflection of trophic morphology. Environmental Biology of Fishes 22: 235-240.
Cole, C. J. 1975. Evolution of parthenogenetic species of reptiles, Pages 340-355 in R. Reinboth, editor. Intersexuality in the animal kingdom. Springer-Verlag, New York, New York, USA.
Cuellar, O. 1977. Animal parthenogenesis. Science 197:837-843.
Das, M. K., and J. S. Nelson. 1989. Hybridization between northern redbelly dace, Phoxinus eos, finescale dace, Phoxinus neogaeus, and their hybrids in Upper Pierre Grey Lake, Alberta. Canadian Field Naturalist 104:409-413.
Dawley, R. M. 1989. An introduction to unisexual vertebrates. Pages 1-18 in R. M. Dawley and J. P. Bogart, editors. Evolution and ecology of unisexual vertebrates. New York State Museum Bulletin 466, Albany, New York, USA.
Dawley, R. M., R. J. Schultz, and K. A. Goddard. 1987. Clonal reproduction and polyploidy in unisexual hybrids of Phoxinus eos and Phoxinus neogaeus (Pisces: Cyprinidae). Copeia 1987:275-283.
Doeringsfeld, M. R. 1996. Niche relationships and morphology of coexisting northern redbelly (Phoxinus eos) and finescale (P. neogaeus) dace with their hybrids. Thesis. University of North Dakota, Grand Forks, North Dakota, USA.
Echelle, A. A., A. F. Echelle, and D. P. Middaugh. 1989. Evolutionary biology of the Menidia clarkhubbsi complex of unisexual fishes (Atherinidae): origins, clonal diversity, and mode of reproduction. Pages 144-152 in R. M. Dawley and J. P. Bogart, editors. Evolution and ecology of unisexual vertebrates. New York State Museum Bulletin 466, Albany, New York, USA.
Elder, J. F., and I. J. Schlosser. 1995. Extreme clonal uniformity of Phoxinus eos/neogaeus gynogens (Pisces: Cyprinidae) among variable habitats in northern Minnesota beaver ponds. Proceedings of the National Academy of Sciences 92:5001-5005.
Ghiselin, M. T. 1974. The economy of nature and the evolution of sex. University of California Press, Berkeley, California, USA.
Goddard, K. A., and R. M. Dawley. 1990. Clonal inheritance of a diploid nuclear genome by a hybrid freshwater minnow (Phoxinus eos-neogaeus, Pisces: Cyprinidae). Evolution 44:1052-1065.
Goddard, K. A., R. M. Dawley, and T. E. Dowling. 1989. Origin and genetic relationships of diploid, triploid, and diploid-triploid mosaic biotypes in the Phoxinus eos-neogaeus unisexual complex. Pages 268-280 in R. M. Dawley and J. P. Bogart, editors. Evolution and ecology of unisexual vertebrates. New York State Museum Bulletin 466, Albany, New York, USA.
Grant, P. R. 1994. Ecological character displacement. Science 266:746-798.
Gray, J. R. 1973. The origin of weediness. Transactions of the Illinois Academy of Sciences 66:94-98.
Hubbs, C. L., and K. F. Lagler. 1958. Fishes of the Great Lakes region. University of Michigan Press, Ann Arbor, Michigan, USA.
He, X., and D. M. Lodge. 1990. Using minnow traps to estimate fish population size: the importance of spatial distribution and relative species abundance. Hydrobiologia 190:9-14.
Johnson, R. A., and D. W. Wichern. 1992. Applied multivariate statistical analysis. Prentice-Hall, Englewood Cliffs, New Jersey, USA.
Johnston, C. A., and R. J. Naiman. 1987. Boundary dynamics at the aquatic-terrestrial interface: the influence of beaver and geomorphology. Landscape Ecology 1:47-57.
Litvak, M. K., and R. I. C. Hansell. 1990a. A community perspective on the multidimensional niche. Journal of Animal Ecology 59:931-940.
Litvak, M. K., and R. I. C. Hansell. 1990b. Investigation of food habit and niche relationships in a cyprinid community. Canadian Journal of Zoology 68:1873-1879.
Lynch, M. 1984. Destabilizing hybridization, general-purpose genotypes and geographic parthenogenesis. Quarterly Review of Biology 59:257-290.
Maslin, T. P. 1968. Taxonomic problems in parthenogenetic vertebrates. Systematic Zoology 17:219-231.
-----. 1971. Parthenogenesis in reptiles. American Zoology 11:361-380.
Moore, W. S. 1984. Evolutionary ecology of unisexual fishes. Pages 329-398 in B. J. Turner, editor. Evolutionary genetics of fishes. Plenum, New York, New York, USA.
Naiman, R. J., C. A. Johnston, and J. C. Kelly. 1988. Alteration of North American streams by beaver. BioScience 38:753-762.
Neff, N. A., and L. F. Marcus. 1980. A survey of multivariate methods for systematics. American Museum of Natural History, New York, New York, USA.
Neff, N. A., and G. R. Smith. 1979. Multivariate analysis of hybrid fishes. Systematic Zoology 28:176-196.
Parker, E. D., Jr. 1979. Phenotypic consequences of parthenogenesis in Cnimidophorus lizards. I. Variability in parthenogenetic and sexual populations. Evolution 33:1150-1166.
Rahel, F. 1994. Foraging in a lethal environment: fish predation in hypoxic waters of a stratified lake. Ecology 75: 1246-1253.
Robinson, B. T., and D. S. Wilson. 1994. Character release and displacement in fishes: a neglected literature. American Naturalist 144:596-627.
Rohlf, F. J., and F. L. Bookstein. 1987. A comment on shearing as a method for "size correction." Systematic Zoology 36:356-367.
Roughgarden, J. 1972. Evolution of niche width. American Naturalist 196:683-718.
Schafer, R., H. Zischler, and J. T. Epplen. 1988. [(CAC).sub.5] is a very informative oligonucleotide probe for DNA fingerprinting. Nucleic Acids Research 16:5196.
Schenck, R. A., and R. C. Vrijenhoek. 1986. Spatial and temporal factors affecting coexistence among sexual and clonal forms of Poeciliopsis. Evolution 40:1060-1070.
Schenck, R. A., and R. C. Vrijenhoek. 1989. Coexistence among sexual and asexual Poeciliopsis: foraging behavior and microhabitat selection. Pages 39-48 in R. M. Dawley and J. P. Bogart, editors. Evolution and ecology of unisexual vertebrates. New York State Museum Bulletin 466, Albany, New York, USA.
Schluter, D. 1994. Experimental evidence that competition promotes divergence in adaptive radiation. Science 266: 798-801.
Schultz, R. J. 1969. Hybridization, unisexuality and polyploidy in the teleost Poeciliopsis (Poeciliidae) and other vertebrates. American Naturalist 103:605-619.
-----. 1971. Special adaptive problems associated with unisexual fishes. American Zoology 11:351-360.
-----. 1977. Evolution and ecology of unisexual fishes. Evolutionary Biology 10:277-331.
Schultz, R. J., and E. Fielding. 1989. Fixed genotypes in variable environments. Pages 32-38 in R. M. Dawley and J. P. Bogart, editors. Evolution and ecology of unisexual vertebrates. New York State Museum Bulletin 466, Albany, New York, USA.
Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Fisheries Research Board of Canada, Ottawa, Canada.
Semlitsch, R. D. 1993. Asymmetric competition in mixed populations of tadpoles of the hybridogenetic Rana esculenta complex. Evolution 47:510-519.
Sokal, R. R., and F. J. Rohlf. 1981. Biometry. W. H. Freeman, New York, New York, USA.
Somers, K. M. 1986. Multivariate allometry and removal of size with principal components analysis. Systematic Zoology 35:359-368.
Stasiak, R. H. 1980a. Phoxinus eos Cope. Page 336 in D. S. Lee, C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D. E. McAllister, and J. R. Stauffer, editors. Atlas of North American freshwater fishes. North Carolina State Museum of Natural History, Raleigh, North Carolina, USA.
Stasiak, R. H. 1980b. Phoxinus neogaeus Cope. Page 338 in D. S. Lee, C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D. E. McAllister, and J. R. Stauffer, editors. Atlas of North American freshwater fishes. North Carolina State Museum of Natural History, Raleigh, North Carolina, USA.
Tautz, D. 1989. Hypervariability of simple sequences as a general source of polymorphic DNA markers. Nucleic Acids Research 17:6463-6469.
Tunner, H. G., and H. Nopp. 1979. Heterosis in the common European water frog. Naturwissenschaften 66:268-269.
Turner, B. J., J. F. Elder, and T. F. Laughlin. 1989 DNA fingerprinting of fishes: a general method using oligonucleotide probes. DNA Fingerprint News 1:15-16.
Turner, B. J., J. F. Elder and T. F. Laughlin. 1990. Genetic variation in clonal vertebrates detected by simple sequence oligonucleotide probes. Proceedings of the National Academy of Sciences 87:5653-5657.
Uzzek T. M., and I. S. Darevsky. 1975. Biochemical evidence for the hybrid origin of the parthenogenetic species of the Lacerta saxicola complex (Sauria: Lacertidae), with a discussion of some ecological and evolutionary implications. Copeia 1975:204-222.
Vasil'ev, V. P., E. D. Vasil'eva, and A. G. Osinov. 1989. Evolution of a diploid-triploid-tetraploid complex in fishes of the genus Cobitis (Pisces, Cobitidae). Pages 153-169 in R. M. Dawley and J. P. Bogart, editors. Evolution and ecology of unisexual vertebrates. New York State Museum Bulletin 466, Albany, New York, USA.
Vrijenhoek, R. C. 1978. Coexistence of clones in a heterogeneous environment. Science 199:549-552.
-----. 1979. Factors affecting clonal diversity and coexistence. American Zoologist 19:787-797.
-----. 1984a. Ecological differentiation among clones: the frozen niche variation model. Pages 217-231 in K. Wohrmann and V. Loeschke, editors. Population biology and evolution. Springer-Verlag, Berlin, Germany, USA.
-----. 1984b. The evolution of clonal diversity in Poeciliopsis. Pages 399-429 in B. J. Turner, editor. Evolutionary genetics of fishes. Plenum, New York, New York, USA.
1989. Genetic and ecological constraints on the origins and establishment of unisexual vertebrates. Pages 24-31 in R. M. Dawley and J. P. Bogart, editors. Evolution and ecology of unisexual vertebrates. New York State Museum Bulletin 466, Albany, New York, USA.
-----. 1994. Unisexual fish: model systems for studying ecology and evolution. Annual Review of Ecology and Systematics 25:71-96.
Weeks, S. C. 1993a. The effects of recurrent clonal formation on clonal invasion patterns and sexual persistence: a Monte Carlo simulation of the frozen niche-variation model. American Naturalist 141:409-427.
-----. 1993b. Phenotypic plasticity of life-history traits in clonal and sexual fish (Poeciliopsis) at high and low densities. Oecologia 93:307-314.
Weeks, S. C., and J. M. Quattro. 1991. Life-history plasticity under resource stress in a clonal fish (Poeciliidae: Poeciliopsis). Journal of Fish Biology 39:485-494.
Weeks, S. C., O. E. Gaggiotti, R. A. Schenck, K. P. Spindler, and R. C. Vrijenhoek. 1992. Feeding behavior in sexual and clonal strains of Poeciliopsis. Behavioral Ecology and Sociobiology 30:1-6.
Werner, E. E. 1977. Species packing and niche complementarity in three sunfishes. American Naturalist 111:553-578.
Wetherington, J. D., K. E. Kotora, and R. C. Vrijenhoek. 1987. A test of the spontaneous heterosis hypothesis for unisexual vertebrates. Evolution 41:721-731.
Wetherington, J. D., S. C. Weeks, K. E. Kotora, and R. C. Vrijenhoek. 1989. Genotypic and environmental components of variation in growth and reproduction of fish hemiclones (Poeciliopsis: Poecilidae). Evolution 43:635-645.
White, M. J. D. 1978. Modes of speciation. W. H. Freeman, San Francisco, California, USA.
Wright, J. W., and C. H. Lowe. 1968. Weeds, polyploids, parthenogenesis and the geographical and ecological distribution of all-female species of Cnemidophorus. Copeia 1968:128-138.
|Printer friendly Cite/link Email Feedback|
|Author:||Schlosser, Isaac J.; Doeringsfeld, Matthew R.; Elder, John F.; Arzayus, Luis F.|
|Date:||Apr 1, 1998|
|Previous Article:||Density-dependent antipredator tactics and habitat selection in juvenile pollock.|
|Next Article:||Demographic responses to habitat fragmentation: experimental tests at the landscape and patch scale.|