Provincialism in plankton: endemism and allopatric speciation in Australian Daphnia.
Growing evidence exists that many zooplankton species have limited distributions. It is accepted that some groups such as the calanoid copepods show more endemism than others, such as the cyclopoid copepods or cladocerans (Hutchinson 1967; Williams 1981; Williams and Allen 1987). Further studies have suggested that these apparent differences in endemism may simply reflect the differing maturity of taxonomic work. A critical taxonomic examination of chydorid cladocerans has suggested, for example, that most taxa with broad distributions are species complexes (Frey 1982a). It remains unclear if the incidence of endemism is linked to subtle variation in dispersal ability. The endemicity of chydorids may result from the attachment of their diapausing eggs onto substrate, in contrast to the floating eggs produced by many other cladoceran genera (Frey 1982a). This example does make clear the need for both a broader examination of endemicity in zooplankton genera and for a more direct assessment of dispersal ability.
Recent genetic studies have provided a more quantitative basis for the estimation of gene flow. Allozyme studies have revealed large gene frequency differences among zooplankton populations from neighboring habitats, which appear to reflect residual founder effects (Hebert 1987; Boileau et al. 1992). Local divergence of this type is, however, of limited evolutionary significance unless habitats persist for sufficient time to allow speciation, which rarely seems to be the case. For example, no evidence exists that temperate zone zooplankters, occupying individual lakes since the Pleistocene deglaciation, have undergone speciation. If speciation does not ordinarily occur in single habitats, then it must occur in lake or pond districts, which might persist for longer time periods than any single habitat and be hundreds to thousands of kilometers from other such districts. However, little effort has been directed towards an examination of the extent of regional gene-pool divergence in freshwater zooplankton. The existing data, which derive almost entirely from the study of populations in recently deglaciated areas, suggest that gene frequency differences are often modest, even on a continental scale (Hebert et al. 1993). The present study builds upon these earlier efforts by examining patterns of gene frequency divergence among members of the genus Daphnia in a more mature landscape - the southeastern quadrant of the Australian continent.
Much of the Australian continent represents a landmass of exceptional stability. The southeastern quarter of the continent shows the most varied topography, but the dominant landscape features have been established for at least 20 million yr (Frakes et al. 1987). The primary topographic feature is the Great Dividing Range, a montane barrier 200 km in width with a peak elevation of 2300 m, which runs parallel to the coast and separates a mesic coastal strip from an arid interior. Climate changes in southeastern Australia have also been modest, and the present dichotomy between an arid interior and a mesic coastal zone was established at least 6.6 M.Y.B.P. (Frakes et al. 1987). During the Pleistocene, the seasonality of climates was enhanced, and snow-lines descended, but permanent ice fields surrounded only the highest peaks (Bowler and Wasson 1983). Rainfall varied as well, but the general pattern of more rain in the coastal than inland zone was sustained.
The presence of a montane barrier coupled with long-term climatic stability might be expected to foster geographic isolation of populations and subsequent speciation. In fact, many groups of Australian organisms show evidence of speciation coincident with the Great Dividing Range, with separation of the fauna into pairs of closely related coastal and interior species (Heatwole 1987). This pattern seems less prevalent in freshwater zooplankton from Australia, where many species are thought to have distributions spanning both the coastal and interior zones. The cladoceran Daphnia carinata s.l. was long thought to be the sole daphniid present in intermittent ponds throughout southeastern Australia (Sars 1914; Bayly and Williams 1973). This species was known to display remarkable morphological variation, which was attributed to phenotypic plasticity (Sars 1914; Grant and Bayly 1981). However, allozyme and morphological studies suggested that this taxon was a complex of at least nine species, one of which included populations reproducing by obligate parthenogenesis (Hebert 1977a, 1981). It was subsequently shown that one species represented [F.sub.1] interspecific hybrids (Hebert 1985), and later allozyme work suggested that taxon diversity had been exaggerated, as Benzie (1988a,b,c) found evidence for the occurrence of only three species. This continuing uncertainty suggests that the D. carinata complex represents a paradigmatic example of the taxonomic confusion that accompanies the joint occurrence of apomixis, hybridization, and phenotypic plasticity.
We aimed both to resolve this taxonomic conflict and to determine parameters important in understanding speciation processes in this complex. A hierarchical approach was employed to disentangle the genetic complexities derived from the coupling of uncertain species boundaries with hybridization and apomixis. The first phase of the analysis focused on genetic comparisons among populations whose genotypic composition approximated that expected for a sexually reproducing taxon. The understanding gained from analysis of these single-species populations was subsequently used to resolve more complex genotypic arrays. Once completed, the analysis provided information not only on species diversity and distributions but also revealed the extent of interspecific hybridization and of regional gene-pool divergence in each species.
Materials and Methods
Pond habitats were sampled for members of the Daphnia carinata complex during January to March of 1990 at 22 sites in New South Wales, Victoria, and the Australian Capital Territory. Sites sampled were Albury (ALB), Armidale (ARM), Bairnsdale (BAI), Berridale (BER), Bombala (BOM), Broken Hill (BRO), Canberra (CAN), Horsham (HOR), Ivanhoe (IVA), Kerang (KER), Mt. Koskciusko (KOS), Maitland (MAI), Mildura (MIL), Moree (MOR), Narrandera (NAR), Numeralla (NUM), Nyngan (NYN), Parkes (PRK), Paroo (PAR), Sydney (SYD), Tibooburra (TIB), and Whittlesea (WHI). Animals were collected from both farm dams and natural ponds using a 400-[[micro]meter] mesh throw net and, when possible, 100 individuals from each pond were frozen in liquid nitrogen immediately after their collection. As well, samples of similar size were preserved in 70% ethanol for subsequent morphological analysis. The most intensive sampling was performed near Canberra and Sydney, where 40 and 16 populations were obtained, respectively. An effort was made to sample ten populations at each of the
remaining sites but, because of drought in some areas and flooding in others, this was not always possible. Prior work has shown that seasonal shifts occur in the abundance of species in the D. carinata complex, but have provided no evidence of complete species replacements during an annual cycle (Hebert 1977b). Our survey undoubtedly failed to reveal the complete species assemblage at some sites as a result of both the limited number of habitats surveyed and seasonal shifts in their abundance. Aside from analysis of members of the D. carinata complex, two populations of the phenotypically distinctive species, Daphnia lumholtzi, were obtained from Centennial Park, Sydney.
Conspicuous variation existed in helmet size and shape among populations from different ponds, and also, in some cases, among individuals from single habitats. As prior work on the D. carinata complex indicated that such variation is important in taxon recognition (Hebert 1977a; Benzie 1988a), an assessment of helmet shape was made prior to allozyme analysis. In many cases, all individuals from a single population shared a similar helmet morphology, and no subdivision was possible. In other populations, two or three distinct helmet sizes (none, small, and large) were present. In these cases, each individual was assigned to one of these categories prior to allozyme analysis. Subsequent to this helmet characterization, 22 individuals from each population were examined, along with reference standards, for variation at ten enzyme loci: aspartate aminotransferase (AAT), aldehyde oxidase (AO), arginine kinase (ARK), fumarate hydratase (FUMH), glucose-6-phosphate isomerase (GPI), lactate dehydrogenase (LDH), malic enzyme (ME), mannose phosphate isomerase (MPI), and two phosphoglucumutase (PGM) loci. Twenty-two other individuals were subsequently analyzed for enzymes found to be polymorphic in the initial run for each population. Hebert and Beaton (1989) provide both E.C. designations for these enzymes and a description of the cellulose acetate gel methodology used for electrophoretic analysis. This technique permitted characterization of single individuals for all ten loci, allowing determination of the number and frequencies of multilocus genotypes in each population.
Genotypic Characterizations of Individual Populations
The genotypic information on each population was summarized by calculating three parameters: the genotypic diversity ratio (GDR), the probability that genotypic frequencies were in Hardy-Weinberg (HW) equilibrium, and the mean fixation index [Mathematical Expression Omitted]. The methods of calculation and rationale for use of the first two parameters are discussed in Hebert et al. (1988). The GDR for each population was calculated by comparing the observed number of genotypes against the mean number expected in a sexual population with the same allele frequencies, based on 100 Monte Carlo simulations. This ratio approximates unity for populations whose genotypic frequencies are both in Hardy-Weinberg and linkage equilibrium, but is much less in obligately asexual populations, multispecies assemblages, or hybrid populations. The extent of deviations from Hardy-Weinberg equilibrium was summarized by calculating the [log.sub.10] probability of the genotypic arrays at each polymorphic locus. The fixation index, F = (He - Ho)/He, varies from a value of -1 for populations that consist solely of heterozygotes to a value of 1 for populations that show polymorphism but lack heterozygotes. For populations polymorphic at several loci, [Mathematical Expression Omitted] represents the arithmetic mean of the individual F-values.
Patterns of Genetic Divergence among Single-Species Populations
Because of the taxonomic uncertainty, it was necessary to employ a genetic criterion for the diagnosis of single-species populations. The initial approach was similar to that employed by Stoddard and Taylor (1988) and involved comparison of the observed number of multilocus genotypes with the expected number in a sexual population derived from 100 simulation runs, using a t-test for comparison of single samples against the corresponding means and variances (Sokal and Rohlf 1981). Populations with significantly fewer (P [less than] 0.05) observed genotypes than expected were temporarily excluded at this stage of the analysis. This approach leads reliably to the exclusion of populations that contain a substantial incidence of obligately asexual forms, interspecific hybrids, or cooccurring species, as in each of these cases the observed number of genotypes is fewer than the number expected in a sexual population with the same allele frequencies. As the number of hybrids declines or the abundance of a second species decreases, the power of the test is reduced. The test is especially weak at detecting those cases in which an inter-specific hybrid occurs at low frequency with one of its parental species. As a guard against Type II errors introduced by these cases, populations containing genotypes heterozygous at five or more of the ten loci were excluded on the presumption that such highly heterozygous genotypes might represent interspecific hybrids. Populations that met both criteria (nonsignificant differences between observed and expected numbers of genotypes and containing genotypes heterozygous at fewer than five of the ten loci) were assumed to represent single-species populations whose genotypic frequencies were in close approximation to Hardy-Weinberg and linkage equilibrium. Additional populations that were homozygous at all ten loci were also assumed to represent single-species populations. It is important to emphasize that, at this phase of the analysis, information on morphological diversity in populations was not used to influence the decisionary process. Single-species populations were designated solely on the basis of their satisfying these genetic criteria. Subsequent to their designation, genetic distances among both these populations and their component genotypes were assessed using Hillis' modification (D*) of Nei's unbiased genetic distance (Hillis 1984). The pattern of genetic distances was then summarized with UPGMA clustering, using a modified version of BIOSYS 1.7 (Swofford and Selander 1989). As the genetic distance analysis indicated the occurrence of eight genetically distinctive groups, the morphological attributes of populations belonging to each group were then compared with those of members of the D. carinata complex (Hebert 1977a). Because of the close correspondence with recognized species, each group was assigned a specific epithet, which is employed throughout the remainder of the paper.
Genetic Analysis of Deviant Populations
The genetic information obtained from single-species populations provided a tool for probing the genetic structure of the deviant populations (those with low genotypic diversity or highly heterozygous genotypes). Specifically, the gene frequency information on each species made possible the determination of both diagnostic alleles and large gene frequency shifts, which were used to identify [F.sub.1] hybrids. The observed genotypic arrays for each species also provided an initial basis for determining the presence in each of the deviant populations of one or more of the taxa detected in single-species populations. Any genotype from a deviant population that was genetically identical to one from a recognized, single-species population was assigned to that taxon. As well, unresolved genotypes that differed from recognized genotypes by only one allele across the ten loci were also grouped with that taxon.
Even after the identification of [F.sub.1] hybrids and the recognition of single species genotypes and their close derivatives, the affinities of several genotypes in the deviant populations remained unclear. To clarify the relationships of these genotypes, metric multidimensional scaling (MDS) was performed for each site with unresolved genotypes. The utility of this technique for genetic distance analysis has been demonstrated by Lessa (1990). Cavalli-Sforza and Edwards' (1967) chord distance was used to summarize genetic divergence among genotypes, as this measure is metric and uses assumptions similar to those for metric MDS algorithms (Wright 1978). To provide an indication of relatedness, genotypes from a specific site with a chord distance of 0.35 or less were encircled following MDS.
Extent of Genetic Divergence among Conspecific Populations
Once the genetic characterization of populations was complete, it was possible to ascertain both the distributional patterns of single species and the extent of genetic divergence among conspecific populations. The extent of variance in gene frequencies (Wright 1978) was determined both among local populations ([F.sub.DS]) and among populations from different sites ([F.sub.ST]). [F.sub.DS] values were calculated for all polymorphic loci for each site where two or more populations of a species were collected. Loci were treated as polymorphic when the frequency of their most common allele was [less than or equal to] 0.99. Five of the species were found at one to three sites, but three other species (carinata, longicephala, and cephalata) were more broadly distributed. A weighted average of [F.sub.ST] across alleles was calculated for each polymorphic locus, and a mean value was then determined for all polymorphic loci.
Genotypic Characterization of Populations
The 187 populations of the Daphnia carinata complex obtained from the 22 collecting sites showed extensive morphological and genetic variation. The three-dimensional representation of the genotypic characteristics of individual populations provides an overview of the extent of this divergence in genotypic arrays. Figure 2A contrasts the genotypic characteristics of a single-species population with those of populations consisting solely of interspecific hybrids or of two noninterbreeding species. The genotypic characteristics of populations in the D. carinata complex showed a V-shaped distribution spanning these idealized states. Many populations possessed the genotypic characteristics of a single species population, but others showed varying levels of heterozygote excess or deficiency. In most cases, these latter populations were also associated with low values of genotypic diversity ratio (GDR) and marked Hardy-Weinberg disturbances.
Patterns of Genetic Divergence in Single-Species Populations
Eighty-nine of the 187 populations contained significantly fewer multilocus genotypes at the [Alpha] = 0.05 level than would be expected in a sexual population. These populations were classified as "deviant" and temporarily excluded from further analysis. Among the other populations, 66 showed no significant difference from expected values, and 32 were monomorphic across all ten loci. Examination of genotypic arrays in these 98 populations showed that genotypes heterozygous at five or more loci occurred in only four populations, which were added to the pool of deviant populations. The remaining 94 populations contained a total of 222 unique genotypes. Individuals within each of these 94 populations shared a similar helmet morphology, whereas 53 of the 93 deviant populations contained individuals with two or three different helmet shapes. Genetic distance analysis on the single-species populations and their component genotypes, followed by UPGMA clustering, showed that populations in the D. carinata complex were separable into eight groups with a genetic distance of 0.20 or greater. A similar pattern was evident at the genotypic level with genetic distances averaging 0.086 among genotypes assigned to a particular group but 0.771 (0.623 excluding Daphnia lumholtzi) among genotypes in different groups. Populations belonging to a specific group invariably showed close morphological similarity, and each corresponded to a different species, with the exception of two clusters, which both possessed morphological features attributed to Daphnia cephalata. Populations in these two groups were allopatric and are hereafter treated as inland/montane (A) and coastal (B) subspecies of D. cephalata. It is worth emphasizing that morphological distinctions among the species in the carinata complex are largely reliant on differences in helmet morphology, although body size is also distinct in Daphnia projecta (Hebert 1977a).
Analysis of Deviant Populations
The initial screening showed that approximately 50% of the populations either contained significantly fewer genotypes than expected in a sexual population or contained some highly heterozygous genotypes. Seventy-eight of these 93 deviant populations contained one or more genotypes that were either identical to or differed by no more than one allele from genotypes in the single-species populations. Deviant populations were also screened for the presence of [F.sub.1] hybrids between each of the possible species pairs, using allelic profiles from single-species populations, which led to the recognition of hybrids in 30 populations. The dual screening process led to the assignment of all genotypes to either a parent species or [F.sub.1] hybrid at 10 sites, but one or more genotypes of unresolved affinity were present at the other 12 sites. By employing metric multidimensional scaling (MDS) to clarify genotypic relationships at each of the 15 sites in [TABULAR DATA OMITTED] which more than one species was present, unidentified genotypes were resolved at all but four sites.
Most genotypes at each of the four coastal sites were readily assigned to known species, although hybrids were present at the two most southerly sites. A single backross genotype was also apparent at Whittlesea, whereas the affinities of one group of four genotypes at Bairnsdale remained unclear following MDS. Patterns of genotypic diversity at two of the montane sites were also straightforward. Only a single species was detected in the alpine grasslands near Mt. Kosciusko, and genotypes at Armidale were readily [TABULAR DATA OMITTED] [TABULAR DATA OMITTED] assigned to two species and their [F.sub.1] hybrids. The situation was much more complex at the other three montane sites. Hybrids between D. carinata and one or more species were common at each site, and the dominant species (thomsoni) at both Bombala and Berridale was itself fragmented into two or three genetically distinct sub-groups. As well, one or more group of genotypes appeared at each site whose affinities were unclear after MDS. Genotypic arrays at the 13 inland sites were much less complicated. Daphnia longicephala was the sole species at six sites and was common at most others although three other species (carinata, cephalata, and projecta) were also detected. Although hybrids between D. carinata and D. longicephala were common at some sites, no inland genotypes remained unassigned following MDS.
None of the seven species in the D. carinata complex were distributed over the whole of southeastern Australia. Three species were more broadly distributed than the rest, with both D. carinata and D. cephalata being found at ten sites, and D. longicephala at 15 sites. The first two species were abundant from coastal sites to low-elevation habitats on both the eastern and western slopes of the Great Dividing Range. Daphnia longicephala was the sole species detected at several inland sites and was common at all sites on the western slopes of the Great Dividing Range but was rare or absent from coastal sites. Three other species were locally abundant, but narrowly distributed. Daphnia projecta was found only at two inland sites in northern New South Wales but was dominant at one site (NYN). Daphnia thomsoni dominated two high elevation sites (BER, BOM) but was otherwise found only at one nearby site (CAN). Daphnia nivalis was the sole daphniid in the highest elevation lakes and ponds (KOS) and was not detected outside these habitats. The remaining species, Daphnia magniceps was obtained at both a coastal and a montane site, but appeared uncommon as it was observed only in a total of five ponds.
TABLE 3. Percentage of individual heterozogosity and frequency of polymorphic loci showing significant (p [less than] 0.05) heterozygote excess or deficit in eight taxa of the Daphnia carinata complex. N, number of unique genotypes recognized for each taxon; Total, total number of polymorphisms examined. Hetero- Heterozygote zygosity Species N (%) Total Excess Deficit carinata 74 .063 90 .19 .03 cephalata A 18 .031 15 .13 .00 cephalata B 15 .030 17 .29 .00 longicephala 98 .076 120 .15 .05 magniceps 1 .000 0 - - nivalis 10 .100 3 .00 .00 projecta 30 .102 23 .13 .13 thomsoni 57 .131 51 .19 .04
Genotypic Characteristics of Taxa
The nature and extent of genotypic diversity in each species was determined by examining all populations in which allozyme data were available for 20 or more individuals. Individual heterozygosities ranged among species from zero in D. magniceps to a high of 0.13 in D. thomsoni. Some populations of the latter species showed signs of polyploidy, as evidenced by the presence of three banded heterozygote phenotypes for loci producing monomeric enzymes. Among the three common species, D. cephalata had a heterozygosity less than half those of D. carinata and D. longicephala. Hardy-Weinberg deviations were common, with 17.2% of polymorphic loci showing a significant heterozygote excess and 4.4% a heterozygote deficit. Variation appeared among species in the pattern of deviation. Populations of D. projecta showed an equal incidence of heterozygote deficit and excess, but [TABULAR DATA OMITTED] the remaining species, excepting D. nivalis for which sample sizes were small, showed a far greater incidence of heterozygote excesses with the highest incidence in coastal populations of D. cephalata.
Genetic Divergence among Conspecific Populations
The extent of gene frequency divergence among local populations within the D. carinata complex was determined whenever two or more polymorphic populations of a species were collected at a site. This analysis indicated the presence of significant gene frequency variation among local populations of all taxa, excepting D. magniceps, which could not be tested as it lacked variation.
The three common species had contrasting patterns of regional gene-pool divergence. Populations of D. carinata showed little divergence with a peak genetic distance of 0.12 noted for two sites represented by single populations (and hence prone to sampling error). By contrast [F.sub.ST] values for both D. longicephala and D. cephalata showed regional differentiation. Populations of D. longicephala from inland NSW and Victoria showed little divergence, but those from two sites along its southern range boundary were distinctive, possessing a mean genetic distance of 0.15 from other populations. Gene frequency divergence was most pronounced among populations of D. cephalata, with the species divided into two major groups showing a mean genetic distance of 0.53. Within each group, there was also evidence of much greater gene frequency divergence between sites than across the entire range of D. carinata.
Patterns in the Incidence of [F.sub.1] Hybrids and Unassigned Genotypes
As allele substitutions were present between each of the species, recognition of [F.sub.1] hybrids was straightforward. Of the 28 combinations possible among the eight taxa, only seven different pairs of taxa were found to cooccur, and hybrids were detected in five of these instances. Daphnia carinata was the common parent in each case and had hybridized with both subspecies of D. cephalata, as well as with D. longicephala, magniceps, and thomsoni. Because hybrids often occurred with one parent or in isolation, the number of habitats containing hybrids was often greater than the number with both parental taxa. The incidence of hybrids varied among sites, with the lowest incidence in inland areas. Four of the five hybrid types were detected near Canberra, but there was never more than one type of hybrid present in a single pond. A marked difference appeared in the incidence of hybrids between D. carinata and the two sub-species of D. cephalata. Only a single hybrid genotype was detected between D. carinata and the coastal subspecies of D. cephalata, despite their frequent cooccurrence. By contrast, many different [F.sub.1] hybrids were detected between the inland/montane form D. cephalata and D. carinata. Hybrids were occasionally detected from sites where one of their parental species appeared absent. Thus, D. carinata was absent from three of six localities with carinata x longicephala hybrids, and the hybrid clone in the Paroo area (PAR) was 300 km from the nearest locality at which D. carinata was collected.
Of the 81 genotypes that were initially unassigned, MDS indicated that 40 were genetically very similar to one of the recognized species or their [F.sub.1] hybrids. Eight other genotypes were intermediate to D. carinata and one of its three different [F.sub.1] hybrids (with cephalata, longicephala, or thomsoni), suggesting that they represented backcrosses. The remaining 33 genotypes showed clear genetic divergence from any known species. All but four of these genotypes occurred at the three southern montane sites and showed closest affinities to the genotypically [TABULAR DATA OMITTED] diverse D. thomsoni, which was also restricted to these sites.
In contrast to most speciation studies that consider groups with a mature taxonomy, this analysis confronted a horrible (sensu Diamond 1992) species complex with a volatile taxonomy. The protocols used to resolve this uncertainty, involving an initial characterization of genetically simple populations, followed by use of the resulting information to probe the origins of genotypic diversity in more complex populations, revealed that three processes were responsible for taxonomic confusion in the Daphnia carinata group. First, analysis of widely separated populations revealed regional gene-pool divergence in two of the three broadly distributed species, and it seems likely that this divergence extends to genes controlling morphology as well as allozymes. Second, although breeding studies were not conducted, genotypic characteristics suggested that populations of two species have made the transition to obligate asexuality. Prior work (Hebert 1981) has shown some obligately parthenogenetic populations of Daphnia cephalata near Sydney, and this study confirmed the prevalence of populations showing fixed heterozygosity in this species. These asexuals do not seem to have arisen via interspecific hybridization as their average heterozygosities are very low. Stronger evidence for obligate asexuality was obtained for Daphnia thomsoni, as some populations showed allozyme phenotypes characteristic of polyploids, and all other known cladoceran polyploids are asexual (Hebert 1987; Weider et al. 1987). Hybridization was the final source of taxonomic complexity, with hybrids between D. carinata s.s. and four other species (magniceps, cephalata, longicephala, and thomsoni) being detected. Hybrids ordinarily occurred at sites where both their parent taxa were observed, but hybrids between D. carinata and D. longicephala at sites in inland New South Wales remote from any known locality for D. carinata, may represent their long-distance dispersal via flooding (Hebert 1977a). The likelihood of such dispersal is supported because Daphnia projecta was also first described from a site near the confluence of the Darling and Paroo Rivers, but later surveys failed to relocate it. Instead, this study shows that D. projecta is abundant in the headwater areas of the Darling River, suggesting its first discovery was linked to the transient establishment of a population. The regular involvement of D. carinata s.s. in hybridization suggests that this species plays a special role in this process. As males are unlikely to be selective in mate choice, it seems probable that female D. carinata lack isolating mechanisms, which those of the other species possess - a possibility that could be confirmed through mitochondrial DNA (mtDNA) analysis (Taylor and Hebert 1993). It is worth noting that the incidence of both hybrids and apomictic forms was greatest at sites on the Great Dividing Range, which were likely most impacted by Pleistocene glaciations.
The present study indicated that no member of the D. carinata complex is distributed over the whole of southeastern Australia. Several species appear to have narrow distributions, with D. projecta restricted to north-central New South Wales, Daphnia nivalis to habitats in alpine grasslands, and D. thomsoni to sites on the Monaro Plateau. Daphnia magniceps appeared to be more broadly distributed, although less so than the other three species (carinata, cephalata, and longicephala). Allozyme studies indicated that the gene pools of these three broadly distributed species varied regionally, although the extent of differentiation was not constant. Populations of Daphnia longicephala throughout inland New South Wales and Victoria showed only minor gene frequency differences, but those at the two most southwestern sites were more divergent. The similarity of inland populations is compatible with the view that gene flow is extensive because of flood-mediated dispersal of resting eggs across the Murray-Darling watershed (Williams 1981). By contrast, populations of D. cephalata were divisible into two genetically distinctive subspecies - one with a coastal and the other with an inland/montane distribution. Allopatric divergence was evident within each subspecies, with populations at Sydney distinct from those at Maitland 100 km to the north or at Bairnsdale 400 km to the south. These results indicate that, in an area of complex topography, local differentiation in daphniid populations can become pronounced. Occupancy of such habitats is not, however, sufficient to ensure divergence, as populations of D. carinata showed much less differentiation than those of D. cephalata, despite their range overlap. This difference seems unlikely to be linked to differences in dispersal ability, as their diapausing eggs are similar in morphology. It might reflect the recent range extension of D. carinata, as a similar lack of gene-pool divergence in areas of complex topography occurs in daphniids from glaciated portions of North America (Hebert et al. 1993).
This study suggests that dispersal is often insufficient to curb local and regional divergence in the gene pools of zooplankton. The apparently broad distribution of some zooplankters is likely ordinarily a consequence of flawed taxonomy - genetic analyses have shown that many taxa are species assemblages whose component species have provincial distributions (Boileau 1991; Hebert and Finston 1993). However, this explanation may not be universal as there are cases of morphologically distinct species such as Daphnia lumholtzi that occur on several continents. Some workers have attributed these broad distributions to vicariance rather than dispersal. The Australasian distribution of the copepod Boeckella triarticulata has for example, been attributed to its origin more than 100 M.Y.A. prior to the fragmentation of Gondwanaland (Bayly and Morton 1979; Maly and Bayly 1991). However, the phenotypic cohesion of populations over this interval in the absence of gene flow seems improbable. It is more likely that such cases are the result of rare, long-distance dispersal events. Upon establishment, the new population is severed from contact with its parental taxon, setting the stage for increasing genetic and phenotypic divergence through time. Species such as B. triarticulata and D. lumholtzi likely reflect recent range extensions, whereas groups showing endemic forms on each continent owe their divergence to more ancient dispersal or vicariance.
This present study has provided insight into factors important in understanding the evolution of the daphniid fauna of southeastern Australia. In contrast to most other zoogeographical regions where at least four to five species groups cooccur, the daphniid fauna in southeastern Australia is dominated by members of a single-species complex. Such low diversity suggests the lack of successful colonization of southeastern Australia by migrants from other biogeographic regions and supports the conclusion that species comprising the carinata complex have arisen in situ. The distributional patterns of these species indicate that the fauna can be separated into inland and coastal or montane elements. This distributional dichotomy parallels that shown by many genera in southeastern Australia and formally recognized by the division of southeastern Australia into the Bassian and Eyrean zoogeographic sub-regions (Littlejohn 1981). It remains unclear if Daphnia species have arisen as a result of physiographic barriers or whether the distributions of species, which have arisen in other fashions, are determined by factors impacted by these barriers. Gene-pool isolation ordinarily seems to be a consequence of divergence in habitat use, as only 5 of 21 possible species pairs were found to cooccur. Shifts in habitat use seem to precede the acquisition of reproductive isolation as four of five cooccurring pairs produced [F.sub.1] hybrids. The slow acquisition of isolating mechanisms may be partially a consequence of the lack of a chromosomal sex-determination mechanism, as studies on other groups indicate that sex-linked genes are often the first to show such divergence (Coyne and Orr 1989). Although [F.sub.1] hybrids were common, they do not now appear to play a key role in gene-pool dynamics. Hybrids between the most genetically divergent species pair (carinata x magniceps) were apparently incapable of sexual reproduction, and although backcross progeny were detected for each of the other three [F.sub.1]s, the presence of diagnostic alleles between all of these species confirms there is little introgression. The selfed progeny of [F.sub.1] hybrids between D. carinata and D. cephalata are known to have low fitness (Hebert 1985), suggesting that genetic divergence among taxa is now sufficient to ensure its protection from erosion by introgression.
Although their extent and consequences are apparent, the origin of genetic differences that now protect the integrity of species in the D. carinata complex are less clear. The local gene-pool divergence noted in each variable species does offer one potential mechanism. If a small percentage of populations are launched as a result of gene frequency shifts during their founding on a trajectory towards reproductive isolation, new species might arise as single populations and subsequently diffuse from their site of origin (Barton 1989). This founder model is particularly attractive as an explanation of daphniid speciation as single unfertilized females can establish a population because of their ability to produce males asexually. However, the occurrence of founder-effect speciation is likely constrained in ponds by a habitat lifespan of less than 1000 yr. Lynch (1985) has argued for the importance of founder-effect speciation, suggesting that its pace may be accelerated by environmental heterogeneity. However, before focusing upon any single mechanism of speciation, it is necessary to evaluate the likelihood of other processes. The importance of allopatric speciation could, for example, be ruled out if regional patterning of gene pools was absent. However, in contrast to earlier presumptions of its unimportance, this study has shown that regional divergence is pronounced in some members of the carinata complex, indicating that local groups of populations do represent an important unit of evolution. Indeed, the extent of divergence among populations of D. cephalata from different watersheds was sufficient to suggest that they represent species in statu nascendi.
This study has established that allopatric divergence may represent an important vehicle for the origin of new species in the carinata complex. The prevalence of regional gene-pool fragmentation now needs to be evaluated in other zooplankton species. In particular, it seems important to ascertain the relationship between variation in the packaging of diapause eggs, which impact the probability of their dispersal (Frey 1982a; Sergeev 1990), and the extent of regional gene-pool divergence. There also exists a need to evaluate the importance of other speciation processes. The prevalence of interspecific hybrids associated with parthenogenesis suggests the likelihood of introgression and reticulate speciation in both cladocerans and rotifers. One recent study (Taylor and Hebert 1993) has established the importance of this mechanism in the genetic divergence of Daphnia galeata from North America and Europe. Taxonomic studies have established the prevalence of species pairs in several zooplankton genera with one taxon restricted to ponds and the other to lakes, a pattern suggesting the importance of disruptive selection in provoking sympatric speciation. Work on chydorid cladocerans (Frey 1982b; Hann 1982) has revealed the prevalence of species pairs showing divergence in the photoperiodic cues that elicit sex, suggesting that habitat-induced phenological shifts lead to assortative mating and speciation in a fashion analogous to the host-induced shifts that seem to be a key agent of sympatric speciation in insects (Tauber and Tauber 1989). Moreover, there appear hints of more unusual mechanisms - some congeneric calanoid copepods show quantum shifts in genome size that are unassociated with polyploidy (McLaren et al. 1966, 1988). Although detailed genetic studies of these cases are required, it seems unlikely that zooplankton species owe their origin to a single mechanism.
This research program was funded by Natural Sciences and Engineering Research Council (NSERC), but the sampling program and allozyme analysis were performed with facilities provided by the Research School of Biological Sciences at the Australian National University. We thank J. Gibson, D. Shaw, A. Wilks, and P. Wilkinson for both access to equipment and stimulating discussions. We thank T. Finston, T. Yankovitch, and P. Gajda for aid in data analysis and P. Forde for the illustrations. M. Boileau, J. Chaplin, D. Taylor, and several anonymous reviewers provided helpful comments on the manuscript.
Barton, N. H. 1989. Founder effect speciation. Pp. 229-256 in D. Otte and J. A. Endler, eds. Speciation and its consequences. Sinauer, Sunderland, Mass.
Bayly, I. A. E., and D. W. Morton. 1979. Aspects of the zoogeography of Australian microcrustaceans. Verhandlungen der Internationalen Vereinigung fur Theoretische und Angewandte Limnologie 20:2537-2540.
Bayly, I. A. E., and W. D. Williams. 1973. Inland waters and their ecology. Longman, Sydney.
Banarescu, P. 1990. Zoogeography of fresh waters, vol. 1. General distribution and dispersal of freshwater animals. AULA, Wiesbaden.
Benzie, J. A. H. 1988a. The systematics of Australian Daphnia (Cladocera: Daphniidae). Species descriptions and keys. Hydrobiologia 166:95-161.
-----. 1988b. The systematics of Australian Daphnia (Cladocera: Daphniidae). Multivariate morphometrics. Hydrobiologia 166:163-182.
-----. 1988c. The systematics of Australian Daphnia (Cladocera: Daphniidae). Electrophoretic analysis of the Daphnia carinata complex, Hydrobiologia 166:183-197.
Boileau, M. G. 1991. A genetic determination of cryptic species (Copepoda: Calanoida) and their postglacial biogeography in North America. Zoological Journal of the Linnean Society 102:375-396.
Boileau, M. G., P. D. N. Hebert, and S. S. Schwartz. 1992. Nonequilibrium gene frequency divergence: persistent founder effects in natural populations. Journal of Evolutionary Biology 4:25-39.
Bowler, J. M., and R. J. Wasson. 1983. Glacial age environments of inland Australia. Pp. 183-208 in J. C. Vogel, ed. Late Cainozoic palaeoclimates of the Southern Hemisphere. Balkema, Rotterdam.
Cavalli-Sforza, L. L., and A. F. W. Edwards. 1967. Phylogenetic analysis and estimation procedures. Evolution 21:550-570.
Coyne, J. A., and H. A. Orr. 1989. Two rules of speciation. Pp. 180-207 in D. Otte and J. A. Endler, eds. Speciation and its consequences. Sinauer, Sunderland, Mass.
Darwin, C. 1859. The origin of species. J. Murray, London.
Diamond, J. M. 1992. Horrible plant species. Nature 360:627-628.
Frakes, L. R., B. McGowran, and J. M. Bowler. 1987. Evolution of Australian environments. Pp. 1-16 in G. R. Dyne and D. W. Walton, eds. Fauna of Australia, vol. 1A. Australian Government Publishing Service, Canberra.
Frey, D. G. 1982a. Questions concerning cosmopolitanism in the Cladocera. Archiv fur Hydrobiologie 93:484-502.
-----. 1982b. Contrasting strategies of gamogenesis in northern and southern populations of Cladocera. Ecology 63:223-241.
Grant, J. W. G., and I. A. E. Bayly. 1981. Predator induction of crests in morphs of the D. carinata King complex. Limnology and Oceanography 26:201-218.
Hann, B. J. 1982. Two new species of Eurycercus (Bullatifrons) from eastern North America (Chydoridae, Cladocera). Taxonomy, ontogeny, and biology. Internationale Revue der Gesamten Hydrobiologie 67:585-610.
Heatwole, H. 1987. Major components and distributions of the terrestrial fauna. Pp. 156-183 in G. R. Dyne and D. W. Walton, eds. Fauna of Australia, vol. 1A. Australian Government Publishing Service, Canberra.
Hebert, P. D. N. 1977a. A revision of the taxonomy of the genus Daphnia (Crustacea: Daphniidae) in south-eastern Australia. Australian Journal of Zoology 25:371-398.
-----. 1977b. Niche overlap among species in the Daphnia carinata complex. Journal of Animal Ecology 46:399-409.
-----. 1981. Obligate asexuality in Daphnia. American Naturalist 117:784-789.
-----. 1985. Interspecific hybridization between cyclic parthenogens. Evolution 39:216-220.
-----. 1987. Genotypic characteristics of the Cladocera, Hydrobiologia 145:183-193.
Hebert, P. D. N., and M. J. Beaton. 1989. Methodologies for allozyme analysis using cellulose acetate electrophoresis. Helena Laboratories, Beaumont.
Hebert, P. D. N., and T. L. Finston. 1993. A taxonomic reevaluation of North American Daphnia (Crustacea: Cladocera) I. The D. similis complex. Canadian Journal of Zoology 71:908-925.
Hebert, P. D. N., S. S. Schwartz, R. D. Ward, and T. L. Finston. 1993. Macrogeographic patterns of breeding system variation in the Daphnia pulex group. I. Breeding systems of Canadian populations. Heredity 70:148-161.
Hebert, P. D. N., R. D. Ward, and L. J. Weider. 1988. Clonal-diversity patterns and breeding-system variation in Daphnia pulex, an asexual-sexual complex. Evolution 42:147-159.
Hillis, D. M. 1984. Misuse and modification of Nei's genetic distance. Systematic Zoology 33:238-240.
Hutchinson, G. E. 1967. A Treatise on limnology, vol. II. Introduction to lake biology and the limnoplankton. John Wiley, New York.
Lessa, E. P. 1990. Multidimensional analysis of geographic genetic structure. Systematic Zoology 39: 242-252.
Littlejohn, M. J. 1981. The Amphibia of mesic southern Australia: a zoogeographic perspective. Pp. 1305-1330 in A. Keast, ed. Ecological biogeography of Australia. Dr. W. Junk, Boston.
Lynch, M. 1985. Speciation in the Cladocera. Verhandlungen der Internationalen Vereinigung fur Theoretische und Angewandte Limnologie 22:3116-3123.
Maly, E. J., and I. A. E. Bayly. 1991. Factors influencing biogeographic patterns of Australian centropagid copepods. Journal of Biogeography 18:455-461.
Mayr, E. 1963. Animal species and evolution. Belknap Press, Cambridge, Mass.
McLaren, I. A., J. M. Sevigny, and C. J. Corkett. 1988. Body sizes, developmental rates, and genome sizes among Calanus species. Hydrobiologia 167/168:275-284.
McLaren, I. A., S. M. Woods, and J. R. Shea. 1966. Polyteny: a source of cryptic speciation among copepods. Science 153:1641-1642.
Nei, M., J. C. Stephens, and N. Saitou. 1985. Methods for computing the standard errors of branching points in an evolutionary tree and their application to molecular data from humans and apes. Molecular Biology and Evolution 2:66-85.
Sars, G. O. 1914. Daphnia carinata King and its remarkable varieties. Archiv fur Mathematik og Naturvidenskab B 34:1-14.
Sergeev, V. 1990. A new species of Daphniopsis (Crustacea: Anomopoda: Daphniidae) from Australian salt lakes. Hydrobiology 190:1-7.
Sokal, R. R., and F. J. Rohlf. 1981. Biometry, 2d ed. W. H. Freeman, New York.
Stoddard, J. A., and J. F. Taylor. 1988. Genotypic diversity: estimation and prediction in samples. Genetics 118:705-711.
Swofford, D. L., and R. K. Selander. 1989. BIOSYS-1: a computer program for the analysis of allelic variation in population genetics and biochemical systematics, Release 1.7. Illinois Natural History Survey, Urbana.
Tauber, C. A., and M. J. Tauber. 1989. Sympatric speciation in insects: perception and perspective. Pp. 307-344 in D. Otte and J. A. Endler, eds. Speciation and its consequences. Sinauer, Sunderland, Mass.
Taylor, D. J., and P. D. N. Hebert. 1993. Habitat dependent hybrid parentage and differential introgression between neighboringly sympatric Daphnia species. Proceedings of the National Academy of Science, USA 90:7079-7083.
Weider, L. J., M. J. Beaton, and P. D. N. Hebert. 1987. Clonal diversity in high arctic populations of Daphnia pulex: a polyploid apomictic complex. Evolution 41:1335-1346.
Williams, W. D. 1981. The Crustacea of Australian inland waters. Pp. 1103-1138 in A. Keast, ed. Ecological biogeography of Australia. Dr. W. Junk, Boston.
Williams, W. D., and G. R. Allen. 1987. Origins and adaptations of the fauna of inland waters. Pp. 184-201 in G. R. Dyne and D. W. Walton, eds. Fauna of Australia, vol. 1A. Australian Government Publishing Service, Canberra.
Wright, S. 1978. Evolution and the genetics of populations, vol. 4. Variability within and among populations. University of Chicago Press, Chicago.
|Printer friendly Cite/link Email Feedback|
|Author:||Hebert, Paul D.N.; Wilson, Christopher C.|
|Date:||Aug 1, 1994|
|Previous Article:||Genetic, acclimatization, and ontogenetic effects on habitat selection behavior in Daphnia pulicaria.|
|Next Article:||Heterozygosity, aggression, and population fluctuations in meadow voles (Microtus pennsylvanicus).|