• allozymes;
  • mitochondrial DNA;
  • molecular phylogenies;
  • phenotypic evolution;
  • reticulate evolution


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Despite many ecological and evolutionary studies, the history of several species complexes within the freshwater crustacean genus Daphnia (Branchiopoda, Anomopoda) is poorly understood. In particular, the Daphnia longispina group, comprising several large-lake species, is characterized by pronounced phenotypic plasticity, many hybridizing species and backcrossing. We studied clonal assemblages from lakes and ponds comprising daphnids from several species complexes. In order to reveal patterns of reticulate evolution and introgression among species, we analysed three data sets and compared nuclear, mtDNA and morphological divergence using animals from 158 newly established clonal cultures. By examining 15 nuclear and 11 mitochondrial (12S/16S rDNA) genetic characters (allozymes/restriction enzymes), and 48 morphological traits, we found high clonal diversity and discontinuities in genotypic and morphological space which allowed us to group clones by cytonuclear differentiation into seven units (outgroup D. pulex). In contrast to six groups emerging from nuclear divergence (related to three traditional species, D. cucullata, D. galeata, D. hyalina and three pairwise intermediate hybrids), a seventh group of clones was clearly resolved by morphological divergence: distinct mtDNA haplotypes within one nuclear defined cluster, ‘D. hyalina’, resembled traditional D. hyalina and D. rosea phenotypes, respectively. In other nuclear defined clusters, association between mtDNA haplotype and morphology was low, despite hybridization being bidirectional (reciprocal crosses). Morphological divergence was greatest between young sister species which are separated on the lake/pond level, suggesting a significant role for divergent selection during speciation along with habitat shifts. Phylogenetic analyses were restricted to four cytonuclear groups of clones related to species. mtDNA and nuclear phylogenies were consistent in low genetic divergence and monophyly of D. hyalina and D. rosea. Incongruent patterns of phylogenies and different levels of genetic differentiation between traditional species suggest reticulate evolutionary processes.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Interspecific hybridization among closely related species may result in merging of gene pools, gene flow between species (introgression), formation of new gene complexes (speciation) or reinforcement of reproductive barriers ( Arnold, 1992). If recombinant genotypes establish new populations, or genes introgress, species phylogenies will include reticulate elements ( Wendel et al., 1991 ; Rieseberg, 1995). Although introgression is usually assumed to be rare among animals, several instances of mitochondrial DNA introgression into the nuclear background of another species have been found (e.g. Ferris et al., 1983 ; Wilson et al., 1985 ; Harrison et al., 1987 ; Dowling & DeMarais, 1993). Examples of nuclear, but not mitochondrial, DNA introgression have also been detected (e.g. Taylor & Hebert, 1993a; Vazquez et al., 1994 ). A frequently used approach for detecting reticulate evolution relies on incongruent gene phylogenies in closely related or hybridizing species ( Arnold, 1997). However, incongruence may originate from processes other than reticulate evolution. At some loci, alleles may be shared among lineages owing to recent divergence, or mutation rates may vary among genes and may differ greatly between lineages ( Moore, 1995). Because discrimination among competing hypotheses on phylogenetic incongruence is inherently difficult, complementary information on several independent characters such as nuclear, mitochondrial and morphological markers is required to reveal underlying processes (e.g. Dowling & DeMarais, 1993).

Here, we present morphological and genetic data on species and interspecific hybrids from one of the most diverse Daphnia groups in Europe, the Daphnia longispina group. This group comprises several species complexes known to form hybrids across large geographical areas ( Wolf & Mort, 1986; Gießler, 1987; Hebert et al., 1989 ; 1993b; Taylor & Hebert, 1993a; Schwenk & Spaak, 1995, 1997a) and isolation by distance was detected in some parental species and hybrids ( Gießler, 1997b). Most data exist for the traditional species D. cucullata, D. galeata and D. hyalina which are common in large lakes, whereas little is known from D. longispina and D. rosea which are mainly restricted to pond habitats ( Brooks, 1957; Flößner, 1972). Environmentally induced phenotypic plasticity and the formation of local races have complicated the interpretation of morphological data in studies on systematics and biogeography ( Wesenberg-Lund, 1926; Benzie, 1986; Hrbácek, 1987; Flößner, 1993). Species and hybrids frequently occur in syntopy and comprise highly variable phenotypes. Hybrids usually exhibit intermediate morphology to parental species ( Flößner & Kraus, 1986; Gießler, 1987), but sometimes closely resemble one parental species, implying backcrossing ( Flößner, 1993) or nonadditive inheritance. Genetic data on hybrid swarms support backcrossing towards one parental species ( Spaak, 1996), suggest nonrandom mating of parental species ( Spaak, 1996) and suggest unidirectional hybridization ( Schwenk, 1993). Backcrossing and nuclear, but not mitochondrial, introgression have been reported from many lakes in Europe and North America ( Wolf & Mort, 1986; Hebert et al., 1989 ; Taylor & Hebert, 1993a; Schwenk & Spaak, 1995; Gießler, 1997a). The spectrum of hybrids, backcrossed clones and stabilized introgressants may bias phylogenetic analyses if the number of species-specific markers is too restricted ( Wolf & Mort, 1986; Spaak, 1996). However, studies using a multilocus allozyme approach permitted species and interspecific hybrid discrimination in some species complexes ( Gießler, 1997a). In addition, recent mitochondrial DNA information on many species of Daphnia has provided data on phylogenetic relationships and the age of species complexes ( Schwenk, 1993; Taylor & Hebert, 1993a; Colbourne & Hebert, 1996; Taylor et al., 1996 ).

Although genetic data suggest low levels of backcrossing and introgression, gene pools seem distinct (fixed allozyme loci and high mitochondrial divergence; Schwenk, 1993; Spaak, 1996; Taylor et al., 1996 ; Gießler, 1997a), but recent molecular evidence suggests reticulate evolution among some North American Daphnia species ( Colbourne & Hebert, 1996). A previously assumed monophyletic ‘species’ was identified as a stabilized introgressant with specific morphology. One therefore asks: is reticulate evolution more frequent in Daphnia than previously assumed? To what extent are morphological evolution and genetic differentiation of species associated? We study the genetic and morphological divergence of animals from clonal assemblages of the D. longispina group originating from different lakes and ponds, covering a broad range of morphological phenotypes from several species complexes. In case of multiple species complexes, species delineation is highly problematic because hybrids of different origin might interbreed as well. Based on previous results ( Gießler, 1997a), we use multidimensional scaling analysis on multilocus genotypes to discriminate between nuclear most divergent groups (presumed species) and intermediates. To test for reticulations between nuclear defined species, we analyse whether mitochondrial and nuclear characters result in congruent phylogenies using a large set of genetic markers. We also estimate the morphological divergence between nuclear defined species using a large set of morphological traits.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Sample collection and allozyme electrophoresis

To study the impact of interspecific hybridization and introgression between species of the D. longispina group involving lake species (D. cucullata, D. galeata, D. hyalina) and pond species (D. rosea; cf. Brooks, 1957; Flößner, 1972; Gießler, 1997a) we chose animals from a limited geographical area and analysed a large set of morphological and genetic markers. Individuals were collected from Daphnia assemblages in seven subalpine lakes in southern Germany near Chiemsee, in Meerfelder Mar in western Germany, in a small body of water near Chiemsee (Seeleitensee) and in a small pond near Munich (Ismaning pond, Table 1). In most of the lakes two or more species coexist that are known to hybridize (cf. Woltereck, 1919; Baumbach, 1922; Aurich, 1933; Gessner, 1953; Jacobs, 1977; Seitz, 1980; Müller, 1993; Gießler, 1997b), and thus our samples are supposed to include a large number of hybrids. One hundred and fifty-eight laboratory clones were established and raised under standardized conditions (19 °C; 16 : 8 h light/dark; well-fed with the green alga Ankistrodesmus gracilis) in a common environment. For outgroup comparisons we included one laboratory clone from D. pulex. All 158 clones were assayed by starch or cellulose acetate gel electrophoresis ( Harris & Hopkinson, 1976; Hebert & Beaton, 1989) at 15 polymorphic allozyme loci. Two laboratory clones were used as marker clones for standard allozyme electrophoresis (cf. Gießler, 1997a).

Table 1.    Origin, nuclear MDS-cluster membership (divergence in 15-locus genotypes, cf. Fig. 2) and composite mitochondrial DNA haplotypes (mtDNA, four restriction enzymes) of animals from the D. longispina complex. C = D. cucullata, G D. galeata, H = ‘D. hyalina’, CH = D. cucullata × D. hyalina, CG = D. cucullata × D. galeata, GH = D. galeataבD. hyalina’. From 158 clones (plus D. cucullata from the Netherlands and outgroup D. pulex) involved in nuclear, mtDNA and morphological analyses, only 154 cytonuclear genotypes could be obtained. Numbers in parentheses represent number of clones analysed per cluster and lake, the number of clones with missing mtDNA haplotype (?) are in bold type. / = genotypes not involved in nuclear MDS analysis, n = total number of cytonuclear genotypes analysed per lake. mtDNA haplotypes found within the D. longispina group are labelled according to their occurrence in nuclear defined species clusters and monophyly with species-specific haplotypes as c1, c2, g1, g2, h1h5 (cf. Figs 1 and 2); c: D. cucullata haplotype from the Netherlands, p: D. pulex haplotype. All lakes (lake abbreviations in square brackets) are located in Germany, except Lake Vechten and Institute pond (The Netherlands). Thumbnail image of

Mitochondrial DNA analysis

Individual daphnids were transferred to 1.5-mL reaction tubes containing 100 μL H3-buffer (10 m M Tris-HCl, pH 8.3 at 25 °C, 0.05 M potassium chloride, 0.005% Tween-20 and 0.005% NP-40 (Replitherm Reaction Buffer, Biozym) and 15 μg proteinase K), homogenized with a perspex pestle precisely fitting a reaction tube, ground for ≈1 min and incubated overnight at 50 °C with mild shaking (waterbath). Finally, we irreversibly denatured the proteinase K with 10 min incubation at 95 °C. The homogenate was stored at 4 °C before being used in a PCR reaction. We amplified mtDNA (12S/16S rDNA) of about 1900 bp length, using the conserved primers 12S2 (5′-ATC GTG CCA GCC GTC GCG GTT A-3′) based on a conserved 12S region of D. pulex ( Taylor & Hebert, 1993a) and S2 (5′-GGA GCT CCG GTT TGA ACT CAG ATC-3′), a universal primer for the 16S gene. Amplifications of mtDNA were performed in 45-μL reaction volumes containing 5 μL homogenate, 1× reaction buffer (Boehringer Mannheim, BM), 0.6 Units Taq, 3 m M MgCl2, 0.25 m M of each dNTP and 0.5 μM of each primer. A thermocycler (OmniGene, Hybaid) was run at 93 °C for 2 :30 min, at 55 °C for 1 min, at 72 °C for 2 min (one cycle) and at 93 °C for 1 min, at 55 °C for 1 min and at 72 °C for 2 min (40 cycles). PCR fragments were separated on 2% agarose 1× TBE gels (LE agarose, Biozym). We selected four informative restriction enzymes (Dde I, Hinf I, Rsa I and Taq I, New England Biolabs, Table 2, cf. Schwenk, 1997) to analyse all clones. For a detailed cladistic analysis on all different haplotypes in our samples, we applied seven additional restriction enzymes (EcoR I, EcoR V, Hae III, Hind III, Mbo I, Mse I, ScrF I, Table 2) on the mtDNA of a subset of clones. One D. cucullata mtDNA-type from the Netherlands was included in parsimony analysis in order to have all known haplotypes represented (Tables 1 and 2. RFLP patterns of PCR products were used to identify composite haplotypes (Table 2, Appendix 3 1).

Table 2.    Nine mtDNA composite haplotypes (columns, upper half) are derived from RFLP analysis of 12S/16S gene regions in the analysis of 154 cytonuclear genotypes from the D. longispina group using four informative restriction enzymes (marked with asterisks). The frequencies of composite haplotypes within nuclear MDS-defined clusters C = D. cucullata, G D. galeata, H = ‘D. hyalina’ and intermediates (CH = D. cucullata × D. hyalina, CG = D. cucullata × D. galeata, GH = D. galeata × ‘D. hyalina’) are presented (lower half, cf. Fig. 2). Composite haplotypes with respect to 11 restriction enzymes were derived in a subset of clones and used in parsimony analysis (cf. Fig. 3). One additional haplotype, c (in bold type), from a D. cucullata clone from the Netherlands, was included to have all known haplotypes represented; p designates the haplotype from the outgroup D. pulex. The letters a to l represent restriction patterns for each enzyme (see Appendix  1), and ? represents missing data. n: total number of clones analysed. For labelling of composite haplotypes cf. Table 1 and Appendix 3 1. Thumbnail image of
Table 3. Appendix 1 Restriction fragment length (in base pairs) for restriction cuts of 12S/16S rDNA gene region. Fragments smaller than 100 bp were not scored. Restriction patterns (haplotype, HP) were used to construct composite haplotypes as shown in Table 2. Thumbnail image of

Morphological analysis

Morphological divergence between 114 available clones (standardized females in 5th instar) was estimated based on a large set of morphological traits, 41 qualitative variables (cf. Appendix 42), and seven morphometric or meristic variables (cf. Appendix  3) using D. pulex as outgroup. Although our samples are locally restricted, the morphological divergence of species examined here, corresponds to the phenotypic variation of species samples from all over Europe (personal observations of S.G.).

Table 4. Appendix 2 List of 41 qualitative morphological variables which were analysed in standardized Daphnia clones from the D. longispina group and in the outgroup D. pulex, and used in morphological analysis (cf. Fig. 5). Thumbnail image of
Table 5. Appendix 3 Description of seven morphometric and meristic variables which were analysed in standardized Daphnia clones from the D. longispina group and used in morphological analysis (cf. ). Thumbnail image of

Data analysis

Multidimensional scaling (MDS) and UPGMA-clustering based on Euclidean distances between 15-locus allozyme genotypes (ALSCAL, PROXIMITIES, CLUSTER, SPSS 6.1) were used to cluster nuclear genotypes by genotypic relatedness (cf. Gießler, 1997a), and to display associations between nuclear genotype and mtDNA haplotype. We applied both methods on the whole data set including a large number of hybrids, because practical evidence suggests that simple clustering algorithms or MDS-analysis on distance matrices (based on a large number of genetic loci) provide a useful picture of genetic relationships, even when reticulate evolution is involved ( Clegg et al., 1993 ; Gießler, 1997a). Even though hybrids are forced to be grouped to only one parent type, we used the hierarchical approach to be able to display all clones and to decide about problematic delineations of primary groups in subsequent MDS-analyses. We used the spatial arrangement of clones in nonhierarchical MDS-analysis to decide about species and interspecific hybrid genomes. Genotypes were coded using alleles per locus as characters (locus with alleles 1, 2, 3, 4: genotype 12 is coded as 0.5 0.5 0 0, genotype 11 as 1.0 0 0 0, etc.).

Because of severe topological disturbance in trees when including numerous hybrids in cladograms (cf. McDade, 1990, 1992), all phylogenetic analyses were restricted to a selection of clones from presumed species (cf. Fig. 4). Phylogenetic tree construction was based on parsimony analyses using discrete data sets: allozyme ( PAUP 3.1.1, Swofford, 1993; tree not shown) and mtDNA restriction site data (MIX, PHYLIP 3.5, Felsenstein, 1993); data from D. pulex were used for outgroup rooting. The 15 polymorphic loci were treated as independent characters by coding different genotypes as character states (cf. Buth, 1984; Murphy, 1993). PAUP analysis based on the heuristic search, the step matrix option, was applied (cf. Mabee & Humphries, 1993) and invariant characters were excluded (tree not shown here). To enable statistical comparison between mtDNA divergence and mean nuclear divergence per haplotype group, all 61 multilocus genotypes from species clusters were grouped according to their respective mitochondrial DNA haplotype (Table 2). A nuclear phylogram was derived (CONTML, PHYLIP 3.5, Felsenstein, 1993) based on allele frequencies and maximum likelihood procedures. Only clones from presumed species clusters were involved (see above). The congruence of phylogenies was tested applying Mantel tests on underlying distance matrices ( Mantel, 1967). Morphological divergence of clones was estimated as Euclidean distances based on z-standardized variables and displayed in a UPGMA dendrogram (CLUSTER, SPSS 6.1).


Figure 4.   Allozyme maximum likelihood tree (CONTML, PHYLIP 3.5) based on allele frequencies at 15 polymorphic loci. For this analysis, 61 genotypes from presumed MDS-species clusters C (D. cucullata), G (D. galeata) and H (‘D. hyalina’) are grouped according to mitochondrial haplotypes c1, g1, g2, h1 to h5 (plus outgroup p: D. pulex) to enable comparison of nuclear and mtDNA divergence between species from the D. longispina group. Encircled are haplotypes from MDS-defined species clusters (cf. Table 1 and Fig. 2). Cytonuclear genotypes with haplotype c2 are missing because this mtDNA type was only found in hybrid cluster CH.

Download figure to PowerPoint

Because of difficulties in culturing, joint analysis of all nuclear, mtDNA and morphology markers could not be obtained from all clones.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Genetic divergence

Out of all 158 clones (cf. Table 1), 132 differed in nuclear multilocus genotype from any other clone. UPGMA-clustering based on divergence in nuclear 15-locus genotypes revealed six clusters (H, G, GH, C, CG, CH) within the D. longispina group using D. pulex as outgroup (Fig. 1). In three clusters (H, G and GH), clones with the same mtDNA haplotypes were grouped together (Fig. 1). Figure 2 displays the spatial arrangement of clones in two dimensions to identify primary groups (parental species) and hybrids (MDS analysis, D. pulex is not included here). In this plot, clusters separated by largest genotypic distances (primary groups) can be assumed to include clones from different species, and pairwise intermediate distances indicate intermediate genotypes (hybrids). With few exceptions (cf. Fig. 1), clones were grouped into the same six clusters. Mean genotypic distance is largest between three multilocus genotype clusters (C, G, H, Fig. 2) and cluster C is the most distinct among them. While it is easy to delineate cluster C, cluster H is heterogenous and the relationship between clusters G and H is complex. Delineations of cluster G and H (as well as CG/CH) correspond to clusters gained by UPGMA-analysis based on the same distance matrix (cf. Fig. 1). Because the spatial arrangement of clones is similar to previous studies using the same approach and mainly animals from the same region ( Gießler, 1997a), we labelled the nuclear divergent primary groups accordingly: C (D. cucullata), G (D. galeata) and H (‘D. hyalina), and intermediate clusters (secondary groups) correspondingly: CH (D. cucullata × D. hyalina), CG (D. cucullata × D. galeata) and GH (D. galeata × ‘D. hyalina’).


Figure 1 UPGMA analysis of nuclear divergence between 158 clones from the D. longispina complex (Euclidean distance between 15-locus allozyme genotypes; PROXIMITIES, CLUSTER, SPSS 6.1). Using D. pulex as outgroup, six clusters are resolved which are labelled H, G, GH, CH, C and CG. Associated mtDNA haplotypes are displayed as symbols. Mismatching clones with respect to cluster membership in nonhierarchical MDS-analysis (Fig. 2) are marked, # denotes clones within GH-MDS cluster, ## clone within CH-MDS cluster, and ### clone within CG-MDS cluster. For labelling of composite haplotypes c1 to h5, see Table 1. ; ?: indicates missing mtDNA information.

Download figure to PowerPoint

The mitochondrial DNA information was used to determine the maternal species of clones. With few exceptions, animals exhibited mtDNA haplotypes that could be expected from presumed parental mtDNA genomes in all intermediate nuclear MDS-clusters, CH, CG and GH (Fig. 2). Thus, intermediate cytonuclear genotypes represent respective hybrids. Evidence for bidirectional hybridization was found in all intermediate clusters (Fig. 2, Table 2). However, we found a unique haplotype, c2, in MDS-cluster CH and unique parental haplotypes that were not found in hybrids (e.g. haplotype g2 in cluster G, Fig. 2). Lake-specific haplotypes were c2 and h2 from Hartsee, haplotypes h4 and h5 from Ismaning pond and g2 from Obinger See (Table 1). All other haplotypes were found in more than one lake, some (c1, g1 and h3) being distributed over distances of 500 km (Table 1).

Phylogenetic analyses based on mtDNA and allozymes

Parsimony analysis based on mtDNA restriction sites (11 restriction enzymes, cf. Table 2, Fig. 3) involving all nine haplotypes revealed three distinct clades using D. pulex (p) as outgroup: all five haplotypes from cluster H, ‘D. hyalina’, were monophyletic as well as the two haplotypes from species cluster G, D. galeata (cf. Fig. 2). Furthermore, the unique haplotype (c2) from intermediate cluster CH (Fig. 2) and the additional D. cucullata haplotype from the Netherlands (c) both were monophyletic with haplotype c1 in cluster C, suggesting specificity for D. cucullata. Main branches are supported by high bootstrap values, and only comparisons within groups lack sufficient resolution. Branch lengths suggest a closer relationship between clusters D. cucullata and D. galeata than either of them to cluster ‘D. hyalina’.


Figure 3 Phylogenetic tree ( PHYLIP 3. .5, Wagner parsimony) based on mitochondrial DNA restriction sites and 11 restriction enzymes. Subsamples of all nine (plus outgroup p: D. pulex) different mtDNA haplotypes gained in the analysis of clones from the D. longispina complex (based on four restriction enzymes, Table 2) were analysed using seven more restriction enzymes. One additional haplotype, c, from a D. cucullata clone from the Netherlands, was included to have all known haplotypes represented. Encircled are haplotypes associated with MDS-defined species clusters (stippled lines): C (D. cucullata), G (D. galeata) and H (‘D. hyalina’, cf. Fig. 2), and monophyletic haplotype groups (dashed lines). Numbers on branches denote bootstrap values after 100 replicates.

Download figure to PowerPoint

A nuclear phylogram (CONTML, PHYLIP 3.5, Fig. 4) was derived using multilocus genotypes from presumed species clusters which were grouped according to their respective mitochondrial DNA haplotype (cf. Table 2). Using D. pulex as outgroup, we got the same topology as in nuclear MDS analysis comprising all clones:branch lengths suggest that cluster D. cucullata clones were the most distinct and cluster D. galeata was more closely related to cluster ‘D. hyalina’ than to cluster D. cucullata (Fig. 4). In cluster ‘D. hyalina’, divergence in gene frequencies set three haplotype groups (h1, h2, h3) closer together.

Morphological divergence

To compare morphological variation within presumed parental species, nuclear genotypic intermediates, and associations between haplotype and morphology, we estimated the morphological divergence between clones. UPGMA analysis based on 48 morphological traits (Euclidean distances, CLUSTER, SPSS 6.1) resolved three main clusters using D. pulex as outgroup (Fig. 5). In main cluster (1) all clones with C-genomes, D. cucullata, clustered together and were set apart from two clusters comprising CG-clones and CH-clones, respectively. Main cluster (2) was more heterogenous: one set of clones bearing H-genomes, ‘D. hyalina’, clustered together, while clones with G-genomes, D. galeata, and GH-clones were scattered over several subclusters. Main cluster (3) comprised another set of clones with H-genomes. Thus, the most striking result was the split of one nuclear defined group of clones, H, by morphological divergence. One group of clones, H1 (haplotypes h1, h2, h3) was morphologically very different from a second group of clones, H2, with distinct haplotypes (h4, h5, cf. Figs 1, 2 and 5). In all other morphologically defined clusters the association between haplotype and morphology was low. The morphology of clones with hybrid genomes was generally found to be intermediate to that of parentals, but the resolution among G and GH clones was especially poor (Fig. 5). Similarly, the morphologies of clones with CH-genomes were not all strictly intermediate to clones with H-genomes or C-genomes, some of them being morphologically very similar to clones with H-genomes.

Comparison of the morphology of clones with the description of traditional species ( Flößner, 1972) revealed that animals belonging to cytonuclear genotype group H2 represent D. rosea phenotypes, while animals from cytonuclear genotype group H1 resemble D. hyalina phenotypes. Phenotypes within the nuclear defined MDS-cluster C correspond to D. cucullata, and phenotypes within G to D. galeata (cf. Figs 2 and 5). Thus, the morphology of D. cucullata was more distinct from D. hyalina and D. galeata than D. hyalina from D. galeata, while D. rosea was morphologically most divergent from all other species.

Comparison of genetic and morphological divergence

The analysis of allozyme data resulted in six groups, which were first interpreted to represent three species D. cucullata, D. hyalina and D. galeata and corresponding hybrids (CG, GH, CH). Discontinuities in genotypic space (Fig. 1) allowed the grouping of clones into primary groups (species) and secondary groups (hybrids) in nonhierarchical analysis (Fig. 2), although about 60% of the clones bore intermediate genomes. Phylogenetic analysis of mtDNA data also resulted in three groups interpreted to represent monophyletic species (Fig. 3). Haplotypes were specifically associated with only one of the three species clusters and clones from hybrid clusters generally bore haplotypes corresponding to hybrid parentage (Fig. 2). Tree topologies based on mtDNA data (restricted to haplotypes found in species, cf. Fig. 3) and allozymes (restricted to corresponding haplotype groups from species clusters, cf. Fig. 4) were incongruent (Mantel’s test: = 0.20, = 0.21). mtDNA data suggest a closer relationship between C and G than either of them to H (Fig. 3). In contrast, allozyme analysis suggests C-clones were the most distinct and G was more closely related to H than to C (Fig. 4). Nuclear and mtDNA analyses were consistent in that all five mtDNA haplotypes of cluster H formed a monophyletic group.

In contrast to genetic analyses, analysis of morphological divergence between clones resolved more than six groups (Fig. 5): morphological divergence was associated with mtDNA-haplotype in clones bearing H-genomes, resulting in two divergent clusters H1 (D. hyalina phenotypes) and H2 (D. rosea phenotypes). With respect to D. hyalina and D. rosea, high morphological divergence does not correspond to low divergence in mtDNA haplotypes ( Figs 3 and 5), and was much more pronounced than differences in nuclear genotype frequencies suggest (Fig. 4). Morphological divergence between D. hyalina, H1, and D. galeata, was much lower, despite higher cytonuclear divergence ( Figs 345). Nevertheless, with some exceptions (e.g. H2), morphological and genetic divergence of clones was associated and clones with hybrid genomes were found to be of intermediate morphology ( Figs 1 and 5). In only one species pair (D. cucullata, D. hyalina) were divergence estimates from all three data sets concordant.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Our study, based on three independent data sets of nuclear, mtDNA and morphological divergence, extends the data on phylogenetic relationships among traditional species within the D. longispina group. In contrast to past studies which focused on few representatives from presumed species ( Schwenk, 1993; Colbourne & Hebert, 1996; Taylor et al., 1996 ; Colbourne et al., 1997 ), we used a different approach and classified animals from natural clonal assemblages (including a large number of hybrids) a posteriori as belonging to species. We relate primary groups emerging from genetic analyses to ‘species’ because we assume that they represent long-term evolutionary units. Genetic and morphological markers confirm that cytonuclear defined ‘species’ largely fit traditional taxonomic descriptions ( Flößner, 1972; Flößner & Kraus, 1986; Wolf & Mort, 1986). However, we find evidence for reticulate evolution among species. Although only few of recurrently formed hybrids or introgressants might be involved in long-term processes, introgressed genes might be highly adaptive, favouring reticulate evolutionary processes.

Genetic differentiation among species and parentage of interspecific hybrids

Based on nuclear genotypic divergence and discontinuities in genotypic space, individuals from several clonal assemblages of the D. longispina complex could be grouped into six groups: three divergent primary groups (species), C, G, H, and pairwise intermediate secondary groups, CG, CH, GH (hybrids). Primary groups were associated with group-specific haplotypes and secondary groups were associated with corresponding parental haplotypes. Poor resolution between two primary groups (H, G) suggest backcrossing. mtDNA analysis confirmed the definition of primary groups: associated haplotypes were grouped together in three monophyletic clusters. Morphological analysis revealed that particular haplotypes within one of the primary groups, H, were linked with significantly different morphology known from two traditional species: D. hyalina and D. rosea. Another primary group, C, comprised D. cucullata phenotypes, and the third, G, D. galeata phenotypes. The close relationship of D. hyalina and D. rosea, based on allozyme data, was consistent with mtDNA RFLP data. Although D. rosea clones could be discriminated by mtDNA haplotypes (h4 and h5) and differed in mean genotype frequencies from D. hyalina, D. hyalina and D. rosea did not form distinct monophyletic groups in cladistic analysis. The low level of genetic differentiation between D. hyalina and D. rosea is supported by mitochondrial DNA sequence data ( Taylor et al., 1996 ; K. Schwenk, unpublished), nuclear DNA (S. Gießler, unpublished) and allozyme analysis ( Taylor et al., 1996 ). The association with distinct mtDNA haplotypes and low nuclear divergence to D. hyalina is consistent with an estimated separation time of about 600 000 years based on 12S rDNA sequence data ( Taylor et al., 1996 ). The question arises whether D. rosea represents a recently differentiated sister species of D. hyalina, or a conspecific lineage that only differs in certain phenotypic characteristics. The clustering of mtDNA types in morphological analysis suggests reduced gene flow. Different habitat preferences (lakes or ponds) and significant ecological differentiation in life histories (S. Gießler, unpublished) fit assumptions of the speciation by habitat shift hypothesis. This speciation model is favoured for several other sister species which are separated on the lake/pond level ( Taylor et al., 1996 ).

We found no evidence for cytoplasmatic isolation between species in case of hybridization. In three intermediate hybrid groups, CH, CG and GH, hybridization was bidirectional, which shows that previously assumed unidirectional hybridization is not a general rule ( Schwenk, 1993). In general, hybrid taxa combined the alleles of their parents and had mtDNA haplotypes from one parent, suggesting recent hybrid origin. This is consistent with low mtDNA sequence divergence between hybrids and parents ( Schwenk, 1993). However, some hybrids exhibited haplotypes not found in a parental species, and vice versa. This might be due to sampling artefacts, but may also indicate nonrandom mating among parental haplotypes. Few intermediates between D. galeata and D. hyalina/D. rosea were found exhibiting nuclear and morphological traits close to one species but mitochondrial DNA of the other. This is consistent with findings that mtDNA introgression in the D. longispina group is rare ( Schwenk, 1993; Taylor & Hebert, 1993a).

Discordant phylogenies and reticulate evolution

To estimate phylogenies, cladistic, maximum likelihood and several phenetic analyses were applied on each of the independent data sets (allozymes, and mtDNA restriction sites). Since tree topologies gained from a particular data set were generally concordant, independent of analytical technique, not all phylogenies are represented in figures. The nuclear phylogeny based on 15 allozyme loci yielded the same topology as did random amplified polymorphic DNA (RAPD) data ( Schierwater et al., 1994 ): D. hyalina and D. rosea are genetically closely related, D. galeata is more closely related to D. hyalina than to D. cucullata, and D. cucullata is most distinct from all others. In contrast to nuclear data, restriction site data of mitochondrial DNA yielded a different branching pattern: D. galeata is more closely related to D. cucullata than to D. hyalina. This phylogeny derived from our mtDNA RFLP data is supported by phylogenetic reconstructions based on sequence analyses of cytochrome b or 12S/16S rDNA mitochondrial gene regions ( Schwenk, 1993; Taylor et al., 1996 ). The incongruence of nuclear and mtDNA phylogenies suggests unbalanced nuclear and cytoplasmic gene transfer between species, and reticulate evolutionary processes.

Although our analysis of genetic and morphological variation in Daphnia is restricted to animals which did not originate from the entire species range, the results might very well reflect a general pattern. Aside from the traditional species D. hyalina and D. rosea, genetic divergence between species is relatively high, and very similar processes of introgression have been reported from other areas ( Taylor & Hebert, 1992, 1993a). Patterns of reticulate evolution and recent asymmetric gene flow from D. rosea to D. galeata mendotae are found in North America and among the D. longispina complex in northern Germany ( Schwenk et al., 1995 ). Reticulations are expected to be found on different time-scales: contemporary fluid species complexes may result in long-term mtDNA or nuclear introgression between species, and contemporary separated gene pools may have experienced ancient introgression. mtDNA sequence divergence estimates the largest D. longispina radiation to be > 5 Myr ( Schwenk, 1993; Taylor et al., 1996 ). Therefore, all species have experienced several phases of glaciation and deglaciation enabling secondary contact of formerly separated gene pools and introgression.

In some Daphnia species complexes generating hybrid swarms today, species gene pools may not have merged, and isolation of mitochondrial DNA gene pools remains nearly complete. This seems to be valid for D. cucullata and D. galeata: introgression during ancient hybridization events and establishment of stabilized recombinants would explain discordant nuclear and mtDNA phylogenies. Nuclear and mtDNA differentiation between D. galeata, D. hyalina and their genotypic intermediates indicate higher rates of gene flow. Despite contemporary hybridization and backcrossing, genetic divergence between sister species (C, G, H) which live syntopic in lake habitats was much greater than between sister species (R, H) from different habitat types, suggesting reproductive isolation to be reinforced in sympatry. Although introgression might be generally rare, the adaptive significance may have a large effect in taxa reproducing by parthenogenesis ( Arnold, 1997). Our results strongly suggest that reticulate evolution needs to be included in the historical narrative of Daphnia.

Speciation and morphological evolution

It has been emphasized that within the genus Daphnia only few morphological traits are variable and morphology is too constrained to understand phylogenetic relationships ( Colbourne & Hebert, 1996). In addition, phenotypic plasticity in field animals disguises the genetic component of morphological divergence between animals. To overcome some of these problems in our morphological analysis, we chose a very large number of traits in newly established standardized clonal cultures. Although our samples included a large set of hybrids, morphological divergence of clones could be reliably estimated and was to a large extent associated with nuclear differentiation. The morphological divergence between D. cucullata and D. galeata or D. hyalina, H1, corresponded to nuclear relatedness. Genetic and morphological intermediacy of hybrids was generally associated, suggesting that most of the recombinants represent F1 hybrids. This is consistent with morphometric comparisons of artificial laboratory F1 hybrids of D. cucullata and D. galeata which revealed intermediate phenotypes (M. Bijl and K. Schwenk, unpublished results). Some hybrids (e.g. bearing GH genomes) were not intermediate to parental species in morphology. Different levels of morphological asymmetries to parents suggest backcrossing to one parental species, or strong selection against certain phenotypes, or nonadditivity of biparentally inherited traits.

With respect to D. rosea and D. hyalina the high extent of morphological divergence in standardized laboratory clones did not correspond to low nuclear divergence, suggesting a recent split. This finding implies that the rate of morphological evolution may to a large extent depend on habitat shifts. Timing of speciation and phenotype change both are suspected to be strongly associated with habitat shifts in Daphnia ( Colbourne et al., 1997 ). Different habitat types (lakes/ponds) linked to different selection regimes may accelerate the rate of morphological evolution. This is consistent with our data. Divergent selection on morphology might explain why striking morphological differences between D. hyalina and D. rosea from different habitat types were not correlated with divergence in genetic markers. Despite pronounced morphological divergence, molecular similarity could reflect inconsistencies between molecular and morphological data such as homoplastic events (reversals and convergences) in the course of evolution ( Johnson et al., 1977 ; Thorpe, 1982; Hillis et al., 1987 ; Murray et al., 1991 ). Genetic markers and morphological traits might differ in their evolutionary histories and mutation rates, resulting in different patterns of molecular and morphological evolution ( Gorman & Kim, 1976). Genetic changes might be substantial during morphological stasis, and vice versa (e.g. Sturmbauer & Meyer, 1992). If phenotypes respond to selection, then phenotypes may experience strong directional selection. We conclude that, in Daphnia, speciation and morphological evolution need not proceed in synchrony.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This work was supported by DFG research grant (GA 362/3-1) and a fellowship of the Netherlands Academy for Arts and Sciences to S.G. (DIR/BW/7035). Manuscript writing was supported by an HSPII grant to S.G. We thank W. Gabriel, J. Vijverberg and W. van Vierssen for support, A. v. Haeseler for statistical advice, B. Nürnberger, J. van Damme, G. Fryer, P. Spaak, T. Städler, S. McLaughlin and T. Saks for their critical reading of the manuscript, and two anonymous reviewers for greatly improving the manuscript. We are grateful to D. Ebert and J. Müller for supplying Daphnia samples.

  1. These coded for: amylase (AMY, E.C., aldehyde oxidase (AO,, glutamate-oxalacetate transaminase (GOT,, glucose phosphate isomerase (GPI,, hexokinase (HEX,, malate dehydrogenase (MDH,, malic enzyme (ME,, mannose phosphate isomerase (MPI,, four peptidases (PEP1, PEP 2.2, substrate L-leucyl- L-alanine, PEP2.1, substrate L-leucyl-glycyl-glycine, PEP3, substrate L-valyl- L-leucine, E.C. 3.4.11/13), phosphoglucomutase (PGM,, triose phosphate isomerase (TPI,, and xanthine dehydrogenase (XDH,

  2. Because it is impossible to analyse allozyme variation in alcohol samples, the nuclear genotype as defined by RAPDs ( Schwenk, 1997) was used to identify species affiliation.

  3. The 41 qualitative variables included morphological traits traditionally used in Daphnia taxonomic keys (e.g. Flößner, 1972, 1993; Flößner & Kraus, 1986).

  4. All measurements were taken from living animals to the nearest 0.01 mm.

  5. Few mismatching clones with respect to hierarchical UPGMA and nonhierarchical MDS cluster membership which were placed in midst UPGMA clusters (marked with #, cf. Fig. 1) but within different intermediate MDS clusters were assigned to clusters corresponding to MDS-cluster membership for further analyses.

  6. Because D. hyalina is restricted to large lakes (cf. Flößner, 1972) and primary group H comprised clones from pond and lake habitats we labelled it ‘D. hyalina’.

  7. Four allozyme-characterized clones from the total sample are missing, cf. Table 1.

  8. Other methods, such as maximum likelihood (RESTML, PHYLIP, ver. 3.5) or phenetic methods (evolutionary distances of restriction fragments, Nei & Tajima, 1983, and UPGMA), revealed identical branching patterns to those obtained by parsimony procedures (results not shown).

  9. The branching topology was supported by nearly identical trees resulting from phenetic analyses of the allozyme data set based on chord distances ( Cavalli-Sforza & Edwards, 1967) and UPGMA (NEIGHBOR, PHYLIP 3.5) or Fitch-Margoliash (FITCH, PHYLIP 3.5). Furthermore, cladistic analysis (PAUP 3.1.1, Swofford, 1993) supports the nuclear tree topology in that cluster C, D. cucullata, was the most distinct (bootstrap values about 90, trees not shown here).

  10. The subdivision into subclusters H1 and H2 was also evident if morphological divergence was based on qualitative data only (cf. Appendix 4 2) or quantitative data only (cf. Appendix 5 3).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Arnold, M.L. 1992. Natural hybridization as an evolutionary process. Annu. Rev. Ecol. Syst. 23: 237 261.
  • 2
    Arnold, M.L. 1997. Natural Hybridization and Evolution. Oxford University Press.
  • 3
    Aurich, H.J. 1933. Das Zooplankton einiger Seen des Chiemgaus. Seine lokale Verbreitung und Verteilung im Raum und Zeit. Int. Revue ges. Hydrobiol. 29: 295 345.
  • 4
    Baumbach, P. 1922. Studien über das Zooplankton des Chiemsees. In: Archiv der Naturgeschichte, 88. Jahrgang, 7. Heft (E. Strand, ed.), pp. 189–192. Berlin.
  • 5
    Benzie, J.A.H. 1986. Phenetic and cladistic analyses of the phylogenetic relationships within the genus Daphnia worldwide. Hydrobiologia 140: 105 124.
  • 6
    Brooks, J.L. 1957. The species problem in freshwater animals. In: The Species Problem (E. Mayr, ed.), pp. 81–123. American Association for the Advancement of Science, Washington, DC.
  • 7
    Buth, D.G. 1984. The application of electrophoretic data in systematic studies. Annu. Rev. Ecol. Syst. 15: 501 522.
  • 8
    Cavalli-Sforza, L.L. & Edwards, A.W.F. 1967. Phylogenetic analysis: Models and estimation procedures. Evolution 21: 550 570.
  • 9
    Clegg, M.L., Gaut, B.S., Duvall, M.R., Davis, J. 1993. Inferring plant evolutionary history from molecular data. NZ J. Bot. 31: 307 316.
  • 10
    Colbourne, J.K., Hebert, P.D.N., Taylor, D.J. 1997. Evolutionary origins of phenotypic diversity in Daphnia. In: Molecular Evolution and Adaptive Radiation (T. J. Givnish & K. J. Sytsma, eds), pp. 163–188. Cambridge University Press.
  • 11
    Colbourne, J.K. & Hebert, P.D.N. 1996. The systematics of North American Daphnia (Crustacea: anomopoda): a molecular phylogenetic approach. Phil. Trans. R. Soc. Lond. B 351: 349 360.
  • 12
    Dowling, T.E. & DeMarais, B.D. 1993. Evolutionary significance of introgressive hybridization in cyprinid fishes. Nature 362: 444 446.
  • 13
    Felsenstein, J. 1993. PHYLIP (Phylogeny Inference Package), Version 3.5c. Distributed by the author. Department of Genetics, University of Washington, Seattle.
  • 14
    Ferris, S.D., Sage, R.D., Huang, C.H., Nielsen, J.T., Ritte, U., Wilson, A.C. 1983. Flow of mitochondrial DNA across a species boundary. Proc. Natn. Acad. Sci. U.S.A. 80: 2290 2294.
  • 15
    Flößner, D. 1972. Kiemen- und Blattfüßer, Branchiopoda, Fischläuse, Branchiura. In: Die Tierwelt Deutschlands (F. Dahl, ed.), pp. 1–501. Gustav Fischer-Verlag, Jena.
  • 16
    Flößner, D. 1993. Zur Kenntnis einiger Daphnia-Hybriden . Limnologica 23: 71 79.
  • 17
    Flößner, D. & Kraus, K. 1986. On the taxonomy of the Daphnia hyalina-galeata complex (Crustacea: Cladocera). Hydrobiologia 137: 97 115.
  • 18
    Gessner, F. 1953. Die Limnologie des Naturschutzgebietes Seeon. Arch. Hydrobiol. 47: 553 624.
  • 19
    Gießler, S. 1987. Mikroevolution and Populationsgenetik im Daphnia Galeata/Hyalina/Cucullata-Komplex (Crustacea: Cladocera). Eine Freilandanalyse. PhD thesis, University of Munich, Germany.
  • 20
    Gießler, S. 1997a. Analysis of reticulate relationships within the Daphnia longispina complex. Allozyme phenotype and morphology. J. Evol. Biol. 10: 87 105.
  • 21
    Gießler, S. 1997b. Gene flow in the Daphnia longispina hybrid complex (Crustacea, Cladocera) inhabiting large lakes. Heredity 79: 231 241.
  • 22
    Gorman, G.C. & Kim, Y.J. 1976. Anolis lizards of the Eastern Caribbean: a case study in evolution. II. Genetic relationships and genetic variability of the bimaculatus group. Syst. Zool. 25: 62 77.
  • 23
    Harris, H. & Hopkinson, D.A. 1976. Handbook of Enzyme Electrophoresis in Human Genetics. North-Holland Publishers Co., Amsterdam.
  • 24
    Harrison, R.G., Rand, D.M., Wheeler, W.C. 1987. Mitochondrial DNA variation in field crickets across a narrow hybrid zone. Mol. Biol. Evol. 4: 144 158.
  • 25
    Hebert, P.D.N. & Beaton, M.J. 1989. Methodologies for Allozyme Analysis Using Cellulose Acetate Electrophoresis: a Practical Handbook. Helena Laboratories, Beaumont, Texas.
  • 26
    Hebert, P.D.N., Schwartz, S.S., Hrbácek, J. 1989. Patterns of genotypic diversity in Czecheslovakian Daphnia. Heredity 62: 207 216.
  • 27
    Hillis, D.M., Rosenfeld, D.S., Sanchez, M. Jr. 1987. Allozyme variability and heterozygote deficiency within and among morphologically polymorphic populations of Liguus fasciatus (Mollusca: Pulmonata: Bulimulidae). Am. Malacol. Bull. 5: 153 157.
  • 28
    Hrbácek, J. 1987. Systematics and biogeography of Daphnia species in the northern temperate regions. Mem. Ist. Ital. Idrobiol. 45: 37 76.
  • 29
    Jacobs, J. 1977. Coexistence of similar zooplankton species by differential adaptation to reproduction and escape, in an environment with fluctuating food and enemy densities. II. Field data analysis of Daphnia. Oecologia 30: 313 329.
  • 30
    Johnson, M.S., Clarke, B., Murray, J. 1977. Genetic variation and reproductive isolation in Partula. Evolution 31: 116 126.
  • 31
    Mabee, P.M. & Humphries, J. 1993. Coding polymorphic data: examples from allozymes and ontogeny. Syst. Biol. 42: 166 181.
  • 32
    Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Res. 27: 209 220.
  • 33
    McDade, L. 1990. Hybrids and phylogenetic systematics I. Patterns of character expression in hybrids and their implications for cladistic analysis. Evolution 44: 1685 1700.
  • 34
    McDade, L. 1992. Hybrids and phylogenetic systematics II. The impact of hybrids on cladistic analysis. Evolution 46: 1329 1346.
  • 35
    Moore, W.S. 1995. Inferring phylogenies from mtDNA variation: mitochondrial-gene trees versus nuclear-gene trees. Evolution 49: 718 726.
  • 36
    Müller, J. 1993. Räumliche und zeitliche Variabilität der genetischen Struktur natürlicher Cladocerenpopulationen (Crustacea, Cladocera). PhD thesis, J. Gutenberg- Universität, Mainz, Germany.
  • 37
    Murphy, R.W. 1993. The phylogenetic analysis of allozyme data – invalidity of coding alleles by presence/absence and recommended procedures. Biochem. Syst. Ecol. 21: 25 38.
  • 38
    Murray, J., Stine, O.C., Johnson, M.S. 1991. The evolution of mitochondrial DNA in Partula. Heredity 66: 93 104.
  • 39
    Nei, M. & Tajima, F. 1983. Maximum likelihood estimation of the number of nucleotide substitutions from restriction sites data. Genetics 105: 207 217.
  • 40
    Rieseberg, L.H. 1995. The role of hybridization in evolution: old wine in new skins. Am. J. Bot. 82: 944 953.
  • 41
    Schierwater, B., Ender, A., Schwenk, K., Spaak, P., Streit, B. 1994. The evolutionary ecology of Daphnia. In: Molecular Ecology and Evolution: Approaches and Applications (B. Schierwater, B. Streit, G. P. Wagner & R. DeSalle, eds), pp. 495–508. Birkhäuser-Verlag, Basel.
  • 42
    Schwenk, K. 1993. Interspecific hybridization in Daphnia: distinction and origin of hybrid matrilines . Mol. Biol. Evol. 10: 1289 1302.
  • 43
    Schwenk, K. 1997. Evolutionary Genetics of Daphnia species complexes – Hybridism in syntopy. PhD thesis, University of Utrecht, The Netherlands.
  • 44
    Schwenk, K., Ender, A., Streit, B. 1995. What can molecular markers tell us about the evolutionary history of Daphnia species complexes? Hydrobiologia 307: 1 7.
  • 45
    Schwenk, K. & Spaak, P. 1995. Evolutionary and ecological consequences of interspecific hybridization in cladocerans. Experientia 51: 465 481.
  • 46
    Seitz, A. 1980. The coexistence of three species of Daphnia in the Klostersee. I. Field studies on the dynamics of reproduction. Oecologia 45: 117 130.
  • 47
    Spaak, P. 1996. Temporal changes in the genetic structure of the Daphnia species complex in Tjeukemeer with evidence for backcrossing. Heredity 76: 539 548.
  • 48
    Spss software package, version 6.1. 1996. Chicago, Illinois.
  • 49
    Sturmbauer, C. & Meyer, A. 1992. Genetic divergence, speciation and morphological stasis in a lineage of African cichlid fishes. Nature 358: 578 581.
  • 50
    Swofford, D.L. 1993. PAUP: Phylogenetic analysis using parsimony, Version 3.1.1. Illinois Natural History Survey, Champaign.
  • 51
    Taylor, D.J. & Hebert, P.D.N. 1992. Daphnia galeata mendotae as a cryptic species complex with interspecific hybrids. Limnol. Oceanogr. 37: 658 665.
  • 52
    Taylor, D.J. & Hebert, P.D.N. 1993a. Habitat-dependent hybrid parentage and differential introgression between neighboringly sympatric Daphnia species. Proc. Natn. Acad. Sci. USA 90: 7079 7083.
  • 53
    Taylor, D.J. & Hebert, P.D.N. 1993b. Cryptic intercontinental hybridization in Daphnia (Crustacea) – the ghost of introductions past. Proc. Roy. Soc. Lond. B 254: 163 168.
  • 54
    Taylor, D.J., Hebert, P.D.N., Colbourne, J.K. 1996. Phylogenetics and Evolution of the Daphnia longispina group (Crustacea) based on 12S rDNA sequence and allozyme variation. Molec. Phylogenet. Evol. 5: 495 510.
  • 55
    Thorpe, J.P. 1982. The molecular clock hypothesis: biochemical evolution, genetic differentiation and systematics. Annu. Rev. Ecol. Syst. 13: 139 168.
  • 56
    Vazquez, P., Cooper, S.J.B., Gosalvez, J., Hewitt, G.M. 1994. Nuclear DNA introgression across a Pyrenean hybrid zone between parapatric subspecies of the grasshopper Chorthippus parallelus. Heredity 73: 436 443.
  • 57
    Wendel, J.F., Stewart, J.M., Rettig, J.H. 1991. Molecular evidence for homoploid reticulate evolution among Australian species of Gossypium. Evolution 45: 694 711.
  • 58
    Wesenberg-Lund, C. 1926. Contributions to the biology and morphology of the genus Daphnia with some remarks on heredity. Mémoires del l’Académie Royale des Sciences et des Lettres de Danemark, Section des Sciences, 8 XI. 2, Copenhagen, Denmark.
  • 59
    Wilson, A.C., Cann, R.L., Carr, S.M., George, M.Jr , Gyllensten, U.B., Helm-Bychowski, K.M., Higuchi, R.G., Palumbi, S.R., Prager, E.M., Sage, R.D., Stoneking, M. 1985. Mitochondrial DNA and two perspectives on evolutionary genetics. Biol. J. Linn. Soc. 26: 375 400.
  • 60
    Wolf, H.G. & Mort, M.A. 1986. Inter-specific hybridization underlies phenotypic variability in Daphnia populations. Oecologia 68: 507 511.
  • 61
    Woltereck, R. 1919. Variation und Artbildung. Erster Teil, Francke, Berlin.