Phylogenetics and evolution of a circumarctic species complex (Cladocera: Daphnia pulex)

Biological Journal of the Linnean Society (1998), 65: 347–365. With 3 figures

Article ID: bj980251

Phylogenetics and evolution of a circumarcticspecies complex (Cladocera: Daphnia pulex)

J. K. COLBOURNE1, T. J. CREASE1, L. J. WEIDER2, P. D. N. HEBERT1,F. DUFRESNE3, AND A. HOBÆK4

1Department of Zoology, University of Guelph, Ontario NIG 2W1, Canada; 2Max-Planck-Institut fur Limnologie, Postfach 165, D-24302 Plon, Germany; 3Departement de Biologie,Universite du Quebec a Rimouski, Quebec G5L 3A1, Canada; 4Norwegian Institute forWater Research, Nordnesboder 5, N-5005 Bergen, Norway

Received 17 December 1997; accepted for publication 23 May 1998

The evolutionary history of freshwater zooplankton is still relatively unknown. However,studies of the microcrustacean Daphnia have revealed interesting patterns; the daphniids thatdominate ponds and lakes in the northern hemisphere may have recent origins, likelyassociated with the glacial advances and retreats during the Pleistocene. Moreover, theyform species complexes that actively engage in hybridization and introgression. The presentstudy examines the phylogenetic relationships among circumarctic members of the Daphniapulex complex, through the analysis of sequence diversity in 498 nt of the ND5 mitochondrialgene. Our results suggest that the complex is composed of three major clades, two of whichare subdivided into at least eight different lineages. Clearly, species in the complex showgenetic discontinuity. Many lineages originated during the Pleistocene, but at least threelineages diverged during the Pliocene. Two taxa (D. pulex, D. pulicaria), thought to be broadlydistributed in the northern hemisphere, are shown to be endemic to single continents. Ingeneral, the diversification of the pulex complex is characterized by rapidly dispersed lineagesspanning enormous distances and also by endemism in temperate areas. Gene flow amonglineages from the temperate region of different continents are restricted to rare inter-continental migrations across a polar bridge followed by convergent morphological evolution.

1998 The Linnean Society of London

ADDITIONAL KEY WORDS:—Arctic – speciation – cladistics – phenetics – molecularsystematics – hybridization – Pleistocene glaciation – ND5 sequence – mitochondrial DNA.

CONTENTS

Introduction . . . . . . . . . . . . . . . . . . . . . . . 348Methods . . . . . . . . . . . . . . . . . . . . . . . . 350

Daphnia isolates . . . . . . . . . . . . . . . . . . . . 350DNA amplification and sequencing . . . . . . . . . . . . . . 350Phylogenetic analysis . . . . . . . . . . . . . . . . . . 352

Results . . . . . . . . . . . . . . . . . . . . . . . . 354ND5 gene sequence diversity . . . . . . . . . . . . . . . . 354Phenetic analysis of sequence divergence in the ND5 gene . . . . . . 354

∗Correspondence to J. K. Colbourne. E-mail:[email protected]

3470024–4066/98/110347+19 $30.00/0 1998 The Linnean Society of London

J. K. COLBOURNE ET AL.348

Cladistic analysis of sequence divergence in the ND5 gene . . . . . . 356Confidence in clades . . . . . . . . . . . . . . . . . . 359

Discussion . . . . . . . . . . . . . . . . . . . . . . . 360Acknowledgements . . . . . . . . . . . . . . . . . . . . 363References . . . . . . . . . . . . . . . . . . . . . . . 363

INTRODUCTION

The island-like nature of limnetic habitats would seem to favour the rapid evolutionof freshwater zooplankton. Yet, their morphological similarities over vast distancessuggest that speciation of microcrustaceans has been constrained by their unusualdispersal syndrome. Individuals cannot actively move beyond the boundaries ofindividual ponds or lakes, but their diapausing stages are capable of long-distancetransport, a factor which was thought to limit opportunities for gene-pool isolation(Darwin, 1859; Mayr, 1963). However, the few efforts to probe the genetic con-sequences of this dispersal syndrome have revealed an unexpected and complicatedpattern. Local populations of zooplankton—sampled within distances of a fewkilometers—typically show large genetic differences with FST values ranging from0.2 to 0.5 (see De Meester, 1996 for a review), indicating that gene flow is insufficientto erode divergence on this geographic scale. Yet, over large distances, populationsoften show both remarkable similarity in their mean gene frequencies and abruptgenotypic shifts at contact zones between different clades (Crease, Lynch & Spitze,1990; Hebert & Finston, 1996a; Weider & Hobæk, 1997). This pattern suggests thatmodern populations have derived from the rapid spread of a small number ofancestral metapopulations, perhaps originally located in different glacial refugia(Crease et al., 1990; Hebert et al., 1993). However, since a similar pattern is alsoobserved in species from areas unimpacted by glaciation (Hebert & Wilson, 1994;Hebert & Finston, 1996a), it appears that this pattern of gene-pool divergence isnot restricted to a few species with an unusual biogeographic history. There is aneed to further investigate the consequence of long-distance dispersal in structuringthe genetics of zooplankton populations if we are to uncover the importance ofgeographical speciation in these organisms.

No group of freshwater zooplankton has been more intensively studied thanmembers of the Daphnia pulex complex (sensu Colbourne & Hebert, 1996). Thiswork was motivated by their dominance in aquatic environments throughout thenorthern hemisphere (Hrbacek, 1987), by their ability to hybridize, and by theirbreeding system and ploidy-level variation. These traits may all be epiphenomenaof active speciation, as this is the sole daphniid assemblage (so far) in thenorthern hemisphere which shows evidence of a recent radiation. The Nearcticfauna is thought to include at least six species (Hebert, 1995). Daphnia pulex(Leydig) and D. pulicaria (Forbes) are broadly distributed in the temperate zoneand their ranges also extend into the Arctic. In contrast, D. arenata (Hebert) andD. melanica (Hebert) are narrow endemics, restricted to coastal ponds in Oregon.Two other species, D. middendorffiana (Fischer) and D. tenebrosa (Sars), are restrictedto the Arctic. The Palearctic fauna is thought to include the same four broadlydistributed taxa from North America (Hrbacek, 1987), but endemic species mayalso exist, as taxonomic studies have been less thorough in this area. Apart fromthe apparent overlap in species distributions, Holarctic daphniids share othersignificant biological similarities. Temperate-zone populations consist solely of

MOLECULAR SYSTEMATICS OF THE DAPHNIA PULEX COMPLEX 349

diploid lineages, while polyploids dominate polar habitats (Beaton & Hebert,1988; Ward et al., 1994). Polyploid lineages are invariably of hybrid origin,although several species have been involved in their formation. For example, allD. middendorffiana are polyploids that appear to be derived from hybridizationevents between D. pulicaria and other members of the pulex complex (Van Raay& Crease, 1995; Dufresne & Hebert, 1997). Similarly, hybridization between D.tenebrosa and D. pulex seems to always generate introgressed polyploids (Dufresne& Hebert, 1994). Although polyploids are apparently absent from the temperatezone, hybrid production is common. Diploid F1 hybrids between D. pulex andD. pulicaria comprise nearly half of the individuals sampled within temperateCanadian habitats, although they ordinarily reproduce by obligate parthenogenesis(Hebert et al., 1993). Studies of their mitochondrial DNA have so far shownthat D. pulex is always the maternal parent (Crease, Stanton & Hebert, 1989;Crease et al., 1990). By contrast, hybrids from polar regions never have thisspecies as the maternal parent.

Holarctic members of the pulex complex also show breeding system diversity:some lineages reproduce by cyclic parthenogenesis and others by obligateparthenogenesis. The two North American endemics (D. arenata, D. melanica)employ only the former breeding system, but the species restricted to arcticenvironments are either entirely (D. middendorffiana) or largely (D. tenebrosa) asexual.Polar populations of D. pulex and D. pulicaria also reproduce asexually, but thegeographic distribution of their breeding systems in the temperate zone is morecomplicated. For example, populations of D. pulex in the eastern half of NorthAmerica reproduce by obligate parthenogenesis, while those in the west arecyclic parthenogens (Hebert et al., 1993). The asexual lineages of all four taxashow an extraordinary amount of clonal diversity, apparently linked to theirrecurrent generation through hybridization and sex-limited meiosis suppression(Innes & Hebert, 1988).

Because these biological attributes are firmly linked to hybridization, cases ofgene-pool isolation and secondary contact by subsequent dispersion (during glacialadvances and retreats) seem particularly important in forging diversification withinthe pulex complex. In fact, phylogenetic studies using sequence divergence in the12S rDNA have indicated that all North American members of the species complexstem from the Pleistocene (< 2 Myr ago), save D. tenebrosa, which diverged from theothers during the Tertiary (Colbourne & Hebert, 1996). Unfortunately, relationshipsamong the closely related species is uncertain, because of the relative uniformity oftheir sequences in this slowly evolving ribosomal gene.

The present study aims to provide a thorough understanding of phylogeneticrelationships among members of the D. pulex complex, especially in the polar regionsof the northern hemisphere. The analysis extends prior genetic investigations of thegroup, in both analytical approach and geographic scale. Most preceding work hasexamined RFLP variation in the mitochondrial genome, but this study employsdirect sequence analysis of a 498 nt fragment of the rapidly evolving ND5 gene.The present study examines sequence diversity among lineages from sites throughoutEurasia and North America, while past work provided only a local perspective ongenetic affinities. Using the phylogenetic framework granted by this study, we intendto present several papers that investigate the interplay between geographic isolationand long-range transport of propagules in the evolutionary divergence of theseorganisms.


METHODS

Daphnia isolates

Populations of the D. pulex complex were sampled from several hundred rockpools,ponds and lakes at locations throughout the Arctic (Table 1). Populations wereinitially surveyed for variation at six polymorphic allozyme loci. Individuals withdistinct multilocus allozyme phenotypes were then chosen for mitochondrial DNAanalysis in order to maximize the detection of unique haplotypes. Restriction sitevariation in a 2.1 kb fragment of the NADH dehydrogenase subunit 4 (ND4) andsubunit 5 (ND5) genes was surveyed in isolates from 276 populations. Of the 171unique mitochondrial DNA haplotypes that were found in the Arctic, 61 weresubsequently chosen for nucleotide sequencing (Table 1) based on their higherfrequency of occurrence among localities. Because the taxonomy of the D. pulexcomplex is paradoxical, isolates from temperate populations of D. pulex and D.pulicaria in Europe (Cerny & Hebert, 1998) and North America (Hebert et al., 1993)were included in the analysis, including the eastern, western and polar NorthAmerican D. pulicaria lineages (Dufresne & Hebert, 1997). Unlike the arctic popu-lations, the taxonomy of the temperate isolates was confirmed prior to sequencingusing both morphological criteria and fixed allozyme differences (Hebert, 1995).Phylogenetic groups were later assigned species names based primarily on theplacement of these temperate isolates within clades. Daphnia melanica and D. arenatafrom North America (Hebert, 1995) were also included in the analysis.

DNA amplification and sequencing

Total genomic DNA was extracted from single animals or multiple animals fromclonal cultures using the Isoquick kit (Orca Research, Bothell, WA). A 2.1 kbfragment containing part of the mitochondrial ND4 and ND5 genes was amplifiedfrom this genomic template using the primers ND4-new: 5′–ACTCTTCAGT-AGCTCATATGA–3′ and ND5-new: 5′–AAGGAAGAAACCATATTAAACC–3′in a total reaction volume of 50 ll containing 10–100 ng of genomic DNA,0.3–0.5 lM of each primer, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 2 mM MgCl2,200 lM each of dGTP, dATP, dCTP and dTTP, 1% dimethyl sulfoxide and 2.5units of Taq polymerase (Boehringer Mannheim). The amplification reaction wasperformed in a Perkin-Elmer-Cetus thermal cycler (Model 480) under the followingconditions: one cycle of denaturation at 94°C for 1 min, and 35 cycles of denaturationat 92°C for 1 min, annealing at 50°C for 1 min, and extension at 72°C for 2 min.The amplified fragment was electrophoresed on a 1% agarose gel in TAE bufferand stained with ethidium bromide. The DNA fragments were visualized underlong-wave UV light, then purified from the gel using the Gene Clean II kit (Bioclean101). The sequence of 498 nt of this fragment, from the ND5 gene, was determinedusing the primer DpuND5b: 5′–GGGGTGTATCTATTAATTCG–3′. Twenty tofifty nanograms of the purified fragment was sequenced using 3 pmol of the primerand the ABI Prism TaqFS dye terminator kit (Perkin-Elmer). The sequencingreactions were analysed on an ABI 377 automated sequencer.


T 1. List of Daphnia included in the study, and their collection site

Taxon Area Site

DOF 1a District of Franklin, CAN Summerset IslandDOF 2 District of Franklin, CAN Summerset IslandDOF 3 District of Franklin, CAN Summerset IslandDOF 4 District of Franklin, CAN Bathurst IslandDOF 5a District of Franklin, CAN Devon IslandDOF 6 District of Franklin, CAN Summerset IslandDOF 7 District of Franklin, CAN Bathurst IslandDOF 8ab District of Franklin, CAN Summerset IslandDOF 9 District of Franklin, CAN Devon IslandDOF 10 District of Franklin, CAN Cornwallis IslandDOF 11a District of Franklin, CAN Cornwallis IslandDOF 12a District of Franklin, CAN Summerset & Devon IslandsDOF 13 District of Franklin, CAN Bathurst IslandDOF 14 District of Franklin, CAN Summerset IslandDOF 15 District of Franklin, CAN Bathurst IslandDOF 16 District of Franklin, CAN Devon IslandDOF 17 District of Franklin, CAN Devon IslandDOF 18 District of Franklin, CAN Cornwallis IslandDOF 19ab District of Franklin, CAN Summerset IslandESB 1 Eastern Siberia Kolyma DeltaESB 2a Eastern Siberia Wrangel IslandESB 3 Eastern Siberia Kolyma DeltaESB 4a Eastern Siberia NW IndigirkaESB 5 Eastern Siberia OlenekskyiESB 6 Eastern Siberia Wrangel IslandESB 7 Eastern Siberia Kolyma DeltaESB 8a Eastern Siberia Kolyma DeltaESB 9 Eastern Siberia Kolyma DeltaGER 1a Germany Grosser BinnenseeGER 2b Germany Grosser BinnenseeGRL 1 Greenland GodhavnGRL 2 Greenland Sondre StromfjordGRL 3a Greenland GodhavnGRL 4a Greenland GodhavnGRL 5 Greenland Uummannaq IslandICE 1 Iceland WesternICE 2 Iceland WesternICE 3 Iceland WesternICE 4a Iceland WesternIWA 1b Iowa, USA AmanaMAN 1 Manitoba, CAN ChurchillMAN 2 Manitoba, CAN ChurchillMAN 3a Manitoba, CAN ChurchillMAN 4a Manitoba, CAN ChurchillNOR 1 Norway FinmarkNOR 2 Norway FinmarkNOR 3 Norway FinmarkNWT 1ab Northwest Territories, CAN TuktoyaktukNWT 2b Northwest Territories, CAN TuktoyaktukNWT 3 Northwest Territories, CAN TuktoyaktukONT 1ab Ontario, CAN Redchalk LakeONT 2a Ontario, CAN Rondeau ParkORE 1ab Oregon, USA FlorenceORE 2ab Oregon, USA FlorenceORE 3b Oregon, USA Florence; ZoilORE 4ab Oregon, USA ZoilORE 5ab Oregon, USA FlorenceORE 6b Oregon, USA FlorenceORE 7b Oregon, USA Lab clone from M. Lynch

continued


T 1—continued

Taxon Area Site

ORE 8b Oregon, USA ZoilSAS 1 Saskatchewan, CAN Redberry LakeSAS 2 Saskatchewan, CAN Humbolt LakeSVL 1a SvalbardSVL 2a SvalbardSVL 3 SvalbardSVL 4 SvalbardSVL 5 SvalbardSVL 6a SvalbardSWI 1b Switzerland BaselSWI 2ab Switzerland BaselSWI 3ab Switzerland BaselSWI 4ab Switzerland BaselSWI 5ab Switzerland BaselWAS 1a Washington, USA Lake WashingtonWSB 1 Western Siberia Kachkovsky BayWSB 2 Western Siberia Kachkovsky BayWSB 3 Western Siberia Kola/KareliaWSB 4a Western Siberia Western Yamal PeninsulaWSB 5 Western Siberia Kachkovsky BayWSB 6 Western Siberia Western Yamal PeninsulaWSB 7a Western Siberia Kolguyev IslandWSB 8 Western Siberia Belyi Island

a Designates taxa used for preliminary cladistic analysis.b Designates taxa with no restriction site data.

Phylogenetic analysis

The 498 nt sequence of the ND5 gene, coding for 166 amino acids, was alignedfor all taxa by eye using the SeqApp v1.9 sequence editor (Gilbert, 1992). Estimatesof sequence divergence between all pairs of unique haplotypes were calculated usingthe Kimura two-parameter model (Kimura, 1980) in MEGA v1.02 (Kumar, Tamura& Nei, 1993). Phenetic analysis of the resulting divergence matrix was carried outusing the neighbour-joining (N-J) method of Saitou and Nei (1987) in MEGA.

Brower (1994) showed that the rates of nucleotide substitution within mitochondrialgenes of arthropods are relatively constant during the first few million years.However, this substitution rate is not specifically calibrated for particular genes, thuscomplicating historical accounts related to the origin of clades when comparingacross data sets. Nonetheless, his general estimate of 2.3% pairwise sequencedivergence per million years was used to approximate the times of divergence amongspecies whose sequence divergence did not exceed 8%. This clock differs only slightlyfrom the conventional mitochondrial sequence divergence rate of 2% per millionyears (Brown, George & Wilson, 1979).

A cladistic analysis of the phylogenetically informative sites was performed usingmaximum parsimony (MP) in PAUP v3.1.1 (Swofford, 1993). Because of the largenumber of sequences, it was impossible to complete a search for parsimoniouscladograms of all taxa. Therefore, representative taxa of all major groupings identifiedfrom the N-J tree were chosen to first verify that the branching patterns obtainedvia an abridged cladistic analysis were consistent with those of the N-J tree. Then,in an hierarchical fashion, relationships within monophyletic groupings were further


A

B

C

D

E

F

Figure 1. A hypothetical cladogram depicting the hierarchical phylogenetic method used to obtain astepwise agreement tree. Taxa from clades adjacent to the tail end of an arrow are used as a functionaloutgroup to resolve relationships among members within sister clades. The method is described indetail in the text.

analysed by successively choosing more closely related sister group taxa as functionaloutgroups (Fig. 1). More precisely, we selected a most parsimonious cladogram forthe whole data set according to the following steps. (i) Clades which occurred inboth the N-J and the abridged cladistic trees were identified, while disputed branchesbetween the trees were collapsed to form polytomies. (ii) Representatives of the sistergroup of each terminal clade were selected and then used as functional outgroupsto find all equally parsimonious trees describing relationships within these terminalclades (A–C, Fig. 1). (iii) A consensus tree was constructed for each set of equallyparsimonious trees found in step two. (iv) Representative taxa of the next sistergroup to pairs of sister clades were then selected and used as functional outgroupsto find the most parsimonious trees describing alliances among these clades (D–F,Fig. 1). (v) From among the most inclusive equally parsimonious trees found in stepfour (E & F, Fig. 1), a tree preserving the consensus topologies of internal cladesobtained from step three was chosen. Thus, a most parsimonious cladogram includingall taxa, with branching pattern and tree length consistent with (or better than) eachfunctional ingroup analysis was chosen as the best tree. We call this cladogram astepwise agreement tree. The rationale behind tree-building in this fashion is bothpractical and ideal; the method provides an unbiased means by which to draw asingle tree from among multiple equally parsimonious cladograms, while nucleotidecharacters are more likely to be correctly polarized because the probability ofmultiple substitutions per site is reduced when taxa that are relatively recentlydiverged from the ingroup taxa are used to root the trees.

Homoplasy was measured using both the consistency and retention indices (CIand RI: see Wiley et al., 1991), even though they were sometimes inconsistent acrossdata sets (Archie, 1989; Naylor & Kraus, 1995). After generating a stepwise agreementtree, the confidence in each clade was assessed by (i) using PAUP to plot 100 000trees drawn at random from all possible topologies as a function of tree length toobtain an estimate of the g1 skewness statistic (Hillis & Huelsenbeck, 1992); (ii) usingAutoDecay v2.9.6 (Eriksson, 1995) to evaluate the decay index (Bremer, 1994); and


(iii) using Random Cladistics v3.0 (Siddall, 1995a) for bootstrapping (Felsenstein,1985) and measuring the jackknife monophyly index (Siddall, 1995b).

RESULTS

ND5 gene sequence diversity

Seventy-nine unique sequences were found among the 82 isolates listed in Table1, as isolates with different restriction site profiles for the 2.1 kb fragment had thesame sequence for the 498 nt fragment. A total of 182 nt positions were polymorphicand variation at 40 of these sites led to amino acid substitutions. As many as fourdifferent amino acids occurred at some of the polymorphic sites. Further, 153 ntwere informative for cladistic analysis. Corrected pairwise sequence divergenceestimates ranged from 0.01% to 19%. Transitional saturation of sequences amongingroup taxa was not evident and their base composition was similar (test forhomogeneity, v2=40.4, df=234, P>0.9).

Phenetic analysis of sequence divergence in the ND5 gene

The N-J tree constructed from the matrix of sequence divergence shows that thepulex complex can be divided into three major groups (A–C, Fig. 2). Group A iscomprised of six distinct clusters of mitochondrial DNA haplotypes. Taxa that clusteron branches one, two and three (Fig. 2) correspond to lineages that have beenidentified as D. pulicaria based on morphology and allozymes (Dufresne & Hebert,1997). One lineage consists of isolates found in Greenland, Iceland and the polarregions of Europe and Canada (Fig. 2, cluster 1). Another lineage contains isolatesinhabiting western temperate regions of North America (cluster 2), while the mostbasal D. pulicaria lineage consists of isolates with an eastern (eastern Canada,Greenland, Iceland) distribution (cluster 3). The other three clusters within GroupA correspond to D. melanica haplotypes (cluster 4), D. middendorffiana haplotypes(cluster 5), and D. pulex haplotypes (cluster 6). The latter clade includes all isolatesfrom North America that have been identified as D. pulex on the basis of morphologyand allozymes, as well as isolates from Eurasia. Thus, this clade was designated‘panarctic D. pulex’. Surprisingly, the D. arenata isolates (ORE 5, ORE 6) clusterwithin panarctic D. pulex. The sister group to D. pulicaria within this N-J tree is D.melanica, the only temperate Daphnia species known to possess cuticular pig-mentation—a trait which is also common in populations of the arctic species, D.middendorffiana and D. tenebrosa (Hebert, 1995).

Group B is subdivided into two genetically divergent clusters of haplotypes. Onecluster (Fig. 2, cluster 7) consists of isolates found in western Siberia and Svalbard,Norway, but also includes isolates that have traditionally been identified as temperateEuropean D. pulicaria (Dufresne, 1995). The other cluster (8), which is restricted tothe Arctic, is very broadly distributed in both North America and Eurasia and hasbeen identified as D. tenebrosa, based on morphology and allozymes (Dufresne &Hebert, 1995). Group C consists solely of isolates of D. pulex from Europe (SWI)and occupies the most basal position on the tree. Allozyme evidence indicated that


Sequence Divergence

0 1%

C 9

8

SWI 5SWI 4

SWI 3SWI 2

SVL 5SVL 6

SVL 4ESB 9

ESB 8ESB 7ESB 6

ESB 5ESB 4

SVL 3ESB 3

ESB 2WSB 8

WSB 7MAN 1MAN 3

MAN 2MAN 4

ESB 1SVL 2

GER 2,SWI 1GER 1

WSB 6WSB 5

WSB 4NWT 3

ORE 8, IWA 1ONT 2

ORE 7WSB 3

GRL 5

GRL 4

WSB 2ORE 6

ORE 5WSB 1

NOR 3ICE 4

NOR 2

DOF 19DOF 18

NWT 2NWT 1DOF 17

DOF 16DOF 15

DOF 12DOF 14

DOF 13ORE 3

ORE 2ORE 4

ORE 1ICE 3

ICE 2GRL 3

ONT 1WAS 1

SAS 1SAS 2

DOF 11DOF 10

NOR 1GRL 2

SVL 1ICE 1

DOF 9DOF 8

DOF 7DOF 6

DOF 5DOF 3, GRL 1

DOF 4DOF 2DOF 1

1

2

3

4

5

6

7

A

B

European D. pulex

D. tenebrosa

European D. pulicaria

panarctic D. pulex

D. middendorffiana

D. melanica

eastern D. pulicaria

western D. pulicaria

polar D. pulicaria

Figure 2. A neighbour-joining tree of ND5 nucleotide variation in the D. pulex complex. Letters onbranches denote the three major groupings within the complex. Numbers indicate either Daphniaspecies or distinct mitochondrial lineages within a species. The nine lineages identified in this studyare shown on the right side of the figure.

these taxa are typical European D. pulex (Hebert, Schwartz & Hrbacek, 1989). Thus,isolates from Europe that have been identified as D. pulex based on morphologybelong to two very divergent clades (see above).

The branches leading to D. melanica, D. middendorffiana and D. pulicaria are veryshort, suggesting that these species diverged within a relatively brief period of time(Table 2). Brower’s (1994) estimate of 2.3% sequence divergence per million yearssuggests that panarctic D. pulex diverged from the ancestor of (D. melanica, D.middendorffiana, D. pulicaria) about 2.2 million years ago, while the split among these


T 2. Mean sequence divergence of the ND5 mitochondrial gene between phylogenetic groupsidentified with the neighbour-joining method (see Fig. 2). The mean sequence divergence within majorgroups is on the diagonal. The mean sequence divergence between major groups is shown above thediagonal. The estimates were corrected using the Kimura (1980) two-parameter model of molecular

evolution

1 2 3 4 5 6 7 8 9

1 0.010 0.024 0.025 0.036 0.037 0.051 0.158 0.148 0.1782 0.007 0.035 0.047 0.044 0.060 0.168 0.160 0.1763 0.005 0.035 0.029 0.049 0.154 0.147 0.1664 0.015 0.037 0.054 0.158 0.145 0.1825 0.009 0.044 0.160 0.156 0.1806 0.016 0.155 0.158 0.1727 0.014 0.075 0.1658 0.028 0.1759 0.012

latter three species occurred between 1.6 and 1.4 million years ago (Table 2). Usingthe same molecular clock, the two subgroups (clusters 7 and 8) of group B appearto have diverged from one another on the order of 3.2 million years ago. Sequencedivergence between groups A and B is 15.3% while the amino acid divergence is5.7%. So far, there is no calibrated ND5 clock to date the more ancient divergencebetween these two clades.

Cladistic analysis of sequence divergence in the ND5 gene

The phenetic analysis suggested that European D. pulex is the most divergentgroup in the pulex complex. Furthermore, additional 12S rDNA sequence informationconfidently placed European D. pulex at the base of the species complex whenincluded in the Colbourne and Hebert (1996) 12S phylogeny of North Americanmembers of the genus (unpublished data). Consequently, this group was used toroot preliminary cladograms constructed using only a few taxa from each majorlineage (see Table 1). A heuristic search using only informative characters (IC) withinthe ND5 sequences found three most parsimonious trees (length=275; CI=0.60;RI=0.88) that differed from each other only at nodes within the polar D. pulicariaclade. These trees (not presented) show distinct clades corresponding to all the majorgroups and lineages outlined by the phenetic tree, except western D. pulicaria, whichgrouped within polar D. pulicaria. Their topologies are similar to the phenogram,except that D. melanica was the sister clade to D. middendorffiana. The g1 statistic (-0.35) of 100 000 random trees indicates strong phylogenetic signal from the totaldata (P<0.01; Hillis & Huelsenbeck, 1992).

Phylogenetic relationships among all members of group B (Fig. 3) were resolvedby performing three stepwise heuristic searches using various functional ingroupsand outgroups. All characters were unordered and equally weighted. The firstsearch, using three European D. pulicaria as an outgroup (WSB 4, GER 1, SVL 2)found 10 most parsimonious trees describing the relationships only among membersof the D. tenebrosa clade (length=85, from 55 IC). The consensus topology from thisanalysis matched with trees obtained from searches using either three taxa fromgroup A (DOF 3, DOF 16, ORE 8) or the European D. pulex as outgroups to


Eu

rope

an D

. pu

lica

ria

DOF 3. GRL 1

term

inal

D. t

eneb

rosa

cla

de

D. t

eneb

rosa

DOF 16

ORE 8, IWA 1

WSB 4

WSB 5

WSB 6

GER 1

GRF 2, SWI 1

SVL 2

ESB 1

MAN 4

MAN 2

MAN 3

MAN 1

WSB 7

ESB 4

ESB 5

WSB 8

ESB 2

ESB 3

SVL 3

ESB 6

ESB 7

ESB 8

ESB 9

SVL 4

SVL 5100*84*3

61*34*193*20 *1

97*32*2

73*26*1

100*31*2

73*8*0

100*55*3

100*68*4

100*51*2 100*77*3

100*95*8100*97*2

100*97*2100*96*5

100*64*2

100*82*2

100*80*41

99*46*2

SVL 6

Figure 3. A stepwise agreement tree for the taxa in group B of the Daphnia pulex complex (length 171,CI=0.68, RI=0.87 with 102 characters). Characters were unordered and equally weighted. Stepwisecladistic trees were resolved from heuristic searches; taxa were added randomly with 25 replications,with MULPARS and steepest descent options invoked and with branch swapping by the tree bisection-reconstruction algorithm. The jackknife monophyly index is printed at each node, followed by bootstrappercentages from 1000 pseudo-replicates and the decay index. Aligned sequences are available uponrequest.

resolve relationships within all of group B. The former analysis uncovered 20 mostparsimonious trees (length=172, from 102 IC), while the latter found 40 equallyparsimonious trees (length=170; CI=0.71; RI=0.91, from 113 IC). The stepwiseagreement tree in Figure 3 confirms the existence of two major lineages. As in thephenetic analysis, haplotypes previously identified as European D. pulicaria (Germanyand Switzerland) form a monophyletic clade with haplotypes from western Siberiaand Svalbard. The D. tenebrosa clade contains very closely related haplotypes (the‘terminal D. tenebrosa clade’) that cannot be confidently resolved due to the smallnumber of characters distinguishing its members, but which occur in Svalbard andthroughout Siberia. In general, the Canadian members of the D. tenebrosa cladeoccur in basal positions. The g1 statistic for this group is −0.94 (P<0.01).

Three stepwise heuristic searches for cladograms describing relationships among


panarctic D. pulex

D. middendorffiana

D. melanica

western D. pulicaria

polar D. pulicaria

eastern D. pulicaria

DOF 1DOF 2DOF 3. GRL 1DOF 4DOF 5DOF 6DOF 7NOR 1SVL 1GRL 2ICE 1DOF 8DOF 9DOF 10DOF 11SAS 1SAS 2WAS 1ONT 1ICE 2GRL 3ICE 3ORE 1ORE 4ORE 3ORE 2DOF 13DOF 15DOF 16DOF 17NWT 1NWT 2DOF 18DOF 19DOF 12DOF 14NOR 2ICE 4NOR 3WSB 1ORE 5ORE 6WSB 3ORE 8, IWA 1NWT 3ONT 2ORE 7WSB 2GRL 5GRL 4WSB 4WSB 7SVL 5

100*46*1*1

100*37*199*37*1

100*48

100*67*1

100*44*1

100*66*1

100*53*1

100*71*2100*54*39

100*71*2

100*28*1

23*26*0

23*16*0

100*68*1

100*88*na

100*95*4

100*74*2

100*88*223*15*0

100*33*176*18*0

100*90*5

100*96*3

100*95*3

99*61*1

100*53*2

100*56*na

100*56*1

79*27*0

Figure 4. A stepwise agreement tree for taxa in group A of the Daphnia pulex complex (length 220,CI=0.60, RI=0.86 with 110 characters). Stepwise trees were resolved as described in Fig. 3. Becauseof computer memory limitations, multiple-tree hennig (mh) overflow was set at 100 trees and branchbreaking (bb) overflow was set at 1784 trees during the bootstrap analysis using Random Cladisticsv3.0.

taxa in group A (Fig. 4) were also performed using various functional ingroups andoutgroups. The first search using three outgroup taxa from panarctic D. pulex (GRL4, ORE 5, ORE 8) found 10 most parsimonious trees (length=97, from 58 IC),which placed D. middendorffiana basal to a (D. melanica (eastern D. pulicaria (polar D.pulicaria, western D. pulicaria))) clade. A second search, which used taxa DOF 3,ORE 4 and DOF 16 as outgroups to show relationships only among members ofthe panarctic D. pulex clade, revealed six equally parsimonious trees (length=39,from 25 IC). The 70% majority rule consensus topologies from each analysis werefound to occur among many parsimonious trees obtained from a final and inclusivesearch for group A cladograms using three group B isolates as an outgroup (WSB4, WSB 7, SVL 5). This last search uncovered 800 equally parsimonious trees(length=221; CI=0.60; RI=0.86, for 110 IC). As in the phenetic analysis, thestepwise agreement tree unambiguously placed panarctic D. pulex (a paraphyletic


clade with respect to D. arenata) basal to the other lineages within group A (Fig. 4).However, the conclusion that D. middendorffiana is basal to D. melanica and D. pulicariais somewhat contentious, because 640 of the 800 most parsimonious trees joined D.melanica to D. middendorffiana. Even so, all phylogenetic analyses involving less distantlyrelated species as an outgroup provided a clear signal against this alternative alliance.Moreover, a re-analysis of data from Van Raay and Crease (1995) on sequencedivergence in the control region (a more rapidly evolving region than ND5), usingpanarctic D. pulex as the outgroup instead of D. melanica, unambiguously groupedD. melanica with D. pulicaria and not with D. middendorffiana (data not shown). Wetherefore chose the topology of Figure 4 as our phylogenetic hypothesis for groupA. Its g1 statistic is −0.73 (P<0.01).

Other than the distinctness of eastern D. pulicaria, relationships within the D.pulicaria clade are not well resolved. The western haplotypes do ally with one another,but polar D. pulicaria is paraphyletic with regard to this group. With the exceptionof a single haplotype from Greenland, the largest polytomy in this clade containsvery closely related haplotypes from the District of Franklin in Arctic Canada. Theother polar D. pulicaria lineage contains haplotypes from across the Arctic, includingCanada, Iceland, Greenland and Svalbard. Within the panarctic D. pulex clade,ORE 7 was classified as D. arenata and ORE 8 was classified as D. melanica, basedon allozyme electrophoresis. These are only a few examples of taxa with introgressedgenomes, a situation which is not uncommon in this group (Crease et al., 1997).Weider et al. (in prep.) will present, in full detail, the results from a circumarcticallozyme study.

Confidence in clades

The JMI, the bootstrap percentages of 1000 pseudo-replicates and the decayindex for the tree-branching for group B are shown in Figure 3. There is strongsupport for the monophyly of both major clades (7 and 8). The low bootstrap valuesat the roots of lower branches simply reflect the uncertain placement of ESB 1 andMAN 4 at the tree’s base (10 of 20 equally parsimonious trees and 47% of thepseudo-replicates even suggest their cladistic alliance). Confidence in the branchingpatterns within the European D. pulicaria (clade 7) is high, but it abates at moredistal nodes within the D. tenebrosa clade. This finding is not surprising, as each nodeabove (ESB 4, ESB 5) is supported by a single nucleotide character—evidence forthe recent origin of the terminal D. tenebrosa clade.

Results from randomization routines applied to the group A cladogram are shownin Figure 4. In general, the monophyly of all major clades is well supported, exceptfor polar D. pulicaria (clade 1). Although the monophyly of the western lineage ofD. pulicaria is strongly preserved, cladistic analysis was unable to discriminate it frompolar members. The other contentious nodes, which define relationships amonggroups 1–5, result from the uncertain placement of D. melanica within the phylogeny.As discussed earlier, there is some cladistic support for its closer alliance with D.middendorffiana instead of D. pulicaria.

Interestingly, a number of amino acids are conserved and diagnostic for majorclades, supporting relationships identified by the nucleotide characters. For instance,two of the 40 variable amino acid characters are diagnostic for European D. pulex(cluster 9, Fig. 2). Two other amino acid changes are diagnostic for either group A


or group B, while another site has diferent amino acids for each of the two majorclades in group B. All isolates of D. tenebrosa, except the three MAN taxa and ESB1 share another amino acid. Diagnostic characters are also present for the (D.middendorffiana (D. melanica, D. pulicaria)) clade, for western and polar D. pulicaria, andfor the DOF 1 to DOF 7 polytomy in the polar D. pulicaria clade.

DISCUSSION

There is considerable uncertainty as to how cladocerans speciate. Both the insularnature of their habitats and their ability to invade new ponds and lakes by thedispersal of a single resistant egg suggest that founder effect speciation should befrequent (Lynch, 1985). Rare long-distance colonization or continuous range ex-pansion of populations via the passive dispersal of eggs can also initiate speciation(Mayr, 1942). Physical and biological regimes differ immensly among freshwaterhabitats, and provide opportunities for adaptive diversification (Colbourne, Hebert& Taylor, 1997). Although the evolutionary history of Cladocera likely implicates acombination of events, direct evidence for the importance of any of the precedingfactors in speciation is meager. However, it is interesting that the three daphniidcomplexes (pulex, longispina, carinata), which respectively dominate pond and lakeenvironments in the northern hemisphere and the inland waters of Australia, areall actively engaged in hybridization and introgression. By contrast, many of theminor and narrowly endemic elements of the daphniid fauna on these continentsare uninvolved in these processes (Hebert, 1995; Taylor, Hebert & Colbourne, 1996;Hebert, 1998). Taxonomic confusion is not only particularly acute in the dominantelements of the fauna, but this relationship may be causal if hybridization andintrogression spurred morphological and ecological mini-radiations that enableparticipant taxa to gain dominance (Schwenk & Spaak, 1997).

This study has examined the patterning of ND5 sequence divergence in the pulexcomplex, the assemblage of daphniids which dominates aquatic habitats in thenorthern hemisphere. Past failures to establish an accepted taxonomic system forthis group have led some researchers to conclude that efforts to impose taxonboundaries on the complex are misplaced (Lehman et al., 1995). However, thepresent results indicate that this complex is not simply a syngameon whose membersshow genetic continuity. Instead, the nature of sequence variation at ND5 providestwo lines of support for the recognition of several species in the complex. First, theextent (as much as 19%) of ND5 sequence divergence among lineages of the pulexcomplex is greater than that typical of single species. Second and more importantly,genotypes cluster into a limited number of nodes. While a study of the phylogeneticconcordance among other mitochondrial genes, as well as nuclear loci, is neededto critically establish species boundaries, the present results do suggest that theexercise is worth undertaking.

Despite its examination of variation in only a single gene, the present studyprovides new insights concerning the patterning of biological diversity in the pulexcomplex. Our investigation establishes the presence of three major clades (A, B, C)and the subdivision of the former pair of clades into eight lineages that coincidewith recognized morphological and biogeographical assemblages. Group C, whichappears ancestral to the whole complex, includes only populations of D. pulex from


Central Europe. Group B includes two lineages (7, 8) with the first consisting ofpopulations of D. pulicaria from Europe, while the second includes D. tenebrosa froma wide range of arctic sites. The final group (A) is more diverse, including sixdifferent lineages. One of these lineages (6) includes populations of D. pulex fromNorth America and Northern Europe, as well as those of D. arenata. Three otherlineages (1–3) are characteristic of D. pulicaria populations from different geographicalregions in North America, while the remaining lineages (4, 5) constitute populationsof D. melanica and D. middendorffiana respectively. Some isolates of the latter species,identified on the basis of allozymes, possess mitochondrial haplotypes from lineages1–3, an expected result because of its origin through hybridization with NorthAmerican populations of D. pulicaria (Dufresne & Hebert, 1994).

Application of the arthropod mtDNA clock (Brower, 1994) indicates that themembers of group A diverged within the last 2.2 million years, coinciding roughlywith the onset of the Pleistocene about 2.4 million years ago (Webb & Bartlein,1992). Although the error associated with these estimates is unknown, a study ofrestriction site variation in the entire mitochondrial genome provided estimates thatare consistent with our result (Van Raay & Crease, 1995). Thus, the vicariant eventsassociated with multiple cycles of glacial advances and retreats during the Pleistocenemost likely provided the impetus for diversification within this group (see Hewitt,1996). In contrast, the two major lineages in group B may have been in existencelong before Pleistocene ice sheets began to expand. Yet, while the ice ages areshown to be an important time in the phylogenetic history of the pulex complex,large climatic oscillations during the Tertiary may have similarly impacted the floraand fauna in the northern hemisphere (Stanley, 1987).

The present study has revealed serious flaws in past taxonomic assignments formembers of the complex. Two of the taxa (D. pulex, D. pulicaria), thought to bebroadly distributed in the northern hemisphere, are now shown to be endemic tosingle continents. Hence this study shows that populations of ‘D. pulex’ from NorthAmerica and Eurasia show 17% sequence divergence at ND5 from some of theirEuropean counterparts. This pair of taxa also shows 6% divergence at 12S rDNA(unpublished data), suggesting their separation some 3 million years ago andsupporting their recognition as distinct species. A parallel situation occurs in D.pulicaria, whose populations show only modest divergence across North America(<3.5%), but 15% sequence divergence at ND5 from their European counterparts.Indeed, the European populations show much greater similarity to D. tenebrosa (cluster8, Fig. 2) and may well have originated from populations of this species during thePliocene. These conclusions, based on mtDNA analysis, gain independent supportfrom allozyme studies, which have confirmed the marked genetic divergence betweenpopulations of D. pulex and D. pulicaria from North America and Central Europe(Cerny & Hebert, 1998). This taxonomic reappraisal provides an explanation forsome puzzling observations. For example, F1 hybrids between D. pulex and D.pulicaria are common in North America, but absent in Europe. The present resultsexplain the differential formation of hybrids by showing that North Americanlineages of these two species are closely allied, belonging to a single major clade (A),while European lineages sharing these species epithets are very divergent, belongingto different major clades (B, C).

The present analysis also raises questions concerning the taxonomic status of D.arenata, a group restricted to ponds west of the Cascade Mountains in Oregon. TheND5 sequence from this species showed little divergence from those of North


American D. pulex. In some cases, this similarity may reflect the past occurrence ofmtDNA introgression. Certainly, many populations of D. pulicaria in North Americaalso contain mtDNA genomes derived through introgression with D. pulex (Hebertet al., 1989). However, D. arenata does show reasonably marked allozyme divergencefrom D. pulex and its predominant mtDNA haplotypes do form a monophyleticgroup (Crease et al., 1997) suggesting that it may merit sustained recognition as adistinct species. The paraphyly of panarctic D. pulex may simply suggest that veryrecent glacial periods have seeded the speciation of this clade from a peripheral D.pulex population that was isolated by the Cordilleran ice sheet (see Avise, 1994).

By determining the sequence differences which distinguish the nine major lineagesin the pulex complex, the present study facilitates the identification of restriction siteswithin the ND4–ND5 gene fragment that allow discrimination among groups. Theexploitation of such markers makes it possible to rapidly assign isolates to a specificlineage, allowing characterization of the geographic distribution of individual lineages,and enabling detailed phylogeographic studies on the complex. While a companionpaper examines this issue in more detail (Weider et al., in review), the present studydoes provide some general insights concerning the phylogeography of the pulexcomplex. Because of their dessication resistant eggs, it has long been assumed thatthe dispersal abilities of cladocerans are strong (see Talling, 1951). The limiteddiversification, noted in this study, among conspecific lineages of the pulex complexfrom single continents reinforces this conclusion suggesting that gene flow is sub-stantial. However, the extent of gene flow among daphniid populations from differentcontinents seems more variable. Two of the three major clades in the pulex complexdo occur in both North America and Eurasia, but only four of their nine componentlineages have Holarctic distributions. Moreover, two patterns of dispersal areapparent. Lineages of the pulex complex characteristic of polar regions are invariablypresent in both the Nearctic and Palearctic regions, but their temperate zonecounterparts are restricted to a single region. Polar taxa apparently serve as thephylogenetic bridge between the faunas on these continents, reflecting both therelative proximity of polar habitats and the lack of broad zones of unfavourablehabitat which must be crossed in colonizing a new continent. By contrast, taxa fromthe temperate zone must cross either oceanic barriers or polar environments toextend their distribution from one continent to another. The pulex complex is atypicalof most other cladocerans because its members are so abundant in the Arctic. Manyother species complexes of daphniids are restricted to the temperate zone (Hebert& Finston, 1996b; Taylor, Finston & Hebert, 1998), suggesting that faunal exhangein these groups will be very rare, a conclusion reinforced by the endemicity of alltheir component taxa to a single continent. The results suggest that, with theexception of a few polar taxa, gene flow among cladoceran lineages from differentcontinents is restricted. Sweepstakes dispersal may occur, but such interchanges areundoubtedly so rare that gene pool divergence and speciation will follow when thecolonizers persist.

The present results extend our understanding of the processes important inevolutionary divergence of daphniids. There is now strong evidence that rates ofmolecular and morphological evolution in Daphnia are discordant. Following aninitial burst of divergence shortly after its origins in the Mesozoic, morphologicalchange in the genus slowed (Colbourne et al., 1997). Taxonomic confusion in the pulexcomplex reflects, in part, the lack of morphological innovation which characterizes thegenus and its commitment to convergent evolution. Hence, genetically divergent


‘D. pulex’ from North America and Europe share morphological features that likelyreflect convergence linked to their occupancy of similar habitats. However, theresolution of taxonomic boundaries in the pulex complex is further complicated bythe involvement of its members in hybridization, introgression, asexuality andpolyploidy. Despite these factors, the present study suggests that genetic discontinuitiesdo exist and that the complex can be partitioned into a modest number of lineageswith biogeographic and morphological integrity. This diversification seems to haveoccurred over a relatively brief interval, suggesting that further genetic charac-terization of the pulex complex may provide an understanding of the processesimportant in countering the evolutionary stasis that has constrained daphniiddiversification since the Mesozoic.

ACKNOWLEDGEMENTS

Logistic support in the Arctic was provided by J. MacDonald at the EasternArctic Scientific Resource Center (Igloolik, N.W.T.), by L. Skytte at the Arctic FieldStation (University of Copenhagen in Godhavn, Greenland) and by the staff at thePolar Continental Shelf Project (Resolute and Tuktoyaktuk, N.W.T.). We are gratefulto the Swedish Polar Research Committee and the Russian Academy of Sciencefor organizing the Tundra Ecology 1994 expedition to Siberia, and to V. Abakumovat the Institute for Global Climate and Ecology (Moscow) for organizing the samplingtrip to Karelia and the Kola Peninsula. Thanks to J. Bull, A. Cox, T. Little, H.Stibor and D. Taylor for their field assistance, and to N. Lehman, T. Little and M.Pfrender for donating samples. Also thanks to A. Boldt, E. Geißler, A. Holliss, C.Rowe and I. Senkpiehl for their technical assistance in the laboratory. This workwas funded by NSERC (Canada), the Ontario Graduate Fellowship Program, theNorthern Studies Training Program (Canada), the Polar Continental Shelf Program(Canada), the NATO Collaborative Research Grants Program, the Max-Planck-Gesellschaft (Germany), the Norwegian Polar Research Institute, the NorwegianResearch Council, the Swedish Polar Research Council, and the Russian Academyof Sciences. We thank M. Lynch and an anonymous reviewer for helpful commentson the manuscript.

REFERENCES

Archie JW. 1989. Homoplasy excess ratios: new indices for measuring levels of homoplasy inphylogenetic systematics and a critique of the consistancy index. Systematic Zoology 38: 253–269.

Avise JC. 1994. Molecular markers, natural history and evolution. New York: Chapman and Hall.Beaton MJ, Hebert PDN. 1988. Geographical parthenogenesis and polyploidy in Daphnia pulex.

American Naturalist 132: 837–845.Bremer K. 1994. Branch support and tree stability. Cladistics 10: 295–304.Brower AVZ. 1994. Rapid morphological radiation and convergence among races of the butterfly

Heliconius erato inferred from patterns of mitochondrial DNA evolution. Proceedings of the NationalAcademy of Sciences USA 91: 6491–6495.

Brown WM, George Jr. M, Wilson AC. 1979. Rapid evolution of animal mitochondrial DNA.Proceedings of the National Academy of Sciences USA 76: 1967–1971.

Cerny M, Hebert PDN. 1998. Intercontinental allozyme differentiation among four Holarctic Daphniaspecies. Limnology and Oceanography: in press.


Colbourne JK, Hebert PDN. 1996. The systematics of North American Daphnia (Crustacea:Anomopoda): a molecular phylogenetic approach. Philosophical Transactions of the Royal Society of LondonB 351: 349–360.

Colbourne JK, Hebert PDN, Taylor DJ. 1997. Evolutionary origins of phenotypic diversity inDaphnia. In: Givnish TJ, Sytsma KJ, eds. Molecular evolution and adaptive radiation. Cambridge:Cambridge University Press, 163–188.

Crease TJ, Lee S-K,Yu S-L, Spitze K, Lehman N, Lynch M. 1997. Allozyme and mtDNAvariation in populations of the Daphnia pulex complex from both sides of the Rocky Mountains.Heredity 79: 242–251.

Crease TJ, Lynch M, Spitze K. 1990. Hierarchical analysis of population genetic variation inmitochondrial and nuclear genes of Daphnia pulex. Molecular Biology and Evolution 7: 444–458.

Crease TJ, Stanton DJ, Hebert PDN. 1989. Polyphyletic origins of asexuality in Daphnia pulex. II.Mitochondrial-DNA variation. Evolution 43: 1016–1026.

Darwin C. 1859. On the origin of species by means of natural selection. London: John Murray.De Meester L. 1996. Local genetic differentiation and adaptation in freshwater zooplankton

populations: patterns and processes. Ecoscience 3: 385–399.Dufresne F. 1995. Genetic relationships among arctic Daphnia: implications for the origins and

evolutionary significance of polyploidy. Unpublished D. Phil Thesis, University of Guelph.Dufresne F, Hebert PDN. 1994. Hybridization and origins of polyploidy. Proceedings of the Royal

Society of London B 258: 141–146.Dufresne F, Hebert PDN. 1995. Polyploidy and clonal diversity in an arctic cladoceran. Heredity

75: 45–53.Dufresne F, Hebert PDN. 1997. Pleistocene glaciations and polyphyletic origins of polyploidy in an

arctic cladoceran. Proceedings of the Royal Society of London B 264: 201–206.Eriksson T. 1995. AutoDecay v.2.9.8. Program available via the World Wide Web at http://www.botan.

su.se/systematik/folk/Torsten.html.Felsenstein J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution

39: 783–791.Gilbert D. 1992. SeqApp version 1.9a. A multiple sequence editor for Macintosh computers. Published electronically

on the Internet by ftp at fly.bio.indiana.edu.Hebert PDN. 1995. The Daphnia of North America: an illustrated fauna. CD-ROM. Distributed by the

author. Department of Zoology, University of Guelph.Hebert PDN. 1998. The Daphnia of Australia: an illustrated fauna. CD-ROM. Distributed by the author.

Department of Zoology, University of Guelph.Hebert PDN, Beaton MJ, Schwartz SS, Stanton DJ. 1989. Polyphyletic origins of asexuality in

Daphnia pulex. I. Breeding-system variation and levels of clonal diversity. Evolution 42: 1004–1015.Hebert PDN, Finston TL. 1996a. Genetic differentiation in Daphnia obtusa: a continental perspective.

Freshwater Biology 35: 311–321.Hebert PDN, Finston TL. 1996b. A taxonomic reevaluation of North American Daphnia (Crustacea:

Cladocera). II. New species in the Daphnia pulex group from the south-central United States andMexico. Canadian Journal of Zoology 74: 632–653.

Hebert PDN, Schwartz SS, Hrbacek J. 1989. Patterns of genotypic diversity in CzechoslovakianDaphnia. Heredity 62: 207–216.

Hebert PDN, Schwartz SS, Ward RD, Finston TL. 1993. Macrogeographic patterns of breedingsystem diversity in the Daphnia pulex group. I. Breeding systems of Canadian populations. Heredity70: 148–161.

Hebert PDN, Wilson CC. 1994. Provincialism in plankton: endemism and allopatric speciation inAustralian Daphnia. Evolution 48: 1333–1349.

Hewitt GM. 1996. Some genetic consequences of ice ages, and their role in divergence and speciation.Biological Journal of the Linnean Society 58: 247–276.

Hillis DM, Huelsenbeck JP. 1992. Signal, noise, and reliability in molecular phylogenetic analyses.Journal of Heredity 83: 189–195.

Hrbacek J. 1987. Systematics and biogeography of Daphnia species in the northern temperate region.Memorie Dell’Istituto Italiano Di Idrobiologia 45: 37–76.

Innes DJ, Hebert PDN. 1988. The origin and genetic basis of obligate parthenogenesis in Daphniapulex. Evolution 42: 1024–1035.

Kimura M. 1980. A simple method for estimating evolutionary rate of base substitutions throughcomparative studies of nucleotide sequences. Journal of Molecular Evolution 16: 111–120.


Kumar S, Tamura K, Nei M. 1993. MEGA: Molecular Evolutionary Genetics Analysis. version 1.02.Distributed by the authors. Pennsylvania State University, University Park, PA.

Lehman N, Pfrender ME, Morin PA, Crease TJ, Lynch M. 1995. A hierarchical molecularphylogeny within the genus Daphnia. Molecular Phylogenetics and Evolution 4: 395–407.

Lynch M. 1985. Speciation in the Cladocera. Verhandlungen Internationalen Vereinigung fur Theoretische undAngewandte Limnologie 22: 3116–3123.

Maddison WP, Maddison DR. 1992. MacClade version 3.0.4. Analysis of phylogeny and character evolution.Sunderland; Sinauer.

Mayr E. 1942. Systematics and the origin of species, from the viewpoint of a zoologist. New York: ColumbiaUniversity Press.

Mayr E. 1963. Animal species and evolution. Cambridge: Belknap Press of Harvard University Press.Naylor G, Kraus F. 1995. The relationship between s and m and the retention index. Systematic

Biology 44: 559–562.Saitou N, Nei M. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic

trees. Molecular Biology and Evolution 4: 406–425.Schwenk K, Spaak P. 1997. Ecology and genetics of interspecific hybridization in Daphnia. In: Streit

B, Stadler T, Lively CM, eds. Evolutionary Ecology of Freshwater Animals. Basel: Birkhauser Verlag,199–229.

Siddall ME. 1995a. Another monophyly index. Revisiting the jackknife. Cladistics 11: 33–56.Siddall ME. 1995b. Random Cladistics. version 3.0. Published electronically on the Internet by ftp at

zoo.toronto.edu.Stanley SM. 1987. Extinction. New York: Scientific American Library, W.H. Freeman.Swofford DL. 1993. Phylogenetic Analysis Using Parsimony. version 3.1.1. Distributed by the author. Illinois

Natural History Survey, Champaign.Talling JF. 1951. The element of chance in pond populations. The Naturalist 1951: 157–170.Taylor DJ, Finston TL, Hebert PDN. 1998. Biogeography of a widespread freshwater crustacean:

pseudocongruence and cryptic endemism in the North American Daphnia laevis complex. Evolution:in press.

Taylor DJ, Hebert PDN, Colbourne JK. 1996. Phylogenetics and evolution of the Daphnia longispinagroup (Crustacea) based on 12S rDNA sequence and allozyme variation. Molecular Phylogenetics andEvolution 5: 495–510.

Van Raay TJ, Crease TJ. 1995. Mitochondrial DNA diversity in an apomictic Daphnia complexfrom the Canadian High Arctic. Molecular Ecology 4: 149–161.

Ward RD, Bickerton MA, Finston T, Hebert PDN. 1994. Geographical cline in breeding systemsand ploidy levels in European populations of Daphnia pulex. Heredity 73: 532–543.

Webb T, Bartlein PJ. 1992. Global changes during the last 3 million years: climatic controls andbiotic responses. Annual Review of Ecology and Systematics 23: 141–173.

Weider LJ, Hobœk A. 1997. Postglacial dispersal, glacial refugia, and clonal structure in Russian/Siberian populations of the arctic Daphnia pulex complex. Heredity 78: 363–372.

Wiley EO, Siegel-Causey D, Brooks DR, Funk VA. 1991. The compleat cladist. A primer of phylogeneticprocedures. University of Kansas, Special Publications 19: 1–158.

Phylogenetics and evolution of a circumarctic species complex (Cladocera: Daphnia pulex)

Documents