Molecular Ecology (2009) 18, 5161–5179 doi: 10.1111/j.1365-294X.2009.04422.x

The noncosmopolitanism paradigm of freshwater zooplankton: insights from the global phylogeography of the predatory cladoceran pediculus (Linnaeus, 1761) (Crustacea, )

S. XU,* P. D. N. HEBERT,† A. A. KOTOV‡ and M. E. CRISTESCU* *Great Lakes Institute for Environmental Research, University of Windsor, Windsor, ON, Canada N9B 3P4, †Biodiversity Institute of Ontario, University of Guelph, Guelph, ON, Canada N1G 2W1, ‡A. N. Severtsov Institute of Ecology and Evolution, Leninsky Prospect 33, Moscow 119071, Russia

Abstract A major question in our understanding of eukaryotic biodiversity is whether small bodied taxa have cosmopolitan distributions or consist of geographically localized cryptic taxa. Here, we explore the global phylogeography of the freshwater cladoceran (Linnaeus, 1761) (Crustacea, Onychopoda) using two mitochon- drial genes, cytochrome c oxidase subunit I and 16s ribosomal RNA, and one nuclear marker, 18s ribosomal RNA. The results of neighbour-joining and Bayesian phylogenetic analyses reveal an exceptionally pronounced genetic structure at both inter- and intra- continental scales. The presence of well-supported, deeply divergent phylogroups across the Holarctic suggests that P. pediculus represents an assemblage of at least nine, largely allopatric cryptic species. Interestingly, all phylogenetic analyses support the reciprocal paraphyly of Nearctic and Palaearctic clades. Bayesian inference of ancestral distribu- tions suggests that P. pediculus originated in North America or East Asia and that European lineages of Polyphemus were established by subsequent intercontinental dispersal events from North America. Japan and the Russian Far East harbour exceptionally high levels of genetic diversity at both regional and local scales. In contrast, little genetic subdivision is apparent across the formerly glaciated regions of Europe and North America, areas that historical demographic analyses suggest that were recolonized just 5500–24 000 years ago.

Keywords: , endemism, glacial refugia, intercontinental dispersal, speciation, zoo- Received 7 August 2009; revision received 25 September 2009; accepted 6 October 2009

sizes and strong dispersal abilities of micro-eukaryotes Introduction maintain genetic homogeneity across their broad distri- A major uncertainty in our understanding of biodiver- butions. The freshwater zooplankton represents one sity is whether unicellular and small-bodied taxa have group that was historically thought to be dominated by cosmopolitan distributions or consist of geographically cosmopolitan species. This paradigm of cosmopolitanism structured cryptic taxa. Even less clear is the scale of stemmed from the assumption that the dispersal of these geographic structure, when it exists. The classical cosmo- organisms via the passive transport of resting stages politanism view, the ‘everything is everywhere’ hypoth- through vectors such as waterfowl, wind and water esis (Baas-Becking 1934), holds that the large population currents mediated extensive gene flow (Mayr 1963). Furthermore, the cosmopolitan nature of freshwater Correspondence: Sen Xu, Fax: (519) 971-3616; zooplankton is supported by the observed morphologi- E-mail: [email protected] cal homogeneity among widely separated populations.

2009 Blackwell Publishing Ltd 5162 S. XU ET AL.

Recently, the classical view of cosmopolitanism has climatic and geologic histories, and to hierarchically been challenged by detailed morphological and genetic explore the evolutionary forces that have shaped pat- studies (e.g. Frey 1982; Hebert & Wilson 1994; Forro´ terns of genetic diversity from global to continental and et al. 2008). These studies clearly show that many so- regional scales. Moreover, due to the limited taxonomic called conspecific populations of zooplankton species scope of previous studies on taxa such as rotifers and including rotifers and cladocerans display strong , our understanding of the phylogeography for genetic divergence not only at global geographic scales, freshwater zooplankton with dissimilar biological and but even at regional levels (Gomez et al. 2000; Hebert ecological attributes appear to be limited (but see Cox et al. 2003; Penton et al. 2004). In fact, many widely dis- & Hebert 2001; Petrusek et al. 2004; Dooh et al. 2006; tributed species have now been recognized as cryptic Rowe et al. 2007). species assemblages, which appear to be tens of mil- In this study, we examine the phylogeography of the lions of years old (Colbourne & Hebert 1996; Colbourne predatory cladoceran, Polyphemus pediculus (Linnaeus, et al. 1998; Taylor et al. 1998; Gomez et al. 2002; Ishida 1761) (Crustacea, Onychopoda) across the entirety of its et al. 2006). Holarctic range (Rivier 1998; Korovchinsky 2006). Poly- The replacement of cosmopolitanism by continen- phemus pediculus is well known for its exceptional eco- tal ⁄ regional endemism has greatly facilitated efforts to logical plasticity. It occupies not only ephemeral probe the evolutionary factors underlying the distribu- habitats such as small ponds but also large lakes and tion of genetic diversity in freshwater zooplankton. The brackish estuaries such as the Gulf of Finland (Butorina genetic differentiation between nearby populations has et al. 1975) where it not only occurs in the littoral zone, been proposed as a consequence of founder events fol- but also far offshore and at great depths. Polyphemus lowed by rapid local adaptation. Moreover, the pres- shares important life history attributes with other pred- ence of large, resilient resting propagule banks of local atory cladocerans of the Family Onychopoda (Rivier populations buffers against the establishment of newly 1998; Cristescu & Hebert 2002). As with all other ony- invading genotypes, thus restricting the extent of gene chopods, Polyphemus lacks the protective, ephippial case flow in the face of frequent dispersal events (Monopoli- that encapsulates the resting eggs of most Cladocera, zation Hypothesis; De Meester et al. 2002, 2006). Com- but its resting eggs can resist extensive periods of freez- parative phylogeographic analyses reveal that ing and desiccation (Butorina 1998; Rivier 1998), provid- topographic barriers to gene flow generate predictable ing a capacity for passive dispersal and persistence in patterns of population divergence in zooplankton, sug- ephemeral habitats. gesting that allopatric speciation has played an impor- In this study, we employ two mitochondrial genes, tant role (Hebert et al. 2003). For example, the cytochrome c oxidase subunit I (COI) and 16s ribosomal Appalachian mountain range in North America has RNA, and one nuclear gene, 18s ribosomal RNA (18s), been shown to act as a severe barrier to gene flow as well as an intense sampling strategy across its range between populations of Daphnia laevis (Taylor et al. in North America and Eurasia to examine the genetic 1998), Daphnia ambigua (Hebert et al. 2003) and structure of P. pediculus and ultimately to test its cosmo- crystallina (Cox & Hebert 2001). Furthermore, palaeocli- politanism. Furthermore, we investigate the evolution- matic changes such as the Pleistocene glaciations have ary history underlying the Holarctic distribution of this had a significant impact on the genetic divergence and species and test the roles of intercontinental dispersal persistence of endemic populations. For example, the and habitat shifts between permanent lakes and ephem- North American (Cox & Hebert 2001) eral ponds in its evolutionary diversification. Finally, and the European (De Gelas & De Me- we use a combination of phylogeographic and popula- ester 2005) represent assemblages of divergent lineages tion demographic analyses to investigate the effects of that retreated to glacial refugia and recolonized recently the Pleistocene glaciations on the genetic structure and deglaciated habitats. population dynamics of the North American and Eur- Although substantial insights have been gained con- asian populations. cerning the genetic diversification of freshwater zoo- plankton, few studies have investigated their Materials and methods phylogeographic patterns on a global scale (Adamowicz et al. 2009). Yet, a global phylogeographic perspective is Sampling and DNA amplification essential for understanding the evolutionary history of freshwater zooplankton and testing the cosmopolitan- Polyphemus pediculus was sampled from 106 freshwater ism hypothesis. This approach also allows us to com- habitats across the Holarctic: 56 sites in Nearctic and 50 paratively analyse phylogeographic patterns in sites in Palaearctic (Fig. 1 and Table 1). Once collected, geographic regions characterized by dissimilar palaeo- samples were sorted, preserved in 95% ethanol and

2009 Blackwell Publishing Ltd GLOBAL PHYLOGEOGRAPHY OF POLYPHEMUS PEDICULUS 5163

180° NA1 NA2 NA3 150° W 150° E NA4 PA1 PA2 PA3 PA4 PA5

120° W 120° E

90° W 90° E

60° W 60° E

30° W 30° E

0°

Fig. 1 Sampling locations and geographic distribution of phylogroups for Polyphemus pediculus. Different shapes identify the nine reconstructed phylogroups (Fig. 2).

stored at 4 C. Genomic DNA was extracted from single and SSU-531 and SSU-1085R (Crease & Colbourne 1998) individuals using a modified proteinase K method (Sch- were used to amplify a 680-base pair (bp) fragment of wenk et al. 1998). the COI gene, a 513-bp fragment of the 16s gene and a A total of 426 individuals (4–5 individuals per habi- 464-bp fragment of the 18s gene respectively. The 25-lL tat) were analysed for the mitochondrial gene cyto- polymerase chain reactions consisted of 2 lL genomic chrome c oxidase subunit I (COI). A representative DNA, 1· PCR buffer, 2 mM MgCl2, 0.2 mM dNTPs, subset of these individuals (i.e. members of the major 0.4 lM of each primer and 1 unit of Taq DNA polymer- COI clades) were examined for the mitochondrial 16s ase (Bioline). The thermal regime for amplifying the rRNA gene (n = 108) and the nuclear 18s rRNA gene COI gene consisted of an initial denaturing step of (n = 124). The primer pairs LCO1490 and HCO2198 3 min at 94 C; 5 cycles of 45 s at 94 C, 45 s at 45 C (Folmer et al. 1994), 16Sar and 16Sbr (Palumbi 1996), and 45 s at 72 C; 30 cycles of 45 s at 94 C, 45 s at

2009 Blackwell Publishing Ltd 5164 S. XU ET AL.

Table 1 A summary of sampling locations with geographic position (in decimal degrees), identified COI haplotypes, habitat types (L – lake, P – pond, R – river) and phylogroup assignment based on the COI phylogeny

Site Code Sampling location Latitude Longitude COI haplotypes Habitat Clade

Yukon-Mackenzie (n = 23) YT1 North Folk, Yukon, CAN 64.58 )138.30 C189 — NA1 YT2 Tatchum Lake, Yukon, CAN 62.28 )136.17 C190 L NA1 YT3 An unnamed pond, Yukon, 63.68 )135.84 C189, C191, C192, P NA1 CAN C193 NT1 Small Frog Lake, Northwest 67.37 )134.15 C71, C72, C73 L NA1 Territories, CAN NT2 Eskimo Lake, Northwest 69.10 )132.38 C74, C75 L NA1 Territories, CAN AK Connors Lake, Alaska, USA 61.20 )149.93 C5, C6 L NA4 Cascadia (n = 46) BC1 Kennedy, British Columbia, 55.10 )122.78 C7, C8, C9, C10 — NA2 CAN BC2 Purden Lake, British Columbia, 53.90 )121.95 C8, C11, C12, C13, L NA2, NA4 CAN C104, C105 BC3 Water Lily Lake, British 54.05 )124.68 C8, C12, C13, C14 L NA2, NA4 Columbia, CAN BC4 Tadpole Lake, British Columbia, 55.33 )128.08 C8, C15, C16, C17 L NA2, NA4 CAN BC5 Oliver Lake, British Columbia, 54.27 )130.27 C8, C15 L NA2, NA4 CAN BC6 Hodder Lake, British Columbia, 56.43 )129.46 C17, C106 L NA2 CAN OR1 Lake of the Woods, Oregon, 42.37 )122.20 C101, C102, C108 L NA2 USA OR2 Fish Lake, Oregon, USA 45.03 )117.08 C103 L NA2 Hudson Bay (n = 30) MB1 Strong Lake Creek, Manitoba, 52.23 )98.90 C46 R NA2 CAN MB2 Burtwood River, Manitoba, CAN 56.05 )96.83 C47, C48 R NA3 MB3 Churchill River, Manitoba, CAN 58.70 )94.17 C47, C49, C50, R NA2, NA3 C51, C52 MB4 Paint Lake, Manitoba, CAN 55.48 )98.02 C47 L NA3 MB5 Goose Creek, Churchill, 59.12 )94.23 C50 R NA2 Manitoba, CAN SK Cowan Lake, Saskatchewan, 53.82 )107.03 C188 L NA2 CAN AB1 Jasper National Park, Alberta, 53.20 )117.92 C1, C2 L NA2 CAN AB2 Medicine Lake, Alberta, CAN 52.83 )117.72 C2, C3, C4 L NA2 AB3 Maligue Lake, Alberta, CAN 52.72 )117.63 C1, C2, C3 L NA2 AB4 Sturgeon Lake, Alberta, CAN 55.07 )117.55 C3 L NA2 AB5 Unnamed pond, Alberta, CAN 54.55 )117.02 C1 P NA2 Great Lakes (n = 53) ON1 Lake Ontario, Ontario, CAN 43.32 )79.67 C79 L NA3 ON2 Doe Lake, Ontario, CAN 44.92 )79,27 C86 L NA3 ON3 Weismuller Lake, Ontario, CAN 44.93 )79.20 C87 L NA3 ON4 Unnamed pond, Ontario, CAN 45.48 )78.22 C88, C89 P NA3 ON5 Chub Lake, Ontario, CAN 47.67 )91.63 C90 L NA3 ON6 Rosseau Lake, Ontario, CAN 45.17 )79.57 C86, C91, C92 L NA3 ON7 Bonaparte Lake, Ontario, CAN — — C93, C94, C95, L NA3 C96, C97 ON8 Unnamed pond, Guelph, 43.53 )80.25 C98, C99, P NA2 Ontario, CAN ON9 Temagami, Ontario, CAN 47.05 )79.78 C52, C100 — NA3 ON10 Unnamed pond, Ontario, CAN 50.97 )90.38 C47, C80 P NA3

2009 Blackwell Publishing Ltd GLOBAL PHYLOGEOGRAPHY OF POLYPHEMUS PEDICULUS 5165

Table 1 Continued

Site Code Sampling location Latitude Longitude COI haplotypes Habitat Clade

ON11 Swan Lake, Ontario, CAN 48.23 )80.25 C47, C81 L NA3 ON12 Portage Lake, Ontario, CAN 53.87 )90.17 C82, C83, C84 L NA3 ON13 McCaulay Creek, Ontario, CAN 48.72 )91.90 C47, C79 R NA3 ON14 Boss Lake, Ontario, CAN 44.87 )79.20 C85 L NA3 ON15 Sandlink Pond, Ontario, CAN 49.77 )86.32 C56 P NA3 NL Duck Pond, Newfoundland, 47.57 )52.70 C67 P NA3 CAN QC1 Lake Parent, Quebec, CAN 48.43 )77.20 C56 L NA3 QC2 Unnamed pond, Quebec, CAN 48.78 )77.10 C56, C109, C110 P NA3 QC3 Chibougamau, Quebec, CAN 49.90 )74.35 C111 — NA3 QC4 Lake Blanche, Quebec, CAN 49.50 )74.17 C112, C113 L NA3 QC5 Small lake, Quebec, CAN 47.28 )71.18 C114 L NA3 MN Vermillion Lake, Minnesota, 47.85 )92.68 C47, C55 L NA3 USA North Appalachian (n = 50) NB Chamcook Lake, New 45.13 )67.07 C56, C57, C58 L NA3 Brunswick, CAN ME Unnamed Stream, Maine, USA 45.00 )69.00 C53, C54 R NA3 NY1 Sandy Pond, New York, USA 40.89 )72.73 C76 P NA3 NY2 Swan Pond, New York, USA 40.91 )72.80 C76, C77 P NA3 NY3 Round Pond, New York, USA 40.99 )72.29 C78 P NA3 CT1 Beach Pond, Connecticut, USA 41.58 )71.72 C18, C19, C20, C21, P NA3 CT2 Hell Hollow Pond, Connecticut, 41.64 )71.87 C22, C23 P NA3 USA MA Fresh Pond, Massachusetts, USA 41.67 )70.15 C43, C44, C45 P NA3 NH Sunapee Lake, New Hampshire, 43.38 )72.07 C60, C61, C62, C63, L NA3 USA C64, C65, C66 Western Europe (n = 32) Ned Lake Naardemeer, 52.30 5.12 C59 L PA1 The Netherlands Ger1 Schusee, Plo¨n, Germany 54.15 10.40 C27, C28 — PA1 Ger2 Trinkwassertalsperre Frauenau, 49.02 13.33 C29, C30 — PA1 Germany Ger3 Branderburg, Langersee, 52.88 14.60 C31 — PA2 Germany Den1 Ovresdam Pond, Hillerød, 55.92 12.30 C24 P PA1 Denmark Den2 Unnamed pond, Denmark 55.00 12.00 C24 P PA1 Nor Troms, Norway 69.11 20.72 C69, N70 — PA1 Fin1 Lapin La¨a¨ni, Kilpisja¨rvi, Finland 69.06 20.78 C25, C26 — PA1 Fin2 Lapin La¨a¨ni, Pallasja¨rvi, Finland 68.04 24.14 C200 — PA1 Eastern Europe (n = 74) Pol Jegocin Lake, Poland 53.65 21.73 C107 L PA1 Rus1 Volgograd Reservoir, Volgograd 48.72 44.72 C115, C116 L PA1 Area, Russia Rus2 Lake Glubokoe, Moscow Area, 55.73 37.60 C154, C155 L PA1 Russia Rus3 Rybinsk Reservoire, Yaroslavle 58.07 38.27 C69, C171, L PA1 Area, Russia C172, C173 Rus4 A pond, Ruza District, Moscow 55.75 36.50 C118 P PA2 Area, Russia Rus5 Lake Beloe-Bordukovskoe, 55.63 39.73 C118, C182 L PA2 Moscow Area, Russia Rus6 Lake Galkino, Kaluga Area, 54.77 35.80 C183, C184 L PA3 Russia

2009 Blackwell Publishing Ltd 5166 S. XU ET AL.

Table 1 Continued

Site Code Sampling location Latitude Longitude COI haplotypes Habitat Clade

Rus7 Srednee Suzdal’skoe Lake, St. 60.03 30.30 C118, C185, C186 L PA2 Petersburg, Russia Rus8 Verkhnee Suzdal’skoe Lake, St. 60.04 30.30 C187 L PA1 Petersburg, Russia Rus9 Nizhnee Bol’shoe Suzdal’skoe 60.04 30.30 C118 L PA2 Lake, St. Petersburg, Russia Rus10 Kurgolovskoe Lake, Leningrad 60.17 30.52 C117, C118, C119 L PA2 Area, Russia Rus11 Lake Vendcherskoje, Karelian 62.18 33.27 C118 L PA2 Autonomous Republic, Russia Rus12 Unnamed lake, Karelian 62.20 33.57 C118, C120, C121 L PA2 Autonomous Republic, Russia Rus13 Lake Gomsel’ga, Karelian 62.05 34.03 C118, C122, C123 L PA2 Autonomous Republic, Russia Rus23 Lake Gluhoe, Pskov Area, Russia 56.28 31.35 C118, C162, C163 L PA1, PA2 Rus27 Granite pool, Kastjan Island, 66.48 33.37 C167 PA3 Karelian Autonomous Republic, Russia Rus28 Lake Mertvoe, Smolensk Area, 55.52 31.65 C168, C169 L PA3 Russia Rus31 Lake Pesno, Tver Area, Russia 56.32 31.92 C118 L PA2 Rus32 Lake Sinovets, Pskov Area, 56.58 29.00 C118, C175 L PA2 Russia Rus33 Unnamed lake near Glubokoe, 56.60 29.01 C69, C118, C176, C177, L PA1, PA2 Pskov Area, Russia C178, C179 Rus39 An unnamed lake, Astrakhan 47.02 47.58 C181 L PA2 Area, Russia Siberia (n = 80) Rus14 Lake Bol’shoe Purul’do, Tomsk 57.87 84.37 C118, C124 L PA2 Area, Russia Rus15 Affluent of River Chulym, 57.80 84.27 C118, C120, C125 R PA1, PA2 Tomsk Area, Russia Rus16 Unnamed Lake, Tomsk Area, 57.80 84.18 C118 L PA2 Russia Rus17 Petrovskaja Protoka, Tomsk 56.58 84.83 C118, C126, C127, R PA2 Area, Russia C128, C129, C130, C131, C132, C133 Rus18 Lake Kurok, Tomsk Area, Russia 56.80 84.55 C118, C132, C134, L PA2 C135, C136, C137 Rus19 Unnamed lake, Tunaev Island, 56.42 84.05 C138, C139, C140, L PA1, PA2 Tomsk Area, Russia C141, C142, C143, C144, C145, C146, C147, C148, C149, C150, C151, C152, C153 Rus20 Unnamed lake, Mel’nikovo, 56.53 84.15 C156 L PA1 Tomsk Area, Russia Rus21 Sukhozhinskoe Lake, Parabel’, 58.68 81.50 C69, C157, C158, C159 L PA1 Tomsk Area, Russia Rus22 An oxbow lake in Surgut, 61.23 73.42 C118, C132, C160, L PA1, PA2 Khanty-Mansi Autonomous C161 Area, Russia Rus24 River Atlian, Chelyabinsk Area, 55.00 59.9 C164 R PA2 Russia Rus25 River Osh, Znamenskoe, Omsk 57.12 73.75 C165 R PA1 Area, Russia Rus26 River Ishim, Omsk Area, Russia 57.63 71.17 C118, C166, C194 R PA2

2009 Blackwell Publishing Ltd GLOBAL PHYLOGEOGRAPHY OF POLYPHEMUS PEDICULUS 5167

Table 1 Continued

Site Code Sampling location Latitude Longitude COI haplotypes Habitat Clade

Rus29 Nadym River, Yamalo-Nentrs 66.05 72.03 C69, C170, C201 R PA1, PA2 Autonomour Area, Russia Rus30 Unnamed lake in Nadym 66.22 72.03 C69, C174 L PA1 District, Yamalo-Nentrs Autonomour Area, Russia Russian Far East (n = 15) Rus34 A roadside ditch, Jewish 48.50 134.90 C195, C179 P PA2, PA4 Autonomous Area, Russia Rus35 A mine lake, Sakhalin Area, 47.32 142.70 C180 L NA4 Russia Rus36 Island of Bol’shoy Ussurijsky, 48.35 134.83 C196 L PA4 Khabarovsk Territory, Russia Rus37 Lake Gassi, Khabarovsk 49.05 136.52 C197, C198 L PA4 Territory, Russia Rus38 Reservoir of Luchegorskaya, 46.45 137.28 C199 L PA4 Primorski Territory, Russia Japan (n = 23) Jpn Lake Aoki, Nagano Prefecture, 36.67 137.85 C32, C33, C34, C35, L PA5 Japan C36, C37, C38, C39, C40, C41, C42

Sampling locations are grouped according to major geographic regions and recognized freshwater biogeographic provinces in North America (Burr & Mayden 1992).

50 C and 45 s at 72 C; and a final extension of 72 C Crandall 1998) was used to select the best-fit model of for 4 min. The thermal regime for amplifying the 16s sequence substitution (COI: TrN + I + G; 16s: K81uf + I; and 18s genes consisted of two cycles of 30 s at 94 C, 18s: GTR + I) under the Akaike Information Criterion 45 s at 60 C and 45 s at 72 C; five cycles of 30 s at (Posada & Buckley 2004). Neighbour-joining phyloge- 93 C, 45 s at 55 C and 45 s at 72 C; followed by 29 netic analysis was conducted in MEGA4 (Tamura et al. cycles of 30 s at 93 C, 1 min at 50 C and 1 min at 2007), based on nucleotide distances corrected using the 72 C. PCR products were purified using the Solid Tamura–Nei model (Tamura & Nei 1993) with or with- Phase Reversible Immobilization method (Deangelis out a gamma rate distribution. Clade support was esti- et al. 1995). Sequencing reactions were performed on mated using bootstrap analyses with 1000 replicates purified PCR products using the forward primers and (Felsenstein 1985). Bayesian phylogenetic analyses were BigDye Terminator 3.1 chemistry, and an ABI 3130XL performed in MrBayes v3.1 (Ronquist & Huelsenbeck automated sequencer was used for sequencing the PCR 2003) with priors left at default values. Bayesian analy- products (Applied Biosystems). All sequences that con- ses for the concatenated COI + 16s sequences were per- tained ambiguous sites were subsequently sequenced formed on both the full alignment and reduced dataset with the corresponding reverse primers. with one representative sequence from each clade to overcome the effects of sampling bias on phylogenetic reconstructions. Settings of the sequence substitution Phylogenetic analyses parameters followed those specified in the best-fit mod- Sequences for the three genes (COI, 16s and 18s) were els as closely as possible. All searches used random aligned in CodonCode Aligner version 2.0.6 (Codon- starting trees and employed four independent runs, Code Corporation). The alignments were inspected each with one cold chain and three incrementally visually and adjusted manually. We performed separate heated chains. Trees were sampled every 100 genera- phylogenetic analyses on the COI (File S1), concate- tions for 2 million generations and the first 25% of all nated COI plus 16s (File S2) and 18s (File S3) align- the trees sampled before convergence were discarded ments using the onychopods leuckartii and as burn-in. The 50% majority rule consensus tree was as outgroups. Phylogenetic analyses generated from the remaining trees and the posterior were performed using neighbour-joining and Bayesian probability of each node was calculated as the percent- inference methods. ModelTest (version 3.7; Posada & age of trees recovering any particular node. Genetic

2009 Blackwell Publishing Ltd 5168 S. XU ET AL. diversity within each major phylogroup was character- background selection and hitchhiking associated with ized by the standard indices of haplotype diversity (h) selective sweeps. Tajima’s D and Fu’s Fs tests were per- and nucleotide diversity (p) using DnaSP version 4.50 formed in Arlequin 3.1 (Excoffier et al. 2005) with (Rozas et al. 2003). The mean genetic sequence diver- 10 000 simulated samples to generate statistical confi- gence between major phylogroups was calculated in dence. MEGA4 using the Tamura–Nei model. Mismatch distribution was used to investigate the The ancestral (internal) vs. derived (tip) relationships demographic history of the recovered phylogroups of the COI haplotypes was analysed using a network (Rogers & Harpending 1992). This distribution is usu- approach, taking into account the possibility of low ally multimodal in samples drawn from populations at sequence divergence, multifurcated genealogies, persis- demographic equilibrium, reflecting the highly stochas- tence of ancestral haplotypes and reticulate relation- tic shape of gene trees, but it is unimodal in popula- ships within population genealogies (Posada & Crandall tions that have passed through a recent demographic 2001). The program Network (version 4.5 available at expansion (Slatkin & Hudson 1991; Rogers & Harpend- http://www.fluxus-technology.com) was employed to ing 1992) or through a range expansion with high levels construct a median-joining haplotype network that is of migration between neighbouring demes (Ray et al. based on minimum spanning trees in which median 2003; Excoffier 2004). The demographic parameter Tau vectors representing extinct ancestral or unsampled (s) was estimated using a generalized nonlinear least- haplotypes were added following the maximum-parsi- square approach, and the confidence interval of this mony principle (Bandelt 1999). Ancestral haplotypes parameter was computed using a parametric bootstrap were identified by their internal ⁄ central positions in the with 10 000 replicates in Arlequin 3.1. The parameter s network, their high frequencies and the number of low- was transformed to estimate the real time since expan- frequency haplotypes derived from them. sion (t) with the equation s =2ut, where u =mTl,mT the number of nucleotides under study, and l the mutation rate for the COI gene per generation (Rogers Bayesian inference of ancestral distributions & Harpending 1992). To test whether the observed mis- Bayesian inference of the ancestral distributions at match distribution fits the simulated sudden expansion major internal nodes on the recovered COI phylogeny model defined by the estimated parameters, sum of was performed using the software MrBayes v3.1 (Huel- square deviations (SSD) between the observed and sim- senbeck & Bollback 2001; Ronquist 2004). Compared to ulated distributions was used, with P-values calculated mapping the evolution of characters on a single phylog- as the proportion of simulations producing an SSD lar- eny under the principle of parsimony, Bayesian infer- ger than or equal to the observed SSD. ence has the advantage of simultaneously accounting for both phylogenetic and mapping uncertainties (Ron- Molecular clock quist 2004). In order to increase computational effi- ciency, each recognized phylogroup was represented by Rates of sequence divergence for COI have been esti- one of its haplotypes and the distribution data (pres- mated to range from 1.4% to 2.6% per million years ence in North America, East Asia or Europe) were (Knowlton et al. 1993; Knowlton & Weigt 1998; Schu- entered in a separate partition. A constrained Bayesian bart et al. 1998). An intermediate rate of 2% divergence inference was run for 2 million generations to infer the per million years, which is consistent with the arthro- ancestral distribution for each internal node of interest. pod mtDNA clock (Brower 1994), was used to infer divergence time between major clades. Demographic analyses Results The demographic histories of the main phylogroups were investigated using Tajima’s D (Tajima 1989), Fu’s Sequence diversity Fs (Fu 1997) and mismatch distributions of pairwise dif- ferences based on the COI sequences (Rogers & Har- The 558-bp alignment for the COI gene included 218 pending 1992; Ray et al. 2003; Excoffier 2004). Tajima’s variable sites of which 196 were parsimony-informative.

D and Fu’s Fs were used to test whether the COI gene Third codon positions accounted for 77% of the vari- evolved under neutrality and whether populations have able sites, whereas first and second codon positions rep- evolved according to the Wright–Fisher model with a resented 16% and 7% of the variable sites respectively. constant effective population size. Negative values of Nonsynonymous substitutions resulted in 18 amino these statistics indicate non-neutral evolution and possi- acid changes. The 426 sequences of the COI gene con- ble population expansion; alternative explanations are tained 201 unique haplotypes (GenBank accession nos

2009 Blackwell Publishing Ltd GLOBAL PHYLOGEOGRAPHY OF POLYPHEMUS PEDICULUS 5169

GQ406848–GQ407048), with an overall haplotype diver- The mitochondrial and nuclear phylogenies using sity (h) of 0.98 and nucleotide diversity (p) of 0.12. The both the neighbour-joining and Bayesian methods 458-bp sequence alignment for the 16s gene contained revealed largely congruent topologies (Figs 2 and 3). 89 variable sites of which 78 were parsimony informa- All phylogenetic reconstructions strongly supported the tive. The 108 sequences of the 16s fragment contained reciprocal paraphyly of Nearctic and Palaearctic lin- 36 unique haplotypes (accession nos GQ407049– eages. Deep-level phylogenetic relationships were gen- GQ407084) (h = 0.94, p = 0.06). The alignment for the erally well resolved. Highly supported sister group 18s gene was 421 bp and contained 33 variable sites of relationships were evident between the PA1 and PA2 which 26 were informative based on the parsimony cri- and between the NA1 and PA3 clades. Moreover, both terion. A total of 10 haplotypes (accession nos the mitochondrial and nuclear phylogenies supported GQ407085–GQ407094) were identified among the 124 the monophyletic clade comprised of the NA1, PA3 and analysed sequences (h = 0.86, p = 0.03). NA2 group, as well as the monophyletic clade consist- ing of the Pacific NA4, PA4 and PA5 group (Figs 2 and 3). However, the mitochondrial and nuclear phyloge- Phylogeographic patterns nies were discordant with respect to the phylogenetic Neighbour-joining and Bayesian phylogenetic analyses position of the Pacific and NA3 clades. Although both of the mitochondrial sequences (COI and concatenated nuclear phylogenies (neighbour-joining and Bayesian) COI plus 16s) consistently recovered nine highly diver- supported the basal position of the NA3 clade, the gent and well-supported monophyletic clades: four mitochondrial phylogenies placed the Pacific clade in Nearctic (NA) clades and five Palaearctic (PA) clades the basal position (Fig. 3). Plotting the habitat types (Figs 2 and 3). Most clades showed well-defined geo- (lake vs. pond) for individual haplotypes (Table 1) on graphic ranges and largely allopatric distributions. The the obtained COI phylogeny showed that members of Nearctic 1 (NA1) clade was exclusively distributed in the NA4 and PA5 clades were exclusively found in northwest Canada covering the Beringian biogeographic large lakes. All other clades were collected from both province, the NA2 clade covered a vast area between lakes and ponds with no obvious phylogenetic division the Pacific coast and the Great Lakes region (Cascadia between the two habitats. and Hudson Bay biogeographic provinces), whereas the Haplotype diversity and nucleotide diversity were NA3 clade included regions from the Great Lakes and estimated for all nine phylogroups (Table 2). Haplotype Northern Appalachian to the Atlantic province (Fig. 1). diversity showed little variation among the phylo- The NA4 clade consisted of haplotypes identified from groups ranging between 0.68 and 0.96, whereas the the coastline of British Columbia and Alaska as well as nucleotide diversity varied dramatically. The nucleotide one haplotype from the Sakhalin Island, in the Far East diversity of the PA5 group from the Japanese Lake Aoki of Russia. The sympatric occurrence of haplotypes (0.014) ranked among the highest along with the Pacific belonging to the NA2 and NA3 clades and the NA2 NA4 (0.020), west Nearctic NA2 (0.016) and Eurasian and NA4 clades was observed in Manitoba and British PA1 (0.015) clades. The lowest nucleotide diversity was Columbia respectively. Strong phylogeographic subdivi- found in the northwest Nearctic NA1 clade (0.005). sion was observed within most of the Nearctic clades. The corrected Tamura–Nei COI nucleotide diver- The Oregon region harboured a well-defined clade gences among the nine phylogroups ranged from 6.3% within the west Nearctic clade NA2, whereas sites in to 24.4%, whereas genetic divergences for the more Massachusetts possessed a clade divergent from the rest conserved 16s and 18s genes ranged from 3.1% to of the east Nearctic clade NA3 (Fig. 2). Three endemic 11.5% and 0.2% to 4.7% respectively (Table S1). The groups corresponding to Alaska, British Columbia and mean COI genetic divergences within the identified the Sakhalin Island were identified within the Pacific phylogroups ranged from 0.8% to 3.6% (Table S1). For NA4 clade. Within the Palaearctic (PA), two geographi- the more slowly evolving nuclear 18s gene that is usu- cally extensive clades (PA1, PA2) were largely sympat- ally used for investigating deep phylogeny beyond the ric across Russia. The PA1 clade included haplotypes level, moderate sequence divergence (2% to found in Europe and Siberia, whereas the PA2 clade 5%) was observed among the nine recovered lineages consisted of haplotypes from European Russia and Sibe- that generally correspond to the monophyletic groups ria (Figs 1 and 2). The glaciated regions of Europe were recovered by the mitochondrial markers (Table S1). predominantly occupied by haplotypes of the PA1 clade. The more restricted PA3 phylogroup was only Ancestral distributions and intercontinental dispersal detected in three habitats in northern European Russia. Two additional clades were detected in the Russian Far The Bayesian reconstructions of the ancestral distribu- East (PA4) and Japan (PA5). tions of major internal nodes suggested that at least

2009 Blackwell Publishing Ltd 5170 S. XU ET AL.

50/69 C25: Fin1 (1) 100/89 PA1 C200: Fin2 (1) G C107: Fin2 (3) 86% C59: Ned (5) 61/- PA2 C115: Rus1 (1) C74: NT2 (1) C116: Rus1 (1) C75: NT2 (1) C69: Nor, Rus3, 21, 29, 30, 33 (12) C190: YT2 (1) C189: YT1, YT3 (5) C171: Rus3 (1) C193: YT3 (2) C26: Fin1 (1) C73: NT1 (1) NA1 C161: Rus22 (1) C71: NT1 (1) 99/90 C172: Rus3 (1) 91/86 D C72: NT1 (1) 100/93 F C191: YT3 (6) C154: Rus2 (2) 70% C192: YT3 (2) C155: Rus2 (1) 81% 99/97 C167: Rus27 (2) C168: Rus28 (3) C170: Rus29 (1) C169: Rus28 (1) PA3 C174: Rus30 (1) C183: Rus6 (8) C176: Rus33 (1) C184: Rus6 (1) C16: BC4 (1) C177: Rus33 (1) C7: BC1 (1) C27: Ger1 (5) C104: BC2 (1) C28: Ger1 (1) C17: BC4, BC6 (2) C159: Rus21 (1) 85/83 E C46: MB1 (4) C13: BC2, BC3 (2) C173: Rus3 (1) 79% C106: BC6 (1) C187: Rus8 (1) PA1 C8: BC1-5 (10) C10: BC1 (1) C158: Rus21 (1) C 85/- C9: BC1 (1) C24: Den1, Den2 (6) 72/68 C1: AB1, AB3, AB5 (3) C29: Ger2 (1) C4: AB2 (1) C30: Ger2 (5) 70% C98: ON8 (2) C99: ON8 (1) NA2 C157: Rus21 (1) C3: AB1-4 (4) C165: Rus25 (1) C188: SK (4) 100/89 C49: MB3 (1) C70: Nor (3) C50: MB3, MB5 (2) C152: Rus19 (1) C51: MB3 (1) C139: Rus19 (1) C14: BC3 (1) C2: AB1, AB2, AB3 (3) C148: Rus19 (1) C102: OR1 (6) C144: Rus19 (1) 99/92 C108: OR1 (1) C147: Rus19 (1) C101: OR1 (1) C103: OR1, OR2 (7) C150: Rus19 (1) 84/87 C43: MA (5) C145: Rus19 (1) C45: MA (1) C162: Rus23 (1) C44: MA (1) 100/79 C67: NL (1) C125: Rus15 (1) C68: NL (1) C141: Rus19 (1) C63: NH (2) 50/69 C160: Rus22 (1) C22: CT2 (2) 75/- C23: CT2 (1) C156: Rus20 (1) C61: NH (2) C153: Rus19 (1) C60: NH (2) C149: Rus19 (1) C76: NY1, NY2 (7) C77: NY2 (2) C151: Rus19 (1) C78: NY3 (1) C123: Rus13 (1) C19: CT1 (2) C138: Rus19 (1) C18: CT1 (1) C64: NH (3) C143: Rus19 (1) C66: NH (1) C140: Rus19 (1) C65: NH (1) 61/- C194: Rus26 (1) C62: NH (4) C20: CT1 (1) C178: Rus33 (1) C58: NB (2) C185: Rus7 (1) C111: QC3 (1) C130: Rus17 (1) C56: NB, ON15, QC1, QC2 (7) C112: QC4 (1) C31: Ger3 (1) 99/76 A C113: QC4 (1) C201: Rus29 (1) C53: ME (1) C129: Rus17 (1) 55% C54: ME (1) C85: ON14 (2) C126: Rus17 (1) C109: QC2 (1) NA3 C128: Rus17 (1) C96: ON7 (1) C137: Rus18 (1) C97: ON7 (1) 0.02 C110: QC2 (2) C131: Rus17 (1) C93: ON7 (1) C133: Rus17 (1) C94: ON7 (2) C135: Rus18 (1) C95: ON7 (1) C21: CT1 (2) C136: Rus18 (1) C83: ON12 (2) C164: Rus24 (1) C88: ON4 (1) C175: Rus32 (1) C89: ON4 (1) PA2 C87: ON3 (1) C186: Rus7 (1) C79: ON1, ON13 (2) C146: Rus19 (1) C84: ON12 (1) C121: Rus12 (1) C100: ON9 (1) C52: MB3, ON9 (2) C142: Rus19 (2) C81: ON11 (2) C179: Rus34 (1) C55: MN (1) C127: Rus17 (1) C57: NB (1) C90: ON5 (5) C122: Rus13 (1) C82: ON12 (1) C134: Rus18 (1) C47: MB2-4, MN, ON10, 11,13 (11) C132: Rus17, Rus18, Rus21 (3) C48: MB2 (1) C80: ON10 (1) C120: Rus12, Rus15 (2) C114: QC5 (1) C182: Rus5 (1) C92: ON6 (1) C181: Rus39 (1) C86: ON2, ON6 (2) C91: ON6 (1) C117: Rus10 (1) C196: Rus36 (1) C119: Rus10 (1) C199: Rus38 (1) C118: Rus4, 5, 7, 9-18, 22, 23, 26, 31-33 (54) 99/93 C197: Rus37 (1) PA4 C195: Rus34 (3) C124: Rus14 (1) C198: Rus37 (1) C163: Rus23 (1) C11: BC2 (1) C166: Rus26 (1) 92/- C105: BC2 (1) 93/- C15: BC4, BC5 (3) B87/- C12: BC2, BC3 (5) 99/87 C180: Rus35 (7) NA4 87% 100/96 C5: AK (1) C6: AK (1) Pacific C39: Jpn (1) C40: Jpn (1) 66/90 C41: Jpn (1) C35: Jpn (5) 100/83 C42: Jpn (1) C34: Jpn (1) C37: Jpn (1) PA5 0.02 C32: Jpn (9) 98/95 C36: Jpn (1) C33: Jpn (1) C38: Jpn (1) Cercopagis pengoi Podon leukarti

Fig. 2 Neighbour-joining (NJ) tree based on 201 haplotypes of the cytochrome c oxidase subunit I (COI) gene. Haplotype notation identifies the site code and the number of individuals showing a given haplotype. Bootstrap values for NJ reconstructions and pos- terior probabilities for Bayesian inferences (in percentage) are shown above each major internal branch. The Bayesian ancestral distri- butions for nodes A–G are shown with posterior probabilities below the branches: white circle – North America; grey circle – East Asia; black circle – Europe. Three lineages derived from inferred intercontinental dispersal events are indicated by bold lines. three independent intercontinental dispersal events appears to be connected to East Asia. The MRCA of occurred during the evolution of Polyphemus (Fig. 2). the PA4, NA4 and PA5 clades (node B) was most Although the centre of origin, the most recent common likely a resident of East Asia (87%), whereas the ancestor (MRCA) of Polyphemus (node A) was ambigu- MRCA of the PA1, PA2, PA3, NA1, NA2 and NA3 ously assigned to North America (posterior probability clades (node C) was likely distributed in North Amer- of 55%), the early evolution of Polyphemus also ica (70%). An intercontinental dispersal event from

2009 Blackwell Publishing Ltd GLOBAL PHYLOGEOGRAPHY OF POLYPHEMUS PEDICULUS 5171

M55: Rus21 (1) N1: Rus3, 19, 21, 34 (10) (A) COI+16s M12: Ned (1) (B) 18s M60: Rus33 (1) PA1 M54: Rus21 (1) 95/100 M53: Rus21 (1) M2: Ger1 (1) PA2 N2: Den1, Rus10 M52: Rus21 (1) PA1 M50: Rus19 (1) 14, 17, 19, 21, 33 (24) 68/99 M48: Rus19 (1) M51: Rus19 (1) 100/100 M49: Rus19 (1) M47: Rus19 (1) M61: Rus33 (1) M46: Rus17 (1) 84/- N3: NT1, YT2 (7) M43: Rus17 (1) M45: Rus17 (1) NA1 63/100 M44: Rus17 (1) PA2 M42: Rus14 (1) M41: Rus14 (1) M40: Rus14, Rus17 (4) N4: Rus27, 51/80 100/100 M36: NT1 (1) Rus28 (4) M35: NT1 (1) NA1 99/100 M13: NT1 (2) PA3 100/100 M69: Rus6 (1) 99/100 M68: Rus6 (8) M59: Rus28 (1) 100/100 M58: Rus28 (1) PA3 N5: Rus6 (4) 100/100 M57: Rus27 (1) M56: Rus27 (1) M4: MB3 (1) 86/99 71/87 M5: MB3 (1) 68/96 M6: MB3 (1) M70: SK (4) 87/67 N6: BC2 (3) 92/87 M16: AB1 (1) M1: BC1 (1) M22: BC6 (1) NA2 100/100 M3: MB1 (1) NA2 76/100 M18: BC2 (1) M39: OR2 (4) N7: BC6, MB3, OR1, OR2, AB1 (23) 100/99 M37: OR1 (1) M38: OR1 (4) M15: ON7 (1) 100/100 M7: MB3 (1) 100/100 M14: ON1 (1) M8: NB (3) NA3 M9: NB (1) N8: BC2, BC5, 100/100 M10: NB (1) PA4 Rus34-38 (15) M11: NB (1) M67: Rus38 (1) 100/100 M64: Rus36 (1) Pacific 99/99 NA4 M65: Rus37 (1) PA4 81/- M66: Rus37 (1) 100/100 M63: Rus35 (1) PA5 N9: Jpn (23) 99/98 M62: Rus35 (6) M17: BC2 (3) NA4 83/59 M21: BC5 (1) 100/100 M20: BC5 (1) M31: Jpn (1) Pacific M33: Jpn (1) 94/99 M32: Jpn (1) NA3 N10: MB3, NB, ON3 (11) M27: Jpn (5) 100/99 M34: Jpn (1) M26: Jpn (1) M29: Jpn (1) PA5 97/97 M28: Jpn (1) M25: Jpn (4) 0.02 M24: Jpn (4) 0.02 M23: Jpn (1) M30: Jpn (1) Cercopagis pengoi Cercopagis pengoi Podon leukarti Podon leukarti

Fig. 3 Neighbour-joining tree based on (A) concatenated mitochondrial cytochrome c oxidase subunit I (COI) and 16s rRNA sequences and (B) nuclear 18s rRNA sequences. Haplotype notation identifies the site code and the number of individuals showing a given haplotype. Bootstrap values for neighbour-joining reconstructions and posterior probabilities (in percentage) for Bayesian infer- ences are shown for each major internal branch.

Table 2 The test results of Tajima’s D, Fu’s Fs and mismatch distributions, with sample size (n), number of haplotypes (NH), haplotype diversity (h) and nucleotide diversity (p) for each identified clade

Mismatch

Tajima’s D Fu’s Fs distributions

Clade n NH h p DPFs P Tau (s) PSSD

Nearctic NA1 21 10 0.87 0.005 )0.64 0.29 )2.60 0.08 3.48 0.39 NA2 62 25 0.94 0.016 )0.41 0.39 )2.99 0.20 2.46 0.32 NA2* 47 11 0.93 0.009 )1.28 0.08 )6.58 0.02 5.16 0.65 NA3 107 56 0.98 0.012 )1.41 0.05 )24.92 0.00 5.57 0.82 NA4 19 7 0.80 0.020 )0.07 0.52 5.64 0.97 36.85 0.09 Palaearctic PA1 77 44 0.96 0.015 )1.44 0.05 )22.27 0.00 4.89 0.87 PA2 95 38 0.68 0.008 )1.90 0.01 )2.40 0.00 0.00 0.00 PA3 15 5 0.70 0.010 )0.43 0.37 3.51 0.93 17.82 0.09 PA4 7 5 0.86 0.010 )0.93 0.21 0.45 0.55 7.98 0.05 PA5 23 11 0.82 0.014 1.34 0.93 )0.02 0.52 15.07 0.04

*Haplotypes from Oregon were excluded from the analyses.

2009 Blackwell Publishing Ltd 5172 S. XU ET AL.

East Asia to North America was inferred to account for NA3 were occupied by two closely related haplo- for the distribution of the NA4 clade in North Amer- types, C47 and C56. Although the C47 haplotype was ica. Based on the ancestral distribution of nodes D widespread in the Great Lakes region (Manitoba, Min- (MRCA of PA1, PA2, NA1, PA3 and NA2), E (MRCA nesota and Ontario), the C56 haplotype was distributed of NA1, PA3 and NA2) and F (MRCA of NA1 and in northeastern North America (New Brunswick, PA3) in North America (posterior probability 70%, Ontario and Quebec). The likely ancestral haplotype for 79% and 81% respectively), two independent intercon- the PA1 clade (C69) was common in western Europe, tinental dispersal events from North America to Eur- European Russia and Siberia, whereas the central haplo- ope were inferred to explain the distribution of PA1, type for the PA2 clade (C118) was found in European

PA2 and PA3 clades in Europe (Fig. 2). Russia and Siberia. Tajima’s D, Fu’s Fs and mismatch distribution analysis supported the population expan- sion hypothesis for the NA1, NA2 (haplotypes from Demographic history Oregon were excluded due to their genetic divergence The median-joining networks for the 201 COI haplo- from the rest of the clade), NA3 and PA1 clades types corroborated the neighbour-joining and Bayesian (Table 2 and Fig. 6). For the PA2 clade, although Taj- phylogenetic reconstructions, with all the recovered ima’s D and Fu’s Fs indicated violation of neutral evolu- phylogroups forming separate clusters (Fig. 4). The net- tion, the population expansion hypothesis was not works for the NA3, PA1 and PA2 clades showed star- accepted because of lack of support from mismatch dis- like topology, suggesting recent population expansion tribution analysis (Table 2). Assuming a COI molecular events (Fig. 5). The centres of the haplotype network clock of 2% sequence divergence per million years and

A B

Fig. 4 Median-joining cytochrome c oxidase subunit I (COI) haplotype network (A) and neighbour-joining tree (B). Individual haplo- types are represented by grey circles and the circle area is proportional to the observed frequency of a given haplotype. Median vec- tors, which represent either extant unsampled sequences or extinct ancestral sequences, are indicated by small white circles. Connecting branches are proportional to the inferred mutation steps between haplotypes and the number of mutation steps between major clades is labelled.

2009 Blackwell Publishing Ltd GLOBAL PHYLOGEOGRAPHY OF POLYPHEMUS PEDICULUS 5173

Fig. 5 Expanded median-joining cytochrome c oxidase subunit I (COI) haplotype network. Median vectors are indicated by small black circles. High-frequency haplotypes are labeled. Colour code legends: black – European haplotypes; dark grey – European Russian haplotypes; light grey – Siberian haplotypes; cross – Russian Far East haplotypes.

10 – 20 generations per year, population expansion of Cryptic species complex the NA1, NA2, NA3 and PA1 clades was estimated to have occurred c. 7450–14 900, 5500–11 000, 12 000– The phylogenetic analyses of 426 specimens from 106 24 000 and 10 500–21 000 years ago respectively. habitats using mitochondrial (COI, 16S) and nuclear markers (18S) reveal an exceptionally deep genetic structure in the widely distributed cladoceran, Polyphe- Discussion mus pediculus. Nine well-supported, largely allopatric This work extends the ongoing reappraisal of the distri- clades were identified across the Holarctic. Exception- bution patterns of genetic diversity in micro-eukaryotes. ally high genetic divergences were observed among the Although cosmopolitanism has been falsified in a range four major North American clades (13–24% COI diver- of widely distributed aquatic micro-eukaryotes includ- gences). These levels far exceed the intraspecific ing bryozoans (Nikulina et al. 2007), diatoms (Medlin COI divergences observed in other North American 2007) and a few rotifer taxa (Fontaneto et al. 2008; Se- cladocerans such as Sida crystallina (2–7%), Daphnia gers & De Smet 2008), it still holds true in most protists ambigua (3–5%) and Daphnia obtusa (1.2–15.5%) (Cox & and rotifers (Finlay 2002; Finlay et al. 2006; Fontaneto Hebert 2001; Hebert et al. 2003; Penton et al. 2004). Fur- et al. 2008). Extensive genetic surveys on freshwater thermore, the COI genetic divergences of Polyphemus on zooplankton have contributed to a recent biogeographic a global scale (6.3–24.4%) are much more pronounced paradigm shift from cosmopolitanism to pronounced than the global divergences found in other taxa such as provincialism (Hebert & Wilson 1994; Colbourne et al. Daphnia rosea s.l. (1–8%) (Ishida & Taylor 2007b). Most 1998; Taylor et al. 1998; Cox & Hebert 2001; Hebert importantly, the genetic divergences among the main et al. 2003; Adamowicz et al. 2004; Penton et al. 2004; phylogroups (COI: 6.3–24.4%, 16s: 3.1–11.5%, 18s: 0.2– Petrusek et al. 2004; Rowe et al. 2007). However, the 4.7%) reach levels characteristic of congeneric species, extent of cryptic genetic diversity at a global scale and and are comparable to the 15–23% COI sequence diver- the major evolutionary forces that shape this diversity gence often reported among various chydorid species are still poorly understood. (Sacherova & Hebert 2003), the 3–15% mitochondrial

2009 Blackwell Publishing Ltd 5174 S. XU ET AL.

0.30 0.20 NA1 NA2 0.25 0.15 0.20

0.15 0.10 Frequency 0.10 0.05 0.05

0.00 0.00 0123456789 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Pairwise differences Pairwise differences

0.15 0.15 NA3 PA1

0.10 0.10 Frequency 0.05 0.05

0.00 0.00 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 0 1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728293031 Pairwise differences Pairwise differences

Fig. 6 Pairwise haplotype mismatch distributions for the NA1, NA2, NA3 and PA1 clades. Bars represent the observed distribution of within-clade pairwise COI differences, whereas the solid line represents the expected pairwise mismatch distributions under the sudden expansion model.

ND5 gene divergence within the species the Pacific clade, branched off sometime during mid- complex (Colbourne et al. 1998), and the 5% 16s gene Miocene (10 Ma), and the subsequent cladogenetic divergence within the Daphnia laevis complex (Taylor events leading to the major clades proceeded from late et al. 1998). Miocene until early Pliocene (10–4 Ma). Although cryptic endemism is common in freshwater Cladocerans are characterized by striking variation zooplankton with wide distribution, only the rotifer in rates of phenotypic evolution (Rivier 1998; Benzie Brachionus plicatilis complex (Gomez et al. 2002) and chy- 2005). Although predatory cladocerans of the order dorid Chydorus sphaericus (Belyaeva & Taylor 2009) are Onychopoda represent a morphologically innovative similar to Polyphemus in terms of the exceptional cryptic group (Cristescu & Hebert 2002), Polyphemus is likely diversity. Nine cryptic species with 15–22% interspecific characterized by unusually pronounced morphological COI gene divergence have been recognized over the glo- stasis. With the exception of Polyphemus exiguus ende- bal distribution of B. plicatilis, whereas seven cryptic spe- mic to the (Rivier 1998), no other morpho- cies possessing 5–25% COI gene divergence have been species has been recognized in this genus (but see detected in the C. sphaericus complex across the northern Herrick 1879, 1884) despite the substantial molecular hemisphere. The deep genetic divergences among its divergence detected in P. pediculus. It is also likely main phylogroups strongly suggest that Polyphemus is an that, similar to all other Onychopods (Cercopagis, By- ancient species complex comprised of at least nine, tothrephes), Polyphemus have pronounced phenotypic highly divergent and largely allopatric cryptic species. plasticity that could obscure potential diagnostic Based on the COI phylogeny, the earliest derived clade, morphological differences between species. Detailed

2009 Blackwell Publishing Ltd GLOBAL PHYLOGEOGRAPHY OF POLYPHEMUS PEDICULUS 5175 morphological studies using males and gametogenetic prompted its rapid range expansion into northern, females that possess more diagnostic morphological harsher environments and facilitate long-distance inter- characteristics than the commonly studied parthenoge- continental dispersal via arctic corridors. In Polyphemus netic females (Kotov et al. 2009) are necessary to allopatric divergence following intercontinental dis- examine the extent of morphological stasis of this eco- persal accounts for about one-third of the speciation logically plastic species complex. events. Intercontinental dispersal followed by allopatric divergence has been proposed as an important specia- tion mechanism for freshwater zooplankton (Adam- Intercontinental dispersal owicz et al. 2009). A recent genetic survey of the Based on the results of the Bayesian inference of ances- genus Daphnia reveals that 30% of the speciation tral distributions, Polyphemus appears to have originated events can be attributed to allopatric speciation at an in North America and ⁄ or East Asia and at least three intercontinental scale (Adamowicz et al. 2009). major intercontinental dispersal events have been Phylogeographic patterns in Polyphemus indicate that involved in its subsequent allopatric divergence. One allopatric divergence is important not only at interconti- dispersal event from East Asia to North America (c. nental scale but also on single continents. Within North 4 Ma) was likely responsible for a recolonization of the America, the demographic analyses and nonoverlap- North American continent while two intercontinental ping distributions of the NA1, NA2 and NA3 clades dispersal events from North America to Europe (c.8 suggests their derivation from different glacial refugia. and 3 Ma) led to the colonization of Eurasia (Fig. 2). In the Russian Far East and Japan, the disjunct distribu- Few cases of relatively recent (possibly after the last tion of three endemic clades is also concordant with an glacial retreat) dispersal from North America to Europe allopatric speciation model. Less clear is the origin of have been documented in northern , the PA1 and PA2 clades in Europe and Siberia which suggesting that the arctic and ⁄ or northern Atlantic is an overlap extensively. The demographic analyses indicate important dispersal corridor via vectors such as migra- that the PA1 clade experienced a recent population tory birds and ice floes (Weider et al. 1999; Markova expansion c. 10 500–21 000 years ago (Tajima’s ) ) et al. 2007). The relatively low sequence divergences D = 1.44, P = 0.05; Fu’s Fs = 22.27, P = 0.00; mis-

(<5% at COI) between Europe and North America in a match analysis, s = 4.89, PSSD = 0.87). This population great number of taxa including gibberum expansion probably involved range expansions follow- (Rowe et al. 2007), s.l. (Ishida & Taylor ing the last glacial retreat, which could have driven the 2007a), D. rosea s.l. (Ishida & Taylor 2007b), D. pulex PA1 clade into secondary contact with the PA2 clade. (Colbourne et al. 1998) and rotifers (Fontaneto et al. 2008) suggest relatively recent intercontinental dispersal Effects of the Pleistocene glaciations (Adamowicz et al. 2009). The extensive genetic survey in the two major regions affected by the Pleistocene glaciations (North America Evolutionary radiation of Polyphemus and Europe) and in the unglaciated regions of the Rus- Although previous taxonomic studies identified two sian Far East and Japan enables evaluation of the effects ecological morphospecies including P. pediculus inhab- of Pleistocene glaciations on their respective phylogeo- iting lakes and Polyphemus stagnalis Herrick 1884 graphic structures. The general view is that genetic described from a shallow marsh (Herrick 1879, 1884), diversity of cladoceran populations in unglaciated our phylogenetic analyses do not show clear phyloge- regions is significantly higher than in formerly glaciated netic division between lake and pond haplotypes in all regions (Hewitt 1996; Ishida & Taylor 2007b). In Poly- the clades except for the NA4 and PA5 clades where phemus, two well-supported phylogroups are found in sampling efforts were exclusively focused on lakes, the Japanese Lake Aoki with 3% COI divergence, suggesting that habitat isolation is not an important whereas little phylogenetic subdivision is observed barrier to gene flow. In contrast, geographic isolation across Europe. Moreover, the nucleotide diversity in plays a dominant role. Considering that all other Ony- Lake Aoki (p = 0.014) is comparable to that of the entire chopods except Polyphemus (i.e. cercopagids and podo- PA1 clade (p = 0.015), which occurs across all of Europe nids) live either in large lakes or marine environments (Table 2). The Pleistocene glaciations have also left clear (Rivier 1998), the most recent common ancestor of signatures on the genetic structure of North American Onychopods most likely lived in permanent habitats. phylogroups. The shallow COI sequence divergences We suggest that adaptations to the temporal pond within the NA1, NA2 (excluding Oregon) and NA3 habitat might have occurred very early in the evolu- groups (0.8%, 1.8% and 1.4% respectively) suggest tionary history of Polyphemus and might have recent, rapid re-colonization of previously unsuitable

2009 Blackwell Publishing Ltd 5176 S. XU ET AL. habitats, which is estimated to have occurred c.7450– tic NA2 and NA3 clades of Polyphemus. The endemic 14 900, 5500–11 000 and 12 000–24 000 years ago respec- lineage in Oregon is divergent from the rest of the NA2 tively, after the retreat of the last glacial maximum. clade by 3% COI gene divergence, and the same Phylogeographic and network analyses allow us to divergence is found between the Massachusetts lineage pinpoint potential glacial refugia for North American and other haplotypes in the NA3 clade. The discovery and European Polyphemus. The NA1 phylogroup exclu- of an endemic lineage restricted to Oregon strengthens sively distributed in Yukon-Mackenzie appears to have the view that this region is a biodiversity hotspot for derived from Beringian, an important glacial refugium freshwater cladocerans. Oregon has been recognized as for cladocerans including D. rosea s.l., D. galeata s.l., Ho- a major glacial refugium for a number of cladocerans lopedium gibberum (Hebert et al. 2007) and the D. pulex such as Sida crystallina, Daphnia laevis, and D. ambigua complex (Weider & Hobaek 2003) that remained ice-free which have endemic phylogroups in Oregon (Taylor during the last glacial maximum (Pielou 1991). For the et al. 1998; Cox & Hebert 2001; Hebert et al. 2003). Fur- NA3 clade the central haplotypes C47 and C56 are dis- thermore, a variety of Daphnia species including D. are- tributed in the Great Lakes region and eastern Canada nata, D. melanica, D. oregonensis, D. villosa, D. latispina suggesting that this clade is derived from a refugium in and even an endemic cladoceran family, the Dumontii- eastern Canada, the centre of expansion for North dae (Santos-Flores & Dodson 2003) are restricted to American Daphnia pulex (Paland et al. 2005). For the Oregon. NA2 clade, the ancestral population most likely sought refuge along the west coast of North America where Conclusions many ice-free areas existed during the last glacial maximum (Pielou 1991). This study provides evidence for exceptionally high In Europe, the distribution range of the central hap- cryptic diversity in the predatory cladoceran Polyphemus, lotype C69 covers regions that were ice-free during the suggesting that continental and regional endemism are last glacial maximum such as the Russian Yaroslave common among freshwater zooplankton with interconti- region (Rus3) and Siberia (Rus21, Rus29, Rus30, nental distributions. Cosmopolitanism in freshwater Rus33), suggesting a northern glacial refugium. Palaeo- zooplankton often appears to represent an artefact due reconstructions of Northern Eurasia indicate that to morphological stasis, extensive , although ice sheets extended deep onto the mainland and poorly resolved rather than biological during the Quaternary (Mangerud et al. 2004), exten- reality. Our phylogeographic results indicate that multi- sive palaeolakes covering large parts of the White Sea ple intercontinental dispersal events from East Asia to basin and the West Siberia formed south of these ice North America and from North America to Europe have sheets due to the obstruction of the north-flowing riv- played an important role in the diversification of Poly- ers. The drainage of the main rivers was subsequently phemus. Further, the Pleistocene glaciations greatly diverted from the northern ice-dammed lakes towards reduced the genetic diversity of Polyphemus phylogroups the Ponto-Caspian basin, offering a possibility for in Europe and North America, and large parts of Europe southern dispersal to aquatic lineages trapped in the and North America became re-colonized after the last northern ice-dammed refugia. glacial maximum (5500–24 000 years ago). Recent phylogeographic studies on fishes, amphibi- ans, and insects have revealed a great number of Acknowledgements refugia outside the more conventional southern refu- gia in regions including Scandinavia, Alps, Danube We thank Z.A. Antipushina, M.A. Belyaeva, L.G. Butorina, J.R. and Ponto-Caspian region (Benke et al. 2009). In con- deWaard, R.T. Dooh, Y.R. Galimov, S.M. Glagolev, M.J. Gry- trast, little is known about the glacial refugia for gier, T. Hanazato, H. Ketelaars, N.M. Korovchinsky, O.A. Kry- lovich, Y. Kusuoka, A. Petrusek, N.M. Reshetnikova, V. European freshwater zooplankton. The few well-estab- Sacherova, A.B. Savinetsky, L.E. Savinetskaya, V.Y. Semashko, lished examples include refugia in the central Iberian X-L. Shi, M. Sakamoto, D.J. Taylor, A.J. Tessier and J.D.S. Witt Peninsula for the rotifers Branchionus plicatilis s.s. and for generously providing specimens. We also thank S.J. Adam- Branchionus manjavacas (Gomez et al. 2000, 2007). Fur- owicz, A. Zhan, J. Vaillant, L.J. Weider and three anonymous ther work is needed to explore how freshwater zoo- reviewers for providing constructive comments on an early plankton responded to the Pleistocene glaciations in version of this manuscript. This work was supported by Uni- Europe. versity of Windsor doctoral scholarships to SX, by Russian Foundation for Basic Research (Grants 09-04-00201-a and 09-04- The vicariance effects of the Pleistocene glaciations 10083-k) and Russian Biodiversity Program to AAK, and by have played a major role in shaping the intraspecific research grants from Natural Sciences and Engineering divergence in freshwater zooplankton (Cox & Hebert Research Council of Canada to PDNH and MEC. 2001). These effects are readily discerned in the Nearc-

2009 Blackwell Publishing Ltd GLOBAL PHYLOGEOGRAPHY OF POLYPHEMUS PEDICULUS 5177

References De Gelas K, De Meester L (2005) Phylogeography of Daphnia magna in Europe. Molecular Ecology, 14, 753–764. Adamowicz SJ, Hebert PDN, Marinone MC (2004) Species De Meester L, Gomez A, Okamura B, Schwenk K (2002) The diversity and endemism in the Daphnia of Argentina: a Monopolization Hypothesis and the dispersal-gene flow genetic investigation. Zoological Journal of the Linnean Society, paradox in aquatic organisms. Acta Oecologica, 23, 121–135. 140, 171–205. De Meester L, Vanoverbeke J, De Gelas K, Ortells R, Spaak P Adamowicz SJ, Petrusek A, Colbourne JK, Hebert PDN, Witt (2006) Genetic structure of cyclic parthenogenetic JDS (2009) The scale of divergence: a phylogenetic appraisal zooplankton populations – a conceptual framework. Archiv of intercontinental allopatric speciation in a passively fu¨r Hydrobiologie, 167, 217–244. dispersed freshwater zooplankton genus. Molecular Deangelis MM, Wang DG, Hawkins TL (1995) Solid-phase Phylogenetics and Evolution, 50, 423–436. reversible immobilization for the isolation of PCR products. Baas-Becking LGM (1934) Geobiologie of inleiding tot de Nucleic Acids Research, 23, 4742–4743. milieukunde. W.P. van Stockum and Zoon, The Hague, The Dooh RT, Adamowicz SJ, Hebert PDN (2006) Comparative Netherlands. phylogeography of two North American ‘glacial relict’ Bandelt HJ (1999) Median-joining networks for inferring . Molecular Ecology, 15, 4459–4475. intraspecific phylogenies. Molecular Biology and Evolution, 16, Excoffier L (2004) Patterns of DNA sequence diversity and 37–48. genetic structure after a range expansion: lessons from the Belyaeva M, Taylor DJ (2009) Cryptic species within the infinite-island model. Molecular Ecology, 13, 853–864. Chydorus sphaericus species complex (Crustacea: Cladocera) Excoffier L, Laval G, Schneider S (2005) ARLEQUIN version 3.0: revealed by molecular markers and sexual stage morphology. an integrated software package for data Molecular Phylogenetics and Evolution, 50, 534–546. analysis. Evolutionary Bioinformatics Online, 1, 47–50. Benke M, Brandle M, Albrecht C, Wilke T (2009) Pleistocene Felsenstein J (1985) Confidence limits on phylogenies – an phylogeography and phylogenetic concordance in cold- approach using the bootstrap. Evolution, 39, 783–791. adapted spring snails (Bythinella spp.). Molecular Ecology, 18, Finlay BJ (2002) Global dispersal of free-living microbial 890–903. eukaryote species. Science, 296, 1061–1063. Benzie JAH (2005) Cladocera: The Genus Daphnia (Including Finlay BJ, Esteban GF, Brown S, Fenchel T, Hoef-Emden K Daphniopsis) (: ) Backhuys, Leiden. (2006) Multiple cosmopolitan ecotypes within a microbial Brower AVZ (1994) Rapid morphological radiation and eukaryote morphospecies. Protist, 157, 377–390. convergence among races of the butterfly Heliconius erato Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R (1994) DNA inferred from patterns of mitochondrial-DNA evolution. primers for amplification of mitochondrial cytochrome c Proceedings of the National Academy of Sciences, USA, 91, 6491– oxidase subunit I from diverse metazoan . 6495. Molecular Marine Biology and Biotechnology, 3, 294–299. Burr BM, Mayden RL (1992) Phylogenetics and North Fontaneto D, Barraclough TG, Chen K, Ricci C, Herniou EA American freshwater fishes. In: Systematics, Historical Ecology, (2008) Molecular evidence for broad-scale distributions in and North American Freshwater Fishes (eds Mayden RL). pp. bdelloid rotifers: everything is not everywhere but most 18–75, Stanford University Press, Stanford, California. things are very widespread. Molecular Ecology, 17, 3136–3146. Butorina LG (1998) Resting eggs and hatching of young Forro´ L, Korovchinsky NM, Kotov AA, Petrusek A (2008) Polyphemus pediculus (Crustacea, Onychopoda). Ergebnisse Der Global diversity of cladocerans (Cladocera; Crustacea) in Limnologie, 52, 521–534. freshwater. Hydrobiologia, 595, 177–184. Butorina LG, Sergeev VN, Kartunova TA (1975) On the Frey DG (1982) Questions concerning cosmopolitanism in occurrence of Polyphemus pediculus (L.) in brackish waters. Cladocera. Archiv fu¨r Hydrobiologie, 93, 484–502. Informatsionnyi Bulleten ‘‘Biologiya vnitrennikh vod’’, 27, 28–31. Fu YX (1997) Statistical tests of neutrality of mutations against Colbourne JK, Hebert PDN (1996) The systematics of North population growth, hitchhiking and background selection. American Daphnia (Crustacea: Anomopoda): A molecular Genetics, 147, 915–925. phylogenetic approach. Philosophical Transactions of the Royal Gomez A, Carvalho GR, Lunt DH (2000) Phylogeography and Society of London, Series B: Biological Sciences, 351, 349–360. regional endemism of a passively dispersing zooplankter: Colbourne JK, Crease TJ, Weider LJ et al. (1998) Phylogenetics mitochondrial DNA variation in rotifer resting egg banks. and evolution of a circumarctic species complex (Cladocera: Proceedings of the Royal Society of London, Series B: Biological Daphnia pulex). Biological Journal of the Linnean Society, 65, Sciences, 267, 2189–2197. 347–365. Gomez A, Serra M, Carvalho GR, Lunt DH (2002) Speciation in Cox AJ, Hebert PDN (2001) Colonization, extinction, and ancient cryptic species complexes: evidence from the phylogeographic patterning in a freshwater crustacean. molecular phylogeny of Brachionus plicatilis (Rotifera). Molecular Ecology, 10, 371–386. Evolution, 56, 1431–1444. Crease TJ, Colbourne JK (1998) The unusually long small- Gomez A, Montero-Pau J, Lunt DH, Serra M, Campillo S subunit ribosomal RNA of the crustacean, Daphnia pulex: (2007) Persistent genetic signatures of colonization in sequence and predicted secondary structure. Journal of Brachionus manjavacas rotifers in the Iberian Peninsula. Molecular Evolution, 46, 307–313. Molecular Ecology, 16, 3228–3240. Cristescu MEA, Hebert PDN (2002) Phylogeny and adaptive Hebert PDN, Wilson CC (1994) Provincialism in plankton – radiation in the Onychopoda (Crustacea, Cladocera): endemism and allopatric speciation in Australian Daphnia. evidence from multiple gene sequences. Journal of Evolution, 48, 1333–1349. , 15, 838–849.

2009 Blackwell Publishing Ltd 5178 S. XU ET AL.

Hebert PDN, Witt JDS, Adamowicz SJ (2003) Nikulina EA, Hanel R, Schafer P (2007) Cryptic speciation and Phylogeographical patterning in Daphnia ambigua: regional paraphyly in the cosmopolitan bryozoan Electra pilosa – divergence and intercontinental cohesion. Limnology and impact of the Tethys closing on species evolution. Molecular Oceanography, 48, 261–268. Phylogenetics and Evolution, 45, 765–776. Hebert PDN, Rowe CL, Adamowicz SJ (2007) Life at low Paland S, Colbourne JK, Lynch M (2005) Evolutionary history temperatures: a novel breeding-system adjustment in a polar of contagious asexuality in Daphnia pulex. Evolution, 59, 800– cladoceran. Limnology and Oceanography, 52, 2507–2518. 813. Herrick CL (1879) Microscopic Entomostraca. Geological and Palumbi SR (1996) Nucleic acids II: the polymerase chain Natural History Survey of Minnesota, 7th Annual Report, reaction. In: Molecular Systematics (eds Hillis D, Mable B, pp. 81–123, Minneapolis, MN, USA. Moritz C), pp. 205–247. Sinauer, Sunderland, MA. Herrick CL (1884) A final report on the Crustacea of Penton EH, Hebert PDN, Crease TJ (2004) Mitochondrial DNA Minnesota. Geological and Natural History Survey of variation in North American populations of Daphnia obtusa: Minnesota, 12th Annual Report, pp. 1–191, St. Paul, MN, continentalism or cryptic endemism? Molecular Ecology, 13, USA. 97–107. Hewitt GM (1996) Some genetic consequences of ice ages, and Petrusek A, Cerny M, Audenaert E (2004) Large their role in divergence and speciation. Biological Journal of intercontinental differentiation of micrura (Crustacea: the Linnean Society, 58, 247–276. Anomopoda): one less cosmopolitan cladoceran? Huelsenbeck JP, Bollback JP (2001) Empirical and hierarchical Hydrobiologia, 526, 73–81. Bayesian estimation of ancestral states. Systematic Biology, 50, Pielou EC (1991) After the Ice Age: The Return of Life to Glaciated 351–366. North America. University of Chicago Press, Chicago, IL. Ishida S, Taylor DJ (2007a) Quaternary diversification in a Posada D, Buckley TR (2004) Model selection and model sexual Holarctic zooplankter, Daphnia galeata. Molecular averaging in phylogenetics: advantages of Akaike Ecology, 16, 569–582. Information Criterion and Bayesian approaches over Ishida S, Taylor DJ (2007b) Mature habitats associated with likelihood ratio tests. Systematic Biology, 53, 793–808. genetic divergence despite strong dispersal ability in an Posada D, Crandall KA (1998) MODELTEST: testing the model of . BMC Evolutionary Biology, 7, 52. DNA substitution. Bioinformatics, 14, 817–818. Ishida S, Kotov AA, Taylor DJ (2006) A new divergent lineage Posada D, Crandall KA (2001) Intraspecific gene genealogies: of Daphnia (Cladocera: Anomopoda) and its morphological trees grafting into networks. Trends in Ecology & Evolution, and genetical differentiation from Daphnia curvirostris 16, 37–45. Eylmann, 1887. Zoological Journal of the Linnean Society, 146, Ray N, Currat M, Excoffier L (2003) Intra-deme molecular 385–405. diversity in spatially expanding populations. Molecular Knowlton N, Weigt LA (1998) New dates and new rates for Biology and Evolution, 20, 76–86. divergence across the Isthmus of Panama. Proceedings of the Rivier IK (1998) The Predatory Cladocera (Onychopoda: , Royal Society of London, Series B: Biological Sciences, 265, 2257– Polyphemidae, Cercopagidae) and Leptodorida of the World. 2263. Backhuys, Leiden, The Netherlands. Knowlton N, Weigt LA, Solorzano LA, Mills DK, Bermingham Rogers AR, Harpending H (1992) Population growth makes E (1993) Divergence in proteins, mitochondrial-DNA, and waves in the distribution of pairwise genetic differences. reproductive compatibility across the Isthmus of Panama. Molecular Biology and Evolution, 9, 552–569. Science, 260, 1629–1632. Ronquist F (2004) Bayesian inference of character evolution. Korovchinsky NM (2006) The Cladocera (Crustacea : Trends in Ecology & Evolution, 19, 475–481. ) as a relict group. Zoological Journal of the Ronquist F, Huelsenbeck JP (2003) MRBAYES 3: Bayesian Linnean Society, 147, 109–124. phylogenetic inference under mixed models. Bioinformatics, Kotov AA, Ishida S, Taylor DJ (2009) Revision of the genus 19, 1572–1574. Baird, 1845 (Cladocera: ), based on Rowe CL, Adamowicz SJ, Hebert PDN (2007) Three new evidence from male morphological characters and molecular cryptic species of the freshwater zooplankton genus phylogenies. Zoological Journal of the Linnean Society, 156,1– Holopedium (Crustacea : Branchiopoda: ), revealed 51. by genetic methods. Zootaxa, 1656, 1–49. Mangerud J, Jakobsson M, Alexanderson H et al. (2004) Ice- Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R (2003) dammed lakes and rerouting of the drainage of northern DnaSP, DNA polymorphism analyses by the coalescent and Eurasia during the last glaciation. Quaternary Science Reviews, other methods. Bioinformatics, 19, 2496–2497. 23, 1313–1332. Sacherova V, Hebert PDN (2003) The evolutionary history of Markova S, Dufresne F, Rees DJ, Cerny M, Kotlik P (2007) the (Crustacea : Cladocera). Biological Journal of Cryptic intercontinental colonization in water fleas Daphnia the Linnean Society, 79, 629–643. pulicaria inferred from phylogenetic analysis of Santos-Flores CJ, Dodson SI (2003) Dumontia oregonensis n. mitochondrial DNA variation. Molecular Phylogenetics and fam., n. gen., n. sp., a cladoceran representing a new family Evolution, 44, 42–52. of ‘Water-fleas’ (Crustacea, Anomopoda) from USA, with Mayr E (1963) Species and Evolution. Harvard University notes on the classification of the Order Anomopoda. Press, Cambridge, MA. Hydrobiologia, 500, 145–155. Medlin LK (2007) If everything is everywhere, do they share a Schubart CD, Diesel R, Hedges SB (1998) Rapid evolution to common gene pool? Gene, 406, 180–183. terrestrial life in Jamaican crabs. Nature, 393, 363–365.

2009 Blackwell Publishing Ltd GLOBAL PHYLOGEOGRAPHY OF POLYPHEMUS PEDICULUS 5179

Schwenk K, Sand A, Boersma M et al. (1998) Genetic markers, genealogies and biogeographic patterns in the cladocera. Sen Xu is a PhD student working in Cristescu’s lab on the evo- Aquatic Ecology, 32, 37–51. lution of reproductive strategies in aquatic crustaceans. Paul Segers H, De Smet WH (2008) Diversity and endemism in Hebert’s research interests include molecular evolution, biodi- Rotifera: a review, and Keratella Bory de St Vincent. versity and DNA based identification systems. Alexey Kotov is Biodiversity and Conservation, 17, 303–316. interested in the systematics, phylogenetics, and biogeography Slatkin M, Hudson RR (1991) Pairwise comparisons of of freshwater cladocerans. Melania Cristescu uses various mitochondrial DNA sequences in stable and exponentially molecular tools to investigate the evolutionary causes and con- growing populations. Genetics, 129, 555–562. sequences of aquatic invasions. Tajima F (1989) Statistical-method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics, 123, 585–595. Tamura K, Nei M (1993) Estimation of the number of Supporting information nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Molecular Additional supporting information may be found in the online Biology and Evolution, 10, 512–526. version of this article. Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: Table S1 The mean COI, 16s, 18s sequence divergences molecular evolutionary genetics analysis (MEGA) between the identified phylogroups estimated using Tamura– software version 4.0. Molecular Biology and Evolution, 24, Nei model 1596–1599. Taylor DJ, Finston TL, Hebert PDN (1998) Biogeography of a File S1. Alignment of 201 unique COI haplotypes in NEXUS widespread freshwater crustacean: pseudocongruence and format. cryptic endemism in the North American Daphnia laevis File S2. Alignment of 69 concatenated COI and 16s haplotypes complex. Evolution, 52, 1648–1670. in NEXUS format. Weider LJ, Hobaek A (2003) Glacial refugia, haplotype distributions, and clonal richness of the Daphnia pulex File S3. Alignment of ten unique 18s haplotypes in NEXUS format. complex in arctic Canada. Molecular Ecology, 12, 463–473. Please note: Wiley-Blackwell are not responsible for the content Weider LJ, Hobaek A, Colbourne JK et al. (1999) Holarctic or functionality of any supporting information supplied by the phylogeography of an asexual species complex I. authors. Any queries (other than missing material) should be Mitochondrial DNA variation in arctic Daphnia. Evolution, 53, directed to the corresponding author for the article. 777–792.

2009 Blackwell Publishing Ltd