Molecular Phylogenetics and Evolution 37 (2005) 474–483

www.elsevier.com/locate/ympev

The phylogeny and taxonomy of austral monodontine topshells (Mollusca: Gastropoda: Trochidae), inferred from DNA sequences �

Kirsten M. Donald ¤, Martyn Kennedy, Hamish G. Spencer

Allan Wilson Centre for Molecular Ecology and Evolution, Department of Zoology, University of Otago, P.O. Box 56, Dunedin, New Zealand

Received 15 December 2004; revised 17 March 2005Available online 3 June 2005

Abstract

The systematics of topshells (family Trochidae) is currently unresolved: at present even the generic boundaries within this groupare poorly deWned. In this study, we used sequence data of two mitochondrial genes (16S and cytochrome oxidase 1, COI) and onenuclear gene (actin) to resolve the phylogeny of a closely related subgroup of the Trochidae, 30 species of largely Southern Hemi-sphere monodontine topshells. The phylogenies constructed revealed Wve well-supported generic clades: a South African clade (genusOxystele Philippi, 1847), which lay basally to four internal PaciWc clades (genera Chlorodiloma Pilsbry, 1889; Monodonta Lamarck,1799; Austrocochlea Fischer, 1885; and Diloma Philippi, 1845). The molecular phylogenies constructed in this study shed light on pre-viously unresolved relationships between diVerent groups of topshells, allowing for the Wrst time assignation (based on DNAsequence) of clearly deWned, well-supported taxonomic and nomenclatural classiWcation of monodontine topshells species. Austroc-ochlea crinita (Philippi, 1849), A. odontis (Wood, 1828), A. adelaidae (Philippi, 1849), and A. millelineata (Bonnet, 1864) are placed inthe genus Chlorodiloma, which we resurrect from synonymy with Austrocochlea. The Japanese M. confusa Tapparone-Canefri, 1874is treated as a separate species from M. labio (Linné, 1758). Melagraphia Gray, 1847 is synonymised with Diloma and its sole mem-ber, M. aethiops (Gmelin, 1791), along with A. concamerata (Wood, 1828), is transferred to that genus. The Juan Fernandez endemicD. crusoeana (Pilsbry, 1889) (Pilsbry, 1889) is synonymised with D. nigerrima (Gmelin, 1791). We Wnd that morphologically crypticspecies are not necessarily close genetically. 2005 Elsevier Inc. All rights reserved.

Keywords: Generic boundaries; Molecular phylogeny; Monodonta; Diloma; Austrocochlea; Chlorodiloma; Melagraphia; Oxystele; Osilinus

1. Introduction the South PaciWc Ocean (Dell, 1971; Morton and Miller,

Monodontine topshells currently ascribed to the gen-era Melagraphia Gray, 1847; Diloma Philippi, 1845;Austrocochlea Fischer, 1885; and Monodonta Lamarck,1799 are often the dominant gastropod, in terms of num-bers or biomass, in the intertidal zone in many areas of

� Nucleotide sequence data reported in this paper are available in theGenBank, EMBL, and DDBJ databases under the Accession Nos. areAY855315–AY855336, AY858057–AY858100, and DQ061080–DQ061104.

* Corresponding author. Fax: +64 03 479 7584.E-mail address: kirsten.donald@stonebow.otago.ac.nz (K.M.

Donald).

1055-7903/$ - see front matter 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.ympev.2005.04.011

1968; Powell, 1979; Schwabe and Barclay, 2003; Wilson,1993). Most recent classiWcations (e.g., Hickman, 1998;Hickman and McLean, 1990) place the species tradition-ally considered as belonging to the subfamily Mono-dontinae into the tribe Gibbulini, within the subfamilyTrochinae. Nevertheless, in keeping with traditionalusage, we retain the descriptor “monodontine” for thosespecies that Hickman (1998) called “Monodonta-like”and which have at various times been classiWed withinthat genus. In spite of the revised taxonomy, the mon-odontine topshells are clearly closely related.

In New Zealand seven monodontine species are cur-rently recognised, one of which is placed in the genusMelagraphia and the remaining six in the genus Diloma

K.M. Donald et al. / Molecular Phylogenetics and Evolution 37 (2005) 474–483 475

(Spencer and Willan, 1996). All the New Zealand speciesare endemic with the intriguing exception of Diloma nig-errima, which also inhabits the Chilean coast (Dell, 1971;Powell, 1979). Four additional Diloma species are foundon PaciWc islands: D. radula, which inhabits a large areaof the PaciWc stretching from Samoa to Japan (Schwabeand Barclay, 2003), D. samoaensis, a Samoan endemic(Schwabe and Barclay, 2003), D. constellatus, anendemic to New Caledonia which was not sampled inthis study, and D. crusoeana, a species whose distribu-tion is thought to be limited to the Juan FernandezIslands, oV the coast of Chile. Related forms, mostrecently ascribed to the genus Austrocochlea are com-mon in Australia (Wilson, 1993), with about 10 speciescurrently recognized. Several species, belonging to thegenus Monodonta also inhabit the coasts of eastern Asia(Higo et al., 1999). Topshell distribution is not restrictedto the PaciWc, indeed it is worldwide and this studyincludes Wve topshell species belonging to the genusOxystele (sometimes treated as a subgenus of Diloma),which are common around the coast of South Africa(Heller and Dempster, 1991), as well as a representativeof the Afro-European genus Osilinus.

Despite their ubiquitous nature, the systematics ofthis group of snails is poorly resolved at present withdiVerent authors grouping diVerent species together,largely because shell characters used in traditional classi-Wcations appear to show signiWcant homoplasy (Hick-man, 1998). For example, some Australian workers (e.g.,Cotton, 1959) have placed species usually considered tobe members of Austrocochlea in Fractarmilla (which hasa New Zealand type), and used this taxon at the genericlevel, a practice never followed in New Zealand, where itis viewed as an endemic subgenus of Diloma. At the spe-cies level, most taxa have a plethora of available namesdating from the 18th and 19th centuries, the validity ofwhich has only become apparent within the last fewdecades. In addition, genetic work has revealed severalcryptic species in Australia (Parsons, 1996; Parsons andWard, 1994) and South Africa (Heller and Dempster,1991).

To elucidate relationships between the various mon-odontine taxa we obtained molecular data to constructclearly resolved, well-supported phylogenetic trees.These molecular phylogenies allow us to assign each ofthese ecologically important species the generic namemost appropriate to its position in the phylogenetictree.

2. Materials and methods

2.1. Sample collection

Trochid snails, representing 30 species, were collectedfrom intertidal sites in New Zealand, Australia, Japan,

Samoa, Chile, South Africa, and the United Kingdom(see Table 1 for species collected and sampling loca-tions). The neozelanic species Micrelenchus huttonii andCantharidella tesselata, also members of the Trochinae,were also collected for use as outgroups. Hickman andMcLean (1990) placed Micrelenchus in the tribe Cantha-ridini, whereas Cantharidella was placed in the Gibbu-lini, although it is not a monodontine. Live snails wereimmediately crushed upon collection, preserved in 70%ethanol, and returned to the laboratory where they werestored at 4 °C.

2.2. DNA extraction and PCR ampliWcation

A 2 mm3 piece of foot tissue was dissected from eachtopshell and rinsed in distilled water to remove any etha-nol. This tissue was then placed in a 5% Chelex 100(Walsh et al., 1991) solution, which was incubated over-night at 65 °C. Following brief vortexing, the solutionwas boiled for 10 min and then centrifuged at 15,000g for10 min. The DNA in the supernatant was used in subse-quent PCR ampliWcations.

Regions of the mitochondrial genes, 16S and cyto-chrome oxidase I (COI), and the nuclear gene, actin, werePCR-ampliWed from the topshell DNA following the pro-cedure of Donald et al. (in press). Annealing temperaturesof 48°C were used for 16S ampliWcation and 50 °C forCOI and actin ampliWcation. The universal 16S primers,16Sar (5�-CGCCTGTTTATCAAAAACAT-3�) and 16Sbr(5�-CCGGTCTGAACTCAGATCACGT-3�), describedby Simon et al. (1994), were used to amplify 16S. The uni-versal invertebrate COI primer LCO1490 (5�-GGTCAACAAATCATAAGATATTGG-3�) (Folmer et al., 1994)and the invertebrate COI primer H7005 (5�-TAACTTCAGGGTGACCAAAAAATCA-3�) (Hafner et al.,1994) were routinely used to amplify COI. If COI ampliW-

cation was unsuccessful under these conditions, the H7005primer was replaced with the alternative H7005-mod1primer (5�-ARTGNGCNACNACRTARTANGTRTCRTG-3�) (Donald et al., in press). Actin primers, Act.1F(5�-TATGTTGGTGATGAGGCTCAG-3�) and Act.1R(5�-TGGGAACATGGTGGAACC-3�), deWned fromplanorbid snail Biomphalaria glabrata actin sequence(Adema, 2002; AF329436), were used to amplify actin. ToconWrm that ampliWcations were successful, 5�l aliquotsof PCR ampliWcations was visualised by agarose gel elec-trophoresis.

The PCR products were separated from excess prim-ers and oligonucleotides in High Pure PCR PuriWcationColumns (Roche). PuriWed DNA was quantiWed usingagarose gel electrophoresis and subsequently sent to theAllan Wilson Centre Genome Sequencing Service forsequencing in an ABI3730 automated sequencer.Sequencing reactions used the original PCR primers,diluted to a Wnal concentration of 0.213�M, as sequenc-ing primers for 16S, COI, and actin. To check that the

476K

.M. D

onald et al. / Molecular P

hylogenetics and Evolution 37 (2005) 474–483

Table 1Topshell speci

Topshell speci der

Micrelenchus S. Miller, University of Otago, New ZealandCantharidella nd H. Spencer, University of Otago, New ZealandOsilinus lineat niversity of Otago, New ZealandChlorodiloma and Y. Bone, University of Adelaide, AustraliaC. odontis (W C Rowley, Victoria Museum, AustraliaC. crinita (Phi s and S. Slack-Smith, Western Australian Museum, AustraliaAustrocochlea d B. Russell, Museum and Art Gallery of the Northern TerritoryA. rudis (Gray th and C. Whisson, Western Australian Museum, AustraliaA. constricta ( niversity of Otago, New ZealandA. porcata (A. niversity of Otago, New ZealandA. brevis (Pars niversity of Otago, New ZealandOxystele sinen atal Museum, South AfricaO. tabularis (K atal Museum, South AfricaO. tigrina (An niversity of Cape Town, South AfricaO. variegata (A niversity of Cape Town, South AfricaO. impervia (M niversity of Cape Town, South AfricaMonodonta la d B. Russell, Museum and Art Gallery of the Northern TerritoryM. labio (Linn eikai National Fisheries Research Institute, JapanM. confusa (T aido University, JapanM. canalifera eikai National Fisheries Research Institute, JapanM. perplexa p aido University, JapanDiloma concam niversity of Otago, New ZealandD. radula (Phi eikai National Fisheries Research Institute, JapanD. samoaensis Zoologische Staatssammlung München, GermanyD. coracina (P ucklandD. bicanalicula ucklandD. nigerrima ( S. Miller, University of Otago, New ZealandD. nigerrima ( iversidad Catolica de la Santisima Concepción, ChileD. crusoeana ( J. Gerber, The Field Museum of Natural History, USAD. zelandica (Q S. Miller, University of Otago, New ZealandD. arida (Finl S. Miller, University of Otago, New ZealandD. subrostrata S. Miller, University of Otago, New ZealandD. aethiops (G S. Miller, University of Otago, New Zealand

es collected and their sampling sites

es Sampling site Collector/sen

huttonii (Smith, 1876) Waitati Inlet, Otago, New Zealand A. Miller andtesselata (Adams, 1851) Allans Beach, Otago, New Zealand K. Donald aus (da Costa, 1778) Wembury Beach, Devon, United Kingdom K. Donald, Uadelaidae (Philippi, 1849) Adelaide South Australia, Australia E. Campbellood, 1828) Melbourne, Victoria, Australia R. Burn andlippi, 1849) Cape Leeuwin, Western Australia, Australia A. Cumming diminuta (Hedley, 1912) Darwin, Northern Territory, Australia S. Horner an, 1826) Perth, Western Australia, Australia S. Slack-SmiLamarck, 1822) Porpoise Hole, Tasmania, Australia K. Donald, U Adams, 1853) Manly, New South Wales, Australia K. Donald, Uons and Ward, 1994) Pitt Water, Tasmania, Australia K. Donald, Usis (Gmelin, 1791) Glen GariV, East London, South Africa M. Bursey, Nrauss, 1848) Glen GariV, East London, South Africa M. Bursey, N

ton, 1838) Cape Town, South Africa G. Branch, Unton, 1838) Cape Town, South Africa G. Branch, Uenke, 1843) False Bay, South Africa G. Branch, U

bio (Linné, 1758) Darwin Harbour, Australia S. Horner ané, 1758) Okinawa, Japan Y. Takada, S

apparone-Canefri, 1874) Kattoshi, Hokkaido, Japan A. Ito, Hokk(Lamarck, 1822) Okinawa, Japan Y. Takada, Serplexa (Pilsbry, 1889) Cape Tachimachi, Hokkaido, Japan A. Ito, Hokk

erata (Wood, 1828) Taranna, Tasmania, Australia K. Donald, Ulippi, 1849) Okinawa, Japan Y. Takada, S (Schwabe and Barclay, 2003) Tutuila Island, American Samoa (ZSM Moll 20020495) E. Schwabe, hilippi, 1851) Boulder Beach, Auckland, New Zealand M. Morley, Ata Dunker in (Philippi, 1844) Boulder Beach, Auckland, New Zealand M. Morley, A

Gmelin, 1791) Menzies Bay, Banks Peninsula, New Zealand A. Miller andGmelin, 1791) Concepción, Chile V. Lobos, UnPilsbry, 1889) Juan Fernandez Islands (lot. no. FMNH-301999) R. Bieler anduoy and Gaimard, 1834) Warrington, Otago, New Zealand A. Miller and

ay, 1926) Purakaunui Bay, Otago, New Zealand A. Miller and (Gray, 1835) Heathcote Estuary, Christchurch, New Zealand A. Miller andmelin, 1791) Purakaunui Bay, Otago, New Zealand A. Miller and

K.M. Donald et al. / Molecular Phylogenetics and Evolution 37 (2005) 474–483 477

correct gene had been ampliWed, BLAST searches (Altsc-hul et al., 1990) were performed to compare ampliWedsequences with all sequences stored in the EMBL data-base.

In all cases several individuals of each species weresequenced, in all cases very little sequence divergence wasdetected within each species (<1% for all genes). Conse-quently only a single randomly chosen individual of eachspecies was represented in the subsequent analysis.

2.3. Phylogenetic analyses

Sequences were aligned using the ClustalX alignmentprogram, run at default parameters (Thompson et al.,1997). The partition-homogeneity test (SwoVord, 2002;the ILD of Farris et al., 1995) was performed to testwhether the 16S, COI, and actin sequences contained thesame signal and could thus be analysed as a single data-set (taxa for which a gene region was unable to besequenced were excluded from this analysis). Subsequentphylogenetic analyses were performed with PAUP* ver-sion 4b10 (SwoVord, 2002) for neighbor joining (NJ)(using GTR corrected distances), weighted maximumparsimony (MP) searches (with a 4:1 transversion:transi-tion ratio, estimated by maximum likelihood on the treefound from an equally weighted MP search), and boot-strap values (Felsenstein, 1985, 1988). MrBayes v.3.0(Huelsenbeck and Ronquist, 2001; Ronquist and Huel-senbeck, 2003) was used to estimate Markov-chainMonte-Carlo Bayesian posterior probabilities. Manyauthors have recently expressed concerns—reviewed inSimmons and Pickett (2004)—that posterior probabili-ties overestimate branch support. For example, a poster-ior probability of 1.00 does not mean that the groupingis certain; the probability is conditional on the data, theprior (in our case, uniform) distribution of trees and,most importantly, the likelihood model of molecularevolution. To reduce the sensitivity of the Bayesianapproach to the incorrect model being speciWed, Huel-senbeck and Rannala (2004) suggest using the mostcomplex model available (e.g., partitioning the dataset soas to parameterise the molecular evolution model foreach gene separately), a recommendation we follow here.The robustness of our data to diVerent methods of phy-logenetic reconstruction gives an informal measure ofthe reliability of the resultant trees (Kim, 1993). Forvisualisation purposes, the related non-monodontinetopshells Micrelenchus huttonii and Cantharidella tesse-lata were deWned as outgroup taxa for all genes analysed.

The NJ bootstrap analyses consisted of 10,000 repli-cates. The weighted MP bootstrap analyses consistedof 1000 replicates using a heuristic search (with 10 ran-dom addition sequence replicates and TBR branch-swapping). The models of nucleotide substitution forthe Bayesian analysis were selected using Modeltest(Posada and Crandall, 1998). The model selected for

each gene region was either GTR + I + G or a sub-model of this model, and thus we used GTR + I + G forall three genes. Bayesian analysis was performed usingMrBayes v3.0 with the following settings for each ofthe three partitions (i.e., genes). The maximum-likeli-hood model employed six substitution types(“nst D 6”). Rate variation across sites was modelledusing a gamma distribution, with a proportion of thesites being invariant (“rates D invgamma”). The shape,proportion of invariable sites, state frequency, and sub-stitution rate parameters were estimated for each parti-tion separately. The Markov-chain Monte-Carlosearch was run with four chains for 2,000,000 genera-tions, with trees being sampled every 100 generations(the Wrst 5000 trees, i.e., 500,000 generations, were dis-carded as “burnin”).

3. Results

3.1. Molecular data

PCR ampliWcation and sequencing routinely yieldedapproximately 565 bp of 16S, 998 bp of COI, and 764 bpof actin sequence. Subsequent BLAST searches demon-strated that all gene sequences obtained displayed high-est identity with 16S, COI, and actin sequences fromother gastropod species. Our 16S sequences were mostsimilar to the GenBank 16S sequence from the topshellsPseudotalopia sakurai (AY163409) (E value < 3 £ 10¡79)and Gibbula cineraria (AY163410) (E value < 1 £ 10¡72).COI sequences showed highest identity with COIsequence data from the turbinid Turbo (Batillus) cornu-tus (D89277) (E value <1 £ 10¡105) and also displayedhigh identity with COI sequence from the topshellTegula viridula (AF132340) (E value <6 £ 10¡78). Actinsequences showed highest identity with actin sequencedata from the abalone Haliotis discus hannai(AY380809) (E value »0.0) and the seahare Aplysia cali-fornica (U01352) (E value < 1 £ 10¡176). These BLASTsearches indicated that in all cases the correct topshellgenes had been sequenced.

3.2. Molecular phylogeny

Gastropods may possess multiple actin isoforms(Adema, 2002) and we cannot rule out the possibilitythat diVerent actin sequences were derived from paralo-gous actin genes. Nevertheless, the primers weredesigned from a single form and our actin topology wasconcordant with the topologies constructed for both 16Sand COI mitochondrial genes. Indeed, a partition-homo-geneity test conWrmed that there was no signiWcantdiVerence in the phylogenetic signal among thesesequences (100 replicates, P D 0.06) and thus the threegenes were subsequently concatenated and analysed as a

478 K.M. Donald et al. / Molecular Phylogenetics and Evolution 37 (2005) 474–483

single dataset consisting of 2169 characters. The com-bined dataset had 681 variables and 605 parsimonyinformative characters.

The trees produced by weighted MP and Bayesiananalysis of the combined dataset are shown in Fig. 1. MPanalysis yielded three very similar trees, which onlydiVered within the D. nigerrima and D. crusoeana node,the three organisms within this node are thus repre-sented as a polytomy (Fig. 1A). At the generic level theNJ (not shown) and Bayesian analyses produced verysimilar topologies, while the weighted MP topology onlydiVered from these in the relative position of genus Aust-rocochlea. All the phylogenies provide evidence that themonodontine southern hemisphere topshells analysed inthis study could be separated into Wve well-supportedgroups or genera. The level of support for each genusdiVered between methods: 58–100% bootstrap values for

NJ (four out of Wve genera had >98% support), 81–100%bootstrap values for weighted MP, and all posteriorprobabilities of 1.00 for the Bayesian topology.

Genetic distances between the Wve genera ranged from16.3 to 19.4%, whereas genetic distances within generaranged from 5.7 to 12.1% (Table 2). Despite considerablegenetic distance being observed within several of the gen-era studied, an observation that was unsurprising giventhe considerable interspeciWc morphological diVerenceswithin each genus, between-genus genetic distances wereconsistently higher than within-genus distances.

Although the European Osilinus lineatus is classiWed asa monodontine, NJ and MP analysis provided no evi-dence that this species could be grouped with any of thesouthern hemisphere species (bootstrap support <50%)on the basis of its DNA sequence. Bayesian analysis pro-vided a moderate posterior probability (0.86) to suggest

Fig. 1. (A) The weighted maximum parsimony (MP) phylogram (using a 4:1 transversion:transition ratio, estimated by maximum likelihood) gener-ated from 2169 bp combined 16S, COI, and actin sequence data. MP analysis yielded three very similar trees, which only diVered within the D. nigerr-ima and D. crusoeana node; the tree selected to be shown (one of the three MP trees) represents this cluster as a polytomy. Numbers in roman font atnodes are weighted MP bootstrap values (>50%) calculated from 1000 replicates using a heuristic search (with 10 random addition sequence repli-cates and TBR branch-swapping). Italic numbers are neighbor-joining (NJ) bootstrap values (>50%) from 10,000 replicates generated using GTR-corrected distances from the same data. (B) The Bayesian phylogram, estimated using Markov-chain Monte-Carlo (MCMC) Bayesian posteriorprobabilities, generated from 2169 bp combined 16S, COI, and actin sequence data. The dataset was partitioned to allow model parameters to be esti-mated for each gene separately. Numbers at nodes are Bayesian posterior probabilities.

K.M. Donald et al. / Molecular Phylogenetics and Evolution 37 (2005) 474–483 479

that O. lineatus was more closely related to the SouthAfrican species (genus Oxystele) than the PaciWc species.All three methods displayed strong support for the mono-phyly of Oxystele (NJ, 100%; MP, 100%; and Bayesian,1.00). All analyses found Oxystele to be situated basal tothe PaciWc topshells, which fell into four separate genera.Most basal of the PaciWc genera was the well-supportedgroup Chlorodiloma (100%, 100%, and 1.00), which com-prised three Australian species, most recently placed inAustrocochlea (Wilson, 1993). Sister to Chlorodiloma werethe three genera Monodonta (98%, 81%, and 1.00), Aust-rocochlea (58%, 84%, and 1.00) and Diloma (100%, 100%,and 1.00). Monodonta comprised four Japanese and oneAustralian species, Austrocochlea consisted of Wve Austra-lian species and Diloma included all New Zealand, bothChilean, one Australian and two PaciWc topshell species.The various generic and subgenera names used in NewZealand (Melagraphia, Fractarmilla, Cavodiloma, Anisodi-loma, and Zediloma) are thus all synonyms of Diloma. Atthe generic level the sole diVerence between the threetopologies was the relative position of Austrocochlea. NJand Bayesian analysis suggested that this group was mostclosely related to Diloma whereas MP analysis weaklyimplied it was most closely related to Monodonta.

At the species level, within Diloma, the Juan Fernan-dez endemic D. crusoeana falls within D. nigerrima fromNew Zealand and mainland Chile. Within Monodonta,the two Japanese taxa most commonly referred to asM. labio labio and M. labio confusa are two geneticallydistinct taxa, which we therefore treat as separate spe-cies; M. labio and M. confusa, respectively. The M. labioindividuals sampled from the Northern Territory, Aus-tralia, are genetically distinct from the JapaneseM. labio, suggesting that this Indo-PaciWc taxon is in facta species complex. An abbreviated synonymy, summaris-ing the classiWcation derived from our phylogeny, isshown in Appendix A.

4. Discussion

This research used molecular techniques and phyloge-netic analyses to elucidate relationships between Wveecologically important topshell genera, ultimately allow-ing us to determine the best-supported grouping of top-shells, assign each taxon the speciWc and generic name

most appropriate to its position in the tree and conWrmwhether current taxonomic assignments are valid.

The relative positions of the monophyletic genericgroups in the molecular phylogenies were similar irre-spective of the tree-building method used (which givesconWdence in the results, Kim, 1993). All the topologiescomprised Wve monophyletic groups of species, one ofwhich was composed of South African species while theremainder comprised the PaciWc species, to which weassigned generic names. Oxystele should be used at thegeneric level for the monophyletic South African groupof species rather than as a subgenus of Diloma. Chlorodi-loma (resurrected from synonymy with Austrocochlea),Monodonta, Austrocochlea, and Diloma ( D Melagra-phia) are appropriate for the four PaciWc groups. Thisarrangement necessitates name changes for some species,as detailed below.

4.1. Genus Oxystele

All topologies unequivocally placed the South Afri-can species monophyletically and basal to the PaciWcgroups, to which they bear little genetic resemblance.Thus Oxystele is more appropriate used at the genuslevel rather than as a subgenus of Diloma. We note alsothat the Oxystele species O. variegata and O. imperviaare genetically relatively alike (see Fig. 1). Indeed, formany years these two South African topshell specieswere considered to be a single species, O. variegata. Onthe basis of allozyme data, shell colour pattern andproperties of the central tooth of the radula, however,Heller and Dempster (1991) suggested that this taxonwas in reality two separate species, each inhabitingdiVerent areas of the intertidal zone. The level of geneticdistinctiveness we Wnd for these taxa (1.96% sequencediVerence) supports Heller and Dempster’s (1991) Wnd-ings that, although similar, O. variegata and O. imperviaare indeed two separate species. The West African O. ful-garata, which we did not sample, is also a member of thisgenus, possibly synonymous with one of these last twospecies (Heller and Dempster, 1991).

4.2. Genus Chlorodiloma

The PaciWc species comprised four genera, all of whichlie interior to the basal South African Oxystele genus.

Table 2Genetic distances within and between Wve topshell genera

Figures are the average percent sequence diVerence (§standard deviation) for COI calculated using the general time-reversible matrix. Bold valuesindicate average within-genus distances.

Topshell genus 1 2 3 4 5

1. Diloma Philippi, 1845 10.0 § 2.9% — — — —2. Austrocochlea Fischer, 1885 16.3 § 2.4% 9.1 § 6.4% — — —3. Chlorodiloma Pilsbry, 1889 16.7 § 2.0% 17.3 § 0.7% 5.7 § 3.6% — —4. Monodonta Lamarck, 1799 16.8 § 2.3% 17.4 § 0.8% 16.8 § 1.0% 12.1 § 3.0% —5. Oxystele Philippi, 1847 18.6 § 2.9% 19.4 § 2.9% 17.4 § 2.3% 19.0 § 2.9% 12.0 § 4.5%

480 K.M. Donald et al. / Molecular Phylogenetics and Evolution 37 (2005) 474–483

Chlorodiloma, the most basal of the Australasian genera,is comprised of the species C. adelaidae, C. crinita, andC. odontis, and has most recently been treated as a syno-nym of Austrocochlea (e.g., Wilson, 1993). As these threespecies are genetically distinct from the remainder of theAustralian Austrocochlea species (see Fig. 1), it is clearthat Chlorodiloma should be resurrected from synonymywith Austrocochlea. Chlorodiloma adelaidae, C. crinita,and C. odontis are conchologically similar and their distri-bution overlaps to an extent, stretching from WesternAustralia (C. adelaidae and C. crinita) through SouthAustralia, Victoria, and Tasmania (C. adelaidae andC. odontis) (Wilson, 1993). Chlorodiloma millelineata,from Torres Strait, which we were unable to obtain, isalso apparently congeneric (Cotton, 1959).

4.3. Genus Monodonta

Our Monodonta sampling consisted of one Australianand the four Japanese populations. We note that, basedon their DNA sequence data, M. labio from northernAustralia, and M. l. labio and M. l. confusa from Japanare all genetically divergent (see Fig. 1). Thus, we arguethat all three are distinct species. It seems that the taxonM. labio, which has an Indo-PaciWc wide distribution, isactually a species complex and requires further investi-gation. Interestingly, Takenouchi (1985) found thatthree colour forms of M. labio from Hong Kong haddiVerent abundances according to the degree of waveexposure of the rocky shore habitat. This pattern wasreplicated in Japan (Takenouchi, 1985). Our Japanesesamples came from a boulder shore protected from theopen ocean by a fringing coral reef (Y. Takada, pers.comm.), whereas our Australian samples were collectedfrom a semi-sheltered locality (R.C. Willan, pers.comm.). All this evidence, paralleling the case of the pre-viously undistinguished Oxystele variegata andO. impervia, suggests that these diVerent forms may bespeciWcally distinct.

The validity of the subgenus Neomonodonta Kurudaet al., 1971; which contains the closely related species M.perplexa and M. neritoides, also depends on which of thetwo forms of M. labio is that of Linné. We intend tosample populations of M. labio from several sitesthroughout the PaciWc to determine how many distinctgenetic groups comprise this apparent species complex.Our sampling of this group was not exhaustive and thereare many other species referred to as Monodonta fromvarious tropical locations (see Appendix A), whichclearly should be examined genetically.

4.4. Genus Austrocochlea

We restrict Austrocochlea to just Wve of the Australianspecies studied, as well as the unsampled West Austra-lian A. zeus, sometimes placed in Chrysostoma Swainson,

1844. Three of these species, A. constricta, A. porcata,and A. brevis, were expected to be closely related as theyare diYcult to separate morphologically (Liddell, 1950)and until fairly recently were classiWed as a single vari-able species: ‘A. constricta species complex’ (Macpher-son, 1961). Measurements of allele frequencies at Wvepolymorphic loci, and subtle anatomic diVerencesrevealed them to be three separate species (Parsons,1996; Parsons and Ward, 1994), a Wnding reinforced bythis study. The high level of genetic similarity betweenA. constricta and A. rudis in the molecular phylogeny(see Fig. 1) is surprising since these two species are mor-phologically distinct from each other.

4.5. Genus Diloma

Diloma consists of one Australian species and theremainder of the PaciWc species, collected from NewZealand, Japan, Samoa, and Chile. All but two of thesespecies are already commonly referred to as Diloma.These exceptions are the neozelanic D. aethiops (previ-ously placed in the monotypic Melagraphia) and theAustralian D. concamerata (previously considered aspecies of Austrocochlea). Our genetic data also con-Wrms the arguments of Dell (1971), that the mainlandChilean and New Zealand populations of D. nigerrimaare conspeciWc (see Fig. 1). Moreover, the sole Dilomafound on the Juan Fernandez Islands, D. crusoeana,also groups within this cluster (see Fig. 1). Pilsbry(1889) claimed that there were distinct shell structurediVerences between these last two taxa, but we wereunable to see these diVerences ourselves in the speci-mens we examined (although we did not examine thetype). This Wnding, together with the genetic evidence,suggests all these populations should be referred to asD. nigerrima. The unsampled Japanese species D. nanaand D. piperinus, apparently also belong in Diloma(Higo et al., 1999). We did not sample the Japanese spe-cies D. (Pictodiloma) suavis or the New Caledonianspecies D. constellatus and so cannot comment on theplacement of these species or the validity of the subge-nus Pictodiloma.

4.6. Biogeography

The topology of our phylogeny, with the African spe-cies lying basally to the Australian species, which liebasally to the New Zealand and South American species,is representative of the inverted southern pattern (ISP), apattern incongruent with vicariance and typicallyexplained by dispersal in the west wind drift (Donaldet al., in press; Glasby and Alvarez, 1999; Renner et al.,2000; Swenson and Bremer, 1997). Also, the complicatedparaphyletic distribution in the molecular phylogeny ofthe Australian species, in which the most basal PaciWcclade is comprised solely of Australian species, implies

K.M. Donald et al. / Molecular Phylogenetics and Evolution 37 (2005) 474–483 481

that topshells have dispersed from Australia to otherareas of the PaciWc (both within and outside Australia) onseveral occasions. Owing to the short-lived lecithotrophicnature of topshell larvae (Hickman, 1992), rafting ofadults is a more plausible method of long-distance dis-persal than passive larval drift (Donald et al., in press)and macroalgae has been suggested as a suitable raftingplatform. The predominant natural habitat of D. nigerr-ima (including its synonym D. crusoeana) is decayingbeachcast bull kelp, Durvillaea spp. (Morton and Miller,1968). Long-distance eastward rafting of intertidal marineinvertebrates has been suggested as a plausible means ofdispersal for a variety of species including topshells (Don-ald et al., in press), oysters (Ó Foighil et al., 1999), and seastars (Waters and Roy, 2004).

Nevertheless, the restriction of Oxystele to southernAfrica and Chlorodiloma and Austrocochlea to Australia,as well as the speciose nature of Diloma in New Zealandimplies that much (and probably most) monodontinecladogenesis occurs without long-distance dispersal.Indeed, the current distributions of species in these gen-era show much overlap, especially in New Zealand,where all seven species are found throughout the threemain islands (Powell, 1979).

4.7. Cryptic species

As noted above, genetic analysis has revealed crypticspecies within several monodontine genera. Therelationships of these species to others in their respectivegenera is not the same in all cases, however. In Oxystele,for instance, the outward appearance of the cryptic pairsimply mirrors the genetic data: the morphologicallysimilar O. impervia and O. variegata form a monophy-letic group and are just 1.96% genetically divergent. Incontrast, in Austrocochlea, the conchologically divergentA. rudis is one of the two most derived species in theclade otherwise consisting of the three species once con-sidered ecotype variants of A. constricta. Similarly, theputative cryptic species we identiWed within Monodontalabio are neither sister taxa nor genetically close.

4.8. Summary

Our phylogenetic analysis of the monodontine top-shells reveals several strongly supported monophyleticclades, which we treat as being of generic status. Oxys-tele is endemic to southern Africa, and Austrocochleaand Chlorodiloma are endemic to Australia, whereasMonodonta and Diloma are widespread in the Indo-PaciWc and PaciWc, respectively. Our data implies thatthe widely distributed Indo-PaciWc taxon M. labioharbors cryptic species. In contrast, D. nigerrimaincludes populations previously considered diVerentspecies, from New Zealand to the Juan FernandezIslands and Chile, a Wnding that suggests long-distance

dispersal is an important factor in monodontine evolu-tion. Several recently described cryptic species inOxystele and Austrocochlea are also distinct at theDNA level, but, surprisingly, are not always eachother’s closest relatives.

Acknowledgments

We are grateful to the following people and museumswho collected or sent many of the topshells analysed inthis study: Rudiger Bieler, Yvonne Bone, GeorgeBranch, Bob Burn, Mary Bursey, Liz Campbell, AndrewCummings, Jochen Gerber, Sue Horner, Atsushi Ito,Victor Lobos, G. Philbey, Aroha Miller, Sonja Miller,Margaret Morley, Chris Rowley, Barry Russell, EnricoSchwabe, Shirley Slack-Smith, Yoshitake Takada, CoreyWhisson, Richard Willan, The Field Museum of NaturalHistory, Museum and Art Gallery of the Northern Ter-ritory, Museum of Victoria, Natal Museum, WesternAustralian Museum, Zoologische StaatssammlungMünchen. Studies like this one would be impossiblewithout such international goodwill. Bruce MarshallidentiWed several of the Japanese Monodonta species.This research was supported by the University of OtagoResearch Committee and the Allan Wilson Centre forMolecular Ecology and Evolution.

Appendix A. Revised classiWcation scheme of the monodontine genera Monodonta, Oxystele, Osilinus, Diloma, Austrocochlea, and Chlorodiloma

Species not included in our study are prefaced by an asterisk (*).They are included for completeness and their placement is provisional,reXecting current taxonomy. Only new or previously obscure synony-mies are shown; new synonymies, ranks, and generic placements areunderlined. For detailed synonymies of the New Zealand species, seePowell (1979) and Willan (1979); for Australian species see Cotton(1959), Wilson (1993), and Parsons and Ward (1994). For a summaryof the related European genera, Osilinus and Phorcus see Gofas andJabaud (1997). Approximate ranges are given for each taxon in squarebrackets ([]).

Family Trochidae RaWnesque, 1815Subfamily Trochinae RaWnesque, 1815Tribe Gibbulini Stoliczka, 1868 (D Monodontini Cossmann in Coss-mann & Peyrot, 1916)Genus Oxystele Philippi, 1847Type species: Trochus sinensis Gmelin, 1791Included species: *O. fulgurata (Philippi, 1848) [central West

Africa], O. impervia (Menke, 1843) [South Africa], O. sinensis(Gmelin, 1791) [South Africa], O. tabularis (Krauss, 1848)[South Africa to Mozambique], O. tigrina (Anton, 1838) [SouthAfrica], O. variegata (Anton, 1838) [South Africa]

Genus Chlorodiloma Pilsbry, 1889 (New rank)Type species: Monodonta crinita Philippi, 1849Included species: C. adelaidae (Philippi, 1849) (New combination)

[southern Australia], C. crinita (Philippi, 1849) (New combina-tion) [Western Australia], *C. millelineata (Bonnet, 1864) (New

482 K.M. Donald et al. / Molecular Phylogenetics and Evolution 37 (2005) 474–483

combination) [Torres Strait], C. odontis (New combination)(Wood, 1828) [southern Australia]

Genus Austrocochlea P. Fischer, 1885Type species: Monodonta constricta Lamarck, 1822 (ICZN Opin-

ion, 1930)Included species: A. brevis Parsons and Ward, 1994 [Tasmania], A.

constricta (Lamarck, 1822) [southern Australia], A. diminuta(Hedley, 1912) [northern Australia], A. porcata (A. Adams,1853) [southern Australia], A. rudis (Gray, 1826) [southern Aus-tralia], *A. zeus P. Fischer, 1874 [Western Australia]

Genus Monodonta Lamarck, 1799Type species: Trochus labio Linné, 1758Included species: *M. australis Lamarck, 1822 [western Indian

Ocean], M. canalifera Lamarck, 1822 [Thailand to Japan], M.confusa Tapparone-Canefri, 1874 (New rank) [Singapore toJapan], *M. gibbula Thiele, 1925 [South Africa], *M. glabrataGould, 1861 [South China and East China Seas], M. labio(Linné, 1758) [western PaciWc Ocean, northern Australia toJapan], *M. nebulosa (Forsskål in Niebuhr, 1775) (D dama Phi-lippi, 1848) [Indian Ocean], *M. vermiculata (P. Fischer, 1874)[western Indian Ocean], *M. viridis Lamarck, 1816 [Japan]

Subgenus Neomonodonta Kuroda & Habe, 1971Type species: Trochus neritoides Philippi, 1849Included species: *M. neritoides (Philippi, 1849) [Japan to southern

China], M. perplexa perplexa Pilsbry, 1889 [Japan to southernChina], *M. perplexa boninensis Asakura & Nishimura, 1987[Ogasawara Islands]

Genus Diloma Philippi, 1845Type species: Turbo nigerrimus Gmelin, 1791Synonyms: Melagraphia Gray, 1847 (Type Turbo aethiops Gme-

lin, 1791) (New synonymy), Neodiloma P. Fischer, 1885 (TypeTurbo aethiops Gmelin, 1791) (New synonymy), AnisodilomaFinlay, 1926 (Type Trochus lugubris Gmelin, 1791) (New syn-onymy), Cavodiloma Finlay, 1926 (Type Trochocochleaexcavata Adams & Angas, 1864) (New synonymy), Fractarm-illa Finlay, 1926 (Type Labio corrosa A. Adams, 1853) (Newsynonymy), Zediloma Finlay, 1926 (Type Z. digna Finlay,1926)

Included species: D. aethiops (Gmelin, 1791) (New combination)(D lugubris Gmelin, 1791, not of subsequent authors, see Willanet al., 1980) [New Zealand], D. arida (Finlay, 1926) [New Zea-land], D. bicanaliculata (Dunker in Philippi, 1844) (D leniorFinlay, 1926) [New Zealand], D. concamerata (Wood, 1828)(New combination) [southern Australia], *D. constellatus (Sou-verbie in Souverbie & Montrouzier, 1863) [New Caledonia],D. coracina (Philippi, 1851) (D excavata Adams & Angas, 1864)[New Zealand], *D. nana Gould, 1861 [Okinawa], D. nigerrima(Gmelin, 1791) (D digna Finlay, 1926 D crusoeana Pilsbry, 1889(New synonymy)) [New Zealand, western South America], *D.piperinus (Philippi, 1849) [tropical PaciWc Ocean to Japan], D.radula (Philippi, 1849) [tropical PaciWc, Philippines, Japan], D.samoaensis Schwabe and Barclay, 2003 [Samoa], D. subrostrata(Gray, 1835) (D corrosa A. Adams, 1853 D novazelandiaeAnton, 1839 (New synonymy)) [New Zealand], D. zelandica(Quoy and Gaimard, 1834) [New Zealand]

Subgenus Pictodiloma Habe, 1946Type species: Trochus suavis Philippi, 1849Included species: *D. suavis (Philippi, 1849) [Japan]Genus Osilinus Philippi, 1847 (D Trochocochlea Mörch, 1852)Type species: Trochus turbinatus Born, 1778 (ICZN Opinion, 1930)Included species: *O. articulatus (Lamarck, 1822) [Mediterra-

nean], *O. atratus (Wood, 1828) [Portugal to West Africa], O.lineatus (da Costa, 1778) [Britain to Morocco], *O. kotschyi(Philippi, 1849) [Indian Ocean], *O. punctulatus (Lamarck,1822) [West Africa], *O. sauciatus (Koch, 1845) (D colubrinusGould, 1849) [West Africa], *O. turbinatus (Born, 1778) [Med-iterranean]

Excluded speciesMany species were originally described or subsequently placedin Monodonta. This list is not exhaustive.Monodonta angulifera A. Adams, 1853 is the type of the chil-odontine genus Perrinia H. & A. Adams, 1854 (Herbert, 1996).M. clathrata A. Adams, 1853 and M. foveolata A. Adams, 1853,as well as M. exasperata A. Adams, 1853 and M. Wscheri Mon-trouzier in Souverbie & Montrouzier, 1866 are chilodontines,probably species of Vaceuchelus Iredale, 1929 and HerpetopomaPilsbry, 1889, respectively (Herbert, 1996).M. (Austrocochlea) quadrasi G.B. Sowerby III, 1889 is a syno-nym of Thaliota tricingulata Adams, 1853 (Herbert, 1994)M. obscura (Wood, 1828) is the type species of PriotrochusFischer, 1879.M. richardi Payraudeau, 1826 and Osilinus mutabilis (Philippi,1846) (also referred to as M. mutabilis) are species of PhorcusRisso, 1826 (Gofas and Jabaud, 1997).

References

Adema, C.M., 2002. Comparative study of cytoplasmic actin DNAsequences from six species of Planorbidae (Gastropoda: Basom-matophora). J. Moll. Stud. 68, 17–23.

Altschul, S.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J., 1990.Basic local alignment search tool. J. Mol. Biol. 215, 403–410.

Cotton, B.C., 1959. South Australian Molluscs—Archaeogastropoda.Government Printer, Adelaide.

Dell, R.K., 1971. The marine Mollusca of the Royal Society Expeditionto Southern Chile, 1958–1959. Rec. Dom. Mus. 7, 155–233.

Donald, K.M., Kennedy, M., Spencer H.G., in press. Cladogenesis asthe result of long-distance rafting events in South PaciWc topshells(Gastropoda, Trochidae). Evolution.

Farris, J.S., Källersjö, M., Kluge, A.G., Bult, C., 1995. Constructing asigniWcance test for incongruence. Syst. Biol. 44, 570–572.

Felsenstein, J., 1985. ConWdence limits on phylogenies: an approachusing the bootstrap. Evolution 39, 783–791.

Felsenstein, J., 1988. Phylogenies from molecular sequences: inferenceand reliability. Annu. Rev. Genet. 22, 521–565.

Folmer, O., Black, M., Hoeh, W., Lutz, R., Vrijenhoek, R., 1994. DNAprimers for ampliWcation of mitochondrial cytochrome c oxidasesubunit I from diverse metazoan invertebrates. Mol. Mar. Biol. Bio-tech. 3, 294–299.

Glasby, C.J., Alvarez, B., 1999. Distribution patterns and biogeographicanalysis of austral Polychaeta (Annelida). J. Biogeog. 26, 507–533.

Gofas, S., Jabaud, A., 1997. The relationships of the Mediterraneantrochid gastropods ‘Monodonta’ mutabilis (Philippi, 1846) and ‘Gib-bula’ richardi (Payraudeau, 1826). J. Moll. Stud. 63, 57–64.

Hafner, M.S., Sudman, P.D., Villablance, F.X., Spradling, T.A., Demas-tes, J.W., Nadler, S.A., 1994. Disparate rates of molecular evolution incospeciating hosts and parasites. Science 365, 1087–1090.

Heller, J., Dempster, Y., 1991. Detection of 2 coexisting species ofOxystele (Gastropoda, Trochidae) by morphological and electro-phoretic analysis. J. Zool. 223, 395–418.

Herbert, D., 1994. Notes on synonymy within the genus Priotrochus Fischer,1879 (Prosobranchia: Trochidae). Ann. Natal. Mus. 35, 139–151.

Herbert, D., 1996. A critical review of the trochoidean types in the Mus-éum d’Histoire naturelle, Bordeaux (Mollusca, Gastropoda). Bull.Muséum national d’Histoire naturelle, 4e série, Section A, 18, 409–445.

Hickman, C.S., 1992. Reproduction and development of trochaceangastropods. Veliger 35, 245–272.

Hickman, C.S., 1998. Superfamily Trochoidea. In: Beesley, P.L., Ross,G.J.B., Wells, A. (Eds.), Mollusca: The Southern Synthesis. CSIRO,Melbourne, pp. 671–692.

Hickman, C.S., McLean, J.H., 1990. Systematic revision and supragen-eric classiWcation of trochacean gastropods. Science Series, NaturalHistory Museum of Los Angeles County 35, 1–169.

K.M. Donald et al. / Molecular Phylogenetics and Evolution 37 (2005) 474–483 483

Higo, S., Callomon, P., Goto, Y., 1999. Catalogue and Bibliography ofthe Marine Shell-bearing Mollusca of Japan. Elle ScientiWc Publica-tions, Japan.

Huelsenbeck, J.P., Rannala, B., 2004. Frequentist properties of Bayes-ian posterior probabilities of phylogenetic trees under simple andcomplex substitution models. Syst. Biol. 53, 904–913.

Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inferenceof phylogenetic trees. Bioinformatics 17, 754–755.

International Commission on Zoological Nomenclature. Opinion1930. Osilinus Philippi, 1847 and Austrocochlea Fischer, 1885 (Mol-lusca, Gastropoda): conserved by the designation of Trochus tur-binatus Born, 1778 as the type species of Osilinus. Bull. Zool.Nomen. 56, 202–203.

Kim, J.H., 1993. Improving the accuracy of phylogenetic estimation bycombining diVerent methods. Syst. Biol. 42, 331–340.

Liddell, J., 1950. An examination of the genus Austrocochlea Fischer,M.Sc. Thesis, University of Sydney, Sydney.

Macpherson, J.H., 1961. Trochus obtusa confusion. Memoirs of theNational Museum of Victoria 27, 201–263.

Morton, J.E., Miller, M., 1968. The New Zealand Sea Shore. Collins,London.

Ó Foighil, D., Marshall, B.A., Hilbish, T.J., Pino, M.A., 1999. Trans-PaciWc range extension by rafting in inferred for the Xat oysterOstrea chilensis. Biol. Bull. 196, 122–126.

Parsons, K.E., 1996. Discordant patterns of morphological and geneticdivergence in the ‘Austrocochlea constricta’ (Gastropoda: Trochi-dae) species complex. Mar. Freshwater Res. 47, 981–990.

Parsons, K.E., Ward, R.D., 1994. Electrophoretic and morphologicalexamination of Austrocochlea constricta (Gastropoda: Trochidae):A species complex. Aust. J. Mar. Freshwater Res. 45, 1065–1085.

Pilsbry, H.A., 1889. Manual of Conchology. Vol. XI. Trochidae, Stom-atiidae, Pleurotomariidae, Haliotidae. Academy of Natural Sci-ences, Philadelphia.

Posada, D., Crandall, K.A., 1998. Modeltest: Testing the model ofDNA substitution. Bioinformatics 14, 817–818.

Powell, A.W.B., 1979. New Zealand Mollusca: Marine, Land andFreshwater Shells. Collins, Auckland.

Renner, S.S., Foreman, D.B., Murray, D., 2000. Timing transantarcticdisjunctions in the Atherospermataceae (Laurales): evidence fromcoding and noncoding chloroplast sequences. Syst. Biol. 49, 579–591.

Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phyloge-netic inference under mixed models. Bioinformatics 19, 1572–1574.

Schwabe, E., Barclay, D., 2003. A new species of Diloma Philippi, 1845(Trochidae: Trochinae: Gibbulini) from the Samoa Islands (Mol-lusca: Gastropoda). Novapex 4, 29–32.

Simmons, M.P., Pickett, K.M., 2004. How meaningful are Bayesiansupport values?. Mol. Biol. Evol. 21, 188–199.

Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H., Flook, P., 1994.Evolution, weighting and phylogenetic utility of mitochondrialgene sequences and a compilation of conserved polymerase chainreaction primers. Ann. Entomol. Soc. Am. 87, 651–701.

Spencer, H.G., Willan, R.C., 1996. The Marine Fauna of New Zealand:Index to the Fauna 3. Mollusca. New Zealand Oceanographic Insti-tute, Wellington.

Swenson, U., Bremer, K., 1997. PaciWc biogeography of the Asteraceaegenus Abrotanella (Senecioneae, Blennospermatinae). Syst. Bot. 22,493–508.

SwoVord, D.L., 2002. PAUP* 4.0b10, Sinauer Associates, Sunderland,MA.

Takenouchi, K., 1985. A boulder shore gastropod fauna in HongKong. In: Morton, B., Dudgeon, D. (Eds.), The Malacofauna ofHong Kong and Southern China II, Vol. 2. Hong Kong UniversityPress, pp. 413–419.

Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins,D.G., 1997. The ClustalX windows interface: Xexible strategies formultiple sequence alignment aided by quality analysis tools.Nucleic Acids Res. 24, 4876–4882.

Walsh, P.S., Metzger, D.A., Higuchi, R., 1991. Chelex 100 as a mediumfor simple extraction of DNA for PCR-based typing from forensicmaterial. Biotechniques 10, 506–513.

Waters, J.M., Roy, M.S., 2004. Out of Africa: the slow train to Austral-asia. Syst. Biol. 53, 18–24.

Willan, R.C., 1979. A key to the intertidal trochid molluscs of thegenera Diloma and Melagraphia in New Zealand. Tane 25, 219–224.

Willan, R.C., Marshall, B.A., Climo, F.M., Cernohorsky, W.O., 1980.RectiWcation of nomenclature for Melagraphia aethiops (Gmelin)and Diloma bicanaliculata (Dunker) (Mollusca: Trochidae). N.Z.J.Mar. Freshwater Res. 14, 413–415.

Wilson, B., 1993. Australian Marine Shells: Part 1 Prosobranch Gas-tropods. Odyssey, Kallaroo, Western Australia.