Journal of Biogeography (J. Biogeogr.) (2013) 40, 594–607

ORIGINAL Historical biogeography and phylogeny ARTICLE of cave (: ) based on mitochondrial and nuclear data Alejandro Botello1,2, Thomas M. Iliffe3, Fernando Alvarez2, Carlos Juan4*, Joan Pons1 and Damia` Jaume1

1IMEDEA (CSIC-UIB), Mediterranean ABSTRACT Institute for Advanced Studies, Esporles, Aim Our aim was to produce a dated phylogeny of Typhlatya, a stygobiont Balearic Islands, Spain, 2National Collection with an extremely disjunct localized distribution across the of , Institute of Biology, National Autonomous University of Mexico, Mexico, Mediterranean, the central Atlantic and eastern Pacific. Using phylogenetic 3Department of Marine Biology, Texas A&M analyses, we examine the role of dispersal and plate tectonics in determining its University at Galveston, TX, USA, distribution. 4 Department of Biology, University of the Location Western Mediterranean, Ascension Island, Bermuda, Bahamas, Balearic Islands, Palma, Balearic Islands, Yucata´n, Caribbean, Gala´pagos, Western Australia. Spain Methods Thirteen of the 17 species of Typhlatya were analysed, using Stygioc- aris, and Antecaridina as outgroups. Fragments of three mito- chondrial and three nuclear genes were combined into a data set of 2449 mitochondrial and 1374 nuclear base pairs.

Results Phylogenetic trees clearly showed Typhlatya to be paraphyletic, with the Gala´pagos species clustering with Antecaridina. Only the phylogenetic posi- tion of (Hispaniola and Puerto Rico) showed some uncer- tainty, appearing as the sister group to the Australian genus on the most likely topology. We estimated an average age of 45 Myr (30.6–61.1 Myr) for the most recent common ancestor of Typhlatya + Stygiocaris + Antecaridina + Halocaridina. All Typhlatya (except Typhlatya galapagensis) + Stygiocaris derived from a node dated to 35.7 Ma (25.7–47.0 Ma), whereas the ancestor of all Typhlatya species (excluding T. monae and T. galapagensis) lived 30.7 Ma (21.9–40.4 Ma). Main conclusions Typhlatya is paraphyletic and apparently absent from the eastern Pacific, with T. galapagensis clustering with Antecaridina. The remain- ing Typhlatya species form a robust monophyletic group with Stygiocaris, and both molecular and morphological evidence support the recognition of three sublineages: (1) Typhlatya s. str., Atlantic–Mediterranean, embracing all Typhlatya species minus T. monae; (2) Stygiocaris, limited to north-western Australia; and (3) T. monae (Caribbean), for which a new genus could be erected. No congruence was found between temporal and geographical projec- tions of cladogenetic events within Typhlatya/Stygiocaris and the major plate

*Correspondence: Carlos Juan, Department of tectonic events underlying Tethyan history. Biology, University of the Balearic Islands, cta. Keywords Valldemossa km 7.5, E-07122 Palma (Balearic COI b Islands), Spain. 16S rRNA, 18S rRNA, 28S rRNA, , cyt , histone H3A, molecular clock, E-mail: [email protected] stygofauna, Tethyan relicts.

594 http://wileyonlinelibrary.com/journal/jbi ª 2012 Blackwell Publishing Ltd doi:10.1111/jbi.12020 Zoogeography of Typhlatya cave shrimps

tions should have an age that at least precedes the establish- INTRODUCTION ment of deep-water conditions in the north-central Atlantic Many aquatic subterranean crustaceans (stygobionts) exhibit Ocean. They must also have persisted since then in a state broad transoceanic disjunct distributions throughout tropical of morphological stasis, or alternatively converged morpho- and subtropical latitudes, so that different congeneric species logically under the shared selection pressures posed by the may be isolated on continents or islands half the world subterranean habitat (Barr & Holsinger, 1985; Hart & Man- apart (Stock, 1993). This pattern is repeated in a diverse set ning, 1986). of taxonomic groups including the remipedes, thermosbae- The major drawback to the hypothesized Tethyan origin naceans, amphipods, isopods, decapods, copepods and os- for these taxa and their vicariance by plate tectonics is their tracods, and it has been explained by the fragmentation of frequent occurrence on relatively young oceanic islands that the continuous ranges of their ancestors by a series of have never been connected to continental shelves. Hart et al. shared isolation events (Stock, 1993; Wagner, 1994). These (1985) proposed that representatives of these lineages on so-called ‘Tethyan’ distribution patterns are best explained Atlantic islands might be survivors from the time when the in terms of the vicariant isolation of the ancestral lineages Atlantic Ocean was very narrow, and that the forerunners of coincident with the fragmentation in the late Mesozoic and these islands were in contact with or close to both shores of Tertiary of the Tethys Sea, a predominantly shallow-water the ocean. Deep-sea dispersal along the crevicular medium circumtropical ocean that existed from the Middle Jurassic associated with the circumglobal system of spreading zones until 20 million years ago (Ma) (Sterrer, 1973; Stock, 1993). also represents a feasible alternative explanation of the pres- The progressive break-up of this east–west palaeoseaway ence of some of these taxa on geologically young oceanic with the collision of continental landmasses and the forma- islands (Boxshall, 1989). tion of broad, deep oceanic basins could have resulted in Typhlatya Creaser, 1936 is a stygobiont genus of atyid the allopatric diversification of the ancestors of present spe- shrimp with a punctuated distribution throughout coastal cies, which subsequently became stranded in inland aquifers. continental and insular ground-waters of the Mediterranean, It follows, therefore, that genera displaying such distribu- north-central Atlantic and east Pacific (Fig. 1). This taxon,

Figure 1 Global distributions of Typhlatya and Stygiocaris. The inset shows the distribution of Typhlatya in the Caribbean region. See Table 1 for the precise distribution of each taxon. Typhlatya mitchelli and T. pearsei are broadly distributed throughout the northern Yucata´n Peninsula, and are shown schematically.

Journal of Biogeography 40, 594–607 595 ª 2012 Blackwell Publishing Ltd A. Botello et al. which has never been reported in open marine habitats but MATERIALS AND METHODS with most of the species inhabiting anchialine waters, has an uncertain biogeographical history (see, for example, Croizat Material examined et al., 1974; Monod, 1975; Rosen, 1975; Buden & Felder, 1977; Iliffe et al., 1983; Hart et al., 1985; Manning et al., The 17 species of Typhlatya currently recognized are found 1986; Stock, 1993; Sanz & Platvoet, 1995 for proposals based on eastern Pacific and mid-Atlantic islands, the coasts of the on vicariance; and Chace & Hobbs, 1969; Monod & Cals, Caribbean, the Antillean Arch, the Bahamas and the western 1970; Chace & Manning, 1972; Peck, 1974; Iliffe, 1986; Mediterranean (Fig. 1, Table 1); 13 were included in the Stock, 1986; Banarescu, 1990 for alternative dispersalist analysis. A single population of each species was analysed, explanations). The broad distribution of Typhlatya has been except for Typhlatya galapagensis, Typhlatya consobrina, described elsewhere as the result of Tethys fragmentation Typhlatya miravetensis and Typhlatya monae (see Table 2). (Buden & Felder, 1977; Stock, 1993). However, the ability of The two species of Stygiocaris Holthuis, 1960; plus Hal- some members of the (typically freshwater) family Atyidae to ocaridina rubra and Antecaridina lauensis,wereincludedin undertake part of their life cycle in the marine environment the data set because of their demonstrated relationship to (diadromy) and the presence of members of Typhlatya and Typhlatya (Monod & Cals, 1970; Page et al., 2008). Other other closely related genera on young oceanic islands also analyses using only 16S rRNA (rrnL), 28S rRNA (LSU) support explanations based on marine dispersal (Smith & and histone H3A sequences were performed in conjunction Williams, 1981; Russ et al., 2010). Recently, divergent phylo- with GenBank sequences from the closest relatives of the geographical patterns among anchialine shrimp have been Typhlatya/Stygiocaris cluster (von Rintelen et al., 2012). related to differences in the duration of their respective planktonic larval (dispersive) phases (Santos, 2006; Craft Sequences and alignments et al., 2008; Russ et al., 2010). Several molecular phylogenetic analyses have dealt with Genomic DNA was isolated from whole specimens using the Typhlatya species, but none has addressed the phylogeny and DNeasy Tissue Kit (Qiagen, Hilden, Germany). Polymerase biogeography of the genus as a whole. Thus, Hunter et al. chain reaction (PCR) was used to amplify fragments of the (2008) investigated the phylogeography of three species mitochondrial cytochrome c oxidase subunit I (COI; two from Yucata´n, the Caicos Islands and Bermuda. Zaksˇek et al. non-overlapping fragments), cytochrome b (cyt b), and rrnL (2007) analysed the molecular phylogeny of the stygobiont genes using the primers shown in Table 3. Fragments of genus Dormitzer, 1853, using species of Typhlatya three other nuclear genes were also amplified: histone H3A, from Spain and Yucata´n as outgroups. Page et al. (2008) 18S rRNA (SSU) and LSU (Table 3). The combined data set found that Typhlatya pearsei (Yucata´n) – the only Typhlatya consisted of 3823 bp (2449 bp of the mitochondrial and included in their analysis – was recovered as the closest rela- 1374 bp of the nuclear genome). tive of the endemic Western Australian subterranean genus PCR was performed in a reaction containing (19)NH4

Stygiocaris, and suggested that they may have descended from buffer, 3.5–5.0 mm MgCl2,0.2mm of each dNTP, 0.2–0.4 lm a common ancestor that lived in the coastal marine habitat each primer, 0.5 U of Taq DNA polymerase and 1–5 lLof of the ancient Tethys Sea and were subsequently separated DNA template, in a final volume of 25 lL. The amplification by tectonic plate movements. Five Typhlatya species were conditions consisted of one cycle of 94 °C for 2 min and 35 also included in a recent molecular phylogeny of the family cycles of 94 °C for 30 s, 47–55 °C for 30 s and 72 °Cfor Atyidae (von Rintelen et al., 2012). Using a relaxed molecu- 1 min, followed by a final extension step at 72 °C for 10 min. lar clock and different calibration priors, these authors esti- The amplified fragments were sequenced in both directions mated an age range from Early Cretaceous to Palaeogene using the ABI Prism BigDye Reaction Kit v. 2.0 and an ABI for what they defined informally as the ‘Typhlatya group’ 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA, (Antecaridina, Halocaridina, Halocaridinides, Stygiocaris and USA). Nucleotide sequences were aligned using mafft 4.0 soft- Typhlatya). ware, taking into account the RNA secondary structure of ribo- Here, we present the first genetic survey undertaken to somal genes (Katoh et al., 2005). reconstruct the phylogeny of Typhlatya, based on 4 kb of nuclear and mitochondrial sequences and a geographically Phylogenetic analyses representative sample. Our aim was to use phylogenetic anal- yses to test the roles of dispersal and of plate tectonics in The program jModelTest (Posada, 2008) was used to select generating the distribution of Typhlatya. With transoceanic the best evolutionary model for each partition, according to dispersal, we would expect disparate estimates for divergence the Bayesian information criterion (BIC). The best model times, inconsistent with those of Tethys’ fragmentation. In was HKY+Γ, except for the SSU+LSU partition, for which contrast, large divergences, preceding the establishment of GTR+Γ was the best. Incongruence length difference (ILD) deep-water conditions in the Atlantic Ocean, would be antic- tests (Farris et al., 1995) were performed with paup* 4.0b10 ipated if the distribution pattern of Typhlatya was better (Swofford, 2002) to check for incongruence among genes. explained by ancient vicariance. We implemented different evolutionary models, data

596 Journal of Biogeography 40, 594–607 ª 2012 Blackwell Publishing Ltd Zoogeography of Typhlatya cave shrimps

Table 1 Typhlatya diversity and distribution. Asterisks denote species retaining pigmented eyespots.

Species Distribution References

T. arfeae Jaume & Bre´hier, 2005 France Jaume & Bre´hier (2005) T. campechae Hobbs & Hobbs, 1976 Yucata´n Peninsula (Mexico) Hobbs & Hobbs (1976) T. consobrina Botosaneanu & Holthuis, 1970 Cuba Botosaneanu & Holthuis (1970) T. dzilamensis Alvarez et al., 2005 Yucata´n Peninsula (Mexico) Alvarez et al. (2005) T. elenae Juarrero, 1994 Cuba Juarrero (1994) T. galapagensis Monod & Cals, 1970 Santa Cruz and Isabela Monod & Cals (1970) Islands (Gala´pagos) T. garciadebrasi Juarrero & Ortiz, 2000 Cuba Juarrero & Ortiz (2000) T. garciai Chace, 1942 Cuba; Caicos Islands Botosaneanu & Holthuis (1970); Buden & Felder (1977); Chace (1942); Holthuis (1977) * T. iliffei Hart & Manning, 1981 Bermuda Hart & Manning (1981) * T. kakuki Alvarez et al., 2005 Acklins (Bahamas) Alvarez et al. (2005) T. miravetensis Sanz & Platvoet, 1995 Spain Sanz & Platvoet (1995) T. mitchelli Hobbs & Hobbs, 1976 Yucata´n Peninsula (Mexico) Hobbs & Hobbs (1976) * T. monae Chace, 1954 Puerto Rico; Dominican Republic; Mona Island Chace (1954, 1975); Debrot (Puerto Rico); Barbuda (Lesser Antilles); (2003); Sket (1988) Curac¸ao (Netherlands Antilles); San Andre´s Island (Colombia) T. pearsei Creaser, 1936 Yucata´n Peninsula Ca´rdenas (1950); Creaser (Mexico) (1936, 1938); Hobbs & Hobbs (1976) * T. rogersi Chace & Manning, 1972 Ascension Island Chace & Manning (1972) T. taina Estrada & Go´mez, 1987 Cuba Estrada & Go´mez (1987) * T. utilaensis Alvarez et al., 2005 Utila Island (Honduras) Alvarez et al. (2005) Typhlatya sp. Belize T. Iliffe, pers. obs. Typhlatya sp. Aruba (Netherlands Antilles) L. Botosaneanu, Zoo¨logisch Museum, Amsterdam, pers. comm. Typhlatya sp. Bonaire (Netherlands Antilles) L. Botosaneanu, pers. comm.

partitioning strategies, tree construction methods and clock which ran for two million generations, sampled at intervals estimation methods to assess their effect on tree topologies, of 1000 generations. All parameters were unlinked and rates branch lengths and evolutionary rates (Phillips, 2009). We were allowed to vary freely over partitions. The convergence explored five different partitioning strategies: (1) seven parti- of all parameters of the two independent runs was assessed tions: considering first, second and third codon positions of in MrBayes 3.1.2 and Tracer 1.5, obtaining effective sam- mitochondrial DNA (mtDNA) as three different partitions, ple sizes of more than 200 (Rambaut & Drummond, 2009). plus rrnL, histone H3A, SSU and LSU as individual parti- After the 10% burn-in samples, the remaining trees from the tions; (2) six partitions: as above, but combining the first two independent runs were combined into a single majority and second mtDNA positions into a single partition; (3) five consensus topology, and the frequencies of the nodes in the partitions – as in (2), but with the nuclear ribosomal genes majority rule tree were taken as the posterior probabilities merged into a single partition; (4) each gene as an indepen- (Huelsenbeck & Ronquist, 2001). dent partition (six partitions); and (5) the mitochondrial Maximum likelihood (ML) analyses using the partition and nuclear sequences treated as two different partitions. schemes described above were performed using RAxML 7.0.4 The competing partition strategies were compared using (Stamatakis et al., 2005). Bootstrap support values were esti- Bayes factors (Brown & Lemmon, 2007). Marginal likeli- mated using the fast bootstrapping method, with 500 repli- hoods and harmonic means were estimated using Tracer cates. 1.5 (Rambaut & Drummond, 2009). The best partition scheme among the five tested was option (3). Molecular clock analyses Bayesian phylogenetic analyses were conducted in the par- allel version of MrBayes 3.1.2 (Huelsenbeck & Ronquist, We estimated node ages using beast 1.6.0 (Drummond & 2001). In each Bayesian search, two independent runs were Rambaut, 2007), enforcing a relaxed molecular clock with performed, starting with the default prior values, random an uncorrelated lognormal distribution and a Yule specia- trees and four Markov chains, three heated and one cold, tion model. For tree calibration, we used the known age

Journal of Biogeography 40, 594–607 597 ª 2012 Blackwell Publishing Ltd 598 Botello A.

Table 2 Collection sites and EMBL accession numbers of Antecaridina, Halocaridina, Stygiocaris and Typhlatya species included in this analysis. Accession numbers in bold correspond to sequences obtained from GenBank (Ivey & Santos, 2007; Page et al., 2008). al. et

EMBL accession numbers

Species Collection site COI 5′ COI 3′ cyt b rrnL (16S) Histone H3A SSU (18S) LSU (28S)

Antecaridina lauensis North East Point, Christmas Island (SE Indian Ocean) HE800898 HE800919 n.a. EU123851 HE800965 HE801016 n.a. Kohala District (Hawaii Is.; Hawaii): anchialine pool PT DQ917432 DQ917432 DQ917432 FN995368 HE800964 HE801015 HE801036 (Santos, 2006) Stygiocaris lancifera Cape Range (W Australia): Tulki well HE800901 HE800922 HE800948 EU123827 HE800968 HE801019 HE801039 Stygiocaris stylifera Cape Range (W Australia): Kuddamurra well (Palms) n.a. HE800923 n.a. EU123836 HE800969 HE801020 HE801040 Typhlatya arfeae Salses (Perpignan; France): Font Estramar HE800906 HE800929 HE800954 HE801000 HE800975 HE801025 HE801045 Typhlatya consobrina Bolondro´n (Matanzas; Cuba): Cueva Chicharrones HE800910 HE800933 HE800956 HE801004 HE800979 HE801028 HE801048 El Veral (Guanahacabibes Peninsula; W Cuba): Cueva del Agua HE800915 HE800940 HE800962 HE801011 HE800986 HE801034 n.a. Typhlatya dzilamensis Dzilam de Bravo (Yucata´n; Mexico): Cenote Cervera n.a. HE800926 HE800951 HE800997 HE800972 n.a. n.a. Typhlatya galapagensis Sta Cruz Is. (Gala´pagos) HE800899 HE800920 HE800946 HE800991 HE800966 HE801017 HE801037 Isabela Is. (Gala´pagos) HE800900 HE800921 HE800947 HE800992 HE800967 HE801018 HE801038 Typhlatya garciai Providenciales (Caicos) HE800909 HE800932 HE800955 HE801003 HE800978 n.a. n.a. Typhlatya iliffei Bermuda: Tucker’s Town Cave HE800904 HE800927 HE800952 HE800998 HE800973 HE801023 HE801043 Typhlatya kakuki Salinas Point (Acklins Is.; Bahamas): Shrimp Hole n.a. HE800941 n.a. HE801013 HE800988 n.a. n.a. Typhlatya miravetensis Pla de Cabanes (Castello´n; Spain): Ullal de la Rambla de Miravet HE800905 HE800928 HE800953 HE800999 HE800974 HE801024 HE801044 Well at Pen˜ı´scola (Castello´n; Spain) HE800916 n.a. HE800963 HE801012 HE800987 HE801035 n.a. Well at Alcala´ de Xivert (Castello´n; Spain) HE800917 n.a. n.a. n.a. n.a. n.a. n.a. Typhlatya mitchelli Hoctu´ n (Yucata´n; Mexico): Cenote de Hoctu´ n HE800902 HE800924 HE800949 HE800995 HE800970 HE801021 HE801041 Typhlatya monae Well at Juan Dolio (Dominican Rep.) HE800907 HE800930 n.a. HE801001 HE800976 HE801026 HE801046 Jaragua NP (Oviedo; Pedernales; Dominican Rep.): Pozima´n Cadena n.a. HE800935 n.a. HE801006 HE800981 n.a. HE801050 Cave at Bosque Gua´nica (SW Puerto Rico) HE800912 HE800936 HE800958 HE801007 HE800982 HE801030 n.a. Cave at Bosque Gua´nica HE800913 HE800937 HE800959 HE801008 HE800983 HE801031 n.a. Cave at Bosque Gua´nica HE800914 HE800938 HE800960 HE801009 HE800984 HE801032 n.a. Cave at Bosque Gua´nica n.a. HE800939 HE800961 HE801010 HE800985 HE801033 n.a.

ora fBiogeography of Journal Samana´ Peninsula (Dominican Rep.): well at Playa del Fronto´n n.a. HE800942 n.a. HE801014 n.a. n.a. n.a. Well at Juan Dolio (Dominican Rep.) n.a. HE800943 n.a. n.a. n.a. n.a. n.a.

ª Well at Juan Dolio n.a. HE800944 n.a. n.a. n.a. n.a. n.a. 02BakelPbihn Ltd Publishing Blackwell 2012 Typhlatya pearsei Sacalum (Yucata´n; Mexico): Cenote Nohchen HE800903 HE800925 HE800950 HE800996 HE800971 HE801022 HE801042 Typhlatya rogersi Anchialine pool at Ascension Is. n.a. HE800931 n.a. HE801002 HE800977 HE801027 HE801047 Typhlatya taina Puerto Escondido (Sta Cruz del Norte; La Habana; Cuba): HE800908 HE800934 HE800957 HE801005 HE800980 HE801029 HE801049 Cueva de la India

n.a., not applicable. 40 594–607 , Zoogeography of Typhlatya cave shrimps

Table 3 Primers used to amplify different mitochondrial and nuclear fragments.

Amplified fragment Primer Primer sequence (5′–3′) Reference

Mitochondrial COI LCO1490 GGTCAACAAATCATAAAGATATTGG Folmer et al. (1994) HCO2198 TAAACTTCAGGGTGACCAAAAAATCA Folmer et al. (1994) NANCY CCYGGTAAAATTAAAATATAAATCTC Simon et al. (1994) Pat TCCAATGCACTAATCTGCCATATTA Simon et al. (1994) Jerry CAACATTTATTTTGATTTTTTGG Simon et al. (1994) COIF1 AAAAAAGAAACMTTYGGYACNYTAGG This study COIR1 TTNARDCCTARGAARTGYTGRGG This study F12 GCCTTCCCCCGGATRAAYAAYAT This study R27 CGGTCGGTCAGCAGYATNGTRATNGC This study cyt b CB1 TATGTACTACCATGAGGACAAATATC Barraclough et al. (1999) CB4 AAAAGAAARTATCATTCAGGTTGAAT Barraclough et al. (1999) rrnL (16S) M14 CGCCTCTTTATCAAAAACAT Xiong & Kocher (1991) M74 CTCCGGTTTGAACTCAGATCA Xiong & Kocher (1991) Nuclear Histone H3A H3aF ATGGCTCGTACCAAGCAGACVGC Colgan et al. (1998) H3Ar ATATCCTTRGGCATRATRGTGAC Colgan et al. (1998) SSU (18S rRNA) 18S3′ CACCTACGGAAACCTTGTTACGAC Shull et al. (2001) 18S2.0 ATGGTTGCAAAGCTGAAAC Shull et al. (2001) LSU (28S rRNA) Ver28Sf CAAGTACCGTGAGGGAAAGTT Lefe´bure et al. (2006) Ver28S2 GTTCACCATCTTTCGGGTC Lefe´bure et al. (2006)

ranges of three major events affecting the diversification of sequences used in this study are shown in Table 2. Note that particular lineages as flat priors: (1) the isolation of the it was not possible to recover the entire sequences of some populations of T. galapagensis from Santa Cruz and Isabela gene fragments for some populations. islands in the Gala´pagos, which cannot be older than the age of the Cocos Ridge and associated seamounts. These Intraspecific divergences now-submerged structures probably formed when the oce- anic crust moved over the Gala´pagos hotspot, and it is Typhlatya galapagensis from Santa Cruz and Isabela islands probable that an archipelago has existed continuously above showed a considerably higher pairwise COI genetic distance the current Gala´pagos area for the past 14.5 Myr (see (8%) than those found among populations of T. monae or Werner et al., 1999, and references therein), so the interval T. miravetensis, or between Typhlatya garciai and Typhlatya 5–14 Ma has been proposed for the separation of the two kakuki (see Appendix S1 in Supporting Information). This populations; (2) the isolation of the ancestor of Stygiocaris suggests that these two island populations are differentiated lancifera and Stygiocaris stylifera after the emergence of the at the species level. Cape Range anticline in north-western Australia (7–10 Ma; The three different T. monae populations from Hispani- Page et al., 2008; see above); and (3) the occlusion of the ola, located at opposite corners of the Dominican Republic, Havana–Matanzas Channel in Cuba at 5–6 Ma (Iturralde- and the population from Bosque Gua´nica in Puerto Rico Vinent et al.,1996),whichcouldhavetriggeredtheisola- showed very low genetic divergences (< 0.5% for COI). tion of the ancestors of the sister species T. consobrina and Moreover, a comparison of T. monae from Playa Fronto´ n Typhlatya taina. (Samana´ Peninsula, northern Dominican Republic) and We assumed three independent substitution rates, imple- Typhlatya utilaensis fromthesinglelocalityknownthusfar mented as three clocks: a rate for the mitochondrial protein- showed that two of the four diagnostic morphological coding genes (COI, cyt b), another for rrnL, and a third for the characters considered for the latter species (Alvarez et al., nuclear data set (histone H3A, SSU and LSU). beast analyses 2005) are similar in both taxa. Unfortunately, the single were run for 50 million generations, sampling every 1000 gen- T. utilaensis specimen available proved to be useless for erations. The outputs were analysed with Tracer 1.5 and molecular analysis. Sequencing further samples could con- TreeAnnotator 1.6.0 (Drummond & Rambaut, 2009), after firm the conspecific status of these two taxa in the future. the first 5 million generations had been discarded. Typhlatya garciai from Providenciales (Caicos Islands) and T. kakuki from Acklins Island (Bahamas; see Table 2) showed identical histone H3A sequences and low divergences RESULTS for COI and rrnL(< 0.6%). We consider here that T. kakuki Data regarding the species, populations, collection sites is only a population of T. garciai that has a completely and corresponding EMBL accession numbers of the DNA regressed cornea.

Journal of Biogeography 40, 594–607 599 ª 2012 Blackwell Publishing Ltd A. Botello et al.

Three populations of T. miravetensis, separated by up to Based on the most probable topology and molecular rates, 40 km in eastern Spain, showed a divergence in the mito- and using the three palaeogeographical events as calibration chondrial markers of 1–3%, but their nuclear sequences were points, a relaxed molecular clock estimated an age of 30.6– identical. The two Cuban populations of T. consobrina 61.1 Myr for the most recent common ancestor (MRCA) of included here showed significant divergences in rrnL (2.3%), Typhlatya + Stygiocaris + Antecaridina + Halocaridina (Figs 4 cyt b (6.1%) and COI (4.7%). & 5, Appendix S2). This analysis also estimated an age of 5.0 –7.7 Myr for the ancestor of the divergent populations of T. galapagensis from Santa Cruz and Isabela (node ‘f’ in Phylogenetic analyses Fig. 4). The age of the MRCA of all Typhlatya species (minus A preliminary Bayesian analysis including our species data T. galapagensis) + Stygiocaris (node ‘b’ in Fig. 4) was 25.7– set plus a selection of the taxa considered by von Rintelen 47.0 Myr, whereas the ancestor of all Typhlatya species et al. (2012),andusingthesamethreegenemarkersas (minus T. monae and T. galapagensis) (node ‘e’ in Fig. 4) those authors, showed that the cluster Antecaridina–Haloca- lived 21.9–40.4 Ma. ridina–T. galapagensis is the monophyletic sister group of the remaining Typhlatya/Stygiocaris species (Fig. 2). ILD DISCUSSION tests of our complete data set indicated that the six parti- tions were not incongruent (P > 0.13). Total evidence Molecular dating using the node age priors separately or in derived from the Bayesian and maximum likelihood phylo- combination produced compatible age estimates, particularly genetic trees corroborated the paraphyly of the genus Typ- for the most recent nodes (< 20 Myr; see Fig. 5 and Appendix hlatya because the Gala´pagos species clustered with S2). We found a marked inconsistency between the divergence A. lauensis with a high posterior probability (PP = 1.0; time estimates in our phylogeny and one of the major palaeo- Fig. 3). The Australian genus Stygiocaris was nested within geographical events in Tethys’ history – the establishment of Typhlatya, as suggested by Page et al. (2008). The only tree deep water between the two shores of the north-central Atlan- node showing weak support involved T. monae,which tic Ocean at about 110 Ma (Sclater et al., 1977; Jones et al., appeared as the sister group to Stygiocaris (PP = 0.93; 56% 1995). Our estimates date the separation between the western bootstrap support in the ML analysis). However, this spe- Atlantic/Caribbean (minus T. monae) and the Mediterranean cies appeared basal to the rest of Typhlatya + Stygiocaris lineage of Typhlatya at 21.9–40.4 Ma (see Fig. 4), which is (PP = 0.87) in an analysis that included additional out- much later than the disruption of the shallow-water connec- group species and a reduced (three genes) data set (Fig. 2). tions between the two shores of the Atlantic. Thus, the distri- Shimodaira–Hasegawa tests revealed no significant differ- bution of Typhlatya/Stygiocaris cannot be explained solely by ences between the two alternative topologies. the vicariant isolation that accompanied the fragmentation of

Figure 2 Bayesian phylogram of Typhlatya/Stygiocaris and related genera based on rrnL, LSU and histone H3A sequences. Numbers beside nodes show Bayesian posterior probabilities.

600 Journal of Biogeography 40, 594–607 ª 2012 Blackwell Publishing Ltd Zoogeography of Typhlatya cave shrimps

Figure 3 Bayesian phylogram showing the relationships among the Typhlatya/Stygiocaris species based on rrnL, COI, cyt b, LSU, SSU and histone H3A sequences, with Halocaridina rubra and Antecaridina lauensis as the outgroups. The numbers above the nodes show the Bayesian posterior probabilities, and those below the nodes show the bootstrap support values estimated with maximum likelihood. the Tethys Sea. We suggest that this disjunct amphi-Atlantic divergence of the Mediterranean and western Atlantic distribution could be the result of the extinction of species Typhlatya lineages (21.6–44.4 Ma; see Fig. 4). from central and eastern Atlantic archipelagos, and that new The sister relationship found between Typhlatya and the Typhlatya species might even await discovery in the Macaro- Australian genus Stygiocaris (with the caveat that corrobora- nesian islands. There is compelling geological evidence for the tion is needed from additional molecular evidence) is hardly presence of drowned archipelagos and seamounts in the compatible with their presumed vicariant divergence due to central East Atlantic Ocean from at least 60 Ma (Geldmacher the occlusion of the connection between the Mediterranean et al., 2001, 2005; Ferna´ndez-Palacios et al., 2011). These and the Indian Ocean (Page et al., 2008). The timeframe Palaeo-Macaronesian islands were located much closer to the established for the collision of the Arabian Plate with Anato- western Mediterranean than they would be today and were lia (16–20 Ma; Meulenkamp & Sissingh, 2003) not only affected by the east-to-west warm circumequatorial marine post-dates our age estimate for the divergence of Typhlatya Tethys Sea current (Ferna´ndez-Palacios et al., 2011). The exis- s. str. and Stygiocaris (25.7–47.0 Ma; Fig. 4), but also (and tence of these vanished archipelagos supports the potential more relevantly) the divergence of the sister taxa T. monae presence of Typhlatya in the area and also the relatively recent and Stygiocaris (22.0–42.3 Ma; Fig. 4).

Figure 4 Chronogram showing the estimated age ranges (Ma; 95% high posterior density limits as confidence intervals) of the cladogenetic events within the Typhlatya–Stygiocaris–Antecaridina–Halocaridina lineage. Asterisks indicate the nodes used as calibration points (see text for details).

Journal of Biogeography 40, 594–607 601 ª 2012 Blackwell Publishing Ltd A. Botello et al.

Figure 5 Plot showing the mean age estimates (Myr) for the nodes shown in Fig. 4, obtained with three different calibration points (see text for details). Black broken lines show the 95% high posterior density limits for the node ages using the three-point combined calibration.

sequence data set, they found that the Mexican T. pearsei Paraphyly of Typhlatya was the sister taxon of Stygiocaris, rather than any surface or Monod & Cals (1970) assigned a series of juvenile blind aty- cave atyids from Australia or the Indo-Pacific region. ids from Santa Cruz and Isabela in the Gala´pagos to Typ- Typhlatya monae, with an apparently broad distribution in hlatya, although they noted some morphological similarities the Caribbean, is unique in displaying a uniramous pereio- to Antecaridina. In our phylogenetic analyses, T. galapagensis pod (the fifth), whereas the other Typhlatya species display a is placed as the sister taxon to Antecaridina lauensis with well-developed exopod on all pereiopods. The introduction strong statistical support, rendering the genus Typhlatya of a new genus to accommodate T. monae on the basis of paraphyletic in its current conception. Our study also sug- this feature should be considered, and would give taxonomic gests a possible species-level differentiation between the pop- relevance to the broad molecular divergence of the Typhlatya ulations of T. galapagensis from the islands of Santa Cruz s. str., Stygiocaris and T. monae lineages. and Isabela. The COI genetic distance between these two Zaksˇek et al. (2007) and Sket & Zaksˇek (2009) have, how- populations (8%) approaches the minimum of 10% found in ever, recently challenged the relevance of features such as the our study to distinguish different Typhlatya species (Appen- presence of certain spines on the anterior margin of the ceph- dix S1), but is considerably higher than the maximum inter- alothorax or the absence of exopods on the pereiopods in populational distances we identified for T. monae, T. garciai/ distinguishing atyid genera. These authors, based on molecu- kakuki and T. miravetensis. lar markers, have shown that a presumed Typhlatya from the Balkans is actually a modified Troglocaris, which displays smooth anterior margins on its cephalothorax and uniramous Stygiocaris and Typhlatya monae pereiopods. Our own data for Antecaridina lend support to Stygiocaris and T. monae are morphologically peculiar, even this hypothesis, because Typhlatya galapagensis (which lacks though they cluster with the rest of the Typhlatya species spines on the anterior margin of the cephalothorax) occurs in in a robust monophyletic group (Fig. 3). The position of our phylogram as a sister group to A. lauensis, a taxon dis- T. monae is not fully established on the tree, but occurs playing both suborbital and pterygostomial spines. Our own either as the sister taxon to the rest of Typhlatya or as the observations of the shape of the sternal process in Typhlatya sister taxon to Stygiocaris (as shown in Figs 2 & 3, respec- s. str. also indicate that it is identical to Stygiocaris, support- tively). In any event, our analysis confirms the long-indepen- ing the congeneric status of the two taxa. dent evolution of these three sublineages (i.e. Typhlatya s. str., Stygiocaris and T. monae; see Fig. 3). Dispersal, population structure and divergence Stygiocaris is a stygobiont genus endemic to north-western of Typhlatya species Australia, composed of three species, only two of which have been formally described (Holthuis, 1960; Page et al., 2008). The distribution patterns of the Atyidae are dependent on Page et al. (2008) have already noted the sister relationship life-history traits, such as their dispersal capacities and spe- between Stygiocaris and Typhlatya based on molecular cies-specific tolerance to local conditions (Page & Hughes, evidence. Using a combined nuclear and mitochondrial 2007). Typhlatya species are usually very localized, limited in

602 Journal of Biogeography 40, 594–607 ª 2012 Blackwell Publishing Ltd Zoogeography of Typhlatya cave shrimps most instances to a single island or narrow portion of coast. marine habitats. Hunter et al. (2008) favoured an alternative However, several Typhlatya species display relatively broad scenario, where mid-Atlantic species reached their present distributions that, in some instances, include territories sepa- distributions by transoceanic dispersal of a shallow-water rated by stretches of sea. Thus, T. monae is known from ancestor. These researchers identified a sister relationship Mona Island, Puerto Rico and Hispaniola (Greater Antilles), between T. iliffei from Bermuda and T. garciai from Provi- the more distant Barbuda (Lesser Antilles) and Curac¸ao and denciales (Caicos Islands) based on molecular evidence, and the San Andre´s islands, the last two at opposite sides of the suggested that the taxon from Bermuda might have derived Caribbean (see Table 1). Our own data for three different from a Bahamian ancestor dispersed via the Gulf Stream. populations from Hispaniola and one from Puerto Rico sug- The most common recent ancestor of the Bahamian, gest the occurrence of panmixia (Appendix S1). Ascension and the Bermudan taxa lived 18.6–33.9 Ma (see Typhlatya garciai/kakuki is known from north-eastern and Fig. 4), which is compatible with the colonization of north-western Cuba, Providenciales (Caicos Islands) and Bermuda by overseas dispersal of a Bahamian ancestor, as Acklins Island (Bahamas; see Table 1). Our own observations proposed by Hunter et al. (2008) (the age of Bermuda is 40– of the latter two populations, separated by a deep-water sea 36 Myr; see above). Our estimate for the divergence of the arm of 173 km, indicated very low molecular divergence, Bermuda–Ascension lineages (1.9–5.3 Ma; see Fig. 4) is also which could be explained either by continuous gene flow compatible with the age of Ascension (2.5 Myr; see above). through dispersal over sea or, more likely given the separa- Although both islands are separated by a huge expanse of tion of the populations, by recent colonization and subse- ocean and the prevailing equatorial currents would make the quent isolation. derivation of one from the other by long-distance over-sea The high dispersal potential of Typhlatya across subterra- dispersal untenable, the most likely explanation for their ori- nean waters has already been pointed out by Hunter et al. gin is that they derived from a diadromous Bahamian lineage (2008) on the Yucata´n Peninsula, where haplotypes of Typ- which colonized Bermuda first and subsequently colonized hlatya mitchelli are shared between populations separated by Ascension Island. up to 235 km. However, we found significant isolation among the populations of T. consobrina (sampled from two CONCLUSIONS locations in Cuba about 330 km apart) and T. miravetensis (three populations separated by up to 40 km in eastern In this study, we examined the molecular phylogeny of Typ- Spain) (Appendix S1). hlatya shrimps using nuclear and mitochondrial gene sequences and a relaxed molecular clock. These stygobiont atyids show an extremely disjunct distribution, which has Transoceanic dispersal of Typhlatya been suggested to derive from plate tectonic vicariance. Our Two closely related species, Typhlatya iliffei (Bermuda) and results confirm the paraphyly of Typhlatya, because T. gala- Typhlatya rogersi (Ascension), are found on mid-oceanic pagensis from the Gala´pagos Islands is the sister taxon to An- islands in the Atlantic. Bermuda is the cap of a mid-plate tecaridina. Furthermore, the Greater Antillean T. monae rise in a sector of the north-western Atlantic with no other probably represents an independent sister lineage to the Aus- seamounts or ridges that could have harboured members of tralian genus Stygiocaris. We have analysed the relaxed Typhlatya in the past. The pillow lavas that formed the origi- molecular clock for Typhlatya using three different calibra- nal Bermuda shield volcano are no older than 47–40 Myr, tion points based on three independent palaeogeographical and at 40–36 Ma, the Bermuda platform had already risen to events. We show that in Typhlatya/Stygiocaris, the ages of the sea level (Vogt & Jung, 2007). Ascension, located about corresponding subclades post-date the establishment of deep 7000 km to the south-east of Bermuda, occurs 90 km west water between the north-central Atlantic Ocean shores. In of the Mid-Atlantic Ridge on 7-Myr-old oceanic lithosphere. addition, the divergence of the T. monae lineage from the Its oldest subaerial lava flows have been dated recently at rest of the genus Typhlatya preceded the cladogenesis of Typ- 2.5 Ma (Minshull et al., 2010). hlatya s. str. into a Mediterranean and a Caribbean/Mid- Iliffe et al. (1983) proposed that T. iliffei and T. rogersi Atlantic clade. Therefore, our results are inconsistent with a represent an ancient atyid stock that survived on submerged simple explanation of the origin of the group based on plate and emergent seamounts along or associated with the Mid- tectonic vicariance. Atlantic Ridge. Alternatively, Hart et al. (1985) and Manning et al. (1986) proposed that the ancestral form of Typhlatya ACKNOWLEDGEMENTS was a deep-sea benthic organism that originally entered the cave environment directly from deep water via cracks and We are greatly indebted to several colleagues for the provi- fissures on submerged seamount slopes during the Mesozoic. sion of specimens or for sharing fieldwork, namely Sammy Opposing this view is the alleged primary freshwater condi- de Grave (Typhlatya from Ascension); Bill Humphreys and tion of the family Atyidae, which already included limnic Tim Page (Antecaridina and Stygiocaris from Christmas representatives by the middle Cretaceous (Rabada`, 1993), Island and Cape Range Peninsula, Western Australia); Jose´ and the fact that Typhlatya has never been recorded in open A. Ottenwalder, Tonyo Alcover and Maria del Mar Bauza`-

Journal of Biogeography 40, 594–607 603 ª 2012 Blackwell Publishing Ltd A. Botello et al.

Ribot (Typhlatya from the Dominican Republic); Scott R. Chace, F.A. (1954) Two new subterranean shrimps (Deca- Santos (Halocaridina from Hawaii); Franck Bre´hier (Typ- poda: ) from Florida and the West Indies, with a hlatya from France); Alberto Sendra (Typhlatya from Spain); revised key to the American species. Journal of the Wash- and Jose Luis Villalobos, Rocı´o Gonza´lez, Antonio Celis, Tu- ington Academy of Sciences, 44, 318–324. lio del Angel and Rolando Mendoza (Typhlatya from the Chace, F.A. (1975) Cave shrimps (Decapoda: Caridea) from Yucata´n Peninsula). A.B. benefited from a CONACYT (Mex- the Dominican Republic. Proceedings of the Biological ico) post-doctoral fellowship during the completion of this Society of Washington, 88,29–44. study (reg. number 76128). This is a contribution to Spanish Chace, F.A. & Hobbs, H.H. (1969) The freshwater and MCINN project CGL2009-08256, partially financed with EU terrestrial decapod crustaceans of the West Indies with FEDER funds. National Science Foundation grants to T.I. special reference to Dominica. Bulletin of the United States supported collection of Typhlatya specimens from the Gala´- National Museum, 292,1–258. pagos (BSR-8417494) and the Bahamas (DEB-0315903). The Chace, F.A. & Manning, R.B. (1972) Two new caridean Bermuda Aquarium, Museum and Zoo and the Bermuda shrimps, one representing a new family, from marine Department of Conservation Services assisted T.I. with pools on Ascension Island (Crustacea: Decapoda: Natan- collection of specimens from Bermuda. Brett Dodson and tia). Smithsonian Contributions to Zoology, 131,1–18. Gil Nolan provided diving assistance in the Bahamas and Colgan, D.J., Mclauchlan, A., Wilson, G.D.F., Livingston, S.P., Bermuda, respectively. We thank Smithsonian Journeys and Edgecombe, G.D., Macaranas, J., Cassis, G. & Gray, M.R. Celebrity Cruise Lines for assisting with travel arrangements (1998) Histone H3 and U2 snRNA DNA sequences and to Puerto Rico. Tom Miller of the University of Puerto Rico molecular evolution. Australian Journal of Zool- generously provided information on cave locations. Funding ogy, 46, 419–437. for fieldwork in Puerto Rico came from a NOAA Ocean Craft, J.D., Russ, A.D., Yamamoto, M.N., Iwai, T.Y. Jr, Hau, S., Exploration grant to T.I. Thanks are due to Timothy Page Kahiapo, J., Chong, C.T., Ziegler-Chong, S., Muir, C., Fujita, and two anonymous referees for constructive comments on a Y., Polhemus, D.A., Kinzie, R.A., III & Santos, S.R. (2008) previous version of this manuscript. Islands under islands: the phylogeography and evolution of Halocaridina rubra Holthuis, 1963 (: Decapoda: REFERENCES Atyidae) in the Hawaiian archipelago. Limnology and Ocean- ography, 53,675–689. Alvarez, F., Iliffe, T.M. & Villalobos, J.L. (2005) New species Creaser, E.P. (1936) Crustaceans from Yucatan. The cenotes of of the genus Typhlatya (Decapoda: Atyidae) from anchia- Yucatan: a zoological and hydrographic survey (ed. by A.S. line caves in Mexico, the Bahamas and Honduras. Journal Pearse, E.P. Creaser and F.G. Hall), pp. 117–132. Carnegie of Crustacean Biology, 25,81–94. Institution, Washington, DC. Banarescu, P. (1990) Zoogeography of freshwaters. Vol. 1. Creaser, E.P. (1938) Larger cave Crustacea of the Yucatan General distribution and dispersal of freshwater . Peninsula. Fauna of the caves of Yucatan (ed. by A.S. Aula-Verlag, Wiesbaden, Germany. Pearse), pp. 159–164. Carnegie Institution, Washington, Barr, T.C. & Holsinger, J.R. (1985) Speciation in cave faunas. DC. Annual Review of Ecology and Systematics, 16, 313–337. Croizat, L., Nelson, G. & Rosen, D.E. (1974) Centers of ori- Barraclough, T.G., Hogan, J.E. & Vogler, A.P. (1999) Testing gin and related concepts. Systematic Zoology, 23, 265–287. whether ecological factors promote cladogenesis in a group Debrot, O. (2003) The freshwater shrimps of Curac¸ao, West of tiger beetles (Coleoptera: Cicindelidae). Proceedings of Indies (Decapoda, Caridea). Crustaceana, 76,65–76. the Royal Society B: Biological Sciences, 266, 1061–1067. Drummond, A.J. & Rambaut, A. (2007) BEAST: Bayesian Botosaneanu, L. & Holthuis, L.B. (1970) Subterranean evolutionary analysis by sampling trees. BMC Evolutionary shrimps from Cuba (Crustacea Decapoda Natantia). Trav- Biology, 7, 214. aux de l’Institut de Spe´ologie ‘Emile Racovitza’, 9, 121–133. Estrada, A. & Go´mez, O. (1987) Una nueva especie del ge´n- Boxshall, G.A. (1989) Colonization of inland marine caves ero Typhlatya (Decapoda: Atyidae) de Cuba. Poeyana, 355, by misophrioid copepods. Journal of Zoology, 219, 521–526. 1–12. Brown, J.M. & Lemmon, A.R. (2007) The importance of data Farris, J.S., Ka¨llersjo¨, M., Kluge, A.G. & Bult, C. (1995) Con- partitioning and the utility of Bayes factors in Bayesian structing a significance test for incongruence. Systematic phylogenetics. Systematic Biology, 56, 643–655. Biology, 44, 570–572. Buden, D.W. & Felder, D.L. (1977) Cave shrimps in the Cai- Ferna´ndez-Palacios, J.M., de Nascimento, L., Otto, R., Del- cos Islands. Proceedings of the Biological Society of Washing- gado, J.D., Garcı´a-del-Rey, E., Are´valo, J.R. & Whittaker, ton, 90, 108–115. R.J. (2011) A reconstruction of Palaeo-Macaronesia, with Ca´rdenas, M. (1950) Informe hidrobiolo´gico y faunı´stico de particular reference to the long-term biogeography of the Yucata´n. Boletı´n de la Sociedad Mexicana de Geografı´ay Atlantic island laurel forests. Journal of Biogeography, 38, Estadı´stica, 69, 135–159. 226–246. Chace, F.A. (1942) A new cave shrimp from Cuba. Proceed- Folmer, O., Black, M., Hoeh, W., Lutz, R. & Vrijenhoek, R.C. ings of the New England Zoological Club, 19,99–102. (1994) DNA primers for amplification of mitochondrial

604 Journal of Biogeography 40, 594–607 ª 2012 Blackwell Publishing Ltd Zoogeography of Typhlatya cave shrimps

cytochrome c oxidase subunit I from diverse metazoan rubra Holthuis, 1963 (Crustacea: Decapoda: Atyidae). invertebrates. Molecular Marine Biology and Biotechnology, Gene, 394,35–44. 3, 294–299. Jaume, D. & Bre´hier, F. (2005) A new species of Typhlatya Geldmacher, J., Hoernle, K., van den Bogaard, P., Zankl, G. (Crustacea: Decapoda: Atyidae) from anchialine caves on &Garbe-Scho¨nberg, D. (2001) Earlier history of the > the French Mediterranean coast. Zoological Journal of the 70-Ma-old Canary hotspot based on temporal and Linnean Society, 144, 387–414. geochemical evolution of the Selvagems Archipelago and Jones, E.J.W., Cande, S.C. & Spathopoulos, F. (1995) Evolu- neighbouring seamounts in the eastern North Atlantic. tion of a major oceanographic pathway: the equatorial Journal of Volcanology and Geothermal Research, 111, Atlantic. Geological Society, London, Special Publications, 55–87. 90, 199–213. Geldmacher, J., Hoernle, K., van den Bogaard, P., Duggen, S. Juarrero, A. (1994) Nueva especie de camaro´ncavernı´cola & Werner, R. (2005) New 40Ar/39Ar age and geochemical (Decapoda: Atyidae: Typhlatya)deCuba.Avicennia, 1,57–66. data from seamounts in the Canary and Madeira volcanic Juarrero, A. & Ortiz, M. (2000) El ge´nero Typhlatya (Crusta- provinces: support for the mantle plume hypothesis. Earth cea: Decapoda: Atyidae) en Cuba, con la descripcio´nde and Planetary Science Letters, 237,85–101. una nueva especie. Avicennia, 12(13), 45–54. Hart, C.W. & Manning, R.B. (1981) The cavernicolous cari- Katoh, K., Kuma, K.-I., Toh, H. & Miyata, T. (2005) MAFFT dean shrimps of Bermuda (, Hippolytidae, and version 5: improvement in accuracy of multiple sequence Atyidae). Journal of Crustacean Biology, 1, 441–456. alignment. Nucleic Acids Research, 33, 511–518. Hart, C.W. & Manning, R.B. (1986) Two new shrimps Lefe´bure, T., Douady, C.J., Gouy, M., Trontelj, P., Briolay, J. & (Procarididae and Agostocarididae, new family) from mar- Gibert, J. (2006) Phylogeography of a subterranean amphi- ine caves of the western North Atlantic. Journal of Crusta- pod reveals cryptic diversity and dynamic evolution in cean Biology, 6, 408–416. extreme environments. Molecular Ecology, 15, 1797–1806. Hart, C.W., Manning, R.B. & Iliffe, T.M. (1985) The fauna Manning, R.B., Hart, C.W. & Iliffe, T.M. (1986) Mesozoic of Atlantic marine caves: evidence of dispersal by sea floor relicts in marine caves of Bermuda. Stygologia, 2, 156–166. spreading while maintaining ties to deep waters. Proceed- Meulenkamp, J.E. & Sissingh, W. (2003) Tertiary palaeogeog- ings of the Biological Society of Washington, 98, 288–292. raphy and tectonostratigraphic evolution of the Northern Hobbs, H.H. III & Hobbs, H.H. Jr (1976) On the troglobitic and Southern Peri-Tethys platforms and the intermediate shrimps of the Yucatan Peninsula, Mexico (Decapoda: domains of the African–Eurasian convergent plate bound- Atyidae and Palaemonidae). Smithsonian Contributions to ary zone. Palaeogeography, Palaeoclimatology, Palaeoecology, Zoology, 240,1–23. 196, 209–228. Holthuis, L.B. (1960) Two new species of atyid shrimps from Minshull, T.A., Ishizuka, O. & Garcia-Castellanos, D. (2010) the subterranean waters of NW Australia (Decapoda Nat- Long-term growth and subsidence of Ascension Island: antia). Crustaceana, 1,47–57. constraints on the rheology of young oceanic lithosphere. Holthuis, L.B. (1977) On some freshwater and terrestrial Geophysical Research Letters, 37, L23306. Crustacea Decapoda from Cuba. Resultats des expe´ditions Monod, T. (1975) Sur la distribution de quelques Crustace´s biospe´ologiques cubano-rumaines a` Cuba (ed. by T. Orghi- Malacostrace´s d’eau douce et saumaˆtre. Me´moires du dan), pp. 271–275. Editura Academiei Republicii Socialiste Muse´um National d’Histoire Naturelle (Serie A, Zoologie), Romaˆnia, Bucarest. 88,98–103. Huelsenbeck, J.P. & Ronquist, F. (2001) MRBAYES: Monod, T. & Cals, P. (1970) Sur une espe`ce nouvelle de cre- Bayesian inference of phylogenetic trees. Bioinformatics, vette cavernicole: Typhlatya galapagensis (Decapoda Natan- 17, 754–755. tia; Atyidae). Mission Zoologique Belge auxˆ Iles Galapagos Hunter, R.L., Webb, M.S., Iliffe, T.M. & Alvarado Bremer, J.R. et en Ecuador, 2,57–103. (2008) Phylogeny and historical biogeography of the cave- Page, T.J. & Hughes, J.M. (2007) Radically different scales of adapted shrimp genus Typhlatya (Atyidae) in the Caribbean phylogeographic structuring within cryptic species of Sea and western Atlantic. Journal of Biogeography, 35,65–75. freshwater shrimp (Atyidae: Caridina). Limnology and Iliffe, T.M. (1986) The zonation model for the evolution of Oceanography, 52, 1055–1066. aquatic faunas in anchialine caves. Stygologia, 2,2–9. Page, T.J., Humphreys, W.F. & Hughes, J.M. (2008) Shrimps Iliffe, T.M., Hart, C.W. & Manning, R.B. (1983) Biogeogra- down under: evolutionary relationships of subterranean phy and the caves of Bermuda. Nature, 302, 141–142. crustaceans from Western Australia (Decapoda: Atyidae: Iturralde-Vinent, M., Hubbell, G. & Rojas, R. (1996) Catalog Stygiocaris). PLoS ONE, 3, e1618. of Cuban fossil Elasmobranchii (Paleocene–Pliocene) and Peck, S.B. (1974) The invertebrate fauna of tropical Ameri- paleogeographic implications of their Lower–Middle Mio- can caves, part II: Puerto Rico, an ecological and zoogeo- cene occurrence. Journal of the Geological Society of graphic analysis. Biotropica, 6,14–31. Jamaica, 31,7–21. Phillips, M.J. (2009) Branch-length estimation bias misleads Ivey, J.L. & Santos, S.R. (2007) The complete mitochondrial molecular dating for a vertebrate mitochondrial phylogeny. genome of the Hawaiian anchialine shrimp Halocaridina Gene, 441, 132–140.

Journal of Biogeography 40, 594–607 605 ª 2012 Blackwell Publishing Ltd A. Botello et al.

Posada, D. (2008) JModelTest: phylogenetic model Atyidae) in the Solomon Islands: an intriguing biogeo- averaging. Molecular Biology and Evolution, 25, graphical problem. Hydrobiologia, 85,49–58. 1253–1256. Stamatakis, A., Ludwig, T. & Meier, H. (2005) RAxML-III: Rabada`, D. (1993) Crusta´ceos deca´podos lacustres de las cali- a fast program for maximum likelihood-based inference zas litogra´ficas del Creta´cico inferior de Espan˜a: Las Hoyas of large phylogenetic trees. Bioinformatics, 21, 456–463. (Cuenca) y el Montsec de Ru´ bies (Lleida). Cuadernos de Sterrer, W. (1973) Plate tectonics as a mechanism for dis- Geologı´a Ibe´rica, 17, 345–370. persal and speciation in interstitial sand fauna. Netherlands Rambaut, A. & Drummond, A.J. (2009) Tracer v. 1.5. Avail- Journal of Sea Research, 7, 200–222. able at: http://tree.bio.ed.ac.uk/software/tracer/ (accessed Stock, J.H. (1986) Caribbean biogeography and a biological 12 July 2011). calendar for geological events. Crustacean biogeography von Rintelen, K., Page, T.J., Cai, Y., Roe, K., Stelbrink, B., (ed. by R.H. Gore and K.L. Heck), pp. 195–203. A.A. Balk- Kuhajda, B.R., Iliffe, T.M., Hughes, J. & von Rintelen, T. ema, Rotterdam. (2012) Drawn to the dark side: a molecular phylogeny of Stock, J.H. (1993) Some remarkable distribution patterns freshwater shrimps (Crustacea: Decapoda: Caridea: Atyi- in stygobiont Amphipoda. Journal of Natural History, 27, dae) reveals frequent cave invasions and challenges current 807–819. taxonomic hypotheses. Molecular Phylogenetics and Evolu- Swofford, D.L. (2002) PAUP*: phylogenetic analysis using tion, 63,82–96. parsimony (and other methods), version 4.0b10. Sinauer Rosen, D.E. (1975) A vicariance model for Caribbean bioge- Associates, Sunderland, MA, USA. ography. Systematic Zoology, 24, 431–464. Vogt, P.R. & Jung, W.-Y. (2007) Origin of the Bermuda vol- Russ, A., Santos, S.R. & Muir, C. (2010) Genetic population canoes and the Bermuda Rise: history, observations, mod- structure of an anchialine shrimp, Metabetaeus lohena els, and puzzles. Geological Society of America Special (Crustacea: Alpheidae), in the Hawaiian Islands. Revista de Papers, 430, 553–591. Biologı´a Tropical, 58, 159–170. Wagner, H.P. (1994) A monographic review of the Thermos- Santos, S.R. (2006) Patterns of genetic connectivity among baenacea (Crustacea: Peracarida): a study on their mor- anchialine habitats: a case study of the endemic Hawaiian phology, , phylogeny and biogeography. shrimp Halocaridina rubra on the island of Hawaii. Molec- Zoologische Verhandelingen, 291,1–338. ular Ecology, 15, 2699–2718. Werner, R., Hoernle, K., van den Bogaard, P., Ranero, C., Sanz, S. & Platvoet, D. (1995) New perspectives on the evo- von Huene, R. & Korich, D. (1999) Drowned 14-m.y.-old lution of the genus Typhlatya (Crustacea, Decapoda): first Gala´pagos archipelago off the coast of Costa Rica: implica- record of a cavernicolous atyid in the Iberian Peninsula, tions for tectonic and evolutionary models. Geology, 27, Typhlatya miravetensis n. sp. Contributions to Zoology, 65, 400–502. 77–99. Xiong, B. & Kocher, T.D. (1991) Comparison of mitochon- Sclater, J.G., Hellinger, S. & Tapscott, C. (1977) The paleo- drial DNA sequences of seven morphospecies of black flies bathymetry of the Atlantic Ocean from the Jurassic to the (Diptera: Simuliidae). Genome, 34, 306–311. present. Journal of Geology, 85, 509–552. Zaksˇek, V., Sket, B. & Trontelj, P. (2007) Phylogeny of the Shull, V.L., Vogler, A.P., Baker, M.D., Maddison, D.R. & cave shrimp Troglocaris: evidence of a young connection Hammond, P.M. (2001) Sequence alignment of 18S ribo- between Balkans and Caucasus. Molecular Phylogenetics somal RNA and the basal relationships of adephagan and Evolution, 42, 223–235. beetles: evidence for monophyly of aquatic families and the placement of Trachypachidae. Systematic Biology, 50, 945–969. SUPPORTING INFORMATION Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H. & Flock, P. (1994) Evolution, weighting, and phylogenetic Additional Supporting Information may be found in the utility of mitochondrial gene sequences and a compilation online version of this article: of conserved polymerase chain reaction primers. Annals of Appendix S1 Percentage uncorrected pairwise genetic dis- the Entomological Society of America, 87, 651–701. tances among the COI sequences. Sket, B. (1988) Fauna of the hypogean waters of San Appendix S2 Mean ages of the tree nodes in Fig. 4 and Andres (Saint Andrews) Island, Colombia. Bioloski Vestnik, molecular evolutionary rates, calculated using different cali- 36,77–82. bration points. Sket, B. & Zaksˇek, V. (2009) European cave shrimp species (Decapoda: Caridea: Atyidae), redefined after a phyloge- As a service to our authors and readers, this journal pro- netic study; redefinition of some taxa, a new genus and vides supporting information supplied by the authors. Such four new Troglocaris species. Zoological Journal of the Lin- materials are peer-reviewed and may be re-organized for nean Society, 155, 786–818. online delivery, but are not copy-edited or typeset. Technical Smith, M.J. & Williams, W.D. (1981) The occurrence of support issues arising from supporting information (other Antecaridina lauensis (Edmondson) (Crustacea, Decapoda, than missing files) should be addressed to the authors.

606 Journal of Biogeography 40, 594–607 ª 2012 Blackwell Publishing Ltd Zoogeography of Typhlatya cave shrimps

BIOSKETCH

Alejandro Botello is a carcinologist who shares with the other authors a vivid interest on the evolution and biogeography of anchialine crustaceans. His research is mainly focused on the cave decapod fauna of the Yucata´n Peninsula. Author contributions: A.B., C.J. and D.J. conceived the ideas; A.B., F.A., T.I. and D.J. collected the data; A.B, J.P. and C.J. analy- sed the data; and A.B., C.J. and D.J. led the writing.

Editor: John Lambshead

Journal of Biogeography 40, 594–607 607 ª 2012 Blackwell Publishing Ltd