Two Genetically Distinct Populations of Bobtail Squid, Euprymna seolopes, Exist on the Island ofO'ahu1
J. R. Kimbell,2 M. J. MeFall-Ngai,2 and G. K. Roderiek3
Abstract: Population structure of the endemic Hawaiian bobtail squid, Eu prymna seolopes, was examined using both morphological and genetic data. Al though allozyme polymorphism was negligible, measurements ofeggs, juveniles, and adults, as well as genetic data sequences of mitochondrial cytochrome oxi dase I, demonstrated highly significant population structuring between two populations found on the northeastern and southern coasts of the island of O'ahu. These data suggest that extremely low levels of gene flow occur among these populations. Population subdivision of marine shallow-water invertebrates in Hawai'i is not expected based on earlier surveys, but may reflect a more general pattern for organisms, both marine and terrestrial, that exhibit limited dispersal. The subdivision also provides insight into the pathway through which coevolution between E. seolopes and its internal symbiont, Vibrio fiseheri, may proceed. THE HAWAIIAN ARCHIPELAGO presents thought to result from the extremely het unique opportunities for the study of evo erogeneous nature of the terrestrial island lutionary biology because of its relative iso landscape, which presents an abundance of lation, high levels of endemism, and the spatially limited, yet distinct, niches ripe for explosive radiation of a number of taxa (Si adaptive radiation (Gillespie et al. 2001). mon 1987, Wagner and Funk 1995). For Marine invertebrates, however, have much example, the terrestrial invertebrate biota, lower rates of endemism than their terrestrial which has been particularly well studied in counterparts, with endemism ranging from 2 genera such as Drosophila flies, Tetragnatha to 20% for nearshore gastropods, brachyuran spiders, and Achatinellinae tree snails, typi decapods, and certain polychaetes and echi cally exhibits extensive adaptive radiations noderms of the central Pacific (Scheltema comprising species endemic to the archipel 1986, Kay and Palumbi 1987). These benthic ago (Roderick and Gillespie 1998, Thacker marine invertebrates appear instead to repre and Hadfield 2000). These patterns are sent attenuated Indo-Pacific fauna, and the marine endemic species that do exist are usu- ally distributed throughout all of the islands 1 This work was supported by NSF grants IBN 99- of the archipelago (Scheltema 1986). The 04601 to M.M.-N. and E. G. Ruby and DEB 0096189 wide geographic distributions of these marine to R. G. Gillespie and G.K.R., NIH grant RRI0926 to invertebrate species indicate that fairly high E. G. Ruby and M.M.-N., NOAA National SeaGrant gene flow must be occurring between pop RlFM-5 to G.K.R. and C. Smith, and NOAA HCRI grant NA870A0381 to G.K.R. Manuscript accepted 28 ulations within species (Palumbi 1992). Such November 2001. gene flow is presumably facilitated by long 2 Pacific ~iomedical Research Center, University of planktonic larval phases, which ensure high -----Hawai'i, 41 'Ahui Street,-Honolulu,-Hawai'i-96816.. ------disp~-rsat-Yet,within-sudrwidespre-adsp-ecies---- 3 Corresponding author: Department of Environ- mental Science, Policy, and Management, University of some population structure may nevertheless California, 201 Wellman Hall MC 3112, Berkeley, Cali- persist (Barber et al. 2000). fornia 94708-3112 (phone: 510-642-3327; fax: 510-642- Although much is known about marine 7428; E-mail: [email protected]). invertebrate species with extended larval life histories, less is understood about species that Pacific Science (2002), vol. 56, no. 3:347-355 do not have broadly dispersing larval or juve © 2002 by University of Hawai'i Press nile stages. One such species is the bobtail All rights reserved squid, Euprymna seolopes, which is endemic to
347 348 PACIFIC SCIENCE· July 2002
the Hawaiian Islands (Singley 1983). Female E. seolopes lay clutches of eggs on hard sub N strates in the shallow waters of the back reef. ,//"(\. No true larval stages occur in this species (i.e., i the juveniles hatch with a morphology and / .....•...... •...... behavior very similar to those of the adult [Arnold et al. 1972]). Specifically, after a brief planktonic stage of hours to a few days, the I juvenile squid take on the diel pattern of be Kaneohe Bay , . havior characteristic of the species (i.e., they bury in the sand during the day and forage in the water column at night). Previously, it was observed that E. seolopes is notable for its habitat loyalty, and it has been suggested that two reproductively t isolated populations exist on northeastern Niu Valley and southern coasts of the island of O'ahu (Singley 1983). The E. seolopes-Vibrio fiseheri FIGURE 1. Collection sites on O'ahu, Hawaiian Archi association has become a model for animal pelago, for specimens examined in this smdy. bacterial symbiosis and in particular the per sistent colonization of animal epithelia by extracellular bacteria (McFall-Ngai and Ruby windward O'ahu (Figure 1). Specimens were 1998). Recent studies of the squid-vibrio maintained in running seawater tables at 24°C symbioses at the level of species have pro at the Kewalo Marine Laboratory, University vided strong evidence for coevolution of the of Hawai'i. host and symbiont (Nishiguchi et al. 1998). Whether population subdivision of E. seolopes Morphometric Data exists is critical to understanding the co evolution of this species with its bacteria Measurements of morphological characters partners-how isolated are populations of E. were performed to determine whether signif seolopes and at what scale would population icant phenotypic differences occur between specific strains of V. fiseheri be expected to Kane'ohe Bay and Niu Valley animals. Three evolve? In this study we tested the hypothesis characters were measured on specimens from of population subdivision of E. seolopes with each of these two sites: female mantle length, both morphological and genetic data. Our egg diameter, and hatchling mantle length. results indicate that populations of E. seolopes The latter two measurements were made on O'ahu are indeed distinct. Further, the using a dissection microscope (Wild M5) at results demonstrate that at least some species lOx with an ocular micrometer. Eggs were of Hawaiian marine invertebrates are more measured on the day that they were laid. The subdivided than previously realized, likely as hatchlings were first anesthetized in a 50: 50 a consequence of a life history lacking dis solution of 0.3 7 M MgClz:seawater before persive egg or larval stages. the mantle-length measurements were taken...... -- ..------:A:pproximately-75--specimenswere-examirred for each character per locality. MATERIALS AND METHODS Capture and Maintenance ofSpecimens Allozymes: Cellulose Acetate Gel Electrophoresis Adult Euprymna seolopes were collected in Variation at allozyme loci was examined 1997 and 1998 from two sites, the shallow through cellulose acetate gel electrophoresis, sand flats offNiu Valley along leeward O'ahu a method that requires very little tissue and along the shores of Kane'ohe Bay on (Richardson et al. 1986, Hebert and Beaton Two Genetically Distinct Populations of Bobtail Squid on O'ahu . Kimbell et at. 349
1989). Squid tissue was frozen in liquid ni DNA, 0.4 !J,M of each primer, 5 ~l of 8 mM trogen and stored at -80°C until used in dNTPs, and 0.12 ~l ofTaq polymerase (Per analyses. Mantle tissue, because it is of low kin Elmer, Norwalk, California). Amplifica yield, and light organs, because they contain tions were performed using a thermocycler symbionts, were removed before homogeni (GeneAmp PCR 9600). The PCR conditions zation. Tissue samples (0.5 g) were placed on were as follows: 35 cycles of 94°C for 30 sec, ice and homogenized in an equal volume of 47°C for 30 sec, and noc for 60 sec. The distilled water. The homogenates were cen primers used to amplify COl were CI-J-I718 trifuged for 5 min at 12,000x g at 4°C to (26 mer, 5'-GGAGGATTTGGAAATT pellet tissue debris. Supernatant fluids were GATTAGTTCC-3') and CI-N-2191 (alias placed in microcentrifuge tubes on ice until "Nancy" 26 mer, 5'-CCCGGTAAAAT use. Allozyme electrophoresis was run on cel T AAAATATAAACTTC-3') (primer desig lulose acetate gels in electrophoresis cham nations given in Simon et al. [1994]). The bers (Zip-zone, Helena Laboratories Inc., resulting PCR products were purified with a Beaumont, Texas). The histochemical stain GeneClean Kit (Bio 101 Inc., La Jolla, Cali ing recipes used were given in Hebert and fornia). Seven microliters of the resultant Beaton (1989). Two buffer systems resulted in DNA and 4 ~l of 1 !J,M CI-J-1718 primer the best resolution: "tris-glycine"-0.4 M were used for cycle sequencing using the Tris with 2 M glycine, pH 8.5, with gellbuffer PRISM Ready Reaction Terminator Cycle dilution 1: 9; and "tris-maleate"-0.1 M tris Sequencing Kit with an ABI Model No. 377 maleate, pH 7.8. Of the 18 presumed en automated sequencer (PE Applied Biosys zymes assayed 7 yielded acceptable enzyme terns, Foster City, California). A total of 18 activity and clear resolution: isocitrate de COl sequences was obtained from 10 indi hydrogenase (IDH; E.c. 1.1.1.42), malate viduals from Kane'ohe Bay and 8 individuals dehydrogenase (MDH; E.C. 1.1.1.37), malate from Niu Valley. All DNA sequences were dehydrogenase NADP+ (ME; E.C. 1.1.1.40), examined and aligned using Sequencher 3.1 glucose-6-phosphate isomerase (GPI; E.C. (GeneCodes, Ann Arbor, Michigan). 5.3.1.9), mannose-6-phosphate isomerase (MPI; E.C. 5.3.1.8), phosphoglucomutase Phylogeog;raphic Analysis (PGM; E.C. 5.4.2.2), and aspartate amino transferase (AAT; E.C. 2.6.1.1). The tris The evolutionary history of the alleles at a glycine buffer was used for PGM and ME, locus can be represented in a phylogenetic and the remaining loci were run with tris tree, which can then be interpreted in light of maleate. Horizontal gels were run at 180 V the geographical distribution of those alleles; and 3 rnA for 45 min at 4°C. Approximately this approach is termed phylogeography (Avise 30 alleles were scored at each locus for 20 2000). The interpretation of such an analysis individuals from each locality. can be complicated for nuclear loci because of recombination (Begun and Aquadro 1992, Mitochondrial DNA: Cytochrome Oxidase I Hudson 1994) and for both nuclear and mi tochondrialloci because of the persistence of Squid DNA was extracted from frozen testis ancestral alleles (Templeton and Sing 1993). tissue (25 mg) with a Qiagen QIAamp Tissue There is no evidence that mitochondrial -----Rir-tc-at:-n-o-:-z9t(4):-1'lre-rsuhrt-eLl-g-enomlc---sequences--m-dUs species recombine, yet DNA was reprecipitated with 95% ethanol, persistence of ancestral alleles is an issue. washed with 70% ethanol, and resuspended Templeton et al. (1992) and Templeton and in TE buffer (10 mM Tris-HCl and 1 mM Sing (1993) provided a method for examining EDTA, pH 8.0). A polymerase chain reaction the recent history of alleles at a locus. The (PCR) was used to produce DNA templates method focuses phylogenetic reconstruction for the sequencing of the mitochondrial cy- on the most recent substitutions before at tochrome oxidase I (COl) gene. Each 50-~l tempting to resolve deeper nodes. Assuming PCR reaction contained 4 ~g of template that more recent changes are less likely to in- 350 PACIFIC SCIENCE· July 2002
dude homoplasy or recombination, sequences plications of differing egg and female body separated by one substitution are linked first, sizes among these populations have not yet the resulting groups form the basis for linking been examined. alleles separated by two substitutions, and The seven enzyme stains resolved nine so on. Details of the method are provided enzyme-coding loci (lDH and AAT each by Templeton and Sing (1993), Templeton stained for two loci). AlI loci examined were (1998), Davies et al. (1999), and Posada and monomorphic. Though surprising, this result Crandall (2001). is not unexpected for squid-other intra! interspecific studies ofsquid population struc Population Analysis ture have yielded negligible allozyme poly morphism and in some cases monomorphism Population structure was determined through (Brierley et al. 1993, 1995, lzuka et al. 1996). the analysis of molecular variation (AMOVA) However, other allozyme studies have dis method of Excoffier et al. (1992) as imple tinguished among closely related squid spe mented by the computer program Arlequin cies (lzuka et al. 1994) and seasonal cohorts 2.0 (Schneider et al. 2000). Indices of popu (Kang et al. 1996). Although allozyme data lation subdivision analogous to standard F collected here contribute no information con statistics (Wright 1951,1965) were calculated cerning the hypothesis ofdistinct populations, with significance assessed using permutation they do raise a question for future work: to procedures that randomly redistribute indi what extent does selection contribute to the viduals among populations. This general ap lack of variation in squid electrophoretic proach was used to estimate F statistics based proteins? upon (1) allele frequencies only, and (2) the Partial Cal sequences of approximately genetic distance among alleles as well as their 430 base pairs from both Niu Valley and frequencies. The F statistics used were (1) () Kane'ohe Bay animals were obtained. Ten (Weir and Cockerham 1984, see Excoffier et sequences from Kane'ohe Bay and eight from al. 1992), and (2) <1>, calculated using the av Niu Valley were resolved. Measures of gene erage Kimura 2-parameter distance observed diversity, mean number of pairwise differ between alleles. Exact tests were performed ences, and base composition were very similar based on the distribution among populations in the two populations (Table 1). Our results of (1) alleles and (2) phylogenetically derived yielded six distinct haplotypes distinguished allele groups (TFPGA, Miller 1998). In the by seven polymorphic sites, with only one latter case, alleles in the same one-step dade haplotype being shared among the two pop of the Templeton networks were pooled. ulations (Figure 3). Statistical analysis revealed significant pop ulation structure between the populations. RESULTS Both frequency- and distance-based F statis Significant differences were found among tics were higWy significant, with the distance populations for both morphological and ge- based estimates, (0.60), being considerably netic characters. Niu Valley females were larger than the frequency-based () (0.49) (Ta higWy significantly larger and produced higWy ble 2), illustrating that the sequence data do significantly larger eggs and hatchlings than provide additional information with regard to ----Rarre'uhe-B-ay-femates-tpigure-2-)-;-suppurtirrg----population-structuringin-this-species~---- Singley's (1983) earlier observations. Al- The values ofboth F statistics are extremely though phenotypic differences between the high, especially considering the small distance two populations do not provide direct evi- between populations. These data, coupled dence of genetic, or inherited, differences with the observation of only one shared hap between the two populations, such data are lotype between the populations (frequency of certainly consistent with this hypothesis of approximately 10% in each population), sug genetic differentiation. The life history im- gest that gene flow is minimal between these Two Genetically Distinct Populations of Bobtail Squid on O'ahu . Kimbell et al. 351
Kaneohe Bay D Niu Valley 3.5 35 -E 3.0 -E 30 E E 2.5 25 -CD -CD .~ ,~ tJ) 2.0 tJ) 20 .! .! .-c: as CD 1.5 E 15 CD ~> .:!::: '-'" 1.0 -~ 10 en "C wen « 0.5 5
0.0 0 Egg Hatchling Female diameter mantle mantle FIGURE 2. Morphological differences among individuals collected in Kane'ohe Bay and Niu Valley locations. Sample sizes were for eggs, hatchlings, and females (78, 62, and 30, respectively) for each population. All comparisons were highly statistically significant (t-tests, P < 0.001). Female age was uncertain; females were measured the day after dying in the laboratory.
two populations. Using the standard migra- DISCUSSION tion estimator for mitochondrial data (Slatkin 1994), migration between these two pop- The analysis of both morphological and ge ulations is calculated as approximately one netic data indicates that populations of E. migrant exchanged between the populations seolopes on O'ahu are extremely subdivided, only every two to three generations, with a likely exchanging very few, if any, migrants, generation time of 3 to 4 months. This effec- each generation. The extent to which other tive rate of migration is very low, yet it is still populations of E. seolopes are so subdivided in _. --sufficient-to-prevent-complete-divergence-be~---Hawai!.i-is-clearly-important-and-weH-worth--- tween the populations by genetic drift alone. investigating. Kane'ohe Bay is surrounded by It should be noted that this estimate of gene the only true barrier reef in the Hawaiian Is- flow should be considered an upper estimate- lands (Kay and Palumbi 1987), and it may be the two populations may also show some ge- that the well-studied Kane'ohe Bay popu netic similarity, if they share a recent history lation is an anomaly. Yet, water does flow (see Bohonak 1999, Bohonak and Roderick actively in and out of the bay. 2001). Subdivision of the E. seolopes populations 352 PACIFIC SCIENCE· July 2002
Kaneohe Bay ditions, one may be able to pinpoint those traits (likely to be cell surface determinants o NiuValley on the host or symbiont) that change as the partners coevolve and become, as a unit, ge netically distinct from their congeneric squid bacteria counterparts. For example, symbionts ofJapanese squid E. morsei infect Hawaiian E. seolopes easily by themselves, but quickly lose in competition with the native Hawaiian strain (Nishiguchi et al. 1998). This compet itive difference was easily determined because E. morsei strains of V fiseheri are brightly lu minous in culture, whereas E. seolopes strains are dim in culture. Thus, when symbiotic FIGURE 3. Mitochondrial haplotype network and local ities as estimated by the method described by Templeton organs of such competition experiments are (1998). Haplotype identities are noted within circles with plated on agar, the ratio of the two strains is a letter, and distances between haplotypes are propor easily determined. Unfortunately, all isolated tional to the number of base pair differences (one or two strains of V fiseheri from Hawai'i are dim in steps) between haplotypes. The size of the circles is pro culture and, therefore, to compete the sym portional to the number of individuals (n = 18) sharing a particular haplotype. bionts of Hawaiian populations of E. seolopes would require genetic engineering of V fiseheri strains to introduce distinguishing has implications beyond the species. If the genetic markers. level of subdivision observed here is found The extent of population structure is also elsewhere, one can also ask, at what level of important in determining whether host isola divergence does one see a concomitant change tion occurs before symbiont isolation or vice in the symbiont Vibrio fiseheri? At the mecha versa, or whether host and symbiont track nistic level, under these very refined con- each other closely. The numbers of V fiseheri
TABLE 1 Within- and Between-Population Comparisons ofMitochondrial COl DNA Sequences
Locality
Kane'ohe Niu Valley Bay (n = 10) (n = 8)
Haplotype Data Frequency Frequency 101
A 0.700 0.000 G B 0.100 0.125 C 0.100 0.000 D 0.100 0.000 ----E------o~OOO---O;iSO------A F 0.000 0.125 Gene diversity 0.53 ± 0.18 0.46 ± 0.20 Mean no. pairwise differences 1.32 ± 0.89 0.93 ± 0.71 Nucleotide composition (%) C 20.7 20.4 T 35.8 36.0 A 27.4 27.3 G 16.1 16.2
a Differences in base positions are shown relative to a reference haplotype (A), with identities shown by".". Two Genetically Distinct Populations of Bobtail Squid on O'abu . Kimbell et al. 353
TABLE 2 Wallace's line? Nature (Lond.) 406:692 F Statistics and Exact Tests for Population 693. Differentiation Begun, D.]., and C. F. Aquadro. 1992. Levels ofnaturally occurring DNA polymorphism Method F p correlate with recombination rates in D. melanogaster. Nature (Lond.) 356:519-520. Frequency (0) .492 <0.00001 Bohonak, A. J. 1999. Dispersal, gene flow, Distance (<1» .604 <0.00001 and population structure. Q. Rev. BioI. Note: Statistics are presented that are based solely on the fre 74:21-45. quency of alleles (IJ) and that also consider the genetic distance among alleles (<1». Bohonak, A. ]., and G. K Roderick. 2001. Dispersal of invertebrates among tempo rary ponds: Are genetic estimates accurate? in the water are strongly dependent on the 1sr. J. Zoo1. (in press). density ofhost squid in the environment (Lee Brierley, A. S., P. G. Rodhouse,J. P. Thorpe, and Ruby 1994). Thus, at the mouth of Ka and M. R. Clarke. 1993. Genetic evidence ne'ohe Bay, where there are few hosts, there of population heterogeneity and cryptic are very few V. fiseheri. There may also be little speciation in the ommastrephid squid Mar genetic exchange between symbiont popula tialia hyadesi from the Patagonian Shelf tions. and Antarctic Polar Frontal Zone. Mar. In sum, populations of E. sealapes on two BioI. (Ber1.) 116:593-602. sides of O'ahu are more distinct than pre Brierley, A. S., J. P. Thorpe, G. J. Pierce, M. viously realized, suggesting that gene flow R. Clarke, and P. R. Boyle. 1995. Genetic may be substantially limited in some marine variation in the neritic squid Laligo fobesi species in Hawai'i. This result likely reflects (Myopsida: Loliginidae) in the northeast the lack of long-lived pelagic larvae. Further, Atlantic Ocean. Mar. BioI. (Ber1.) 122:79 the Hawai'i populations ofE. sealopes and their 86. symbionts offer some interesting opportunities Davies, N., F. X. Villablanca, and G. K Ro to gain insight into the coevolution of bacte derick. 1999. Bioinvasions of the medfly, ria and their animal hosts. Evidence that the Ceratitis eapitata: Source estimation using host populations can be extremely subdivided DNA sequences at multiple intron loci. is an intriguing start. Genetics 153:351-360. Excoffier, L., P. Smouse, and J. Quattro. 1992. Analysis of molecular variance in ACKNOWLEDGMENTS ferred from metric distances among DNA haplotypes: Application to human mito We thank two anonymous reviewers for chondrial DNA restriction data. Genetics helpful and insightful comments. 131:479-491. Gillespie, R. G., F. G. Howarth, and G. K Literature Cited Roderick. 2001. Adaptive radiation. Pages 25-44 in S. A. Levin, ed. Encyclopedia of Arnold, ]., C. Singley, and L. Williams biodiversity. Academic Press, New York. Arnold. 1972. Embryonic development Hebert, P. D. N., and M. J. Beaton. 1989. ------ana post-natcl:iiD.g survival ottneSepiol1d MethOdOlogies for allozyme analysis using squid Euprymna seolopes under laboratory cellulose acetate electrophoresis. Helena conditions. Veliger 14:361-364. Laboratories, Beaumont, Texas. Avise, J. c. 2000. Phylogeography: The his Hudson, R. R. 1994. How can the low levels tory and formation of species. Harvard of DNA sequence variation in regions of University Press, Cambridge, Massachu the Drosophila genome with low recom setts. bination rates be explained. Proc. Natl. Barber, P. H., S. R. Palumbi, M. V. Erd Acad. Sci. U.S.A. 91:6815-6818. mann, and M. K Moosa. 2000. A marine 1zuka, T., S. Segawa, T. Okutani, and K-I. 354 PACIFIC SCIENCE· July 2002
Numachi. 1994. Evidence on the existence Speciation and phylogeography of Ha of three species in the oval squid Sepioteu waiian terrestrial arthropods. Mol. Ecol. this lessoniana complex in Ishigaki Island, 7:519-531. Okinawa, southwestern Japan, by isozyme Scheltema, R. S. 1986. Long-distance dis analyses. Venus Jpn. J. Malacol. 53 :217 persal by planktonic larvae of shoal-water 228. benthic invertebrates among central Pa Izuka, T., S. Segawa, and T. Okutani. 1996. cific islands. Bull. Mar. Sci. 39:241-256. Biochemical study of the population het Schneider, S., D. Roessli, and L. Excoffier. erogeneity and distribution of the oval 2000. Arlequin: A software for population squid Sepioteuthis lessoniana complex in genetic data analysis. v. 2.00. University southwestern Japan. Am. Malacol. Bull. of Geneva (distributed on web: http:// 12:129-135. anthropologie.unige.chlarlequin,). Kang, Y-]., Y-H. Kim, Y.-K Hong, J.-Y. Simon, C. 1987. Hawaiian evolutionary biol Park, and K-Y Park. 1996. A population ogy: An introduction. Trends Ecol. Evol. genetic analysis of the common squid, 2:175-178. Todarodes pacificus Steenstrup in the Ko Simon, c., F. Fratti, A. Beckenbach, B. rean waters. J. Korean Fish. Soc. 29:320 Crespi, H. Liu, and P. Flook. 1994. Evo 331. lution, weighting, and phylogentic utility Kay, A. E., and S. R. Palumbi. 1987. En of mitochondrial gene sequences and a demism and evolution in Hawaiian marine compilation ofconserved polymerase chain invertebrates. Trends Ecol. Evo!. 2: 183 reaction primers. Ann. Entomol. Soc. Am. 186. 87:651-701. Lee, K-H., and E. G. Ruby. 1994. Effect of Singley, C. T. 1983. Euprymna scolopes. Pages the squid host on the abundance and dis 69-74 in P. R. Boyle, ed. Cephalopod life tribution of symbiotic Vibrio fischeri in na cycles. Academic Press, London. ture. Appl. Environ. Microbiol. 60:1565 Slatkin, M. 1994. Gene flow and population 1571. structure. Pages 3-17 in L. A. Real, ed. McFall-Ngai, M. ]., and E. G. Ruby. 1998. Ecological genetics. Princeton University Sepiolids and vibrios: When first they Press, Princeton, New Jersey. meet. BioScience 48:257-265. Templeton, A. R. 1998. Nested clade analysis Miller, M. P. 1998. Tools for population ge of phylogeographic data: Testing hypoth netic analysis. v. 1.3. Northern Arizona eses about gene flow and population his State University, Flagstaff. tory. Mol. Ecol. 7:381-397. Nishiguchi, M. K, E. G. Ruby, and M. J. Templeton, A. R., and C. F. Sing. 1993. A McFall-Ngai. 1998. Competitive domi cladistic analysis ofphenotypic associations nance among strains of luminous bacteria with haplotypes inferred from restriction provides an unusual form of evidence for endonuclease mapping. IV. Nested analy parallel evolution in the sepiolid squid ses with cladogram uncertainty and re vibrio symbiosis. Appl. Environ. Micro combination. Genetics 134:659-669. bioI. 64:3209-3213. Templeton, A. R., K A. Crandall, and C. F. Palumbi, S. R. 1992. Marine speciation on a Sing. 1992. A cladistic analysis of pheno small planet. Trends Ecol. Evo!. 7:114 ~ic associations with haplotypes inferred 118. fromrestriction enClonuclease mappmg and- Posada, D., and K A. Crandall. 2001. Intra DNA sequence data. III. Cladogram esti specific gene genealogies: Trees grafting mation. Genetics 132:619-633. into networks. Trends Ecol. Evol. 16:37 Thacker, R. W., and M. G. Hadfield. 2000. 45. Mitochondrial phylogeny of extant Ha Richardson, B. J., P. R. Baverstock, and M. waiian tree snails (Achatinellinae). Mol. Adams. 1986. Allozyme electrophoresis. Phylogenet. Evol. 16:263-270. Academic Press, New York. Wagner, W. L., and V. A. Funk, eds. 1995. Roderick, G. K, and R. G. Gillespie. 1998. Hawaiian biogeography: Evolution on a Two Genetically Distinct Populations of Bobtail Squid on O'ahu . Kimbell et al. 355
hot spot archipelago. Smithsonian Institu Wright, S. 1951. The genetical structure of tion Press, Washington. populations. Ann. Eugen. 15:323-354. Weir, B. S., and C. C. Cockerham. 1984. ---. 1965. The interpretation of popula Estimating F-statistics for the analysis of tion structure by F-statistics with special population structure. Evolution 38:1358 regard to systems of mating. Evolution 1370. 19:395-420.