You say Conchaphila, I say Lurida: molecular evidence for restricting the Olympia oyster (Ostrea lurida Carpenter 1864) to temperate western North America.ABSTRACT The western North American bivalve mollusc known as the Olympia oyster, long known as Ostrea lurida Carpenter 1864 ([dagger]), is a historically exploited native species that has been largely displaced by larger nonnative oysters. There is much renewed interest in documenting and restoring its native populations and recent successful culturing has attracted a specialty market for these oysters. Yet its name was called into question when it was synonymized with O. conchaphila Carpenter 1857, an oyster whose type locality is Mazatlan, Sinaloa, Mexico. Others have considered it more plausible that the Olympia oyster is a more northern species, distinct from O. conchaphila, but morphological or molecular evidence either way has been lacking. Here we used a molecular approach to test the single versus two-species hypotheses with samples from Sinaloa, Mexico, near the type locality of O. conchaphila (Mazatlan, Mexico), and samples from Willapa Bay, WA, the type locality of O. lurida, as well as samples from intermediate locations. Based on our combined and separate analyses of two mitochondrial DNA (mt DNA) markers, 16S ribosomal RNA (16S) and cytochrome oxidase III (CO3), native Ostrea from Sinaloa, Mexico are reciprocally monophyletic with a clade from multiple other localities between Baja California, Mexico and British Columbia, Canada, including Willapa Bay, WA. Corrected pairwise sequence comparisons for 16S indicate these two groups last shared a common ancestor 1.5-3.9 mya (2.06% sequence divergence). Based on these results and assuming that the Sinaloa group represents the true O. conchaphila, molecular evidence supports O. conchaphila and O. lurida as separate species. Posthoc morphological comparisons uncovered no significant support for morphological distinction between the two taxa, underscoring the difficulty associated with using morphology alone to distinguish closely related oyster species. Despite the present lack of any morphological diagnostic differences for separating these nominal species, the molecular data are not consistent with the synonymy of the species and support the reinstatement of O. lurida from all the localities north of central Baja California.
KEY WORDS: Olympia oyster, Ostrea conchaphila, Ostrea lurida, Ostreidae, systematics, phylogeny, cryptic species
Many marine taxa are known to exhibit a high degree of phenotypic plasticity that challenges systematic biologists and creates problems for conservation efforts. Knowlton (1993) described the abundance of cryptic marine species, which are characterized as being almost impossible to distinguish based on morphology alone. Species definitions have been the subject of many debates (Donoghue 1985) and, in recent decades, even the most popular species concepts have been criticized, such as the biological species concept (Mishler & Donoghue 1982, Donoghue 1985, Cracraft 1987, Kluge 1990) and the phylogenetic species concept (Avise 2004). Regardless of one's species definition, the use of molecular markers has enabled scientists to identify previously cryptic species and to resolve taxonomic inconsistencies (Knowlton 1993, Simison & Lindberg 1999, Dawson & Jacobs 2001, Crummett & Eernisse 2007, Shilts et al. 2007, Hewson 2008).
Recent efforts to restore populations of the Olympia oyster ([double dagger]), Ostrea conchaphila, some involving transplants of oysters between estuaries separated by large distances, have been conducted despite lingering uncertainty about the oyster's true identity. The taxonomic problem is whether Ostrea lurida Carpenter 1864, the name once widely used for the Olympia oyster, should be considered a junior synonym of O. conchaphila Carpenter 1857. This synonym was proposed by Harry (1985) without explanation but is presumed to be based on morphological examination of museum specimens combined with the notion that the species was more widely ranging than previously thought. Prior to Harry's study, O. lurida and O. conchaphila were thought to be separate species, with the range of O. conchaphila from Mazatlan, Mexico to Panama (Dall 1914) and the range of O. lurida from Sitka, AK, USA to Cabo San Lucas, Baja California Sur, Mexico. Neither species was thought to exist in the contiguous coastline between the two species' ranges, including all of the Gulf of California, Mexico (Dall 1914). An exception (although not addressing the Gulf of California) was Hertlein's (1959) discussion of a possible transition zone between these species based on intermediate-appearing oysters from San Diego, CA and northern Baja California, Mexico, which would necessitate a considerable northern extension of the range of O. conchaphila. After Harry (1985) proposed this was all a single species, some (e.g., Baker 1995) have questioned the synonymy whereas others have indiscriminately used one name or the other, mostly in reference to oysters from Washington or British Columbia.
A separate issue is Harry's (1985) revival of Ostreola Monterosato 1884 (type species Ostrea stentina [Payraudeau, 1826]), which he also extended to include O. conchaphila. Subsequent morphological (Coan et al. 2000) and molecular (Kirkendale et al. 2004, Lapegue et al. 2006, Shilts et al. 2007) studies have not supported Ostreola as distinct from Ostrea. In particular, Shilts et al. (2007) have called for the dissolution of Ostreola and reassignment to Ostrea.
Until the present study, no one has tested Hertlein's (1959) predicted zone of overlap in southern California and along the Baja California peninsula, and this information would be especially useful considering ongoing native oyster restoration efforts. For example, government agencies such as NOAA and nongovernmental organizations such as The Nature Conservancy are interested in expanding restoration projects to locations south of San Francisco Bay, CA (Dick Vander Schaaf, The Nature Conservancy, pers. comm.) to include Hertlein's (1959) zone of overlap.
Given the high phenotypic plasticity of these oysters, the use of molecular markers applied in a geographic context could potentially be useful for resolving whether the native oyster is one or more species. Parallel studies with oysters in different ocean basins (Lam & Morton 2006, Lapegue et al. 2006, Shilts et al. 2007, Wang & Guo 2008) have already demonstrated the usefulness of molecular markers. Here, for the first time, we have tested Harry's (1985) synonymy of the two species with DNA sequence comparisons, followed by morphological comparisons that were posthoc with respect to groupings identified by our molecular analyses. We analyzed two mitochondrial DNA (mtDNA) markers, 16S ribosomal DNA (16S) and cytochrome oxidase subunit III (CO3), for oysters collected at multiple localities between Willapa Bay, WA (the type locality of O. lurida) and Mazatlan, Mexico (the type locality for O. conchaphila). Our sampling sites also included seven localities between southern California, USA and the Pacific coast of Baja California, Mexico to test for the presence of Hertlein's (1959) predicted zone of overlap.
MATERIAL AND METHODS
All samples collected whose sequences were used in the analysis, along with localities and sample numbers, are listed in Table 1. Samples from Willapa Bay, WA were collected and identified as the Olympia oyster by Russell Rogers (WA Department of Fish and Wildlife) and were given to us as soft tissue only. One ingroup sequence O. "conchaphila" from Barkley Sound, B.C., Canada, and all outgroups used in the analyses were downloaded from GenBank (Table 2). Some of the names listed in Table 2 reflect recent taxonomic changes, such as favoring Ostrea over Ostreola for those species commonly assigned to the latter (see Introduction). When collected in the field, all oysters were opened and then immediately preserved in 95% ethanol. We randomly selected a subset of samples collected from each site for analysis, and these were each given unique voucher labels. Because samples from Willapa Bay, WA were soft tissue only, the entire sample was extracted.
DNA Extraction and "Touchdown" Polymerase Chain Reaction (PCR)
For the DNA extraction and purification, we used DNeasy tissue kits from Qiagen (Valencia, CA) and followed the manufacturers' recommendations. For the tissue, approximately 25 mg of the adductor muscle was dissected and digested in a Proteinase-K and lysis buffer bath overnight at 56[degrees]C. Double-stranded products of partial 16S were amplified via PCR with universal animal primers (Palumbi 1996): 16Sar: 5'--CGC CTG TTT ATC AAA AAC AT--3', 16Sbr: 5'--GCC GGT CTG AAC TCA GAT CAC GT--3'. Double-stranded products of CO3 were amplified via PCR with primers modified for oysters (P. Baker unpublished): CO3F: 5'--AAA AGT TCA AAG CGG TCT TA--3', CO3R: 5'--AGC TAA CAT ACG AAC AAG GC--3'.
A HotStarTaq kit from Qiagen was used to prepare 50 [micro]L reactions for amplification of 16S and CO3. Each reaction included 4.0 [micro]L of 25 mM Mg[Cl.sub.2] (CO3 only), 5 [micro]L of 10 x buffer, 5 [micro]L of dNTPs (10 mM), 2 [micro]L of primer (1 [micro]L forward/1 [micro]L reverse, 10 [micro]M), 1 [micro]L genomic DNA template, 0.25 [micro]L HotStarTaq polymerase, and deionized sterile water to bring the reaction volume to 50 [micro]L per reaction. Reactions were carried out via a GeneAmp PCR System 2400 (Perkin-Elmer) thermocycler. HotStarTaq activation and initial denaturation temperatures were 95[degrees]C for 15 min (a step specific to the HotStarTaq) and then 94[degrees]C for 1 min thereafter (denaturation only). For 16S, initial annealing temperatures of 53[degrees]C for eight cycles of 1 min each (decreased by 1[degrees]C each cycle) were used; final annealing temperatures of 48[degrees]C for 25 cycles of 1 min each were used and an extension temperature of 72[degrees]C was used for 1 min per cycle and for the final extension period of 10 min, followed by a hold at 10[degrees]C until samples were removed to storage at 5[degrees]C. For CO3, initial/final annealing temperatures of 50/45[degrees]C for 30 cycles of 1 min each were used and an extension temperature of 72[degrees]C was used for 1 min per cycle and for the final extension period of 10 min. PCR product was visualized with agarose-gel electrophoresis (1.5%), ethidium bromide staining, and UV illumination to confirm successful amplification of the mtDNA region of interest.
DNA Sequencing Alignment and Analyses
PCR product cleanup and sequencing were performed commercially at the Duke University Institute for Genome Sciences and Policy DNA sequencing facility (Duke IGSP, Durham, NC). Each sequence contig was assembled from a pair of sequences from opposite strands using CodonCode Aligner software (CodonCode Corp, Dedham, MA). All new sequences were submitted to GenBank and have been assigned the accession numbers FJ768501-FJ768589 (16S) and FJ768590-FJ768673 (CO3).
All sequences were aligned by eye with the use of MANIA software (D. L. Swofford & D. J. Eernisse, unpublished). There was very little sequence length variation, and so alignment was generally not problematic. The DNA sequence alignments of 16S and CO3 were 502 and 417 sites, respectively. Heuristic searches were performed for the combined, and separate, data set, or sets, using maximum parsimony (MP) and maximum likelihood (ML) optimality criteria. Altogether, 161 specimens were in our combined data set, including 111 for 16S and 86 for CO3. 16S sampling emphasized the contrast between Ostrea conchaphila and O. lurida and tested for Hertlein's (1959) proposed region of overlap whereas CO3 sequencing emphasized tests of phylogeographic structure within O. lurida. The larger 16S data set included 22 sequences downloaded from GenBank, whereas only two of these (Crassostrea virginica and C. gigas) were available for inclusion in our CO3 data set. The other 84 CO3 sequences were either Ostrea "lurida" (n = 24 from 8 widely separated localities) or O. conchaphila (n = 7 from 2 localities in close proximity). To combine the two data sets, 50 ingroup specimens with CO3 sequences but without 16S sequences were deleted from the analysis (i.e., the combined analysis included the same 111 taxa as in the 16S-only analysis). This avoided confounding tree searching by having a "missing data" set of taxa from one data set that shared no characters with a similar set from the other.
For the MP criterion, we used the PAUP* 4.0b10 software package (Swofford 2003). All gaps were treated as missing data to avoid treatment of adjacent gaps as separate characters. The two-part search strategy used by Kelly and Eernisse (2008) was used to improve the efficiency of searching, specifically to avoid having searches become unnecessarily trapped in nonoptimal tree islands. The first part involved 1,000 replicate searches with only 10 trees held per replicate, keeping all minimum length trees found. Then the second part involved swapping on all minimum-length trees found with the trees-held restriction removed, starting with trees already in memory and with maxtrees set to 1,000. Branch robustness was estimated by 1,000 bootstrap replicate searches, each with 10 replicate random stepwise addition sequence heuristic searches.
For the ML criterion, we used genetic algorithm for rapid likelihood inference (GARLI) v. 0.951 (Zwickl 2006). Because this program searches for trees with a fast and accurate genetic algorithm whose speed is not substantially influenced by the number of taxa, it can handle much larger data sets than are conventionally analyzed with the ML criterion. The version of GARLI used has automatic model selection, which is optimized for the particular data set during analysis. (A newly released v. 0.96 of the program adds additional model assignment options that were not previously available.) Ten separate GARLI searches were performed and the tree selected was the one with the most optimal (closest to zero) final likelihood score. Branch robustness was estimated with 100 bootstrap replicate searches using GARLI. To check whether these results agreed with a more conventional ML search, we performed such a search with PAUP* with the number of taxa greatly reduced to 29 taxa to allow the feasible completion of searches. For this analysis, we used FINDMODEL (Tao et al. 2005) to identify an appropriate model for each data set. Based on this selection, we performed a heuristic search for the combined data set using the general time-reversible plus gamma (GTR + [GAMMA]) model with optimality criterion set to minimum evolution and tree-bisection-reconnection (TBR) branch swapping algorithm.
Corrected pairwise differences within and between species were calculated using Arlequin version 3.11 (Excoffier et al. 2005) with 1,000 permutations, transitions and transversions equally weighted, and acceptable level of missing data set to 5 percent. Corrected pairwise differences were used to estimate intra and interspecific percent sequence divergence by dividing the mean pairwise differences between species by the number of base pairs used in the pairwise analysis.
To verify the identities of the samples collected from and around Mazatlan, several morphological characters previously identified as potentially diagnostic within the Ostreidae (Keen 1958, Torigoe 1981, Harry 1985, Castillo Rodriguez & Garcia-Cubas 1986, Quayle 1988, Coan et al. 2000) were considered. The morphologies of the samples from Mexico were examined in a "blind" manner, with no a priori knowledge of the molecular outcomes. Characters were examined in selected specimens from mainland Mexico from which DNA was extracted (Table 3). Post-hoe examinations of nine specimens (Table 3) from Hertlein's (1959) zone of overlap (Southern California, USA and Baja California, Mexico) were also completed and compared with the blind-examined specimens in an effort to validate findings from the molecular analyses. Also examined was the shell of a syntype specimen of Ostrea lurida from the Redpath Museum at McGill University, Montreal, Canada. The specimen, RMM 125, is presumed to be part of the type lot described by Philip Carpenter. The only reason that we have avoided designating this specimen the lectotype for O. lurida is our lack of knowledge of other syntype specimens. The syntypes of Ostrea conchaphila are housed at The Natural History Museum in London, England. Images of one of these specimens selected by Kathie Way, as a candidate for lectotype designation, were obtained, but too late to include in this manuscript.
Shell and soft-tissue characters used in the morphological examinations are described individually later. For illustrations of some of these characters see Harry (1985).
Chomata Sweep and Chomata Shape
Chomata are small (often under 1 mm), repeating, tubercule or denticle-like shell features on the internal shell margin on either side of the hinge. Pits (catachomata) lie in the left, or attached valve, and matching denticles (anachomata) lie in the right valve (Torigoe 1981). Chomata shape may be rounded (less than twice as long as wide) or elongate (more than twice as long as wide).
Chomata extend (sweep) in a series from the hinge area around the margin towards the ventral side. Chomata can be a useful character to distinguish the genus Crassostrea from Ostrea, as Crassostrea are believed to lack this character (Harry 1985). In some taxa (e.g., Saccostrea) the chomata extend all the way to the ventral margin (100% sweep), but, in other taxa, the chomata extend only a fraction of that distance.
Plicae (plications) are matching folds in both shell valves that are expressed in a zigzag manner, particularly near the ventral shell margin, so that the valves close precisely. Individual plica-like folds can be mimicked by the shell conforming to substratum topology.
Shell color, both internal and external, is sometimes listed as a character for oysters, especially for Crassostrea (e.g., Keen 1958, Quayle 1988, Coan et al. 2000), where the color is typically described as chalky white, but these same sources admit its high variability. Furthermore, shell color can be altered or eliminated by weathering and preservation techniques (Patrick Baker, pers. obs.). Descriptions of Ostrea conchaphila (and lurida) often note an olive-green interior color (e.g., Keen & McLean 1971, Hertlein 1959).
Other Shell Characters
External shell sculpture, in the form of foliations or spines, has been described for some taxa (Torigoe 1981). Secondary mantle retractors occur in the genus Saccostrea and appear on the shell as a series of small oval muscle scars paralleling the pallial line (Torigoe 1981).
Soft Tissue Characters
Torigoe (1981) and Harry (1985) used mantle tentacles and the morphology of anal papillae as diagnostic characters although Coan et al. (2000) questioned their value and, likewise, Baker (pers. obs.) observed that preservation techniques might change their appearance. Therefore, characters were reduced to polymorphism of tentacles on the inner two mantle folds (those with the most distinct tentacles), and the presence or absence of an external anal papilla.
Trees that were optimal for the maximum parsimony (MP) and maximum likelihood (ML) optimality criteria, for the combined and separate 16S and CO3 datasets, all resulted in similar topologies: a northern group (O. "lurida" from British Columbia to Baja California) and a southern group (O. "conchaphila" from EdPn and EdPs in Sinaloa, combined as EdP hereafter) are reciprocally monophyletic with respect to all other included taxa. Variation among the first 1,000 MP trees found is adequately represented by one of these and was selected arbitrarily (Fig. 1). The sampling scheme for 16S (n = 111) and CO3 (n = 84) data sets differed considerably in the localities sampled (see Methods and Table 1). Only 36 of the 86 taxa with CO3 data were included in our combined analysis of 111 total taxa (Fig. 1 and Fig. 2). Analysis of the separate CO3 data set with either only these 36 included taxa or else with the complete 86-taxon CO3 data set produced results entirely consistent with analysis of the 16S data set and the combined set. For example, for our MP analysis, the nodes for each focal species or for both sister species together in the CO3-only data set are supported with high (>95%) bootstrap support. These results agree with the 16S-only and combined sets, though the latter produced somewhat lower bootstrap support (e.g., Fig. 1). The MP and ML bootstrap support values mostly agreed with each other, but with somewhat lower support for nodes with ML. Results of a PAUP * search using the ML criterion, with a necessarily reduced set of taxa, were consistent with those from the more extensive GARLI analysis (Fig. 2). This implies that the particular topological results emphasized here were independent of the optimality criterion (MP or ML) or ML approach (PAUP* or GARLI) used. The best MP and ML trees did differ somewhat in the topology of the most proximal outgroups for O. conchaphila + O. lurida (Fig. 1 and Fig. 2, respectively). In our best ML tree, a greater proportion of the total branch length is found in the separation of our ingroup and the most proximal outgroups (Ostreinae + Lophinae) from our more distal outgroups from Saccostrea and Crassostrea (both in Crassostreinae). Although Ostreinae is paraphyletic to Lophinae in our best trees (as indicated by the quotation marks around Ostreinae in Fig. 2A), there is no strong bootstrap support for such nesting of Lophinae within Ostreinae, so it is possible that further resolution will support their reciprocal monophyly.
The bootstrap support supporting the monophyly of O. lurida and O. conchaphila are each high for both MP and ML (Figs. 1 and 2B). However, the joining of these as sister species is more equivocal. Our best trees for MP and ML resolves them as sister taxa but with low bootstrap support. The bootstrap proportions supporting the node joining O. conchaphila + O. lurida is higher for MP (69%) than for ML (43%). The only alternative grouping observed for MP and ML with >10% bootstrap support had O. lurida as sister taxon of all other Ostreinae + Lophinae species, including O. conchaphila. This arrangement was supported in 11% of the MP bootstrap replicates and 35% of the ML replicates.
[FIGURE 1 OMITTED]
The average percent sequence divergence between paired combinations of 16S sequences for O. conchaphila versus O. lurida was 2.06%. Samples with 16S only from Mazatlan (Maz) and Bahia de Kino (BdK) did not group with either Ostrea species. All Maz samples and one BdK sample grouped with Saccostrea spp., whereas the other BdK sample grouped with Crassostrea spp. (almost certainly C. gigas, assuming the sequence from GenBank is correctly identified as from this species). Independent morphological examinations only partially supported their assignment to these genera as described later. We used morphology to identify one of the specimens (see later) as S. palmula (Carpenter, 1857), and this was later resolved as Saccostrea with DNA. Because this is the only recognized species of Saccostrea from the Panamic region of the eastern Pacific (P. Valentich-Scott, pets. comm.), we have provisionally assigned all of the specimens resolved with DNA as Saccostrea to this species.
Separate analysis of the complete CO3 data set (as well as especially our MP combined data set analysis) supported a phylogeographic separation within O. lurida. Ignoring autapomorphic substitutions in six of the 84 total ingroup taxa, the entire CO3 data set reduces to the variable sites summarized in Figure 3 and the corresponding haplotype network in Figure 4. In particular, sites for either side of the southern end of Vancouver Island (B.C., Canada) were identical to each other but separated from Willapa Bay (southwestern Washington, USA). All other O. lurida sites from Oregon, CA, and Baja California, Mexico were identical to the Willapa Bay site, ignoring autapomorphies. The southern and northern haplotype groups within O. lurida are equally separated from the much more distant O. conchaphila, with each having an identical nucleotide state in common with O. conchaphila, relative to the other at two sites. Because two highly divergent CO3 sequences for Crassostrea spp. were the only available outgroups for our CO3 data set, these did not provide a useful rooting for inferring whether these alternative matching states are plesiomorphic or derived. Because of this lack of proximal outgroups and because O. conchaphila is relatively far from O. lurida, the rooting depicted in Figure 4 for O. conchaphila midway between the Vancouver Island and more southern haplotype group is arbitrary.
[FIGURE 2 OMITTED]
The phylogeographic contrast discovered for CO3 was not seen in the three 16S sequences obtained from the Vancouver Island populations versus the other 108 more southern sequences. This is probably best explained by the generally slower substitution rate of the 16S gene region.
Independent Morphological Identification of Mainland Mexico Samples
The findings from the independent morphological examinations of samples from Mexico were in agreement with the molecular results in most cases (Table 3). All samples from two adjacent sites within Ensenada del Pabellon, Sinaloa, Mexico were tentatively identified as Ostrea conchaphila based on the presence of a rounded or slightly elongate chomata that extended <50% along the internal shell margin on either side of the hinge (Table 3); this was in agreement with the molecular results. In contrast, one sample from Mazatlan (Maz 0503) was identified as Saccostrea palmula based on the presence of chomata all the way around the shell, a purple interior margin and the absence of anal papilla and secondary mantle retractor scars. The identifications of the other two samples from Mazatlan could not be positively resolved using morphological characters. All three Mazatlan samples grouped with Saccostrea sp. in the molecular results. From Bahia de Kino, one specimen (BdK 0701) was identified as Crassostrea sp. based on the absence of chomata and the thick, chalky shell, and this finding was congruent with the molecular results. Despite the presence of chomata in the other BdK sample examined (BdK0702), this sample was also tentatively identified as Crassostrea sp. based on the thick, chalky shell (Table 3), however, this sample grouped with Saccostrea sp. in the molecular findings.
Post-Hoc Morphological Examination of Samples from Hertlein's Zone of Overlap
Our post-hoc morphological examinations were not able to resolve any consistent differences between the putative northern O. lurida and the southern O. conchaphila (Table 3). For each character there was at least one sample that differed from the mode (e.g., chomata shape, Table 3). Morphological distinctions were also not apparent from the examination of a syntype specimen of Ostrea lurida.
Specific observations from the independent and post-hoc examinations are reported later for each character.
Chomata Sweep and Chomata Shape
Chomata were present in all specimens except BdK 0701 (Table 3). In some specimens, this chomata shape was intermediate and/or variable, designated as round-elongate in Table 3. In none of the specimens examined, however, was the presence or absence of chomata ambiguous, at least under low-power magnification.
Figure 3. Summary of observed CO3 variation by three haplotype groups. Variation was consistent with dividing localities into southernmost, northernmost, and all other localities in between, as listed in the CO3 column of Table 1. Oco-Southern (S) includes both Ostrea conchaphila sites, EdPn and EdPs; Olu-Northern (N) includes both Vancouver Island sites for O. lurida, LH and AI. Olu-All Other (A) includes all other O. lurida localities. Sample sizes for three observed haplotypes: S = 7; N = 7; A = 70. This includes three individuals (two S and one A) that differed from the other normal haplotype patterns by only a single unique base pair difference, but none of this variation occurred at sites a-m. The similarity "Pattern" tallies to 12 for A + N ("1") and two each for A + S and N + S ("2" and "3"). No assumption was made to root the similarity matching patterns because the only available outgroups (Crassostrea spp.) were highly divergent from these three haplotypes. Alignment sites summarized above were as follows, relative to the complete mitochondrial genome for Crassostrea gigas, AF177226: a = 190; b 210; c = 231; d = 258; e = 309; f = 366; g = 395; h = 399; i = 435; j = 453; k = 459; I = 495; m = 501; n = 519; o = 558; and P = 582. Of these, only site 'g' represents a nonsynonymous substitution. CO3Site abcdefghijklmnop Oco-Southern TGTCACATGTTTGTAA Olu-Northern AACTGTGTAGTCACGC Olu-Allother AACTGTACGGCCACGC Pattern 1111112321311111
[FIGURE 4 OMITTED]
Plicae were scored as absent if they did not appear to arise independently of underlying substratum topology. In some specimens, plicae were numerous and regular, but in others they were irregular or few in number (Table 3).
Internal shell color was ambiguous in all samples except for Maz 0503 in which it was deep purple; internal shell color is not included in Table 3. External shell color, where it differed unambiguously from white or off-white, was recorded in Table 3.
Other Shell Characters
External shell sculpture in the form of foliations or spines was not present in any of the specimens examined in this study. Secondary mantle retractor scars were present in specimen Maz 0503, which helped in diagnosing it as Saccostrea. Both samples from BdK had relatively thick shells throughout (>2 mm) and were chalky in texture, which is a common description to Crassostrea shells.
Soft Tissue Characters
For some of the samples, characters such as anal papillae and mantle tentacles could not be defined precisely because of tissue damage, poor preservation, and possible environmental plasticity. The presence/absence of an anal papilla and the relative size of mantle tentacles on the two inner folds were variable within suspect taxa.
The combined and separate analyses of the 16S and CO3 sequence data sets do not support Harry's synonymy of Ostrea lurida with Ostrea conchaphila, but instead support the original classifications by Carpenter (1857, 1864), who treated them as separate oyster species. Our evidence thus provisionally supports the reinstatement of Ostrea lurida as the Olympia oyster. From the analysis of the 16S data set, samples from the type locality of O. lurida, Willapa Bay, WA, along with seven other locations in Southern California and Baja California are unambiguously supported as a monophyletic grouping, within which we noted only minor variation. Likewise, 16S sequences of oysters from two adjacent locations within Ensenada del Pabellon (EdP in Fig. 1), Sinaloa, Mexico, near the type locality of O. conchaphila, group as a well-supported phylogenetic lineage. Relative to the most proximal available outgroups, specifically various Ostrea spp. listed in Table 2, O. lurida and O. conchaphila are weakly supported as sister species. However, Keen and McLean (1971) note the presence of at least 3 other putative Ostrea species in the Gulf of California, for which molecular data are unavailable, but who might serve as sister taxa to either focal species in this study.
Bootstrap support for the sister relationship between the focal species was higher for MP (68%) than for ML (43%). For either MP or ML, the highest and lowest bootstrap support proportions were observed for the CO3-only and 16S-only data sets, respectively, but even the latter was reasonably well supported in the case of MP. As detailed earlier in our results, alternative grouping with the next highest bootstrap proportion for both MP (11%) and ML (35%) placed O. lurida as basal within the "Ostreinae" + Lophinae monophyletic grouping. Still, the best trees for either optimality criterion places O. conchaphila and O. lurida as sister species. The lack of bootstrap support could be caused by the considerable branch length divergence of the involved nodes in their separation from the next available outgroups (Saccostrea and Crassostrea spp.), or the presently sparse intraspecific sampling for our other "Ostreinae" + Lophinae outgroups. Alternatively, the low bootstrap support might reflect the phylogenetic possibility that these native North American oyster species might not be sister species. For example, we have not included any extant or extinct Ostrea spp. from the northwestern Pacific and perhaps one or more of these could have a closer relationship to the similarly cooler water O. lurida.
Interesting phylogeographic structure was observed within O. lurida for the CO3 data set and this coincided with the geographic separation between Willapa Bay, WA and the southern part of Vancouver Island, B.C., Canada. A break in this vicinity is unexpected. For marine organisms, most have found little evidence of a biogeographic break in this area (e.g., Kelly & Eernisse 2007), although some have found a break further north (e.g., Hart et al. 2005) or coinciding with the exposed coast versus inland passage discontinuity between British Columbia and Alaska (e.g., Lindstrom et al. 1997). It is possible that the Vancouver Island localities are populated with oysters that are relicts of Pleistocene glacial refugia, which have been inferred for areas north of these Vancouver Island localities. Specifically, glacial refugia during the late Wisconsin glaciation have been supported for the Queen Charlotte Islands, northern British Columbia (Hetherington et al. 2003) and in the Alexander Archipelago of southeastern Alaska (Carrara et al. 2007). At least one terrestrial species, the lodgepole pine, might have relicts in glacial refugia on the outer coast of Vancouver Island (Godbout et al. 2008). However, the present evidence for phylogeographic separation needs further study with more localities and more gene regions before such historical hypotheses can be accurately inferred.
Because of their extremely plastic morphologies, oysters present difficult challenges for identifications based solely on shell characteristics in the field and such identifications can be especially challenging or even impossible, even after careful examination of internal anatomy and shell structure in the laboratory. For example, the two samples from Bahia de Kino were identified as Ostrea conchaphila in the field, and were reidentified as Crassostrea sp. based on shell morphology in the laboratory. The subsequent molecular analyses grouped only one of the two samples as nearly identical to C. gigas with 100% MP bootstrap support, the other grouped with our other specimens presumed to be Saccostrea palmula (Fig. 1, Table 2). The need, in some cases, to rely on molecular analyses to positively identify specimens underscores the challenges associated with resolving the systematics of this troublesome group. Regarding the morphological characters examined, the presence/absence of chomata was the only unambiguously diagnostic character and it was only informative at a generic level, not between species of Ostrea.
Aside from the presence/absence of chomata, no other shell or soft tissue characters were 100% congruent with molecular data. Only one of the three Saccostrea samples examined had unambiguously elongate chomata all the way around the shell margin. Whereas all Ostrea samples examined had chomata, some were rounded whereas others were slightly elongate. Specimens identified as O. lurida from molecular study (Tables 1 & 3) were nearly as likely to have external shell color or shell plications as not. Soft tissue characters were more consistent, but for each character, at least one specimen, even in this small sample size, differed from the mode. The potential value of anal papillae was further limited by tissue damage. The variability we observed is consistent with comments by Keen (1958) and Coan et al. (2000) on Ostreidae morphology; Coan et al. (2000) emphasize the need for molecular data for more reliable identification. In all, there was considerable intraspecific and interpopulation variation that confounded any attempt to separate the genetically distinct lineages of O. lurida from O. conchaphila by morphology alone.
For the combined analyses of the 16S and CO3 datasets, O. conchaphila and O. lurida were well supported as distinct taxa (Figs. 1-2). Unfortunately, we were unable to find any examples of a calibrated molecular clock for either 16S or CO3 for other bivalves, but an approximate comparison is available for a molecular clock calibrated for 16S in Pacific porcelain crabs (Stillman & Reeb 2001), based on this, the range of estimated pairwise divergence times for O. conchaphila versus O. lurida is about 1.5-3.9 mya. The evidence for the reciprocal monophyly of O. conchaphila and O. lurida is indicative of a significant absence of gene flow between the two groups. However, again, we found no diagnostic morphological character corresponding to either the northern or southern groups (Table 3). Although we have contrasted oysters from near the type locality of O. conchaphila with oysters from the type locality of O. lurida, much further north in Washington, we have not carefully studied the range of morphological variation across our O. lurida localities with enough detail to determine whether there might be diagnostic shell or anatomical traits to separate O. lurida from O. conchaphila. Moreover, we have only collected O. conchaphila from one pair of nearby sites near Mazatlan. Clearly, further morphological study, including more intensive sampling throughout the range of O. conchaphila, is needed.
Based on our molecular evidence, we only found O. lurida within Hertlein's (1959) putative zone of overlap that is supposed to exist between these two species. Samples from the seven localities in southern California and northern Baja California (Table 1) all were effectively indistinguishable from O. lurida at the type locality, Willapa Bay, WA, and this was true for both mtDNA gene regions. In the future, sampling from other locations is warranted to determine the existence, if any, of a zone of overlap, and any such zone is predicted to occur further south than where we sampled, perhaps on the outer coast of Baja California Sur. A transition zone might also exist within the Gulf of California, however, previously published reports imply that neither species occurs within the Gulf of California (e.g., Keen & McLean 1971). Still, this could be because of lack of study, because molecular data on Gulf of California oyster species are mostly lacking. Thus, the only stretch of ostensibly inhabited coastline that was not sampled in this study extends from locations south of Bahia San Quintin, Baja California to the previously reported southern range endpoint for Ostrea lurida at Cabo San Lucas, Baja California Sur. Recent extensive intertidal searches in Cabo San Lucas failed to locate any oysters (Polson & Zacherl, this issue). The only mention of the existence of a zone of overlap was by Hertlein (1959), who also described three separate morphotypes found in southern California. Considering the high degree of phenotypic plasticity in oyster shell morphology, a zone of overlap more likely never was present in southern California and northern Baja California. Instead, Hertlein (1959) might have documented evidence of the high degree of phenotypic plasticity exhibited by O. lurida.
To bridge the 2 most popular species concepts, the biological species concept and the phylogenetic species concept, the concordance principle (CP) introduced by Avise & Ball (1990), recognizes a species based on the phylogenetic concordance of multiple independent molecular markers. Moreover, CP calls for reproductive isolation but is based on intrinsic rather than extrinsic (i.e., geographic) forces. With this study, we have provided molecular evidence across two nonindependent (i.e., tightly linked) mtDNA markers in a geographic context. Our phylogenetic estimates support O. lurida and O. conchaphila as two distinct taxa, with data still lacking for any transitional localities, should they exist. Finding a zone of overlap (sympatry) within which there is little evidence for ongoing hybridization would have provided stronger support that O. lurida and O. conchaphila are separate species and not simply extremes of a latitudinal cline. Future studies would benefit from sampling variable nuclear markers to test for such hybridization. Recently Wang and Guo (2008) found ITS length variation across multiple species of Crassostrea but unfortunately no variation was detected across the Ostrea spp. compared.
The provisional decision to again regard these clades as separate species here, reviving Carpenter's Ostrea lurida and rejecting Harry's (1985) synonymy, is based on the evidence from this study for historical separation of the oysters from the Panamic versus the cooler northeastern Pacific. Whereas no one has yet demonstrated diagnostic shell or anatomical characters, this situation is in no small part caused by their notorious plasticity, and having the present phylogenetic estimate will help define where morphological contrasts with a genetic basis are likely to be found. With roughly 1.5-3.9 my separation, our mitochondrial markers were successful in distinguishing these species. The Olympia oyster becomes a sister species whose southern range limit has yet to be carefully explored along the Baja California Peninsula, and much less is yet known about O. conchaphila. In addition to molecular distance, this pair of oyster species is separated both by geography and contrasting climates. Although other Ostrea spp. and other locations in Mexico need to be sampled, evidence to date implies a sibling species pair with one in the subtropical Panamic marine biogeographical province and the other extending along the cooler Californian and Oregonian temperate provinces to the north. Our preliminary evidence for further phylogeographic structure within O. lurida is potential evidence that such historical processes might be ongoing. Thus, calling for the reinstatement of Ostrea lurida as it was originally described and once widely accepted will help focus research on the remaining mysteries of the Olympia oyster throughout its range.
The authors thank Dr. Michel Hendrickx and Jose "Pepe" Salgado (Universidad Nacional Autonoma de Mexico, Instituto de Ciencias del Mar y Limnologia, Mexico) for collecting samples from Sonora and Sinaloa, Mexico and for all their help with our sampling in Mazatlan. The authors also thank Russell Rogers (Washington Department of Fish and Wildlife) for the Willapa Bay samples and Dr. Anthony Ricciardi from the Redpath Museum, Canada for lending us a syntype of Ostrea lurida. Paul Valentic-Scott (Santa Barbara Museum of Natural History) helped with acquisition of our vouchers into that museum and provided valuable information on Panamic Ostreidae. The authors also appreciate the help of Ms. Lisa Bukovnik at Duke University's IGSP for her help with sequencing our samples. Financial support was provided by grants to D. Zacherl from the CSU Special Fund for Research, Scholarship, and Creative Activity and from NSF-OCE # 0351860, and grants to M. Poison from the Department of Biological Science and the Dr. and Mrs. Donald B. Bright Environmental Scholarship. D. Eernisse acknowledges sabbatical support by CSUF and a fellowship from the National Evolutionary Synthesis Center (NESCent) at Duke University, NSF-EF#0423641.
Avise, J.C. 2004. Molecular markers, natural history, and evolution. Second edition. Sinauer Associates, Inc. Publishers. Sunderland, Massachusetts. 684 pp.
Avise, J. C. & R. M. Ball. 1990. Principles of genealogical concordance in species concepts and biological taxonomy. Oxford Surv. Evol. Biol. 7:45-67.
Baker, P. 1995. Review of ecology and fishery of the Olympia oyster, Ostrea lurida, with annotated bibliography. J. Shellfish Res. 14:501-518.
Carrara, P. E., T. A. Ager & J. F. Baichtal. 2007. Possible refugia in the Alexander Archipelago of southeastern Alaska during the late Wisconsin glaciation. Can. J. Earth Sci. 44:229-244.
Castillo Rodriguez, Z. G. & A. Garcia-Cubas. 1986. Taxonomia y anatomia comparada de las ostras en las costas de Mexico. An. Inst. Cienc. Mar Linmol. Univ. Nac. Aulon. Mex. 13:249-314.
Clement, M., D. Posada & K. Crandall. 2000. TCS: a computer program to estimate gene genealogies. Mol. Ecol. 9:1657-1660.
Coan, E.V., P. Valentich-Scott & & F.R. Bernard. 2000. Bivalve seashells of western North America: marine bivalve mollusks from Arctic Alaska to Baja California. Natural History Monographs No. 2, Santa Barbara Museum of Natural History, Santa Barbara, California, USA. 764 pp.
CodonCode Corporation. Software for DNA sequencing, www.codoncode. com.
Cracraft, J. 1987. Species concepts and the ontology of evolution. Biol. Philos. 2:329-346.
Crummett, L. T. & D. J. Eernisse. 2007. Genetic evidence for the cryptic species pair, Loilia digitalis, and Lottia austrodigitalis and microhabitat partitioning in sympatry. Mar. Biol. 152:1-13.
Dall, W. H. 1914. Notes on west American oysters. Nautilus 28:1-3.
Dawson, N. D. & D. K. Jacobs. 2001. Molecular evidence for cryptic species of Aurelia aurita (Cnidaria, Scyphozoa). Biol. Bull. 200:92-96.
Donoghue, M. J. 1985. A critique of the biological species concept and recommendations for a phylogenetic alternative. Bryologist 88:172-181.
Excoffier, L., G. Laval & S. Schneider. 2005. Arlequin ver. 3.0: an integrated software package for population genetics data analysis. Evol. Bioinform. Online 1:47-50.
Godbout, J., A. Fazekas, C. H. Newton, F. C. Yeh & J. Bousquet. 2008. Glacial vicariance in the Pacific Northwest: evidence from a lodgepole pine mitochondrial DNA minisatellite for multiple genetically distinct and widely separated refugia. Mol. Ecol. 17:2463-2475.
Harry, H. 1985. Synopsis of the supraspecific classification of living oysters (Bivalvia: Gryphaeidae and Ostreidae). Veliger 28:121-158.
Hart, M. W., J. Sunday & C. C. Keever. 2005. Dispersal and climate history in sea star population genetic structure. Integr. Comp. Biol. 45:1009. [abstract].
Hetherington, R., J. Vaughn Barrie, R. G. B. Reid, R. MacLeod, D. J. Smith, T. S. James & R. Kung. 2003. Late Pleistocene coastal paleogeography of the Queen Charlotte Islands, British Columbia, Canada, and its implications for terrestrial biogeography and early postglacial human occupation. Can. J. Earth Sci. 40:1755-1766.
Hertlein, L. 1959. Notes on California oysters. Veliger 2:5-10.
Hewson, W.E. 2008. Biogeography of Lottia strigatella (Carpenter 1864) and a description of a cryptic species. Master thesis. California State University, Fullerton, CA. 82 pp.
Keen, A. M. 1958. Sea shells of tropical west America. Stanford, California: Stanford. 624 pp.
Keen, A. M. & J. H. McLean. 1971. Sea shells of tropical west America. Second edition. Stanford, California: Stanford University Press. 1,064 pp.
Kelly, R. P. & D. J. Eernisse. 2007. Southern hospitality: A latitudinal gradient in gene flow in the marine environment. Evolution hit. J. Org. Evolution 61:700-707.
Kelly, R. P. & D. J. Eernisse. 2008. Reconstructing a radiation: The chiton genus Mopalia in the north Pacific. Invertebr. Syst. 22:1-12.
Kirkendale, L., T. Lee, P. Baker & D.O. Foighil. 2004. Oysters of the Conch Republic (Florida Keys): a molecular phylogenetic study of Parahyotissa megintyi, Teskeyostrea weberi and Ostreola equestris. Malacologia 46:309-326.
Kluge, A. G. 1990. Species as historical individuals. Biol. Philos. 5:417-431.
Knowlton, N. 1993. Sibling species in the sea. Atom. Rev. Ecol. Syst. 24:189-216.
Lam, K. & B. Morton. 2006. Morphological and mitochondrial-DNA analysis of the Indo-West Pacific rock oysters (Ostreidae: Saccostrea species). J. Molluscan Stud. 72:235-245.
Lapegue, S., I. B. Salah, F. M. Batista, S. Heurtebise, L. Neifar & P. Boudry. 2006. Phylogeographic study of the dwarf oyster, Ostreola stentina, from Morocco, Portugal and Tunisia: evidence of a geographic disjunction with the closely related taxa, Ostrea aupouria and Ostreola equestris. Mar. Biol. 150:103-110.
Lindstrom, S. C., J. L. Olsen & W. T. Stare. 1997. Postglacial recolonization and the biogeography of Palmaria mollis (Rhodophyta) along the Northeast Pacific coast. Can. J. Zool. 75:1887-1896.
Mishler, B. D. & M. J. Donoghue. 1982. Species concepts: a case for pluralism. Syst. Zool. 31:491-503.
Palumbi, S. 1996. Nucleic Acids II: the polymerase chain reaction. In: D. M. Hillis, C. Mortiz & B. K. Mable. editors. Molecular Systematics, 2nd edition. Sunderland, MA: Sinauer and Assoc. Inc. pp. 205-246.
Quayle, D.B. 1988. Pacific oyster culture in British Columbia. Canadian Bulletin of Fisheries and Aquatic Sciences 218. Department of Fisheries and Oceans, Ottawa, Ontario, Canada. 192 pp.
Shilts. H. M., M. S. Pascual & D.O. Foighil. 2007. Systematic, taxonomic and biogeographic relationship of Argentine flat oysters. Mol. Phylogenet. Evol. 44:467-473.
Simison, W. B. & D. R. Lindberg. 1999. Morphological and molecular resolution of a putative cryptic species complex: a case study of Notoacmea fascicularis (Menke, 1851) (Gastropoda: Patellogastropoda). J. Molluscan Stud. 65:99-109.
Stillman. J. H. & C. A. Reeb. 2001. Molecular phylogeny of eastern Pacific porcelain crabs, genera Petrolisthes and Pachycheles based on mtDNA 16S rDNA sequence: phylogeographic and systematic implications. Mol. Phylogenet. Evol. 19:236-245.
Swofford, D. L. 2003. PAUP*. Phylogenetic Analysis Using Parsimony (* and Other Methods). Version 4.0b10. Sinauer Associates, Sunderland, Massachusetts.
Tao, N., W. J. Bruno, W. Abfalterer, B. M. E. Moret, T. Leitner & C. Kuiken. 2005. FINDMODEL: a tool to select the best-fit model of nucleotide substitution, http://hcv.lanl.gov/content/hcv-db/findmodel/findmodel.html.
Torigoe, K. 1981. Oysters in Japan. J. Sci. Hiroshima Univ., B. 1. 29:291-481.
Wang, Y. P. & X. M. Guo. 2008. ITS length polymorphism in oysters and its use in species identification. J. Shellfish Res. 27:489-493.
Zwickl, D.J. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. PhD thesis. University of Texas, Austin. 125 pp.
MARIA P. POLSON, (1) WILLIAM E. HEWSON, (1) DOUGLAS J. EERNISSE, (1) PATRICK K. BAKER (2) AND DANIELLE C. ZACHERL (1) *
(1) Department of Biological Science, California State University Fullerton, P.O. Box 6850, Fullerton, California 92834; (2) Department of Fisheries and Aquatic Sciences, University of Florida, 7922 NW 71st Street, Gainesville, Florida 32653
* Corresponding author. E-mail: firstname.lastname@example.org
([dagger]) The taxonomy of the Olympia oyster has been in dispute since Harry (1985) proposed synonymy of Ostrea lurida Carpenter 1864 and Ostrea conchaphila Carpenter 1857. Poison et al. 2009 provide molecular evidence that the Olympia oyster refers to the nominal species, Ostrea lurida Carpenter 1864. In view of their genetic data, and for consistency, the original taxon, Ostrea lurida, is used throughout this volume to refer to the Olympia oyster, which is distributed from approximately Baja California (Mexico) to southeast Alaska.
([double dagger]) The common name as sanctioned by the American Fisheries Society is Olympia oyster, though other names are commonly used including, "native oyster," "California oyster," "Yaquina oyster."
TABLE 1. Sampling localities for sequenced western North American oysters with the number of samples sequenced for each gene. Names are as determined in this study. 16S sampling emphasized the contrast between Ostrea conchaphila and O. lurida and tested for Hertlein's (1959) proposed region of overlap whereas C03 sequencing emphasized tests of phylogeographic structure within O. lurida. Species Locality (Code) Ostrea lurida Ladysmith Harbor, Vancouver Id.. B.C. (LH) Ostrea lurida Ahmah Island, Vancouver Id., B.C. (AI) Ostrea lurida Willapa Bay, WA (WB) Ostrea lurida Yaquina Bay, OR (YB) Ostrea lurida Coos Bay, OR (CB) Osirea lurida Humboldt Bay, CA (HB) Ostrea lurida Tomales Bay West, CA (TB) Ostrea lurida Drakes Estero, CA (DE) Ostrea lurida Pt, San Quentin, San Francisco Bay, CA (SF) Ostrea lurida Elkhorn Slough, Monterey Co., CA (ES) Ostrea lurida Mugu Lagoon. CA (ML) Ostrea lurida Alamitos Bay, CA (AL) Ostrea lurida Newport Bay, CA (NB) Ostrea lurida Aqua Hedionda, CA (AH) Ostrea lurida Batiquitos Lagoon, CA (BL) Ostrea lurida Scripps Pier, La Jolla, CA (LJ) Ostrea lurida Mission Bay, CA (MB) Ostrea lurida San Diego Bay, CA (SD) Ostrea lurida Bahia San Quintin, Baja California (SQ) Ostrea conchaphila Ensenada del Pabellon (N), Sinaloa (EdPn) Ostrea conchaphila Ensenada del Pabellon (S), Sinaloa (EdPs) Saccostrea spp. Urias Lagoon, Mazatlan, Sinaloa (Maz) Saccostrea sp. Bahia de Kino, Sonora (BdK) Crassostrea gigas Bahia de Kino, Sonora (BdK) Species GPS Coordinates Ostrea lurida N 48[degrees]59.321', W 123[degrees]48.241' Ostrea lurida N 48[degrees]56.911', W 125[degrees]5.239' Ostrea lurida N 46[degrees]29.916', W 124[degrees]01.766' Ostrea lurida N 44[degrees]36.807', W 124[degrees]4.799' Ostrea lurida N 43[degrees]19.313', W 124[degrees]23.156' Osirea lurida N 40[degrees]45.225', W 124[degrees]12.92' Ostrea lurida N 38[degrees]13.763', W 122[degrees]58.556' Ostrea lurida N 38[degrees]0.302', W 123[degrees]0.307' Ostrea lurida N 37[degrees]56.508', W 122[degrees]30.315' Ostrea lurida N 36[degrees]48.631', W 121[degrees]46.996' Ostrea lurida N 34[degrees]6.104', W 119[degrees]5.882' Ostrea lurida N 33[degrees]44.264', W 118[degrees]7.114' Ostrea lurida N 33[degrees]37.173', W 117[degrees]53.542' Ostrea lurida N 33[degrees]08.579', W 117[degrees]19.370' Ostrea lurida N 33[degrees]08.579', W 117[degrees]19.370' Ostrea lurida N 32[degrees]52.087', W 117[degrees]15.307' Ostrea lurida N 32[degrees]47.673', W 117'13.468' Ostrea lurida N 32[degrees]42.505', W 117[degrees]10.260' Ostrea lurida N 30[degrees]27.858', W 115'57.834' Ostrea conchaphila N 24[degrees]30.063', W 107[degrees]41.245' Ostrea conchaphila N 24[degrees]28.843', W 107[degrees]33.396' Saccostrea spp. N 23[degrees]12.33', W 106[degrees]24.315' Saccostrea sp. N 28[degrees]45.55', W 111[degrees]54.8' Crassostrea gigas N 28[degrees]45.55', W 111[degrees]54.8' Species 16S Seqs CO3 Seqs Ostrea lurida 2 4 Ostrea lurida 1 3 Ostrea lurida 21 5 Ostrea lurida 0 5 Ostrea lurida 0 5 Osirea lurida 0 5 Ostrea lurida 0 5 Ostrea lurida 0 4 Ostrea lurida 0 4 Ostrea lurida 0 3 Ostrea lurida 0 4 Ostrea lurida 5 5 Ostrea lurida 4 5 Ostrea lurida 4 3 Ostrea lurida 4 5 Ostrea lurida 5 1 Ostrea lurida 4 5 Ostrea lurida 4 3 Ostrea lurida 4 3 Ostrea conchaphila 2 2 Ostrea conchaphila 7 5 Saccostrea spp. 24 0 Saccostrea sp. 1 0 Crassostrea gigas 1 0 TABLE 2. GenBank sequences used in analyses (16S except as noted). Species Locality in GenBank Accession Ostrcct lurida * Barkley Sound, BC. AF052071 Canada Outgroups Ostrea algoensis Port Alfred, AF052062 South Africa Ostrea angasi St. Helen. Tasmania. AF052063 Australia Ostrea auporia Hauraki Gulf, AF052064 New Zealand Ostrea chilensis Hauraki Gulf, AF052065 New Zealand Ostrea denselamellosa South Korean AF052067 hatchery stock Ostrea edidis France DQ280032 Ostrea equestris Big Pine AF052074 Key, FL, USA Ostrea puelchana San Matias AF052073 Gulf, Argentina Ostrea spreta Argentina DQ640402 Ostrea weberi Florida, USA AY376601 Alectryonella plicatula Pohnpei, Micronesia AF052072 Cryrptostrea permollis Florida panhandle, USA AF052075 Dendostrea folium Aitutaki, Cook AF052069 Islands Dendostrea frons Carrie Bow Cay, AF052070 Belize Lopha cristagalli Guam AF052066 Saccostrea commiercialis Moreton Island, AF353100 Queensland, Australia Saccostrea cuccullata Australia AF458901 Crassostrea ariakensis China AY160757 Crassostrea rhizophorae Brazil AJ312938 Crassostrea gigas Bangor, U.K. (cultured) AJ553903 Crassostrea gigas(CO3) None listed AF177226 (native to northwestern Pacific) Crassostrea virginica None listed AF092285 (native to northwestern Atlantic) Crassostrea virginica Delaware Bay. USA AY905542 (CO3) * This ingroup sequence from GenBank. O. "conchaphila," was reidentified in the current study as O. lurida. TABLE 3. Morphological characters examined among a subset of samples used in the molecular analyses (see Table 1 and Figure 1 for locality codes in the first column). ID Molecular Specimen ID Field * ID Laboratory ** ([section]) BdK 0701 O. conchaphila Crassostrea sp. (a) Crassostrea sp. BdK 0702 O. conchaphila Crassostrea sp. (a) Saccostrea sp. Maz 0503 O. conchaphila Saccostrea sp. (a) Saccostrea sp. Maz 0510 O. conchaphila Unresolved (a) Saccostrea sp. Maz 0512 O. conchaphila Unresolved (a) Saccostrea sp. EdPn 0702 O. conchaphila O. conchaphila (a) O. conchaphila EdPs 0701 O. conchaphila O. conchaphila (a) O. conchaphila EdPs 0703 O. conchaphila O. conchaphila (a) O. conchaphila EdPs 0704 O. conchaphila O. conchaphila (a) O. conchaphila SD 0601 O. conchaphila O. conchaphila (a) O. lurida SD 0603 O. conchaphila O. conchaphila (a) O. lurida AH 0505 O. conchaphila O. conchaphila (a) O. lurida NB 0601 O. conchaphila O. conchaphila (a) O. lurida NB 0602 O. conchaphila O. conchaphila (a) O. lurida SQ 0504 O. conchaphila O. conchaphila (a) O. lurida SQ 0505 O. conchaphila O. conchaphila (a) O. lurida LJ 0503 O. conchaphila O. conchaphila (a) O. lurida LJ 0505 O. conchaphila O. conchaphila (a) O. lurida Chomata Chomata Specimen Sweep Shape Plicae BdK 0701 None Absent No BdK 0702 <50% Round Yes Maz 0503 100% Elongate Yes Maz 0510 <50% Round Yes Maz 0512 <50% Round Yes EdPn 0702 <50% Round-elongate Yes EdPs 0701 <50% Round-elongate Yes EdPs 0703 <50% Round-elongate Yes EdPs 0704 <50% Round-elongate Yes SD 0601 <50% Round Yes SD 0603 <50% Round Yes AH 0505 <50% Round Yes NB 0601 <50% Round No NB 0602 <50% Round No SQ 0504 <50% Round Yes SQ 0505 <50% Round Yes LJ 0503 <50% Round No LJ 0505 <50% Round-elongate No External Mantle Anal Specimen Color Tentacles Papilla BdK 0701 No Monomorphic NA BdK 0702 No Polymorphic Internal Maz 0503 Yes Monomorphic None Maz 0510 No Monomorphic None Maz 0512 No Polymorphic Simple EdPn 0702 Yes Polymorphic Simple EdPs 0701 No Monomorphic None EdPs 0703 Yes Polymorphic Simple EdPs 0704 Yes Polymorphic Simple SD 0601 Yes Polymorphic NA SD 0603 Yes Polymorphic Simple AH 0505 Yes Polymorphic NA NB 0601 No Polymorphic Simple NB 0602 No Polymorphic Simple SQ 0504 Yes Polymorphic NA SQ 0505 Yes Polymorphic Simple LJ 0503 No monomorphic NA LJ 0505 No polymorphic NA * Taxon ID based upon field identification using external morphological characters. ** Taxon ID based upon both (a) pre and (b) posthoc laboratory examination of internal and external morphological characters. ([section]) Taxon ID after molecular results; Saccostrea sp. is likely S. palmula, the only Panamic member of this genus.