Portal de Periódicos da CAPES

Ecological and genetic variation in reef-building corals on four Society Islands

2016; Wiley; Volume: 61; Issue: 2 Linguagem: Inglês

10.1002/lno.10231

1939-5604

Peter J. Edmunds, James J. Leichter, Erika C. Johnston, Eric Tong, Robert J. Toonen,

Marine Biology and Ecology Research

Limnology and OceanographyVolume 61, Issue 2 p. 543-557 ArticleFree Access Ecological and genetic variation in reef-building corals on four Society Islands Peter J. Edmunds, Corresponding Author Peter J. Edmunds Department of Biology, California State University, Northridge, CaliforniaCorrespondence: peter.edmunds@csun.eduSearch for more papers by this authorJames J. Leichter, James J. Leichter Scripps Institution of Oceanography, University of California at San Diego, La Jolla, CaliforniaSearch for more papers by this authorErika C. Johnston, Erika C. Johnston School of Ocean and Earth Science and Technology, Hawai'i Institute of Marine Biology, University of Hawai'i at Mānoa, Kāne'ohe, Hawai'iSearch for more papers by this authorEric J. Tong, Eric J. Tong School of Ocean and Earth Science and Technology, Hawai'i Institute of Marine Biology, University of Hawai'i at Mānoa, Kāne'ohe, Hawai'iSearch for more papers by this authorRobert J. Toonen, Robert J. Toonen School of Ocean and Earth Science and Technology, Hawai'i Institute of Marine Biology, University of Hawai'i at Mānoa, Kāne'ohe, Hawai'iSearch for more papers by this author Peter J. Edmunds, Corresponding Author Peter J. Edmunds Department of Biology, California State University, Northridge, CaliforniaCorrespondence: peter.edmunds@csun.eduSearch for more papers by this authorJames J. Leichter, James J. Leichter Scripps Institution of Oceanography, University of California at San Diego, La Jolla, CaliforniaSearch for more papers by this authorErika C. Johnston, Erika C. Johnston School of Ocean and Earth Science and Technology, Hawai'i Institute of Marine Biology, University of Hawai'i at Mānoa, Kāne'ohe, Hawai'iSearch for more papers by this authorEric J. Tong, Eric J. Tong School of Ocean and Earth Science and Technology, Hawai'i Institute of Marine Biology, University of Hawai'i at Mānoa, Kāne'ohe, Hawai'iSearch for more papers by this authorRobert J. Toonen, Robert J. Toonen School of Ocean and Earth Science and Technology, Hawai'i Institute of Marine Biology, University of Hawai'i at Mānoa, Kāne'ohe, Hawai'iSearch for more papers by this author First published: 04 February 2016 https://doi.org/10.1002/lno.10231Citations: 25AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract We quantified benthic community structure on shallow (10 m isobath) reefs separated by 3–130 km on four islands in the south Pacific, and evaluated the roles of disturbances vs. coral recruitment as causes of spatial heterogeneity. Reefs were surveyed in 2013 on Moorea, Tahiti, Tetiaroa, and Maiao, with community structure sampled at two sites on each island using photoquadrats. The effects of coral recruitment on population structure were evaluated through genetic analyses of Pocillopora on three of the islands. Benthic community structure with functional group resolution differed among islands and generally among sites, but coral community structure (generic resolution) differed among islands, but generally not among sites. Genetic analyses of Pocillopora using the open reading frame of host mtDNA revealed varying relative abundances of Pocillopora meandrina/Pocillopora eydouxi, Pocillopora verrucosa, Pocillopora effuses, and two unnamed haplotypes on each island. These results suggest that corals on each island represent unique samplings of genetically discrete larval assemblages rather than random samplings of a single larval assemblage. Together, our findings emphasize the extent to which coral community structure varies over a scale of <200 km, and suggests that recruitment from spatially discrete pools of coral larvae plays an important role in creating spatial variation in community structure, even where reefs are connected by prevailing currents. Spatial variation in community structure is engrained in ecological science (Levin 1992), and it reflects the occurrence of dissimilar communities in different locations across a variety of spatial scales. Disturbances are a leading cause of such variation (Pickett and White 1985), but community structure also is an emergent property of population growth facilitated by the exchange of propagules within metapopulations (Gilpin and Hanski 2012). The magnitude of these exchanges depends on the degree to which populations are open or closed (Cowen et al. 2000), but the effects can be large. For example, when the flux of propagules involves ecologically important species with a strong capacity to acquire space, recruitment can initiate benthic community development that is unique to the time of arrival of the propagules (Connell and Slatyer 1977). Together the aforementioned effects make it difficult to determine the causes of spatial heterogeneity in community structure, particularly in aquatic environments where many organisms produce long-lived pelagic larvae that can settle in distant locations (Cowen and Sponaugle 2009), and variation in community structure can be largely uncoupled from recruitment (Caley et al. 1996). Tropical coral reefs provide good examples of the challenges described above, because their communities differ dramatically across space (Done 1982; Connell 1997), and their ecosystem engineers (i.e., scleractinians) rely on pelagic larvae for recruitment (Harrison and Wallace 1990). Surprisingly, however, the role of coral recruitment in creating spatial variation in community structure is not completely understood (Arnold et al. 2010). The very occurrence of spatial variation in reef communities motivates investigation of the causal processes, but the advantages of doing so are accentuated by the potential also to explain why reefs vary over time. Substituting space-for-time to investigate temporal variation in community structure has limitations (Pickett 1989), but nevertheless it is a well-established approach for this purpose (Blois et al. 2013; Longenecker et al. 2014). The advantages are especially clear for longevous taxa like many corals, for it can improve understanding of past, present, and future communities without requiring studies of a length commensurate with the life span of corals. Given the negative changes that have affected coral reefs since the 1970s (Bruno and Selig 2007; Jackson et al. 2014), and the likelihood that these will continue (van Hooidonk et al. 2014), there is much to be gained by exploiting space-for-time constructs to better understand the changes affecting benthic community structure. There is a rich empirical and theoretical consideration of the causes of variation in coral reef communities (Connell 1978; Edmunds and Bruno 1996; Hughes et al. 2012), and it is likely that the spectrum of causes remains largely unchanged in the current vs. the previous century. However, most present-day reefs are different compared to the recent past (Jackson et al. 2014), and it is reasonable to posit that the causes of variation in community structure now act in modified ways compared to the past. For instance, with reduced population sizes of vertebrates (Jackson 1997) and acroporid corals (Aronson and Precht 2001), the synecology of present-day reefs probably differs from that prevailing a century ago (Jackson 1997). Changing aspects of the physical [e.g., seawater temperature (Tierney et al. 2015)] and chemical [e.g., seawater pH (Doney et al. 2009)] environment provide reasons to expect the modes of action of ecological processes like larval substrate selection (Gleason and Hofmann 2011) or competitive hierarchies among scleractinians (Chadwick and Morrow 2011) to have changed. Ecological surveys are effective in describing the changes in community structure that can result from these effects, and when repeated over time they can quantify trajectories of change. They cannot, however, address the movement of propagules (i.e., larvae) among locations as a potential mechanism determining population growth, which means they cannot evaluate the capacity for distant populations to modulate local-scale (i.e., over < 20 km) population dynamics. These issues are integral to metapopulation theory (Gilpin and Hanski 2012), and while analyses of seawater currents can elucidate the potential for larvae to be transported among populations (Kough and Paris 2015), ecologically meaningful connectivity among populations can only be determined through analyses of genetic population structure (Selkoe and Toonen 2011). In this study, we address spatial variation in coral reef community structure, and genetic variation for the dominant coral genus Pocillopora on outer reefs of Tahiti, Moorea, Tetiaroa, and Maiao in the south Pacific. Benthic sampling was used to quantify community structure at 10 m depth, and genetic population analyses of Pocillopora spp. were used to evaluate the capacity for coral larvae to travel among islands. We assumed that the coral reefs would differ among locations (Done 1982; Karlson and Hurd 1993), and designed the sampling to quantify this variation through two analyses resolving benthic organisms to functional group and also to genera of scleractinians and Millepora. Genetic analyses were implemented to evaluate the extent of connectivity among coral populations on the islands, but this objective proved beyond the scope of the study given the taxonomic resolution we ultimately obtained (described below). Instead of addressing connectivity directly, we sequenced the open reading frame (ORF; Flot and Tillier 2007) of the host mitochondrial DNA in Pocillopora, and used the results to consider the likelihood that pocilloporids on multiple islands were randomly drawn from a common pool of larvae. The ecological and genetic analyses were used to evaluate the goodness of fit of our results to two domains of theory explaining how spatial variation in coral reef community structure is maintained. One domain posits that a spatio-temporal mosaic of disturbances leaves local coral reef communities representing the time since the last disturbance (Karlson and Hurd 1993); the other domain posits that heterogeneous larval supply, self-seeding, and closed population structures (Cowen and Sponaugle 2009), cause reefs to follow dissimilar community trajectories. In reality, it is likely that both domains of theory have value in explaining coral reef community structure, and the more salient question is what set of factors determines the relative importance of each domain in controlling community structure at specific places and times. Methods Overview Coral reef community structure on four islands was sampled photographically along the 10 m isobath in April and May 2013. Two sites were sampled on Moorea that are permanently marked as part of the time-series analyses of the Long-Term Ecological Research project (Adam et al. 2011), and two sites on each of Tahiti, Tetiaroa, and Maiao (Fig. 1) that are not permanently marked and were sampled for the present project alone. The sites on Moorea (LTER1 and LTER2) were selected haphazardly in 2005 to sample the north shore, outer reef; all other sites were selected haphazardly in 2013, but also are on outer reefs (Fig. 1). Between April and July 2013, pocilloporid corals were sampled for genetic analysis along the 10 m isobath on the outer reefs of Moorea, Maiao, and Tetiaroa to quantify population genetic structure within a dominant coral genus. Figure 1Open in figure viewerPowerPoint Locations of study sites on Moorea, Tahiti, Maiao, and Tetiaroa in French Polynesia; benthic sampling took place in April and May 2013. Insets show sampling sites around each island: (A) Maiao: Site 1 (17°37.917′S, 150°37.701′W) and Site 2 (17°38.006′S, 150°38.227′W); (B) Moorea: Site 1 (17°28.435′S, 149°50.222′W) and Site 2 (17°28.242′S, 149°48.472′W); (C) Tahiti: Site 1 (17°35.553′S, 149°37.597′W) and Site 2 (17°36.591′S, 149°37.351′W); and (D) Tetiaroa: Site 1 (16°58.882′S, 149°34.121′W) and Site 2 (16°59.176′S, 149°34.954′W). Dashed lines in the insets mark the reef crest; all sampling occurred on the outer reef. Benthic community structure Benthic community structure was sampled along a 50 m transect placed along the 10 m isobath at each site. Photoquadrats (0.5 × 0.5 m, n ∼40) were recorded at positions randomly determined along each transect using a camera mounted on a framer that held it perpendicular to the reef (images available at http://mcr.lternet.edu). Photographs were recorded with a Nikon D7000 camera (16 megapixels) and 18–105 mm DX Nikon lens contained in a waterproof housing and attached to two strobes (Nikon SB105). Objects as small as 3–5 mm diameter could be identified with this system. Photoquadrats were analyzed using CPCe software (Kohler and Gill 2006), and benthic substrata were identified beneath 200 randomly located points in each image. First, a functional group analysis was completed in which scleractinians, macroalgae (algae > 1 cm tall that consisted mostly of Dictyota, Halimeda, and Asparagopsis), and a combined group consisting of crustose coralline algae, bare space, and algal turf (<1 cm tall, defined as crustose coralline algae, algal turf, and bare space [CTB]) were scored. In some locations an encrusting form of Lobophora was encountered and this was scored as macroalgae. CTB was scored as a single group, as the component substrata could not be distinguished in planar images, and because a finer resolution was not required for the questions being addressed. Second, a fine-resolution analysis was conducted in which scleractinians were scored by genus together with the hydrocoral Millepora, which was abundant in some quadrats. Statistical analysis of benthic community structure Benthic community structure was displayed with bar graphs and non-metric multi-dimensional scaling (MDS) plots, and statistically compared with multivariate tools and univariate techniques, in both cases testing for effects of islands and sites. Analyses were completed first by functional groups, and second, by genera of scleractinians and Millepora. Univariate analyses were applied to all three functional groups, and to the eight most common genera. Bar graphs were used to display mean and standard errors using untransformed data with photoquadrats as replicates. MDS was used to visualize patterns of community similarity among sites based on mean values (by site) of each category of benthic organism. To prepare MDS plots, data were square-root (functional groups) or Log (x + 1) (coral genera) transformed to adjust for unequal representation of benthic organisms, and converted to a matrix of Bray–Curtis similarities. Multiple restarts of 100 iterations were used until stress stabilized and ordinations were repeatable. Site symbols were scaled as circles representing the percentage cover of categories of benthic taxa. Multivariate inferential tests were completed with ANOSIM, SIMPER, and PERMANOVA (Anderson et al. 2008). To compare among islands and sites (within islands) using photoquadrats as replicates, ANOSIM was used with islands as a fixed effect and sites nested in islands. To test the hypothesis that coral reef community structure differed among islands using the percent cover of Pocillopora, Porites, Astrea, Montipora, Leptastrea, Acropora, Pavona, and Millepora as dependent variables, univariate, two factor mixed model ANOVAs were used in which island was a fixed effect and sites were random and nested within islands. The assumptions of normality and equality of variance were tested through graphical analyses of residuals. Statistical analyses were completed using Primer-E with the PERMANOVA add on (for multivariate analyses) or Systat 11 (for univariate analyses) software. Scleractinian genetics To explore the genetic population structure of corals across sites and islands, Pocillopora tissue was sampled from the outer reefs of Moorea, Tetiaroa, and Maiao; logistical constraints prevented sampling of the outer reef of Tahiti. Sampling targeted coral colonies selected haphazardly based on their conformation to the classic morphology of Pocillopora meandrina (sensu Dana 1846). Colonies were sampled either along each transect that was used for the photographic sampling (Tetiaroa and Maiao), or from an outer reef site on Moorea. Sampled colonies were < ∼10 cm diameter, which suggests that they were < 3 yr old based on growth rates we have recorded at 10 m depth on the north shore of Moorea (P. J. Edmunds, unpubl.). At least on the north outer shore of Moorea, the destructive effects of cyclone Oli in February 2010 demonstrate that none of the Pocillopora spp. colonies were > 38 months old when sampled in April 2013 (because the outer reefs at 10 m depth on the north shore were reduced to < 1.9% cover by April 2010). Annual photographic sampling between 2005 and 2014 suggested that Pocillopora verrucosa and Pocillopora eydouxi also are common on the reefs of Moorea, and that Pocillopora woodjonesi is encountered occasionally (P. J. Edmunds personal observation). Because congeneric pocilloporids are difficult to identify underwater, especially when colonies are small (Pinzón et al. 2013), we assumed our sampling included P. verrucosa and P. eydouxi [which are thought to be common in Moorea (Bosserelle et al. 2014)] even though P. meandrina was targeted for sampling. Approximately 25 colonies of putative P. meandrina were sampled at each of the two sites on Tetiaroa and Maiao (Fig. 1) so that ∼50 colonies were collected from each island. Sampling occurred at the same time as the photoquadrats were recorded (April 2013), and wire cutters were used to collect tissue biopsies (∼5 × 10 mm) from each colony. Biopsies were bagged individually, and the cutters were rinsed with seawater between samplings to reduce the possibility of cross contamination. In July 2013, ∼50 biopsies (1–3 cm branch tips) were similarly sampled from colonies of P. meandrina morphology at one outer reef site (10 m isobath) on Moorea (the LTER0 site; Fig. 1). Biopsies from the April and July samplings were placed on ice and processed within ∼2–3 h of collection by trimming to ∼125 mm3 and transferring to a salt-saturated dimethyl sulfoxide solution for preservation at room temperature (∼12 h), followed by long-term storage at ∼4°C (Gaither et al. 2011). Samples were shipped to Hawai'i for DNA extraction and processing. DNA was extracted using the DNeasy 96 Blood & Tissue Kit (Qiagen, Venlo) following the manufacturer protocol. The second elution was diluted 1 : 10 with sterile distilled water and used as template DNA for polymerase chain reaction (PCR) amplification with the mitochondrial ORF primers Fatp6.1 TTTGGGSATTCGTTTAGCAG, and RORF SCCAATATGTTAAACASCATGTA (Flot et al. 2008). PCR mixes contained 5 μL of BioMix Red, ready-to-use 2× reaction mix (Bioline Ltd., London), 0.13 μL of each forward and reverse primer (10 μM), and 1 μL of template DNA (5–50 ng) with deionized water to 10 μL final volume. Each PCR reaction followed the cycling protocol of Flot et al. (2008), with a denaturation step of 60 s at 94°C, followed by 40 cycles (30 s at 94°C, 30 s at 53°C, and 75 s at 72°C), with a final elongation step of 5 min at 72°C. PCR products were visualized through electrophoresis on a 1% w/v TAE agarose gel and purified by incubating 6 μL of PCR product with 0.82 μL of exonuclease I (ExoI, Fermentas) and 0.08 μL of fast alkaline phosphatase (Fermentas) at 37°C for 60 min, followed by deactivation at 85°C for 15 min. Samples were then sequenced on an Applied Biosystems 3730XL Genetic Analyzer (Maiao N = 34, Tetiaroa N = 46, Moorea N = 55) at the Center for Advanced Studies of Genomics, Proteomics, and Bioinformatics core facility at the University of Hawaii at Manoa. To assess the molecular diversity of colonies with P. meandrina morphology among islands, a 434 bp sequence of the mitochondrial open reading frame (mtORF) from each individual was aligned in Geneious v.6.1 (Biomatters). An Analysis of Molecular Variance (AMOVA) was calculated in Arlequin v3.5 to partition molecular variance by island. Exact tests of sample differentiation (Raymond and Rousset 1995), using a Markov Chain length of 1,000,000 steps in Arlequin v3.5, were used to compare Pocillopora spp. assemblages by island and to evaluate the likelihood that the samples were drawn randomly from a common pool of pelagic larvae. Results Overview The underwater geomorphology of the outer reefs at the four islands was similar, with all having a steep profile and deep water (>500 m depth) within 100 m of the reef crest. These shallow reefs developed in the last ∼7000 yr as sea level rose following the end of glaciation in the Holocene and inundated the antecedent reef framework (Bard et al. 1990). The reefs off Moorea, Maiao, and Tetiaroa were on northern shores and exposed to waves from the north during the Austral summer, but probably were sheltered from waves from the south during the Austral summer. The reef off Tahiti was on the western side, and sheltered from northern waves but exposed to southern waves. Moorea and Tahiti provided high-island backdrops (1207 m and 2241 m high, respectively) to their reefs, Maiao provided a low island backdrop (154 m high), and Tetiaroa provided an atoll with a mean height above sea level of only a few meters. Qualitatively, the coral reef community structure at the four islands differed, although all had < 25% coral cover and large amounts of crustose coralline algae, algal turf, macroalgae, or exposed carbonate rock. The reefs of Moorea had large numbers of juvenile scleractinians (often >10–20 m−2) that were mostly pocilloporids. The reef off Tetiaroa had only a low percent cover of scleractinians, and was characterized by a pavement of carbonate rock. The reef off Tahiti had the highest percent cover of scleractinians (∼20%) with large colonies (10–20 cm diameter) of mostly pocilloporids, acroporids, and poritids. The reef off Maiao had intermediate coral cover (∼15%), a coral community with relatively large numbers of colonies (mostly pocilloporids and acroporids) in the ∼10 cm diameter size class, and a substratum heavily populated with crustose coralline algae. Descriptive benthic community structure Along the 10 m isobath, mean coral cover ranged from 1.4% ± 0.2% (Tetiaroa 2) to 27.1% ± 2.3% (Tahiti 2), CTB ranged from 44.6% ± 1.7% (Maiao 2) to 91.4% ± 0.7% (Tetiaroa 2), and macroalgal ranged from 5.4% ± 0.8% (Tahiti 2) to 38.6% ± 2.2% (Maiao 2). Variation in community structure among sites within islands sometimes was large, notably for macroalgae at Maiao (8.7% ± 1.2% vs. 38.6% ± 2.2% at the two sites on this island) and Tahiti (14.5% ± 1.4% vs. 5.4% ± 0.8%), and CTB at Maiao (73.8% ± 2.6% vs. 44.6% ± 1.7%; Fig. 2). The MDS plot based on scleractinians, macroalgae, and CTB, graphically displays the multivariate relationships among sites and islands (Fig. 2D), and reveals differences between Tetiaroa and the other islands. On a relative scale, these differences are not large and the sites do not group in significantly different clusters (SIMPROF, p ≥ 0.38) as revealed by their placement within the 80% similarity contour. Figure 2Open in figure viewerPowerPoint Coral reef community structure on the outer reefs of Moorea, Tahiti, Maiao, and Tetiaroa, with two sites analyzed on each island (Fig. 1). (A–C) Mean percentage cover (± SE, n = 38–43) of scleractinians, macroalgae, and a combined functional group consisting of crustose coralline algae, algal turf, and bare space (CTB). (D) Multi-dimensional scaling (MDS) plot prepared from the mean percentage cover of scleractinians, macroalgae, and CTB (square root transformed) at all site and island combinations, with circles scaled to represent the site means for scleractinians. Similarity contours display fixed resemblance levels determined using Bray–Curtis similarities, although none of the clusters are statistically significant (SIMPROF, p ≥ 0.38). Eighteen scleractinian genera and Millepora were encountered, seven of the scleractinian genera had mean covers ≥ 0.6% (averaged over all surveys), and Millepora had a mean cover of 1.2%. The remaining 11 scleractinian genera covered < 0.1% of the reefs (averaged over all surveys). For the most common genera: Pocillopora ranged from 0.2% ± 0.1% (Tetiaroa 2) to 11.6% ± 1.9% (Tahiti 2) cover; Porites ranged from 0.3% ± 0.1% (Tetiaroa 2) to 9.5% ± 1.3% (Tahiti 1); Astrea ranged from 0.1% ± < 0.1% (Tetiaroa 2) to 1.8% ± 0.6% (Maiao 2); Montipora ranged from 0.1% ± < 0.1% (Tetiaroa 1) to 3.6% ± 0.7% (Maiao 1); Leptastrea ranged from < 0.1% ± < 0.1% (Moorea 2) to 2.5% ± 0.7% (Maiao 2); Acropora ranged from 0% (Tetiaroa 2) to 2.9% ± 0.7% (Tahiti 2); Pavona ranged from < 0.1% ± <0.1% (Tahiti 2) to 1.5% ± 0.3% (Maiao 1); and Millepora ranged from 0% (Tetiaroa 2) to 2.3% ± 1.5% (Maiao 1; Fig. 3; all mean ± SE). The spatial structuring of the coral communities around these islands is revealed in MDS plots. Sites on each island were more similar to one another than to sites on other islands, and two clusters of sites were statistically significant: (Maiao + Tahiti + Moorea) vs. Tetiaroa, and Maiao vs. (Tahiti + Moorea; SIMPROF, p ≤ 0.002; Fig. 4). Figure 3Open in figure viewerPowerPoint Coral community structure on the outer reefs of Moorea, Tahiti, Maiao, and Tetiaroa, with two sites analyzed on each island (Fig. 1). Results shown for the seven most abundant scleractinian genera and Millepora. Mean ± SE (n = 38–43) show for each genus panel. Figure 4Open in figure viewerPowerPoint Multi-dimensional scaling (MDS) plot based on site means of all scleractinians genera and Millepora (Log (x + 1) transformed), with circles scaled to mean cover of Pocillopora (A) or Porites (B). Similarity contours display fixed resemblance levels determined using Bray–Curtis similarities; Maiao vs. (Tahiti + Moorea) is significant (SIMPROF, p = 0.002) as is Maiao + Tahiti + Moorea vs. Tetiaroa (SIMPROF, p < 0.001). Inferential contrasts among islands and sites Coral reef community structure resolved to functional group varied among islands (ANOSIM, R = 0.521, p = 0.048), although no pairwise contrasts were significant (p ≥ 0.333), and it differed between sites nested within islands (R = 0.200, p = 0.001). A similar result was obtained for the separate analysis of the 18 coral genera and Millepora, which revealed differences among islands (ANOSIM, R = 0.542, p = 0.010) and between sites within islands (R = 0.044, p = 0.001). In univariate analyses, the cover of scleractinians, macroalgae, and CTB differed among islands, and macroalgae and CTB differed between sites for each island (Table 1, p < 0.001). The seven common genera of scleractinians and Millepora also varied among islands (p < 0.001), but only Acropora differed between sites within islands (Table 2). Table 1. Statistical analysis of coral reef community structure at 10 m depth on four islands with two sites island−1 (Fig. 1). The percentage cover of scleractinians corals, macroalgae, and CTB was compared in a two factor ANOVA with sites as a random factor nested in islands. Data were arcsine transformed prior to analysis using photoquadrats as replicates. Dependent variable Factor df MS F p Scleractinians Island 3 2.060 129.232 <0.001 Sites(Islands) 4 0.028 1.740 0.141 Error 314 0.016 Macroalgae Island 3 0.540 27.367 <0.001 Sites(Islands) 4 1.036 52.534 <0.001 Error 314 0.020 CTB Island 3 2.105 122.874 <0.001 Sites(Islands) 4 0.578 33.728 <0.001 Error 314 0.0171 Table 2. Statistical analysis of community (scleractinians and Millepora) structure by common genera on four islands with two sites island−1 (Fig. 1). Percentage cover by genus was compared in a two factor ANOVA with sites as a random factor nested in islands. Data were arcsine transformed and the analysis completed using photoquadrats as replicates. Dependent variable Factor df MS F p Pocillopora Island 3 1.166 96.167 <0.001 Sites(Islands) 4 0.017 1.394 0.236 Error 314 Porites Island 3 0.773 62.680 <0.001 Sites(Islands) 4 0.017 1.413 0.229 Error 314 Astrea Island 3 0.122 29.730 <0.001 Sites(Islands) 4 0.005 1.207 0.308 Error 314 Montipora Island 3 0.246 33.805 <0.001 Sites(Islands) 4 0.011 1.476 0.209 Error 314 Millepora Island 3 0.018 2.204 0.088 Sites(Islands) 4 0.012 1.453 0.216 Error 314 Leptastrea Island 3 0.100 20.601 <0.001 Sites(Islands) 4 0.008 1.667 0.157 Error 314 Acropora Island 3 0.129 33.228 <0.001 Sites(Islands) 4 0.029 7.422 <0.001 Error 314 Pavona Island 3 0.065 21.987 <0.001 Sites(Islands) 4 0.003 1.108 0.353 Error 314 Scleractinian population genetics Sampling of putative P. meandrina at Tetiaroa, Maiao, and Moorea yielded 46, 34, and 55 biopsies (respectively) from different colonies that were processed for scleractinian host genotypes. Five discrete mitochondrial lineages of Pocillopora spp. were found: (1) P. verrucosa, (2) a lineage shared by individuals identified as either P. meandrina or P. eydouxi as adults, (3) Pocillopora effusus [species identifications after Schmidt-Roach et al. (2014)] and two haplotype lineages for which no consistent morphological identification has been made of adults, designated haplotype 8 and 11 (after Forsman et al. 2013; Pinzón et al. 2013; Figs. 5, 6). The proportion of mitochondrial lineages varied among the islands, with higher haplotype diversity at M