Microsatellite loci isolated from the Caribbean coral,Montastraea annularis

We report the isolation and characterization of seven microsatellite loci from the Caribbean reef‐building coral, Montastraea annularis. All loci are polymorphic with allele numbers ranging from five to 31 and observed heterozygosities from 0.17 to 0.89. These loci can be used in assessing gene flow patterns and diversity of this stony coral species both for local coral reef management purposes as well as for elucidating population connectivity within the greater Caribbean basin. These markers should also be applicable to other species of Montastraea and for resolving taxonomic relationships within the M. annularis species complex.     

Quantifying the colony shape of the Montastraea annularis species complex

The reef coral Montastraea annularis has been used in a wide range of investigations. Recently, it has been recognized as a complex of three species based on field observations of the variation in colony shape. These observations have also been confirmed by molecular methods as well as morphometrics on individual corallites in the colonies. This paper presents a new quantitative method for measuring overall colony shape based on geostatistics. Seventeen colonies collected from San Blas, Panama in 1995 and 1996 were examined using the Polhemus 3SPACE FASTRAK system to construct three-dimensional coordinates of the center of several hundred corallites in each colony. This method measures the larger bumps or “ridges” as well as the smaller bumps or “lumps” of the colonies. Variograms were then calculated for all the specimens and the lag distances and the values of the variogram were used in a multivariate statistical analysis. Overall, this method indicated that M. annularis and Montastaea faveolata overlap in their relative amount of bumpiness in colony shape while Montastraea franksi is distinct from the other two species. This method has implications for both the modern and fossil record of Montastraea as well as other organisms with similar shapes.     

Herbivory versus corallivory: are parrotfish good or bad for Caribbean coral reefs?

With coral cover in decline on many Caribbean reefs, any process of coral mortality is of potential concern. While sparisomid parrotfishes are major grazers of Caribbean reefs and help control algal blooms, the fact that they also undertake corallivory has prompted some to question the rationale for their conservation. Here the weight of evidence for beneficial effects of parrotfishes, in terms of reducing algal cover and facilitating demographic processes in corals, and the deleterious effects of parrotfishes in terms of causing coral mortality and chronic stress, are reviewed. While elevated parrotfish density will likely increase the predation rate upon juvenile corals, the net effect appears to be positive in enhancing coral recruitment through removal of macroalgal competitors. Parrotfish corallivory can cause modest partial colony mortality in the most intensively grazed species of Montastraea but the generation and healing of bite scars appear to be in near equilibrium, even when coral cover is low. Whole colony mortality in adult corals can lead to complete exclusion of some delicate, lagoonal species of Porites from forereef environments but is only reported for one reef species (Porites astreoides), for one habitat (backreef), and with uncertain incidence (though likely <<10%). No deleterious effects of predation on coral growth or fecundity have been reported, though recovery of zooxanthellae after bleaching events may be retarded. The balance of evidence to date finds strong support for the herbivory role of parrotfishes in facilitating coral recruitment, growth, and fecundity. In contrast, no net deleterious effects of corallivory have been reported for reef corals. Corallivory is unlikely to constrain overall coral cover but contraints upon dwindling populations of the Montastraea annularis species complex are feasible and the role of parrotfishes as a vector of coral disease requires evaluation. However, any assertion that conservation practices should guard against protecting corallivorous parrotfishes appears to be unwarranted at this stage.     

Parrotfish abundance and selective corallivory on a Belizean coral reef

Parrotfish are important members of coral reef communities because they consume macroalgae that would otherwise outcompete reef-building corals for space. However, some Caribbean parrotfish species also feed directly on live corals, and thus have the potential to negatively impact coral fitness and survival. This study investigates selective grazing by parrotfish on particular coral species, differences in grazing incidence among reef habitats and intraspecific discrimination among colonies of several coral species. We also investigate spatial and temporal patterns of parrotfish species abundance across habitats on the Belize barrier reef, and examine correlations between parrotfish abundance and grazing intensity across reef habitats. We found that members of the Montastraea annularis species complex, major builders of Caribbean reefs, were preferred targets of parrotfish grazing across all reef habitats, while M. cavernosa, Agaricia agaricites, Diploria strigosa, Porites astreoides and Porites porites were not preferred; Siderastrea siderea was preferentially grazed only in the spur and groove habitats. Parrotfish grazing preferences varied across habitats; M. annularis was grazed most often in shallow habitats, whereas M. franksi was consumed more at depth. Although it was not possible to directly observe parrotfish grazing on corals, we did find a positive correlation between Sparisoma aurofrenatum abundance and M. franksi grazing incidence across habitats. Finally, when we compared our results to parrotfish abundances measured by a previous study, we found that Sparisoma viride and Sp. aurofrenatum, two species known to be corallivorous, had increased abundances between 1982 and 2004. In light of escalating threats on Caribbean reef corals, it would be important for future studies to evaluate the impact of parrotfish corallivory on coral survival.     

Pleistocene reef environments, constituent grains, and coral community structure: Curaçao, Netherlands Antilles

We investigated the degree to which component grains vary with depositional environment in sediments from three reef habitats from the Pleistocene (125?ka) Hato Unit of the Lower Terrace, Curaçao, Netherlands Antilles: windward reef crest, windward back reef, and leeward reef crest. The windward reef crest sediment is the most distinctive, dominated by fragments of encrusting and branching coralline red algae, coral fragments and the encrusting foraminiferan Carpenteria sp. Windward back reef and leeward reef crest sediments are more similar compositionally, only showing significant differences in relative abundance of coral fragments and Homotrema rubrum. Although lacking high taxonomic resolution and subject to modification by transport, relative abundance of constituent grain types offers a way of assessing ancient skeletal reef community composition, and one which is not limited to a single taxonomic group. The strong correlation between grain type and environment we found in the Pleistocene of Curaçao suggests that constituent grain analysis may be an effective tool in delineating Pleistocene Caribbean reef environments. However, it will not be a sufficient indicator where communities vary significantly within reef environments or where evolutionary and/or biogeographical processes lead to different relationships between community composition and reef environment. Detailed interpretation of geological, biological, and physical characteristics of the Pleistocene reefs of Curaçao reveals that the abundance of the single coral species, Acropora palmata, is not a good predictor of the ecological structure of the ancient reef coral communities. This coral was the predominant species in two of the three reef habitats (windward and leeward reef crest), but the taxonomic composition (based on species relative abundance data) of the reef coral communities was substantially different in these two environments. We conclude that qualitative estimates of coral distribution patterns (presence of a key coral species or the use of a distinctive coral skeletal architecture), when used as a component in a multi-component analysis of ancient reef environments, probably introduces minimal circular reasoning into quantitative paleoecological studies of reef coral community structure.     

Growth of Pleistocene massive corals in south Florida: low skeletal extension-rates and possible ENSO,decadal, and multi-decadal cyclicities

Skeletal extension-rates and their variability are significantly lower in Pleistocene massive reef-building corals (Montastraea annularis group) in south Florida as compared with modern corals of the same taxa in the same study area. We analyzed 1,429 annual increments in 18 cores of Montastraea colonies from Windley Key and the Key Largo Waterway in the late Pleistocene Key Largo Limestone, which was deposited during marine isotope stage 5e (ca. 125 kyrs BP). The average extension-rate is 5.2 mm/year, which is about half the value known for modern Montastraea in shallow water reef environments. With an average standard deviation (SD) of 1.01, the variability of extension-rates is at the lower range limit of modern Montastraea in south Florida (SD = 1–3.6). Due to the higher sea level, the Pleistocene Key Largo patch reef trend was located on a large carbonate platform. Unlike today, the island chain of the Florida Keys, which function as a shelter for the Florida Reef Tract from inimical bank water, was not in existence. Corals probably grew under higher-than-present sea surface temperatures, which resulted in comparably low skeletal extension-rates. The detection of 3–7 year, decadal, and multi-decadal cyclicities in extension-rate time series suggests that the major modes of modern tropical climate variability such as the El Niño-Southern Oscillation (ENSO) and possibly the Atlantic Multidecadal Oscillation (AMO) were in effect during the last interglacial.     

Response of Pleistocene Coral Reefs to Environmental Change Over Long Temporal Scales

SYNOPSIS. TWO studies from the Pleistocene coral reef fossilrecord demonstrate the sensitivity of reef communities to bothlocal environmental parameters and habitat reduction. In thefirst study, Pleistocene reef coral assemblages from Papua NewGuinea show pronounced constancy in taxonomic composition andspecies diversity between 125 and 30 ka (thousand years). Spatialdifferences in reef coral community composition during successivehigh stands of sea level were greater among sites of the sameage than among reefs of different ages, even though global changesin sea level, atmospheric CO₂ concentration, tropical benthichabitat area, and temperature varied at each high sea levelstand. Thus, local environmental variation associated with runofffrom the land had greater influence on reef coral communitycomposition than variation in global climate and sea level.Proportional sampling from a regional species pool does notexplain the temporal persistence and local factors likely playeda major role. Examination of coral reef response to global changeshould not only involve regional diversity patterns but alsolocal ecological factors, and the interactive effects of localand global environmental change. In the second study, Pleistocene extinction of two widespread,strictly insular species of Caribbean reef corals, Pocilloporacf. palmata (Geister, 1975) and an organ-pipe growth form ofthe Montastraea "annularis" species complex, was natural anddid not involve gradual decrease in range and abundance, butwas sudden (thousands of years) throughout the entire range.One explanation is that sea level drop at the Last Glacial Maximum(LGM—18 ka) resulted in a threshold of habitat reduction,and caused disruption of coral metapopulation structure. Thresholdeffects predicted by metapopulation dynamics may also explainthe apparent paradox of the large amount of degraded modernreef habitat without any known modern-day reef coral extinctions.The rapid extinction of widespread Pleistocene species emphasizesthe vulnerability of reef corals in the face of present rapidenvironmental and climatic change.     

Analysis of complete mitochondrial DNA sequences of three members of the Montastraea annularis coral species complex (Cnidaria, Anthozoa, Scleractinia)

Complete mitochondrial nucleotide sequences of two individuals each of Montastraea annularis, Montastraea faveolata, and Montastraea franksi were determined. Gene composition and order differed substantially from the sea anemone Metridium senile, but were identical to that of the phylogenetically distant coral genus Acropora. However, characteristics of the non-coding regions differed between the two scleractinian genera. Among members of the M. annularis complex, only 25 of 16,134 base pair positions were variable. Sixteen of these occurred in one colony of M. franksi, which (together with additional data) indicates the existence of multiple divergent mitochondrial lineages in this species. Overall, rates of evolution for these mitochondrial genomes were extremely slow (0.03–0.04% per million years based on the fossil record of the M. annularis complex). At higher taxonomic levels, patterns of genetic divergence and synonymous/nonsynonymous substitutions suggest non-neutral and unequal rates of evolution between the two lineages to which Montastraea and Acropora belong.     

Population structure and connectivity of the mountainous star coral,Orbicella faveolata,throughout the wider Caribbean region

As coral reefs continue to decline worldwide, it becomes ever more necessary to understand the connectivity between coral populations to develop efficient management strategies facilitating survival and adaptation of coral reefs in the future. Orbicella faveolata is one of the most important reef‐building corals in the Caribbean and has recently experienced severe population reductions. Here, we utilize a panel of nine microsatellite loci to evaluate the genetic structure of O. faveolata and to infer connectivity across ten sites spanning the wider Caribbean region. Populations are generally well‐mixed throughout the basin (F_ST = 0.038), although notable patterns of substructure arise at local and regional scales. Eastern and western populations appear segregated with a genetic break around the Mona Passage in the north, as has been shown previously in other species; however, we find evidence for significant connectivity between Curaçao and Mexico, suggesting that the southern margin of this barrier is permeable to dispersal. Our results also identify a strong genetic break within the Mesoamerican Barrier Reef System associated with complex oceanographic patterns that promote larval retention in southern Belize. Additionally, the diverse genetic signature at Flower Garden Banks suggests its possible function as a downstream genetic sink. The findings reported here are relevant to the ongoing conservation efforts for this important and threatened species, and contribute to the growing understanding of large‐scale coral reef connectivity throughout the wider Caribbean.     

Decadal‐scale rates of reef erosion following El Niño‐related mass coral mortality

As the frequency and intensity of coral mortality events increase under climate change, understanding how declines in coral cover may affect the bioerosion of reef frameworks is of increasing importance. Here, we explore decadal‐scale rates of bioerosion of the framework building coral Orbicella annularis by grazing parrotfish following the 1997/1998 El Niño‐related mass mortality event at Long Cay, Belize. Using high‐precision U‐Th dating and CT scan analysis, we quantified in situ rates of external bioerosion over a 13‐year period (1998–2011). Based upon the error‐weighted average U‐Th age of dead O. annularis skeletons, we estimate the average external bioerosion between 1998 and 2011 as 0.92 ± 0.55 cm depth. Empirical observations of herbivore foraging, and a nonlinear numerical response of parrotfish to an increase in food availability, were used to create a model of external bioerosion at Long Cay. Model estimates of external bioerosion were in close agreement with U‐Th estimates (0.85 ± 0.09 cm). The model was then used to quantify how rates of external bioerosion changed across a gradient of coral mortality (i.e., from few corals experiencing mortality following coral bleaching to complete mortality). Our results indicate that external bioerosion is remarkably robust to declines in coral cover, with no significant relationship predicted between the rate of external bioerosion and the proportion of O. annularis that died in the 1998 bleaching event. The outcome was robust because the reduction in grazing intensity that follows coral mortality was compensated for by a positive numerical response of parrotfish to an increase in food availability. Our model estimates further indicate that for an O. annularis‐dominated reef to maintain a positive state of reef accretion, a necessity for sustained ecosystem function, live cover of O. annularis must not drop below a ~5–10% threshold of cover.     

FISH-Flow: a quantitative molecular approach for describing mixed clade communities of Symbiodinium

Our understanding of reef corals and their fate in a changing climate is limited by our ability to monitor the diversity and abundance of the dinoflagellate endosymbionts that sustain them. This study combined two well-known methods in tandem: fluorescent in situ hybridization (FISH) for genotype-specific labeling of Symbiodinium and flow cytometry to quantify the abundance of each symbiont clade in a sample. This technique (FISH-Flow) was developed with cultured Symbiodinium representing four distinct clades (based on large subunit rDNA) and was used to distinguish and quantify these types with high efficiency and few false positives. This technique was also applied to freshly isolated symbionts of Orbicella faveolata and Orbicella annularis. Isolates from acutely bleached coral tissues had significantly lower labeling efficiency; however, isolates from healthy tissue had efficiencies comparable to cultured Symbiodinium trials. RNA degradation in bleaching samples may have interfered with labeling of cells. Nevertheless, we were able to determine that, with and without thermal stress, experimental columns of the coral O. annularis hosted a majority of clade B and B/C symbionts on the top and side of the coral column, respectively. We demonstrated that, for cultured Symbiodinium and Symbiodinium freshly isolated from healthy host tissues, the relative ratio of clades could be accurately determined for clades present at as low as 7 % relative abundance. While this method does not improve upon PCR-based techniques in identifying clades at background levels, FISH-Flow provides a high precision, flexible system for targeting, quantifying and isolating Symbiodinium genotypes of interest.     

Multimodal Optical Microscopy Methods Reveal Polyp Tissue Morphology and Structure in Caribbean Reef Building Corals

An integrated suite of imaging techniques has been applied to determine the three-dimensional (3D) morphology and cellular structure of polyp tissues comprising the Caribbean reef building corals Montastraeaannularis and M. faveolata. These approaches include fluorescence microscopy (FM), serial block face imaging (SBFI), and two-photon confocal laser scanning microscopy (TPLSM). SBFI provides deep tissue imaging after physical sectioning; it details the tissue surface texture and 3D visualization to tissue depths of more than 2 mm. Complementary FM and TPLSM yield ultra-high resolution images of tissue cellular structure. Results have: (1) identified previously unreported lobate tissue morphologies on the outer wall of individual coral polyps and (2) created the first surface maps of the 3D distribution and tissue density of chromatophores and algae-like dinoflagellate zooxanthellae endosymbionts. Spectral absorption peaks of 500 nm and 675 nm, respectively, suggest that M. annularis and M. faveolata contain similar types of chlorophyll and chromatophores. However, M. annularis and M. faveolata exhibit significant differences in the tissue density and 3D distribution of these key cellular components. This study focusing on imaging methods indicates that SBFI is extremely useful for analysis of large mm-scale samples of decalcified coral tissues. Complimentary FM and TPLSM reveal subtle submillimeter scale changes in cellular distribution and density in nondecalcified coral tissue samples. The TPLSM technique affords: (1) minimally invasive sample preparation, (2) superior optical sectioning ability, and (3) minimal light absorption and scattering, while still permitting deep tissue imaging.     

Use of fractal dimensions to quantify coral shape

A morphometrical method to quantify and characterize coral corallites using Richardson Plots and Kaye’s notion of fractal dimensions is presented. A Jurassic coral species (Aplosmilia spinosa) and five Recent coral species were compared using the Box-Counting Method. This method enables the characterization of their morphologies at calicular and septal levels by their fractal dimensions (structural and textural). Moreover, it is possible to determine differences between species of Montastraea and to tackle the high phenotypic plasticity of Montastraea annularis. The use of fractal dimensions versus conventional methods (e.g., measurements of linear dimensions with a calliper, landmarks, Fourier analyses) to explore a rugged boundary object is discussed. It appears that fractal methods have the potential to considerably simplify the morphometrical and statistical approaches, and be a valuable addition to methods based on Euclidian geometry.     

Interspecific variation in distributions and diets of coral reef butterflyfishes (Teleostei: Chaetodontidae)

This study examined within‐reef distributions for 19 species of butterflyfishes (Chaetodontidae) at Lizard Island, northern Great Barrier Reef, Australia, and compared spatial patterns of abundance among species with contrasting diets. Spatial variation in abundance of butterflyfishes was most prominent among physiognomic reef zones mainly due to significant zonation of eight species, including four obligate hard‐coral feeders (Chaetodon trifascialis, Chaetodon baronessa, Chaetodon plebeius and Chaetodon lunulatus) and four generalist species (Chaetodon auriga, Chaetodon citrinellus, Chaetodon kleinii and Chaetodon rafflesi). Distributions of obligate hard‐coral feeders were closely associated with spatial variation in percentage cover of scleractinian corals, but no more restricted compared with facultative hard‐coral feeders or non‐coral feeders. Species with highest dietary specialization (C. trifascialis and C. baronessa), however, exhibited the most pronounced zonation patterns and were restricted to habitats with greatest abundance of their preferred prey. While there are conspicuous links between dietary specialization v. spatial patterns in abundance of butterflyfishes, it remains unclear whether dietary specialization is the cause or consequence of more restricted distributions.     

Community patterns on a submerged barrier reef at Barbados,West Indies

A quantitative survey of a submerged barrier reef was undertaken in Barbados, West Indies, over a two year period (1971–73). Photo-line transects were employed to obtain coverage data on corals and other benthic organisms. Light, sedimentation, currents, oxygen, temperature and salinity were also monitored. Results indicate corals cover about 30 per cent of the bottom with living colonies; another 7 per cent is contributed by other zoobenthos. The most abundant coral species are Montastrea annularis, M. cavernosa and Siderastrea siderea, each contributing between 4 and 5 per cent of bottom cover. Light is the only physical factor monitored that correlated significantly with biomass; sedimentation may have a secondary effect. Most of the barrier reef is composed of mixed coral associations forming a biologically accommodated community. Comparisons are made between the barrier reef in Barbados and deeper reefs in Jamaica and Curacao. Reefs are, in the main, similar but coral species and community structure differences do occur.     

The Relative Significance of Host–Habitat,Depth, and Geography on the Ecology,Endemism, and Speciation of Coral Endosymbionts in the Genus Symbiodinium

Dinoflagellates in the genus Symbiodinium are among the most abundant and important group of eukaryotic microbes found in coral reef ecosystems. Recent analyses conducted on various host cnidarians indicated that Symbiodinium assemblages in the Caribbean Sea are genetically and ecologically diverse. In order to further characterize this diversity and identify processes important to its origins, samples from six orders of Cnidaria comprising 45 genera were collected from reef habitats around Barbados (eastern Caribbean) and from the Mesoamerican barrier reef off the coast of Belize (western Caribbean). Fingerprinting of the ribosomal internal transcribed spacer 2 identified 62 genetically different Symbiodinium. Additional analyses of clade B Symbiodinium using microsatellite flanker sequences unequivocally characterized divergent lineages, or “species,” within what was previously thought to be a single entity (B1 or B184). In contrast to the Indo-Pacific where host-generalist symbionts dominate many coral communities, partner specificity in the Caribbean is relatively high and is influenced little by the host’s apparent mode of symbiont acquisition. Habitat depth (ambient light) and geographic isolation appeared to influence the bathymetric zonation and regional distribution for most of the Symbiodinium spp. characterized. Approximately 80% of Symbiodinium types were endemic to either the eastern or western Caribbean and 40–50% were distributed to compatible hosts living in shallow, high-irradiance, or deep, low-irradiance environments. These ecologic, geographic, and phylogenetic patterns indicate that most of the present Symbiodinium diversity probably originated from adaptive radiations driven by ecological specialization in separate Caribbean regions during the Pliocene and Pleistocene periods.     

Parrotfish predation on massive Porites on the Great Barrier Reef

Parrotfish grazing scars on coral colonies were quantified across four reef zones at Lizard Island, Northern Great Barrier Reef (GBR). The abundance of parrotfish grazing scars was highest on reef flat and crest, with massive Porites spp. colonies having more parrotfish grazing scars than all other coral species combined. Massive Porites was the only coral type positively selected for grazing by parrotfishes in all four reef zones. The density of parrotfish grazing scars on massive Porites spp., and the rate of new scar formation, was highest on the reef crest and flat, reflecting the lower massive Porites cover and higher parrotfish abundance in these habitats. Overall, it appears that parrotfish predation pressure on corals could affect the abundance of preferred coral species, especially massive Porites spp, across the reef gradient. Parrotfish predation on corals may have a more important role on the GBR reefs than previously thought.     

Coral mass- and split-spawning at a coastal and an offshore Venezuelan reefs,southern Caribbean

This study aimed to evaluate potential differences in coral spawning behavior between a fringing coastal reef and an offshore reef in the southern Caribbean. For this, scleractinian and gorgonian colonies (N = 324) of 21 species were mapped along eight transects, each 10-m long, at two study sites located in Morrocoy and Los Roques National Parks, Venezuela. Observations were made between 19:30 and 23:00 from August 23 to 30 and from September 26 to 30, 2002. Ninety one colonies belonging to six hard coral and seven octocoral species spawned or planulated during this period. We were unable to observe any signs of reproductive activity in 95 colonies of nine species different from those that reproduced. Despite the differences in environmental conditions between the two sites, we observed high synchrony in the spawning behavior of seven coral species common to both reefs. The most striking result was the ability of colonies of Montastraea faveolata and Eusmilia fastigiata to split spawn up to three times, either in consecutive nights or in different months.     

Modeling vertical coral connectivity and mesophotic refugia

Whether mesophotic reefs will behave as refugia for corals threatened by global climate change and coastal development depends on vertical exchange of larvae between diverse habitats. Here we use a biophysical model of larval dispersal to estimate vertical connectivity of a broadcasting (Orbicella faveolata) and a brooding (Porites astreoides) species of coral in the US Virgin Islands. Modeling predicts subsidy to shallow areas by mesophotic larvae of both species based on local hydrology, adult reproductive characteristics, larval traits, and a wide range of scenarios developed to test depth-sensitive factors, such as fertilization rates and post-settlement survivorship. In extreme model scenarios of reduced fertilization and post-settlement survivorship of mesophotic larvae, 1–10 % local mesophotic subsidy to shallow recruitment is predicted for both species, which are demographically significant. Although direct vertical connectivity is higher for the broadcaster, the brooder demonstrates higher local multigenerational vertical connectivity, which suggests that local P. astreoides populations are more resilient than those of O. faveolata, and corroborates field studies. As shallow habitat degrades, mesophotic–shallow subsidy is predicted to increase for both species. This study is the first of its kind to simulate larval dispersal and settlement between habitats of different depths, and these findings have local, regional, and global implications for predicting and managing coral reef persistence in a changing climate.     

The effect of Hurricane Allen on the Bellairs fringing reef,Barbados

On August 4, 1980 Hurricane Allen damaged the northern fringing reef located just offshore from the Bellairs Research Institute of McGill University on the west coast of Barbados. During the summer and one winter month of 1981 a resurvey of the reef was done and the results were compared with a similar survey made in 1974. On a reefal scale the changes in the proportion of substrates before and after Hurricane Allen were statistically significant for all substrates and species with the exception of Millepora spp. Among the corals, Porites porites has been most affected by the catastrophe: its coverage was reduced by 96%. The other corals (Siderastrea sidera, Porites astreoides, Montastrea annularis, and Agaricia agaricites) were reduced by 75 to 25%. Onshore transport is suggested by the high abundance of Acropora cervicornis (20%) and Madracis mirabilis (30%) in the rubble on the fringing reef (the former only grows seaward of the reef). Offshore transport is suggested by the relatively poor representation of Porites porites (32%) in the rubble. Shannon-Weaver diversity indices of the corals dropped from 1.61 to 1.26. The changes in diversity are best explained by the intermediate disturbance hypothesis.