Safety in numbers and the spatial scaling of density-dependent mortality in a coral reef fish

In coral reef fishes, density-dependent population regulation is commonly mediated via predation on juveniles that have recently settled from the plankton. All else being equal, strong density-dependent mortality should select against the formation of high-density aggregations, yet the juveniles of many reef fishes aggregate. In light of this apparent contradiction, we hypothesized that the form and intensity of density dependence vary with the spatial scale of measurement. Individual groups might enjoy safety in numbers, but predators could still produce density-dependent mortality at larger spatial scales. We investigated this possibility using recently settled juvenile bluehead wrasse, Thalassoma bifasciatum, a small, aggregating reef fish. An initial caging experiment demonstrated that juvenile bluehead wrasse settlers suffer high predation, and spatial settlement patterns indicated that bluehead wrasse juveniles preferentially settle in groups, although they are also found singly. We then monitored the mortality of recently settled juveniles at two spatial scales: microsites, occupied by individual fish or groups of fish and separated by centimeters, and sites, consisting of approximately 2400-m2 areas of reef and separated by kilometers. At the microsite scale, we measured group size and effective population density independently and found that per capita mortality decreased with group size but was not related to density. At the larger spatial scale, however, per capita mortality increased with settler density. This shift in the form of density dependence with spatial scale could reconcile the existence of small-scale aggregative behavior typical of many reef fishes with the population-scale density dependence that is essential to population stability and persistence.     

Density dependence in marine fish populations revealed at small and large spatial scales

Experimental manipulation of population density has frequently been used to demonstrate demographic density dependence. However, such studies are usually small scale and typically provide evidence of spatial (within-generation) density dependence. It is often unclear whether small-scale, experimental tests of spatial density dependence will accurately describe temporal (between-generation) density dependence required for population regulation. Understanding the mechanisms generating density dependence may provide a link between spatial experiments and temporal regulation of populations. In this study, I manipulated the density of recently settled kelp rockfish (Sebastes atrovirens) in both the presence and absence of predators to test for density-dependent mortality and whether predation was the mechanism responsible. I also examined mortality of rockfish cohorts within kelp beds throughout central California to evaluate temporal (between-generation) density dependence in mortality. Experiments suggested that short-term behavioral responses of predators and/or a shortage of prey refuges caused spatial density dependence. Temporal density dependence in mortality was also detected at larger spatial scales for several species of rockfish. It is likely that short-term responses of predators generated both spatial and temporal density dependence in mortality. Spatial experiments that describe the causal mechanisms generating density dependence may therefore be valuable in describing temporal density dependence and population regulation.     

Historical population demography of red snapper (<Emphasis Type="Italic">Lutjanus campechanus</Emphasis>) from the northern Gulf of Mexico based on analysis of sequences of mitochondrial DNA

We evaluated stock structure and demographic (population) history of red snapper (Lutjanus campechanus) in the northern Gulf of Mexico (Gulf) via analysis of mitochondrial (mt)DNA sequences from 360 individuals sampled from four cohorts (year classes) at three localities across the northern Gulf. Exact tests of genetic homogeneity and analysis of molecular variance both among cohorts within localities and among localities were non-significant. Nested clade analysis provided evidence of different temporal episodes of both range expansion and restricted gene flow due to isolation by distance. A mismatch distribution of pairwise differences among mtDNA haplotypes and a maximum-likelihood coalescence analysis indicated a population expansion phase that dated to the Pleistocene and probably represents (re)colonization of the continental shelf following glacial retreat. The spatial distribution of red snapper in the northern Gulf appears to have a complex history that likely reflects glacial advance/retreat, habitat availability and suitability, and hydrology. Habitat availability/suitability and hydrology may partially restrict gene flow among present-day red snapper in the northern Gulf and give rise to a metapopulation structure with variable demographic connectivity. This type of population structure may be difficult to detect with commonly used, selectively neutral genetic markers.Communicated by P.W. Sammarco, Chauvin     

Population dynamics of natural colonies of <Emphasis Type="Italic">Aurelia</Emphasis> sp. scyphistomae in Tasmania,Australia

The aim of this study was to identify potential environmental controls of the asexual phases of reproduction by measuring the rates of asexual reproduction (budding and strobilation) and mortality in naturally occurring populations of Aurelia sp. scyphistomae at different spatial and temporal scales. The percentage cover, density of colonies of Aurelia sp. scyphistomae, and density of the population of two naturally occurring colonies of Aurelia sp. scyphistomae were examined over 2 years in southern Tasmania. Artificial substrates were also deployed to investigate colony dynamics when density dependent effects were reduced. Clear spatial and temporal differences in the population dynamics of the colonies were observed. Density dependent effects controlled budding and recruitment of new scyphistomae to the substrate when populations were dense and space limiting. In contrast, environmental controls of budding and strobilation were more apparent in a colony with significantly greater area of bare substrate and hence room for expansion. Water temperature and rainfall (as a proxy for salinity) were linked to changes in population size. Annual and seasonal differences in population dynamics were not observed in a colony limited by space but were apparent in a colony where space was not limited. When space was removed as a limiting factor by deploying artificial substrates, a seasonal environmental effect on the rate of growth of the colony was observed. These studies suggest that the growth, survival and reproduction of the sessile colonial phase of Aurelia sp. is regulated by a combination of density dependent factors and environmental conditions, which are consequently important to the formation of jellyfish blooms.     

Population dynamics of the acorn barnacles, Tetraclita squamosa and Tetraclita japonica (Cirripedia: Balanomorpha), in Hong Kong

Tetraclita squamosa and Tetraclita japonica are common, intertidal barnacles. In Hong Kong, the population dynamics of the two species exhibited spatial and temporal variation on two semi-exposed shores. T. squamosa produced egg masses from May–June and annual settlement and recruitment occurred from June–July. In contrast, settlement and recruitment of T. japonica was sparse from March–May, but intensified in July–October. Mature gonads and egg masses were, however, only present from September–November, suggesting the larvae from the two settlement pulses originated from other locations as well as Hong Kong populations. Settlement intensity and post-recruitment mortality of the two species varied between sites, possibly due to spatial variation in free rock space, physical transport of larvae and abundance of food in the water column.Communicated by T. Ikeda, Hakodate     

Naticid snail predation on early post-settlement surfclams (<Emphasis Type="Italic">Spisula solidissima</Emphasis>) on the inner continental shelf of New Jersey,USA

Naticid gastropod predators leave characteristic boreholes in the shells of their bivalve prey that allow the evaluation of spatial differences in occurrence and intensity of predation. This approach, extensively used in paleobiological research, was used to compare spatial variation in predation by moonsnails (Euspira heros Say) on initial recruits of the Atlantic surfclam (Spisula solidissima Dillwyn). A four-year sampling at two 12-m deep stations on either flank of Beach Haven Ridge on the inner continental shelf detected temporal and spatial differences in surfclam densities following a large settlement pulse in early July. Between-station density differences were large during the first peak in surfclam density, but decreased in August-September, when densities at both stations also decreased. Surfclam seasonal peaks were followed by peaks in density of articulated shells (recent mortality due to non-crustacean predators and/or intrinsic mortality) and surfclams with boreholes (mortality due to naticids). The proportion of surfclams with boreholes was consistently higher (more than twice, on average) at the station where the highest densities of surfclams were also detected. These spatial differences are interpreted as differences in predation intensity resulting from a corresponding higher abundance of moonsnails at the same station. Although naticid predation is not the primary source of surfclam mortality, it consistently contributes to the reduction of spatial differences in density initially created by dissimilar levels of larval surfclam settlement.     

Time and space: genetic structure of the cohorts of the common sea urchin <Emphasis Type="Italic">Paracentrotus</Emphasis><Emphasis Type="Italic">lividus</Emphasis> in Western Mediterranean

Spatio-temporal variability in settlement and recruitment, high mortality during the first life-history stages, and selection may determine the genetic structure of cohorts of long-lived marine invertebrates at small scales. We conducted a spatial and temporal analysis of the common Mediterranean Sea urchin Paracentrotus lividus to determine the genetic structure of cohorts at different scales. In Tossa de Mar (NW Mediterranean), recruitment was followed over 5 consecutive springs (2006–2010). In spring 2008, recruits and two-year-old individuals were collected at 6 locations along East and South Iberian coasts separated from 200 to over 1,100 km. All cohorts presented a high genetic diversity based on a fragment of mtCOI. Our results showed a marked genetic homogeneity in the temporal monitoring and a low degree of spatial structure in 2006. In 2008, coupled with an abnormality in the usual circulation patterns in the area, the genetic structure of the southern populations studied changed markedly, with arrival of many private haplotypes. This fact highlights the importance of point events in renewing the genetic makeup of populations, which can only be detected through analysis of the cohort structure coupling temporal and spatial perspectives.     

Effect of prey densities on cannibalism in Cape hake (Merluccius capensis) off Namibia

The preference of large (>60 cm total length) Cape hake (Merluccius capensis Castelnau, 1861) for small conspecifics under natural conditions was analyzed off southern Namibia (S.E. Atlantic) between 1983 and 1989. The results indicate that cannibalism is not related to the density of the small conspecifics nor to the density of alternative prey, but rather that large hake have a dietary preference for small conspecifics. The main consequence of such cannibalism is a lack of density-dependent regulation of the M. capensis population.     

Temporal and spatial variability in settlement of the sea urchin<Emphasis Type="Italic"> Paracentrotus lividus</Emphasis> in the NW Mediterranean

Settlement into the benthic habitat may be an important process in regulating sea urchin abundance, which potentially modifies the structure of benthic communities. Strong settlement events may increase sea urchin abundance beyond a certain threshold, leading to the formation of coralline barrens (overgrazed communities with a dominance of encrusting coralline algae). To understand the role of settlement in regulating sea urchin populations we first need to determine settlement variability. Temporal variation in settlement of the sea urchin Paracentrotus lividus was monitored at three sites in the Medes Islands, NW Mediterranean, during three settlement seasons (March 1998 through October 2000). Spatial variation in settlement was studied in 1999 at 50 sites along a gradient of exposures to waves and currents, inside and outside the archipelago, and separated by distances from tens to thousands of meters. Bathymetric distribution of settlement was also studied in 2000 at six sites at 5, 10, 15 and 20 m depths. Settlement of P. lividus occurred in a single annual peak within 3 weeks in May–June. Differences in settlement between years were more than two orders of magnitude. Spatial variability was found at all scales investigated, showing strong patchiness at the smallest spatial scales (tens of meters). Sea urchins settled preferentially at depths between 5 and 10 m. Substratum type, level of protection, and adult population densities were not significant in determining settlement. However, settlement was found to be related to the degree of exposure to waves and currents, indicating that physical processes are very important at the spatial scales investigated. This greatly variable settlement is a necessary, although not sufficient, condition to create gradients of adult P. lividus abundance. Further studies should be designed to investigate the interaction between settlement strength and post-settlement mortality.Communicated by O. Kinne, Oldendorf/Luhe     

Vertebrate herbivory impacts seedling recruitment more than niche partitioning or density-dependent mortality

In tropical forests, resource-based niches and density-dependent mortality are mutually compatible mechanisms that can act simultaneously to limit seedling populations. Differences in the strengths of these mechanisms will determine their roles in maintaining species coexistence. In the first assessment of these mechanisms in a Congo Basin forest, we quantified their relative strengths and tested the extent to which density-dependent mortality is driven by the distance-dependent behavior of seed and seedling predators predicted by the Janzen-Connell hypothesis. We conducted a large-scale seed addition experiment for five randomly selected tropical tree species, caging a subset of seed addition quadrats against vertebrate predators. We then developed models to assess the mechanisms that determine seedling emergence (three months after seed addition) and survival (two years after seed addition). As predicted, both niche differentiation and density-dependent mortality limited seedling recruitment, but predation had the strongest effects on seedling emergence and survival. Seedling species responded differently to naturally occurring environmental variation among sites, including variation in light levels and soil characteristics, supporting predictions of niche-based theories of tropical tree species coexistence. The addition of higher densities of seeds into quadrats initially led to greater seedling emergence, but survival to two years decreased with seed density. Seed and seedling predation reduced recruitment below levels maintained by density-dependent mortality, an indication that predators largely determine the population size of tree seedlings. Seedling recruitment was unrelated to the distance to or density of conspecific adult trees, suggesting that recruitment patterns are generated by generalist vertebrate herbivores rather than the specialized predators predicted by the Janzen-Connell hypothesis. If the role of seed and seedling predation in limiting seedling recruitment is a general phenomenon, then the relative abundances of tree species might largely depend on species-specific adaptations to avoid, survive, and recover from damage induced by vertebrate herbivores. Likewise, population declines of herbivorous vertebrate species (many of which are large and hunted) may trigger shifts in species composition of tropical forests.     

Population structure and genetic variation of lane snapper (<Emphasis Type="Italic">Lutjanus synagris</Emphasis>) in the northern Gulf of Mexico

Lane snappers (Lutjanus synagris), sampled from eight localities in the northern Gulf of Mexico (Gulf) and one locality along the Atlantic coast of Florida, were assayed for allelic variation at 14 nuclear-encoded microsatellites and for sequence variation in a 590 base-pair fragment of the mitochondrially encoded ND-4 gene (mtDNA). Significant heterogeneity among the nine localities in both microsatellite allele and genotype distributions and mtDNA haplotype distributions was indicated by exact tests and by analysis of molecular variance (AMOVA). Exact tests between pairs of localities and spatial analysis of molecular variance (SAMOVA) for both microsatellites and mtDNA revealed two genetically distinct groups: a Western Group that included six localities from the northwestern and northcentral Gulf and an Eastern Group that included three localities, one from the west coast of Florida, one from the Florida Keys, and one from the east (Atlantic) coast of Florida. The between-groups component of molecular variance was significant for both microsatellites (Φ _CT = 0.016, P = 0.009) and mtDNA (Φ _CT = 0.208, P = 0.010). Exact tests between pairs of localities within each group and spatial autocorrelation analysis did not reveal genetic heterogeneity or an isolation-by-distance effect among localities within either group. MtDNA haplotype diversity was significantly less (P < 0.0001) in the Western Group than in the Eastern Group; microsatellite allelic richness and gene diversity also were significantly less in the Western Group (P = 0.015 and 0.013, respectively). The difference in genetic variability between the two groups may reflect reduced effective population size in the Western Group and/or asymmetric rates of genetic migration. The relative difference in variability between the two groups was substantially greater in mtDNA and may reflect one or more mtDNA selective sweeps; tests of neutrality of the mtDNA data were consistent with this possibility. Bayesian analysis of genetic demography indicated that both groups have experienced a historical decline in effective population size, with the decline being greater in the Western Group. Maximum-likelihood analysis of microsatellite data indicated significant asymmetry in average, long-term migration rates between the two groups, with roughly twofold greater migration from the Western Group to the Eastern Group. The difference in mtDNA variability and the order-of-magnitude difference in genetic divergence between mtDNA and microsatellites may reflect different demographic events affecting mtDNA disproportionately and/or a sexual and/or spatial bias in gene flow and dispersal. The spatial discontinuity among lane snappers in the region corresponds to a known zone of vicariance in other marine species. The evidence of two genetically distinct groupings (stocks) has implications for management of lane snapper resources in the northern Gulf.     

Predation, habitat complexity, and variation in density-dependent mortality of temperate reef fishes

Density dependence in demographic rates can strongly affect the dynamics of populations. However, the mechanisms generating density dependence (e.g., predation) are also dynamic processes and may be influenced by local conditions. Understanding the manner in which local habitat features affect the occurrence and/or strength of density dependence will increase our understanding of population dynamics in heterogeneous environments. In this study I conducted two separate field experiments to investigate how local predator density and habitat complexity affect the occurrence and form of density-dependent mortality of juvenile rockfishes (Sebastes spp.). I also used yearly censuses of rockfish populations on nearshore reefs throughout central California to evaluate mortality of juvenile rockfish at large spatial scales. Manipulations of predators (juvenile bocaccio, S. paucispinus) and prey (kelp, gopher, and black-and-yellow [KGB] rockfish, Sebastes spp.) demonstrated that increasing the density of predators altered their functional response and thus altered patterns of density dependence in mortality of their prey. At low densities of predators, the number of prey consumed per predator was a decelerating function, and mortality of prey was inversely density dependent. However, at high densities of predators, the number of prey killed per predator became an accelerating response, and prey mortality was directly density dependent. Results of field experiments and large-scale surveys both indicated that the strength of density-dependent mortality may also be affected by the structural complexity of the habitat. In small-scale field experiments, increased habitat complexity increased the strength of density-dependent mortality. However, at large scales, increasing complexity resulted in a decrease in the strength of density dependence. I suggest that these differences resulted from scale-dependent changes in the predatory response that generated mortality. Whether increased habitat complexity leads to an increase or a decrease in the strength of density-dependent mortality may depend on how specific predatory responses (e.g., functional or aggregative) are altered by habitat complexity. Overall, the findings of this study suggest that rates of demographic density dependence and the resulting dynamics of local populations may largely depend upon attributes of the local habitat.     

Seasonal nestmate recognition in the ant Formica exsecta

Under favorable conditions, the mound-building ant Formica exsecta may form polydomous colonies and can establish large nest aggregations. The lack of worker aggression towards nonnestmate conspecifics is a typical behavioral feature in such social organization, allowing for a free flux of individuals among nests. However, this mutual worker toleration may vary over the seasons and on spatial scales. We studied spatio-temporal variation of worker–worker aggression within and among nests of a polydomous F. exsecta population. In addition, we determined inter- and intracolony genetic relatedness by microsatellite DNA genotyping and assessed its effect on nestmate recognition. We found significant differences in the frequency of worker exchange among nests between spring, summer, and autumn. Moreover, we found significant seasonal variation in the level of aggression among workers of different nests. Aggression levels significantly correlated with spatial distance between nests in spring, but neither in summer nor in autumn. Multiple regression analysis revealed a stronger effect of spatial distances rather than genetic relatedness on aggressive behavior. Because nestmate discrimination disappeared over the season, the higher aggression in spring is most plausibly explained by cue intermixing during hibernation.     

Influence of substratum heterogeneity and settled barnacle density on the settlement of cypris larvae

On the Atlantic coast of Canada, Semibalanus balanoides (L.) is widely distributed in the mid-intertidal zone, whereas in the Gulf of St. Lawrence, this species is mostly limited to crevices. We tested the hypothesis of regional differences in microhabitat selection by barnacle larvae at settlement in 1984 and 1985 at St. Andrews, New Brunswick, Canada. Since larvae settle in microhabitats already colonized by adults, the relative influence of settled barnacle density and of different scales of substratum heterogeneity on settlement were evaluated experimentally at Capucins, Québec, (Gulf of St. Lawrence) and at St. Andrews, New Brunswick (Atlantic coast). On a large scale (>10 cm deep crevices) of heterogeneity, results show that, in the Gulf, cypris larvae settled nearly exclusively (93%) in natural crevices rather than on adjacent horizontal surfaces. On the Atlantic coast, settlement was more important outside than inside of crevices, when the substrata were either natural or artificial. This result is unique and contrasts sharply with all known reports on barnacle settlement in relation to surface contour. The influence of barnacle density on settlement was greater than that of large scale heterogeneity. On a small scale (<1.5 cm deep cracks), the presence of conspecifics had a stronger effect on settlement than heterogeneity in both regions. Field observations showed a relationship between larval settlement density and percentage of adult cover. Settlement increased up to 22 or 30% (Gulf and Atlantic coast) of adult cover and decreased afterwards. The results confirm the hypothesis of larval selection for cryptic habitats in the Gulf and the opposite behaviour (preferences for horizontal surfaces) on the Atlantic coast. This microhabitat selection is apparent at large scales of heterogeneity, whereas at small scales, the presence of conspecifics is the predominant factor.Contribution to the programme of GIROQ (Groupe interuniversitaire de recherches océanographiques du Québec)     

Spatial variation in population dynamics in a colonial ascidian (Botryllus schlosseri)

Recent interest in the dynamics of marine invertebrate populations has focused largely on taxa with an open population structure. However, in many colonial taxa with limited larval dispersal, settlers may be locally derived. Consequently, dynamics may vary among sites that are separated by relatively short distances. This study explored spatial variation in temporal dynamics of colonial ascidians (Botryllus schlosseri Pallas) inhabiting five sites distributed along a ≈ 17-km temperature and phytoplankton gradient in the Damariscotta River estuary, Maine, USA. Settlement and population densities and sexual reproductive status were assayed throughout the summer seasons of 1996 and 1997. Sexual reproduction and larval settlement commenced earlier in the summer in up-river populations, which subsequently underwent a seasonal population explosion that was much smalier in down-river populations. Two peaks in settlement density up-river (in early July and early September) suggest that colonies there may have completed two sexual generations, in contrast to a single generation at down-river sites. Similar spatial variation is expected among populations of other taxa with limited larval dispersal when they are distributed across environmental gradients. Published online: 18 September 2002     

Parasitism and a shortage of refuges jointly mediate the strength of density dependence in a reef fish

Various predator-prey, host-pathogen, and competitive interactions can combine to cause density dependence in population growth. Despite this possibility, most empirical tests for density-dependent interactions have focused on single mechanisms. Here we tested the hypothesis that two mechanisms of density dependence, parasitism and a shortage of refuges, jointly influence the strength of density-dependent mortality. We used mark-recapture analysis to estimate mortality of the host species, the bridled goby (Coryphopterus glaucofraenum). Sixty-three marked gobies were infected with a copepod gill parasite (Pharodes tortugensis), and 188 were uninfected. We used the spatial scale at which gobies were clustered naturally (approximately 4 m2) as an ecologically relevant neighborhood and measured goby density and the availability of refuges from predators within each goby's neighborhood. Goby survival generally declined with increasing density, and this decline was steeper for gobies with access to few refuges than for gobies in neighborhoods where refuges were common. The negative effects of high density and refuge shortage were also more severe for parasitized gobies than for gobies free of parasites. This parasite has characteristics typical of emerging diseases and appears to have altered the strength of a preexisting density-dependent interaction.     

Settlement, recruitment and potential predators and competitors of juvenile echinoderms in the rocky subtidal zone

Recruitment patterns of marine invertebrates are affected both by settlement and early post-settlement events. This study examined the settlement and recruitment patterns of echinoderms at three sites in the rocky subtidal zone of Bocabec Cove, Bay of Fundy, Canada using artificial turf collectors and quadrats on the natural substrate. Potential predators were quantified at two of the sites along transects and in 1-m² quadrats. Both potential predators and competitors were quantified in 0.0625-m² quadrats. Settlement varied across sites (1.5–3 km apart) and two years of sampling (2004, 2005). The site of most potential settlement differed for the three groups of echinoderms: sea urchins (Strongylocentrotus droebachiensis), sea stars (Asterias spp.) and sea cucumber (Psolus fabricii). Settlement densities on the artificial turf collectors tended to be greater than the densities of settlers on the natural substrate. On the natural substrate, the only significant difference between densities of juveniles over time was that newly settled sea stars were found in July and were not found the following October. Large lobsters and carnivorous worms were potential predators with densities that varied between sites. Potential competitors that differed in abundance between sites were herbivorous gastropods and conspecifics for sea urchins; and carnivorous worms for sea stars. This study suggests that patterns of recruitment are either set up by patterns of settlement or by events during the first few weeks/months on the benthic substrate for these echinoderms.     

Comparison of larval duration and pre- and post-settlement growth in two species of damselfish,Chromis atripectoralis andPomacentrus coelestis (Pisces: Pomacentridae), from the Great Barrier Reef

Light traps were used to capture larval fishes, immediately before settlement, at two localities 500 km apart on the Great Barrier Reef (GBR) in December, 1987. Samples from Lizard Island, in the northern GBR, and Davies Reef, in the central GBR, were dominated by two species of damselfish:Chromis atripectralis andPomacentrus coelestis. Analysis of otoliths revealed significant differences in both size and age at settlement between the two localities forP. coelestis, but not forC. atripectoralis. Growth rates determined for pre- and post-settlementP. coelestis suggested a sigmoidal growth trajectory through the larval life, with growth slowing as fishes approached the time of settlement. Post-settlement growth rates were faster than growth prior to settlement in both species. Growth in both species was, however, similar between localities. The relationship between fish size and otolith size was complex, varying both between pre- and post-settlement fishes, and among localities. This emphasizes the need to validate the relationship between fish size and otolith size before otoliths may be used to back-calculate individual growth trajectories.Contribution No. 500 from the Australian Institute of Marine Science     

Genetic structure of the subtidal red alga Delisea pulchra

We used random amplified polymorphic DNA (RAPDs) to examine small-scale spatial genetic structure in the red alga Delisea pulchra (Greville) Montagne at two locations near Sydney, Australia. We examined genetic structure among plants at four spatial scales ranging from 2 km apart down to <50 cm apart between locations, among sites within locations, among quadrats within sites, and among plants within quadrats. Haploid stages of D. pulchra were absent from the populations studied, suggesting that they are maintained through asexual reproduction of diploid plants. Consistent with this, we found that 19 RAPD phenotypes scored in this study had multiple individuals, indicating the presence of clones in these populations. However, there were no RAPD phenotypes common to two locations separated by only 2 km. Analysis of molecular variance revealed that strong genetic differences occurred between plants from these two locations, with 46.3% of the total genetic variation occurring at this scale, most probably reflecting limited gene flow. Within each location, <25% of the genetic variation was attributable to differences among sites or quadrats, indicating gene flow at those smaller scales. Most of the variation within each location occurred at the smallest spatial scale, among plants within 0.25 m² quadrats. Nonetheless, some pairwise genetic distances (φST) between sites or quadrats within locations were large, indicating some genetic divergence on smaller scales. Genetic distance was independent of spatial distance within both locations, suggesting that fine-scale differences within locations were most probably caused by variation in fine-scale patterns of water movement or fine-scale natural selection. We assessed the impact of one potential selective agent, grazing sea urchins, on the fine-scale genetic structure of D. pulchra. There was no evidence that grazing by sea urchins affected the genetic structure of D. pulchra. In combination with demographic data, our results indicated that local populations of D. pulchra within locations were relatively open and that fine-scale genetic structure was probably constrained by gene flow. At the larger scale however, strong genetic differentiation indicated little gene flow between locations and restricted dispersal of spores. Received: 22 April 1999 / Accepted: 29 November 1999