Changes in survivorship,behavior, and morphology in native soft‐shell clams induced by invasive green crab predators

Many studies on invasive species show reduced native densities, but few studies measure trait‐mediated effects as mechanisms for changes in native growth rates and population dynamics. Where native prey face invasive predators, mechanisms for phenotypic change include selective predation, or induced behavioral or morphological plasticity. Invasive green crabs, Carcinus maenas, have contributed to declines in native soft‐shell clams, Mya arenaria, in coastal New England, USA. We tested the hypothesis that clam ability to detect chemical cues from predators or damaged conspecifics would induce greater burrowing depth as a refuge from invasive crabs, and greater burrowing would require increased siphon growth. To determine how crab predation affected clam survivorship and phenotypic traits in the field, clams in exclosure, open, and crab enclosure plots were compared. Crab predation reduced clam density, and surviving clams were deeper and larger, with longer siphons. To determine whether the mechanism for these results was selective predation or induced plasticity, phenotypes were compared between clams exposed to chemical cues from crab predation and clams exposed to seawater in laboratory and field experiments. In response to crab predation cues, clams burrowed deeper, with longer siphons and greater siphon mass. Overall, crab predation removed clams with shorter siphons at shallow depths, and crab predation cues induced greater burrowing depths and longer siphons. Longer siphons and greater siphon mass of deeper clams suggests clams may allocate energy to siphon growth in response to crabs. By determining native behavior and morphological changes in response to an invasive predator, this study adds to our understanding of mechanisms for invasive impacts and illustrates the utility of measuring trait‐mediated effects to investigate predator–prey dynamics.     

Trade-off between safety and feeding in the sea anemone Anthopleura aureoradiata

After a physical disturbance, such as a predatory stimulus, several marine organisms withdraw into a refuge to ensure survival. Consequently, they cannot engage in other essential activities. Re-emergence reflects a trade-off between avoidance of disturbance and lost feeding opportunities. The New Zealand sea anemone (Anthopleura aureoradiata) avoids predators by withdrawing its tentacles into its upper body. In a field experiment, responses of 480 A. aureoradiata to a physical disturbance were studied at three locations to assess the influence of the following factors on hiding time: physical contact with a neighbour; density of nearby conspecifics; total density per pool; and body size. Withdrawal times were significantly shorter for anemones in contact with a conspecific, occurring in high-density patches or of larger size. These patterns are consistent with rapid re-emergence being stimulated by competition for food from immediate neighbours and/or a lower risk of predation when in groups (i.e. a dilution effect), with larger size also potentially reducing vulnerability to predation.     

Are all patterns created equal? Cooperation is more likely in spatially simple habitats

Cooperative behaviours, such as aggregation with neighbouring conspecifics, can enhance resilience in habitats where risks (i.e. predation, physical disturbances) are high, exerting positive feedback loops to maintain a healthy population. At the same time, cooperation behaviours can involve some extra energy expenditures and increasing resource competition. For sessile reefs, like mussels, simulation models predict increased cooperation under increasing levels of environmental stress. Predation risk is viewed as a behaviour‐modifying stressor, but its role on cooperation mechanisms, such as likelihood of reciprocity, has not yet been empirically tested. This study harnesses this framework to understand how cooperation changes under different perceived levels of predation risk, using mussel beds as model of a complex “self‐organised” system. Hence, we assessed the context dependency of cooperation response in different “landscapes of fear,” created by changes in predator cues, substratum availability and body size. Our experiments demonstrated that i) cooperation in a mussel bed system increases when predator cues are present, but that this relationship was found to be both, ii) strongly context‐dependent, particularly upon substratum availability and iii) size‐dependent. That is, while cooperation is in general greater for larger individuals, the response to risk results in greater cooperation when alternative attachment substratum is absent, meaning that simpler landscapes may be perceived as riskier. The context dependency of structural complexity is also an essential finding to consider in a changing world where habitats are losing complexity and cooperative strategies should be maximised.     

Scales of predator detection behavior and escape in Fissurella limbata: A field and laboratory assessment

The consumptive effects of predators are widely acknowledged, but predation can also impact prey populations through non‐consumptive effects (NCEs) such as costly antipredator behavioral responses. The magnitude of antipredator behavioral responses by prey is determined by an assessment of risk using sensory cues, which in turn is modulated by the environmental context. We studied the detection behavior and escape response of the keyhole limpet Fissurella limbata from the predatory sea star Heliaster helianthus. Through laboratory and field experimental trials, we quantified the distance and time of predator detection behavior by the prey, and measured their active escape responses when elicited. We found that predator detection by the limpet was chiefly mediated by distance, with experimental individuals capable of detecting predator presence effectively up to distances of at least 50 cm in the field and 70 cm under laboratory conditions. Our results indicate that this prey species is able to evaluate the proximity of its predator and use it as an indication of predation risk; therefore, predator–prey distance appears to be a primary predictor of the magnitude of the antipredator response. Given the tight relationship between predator distance and prey movement and the important role herbivores can play, particularly in this ecosystem, we expect that NCEs will cascade to the patterns of abundance and composition of rocky shore communities through changes in prey foraging behavior under risk.     

Sex differences in body morphology and multitrophic interactions involving the foraging behaviour of the crab Carcinus maenas

Many animals are sexually dimorphic, but the underlying evolutionary causes and ecological consequences of sexual dimorphism are not fully understood. One predicted consequence for sexual dimorphism is that different sexes show niche differentiation. If sexual dimorphism is in feeding appendages, then differences may be manifested by different diets and thus contrasting behavioural responses to potential prey. Sexual dimorphisms in feeding appendages may also result in different handling times, which may then be correlated with differences in exposure, and, hence predation risk to the predator while feeding. In addition, the prey of the sexually dimorphic predator may respond differently to cues from each sex according to the predation risk each presents to the prey. We tested these predictions using a crab (Carcinus maenas) with sexual dimorphism in chelae dimensions, its predator the cuttlefish Sepia officinalis and prey with known differences in handling times; the gastropod molluscs Gibbula umbilicalis and Littorina littorea. We demonstrated that male C. maenas orientated more frequently to cues from L. littorea whereas females orientated more towards G. umbilicalis in contradiction of patterns predicted by handling times. Male crabs had a faster heart rate than females but this was not influenced by food‐based cues. We also showed no difference in foraging times with respect to changing levels of predator risk and also no differences in gastropod responses to odours from male or female crabs. Our results showed that predictions of handling time and sexual dimorphism are not associated. The experiments indicated the male and female crabs are probably ecological equivalents and thus niche differentiation is less likely.     

Energy budgets and predation impact of dominant epibenthic carnivores on a shallow soft bottom community at the Swedish west coast

The energy flow through the epibenthic community on a shallow soft bottom in Gullmar Fjord, Sweden, is evaluated. The interaction between the dominant epibenthic predators: brown shrimp Crangon crangon L., juvenile plaice Pleuronectes platessa L. and sand goby Pomatoschistus minutus (Pallas), and their food supply is examined, and the fate of the biomass produced within the shallow water ecosystem is described. A steady state model is presented which describes the amounts, pathways and annual rates of energy flowing through the shallow water community. Field and laboratory studies are performed in order to validate the consumption rates of the fish and shrimp species. The brown shrimp occupies a key position in this ecosystem. It represents food, consisting of newly settled juveniles, both to its conspecifics and to other epibenthic species in shallow water. Sixty to eighty per cent of the annual production of shrimps in shallow water is consumed by the epibenthic predators. The brown shrimp constitutes a main predator on other epibenthic species and, by providing such feedback loops, the shrimp constitutes a highly significant regulator of epibenthic carnivores in shallow water. A loss from the benthic community will occur due to predation, maximally amounting to 24–34% of the annual production, which is a high value compared to similar areas. A dynamic simulation is also performed in order to calculate the seasonal change of energy in the epibenthic community in shallow water. Based on mortality calculations, the total output from the shallow nursery should amount to 0·6 g m⁻² year⁻¹ (dry wt).     

Predatory Promesostoma species (Plathelminthes, Rhabdocoela) in the Wadden Sea

Laboratory experiments were performed on the food ecology of four congeneric species of free-living plathelminths, Promesostoma caligulatum, P. marmoratum, P. rostratum, and P. meixneri, all inhabiting an intertidal sandflat near the island of Sylt (North Sea). Their prey spectrum is within the microcrustaceans: P. caligulatum preferred ostracods, while the other three species favoured copepods, with species-specific differences for copepod species and size classes. Daily consumption of prey species varied with the size of both the predator and the prey. On average, P. marmoratum consumed 0.76 Harpacticus flexus per day while this rate decreased to 0.06 in P. meixneri, the smallest predator. When these Promesostoma species were fed with Tachidius discipes, a smaller prey species, their predation rates were about 25% higher. While the larger predators preferred the larger harpacticoids as prey, the small P. meixneri preferred small cyclopoids over larger harpacticoids. In terms of biomass, P. marmoratum's mean consumption of T. discipes per day was about half the predator's own weight. This average varied with prey density and temperature. A comparison of these consumption rates with the field densities of the predators and their prey shows that the plathelminth predators may consume as much as 10% per day of their copepod prey populations, thus strongly influencing these prey populations on these sandflats. The predation pressure of P. caligulatum on ostracods was about 1% per day of the prey population. Since ostracods usually have fewer generations per year, the total effect on the population dynamics may be similar to that on copepods. Therefore, nocturnal swimming of copepods in the water column may be interpreted as an attempt to escape plathelminth predators.     

Temporal variation in cannibalistic infanticide by the shore crab Hemigrapsus oregonensis: implications for reproductive success

Larvae of benthic marine organisms are released amid high densities of suspension feeding and predatory adults and are highly subject to being consumed, even by conspecifics or their own parent. During laboratory feeding trials conducted in June 2006, female shore crabs (Hemigrapsus oregonensis) from Stege Marsh in San Francisco Bay (37°54.530′ N, 122°19.734′ W) that released their larvae during the previous 24 h ate fewer conspecific larvae than females that had not recently released larvae, though the behavior was not repeated during similar trials in 2007. Additionally, the number of larvae eaten increased with increasing starvation time, and hungrier females showed a trend toward eating more larvae from a different species (Carcinus maenas) than larvae of conspecifics. Thus, suppression of suspension feeding may reduce conspecific predation of newly released larvae, but this response partially depends on hunger level. This is the first time crabs have been shown to suppress feeding to reduce cannibalism of larvae, and this behavior could affect reproductive success and population dynamics.     

Predominance of parallel‐ and cross‐predation in anglerfish

Several studies in the last 20 years have revealed that morphological asymmetry in fish can be characterized as ‘antisymmetry’. Antisymmetry is a lateral dimorphism in which each population consists of individuals with well‐developed left sides (lefties) and well‐developed right sides (righties). This dimorphism influences predator–prey interactions. In some piscivorous fishes, it has been found that predators can catch more prey of the opposite morphological type to themselves (cross‐predation) than of the same morphological type (parallel‐predation). Our previous work clarified that the predominance of cross‐predation is caused by lateralized behaviors of predators and prey that correspond to their morphological antisymmetry. Moreover, based on the results of our behavioral observations, we hypothesized that parallel‐predation can predominate when predators encounter the potential prey frontally. To test this hypothesis, in the present study we investigated the relationship between lateral morphological types of anglerfish (Lophiomus setigerus) and those of the prey fishes found in their stomachs. Anglerfish attract potential prey using their first dorsal fin (illicium) as a lure, and their frontal encounters with potential prey fishes were photographed in situ and observed in an aquarium. The results of a stomach contents analysis indicated that parallel‐predation predominated in five benthopelagic prey fish species (perches and eels). By contrast, five benthic prey fishes (gobies and weevers) exhibited the predominance of cross‐predation. These results not only demonstrate the predominance of parallel‐predation in a natural fish community, but also suggest that the relationship between morphological types of predator and prey species can be reversed depending on the lifestyle of prey.     

Impact of predation on the polychaete Hediste diversicolor in estuarine intertidal flats

In estuarine sediment flats benthic macroinvertebrates are intensively consumed by a variety of predators, such as aquatic birds and nekton (mostly fish and crustaceans). However, there is still a lack of conclusive studies that evaluate if this predation has a relevant impact on the populations of those invertebrates, which are a key element of the estuarine food chain. In the Tagus estuary we experimentally tested and quantified the impact of predation on the polychaete Hediste diversicolor, one of the most important prey for a variety of predators in many estuaries. Using an exclusion experiment, we compared the seasonal variation in the densities of H. diversicolor from February to November in sediment plots (1) available to both bird and nekton predators, (2) just to nekton, and (3) without predators. We also followed changes in the abundance of potential predators throughout the study. The lowest densities were systematically observed in the plots accessible to all predators, followed by those which excluded just birds, and finally by those that excluded all predators. The exclosures were in place for 9 months, at the end of which the average density of H. diversicolor in the plots protected from all predators was eight times greater than in those without any protection. These results demonstrate that predation had a major impact on the densities of H. diversicolor. The relative importance of bird and nekton predation varied along the study, and this seems to be determined by different peaks of abundance of the two types of predators. However, when present in high densities, birds and nekton seem to have a similar impact on H. diversicolor. Our results suggest that predation is a key factor on the population dynamics of H. diversicolor. In addition, the levels of predation that we observed suggest that this polychaete can be a limited resource, and this could have major ecological consequences for predators for which it is a key prey.     

Predator and scavenger aggregation to discarded by-catch from dredge fisheries: importance of damage level

Predator and scavenger aggregation to simulated discards from a scallop dredge fishery was investigated in the north Irish Sea using an in situ underwater video to determine differences in the response to varying levels of discard damage. The rate and magnitude of scavenger and predator aggregation was assessed using three different types of bait, undamaged, lightly damaged and highly damaged individuals of the great scallop Pecten maximus. In each treatment scallops were agitated for 40 minutes in seawater to simulate the dredging process, then subjected to the appropriate damage level before being tethered loosely in front of the video camera. The density of predators and scavengers at undamaged scallops was low and equivalent to recorded periods with no bait. Aggregation of a range of predators and scavengers occurred at damaged bait. During the 24 hour period following baiting there was a trend of increasing magnitude of predator abundance with increasing damage level. However, badly damaged scallops were eaten quickly and lightly damaged scallops attracted a higher overall magnitude of predator abundance over a longer 4 day period. Large scale temporal variability in predator aggregation to simulated discarded biota was examined by comparison of results with those of a previous study, at the same site, 4 years previously. The community of predators and scavengers present at the bait was substantially different during the 2 studies. These changes in scavenger community were only partially related to changes in the mean background density of scavenging species.     

Food web and fish stock changes in central Chile: comparing the roles of jumbo squid (Dosidicus gigas) predation,the environment,and fisheries

We analyzed recent food web and fish stock changes in the central Chile marine ecosystem, comparing the roles of jumbo squid (Dosidicus gigas) as predator, the environment, and fishing. To accomplish this we used food web modeling and the Ecopath with Ecosim software (EwE). The principal fish stocks have experienced wide decadal fluctuations in the past 30 years, including stock collapses of horse mackerel (Trachurus murphyi) and hake (Merluccius gayi), and there was a large influx of jumbo squid during the mid-2000s. We used two EwE models representing the food web off central Chile to test the hypothesis that predation by jumbo squid has been significant in explaining the dynamics of the main fishing resources and other species in the study area. Results indicate that predation by jumbo squid on fish stocks is lower than that of other predators (e.g. hake) and the fishery. Long-term fluctuations (1978–2004) in the biomass of the main fish stocks (as well as other components of the food web) seem to be related to fishing and to variation in primary production, rather than to predation by jumbo squid alone. Jumbo squid seems to play a role as predator rather than prey in the system, but its impacts are low when compared with the impacts of other predators and fishing. Therefore, we conclude that jumbo squid predation on its prey was not the primary force behind the collapse of important fish stocks off central Chile. Future efforts should be directed to better understanding factors that trigger sudden increases in jumbo squid abundance off central Chile, as well as modeling its trophic impacts.     

Effects of habitat complexity on cannibalism rates in European green crabs (Carcinus maenas Linnaeus, 1758)

The habitat in which predator–prey interactions take place may have a profound influence on the outcome of those interactions. Cannibalism is an intriguing form of predation whereby foraging by predators may contribute to the regulation of their own populations.This is particularly interesting in the case of invasive species, like the widely distributed European green crab (Carcinus maenas). This study explores how habitat complexity influences cannibalism rates in green crab populations of Prince Edward Island, Atlantic Canada. Both laboratory and field experiments were conducted to measure feeding rates by individual adult green crabs on a standard number of smaller conspecifics. In the laboratory, experimental treatments mimicked unstructured to increasingly structured habitats: water, sandy bottom, oyster shells, mussel shells, oyster shells with sandy bottom and mussel shells with sandy bottom. In those trials, adult green crabs consumed several times more juveniles on unstructured habitats than on the most structured ones, with a gradual decrease in predation rates across increasingly complex habitats. Field inclusion experiments used the same approach and were conducted in sandy bottoms, sandy bottoms with a layer of oyster shells and sandy bottoms with a layer of mussel shells. These trials showed similar patterns of decreasing feeding rates across increasingly complex habitats, but differences among treatments were not significant. These results support the idea that complex habitats have the potential to mediate predator–prey interactions, including adult–juvenile cannibalism in green crabs.     

Skin toxins in coral‐associated Gobiodon species (Teleostei: Gobiidae) affect predator preference and prey survival

Predation risk is high for the many small coral reef fishes, requiring successful sheltering or other predator defence mechanisms. Coral‐dwelling gobies of the genus Gobiodon live in close association with scleractinian corals of the genus Acropora. Earlier studies indicated that the low movement frequency of adult fishes and the development of skin toxins (crinotoxicity) are predation avoidance mechanisms. Although past experiments showed that predators refuse food prepared with goby skin mucus, direct predator–prey interactions have not been studied. The present study compares the toxicity levels of two crinotoxic coral gobies – Gobiodon histrio, representative of a conspicuously coloured species, and Gobiodon sp.3 with cryptic coloration – using a standard bioassay method. The results show that toxin levels of both species differ significantly shortly after mucus release but become similar over time. Predator preferences were tested experimentally in an aquarium in which the two gobies and a juvenile damselfish Chromis viridis were exposed to the small grouper Epinephelus fasciatus. Video‐analysis revealed that although coral gobies are potential prey, E. fasciatus clearly preferred the non‐toxic control fish (C. viridis) over Gobiodon. When targeting a goby, the predator did not prefer one species over the other. Contrary to our expectations that toxic gobies are generally avoided, gobies were often captured, but they were expelled quickly, repeatedly and alive. This unusual post‐capture avoidance confirms that these gobies have a very good chance of surviving attacks in the field due to their skin toxins. Nonetheless, some gobies were consumed: the coral shelter may therefore also provide additional protection, with toxins protecting them mainly during movement between corals. In summary, chemical deterrence by crinotoxic fishes seems to be far more efficient in predation avoidance than in physical deterrence involving body squamation and/or strong fin spines.     

Interpreting odours in hermit crabs: A comparative study

Odours of different sources can indicate to hermit crabs the availability of empty shells, crucial resources for the life cycle of almost all of them. Here, we compared Clibanarius erythropus and Pagurus bernhardus for the intensity of investigative behaviour exhibited towards an empty, well-fitting shell in the presence of (1) plain seawater as control and seawater conditioned by (2) dead and live snails, (3) dead and live conspecifics, (4) live predators, and (5) food. During 10 min of observation, we recorded latency (the time until the first contact with the shell), and the number and duration of shell investigation bouts. The two species behaved similarly when exposed to the odours of food, live snails, and predators, while a more intense shell investigation was induced by dead snail odour in C. erythropus and by dead or live conspecific odour in P. bernhardus. Further studies should investigate the influence of phylogeny and ecology on this interspecific difference.     

Intraguild predation and partial consumption of blue sharks Prionace glauca by Cape fur seals Arctocephalus pusillus pusillus

The top-down effects of predators on ecosystem structure and dynamics have been studied increasingly. However, the nature and consequence of trophic interactions between upper-trophic-level predators have received considerably less attention. This is especially the case in marine systems due to the inherent challenges of studying highly mobile marine species. Here we describe the first documentation of asymmetrical intraguild predation by a pinniped predator on a mid-sized predatory shark. The report is based on direct observations in South African waters, in which free-swimming blue sharks Prionace glauca were captured and partially consumed by Cape fur seals Arctocephalus pusillus pusillus. These observations are important not just for understanding the interactions between these two species but more broadly for their implications in understanding the trophic ecology of pinnipeds, many populations of which have increased while numerous shark populations have declined.     

Small macro‐benthic invertebrates influence the survival,growth and behaviour of juvenile green sea urchins Strongylocentrotus droebachiensis in the laboratory

Recent settlers of many marine benthic invertebrates are cryptic, which exposes them to a suite of animals that differs from those they may experience as adults, potentially resulting in interactions causing mortality and/or reducing growth. Previous field experiments have indicated that such is the case with small juvenile green sea urchins Strongylocentrotus droebachiensis but which taxa are responsible for the mortality and reduced growth was not determined. A laboratory study was conducted to examine the effects of small macro‐benthic invertebrates, specifically chitons, scaleworms and larger juvenile conspecifics, as well as the full suite of cobble‐dwelling organisms, on the mortality, growth and behaviour of small (<3 mm) juvenile sea urchins. The likelihood of survival of small juvenile sea urchins was lower in the presence of larger juvenile sea urchins or with the full suite of cobble‐dwelling organisms than in the absence of animals. The small juvenile sea urchins survived and grew the best when they were with chitons and scaleworms. The behaviour of small sea urchins with the full suite of cobble‐dwelling organisms was more cryptic than the behaviour of urchins with scaleworms. This study indicates that interactions with the suite of small organisms living amongst cobbles can affect survival, growth and behaviour of small juvenile sea urchins, and that larger juvenile sea urchins can be a source of mortality for smaller conspecifics.     

Feeding ecology of whelks on an intertidal sand flat in north‐eastern New Zealand

The dispersion patterns and feeding behaviour of intertidal whelks were investigated in north‐eastern New Zealand. Aggregations of whelks feeding on clams occurred regularly, with an average of 15–20 individuals per aggregation. Death of clams was attributable to predation in up to 50% of the aggregations, although scavenging of dead and moribund animals was the main activity of whelks. Laboratory and field studies showed that Lepsiella scobina (generally not considered a soft shore species) was the main whelk predator at Lews Bay, Whangateau Harbour, drilling a small neat hole through the clam shell. The distinction between Cominella spp. as predators or scavengers is not so clear. Reseeding of intertidal clams, Austrovenus stutchburyi, has been proposed as a technique for reestablishing populations in some degraded New Zealand estuaries. The impact of whelk predation has many implications for A. stutchburyi reseeding. Both L. scobina and Cominella adspersa may potentially prey on newly reseeded clams as they preferentially attacked small clams in the laboratory. On‐growing to a larger size before reseeding may be advantageous although large size was not found to protect prey from predation. Additionally, no seasonal trend in whelk activity was found.     

The role of macrophytes as a refuge and food source for the estuarine isopod Exosphaeroma hylocoetes ( )

The role of submerged macrophytes as refugia from fish predation and as possible food sources for the estuarine isopod Exosphaeroma hylocoetes ( Barnard, K.H., 1940) was investigated. Laboratory experiments tested the effectiveness of artificial vegetation, replicating submerged vegetation, in enabling isopods to elude selected fish predators Rhabdosargus holubi, Glossogobius callidus, Monodactylus falciformis and Clinus cottoides. Isopods preferentially hid in the vegetation (>90%), even in absence of fish. The predatory fish had varying success in finding isopods within the vegetation. Isopod mortality ranged from 2% (R. holubi) to a maximum of 87% (C. cottoides) within vegetation, depending on the fish predator present. Stable isotope and fatty acid analyses ruled out the submerged macrophyte Ruppia maritima and inundated fringing grasses as direct food sources, but highlighted the epiphytic biota (mainly diatoms) found on the submerged vegetation and sediments as more likely food sources. These findings are consistent with gut content analyses. The results suggest that the close association of E. hylocoetes with R. maritima is the result of the vegetation providing the isopod with a refuge against fish predation as well as areas of increased food availability.     

Persistent alternate abundance states in the coral reef sea urchin Diadema savignyi: evidence of alternate attractors

Alternate attractors have been shown to exist in a variety of terrestrial and aquatic systems, e.g. temperate forests, savannas, shallow lakes, wetlands, coral reefs, kelp forests. The shift from one attractor to another, also referred to as a regime shift, is thought to occur when a system passes some critical threshold such that the trajectory of the system changes direction. Alternate attractors in population dynamics can also exist, leading to alternate stable states in the population abundance of a species. This study explored alternate attractors in the population dynamics of the Indo‐Pacific sea urchin Diadema savignyi and the potential underlying mechanisms that promote its bi‐stability. In Moorea, French Polynesia, the local abundance of D. savignyi, a functionally important herbivore in lagoon habitats, occurs in two states: (i) solitary individuals that occupy crevices in low densities and (ii) aggregations of tens to hundreds of individuals. These different states are temporally stable and are not explained by spatial differences in recruitment rates of juveniles. A field experiment revealed that the per capita mortality rate of adult D. savignyi was substantially lower at sites where urchins occurred in aggregations compared with sites at which they were solitary individuals. An additional experiment showed that per capita mortality decreased with increasing aggregation size. Individuals in high‐density aggregations, however, had significantly smaller test diameters than solitary individuals, indicating that individuals in aggregations may be food limited. Collectively, the evidence suggests that the two different local abundance states of D. savignyi result from negative feedback loops where high local density can be maintained by aggregative behavior that greatly reduces per capita risk of predation when the local number of adult sea urchins is sufficiently large; sites with few sea urchins remain at low density because individuals are more susceptible to predation when crevices are occupied but there are not enough individuals to form large aggregations. Thus, there may be alternate attractors in the population dynamics of D. savignyi that can produce either persistently low or high local population densities.