Cooperative prey capture by young subsocial spiders: II. Behavioral mechanism

Subsocial spiders demonstrate an intermediate stage in the evolution of permanent sociality. Cooperative hunting is an important attribute of their sociality, but has not been documented in subsocial arthropods. After cannibalizing their mother, young of the subsocial spider Amaurobius ferox (Araneae, Amaurobiidae) remain together for several instars and feed communally. We monitored the collective prey capture behavior of the spiderlings. All the clutches showed collective capturing sequence (latency–orientation–moving–touching–seizing–feeding) toward the prey that was 10 times more massive than each individual. The first three individuals that exhibited attacking behavior were responsible for 90% of the total number of attacks, while 68% of the individuals within the group never exhibited attacks during the first 10 min following the introduction of prey into the communal web. First arriving individuals at the prey most often seized the antennae and legs of the prey, which probably facilitate access to the prey for subsequent individuals. The spiderlings that arrived later occupied more likely the abdomen and thorax, which contain more nutrition than the extremities occupied earlier. The individual apportionment of collective hunting behavior suggests a coordinated teamwork among individuals.     

Dispersal behaviour in a subsocial spider: group conflict and the effect of food availability

This study investigates proximate factors influencing dispersal behaviour in the subsocial spider Amaurobius ferox. Dispersal of spiderlings from the natal web occurred as a progressive sigmoidal function (mean duration of dispersal period=31 days), with considerable individual variability in developmental instar and body mass within the clutch at the time of dispersal. The spiderlings showed a significant decrease in group cohesion on the 6th day after their second moult, which corresponded to the beginning of the dispersal period. Mutual aggressiveness appeared when the spiderlings began to show predatory behaviour (4th day post-second moult) and increased over the course of the dispersal period. While lack of prey in the maternal nest accelerated dispersal behaviour, the addition of prey items lengthened the dispersal period in previously non-fed clutches. Individuals that dispersed were smaller than the remaining individuals when measured on the day 50% of the clutch had dispersed. Timing of the appearance of the developmental characters (second moult, predation activity, agonistic behaviour against siblings, reduction of group cohesion, dispersal) suggests that the dispersal trait might have evolved in consequence of these different functional behaviours. Received: 10 August 1999 / Received in revised form: 15 March 2000 / Accepted: 1 April 2000     

Exploring the effects of individual traits and within-colony variation on task differentiation and collective behavior in a desert social spider

Social animals are extraordinarily diverse and ecologically abundant. In understanding the success of complex animal societies, task differentiation has been identified as a central mechanism underlying the emergence and performance of adaptive collective behaviors. In this study, we explore how individual differences in behavior and body size determine task allocation in the social spider Stegodyphus dumicola. We found that individuals with high body condition indices were less likely to participate in prey capture, and individuals’ tendency to engage in prey capture was not associated with either their behavioral traits or body size. No traits were associated with individuals’ propensity to participation in web repair, but small individuals were more likely to engage in standard web-building. We also discovered consistent, differences among colonies in their collective behavior (i.e., colony-level personality). At the colony level, within-colony variation in behavior (aggressiveness) and body size were positively associated with aggressive foraging behavior. Together, our findings reveal a subtly complex relationship between individual variation and collective behavior in this species. We close by comparing the relationship between individual variation and social organization in nine species of social spider. We conclude that intraspecific variation is a major force behind the social organization of multiple independently derived lineages of social spider.     

Behavioral correlates of non-random mortality among free-ranging female vervet monkeys

Summary Over a 3.5 year period, illness and predation operated in a non-random manner on free-ranging vervet monkeys in Amboseli National Park. As a result, there was no correlation among adult females between dominance rank and reproductive success. Deaths due to illness were concentrated among low-ranking individuals, and appeared to occur as a result of restricted access to food and water during the dry season. In contrast, deaths due to predation were concentrated among high-ranking individuals. The precise cause of such increased vulnerability could not be determined.High-ranking females alarm-called at higher frequencies than low-ranking females, and were also more aggressive than low-ranking females during intergroup encounters. In contrast, low-ranking females were more likely to initiate friendly interactions with the members of other groups. The non-random distribution of causes of mortality suggests that individuals living in the same social group may confront different selective pressures. Perhaps as a result, individuals appear to respond differently to similar social and environmental variables.     

Multiple techniques confirm elevational differences in insect size that may influence spider sociality

Social and subsocial spiders of the genus Anelosimus exhibit an altitudinal pattern in their geographic distribution at tropical latitudes in the Americas. Social species, which capture prey cooperatively, occur primarily in the lowland rain forest and are absent from higher elevations, whereas subsocial species are common at higher elevations but absent from the lowland rain forest. Previous studies have suggested that differences in the size of potential insect prey along altitudinal gradients may explain this pattern as insects were found to be, on average, larger in lowland rain forests than at higher elevations. These studies, however, may have under-sampled the insect size composition of each habitat because only one sampling technique was used. Using a number of collection methods we sampled the insect size composition in the environments of social and subsocial spiders in this genus. We found that the average insect size in lowland rain forest habitats was indeed larger than at high-elevation cloud forests in eastern Ecuador. We also found that, even though the various techniques differed in the size of the insects they captured (visual searching and blacklighting yielding larger insects than beating, sweeping, or malaise trapping), they all caught, on average, larger insects in the lowlands. Overall, spider colonies in the lowlands caught larger prey than did spider colonies at higher elevations, paralleling differences in insect size distribution obtained by the various techniques in their respective environments.     

A laboratory study of predation by Aurelia aurita on larval herring (Clupea harengus): Experimental observations compared with model predictions

Predation by the medusa Aurelia aurita L. on early first-feeding stage larvae of the herring clupea harengus L. was studied in the laboratory. The medusae were captured in Loch Etive, Scotland. Herring larvae were reared from the extificially fertilized eggs of spawning Clyde herring caught in March, 1982. Swimming speeds, volume searched”, capture efficiency and predation rates increased as medusa size increased. Predation rates on fish larvae increased with prey density, but appeared to approach a maximum at high prey densities; in 1 h experiments, a maximum rate of predation of 6.64 larvae h^-1 was estimated by fitting an Ivlev function. A model to predict predation rates was constructed from swimming speeds, sizes and densities of medusae and larvae, and capture efficiency. The rates of predation predicted from the model fell within the range of experimental data, but tended to underestimate rates and did not account for saturation of medusae. Swimming patterns of medusae changed after prey capture: (a) before capture, encounter rates were low and medusae were relatively less active; (b) after capture of 1 larva, encounter rates doubled, with the stimulated medusae exhibiting increased activity and an aftered “searching” path; and (c) after capture of many larvae, swimming speeds and encounter rates of medusae decreased.     

Determinants of vigilance behavior in the ring-tailed coati (Nasua nasua): the importance of within-group spatial position

Individuals living in social groups are predicted to live under unequal predation risk due to their spatial location within the group. Previous work has indicated that individuals located at the edge of groups have higher “domains of danger”, thus are more likely to engage in vigilance or antipredator behavior. We studied the determinants of vigilance behavior in two groups of ring-tailed coatis in Iguazu National Park, Argentina. In addition to the expected pattern that coatis were more vigilant at the edge of the group, we found that individuals were particularly vigilant at the front edge of the group. This pattern conforms to predictions of differing predation risk caused by sit-and-wait predators with respect to mobile animal groups. In addition, coatis exhibited less vigilance when the number of neighbors within 5 m and group size increased. Of the three spatial variables tested, within-group spatial position was the most important predictor variable determining vigilance levels. These results confirm that spatial position has major effects on vigilance behavior, and that group directionality is an important factor which should be taken into account when measuring vigilance behavior. Coatis were more vigilant when juveniles less than 6 months old were in the groups. The presence of these young juveniles also affected the relationship between alarm response and vigilance levels. Coatis were more vigilant after strong alarm reactions, but only when young juveniles were not present in the groups. This may indicate that coatis give differential responses to alarm calls depending on the age of the caller. A comparison of antipredator vigilance between coatis and sympatric capuchin monkeys is consistent with the hypothesis that terrestriality leads to higher perceive predation risk for coatis.     

Malaria and risk of predation: a comparative study of birds

Predators have been hypothesized to prey on individuals in a poor state of health, although this hypothesis has only rarely been examined. We used extensive data on prey abundance and availability from two long-term studies of the European Sparrowhawk (Accipiter nisus) and the Eurasian Goshawk (Accipiter gentilis) to quantify the relationship between predation risk of different prey species and infection with malaria and other protozoan blood parasites. Using a total of 31 745 prey individuals of 65 species of birds from 1709 nests during 1977-1997 for the Sparrowhawk and a total of 21 818 prey individuals of 76 species of birds from 1480 nests for the Goshawk during 1977-2004, we show that prey species with a high prevalence of blood parasites had higher risks of predation than species with a low prevalence. That was also the case when a number of confounding variables of prey species, such as body mass, breeding sociality, sexual dichromatism, and similarity among species in risk of predation due to common descent, were controlled in comparative analyses of standardized linear contrasts. Prevalence of the genera Haemoproteus, Leucocytozoon, Plasmodium, and Trypanosoma were correlated with each other, and we partitioned out the independent effects of different protozoan genera on predation risk in comparative analyses. Prevalence of Haemoproteus, Leucocytozoon, and Plasmodium accounted for interspecific variation in predation risk for the two raptors. These findings suggest that predation is an important factor affecting parasite-host dynamics because predators tend to prey on hosts that are more likely to be infected, thereby reducing the transmission success of parasites. Furthermore, this study demonstrates that protozoan infections are a common cause of death for hosts mediated by increased risk of predation.     

Smooth-billed ani (Crotophaga ani) predation on butterflies in Mato Grosso, Brazil: risk decreases with increased group size

Although most insects are vulnerable to predation by a variety of predators, including birds, there are few direct observational studies in the wild of avian predation on adult butterflies. We examined the predatory behavior of smooth-billed anis (Crotophaga ani) on butterflies, and the spacing behavior of the butterflies which were concentrated on a mineral-rich beach on the Cristalino River, in Mato Grosso, Brazil. We studied eight of the most regularly occurring butterfly species which came each morning to engage in "puddling." Most species of butterfly were closely associated with conspecifics, although nearest-neighbor distance varied among species. The pierids - "yellows" (Aphrissa statira, Phoebis trite), "oranges" (Phoebis argante), and sometimes "whites" (Daptoneura leucadia) - formed very dense groups (or clusters) of up to 1,000 individuals occasionally joined by a few kite swallowtails (Eurytides spp.). Most other butterfly species formed small groups (e.g., daggerwings, Marpesia spp.) or were dispersed individually and non-clumped over the beach (e.g., dingy purplewing, Eunica monima). Anis foraged using two strategies: rapid frontal attack on dense groups of butterflies (yellows, oranges, whites), and a stealthy approach to solitary butterflies (mainly purplewings) or those in small groups. For yellows, the most common butterfly caught by anis, the capture rate reached over 6 per 15 min per ani, and about 8% of those captured managed to escape. Capture rates were much lower for the other species. Time of day, age of the ani (adult or young-of-the-year), and total number of each butterfly species present accounted for variation in the number of each species captured by anis. The number of butterflies captured per 15 min increased as the number of butterflies present increased, but reached a threshold beyond which the capture rate did not increase. The capture rate per individual butterfly (individual risk) decreased with group size up to a group of 40 individuals and remained low with further increases. Thus a butterfly in a group of 100 was no less likely to be eaten than one in a group of 40. For individual ani forays into dense groups of pierids, an individual ani was unable to catch more than 16, regardless of group size. These data confirm the dilution effect of group size for butterflies; each individual yellow or orange was less at risk from ani predation when in a group.     

Context-dependent acceptance of non-nestmates in a primitively eusocial insect

The benefits of cooperation are essential in driving group formation. However, an individual can gain significant benefits by acting selfishly at a substantial cost to others in the group. Thus, group members must find a balance between accepting and rejecting potential new members. Here, I explore the factors that mediate acceptance of non-related individuals during the period of group establishment in the primitively eusocial wasp Mischocyttarus mexicanus. In this species, group composition changes during establishment, with non-related females (non-nestmates) sometimes accepted into a foreign colony. By experimentally introducing non-nestmates to newly established colonies, I test the hypothesis that acceptance threshold of nestmates towards non-nestmates shifts depending on the ecological context, as predicted by the Optimal Acceptance Threshold Model. I explored how non-nestmate age (young vs. old), stage of colony establishment (early vs. late), initial behavior of the non-nestmates (non-aggressive vs. aggressive), and the behavioral response by nestmates (non-aggressive vs. aggressive) affected the rates of acceptance. My results show an effect of both non-nestmate age and stage of colony development on non-nestmate acceptance. Young non-nestmates were more frequently accepted in early than in late colonies. Late colonies more frequently rejected both young and old non-nestmates, suggesting a cost of accepting potential usurpers into late colonies. Surprisingly, non-nestmate aggressive behavior did not have a direct effect on their acceptance, but it triggered an aggressive response from nestmates. These findings reveal a shift in the acceptance threshold, suggesting an effect of the social context and the specific needs of a colony on non-nestmate acceptance.     

Do social spiders cooperate in predator defense and foraging without a web?

Nearly all social spiders spin prey-capture webs, and many of the benefits proposed for sociality in spiders, such as cooperative prey capture and reduced silk costs, appear to depend on a mutually shared web. The social huntsman spider, Delena cancerides (Sparassidae), forms colonies under bark with no capture web, yet these spiders remain in tightly associated, long-lasting groups. To investigate how the absence of the web may or may not constrain social evolution in spiders, we observed D. cancerides colonies in the field and laboratory for possible cooperative defense and foraging benefits. We observed spiders’ responses to three types of potential predators and to prey that were introduced into retreats. We recorded all natural prey capture over 447 h both inside and outside the retreats of field colonies. The colony’s sole adult female was the primary defender of the colony and captured most prey introduced into the retreat. She shared prey with younger juveniles about half the time but never with older subadults. Spiders of all ages individually captured and consumed the vast majority of prey outside the retreat. Young spiders benefited directly from maternal defense and prey sharing in the retreat. However, active cooperation was rare, and older spiders gained no foraging benefit by remaining in their natal colony. D. cancerides does not share many of the benefits of group living described in other web-building social spiders. We discuss other reasons why this species has evolved group living.     

Foraging, prey capture, and gut contents of the mesopelagic narcomedusa Solmissus spp. (Cnidaria: Hydrozoa)

Narcomedusae are the most common group of medusae in the mesopelagic depths of Monterey Bay, California. Numerous capture events of various prey taxa were recorded in situ and analyzed using the Monterey Bay Aquarium Research Institute's remotely operated vehicle "Ventana". In situ video analysis of the stomach contents of 82 Solmissus incisa and S. marshalli revealed 88 identifiable prey items. Most (88%) were gelatinous animals. Of these gelatinous prey, 60.3% were ctenophores, 20.5% were cnidarians, 12.8% were salps, 3.8% were chaetognaths, and 2.6% were polychaetes. Euphausiids accounted for 11.4% of the diet, but they were probably captured adventitiously, as the gut contents of ctenophore prey. The tentacle-first foraging behavior of the narcomedusae is an effective way to capture large, comparatively fast-moving prey, because the fluid disturbance caused by swimming is decoupled from the area of prey capture. This behavior contrasts with the prevailing models of feeding behavior in medusae. Stealth predation may be the dominant mode of capturing prey by medusae in the mesopelagic depths of the oceans.     

Seasonal and spatial differences in the fish assemblage of the non-estuarine Sulaibikhat Bay,Kuwait

In the period September 1986 to August 1987 fish were captured once a month, using an otter trawl, from the intertidal and subtidal regions of Sulaibikhat Bay, Kuwait. Correlation analysis revealed that numbers were not directly related to temperature at the time of sampling. The recruitment of large numbers of young of the year Liza carinata (Valenciennes) during March was related to low temperatures in November, the probable time of spawning of this species. Salinity at the time of sampling was inversely correlated with numbers. This result indicates that large numbers of 0+fish recruit to the Bay during the period of maximum fresh water outflow through the Shatt-al-Arab. Numbers of fish were significantly greater in the intertidal region, where they were present almost entirely as 0 group fish, than in the subtidal. The three dominant species of the assemblage are shown to use the two depth intervals in differnt ways. L. carinata was capture a almost exclusively in the intertidal region as 0+ fish. Pomadasys stridens (Forsskål) was captured as 0+ fish in both the intertidal and subtidal regions but in greater numbers in the intertidal region. Leiognathus brevirostris (Valenciennes) was captured as small, mostly 0+ individuals in the intertidal region and as larger fish in the subtidal region. The use made by the smaller fish of the intertidal region is related to the avoidance of sublittoral predators and reduction of intraspecific competition, whilst larger fish in the sublittoral region may be avoiding predation by piscivorous birds.     

Prey vulnerability in relation to sexual coloration of prey

Sexual selection that results in the evolution of exaggerated secondary sexual characters has been hypothesized to impose production and maintenance costs of such traits on their bearers. Costs arising from sexual selection could increase the intensity of predator-mediated natural selection, leading to the prediction that species with exaggerated secondary sexual characters should be particularly susceptible to predation. We tested this prediction in a comparative analysis based on 31,745 prey individuals belonging to 66 species of birds collected from a total of 937 breeding events by 33 to 66 different pairs of European sparrowhawks Accipiter nisus annually during a period of 21 years. To assess vulnerability of different species we estimated a prey vulnerability index based on the difference in the logarithmically transformed absolute abundance of prey minus the logarithmically transformed expected abundance as determined by population density of breeding birds. The prey vulnerability index was predicted by sexual dichromatism, accounting for 23% of the variance in risk of predation among species, even when considering similarity in phenotype among species due to common descent (in the latter case explaining 12% of the variance). This finding suggests that sexual selection is an important evolutionary force-affecting predator–prey interactions.Electronic Supplementary Material Supplementary material is available for this article at     

Correlates of boldness in three-spined sticklebacks (<Emphasis Type="Italic">Gasterosteus aculeatus</Emphasis>)

Behavioural variation is known to occur between individuals of the same population competing for resources. Individuals also vary with respect to their boldness or shyness. An individuals position along the shy-bold axis may be defined as the extent to which it is willing to trade off potentially increased predation risks for possible gains in resources. Similarly, group living may be interpreted as a trade-off between anti-predatory tactics and foraging efficiency. The responses of three-spined sticklebacks (Gasterosteus aculeatus) were tested across four social contexts to assess relative boldness or shyness and to further examine whether their behaviour would be consistent within and between contexts. Individuals displayed consistent responses within and between the first two contexts: those individuals which resumed foraging rapidly after a simulated aerial predator attack also displayed low shoaling tendencies. Such fish were deemed to be bold, whilst those which displayed the converse behaviour, slow resumption of foraging and a high shoaling tendency, were deemed to be shy. In a third context, bold individuals out-competed shy conspecifics for food. Boldness was also positively correlated with growth over a 6-week period. The position adopted by an individual within a group is usually interpreted as a trade-off between predation risk and foraging efficiency—both are greater at the front of a mobile group. Bold individuals showed significantly stronger tendencies towards front positions than shy conspecifics. The results suggest that, contrary to some previous studies on other animals, bold or shy behaviour in sticklebacks is consistent between contexts.Communicated by T. Czeschlik     

Is boldness affected by group composition in young-of-the-year perch (Perca fluviatilis)?

In young-of-the-year perch (Perca fluviatilis), individuals within groups differed in the degree of boldness, estimated by habitat utilisation and feeding activity in visual contact with a potential predator. We looked at changes in individual behaviour in connection with change of group composition. During the first period, perch were randomly assigned to groups, and time spent in open habitat versus in vegetation and number of prey attacks were registered. The perch were then categorised into personality types (shy, bold, intermediate) according to their behaviour. During the second period, fish were observed when sorted into new groups, each containing only one personality type. Shy individuals showed the largest changes in behaviour, and increased both the time spent in the open and the number of prey attacks when placed into the new groups. Feeding activity in shy fish during the second period was affected by group composition during the first period. After regrouping, bold individuals decreased their time in the open, whereas intermediate individuals did not change behaviour. Time in the open habitat was, to some extent, influenced by the behaviour of the other members of the group, but number of prey attacks was not. The behaviour of fish of the different personality types we have defined in this study seemed to be based on innate traits, but also modified by the influence of other group members and by habituation to the environment.Communicated by J.Krause     

Species-specific influence of group composition on collective behaviors in ants

The success of a social group is often driven by its collective characteristics and the traits of its individuals. Thus, understanding how collective behavior is influenced by the behavioral composition of group members is an important first step to understand the ecology of collective personalities. Here, we investigated how the efficiency of several group behaviors is influenced by the aggressiveness of its members in two species of Temnothorax ants. In our manipulation of group composition, we created two experimentally reconstituted groups in a split-colony design, i.e., each colony was split into an aggressive and a docile group of equal sizes. We found strong species-specific differences in how collective behaviors were influenced by its group members. In Temnothorax longispinosus, having more aggressive individuals improved colony defense and nest relocation efficiency. In addition, source colony identity strongly influenced group behavior in T. longispinosus, highlighting that manipulations of group compositions must control for the origin of the chosen individuals. In contrast, group composition and source colony did not influence collective behaviors in Temnothorax curvispinosus. This suggests that the mechanisms regulating collective behaviors via individual differences in behavior might differ among even closely related species.     

The use of Doppler-shifted echoes as a flutter detection and clutter rejection system: the echolocation and feeding behavior of Hipposideros ruber (Chiroptera: Hipposideridae)

Summary Hipposideros ruber use CF/FM echolocation calls to detect the wing flutter of their insect prey. Fluttering prey were detected whether the insects were flying or sitting on a surface, and prey in either situation were captured with equal success (approximately 40% of capture attempts). Stationary prey were ignored. The bats did not use visual cues or the sounds of wing flutter to locate their prey. Wing flutter detection suggests that H. ruber exploit the Doppler-shifted information in echoes of their echolocation calls. These bats fed primarily upon moths, usually those of between 10 and 25 mm wingchord, although moths of less than 5 mm and greater than 40 mm wingchord were also attacked and captured. They showed no evidence of selecting moths on the basis of species or other taxonomic distinction, and occasionaly captured other insects.     

Diet and seasonal prey capture rates in the Mediterranean red coral (Corallium rubrum L.)

Gorgonians are passive suspension feeders, contributing significantly to the energy flow of littoral ecosystems. More than in active suspension feeders (such as bivalves, ascidians and sponges) their prey capture is affected by spatial and temporal prey distribution and water movement. Corallium rubrum is a characteristic gorgonian of Mediterranean sublittoral hard bottom communities. This study found a high variability in the annual cycle of prey capture rate, prey size and ingested biomass, compared to other Mediterranean gorgonians. Detrital particulate organic matter (POM) was found throughout the year in the polyp guts and constituted the main proportion of the diet (25–44%). Crustacean fragments and copepods (14–46%) accounted for the second major proportion, while invertebrate eggs (9–15%) and phytoplankton (8–11%) constituted the smallest part of the diet. To verify the importance of detrital POM in the energy input of this precious octocoral species, in situ experiments were carried out during the winter–spring period. The results confirm the importance of detrital POM as the main source of food for C. rubrum [0.13±0.04 μg C polyp⁻¹ h⁻¹ (mean±SD)]. This study also compares the prey capture rates of two colony size classes and two depth strata: Within the same patch, small colonies (<6 cm height) captured significantly more prey per polyp (0.038±0.09 prey polyp⁻¹ h⁻¹) than larger colonies (>10 cm high) (0.026±0.097 prey polyp⁻¹ h⁻¹) and showed a higher proportion of polyps containing prey (17% compared to 10%). Comparing colonies of similar size (<6 cm height) revealed that the colonies situated at 40 m depth captured significantly more prey (0.038±0.09 prey polyp⁻¹ h⁻¹) than the ones at 20 m (0.025±0.11 prey polyp⁻¹ h⁻¹). One pulse of copepods was recorded that constituted 16% of all captured prey during the 15-month period studied in one of the sampled populations. The data suggest that the variability of hydrodynamic processes may have a higher influence on the feeding rate than seasonal changes in the seston composition. The carbon ingestion combined with data on the density of the exploited population results in 0.4–9.6 mg C m⁻² day⁻¹. The grazing impact of current, heavily exploited and small-sized populations is comparable to that of larger Mediterranean gorgonians, suggesting that unexploited red coral populations may have a high impact compared with other passive suspension feeders.     

Zooplankton capture by two scleractinian corals,Madracis mirabilis andMontastrea cavernosa,in a field enclosure

Capture of zooplankton by scleractinian corals has been noted for several species, yet quantitative information on rates of capture and differential capture by prey taxon has been lacking. We used field enclosures to examine prey capture for two coral species,Madracis mirabilis (Duchassaing and Michelotti) andMontastrea cavernosa (Linnaeus), on the north coast of Jamaica (Discovery Bay) in November 1989, February and March 1990, and January 1992.M. mirabilis has small polyps and a branching colony morphology (high surface/volume ratio), whereasM. cavernosa has large polyps and mounding colonies (low surface/volume ratio). Corals were isolated front potential prey, then were introduced into enclosures with enhanced zooplankton concentrations for 15- to 20-min feeding periods. Corals were fixed immediately after the experiment to prevent digestion, and coelenteron contents were examined for captured zooplankton. Plankton pumps were used to sample ambient zooplankton in the enclosures near the end of each run. Selectivity and capture rates were calculated for each prey taxon in each experiment; both indices were high for relatively uncommon large prey, and low for copepods, which were often the most common items in the plankton. Sizes of zooplankton captured by both species were generally larger than those available considering all prey taxa combined, but were almost the same for both coral species, even though the corals' polyp sizes are very different. This occurred primarily because small copepods, with low capture rates, dominated most plankton samples. For specific prey species, or group of species, there were few significant differences in size between the prey available and the prey captured.M. mirabilis, with small polyps, also captured far more prey per unit coral biomass than didM. cavernosa, with much larger polyps. We hypothesize that the large differences in capture rate of prey taxa are related to escape or avoidance behavior by those potential prey, and to the mechanics of capture, rather than to any selectivity by the corals.