Food and microhabitat resource use by vermivorous gastropods on Heron reef crest was analyzed from 3 samples, separated by 12 and 20 month periods. The assemblage is typical of those of other Indo-Pacific sites with respect to mean shell size and population density. These 2 parameters varied among samples for some species, although no general trends occurred. Fluctuations in species' size and density are assumed not to have significantly influenced availability of prey or microhabitat types.
Colwell and Futuyma's (l971) method of ecological scaling of resource use data, was used for microhabitat niche breadth and overlap estimation, allowing between-sample comparisons. Data on gastropod activity in aquaria, together with microhabitat use patterns, demonstrated significant day-night differences in microhabitat resource use. Gastropods occurred mainly in the structurally complex, 'refuge' microhabitats during the day and showed an increased abundance in smooth, exposed, 'foraging' microhabitat at night compared with the day. Nassarius gaudiosus is the most extreme microhabitat specialist during the day and the most extreme microhabitat generalist at night. A similar, although less extreme trend was exhibited by other gastropods.
Microhabitat #4 (sand under rocks) and #2 (sand in depressions) are the major contributors to structural heterogeneity in the gastropods' environment. Multiple regression analysis demonstrated a significant relationship between pooled gastropod abundance (= pooled abundance of N. gaudiosus, Vasum turbinellus, Conus miliaris, C. coronatus, C. flavidus) and the abundance of microhabitat #2. The abundances of N. gaudiosus, C. coronatus and C. flavidus each showed a significant relationship with the abundance of microhabitat #4. Therefore N. gaudiosus is influenced, directly or indirectly, by the availability of its chief refuge microhabitat. Since C. coronatus and C. flavidus occur mainly in microhabitat #2 during the day, the cause of the relationship between the abundance of microhabitat #4 and the abundance of these 2 species is unclear. The availability of the chief foraging microhabitat #7 (smooth limestone with a thin layer of algal-bound sand) does not limit gastropod abundance.
Microhabitat niche overlap was high for C. coronatus, C. miliaris, C. flavidus, V. turbinellus and N. gaudiosus at night and was also high during the day for all these species, except N. gaudiosus. These data are consistent with Pianka's (1976) hypothesis that maximum tolerable niche overlap increases as resource availability increases.
There were 8 gastropod species for which 5 or more diet records were obtained and two distinct feeding strategies are discernable within the gastropod assemblage. Members of the first group consume only errant polychaetes (mostly eunicids) and, in order of increasing food niche breadth, are Conus miles, C. miliaris, C. sponsalis, C. coronatus and N. gaudiosus. These species show a negative relationship between predator body size and food niche breadth, a result consistent with Leviten's (1976) model of Conus feeding strategies. Species of the second group consume sedentary polychaetes as part, or all of their diet and are Peristernia nassatula, C. flavidus and V. turbinellus, in order of increasing food niche breadth. For these species, predator body size and food niche breadth are positively related, which is consistent with Schoener's (1971) feeding strategy model. Because errant polychaetes were relatively more abundant in substratum samples than were sedentary worms, the observed dichotomy in feeding strategies of vermivorous gastropods probably results from the difference in availability of prey for the two groups. As the body size of gastropods that consume sedentary worms increases, relative prey availability is reduced and therefore food niche breadth increased. In contrast, the consumers of the very abundant errant polychaetes are not prey limited as gastropod body size increases, and the narrow food niche breadths of large-bodied Conus result from increasing minimum prey size. Data on abundance and spatial predictability of polychaetes, combined with the gastropod diet data, are consistent with the predictions of optimal foraging theory, that abundant food supplies will lead to selective foraging and narrow food niche breadths, particularly in the case of C. miles and C. miliaris which specialize on the very abundant eunicid worms and P. nassatuta which specializes on chaetopterids, the most abundant sedentary polychaetes.
For Heron Reef vermivorous gastropods, the microhabitat data are consistent with the intermediate disturbance hypothesis of community structure maintenance (Connell, 1978), which assumes the assemblage is in a non-equilibrium state, and the food data are consistent with the niche diversification hypothesis (Connell, 1978), which assumes the assemblage is in an equilibrium state. This apparent paradox may result from multiple factors operating to organize communities at different spatial and temporal scales,