Food resource partitioning in a rocky intertidal fish assemblage

Abstract

A resident intertidal fish assemblage at Dillon Beach, California, U.S.A., was persistent and resilient for a 42‐month period. Data from a previous study indicated that productivity was the best predictor of several assemblage characteristics; hence, food resources may have been limiting in this assemblage. To determine whether food resource partitioning was a potential organizational mechanism for this assemblage, the following aspects of trophic organization were quantified for 1,858 individuals of 16 species, collected over a 31‐month period: (1) annual, seasonal, and size‐related dietary changes, (2) annual and seasonal patterns of interspecific dietary overlap, (3) delineation of the relationship between morphological similarity and dietary similarity and (4) estimates of annual prey consumption rates by fish predators.Fishes were classified into seven basic feeding groups: crab feeders, crab‐shrimp feeders, shrimp‐hermit crab feeders, polychaete‐shrimp feeders, Idotea feeders, algal‐sea anemone feeders, and gammarid feeders. Major prey taxa consumed by these groups were: Cancer spp., Pugettia producta, Heptacarpus taylori, Pagurus hemphilli, Idotea spp., Anthopleura spp., Viva spp., Porphyra spp., gammarids and polychaetes. Based on dietary breadth estimates using Levins' index (Levins, 1968), species possessed either narrow (1 08‐2‐30) or wide (7–62‐9‐71) dietary niches, with only one species displaying an intermediate value. High values were obtained for estimates of intraspecific dietary overlap between seasons (x̄± IS.D. = 0–678 ± 0–197) and among years (x̄± IS.D. = 0–634 ± 0178). Hence, species generally exhibited stable patterns of food resource utilization. Many fishes showed size‐related dietary changes. These generally represented switches from smaller (e.g. gammarids) to larger (e.g. crabs, shrimp, polychaetes) prey.Interspecific dietary overlap was very low and averaged 0–275 ± 0–255 for all pairwise comparisons. Nine of 12 species did not display seasonal differences in interspecific dietary overlap, and the three remaining species exhibited lower overlap during the season of lowest productivity. There was a significantly greater number of low overlap values (i.e. < 0–400) than could be expected by chance alone. In addition, species which displayed the lowest amount of dietary overlap were those for which food limitation was probably most severe (i.e. resident‐resident overlaps < resident‐seasonal overlaps < seasonal‐seasonal overlaps). These data indicate that food resource partitioning probably was operating within this assemblage. This mechanism is the most likely process responsible for assemblage organization because predators, competition for space and environmental disturbances do not seem to have a strong effect on this assemblage.Morphological and dietary similarity were assessed by conducting a multivariate analysis on nine variables associated with prey capture: standard length, head length, head height, jaw length, jaw width, mouth protrusibility, mouth orientation, eye position and the number of gill rakers on the first gill arch. A principal component analysis of these data extracted two components which accounted for 84% of the variance and indicated that species comprised morphologically distinct groups. The first principal component was a general size‐related component, and the second was a trophically linked component on which mouth orientation, eye position and the number of gill rakers loaded most heavily. MANOVA indicated that significant differences existed among species' morphologies and a discriminant function analysis (DFA) correctly assigned individuals to species 96% of the time. To test the ecomorphological hypothesis, Mahalanobis distances derived from the DFA were regressed against interspecific dietary overlap. None of ten regressions, representing all possible combinations of residency pattern and season, yielded correlations which were significantly different from zero. These results are discordant with the predictions of the ecomorphological hypothesis and suggest that behaviour is more important than morphology in determining prey utilization.Estimates of prey consumption rates were quantified by deriving an estimate of the mean number of each prey species consumed per day by each fish species. Annual consumption rates summed over all predators for various prey species ranged from 625 to 13,826 individuals consumed per year in the 19 m by 30 m study site. These data indicate that fish predation may exert a strong selective pressure on intertidal invertebrates.The Dillon Beach rocky intertidal fish assemblage was organized through deterministic processes (i.e. food resource partitioning) over a 42‐month period. This mechanism is different from the organizational mechanism exhibited by intertidal algal and invertebrate assemblages which are primarily affected by stochastic forces. The evolutionary forces responsible for assemblage organization affect different taxa in different ways; hence, habitat‐wide extrapolations regarding community organization are probably unwise.