Analyses of paired sedlment and gut samples from specimens of the bathyal trochid gastropod Bathybemblx aeola (Watson) reveal selective ingestion of fine particles from a mixed mud and sand substrate. The radula is the selecting organ. Results contradict the prevalent notion that sediment-ingesting gastropods are lndiscnminant. The standard concept of the unselective deposit feeder as a primitive, generalist vacuum cleaner of the sea floor is difficult to accept from a morpholog~cal viewpoint because it requires perfecting a feeding apparatus capable of taking an unbiased sample. Results are con~patible wlth the hypothesis that se lec t~ve ingestion of fine particles with greater surface area of organic coating is energetically efficient. Selective feeding is a theoreticdlly appealing adaptive solution to dividing and efficiently using limited resources among the diversity of benthic invertebrates known to coexlst in the deep sea (Sanders, 1968; Grassle and Sanders, 1973). In deposit-feeding organisms, however, selectivity may not be so much an ecological problem related to conlpetition and resource partitioning as it is a problem relatlng the sizes and qualities of particles in the substrate to the sizes of organisms and the biomechanics and morphology of their feeding apparatus. From the point of view of the organism living in an environment in which organic matter is relatively scarce (Filatova, 1969; Sanders and Hessler, 1969), plasticity of feedlng behavior may be advantageous (Gage, 1978), and the ingestion of mud may be viewed as a problem in maximizing net energy gain (Taghon et al., 1978). Instances of selective feeding in the deep sea have been difficult to establish. Inferences in the literature are based on particle-size ranges and types in gut contents (Menzies, 1962; Carey, 1972). However, dramatic differences in gut contents of organisms coexisting in the same environment may reflect microhabitat selection (as in surface vs. subsurface feeding) rather than a morphologically based system of particle selection and rejection. Dredge hauls and grab O Inter-Research/Printed in F. R Germany samples mix sediments and separate organisms from feeding microhabitats. The difficulty of sampling an organism together with the substrate upon which it is feeding (particularly on and in thinly-stratified sediments) strictly limits inferences of selectivity. Fortuitous mini substrate samples from the vicinity of the mouth, taken by 2 llving specimens of the large (shells > 30 mm high) sediment-ingesting trochid gastropod Bathybernbix aeola (Watson) at the t ~ m e of their collection (R/V Soyo Maru sta. T 2 3 , 3504.6'N, 13929.0'E, 700 m) provide nearly ideal data for comparison with gut contents. When disturbed, a gastropod crawling on a soft substrate may retract into its shell carrying sediment that is trapped behind the tight-fitting operculum. An individual is thus occasionally capable of providing 2 samples of its feeding substrate: one taken by the retracting foot and the other by the feeding organ or radula. Paired, snail-collected sediment and gut samples from the two individuals were analyzed separately (4 runs) with a zone-sensing particle-size analyzer, revealing replicate distributions of particle size strongly skewed toward fine sand in the 2 substrate samples, with differential accun~ulation of the fine fraction in the gut samples. Fig. 1 shows a distinctly~ bimodal distribution of particles in the gut, with the largest mode in the clayand silt-size range. The virtual identity of grain-size distributions from the 2 specimens is remarkable. Inferences of selectivity are strengthened by examination of the radula as a morphological basis of particle selection. Bathybembix and a n allied group of deep-water deposit-feeding genera (Calliotropis, Cidarina, Ginebis, Convexia) have a distinctive radula (Fig. 2a, b, c) with several features that represent good design for ingesting fine particles from a mixed sediment (Hickman, 1980). The radula is of the rhipidoglossate type, with a longitudinal band of robust central 340 Mar. Ecol. Prog. Ser 6: 339-342, 1981
Read full abstract