Food Habits of the Gastropod litterata lamarck: Relation to Trophic Structure of the Intertidal Marine Bench Community in Hawaii' ALAN J. KOHN2

ABSTRACT: The stenoglossan gastropod Mitra litterata Lamarck, common on intertid al solution benches on Oahu, preys on sipunculids occurring in burrows in reef limestone. Phascolosoma scolops Selenka and DeMan is the main component of the diet. Phascolosoma sp. cf. P. beronis Edmonds and Aspidosiphon elegans Chamisso and Eysenhardt are also eaten. This is the first report of the nature of the food of a member of the family . Mit ra litterata belongs to a distinct trophic subweb from other co-occurring predatory gastropods of similar size and population density-Morula gramtlata (Duclos), which preys mainly on herbivorous gastropods, and Conus spp., which eat herbivorous polychaetes.

RELATIVELY SMALL to moderate-sized predatory abundant predatory gastropod . Two small quan­ gastropod molluscs of the genera Conus, Momla, titative samples, at Kahuku and Kapoho Point, and Mitra, all belonging to the prosobranch Oahu, contained 3 individuals in 65 m2 (0 .051 suborder Stenoglossa, are abundant and promi­ m2) and 2 individuals in 9.3 m2 (0.22 /m2), nent carnivores in the biotic community of inter­ respectively. This note reports a study of the tidal solution benches that comprise almost trophic position of M. litterata in the solution one-third of the shoreline of Oahu, Hawaii bench community. M. litterata probably preys (Wentworth, 1939). Up to eight species of exclusively on sipunculids, thus exploiting a Conus (superfamily Toxoglossa), with mean resource not known to be utilized by any other shell lengths 20-35 mm, occur at densities of co-occurring gastropods. Little is known of other 0.2-0.6/m2. Most are specialized predators on predators of sipunculids: Fischer (1925) lists different species of Eunicidae and Nereidae, anemones (Sagartia and others), crabs (Car­ families of herbivorous polychaetes (Kohn, cinus and Eupagurus) , and fishes (plaice and 1959). cod), but he gives no further information. In order to determine the position of the Virtually no information exists on the nature other predatory gastropods in the trophic struc­ of food and feeding in the Mitridae, although ture of the community and to relate this to their members of the suborder Stenoglossa are gen­ taxonomic position, their feeding activities have erally assumed to be carnivores. Cernohorsky been examined. Morula granulata (Duclos) (1965) noted that spp. are attracted (superfamily Muricacea) is smaller than the by carrion, on which they feed. In a popular Conus spp. (mean shell length 14 mm) but book, Abbott (1962) mentioned worms and more numerous (81m2), and it eats only other clams as food of mitrids but gave no further molluscs, mainly small browsing gastropods of information. Risbec (1955) proposed a toxo­ the genera Ceritbium, Rissoina, Heliacus, and glossan feeding mechanism in the Mitridae, but Bittium, and the mussel H ormomya (Miller, this is incorrect. The mitrid radula is typically 1968). rachiglossan and is borne at the end of an ex­ Mitra litterata (superfamily Volutacea) tremely long pleurembolic proboscis, which may (mean shell length 17 mm) is the second most extend the length of the shell. In , the subgenus to which Mitra litterata belongs, 1 Contribut ion N o. 356, Hawaii Institute of Marine the radular teeth are laterally elongate. Cerno­ Biology. Manuscript received February 25, 1970. horsky (1966) described and figured the radula 2 Department of Zoology, University of Washing­ ton, Seattle, Washington 98105 . of M. litterata. The rachidian teeth of the sped- 483 484 PACIFIC SCIENCE, Vol. 24, October 1970

mens I examined had four, and the lateral teeth In no case was the "accessory proboscis" visi­ seven or eight conical denticles of nearly uni­ ble during the feeding process. This muscular, form height. vermiform structure, considered by Risbec (1955) to release a venom, originates in the proboscis behind the radula and was observed OB SERVATIONS to extend and retract through the mouth during Feeding dissection of living specimens. Its function re­ The feeding process of M . litterata has been mains unknown. observed in the laboratory, specimens have been The prob oscis and radula appear to fun ction observed feeding in nature, and the food of exclusively as conveyors, capturing the sipun­ others in nature has been identified from the culid and moving it, intact and undamaged, examination of alimentary tracts of preserved into the mid-esophagus and stomach. In one specimens and of fecal material recovered from instance, Mrs. S. Miller collected a Mitra lit­ specimens kept alive in the laboratory after terata in the act of swallowing a Phascolosoma collection. In every case, the food has been a scolops Selenka and DeMan . The predator was sipunculid worm of species that characteristically dissected five hours later, revealing the prey occur in burrows in calcareous rock. The papillae in the stomach, markedly distending the stomach and hooks of sipunculids pass undigested wall. The trunk of the sipunculid was mostly through the gut of Mitra, facilitating identifica­ intact, the introvert mostly digested; all damage tion of the prey. Species identifications were appeared to have been caused by chemical diges­ made primarily from the monograph of Selenka, tion. The sipunculid was removed from the DeMan, and Bulow (1 883). stomach and remained alive until it was fixed Observations on feeding in the laboratory two hours later. were made at night, when M . litterata is nor­ Food in Natnre mally active and when the intro verts of burrow­ ing sipunculids are extended. When placed on Results of examination of the alimentary tracts a piece of reef limestone containing one or more and feces of Mitra litterata are summarized in specimens of Aspidosiphon elegans Chamisso Tabl e 1. Of the 24 specimens examined , each of and Eysenhardt, Mytra litterata extends the si­ 16 contained the remains of one sipunculid. All phon along the substrate, waving it from side to specimens were collected during the morning or side. When the siphon contacts a sipunculid, the late at night, and it is likely that they fed at head is extended and also moved from side to night. Furthermore, attempts to induce feeding side until one of the extended tentacles contacts in the laboratory succeeded only at night. The the worm . Usually the proboscis extends only after both siphon and tentacle have touched the T ABLE 1 sipunculid. The padlike proboscis lip, bearing G UT AND F ECES CONTENTS OF 24 Mitra litterata projecting, nonmotile, and evidently sensory INDIVIDUALS ON I NTERTIDAL BENCHES AT cilia, then contacts the sipunculid and moves KAH UKU, K APO HO POINT, AN D H AU ULA , OAHU over its exposed body surface. The proboscis begins to engulf the sipuncul id when the lips No. are affixed to the introvert, or to the anterior FOOD IT EM F OUND M . litterata end of the trunk if the introvert is withdrawn. Sipunculids The rate of ingestion probably varies with the Phascolosoma scolops size of the sipunculid. In laboratory observa­ (Selenka and D eMan) 9 tions, M. litterata did not succeed in extracting Phascolosoma sp. d . A spidosipbon elegans from its burrow; it re­ P. beronis Edmonds 3 quired several minutes to ingest a small speci­ Aspidosiph on elegans (C hamisso and Eysenhardt ) 4 men of A . elegans removed from its burrow and placed in a dish. The distal end of the proboscis Total sipunculids 16 is not expansible and probably limits the diam­ Minute, unidentifiable, eter of prey that can be eaten. soft fecal material 3 Food H abits of M itra litterata-KoHN 485

rate of passage of food through the gut is not minimal data presented above, is about 0.08 known, but, if the mean number of food items sipunculid eaten per square meter per day. M. per gut is taken as the daily feeding rate, as litterata would thus consume about 5 percent of seems reasonable from studies of Conus of sim­ the standing crop of sipunculids per year, an ilar size (Kohn, 1959) , this rate is about 0.7 estimated rate quite similar to the feeding rate sipunculid per M itra litterata per day. of Conus on polychaetes at Kahuku (Kohn, The sipunculid fauna of Hawai i is incom­ 1959) . pletely known . Edmonds ( in preparation) lists

four species of Pbascolosoma, but not P. scolops, DISCUSSION the commonest prey of M itra litterata on inter­ tidal benches. Except for slight differences in M itra litterata preys on Phascolosoma and the platelets of the trunk papillae, the specimens Aspidosipbon on intertidal solution benches, ap­ studied agree in all details with the original parently extracting the sipunculids from their description of Phascolosoma scolops. However, burrows with the long proboscis and radula and the identification should be regarded as tenta­ conveying the prey, intact and alive, to the tive because the of this group of capacious stomach. It thus exploits a food re­ sipunculids is not settled . Th e A spid osi­ source that is not utilized at all by M omla ph on has been reported from Hawaii by Ed­ gramtlata and Conus spp., co-occurring preda­ mondson (1946) and Edmonds (in prepara­ tory stenoglossan gastropods of roughly similar tion) , but neither author identified the species. size and population density. The diets of the The commonest sipunculids in areas where three genera, belonging to three different super­ Mitra litterata occurs are Phascolosoma scolops families, do not overlap, and the different spe­ and A spidosiphon elegans, found in burrows, cies are the primary carnivores in distinct food probably of their own construction, in the reef subwebs. limestone . The species listed as Phascolosoma Paine ( 1963) described in detail a subweb of sp. d. P. beronis Edmonds (Table 1) differs predatory gastropods on an intertidal sand bar from the original description of that species only in Florida. On Hawaiian solution benches, how­ in the pigmentation of the tentacles, the nephri ­ ever, there is no equivalent of the large fasci­ dium, and the spots on the trunk. A fourth olariid gastropods foun d there, that prey gener­ sipunculid, Phascolosoma pelma Selenka and ally on smaller carnivorous gastropo ds. This DeMan, previously reported from Hawaii by is probably because of the hard, rather smooth Edmondson (1946) and Edmonds (in prepara­ topography of solution benches and their ex­ tion) , was found in the same habitat but was posure to desiccation at low tide and to violent not eaten by any of the M itra litterata examined. wave action at high tide (Kohn, 1967) . The Phascolosoma spp. share the same microhabitat habitat does not provide shelter suitable for as A spidosiphon but may be foun d also in sand large, motile invertebrates, and selection appears bound by algae and on the surface of to favor small body size (Kohn, 1968) . Th e the reef rock, and in crevices. Several species of harshness of the environment may thus effec­ sipunculids of these genera frequently co-occur tively preclude a secondary carnivore level in in adjacent burrows in the same rock (Rice, the trophi c structure of the benthic community. 1969). Four 100-625-cm2 samples from the solution ACKNOW LEDGMENTS bench at Kahuku (Sta. 5 of Kohn, 1959) con­ tained 6 to 26 sipunculids each and gave a mean This study was supported by National Science density of 470j m 2• Phascolosoma scolops was Foundation grant GB-6134, and in part by GB­ the most abundant species. A substrate sample 5942X. I thank participants in the Summer of 270 em- from the solution bench at Kapoho Training Program, Hawaii Institute of Marine Point contained 20 sipunculids, of which 19 Biology, for aid with the field work; Dr. Mary were Aspid osipb on elegans, suggesting a density E. Rice for aid in identifying the sipunculids ; of 740 jm2. A crude estimate of the feeding rate Dr. Vera Fretter and Dr.R. T. Paine for dis­ of the M itra litterata population, based on the cussion and criticism of the manuscript; and 486 PACIFIC SCIENCE, Vol. 24, October 1970

Dr. 1. G. Eldredge for providing the manu­ on Indo-West Pacific reef platforms. Ameri­ script of the second edition of Reef and Shore can N aturalist, vol. 101, pp. 251-259. Fauna of Hawaii. --- 1968. Microhabitats, abundance and food of Conus ( ) on atoll reefs in the Maldive and Chagos islands. Ecology, LITERATURE CITED vol. 49, pp . 1046-1062. ABBOTT, R. T . 1962. Sea shells of the world. MILLER, A. C. 1968. The ecology of M orula New York, Golden Press. uva and M orula granulata on Checker Reef. CERNOHORSKY, W. O. 1965. The Mitridae of Unpublished report, Hawaii Institute of Fiji. Veliger, vol. 8, pp. 45- 61. Marine Biology. CERNOHORSKY, W. O. 1966. Radulae of four PAINE, R. T. 1963. Trophic relationships of 8 species of Mi trid ae ( : Gastropoda). sympatric predatory gastropods. Ecology, vol. Journal of Conchology, vol. 26, pp. 15-18. 44, pp. 63-73. EDMONDS, S. J. In preparation. Phylum Sipun­ RICE, M. E. 1969. Possible boring structures of cula and Phylum Echiura. In: 1. G. Eldredge, sipunculids. American Zool ogist, vol. 9, PI' ed., Reef and shore faun a of Hawaii. 2nd ed. 803-812. Honolulu, B. P. Bishop Museum. RISBEC, J. 1955. Considerations sur l'anatomie EDMONDSON, C. H . 1946. Reef and shore fauna comparee et la classification des gasteropodes of Hawaii. Honolulu, B. P. Bishop Museum. prosobranches. Journal de Conchyliologie, vol. FISCHER, W . 1925. Echiuridae, Sipunculidae, 95, pp. 45-82. Priapulidae. Tierwelt der Nord- und Ostsee, SELENKA, E., G. D EMAN, and C. BULOW. Lief. I, Teil VId. 1883-1884. Die Sipunculider. In: C. Semper, KOHN, A. J. 1959. The ecology of Conus in ed., Reisen im Archipel der Ph ilippinen, Hawaii. Ecological Monographs, vol. 29, Teil II, Bd. 4, Apt. 1, pp. 1-131. pp. 47- 90. WENTWORTH, C. K. 1939. Marine bench-form­ - -- 1967. Environmental compl exity and ing processes. II. Solution benching. Journal species diversity in the gastropod genus Conus of Geomorphology, vol. 2, pp. 3-25.