Foraging Behavior of the West Atlantic Trumpetfish, <I

166 . IlULLETINOFMARINESCIENCE.VOL.33. NO.I. 1983

Channel). Harpiliopsis beaupresii (Audouin, 1852)-(Mauritius). Harpi/iopsis depressa (Stimpson, 1866)-(Zanzibar). Harpiliopsis spinigera (Ortmann, 1870)-(Farquhar Is., Seychelle Islands).· Ha- modactyloides incompletus (Holthuis, 1958)-(Zanzibar).·

LITERATURE CITED

Bruce, A. J. 1978. A report on a collection of pontoniine shrimps from Madagascar and adjacent waters. Zool. Joum. Linn. Soc. 62: 205-290, figs. 1-44. . --. 1979. Onycocarisfurculata sp. nov., a new pontoniine shrimp from La Reunion. Cahiers Indo-Pacifique 1: 323-334, figs. 1-4. --. 1980. Notes on some Indo-Pacific Pontoniinae. XXXII. The occurrence of Paratyptol1 siebenrocki Balss on La Reunion. Crustaceanana 38: 237-246, figs. 1-4. Ribes, S. 1978. La macrofaune vagile associee ii la partie vivante des scleractinaires sur un recif frangeant de I'Be de la Reunion (Ocean Indien). These de doctorat de 3 erne cycle en oceanologie, Universite Aix-Marseille 2.

DATEACCEPTED: October 21, 1982.

ADDRESS: Division of Natural Sciences, The Northern Territory Museum. P.O. Box 4646. Darwin, Australia. 5794.

IH'LI.ETINOFMARINESCIENCE.33(1): 166-171. 1983

FORAGING BEHAVIOR OF THE WEST ATLANTIC TRUMPETFISH, AULOSTOMUS MACULA TUS: USE OF LARGE, HERBIVOROUS REEF FISHES AS CAMOUFLAGE

Richard B. Aronson

A limit to the density of Batesian mimics of an unpalatable species has been postulated, due to predator learning (Carpenter, 1936; Rothschild, 1971). Simi- larly, one might expect a limit to the frequency with which a predator camouflages itself in a particular manner in order to capture visually-oriented prey. The west Atlantic trumpetfish, Aulostomus maculatus Valenciennes, 1842 (Aulostomidae) is a coral reef species that feeds primarily on fishes (Randall, 1967), exhibiting several modes of foraging that employ camouflage. The evolutionary problem of prey recognition may have selected for a limit to the frequencies with which A. maculatus utilize their various forms of camouflage. Here I discuss one type of camouflage: A. maculatus associate with large, herbivorous fishes, apparently to disguise themselves. A detailed examination of the associations was undertaken to determine whether they are, in fact, modes of foraging. The prediction of a limit to the use of particular modes of predator camouflage was tested by observing the frequencies of those modes. Aulostomus maculatus commonly hunt by ambush. They suspend their elongate bodies, head down, over the bottom and dart down to attack prey (Randall, 1968). They are known to align with gorgonians, long sponges and elongate artificial objects (e.g. ropes). Such behavior apparently conceals trumpetfish from their intended prey (Eibl-Eibesfeldt, 1955; Randall, 1968; Scaff, 1980). Randall (1968) describes A. maculatus as usually "brown or reddish brown, with lengthwise pale lines, scattered small black spots, and a black streak on the upper jaw .... " They are, however, able to change color. Randall (1968) reports color phases in which the upper half of the head is yellow or purple. Color change NOTES 167 may aid A. maculatus in foraging by providing camouflage. For example, individuals that follow, and hunt in, schools of blue chromis, Chromis cyanea, often adopt bright blue snouts (Kaufman, 1976). In another form of camouflage (the subject of this paper), A. maculatus align with, or swim in schools of, other fishes of approximately the same length. Trum- petfish have been described as aligning with parrotfishes (Scaridae; Eibl-Eibesfeldt, 1955; Scarr, 1980), groupers (Serranidae; Eibl-Eibesfeldt, 1955; Collette and Tal- bot, 1972; Scarr, 1980) and the Spanish hogfish, Bodianus rufus (Collette and Talbot, 1972). When aligning with another fish, " ... the trumpetfish's body arches in line with the other fish's spine" (Scarr, 1980). Collette and Talbot (1972) noted that an A. maculatus aligning with a graysby, Epinephelus cruentatus, was brown, like the grouper. A yellow phase individual was observed aligning with a red and yellow B. rufus (Collette and Talbot, 1972). Kaufman (1976) found many blue-snouted A. maculatus swimming in groups of blue tang, Acanthurus coeruleus. Eibl-Eibesfeldt (1955) observed A. maculatus attacking prey from their po- sitions of alignment with parrotfishes, and felt certain that the purpose of such alignment was camouflage. Hobson (1968) noted that Pacific cornetfish, Fistularia petimba, also use parrot fishes and groupers as camouflage to approach prey.

METHODS AND STUDY AREA

Field observations were made during 19-26 January 1981 while snorkeling in and around a shallow (-1.5 to -3.0 m) reefa few meters off the west (leeward) coast of Bonaire, Netherlands Antilles. The reef is at Plaj'i Lechi, just north of the town of Kralendijk, at approximately 12°10' IO"N latitude and 68°17' IO"W longitude. It consists primarily of the branching corals Acropora cervicornis and A. pal- ma/a. and vertical sheets of the hydrocoral Millepora complana/a. Aulos/omus macula/us were commonly seen around this reef. Formal observations (behavioral sequences and reef surveys) were made between 0830 and 1100 h. Incidental observations were made at all hours of the day. Eighteen behavioral sequences of A. macula/us were recorded over a period of 5 d. After standard length (SL) was estimated and color noted, each individual was watched until lost from view, or for a maximum of 5 min. All changes in behavior, color and orientation (horizontal vs. vertical) were recorded. Prey items and intended prey items were identified when possible. On two mornings (19 and 20 January), visual surveys were conducted to determine the proportions of A. macula/us engaged in various modes offoraging. These surveys were made by swimming around the reef and noting the activity of each A. maculatus at the moment of first sighting.

RESULTS Three Aulostomus maculatus were observed aligning with parrot fishes during the recording of behavioral sequences, two during the reef surveys and seven during incidental observation. Six individuals aligned with terminal phase queen parrotfish, Scarus vetula, two with initial phase Scarus vetula, and four with initial phase stoplight parrotfish, Sparisoma viride. The interactions were much as described by Eibl-Eibesfeldt (1955) and Scarr (1980), with the A. maculatus generally conforming very closely to the dorsal profiles of the parrotfishes. The colors of the trumpetfish conformed, in most cases, to those of the parrotfishes (Ta- ble 1). On two occasions A. maculatus were observed associating with terminal phase Scarus vetula for > I min. These A. maculatus foraged when the Scarus vetula stopped to feed on algae; the trumpetfish examined small cavities in the reef, and, one, made strikes at several Eupomacentrus partitus. The other encounters were considerably shorter (.::5 30 s). In these instances, the trumpetfish aligned briefly and then swam off. One individual made an unsuccessful strike at an unidentified goby from its position of alignment with an initial phase Sparisoma viride. In another case, a terminal phase S. vetula chased the aligning A. maculatus and 168 BUllETIN OF MARINE SCIENCE. VOL 33. NO. I. 1983

Table I. Summary of color patterns of A ulostomus maculatus aligning with parrotfishes and swimming in schools of Acanthurus spp. Colors reported for Aulostomus maculatus are thosc recorded at moment of first observation (g-s, gray with longitudinal, silver stripes; greens and blues are irridescent shades for A. maculatus)

Number Herbivore(s) Aulostomus maculalus Color Observed Terminal phase Scarus vetula g-s venter, light green dorsum and snout 2 (blue, pink and green pastel gray body, blue snout 2 shades) g-s venter, light green dorsum, blue snout I g-s body, light green snout I Initial phase S. vetula g-s body and snout 2 (black, white and gray) Initial phase Sparisoma viride g-s body and snout 2 (red and gray) brown and silver lateral stripes, gray snout I gray body, blue snout I

School of Acanthurus spp. gray body, blue snout 7 (gray and irridescent blue) dark gray body, blue snout 2 g-s body and snout 2 g-s body, blue snout I gray body and snout I g-s body, light green snout I nipped it. An initial phase Sparisoma viride did the same, and another initial phase S. viride chased away an A. maculatus with a flick of its tail. In at least two observed cases, trumpetfish were in direct contact with parrotfishes, and appeared to be riding their dorsa and/or leaning on their extended pectoral fins. A group consisting of 30 to 50 ocean surgeons, Acanthurus bahianus. 10 to 15 doctorfish, A. chirurgus, and/or blue tang, A. coeruleus, and two to five goatfish, Mulloidichthys martinicus and Pseudupeneus maculatus, was seen on the reef on numerous occasions; this appeared to be the same school in all cases. The school was seen 11 times during formal and incidental observations. It contained no Aulostomus maculatus on four occasions. Twice the school had one trumpetfish, three times it contained two individuals, and twice there were three A. maculatus present. The herbivores moved through the reef, invading damselfish (Pomacen- tridae) territories and cropping their algal lawns (see Ogden and Lobel, 1978 for review). The trumpetfish searched while the herbivores fed; when the herbivore moved on, the A. maculatus rejoined the group. Two trumpetfish made strikes at unidentified objects while they were swimming with the school. Aulostomus maculatus which swam with the school generally had bodies that were solid gray or gray with longitudinal, silver stripes, and most had irridescent blue snouts (Table 1). The gray coloration matched the bodies of Acanthurus bahianus and A. chirurgus, and the blue snouts conformed to the irridescent margins of the dorsal and anal fins of A. bahianus. as well as to the overall body coloration of A. coeruleus. Support for the idea of camouflage comes from a behavioral sequence in which two Aulostomus maculatus actually changed color to match changes in the fishes with which they were swimming. The two trumpetfish were light gray (like the Acanthurus spp.) with blue snouts. At one point during the observations, the Acanthurus spp. suddenly turned dark gray, and the A. maculatus immediately did the same (they retained their blue snouts). One minute later, the Acanthurus spp. returned to their light gray coloration, and the two trumpetfish did as well. The most common mode of foraging was "solitary" foraging, in which an A. NOTES 169

Table 2. Frequency distribution of Aulostomus maculatus foraging modes observed during two surveys of the study area

Foraging Mode Frequency "Solitary" foraging 33 Swimming in school of Acanthurus spp. 5 Aligning with parrotfishes 2 Swimming in school of Chromis multilineatus 1 x' = 68.17; df= 3; P < 0.005 N=41 maculatus swam 0.5 m or less above the substrate, stopping occasionally to search. If a small fish was spotted on the substrate, the A. maculatus would suspend itself in the water column, either horizontally or vertically. Vertical suspension generally indicated an incipient lunge, and attacks were as described by Randall (1968). Most solitary foragers were gray with longitudinal, silver stripes. The frequencies of the various foraging modes observed during the surveys are presented in Table 2. There is a significant deviation of the results from random expectation. Thus, solitary foraging was very common, whereas the other three modes were relatively rare.

DISCUSSION AND CONCLUSIONS It seems clear that Aulostomus maculatus swim with large, herbivorous reef fishes in order to camouflage themselves from their prey, Of particular significance are the foraging behaviors oftrumpetfish and their color changes when associated with parrotfishes and Acanthurus spp. In general, A. maculatus changed color to conform to the colors of the fishes with which they swam. Alternatively, they swam with herbivores that were shades of gray (the solitary foraging coloration of A. maculatus; Table I). A second function of the associations may be the concealment of A. maculatus from their predators (Randall, 1967) and/or competitors (e.g. other A. maculatus; see below). Robertson et al. (1976) list three suggested functions of the association ofpred- ators with schools of the striped parrotfish, Scarus croicensis: (1) predation on animals disturbed by the feeding herbivores (suggested for A', maculatus and the barred hamlet, Hypoplectrus puella), (2) predation on damselfishes when their territories are invaded by the herbivores (suggested for H. puella), and (3) predation on the S. croicensis themselves (suggested for A. maculatus). The first two of these may apply to A. maculatus swimming with Acanthurus spp. schools as well; however, all of the fishes in the school studied seemed much too large for consumption by trumpetfish, rendering the third function highly unlikely. Only 17% of the trumpetfish sighted were associated with parrotfishes and Acanthurus spp. during the surveys. G. Helfman (pers. comm.) has observed an even lower frequency of these behaviors at St. Croix, United States Virgin Islands. There are several possible explanations for the relative rarity ofthese camouflage phenomena, and they are not mutually exclusive. The first concerns the abundance of the large herbivores: there was only one school of Acanthurus spp. on the reef, probably limiting the number of A. maculatus that could use those fishes for camouflage (Table 2). On the other hand, there were numerous parrotfishes present without associated A. maculatus. Second, aggression by parrotfishes toward aligning A. maculatus is a likely reason for the rarity of that association. The aggressive parrotfishes may have been attempting to prevent trumpetfish from riding them, such riding causing 170 BULLETIN OF MARINE SCIENCE. VOL. 33. NO. I. 1983 unnecessary energy expenditure by the former. Also, aligning A. maculatus some- times interfered with parrot fish feeding; at times their snouts were so near to the parrotfishes' mouths that the latter ceased feeding (pers. obs.). Third, a high frequency of trumpetfish associating with large herbivores might allow fish prey species to learn to recognize the camouflaged predators. Alterna- tively, prey species might simply learn to avoid all parrotfishes and all schools of Acanthurus spp. Dill's (1973) theoretical considerations indicated that an increase in fish prey reactive distance due to experience may increase fish predator pursuit time and decrease capture success. As mentioned previously, the hypothesized effect is analogous to the expectation that an overabundance of Batesian mimics would obviate the adaptive value of the mimicry as a result of predator experience (Carpenter, 1936; Rothschild, 1971). If A. maculatus maintain feeding territories (I have seen individuals chase conspecifics from areas of reef, to which the aggressive A. maculatus returned; also G. Helfman, pers. comm.), then, by this explanation, the use oflarge herbivores as camouflage for a relatively small fraction of the time is advantageous in terms of individual selection; the frequencies of different modes of foraging ultimately may be explicable by optimality theory (Marten, 1973). If A. maculatus are not territorial, then this prey learning explanation does not apply. At any time, any individual can move through an area and use the camouflage afforded by large herbivores to capture naive prey or, at least, cause them to become experienced. The prey learning explanation might be tested by a removal experiment. A small reef could be kept free of parrotfishes and Acanthurus spp. for the generation time of a benthic fish species commonly eaten by camouflaged A. maculatus. Reintroduction of the large herbivores should result, initially, in a higher strike efficiency for camouflaged trumpetfish feeding on the prey species on the exper- imental reef than on a control reef. Further work on A. maculatus foraging behavior will provide insights into the advantages of, and constraints on, the use of camouflage. The findings will, no doubt have implications for mimicry theory and optimal foraging theory.

ACKNOWLEDGMENTS

I thank W. L. Fink, G. Helfman, L. S. Kaufman, K. P. Sebens, and two anonymous reviewers for their advice and criticism. This research was conducted as part of Harvard University's 1981 Tropical Ecology course, and was supported by the Atkins Garden Fund of Harvard University and a National Science Foundation Graduate Fellowship to the author.

LITERA TURE CITED

Carpenter, G. D. H. 1936. The facts of mimicry still require natural selection for their explanation. Proc. R. Soc. Lond., Ser. B 121: 65-67. Collette, B. B. and F. H. Talbot. 1972. Activity patterns of coral reef fishes with emphasis on nocturnal-diurnal changeover. Pages 98-124 in B. B. Collette and S. A. Earle, eds. Results of the Tektite Program: ecology of coral reef fishes. Nat. Hist. Mus. L.A. Cty. Sci. Bull. 14. Dill, L. M. 1973. An avoidance learning submodel for a general predation model. Oecologia 13: 291-312. Eibl-Eibesfeldt, I. 1955. Uber Symbiosen, Parasitismus und andere besondere zwischenartliche Be- ziehungen tropischer Meerfische. Z. Tierpsychol. 12: 203-219. Hobson, E. S. 1968. Predatory behavior of some shore fishes in the Gulf of California. U.S. Dep. Interior, Bur. Sport Fish. Wildl. Res. Rep. 73: 1-92. Kaufman, L. 1976. Feeding behavior and functional coloration of the Atlantic trumpetfish, Aulo- stomus maculatus. Copeia 1976: 377-378. Marten, G. G. 1973. An optimization equation for predation. Ecology 54: 92-101. Ogden, J. C. and P. S. Lobel. 1978. The role of herbivorous fishes and urchins in coral reef com- munities. Env. BioI. Fish. 3: 49-63. NOTES 171

Randall, J. E. 1967. Food habits of reef fishes of the West Indies. Stud. Trop. Oceanogr. 5: 665- 847. ---. 1968. Caribbean reef fishes. T. F. H. Publications, Hong Kong. Robcrtson, D. R., H. P. A. Sweatman, E. A. Hetcher and M. G. Cleland. 1976. Schooling as a mechanism for circumventing the territoriality of competitors. Ecology 57: 1208-1220. Rothschild, M. 1971. Speculations about mimicry with Henry Ford. Pages 202-223 in R. Creed, ed. Ecological genetics and evolution: essays in honour of E. B. Ford. Blackwell Scientific Pub- lications, Oxford. Scarr, D. 1980. Trumpetfish: the advantages of being thin. Sea Frontiers 26: 160--164.

DATEACCEPTED: October II, 1982.

ADDRESS: The Biological Laboratories. Harvard University. Cambridge. l'vfassachusetts 02 J 38.

IJllLl.ETIN OF MARINE SCIENCE, 33(1): 171-172, 1983

EXTRAORAL FEEDING IN LUIDIA CLATHRATA (SAY) (ECHINODERMATA: ASTEROIDEA)

James B. McClintock, Thomas S. Klinger and JohnM. Lawrence

One of the most important evolutionary steps of the Asteroidea was the tran- sition from intraoral to extraoral feeding which allowed access to many more food sources. Extraoral feeders may feed on detritus in addition to macrofauna. In- traoral feeders are generally limited to soft substrata while extraoral feeders may utilize both hard and soft substrata (Jangoux, 1982). Members of the Luidiidae are considered to be intraoral feeders exclusively. This family is believed to have a number of primitive traits including an intraoral mode of feeding, Luidia clathrata has been consistently reported as an intraoral feeder (Hulings and Hemlay, 1963; Lawrence et a1., 1974; Lawrence and Dehn, 1979; McClintock and Lawrence, 1981). The purpose of this study was to investigate the ability of this asteroid to feed via extrusion of the cardiac stomach.

MATERIALS AND METHODS

Field observations were made in December 1981 in Old Tampa Bay, Horida in 3 m of water. Using SCUBA a total of231 individuals were noted for their level of activity (n = 160 unburied and moving on the substratum and n = 71 buried and immobile) and then removed from the substratum and the position of the cardiac stomach noted. To investigate whether organic-rich sediment could promote stomach eversion, 10 individuals were placed in an aquarium containing sediment fortified with a fine organic powder (30% protein, dry wt.) dyed with Nile blue (100:1, sediment: powder, w:w). Ten other individuals were placed in an aquarium containing sediment with no added organic material. Individuals were checked for stomach eversion at 3-h intervals over a 12-h period. At the end of this 12-h period, individuals from the aquarium with dyed enriched sediment were dissected and the contents of the cardiac and pyloric stomachs were observed under a dissecting microscope.

RESULTS The stomach was extruded and applied to the substratum in 32% of the individuals observed in the field. This included 62% of the buried individuals, but only 18% of the individuals on the surface of the substratum. Organic-rich sediment promoted stomach eversion in the individuals in aquaria.