BULLETIN OF MARINE SCIENCE, 28(1): 181-188, 1978 CORAL REEF PAPER

INTRASPECIFIC AGONISTIC BEHAVIOR IN THE ROCK-BORING LUCUNTER (L.) (ECHINODERMATA: ECHINOIDEA)

Herman Griinbaum, Glen Bergman, Donald P. Abbott, and John C. Ogden

ABSTRACT The agonistic behavior of the sea urchin (L.) was studied in the algal ridge reefs of Boiler Bay, St. Croix, U.S. Virgin Islands, where it inhabits burrows in the reef surface. Sixty-four encounters were set up by placing an urchin (intruder) at the opening of an occupied burrow (host). In 46 cases, agonistic behavior between the host and intruder resulted. Most of the interactions involved pushing, but biting was fre- quently observed. The encounter usually resulted in the eviction of the intruder from the burrow, but occasionally the host was evicted or the urchins cohabited the burrow. Burrow occupants successfully evicted all intruding urchins of their own or slightly larger size, and regained possession of their burrows after being experimentally replaced by an urchin of equal size. As Echinometra live in dense aggregations in wave-swept areas, this behavior may be significant in defending protected living space and access to a food sup- ply.

Many echinoids in the sea today are gre- in a variety of habitats. It is most abundant garious and form homotypic and heterotypic on tidal terraces, beachrock, and other areas aggregations which have been considered to of rocky shore from just below the low tide be the result of individual reactions to the level to depths of a few meters, especially in physical environment or short-term aggrega- regions swept by surge and surf (Ginsburg, tions for reproductive activities, and not the 1953; Kaye, 1959; Lewis, 1960; McPher- result of any social behavior (Reese, 1966). son, 1969; Hunt, 1969; Teytaud, 1971; Ab- Pearse and Arch (1969), however, con- bott et aI., 1974; and earlier authors cited in sider that the large groups of diademid Mortensen, 1943). Here it usually inhabits echinoids common in tropical seas do in- holes and crevices that afford protection volve true social behavior. They present against dislodgment. evidence to support this view consisting of The northeast coast of St. Croix, U.s.V.I., group responses to disturbance, and aggre- in the area of Boiler Bay, has well developed gation behavior involving particular patterns algal ridge reefs and beachrock benches of spine contact. Other authors have exam- which are extensively tunnelled and bur- ined different sorts of aggrega- rowed by E. lucunter (Adey, 1975 and Ab- tions and have concluded that some form of bott et aI., 1974 for detailed maps of the social interaction is involved (Dix, 1969; area). Within this area, as elsewhere, most Branham et aI., 1971; Warner, 1971). 80- are in individual pits or burrows cial behavior involving agonistic encounters (Fig. 1). Others are arranged along con- has been recently reported for all three or- tinuous crevices, and in areas where urchin ders of sea stars by Wobber (1975) and de- boring has been extensive, adjacent bur- scribed for a few . rows have run together and a small group Echinometra lucunter (L.), a widely dis- of animals occupies a larger open depres- tributed sea urchin in the tropical and sub- sion. However, even in situations where ad- tropical Atlantic (Mortensen, 1943; Clark, jacent animals are not separated by obvious 1954) is found abundantly in shallow waters physical barriers, they are always out of 181 182 BULLETIN OF MARINE SCIENCE, VOL. 28, NO. I, 1978

Figure 1. Echillometra [ucullter in burrows in one of the algal ridge reefs at Boiler Bay, St. Croix. Note close spacing of burrows and benthic algae at burrow openings. reach of one another. Repeated observa- appeared in a preliminary account (Griin- tions on selected individuals, day and night, baum and Bergman, 1974). show that each has its home site, distinct from that of any other individual. It METHODS AND MATERIALS moves short distances, especially at night, From January 11 to 20, 1974, observa- but each animal clearly "homes" in the tions were made on the agonistic behavior of classic sense, and ranges over a narrow area. E. lucunter living in the algal ridges of Boiler One rarely finds two individuals in the same Bay. The ridges, located within 50 m of the burrow, or two with spines, tests, or tube shoreline, afford an excellent site for ob- feet touching; when one does, the situation servation of the urchins in their burrows, as is far from static (Abbott et al., 1974). well as providing protection from the surge Prior to the present study, Griinbaum, characteristic of this bay. Observations were Warner and Ogden (unpubl.) noted the made both day and night by observers rapid disappearance of tagged E. lucunter standing on the bottom near the ridges us- which were released in areas with extensive ing snorkel gear. All burrows chosen for ex- burrows of conspecifics. On closer exam- perimentation were completely submerged ination it was discovered that the tagged during observations. As the ridges are ex- urchins were being forcibly prevented from tensively burrowed, many burrows run to- taking residence in occupied burrows. The gether and thus have multiple entrances, present study documents this behavior and large openings, or complex structures. Ex- assesses its importance in the establishment perimental situations were set up by remov- and stabilization of urchin populations in ing urchins of various sizes from their bur- burrows. Some details of the work have rows and moving them to the entrances of GRONBAUM ET AL.: BEHAVIOR IN THE SEA URCHIN ECHINOMETRA 183

BURROW TOO SMALL TO INTER IN'mUDEA MOVED AWAY

~ NO VISIBU! CONTACT BETWI(N ~ HOST AND INTRUDER NO 'IGHT.. >1 INTRUDER un ~ ," BURROW ~ IPINI tONIACT BITWIIN . HOST AND INTAUDiA

INTRUDIR ATTEMptED FUR'tHEA ADVANCE 1 BUT WAS PUSHED OUT ~ 0' BURROWBYHOS'> HOST ADVANCED AND INTRUDER LEfT PUS~~~R~::RINS' ~ BURROW INTRUDER RUIiAnD WIIHOUT 'URTHER .PUSHING BY HOSt

INnUDII PLACID A'"IN- [),. 'NTRUDER .HTlRED D TRAHC' r~.H6~~T BURROW \)HOST BUIAOW HOST PUSHED .0 BURROW . fiGHT INVOLVED 6 INTRUDER ADVANCED ~ OPENING, LE" BURROW 46 . PUSHINGIONLY ~ AND PUSHED AGAINST HOST ~ BURROW LARG' 'NOUGH '011 ~ two. IHfRUDER DISPLACED 26 HOST BUT BOTH CONtiNUE TO OCCUpy BURROW

FIOHT B RESULTED ':~.~~:~,~~T~:L~~::CTED 80T" HOST AND 'NTAUDER UUD ~'NTRUD'R ACTIVELy.nR'ATID MOUT" (91 FROM HOST, LEFT BURROW PlOHT INVOLVED IITING AS WELL AS PUSHING ONl Y Hon USED HOST ACTIVELY IVlcnD IY MOUTH (7) PUSHING OF INTRUDIR, I UFT IURROW ONl'f INUUI.'IU } USED MOUTH (4) aURROW LARGI ENOUGH fOIl lWOI INTRUDeR I.'IISP1.ACED 1 Hon BUT BC)IH CONTINUE TO 'OCCUpy BURROW

Figure 2. Agonistic behavior patterns between host and intruder. Figures and relative width of arrow indicates number of occurrences of each sequence. See text for details. already occupied burrows. The transferred visible contact with the host, (3) aU of the urchin was called the "intruder" and the oc- remaining 46 trials resulted in agonistic be- cupant of the burrow the "host." Sizes of havior. urchins were taken by measuring the largest The agonistic encounters began when the test diameter at the equator with a vernier intruder moved far enough into the burrow caliper (the test is often markedly oval). to make physical contact with the host, which reacted with immediate movement of RESULTS spines and tube feet. Once aware of the in- trusion, the host usually moved directly to- Sixty-four encounters were set up and ward the intruder, keeping spines and tube various sequences of agonistic behavior were feet in constant motion. In nine out of 20 noted (Fig. 2). Following the initial place- encounters (Fig. 2) in which the host gave ment of the intruder, one of three events an initial push to the intruder, the latter ac- occurred: (1) in three trials the intruder was too large for the opening of the burrow, tively backed off and left the burrow; in the and simply moved away, (2) in another 15 other 11 encounters the intruder resisted the trials there were no agonistic responses; in push but was forced all the way out by the 10 of these cases the intruder retreated im- host. While being evicted the intruder often mediately after making physical contact with flattened its spines against the test. In six the host in the burrow; in the other five cases the results were reversed, the intruder cases the intruder retreated prior to any pushing harder than the host; here the host 184 BULLETIN OF MARINE SCIENCE, VOL. 28, NO. I, 1978

Figure 3. Agonistic encounter between two E. lucunter in an aquarium involving the use of teeth. Photograph by C. Kitting. was either evicted altogether (N == 4) or in host (eight cases) or by cohabiting the bur- the case of larger burrows, was displaced to row. Cohabitation was observed in three in- one side and continued to cohabit the bur- stances; in all cases the cohabiting urchins row with the intruder (N == 2). took up positions well-separated from the Twenty trials resulted in the use of the host in the burrow. teeth by either host or intruder or both (Fig. The duration of the encounters varied 2). Here the attacker rotated its body until considerably. Interactions involving only tbe oral surface faced tbe opponent, then pushing lasted 3 to 30 min. When biting was protruded its Aristotle's lantern and pro- involved, the interactions lasted up to 5 h. ceeded to bite spines off of the opponent The greater the size difference between the (Fig. 3). Subsequent pushing, often with urchins, the shorter the encounter. continued biting, led to the eviction of the The correlation between the size of the opponent or to a counterattack. In the lat- urchins and the outcome of the encounter is ter case the attacked urchin rotated itself shown in Figure 4. Line A is a reference until the two animals were in a mouth-to- line showing host and intruder of equal size. mouth position. Mutual biting in the oral Line B is empirically derived from the ex- area eventually gave way to pushing and at perimental results. At all points below the last one of the urchins was usually evicted. line the intruder dominated the encounter, In all, 11 intruders out of 64 managed to evicting the host. When an intruder was at gain a home, either by evicting the original least 7 mm larger in diam than its opponent, GRONBAUM ET AL.: BEHAVIOR IN THE SEA URCHIN ECHINOMETRA 185 it remained in the burrow. The size range 6 of larger intruders was, of course, limited by the size of the burrow opening. Hosts al- /A ways evicted intruders of their own size or 5 smaller. In 19 cases hosts evicted intruders B o • 0/ E 00/ larger than themselves (Fig. 4), further in- u

4 ;: o 0 dicating that the host has a natural advan- III N o tage. To examine further this "host advan- l/) 00/00:: o 0 0 tage," five encounters were set up between :;; 3 o 0 0 o 0 urchins of equal size. In each, a host urchin o o 0 0 0 z: was removed from its burrow and an intrud- o~ • er was placed within. The host was then 2 o;<~ placed at the entrance of its burrow and thus made to assume the role of intruder. In all 0/ five cases the original host entered its bur- 2 3 row, engaged in battle, and eventually re- 4 5 6 gained possession of its home. This type of INTRUDER SIZE (eml interaction was usually lengthy and involved Figure 4. Size oC host and intruder and outcome biting by both urchins. of encounter. Open circles indicate encounters where host prevailed. Solid circles indicate en- counters where intruder prevailed. Line A plots DISCUSS rON host/intruder of equal size. At all points below line B the intruder dominated the encounter. Intraspecific tolerance of close proximity appears to be the rule in echinoids, and it is believ~d that the example of E. lucunter area of residence (Abbott et aI., 1974) . constitutes the first clear demonstration of Earlier workers usually examined E. lucun- territorial defense and aggressive behavior in ter populations by day, when all animals are sea urchins. There is no mention of such in their burrows. Some movements within behavior in the reviews of Hyman (1955), elongate burrows over a period of weeks Moore (1966), Reese (1966), or Meadows have been recorded (Hunt, 1969). Kier and Campbell (1972), and the only sugges- and Grant (1965) noted that most animals tion of possible territorial behavior we have are not trapped within burrows by narrow encountered in other urchin species is men- entrances, and suggested the possibility of tioned by Forster (1959) where each Echi- noctural forays. McPherson (1969) made nus esculentus in his study site occupied a three separate night observations and found a browsed area of 0.25-0.5 m2• Forster that most E. lucunter remain in essentially notes that it is as yet unknown whether hom- their daytime positions; however, some ing occurs, or whether the browsed areas are emerge partially from their burrows, and a comparable to the territories of limpets. The few come out on the open rock surface un- low population densities found (one per der cover of darkness. Several contempo- 2 4.66 m ) suggest that encounters between rary studies (Abbott et aI., 1974) confirm individuals of E. esculentus are infrequent that most movement occurs within burrows, here. and in the area immediately adjoining the In contrast, Echinometra lucunter may burrow entrances. occur in population densities of up to 240 Dispersion in populations of settled adults individuals per m2 (Warner and Miller, is virtually absent. A 1 X 3 m2 area of 1974). In some areas of high population beachrock in the center of a very dense E. density the burrows are interconnected, yet lucunter population was cleared of urchins even here each urchin has its own defined in October 1973. In September 1974 the 186 BULLETIN OF MARINE SCIENCE, VOL. 28, NO. I, 1978 area was still almost completely bare of succeeds in defending its burrow success- urchins, and its boundaries were still clearly fully in most encounters unless the intruder marked by the urchin populations on all is considerably larger. The nature of the sides (Ogden et aL, 1975). However, under advantage conferred on the host by burrow crowded conditions, periodic contacts be- ownership is by no means clear. An animal tween E. lucunter individuals are inevitable. in its burrow is certainly better braced to Naturally occurring contacts and fights were push than one outside, but the cases where observed on a few occasions, and the fre- hosts evicted intruders after the host had quency with which spines occur in the gut been artificially removed from its burrow, contents (Abbott et aL, 1974) suggests en- an intruder inserted in the burrow, and the counters are not uncommon. host then replaced at the entrance of its for- Fighting is a means of spacing animals mer burrow, suggest that the phenomenon is within a given area (Manning, 1967); it not so simple. Perhaps, the host advantage permits individuals to defend resources es- may be explained by the familiarity of the sential to them. In the case of E. lucunter host with the structure of the burrow. the essential resources are shelter and food. It is likely that E. lucunter will also be These urchins live in wave-swept areas, and found to defend its burrow against intruders the sheltering offered by a burrow or crevice of other species. Preliminary work (Gri.in- is first and foremost important as a protec- baum, unpubL) showed eviction of Dia- tion against dislodgement. It seems likely dema antillarum Philippi and Tripneustes that the burrowing habit is also of importance ventricosus (L.) from the mouths of E. lu- in protecting the urchins from predators. cunter burrows. Randall (1967) recorded 26 species of West In E. lucunter, E. viridis, and the closely Indian fishes which prey, at least in part, on related E. mathaei, Mortensen (1943) has sea urchins. Seven of these species were shown that· each of the five auricles sur- found to contain remains of E. lucunter, or rounding the Aristotle's lantern in the test unidentified species of Echinometra. Glynn bears a well-developed flange or tag which ( 1968) reported that E. lucunter is also provides for attachment of the lantern re- preyed upon by a shore bird, the ruddy tractor muscles. Mortensen called attention turnstone (Arenaria interpres). The conch to the unusually large size of these muscles may prey on this urchin as in E. lucunter, and to the fact that their in- well (Hughes and Hughes, 1971). Exposed sertion grooves on the lantern pyramids are urchins are vulnerable, but urchins firmly rugose and larger than in most Echinometra anchored in their burrows, with spines species. He suggested that these features are braced against the walls or projecting out- probably not related to rock boring, as they ward, are surely more difficult objects for are lacking in such good rock-boring species a predator to dislodge. Finally, fighting may as E. vanbrunti. We agree, and suggest as also be of some advantage in insuring that an alternative that these features are related each urchin, in its burrow, has unobstructed to fighting. In E. lucunter the mobility of access to food. This includes not only the the lantern, jaws, and teeth we have ob- benthic plants growing in and around each served in animals engaged in fighting is burrow entrance but also the drifting macro- truly startling. The species E. viridis, a rock scopic plants which are very effectively cap- borer which is sometimes sympatric with E. tured by the urchins as they float past the lucunter, also displayed agonistic behavior mouths of burrows. in preliminary experiments (Gri.inbaum, un- While most encounters result in fights, the publ.). Preliminary investigation (Ogden, physical damage caused by the combatants unpubl.) of burrow defense in E. mathaei to each other is relatively minor. The host on the Kana Coast of Hawaii, has shown GRONBAUM ET AL.: BEHAVIOR]N THE SEA URCH]N ECHINOMETRA 187

that E. mathaei will evict conspecific intrud- mata: Echinoidea) at St. Croix, U.S.V.I. ers from its burrows in large lava blocks by West Indies Lab. Spec. Publ. 4: 87-95. Hughes, R. N., and H. P. I. Hughes. 1971. A pushing; however, no biting was observed. study of the gastropod Cassis tllberosa (L.) preying upon sea urchins. J. Exp. Mar. BioI. ACKNOWLEDGMENTS 7: 305-314. Hunt, M. 1969. A preliminary investigation of We thank I. A. Abbott and the Problems in Ma- the habits and habitat of the rock boring rine Biology Class (January, 1974) at the West urchin, Echinometra lucunter, near Devon- Indies Laboratory for their contributions to this shire Bay, Bermuda. Pages 35-40 in R. N. work. Thanks are due also to G. Grimm, P. Mal- Ginsberg, and P. Garrett, eds. The 1968 Sem- pass and K. Warner for their helpful advice in the inar on organism-sediment relationships. Ber- preparation of an earlier draft of the manuscript. muda BioI. Sta., Spec. Publ. No.2: 153 pp. W. B. Gladfelter read the final draft and made Hyman, L. H. 1955. The Invertebrates, Vol. 4. many helpful comments. This is contribution num- Echinodermata. McGraw Hill, N.Y. 763 pp. ber 6 of the West Indies Laboratory, Fairleigh Kaye, C. A. 1959. Shoreline features and Qua- Dickinson University, St. Croix, U.S. Virgin Is- ternary shoreline changes, Puerto Rico. U.S. lands. Geol. Survy Prafes. Paper 317-B. 49-140. U.S. Gov't Printing Office, Wash. D.C. Kier, P. H., and R. E. Grant. 1965. Echinoid LITERATURE C]TED distribution and habits, Key Largo Coral Reef Preserve, Florida. Smithsonian Misc. Coli. Abbott, D. P., J. C. Ogden, and I. A. Abbott, eds. 149: 68 pp. 1974. Studies on the activity pattern, be- Lewis, J. B. 1960. The fauna of the rocky havior and food of the echinoid Echinometra shores of Barbados, West Indies. Canad. 1. lucunter (L.) on beachrock and algal reefs at Zoo I. 38: 391-435. St. Croix, U.S.v.J. West Indies Lab. Spec. Manning, A. 1967. An introduction to animal Publ. 4: 111 pp. behavior. Addison-Wesley Publ. Co., Read- Adey, W. H. 1975. The algal ridges and coral ing, Mass. 208 pp. reefs of St. Croix: their structure and Holo- McPherson, B. F. 1969. Studies on the biology cene development. Atoll Res. Bull. 187: 67 of the tropical sea urchins Echinometra lucun- pp. ter and . Bull. Mar. Sci. Branham, J. M., S. A. Reed, J. H. Bailey, and J. Gulf Carib. 19: 195-213. Caperon. 1971. Coral-eating sea stars Meadows, P. S., and J. I. Campbell. 1972. Hab- Acanthaster plancii in Hawaii. Science 172: itat selection by aquatic invertebrates. Adv. ] ]55-1157. Mar. BioI. ]0: 27]-382. Clark, A. H. 1954. (other than Moore, H. B. 1966. Ecology of echinoids. holothurians) of the Gulf of Mexico. Pages Pages 73-85 in R. A. Boolootian, ed. Phys- 373-379 in P. S. Galtsoff, ed. Gulf of Mex- iology of Echinodermata. Interscience PubI., ico, its origin, waters and marine life. Fishery N.Y. 822 pp. Bull. U.S. Fish and Wildlife Service 55: 604 Mortensen, Th. 1943. A monograph of the pp. Echinoidea. III. 3 . II. Echin- Dix, T. G. 1969. Aggregating in the echinoid idae, Strongylocentrotidae, Parasaleniidae, Evechinus chloroticlls. Pac. Sci. 23: 123- . C. A. Reitzel, Copenhagen. 124. Ogden, J. C., D. P. Abbott, and I. A. Abbott. Forster, G. R. 1959. The ecology of Echinlls 1975. Behavioral and ecological studies on esculentus L. Quantitative distribution and the echinoid Echinomelra lucunler (L.) at rate of feeding. J. Mar. BioI. Ass. U.K. 38: St. Croix. Abstract, 11th meeting Assoc. Is- 361-367. land Mar. Labs. Carib. St. Croix.

Teytaud, A. R. 1971. Food habits of the goby, Boiler Bay, 81. Croix. West Indies Lab. Spec. Ginsburgellus novemlineatus, and the cling- Publ. 4: 27-34. fish, Arcos rubiginosus, associated with echi- Wobber, D. R. 1975. Agonism in asteroids. noids in the Virgin Islands. Carib. J. Sci. 11: BioI. Bull. 148: 483-496. 41-45. DATE ACCEPTED: January 12, 1977. Warner, G. F. 1971. On the ecology of a dense bed of the brittlestar Ophiothrix fragi/is. 1. ADDRESSES: (H.G., G.B. and J.C.O.) West In- Mar. BioI. Assn. U.K. 51: 267-282. dies Laboratory, Fairleigh Dickinson University, Warner, K., and R. Miller. 1974. Population Box 40/0, Christiansted, St. Croix, U.S. Virgin density and size class distribution of Echino- Islands 00820. (D.P.A.) Hopkins Marine Station, metra lucunter on selected algal ridges of Pacific Grove, California 93950.