J. Exp. Mar. Biol. Ecol., 1990, Vol. 140, pp. 197-207
Aggression among sea urchins on Caribbean coral reefs M.J. Shulman Smithsonian Tropical Research Institute, Balboa, Republic of Panama
(Received 25 May 1988; revision received 29 September 1989; accepted 2 April 1990) Abstract: The existence and nature of intra- and interspecific aggression were examined for five species of sea urchins inhabiting Caribbean coral reefs. The studies took place in the San Blas Islands of Panama and involved the following species: Echinometra lucunter (Linnaeus), E. viridis Agassix, Diadema antillarum Phillipi, Lytechinus williamsiChesher, and Eucidaris tribuloides (Lamarck). An intruder was placed next to an undisturbed resident and the behaviors and responses of both individuals were followed. All pairwise combinations of resident species and intruder species were tested except for combinations involving D. antillarum as intruder. For E. viridis and E. lucunter, agonistic interactions occurred commonly (46-79x of trials) between conspecifics and congeners. Almost all of the agonistic encounters involved pushing. Additionally, biting occurred in 8-25 % of the trials. Residents were most ofken the aggressors and usually succeeded in retaining their location. Intruders only succeeded in forcing residents out of their positions if the intruder was equal to or larger in size than the resident. Surveys of an undisturbed population of E. viridzkduring the daytime indicated that 16% of the individuals were engaged in intraspecific agonistic interactions at any one time. D. antillarum exhibited biting behavior against both species ofEchinometra in 23-24% of the trials. Biting attacks against L. williamsiand E. tribuloides occurred rarely. L. williamsionly once demonstrated pushing behavior and never was observed biting another sea urchin. E. tribuloidesoccasionally exhibited pushing and biting behaviors, both as residents and as intruders, and was twice observed biting Echinometra. These studies suggest that two kinds of agonistic interactions may commonly occur among Caribbean reef-dwelling sea urchins: (1) intraspecific and interspecific aggression among Echinometra, and (2) predatory/aggressive attacks against Echinometra by D. antillarum. The former may result in greater dispersion ofEchinometra relative to food and shelter resources and control the spatial distribution and concentration of Echinometra grazing pressure within an area. The attacks by D. antillarum may result in the restriction to, or higher densities of, Echinometra in crevices and rugose microhabitats that provide shelter from the larger-bodied D. antillarum. Key words: Echinoid; Predation; Territoriality; Tropics
Resource defense occurs commonly among organisms (Davies & Houston, 1984). A variety of resources are known to be defended: food, water, shelter, and mates (Kaufmann, 1983), and a multitude of resource defense mechanisms have been described. These mechanisms include aggression, preying on competitors (Park et al., 1965), and rendering the resource undesirable (Harper, 1977), inaccessible, or Correspondence address: M. J. Shuhnan, Department of Biology, University of California at Los Angeles, Los Angeles, CA 90024, USA. 0022-0981/90/$03.50 0 1990 Elsevier Science Publishers B.V. (Biomedical Division)
& Krebs, 1978). Defense of resources
effects at the population access ofconspecifics and reproduction Similarly,
may affect population
of some or all members
defense of resources
level. Behaviors density, dispersion
of the population
against heternspecific acquisition
that limit the resource and the growth, survi\ al
(Davies & Houston,
may affect distribu-
of other species (c.g. Robertson
et al., 1976). Resource defense mechanisms have been described in a number of groups of marine organisms. Space limited species use allelochemicals, behavioral aggression, overgrowth, and undercutting to acquire or defend space (Jackson, 1977). Mobile animals, such as fishes, limpets, and octopuses, generally use agonistic behaviors to defend food, shelter and/or mates (Stimpson, 1970; Low, 1971; Branch, 1981). In this study, the existence and nature of aggression and resource defense are examined for five species of sea urchins inhabiting Caribbean coral reefs. Intraspecific agonistic behavior has previously been described in three species of sea urchins: Echinometra lucunter et al., 1978), E. mathaei (Tsuchiya & Nishihira, 1985), and Strongylo& Roe, 1983) and attributed to defense of a vital shelter resource. Additionally, predatory attacks by the temperate sea urchin Lytechinus unamesus against S. purpuratus and S. franciscanus have been reported (Coyer et al., 1987). Here I examine both intra- and interspecific aggressive interactions among five sea urchin species and discuss their potential effects on foraging behaviors, resources, and the distribution of competing species. (Grunbaum
centrotus purpurcrtus (Maier
STUDY AREA AND ORGANISMS
in the San Blas
Islands of Panama from March to September 1984 and from May to September 1985. The five species of sea urchins studied inhabit reefs throughout the Caribbean (Colin, 1978). My experiments were conducted on several patch reefs in the vicinity of San Blas Point (9’ 33’ N, 78” 57’ W). These reefs extend from depths of 0.5-7 m and are partially covered by colonies of Agaricia agaricites (Linnaeus), A. tenutfilia Dana, Millepora complanatu Lamarck,
Porites furcata Lamarck, and P. astreiodes Lesueur. The five species of sea urchins studied were Echinometra lucunter (Linnaeus), E. viridis Agassiz, Diadema antiliarum Phillipi, Lytechinus williamsi Chesher, and Eucidaris tribuloides (Lamarck). E. Zucunter inhabited shallow areas of these reefs, generally at depths of < 1 m. E. viridis co-occurred with E. lucunter at those shallow depths but was also very common (up to 60 . m - ‘) along the entire depth gradient. The other three species were found deeper than 0.5 m.
E. lucunter and E. viridis are small sea urchins (mean test diameter of individuals in the study was 2 1 mm for E. lucunter and 20 mm for E. viridis). They have medium length (2 = 13 mm), stout spines. D. antillarum is a large sea urchin (X = 47 mm test diameter for individuals used in the study) with long (X = 77 mm), hollow spines. L. wilhizmsi is
AMONG SEA URCHINS
a small sea urchin (Sz= 15 mm test diameter), with very short (X = 4 mm), relatively thin spines. Eucidaris tribuloides is a small-sized sea urchin (S;:= 19 mm test diameter) with medium length (X = 23 mm), extremely stout, solid, blunt-ended spines.
The following experiments were performed to determine whether the various sea urchins were capable of intra- and interspecific aggression. A sea urchin, denoted the “intruder”, of the appropriate species was chosen haphazardly from those available on the reef, and its maximum test diameter and spine length were measured. A “resident” sea urchin of the appropriate species was haphazardly chosen from those a minimum distance of 1 m from the original position of the intruder. The intruder was then placed next to the resident, such that a distance of z 1 cm separated the tips of their spines. The subsequent movements and behaviors of both sea urchins were then recorded until the interaction came to an end. The end of the interaction was arbitrarily defined as having taken place if neither sea urchin had moved within the previous 10 min or the individuals became separated by a distance of > 15 cm. At this point, the test diameter and spine length of the resident were measured. With one exception (described below), all possible pairwise combinations of resident species and intruder species were tested. 20-40 trials were made for each combination of resident species and intruder species, except for the E. viridis/E. viridis combination, of which 158 trials were made. No individual sea urchin was used in more than one trial. D. antillarum, when detached from the substratum, would begin rapidly moving all of its spines. When placed back on the reef, it would not remain in one spot but moved away very quickly. Because of this behavior, it was not possible to test the responses of any species to D. antillantm intruders. Further observations were made to determine the natural frequency of aggressive encounters among E. viridis. On patch reef “Porvenir 26” (Robertson, 1987, Fig. l), the densities and number of E. viridis involved in aggressive interactions were counted in 20 haphazardly placed 1-m’ quadrats. All of the quadrats were censused during daylight hours. A sea urchin was scored as being involved in an agonistic interaction if: (a) it was pushing or being pushed by another E. viridis and/or (b) it was biting or being bitten by another E. viridis.
The manner and frequency in which aggressive behaviors were exhibited differed between species. These behaviors and their associated frequencies are described below for each species studied.
M. J. SHULMAN
Percentage of experimental interactions between sea urchins that resulted in pushing and/or biting by either resident or intruder. Number of trials is shown in parentheses. Species codes: El, E. lucunfer; Ev, E. viridis; D. D. untillawn; Lw, L. williamsi; Et, E. rribuloides.
50 (20) 19 (28) 5 (20) 25 (20)
El Ev Lw Et
24 (38) 28 (40) 5 (20) 10 (20)
65 (20) 46 (158) 10 (40) 40 (20)
10 (20) 0 (20) 5 (20) 5 (20)
5 (20) 15 (20) 0 (20) 5 (20)
AND E. VIRIDIS
congeneric intruders, agonistic interactions occurred in 46-79% of the trials (Table I). In a large majority of the cases, the resident was the aggressor (Table II). Residents generally responded to intruders by moving towards them, contacting and intermeshing spines and then pushing the intruder
away by moving in its direction.
Pushing went on
TABLE II Percentage of experimental interactions between sea urchins that resulted in either biting or pushing between resident and intruder. Numbers shown are: % interactions in which resident was aggressive/% interactions in which intruder was aggressive. Species codes and number of trials are same as those shown in Table I. Intruder
Resident El Push
El Ev Lw Et
40/5 64111 5/O 5/10
Et _____~ Push Bite
6015 35/l 5/O 3515
lo/l5 8/O 8/O 5/20
3/O 313 O/O o/5
24/O 23/O 5/O 5/O
o/10 O/O 5/O O/5
O/O O/O O/O O/O
O/5 10/5 O/O 5/o
11/l O/O O/5
O/O o/10 O/O o/o
for < 1 to 150 min (X = 11.5 min) and within that time the pushed sea urchin would move 0.5-10 cm (X = 2.9 cm) (Table III). The sizes of E. viridis residents and intruders in trials that resulted in pushing are shown in Fig. 1; the winner of the encounter is indicated by the symbol used. (Winning is defined as retaining or acquiring occupancy of the resident’s original location.) Intruders only won encounters when the resident : intruder size ratio was < 1.14 (shown as the dashed line in Fig. 1). Residents won encounters at all size ratios.
AMONG SEA URCHINS TABLE III
Mean distance pushed (cm) for those experimental interactions that resulted in one sea urchin pushing other. Number of trials involving pushing is shown in parentheses. Species codes are same as in Table I. Resident
El Ev Lw Et
2.4 (9) 2.5 (22) 3.0 (1) 1.3 (4)
2.0 (13) 3.3 (66) 1.7 (3) 1.2 (8)
6.0 (1) 6.0 (2) - (0) 8.0 (1)
1.5 (2) - (0) 5.0 (1) 2.0 (1)
Et 2.0 2.0 0.5
(1) (3) (0) (1)
TEST DIAMETER (mm)
Fig. 1. Sizes of E. viridis residents and intruders for those experimental interactions which resulted in at least one of individuals pushing other. Winning encounter was defined as retaining or acquiring occupancy of resident’s original location. Winner of each interaction is indicated by symbol: closed circle, resident wins; open triangle, intruder wins; open circle, no winner (neither contestant occupies resident’s original position). Intruders only won encounters when resident: intruder size ratio was < 1.14 (shown as dashed line). Residents won encounters at all size ratios.
only occurred in 8-25 % of EchinometralEchinometra encounters (Table II), and almost always (79% of cases) after an initial bout of pushing had taken place. However, only 21% of pushing encounters were following by biting behavior. During the biting behavior, the attacking Echinometra would rotate onto its equatorial side, with the oral surface facing the other sea urchin, and begin biting the spines of the opponent. This biting behavior continued for l-38 min (X = 10 min) and was frequently accomBiting
minutes of the commencement of the biting behavior (67 “; of time). Examination urchins that had been bitten showed that spine tips had been broken off
Echinometru (Student’s (Student’s
t test: P < 0.01); this was true for both residents
r tests: P < 0.05). The smallest
was 2.0 cm in diameter
the smallest biting intruder was 2.3 cm. Residents only bit in trials in which they were equal to or larger in size than the intruder; intruders only bit when they were larger than the resident. The nature of the intra- and interspecific aggressive behavior was similar in the two species of Echinometrcr. The frequencies of these interactions, however, differed. The frequency of interactions which resulted in pushing and/or biting was higher in E. lucunter/E. viridis trials than it was in conspecific trials (x2 test: P < 0.01). Biting occurred more frequently in encounters between individuals ofthe different species than they did between conspecifics, though these differences were not significant (x2 test; 0.10 > P > 0.05). In trials with combinations of E. lucunter and E. viridis, intruders were the biters in four out of the eight trials in which biting occurred, a significantly different pattern than that seen in conspecific trials, in which only residents were ever observed biting (Fisher’s exact test; P < 0.025). Additionally, in conspecific biting interactions. the biter was never forced to retreat, while in heterospecilic interactions, the biter retreated in four out of eight cases (a significant difference: Fisher’s exact test; P < 0.025). Both species of Echinometru were more aggressive towards conspecifics, congenerics. and E. tribuloides than they were towards L. williamsi (Table II). However, both species of Echinometru residents were occasionally observed pushing L. williamsi (Table II). In a few trials, resident E. viridis bit L. williamsi and E. tribuloides (Table 11). Determination of the frequency of naturally occurring interactions among E. viridis showed that O-27’:;, (S = 15.6”/,) of the individuals in the quadrats on “Porvenir 26” were involved
during daylight hours.
D. antillantm rarely pushed
and were only rarely pushed
intruder (Table II). Biting by resident D. antillarum, however, was observed in z l/4 of the interactions with intruding Echinomefru and occasionally in interactions with intruding L. williumsi and E. tribuloides (Table II). In these attacks, the D. antillarum approached and placed some oral spines over the intruder and then simultaneously drew the intruder under its body and climbed on top. This was followed by biting the intruder’s spines. The biting pressure was quite intense: the breaking of the spine tips was audible and accompanied by a sudden jerk of the intruder’s test. In attacks on Echinometra the biting went on for l-20 min (X = 11 min). It almost always ended with the intruder escaping from the D. antillarum. However, in one instance, an E. viridis did
not escape from the D. antillarum; an area of spines was completely bitten off down to the test, and a patch of test chewed through. This interaction only ended after the investigator removed the E. viridis after 20 min of biting by the D. antillarum. L . WILLIAMSI
L. williamsi rarely had aggressive encounters with con- or heterospecific sea urchins (Table I). In only one out of 20 trials did a resident L. williamsi push a conspecilic intruder; intruding L. williamsi never pushed conspecific residents and only rarely pushed heterospecilic residents (Table II). L. williamsi were never observed biting other sea urchins (Table II). As intruders, L. williamsi received less aggression from other sea urchin species than did the other species studied (Table II). E. TRIBULOIDES
E. tribuloides occasionally exhibited pushing behavior, both as residents and as intruders. This behavior was shown in 5-10% of interactions with conspecilics and heterospecilics (Table II). While pushing, E. tribuloideswould use their spines to provide the leverage. This was unlike Echinometra, which would depend upon tube feet and only occasionally on the purchase provided by spines against substratum. E. tribuloides was twice observed biting Echinometra: a resident bit an E. viridis and an intruder bit an E. lucunter (Table II). E. tribuloides was pushed in 10-357, of trials with resident Echinometra (Table II) and occasionally bit by E. viridis(20% of trials) and D. antillarum (5% of trials) (Table II).
In these experiments, all live species had at least one individual that, as a resident, responded aggressively to an intruder. However, only three species: E. lucunter, E. viridis, and D. antillarum, commonly exhibited aggressive behavior towards one or more species. The behaviors of Echinometra spp. and D. antillarum were very ditferent. The most common aggressive response in Echinometra was pushing of an opponent; biting occurred more rarely (6-3 1% of encounters) and was usually exhibited only by large individuals. Echinometra aggression was directed most frequently towards species similar in size and behavior, that is, conspecilics, congeners, and E. tribuloides. Intraspecific aggressive behaviors among E. viridis appear to be common under natural field conditions. The censuses performed here suggest that, on average, E. viridis may be spending 16% of daytime in aggressive interactions. In contrast to the behavior of Echinometra, aggressive behavior in D. antillarum consisted almost entirely of biting; almost all of the victims were Echinometra. How D. antillarum, particularly solitary individuals inhabiting crevices, might respond to conspecific intruders is not known. L. williamsi was nonaggressive: only one individual was ever observed pushing
and no biting was ever observed.
E. trihuloides exhibited
slightly higher levels
of aggressive behaviors, both as residents and intruders. These behaviors, which included both pushing and biting, were most frequently directed towards Ec~hinometra. What might be the functions suggested that the intraspecific
of these aggressive behaviors? agonistic
of E. lucunter functioned
et al. (197X) to defend
burrows; Maier & Roe (1983) offered the same hypothesis to explain aggressive behavior among the temperate zone sea urchin S. purpuratus. For both of these species the populations under study inhabited burrows in wave-exposed areas: alga1 ridges in the case of E. lucunter and an intertidal pool in the case of S. purpuratus. Presumably the burrows provided protection from dislodgement by the surge. The populations of Echinometra studied here occupied a calmer habitat, and the sea urchins sheltered in much less protected crevices rather than in burrows. Aggressive behavior was seen in Echinometra residents that were out in the open, as well as in those occupying holes. Similarly, natural aggressive interactions frequently took place some distance from shelter sites. Presumably those sea urchins in the open were not defending a shelter site, but may have been defending a fixed or moving feeding area. The biting behavior of D. antillarum towards Echinometra may represent attempts (sometimes successful) at predation. Quinn (1965) reported instances of starved D. antillarum in aquaria feeding on the sea urchin Tripneustes esculentus and the sand dollar Clypeaster rosaceus. Field observations of predation by one sea urchin species on others have been reported previously. Coyer et al. (1987) observed attacks by Lytechinus anamesus against S. purpuratus and S. franciscanus. These attacks only occurred during seasons of shortage in available behavior of D. antillarum against
algae, which were the preferred food. The biting Echinometra may, alternatively or additionally, be
defense offeeding areas against a competing herbivore. When attacked by 1). antillarum. the Echinometra did flee open areas and eventually took refuge in crevices that were too small for the larger D. antillarum to enter. It is possible that D. antilfarum, with its brittle spines, is unable to use pushing as a method of defending a feeding area. Are there rules that govern the pushing and biting behavior in Echinometra, determining when these behaviors will be exhibited and which contestant will win the encounter? Pushing and biting were much more likely to be exhibited by large than small Echinometra. The size of the smallest E. viridis observed pushing was 14 mm and biting was 20 mm (the smallest individual in the trials was 10 mm). Residents only bit intruders when they were equal or larger in size than their opponents. Intruders only bit when they were larger than the residents. These results suggest that only individuals that are larger than their opponents are in a situation where they can or will bite. There could be physical restrictions against biting a larger opponent or, possibly, due to the risks involved in biting a larger opponent, selection may have only favored biting when the opponent was of similar or smaller size. In the experiments described here, E. viridis residents were able to win aggressive encounters against both smaller and larger intruders. Intruders were only able to win encounters if they were very similar in size or larger than the resident. The latter result
AMONG SEA URCHINS
differs to some degree from that found by Grunbaum et al. (1978) in their study on intraspecific aggression among E. lucunter. They found that only when the intruder was > 39% larger than the resident would the intruder win the fight. The difference between the two studies may be due to the differences in habitat. In the St. Croix study, the cost of being evicted from a burrow in the exposed algal ridges was presumably very high. In the less exposed patch reefs in the San Blas Islands the cost is probably lower. Therefore, even though larger intruders may have been capable of evicting smaller residents, the benefits of doing so may not have been very high, and not worth the costs of fighting. Additionally, in both studies, the intruder had been disturbed by the experimenters while the resident had not; this difference may have resulted in a decreased likelihood of intruders engaging in, or winning, an aggressive interaction. However, studies on a number of other territorial animals have revealed that residents frequently win territorial battles, in which, on purely size or strength grounds, they would be expected to lose. This ‘resident advantage’ has been hypothesized to be due either to the higher value of familiar resources relative to unfamiliar resources or a greater resource holding potential for the resident (Riechert, 1978; Hammerstein, 198 1; Parker, 1984). Do aggressive species of sea urchins show special adaptations for the behaviors involved? Grunbaum et al. (1978) suggested that unusual characteristics of the Aristotle’s lantern in E. lucunter are adaptations for biting during agonistic interactions. These features, which were described by Mortensen (1943), include well-developed tags or flanges on the auricles surrounding the lantern and broad, rugose lateral furrows on the lantern pyramids. Both the auricle tags and the pyramid furrows provide attachment points for the unusually large lantern muscles. However, among the Echinometru, aggressive biting is not unique to E. lucunter; both E. viridis (this report) and E. mathaei (Tsuchiya & Nishihira, 1985) exhibit this behavior. In E. viridis and E. mathaei, the lantern muscles are smaller, the auricle tag is present but less developed, and, in E. mathaei, the lateral furrow on the pyramids is narrower and smoother than is seen in E. lucunter (Mortensen, 1943). The unique morphological features seen in the lantern apparatus of E. lucunter, therefore, do not appear to be associated with a unique aggressive biting behavior. Additionally, aggressive biting has been seen in species of sea urchins with very different mouth morphologies. Both D. antillarum and E. tribuloides were observed biting opponents in the experimental encounters. Neither of these species share the specialized morphological characteristics of the Aristotle’s lantern which some Echinometra exhibit. E. tribuloides, in fact, as is characteristic of the primitive cidarids (Smith, 1984), has only extremely limited lateral motion of the Aristotle’s lantern. Aggressive biting, therefore, is possible without the specialized mouth morphologies seen in E. lucunter, and, to a lesser extent, E. viridisand E. mathaei. What are the possible ecological consequences of the observed intra- and interspecific aggression? For Echinometra, a likely result is a more even spacing of individuals relative to shelter/food resources. Additionally, small individuals, which are aggressively inferior, may get excluded from higher quality habitat (including both feeding areas and
shelters). The defense ofa feeding resource also suggests that Echinometrcr forage within a very circumscribed area; data on movements of both Echinometru species indicate that daily movement is generally very restricted (Shulman, in prep.). If Echinomrtm are more evenly
due to intraspecific
for the ecology of the sessile assemblage.
whose daytime aggregativc behavior
Unlike I). ~/nti/larzzm.
is likely to produce a patchy distribution
pressure, the effects of Echinometrrr grazing would be more evenly distributed. The interspecific aggression between E. lucunter and E. viridis presumably reflects the great ecological similarity between the two species. In habitats in which they overlap, they will presumably be using, and possibly competing for, the same food and shelter resources (McPherson, 1969; Shulman, in prep.). The data do not provide any evidence for the aggressive dominance of one species over another. Instead, dominance is determined by the relative sizes and possibly the resident status of the opponents. The aggression/predation of D. untillurum against the two species of Echinometru may have a number of ecological effects. The three species are all herbivores and it is possible that D. untillurum may be able to exclude Echinometm from more open, less rugose portions of the substratum and restrict them to areas having smaller crevices within which the Echinometrrr can take refuge. In fact, E. viridis is more common on rugose substratum (Shulman, unpubl. data); however, this may be due to a requirement for crevices that will provide protection from other predators, such as fishes (Randall, 1967 ; Parker & Shulman, 1986) rather than refuges from D. antillarum. Williams (1980. 1981) studied competition among D. untillurum. E. viridis and the territorial damselfish Stegastesplanijwzs ( = Ezlpomacentrz~splan~~kms). She suggested that exploitation competition existed between the two sea urchins and that D. antillurum, because of its larger size and resulting greater consumption of food, was the superior competitor. Additionally. Williams suggested that, within the Acropora cervicornis habitat, the greater aggression of S. p1uniJLon.stowards D. antillurum permitted coexistence of the competitively inferior E. 1iridi.v. In contrast, these behavioral studies suggest that interference competition
may occur between these two sea urchins,
but that E. viridis has refuge in
crevices that D. antillurum cannot enter and in rugose microhabitats because of its large size, cannot forage in effectively.
that 1). ~mtillwwm.
I thank D. Pilson, D. Parker and H. Hess for assistance in the field and M. Grober, H. Lessios, and J. Morin for comments on the manuscript. I am grateful to the Kuna Indians and the Republic of Panama for granting permission for field work in the San Blas Islands and the Smithsonian Tropical Research Institute for providing facilities and logistical support. This work was funded by a grant from the Scholarly Studies Program of the Smithsonian Institution.
AMONG SEA URCHINS
REFERENCES Andersson, M. &J. Krebs, 1978. On the evolution ofhoarding behavior. Anim. Be&., Vol. 26, pp. 707-711. Branch, G.M., 1981. The biology of limpets: physical factors, energy flow, and ecological interactions. Oceanogr. Mar. Biol. Annu. Rev., Vol. 19, pp. 235-380. Colin, P. L., 1978. Caribbean reef invertebrates and plants. T.F.H. Publications, Neptune City, 512 pp. Coyer, J.A., J.M. Engle, R. F. Ambrose & B.V. Nelson, 1987. Utilization of purple and red sea urchins (Strongylocentrotus purpuratus Stimpson and S.jianciscanus Agassiz) as food by the white sea urchin (Lyrechinm anamesus Clark) in the field and laboratory. J. Exp. Mar. Biol. Ecol., Vol. 105, pp. 21-38. Davies, N. B. & A. I. Houston, 1984. Territory economics. In, Behavioral ecology, edited by J. R. Krebs & N. B. Davies, Sinauer Associates, Sunderland, pp. 148-169. Grunbaum, H., G. Bergman, D.P. Abbott & J.C. Ogden, 1978. Intraspecific agonistic behavior in the rock-boring sea urchin Echinometra lucunter (L.) (Echinodermata : Echinoidea). Bull. Mar. Sci., Vol. 28, pp. 181-188. Hammerstein, P., 1981. The role of asymmetries in animal contests. Anim. Behav., Vol. 29, pp. 193-205. Harper, J. L., 1977. The population biology ofplants. Academic Press, London, 892 pp. Jackson, J.B.C., 1977. Competition on marine hard substrata: the adaptive significance of solitary and colonial strategies. Am. Nat., Vol. 111, pp. 743-767. Kaufmann, J. H., 1983. On the definitions and functions of dominance and territoriality. Biol. Rev., Vol. 58, pp. l-20. Low, R.M., 1971. Interspecific territoriality in a pomacentrid reef fish, Pomacentrusflavicauda Whitley. Ecology, Vol. 52, pp. 648-654. Maier, D. & P. Roe, 1983. Preliminary investigations of burrow defense and intraspecific aggression in the sea urchin, Strongylocentrotus purpuratus. Pac. Sci., Vol. 37, pp. 145-149. McPherson, B.F., 1969. Studies on the biology of the tropical sea urchins, Echinometra lucunter and Echinometra viridis. Bull. Mar. Sci., Vol. 19, pp. 194-213. Mortensen, T., 1943.A monograph of the Echinoidea. III. 3 Camarodonta. II. Echinidae. Strongvlocentrotidae. Parasaleniidae, Echinometridae. C.A. Reitzel, Copenhagen, 446 pp. Park, T., D. B. Mertx, W. Grodzinski & T. Prus, 1965. Cannibalistic predation in populations offlour beetles. Physiol. Zool., Vol. 38, pp. 289-321. Parker, G.A., 1984. Evolutionarily stable strategies. In, Behavioral ecology, edited by J.R. Krebs & N. B. Davies, Sinauer Associates, Sunderland, pp. 30-61. Parker, D.A. & M. J. Shulman, 1986. Avoiding predation: alarm responses of Caribbean sea urchins to simulated predation on conspecific and heterospecific sea urchins. Mar. BioL, Vol. 93, pp. 201-208. Quinn, B.G., 1965. Predation in sea urchins. Bull. Mar. Sci., Vol. 15, pp. 259-264. Randall, J. E., 1967. Food habits ofreeffishes of the West Indies. Stud. Trop. Oceanogr., Vol. 5, pp. 665-847. Reichert, S.E., 1978. Games spiders play: behavioral variability in territorial disputes. Behav. Ecol. Sociobiol.. Vol. 3, pp. 135-162. Robertson, D.R., 1987. Responses of two coral reef toadfishes (Batrachoididae) to the demise of their primary prey, the sea urchin Diadema antillarum. Copeia, Vol. 1987, pp. 637-642. Robertson, D. R., H.P.A. Sweatman, E.A. Fletcher & M.G. Cleland, 1976. Schooling as a mechanism for circumventing the territoriality of competitors. Ecology, Vol. 57, pp. 1208-1220. Smith, A., 1984. Echinoid palueobiology. George Allen & Unwin, London, 190 pp. Stimpson, J., 1970. Territorial behavior of the owl limpet, Lottia giguntea. Ecology, Vol. 51, pp. 113-118. Tsuchiya, A. & B. Nishihira, 1985. Agonistic behavior and its effect on the dispersion pattern in two types of the sea urchin, Echinometra mathaei (Blainville). Galaxea, Vol. 4, pp. 37-48. Williams, A.H., 1980. The threespot damselfish: a noncarnivorous keystone species. Am. Nat., Vol. 116, pp. 138-142. Williams, A. H., 1981. An analysis of competitive interactions in a patchy back-reef environment. Ecology, Vol. 62, pp. 1107-l 120.