Trapezia and T etralia (, Brachyura, Xanthidae) as Obligate

Ectoparasites of Pocilloporid and Acroporid Corals': 2

JENS W . KNUDSEN3

THE OCCURRENCE of marine invertebrates in was found once in Seriatopora, a finely branch­ the branches of living and dead corals has long ing member of the Pocilloporidae in which the been recognized. Two genera, Trapezia spatial relationships found in the Acroporidae and T etralia, of the family Xan thidae are deter­ obtain." Thus Garth suggests a basis of crab mined by Garth ( 1964) as being obligate com­ size and coral spatial relationships as a possible mensals of the coral families Pocilloporidae and basis of this "obligatory commensalism." Acroporid ae, respectively. Crane (1947) lists The writer spent four months at the Eniwetok species of the genus T rapezia as being found Marine Biological Laboratory, from February only in pocilloporid corals along the west coast through May 1965, in order to work on crab of tropical America. Miyake (1939), in listing ecology and to determine the possible basis for the Brachyura of Micronesia, records T rapezia this seemingly ironclad crab-coral relationship. cymodoce as collected from Stylophora, a pocil­ Several possible theories seemed worth inves­ loporid coral. Garth 's original collecting tech­ tigating in order to ascertain the factors upon niques used at Eniwetok Atoll, Marshall Islands, which this commensal-host relationship is main­ were refined in his later collecting in July 1959, tained: (1 ) the crab-size coral-space relation­ at which time he segregated each collection of ship suggested by Garth ; (2) that some ocean­ coral by species to avoid mixing coral commen­ ographic condition (water temperature, sals found therein. currents, wave action, etc.) is coincidentally From 10 collections of acroporid coral Garth required by both the crab genus and its respec­ extracted two species of T etralia (T. glaberrima, tive host coral, therefore making the relation­ also taken once on Pocillopora damicornis, and ship one of convenience; (3 ) that the crab T . heterodactyla) , but found no specimens of genera may prove to be filter-feeders utilizing T rapezia. Conversely, from 14 collections of the same food required by their host corals, thus pocilloporid corals he obtained five species of making the relationship one of simple con­ T rapezia (cymodoce, f . ferrttginea, digitalis, venience; (4) that the host corals provide some danai, rttfoptlnctata) , but obtained only one special form of protection, in addition to simple specimen of Tetralia. Because of this rather hiding places, which exclusively attract the crab exclusive distribution of the species of Trapezia genera ; (5) the possibility that more collections and Tetralia, Garth rightly concludes that these would reveal that both genera of would be are obligate commensals of long standing. Garth found almost equally on the Pocilloporidae and states ( 1964: 142), "In general, the larger Acroporidae. forms were found in the more robust Pocil­ These five suppositions served to initiate field loporidae, the smaller forms in the more delicate research. Th e first three (space-size, oceanog­ Acroporidae. Thus, the T rapezia species oc­ raphy, food requirements) could be the key to curred in the pocilloporid corals, the T etralia the coral-crab relationship either with live or species in the acroporid corals, although Tetralia dead corals of the proper families, providing such corals were not overgrown with algae. However, supposition number four (special 1 Sponsored by NSF Grant G-2412. protection) could function only with live corals. 2 Facilities and sponsorship at the Eniwetok Ma­ rine Biological Laboratory by the Atomic Energy Garth's term of obligate commensal was defined Commission. for our work as a situation where the crabs 3 Pacific Lutheran University, Tacoma, W ashington . in question are obligated to live with their host Manuscript received Mar ch 21, 1966. corals in order to receive some benefit which 51 52 PACIFIC SCIEN CE, Vol. XXI , January 1967

they could not otherwise obtain, but are not opposed to the white color of dead coral) is harmful to their coral hosts. required for pro tection. Crane (1947:83) and Garth (19 64: 142) To test the premise that coral color may be note that Trapezia and/or T etralia are limited an important factor (Exp eriment N o. 2) , new in number to a single mated pair (with their corals were killed, cleaned, washed, dyed with young, Garth adds) on smaller coral heads, vital stains of an appropriate color value so as establishing a territory which they somehow to match the live corals, stocked with newly protect (Garth). It must be noted, however, collected crabs of the proper genera, and placed that the crabs have normal pelagic larvae, the back in the Eniwetok Island pool. Controls megalops of which must seek out the proper were used as in Experiment 1. After three days species of host coral before settling and under­ the stained acroporid heads lacked the T etralia going metamorphosis. Thus, hundreds of corals crabs and the pocilloporid heads lacked the must be rejected by each megalops until the Trapezia specimens, though some galatheid proper species is found. Gurn ey (1938:76-77) crabs and a few shrimps had taken up residence. described the zoeal stages of Trapezia cymodo ce All but one of the control crabs were found on and T etralia glaberrima which he reared from their respective corals. adults in captivity, but made no note of the The results of Experiments 1 and 2 suggest larvae being photonegative. Thus, one may that size-space relationship (which is the same assume that normal photopositive larvae would in dead and live coral) is not an exclusive factor. be attracted away from the coral host of the Also, the oceanographic conditions of currents, adult crab, to take up a normal pelagic existence. wave action, water temperature, and dissolved Furthermore, of the 306 clutches of eggs that oxygen were constant in the environment of the we hatched out at the Eniwetok laboratory, 22 live and dead corals, and thus oceanographic belonged to species of Trapezia and Tetra lia. factors and the availability of water-borne food All of their larvae proved to be photopositive, were not exclusively responsible for the com­ thus indicating that they are pelagic in their mensal association. Any of these, or other, fac­ larval development. tors could be critical in the distribution of these To test the supposition that size-space, ocean­ crabs, but the requisite of live coral appeared ographic, or food relationships may serve as a to be important. key factor in commensalism, Experiment No. 1 To check this premise, selective collections was conducted in a deep pool on the ocean side of corals (Experiments 3-5) were made on the of the north end of Eniwetok Island. Pocillo­ outer reef at the north end of Eniwetok Island. porid and acroporid corals attached to large but Pocillopora danae heads were obtained just in­ movable rocks were collected and killed by air side the algal ridge where this coral abounds. drying. Some of these corals were dried for Six buckets of completely dead coral heads of three days while others were cleaned by rotting this species (Experiment 3, station 58) were the tissue, washing, soaking out, and then air collected from situations that were totally iso­ drying. N ext, these corals were tagged, restocked lated from living corals. Th ese corals yielded with newly collected crabs (Trapezia on pocil­ 16 species of crabs (to be described in another loporid coral and T etralia on acroporid coral, paper), but contained no specimens of Trapezia the normal hosts) , and placed in the bottom of or T etralia. the pool. Controls consisted of tagged live N ext, live heads of Pocillop ora danae which corals which were stocked with newly collected were partially dead and overgrown with algae, crabs of the proper species. were processed as follows: Each head was After three days it was observed that all of snapped loose from the reef flat with a geol­ the experimental crabs were missing from the ogist's hammer and instantly lifted from the dead corals (though some galatheid crabs had water and cleaved. Th e live portion with small taken up residence). All crabs on the control areas of dead coral (Experiment 4, station 59 corals were present (even when rechecked sev­ from which 14 species of crabs were obtained) eral weeks later) . Thi s suggested that either was placed in one bucket, the totally dead por­ live coral or coral with the natural color (as tion (Experiment 5, station 60 with 18 species Trapezia and Tetralia Bctoparasites-c-Kx uosnx 53 of crabs) was placed in another bucket. Th is specimens of Trapezia. These experiments indi­ process was completed swiftly so as to prevent cate that normal filtered food, meat baits, and crab movement from one part of the head to algae are not the normal diet of the Trapeziinae. another. Both the live and dead portions of N ext (Experiment 9), crabs were collected in coral included specimens of Trapezia. the field from their coral hosts, killed by cutting Thus, Experiments 3, 4, and 5 suggest that the carapace (to insure quick preservation), Trapezia will dwell in dead Pocilloporidae only and dropped into solutions of 5% formalin or if a live portion is present, again suggesting 75% alcohol. Under microscopic examination that live corals are essential for the commensal (430 diameters), without stain, no specimens relationship. Furthermore, totally live heads of of phytoplankton or zooplankton could be recog­ Pocilloporidae (Experiment 6, station 63) nized. Instead, small round globules of some yielded 17 species of crabs including 5 species sort were present in the stomachs. This material of Trapezia, as would be expected from Garth's was not identified or classified. account and from the control results of Experi­ Concurrently, a series of laboratory experi­ ments 1 and 2. ments (Experiments 10-13) were initiated to Acroporid and pocilloporid coral heads are gain better insight into the possible host-speci­ often found side by side in a given habitat, to­ ficity exhibited by the Trapeziinae, and to gether with their commensals, T etralia and determine the nature of their feeding habits. Trapezia, respectively. The very low incidence For Experiment 10, two aquaria measuring of mismatched crabs and corals recorded at the 10 by 24 inches, by 18 inches deep, were set time of our experimentation (one record of up and suppl ied with running seawater. Each T etralia on a pocilloporid coral, Garth, 1964 : tank was devoid of foreign material but was 142) prompted further field studies to deter­ provided with one live pocilloporid head with mine if these crab genera display a true specific­ three T etralia and one live acroporid head with ity for their coral families. Live acroporid and three Trapezia (these crabs were switched from pocilloporid corals were collected (Experiment their "preferred hosts") . Th e two coral heads 7), their crabs were exchanged (Tetralia placed in each aquarium were placed about 10 inches on Pocilloporidae and Trapezia on Acropori­ apart. After 24 hours the collective results from dae) , and then were placed back in the deep the two tanks showed that three T etralia had reef pool. After 3 days the crabs were absent, migrated to acroporid coral, one remained on suggesting that a true specificity does exist. a pocilloporid coral, one was dead and the last The feeding habits of the Trapeziinae were was missing; all six of the Trapezia had moved checked in the laboratory on captive to pocilloporid corals. This experiment demon­ used in reproductive studies. Attempts to serve strated a distinct prefe rence on the part of these as food either cut fish or algae proved unsuccess­ crabs to seek out their preferred host coral. The ful for maintaining breeding animals. Brine data are not conclusive, but may suggest that shrimp (Artemia) nauplii were offered to the this preference is slightl y stronger in the crabs with only apparent good results. Experi­ Trapezia genus. mentally, however, crabs were starved 3 days Experiment 11 utilized two identical running­ in clean aquaria (Experiment 8) and then of­ seawater aquaria devoid of all foreign material. fered tremendous numbers of live Artemis A small head of acroporid coral with five nauplii for filter feeding. Half of these crabs Trapezia was placed in one tank, while a pocil­ were fed at night , the other half during the loporid coral with five T etralia was placed in daytime. After a suitable feeding period had the second tank. N o other corals were available passed, the crabs were killed and the stomach to these crabs.After 24 hours five Trapezia contents examined. In no instance could Artemia remained on the acroporid coral (although one nauplii, or their fragments, or other plankton, was dead) ; four Tetralia remained on the pocil­ be recognized. On the other hand , Artemia eggs, loporid coral while a fifth crab (dead) was which happened to be introduced with the found a few inches away. Th ese results tend nauplii and which sank to the bottom of the to suggest that physical protection was sought aquaria, were found in the stomachs of two here in the "wrong" coral since the proper 54 PACIFIC SCIENCE, Vol. XXI, January 1967

coral was not available. Since the writer can specimens were starved for 3 days, then returned distinguish between most live acroporid and to live corals in small aquaria. When accus­ pocilloporid corals on the basis of odor alone, tomed to the aquaria many crabs began what he assumes here that chemical "odor, " or the turned out to be feeding activities. The follow­ absence of it, may cause these crabs to seek ing is an account of the typical behavior dis­ out their preferred host corals when they are played : present. The crab climbed into the coral branches, Experiment 12 repeated Experiment 10 but then placed the dactyli of the walking legs arranged the corals in a definite upstream­ (WL) 3 and 4 into polyp cups, depressing downstream relationship within the four run­ the polyps. Next, WL-l were inserted into ning-seawater aquaria used. Each of the four other polyp cups between the tentacles of the aquaria had one acroporid coral with 4 Trapezia polyp, and "scratched" back and forth at a (a total of 16 Trapezia), and one pocilloporid rate of about 4 strokes per second, for about coral with 4 Tetralia (a total of 16 Tetralia) . 4 seconds. Th e tips of WL-l were then alter­ Corals were placed about 10 inches apart (b e­ nately cleaned by the mouthparts of the crab. tween the nearest opposing borders). In aquaria During the cleaning operation, WL-2 were used A and B the acroporid corals were placed up­ to scratch new polyps, then were cleaned by stream to the pociIloporid corals. The converse the mouthparts. Mucoid material could be seen was true with aquaria C and D. clinging to the tips of the walking legs during The results of Experiment 12 A and B, after this procedure. 24 hours were as follows: 6 Trapeziamoved up­ Periodically material was transferred from stream to their preferred pociIloporid corals, WL-3 or 4 to WL-2 or 1, and then brought to while 2 Trapezia remained on or under the the mouth. Occasionally the chelipeds were acroporid coral ; 4 T etralia migrated downstream moved over the coral epidermis between the to their preferred acropid corals, while 4 were polyps. The fingers subsequently were cleaned missing altogether. in the mouthparts of the crab. These activities Th e results, of experiment 12 C and D, after were repeated over and over again, as the crab 24 hours were as follows: only 2 Trapezia slowly moved up through the coral branches. migrated downstream to the pocilloporid corals, while 6 remained on or under the acroporid Upon examining the dactyli of Trapezia f. coral; 4 Tetralia migrated upstream to the pre­ ferrttginea, it was noted that a special brush ferred acroporid coral, 2 remained with the and comb is present on the terminal segments of pocilloporid coral, and 2 were lost. each leg (referred to henceforth as the food The combined results of experiment 12 show brush and food comb) . The food brushes are that, of the 16 crabs that did migrate, 10 situated at the distal end of each dactylus (Fig. moved upstream to their preferred host while 6 IE) and consist of several short, stout, blunt­ migrated downstream. Again, these results are ended spines for agitating the coral polyp, and not conclusive but suggest that chemical odors a dense tuft of bristles for collecting mucus, may enhance the location of the preferred coral , and other debris. Th e bristle tuft is host. fully developed in walking leg 1 but is pro­ In Experiment 13, 3 T etralia specimens were gressively less well represented posteriorly (Fig. placed in each of four large runn ing-seawater 1 A-D) . Borradaile (1 903 :240) figures the aquaria which also contained some nylon mesh terminal spines of Trapezia f. ferrt/ginea and netting soaked with mucus from live acroporid suggests that "the remarkable ending of its legs coral. Only 4 of the 12 test animals located is in some way connected . . . " with its life the mucus-gauze "bait" after 24 hours. These in the coral. The terminus of each dactylus experiments were considered incomplete, how­ protrudes ventrally, thus forming a concavity ever, and will be continued in the future. on the ventral surface of the immediate proxi­ To test more fully the reactions of Trapezia mal part of the dactylus. The food combs, shown (Experiment 14, using T. f. ferrttgillea and T . in the posterior view of the left dactyli (Fig. f. areolata) in its host coral Pocilloporn, crab 1 A- D) , consists of from 3 to 6 rows of Trapezia and Tetralia Ectoparasites-KNUDSEN 55

F ~--~

E

M ~

I MM

FIG. 1. Trapezia f. [erruginea: A-D, Dactyli 1- 4; E, tip of dactylus 1; F, a bristle from food comb; G, 2nd maxilliped. Tetralia beterodactyla: H-K, Dactyli 1-4 ; L, 2nd maxilliped ; M , a bristle from maxiIIiped food brush. Anatomy: 1, food brush ; 2, food comb; 3, groove. (All drawings to the same scale except E.) 56 PACIFIC SCIENCE, Vol. XXI, January 1967

feathered bristles (Fig. IF) which extend the stomachs of crabs offered Artemia nauplii under each dactylus and proceed an equal dis­ for filter feeding . Presumably the eggs were tance up the anteroventral surface of each dac­ transferred from the aquarium floor to the crabs' tylus. The combs are well developed and are mouthparts after they had been picked up on presumably used in concentrating mucus from the food brushes. other legs, and in transferring mucus to other In a search of the literature for similar coral legs or to the mouthparts as described above. parasites, Gerlach's paper (1961 :3) describes The endopodites of maxillipeds 3 are usually "an as yet unidentified aberrant copepod with well bristled, and those of maxillipeds 2 have a worm-like body which apparently lives on a fan of spines and dense tufts of bristles (Fig. coral, mainly Pocilloporidae . . . . These ani­ 1 G) for combing food into the mouth (these mals could be observed as they crawled about are to be known as maxilliped food brushes) . on the surface of the coral and slashed at the Furth ermore, the longitudinally bifurcated basal tissues of the coral polyps with the sharp claws exopodite segment of maxillipeds 1 of Trapezia of the first pair of legs. Here the point to may have a valve function to prevent the loss be considered is that this is a form which has of food during food-transfer operations. become particularly adapted to a mode of life Th e dactyli of T etralia heterodactyla (Fig. 1 parasitic on coral." This behavior parallels the H-K) have relatively small food brushes which behavior of Trapeziinae described herein. are almost equally developed on all legs. The As to the degree of parasitism, that is, the ventral concavity is very shallow and contains effect of the crabs' parasitism upon pocilloporid two rows of food combs consisting of flat, corals, no data are available. Th e parasites are blunt, unfeathered bristles. These combs are probably quite effi cient, that is, they do not almost restricted to the ventral surface of the quickly kill or greatly harm their host. Other­ dactylus. A conspicuous groove is also present, wise, every case where numerous crabs are proximal to the end of each dactylus. These found occupying a coral head would result in grooves are better developed on WL -l and 4. the rapid destruction of the crabs' microhabitat. Th e maxilliped food brushes of Tetralia are The amount of food produ ced for crab con­ very well developed only on maxillipeds 2 sumption (or the number of polyps per head) (Fig. 1 L ) , and consist of dense masses of probably serves as a basis for territoriality ob­ bristles which are feathered (Fig. 1 M) dis­ served by Garth (1964:142). tally. The basal segment of the exopodite of The coral-host preference of the two genera maxillipeds 1 are also bifurcated as described of crabs may well be correlated with the rela­ for T rapezia. tive difference in size, and thus efficiency, of Live corals, in seawater, were examined under the food brushes and combs. This conclusion a dissecting microscope (Experiment 15), then is based on the fact that when live acroporid "scratched" as described for Trap ezia, with a corals are removed from seawater and placed dull probe. The probe was repeatedly coated in the shade they secrete vast quantities of with mucus and debris from the coral . mucus, while pocilloporid corals secrete little When examined at 430 diameters, this material mucus under the same condition. T etralia, with proved to be identical with that material on the the smaller and less effi cient brushes and combs, crabs' food brushes (Experiment 14) and with thus takes advantage of a coral family which material in the stomachs of newly-killed and presumably is capable of secreting the greatest also field-preserved specimens. amount of mucus. Tr apezia, on the other hand , These experiments demonstrate that T rapezia has larger brushes and combs but lives with a f. ferl'llgil1 ea is actually a parasite with a strong less "productive" coral family. More research host specificity, at least on the coral family is being done on this aspect. level. The presence of food brushes on other The exact basis of the host specificity dis­ Eniwetok Trapezia species and on T etralia played by these genera of crabs may well be species warrants recognizing them as parasites. related to the crab-size coral-space premise The use of the food brushes in feeding would suggested by Garth, or to the distinct difference also explain why A rtemia eggs were found in in the chemical odor and probable chemical T rapezia and T etralia Ectoparasites-KNUDSEN 57 makeup of the chief source of food, that is, Commission, for support and assistance ; to the mucus secreted by the coral animals. In our Dr. John S. Garth, who worked long hours to collections at Eniwetok Marine Laboratory we help review and identify Eniwetok crab speci­ isolated a large number of corals by species and mens ; and to my students, Jack Shannon and carefully separated the parasitic and commensal Dave Pearson, who diligently aided in our field crabs found therein. The identification of these research. animals may well reveal a more definite species­ to-species relationship between crabs and corals. Th ese data and others will be recorded in an­ REFERENCES other paper, along with additional experimenta­ BORRADAILE, L. A. 1903. Marine . tion to be conducted at the Eniwetok Marine III. The Xanthidae and some other crabs. Biological Laboratory. In : J. Stanley Gardiner, The Fauna and Geography of the Maldive and Laccadive CONCLUSIONS Archipelagoes 1(3) :237- 271. CRANE, 1947. Eastern Pacific Expeditions of The literature, our field collections, and field J. the New York Zoological Society. XXXVIII. experimentation confirmed that an obligate host Intertidal brachygnathous crabs from the west specificity exists between the crab genus T rapezia coast of tropical America, with special refer­ and Pocilloporidae corals, and between the crab ence to ecology. Zoologica N . Y. 32(2) :69­ genus Tetralia and Acroporidae corals. Further­ 96. more, these crabs require living corals as a GARTH , J. S. 1964. The Crustacea Decapoda source of food, in addition to protection, and (Brachyura and Anomura) of Eniwetok should be recognized as obligate ectoparasites Atoll, Marshall Islands, with special refer­ of their respective host corals. There may be a ence to the obligate commensals of branching relationship between the food brush and comb corals. Micronesica 1(1-2) :137- 144. size and the apparent ability of the two coral GERLACH, S. A. 1961. The Tropical Coral Reef families to secrete mucus which governs the host as a Biotope. Atoll Research Bull., Pt. 80, preferen ce displayed by Trapezia and T etralia. pp. 1-6. MIYAKE, S. 1939.N otes on Crustacea Brachy­ ACKNOW LEDGME NTS ura collected by Professor Teiso Esaki's The writer is indebted to the N ational Sci­ Micronesia Expeditions of 1937-1938, to­ ence Found ation and to Dr. Robert Hi att, gether with a check list of Micronesian Director of the Eniwetok Marine Biological Brachyura. Rec. Oceanogr. Work Japan Laboratory, sponsored by the Atomic Energy 10(2) :168-247.