A Review of the Labrid Fish Genus Labroides, with Descriptions of Two New Species and Notes on Ecology!

JOHN E. RANDALL

AN INDO-PACIFIC genus of smalllabrid fishes of nine dorsal spines, and in Riippell's mod­ (wrasses), Labroides Bleeker, is here restricted erately detailed description there is no men­ to four species, two of which are described as tion of a bilobed condition of the lower lip. new from islands of the tropical central Pa­ Fowler (1928: 331) (after Schmeltz) also cific. Underwater observations of all of the placed quadri/ineatus in Labroides and listed it species have revealed the unusual food habit from Samoa . I am dubious of this record . of removal of ectoparasites from other fishes. Possibly Schmeltz obtained the young of This is discussed for each species in the ac­ Labrichthys cyanotaenia Bleeker which have counts following taxonomic considerations. two lengthwise pale bands on the body (al­ though these are lower on the body of cyano­ Gunther (1862: 120) and Fowler (1949: taenia than they are on quadrilineatas and the 117) considered Diproctacanthus xanthums more ventral band is not as distinct). La­ Bleeker in the genus Labroides. I am in brichthys unilineata (Guichenot) from Guam agreement with de Beaufort (1940: 19, 151, is probably a synonym of Labrichthys cyano­ fig. 26) that the monotypic Diproctacantbus taenia, based on this juvenile color pattern. Bleeker is a valid genus. D. xanthums lacks Saville-Kent (1893: 308, col. pl. 16, figs. 4, the characteristic bilobed lower lip of La­ 9) described two species of Labroides, L. broidesand has two instead ofthree anal spines. bicincta and L. auropinna, after seeing them in Smith (1957: 100, 104) established a new coral pools of Lady Elliot Island reef in the genus, Fouilerella, for Labroides bicolor, princi­ Great Barrier Reef of Australia. He obtained pally on the basis of fewer lateral line scales no specimens, but preferred to "provision­ than dimidiatus. In view of the similarity of ally" give the fishes new names . Although bicolor to dimidiatus in other respects than the color drawings on which these names are number of scales, I refer Pow/ere/fa to the based are crude, it seems very likely that the synonymy of Labroides blue and black L. bicincta is a juvenile L. Smith included Labrus quadrilineatus Rup­ dimidiatm. Saville-Kent mentioned its "re­ pell (1835: 6, pl. 2, fig. 1) from the Red Sea semblance to the white and black banded L. with dimidiatus in the genus Labroides even dimidiatus, C. and V." Obviously he was not though its scale counts would seem to ally aware that the true life color of L. dimidiatus it with bicolor. Actually quadrilineatus prob­ is blue and black . The figure of L. auropinna, ably does not belong in Labroides, for this a blue fish with yellow fins, fits no known species has a completely scaled head (whereas species of Labroides. Until a more complete in Labroides the head is naked except for sub­ description with necessary meristic, measure­ orbital, postorbital, and opercular regions), ment, and detailed descriptive data on the a slightly emarginate caudal fin, eight instead species is available, I prefer not to recognize this name . It is possible that L. auropinna is 1 Contribution No. 209 from the Marine Labora ­ tory, University of Miami, Miami, Florida. Manuscript not a species of Labroides, and perhaps not received Ocrober 16, 1956. even a labrid.

327 328 PACIFIC SCIENCE, Vol. XII, October, 1958

The genus Labroides is distinguished from often distinct anteriorly on body and on other genera of the Labridae in having a small head) 2 mouth with thick lips, the upper with a shal­ 2a. Snout long, diameter of eye contained 2.1 low median groove and the lower divided to 2.5 times in snout length; caudal pe­ . into two prominent, anteriorly projecting duncle and caudal fin pale with a promi­ lobes. The inner surface of the lips has fleshy nent black crescent posteriorly in fin, the folds. There is a single pair of large, curved, attenuate ends of which extend to mar­ canine teeth anteriorly in each jaw, the upper gins of caudal peduncle (tropical Pacific) pair fitting inside the widely-spaced lower · Lab roides bicolor pair when the mouth is closed. There is a large, anteriorly directed, canine tooth at the 2b. Snout not long, diameter of eye contained extreme posterior part of the upper jaw, sep­ 1.5 to 1.9 times in snout length; caudal arated by a gap from anterior teeth. The peduncle and caudal fin black with upper remaining teeth are small, those in the lower and lower edges of fin pale 3 jaw in several close-set rows between the ca­ nines and those in the upper jaw forming a 3a. Median lateral black band which extends large mass just behind the anterior canines, pos teriorly from snout continuous with this mass with a marked indentation anteri­ black posterior half of body; no dusky orly in the mid-line. The body is compressed, streak on cheek below eye; pale edges of its width contained about 2,5 to 3 times in caudal fin magenta in life; maxim um the head length, and moderately elongate, its standard length in excess of 80 mm . depth contained about 3.5 to 4 times in the (Hawaiian Islands) . standard length . The caudal fin is truncate or ·. . . .Labroides phthirophagu s, n. sp. slightly rounded . The preopercle is entire, its 3b. Median lateral black band which extends margin smooth. The gill membranes are at­ posteriorly from snout merges gradually tached to the isthmus. The snout is pointed. to broad, pale-brown (dull-orange in life) The snou t, entire dorsal surface of the head, area in center of body; a narrow dusky chin, and throat are scaleless. The lateral line streak on cheek below eye running from is continuous, and the lateral line scales num­ chin to base of pectoral (may be faint in ber 28 or 52 to 54. Fin ray counts are as fol­ small adults ); pale edges of caudal fin lows: D IX, 11 or 12; A III , 10; P 13 (upper­ light lavender in life; maxim um standard most two rays unb ranched). The species are length about 60 mm. (Society Islands small, individual fish usually being less than and Tuamotu Archipelago) . 100 mm . in standard length. · Labroid es rubrol abiatus, n. sp. KEY TO THE SPECIES OF L ABROID ES J uvenile specimens for all species are un­ (Applicable primarily to adults ; available to me; however, all have been see discussion below concerning juveniles) viewed underwater, frequently in the proxim­ ity of adults. J uveniles of all four species ap­ 1a. Lateral line scales 52 to 54; color light pear to have the same basic color pattern, blue with a median lateral black band namely black with a broad band of color from snout to end of caudal fin (this band along the back which extends and narrows on broadening as it passes posteriorly on to snout. On L. dimidiatm and L. rubrolabiatus body) (Indo -Pacific) . this band is brilliant deep blue (these species ...... Lab roides dimidiatu s are readily separable by scale counts; see key lb. Lateral line scales 28; color not as above above); on L. bicolorit is bright yellow; on L. (although a median lateral black band is phthirophagus it is bright purple. Review of Labroides- RANDALL 329

Labroides dimidiatus and Moorea, another color variant was ob­ (Cuvier and Valenciennes) served. Occasional adults have a dull red­ Fig. 1 orange region in the middle of the body be­ low and adjacent to the median black band. Labrus latovittatus Riippell, (non Lacepede), In the Marquesas L. dimidiatus were seen with 1835, Neue Wirbelth., Fische ... , p. 2. a red-orange area which was longer and oc­ Cossyphus dimidiatus Cuvier and Valenciennes, curred above as well as below the black band. 1839, Hist. Nat. des Poiss., vol. 13, p. 136. Barnard (1927: 749) described two color Labroides paradiseus Bleeker, 1851, Natuurk. varieties of the species at Natal, East Africa, Tijdschr. v. Nederland. Indie, vol. 2, p. 249. one with a dark stripe across the base of the Labroides bicincta Saville-Kenr, 1893, The pectoral fin and one without. He stated that Great Barrier Reef of Australia, p. 308, pl. the color of the species in Africa is blue or 16, fig. 4. yellow with a black longitudinal band. Smith Labroides caeruleo-lineatus Fowler, 1945, Acad. (1949: 291), also reporting on the species Nat. Sci. Phila:, Proc., vol. 97, p. 65, fig. 7. from East Africa, described the color as vary­ ing rapidly from light pink through straw TYPE LOCALITY: Mauritius. yellow to dark blue, apparently at will and This blue and black species is the most according to the emotional state of the fish. I common and widespread of the genus,rang­ have never seen L. dimidiatus in the Pacific ing from Africa to the rropical Pacific (where pink or yellow instead of blue, nor have I it is recorded from most major island groups). observed any rapid color changes. With age Labroides paradiseus Bleeker differs from there is a loss of brilliance and a lightening of dimidiatusin having a hooklike ventroanterior the blue color. Below the median black band extension of the broad black band in the large adults are nearly white. caudal fin. I have observed various degrees of A juvenile specimen, 24 mm. in standard intermediacy between typical dimidiatus and length, collected by the author in Tahiti, was rhe paradise!Is form, and I regard the latter as colored in life as follows : black with a band a color variety, as did Gunther (1881: 243). of deep blue about a pupil diameter in width In the Society Islands and Tuamotu Archi­ beginning at upper lip and passing backward pelago I have seen no specimens of the para­ through upper part of eye (lower edge of diseus variety, although collections are not ex­ band at upper edge of pupil) on to nape, tensive. In large collections from the Marshall where it is nearly two pupil diameters in Islands and the Philippines in the United width, and thence on to back, where it gradu­ States National Museum the paradiseus form ally narrows until it terminates dorsally on predominates. caudal peduncle; the black middorsal region, At Makatea and Takaroa in the Tuamotu which is bordered by a blue band on each Archipelago, and to a lesser extent at Tahiti side, is broadest on the head and narrows as

FIG. 1. L abroides dimidiatus, 83 mm. , Society Island s. Caud al fin not expanded, hence ernarginare appearance. 330 PACIFIC SCIENCE, Vol. XII, October, 1958 it passes posteriorly until it ends at rear of off its larger companion's beaklike jaws, teeth dorsal fin; head and thorax below level of and head area." Th ey implied that this un­ lower edge of eye are pale; caudal fin black usual association is limited to just these two with narrow pale-violet edges; dorsal and anal species; however, I have observed (Figs. 2, 3, fins dusky basally and hyaline distally, the 5) L. dimidiatus picking at the heads, bodies, dorsal blackish anteriorly , the anal darker and fins of numerous species of reef fishes posteriorly ; paired fins pale. representi ng many different families, includ­ Labroides caeruleo-lineatus Fowler (1945: 65, ing the carangids and serranids, members of fig. 7) (1946: 159) is a young L. dimidiatus. which habitually prey upon small reef fishes. The description was based on a 28 mm. speci­ None of the above authors mentioned the men, and the figure clearly shows the typical removal of ecroparasites; the use of the term juvenile color pattern as described above. cleansing gives the connotation of removal L. dimidiatus occurs most commonly in of debris or particles of food . coral or coral-sand areas which are protected The stomach contents of two specimens from wave action . Around atolls individuals from the Gilbert Islands and three from the are most often seen in lagoons rather than Marshall Islands consisted of calagoid cope­ outside the peripheral reefs. I have observed pods which are ectoparasitic on fishes; two them in water as shallow as 2 feet and as deep more specimens from the Gilberts had eaten as 120 feet, but they appear to be more small isopods (also fish parasites) along with abunda nt in shallow water. Regan (1908: 230) a few fish scales (Randall, 1955b: 144). Sub­ has recorded the species from 34 fathoms in sequently, a specimen, 72 mm. in standard the Maldives. length, from Hull Atoll, Phoenix Islands, .Of this species de Beaufort (1940: 148) was found to contain three caligid copepods wrote, "In the aquarium of Amsterdam I ob ­ served that Labroides dimidiatus cleans the sur­ rounding of the mouth and the gill openings of large fishes (H et Aquarium IV, 1936: 153) in the same way as has been observed by Beebe for the Atlantic species Iridiobivittatus." Smith (1949: 291) stated that L. dimidiatus "Feeds on minute organisms on rocks and has been observed to nibble over the mouth parts and gill covers of large Rock- cods." Doty and Morrison (1954: 24) observed what they termed an interesting association be­ tween a parrot fish and a fish which is un ­ doubtedly L. dimidiatus? The latter "cleans

2 These authors erroneously refer to this specie s as a blenny-probably because of the confusion that results from the similarity in color pattern of L. dimidiatus to the blenny Aspidontus taeniatus Quoy and Gaimard, a similarity so striking that Barnard (1927: 749) sug­ gested that mim icry might be involved, with perhap s on e or the other species being poisonous. R. W. Hiatt and D. W. Strasburg observed the resemb lance at Arno Atoll in the Marshall Islands and suspected mimicry. FIG. 2.A parrot fish, Scarus sp.(upper fish), and a Randall (1955b: 144), noting it in the Gilbert Islands, surgeon fish, Zebrasoma scopus (lower fish), each being proposed that the blenny might be mimicking the examined for ectoparasires by L abroides dimidiatus. labrid, since the latter might gain protection from Rep roduced from 16 mm. movie film taken in the pred aceou s fishes by virtue of its food habi ts. lagoon of Takaroa, Tuarn otus. Review of Labroides- RANDALL 331

they remain completely motionless in the water, although they may swim slowly. Usu­ ally their fins are fully erected (this may be associated with the high percentage of crusta­ cean parasites which occur on the fins as compared with the rest of the body epider­ mis), and there may be a disorientation of the body to a position out of the vertical plane (i.e., tilted to one side) or with the anterior part of the body elevated. At times color changes are apparent. The black-hued surgeon fish Acanthurus achilles Shaw, for example, may become bright blue when picked over by Labroides. Of all of the species of Labroides, L. dimidi­ atus is the least wary and will often swim up to a human observer, sculling along by pec­ FIG. 3. A pair of L abroides dimidiatus pickin g at the toral fins alone in typical wrasse fashion , as if wrasse Epibulis insidiator. Re produced from 16 mm. movie film taken in the lagoon of Takaro a, Tu amotus. to investigate the possibility of human ecto­ parasites. 1 have experienced a picking at the hairs of my legs by this species and on one and three small fish scales in the stomach and occasion a persistent and vigorous nipping the digested remains of numerous calagoid at a small mole on my chest. copepods in the intestine. A 71 mm . speci­ Randall (op. cit.) noted a peculair mode of men from the Society Islands had two cala­ swimming often displayed by L. dimidiatus goids-in the stomach and four calagoids and when near other fishes, describing it as an one larval gnathiid isopod in the intestine. oscillation of the posterior part of the body One section of the intestine had a bit of soft during slow progression. Doty and Morrison bottom debris including a few fragments of wrote that this species swam "in an odd algae . A 65 mm. specimen from the Society leaping fashion , somewhat like a finch in Islands had an empty stomach and the re­ flight. " 1 presume that this "dancing" about mains of 13 calagoid copepods and 1 lernaeid other fishes by this and other species of copepod in the intestine. A 24 mm . juvenile Labroides, coupled with their remarkable col­ had eaten five small larval gnathiid isopods or, serves to enhance their recognition by the and three immature calagoid copepods. Non­ other fishes. Further, 1 noted in the Society parasitic crustaceans were not found in any of Islands that L. dimidiatus, when near other the specimens examined. The bottom debris fishes, often elevates its dorsal fin, especially in the one specimen may indicate occasional anteriorly, such that the black forward part of feeding on free-living forms, however, the the fin appears like a conspicuous black species appears to be at least dominantly a triangle. feeder on crustacean ecroparasites of fishes. On several occasions 1 have seen Labroides Although Doty and Morrison (op. cit.) dimidiatus enter the buccal and gill cavities of stated that parrot fish were feeding when the larger fishes, ostensibly to remove parasites. labrid cleansed them, host fishes are generally Probably this occurs more often than these not so oblivious to having their parasites re­ few observations suggest, for the proximity moved. Usually the fishes display some dis­ of a swimmer appears to frighten the host tinctive behavior during the process . Often fishes sufficiently to preclude their entering 332 PACIFIC SCIENCE, Vol. XII, October, 1958 into this more complex symbiotic association. Epibulis insidiator (Pallas), the grouper Variola In Paperoai Bay, Moorea, at a depth of 20 louti (Forskdl), and the goatfish Parupeneus feet, about a 200 mm . Parupeneus trifasciatus barberinus (Lacepede) via the gill openi ngs. (Lacepede) approached an adult L. dimidiatus. Labroides dimidiatus are frequentl y seen in As it neared the latter, it changed its color pairs. Also it has been noted that individ ual rapidly from light tan to mottled pink. The fish or pairs of fish appear to remain around Labroides picked over the body and fins of the same small coral head or small section of this goatfish , moving anteriorly. When it a larger reef. The fish described above in as­ reached the forward part of the head , the sociation with Parupeneus trifasciat«: is one of goatfish opened its mouth and the wrasse in­ a pair which has been sighted in the same serted more than half of its body into the area over a period of six months. Other pairs mouth and remained there several seconds . or individual fish have been observed to be Subsequently two adult goatfish of the spe­ restricted to small areas. cies Mulloidichthys samoensis (Gunther) inter ­ If a parasite-feeding fish is resident to a rupted the pair and received the attention of small section of bottom and other fishes in the Labroides. D . W.Strasburg has written the area are nonmigratory (probably true for that he made the same observation with L. the majority of coral reef fishes), its food sup­ dimidiatus and P. trifasciatusat Eniwetok Atoll ply would be insufficient, for it would be in the Marshall Islands . In Teavenui Pass, limited to the ectoparasites offishes in its im­ Bora Bora, Society Islands, at a depth of mediate surroundings. Therefore, it was not about seven feet, I saw a 1240 mm. moray eel surprising to note that fishes come from be­ (Gymnothorax javanicus Bleeker) with its head yond the range of vision to a Labroidessite. A and anterior third of its body projecting from school of subadult Mulloidichthys samoensis beneath a coral ledge. It was being picked was consistently seen to occupy a region of over by a L. dimidiatzis about 50 mm . long. sandy bottom over 50 feet from the shore After about 30 seconds the wrasse devoted its reef where the pair of L. dimidiatus menti oned attention to the eel's head, whereupon the above could be found. As I watched, small latter opened its mouth widely, maintaining groups of about eight of the goatfish moved it in this position for about eight second s toward the Labroides site and remained in the while the labrid picked inside the upper and area until they had been picked over; then lower jaws and then disappeared back into they would return to their usual place. Their the pharynx. A sharp lateral jerk of the eel's swimmin g back and forth from where the head preceded the departure of the wrasse. school congregated to where the Labroides The eel was then speared to permit positive were resident was not haphazard but direc­ identification. A similar observation was made tional, even though the water was not clear at Takaroa in the Tuamotus . The eel ap­ enou gh for me to see the reef from where the peared to be the same species and the L. school was situated. Another species which dimidiatus was a juvenile about 20 mm . in swam to these two labrids from a locality length. At Caroline Atoll (10° S., 150°14 ' W.) previously known to me was a large adult I saw an adult L. dimidiatus slip the anterior A budefduf septemfasciattlS (Cuvier and Valen­ third of its body into the gill cavity of the ciennes), the only one of its size in the area. chaetodontid Hemitaurichthys thompsoni Fowler The rocks along the shore where it habitually by way of the gill open ing. The latter held its hid are 45 feet from the labrid locale. Caranx opercula open without respiring for several melampygus Cuvier and Valenciennes, a foot seconds to accommodate the labrid. In the or more in length, were other visitors which Tuamotus individual L. dimidiatus were ob­ came to be "serviced" and left when the served to enter the gill cavities of the wrasse service had been finished. These fish are what Review of Labroides- RANDALL 333 might be termed roving carnivores and ap­ dividual of this species pecking at the body pear to move over consid erable distances. Yet of the Nassau grouper, Epinephelus striatus they seemed to know the Labroides site, for (Bloch), in Florida, and can thus corroborate they would come up to it from deeper water Longley's observation. In addition, I saw a of the bay. At the atoll of Takaroa in the juvenile of the pomacentrid Microspathodon Tuamotus a four-foot moray eel (Gymnothorax cbrysyrus (Cuvier and Valenciennes) picking javanicus) was seen to leave a hole in the over the fins and bod y of an angel fish, coral, swim 15 feet over the bottom to a small Pomacantbus arcnatus (Linnaeus), which re­ coral head where a Labroides dimidiatus was mained nearly motionless during the process. located. There it was picked over the head and Also suspected of feeding on ectoparas ites body by the labrid, after which it returned to of fishes by Longley are the goby Elacatinus the hole . Other fishes, such as Ctenochaetus oceanopsJordan (p. 226) and the youn g of the striatus (Quoy and Gaimard), which were labrid Thalassoma bifasciatum (Bloch) (p. 198). commonly seen being nibbled at by L. dimidi­ Of the former he wrote , "Wherever found , atus, could not be distinguished from one these small fish slip out from shelter, attach another or identified as having come from a thems elves by their ventral disks to other certain area; nevertheless, it is expected that fishes, and 'creep' over them , presum ably to some means of marking many such fishes in look for parasites. As many as 6 have been the area, such as with different colored tags, seen together on a Mycteroperca venenosa about would demons trate that these fish are aware ~ yard long. .. . Th e boldness of these tiny of the place where the Labroides can be found fish is almost incredible. They will creep over and swim there, perhaps under the stimula­ the teeth of the great Pseudoscarus, or enter tion of being irritated by ectoparasites. Indi­ the mouths of grunts and groupers and ex­ vidual L. dimidiatus are not wholly dependent plore them with unhurried movements. . .. on fishes bearing parasites coming directly to Their atte ntions usually continue until the them. Often the species has been observed larger fishes grow restless, start up abrup tly, swimming over the reef for distances as great or move away and leave them , whereupon as 60 feet "servicing " fishes on the way. This they return to their original stations." was especially true in areas where the high ly Interestingly, the color ofthis goby is simi­ territorial damsel fish Pomacentrus nigricans lar to that of Labroides dimidiatus. E. oceanops (Lacepede) was common. As one damsel fish which I observed in Florida were blue with was being tended, another nearby would as­ a black longitudinal band running from the sume a stationary pose with fins erect and the snout through the lower part of the eye, Labroides would move on to the latter with broadening on the body, and endin g on little hesitation . For further discussion of the ventral half of caudal fin. subject of restricted " home" sites of Labro­ A recent paper by Eibl-Eibesfeldt (1955) ides, as applicable to L. phthirophagus, see the is devoted in a large part to the detailed account of this species. observation of the behavior of fishes which The habit of feeding on ecroparasites is not "cleanse" larger fishes. He observed the fol­ unique to the genus Labroides, as might be lowing species in the Caribbean in symbiotic surmised from the reference above to the association with larger fishes: Eiacatinus oce­ labrid Iridio ( = Halichoeres) bivittatus of the anops, young Thalassoma bifasciatmn, youn g Atlantic. Longley (in Longley and Hilde­ Anisotremus virginicus, young Bodianas rtlfus brand , 1941: 129) reported that small pork­ (Linnaeus) (Labridae), and Gramma bemicbry­ fish, Anisotremus virginictls (Linnaeus), nibb le sos Mowbray (Pseudochromidae). He noted and peck at the surfaces of larger fishes at that the larger fishes do not merely accept the Tortugas.I was fortunate to observe one in- cleansing but invite it by takin g special posi- 334 PACIFIC SCIENCE, Vol. XII, October, 1958

tions . Furthermore he observed that the sites of other California fishes. larger fish seek out the stations of the "clean­ The food habit of removal of ectoparasites ers." He made the followin g observation of from other fishes represents a distinc t bio­ Elacatinu: oceanops and Epinephe/us striates. As logical niche. It is interesting to note how soon as the grouper came close to the coral different species, often from totally different where the gobies were located, the little fish families, have filled this niche in different swam immediately in his direction and started areas of the world ; however , more of them are to clean his body. Sometimes the grouper lay species of Labridae than of any other family. on his side. He allowed the gobies to enter It should be added that primarily tropical and and leave the buccal cavity through his mouth subtropical areas, where labrids are usually and gill openings which he held rigidly open. numerous, have been investigated. After about 30 seconds he respired once or In addition to fishes like Labroides, certain twice and then opened his mouth and elevated shrimps have been observed removing ecro­ his gill covers again . When he wanted to parasites from fishes. In Papetoai Bay and the leave, he made a signal by closing his mouth lagoon of Moorea at depths of about 70 feet sharply, although not completely, and then the author watched a shrimp picking at fishes. opening it widely. At this signal the gobies The shrimp was identified as Hippo/ysmata came out. Before leaving the grouper shook grabhami by Dr. F. A. Chace, Jr. The adult his body. Even if frightened (as by a diver) he shrimp are about 40 to SO mm . in total body still too k time to make the signal to the length , have a brilliant red band along the gobies. Of the small fishes listed above which back which is bisected by a white middorsal cleanse larger fishes, only E/acatinus oceanops line, and long white antennae and antennules. and young Bodianus rttfUS were seen to enter They were seen on the surface of isolated the buccal cavity of larger fishes. Large fish block s of coral. When fish approached they were never observed trying to catch any of waved their antennae and antennules as if to the little fish. Eibl-Eibesfeldt suggests that it attract the fish. Among the fishes seen to be ascertained whether the appetite of pre­ enter into symbiotic association with this daceous fishes is first appeased before they shrimp were: Apogon exostigma (Jordan and allow themselves to be picked over by the Starks), Apogon sp., Anthiassp., juvenile Acan­ smaller fishes. No mention was made of the thttrtls mata (Cuvier), and Gymnothorax flavi­ examination of stom ach contents for ecro­ marginata (Riippell). A shrimp climbed on to parasites by this author. a two-foot specimen of the latter species and Eibl-Eibesfeldt has informed me in a lette r busily picked over the dorsal part of the eel's that he has since observed" Corisgiof rediRisso head . cleaning a Crenilabris" in the Mediterranean south of Naples . Labroides bicolor Fowler and Bean Szidal and Nani (1951: 412) examined the Fig . 4 stomach contents of a Remora remora (Lin­ naeus) from the coast ofArgentina and found Labroidesbicolor Fowler and Bean, 1928, U. S. four parasitic copepods of the family Caligi­ Natl. Mus., Bul. 100, vol. 8: 224, pI. 18. dae, three of them males of Achteinus dentatus TY PE LOCALITY : Port M aricaban, Philippine Wilson and the other a female Pandarus; both Islands . forms are known to parasitize sharks . J udging from the few locality records of Hubbs and Hubbs (1954: 194) stated that the species (Philippines, Java (Hardenberg, the embiotocid Bracbyistias frenatus Gill and 1936), J apan (Kamo hara, 1952), the Gilbert the labrid OX-YiN/is ca/ifornica (Gu nther) have Islands (Randall, 1955b), Aldabra, Indian at times been observed feeding on ectopara- Ocean (Smith, 1955)), L. bicolo r would ap- Review of Labroides - RANDALL 335

FIG. 4. Labroides bicolor, 81 rnm., Gilbert Islands . pear to be rare, although wide-ranging in adult stage. In the following account, the the Indo-Pacific. I can attest to its rarity at color of juveniles is based on underwater ob­ Onotoa Atoll in the Gilbert Islands, where servation. Juveniles as small as about 15 mm . two months of field work resulted in the in standard length have been observed. From sighting ofonly two individuals. Subsequent­ this size to about 30 mm. the fish appear en­ ly in the Society Islands and Tuamotu Archi­ tirely black except for a brilliant yellow band pelago I have found the species surprisingly along the back above the midlateral line of common for the genus . The two fish at the body (this band confluent with one on Onotoa were observed in clear water beyond other side dorsally on caudal peduncle and the wind ward reef of the atoll. Dr. Leonard caudal fin) which extends on to head , includ­ P. Schult z of the United States National Mu­ ing upper edge of eye and narrows as it termi­ seum has informed me that he will record the nates on snout. At a length of about 30 mm. species from the northern Marshall Islands in the same general color obtains but the caudal volume II of Fishes of the Marshall and M ari­ fin and posterior part of caudal peduncle be­ anas Islands. His specimens were taken on come pale yellow or almost white; subse­ reefs exposed to severe wave action. In the quently, at the posterior part ofthe caudal fin, Society Islands and Tuamotus, however, the a trace of the black crescent so typical of species has rarely been observed outside of adults appears. Then the yellow band along the barrier reefs but occurs more often around the back is replaced by light gray (the brilliant reefs in protected lagoon and bay areas. The yellow color persisting longest on the head) , apparent difference in habitat between this and a light-gray area appears ventrally, thus re­ fish in the Gilbert and Marshall islands and stricting the black body color to a broad band the Society Islandsmay be a manifestation of along the middle of the body and head. possible subspecific differentiation. Further Concomitant with the appearance of gray collections and observations are needed to color is the development ofthe black crescent elucidate the problem. Specimens from the in the caudal fin. Anterior to the crescent a Society Islands and Tuamotus have been sent greenish tinge appears , replacing the pale­ to the United States National Museum and yellow color. Curiously, the next trend is a the Natural History Museum, Stanford Uni­ darkening of the light gray above and below versity. L. bicolor was observed but not taken the black band on the head and body and an from the Marguesas Islands. intensification and enlargement of the yellow Although only adults were collected, the area posteriorly on the body. At about 60 mm. abundance ofL. bicolor at the island ofMoorea the yellow area is defined but scales anterior enabled me to perceive marked color changes to the caudal peduncle have blackish centers . which take place from the juvenile to the It is not until a standard length of nearly 336 PACIFIC SCIENCE, Vol. XII, October, 1958

75 mm . is reached that this large elliptical length, have been examined. Four were empty region is bright yellow, free or nearly so of save for a mucuslike substance and, in the scales with dusky centers. The green on the case of two of the specimens , tiny digenetic caudal fin also intensifies and blue appears flukes in the intestine, which were probably posterior to the crescent and on the outer internal parasites of these fish. A 69 mm. part of the dorsal and anal fins. On large fish specimen had two calagoid copepods and one (the species is the largest of the genus, at­ fish scale in the stomach ; 65 and 80 mm . taining a standard length slightly in excess of specimens were empty except for a few fish 100 mm. ) a suffusion of dark blue becomes scales in the · stomach and intestine. The apparent anteriorly . stomach of a 60 mm . specimen was empty, A 98 mm. specimen speared in Moorea but three small larval isopods of the family provided the following color description: Gnathiidae (larvae of this family consist ex­ body black except for a broad yellow area clusively of fish parasites) were found in the posteriorly, the forward end of which is intestine. rounded and reaches a vertical through base of The presence of fish scales in the gut is fourth soft ray of dorsal fin; head dark blue comprehensible when the stout, curved, ca­ and lacking the prominent black lateral band; nine teeth and the vigor with which the little lips dark blue, shading to light blue on inner labrids nip the host fishes are considered. A surfaces; caudal fin with a subterminal black medium-sized L. bicolor was seen to actually crescent, bright green anterior to and blue lift a piece of skin of an adult puffer, A rotbron posterior to this marking; dorsal and anal fins meleagris Bloch and Schneider, into a pro­ black with broad light blue edges ; pectorals nounced peak in its apparent effort to dis­ hyaline; pelvics black. lodge an ectopa rasite. This specimen and another 94 mm. one The host fishes do not always tolerate such with the same color are males. On other indi­ ardent pecking. Often they move their bodies dividuals as large as blue-headed ones seen sharply or swim away when the pecking underwater the dark midlateral band is visible becomes vigorous. on the head and the deep-blue color is lack­ Adult L. bicolor do not display the oscil­ ing , although the head and anterior part of latory swimmin g movement as previously the body above and below the bandare dark described for L. dimidiatlls. The brilliant yel­ grayish blue. The largest of the latter color low and black young do, however. A further variety collected in the Society Islands is distinction between the behavior of youn g 80 mm . in standard length, and it is a female. and adults lies in the tendency of the former It is believed that others may consistently to remain in a very restricted region, fre­ prove to be females too. At times one of each quently a small cave or under a ledge. Adults color variety are seen together as a pair. The are almost constantly on the move and appear above-mentioned color differences, if con­ to cover a larger area than other species of clusively demonstrated to be sexual, are slight Labroides . Still, they seem to remain in the compared to the degree of sexual dichroma­ same general region. In addition to seeing tism known in some of the Labridae (Randall, what appeared to be the same individuals day 1955a). after day at a certain reef, several which were On numerous occasions in the Society wounded with a spear and could be positively Islands and Tuamotus individuals of this identified were repeatedly sighted at appro xi­ species, both juvenile and adult, have been mately the same place. observed picking at the bodies, heads, and As might be surmised from the above, fins ofother fishes. The contents of the gut of adults of this species are more prone to seek eight specimens , 60 to 98 mm . in standard out fishes from which to remove parasites Review of Labroides- RANDALL 337 than the other species of the genus. They have been observed to follow individual fishes closely for 30 feet or more , frequently making contact with the dorsal fin or the dorsal part of the body of these fishes. I have never observed L. bicolor enter the mouth or gill cavities of larger fishes as has been observed for L. dimidiatus, L. phthiropha­ gus, and L. rubro/abiatus. L. bicolor is notice­ ably shyer than the other species, however, and individual fish usually swim away as a swimmer approaches; thus it is possible that such behavior occurs. The three species of Labroides which are found in the Society Islands, Tuamorus, and Marquesas, L. dimidiatus, L. bicolor, and L. rubro/abiatus, are frequently seen near one another and rarely exhibit territoriality. On FIG. S. Labroides bicolor, (on right) being "serviced" one occasion I observed all three together by L. dimidiatus. Reproduced from 16 mm . movie film taken in the lagoon of Takaroa, Tu amorus . around a single small coral head. At Takaroa in the Tuamotus L. dimidiatus and L. rubro­ /abiatus were both seen to pick at the body of or less in standard length have been seen the parrot fish, Searus harid Forskal, at the in association with adult groupers (Cepha/o­ same time. In Moorea a five-foot moray eel pbolis argus Bloch and Schneider) 200 to of the species Gymnothorax javanieus was ob­ 300 mm. in standard length and L. bicolor of served lying nearly motionless entirely in the about the same small size with full-grown open beside a small head of coral and being squirrel fishes (Myripristis spp. ). picked over simultaneously by two adult At Takaroa an adult L. bicolor was ob­ Labroides, one L. dimidiatus and the other L. served to swim after a six-foot shark of the bicolor. Until the small labrids were seen, the species Triaenodon obesus (Riippell) and to eel's position was perplexing, for morays are make a few hasty pecks at the dorsal part of rarely seen free from their holes in the coral the body. during the day unless in rapid transit from one hole to another. L. dimidiatus has thrice Labroides phthirophagus, new speCIes been observed to peck briefly at the bodies of PI. 1A adult L. bicolor; twice the dimidiatllS were of about equal size as the bicolor (Fig. 5), and HOLOTYPE: U. S. Natl. Mus. No. 164466, a the remaining time definitely smaller. male specimen, 68.0 mm. in standard length Generally , individuals of the species of and 84.0 mm. in total length, collected by Labroides tend fishes larger than themselves , J. Randall with a spear in 7 feet of water, juveniles "taking care" of the smaller reef about 100 yards offshore from the Waikiki fishes. It is by no means rare, however, to see branch of the Hawaii Marine Laboratory, species of Labroides picking at fishes smaller Honolulu, Territory of Hawaii , on March 1, than themselves . Conversely, tiny juveniles 1955. have been frequently seen busily working PARATYPES: U. S. Natl. Mus. No. 164469, over the surface of fishes ten times or more 69.0 mm. in standard length, collected by their size. Several times L. dimidiatus 20 mm. J. Randall with a spear in 5 feet of water in 338 PACIFIC SCIENCE, Vol. XII, October, 1958

Kealakekua Bay, Hawaii, on June 15, 1954; large paratype and none in the juvenile speci­ U. S. Nat!' Mus. No. 164467, 83.5 mm. in men). Snout, chin, throat, and dorsal part of standard length, collected with rotenone by head naked. Gill rakers (including rudiments) W. Gosline and class at Diamond Head, on first gill arch 14 (large paratype only). Oahu, on Dec. 22, 1951; U. S. Nat!' Mus. No. Head length 2.77 (2.22- 2.80) ; depth of 164468, 18.8 mm . in standard length, col­ body 3.64 (3.69-3.50) ; snout to anus 1.59 lected with rotenone by W. Gosline and class (1.60-1.61); snout to origin ofpelvic fins 2.72 at Diamond Head, Oahu, on Dec. 23, 1953; (2.77-2.74) ; snout to origin of dorsal fin 2.86 Stanford Univ . Mus. No. SU48445, 2 speci­ (2.51-2.95); length of dorsal fin base 2.06 mens, 64.5 and 69.0 mm . in standard length, (2.09- 1.95) ; length of anal fin base 3.54 collected by J . Randall with a spear in 18 feet (3.76- 3.76)- all in standard length. of water at Waianae, Oahu, on Oct. 14, 1955; Width of body at gill opening 2.34 (3.12­ British Mus. No. 1955.12.12.1, 60.0 mm. in 2.64); least depth of caudal peduncle 1.92 standard length, collected by J. Randall with (2.62- 2.04); snout length 3.46 (3.73- 3.21); a spear in 35 feet of water at Manana Island diameter of eye 5.16 (4.20-5.97) ; width of (Rabbit Island ), Oahu, on Sept. 29, 1955. interorbital 3.64 (4.65- 3.59); length of pec­ In addition to the above , there is a series of toral fin 1.61 (1.87-1.65); length of pelvic fin 12 specimens, 19 to 76 mm. in standard 2.31 (2.80-2.29); width of mouth (rictus to length, in the collection of the University rictus) 5.39 (5.25-5.98); mid-center of upper of Hawaii. lip to rictus 6.13 (5.80-6.15)-all in head D ESCRIPTION: Based on the holotype and length. the smallest (18.8 mm. in standard length, Dorsal spines progressively longer, the 23.2 mm. in total length) and the largest ninth 3.50 (3.90-3.59) in head length; first (83.5 mm. in standard length, 103 mm. in dorsal soft ray 2.45 (2.89-2.54) in head total length) paratypes. Counts and measure­ length ; second anal spine twice as long as ments are recorded for the holorype, followed first; third anal spine twice as long as second; in parentheses by data for the small and large third anal spine 3.36 (3.72-3.32) in head paratypes, respectively. When counts of the length; first anal soft ray 2.29 (2.80-2.56) in two paratypes are the same as the holotype, head length. Caudal fin slightly rounded. only a single number is given. A pair of large curved canine teeth in upper Dorsal fin rays IX , 11; anal fin rays III, 10; jaw, each nearly a pupil diameter in length pectoral fin rays 13 (the uppermost rudimen­ and separated by a distance about equal to the tary, the next unbranched ); pelvic fin rays diameter of one of the teeth at the base. A I, 5; principal caudal rays 14. Lateral line similar pair of canine teeth in lower jaw, scales 28, the last slightly enlarged and with lateral to but nearly touching upper teeth a prominent lateral line tube; lateral line when mouth is closed. Remaining teeth small, partially interrupted in adults (the lateral line except one at angle of jaw, tho se at the in the 20th and 21st scales, the two which symphysis of upper jaw in numerous close­ contain that part of the line which angles set rows, forming a semicircular mass between sharply downward at the level of the posterior the two canine teeth . part of the dorsal fin, is poorly developed; no Color in alcohol : posterior half of body degeneration in the lateral line at this location and caudal fin black except upper and lower is apparent in the 18.8 mm. juvenile speci­ edges of caudal fin and dorsal edge of caudal men , however). Scale rows above lateral line peduncle which are white (magenta in life); to base of first soft dorsal rays 3; scale rows anterior half of body light tan (yellow in life) below lateral line to base of first anal soft with a median dorsal black band (beginning rays 8; median predorsal scales 8 (9 or 10 in in a small spot on median upper portion of Review of Labroides - RANDALL 339

upper lip) and a median lateral black band underwater in Kaneohe Bay, Oahu . running posteriorly from lower corner of up­ Although generally seen around rock or per lip and tips of lobes of lower lip through coral, Labroides phthirophagtls does not appear . eye to merge with black posterior part of restricted to any special habitat. I have ob­ body; dorsal and anal fins pale (light blue in served it in clear water of high coral cover life), except for basal part of spinous portion such as Kealakekua Bay and turbid, at times of dorsal fin which is black (this color con­ brackish, water lacking living coral such as tinuous with that of median dorsal black the Ala Wai Yacht Basin, Honolulu (it is, band ) and a narrow horizontal dark line ex­ however, uncommon in the latter area). It tending from ends of last few dorsal spines occurs in water at least as shallow as 2 feet posteriorly into the anterior half of the soft and has been seen at a depth as great as 90 portion of the fin about three-fourths the feet (it may occur deeper- there have been distance from the base of the fin; paired fins few observations beyond 90 feet). pale. The body of the 18.8 mm . juvenile L. phthirophagllS (Greek : phtheiros, louse; specimen is black except for a broad light­ pbagoi«, eat) is named for its habit of feeding brown band along the back, this band con­ on external parasites of fishes. tinuing on to top of caudal fin; the remainder The food habits of L. phthirophagllS were of the caudal fin is black like the body except suspected from observation of its picking at for lower distal corner which is pale; the head the surface of other fishes (Fig. 6) and from is brown with median dorsal and median the knowledge that L. dimidiatus feeds on lateral black bands. In life the band along the parasitic copepods and isopods from fishes. dorsal part of the head and body is bright .In order to obtain more positive evidence purple . No yellow color is visible on juve­ of such a mode of feeding , the contents of niles. The first indication of the yellow colora­ the alimentary tract of 11 specimens of L. tion is a suffusion of tan over the head and phthirophagllS in the collection of the Uni­ anterior part of the body. This occurs at a versity ofHawaii were examined. None of the length of about 50 mm . (based on estimates specimens represent type material. They were of the size of individuals seen underwater). Labroidesphthirophagtls is known only from the Hawaiian Islands, where it is a common species. That it escaped being described in Jordan and Everrnann's time is probably due mainly to its small size. It is too small to be caught in the usual fish traps or hook and line. An indication of its abundance is its listing [as L. dimidiattls by Brock (1954: 307)] among the species consistently tabulated in his underwater transects of Hawaiian reef fishes. The record of Labroides dimidiatllS by Gunther (188 1: 243) from the Hawaiian Is­ lands is open to question. It seems possible that he might have obtained L. phthirophagtls and confused it with L. dimidiaua, for the latter, if it occurs in Hawaii at all, is certainly FIG. 6. LabroidesphthirophaguJpicking at the body of rare. On one occasion I observed a small a goatfish, P arupeneus multi/aJciatuJ. Reproduced from individual of what I believe to be this species 16 mm . movie film taken off Manana Island, Oahu. 340 PACIFIC SCIENCE, Vol. XII, October, 1958 collected with rotenone from the islands of Chaetodon lanul« (Lacepede), the latter ceased Oahu, Molokai, Maui, and Hawaii. The re­ its gill movements and elevated its operculum. sults of the examination of the gut contents The labrid inserted the anterior half of its of these specimens are given below in Table 1, body into the gill cavity through the gill along with the date of collection and the opening, where, for several seconds, it pre­ standard length and sex of the specimens. sumably searched for parasites." The same be­ It seems evident from Table 1 that L. havior was observed for the goatfish Pam­ phthirophagus feeds primarily on parasitic peneus prophyreus (Jenkins). With the use of copepods and isopods from fishes. The fish self-contained breathing apparatus in clear scales and fragment of fin membrane were water of about 40 feet in depth off Manana probably accidentally ingested in the process Island (Rabbit Island ), Oahu, ail individual of picking at parasites. The cyclopoid cope­ Labroides phthirophagus was seen to enter the pods (lichomolgids and harpacticoids) do not buccal cavity of an adult goatfish, Mulloi­ appear parasitic. They are tiny, and when diehthys samoensis (Gunther), by way of the present in the gut, they are usually numerous, mouth, such that the majority of the body of suggesting that they may have been encoun­ the labrid was lost from view. About five sec­ tered by the labrid in a convenient concentra­ onds elapsed before the Labroides emerged. tion. That parasites are not the sole diet was Frequently the act of removal of parasites further shown by the following observation by Labroides phthirophagus occurs beneath a in Kealakekua Bay, Hawaii. Several nests of ledge or in an interstice in the reef, making the pomacentrid fish Abudefdllf abdominalis observation difficult . Furthermore, the ap­ (Q uoy and Gaimard) were seen at a depth of proach of a swimmer usually interrupts the 90 feet. When the guarding adult was fright­ proceedings. For these reasons an effort was ened from one of the nests , a number of made to brin g living specimens of Labroides other fishes commenced to feed avidly on the into large aquaria where it was hoped detailed eggs, among them an adult L. phthirophagus. observations of behavior could be made. Like Labroides dimidiatus and L. bicolor, L. A total of three were caught underwater phthirophagus is nonspecific in the fishes which with dip nets in Kaneohe Bay and brou ght to it "services." Among the groups ofreef fishes the Honolulu Aquarium. None were main­ commonly seen participating in symbiotic re­ tained in the aquaria lon g enough for normal lationship with L. phthirophagus are acanthu­ feeding behavior to manifest itself; however rids, pomacentrids, scarids, labrids, chaeto­ the reactions of the resident fishes of the dontids, mullids, and plectognaths. aquaria bear mentioning. When placed in a Species of Caranx are not abundant in the tank containing many moorish idols (Zanclus Hawaiian Islands and none were observed in cornutus (Linnaeusj), one balistid (Rhineean­ association with L. phthirophagus. Only one thusaeuleatus (Linnaeusj), and an unidentified carangid, Deeapterus pinnulatus (Eydoux and large scorpaenid, the Labroides swam about Souleyer), was seen to associate with the . the aquarium, paying little attention to the labrid. As a school of this species swam over other fishes. It was evident, however, that the the bottom off Waikiki, Oahu, a single Labroides phthirophagttS swam upward to the 3 I suspect that the parasites removed from the gill school. One of the carangids left the school, cavity by Labroides are still principally calagoi d cope­ pods. In a study of the parasites of the surgeon fish swam slowly in a small circle, and quivered A canthurus triostegus sanduicensis Streets (MS dara), I slightly as the Labroides pecked at its body ob served indiv iduals of the most commo n parasitic surface. copepod of th is fish, an undescribed species of Lepeo­ pbtbeirus, freely enter and leave the gill and bucca l On one occasion when a Labroides was cavities of the fish. The genus Lepeopbtbeirus is rhe mos t tend ing a large butterfly fish of the species abu nda nt of the calago ids listed in Table 1. Review of Labroides- RANDALL 341

TABLE 1 GUT CONTENTS OF SPECIMENS OF Labroidespbtbiropbagus

STANDARD DATE LENGTH SEX GUT CONTENTS COLLECTED (MM.)

Apr. 10, '50 ...... 58 F(ripe) 2 calagoid copepods, 4 larval gna rhiid isopods, fish scales Sept . 21, '51...... 53 F 4 calagoid copepods, 25 harpacticoid copepods* M ay 11, '52 ...... 60 F 3 calagoid copepods, 1 lernaeid copepod], 88 lichomolgid copepodst D ec. 31, '5 2 ...... 42 M 2 calagoid copepods , 1 lernaeid cop epo d t 49 F 14 calagoid copepods, 1 lern aeic! cop epod], 4 larval gnathiid isopods 55 F 10 calagoid copepods, 1lernaeid copepod], 6 larval gnathiid isopods, frag ment of fin membrane 76 M 2 calagoid copepods, 14 digenetic tremato des§ J an. 21, '53...... 39 F? 1 calagoid copepod, 8 larval gnathiid isopods June 19, '53 ...... 60 F 1 calago id copepod, 19 harpacticoid copepod s* July 25, '55...... 58 F 6 calagoid copepods, fish scales Aug. 7, '55...... 63 F 14 calagoid copepods '* T isbe sp .; most species are free-living but at least one is comme nsal ( in a pelecypod ) . t Peniculus sp., generally attached to fins. t Ma ny membe rs of this famil y of cyclopoid copepods Jive commensally with tunicates, pelecypods, flatworms, etc.; how ­ ever, accord ing to Dr . Paul L. Ill g, this species of copepod does not fit well into any genus so far proposed. Alrhough no Iichornolgid has turned up as parasitic on a vertebrate, this one is sufficiently distinctive anatomically ( the arrangement of the maxi llary suggests its use as a substitute masticatory appendage ) th at it might have developed a divergent feeding mechanism. § U ndigested, pr obably a gut parasite of th e Lsbroides. moorish idols and the trigger fish were im­ This poses a question which is as yet un­ mediately 'aware of the new arrival, for they answered. Do these fishes learn the role of reacted as in the normal environment. As the Labroides through experience or is their be­ Labroides swam past individual fish, many havior innate? The answer could be provided were observed to cease swimmin g, and oc­ through observation of Labroides phthiropha­ casional moorish idols stopped their respira­ gus with aquarium-reared fishes or with fishes tory movements and raised their gill covers. taken from an area where species of Labroides The following morning the Lahroides could are known not to occur. If learned through not be foun d in the aquarium. experience, how long is the capacity for The second Labroides was added to a tank symbiotic association retained? containing two parrot fishes (Scarus sp.) about The second Labroides was left overnight in 250 mm. in length and seven surgeon fishes the aquarium, and it also was gone the next (Acanthurtls spp.). Again the Labroides paid day. It is not known whether this and the first little heed to the other fishes, but they reacted Labroides which disappeared were eaten by characteristically. The surgeon fishes swam fish or whether they escaped down the drains about nervously in a restricted area of the (which, it was later noted , were not screened). tank. The behavior of the parrot fish was al­ The third Labroides was placed in an aquar­ most ludicrous. They followed the Labroides ium containing only adulr wrasses of many aroun d the aquarium, stopping occasionally species. It was immediately chased abour the to literally stand on their tails, waggle their tank by several individuals of different spe­ pectoral fins, and seemingly to ogle the cies and ultimately eaten by one of them . labrid, as if to entreat it to perform the ex­ This was totally unexpected, for various pected services. The parrot fishes and surgeon wrasses were among the fishes which were ob­ fishes were recognized as ones which had served being serviced by Labroidesphthiropha­ been in the aquarium at least several months. gus in the reef environment. The immediate 342 PACIFIC SCIENCE, Vol. XII, October, 1958

pursuit of the Labroides in the aquarium sug­ Labroides sites are usually well separated gests that the wrasses might have been from one another. This first became evident conditioned to rapid seizure of food items around the patch reefs in Kaneohe Bay. added to the surface of the tank, and that this Labroides phthirophagus is generally found at reaction may have superseded the one which a depth of about 5 to 20 feet at the steep would occur in the normal habitat. margins of these reefs rather than their shal­ When in the proximity of other fishes, low, truncate tops. The spacing between sites, Labroides phthirophagus exhibits the same un­ at times remarkably regul ar, is now linear and usual oscillatory mode of swimming as seen more readily perceived by an observer swim­ in L. dimidiatus, although not as commonly. ming around such a reef. As previously mentioned, the act of re­ Although occurring singly or in groups of mo val of ectoparasites by L. phthirophagus as many as five individuals, Labroidespbtbiro­ frequently takes place in a region of cover phagus is most commonly seen in pairs . Often such as beneath a ledge . Repeated observa­ groups of three or more contain one or more tions of one such ledge at Manana Island, juveniles in the company of a pair of adults. Oahu, revealed that a pair of this species of When a pair of adults is seen, it is natural to Labroides could invariably be found at or near expect that one would be a male and the this site . These observations continued for a other a female. There is some evidence that period of over three years. Other pairs or thi s is the case. Of the four specimens shown groups of three fish were similarly checked in Table 1 which were collected at a poison over long periods of time at the following station on December 31, 1952, two are males locati ons: Waikiki, Kaneohe Bay, Hanauma and two are females. One of the two para­ Bay, and Ke alakekua Bay, Hawaii. I conclude types which were speared on October 14, from these and lesser observations that adult 1955, at Waianae, Oahu, is a ripe female L. phthirophagus tend to remain in a restricted (standard length 64 mm.; length of ovary in area of the reef. situ 12 mm.) and the other appears to be a As with Labroides dimidiatus, it is believed male. These two fishes were found together that fishes seek out the area where Labroides and were the only two seen in the area phthirophagus occur in order to have their surveyed. parasites remo ved. The first evidence that fishes might come to a Labroides site from be­ Labroides rubrolabiatus, new speCIes yond the immediate vicinity was the sighting PI. 1 R at Manana Island of an adult manini, Acan­ thurus triostegus sandvicensis Streets, swimming HOLOTYPE: U . S. Natl. Mus. No. 164603, from at least 40 feet away direct to the a female specimen, 54 mm. in standard length Labroides ledge where it stopped and erected and 66.5 mm. in total length, collected by its fins. As one of the Labroides swam up to ] . Randall with rotenone in the lagoon at a this fish, the latter was speared and taken de pth of 20 feet next to a large coral head ashore. Macroscopic examination revealed the about 100 feet from shore one-fourth of a presence of three calagoid copepods. Seem­ mile east of the entrance to Papetoai Bay, ingly purposeful swimming by other fishes to Moorea, Society Islands, on May 3, 1956. Labroides areas has been observed. Generally PARATYPES: U. S. Natl. Mus. No. 164604, there is a con centration of reef fishes around a female specimen, 51 mm . in stan dard length, a Labroides site. I have on several occasions collected by ]. Randall with a spear 30 feet off located such a site by first noting a greater the lee reef near the boat entrance of Caroline density of reef fishes in a restricted area than Atoll (100 S. 1500 14' W.) at a depth of about could be explained by any other reason. 15 feet on February 12, 1956; Stanford Univ. A

B

Plate 1 A. Labroides phthirophagus. Holotyp e, Oahu, Hawaiian Islands. Standard length , 68 mm. B. Labroides rubrolabiatus. Holorype, Moorea, Sociery Islands. Standard length, 54 mm.

Review of Labroides- RANDALL 343

Mus. No. 48871, a female specimen, 48 mm . ing length of caudal fin); scale rows above in standard length, collected by J. Randall lateral line to base of first soft dorsal rays 3 with a spear in the Moorea lagoon just east (the upper row extending on to base of dorsal of Tareu pass at a depth of 8 feet on May 22, fin); scale rows below lateral line to base of 1956 (specimen was picking on the body of a first anal soft rays 8 (two additional rows of female Scarus sordidus ForskiH when speared); moderately large scales on base of anal fin at British Mus. No . 1956.8.14.1, a female speci­ this point); median predorsal scales 8 or 9; men with mature gonad, 45 mm . in standard snout, chin, throat, and dorsal part of head length, collected with a spear by J. Randall naked. Gill rakers (including rudiments) on within 75 feet of the locality of the May 22 first gill arch of 48 mm. paratype 15. specimen at a depth of about 10 feet on Head length 2.78 (2.67- 2.77); depth of March 29, 1956. body 3.48 (3.50-3.69) ; snout to anus 1.66 Two more specimens collected by the au­ (1.70-1.69); snout to origin of pelvic fins 2.70 thor, a male 46.5 mm . in standard length, (2.59- 2.66); snout to origin of dorsal fin 2.92 from a depth of 7 feet in the Moorea lagoon (2.74-2.91); length of dorsal fin base 1.99 near the entrance to Paperoai Bay on July 1, (2.08-2.07); length of anal fin base 3.60 1956, and a 54.5 mm . female from 15 feet in (3.89-3 .90)-all in standard length. the pass at Takaroa, Tuamotu Archipelago, Width of body at gill opening 2.70 (2.81­ Nov. 9, 1956, have been placed in the collec­ 2.75); least depth of caudal peduncle 2.16 tion of the Uni versity of Hawaii . Both speci­ (2.33-2.32); snout length 2.85 (2.91- 2.87); mens were speared, the former as it was pick­ diameter of eye 5.11 (4.82-4.98); width of ing at the dorsal fin of Zebrasoma scopas interorbital 3.81 (4.02-3.88); length of pec­ (Cuvier). No external differences could be toral fin 1.50 (1.53 and 1.53) ; length of pelvic ascertained between the one male specimen fin 1.96 (1.96- 1.94)- all in head length. and the female specimens. First dorsal spine 8.64 (9.62-9.20) in head Some proportional measurements of the length; dorsal spines progressively lon ger paratypes were difficult to make because of (filaments of interspinous membranes extend injury from spears. beyond spine tips), the ninth 3.40 (3.75- 3.58) DESCRIPTION: Based on the holotype and in head length; first dorsal soft ray 2.43 (2.42­ the smallest (45 mm. in standard length , 55.3 2.52) in head length; first anal spine nearly mm. in total length) and the largest (51 mm . half as long as second; second anal spine in standard length , 62.5 mm . in total length) about two-thirds as long as third ; third anal paratypes. Mea surements are recorded for the spine 3.63 (3.47-3.57) in head length ; first holorype, followed in parentheses by data for anal soft ray 2.18 (2.00- 2.11) in head length. the small and large paratypes, respectively. Caudal fin slightly rou nded . Meristic data are the same for the holotype Width of mouth (rictus to rictus) equal to and paratypes (with the possible exception of eye diameter. Center of upper lip to rictus median predorsal scales which are obscure about 5 in length of head. Distance between and difficult to count), hence only a single lobes of lower lip about one-third diameter number is given for each count. of eye. Dorsal fin rays IX , 11; anal fin rays III , 10; A pair of curved canine teeth (the uppers pectoral fin rays 13 (the uppermost rudimen­ mo re curved than the lowers) anteriorly in tary, the next unbranched); pelvic fin rays each jaw, about one-fourth eye diameter in I, 5; principal caudal rays 14. Lateral line length. Upper canines fit inside lowers when scales 28, the last enlarged and bearing a jaws are closed (upper and lower canines on prominent lateral line tube (scales extend one side separated by a distance slightly less posterior to lateral line nearly half the remain- than diameter of base of one of the teeth ). A 344 PACIFIC SCIENCE, Vol. XII, October, 1958 large semicircular mass of fused reeth in cen­ midlateral and middorsal bands chartreuse ; ter of upper jaw just posterior to upper ca­ head below midlateral band chartreuse, shad­ nines. This mass is as long as the upper ca­ ing to pale blue ventrally; lips edged in bright nines with which it is in contact basally. It is red; abdomen pale blue shading to dull pale indented anterocentrally. An inconspicuous yellow posteriorly; dorsal fin orange basally, band of small teeth may be seen anteriorly in hyaline distally (only tips of interspinous the lower jaw. A canine tooth at angle of jaw. membranes are hyaline whereas the outer half Color in alcohol of Moorea specimens: of the soft portion of the fin is clear); anal fin light greenish brown, shading anteriorly on divisible by color into three lengthwise bands, head to light bluish gray, with a black band light blue basally, orange in the middle, and running from mouth through eye on to body; hyaline distally ; paired fins pale except for on the body the band broadens as it passes dark edges of pectoral rays. posteriorly and becomes progressively indis­ The brightness ofthe life color ofthe holo­ tinct until, in the median part of the body type had already faded when the photograph beyond the tip of the outstretched pectoral of Plate 1 was taken. The orange of the body fin, it is barely discernible or absent; pos­ and especially the yellow green of the head teriorly it reappears, becoming black on were more brilliant when the fish was alive. caudal peduncle and caudal fin (which is The specimen from Caroline Atoll, al­ completely black except for upper and lower though exhibiting the same general color in edges and the posterior rounded corners alcohol, is more melanistic than the ones from which are white); a broad black middorsal Moorea. The median portion of the body is band runs posteriorly from tip of snout and dusky, almost blackish, and the white mar­ gradually lightens until it disappears at rear gins of the caudal fin are narrower and re­ base of spinous portion of dorsal fin; a di­ stricted to the outer half of the fin. agonal dusky streak from chin across cheek The life colors were approximately the below eye to lower base of pectoral fin (ob­ same. The color on the head was noted in scure in 45 mm. paratype); spinous portion life as dusky iridescent green; the middle of of dorsal fin dusky basally, pale distally; anal the body was dusky orange and the abdomen fin and soft portion of dorsal fin pale with a dusky yellow; the narrow band on the cheek dark horizontal streak anteriorly located about was blackish orange; the edges of the caudal two-thirds the distance from base to margin fin were hyaline violet; the lower lip was of fins; paired fins pale except for dark edges bright red and the upper mottled with red; of pectoral fin rays. the dorsal and anal fins were orange basally The following color note from life was and hyaline distally, and the paired fins pale made of the holotype: body from region of with a purplish hue . end ofpectoral fin to caudal peduncle orange ; One pair of L. rubrolabiatus at Caroline caudal peduncle and caudal fin (except for Atoll was seen in company with a juvenile upper and lower edges which are pale-blue which was iridescent blue with a broad mid­ violet) jet black; a black band passing from lateral black band . One of another pair snout through eye, broadening on to an­ showed intermediate coloration- the anterior terior porti on of body and merging gradually half was bright blue with the black median with dusky orange on median part of body; a band, and an orange-brown area was just similar black band from snout to anterior appearing in the middle of the body. third of back where it also merges with the L. rubrolabiatus is the smallest of the four orange; a narrow blackish line from slightly species of Labroides. Indi viduals seen under­ ventral and posterior to rictus to lower base water rarely appear even slightly larger than of pectoral fin; narrow band on head between the largest specimens collected. An excep- Review of Labroides - RANDALL 345

tionally large one, estimated at between 60 tus (Riippell) via the mouth. and 70 mm . in standard length, was sighted The stomach of the paratype from Caroline on the top of a patch reef in the lagoon Atoll was opened and found to contain one at Takaroa. larval gnathiid isopod. The 45 mm . paratype This species was first seen by me during a from Moorea was similarly examined; this short stay at Caroline Atoll. It was abundant fish had eaten nine larval gnathiid isopods. in the clear water off the lee reef, but was not In the Society Islands and Tuamotus the seen in the lagoon (which lacks a pass). It species is not as common as L. dimidiatus or was often seen in pairs, and commonly two L. bieolor. It occurs outside the reef and in of the little labrids picked rapidly over the lagoons, usually at depths greater th!ln 15 body of a larger fish simultaneously, paying feet, and in areas of good circulation with special heed to the head . Only a slight tend­ much live coral and clear water. ency toward the oscillatory swimming as de­ At the atoll of Tikahau, Tuamotu Archi­ scribed for L . dimidiatus was noted. It was pelago, outside the reef just north of the pass, seen to peck at the bodies of numerous parrot L. rubrolabiatus was seen at a depth of 120 fishes (Searus spp .) and surgeon fishes (Cteno­ feet, in the company of a single L. dimidiatus. ehaetus striatus and Aeanthurus spp .), the No specimens were collected from Tikahau or groupers CephalopholisargusBloch and Schnei­ from Nuku Hiva in the Marquesas, where der, and Plectropomus leopardus (Lacepede) the species was also observed underwater. . (3;.7 feet in length), the giant wrasse Cheilinus Whether the species will be found to be undulatusRiippell (about 4 feet in length), the restricted to southeast Oceania or whether it goatfish Pampeneus ehryserydros (Lacepede), will eventually be taken farther to the west and the lurjanoid Monotaxis grandoeulis (For­ remains for further collections to reveal. skill). Later at Tahiti and Moorea the list of The species is named rubrolabiatus (Latin species seen in association with Labroides rubro, combining form of ruber, red ; labiatus, rubrolabiatus was extended such that it en­ lipped) in reference to the color of the lips compassed nearly all the families of reef in life. fishes in the area. ACKNOWLEDGMENTS L. rubrolabiatus has been observed entering the gill cavities of parrot fishes. This was first I wish to express my sincere thanks to Dr. seen at the entrance to the pass of the atoll of William A. Gosline of the Uni versity of Ha­ Takaroa in the Tuamotus. The labrid inserted waii for generously providin g specimens for the anterior half of its body through the gill the analysis of stomach contents of the Ha­ opening of Searus harid Forskal. The second waiian species and for paratype s of this spe­ observation was made just outside the barrier cies. Thanks are also due Dr. Thomas E. reef on the southeast coast of Moorea at a Bowman, United States National Museum, depth of 10 feet. About a 40 mm . Labroides Washington, D.C., and Dr. Paul L. Illg, Uni ­ was seen to pick several times at the right gill versity of Washington, Seattle, for assistance opening of Searus vermieulatus approximately in the identification of crustacean fish para­ 250 mm. in standard length. The parrot fish sites taken from the stomachs of species of angled its left side downward and elevated its Labroides; and to Dr. W. Klausewit z, Senck­ gill covers. The labrid slipped into the gill enberg Museum, Frankfurt, Germany, D r. D. chamber and was completely lost from view W . Strasburg, Pacific Oceanic Fishery Investi­ for about two seconds. On two occasions gations, Honolulu, Dr. E. Trewavas, British outside the reef at Takaroa L. rubrolabiatm Museum, London, Dr. 1. Eibl-Eibesfeldt, was seen to enter the buccal cavity of adult Max Planck-Institut, Buldern, Germany, Dr. groupers of the species Epinepbeius / useogutta- A. H. Banner, Uni versity of Hawaii, Hono- 346 PACIFIC SCIENCE, Vol. XII, October, 1958 lulu, and D r. F.A. Chace, J r., United States Parasitismus und andere besondere zwi­ N ational M useum, Washington, D.C., for schenartliche Beziehungen tropischer M ee­ pertinent information. resfische. Zeitscbriftf ur Tierpsychologie 12(2) : 203-219, 22 figs . REFERENCES FOWLER, H . W . 1928. The Fishes of Oceania. Bernice P. Bishop Mus., Mem. 10. i- iii + BARNARD, K. H . 1927. A monograph of the 540, 49 pls., 82 text figs . marine fishes of South Africa. Part 2. So. - - - 1945. Fishes from Saipan Island, M i­ African Mus., Ann. 21: 419-1065, 20 pls., ronesia. A cad. Nat. Sci. Phila., Proc. 97: 14 tex t figs. 59- 74, 19 figs. BEAUFORT, 1. F. DE. 1940 . The Fishes of the - -- 1946. A collection of fishes obtained Indo-Australian Archipelago. Vol. 8, xv + in the Riu Kiu Islands by Captain Ernest 508 pp., 56 figs . E.]. Brill, Leiden. R . Tinkham., A.U.S. A cad. Nat. Sci. Phila., BLEEKER, P. 1851. Bijdrage tot de kennis der Proc. 98: 123-218, 76 figs . ichthyologische fauna van de Banda­ --- 1949. The Fishes ofOceania, Supplement eilanden. N atuurk. Tijdschr. v. Nederland. 3. Bernice P. Bishop Mus., Mem. Vol. 12, Indie 2: 225- 261. No.2. 186 pp. - -- 1856 . Beschrijvingen van nieuwe of FOWLER, H. W ., and B. A. BEAN. 1928. Con­ . weinig bekende vischsoorten van Menado tributions to the Biology of the Philippine Arch­ en Makassar grootendeels verzameld op ipelago and Adjacent Regions. U . S. Natl. eene reis naar den Molukschen Archipel in M us ., Bul. Vol. 100, No.7. i-viii + 525, het gevolg van den Gouverneur-Generaal 49 pls , D uymaer van Twist. Soc. Sci. Indo-Neerland., GUNTHER, A. C. 1862. Catalogue ofthe Fishes A cta 1: 1-80 . in the British Museum. Vol. 4, xxi + 534 pp. - -- 1862. A tlas Lcbtbyologique des Indes Taylor and Frances, London. Orientales Neerlandaises Publie Sous les Aus­ --- 1881. Fische der Siidsee, M us. Godef­ pices du Gouoernement Colonial Neerlandais. froy, Jour. 7( 15): 129-256, 57 pls. T . T. Scaroides et Labroides. 168 pp., 48 pls. HARDENBERG, ]. D . F. 1936. Some new or Fred. M uller, Amsterdam . rare fishes ofthe In do-Au stralian Archipela­ BREDER, C. M . 1946 . An ana lysis of the de­ go. V. Treubi« 15: 367-378. [Buitenzorg .] ceptive resemblances of fishes to plant HUBBS, C. 1., and 1. C. H UBBS . 1954. D ata on parts, with critical remarks on protective the life history, variation, eco logy, and re­ coloration, mimicry and adaption. Bing­ lationships of the kelp perch, Bracbyistius ham Oceanog. L ab. Collect., Bul. 10(2): 1- 49, f renatus, an ernbiorocid fish of the Cali­ 3 figs ., 2 pls, fornias. Calif Fish and Game 40( 2) : 183­ BROCK, V. E. 1954. A preliminary report on a 198, 1 fig. method ofestimating reef fish populations. KAMOHARA, T. 1952. Additions to the fish J our. Wildlife M angt. 18(3): 297- 308, fauna of Tosa, Japan. Kochi Univ., Nat. Sci., 1 fig. Rpt. 2: 1-10, 3 figs. CUVIER, G ., and A. VALENCIENNES. 1839. LONGLEY, W . H., and S. F. HILDEBRAND. Histoire N aturelle des Poissons. Vol. 13, xvii 1941. Systematic Catalogue of the Fishes of + 370 pp., 20 pls, Pitoi s-Levrault, Paris. Tortugas, Florida. Tortugas Lab., Papers, D o TY, M . S., and ]. P. E. M ORRISON. 1954. No. 34 (Carnegie Inst, W ash ., Pub . 535). Interrelationships of the organisms on i- xiii + 331, 34 pls . Raroia aside from man . Atoll Res. Bal. 35: RANDALL, ]. E. 1955a. Stethojulis renardi, the 1-61,9 figs . adult male of the labrid fish Stetbojulis EIBL-EIBESFELDT, 1. 1955. Uber Symbiosen, strigiventer. Copeia 1955 (3): 237. Review of Labroides - RANDALL 347

- -- 1955b. Fishes of the Gilbert Islands. Central News Agency Ltd., Cape Town. Atoll Res. Bul. 47. i-v + 243. - - 1955. The fishes of Aldabra . Part IV. REGAN, C. T. 1908. Report on the marine Ann. and M ag. N at. Hist, Ser. 12, 8: 928­ ,fishes collected by Mr. ]. Stan ley Gardiner 937, 1 pl., 1 fig. in the Indian Ocean. Linn. Soc.L ondon, - -- 1957. List of the fishes of the family Trans. 12(3): 217-255, pls. 23-32. Labridae in the western In dian Ocean with RUPP ELL, E. 1835. Fische des rothen Meeres, new records and five new species . Rhodes IN Neue Wirbelthiere zu der Fauna von Abys­ Univ. Ichthyol. Bul. 7: 99-114, 2 col. pls., sinien gehOrig. Pp . 1-148, pls. 1-33. 4 figs. SAVILLE-KENT, W. 18930 The Great Barrier SZIDAL, 1. , and A. NAN!. 1951. Las remoras Reef of A ustralia. 387 pp o, 48 pls., 16 col. del Atlantico Austral ca n un estudio de su pls, W. H . Allen and Co. , London. nutricion natural y de suo parasitos (Pisc. SMITH, ]. 1. B. 1949. The Sea Fishes of Southern Echeneidae). Rev. Inst. Invest. M us. Argent. Africa. xvi + 550 pp ., figs., 103 col. pIs. Gene. Nat. Zoo/. 2(6) : 385-417, 14 figs.