Population Structure,Reproductive Behavior and Protogynous Hermaphroditism in the Angelfish

Japanese Journal of Ichthyology 魚類学雑誌 Vol.25,No.11978 25巻1号1978年

Population Structure,Reproductive Behavior and Protogynous Hermaphroditism in the Angelfish. Centropyge interruptus at Miyake-jima,Japan

Jack T.Moyer and Akinobu Nakazono Received January 14,1978)(

Abstract.Population structure,reproductive behavior and protogynous sex inversion of the angelfish Centropyge interruptus was studied in the waters of Miyake-jima,Japan.A single male dominates a harem of from one to four females,with the size of the harem being dependent upon the nature of the substrate.Females are controlled by aggressive herding by the single male.Six action patterns associated with maintenance of dominance by the male and with courtship are described.These include(1)rushing,(2)circling,(3) soaring,(4)mutual soaring,(5)nuzzling,and(6)spawning.Spawning usually takes place almost every day at approximately sunset during the period from May to October. Removal of the male resulted in sex change by the ranking female.Histological examination of the gonads confirmed field observations of protogynous sex inversion.The histology of the gonads of females,males and intermediate individuals is described.

Studies of reproductive strategies and popu- representatives of the various angelfish genera lation structures of pomacanthid fishes are may eventually provide useful tools for clas- scarce.Lobel(1975),Shen and Liu(1976)and sification of this difficult family. Suzuki et al.(1977)have reported on various The population structure and reproductive aspects of reproduction in the genera Centro- behavior of Centropyge interruptus were studied pyge and Genicanthus.Most of these studies near the Tatsuo Tanaka Memorial Biological have been based mainly on aquarium observa- Station(TMBS)at Miyake-jima(34•‹05'N, tions and do not cover all aspects of repro- 139'30'E),one of the Izu Islands,from early ductive behavior.Population structure re- July,1976,to the end of the spawning season mains unknown for any of the Pomacanthidae in October,1977.Histological studies of

(Ehrlich,1975). gonads were made at the Fishery Research Since the taxonomic revision of the angel- Laboratory of Kyushu University.This paper fishes by Fraser-Brunner(1933),considerable reports the results of these investigations. controversy has appeared in the literature concerning the classification of angelfishes. Methods and materials Yasuda and Tominaga(1970)united Genicanthus The social behavior of Centropyge interruptus with Holacanthus;however,Randall(1975) was investigated daily from July to September, preferred to follow Fraser-Brunner(1933)in 1976-77.Observations were limited to 2-3 separating them into two genera.Centropyge dives per week from September to June in interruptus(Tanaka)was long erroneously both years.More than 300 hours of observa- identified as C.fisheri Snyder.Holacanthus tions in nature were made using SCUBA. interruptus was resurrected by Tominaga and Social interactions were recorded while lying Yasuda(1973);however,Randall and Wass on the bottom within the territory of the (1974)placed the species in Centropyge, social group under study.Individual fish partly on the basis of feeding habits.We were not marked,but could be identified by have followed Randall and Wass(1974)as a location,individual variations in color pat- result of observations reported in this paper. terns,and individual behavior.Water tem- Detailed studies of both the reproductive peratures were read from commercial diving units and courtship behavioral patterns of guages.Forty specimens ranging from 86 to

•\ 25•\— 魚類学雑誌 Japan.J.Ichthyol.25(1),1978

150 mm in standard length were taken with and embedded in paraffin.They were cut screen nets and multi-pronged spears.They at 7p in cross section and stained with were fixed in 10%formalin and measured. haematoxylin and eosin.Their structure and The gonads of all specimens were removed sex were examined under a microscope.All

Fig.1.Map of the study site at Igaya Bay, Miyake-jima,showing relative locations, dimensions,and depths of Centropyge

interruptus territories of the males of Harems A•`O.Note changes in A and

B between July and August,1976,and

between 1976 and 1977(explanation in text).Harems C•`O remained the same

in both years.

•\ 26•\ Moyer and Nakazono:Behavior of Centropyge interruptus fish specimens are deposited at the Tanaka Table 1.Population structure of Centro- Memorial Biological Station(TMBS)and pyge interruptus at Igaya study site, gonad specimens are at the Fishery Research Miyake-jima,Japan,1976 and 1977. Laboratory,Kyushu University.

Results

Centropyge interruptus is a protogynous hermaphrodite that lives in a harem social structure with one dominant male and from 1 to 4 females per reproductive unit.The size of the harem and the intensity of territorial defense is directly influenced by the habitat. Habitat and population structure.Centro- pyge interruptus occurs in a variety of habitats at Miyake-jima.All habitats are dominated by boulders and/or lava cliffs.Although there are often outcroppings of corals within a particular anglelfish's territory,C.interruptus, unlike other members of the genus,is not a coral reel fish.In fact,in areas of extensive Acropora corals,the species is scarce or lack- ing altogether. Three distinct habitat types were noted,i.e. (a)tunnels and arches,(b)volcanic cliffs,and (c)large boulders.The density of populations of Centropyge interruptus appears to be directly influenced by the nature of the habitat. Crevices,cracks,and holes along the ceiling and walls of tunnels and arches provide maximum shelter,and population densities are

greatest at such locations.The main center increased from 9 to 18 m,with shallow points of activity of one male and his harem of at the cliff-top varying considerably between four females at the Abe Tunnel was restricted 5 and 12 m(Fig.1).Including in this area to an area of about 5m•~3m,directly below were a maximum of 15 harems in 1976 and the ceiling of the tunnel,which consisted of 14 harems in 1977.The population structure a complicated network of crevices and cracks. within each harem is shown in Table 1,with This tunnel,a dark cavern 21m•~5m•~14m territorial sizes and respective locations ap- and ranging in depth from 16 to 20 m,con- pearing in Fig.1. tained five harems totaling 22 fish(5 males, Sexual dichromatism.Previously unde- 17 females).Volcanic cliffs are also heavily scribed sexual dichromatism in this species eroded into crevices,holes,and caverns,and, makes sexual discrimination obvious at dis- like arches and tunnels,support dense popula- tances up to 10 m or more,even by the tions of C.interruptus.Boulder habitats human observer(Figs.2,5).The opercular provide less cover and display the lowest region of females is primarily orange,flecked population densities,with the reproductive with small blue dots.The posterior soft unit usually a monogamous,but unstable dorsal and anal fins are dark blue.Males pair. have heavy blue lines in the opercular area, Our chief study site was situated along an making the head appear blue from a distance eroded volcanic cliff;running seaward for of several meters.The posterior soft dorsal 190 m outside the fishing port of Igaya(Fig.1). and anal fins are streaked with broad,irregular Depths at the foot of the cliff gradually bright blue bands on a black background,

•\ 27•\ 魚類学雑誌 Japan.J.Ichthyol.25(1),1978

easily visible from 10 m or more. Fish species with colors resembling C. Territoriality.Males vigorously defend interruptus were often momentarily rushed. their territorial boundaries against neighboring Such encounters apparently resulted from males.Boundaries are easily defined by a mistaken identity,suggested by the fact that particular boulder,crevice,or lava pinnacle, the rushing male would suddenly stop several and are seldom violated.However,occasional- meters from the object of his attack and ly during courtship,a male will enter another return to feeding.Presumably,the rushing male's territory to rush(Fig.4A)a female male mistook the other blue species as a from the neighboring harem.Such intrusions member of his harem.Rather than territorial are met by aggressive chasing by the defend- defense,rushing of this nature is related to ing male.Visiting females from neighboring the establishment and maintenance of domi- harems are rushed and circled(Fig.4B)but nance,and will be discussed below.Species are rarely driven from the territory. rushed in this manner included the damselfish Sex-distinguishing colors play an impotant Pomacentrus coelestis Jordan et Starks,adults of role in courtship behavior;but,in addition, the wrasse Coris gaimardi(Quoy et Gaimard), they appear to elicit aggression between neigh- and the angelfish Centropyge bispinosus boring males.Encounters between rival (Gunther). males at territoral borders often result in With the exception of C.bispinosus,which broadside posturing,head to tail,fins ex- appears a velvet blue at depths of 10 m or tended to clearly show sex-distinguishing color more,other congeners were tolerated.Various patterns. male C.interruptus in the study site shared Females respond somewhat agonistically their territories with C.tibicen(Cuvier),C. toward outside females,but no more so than ferrugatus Randall et Burgess,and C.vrolicki toward other female members of their own (Bleeker). harem.Encounters with females from neigh- Establishment of dominance and sex inver- boring territories are relatively rare,and sion.Observations in the field revealed the occur only when feeding brings them into relationship between sex inversion and the close proximity to each other.Then the social system of this species.On Oct.10, home female rushes and circles the visitor, 1976,aquarium hobbyists collected the male invariably causing the outsider to flee to the and one of four females from Harem A and safety of familiar shelter holes.This cannot both females from Harem B(Fig.1,Table 1). be interpreted as territorial defense by the By Oct.17,the largest of the remaining three home female,due to the fact that identical females at Harem A was aggressively rushing behavior is displayed by dominant(larger) and circling the smaller females.Blue flecks females against subdominant(smaller)females on the operculum were changing into lines. within the same harem. By Oct.22,the soft dorsal and anal fins Interspecific agonism seems to be directed were in typical male color pattern,and the only at the pomacentrid,Eupomacentrus altus opercular region was distinctly more blue. Okada et Ikeda.Aggressive chasing of the This fish was collected on Oct.23,and the damselfish by the male angelfish is common, gonads were removed for histological examina- and often interrupts spawning.Although the tion.Twenty-four hours later,the largest pomacentrid is forced to seek cover from of the remaining two females had begun to attacks by male C.interruptus,the situation rush and circle the smaller female.However, is often reversed when the damselfish chases by Oct.29,the male from Harem B,who a female member of the angelfish harem, had lost his entire harem,had extended his especially if the female is small.Every C. territory to include most of A,establishing interruptus male under observation shared part dominance over the remaining females in the of his territory with at least one E.altus. A harem. Whether the aggressive nature of this rela- Histological examination indicated that the tionship is related to food habits,competition gonads of the dominant A female observed for shelter,or both,is not yet known. in male courtship behavior were transitional

•\ 28•\— Moyer and Nakazono:Behavior of Centropyge interruptus

Fig.2.Sexual dichromastism in Centropyge interruptus.Top:Functional female.Center:Transi- tional male.Botton:Functional male.

•\ 29•\ 魚類学雑誌 Japan-J.lchthyol.25(1),1978

Fig.3.Photomicrographs of the gonads of Cei opyge interruptus.Scales sho 100Fe.A:Ovary TMBS 770509-10).Arrow shows the thickened part of the ovarian membrane.B:Higher( magnification of a thickened part of ovarian membrane(TMBS 761016-2).00,oocyte:OM, ovarian membrane;LS,labyrinth-like structure.C:Ambosexual gonad(TMBS761023-1). Notice spermatogenic cells(arrows),scattered oocytes and the large area of the labyrinth- like structure(LS).FOC,former ovarian cavity.D:Arrows showing the occurrence of the same cells as those of the labyrinth-like structure in a mature testis(TMBS 770706-1).

(Fig.3C).in this case,the breeding season (TMBS 760926•`9 760926---12)were taken the had been over for about three weeks. previous September. Two fish that had formerly been females That sex change is socially controlled,with were collected on July 6,1977,in the peak the highest ranking fish changing from a of the breeding season.One of them,a male functional female to a male is obvious from (TMBS 770706-1.Table 2,Fig.3D),had the above observations.Dominance competi- been the ranking female in the harem of the tion continues within the harem throughout male,TMBS 770528-2(Table 2)that had been the year,especially if the sizes of the male collected 39 days previously.Similarly,TMBS and the ranking female are similar.Harem 770706-2(Table 2)was found to be a perfect E displayed such a situation,with agonistic male only 20 days after the death of the encounters,including aggressive chasing by former male,a large fish of 145-150 mm in the male,often delaying spawning by the standard length that was speared and lost, ranking female until several minutes after the badly wounded,never to be seen again. other females had spawned. Another possible example of sex-inversion After the breeding season,open wounds can caused by collection of the dominant male often be seen directly below the dorsal fin of is TMBS 770509-9(Table 2),a small male of both the male and the ranking female,indicat- 112 mm in standard length,collected in May, ing a continuing struggle for dominance. 1977,from the same location that four fish Fighting behavior was only rarely observed,

•\ 30•\— Moyer and Nakazono:Behavior of Centropyge interruptus

Table 2.Body length and sex of specimens of C.interruptus arranged in size order.The sex of all specimens were examined histologically.,female;F/M,ambosexual defined by Reinboth,1970);6 ,male. (

and although biting below the dorsal fin was males were found only in larger size ranges, noted,it is not known whether such wounds the smallest measuring 112 mm in standard were inflicted by biting or by use of the length.Histological investigations of the opercular spines. gonads of these specimens indicated sex An unusual situation existed in Harem A inversion in this species.The ovarian mem- in July,1976(Fig.1).Two males shared brane of C.interruptus is thickened in several the same harem,although usually only the places and these thick parts have an inner larger of the two was able to spawn.Spawn- cavity(Fig.3A).In this cavity,cells abound ing by both males on the same evening was in the cytoplasm forming a labyrinth-like observed only once,and frequently one or structure(Fig.3B).Spermatogenic cells are the other was interrupted during courtship usually not seen when the ovarian part is display by the aggressive chasing of his rival. functional.However,upon sex inversion, It is possible that collecting by hobbyists at the number of cells in this structure increases the shallow end of the territory,which shrunk and the spermatogenic cells appear among greatly after the disappearance of the smaller these tissues(Fig.3C).The cells in the male,accounted for this rare situation.The labyrinth-like structure were observed even smaller fish disappeared after a few days. in a mature testis as walls of vasa diferentia Histology.The relationship between body and also as supporting tissues for testicular length and sex in 40 specimens of C.inter- lobules(Fig.3D).The above mentioned ruptus is further evidence of sex inversion histological results show that individuals of in this species(Table 2).Smaller fish were C.interruptus are protogynous hermaphrodites, all females with the largest measuring 133 in which the ovarian part matures initially, mm in standard length.On the other hand, with undeveloped testicular parts on the

•\ 31•\— 魚類学雑誌 Japan.J.Ichthyol,25(1),1978 periphery,and when the ovarian parts disap- circling her once,and then moving rapidly pear,the testicular parts begin to mature. on.Rushing(Fig.4A)seems to be elicited Maintenance of dominance and courtship by the blue and orange coloring of the female behavior.As noted above,females are not (see"Territoriality").When the patrolling bound by territorial borders,although fa- male sees a female,he rushes directly at her, miliarity with the home territory tends to swimming at a high speed.They circle each limit their centers of activity.During the other,head-to-tail,abdomens facing each other reproductive season,females may sometimes (Circling:Fig.4B).The male then swims be attracted into neighboring territories by the rapidly away in search of another female, courtship of the neighboring male.For ex- while the female continues to feed. ample,in 1977,the No.2 female from Harem Actual courtship begins in most harems D occasionally moved into Harem E to spawn about 30 min before sunset with"soaring" with the E male(Table 1).Therefore,in by the male(Fig.4C,5A).Soaring continues addition to his own survival,reproductive until about 10 min before sunset,when mutual success of a particular male is dependent to soaring,nuzzling,and spawning follow in a great extent upon his ability to maintain sequence. dominance over his females and to keep them During the soaring phase,the male extends within his territory. all of his fins,leans 45•‹•`90•‹ to one side,

Six easily distinguishable action patterns are fully displaying his sex-identifying fin and involved in the spawning sequence:(a) opercular markings.The soft dorsal and anal rushing,(b)circling,(c)soaring,(d)mutual fin color pattern is especially clear at such soaring,(e)nuzzling,and(f)spawning(Figs. times(Fig.5A).The male remains motion-

4,5).The first two of these,i.e.,rushing less during soaring,except for occasional and circling,are aggressive in nature,and stabilizing movements of his pectoral and therefore cannot be considered as true" court- caudal fins.The behavior of specific male is ship behavior".However,because these consistent.Some males,e.g.C and G,soar actions maintain the male's dominance over very high in the water column,often several his harem in addition to controlling the meters above their mates.Others,such as centers of activity of his females,they are D,consistently soar within 1-2 m of the sub- essential prerequisites to courtship and must strate,directly above the female.After soar- be included in the spawning sequence. ing for a few minutes over a particular

Although both rushing and circling have female,the male may depart to search for been occasionally observed from morning to another. mid-day,normally the male spends most of In the average harem,spawning activities his time early in the day feeding with his start about 10 min before sunset when females females on algae from the volcanic boulders begin to seek out their male,who in turn at the base of the cliff or moving 3•`4 m into may be searching for a receptive female. the water column with his harem to join Attracted by the conspicuous soaring by the heterotypic aggregates that include Chromis male,a female approaches him and soars in flavomaculata Kamohara,C.weberi Fowler, front or beside him(Fig.4D,"mutual soar- C.miyakeensis Moyer et Ida,and Franzia ing").This elicits nuzzling(Figs.4E,5B) squamipinnis(Peters).At such times,the by the male,who attempts to move under angelfish feed on the feces of the numerous his mate and place his snout against her vent. damselfishes and anthiids. Usually nuzzling does not immediately follow Rushing and circling of females by the mutual soaring,but is preceded often by male increases in frequency and intensity in several minutes of" positioning".Females the early afternoon,reaching a peak in most are obviously disturbed by the body contact harems 2.5 hrs•`30 min before sunset.During associated with nuzzling,and frequently move this time,the male stops feeding and swims away,again and again.Finally the pair rather rapidly around the borders of his ter- reaches the nuzzling position,which may be ritory,rushing each female as he sees her, held from 2 to 18 sec.At a particular mo-

•\ 32•\ Moyer and Nakazono:Behavior of Centropyge interruptus

A B

C D

E F

Fig.4.Action patterns of Centropyge interruptus during the spawning sequence.A:Rushing. B:Circling.C:Soaring.D:Mutual soaring.E:Nuzzling.F:Spawning.

•\ 33•\— 魚類学雑誌 Japan.J.Ichthyol.25(1),1978

Fig.5.Various stages of the spawning sequence of Centropyge inrerruptus.A:Male at Harem 0 soaring above a female.B:Male at Harem E(below)nuzzling just prior to spawning. C:Spawning.Note cloud of gametes.D:The E male spawns with one female while the next female soars.Note sex distingushing color patterns in each photo. ment,the male rapidly opens and closes his females have spawned. mouth,accompanied by a quick fluttering of Of 175 spawning observed,most occurred be- the pectoral fins.Spawning follows quickly. tween 10 min before and 5 min after sunset The male dashes toward the substrate im- on clear evenings with good visibility.On mediately after release of gametes.followed dark,overcast or rainy evenings.or when closely by the female(Figs.4F.5C.5D).A water visibility was poor,swawning times short"after-chase"follows,with the male were much earlier,indicating a definite rela- swimming in rapid pursuit of the female, tionship between spawning and light intensity occasionally soaring momentarily again. (Table 3).Individual harems were consistent Spawning time and sites.Females do not in their spawning times(Table 3). spawn in a fixed order,dependent upon Spawnings usually occurred in rapid suc- dominance.The highest ranking female may cession within a harem.For example,on or may not spawn first,dependent upon July 13,1977,all four Harem E females chance locations of individual harem mem- spawned with their male within 8 min.The bers as the short spawning period begins. following evening,the four E females and Frequently dominance competition between the second-ranking D female all spawned with the male and the ranking female may delay the E male within six minutes.On July their spawning until after lower ranking 15,the same five females spawned with the

•\ 34•\ Moyer and Nakazono:Behavior of Centropyge interruptus

Table 3.Spawning times in relation to time of sunset.

E male over a period of eight minutes.Other microlepidotus Lacepede,both species being harems behaved similarly. ignored under normal circumstances.Spawn- Like spawning times,spawning sites were ing is hindered under such conditions for also consistent.For example,the A harem. two reasons:(1)Since the male remains usually spawned about 20 m away from the close to the substrate,his soaring is not cliff and high in the water column.The C, clearly visible to the females in his harem; D,and I harems also moved 10-20 m from and,(2)females are easily disturbed during the cliff,but spawned fairly close to the sub- the nuzzling stage,even under the best of strate.C and G spawned high in the water conditions.The added caution caused by column and close to the cliff,while E spawned cloudy water means that frequently spawnings within 1,-2 m of the top of the lava pinnacle are interrupted in the nuzzling stage by the that made up that territory(Fig.1). close approach of harmless schooling fishes. Frequency of spawning.Although court- Darkness comes rapidly on such occasions, ship began with intensive rushing,circling, and females will retreat to cover for the night and soaring by the male when water tem- before spawning.

peratures reached 21•Ž in late April,1977, The possibility was considered that females,. no spawning was observed until May 28,when inhibited from spawning due to interruptions waters rose to 22•Ž.Except for April 24, by potential predators,might spawn at the when the temperature reached 22•Ž,temper- following dawn.On two occasions,females. atures of 20•Ž or less inhibited courtship until from Harem E were purposely frightened and May 21,when soaring was observed again at prevented from spawning during the usual 21•Ž.A week of varying temperatures in evening hour.Observations beginning before late July,1977,indicated that spawning was sunrise the following morning revealed that, inhibited when temperatures plunged below although Centropyge interruptus is one of the 22•Ž and that no courtship occurred in waters last diurnal fish to take cover in the evening, colder than 21•Ž.Spawning ended in the first its morning activities begin relatively late, week of October in both 1976 and 1977, i.e.after Franzia squamipinnis and various although soaring by males continued for an Chromis spp.,and consist of individual feeding other two weeks.Water temperatures did not with no social interaction. drop below 22•Ž until December in both The moon is a factor in the spawning of seasons. many fishes(Allen,1972;Moyer,1975;and Water clarity seems to be crucial to maxi- others).Spawning of all females within mum spawning.All fish,male and females, individual harems and the total number of remain close to shelter in cloudy water,and spawnings observed were analysed in relation quickly flee for cover at the approach of to moon phases.Our results showed that harmless fishes,such as large schools of the spawning occurs in equal intensities at all surgeonfishes Naso lopezi Herre and Prionurus phases of the moon.

•\ 35•\ 魚類学雑誌 Japan.J.Ichthyol.25(1),1978

dominate all other members of its species Discussion within its range of activity,changes from a Habitat and population density.The im- female that spawns only once in a given portance of shelter and an intimate know- evening into a male that spawns with each ledge of the bottom topography within their of his females.By changing sex,it has territories has been stressed for members of demonstrated a level of genetic fitness above the genus Centropyge(Ehrlich,1975).Our the average member of the population,be- observations suggest that the more fractured cause it is contributing more to the gene and eroded the substrate,the greater the pool of the population than are any of the density of populations of C.interruptus. fish it dominates(see Emlin and Oring,1977, Food does not seem to be a major limiting for definition of"fitness").If this fish can factor in population densities.There appears dominate and therefore spawn with more to be sufficient algae to provide for crowded females than can neighboring males,it has populations.The fact that C.interruptus demonstrated an even higher level of fitness territories are frequently inhabited by grazing by contributing more than its rivals to the Eupomacentrus altus and often by other total gene pool.In order to maintain do- species of Centropyge in addition to the C. minance and therefore spawning rights over interruptus harem is evidence that there is the females in its range of activities,it ample food for dense populations.Lobel's patrols the perimeter of this area,keeping (1975)observations of Centropyge potteri out neighboring males,permitting entrance (Jordan et Metz)are similar with respect to by any neighboring females,and forcing its sufficiency of food supply.Fecal materials own females to remain within its center of from plankton-eating pomacentrids and anthi- activities,which therefore become,by defini- ids seem to provide an important percentage tion,a territory.Though differing in some of the diet of C.interruptus,and may con- respects,e.g.synchrony of female spawnings, tribute in a minor way to increasing popula- the harem social structure of C.interruptus tion densities along cliffs where upwelling serves as an example in a marine environment currents attract enormous heterotypic ag- of"female defense polygyny"as defined by gregates of such plankton-feeders.However, Emlen and Oring(1977),with the availability our observations indicate that relative abun- of adequate shelter being an unevenly distri- dance of shelter holes seems to be more buted critical resource that provides the crucial to population densities in this species environmental potential for polygamy". " than food supply. We have observed similar harem social Dominance,sex inversion and territoriality. structures in Centropyge vrolicki,C.heraldi, Territory itself seems to be entirely the result C.fiavissimus,C.ferrugatus,C.sp.,and C. of a male's attempt to maintain dominance tibicen,although we as yet have no histological over,and therefore spawning rights with the evidence of sex inversion(Table 4).A nearly females in his harem.By rushing and circl- identical reproductive system has been de- ing each of his females from the perimeter scribed by Robertson and Choat(1974)in the of his territory,he tends to herd them back labrid fish Labroides dimidiatus(Valeniennes). into the center of the,territory.This both In this species,harems are established and a prevents females from moving within visual strict hierarchy is maintained based on the range of the neighboring male's courtship dominance of a single male over a group of display and forces all members of his harem females.Romoval of the dominant male into a relatively small area,allowing for a results in sex change by the ranking female. better chance of all females and the single Population structure,reproductive behavior, male finding each other during the short and the problem of phylogenetic relationships, spawning period. Fricke(1973,1976)suggests monogamy for The selective advantages of such a repro- most Chaetodontidae,within which he in- ductive strategy can be explained as follows: cludes the pomacanthids.In his only direct An individual C.interruptus that is able to reference to population structure of a po-

•\ 36•\ Moyer and Nakazono:Behavior of Centropyge interruptus

Table 4.Harem reproductive units in six species of the genus Centropyge.

macanthid species,he reports that the Car- cation)reports a harem social structure in ribean angelfish.Pomacanthus paru(Bloch) Holacanthus trimaculatus Lacepede at Guam. usually lives in pairs(Fricke,1976).Our Randall(1975)suggested sex inversion within observations of Pomacanthus imperator(Bloch) the genus Genicanthus,and this was later in the Bonin Islands and P.semicirculatus demonstrated in aquarium studies by Shen (Cuvier)in the Yaeyama Islands are in agree- and Liu(1976)for Genicanthus semifasciatus ment with Fricke and suggest that the re- (Kamohara)and by Suzuki et al.(1977 and productive unit within the large angelfishes in press)for G.lamarck(Lacepede).Our of the genus Pomacanthus may be monogamous limited field observations of G.semifasciatus pairs.That harems do in fact occur from in the Yaeyama Islands and at Miyake-jima time to time cannot be ruled out at this time. suggest a harem social structure similar to Yasuda(personal communication)reports a C.interruptus. harem-like social structure in the genus Knowledge of sequential behavioral patterns Chaetodontoplus. during courtship and spawning among the Lobel(1975 and in press)reports that pomacanthids awaits further research;how- Centropyge potteri lives in pairs or in small ever,there is some evidence to support a groups in Hawaiian waters,and that the high degree of similarity between species population structure appears to be related to within the family.In addition to our ob- the habitat.Although he refers to"one servations of C.interruptus,we have observed or two males and several females"(Lobel, rushing and circling in C.vrolicki,C.fer- 1975)and to"harem spawings"(Lobel,in rugatus,C.tibicen,C.sp.,and Pygoplites press),the nature of C.potteri's social struc- diacanthus(Boddaert). ture remains unclear.As noted above,we A photograph appearing both in Hiyama have observed harems dominated by single and Yasuda(1971)and in Burgess and Axelrod males in Centropyge vrolicki,C.ferrugatus, (1972)shows a male Genicanthus lamarck C.heraldi,C.flavissimus,C.sp.,and C. displaying above his mate in a posture tibicen(Table 4).Shepard(person communi- identical to soaring in Centropyge interruptus,

•\ 37•\— 魚類学雑誌 Japan.J.Ichthyol.25(1),1978 and Randall(personal communication)reports of TMBS and New College,Florida,for a similar display by G.caudovittatus(Gunther) preparing Figs.1,2 and 4,and to Dr.Hiroshi in the Red Sea.A posture apparently identical Tsukahara of Kyushu University for sugges- to nuzzling in C.interruptus has been described tions and continued encouragement.Contri- for Centropyge bispinosus,C.fisheri,and C. bution number 124,Fishery Research Laborato- flavissimus by Lobel(1975 and in press),by ry,Kyushu University,and number 17,TMBS. Hioki(personal communication)for Genican- thus lamarck,and by Suzuki,et al.(in press) Literature cited for G.semifasciatus. Allen, R. R. 1972. The anemonefishes: their Such limited data make comparative analysis classification and biology. TFH Publications, of behavioral patterns impossible.However, Inc., Neptune City, N. J., 288 pp., 40 figs. some observations seem significant.Centro- Burgess, W. and H. R. Axelrod. 1972. Pacific pyge interruptus,C.ferrugatus,and C.sp.all marine fishes, vol.1. TFH Publications, Inc., display identical sex-distinguishing blue and Neptune City, N. J., 280 pp., 489 figs. black stripes on the soft dorsal and anal fins, Ehrlich, P. R. 1975. The population biology of coral reef fishes. Ann. Rev. Ecol. Syst., 6: unlike sex-identifying features of other angel- 211•`247. fishes,suggesting close relationships between Emlen, S. T. and L. W. Oring. 1977. Ecology, these species(personal observation).Sexual sexual selection, and the evolution of mating dichromatism was previously known in systems. Science, 197 (4300): 215•`223, figs. pomacanthids only in the genus Genicanthus 1•`2. (Randall,1975),and differs in pattern from Fraser-Brunner, A. 1933. A revision of the the Centropyge spp.mentioned above.Un- chaetodont fishes of the subfamily Pomacanth- fortunately,behavioral studies are still un- inae. Proc. Zool. Soc. London, pt. 3: 543•` available for most angelfishes. 599, figs. 1•`28, pl. 1. Fricke, H. W. 1973. Behavior as a part of Acknowledgments ecological adaptation: In situ studies in the coral reef. Helgolander Wiss. Meeresunters, Many people aided in this study.We 24: 120•`144, figs. 1•`18. particularly express our appreciation to John Fricke, H. W. 1976. Bericht aus dem Riff. Piper W.Shepard,Katherine A.Meyer,Scott Verlag, Munich, Germany, 254 pp., 87 figs. Dolginow,and Lynn B.Jenkins of TMBS, Hiyama, Y. and F. Yasuda. 1971. Living fishes and to Dr.Hitoshi Ida of Kitasato University of the Japanese coastal waters. Kodansha, and Mr.Yutaka Yogo of Kyushu University. Tokyo, 337 pp., 527 figs. (In Japanese). Mr.Katsumi Suzuki,Mr.Syo Hioki,and Mr. Lobel, P. S. 1975. Hawaiian angelfishes. Marine Kazuhiro Iwase of the Science Museum,Tokai Aquarist, 6(4): 30•`41, figs. 1•`5. University,Dr.Fu jio Yasuda,Tokyo Uni- Lobel, P. S. (In press). Diel, luner, and seasonal versity of Fisheries,and Dr.John E.Randall, periodicity in the reproductive behavior of the Bishop Museum,Hawaii,generously shared pomacanthid Centropyge potteri and some other reef fishes in Hawaii. research observations with us.Phil S.Lobel, Moyer, J. T. 1975. Reproductive behavior of

Harvard University,provided an important the damselfish Pomacentrus nagasakiensis at manuscript prior to publication.Many people Miyake-jima, Japan. Japan. J. Ichthyol., 22(3):

at the Marine Laboratory,University of Guam, 151-163, figs. 1•`5.

aided us in our research there.Especially, Randall, J. E. 1975. A revision of the Indo- John Eades and John Shepard helped collect Pacific angelfish genus Genicanthus, with de- important specimens for comparative analysis. scriptions of three new species. Bull. Mar. Sci., Mr.Katsumi Suzuki and Dr.Yoshiaki To- 25(3): 293421, figs. 1•`17. minaga read the manuscript and made im- Randall, J. E. and R. C. Wass. 1974. Two new portant suggestions.Dr.Yasuhiko Taki, pomacanthid fishes of the genus Centropyge Tokyo University of Fisheries provided con- from Oceania. Japan. J. Ichthyol., 21(3): 137 144, figs. 1•`2. siderable assistance.A very special thanks Reinboth, R. 1970. Intersexuality in fishes. In

to Kerry C.Sussex of TMBS for many hours Hormones and the environment. Mem. Soc. of field observations,to Patricia A.Hadley Endor., 18: 515•`543, figs. 1•`2.

•\ 38•\— Moyer and Nakazono : Behavior of Centropyge interruptus

Robertson, D. R. and J. H. Choat. 1974. Proto- long-tailed pomacanthine fishes from Miyake-

gynous hermaphroditism and social systems in jima and Okinawa-jima, Japan. Japan. J. labrid fish. Proc. Int. Coral Reef Symp. 2nd., Ichthyol., 14(4): 141-151, figs. 1•`13. 1: 217•`225, figs. 1•`4. (JTM : Tatsuo Tanaka Memorial Biological Station, Shen, S. C. and C. H. Liu. 1976. Ecological and Ako, Miyake-jima, Tokyo 100-12, Japan; AN: morphological study of the fish fauna from the Fishery Research Laboratory, Kyushu University, waters around Taiwan and its adjacent islands. 46-12, Tsuyazaki-cho, Munakata-gun, Fukuoka- 17. A study of sex reversal in a pomacanthid ken 811-33, Japan) fish, Genicanthus semifasciatus (Kamohara). Sci.

Rep. Nat. Taiwan Univ., 6 140•`150, figs. 三宅島におけるレンテンヤッコの社会構造,産卵行動と雌性先熟性転換 1•`8. Suzuki, K., K. Hioki, and K. Iwase. 1977. Jack T.Moyer・中園明信

Suisonai de kansatsu sareta chochouo-ka gyorui 東京都下三宅島において,レンテンヤッコの社会構 no sanran-seitai. I. Tatejima-yakko (Genicanthus 造,産卵行動を潜水観察するとともに,性転換につい lamarck) no sanran to shoki seikatsushi. Ab- て調べた。本種の雄は,普通雌数尾よりなるハレムを stract 269, Annual Meeting of Japanese Society 持っているが,雌の尾数は生息基盤の性質によって異 of Scientific Fisheries: 71. (In Japanses). なり,1～5尾の範囲にあった。本種の産卵は,5月 Suzuki, K., S. Hioki, and K. Iwasa. (In press). ～10月にかけてほぼ毎日日没前後に行なわれたが ,雄 Spawning behavior, early life history, and sex の雌に対する優位性の維持と求婚に関して(1)rush- reversal in Genicanthus lamarck in an aquarium. ing,(2)circling,(3)soaring,(4)mutualsoaring, Suzuki, K., S. Hioki, and Y. Tanaka. (In press). (5)nuzzling,(6)spawningの6型の行動型を認めた。 Spawning behavior, early life history, and sex 本種のハレムより雄を取り除くと,次の順位にある reversal in Genicanthus semifasciatus in an 雌が性転換して雄へと変った、さらに,組織学的に本 aquarium. 種の雌性先熟性転換を確認するとともに,雌,雄およ Tominaga, T. and F. Yasuda. 1973. Holacanthus び移行型の個体の生殖腺構造を記載した. interruptus, a valid pomacanthid species, dis-

tinct from Centropyge fisheri. Japan. J. Ichthyol., (Moyer:100-12東京都三宅島阿古富賀農園田 20(3): 157•`162, fig. 1. 中達男記念生物実験所;中園:811-33福岡県宗像郡津 Yasuda, F. and T. Tominaga. 1970. Two new 屋崎町九州大学農学部附属水産実験所)

―39―