contrasts in social behavior between central american cichlid fishes and coral-reef surgeon fishes

AMER. ZOOL., 14:9-34 (1!)74) .

Contrasts in Social Behavior between Central American

Cichlid Fishes and Coral-reef Surgeon Fishes

GEORGE W. BARLOW

Department of Zoology and Museum of Vertebrate Zoology,University of California, Berkeley, California 94720

SYNOPSIS. The social systems and related behavior of cichiid and surgeon fishes are com-pared in terms of (i) physical spacing, (ii) theoretical spacing (communication), (iii)castes, (iv) group composition, (v) open versus closed groups, and (vi) reproductivebehavior.

Cichlids only recently invaded Central America. Despite the occurrence of about 100species there, most are in one genus, Cichlasoma. Yet, they express a spectrum of feed-ing behavior, ranging from grazing herbivore through omnivore to predator, each ofvarying degrees of specialization. In contrast, their social behavior is remarkably con-servative. There is a tendency for the generally found division of labor, with the femaledoing more of the direct caretaking of the eggs or fry and the male more of the defense,to lead toward polygyny. This is counterbalanced by the need for both parents to de-fend the fry. Communication is most accessible through a study of color patterns. Whileseemingly diverse, there is a common plan that entails the use of some or all of thesame vertical bars and their central spots, and the appearance of yellow orange, red,or black ventrally.

The coral-reef community is one of the oldest in existence. Surgeon fishes are pan-tropical, especially in the Pacific Ocean, and have developed distinct generic groupingswithin a compact family of about 75 species. Most are herbivorous, with some morespecialized than others. The species fall into guilds, within which there is broad overlapin diets. The social systems differ radically, both when breeding and when not, and canbe understood as consequences of their strategies of obtaining food.

Wickler's classification of reproductive types within the Cichlidae is shown to be noadvance over the previous dichotomy of substrate and mouthbreeding species. Postercoloration in surgeon fishes is apparently as important, or more, in extraspecific thanintraspecific aggression, and poster-colored surgeon fishes show pronounced rapid colorchanges when fighting intraspecifically.

INTRODUCTION creating an awareness of the relativelyprimitive understanding we have of fish

An important service rendered by sym- s o c i a l s y s t e m S i n relation to their environ-posia is the identification of areas where m e n t T h i s s t a n d s i n s t a r k c o n t r a s t t 0 t h e

work is needed. This paper will have made prOgress made in recent years in studiesa contribution if it does nothing beyond o £ t h e s o c i a l s y s t e m s of b i r d s a n ( j m ammals

(e.g., Crook on weaver birds, 1964, andSo many individuals have helped in ways large primates, 1970; Estes, 1969, On wildebeest;

and small over the years that they are too numerous Pitelka et al., 1973, On shorebirds; and Tin-to list, but my thanks go to ail of them The manu- b e r g en , 1959, and his colleagues on gulls),script was kindly reviewed by J. R. Bayhs and K. R. ° ° ° 'McKaye. The cichlid work has been generously sup- While there has been no shortage of be-ported by the National Science Foundation, most havioral studies on fishes, most of it eitherrecently through grants GB 13426 and GB 32192X. u u c J • u» i_

N.S.F also funded in part the research on the h a S b e e n fragmented within groups, Or hassurgeon fishes. I am pleased to acknowledge sub- concentrated on certain species. Thus, thestantial support from the Oceanic Institute of Ha- comparative Studies On cichlid fishes bywaii, through its then Director K. S. Norris. The Wickler (e.g., 1966) and in the field byfield trip to Eniwetok was arranged by P. Helfrich, LOWe-McConnell (e.g., 1969) have been Op-Director of the Eniwetok Marine Biology Labora- . . , , ' . . r

tory, with funds provided by the Atomic Energy portunistic rather than programmatic. Ap-Commission. felbach (1969) concentrated on the genus

at Indiana University L

ibrary on May 2, 2014

http://icb.oxfordjournals.org/D

ownloaded from

http://icb.oxfordjournals.org/

10 GEORGE W. BARLOW

Tilapia although he studied but 4 of 11species intensively. The emphasis has beenon reproductive behavior, mostly as seenin the laboratory. (To date, the book byFryer and Illes [1972] has not been avail-able to me.) Obviously, the study of socialsystems in fishes calls for observing all typesof social behavior, ideally in nature. Pro-grammatic studies of this type are nowbeing pursued by Myrberg (e.g., 1972) onpomacentrids and by Ernest Reese onchaetodontids. Okuno (1963) has attempteda synthetic overview.

Being terrestrial, man is ill equipped andreluctant to invade the aquatic realm toobserve fishes there. This attitude has evencarried over to easily observed stream andlakeshore species. With the advent ofSCUBA it has become possible to watchfishes underwater, even at moderate depths.Unfortunately, the observer is usually lim-ited to 1 to 2 hr per dive, and generally tonot more than two dives per day. Watertemperatures are such that even with gooddiving suits a human being experiencesa serious heat loss when lying still. Andpreparation for a dive often consumes con-siderable time. This, plus the basicallyhostile environment, makes underwaterobservations unlikely, though not at all im-possible, between about 5:00 PM and 9:00AM, when much of the most interestingbehavior is occurring. Finally, even thosewho have become good divers have beentrained mostly by ichthyologists who havenever been underwater. Consequently, theproblems inherited by students have usuallybeen formulated by a terrestrially boundprofessor with more conventional views offish biology.

Students embarking on a study of fishbehavior would be best advised to readbeyond the fish literature, studying par-ticularly the recent work done on birds andmammals that lies at the interface of be-havior and ecology. These ideas can betested on fishes and reformulated. Fishes,after all, are the oldest, most diverse, mostspecies-rich, and, in some instances, the mostobservable of all vertebrates. The ability ofthe observer literally to fly about in their

environment creates possibilities unheardof in the terrestrial realm.

In what follows I attempt to bring to-gether some of my thoughts on the socialsystems of freshwater cichlid fishes (Cich-lidae) and marine surgeon fishes (Acan-thuridae). I have been working on cichlidsfor about 14 years, but mostly in the labo-ratory. During the last 7 years I have madea number of field trips to Central Americawhere my students and I have observedcichlid fishes underwater, primarily inNicaragua. We have also watched thesefishes in Panama, Costa Rica, El Salvador,and British Honduras.

The involvement with surgeon fishes ismore recent but more intense and overtlycomparative. I spent 7 weeks at KealakekuaBay, Hawaii, in 1971, and then 3 weeks atEniwetok Atoll with Ken McKaye in 1972.Prior to my interest in this group I madeincidental underwater observations on theirbehavior in places ranging from the Gulfof California through the Galapagos Islandsand numerous Pacific Islands from Oahu tothe Philippines.

Two important points need to be madeabout the work that follows. It is a prema-ture summary with relatively little docu-mentation. Often there will be noncon-trasting information on the two families offishes, and there will be gaping holes in thestory. This summary should be regarded asmuch a proposal of hypotheses as a declara-tion of a state of affairs. I anticipate thatmany of the conclusions here will be re-placed as our knowledge grows.

BACKGROUND INFORMATION

Cichlidae

Cichlids are the most successful percoidfishes in fresh water. They are especiallyspecies-rich in Africa and South America,and also occur in tropical Asia. In thoseareas they are a well-established componentof the mature fauna and have reasonablywell defined genera; often there are manylocally sympatric species.

Apparently the cichlid fishes invadedCentral America recently from South Amer-




ownloaded from


BEHAVIOR OF CICHLID AND SURGEON FISHES 11

ica as the Panamanian Isthmus emerged,probably during the late Pliocene (Myers,1966; Miller, 1966). At the time of theirentry, Central America had a depauperatefish fauna dominated by poeciliids, that is,the mollies, swordtails, and their allies(Myers, 1966). The advancing cichlidsradiated into the different water systemsgiving rise to nearly 100 species. Reflectingtheir recency, almost all the species havebeen placed in Cichlasoma, although severaldifferent lines obviously exist within thatgenus. Three other distinct genera, withbut six nominal species, have been de-scribed.

Geographical isolation has been a de-cisive factor. Each water system is separatedfrom the next by a land barrier, restrictingthe movements of both young and adults.One consequence is the apparent re-evolu-tion of similar types in separate drainagesystems. While some species are widely dis-tributed, there is a high degree of endemism.In no water system is there an excessivenumber of sympatric species; the modalnumber of species appears to be around 10.The extensive Rio Asumacinto system con-tains 44 known species, but there is endem-ism within its boundaries (Miller, 1966).Thus, each faunal assemblage in each riversystem may be thought of as a replicateexperiment with some endemic and a fewubiquitous species.

One of the most well-studied aspects ofcichlid biology is reproductive behavior,probably because of the ease with whichthey breed in aquaria and their well-devel-oped parental care. Generally, the male andfemale form a pair, prepare a nest site, andspawn. The eggs are then tended andfanned, as are the larvae that are kept inprepared places. Protective care continuesfor some weeks after the fry have becomefree swimming.

A major variation on this theme is thedevelopment of mouthbreeding in at leastfour different lines; these are representedby a number of genera that can be groupedwith Tilapia and with Haplochromis inAfrica, and by Geophagus and by Aequi-dens in South America. The other variation

is polygyny, a harem society with one maleand several substrate breeding, parentalfemales, shown in the genus Lamprologus(Wickler, 1965) in Africa and Apisto-gramma (Burchard, 1965) in South Amer-ica. Both mouthbreeding and polygynousspecies were derived from species that hadjoint parental care. Furthermore, mouth-breeding in the New World genera is lesswell developed than in the African forms.

Fundamental to understanding the adap-tiveness of the social system of any animalis a knowledge of its feeding behavior. Un-fortunately, there is little precise informa-tion available from the cichlid fishes inCentral America. Mostly, it is based onwatching the fish feed in nature, oftenunder circumstances that preclude a confi-dent knowledge of what is being eaten.Nonetheless, some differences are obvious,such as those between the herbivores andpiscivores. Feeding adaptations, moreover,are correlated with morphological differ-ences and with the habitat in which thespecies occurs. Hence the feeding habitscan also be judged from morphology andoccurrence.

The fry of all cichlid fishes start as carni-vores. Most feed on plankton at first butsoon take benthic microfauna as well. Theyoung of many species in the New Worldnibble mucus from the parents' body.Virtually nothing is known about this typeof feeding behavior in nature, however,except for the observations made by us inNicaragua (Noakes and Barlow, 1973).

Within the juvenile and adult cichlidsof Central America one finds almost allgeneral feeding types. The most commonlyoccurring type is the omnivore. These fishesmay eat Aufwuchs, plants, invertebrates,and fishes. Within this category there willdoubtless prove to be varying degrees ofspecialization on different kinds of inverte-brates, for example, snails, as well as cich-lids that take more of one kind of materialthan of another. Most species in this groupare middle sized, relatively deep-bodiedfishes with small to modest gapes such asCichlasoma citrinellum, C. beani, C. macu-licauda, C. octojasciatum, and C. cyano-




ownloaded from


12 GEORGE W. BARLOW

C. meeki C. rostra turn

FIG. 1. Eight species of Cichlasoma plus Neetroplus Color patterns are from live specimens; the C.nematopus. The relative sizes have been estimated, labiatum was totally red.




ownloaded from



guttatus (see Fig. 1 for some of the speciesmentioned here).

Many species are largely herbivorous,such as the small monotypic Herotilapiamultispinosa, the middle sized Cichlasornasynspilum, and the large C. tuba. Thesespecies, nonetheless, can be omnivorous.

A much smaller group is represented bythe piscivorous large bass-like species suchas Cichlasoma dovii, C. managuense, C.motaguense, C. friedrichsthalii, and Peteniasplendida. These species take a variety offishes, including other members of theirown family, but prefer unarmored fishessuch as atherinids.

Another group, perhaps more numerousthan the piscivores, are the substrate sifters.Many omnivores sometimes sift the sub-strate for food, as does C. citrinellum, butsome species are specialized in this regard.Included here are members of the subgenusThorichthys (the firemouth group, e.g.,C. meeki) and C. longimanus. These aresmall- to medium-sized species with rela-tively elongate faces. The most extremedevelopment is seen in C. longirostris; thismedium-sized fish has a long pointed snoutthat it plunges into the bottom; it bearsan uncanny resemblance to South Americacichlids of the genus Geophagus.

Close to the sand sifters is the distinctiveCichlasoma nicaraguense (usually labelledas its junior synonym C. spilotum byaquarists). It occurs in mixed rock and sandhabitats. Its feeding behavior is unknown,although it probably takes algae, detritus,and invertebrates while engaging in sandsifting or scraping Aufwuchs. It is a mod-erate-sized species with a distinctivelyrounded shape to its head. In many ways,its morphology is intermediate to the nextcichlid, Neetroplus.

One of the most highly specialized feed-ing types is that of the Aufwuchs scraperNeetroplus (three nominal species). Thissmall, slender species has a down-turnedmouth and can be seen rasping the Auf-wuchs from rocks. On occasion, however,they can be carnivorous. For instance, theyfollow under schools of spawning atherineseating the falling eggs. They also prey on

the fry of other cichlids.The next feeding type is one largely in-

ferred from morphology. Cichlasoma labi-atuin is a slender species of medium sizethat has enormous puffy lips. These evi-dently act as a gasket when the mouth ispushed into crannies (Baylis, personal com-munication), facilitating the extraction ofdetritus and small invertebrates.

The cichlids of Central America thuspresent a broad spectrum of feeding adap-tations. But these adaptations are not pro-found, for the herbivores will feed ascarnivores when suitable prey are presentedto them. Even the most carnivorous specieswill readily take inanimate laboratory food.Nonetheless, the different feeding adapta-tions are reflected in the size and bodyshapes of the fishes. And while the mostcommon feeding type is the omnivore,specializations exist. The most highly spe-cialized of these have given rise to the threegenera, other than Cichlasoma, with butfive species.

Acanthuridae

Surgeon fishes occur throughout the trop-ical seas where coral or rocky reefs providethe appropriate shallow-water habitats. Toan ichthyologist they are a refreshinglycompact family, consisting of about 75 spe-cies in six well-defined genera. While theyare an advanced group of percoid fishes,they are members of the oldest continuallyexisting community, that of the coral reef(Newall, 1971). They must, therefore, inter-act and compete with large numbers ofother types of fishes.

There is some element of instability,however, since they have pelagic eggs andlarvae that are variously vagile. The morewidely distributed species must, in par-ticular, be sufficiently adaptable to copewith the complex of species where theyfind themselves, for once the specializedAcronurus larva has descended to the reef,it is tied to that habitat. Open water formsa barrier to the adults.

Little is known about the spawning be-havior of surgeon fishes. I will return tothis later.




ownloaded from


14 GEORGE W. BARLOW

A.nigrofuscus A. ac hi lies

C. strigosus

FIG. 2. Ten surgeon fishes, illustrating representa- combination of preserved and live soecimens andtives of the feeding guilds (see text). The relative may not be complete or correct in all details.sizes have been estimated. Color patterns are from a




ownloaded from



In contrast to the cichlids, the feedingbiology of adult surgeon fishes is relativelywell known, at least for the species thatoccur in Hawaii (Jones, 1968). The plank-tonic larvae are presumably carnivorous.But as soon as they settle to the reef andmetamorphose, most become herbivorous.As they mature, they can be assigned toone of the feeding guilds. (Some representa-tive species are shown in Fig. 2.)

Among the species tied to the hard reefsare the detritus feeders, all of the genusCtenochaetus. Mostly these are small tomedium in size. One species, however, theHawaiian endemic C. hawaiiensis, is mod-erately large.

Perhaps the largest guild, in number ofspecies, is the reef grazers (Jones consid-ered the reef and sand grazers collectively,but they are better treated as separategroups). Most of the reef grazers are in thegenus Acanthurus. These species are smallto medium in size and seem to feed almostcontinuously by scraping at the reef withtheir mouths. There is broad overlap in thespecies of algae they consume (Jones, 1968).

Some division of resources is effectedamong the reef grazers by some of thespecies exploiting certain parts of the en-vironment more than others. For example,atop the reef flat A. guttatus advances andrecedes with the tide, as does A. triostegus.Being a smaller species, A. triostegus isoften further shoreward and less in thesurf line. At the edge of the reef flat andwhere the water surges a great deal one ismore apt to encounter A. achilles, A. glau-copareius, and A. lineatus. Just off the reefflat, but still in the surgy area, are schoolsof A. lucopareius. Just below the roughwater, and more intimately associated withthe substrate, one finds A. nigrofuscus andA. nigroris. Here, and moving deeper, thedetritus feeders become more abundant. AtEniwetok one species, A. pyroferus, is gen-erally not seen until depths of at least 10 mhave been reached; it is not clear what itis feeding on, however. None of these spe-cies is restricted to one place in the environ-ment; it is common to find them feedingtogether, and in different areas.

Jones placed the two species of Zebra-soma within the grazing guild because ofthe species of algae found in their guts.But their long snouts suggest access to othertypes of food. The high rate of feeding,small size, and lack of mobility of Z.flavescens suggest that it is indeed a reefgrazer. But the larger Z. veliferum spendsless time feeding and it ranges over widedistances on the reef, suggesting that it isa browser (see below) not a grazer.

The next important guild is the sandgrazers. All are in Acanthurus. They tendto be large mobile species, such as A. dussu-mieri and A. mata, who refuge (Hamiltonand Watt, 1970) at selected places on thereef, or range widely among patch reefs.Their strategy is to take refuge on the reefwhen danger threatens and, when not, tomove to nearby sandy areas where theyingest sand. They have a large musculargizzard in which they remove diatoms andalgae from the sand grains (Jones, 1968).These fishes regularly move in schools ofmixed species, and each species stronglyresembles the others. While often busy withfeeding, they seem to have much timeavailable for other kinds of behavior.

Some species are intermediate betweenreef and sand grazers. Acanthurus olivaceusfits this category since it both grazes on thereef and ingests sand. It is intermediate insize, trophic anatomy, color pattern, andbehavior. Acanthurus gahm may also be ofthis type.

Browsers are surgeon fishes who feed onrelatively leafy algae that is apparentlypatchy in occurrence but quickly andeasily ingested. All spend relatively littletime feeding and have much free time.Most browsers are in the genus Naso andare large, highly mobile species. They arefrequent where reefs drop to depths, suchas off headlands, or at pinnacles in atolllagoons. Two species favor shallow water.One is a typical gray Naso, N. unicornis.The other is the atypical N. lituratus.Acanthurus bleekeri shares many featureswith these Nasos, and may be a browser.Zebrasoma veliferum, as noted, may alsobelong here.




ownloaded from


GEORGE W. BARLOW

The most exceptional feeding guildwithin the surgeon fishes is that of theplankton feeders. This type of behaviorhas appeared in at least two differentgenera. Acanthurus thompsoni is a smalltypical Acanthurus except for its moreterete shape. It feeds in schools in deeperwater, say 10 to 20 m, off the face of cliffsand pinnacles. Naso hexacanthus is a largeNaso that is said to feed on plankton(Jones, 1968). As is typical of this genus,it is a large highly mobile species. It has,however, a more streamlined shape, andthe pair of knives on each side of its caudalpeduncle have been reduced to roundedscutes.

These herbivores, then, may be splitinto guilds within which there is broadoverlap in diet. The chief consequencesof membership in a guild are size andmobility, time free from feeding, and ex-tent to which competition exists.

The detritus feeders and reef grazers aremostly small busy feeders with time forlittle else. The reef grazers also competewith many species. All these fishes tend toremain in a relatively small area, althoughsome of them, such as the reef-flat invaders,roam considerably.

At the other extreme are the browsers,large mobile species who feed quickly andhave much open time. Competition be-tween the species seems minimal.

The sand grazers are in many ways inter-mediate between the reef grazers and thebrowsers. They are large, but generally notas large as the browsers, and they are mo-bile. They seem to have an intermediateamount of free time, and they often traveland feed in mixed species groups. There isprobably no competition for sand since itis available in boundless amounts if thefish dare range far enough from the pro-tective reef. To do this, they join companywith other species feeding in the same way.

SOCIAL SYSTEMS

Anyone having experience with the re-cent literature in social systems will beaware that the term means different thingsto different investigators. For that reason,

I will briefly outline what I consider theessential dimensions. Some will not be ap-plicable to certain species, and some arebetter treated together than separately, asoccurs when discussing patterns of spacingand group composition below.

1) Physical spacing: A convenient andunambiguous method of describing socialsystems is to measure the spacing betweenits units. On the one hand, this will consistof measuring the spacing between solitaryanimals or groups that can be consideredthe social units. On the other hand, dis-tance between individuals within a groupconstitutes a direct description of the socialstructure of that group. Home range andterritoriality are included here.

2) Theoretical spacing (communication):Two approaches are conventionally used,the more common being dominance-sub-ordinance relationships. The other measureof theoretical spacing or distance can be therate of exchange and the consequences ofcommunication between individuals. Ifthese are signals communicating dominanceand subordinance relationships, then theapproach is identical to the conventionaldominance-subordinance hierarchy. How-ever, communication networks extend be-yond that into the exchange of signalsbearing other types of information. In whatfollows I will be considering theoreticalspacing primarily in the context of ap-proaching and withdrawing.

3) Castes: Social systems may consist ofgroups in which all individuals are basic-ally the same, that is, of one caste, as incertain schooling fishes. Often, however,individuals differ. The two principle differ-ences are age and sex. But caste can bedistinguished even within these, particu-larly in reference to the phase of repro-ductive cycles, such as courting, parental,and so on. Taken together, age, sex, andreproductive state define an individual'scaste (sensu McBride, 1971).

4) Group composition: Social systems canbe defined by the number of individuals,by caste, of which they are composed. Forexample, one social system in cichlid fishesis the harem (one-male) group, as in Lam-




ownloaded from



prologus. Another social system, as inTilapia, consists of territorial males in agroup receiving females who come indi-vidually, thus a lek society.

5) Open versus closed groups: If animalsmay join and leave a group with relativelylittle disturbance to that group, it is saidto be open. If animals in a group resist thejoining of that group by another indi-vidual, and if the individuals in that grouptend to stay together, the group is said tobe closed.

6) Special attributes of reproduction: Toround out or make intelligible the descrip-tion of a social system, and in particular toappreciate its adaptiveness, it is often neces-sary to take into consideration the specialattributes of reproductive behavior. Insome species of cichlid fishes, for instance,there would be no difference in social sys-tem as defined in the foregoing. Yet onespecies might be a substrate breeder, andthe other a mouthbreeder. Sometimes, too,it is more convenient to consider attributessuch as open versus closed (pair bonding)and communication in the context of re-producing individuals as opposed to thosewho are not.

Cichlidae

Physical spacing. In all species so far in-vestigated, the fry form a dense school ofclosely spaced individuals which stays closeto the parents. As juveniles, they tend toform aggregations, but information is scantyhere. The nonbreeding adults of all speciesare also inclined to congregate, but thereare detectable differences in this behavior.

The small species living in well-articu-lated environments, such as among rockslides or submerged branches, congregatewhere such environments exist, then spacethemselves there. They rely on the environ-ment for protection and do not form con-spicuous groups, although they do some-times move about in groups. Cichlasomanigrofasciatum, as one example, can some-times be observed moving in groups whilefeeding. Individual Ncctroplus, as another,are generally well spaced while associatedwith the bottom. However, I have seen

them in dense schools, numbering perhaps10,000 individuals, hovering over rockyoutcrops at depths of 10 to 15 m in LakeJiloa, Nicaragua. McKaye (personal com-munication) has observed them moving ingroups over the bottom while feeding.

The medium-sized species that live withinan articulated environment are more mo-bile and more apt to travel in schools. Butwhen not moving, they generally spreadout in weakly defined groups or as indi-viduals.

Medium-sized species found in sandy oropen areas, such as over beds of Chara, areapt to maintain group cohesion both whilemoving and feeding (e.g., C. longimanus,C. nicaraguense, and C. maculicauda). How-ever, some of these species also tend eitherto be isolated (C. rostratum), or to move inthe company of other species. Occurringover open bottom, therefore, tends to pro-mote continual cohesion, even to the extentof interspecific associations.

The large piscivorous species are inclinedto live well spaced, solitarily or in pairs, asseen particularly in C. dovii and C. mana-guense. Even these may form small schoolsor groups as adults (Petenia splendida, per-sonal observation; C. dovii in rivers ofCosta Rica, Meral, personal communica-tion; and C. dovii in Lake Jiloa, Baylis,personal communication).

In no case is there adequate informationto make conclusive statements about homerange. Incidental observations suggest, none-theless, that some individuals remain in thesame general area for a period of at leastsome days or weeks.

Territoriality in nonbreeding adults isgenerally transitory and difficult to recog-nize. It is most commonly seen in one fishfeeding on the bottom who, through aggres-sion, maintains space around it free of otherindividuals. At least one species may havea cave or crevice which has been dug outand in which an individual may seek refugewhen threatened with danger (C. citrincl-lurri); however, no defense of this retreathas been seen, nor obvious avoidance byanother potential occupant.

Large predators such as C. managuense




ownloaded from


18 GEORGE W. BARLOW

and C. dovii may hold territories as pairsthe year around. It is not clear whetherthey actually do, and whether they are goingthrough one breeding cycle after another(Bleick, 1970). Adult C. dovii are said notto hold territories in the nonbreeding sea-son in Lake Jiloa (Baylis, personal com-munication).

Theoretical spacing. Dominance-subordi-nance hierarchies are easy to demonstratein the laboratory for a variety of CentralAmerican cichlids. This problem has notbeen explored in nature, however, since nofishes were marked so they could be recog-nized as individuals. However, I doubt theoccurrence of stable dominance hierarchiesin nature because groups change their com-position so readily; they appear to be opengroups, not closed. Dominance-subordi-nance interactions, in contrast, are obviousin the field.

It is more profitable to treat these ap-proach-withdrawal relationships in thecontext of communication. In so doing, itis necessary to consider the three plausiblesensory modalities.

It is now known that some communicateacoustically (Myrberg et al., 1965; Schwarz,unpublished). These signals consist ofpulsed low-frequency sounds, emitted mostlyduring aggression, whether between rivalsor within pairs. While it is too early toassert with confidence, acoustical signalsseem to show little differentiation betweenspecies or between the sexes within the spe-cies after size differences have been takeninto account.

There is even less information about theuse of chemical signals. It is known, none-theless, that some cichlid fishes can recog-nize not only larvae or fry of their ownspecies, but that they can discriminate be-tween the odor of their own young andthat of other young of their own speciesand of the same age (Kiihme, 1963; McKayeand Barlow, unpublished). Since in somany species there is virtually no sexualdimorphism, other than the size relation-ship after pairing, it is likely that sexualdiscrimination is chemical. I also suspectthat chemicals may be the most important

signal for species recognition in reproduc-tive behavior.

When considering visual signals it is con-venient to divide them into three types:(i) the movements performed, (ii) the shapeof the individuals, and (iii) their color pat-terns. Considering first the relatively stereo-typed movements, called here Modal ActionPatterns or MAP's (Barlow, 1968), thesimilarity between the different species isstriking. It is possible to use the descrip-tions in Baerends and Baerends-van Roon's(1950) monograph to deal with most of thespecies. There are clearly some statisticaldifferences, however, in the frequency ofoccurrence and sequencing of the variousMAP's. And there are some patent butsmall differences in the MAP's used by dif-ferent species. For instance, Neetroplussometimes lies on its side in frontal display.And C. meeki threatens frontally withgreatly extended opercles and branchio-stegal membrane. There are also small dif-ferences between the species in the degreeto which the median fins are raised orclosed, and when. Nonetheless, the generalconclusion holds that there is no trenchantdifferentiation among the MAP's in theCentral American cichlids. In fact, theydiffer little from their African substrate-breeding counterparts.

There are obvious differences in shapesbetween the different species (Fig. 1). How-ever, the general plan of the omnivore isso widely distributed that there is conse-quently much similarity between many ofthe species. Several show sexual dimorphismin shape and size with increasing age, orjust during the breeding season: Males maydevelop a large nuchal hump, sometimestogether with swelling of the throat (e.g.,C. citrinellum, C. parma, and C. dovii). Insome species, the nuchal hump is schema-tized into an almost wart-like protuberance,as in C. nicaraguense and in Neetroplus. Insome species, such as C. macracanthum andC. nigrofasciatum, the trailing edges of thedorsal and anal fins are more protractedin the male than in the female. And infreely mated pairs, the male is invariablylarger than the female; the difference may




ownloaded from



be appreciable (e.g., C. maculicauda) orslight (e.g., Petenia).

It is when considering coloration thatthe richest material is found. There is ap-preciable divergence among the species, butit is not as profound as the initial impres-sion might convey. In particular, the blackmarkings on the body are conservative.These start on the fry as a series of blotchesalong the top and sides of the body. Withdevelopment, the blotches break up into aseries of eight vertical bars. As they arebeing elaborated, a distinct black spot de-velops in the middle of each bar. Withfurther development, the now juvenile fishdemonstrate the ability to turn on eitherthe bars or the spots, and to combine thespots into a median stripe running thelength of the body (see Fig. 25 in Baerendsand Baerends-van Roon, 1950).

The adult thus has available to it a selec-tion of spots, bars, and a stripe of varyinglength. For convenience, each bar and itsspot can be enumerated (Fig. 3).

In a large number of species, the adultshave a "neutral" color pattern dominatedby a few spots. Two are usually especiallywell developed, a small black spot at thebase of the tail, and another (often numberfive) just posterior to the center of thebody (Fig. 4).

Within a given species the pattern ofspots or bars varies (Fig. 4) according tothe behavior and environment. In general,

9 \ AjU / /

FIG. 3. Herotilapia tnultispinosa in breeding colora-tion (from Baylis, 1974). The major bars, with vis-ible central spots, are numbered one to eight fromrear to front; the head bars are also numbered.Note, too, the black ventral region.

the more the fish finds itself in open water,often schooling, the greater the tendencyto develop the row of spots into a stripe.While up in the water but over rocks, par-ticularly hovering in groups, the generalpattern is for the appearance of the mid-body and base-of-tail spots. When the fishmove closer to the bottom, often passingin and out among holes or submerged treebranches, one sees a combination of spotsand bars with softly developed edges.

Of more interest to the theme being de-veloped here is the considerable variationin deployment and number of bars and/orspots among the different species. Cichla-soma citrinellum is typical of those speciesthat develop the characteristic pattern ofone spot at the base of the tail and one spotjust to the rear of the middle of the body,plus a few others, while moving about ingroups. In contrast, C. maculicauda has apattern in which the mid-body spot is com-bined with the ventral half of its black bar,giving the fish a vertical slash in the middleof its body; there has also been a fusion ofthe tail spot with the one or two ahead ofit to create a horizontal slash at the base ofthe tail (Fig. 1). As an extreme example,Neelroplus sometimes shows weakly devel-oped barring, but the fish generally have aslaty gray body with but one vertical blackbar, number 5 (Fig. 1).

The various species seem consequentlyto be utilizing an essentially digital code ofblack spots and/or bars. Not all combina-tions are seen, and some combinations arefavored. No analysis has yet been done onthe degree to which sympatric species usedifferent patterns, and allopatric speciesthe same. One of the difficulties is thatinterspecific communication may be im-portant. For instance, the simple two-spotpattern, or longitudinal stripe, may facili-tate interspecific schooling. More criticalto species isolation are the color patternsmanifested during breeding (see below).

There is similarly a certain conservatismin the use of colors on the body as a whole.The first general conclusion is that whenyellow, orange, or red are present, it is com-monly found on the ventral surface of the




ownloaded from


B GEORGE W. BARLOW

FIG. 4. Four color patterns seen in Cichlasoma citrinellum (explained in text).

body. Red is especially characteristic ofeyes. Those species that have an appreciablearea colored yellow, orange, or red are gen-erally found in more turbid waters. Huesof these longer wave lengths penetratemurky water better than do blue or green.An excellent example of a yellow fish inmurky water is Herotilapia, which occursin swamps (Baylis, 1974). Another is theappearance of totally yellow, orange, or redmorphs in C. citrinellum and C. labiatumin the perpetually murky Great Lakes ofNicaragua. On the other hand, C. salviniis brilliantly yellow when breeding yetoccurs in clear waters in British Honduras;it is also an extremely aggressive species.Other brilliantly colored species occur inthose clear waters, such as C. synspilumwhich is a gaudy and variable combinationof orange, yellow, and red with black mark-ings. Likewise, Petenia occasionally hasbrilliant yellow or orange morphs in theclear waters of British Honduras. Suchcorrelations between coloration and thespectral properties of the water are difficultto make because of the diverse waters occu-pied by many species.

The use of blue or green tones in thebody seem to characterize those species thatgenerally occur in clearer waters (e.g., C.urophlhalmus). These species also oftenhave brilliant blue vermiculations on theface, as seen in C. dovii, C. octojasciatum,and Aequidens coeruleopunctatus. In Nica-ragua, the large C. dovii tends to occur inthe clearer bodies of water and has a de-cidedly blue and green cast, whereas itscounterpart C. managuense resides in themurkier areas and has a yellowish tone toits body.

Some species have blue or green on theirtops and sides but are orange or red ven-trally. Good examples are C. longimanus,which occurs in relatively turbid waters,and C. maculicauda, which appears bothin clear and murky waters.

Many cichlids have pearl-like flecks ontheir body or fins. This occurs in fishesthat dwell in open areas, particularly outover the sand or in midwater. Doubtlessthey are showing a partial "mirroring"(Denton and Nicol, 1966) to facilitatecamouflage. Examples are C. rostratum andC. macracanthum, and the species of Geo-




ownloaded from



phagus.Finally, I need mention the recurring

development of black areas on the ventralsurface of some species. It appears duringbreeding (see below) in a variety of species(C. spilurum, C. macracanthum, C. maculi-cauda, C. centrarchus, Herotilapia). Thisis inverse countershading, which makes theparental fish more conspicuous (Albrecht,1962; Baylis, 1974).

Breeding colors. One of the most com-monly recurring changes during breedingbehavior is the intensification of the normalpattern of vertical bars. Contrast is en-hanced by making the bars very dark withsharp edges while the interspaces betweenbars become pale (Fig. 4). There are inter-esting species differences. One species dropsthe barring on the head and the first oneon the body (C. macracanthum). Other spe-cies display the bar that connects the eyesdorsally (C. beani; Herotilapia) (Fig. 3).Yet another species emphasizes the hori-zontal body stripe (C. longimanus). Otherstrategies may also be involved. For in-stance, C. citrinellum and C. nigrofasciatumshow similar black barring when breedingand they are sympatric; however, C. citrinel-lum is a much larger species than C. nigro-fasciatum, and the female of C. nigrofascia-tum develops a large orange blotch on herside. Yet another tactic is shown in Neetro-plus: When the fish breed they merelyreverse the color pattern from a gray bodywith one black bar to a black body but nowwith the bar white.

It is again early to say with confidence,but there seems to be a correlation betweenthe use of bars versus the stripe and thephysical environment. Those species breed-ing among rocks generally wear bars. Thosebreeding more in the open favor the stripe,or some combination of spots plus part ofthe stripe.

There is also a general intensification andspreading of colors, especially yellow,orange, or red, or black, from the ventralsurface up the sides (Fig. 3). Many species,moreover, make the eye more conspicuous,particularly by producing a pale eye againsta dark face (e.g., Neetroplus, C. macracan-

thum).Thus, despite the diversity of color pat-

terns in the cichlid fishes of Central Amer-ica there is a common theme. Mostly itconsists of a simple recombining of thespots and/or bars on the body, togetherwith the manipulation of colors yellow tored, or black, ventrally, and blue or greenrlnrcallv Thprp arp of

~ - " " ~ j ' - - - - - — —, alsoj

remarkable exceptions. For instance, C.tuba displays a large patch of white whenbreeding (G. H. Meral, personal communi-cation). And many species show interestingdetails such as red margins on the fins(C. alfaroi).

Castes. There is no noteworthy differen-tiation into castes. One can distinguishage castes, such as larvae, fry, juveniles, andadults. However, the juveniles are littlemore than small nonbreeding adults. Andthere is not much to distinguish betweenadult males and females when not breed-ing. Even in breeding pairs, dimorphismis often expressed only as the male beingsomewhat larger than the female (seebelow).

Group composition. Group compositionin nonbreeding fishes has been describedin the section on spacing. Generally, groupsare open and highly variable in numbers.Males and females in nonbreeding groupsare indistinguishable in their behavior.

Reproductive behavior. In contrast to thediversity of feeding habits, habitats occu-pied, and morphological differences, thereproductive behavior of these cichlids isnoticeably uniform. The typical patternstarts with pair formation. We are leastcertain as to how pairs form in relation tothe holding of a territory. We now suspectthat either the female attracts the male toa territory which they then defend (as inC. nigrofasciatum; Meral, personal com-munication), or that the pair forms awayfrom the territory and then captures onefor themselves (C. citrinellum; McKaye,personal communication). The two fish al-ternately court in the territory, which isfor reproduction, not feeding. The eggs areplaced in a chamber. The larvae are simi-larly hidden away in cavities or pits.




ownloaded from


22 GEORGE W. BARLOW

In most species the female is clearlysmaller than the male, being about 80 to90% of his weight, and of a similar colorpattern. Females develop full breedingcoloration faster than do males, and oncehaving developed it, are less apt to re-verse it.

There is also appreciable differentiationof roles. The female does most of the directcare of the eggs and larvae, though not allof it in most species, and is more persistentin remaining near the offspring whendanger threatens. The male spends moretime patrolling the territory and drivingoff conspecifics.

When the fry become free swimmingthey form a large, coherent and ball-shapedschool. Both parents vigorously defend theseagainst predation; if the parents are chasedaway, the fry are eaten within minutes byother fishes which are constantly in at-tendance.

Not only do the parents defend the fry,but in many species they also help provide

food. In some species the parents turn overleaves for the offspring (C. nigrofasciatum;G. H. Meral, personal communication), asdo the Asian cichlid Etroplus suratensis(personal observation) and the Africancichlid Pelmatochromis guentheri (Myr-berg, 1965). When the fry are extremelyhungry in almost any Central Americanspecies they will begin to graze on themucus on the sides of their parents (C.nigrofasciatum, C. spilurum, C. macracan-thum, C. friedrichstahlii, C. beani, C. longi-manus). In C. citrinellum the response isapparent even when the fry are only mod-estly hungry (Noakes and Barlow, 1973),as is also true of C. labiatum.

The most noteworthy differences in breed-ing behavior between the different speciesare consequences of the physical environ-ment. Aequidens coeruleopunctatus breedsin streams in Panama where there is littlehard substrate on which they can deposittheir eggs. The female selects a relativelyrare rubbery leaf from among the litter on

FIG. 5. A dutch of eggs of Aequidens caeruleopunc-tata laid on a particularly rubbery leaf in the Agua

Salud, Panama.




ownloaded from



the bottom and carries it past the male dur-ing courtship. Later she spawns on theleaf (Fig. 5). When danger threatens shegrasps the leaf in her mouth and pulls itback under the bank of the stream or intoshallow water. If the leaf is turned upsidedown, she attempts to right it. This is ob-viously an adaptation both to the shortageof spawning sites and to the fact that theyspawn in streams that are subject to pe-riodical torrential flooding. Were the waterto suddenly become fast and high, thefemale could simply back into the quieterwaters on the side of the stream, and returnto the main stream as the water subsided.

The species C. centrarchus and Hero-tilapia both live in habitats where there isconsiderable vegetation. Both of these smallspecies plaster their adhesive larvae on theplants, well off the bottom. Petenia andC. synspihim breed over sand bottom indense vegetation in British Honduras. Theyexpose the roots of these plants, uponwhich they spawn. The larvae are depositedin sand pits nearby.

Cichlasoma maculicauda breeds in LakeGatun in Panama in areas where much ofthe bottom is either clay or covered withdense beds of Chara. The one nest JohnMertz and I found in clay bottom was al-most a perfect cylinder with a diameterequal to the body length of the female, andthe depth about equal to twice her bodylength. The bottom was an enlarged cham-ber with many rootlets on one wall, pre-sumably where the eggs had been laid.About half way up the hole was a smallantechamber in which the larvae had beenplaced (such a construction would simplynever have been seen in the standard aquar-ium setting).

Cichlasoma nicaraguense is one speciesthat is reasonably sexually dichromaticwhen breeding. Living in a mixed rock andsand habitat, it digs a remarkably deephole in the sand by a rock. It lays extremelylarge but few eggs that lack adhesive threads(I thank the hobbyist Dick Stratton for call-ing this to my attention). The eggs are notonly large but rather buoyant, bouncingaround in the bottom of the nest. The ab-sence of adhesive threads, large size, and

buoyancy are adaptations to prevent lossof the eggs in the sand. While the femaledoes some fanning of the eggs, she alsodoes an inordinate amount of taking theeggs into her mouth and spitting them out(Stratton, 1968). Apparently this species isbut one step away from becoming a mouth-breeder. When the fry swim, the male helpsin their defense (personal observation).

There is also a trend toward polygyny inthese cichlids. Aequidens coeruleopunctatusfemales are often found alone, caring fortheir eggs, larvae, or fry. I observed a maleand female defending free-swimming fry inwhich the male ranged up to a meter awayfrom the female, encountered another fe-male, and began courting with her. Meral(personal communication) has made similarobservations from one population of C.nigrofasciatum in Costa Rica. The oftenconfirmed clearly stronger parental responsein the female of most of the Central Ameri-can cichlids suggests that the potential forpolygyny is wide spread.

The general pattern of reproductive be-havior within the cichlid fishes, in conclu-sion, is one of conservatism. Irrespective ofother aspects of their biology, particularlytheir feeding behavior, they tend to haveremarkably similar reproductive behavior.The most pronounced divergence in be-havior is related to differences in the phys-ical environment in which they breed.There is also a trend toward division oflabor. This has meant size dimorphism withthe larger male more apt to defend at theboundary of the territory and the femaleto stay with the offspring. This seems tohave laid the ground work for the develop-ment of polygyny. The major pressure pre-venting the development of polygyny isdoubtless the necessity of two adult fish todefend the fry against the many predatorsthat lurk around them.

Acanthuridae

Physical spacing. Juvenile fishes havebeen observed in only a few species, so in-formation about them is fragmentary. Thepattern, insofar as known, is to estab-lish solitary feeding territories after meta-




ownloaded from


24 GEORGE W. BARLOW

morphosis, as in Acanthurus chrysosoma(Okuno, 1963). Juveniles are sometimes dif-ferently colored than their adults, and thismay be associated with their territorial wayof life. However, the juveniles of somespecies look much like the adults and arealso highly territorial: Randall (1961a) de-scribed well the early behavior of A. tri-ostegus, and I have observed them in somedetail. Evidently unique among the surgeonfishes, these juveniles occupy intertidalpools, and often at relatively high densities;individual territories are fiercely defended.

Among the adults there is a spectrum ofpatterns of spacing. In some species feedingterritories are held by individuals or bypairs, and range from small to exceedinglylarge. Within these same species at anothertime, schooling behavior may prevail andterritoriality be lacking.

It seems more appropriate to these fishesto discuss spacing from the point of viewof territoriality than from the degree ofschooling. Territoriality in these fishes isthe securing of a feeding area. (Similarly,much of the grouping behavior can be ex-plained as feeding strategy.) A difficulty indiscussing territoriality here, as among non-breeding cichlids, is that feeding territoriesare often small and only briefly held, beinggiven up for a new territory. This is not,therefore, the type of territoriality thatstudents of bird and mammal behaviorgenerally refer to when they use the termterritoriality. But this behavior gradessmoothly into well-defined, apparently per-sistent territoriality. In what follows, I willproceed from the species that are highlyterritorial to those that are less so, and onto those that seem to be nonterritorial.

There are many highly territorial smallreef grazers, such as A. nigrofuscus, A.achilles, and A. glaucopareius. These fisheshold territories ranging from roughly 5-20m2. Whether these territories change insize, shape, or location through time is un-known.

More widely seen is the pattern found inmany of the small reef grazers. This is oneof holding territories briefly while feeding,driving away intruders, but then movingon. Frequently these species form schools

while traveling from one place to another.Sometimes they feed together, but they areusually alone, as typified by A. nigroris. In-cluded here also are Zebrasoma flavescensand possibly Naso lituratus (this is thesmallest of the genus Naso and differs fromthe other Naso in many ways).

Many species are weakly, or not at all,territorial. These include a number of smallto moderately large species adapted to in-vading the reef flats at high tide. Since theyhave to leave these areas as the tide ebbs,persisting territoriality would be disruptedby the tidal cycle. Included here is A. gutta-tus, a schooling medium-sized species.Acanthurus triostegus is a smaller speciesthat moves into the yet shallower areas onthe reef flat, also in large schools. Acan-thurus lineatus moves with the tide, buttends to frequent surge channels. Acan-thurus leucopareius forms groups thatmove along the edge of the reef front andshows only sporadic if any territoriality.Interestingly, A. achilles and A. glauco-pareius both hold territories on the topsof the reef flat near the front of the reef,yet individuals must forsake their terri-tories at times; then they form schools ad-jacent to their uncovered territories.

The detritus feeders present some prob-lems in classification, problems that couldprobably be resolved quickly by markingindividuals. The smaller species are evi-dently territorial and may be living ingroups in some species. They are closelybound to the reef and show a fair amountof aggression among themselves, as well asa considerable amount of toleration formembers of their own species in closeproximity. The large species, Ctenochaetushawaiiensis appears to be a special case.Apparently it lives in groups which gatherin numbers of, say, 3 to 20 at what appearto be refuges. Individuals or subgroupsmay leave and return from time to time. Itremains to be determined whether thesegroups have exclusive use of territories.

Most of the weakly or nonterritorial spe-cies are highly mobile, moving about tomore patchily distributed food. These arethe larger species (plus the smaller reef-flatvagrants mentioned above). A transitional




ownloaded from



species is the large Naso iinicornis, the onlygray unicorn fish that feeds in the intertidaland just subtidal areas. Most of the grayNaso are large species that move in schoolswithin a given area but show no terri-toriality. Possibly, however, the males insome Naso may be defending stations, offthe pinnacles and cliffs where they occur,to serve a reproductive function (see below).

The sand grazers, as already described,refuge in mixed species groups on the reefand move out over the sand to feed in inter-specific schools. The sand is essentially alimitless and indefensible resource, and notsurprisingly these fishes show no terri-toriality. They do have a preferred refugein many cases, and one can properly speakof home range.

One species, Zebrasoma veliferum seemsexceptional. Unlike the territorial reefgrazers, it roams considerable distancesalong the reef. Yet it also appears to beterritorial in that it avoids the area occu-pied by other Z. veliferum. I have seen somedisplays occur when the fish meet at whatmight be the boundaries, but no overtaggression. I have also seen this speciesaggregate in as many as six to eight indi-viduals, but at low tide in patch-reef situa-tions where the animals may have beenforced out of their territories.

Finally, it seems likely that the plankti-vorous A. thompsoni is not territorial.These animals live in schools and takerefuge in the reef when danger threatens;conceivably, they could have individualholes to which they flee and defend, asdoes another planktivore, the black triggerfish Melichthys buniva.

Theoretical spacing. Dominance hier-archies have not been studied in the field,again because of the failure to mark indi-viduals. They may exist, however, particu-larly if the species form closed groups. Themost likely candidates would be those thatmost seem to live in groups, such as Cteno-chaetus hawaiiensis and C. strigosus. Dom-inance hierarchies may also exist in theadult Naso lituratus one sometimes finds ingioups and in which considerable fightingis seen. Much aggression is also expressedeven by apparently nonterritorial reef

grazers and by some browsers.There are lather well-defined dominance-

subordinance relationships between species.In some instances, the amount of aggres-sion between certain species pairs is morethan that seen within them. Thus, inter-specific dominance relationships are animportant factor in the social life of reef-dwelling surgeon fishes.

An appreciation of aggression interac-tions is the key to understanding the com-munication network of the social behaviorof surgeon fishes. Almost all the obvioussignals seem to be related to aggressivebehavior.

The only sensory modality that will beconsidered in detail is the visual system. Ihave no knowledge of whether chemicalsare used in communication. As for acousti-cal signals, the twitch-like displays of someof the species may be associated with theproduction of sound, but little more canbe said.

All the surgeon fishes so far observedseem capable at some age of performingMAP's common to aggressive encounters asseen in most teleost fishes. For example,these fishes regularly show lateral displayand tail beating. Also seen are verticalpostures such as standing on the head ortail, and such overtly aggressive behavioras pursuit, ramming, and cutting with thecaudal knife.

There are, nonetheless, some clear di-vergences in the MAP's employed in com-munication in the different species. Acan-thurus achilles can often be observed indisplaying pairs, swimming in parallel orin tight circles; one or the other fish maytwitch all of the median and paired finssynchronously and repeatedly. Naso lit-uratus occasionally performs what lookslike a tail waggle while standing in theparallel position with another of its spe-cies; this is obviously restrained tail beating,and often results in the two fish beatingmore at one another with their heads thanwith their tails; it is commonly terminatedin one deep tail beat (an attempt to cutwith the pair of knives) immediately fol-lowed by a chase. Another divergence is thepronounced "zooming" seen in C. hawaii-




ownloaded from


26 GEORGE W. BARLOW

ensis: These fish, particularly when onereturns to the group, swim at high speedat one another as though to attack. Sud-denly one fish turns off and glides throughthe water. Acanthurus olivaceus shows asimilar behavior.

Although I was able to detect species-typical differences in behavior, all thespecies showed much the same kind of be-havior when interacting interspecifically.That is, the species-typical behaviors appearto be restricted largely to intraspecific com-munication.

There is little differentiation in bodyshape within each of the various species.However, in the genus Naso sexual di-morphism in body shape and size seemsmore the rule. Several species have a medianhorn-like protuberance that extends straightahead from the space above the eyes, whichgave rise to the common name of unicornfishes; only the males have this feature(Fig. 2). Some species lack a horn and havemerely a pronounced hump where the hornwould otherwise be (N. vlamingi). Nasolituratus lacks even this nuchal hump, butthe male has conspicuous streamers trailingout from the top and bottom margins ofits caudal fin (Fig. 2). In all of the Nasothat are sexually dimorphic (N. hexacanthusappears to be isomorphic), the male isusually larger than the female as well.

As with the cichlids, the most workabledifferences appear in the color patterns. Butunlike the cichlids there is no obvious pat-tern for the group as a whole, with oneexception: There is a pronounced tendencyfor either the caudal peduncle or the tailitself, or both, to bear a contrast-rich mark-ing. Doubtless this is a consequence of thecaudal peduncle being armored with asharp, and perhaps poisonous knife (Ran-dall, 1959; Yasumoto et al., 1971). In mostspecies the knife is carried in a sheathedgroove, one on each side of the fish. In Nasothere are two knives on each side, and theseare fixed out and thornlike.

Much of the differentiation in color pat-terns can be related to the feeding guildand thus to the environment. The plankti-vores and certain of the browsers spendmuch of their time out off the reef; all are

principally gray through blue and pro-tectively counter-shaded. The large Nasoin this group commonly form interspecificassociations when in open water (N. brevi-rostris, N. hexacanthus, N. vlamingi).

The sand-grazers are also a uniformgroup. When up over the reef these largefishes are generally black with conspicuoustail markings, such as a white band on thecaudal peduncle. They also tend to havesubtle markings, particularly yellow on theface, but also elsewhere, such as on themedian, pectoral, or tail fin. They are thusmuch alike in the general pattern of ablack fish with a pale tail mark. When outover the pale sand bottom their bodycoloration fades to gray. They so regularlyschool and feed together, that I am temptedto suggest that they are engaging in inter-specific mimicry. To the observer under-water, these fishes are often difficult todistinguish at a distance because the re-semblance is so great. The more subtlecolor markings are apparent only at closerange and probably facilitate species identi-fication.

Acanthurus olivaceus once again revealsits intermediate nature between that of asand and a reef grazer: Its body color variesbetween being all black or bicolor. Thepatch of orange color across its pectoralgirdle is bright but relatively small, and ithas subtle purple facial markings that areapparent only when up close.

Another form of intermediacy is seen inN. lituratus. It is basically black, like thesand grazers. It shares this characteristicnot only with the sand grazers but with agoodly number of moderately large fishesthat regularly hover up over the reef andtake refuge in it, as does the highly mobileN. lituratus. But N. lituratus also has con-siderable ornamentation on the face andmedian fins; its caudal peduncle is brilliantorange, announcing the pair of formidableknives on each side. Thus, it is intermediatein color pattern between the sand grazersand the more colorful reef grazers that willbe considered below.

The detritus feeders are usually drab.Most are dull brown with muted thin linesor blue dots on the body and inconspicuous




ownloaded from



orange dots on the face. Commonly onlyone species is found in a given place on thereef. In Hawaii, however, C. strigosus andC. hawaiiensis occur together. The smallerC. strigosus has a conspicuous orange ringaround its eye. It sometimes chases themuch larger C. hawaiiensis. The largerspecies is black with blue highlights. Inter-estingly, it spends much time well up fromthe reef. Thus, C. hawaiiensis resemblesthe sand grazers both in color and in thisaspect of its behavior.

Among the reef grazers there has been aflowering of diversity in color patterns.Still, one can detect the tendency for tailmarkings that advertise the dangerous knife(although not always), as well as special andin some cases subtle facial markings. Thetypes of coloration range from having thebody a solid conspicuous color, such as yel-low or blue, to marking the body with avariable number of vertical bars or withthin stripes. Some species have patches ofcolor on the body, or polka dots. A commontheme also is bright coloration on the pec-toral fins or around the eyes.

In spite of this diversity and conspicuous-ness, there is evidence of protective colora-tion. For instance, A. guttatus is often foundwhere the water is full of tiny bubbles fromthe surf, and it has numerous white spots.Further, the most common basic colors areinconspicuous, and most species are counter-shaded.

Many of these species conform to theconcept of poster-colored fishes as advancedby Lorenz (1966), and since they range fromhighly aggressive to highly pacific, thisseemed a good opportunity to test hishypothesis. It states that the color patternsassociated with aggression are permanentlyturned on; no further change is necessaryor possible. Further, the hypothesis predictsthat poster-colored coral-reef fishes will di-rect their aggression only to members oftheir own species. A substantiating piece ofevidence cited by Lorenz is that many dam-sel fishes (Pomacentridae) are drab andnonterritorial as adults (my observations onthis correlation are to the contrary), whereastheir poster-colored juveniles are stronglyterritorial.

The surgeon fishes have proved capableof extremely rapid and profound colorchanges, especially in connection with ag-gressive behavior. The basically black C.hawaiiensis sometimes develops a brilliantblue face while chasing. The extremely ag-gressive lavender tang, A. nigrofuscus, pro-duces a dark profile around its head andbody, leaving its center pale; C. strigosusand C. striatus show a similar change whenfighting. The most remarkable color changesare shown by the most brilliantly coloredspecies, ATaso lituratus: When fighting, theforehead becomes brilliant canary yellowas do the pectoral fins. Sometimes during afight the entire body becomes sky blue, onlyto change rapidly back to black. Extraordi-nary and rapid color changes are also seenin the nonposter colored, open-water Nasos.Depending on the species, these fishes turnon a brilliant blue-white vest, bib, or wedgejust behind the head, with similar colorchanges on the tail (see N. tapeinosoma inFig. 6b of Eibl-Eibesfeldt, 1962).

The remarkable fact in common to allthese observations is that the color changesoccur in intraspecific encounters. I haveseen some changes in extraspecific threats,but only when a well-developed fightoccurred (which is infrequent extraspe-cifically).

The one surgeonfish that was never ob-served to change colors when aggressive asan adult was A. triostegus. This is the leastaggressive surgeonfish I have observed.Also, its caudal knife is greatly reduced and"unadvertised." However, the juveniles areaggressive and do show accompanying colorchanges. Thus, this is an example that isexactly contrary to the poster-color hypothe-sis: The least aggressive adult is character-ized by the lack of change in coloration.

I observed what I believe to be prespawn-ing behavior in A. triostegus, C. striatus,and N. brevirostris. In each instance thecolor changes were similar to those seen inaggressive behavior. In the case of A. tri-ostegus, the color changes paralleled thosenoted for aggressive juveniles.

Surgeon fishes can also change their dark-ness to improve background matching and,thus, protection. The gray Nasos are pale




ownloaded from


28 GEORGE W. BARLOW

up in the water but darken as they approachthe reef to feed. Many of the reef dwellingAcanthurus and Ctenochaetus darken orfade as appropriate to their surroundings.

There are several conclusions that canbe drawn from the analysis of color pat-terns in surgeon fishes. First, there is con-siderable species differentiation by color,but where the fish are constrained fromdramatic coloration, as in the open-waterspecies, some of this differentiation hasbeen transferred to body shape. Also, therehas been either convergence or mimicry inthose species that move about together in ahabitat that places constraints on the colorpatterns, i.e., off-reef Nasos and the sandgrazers.

Diversification of color pattern can occurwhen the species are intimately associatedwith the reef that affords protection frompredators (a similar view has been expressedby Hamilton, 1969). The diversification ofcolor patterns here is probably a conse-quence of the abundance of species, sincethe communication involved in aggressionis important in extraspecific encounters aswell as in intraspecific hostile behavior. Re-call that the reef-dwelling detritus feedersare drab in color when they usually face nocompetition from other acanthurid detritusfeeders. This suggests that the diversity ofcolors among reef grazers serves to facilitaterecognition of intra- versus extra-specificcompetitors for food. Their existence in thereef affords a measure of escape from pre-dation. Thus, their relationship to preda-tion is seen as permissive of the conspicu-ousness rather than causative.

There appears to have been selection forextremely rapid color change to communi-cate aggressive intent within species, evenin the poster-colored ones.

Castes. The biology of juvenile surgeonfishes is poorly known. In a few cases, how-ever, the young are clearly different fromthe adult, usually being solitary and ter-ritorial. In at least three widely unrelatedspecies this is associated with the elabora-tion of bright yellow color. Nonetheless, inA. triostegus, in which the young are ex-tremely territorial, the young show markedcolor changes in association with aggression.

Sex is difficult to distinguish externallyin most species, especially in the generaAcanthurus and Ctenochaetus. Some spe-cies, however, live in pairs, suggesting adifferentiation into sexual roles. Includedhere are the very similar A. achilles and A.glaucopareius. Zebrasoma veliferum livesin groups of one male and two females aswell as in pairs; in each instance, the maleis larger than the female. Randall (1961«)reported that A. triostegus may live inschools in which all individuals are maleor all female.

In the Naso group, sexual dimorphism isoften pronounced in shape and in color;the details have already been reported. InN. lituratus, while the unicorn is wantingits signal function appears to be servedby the canary-yellow patch that is set offagainst a black background on the foreheadduring fights.

Group composition will not be discussedseparately here, since all of the necessaryinformation has been touched upon in theforegoing, or will be treated under repro-duction.

Reproductive behavior. The reproductivebehavior of surgeon fishes is poorly known.This is in large part because spawningprobably usually takes place about dusk.The reef inhabitants count among theirnumbers many small planktivorous speciesthat would devour the eggs if they werearound much of the day. Releasing the eggsjust before dark reduces to a brief periodthe time the eggs are exposed to predationfrom those planktivores, and it is also atime when many fish species are takingrefuge in the reef (Hobson, 1972).

The reported cases of spawning (Randall,19616) involve A. triostegus, C. striatus, andZ. scopas (=Z. flavescens). All of those spe-cies were noted to aggregate toward sunset,and to move up in the water to spawn insmall groups. The nonbreeding behavior inthese species probably facilitates groupspawning since each tends to live in groups,although A. triostegus is more apt to formlarge schools, and C. striatus to live in smallaggregations. Zebrasoma scopas may some-times live in pairs.

In spite of this lack of information, there




ownloaded from



is much suggestive evidence about the re-productive behavior of some of the species.For instance, those species that regularlymove about in pairs probably also spawnin pairs, although this is not certain. Zebra-soma veliferum is a relatively uncommonspecies that is thinly distributed in its en-vironment; it could profit by living inpairs. Its congener, Z. flavescens, lives inlarge loose aggregations where it is abun-dant, such as in Hawaii, but is sometimesseen in distinct pairs at Eniwetok where itis much less abundant. Acanthurus achillesand A, glaucopareius are clearly paired onthe reef. Acanthurus achilles appears toform pairs as juveniles that may persistthrough life. In contrast to Z. veliferum,these two Acanthurus are often locallynumerous and have small territories withseveral neighboring pairs visible to them.Furthermore, when disturbed they swim upfrom their territories and form schools. Itis thus an open question as to whether theymight spawn as pairs or within these largergroups.

One of the most interesting of all thesurgeon fishes is N. lituratus. In someplaces, such as at Eniwetok, they exist asdistinct pairs in the lagoon, but they formdense schools at low tide off the seawardreef. In Hawaii I have often observed onemale with one, two, or three females. Oc-casionally, these animals tarry in nonfeed-ing groups in shallow water where thefemales appear to fight over the male.

The shallow-water unicorn fish, N. uni-cornis, seems to have a harem society. InHawaii I have seen one male drive othermales away from groups of presumably fe-males that consisted of about 20 fish.

The other large Naso that live off thepinnacles and cliffs may have yet anotherreproductive strategy. Naso brevirostris, forexample, is a highly dimorphic species inwhich one could anticipate sexual compe-tition. While inconclusive, observations byMcKaye and me suggest that males holdstations along the cliffs or pinnacles, muchas do individuals in a lek society. Nasovlamingi, on the other hand is less dimor-phic, having only a nuchal hump, whichwould suggest reduced sexual competition;

nothing is known of its breeding behavior.Naso hexacanthus is not obviously dimor-phic and is strongly schooling, probablyin relation to its planktivorous way of life.The implication is that there is no sexualcompetition as a consequence, ultimately,of its feeding strategy, and that the fishtherefore spawn in groups. Nonetheless,this is pure speculation since no informa-tion is available.

The evidence suggests a range of repro-ductive behavior from group spawningthrough harem formation and leks, andeven enduring pairs. There are some sug-gestions of ecological correlations withbreeding strategy, but the relationships arenot clear.

DISCUSSION

To recapitulate, the cichlid fishes inCentral America are products of a geologi-cally recent radiation in the near absenceof competition from other types of fishes.They have occupied almost the entiregamut of trophic types, but this divisionhas not been profound. Interestingly, theleast frequently encountered feeding spe-cialization is that of the algal or Aufwuchsscraper. Only Neetroplus relies primarilyon this food source. This may be becausethe poeciliid fishes preceded the cichlidsinto Central America (Myers, 1966), andmany of these are specialized as Aufwuchsor algae feeders.

The surgeon fishes, in contrast, are awell-established family, having about thesame number of species as do the cichlidsin Central America. They exist in a com-plex, mature community with much tro-phic competition within and without thefamily. Other reef grazers include the par-rot fishes (Scaridae), the damsel fishes(Pomacentridae), the siganids, and thekyphosids. Within their specialization ofalgal feeding there is a further divisioninto guilds with narrow feeding habits. Butwithin each guild there is considerableoverlap in diet, and this is reflected in thesocial extraspecific interactions, the typedepending on the guild. Thus, in the reef-grazing guild there is much aggressive




ownloaded from


30 GEORGE W. BARLOW

interaction between species. But among thesand grazers and among the large Naso thatlive off the reef in open water, the fishesnormally move together with little aggres-sive interaction between species.

I will quickly review the highlights ofsome, but not all, of the parameters ofsocial systems as they apply to cichlid fishesand surgeon fishes.

The cichlid fishes space themselves eithersolitarily, in loose groups, or in well-definedschools. The more predatory they are, themore they tend to be solitary, while themore omnivorous, the more they tend toaggregate. But within any species, individ-uals may at some time be solitary and atothers may move in close proximity tomembers of its species.

Within the surgeon fishes there is alsoa spectrum of patterns of spacing. Somespecies maintain relatively persistent feed-ing territories, others have only transientterritories, and others appear never to beterritorial.

Both in the cichlids and the surgeonfishes the smaller species that are morebound to the substrate are in general moreinclined to be persistently territorial andspaced out. Thus, the small species that areterritorial are the most aggressive membersof the family.

Turning to the large species, in theCichlidae these tend to be predators. Theyare generally well spaced and often terri-torial. In the laboratory they are exceed-ingly aggressive, especially C. dovii.

Jn contrast, the large surgeon fishes aregenerally highly mobile and gregarious spe-cies. While aggression is not common, it isseen among station-holding males, andregularly in the reef-dwelling N. lituratus.

Living in the open, as most of these largesurgeon fishes do, has favored schooling asa maneuver to avoid predation. And whenthe species so engaged are not in competi-tion for food, extraspecific schooling be-comes a simple extension of that protectivemaneuver. Such behavior is facilitated by aclose resemblance among the different spe-cies and by reduced aggression.

Something similar may be going onamong the cichlid fishes that occur together

and that feed over open bottom. In Nicara-gua, the three species C. longimanus, C.nicaraguense, and C. rostratum bear a strik-ing resemblance to one another when notbreeding. They commonly school together.However, when feeding there is extraspe-cific aggression, with C. longimanus domi-nating.

The two families differ emphatically inthe use of color in communication. Withinthe New World cichlids there is a commontheme, almost a digital code, in the devel-opment of dark vertical bars and their con-tained black spots. There is also a recurrentpattern in the application ventrally of thecolors yellow through orange and red,sometimes black, and blues or greensdorsally.

The only common theme in the colora-tion of surgeon fishes is the warning colora-tion associated with the dangerous knife atthe base of the tail. Additionally, speciesthat live away from the reef, and thus moreexposed to predation, are colored for con-cealment. In the shelter of the reef, how-ever, the color patterns tend to proliferate,especially among the reef grazers. Thedetritus feeders that live in the same envi-ronment are exceptional in that they arerelatively conservative and cryptic in theircoloration. This may simply reflect thefact that there are generally few sympatricspecies of Ctenochaetus.

Color changes in the surgeon fishes aremore highly developed than in the cichlidsof Central America in that they are somuch more rapid. These conspicuous,contrast-rich, color changes emerge in thecontext of hostile encounters, predomi-nantly intraspecifically, when even the mostbrilliant species change their signals. Thedrab surgeon fishes can also adopt strikingcolor patterns.

Lorenz (1966) was extremely perceptivewhen he recognized the nexus betweencolor patterns as signals, aggression, andthe intense trophic competition on thecoral reef. (It is not clear to me, however,how he decided which species to consideras being poster colored; to me they lie on acontinuum of conspicuousness.) Lackingadequate information, Lorenz presumed




ownloaded from



perfect habitat partitioning among theposter-colored fishes, with no significantextraspecific competition.

In the reef-grazing surgeon fishes, extra-specific and even extrafamilial competitionis an important reality. The conspicuouscolor patterns probably serve as broadcastsignals, loosely addressed both to intra- andto extraspecific competitors. Color changesusually occur only during actual fightingwhere they are addressed to the object ofthe aggression. Since the more intense com-bats are largely reserved to intraspecificencounters, the signals emerging duringsuch fights are primarily for intraspecificcommunication.

There is little worth commenting on withregard to the castes among cichlids andsurgeon fishes. The cichlids are little dif-ferentiated when not breeding. Among thesurgeon fishes a distinct juvenile caste ispresent in many species, and within Nasothe sexes are usually recognizably different.

The problem of group composition hasalready been touched upon under the cate-gory of spacing. It bears repeating that thegroups are variable and relatively nondif-ferentiated in the nonbreeding cichlidfishes. Furthermore, the groups seem to beopen since their members leave and otherjoin. Within the surgeon fishes the samesituation prevails in many species. How-ever, some species exist as pairs and even asthreesomes, which are thus small closedgroups. They may live in even larger groupsthat are closed, in other species, but thisremains to be determined.

Generally the most interesting aspect ofsocial organization is the reproductive be-havior. Indeed, many writers equate it withsocial organization. It is here that it is mostdifficult to make contrasting comments be-cause so little is known about the surgeonfishes. Fragmentary observations suggestthat when their behavior is known we willhave found a diversity ranging from groupspawning through continuously maintainedpairs, lek societies, and harems.

In contrast, more is known about repro-duction in cichlids than about any otheraspect of their social behavior. In CentralAmerica they have proved conservative in

this regard. They follow the pattern of pro-longed courtship with pair bonding, fol-lowed by parental care. The female doesmost of the direct parental care while themale guards the territory or remains nearbyuntil the fry begin swimming. Then bothsexes guard them. There is evidence of aninclination toward polygyny, but this iscounteracted by selective pressure frompredators, requiring both parents for pro-tection of the fry. The greatest differencesin reproductive behavior result from adap-tations to differences in physical environ-ment, particularly the suitability of thesubstrate as a spawning platform.

The only recent attempt to put the com-parative study of reproductive behavior ofcichlid fishes on a more solid footing hasbeen that by Wickler (1962, 1966). He didan excellent job of analyzing the changes inbehavior and egg morphology associatedwith increasing adaptation to spawning inholes. He attempted to relate the methodof spawning and parental care to whetherthe species is monogamous or polygamous,and to whether it is sexually dimorphic orisomorphic. The previous classification di-vided the various species into substratebreeders versus mouthbreeders. Wickler,however, detected similarities betweenmouthbreeders and those substrate breed-ers that hide their eggs and larvae; thesewere called, collectively, "concealmentbreeders." The other cichlid fishes weretermed "open breeders."

A number of difficulties arise when tryingto use his more definitive statement (Wick-ler, 1966). First, he gave no criteria forclearly distinguishing when a fish shouldbe classified as an open or a concealmentbreeder. His student Apfelbach (1969) laterreclassified Tilapia mariae as a concealmentbreeder, whereas Wickler had consideredit an open breeder, although both madethe same basic observations. Likewise, wehave found that C. nigrofasciatum is a hole(concealment) breeder, not an open breeder.In fact, my observations lead me to suspectthat virtually all of the Central Americancichlids are concealment breeders, whichis not to be confused with breeding in theopen, away from cover.




ownloaded from


32 GEORGE W. BARLOW

The second diiiiculty is thai Wickler(1966) gave no workable criterion for de-ciding when a species is dimorphic as op-posed to isomorphic. His definition waspostulational: a species is dimorphic whenthe species itself can immediately recognizethe sex of a conspecific. Evidently in prac-tice he relied on the degree to which hecould distinguish the sexes by color pattern.In so doing, he appeared to recognize smalldifferences in the genus Tilapia, with whichhe is familiar. He overlooked pronounceddifferences in Cichlasoma: he tallied as iso-morphic the obviously dimorphic anddichromatic C. nigrofasciatum.

I have taken the liberty of rearranging,simplifying, and summarizing Wickler'sdata (Table 1). The old classification is ad-jacent to the new. If the classification is animprovement, it should reduce the intra-category diversity present in the originalclassification. As can be seen, there waslittle if any gain. This is especially truewhen one considers that the category"open" may be nonexistent, that errorsexist in the assignment to the categories,and that the criteria employed are almostimpossible for another person to apply.

For analytical purposes the degree ofdimorphism should be regarded as con-tinuous rather than discrete. The sameapplies to the degree to which each speciesis adapted to a particular set of environ-mental conditions. In my experience Cen-tral American cichlids are all at least size

dimorphic lo some degree.Wickler (1966, p. 137) has also com-

mented on the direction of evolution inthe reproductive behavior of cichlid fishes:"Evolution within this family leads awayfrom the highly developed monogamy, andthus runs, to a degree, backward." This isa misreading of the evidence. In all knownfishes that are parental, outside of theCichlidae, the male is the parent. The evo-lutionary progression has been from (i) anexclusively parental male, to (ii) sharedparental care by the male and female, to(iii) a division of roles with the female asthe direct parent and the male as guardian,to (iv) polygyny (Barlow, 1964). This isclearly a forward progression resulting in amore efficient division of labor between themale and female.

Probably the main hurdle to the devel-ment of polygyny within the Cichlidae isthat the school of fry requires the protec-tion afforded by both parents. Polygynywould probably appear more often if pre-dation on the fry were alleviated. Thiscould be done in a number of ways, suchas the fry behaving differently. Wickler(1966) has shown one way in which thismight proceed with the development offewer larger eggs leading to larger, andmore independent fry. Another adaptationwould be for the fry to hug the bottom,thus avoiding many predators. Finally, thecichlids could breed in areas either eco-logically or geographically distant from

TABLE 1. Species oj ciclilid fishes in each category of reproductive behavior, contrasting two systemsof classification."

Monogamous, MonomorphicMonogamous, DimorphicPolygamous, DimorphicPolygamous, Monomorphic

Total species

H (Diversity)6

OriginalSubstrate

22 63%9 26%4 11%0 0%

35

1.14

Type o£ breeding

classification

2161

10

1.84

Mouth

20%10%60%10%

1.13

20000

20

Revised classificationOpen

100%o%0%0%

0

1.71

Concealment

410101

25

16%40%40%4%

1.42

Data from Wickler (1966).

> H = -^-(log2N! — 2log2n!).




ownloaded from



predators. It would be profitable in thisrespect to study the behavior in the fieldof the polygynous Apistogramma and Lam-prologus. Of course it would also be neces-sary to take into consideration the problemof energetics. Polygyny would necessitatethat many animals in the population beready to breed at the same time. Often thisis not the case in the tropics.

It is worthwhile to compare the repro-ductive strategies of the surgeon and cichlidfishes. The acanthurids require no specialsurface for spawning. However, those spe-cies that live as pairs are territorial reefdwellers. Apparently a physical center ofactivity promotes a continued association;it also enables a pair to exclude intruders.

Most likely all the surgeon fishes releasevast numbers of eggs that become part ofthe plankton and are widely dispersed.Equally likely, they spawn repetitively, sothat the number of gametes produced an-nually by each fish must be prodigious.With this simple and relatively primitivemode of reproduction, and a long timespan, they seem to have produced a diversityof social systems.

The substrate-breeding cichlids, in con-trast, require a particular surface on whichto leave their eggs, and subsequently aplace where they can put their helplesslarvae in order to defend them. They thenkeep the few thousand or hundred vulner-able fry (a relatively small number) closeto them, which makes their protectioneasier. One reproductive cycle takes aboutsix to eight weeks; it is doubtless so ener-getically costly that it is not immediatelyrepeated in nature. Such a mode of repro-duction, however, must be advantageousor it would not have evolved. Yet it hasbeen evolved at a cost. It ties the species,especially the smaller ones, to certain habi-tats and thus to a limited number of trophicsituations. Remarkably, most substrate-breeding cichlids the world over have muchthe same social system. (One might haveexpected them to evolve closed groups froman extended family relationship.)

By evolving mouthbreeding some Africangenera appear to have broken away fromprevious constraints. The habitat require-

ments for spawning are minimal, and thefry are more advanced when released. Per-haps as a consequence, there has been agreater trophic radiation, as in the evolu-tion of plankton filtering species.

In both the cichlids and surgeon fishes,nonetheless, the social systems are seen ulti-mately as consequences of their feedingbehavior, but other factors have a profoundand sometimes more proximate influence.In this paper, the physical environmentand the effect of predation were seen as themost important proximate modulators ofthe reproductive behavior in the Cichlidae.In the Acanthuridae, timing of spawningwas viewed as an adaptation to avoidingegg predators. But the basic social organiza-tion of surgeon fishes most clearly reflectstheir food habits and the physical environ-ment in which they engage in feeding.Here the physical environment means essen-tially the degree to which they are exposedto predation. Another important variableshould prove to be amount of open time,which is also a consequence of feeding be-havior.

The social organization of these fishes isnot, therefore, some happy accident of acapricious evolutionary machination, butrather it reflects an ultimate fundamentaladaptation to the bioenergetics of thespecies.

REFERENCES

Albrecht, H. 1962. Die Mitschattierung. Experientia18:284-286.

Apfelbach, R. 1969. Vergleichend qualitative Unter-suchungen des Fortpflanzungsverhaltens brutple-geraono- und -dimorpher Tilapien (Pisces, Cichli-dae). Z. Tierpsychol. 26:692-725.

Baerends, G. P., and J. M. Baerends-van Roon. 1950.An introduction to the study of the ethology ofcichlid fishes. Behaviour Suppl. 1:1-243.

Barlow, G. W. 1964. Ethology of the Asian teleostBadis badis. V. Dynamics of fanning and otherparental activities, with comments on the behaviorof the larvae and postlarvae. Z. Tierpsychol. 21:99-123.

Barlow, G. W. 1968. Ethological units of behavior,p. 217-232. In D. Ingle [ed.], The central nervoussystera and fish behavior. Univ. Chicago Press,Chicago.

Baylis, J. R. 1974. The behavior and ecology ofHerotilapia multispinosa (Pisces, Cichlidae). Z.Tierpsychol. (In press)




ownloaded from


34 GEORGE W. BARLOW

Bleick, C. R. 1970. The behavior of the CentralAmerican fish, Cichlasoma managuense, and thefunctions of its color patterns: a laboratory andfield study. Masters Thesis, Univ. of California,Berkeley.

Burchard, J. E. 1965. Family structure in the dwarfcichlid Apistogramma trifasciatum Eigenmannand Kennedy. Z. Tierpsychol. 22:150-162.

Crook, J. H. 1964. The evolution of social organisa-tion and visual communication in weaver birds.(Ploceinae). Behaviour Suppl. 10:1-178.

Crook, J. H. 1970. The socio-ecology of primates, p.103-166. In J. H. Crook [ed.], Social behaviour inbirds and mammals. Academic Press, New York.

Denton, E. J., and J. A. C. Nicol. 1966. A survey ofreflectivity in silvery teleosts. J. Mar. Biol. Ass.U. K. 46:685-722.

Eibl-Eibesfeldt, I. 1962. Freiwasserbeobachtungen zurDeutung des Schwarmverhaltens verschiedenerFische. Z. Tierpsychol. 19:165-182.

Estes, R. D. 1969. Territorial behavior of the wilde-beest (Connochaetes taurinus Burchell, 1823). Z.Tierpsychol. 26:284-370.

Fryer, G., and T. D. lies. 1972. Cichlid fishes of theGreat Lakes of Africa. Oliver and Boyd, Edin-burgh.

Hamilton, W. J. 1969. Coral fish coloration. 91 p.(Unpublished manuscript)

Hamilton, W. J., and R. E. F. Watt. 1970. Refuging.Ann. Rev. Ecol. Syst. 1:263-287.

Hobson, E. S. 1972. Activity of Hawaiian reef fishesduring the evening and morning transitions be-tween daylight and darkness. Fish. Bull. 70:715-740.

Jones, R. S. 1968. Ecological relationships in Ha-waiian and Johnston Island Acanthuridae (sur-geonfishes). Micronesica 4:309-361.

Kuhme, W. 1963. Chemisch ausgeloste Brutpflege-und Schwarmreaktionen bei Hemichromis bima-culatus (Pisces). Z. Tierpsychol. 20:688-704.

Lorenz, K. 1966. On aggression. Harcourt, Brace,and World, New York.

Lowe-McConnell, R. H. 1969. The cichlid fishes ofGuyana, South America, with notes en their ecol-ogy and breeding behaviour. Zool. J. Linnean Soc.48:255-302.

McBride, G. 1971. Theories of animal spacing: therole of flight, fright, and social distance, p. 53-68.In A. H. Esser [ed.], Behavior and environment.Plenum, New York.

Miller, R. R. 1966. Geographical distribution ofCentral American freshwater fishes. Copeia 1966:773-802.

Myers, G. S. 1966. Derivation of the freshwater fish

fauna of Central America. Copeia 1966:766-773.Myrberg, A. A. 1965. A descriptive analysis of the

behaviour of the African cichlid fish, Pelmato-chromis guentheri (Sauvage). Anim. Behav. 13:312-329.

Myrberg, A. A. 1972. Social dominance and territo-riality in the bicolor damselfish, Eupomacentruspartitus (Poey) (Pisces: Pomacentriaae). Behaviour41:207-231.

Myrberg, A. A., E. Kramer, and P. Heinecke. 1965.Sound production by cichlid fishes. Science 149:555-558.

Newall, N. D. 1971. An outline history of tropicalorganic reefs. Amer. Mus. Novitates (2465): 1-37.

Noakes, D. L. G., and G. W. Barlow. 1973. Ontogenyof parent-contacting behavior in young Cichlasomacitrinellum (Pisces, Cichlidae). Behaviour 46:221-255.

Okuno, R. 1963. Observations and discussions on thesocial behavior of marine fishes. Publ. Seto Mar.Biol. Lab. 11:111-166.

Pitelka, F. A., R. T. Holmes, and S. F. MacLean, Jr.1974. Ecology and evolution of social organizationin arctic sandpipers. Amer. Zool. 14:185-205.

Randall, J. E. 1959. Report of a caudal-spine woundfrom the surgeonfish Acanthurus lineatus in theSociety Islands. Wasmann J. Biol. 17:245-248.

Randall, J. E. 1961a. A contribution to the biologyof the convict surgeon fish of the Hawaiian Is-lands, Acanthurus triostegus sandvicensis. Pac.Sci. 15:215-272.

Randall, J. E. 1961 b. Observations on the spawningof surgeonfishes (Acanthuridae) in the Society Is-lands. Copeia 1961:237-238.

Stratton, E. S. 1968. An ethogram of Cichlasomaspilotum (Pisces: Cichlidae). Masters Thesis, Univ.of California, Berkeley.

Tinbergen, N. 1959. Comparative studies of the be-haviour of gulls (Laridae): a progress report.Behaviour 15:1-70.

Wickler, W. 1962. Zur Stammesgeschichte funk-tionell korrelierter Organ- und Verhaltensmerk-male: Ei-Attrappen und Maulbruten bei afrika-nischen Cichliden. Z. Tierpsychol. 19:129-164.

Wickler, W. 1965. Neue Varianten des Fortpflanz-ungsverhaltens afrikanischer Cichliden (Pisces,Perciformes). Naturwissenschaften 52:219.

Wickler, W. 1966. Sexualdimorphismus, Paarbildungund Versteckbriiten bei Cichliden (Pisces: Perci-formes). Zool. Jahrb. Syst. 93:127-138.

Yasumoto, T., Y. Hashimoto, R. Bagnis, J. E. Ran-dall, and A. H. Banner. 1971. Toxicity of thesurgeon fishes. Bull. Jap. Soc. Sci. Fish. 37:724-734.




ownloaded from


contrasts in social behavior between central american cichlid fishes and coral-reef surgeon fishes

Documents