Bijdragen lol de Dierkunde, 60 (3/4) 225-232 (1990)

SPB Academie Publishingbv, The Hague

Growing knowledge and understanding of dichromatism and sex reversal

in Labridae

Machteld J. Roede

Health Sciences, State University Limburg, P. O. Box 616, 6200 MD Maastricht, The Netherlands

sexual Keywords: Sex reversal, hermaphroditism, diandry, monandry, sex dimorphism, color changes,

behavior, Labridae

Abstract pour l’explication de l’évolution des divers modes typiques de la

biologie reproductive des Labridae.

A brief survey is given of current literature in comparison with former studies. Since the 18th century phenotypic diversity of

Labridae is of At first in dribs and a matter argument. drabs, yet Introduction last decades at a rapid rate, dichromatic adult morphs are sy- nonymized. There are few monochromatic species.

introduction to the marine world was a are protogynous hermaphrodites. Most species are My during

with Jan Stock described as monandric, but a minority (of slightly or extremely student field trip to southernFrance dichromatic)species is diandric,with both primary and second- of the leaders. fieldwork the as one Later, during on

males. Terminal males with but in- ary reproduce single females, "girelle" in Banyuls-sur-Mer and PhD-work on itial phase males spawn in aggregations. Some species are Caribbeanwrasses, Jan happened to be passing by. haremic, and large males are permanently or temporarily ter- he succeeded the late as ritorial. Last decades insight in the various social factors has Finally, professor Engel my

Just while contribution greatly increased, resulting in theoretic models to explain the promotor. preparing a to evolution diverse of labrid reproductive of the typical, modes Stock's farewell, Charnov's analysis of sex reversal biology. in labroid fish (1982) was brought to my notice. I

refrained from an intended paper on present day

work, and became wrapped up once more in the

Résumé fascinating world of "my" wrasses, that I left a

quarter of a century ago. On passe brièvement enrevue la littérature récente, encomparai- When labrid work my was published (1966- son avec les études plus anciennes. Depuis le XVIIIèmesiècle, la

1975) there was hardly any response. Conform to diversité phénotypique des Labridae est un sujet de discussion.

D’abord de façon sporadique, mais ces dernières décennies à un the "ignoration of hermaphroditic fish by com- rythme rapide, les morphes dichromatiques adultes sont syn- parative endocrinologists, behaviorists, and ecolo- onymisées. Il y a peu d’espèces monochromatiques. gists - all of whom should find among these un- Les Labridae sont des hermaphroditesprotogynes. La plupart usual fish exceptions to prove their rules" (Atz, des sont espèces décrites comme étant monandriques, mais il y

or an illustration of the drawback of the atti- a une minorité d’espèces diandriques (faiblement ou extrême- 1965)?

mâles aussi bien tude scientist to stick to the facts and but ment dichromatiques), aux primaires que secon- as nothing daires. Les mâles terminaux se reproduisent avec des femelles the facts, according to Hull's observation (1989) Cer- seules, mais ceux de phase initiale fraient enaggregations. that "Those scientists who have attempted to taines espèces ont des harems, et les mâles de grande taille sont produce theory-free descriptions of natural territoriaux, de façon permanente ou temporaire. Pendant ces have had little the dernières décennies, la connaissance de facteurs sociaux divers phenomena pitifully impact on

of science"? a beaucoup progressé, ce qui a abouti à des modèles théoriques course 226 M.J. Roede - Dichromatism and sex reversal in Labridae

Later, several ichthyologists extended both the mention its relation to sex - a first comparative,

labrid empiric and theoretic knowledge about the swift, demographic study (data on abundance,

of size- the agile fish. Yet, except for the world-topic harem color-, and sex-composition of popula-

Labroidesdimidiatus(cf. Robertson, 1972) and the tions) was performed. Investigations of 5800 speci-

Science coverstory of Thalassoma bifasciatum (cf. mens belonging to one Mediterranean and seven

Warner et al., 1975), the wrasses stayed mainly a Caribbean species conclusively proved reversal of

small in all and in- field of interest for a relatively in-crowd. sex eight species studied, quantified an

However, in view of the models labrids offer terspecific variability with in some species a strict,

population biologists, and the interest of e.g. par- in others a more loose relationbetween color, size,

tial and total outsiders as sociobiologists and and sex (Roede, 1966, 1972, 1975).

and of sexual be- feminists for the evolution of sex The present paper intends to check the universali-

havior in Labridae deserve of this brief of general, the certainly ty variability by a survey current

more general recognition. literature, including theoretical explanations con-

The speciose family of Labridae is most diverse cerning the evolution of dichromatism and sex

in terms of size and form. Closely related species reversal.

often cannot be separated on basis of meristic

Color offers relevant identifi- characters. pattern a

cation mark, though complicated by the fact that Labridae

numerous species display a striking phenotypic

diversity during life. Juveniles may differ from the General information. — (If no specific reference is

adults, and there may be a smaller, plainly colored given, the informationis from Roede, 1972.)

as well as a larger, brightly colored adultphase. The Labrid fishes are (sub)tropical and temperate

marine of clear lattermay exhibit an elongation of parts of the dor- species coastal waters. Typically,

sal fin or of the pelvic fins, a strongly forked caudal they have thick lips and protruding teeth.

fin, or a prominent hump on the forehead. Most wrasses are reef-dwellers, active during

Since Linnaeus, roughly speaking, five steps in daytime, but with approach of sunset or danger

the approach of this dimorphism can be distin- rapidly seeking shelter among the rocks or dense

guished. At first, the adult forms were separated layers of algae, or digging into the sand. Tropical

into different consid- species. Secondly, they were species can be observed the whole year through but

ered to be females and males of one protogynous in colder areas in autumn and winter the fishes

species. This concept was unsettled when among either move to deeper water or hibernateburrowed

the "female" morph functional males were found. under the sand.

The first indications in this direction date back to Various species form abundant aggregations,

the end of the 19th but with crowds labroid century these findings were mingling the of other species.

not taken Skin- and scuba- Less less certainly generally up. common are numerous, homospecific spe-

diving underwater observations on labrid cies at rockless sand in expanded staying cays or seagrass

social and sexual be- beds. In the Labridae ben- systems, including territorial general greedy are

havior. This knowledge formed the base for thophagic carnivores, not overparticular in the

hypotheses to explain labrid dichromatism and choice of their food. Though some interspecific

reproductive in biology the framework of evolu- differences in diet occur, there is a great overlap.

tionary theories; at first, by field researchers (vari- According to size the diet changes from copepods

ous papers by Warner and Robertson c.s.), later to mainly crustaceans and molluscs. Robertson &

also by mere theoretic analysis (Charnov, 1982). Polunin (1981) and Robertson (1984) experimental-

Around — 1960/1963 when for some European ly refuted the traditional view of a food resource

color matter species and sex were a of argument, partition among reef fishes. (The young of) some and for the of Caribbean majority wrasses change species are cleaner fishes, feeding on ectoparasites of color was not or only vaguely described, not to of other fishes. - Bijdragen tot de Dierkunde, 60 (3/4) 1990 227

Color patterns and dichromatism Besides, Randall and others described:

(a) A few monochromatic species: Coris venusta (Randall,

1976), Labroides dimidiatus (Robertson & Hoffman, 1977), Two categories of color changes can be distinguished: Pseudolabrus semifasciatus, P. torotai and P. gayi (but P. color - due to and a. Fast, reversible shadings of aggregations fuentesi is sexually dichromatic) (Russell & Randall, 1980) and dispersions of the pigment within the chromatophores. These Labroides phthirophagus (Randall, 1985). are related to affective behavior. Yet, prudence is called for when uniform coloration is record- b. Morphologicalcolor changes - in which the number of chro- ed, because this can be due to the fact that the relatively more matophores may alter as well as the amount of pigment they

rare TP specimens were just not included in collected samples. contain. These are relatively slowly developing processes of a

As illustrated by Thalassoma cupido, an abundant Japanese rather irreversible character.

that became listed as dichromatic when Katherine Meyer

(1977) bothered to collect a 220 specimens.

(b) Some species showing just a fading of colors in relation to The following refers to the morphological, adult

increase in size, in some combined with alteration of particular color changes. Roede found in relation to increas- spots, e.g. Halichoeres podostigma(Randall, 1980), H. hortula- size of: color ing a variety (a) no striking changes of nus and H. trispilus (Randall & Smith, 1982), H. ornatissimus

(Halichoeres poeyi), (b) only a shading from dark, and Thalassoma ballieui (Randall, 1985). contrasting colors towards soft, pastel hues with no (c) Yet, dozens of species show either a partial dichromatism,

and actual of thebasic e.g. Halichoeres dussumieri H. stigmaticus (Randall et al., changing pattern (H. bivittatus),

1982),or are completely dimorph, e.g. Anampses chrysocepha- (c) conspicuous color changes, varying from local lus, A. neoguinaicus and A. femininus (exceptionally IP more alterations (Hemipteronotus splendens) to more colorful than TP) (Randall, 1972), Coris gaimard (Randall,

radical ones involving all parts of the body (Coris 1976), C. aygula, Stethojulisalbovittata, Halichoeres margina-

and julis, Thalassoma bifasciatum, Halichoeres macu- lus, Anampses meleagrides(Randall, 1983) A. cuvier, Coris

ballieui, C. flavovitta, Stethojulisbalteata, Thalassoma duper- lipinna and H. garnoti ).

rey, Th. purpureum , Th. trilobatum, Pseudojulis cerasinus and Detailed descriptions were given of the smaller, Gomphodus varius (Randall, 1985). more drab (be it in Th. bifasciatum on peak mo- ments first adult or initial dazzlingly yellow) phase, Later publications thus subscribe to the conclusion intermediate color phases and the gaudy terminal that Labridae are mutually divergent in their

Due to reversible the differences phase. shadings phenotypic diversity, even to be found within one between the in- or decrease tem- phases may . porarily.

At first, the two forms were indicated as either The relationship of color to size, and sex "drab and terminal", "drab and gaudy" or "fe- male and male" Roede to denote phases. preferred At first, the distinct phenotypes were a source of the adulthood of also the smaller ones by using - confusionand especially in the 18th and 19th cen-

"first terminal adult and This - phase". terminology tury of extreme controversy. They received has not been generally adopted and "initial phase" different vernacular and scientific names, e.g.

(IP) and "terminal became common phase" (TP) Coris giofredi/C. julis and Thalassomanitidus/Th. that "initial" use. Be aware is not equivalent to bifasciatum.

For detailed court- "juvenile". a survey, including

As early as 1800 Retzius synonymized morphs of Labrus bi- ship alterations, see Robertson et al., 1977.

maculatus. Steindachner (1868) was the first to unite the Numerous reports on labrids have been presented brownish and the orange banded turquoise forms of Coris julis, by Randall, a Flying Dutchman moving from one but his revision of systematics was not generally accepted up to sea to the other, generating new species on the mo- 1955. Longley's uniting(1914,1915)of the yellow and bluehead- ment he dives to be in the Guin- of Thalassoma became down; easily listed green morphs bifasciatum generally ac-

cepted after 1930. (Detailed and references in Roede, ness Record Book on basis of the hours spent under survey 1972.) water, often at considerable depths, the number of

The last decades frequently, and now noiselessly, incorrectly species described, or the number of his color pic- classified dichromatic pairs are linked, such as (IP and TP, tures of the fins erected and the fish, showing essen- respectively) strigiventer/S. renardi (cf. Randall, tial "life" colors. 1955), 228 M.J. Roede - Dichromatism and sex reversal in Labridae

Anampses caeruleopunctatus/A. diadematus (Randall, 1972), in this family (Atz, 1964; Roede, 1973). In ancient

rosea/C. ballieui Halichoeres Coris (Randall, 1976), gym- times sex of labrids already caused confusion; Pliny nocephalus/H. chloropterus and H. timorensis/H. kawarin correctly observed that in the "girelle" (Coris julis) (Randall, 1980), Lepidaplois albomaculatus/Bodianus axillaris, there are more females than males, and sometimes Julis kalosoma/Stethojulisinterrupta, Julis notopsis/Halichoe-

males at all. varialvus/M. no res marginatus, Macropharyngodon bipartitus,

Macro- and of and Julis umbrystygma/Thalassoma purpureum (survey in: micro-inspection gonads gave

rubrocaudatus/A. Randall, 1983) or Anampses chrysocephalus conclusively proof of sex reversal. Generally regres- and Gomphodus varius/G. tricolor (Randall, 1985). Sometimes sion of ovaries takes place in wrasses of intermedi- also intermediates and juveniles have received separate scientific ate sizes; areas of spermiogenesis amid ovarian names. Of the now linked pair Halichoeres leparensis and H. ar-

structure in found in fish in initial the intermediate considered their regression are gus H. fijiensis was even as hybrid (Weber, 1913 and De Beaufort, 1940 in: Randall, 1980), colors (Reinboth, 1961, 1975; Roede, 1966, 1972; while the adult forms Coris eydouxii (IP) and C. lepomis (TP) Meyer, 1977). Randall (1980) reported (gross in- now are synonymized under the juvenile name Coris flavovittata spection) intersexual gonads in intermediate fish of (cf. Randall, 1976, 1985). Halichoeres melanurus, H. chloropterus and H.

timorensis. labrid A second step was to regard adult dimor- The idea of a sex-specific dichromatism proved smaller phism as a sexual dichromatism, the morph to be an over-simplification because in some species being females, the more conspicuous one males. the initial found among morph up to 15-30% were Also in monochromaticspecies females are smaller, to be functional with often males, very mature whereas most large fish are males. This implies a testes. reversal of sex as a common phenomenon.

This for Coris Bacci & Sex reversal was signalized julis (Lo Bianco, 1909;

Razzanti, 1957, 1958; Reinboth, 1957; Roede, 1966),

Crenilabrus ocellatus and C. quinquemaculatus(Soljan, 1930, In the dioecious higher vertebrates, reproducing by 1931), Labrus bimaculatus (Lönnberg & Gustafson, 1937), internal fertilization, spontaneous reversal of sex Thalassoma bifasciatum (Longley et al., 1941; Stoll, 1955;

most It is due to tumor chro- occurs rarely. a or sex Roede, 1972), Halichoeres bivittatus (Roede, 1972) and mosomalaberrations. But in the higher bony fishes (Randall, 1955).

of hermaphroditism (the presence recognizable ovarian and testicular tissue in a single individual;

Atz, 1964) is the rule rather than the exception, es- Diandry pecially among Cyprinodontiformes and Per-

Often ciformes. this hermaphroditism takes the se- Reinboth (1961) distinguished primary males,

i.e. individual quential form, the changes sex some- (male at birthand throughout life; externally to be where during its life. distinguished from IP females by genital papillae)

In the Serranidae both synchronous and pro- and secondary males (formerly females), differing togynous hermaphroditism are well represented. In in vas deferens and structure of the gonad. By now a numberof species the gonads stay functional dur- diandry is amply studied, especially in Thalassoma ing the transition from female towards male; yet, bifasciatum (various publications by Warner and

is self-fertilization not frequent. In Sparidae pro- Robertson c.s., Drew, 1986) and Th. cupido (Mey- togynous, protandrous as well as simultaneous her- er, 1977). maphroditism occur. But unlike the serranid ovo- Yet, in the numerous taxonomie descriptions of testis, the ovarianand the testicular the last decades the of parts are clearly only incidentally occurrence separated connective tissue. Labridae by In pro- males among the initial phase is reported, e.g. for togynous reversal of sex proceeds via a total decline Halichoeres scapularis and H. marginatus (Randall of the transforming ovary. Consequently, a tem- et al., 1982) and Thalassoma trilobatum (Randall, porary functional hermaphroditism does not occur 1985). Bijdragen tot de Dierkunde, 60 (3/4) - 1990 229

few have been Gonadal activity praiseworthy that a specimens ever

collected.

there is relation between dian- Testes of IP males are markedly more developed Summarizing, no than those of TP males. The mature testes of the dry and degree of color change, diandry being

abdominal for species with a (Coris julis, former occupy great part of the cavity, reported great whereas TP testes, also in the minority with ripe Thalassoma bifasciatum, Stethojulis strigiventer,

of Halichoeres or a moderate gonads, are relatively small strings. Most the marginatus ) only (Th. largest terminal males have nonfunctionalgonads, cupido, Halichoeres bivittatus) contrast between

While with local forms of senescence (quantified in initial and terminal phase. the apparently

is for Roede; confirmed, also for diandric parrotfishes, much more common monandry proven spe- by Robertson & Choat, 1975; Warner & Robertson, cies that are monochromatic (Halichoeres poeyi),

1978). monandric species show only slight or moderate

It seems paradoxical that the larger, often ter- (Hemipteronotus splendens and He. martinicen-

between the adult ritorial and aggressive part of the population has sis), or striking differences absolutely and relatively smaller gonads. This in morphs (H. garnoti and many others). contrast to observations by various authors that TP males (at least of Thalassoma bifasciatum) may

40-100 but IP males Labrid social spawn times, just once or systems twice a day.

For an understanding of the basic causes that favor

dichromatismand sequential hermaphroditism var-

Monandry ious aspects of labridsocial systems are relevant. By

and structure now, population density are being

In the major part of at least 90 species they studied even by grid studies and observations of in- described, Randall and co-authors stick to a strict dividual fish (Shapiro, Fitch in: Drew, 1986). color- and size-specific separation of females and males. E.g. for Halichoeres margaritaceus, H. nebulosus and H. miniatus (Kuiter & Randall, Sexual Behavior

1981), the Anampses, Pseudolabrus and most of

diverse. the Halichoeres species mentionedabove, eight spe- Modes of reproduction are In tropical

labrids into the cies of Pseudojuloides (Randall & Randall, 1981) gametes are released freely water, and various Paracheilinus species (Randall & Lub- but temperate wrasses may lay demersal eggs and bock, 1981). exhibit male brood-care (Robertson et al., 1977).

Indeed, next to diandry monandric species (all In all species pair-spawning occurs, a terminal

with females. males being reversed females) exist, as shown by male reproducing single These males demographic studies for Halichoeres garnoti, may defend either permanently or temporarily ter-

Hemipteronotus splendens, Hemipteronotus mar- ritories, and display courting behavior. (Such be-

for dichromat- tinicensis, Bodianus rufus and Clepticus parrae havior gave Randall support linking

(Roede, 1972; Warner et al., 1978). ic pairs.) In diandricspecies, moreover, IP females

and males in Both of The author has strong doubts, though, whether spawn aggregations. types all the species presented elsewhere as sexually spawning are characterised by a rapid up and down dichromatic actually are monandric. Only studies rush to minimize reef-based prédation on the

the Often the fishes from of sufficiently large samples can give a decisive an- spawners or eggs. migrate swer for each species concerned. This is not a nega- their feeding areas to special spawning sites, at tive criticism, since some of the discussed species specific times of the day.

that are rare ones, or common, yet with a predilection Soljan (1930, 1931) reported when large

after for deep waters, or secretive in habits so it is Crenilabrus males, building nests, court the 230 M.J. Roede - Dichromatism and sex reversal in Labridae

females number of females. thus females, small males of similar hues as the Experimentally, even

fertilize small females can be forced to reverse sex rush foreward at the supreme moment to at (Shapiro,

Such 1984, in: Drew, least part of the eggs. "sneaking" (stealing 1986). spawnings with a female waiting to spawn with a

TP male) and "streaking" (joining TP maleand fe- Theoretic models male at the climax of spawning) by IP males is now reported for several other diandric species (Warner Most theories to explain the complex labrid interre- et al., 1978). lations of dichromatismand modes of reproduction Either the fish form loosely organised mating were formulated around the seventies. maintain (Labroides systems, or males harems The conflicting necessities of concealment for dimidiatus, Robertson et al., 1977; Cirrhilabrus predators and of sexual recognition should in- rubriventralisand Paracheilinusspecies, Randall et fluence the degree of dichromatism. Labrids of al., 1981). The latter display a rigid hierarchy, with coral areas tend to be fully dichromatic, whereas a permanently territorial male dominating males for wrasses from seagrass beds cryptic coloration is and a group of females. essential in view of risk more a greater on préda-

tion. In a species with long-term stability and cons-

Abundance and sex ratios tant male-female interactions such as monochro-

matic Labroides dimidiatus, males - all maintain- In all species females are in the majority; the initial ing harems and sexually active - have not to phase strikingly outnumbers intermediates and ter- differentiate themselves from inactive or nonter- minal specimens. Either part of the fish stop grow- ritorial males. Here are no strong selection pres- ing, or there is a sudden size-specific mortality. In sures towards difference in coloration (Robertson this respect the description ofthree types of Y chro- et al., 1977). This is contradicted, though, by other mosomes in Coris julis may matter (Duchác et al., haremic species as Paracheilinus being sexually 1982, in: Drew, 1986). These could control in- dichromatic. differences traspecific in the ability to change sex In 1981 Robertson argued that the evolutionary and colors (see below), but also lifespan. of is rather development dichromatism a response Local differences are analysed in detail, mainly to inter- than intrasexual interactions. The relative for Thalassoma bifasciatum. Over the day sex ra- spatial distributions of feeding and spawning sites tios may temporarily differ. Samples of clusters affect the chance of spawning with a TP male, and consist ready to spawn mainly of mature IP males, the ability of males to monopolize mates. but in sexually inactive groups females strongly Lower vertebrates are strictly dioecious. This outnumber IP males. With increase of reef size leads to the question how natural selection has fa- population (deme) size increases, yet the proportion vored hermaphroditism over dioecy in higher of bluehead(TP) males, thus the numberof territo- teleosts. Most discussions are based on Ghiselin's ries, decreases (Warner & Hoffman, 1980). suggestion (1969) that reversal of sex was favored

when the reproductive success is closely related to

or thus evolved since it was more advanta- Determination of sex and sex reversal age size,

geous for males to be larger. This is complicated,

The discussion is directed to the question whether though, by the phenomenom of diandric species, sex determination is controlled where non-sex males co-exist with strictly by genes, changing sequen- triggered by environmental events, or a combina- tial hermaphrodites. Randall et al. (1963) already tion of both (e.g. an environmentally induced shift remarked that pair spawning is the basic reproduc- in least gene expression). In at some species sex tive pattern for parrot and labrid fishes and ag-

is reversal is socially induced, as shown by change gregate spawning a secondary development. of sex by the removal of a male in haremic species The existence of diandry depends upon opportu-

(Robertson, 1972), or by surpassing a threshhold nities for primary males to breed. Their frequency Bijdragen tot de Dierkunde, 60 (3/4) - 1990 231

1985. Guide Hawaiian reef fishes: 1-79 depends upon population density, a non-per- Randall, J.E., to (Har-

rowood Books, Newton Square PA). manent territorial mating system, and their chances

J.E. & M. 1980. Description of a new and of terminal Randall, Dor, genus to interfere with the spawning large, species of labrid fish from the Red Sea. Israel Journal of have social males. When males, however, a strong , 29: 153-162. is monandric. control over the females, the species Randall, J.E. & R. Lubbock, 1981. Labrid fishes of the genus

from These ideas are checked by Charnov (1982) who Paracheilinus,with descriptions of three new species the

Ichthyology, 19-30. tested the ESS (Maynard Smith's "Evolutionary Philippines. Japan. J. 28:

Randall, J.E. & Helen A. Randall, 1963. The spawning and ear- Stable Strategy" of 1976) allocation to male versus ly development of the Atlantic parrot fish, Sparisoma female function by subjecting labrid data to the rubripinna, with notes on other scarid and labrid fishes. Shaw-Mohler and derived equations. Zoologica, 48: 49-60.

revision of the labrid Summarizing, social systems control mating op- Randall, J.E. & Helen A. Randall, 1981. A

fish genus Pseudojuloides, with descriptions of five new spe- portunities, yet spatial distribution of food and cies. Pacific Science, 35: 51-74. shelter determine social systems. Accordingly, Randall, J.E. & MargaretM. Smith, 1982. A review ofthe labrid herbivore scarids with terri- among the permanent fishes of the genus Halichoeres of the Western Indian Ocean,

tories haremic are much more common than species with descriptions of six new species. Ichthyol. Bull. J. L. B.

Smith Univ. 45: 1-26. among the carnivorous labrids. Inst. Rhodes, Grahamst.,

Reinboth, R., 1961. Natürliche Geschlechtsumwandlung bei

adulten Teleosteern. Zool. Anz., (Suppl.) 24: 259-262.

Reinboth, R. (ed.), 1975. Intersexuality in the kingdom: References 1-450 (Springer-Verlag, Berlin).

Robertson, D.R., 1972. Social control of sex reversal in coral-

reef fish. 177: 1007-1009. Atz, J.W., 1964. Intersexuality in fishes. In: C.N. Armstrong& Science,

of A.J. Marshall (eds.), Intersexuality in vertebrates including Robertson, D.R., 1981a. Coexistence: symbiotic sharing

feeding areas and algal food by some coral reef fishes from man: 145-232 (Academic Press, London/New York).

Atz, J.W., 1965. Hermaphroditic fish. Science, 150 (3697): the Western Indian Ocean. Mar. Biol., 62: 185-195.

789-797. Robertson, D.R., 1981b. The social and mating systems oftwo

maculipinnaand H. garnoti,offthe Charnov, E.L., 1982. The theory of sex allocation. Monographs labrid fishes, Halichoeres

in Population Biology, 18: 1-335 (Princeton University Caribbean coast of Panama. Mar. Biol., 64: 327-340.

territorial Press, Princeton, N.J.) Robertson, D.R., 1984. Cohabitation of competing

reef. Drew, Catherine L., 1986. Dichromatism and diandry in Thalas- damselfishes on a Caribbean coral Ecology, 65:

the Wrasse. Thesis 1121-1135. soma bifasciatum (Labridae), Bluehead

Stat. Univ. New York at Stony brook: 134+25 pp. Robertson, D.R. & J.H. Choat, 1974. Protogynous her-

Ghiselin, M.T., 1969. The evolution of hermaphroditismamong maphroditism and social systems in labrid fishes. Proc. Sec-

. Quart. Rev. Biol., 44: 189-208. ond Int. Symp. Coral Reefs, 1: 217-225.

of female Hull, D.I., 1989. Progress in ideas of progress. In: M.H. Mat- Robertson, D.R. & S.G. Hoffman, 1977. The roles

mate choice and in the of thew (ed.), Evolutionary progress: 27-48 (University of prédation mating systems some

Chicago Press, Chicago). tropical labroid fishes. Zeitschr. Tierpsychologie, 45:

Kuiter, R.H. & J.E. Randall, 1981. Three look-alike Indo- 298-320.

Pacific labrid fishes, Halichoeres margaritaceus, H. nebulo- Robertson, D.R. &N.V.C. Polunin, 1981. Coexistence: symbi-

sus and H. miniatus. Rev. fr. Aquariol., 8: 13-14. otic sharing offeedingareas and algal food by some coral reef

Meyer, Katherine A., 1977. Reproductive behavior and patterns fishes from the Western Indian Ocean. Mar. Biol., 62:

of sexuality in the Japanese labrid fish Thalassoma cupido. 185-195.

Japan. J. Ichthyology, 24: 101-112. Roede, Machteid J., 1966. Notes on the labrid fish Coris julis

Randall, J.E., 1955. Stethojulis renardi, the adult male of the (Linnaeus, 1758) with emphasis on dichromatism and sex. Vie

labrid fish Stethojulis strigiventer. Copeia, 3: 237. Milieu, (A) 17 (3A): 1317-1333.

Machteid Color related to and be- Randall, J.E., 1972. A revision of the labrid fish genus Roede, J., 1972. as size, sex,

Anampses. Micronesica, 8: 151-195. havior in seven Caribbean labrid fish species (genera Thalas-

Randall, J.E., 1976. A review of the Hawaiian labrid fishes of soma, Hallichoeres and Hemipteronotus). Studies Fauna

26: 1-10. the genus Coris. Uo, Jap. Soc. Ichthyol., Curaçao, 42: 1-266.

Randall, J.E., 1980. Two new Indo-pacificlabrid fishes of the Roede, Machteid J., 1973. Intersexuality and sex reversal in fish-

with other of the Intern. Res. Commun. Medicine and genus Halichoeres, notes on species genus. es. Syst. Veterinary

Pacific Science, 34: 415-432. comparative Physiology, (73-9)15-14-12.

Randall, J.E., 1983. Red Sea reef fishes: 1-192 (Immel Publ., Roede, Machteid J., 1975. Reversal of sex in several labrid fish

London). species. Pubbl. Staz. zool. Napoli, 39 (Suppl.): 595-617. 232 M.J. Roede - Dichromatism and sex reversal in Labridae

Russell, B.C. & J.E. Randall, 1980. The labrid fish genus Pseu- determination of mating system and sexual composition in

dolabrus from islands of the Southeastern Pacific, with tropical marine fishes (Thalassoma). Evolution, 34: 508-518.

description of a new species from Rapa. Pacific Science, 34: Warner, R.R. & D.R. Robertson, 1978. Sexual patterns in the

433-440. labroid fishes of the western Caribbean, I: The Wrasses

Soljan, T., 1930. Die Fortpflanzung und das Wachstum von (Labridae). Smiths. Contrib. Zool., 254: 1-27.

Crenilabrus ocellatus Forsk., einem Lippfisch des Mittel- Warner, R.R., D.R. Robertson & E.G. Leigh, 1975. Sex change

meeres. Zeitschr. wiss. Zool., 137: 150-174. and sexual selection. Science, 190: 633-638.

Soljan, T., 1931. Brutpflege durch Nestbau bei Crenilabrus

quinquemaculatus Risso, einem adriatischen Lippfisch.

Zeitschr. Morph. Ekol. Tiere, 20: 132-135. Received: 7 February 1990

R.R. Warner, & S. Hoffman, 1980. Local population size as a Revised: 11 June 1990