sign in sign up
<<
Home , Bodianus, Coral reef fish, Mating, Mating system, Serranus tigrinus

Sigma Xi, The Scientific Research Society

Mating Behavior and Hermaphroditism in Reef : The diverse forms of sexuality found among tropical marine fishes can be viewed as to their equally diverse systems Author(s): Robert R. Warner Source: American Scientist, Vol. 72, No. 2 (March-April 1984), pp. 128-136 Published by: Sigma Xi, The Scientific Research Society Stable URL: http://www.jstor.org/stable/27852523 . Accessed: 24/10/2011 15:46

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Sigma Xi, The Scientific Research Society is collaborating with JSTOR to digitize, preserve and extend access to American Scientist.

http://www.jstor.org Mating Behavior and The diverseforms of sexualityfound in Coral among tropicalmarine fishes can be Hermaphroditism as to their Reef Fishes viewed adaptations equally diverse mating systems Robert R. Warner

The colorful diversity of coral-reef fishes has long been of protogynous are constantly cropping up, and a source of fascination for both scientists and amateurs. the phenomenon may be much more frequent than We now know that this diversity of form and color is previously imagined. matched by an immense variety of social behaviors and Change of from male to female?protandry? sexual life histories, including several kinds of func appears to be less common. It is known in eight families tional hermaphroditism. Recent observations and ex of fishes, three of which?porgies (Sparidae), damsel perimental work suggest that the sexual patterns found fishes (Pomacentridae), and moray (Muraenidae; in fishes may best be viewed as evolutionary responses Shen et al. 1979)?are found on coral reefs. The dam to the species' mating systems, and much of the evidence selfish and porgy families also include species that are an I review here bears out this idea. Since most theory in protogynous. Such variability within a family offers has been derived from studies of important opportunity to test sex-change theory, and terrestrial and , which have strictly deserves further study. separate , the relationships between sexual ex pression and mating behavior in fishes offer new in sights into the role of sexuality in social evolution. Like other vertebrates, most species have sep arate a condition as How sexes, known gonochorism. ever, fishes are by no means restricted to this pattern: in many species individuals are capable of changing sex, a phenomenon sometimes called sequential herma phroditism, and in others fishes can be both sexes at the same time, displaying simultaneous hermaphrodi tism. This sexual flexibility is quite widespread. At least fourteen fish families contain species that exhibit sex as a change from female to male, termed protogyny, normal part of their life histories (see Policansky 1982 for a recent review). Eleven of these families are common in coral-reef areas; and in the (Labridae), par rotfishes (Scaridae), and larger () protogyny occurs in the great majority of the species are studied (Fig. 1). Changes from female tomale also known to occur in (Pomacentridae; Fricke and Holzberg 1974), angelfishes (), gobies (), porgies (Sparidae), emporers (Leth rinidae; Young and Martin 1982), soapfishes (Gram mistidae), and dottybacks (Pseudochromidae; Springer et al. 1977). However, we have no indication of how common sex change might be in these families, since few species have been carefully investigated. New reports

Robert R. Warner isAssociate Professor ofMarine at the Ph.D. University of California, Santa Barbara. He received his from Scripps to Santa Barbara in Institution ofOceanography in 1973. Before coming a Research 1975, he was postdoctoral fellow at the Smithsonian Tropical Institute at Balboa, Panama. His interests include the interactions of life history and behavioral characteristics, the evolution of , the adaptive significance of delayed maturity inmales, and the sexual patterns of tropical marine fishes. Address: Department of Biological Sciences, University of California, Santa Barbara, CA 93106.

128 American Scientist, Volume 72 CL to

a) 0 ml ml ag ml Z ag Sieo dl

a small Figure 2. The expected fertility of female fish (gray line), When large males monopolize mating to the detriment of a measured as the number of eggs produced, usually depends males, however, male fertility rises dramatically at certain point a directly on size and thus shows a steady increase with growth. By in growth (center), and an individual that remains female when a a contrast, a male's expected fertility (colored line), measured as the small and changes to male at large size will be selectively an number of eggs fertilized, is affected by the . When favored. When mating consists of random pairing (right), males and females form monogamous pairs matched by size or individual would do best to function as a male while small (with a when males compete with each other to fertilize eggs and thus to a chance of fertilizing a larger female) and as female when large on a produce the most sperm, both sexes show a similar increase of (capitalizing high capacity for egg production). (After fertility with size (left), and no selection for sex change exists. Warner 1975.)

Species in which individuals are simultaneous maphroditism, or sex change, separately from simulta are , producing eggs and sperm at the same neous hermaphroditism, because they distinct phe time, are common among deep-sea fishes (Mead et al. nomena under the influence of very different selective rare 1964) but quite elsewhere. However, several species regimes. of small sea basses (Serranidae) found in shallow water Ghiselin proposed the "size-advantage model" to sex in tropical areas are known to be simultaneous herma account for many cases of change. The concept is phrodites (Smith 1975). The existence of this sexual simple: if the expected number of (measured, pattern in an abundant species such as the sea basses say, as the number of eggs produced or fertilized) differs poses special evolutionary problems that are discussed between the sexes with size, then an individual that or more below. changes sex at the right size age will have off or Because the warm, clear waters of coral reefs create spring than one that remains exclusively male fe conditions that are nearly ideal for observing male. cause behavior, the social and mating systems of tropical ma What might such sexual differences in the. are im rine fishes have been particularly well studied. It ismy distribution of expected fertility? Two factors intention here first to outline a central hypothesis which portant here: the relative number of male and female relates sexual patterns to social behavior, and then to test produced and the characteristic mating be a this hypothesis at several levels, using information havior of the species. In many cases, female's fertility can or manu available in the literature. is limited by the number of eggs she hold facture, which in turn is controlled by her size, her store of or both. Thus it makes little difference Sex energy, change whether she mates with one male or with many. Male Why should natural selection favor sexual patterns dif fertility, on the other hand, is often limited not by the ferent from pure gonochorism? In other words, under number of sperm an individual can produce but rather what circumstances might we expect hermaphroditism by the number of females with whom he mates and their to be This was first dealt with in a adaptive? question fertility. Because of this, the fertility of males is poten fashion Ghiselin who can comprehensive by (1969, 1974), tially much more variable than that of females, and the conditions under which herma suggested general reach very high levels in certain circumstances (Williams phroditism would be expected to evolve. More recent 1966; Trivers 1972). work has related these to generalities the specific social While size of production sets the stage for of Fischer systems hermaphroditic species (Warner 1975; potential differences in the reproductive success ofmales to 1980; Charnov 1982). It is best view sequential her and females of various sizes, it is often themating system that determines the actual values (Fig. 2). For example, in monogamous species where both members of a pair 1. The Scaridae is one of several common Figure family are normally about the same size, the fertility of males families of coral-reef fishes in which the majority of species and females is the same over their entire include individuals that from female to a approximately change male, size and sex no known as In the range, changing conveys advantage. phenomenon protogyny. striped parrotfish, contrast, coral-reef fishes have Scarus iserti, small females are as females By many mating boldly striped (top)-, inwhich males the grow larger, they change sex and adopt the brilliant blue and systems larger monopolize spawning of females. In this situation smaller males not yellow coloration found in older males (bottom). Groups of may at while females of size have little trouble young females defend territories along the edges of reefs; males all, equivalent and a mate. rate is thus sex-changed females have larger territories which usually finding The spawning of small males encompass several of these harems. (Photos by D. R. Robertson.) lower than that of small females, but large males expe

1984 March-April 129 rience relatively high mating success. Since the distri nov 1982; Goodman 1982). Individuals should change sex a bution of fertility differs between the sexes with size, we sex when the other has higher reproductive would expect sex change to be adaptive: an individual value?that is, higher future expected means that functioned as a female when small and as a male taking into account the probability of death. This sex an after attaining a large size would have more offspring that individuals may (and do) change and suffer over its lifetime than one that remained either male or initial drop in reproductive success, but by making the a female, and thus protogyny should be favored by natural change they increase the probability of attaining high selection (Warner 1975). level of success in the future. These are complications Other mating systems lead to selection for protan we need not consider here, since they do not affect the can dry. Males usually produce millions of sperm, and small general idea that the mating system determine the individuals are physically capable of fertilizing females adaptive value of various forms of sex change. as of almost any size. Thus in mating systems where no Testing an evolutionary idea such this is difficult, are monopolization occurs and where mating consists of since experimental manipulations often impossible one random pairing, it should be advantageous to be a male or exceedingly time-consuming. Typically, must rely when small (since it is probable that any mating will be instead on a search for correlations between the hy as with larger individuals) and a female when large pothesized cause and effect. As long sufficient varia (thereby taking advantage of a high capacity for egg tion exists in the traits in question, the search may take a production). place among unrelated species, within related group, Because the fertilization of eggs occurs outside the or even within a single species. The wide diversity of body in many fish species, spawning is not limited to sexual patterns and behaviors among coral-reef fishes on simple pairs: numerous individuals can mate simulta allows investigation all these levels. In addition, the to ex neously in large groups. Although mating occurs more fact that sex changes are often direct responses as or less at random in such spawning groups, protandry ternal cues makes possible experimental study well. is not necessarily adaptive ifmany males release sperm In this simultaneously. case, competition among sperm in from several males to fertilize eggs creates a situation in Mating sex-changers which male fertility is limited by the number of sperm In general, the mating systems of sex-changing species success sex produced. Such production should increase with size in are those inwhich reproductive varies with a fashion similar to egg production by females, and thus and size. Larger males tend tomonopolize mating, either sexes visit or no fertility differential between the exists. by defending spawning sites that females by The size-advantage model has been refined over the controlling a harem of females. In most of the species, are years to allow for sexual differences both in mortality eggs are simply released into the water, and males to and in the rate at which fertility changes with size free to devote a large amount of time , (Warner et al. 1975; Leigh et al. 1976; Jones 1980; Char spawning, and defense of mating sites.

4-W9

4-~ \Ws,

z-,-a 44 :~ ......

. . ...

. . .. . E W A

as Figure 3. The sexual pattern and mating system of the bluehead interfering with the mating of older males by rushing in sperm as , Thalassoma bifasciatium, is typical of many wrasses. As is released. Young females and young primary males such in most protogynous species, large males play a dominant role. those shown group-spawning are characterized by greenish-black sex These older individuals ?both primary males and sex-changed lateral markings, whereas both older primary males and ? as as with females defend spawning grounds, pair-mating with many changed females a distinctive white band bordered 150 visiting females daily (left). Young primary males engage in black, like the larger individual at the left. (Photos by S. G. group spawning (right) and in "sneaking," the practice of Hoffman.)

130 American Scientist, Volume 72 4. Figure The study of three closely related species of the wrasse seems Bodianus to support the idea that sex change is less common in species in which large males have less opportunity to monopolize mating. Both Bodianus ruf us (top), a species in which large males defend harems of females, and B. diplotaenia one (center), in which large males defend temporary spawning are as sites, protogynous, might be expected in mating systems where large males dominate. By contrast, B. eclancheri (bottom), which in no spawns groups with pattern of domination by large males, is functionally gonochoric, with the sexes existing in equal ratios. (Photos by S. G. Hoffman.)

A good example of such a mating system is found in thewrasse Labrides dimidiatus,which feeds by cleaning the skin, mouth, and of other fishes at specific "cleaning stations" on the reef. The cleaner-wrasses at a station a live in group consisting of a single male and a or harem of five six females. The male actively defends these females and mates with each one every day. This to be a in appears system which there is no advantage to a small being male, and indeed the species is totally protogynous (Robertson 1972). In the last decade, marine biologists have begun to the study mating behavior of a wide variety of proto in gynous species their coral-reef . It is striking that virtually all these species exhibit some form of monopolization of mating by large males, even though they are found in a diverse array of families such as wrasses, , damselfishes, angelfishes, basses, and gobies (see, for example, Moyer and Nakazono 1978a; Robertson and Warner 1978; Cole 1982; and Thresher and Moyer 1983). Haremic mating systems to be most common appear among these protogynous are species. Coral-reef fishes often quite sedentary, and not it is surprising that large males have come to domi nate and defend a local group of females inmany cases. It is this kind of situation just that evolutionarily favors sex change from female tomale. are There also some apparent exceptions to the trend toward the dominance of the large male in pro a togynous species. In number of species of gobies (Lassig 1977), a small bass (Jones 1980), and a wrasse (Larsson 1976), the social system appears to be mono gamous, and thus protogyny would not be adaptive. sex Lassig (1977) suggests that change in the gobies he studied is an to allow reconstitution of a ! mated in case ^ ^* pair of death. This explanation probably * would not apply to the more active wrasses and basses, l?""'-" ?*.^????^ however. For other families inwhich protogyny occurs, such as the groupers and the emporers, we simply lack suffi cient knowledge of themating behavior to state whether the predictions of the size-advantage model hold. Our knowledge of themating habits of protandrous fishes is also of the incomplete. Many species known to lend support to the size-advantage model, approaches be live in schools not protandrous large closely associ from the opposite direction are less satisfactory. For ex ated with the the substrate; however, details of their ample, sex change is not found in every species inwhich behavior have not been The mating reported. size-ad mating ismonopolized by large males. Perhaps this is model that vantage suggests mating might consist of asking toomuch of evolution, since an adaptive situation but this haphazard pairing, prediction remains to be does not guarantee the appearance of a trait. It may tested. The the one anemonefishes, group of protan simply be that the capacity for sexual flexibility has not drous whose species mating system iswell known, fit the yet evolved in some species, or that unknown factors model in a but size-advantage precise unexpected way, reduce the advantage of sex change. Unfortunately, like as is discussed below. many evolutionary arguments, this one is virtually un While most studies of fishes known to change sex testable.

1984 March-April 131 all of whom in fertilization. Comparisons within families males, participate Spawning groups can contain from two to over a hundred males. We can some avoid of the uncertainty inherent in broad For the size-advantage model to hold, variation in comparisons among families by examining the sexual the degree of sex change should correspond to differ and a patterns mating behaviors of group of species ences in the mating systems. Specifically, sex change within a in we sex family which know change iswide should be less common or absent in species where large spread. Using this approach, the absence of sex change males have less opportunity to monopolize mating. where it is theoretically adaptive is less easily dismissed Hoffman's recent work (1980 diss., 1983) on three closely as evolutionary lag. related wrasses of the genus Bodianus provides a clear The wrasses and the (Labridae) parrotfishes (Scar demonstration of this relationship between sexual ex are and well-known families of coral-reef idae) large pression and mating system (Fig. 4). Bodianus rufus of the fishes that include made of both many species up pri Caribbean is haremic, whereas B. diplotaenia of the east males?that fish that remain males for their ern mary is, tropical Pacific defends a spawning site visited by entire lives?and individuals. Both protogynous pri females; thus large males monopolize mating in both males and females can become territo mary dominant, species. Correspondingly, small males are absent in both rial males if The they grow large enough. proportion of species, and all males are the result of sex change in smaller males is a measure of the degree of sex change functional females. On the other hand, the multicolored in the in cases where small are present species: males B. eclancheri of the Galapagos Islands is a group-spawner sex is at a absent, change maximum, and when they form with no apparent pattern of dominance related to size half the the is or sex. population, species essentially gonocho Hoffman could find no evidence of sex change ristic. during adult life in this species. Individuals appear in The of sexual within thewrasses diversity types and stead to be functional gonochores: change from female is reflected in a of behavior parrotfishes diversity mating tomale occurs before maturation, and males are equally Small males either interfere with the (Fig. 3). mating common in all size classes of the population. The pro activities of males in to larger by darting join the duction of small males through prematurational sex at the moment is a spawning couple sperm released, change has also been noted in some species of parrot or practice called "sneaking," they take over a whole fishes (Robertson and Warner 1978). site en masse and with the fe spawning group-spawn What factors lead to changes in themonopolization males that there. In a fe appear group spawning single of mating by large males? We have found that increased male releases her in the midst of an of eggs aggregation population density around spawning sites plays a role in lowering the ability of a male to defend his harem or territory adequately against smaller males (Warner and 0-20% 15-30% 25-50% Hoffman 1980a). In extreme cases, some spawning sites mates mates males primary primary primary can be undefendable and may be entirely abandoned to group-spawners (Warner and Hoffman 1980b). A recent study of the wrasses of the Caribbean (genera Thalassoma, Halichoeres, Bodianus, and Clepticus) revealed that among species living in similar habitats, those with low population densities tended to have few or no small males (Warner and Robertson 1978). Re gardless ofwhether themating system was characterized by harems or spawning-site defense, larger males suc cessfully monopolized mating in these species. In species living at greater densities, the proportion of primary males rose as high as 35%. Among these densely dis tributed species, group spawning as well as territorial mating was seen, with larger males subject to varying amounts of interference from small primary males. population population population The most thoroughly studied species in this group under200 200-400 over400 is the bluehead wrasse, Thalassoma bifasciatum. In this 5. males control the sites Figure The mating system of the bluehead wrasse, which species, large normally spawning on a on smaller reefs where the of the spends its adult life single reef, varies dramatically with the density mating popu size of the reef. In lation is are this diagram of typical mating configurations, low, and small males nearly absent from the predominant direction of the current is indicated by arrows; these local populations (Fig. 5). On large reefs, where sites at the downcurrent end of the reef are preferred by spawning sites are much more crowded, group-spawn females. On small reefs to 600 occurs spawning (2 m2), mating ing aggregations occupy themajor sites and small males with males in their territories blue areas), exclusively large (light are relatively common (Warner and Hoffman 1980b). and primary males are relatively rare. On intermediate-sized Since individuals arrive on reefs as drifting planktonic reefs (600 to 1,000 m2), territorial males continue to occupy the larvae, the precise mechanisms leading to the distribu prime spawning area, but a spawning group of smaller males is tion of small males are not known, but this active just upcurrent (dark blue area). On the largest reefs (above example a serves as a useful illustration of how can affect 1,000 m2), large spawning group of small males occupies the density the major downcurrent spawning site, and territorial males are sex-changing strategy. on relegated to less productive upcurrent sites. Primary males reach Not surprisingly, the effect monopolization of on their highest concentrations such reefs, constituting up to 50% mating ismost pronounced at extreme densities. T. lu of the population. casanum of the eastern tropical Pacific is themost densely

132 American Scientist, Volume 72 6. Figure The most adaptive direction for sex can change depend on the size of the social group. When the group consists of a single pair, both individuals profit if the larger member of the pair is a female, since she could produce more eggs than a smaller individual. Protandry is most adaptive in this and is in case, found the strictly monogamous anemonef ishes of the genus Amphiprion ( top). In larger social groups, the combined egg production of smaller can members easily exceed the egg of production the largest individual, and thus his output is maximized by as a male. functioning Protogyny would be expected here, and is found in the group damself living ishes of the genus Dascyllus (bottom), which are closely related to anemonef ishes. (Photo at top by H. Fricke; photo at bottom by F. Bam.)

distributed wrasse thus far studied; its population is essentially gono choristic, with about 50% primary males (Warner 1982). Large territorial ..-' ... are rare males rather and only mod erately successful in this species, and nearly all mating takes place in groups. Certain characteristics of the that access allow to spawning sites by small males apt to engage in "sneaking" should also affect mate monopolization. These characteris tics are difficult to measure in a quantitative fashion, but some trends are evident. For example, small par rotfishes that live in beds of sea grass near coral reefs have a higher pro portion of small males than species that exist in similar densities on the reefs themselves (Robertson and Warner 1978). In one grass-dwelling species, sex change appears to be entirely absent (Robertson et al. 1982). Sea grasses offer abundant for small fishes, and dominant males in on the fact that the hiding places depends social group is rigidly lim these habitats suffer interference from smaller males in to ited two adults. Ifmore adults were present, themost a of their sexual high proportion . adaptive pattern could instead be protogyny. This the most variation within a is because the a Perhaps telling family largest individual, as male, might be able occurs in the damselfishes where sex to more (Pomacentridae), fertilize eggs than it could produce as a female. was discovered and In change only recently (Fricke Fricke accordance with this, protogyny appears in some re Small damselfishes called or ane 1977). clownfishes lated damselfishes (genus Dascyllus) inwhich the social monef ishes live in or near of adults are (genus Amphiprion) large groups larger (Fig. 6; Fricke and Holzberg anemones in reef areas and thus stinging have extremely 1974; Swarz 1980 and pers. com.; Coates 1982). limited home ranges. They appear to be unaffected by the cells of the and a cer stinging anemone, may enjoy Social control of sex tain amount of protection from the close association change (Allen 1972). An anemonefish consists of two Another of society way testing the size-advantage model is mature individuals and a variable number of an juveniles. through investigation of the dynamics of sex change The are the individual is a within a species protandrous; largest species. So far, I have stressed the importance female, the smaller adult a male and Fricke of the in (Fricke 1977; mating system determining the advantage of and Nakazono The a sex and a Moyer 1978b). per capita production given size. Within mating system, it is often of fertilized eggs is higher when the larger individual relative rather than absolute size that determines re of a is the and is thus ad mating pair female, protandry productive expectations. For example, when dominance to both adults on vantageous (Warner 1978). depends size, the probable mating success of a par Note that the of in this case advantage protandry ticular male is determined by the sizes of the other males

1984 March-April 133 in the local population. Itwould be most adaptive for largest remaining individuals that undergo sex change to to individuals be able change from female to male when the opportunity presents itself. Even when ex when their expectations of successful reproduction as perimental groups consist entirely of small individuals, a male increase Thus a considerably. the removal of sex change can still be induced in the largest individuals a large/dominant male from population should result present, in spite of the fact that they may be far smaller in a sex change of in the next largest individual, but no than the size at which sex change normally occurs change should be expected in the rest of the local pop (Hoffman 1980; Warner 1982; Ross et al. 1983). ulation. The exact behavioral cues used to trigger sex change Such sex social control of change has been noted in appear to differ among species. Ross and his co-workers several of species protogynous coral-reef fishes. Because have shown experimentally that the sex-change re haremic species exist in small, localized groups, they sponse in the Hawaiian wrasse T. duperrey depends have proved to be exceptionally good candidates for solely on relative size and is independent of the sex and studies of this kind. In the cleaner-wrasse L. dimidiatus, coloration of the other individuals in a group, whereas Robertson (1972) found that if themale is removed from Shapiro and Lubbock (1980) have suggested that the the harem, the largest female rapidly changes sex and local sex ratio is the critical factor in the bass Anthias over takes the role of harem-master. Within a few hours squamipinnis. While it is still unclear how sex change is she adopts male behaviors, including spawning with the regulated in fishes that live in large groups, the mech new females. Within ten days this male is producing anisms appear to operate with some precision. Shapiro active sperm. By contrast, the other females in the harem (1980) found that the simultaneous removal of up to nine remain unchanged. male Anthias from a group led to a change of sex in an Social control of sex change has also been found in equivalent number of females. other haremic species (Moyer and Nakazono 1978a; Social control of sex change occurs in protandrous as Hoffman 1980; Coates 1982), well as in species that live fishes as well, and in a pattern consistent with the size in bigger groups with several large males present advantage model. A resident male anemonefish will (Fishelson 1970; Warner et al. 1975; Shapiro 1979; War change sex if the female is removed (Fricke and Fricke ner 1982; Ross et al. 1983). In all cases, it is always the 1977; Moyer and Nakazono 1978b). One of the juve . niles?who apparently are otherwise repressed from maturing?then becomes a functional male and the adult couple is reconstituted.

Simultaneous hermaphroditism In one sense, the adaptive significance of simultaneous hermaphroditism is obvious: by putting most of their energy into egg production and producing just enough sperm to ensure fertilization, a hermaphroditic mating couple can achieve a much higher output of young than a male-female pair (Fig. 7; Leigh 1977; Fischer 1981). The problem, however, rests with the maintenance of si multaneous hermaphroditism in the face of an alterna tive male strategy. Consider an individual that fertilizes the eggs of a , but does not reciprocate by uses producing eggs of its own. Instead, this individual the energy thus saved to find and fertilize other her a maphrodites. This strategy would spread rapidly in purely hermaphroditic population, effectively forcing it to become gonochoristic. Itwould therefore appear that where simultaneous hermaphroditism is present, there should exist some means of preventing this kind of "cheating" (Leigh 1977; Fischer 1981). Among the small coral-reef basses (Serranidae) that 0 optimum 10050 are known to be simultaneous hermaphrodites, two Resources allocated to sperm production (%) types of possible anticheating behavior have been ob common on Figure 7. The number of eggs produced by an individual is served. The hamlets, small basses Caribbean normally directly related to the amount of energy devoted to coral reefs (genus Hypoplectrus), appear to ensure that their manufacture In case {gray line). the of sperm, however, a investments in eggs are kept nearly even between the relatively low output can often produce maximum success members of a what Fischer has - spawning pair by (1980) (colored line) that is, a small amount of sperm can fertilize all called "egg trading." In this behavior, a pair alternates of a eggs, and further investments are partner's superfluous. sex roles over the course of 8). Each time Natural selection favors the individual with the overall mating (Fig. highest an individual functions as a female, it extrudes some, but reproductive rate combining male and female functions (black not all, of its eggs. As a male, it fertilizes the eggs of its line), and in this case the optimum result is obtained by putting who also out in several batches. Thus most of the energy into egg production. A simultaneous partner, parcels eggs a both individuals are forced to demonstrate their commit hermaphrodite following this strategy has much higher success an ment to and reproductive than individual that is exclusively male egg production, neither has the chance or female. (After Fischer 1981.) for an unreciprocated fertilization of a large batch of eggs.

134 American Scientist, Volume 72 Figure 8. Simultaneous hermaphrodites among fishes include the hamlet (genus Hypoplectrus), a small bass that may alternate sexual roles as many as four times in the course of a single mating, by turns offering eggs to be fertilized and fertilizing its partner's eggs. Here the fish acting as the male curves its body around the relatively motionless female, cupping the upward-floating eggs as he fertilizes them. The strategy of parceling out eggs in a number of batches means that the egg contributions of the two partners are roughly equal, and reduces the rewards of "cheating" by fertilizing a large batch of eggs and then refusing to offer eggs for fertilization in return. (Photo by S. G. Hoffman.)

Another method of preventing desertion is to reduce the opportu nities of your partner to find another mate. Some simultaneously herma phroditic species of the genus Serva tius delay their mating until late before These dusk, just nightfall. b species do not engage in egg-trading, but presumably the onset of darkness means that time is quite limited be fore shelter must be taken for the night, and thus further mating is impossible (Pressley 1981). Although these anticheating behaviors are fascinating in their own right, they give us little insight into the of the sexual itself. can see origin pattern Simultaneous cally works against cheating, I no reason why an hermaphroditism has been viewed as adaptation to coral reefs are particularly good places for such pre extremely low population density: if finding mates is vention to come about. it a to to difficult, helps great deal be able mate with Among hermaphroditic groups other than tropical whomever meet our you (Tomlinson 1966; Ghiselin 1969). marine fishes, knowledge of behavior and ecology Thus is many deep-sea fishes, sparsely distributed in their generally insufficient to carry out a similar analysis are habitat, simultaneous hermaphrodites (Mead et al. of the relationship of sexual pattern tomating system. 1964). Perhaps the small serranids evolved from ances Broad surveys of vertebrates (Warner 1978) or tors who lived at low densities, and developed their in general (Policansky 1982) have shown large-scale a anticheating behaviors at later stage when densities tendencies toward hermaphroditism in some groups, were higher. Alternatively, and perhaps more likely, the sporadic appearance of the phenomenon in others, and existence of late-dusk spawning behavior could have a total lack of sexual flexibility in still others. While as allowed the development of hermaphroditism. Unfor major features such the greater complexity of terres we run tunately, again up against the problem of un trial reproduction may help to explain the lack of her testability and limited predictive power: several other maphroditism in some large groups (Warner 1978), mate are coral-reef fishes late in the day, but they not many others must await more thorough investigation simultaneous hermaphrodites. of the life histories and behavior of the organisms in question. Broader On this point, Policansky (1982) has suggested that patterns a major problem of sex-change theory is that among The contrast between sex change and simultaneous closely related species with similar life histories, some sex sex some hermaphroditism is intriguing: change, particularly change and do not. In light of this review, I am protogyny, appears to be a specific adaptation to certain not yet ready to take such a dim view of the size-ad mating systems that happen to be common on coral vantage model. Sexual expression in fishes is extraor reefs. These mating systems may be prevalent in coral dinarily adaptable, and closely related species can have habitats because clear water, relatively low mobility, and quite different sexual patterns that appear to be pre care a the absence of paternal allow greater degree of dictable from their different mating systems. For the dominance by large males. Simultaneous hermaphro fishes, at least, divergent sexual expressions may be no ditism, on the other hand, is theoretically adaptive in a more surprising than differences in coloration. Detailed wider variety of circumstances, but is evolutionarily considerations of the mating systems and life histories unstable unless male-type cheating can be prevented. of other sexually labile groups are needed to test the While the wide dispersion of deep-sea fishes automati hypothesis further.

1984 March-April 135 References Allen, G. R., 1972. The Anemone fishes, Their Classification and Biology. Shapiro, D. Y. 1979. Social behavior, group structure, and the control T. F. H. Publications. Neptune City, NJ: of sex reversal in hermaphroditic fish. Adv. Study Beh. 10:43-102. E. L. 1982. The Sex Allocation. Princeton Univ. - 1980. sex Charnov, Theory of Serial female changes after simultaneous removal Press. of males from social groups of a fish. Science 209:1136 37. Coates, D. 1982. Some observations on the sexuality of humbug dam aruanus selfish, Dascyllus (Pisces, Pomacentridae) in the field. Z. f?r Shapiro, D. Y., and R. Lubbock. 1980. Group sex ratio and sex reversal. Tierpsychol. 59:7-18. /. Theor. Biol. 82:411-26. success Cole, K. S., 1982. Male reproductive behavior and spawning Shen, S.-C, R-P. Lin, and F. C-C. Liu. 1979. Redescription of a pro in a zone Can. 10: temperate goby, Coryphopterus nicholsi. J. Zool. tandrous hermaphroditic moray (Rhinomuraena quaesita Garman). 2309-16. Bull. Instit. Zool. Acad. Sinica 18(2):79-87. E. A. si Fischer, 1980. The relationship between mating system and Smith, C. L. 1975. The evolution of hermaphroditism in fishes. In In multaneous hermaphroditism in the coral reef fish Hypoplectrus terscxuality in the Animal , ed. R. Reinboth, pp. 295-310. nigricans (Seranidae). Anim. Beh. 28:620-33. Springer-Verlag. in a _1981. Sexual allocation simultaneously hermaphroditic Springer, V. G., C. L. Smith, and T. H. Fraser. 1977. Anisochromis straussi, Am. Nat. coral reef fish. 117:64-82. new species of protogynous hermaphroditic fish, and synonomy of and Pseudochromidae. Fishelson, L. 1970. Protogynous sex reversal in the fish Anthias Anisochromidae, Pseudoplesiopidae, Smithsonian Contributions to 252:1-15. squamipinnis (Teleostei, Anthiidae) regulated by presence or absence Zoology Nature of male fish. 227:90-91. Swarz, A. L. 1980. Almost all Dascyllus reticulatus are girls! Bull. Mar. Sei. 30:328. Fricke, H. W., and S. Fricke. 1977. and sex change by ag in gressive dominance coral reef fish. Nature 266:830-32. Thresher, R. E., and J.T. Moyer. 1983. Male success, courtship com and of in three Fricke, H. W., and S. Holzberg. 1974. Social units and hermaphroditism plexity, patterns congeneric species of monochromatic and dichromatic damselfishes in a pomacentrid fish. Naturwissensch. 61:367-68. sexually (Pisces: Pomacentridae). Anim. Beh. 31:113-27. Ghiselin, M. T. 1969. The evolution of hermaphroditism among ani N. 1966. The of mals. Quart. Rev. Bio. 44:189-208. Tomlinson, advantages hermaphroditism and par _ thenogenisis. /. Theoret. Biol. 11:54-58. 1974. The Economy of Nature and the Evolution of Sex. Univ. of California Press. Trivers, R. L. 1972. and sexual selection. In Sexual Selection and the Descent ofMan, 1871-1971, ed. B. Campbell, pp. Goodman, D. 1982. Optimal life histories, optimal notation, and the 136-79. Chicago: Aldine Publishing Co. value of reproductive value. Am. Nat. 119:803-23. Warner, R. R. 1975. The adaptive significance of sequential herma Hoffman, S. G. 1980. Sex-related social, mating, and foraging behavior phroditism in . Am. Nat. 109:61-82. in some sequentially hermaphroditic reef fishes. Ph.D. diss., Univ. _ of California, Santa Barbara. 1978. The evolution of hermaphroditism and unisexuality in _ aquatic and terrestrial vertebrates. In Contrasts in Behavior, ed. E. S. 1983. Sex-related foraging behavior in sequentially herma Reese and F. J. Lighter, pp. 77-101. Wiley. phroditic hogfishes (Bodianus spp.). Ecology 64:798-808. _ sex 1982. Mating systems, change, and sexual demography in Jones, G. P. 1980. Contribution to the biology of the redbanded perch the rainbow wrasse, Thalassoma lucasanum. Copeia 1982:653-61. Ellerkeldia huntii (Hector), with a discussion on hermaphroditism. R. and S. G. 1980a. and the /. Fish Biol. 17:197-207. Warner, R., Hoffman, Population density economics of territorial defense in a coral reef fish. 61: Larsson, H. O. 1976. Field observations of some labrid fishes (Pisces: Ecology 772-80. Labridae). In Underwater 75, vol. 1, ed. John Adolf son, pp. 211-20. _ as a Stockholm: SMR. 1980b. Local population size determinant of a mating system and sexual composition in two tropical reef fishes {Th?lassoma Lassig, B. R. 1977. Socioecological strategies adapted by obligate spp.). Evolution 34:508-18. coral-dwelling fishes. Proc. 3rd Int. Symp. Coral Reefs 1: 565-70. Warner, R. R., and D. R. Robertson. 1978. Sexual patterns in the labroid E. G., Jr. 1977. How does selection reconcile individual advan Leigh, fishes of the western Caribbean, I: The wrasses Smithso with the of the pnas 74:4542-46. (Labridae). tage good group? to 254:1-27. nian Contributions Zoology Leigh, E. G., Jr., E. L. Charnov, and R. R. Warner. 1976. Sex ratio, sex Warner, R. R., D. R. and E. G. 1975. Sex and natural selection, pnas 73:3656-60. Robertson, Leigh, Jr. change change, and sexual selection. Science 190:633-38. Mead, G. W., E. Bertelson, and D. M. Cohen, 1964. Reproduction among Williams, G. C. 1966. Adaptation and Natural Selection. Princeton Univ. deep-sea fishes. Deep Sea Res. 11:569-96. Press. Moyer, J. T., and A. Nakazono. 1978a. Population structure, repro Young, P. C, and R. B. Martin. 1982. Evidence for protogynous her ductive behavior and protogynous hermaphroditism in the angel maphroditism in some lethrinid fishes. /. Fish Biol. 21:475-84. fish interruptus at Miyake-jima, Japan. Japan. J. Ichthyol. 25:25-39. - 1978b. Protandrous hermaphroditism in six species of the anemonefish genus Amphiprion in Japan. Japan. J. Ichthyol. 25: 101-6.

Pressley, P. H. 1981. Pair formation and joint territoriality in a simul taneous hermaphrodite: The coral reef fish tigrinus. Z. f?r Tierpsychol. 56:33-46. Policansky, D. 1982. Sex change in and animals. Ann. Rev. Ecol. Syst. 13:471-95. Robertson, D. R. 1972. Social control of sex reversal in a coral reef fish. Science 1977:1007-9.

Robertson, D. R., R. Reinboth, and R. W. Bruce. 1982. Gonochorism, ...... protogynous sex-change, and spawning in three sparasomatinine parrotfishes from the western Indian Ocean. Bull. Mar. Sei. 32: 868-79...... Robertson, D. R., and R. R. Warner. 1978. Sexual in the labroid ...... patterns .. . fishes of the western Caribbean, II: The parrotfishes (Scaridae). Smithsonian Contributions to Zoology 255:1-26. Ross, R. M., G. S. Losey, and M. Diamond. 1983. Sex change in a coral-reef fish: Dependence of stimulation and inhibition on relative size. Science 221:574-75.

136 American Scientist, Volume 72