Y. Shlesinger á T. L. Goulet á Y. Loya

Reproductive patterns of scleractinian corals in the northern Red Sea

Received: 2 February 1993 /Accepted: 9 March 1998

Abstract The majority of published accounts onscleractinian coral reproduction are from the tropicalPaci®c and Caribbean, with very little informationknown about Red Sea species. This report examinesvariation in reproductive mode in 24 species of her-matypic corals (belonging to seven families) in the Gulfof Eilat, Red Sea. Eighteen species are hermaphroditicbroadcasters, two are hermaphroditic brooders andthree are gonochoric broadcasters. In the Pocillopor-idae, the gonads project into the body cavity, while inthe other six families the gonads reside inside the me-senteries. The number of gonads per polyp in broad-casting species follows family or genus lines. Fecundity(eggs per polyp) increases with polyp size. Broodingspecies usually exhibit one or two gonads per polyp andeach gonad contains only one to three oocytes. Oocytesize varies widely and does not relate to mode of re-production. The largest oocytes (diameter � 450 lm)occur in the brooding coral Alveopora daedalea and inbroadcasting species of the genus Acropora (diame-ter � 420 lm). Gonad morphology and gonochorismversus hermaphroditism appear to be constrained phy-logenetically at the family or genus level. Lastly, thisreport compares the data presented for Red Sea sclera-

ctinian species with the data available on scleractiniancorals from other geographical regions.

Introduction

Knowledge of scleractinian coral reproduction hasprogressively grown in the past 15 years. As Richmondand Hunter stated in a 1990 review, reproductive datawere available for 40% of known species from thetropical Paci®c, 30% of Caribbean coral species andonly 6% of Red Sea species. Since the extensive reviewsby Richmond and Hunter (1990) and Harrison andWallace (1990), additional studies have been publishedon reproduction in corals from the tropical Paci®c(Glynn et al. 1991, 1994, 1996; Bermas et al. 1992; Dai etal. 1992; Shimoike et al. 1992; Stanton 1992; Stobart etal. 1992; Ward 1992, 1995a, b; Tanner 1996; Sakai1997), the Caribbean (Szmant 1991; Johnson 1992; VanVeghel 1993, 1994; Van Veghel and Bak 1994; Van Ve-ghel and Kahmann 1994; Knowlton et al. 1997) and theGulf of Mexico (Gittings et al. 1992). However, infor-mation on the reproductive biology of corals from theRed Sea is very limited (Loya 1976; Rinkevich and Loya1979a, b; Shlesinger and Loya 1985, 1991; Kramarsky-Winter and Loya 1996; Kramarsky-Winter et al. 1997)and still lagging.

Data on scleractinian coral reproductive modes suchas variation in sexuality (gonochorism versus hermaph-roditism), seasonality, and periodicity among coralspecies have been used in an attempt to categorize thereproductive strategies of scleractinian corals. Severalhypotheses have been proposed. One hypothesis statesthat the reproductive variability is due to energeticlimitation (Kojis 1986; Ward 1995a). As reproductiveactivity involves energetic expenditure, under certainenergetic conditions di�erent reproductive pathwaysmay exist. Fecundity and the ratio of male to femalecolonies may be regulated by a variety of internal andexternal processes (Rinkevich and Loya 1987; Ward1995a).

Marine Biology (1998) 132: 691±701 Ó Springer-Verlag 1998

Communicated by O. Kinne, Oldendorf/Luhe

Y. ShlesingerEnvironmental Department Municipality of Eilat,P.O. Box14, Eilat 88100, Israel

T.L. GouletDepartment of Biological Sciences,State University of New York at Bu�alo,Bu�alo, New York 14260, USA

Y. Loya (&)Department of Zoology, The George S. WiseFaculty of Life Sciences and the Porter Super-Centerfor Ecological and Environmental Studies,Tel Aviv University, Ramat Aviv 69978, Israel

Reproductive mode (brooding versus broadcasting)may be determined by the coral's morphological traitssuch as colony size, polyp size, oocyte diameter andgonad location (Rinkevich and Loya 1979a; Szmant1986; Van Veghel and Kahmann 1994). Van Moorsel(1983) hypothesized that reproductive requirements maymodify skeletal characteristics, polyp size and number ofgonads. However, existing data show wide morpholog-ical variation and no linkage to reproductive mode(Kojis and Quinn 1982b; Harriott 1983a; Fadlallah1985; Wallace 1985; Harrison and Wallace 1990).

Reproductive modes may be dictated by phylogeneticconstraints. Harrison (1985) summarized the availabledata on reproductive features of coral species at variouslocations in the tropics, and noted the consistency ofsexuality (gonochorism versus hermaphroditism), gonadarrangement and anatomy within taxonomic groups.His thesis of shared traits due to shared ancestry hasbeen partially supported by detailed descriptions of thescleractinian family Acroporidae (Wallace 1985).

Reproductive mode may be externally controlled byenvironmental factors. Stimson (1978) suggested thatbrooding or broadcasting patterns are determined bydepth below sea surface. Deep sea species are predomi-nantly broadcasters, while species in shallow waterbrood planulae. Further research, however, has revealedthat both brooding and broadcasting species inhabitshallow water (Kojis and Quinn 1982a; Shlesinger andLoya 1985; Harrison and Wallace 1990). Reproductivedi�erences among species may result from habitat con-ditions in di�erent zoogeographic regions (Szmant 1986;Van Woesik 1995; Tanner 1996).

Existing data are insu�cient to strongly support oneof the above theories over the others. Additional com-parative information on sexuality, fecundity and repro-ductive development are needed in order to determinethe factors responsible for the observed reproductivepatterns of stony corals (Harrison and Wallace 1990).

This report summarizes a 10-year study on the re-productive characteristics of 23 Red Sea scleractinianspecies. The studied species, in addition to Stylophorapistillata, contribute approximately 70 to 80% of thetotal living coral cover in the Gulf of Eilat, Red Sea, andare therefore very important in this ecosystem (Loya1972). Furthermore, the present study adds much neededdata about scleractinian coral reproduction in the RedSea. The results are discussed in light of di�erent repro-ductive pathways, with special emphasis on the questionof phylogenetic constraints on gonad parameters, andthe relationship between reproductive strategy (broodingversus broadcasting) and gonad morphology.

Materials and methods

This study was conducted on fringing reefs in the northern Gulf ofEilat, Red Sea (29°30¢N; 35°55¢E). Samples were collected from a5 km stretch of coral reef approximately 10 km south of the city ofEilat. The reproductive ecology of 23 common species of sclera-ctinian corals was examined over 10 years (1980 to 1990). In order

to obtain extensive comparative information, each species wassampled during several reproductive seasons.

Data collected during the ®rst 4 years of the study (1980 to1984) were used to determine the spawning and gonad developmentperiods of the coral species. Six large colonies (>15 cm diameter)from each species were selected randomly on the reef and sampledmonthly for 2 to 4 consecutive years. For branching corals, 5-cmlong branches were collected. For massive corals, live tissue andskeleton approximately 10 cm2 in surface area were removed with ahammer and chisel. Freshly collected samples were examined undera Nikon dissecting microscope to determine oocyte number, colorand size in each gonad. These samples were then ®xed for 24 h in4% formaldehyde solution in sea water, and preserved in 70%alcohol. Corals were decalci®ed using formic acid/sodium citratesolution (after Rinkevich and Loya 1979a). Decalci®ed tissue wasembedded in para®lm, cut into serial sections (6 lm thick), andstained with Dela®eld hematoxylin and eosin.

In addition to the monthly monitoring, in order to detect theprocess of gamete spawning or planula shedding, we conducteddaily in situ monitoring of ten tagged colonies of each speciesduring their respective breeding seasons. Observations on releasedgonad products followed Shlesinger and Loya (1985). Based on thedata obtained in the ®rst 4 years, during the following 6 years, 15colonies of each species were sampled and examined only prior totheir respective spawning periods. In order to quantify data ongonad parameters, polyp and oocyte sizes were assessed from his-tological slides of decalci®ed polyps using a micrometer objectivelens. Species were divided into three size classes: those with smallpolyps (0.7 to 3.0 mm in diameter) in which a histological sectioncontained 10 to 20 polyps each; species with medium-size polyps(3.0 to 6.0 mm in diameter) whose sections contained two to fourpolyps; and those with large polyps (6.0 to 20.0 mm in diameter)consisting of only one polyp per sample. We examined one sectionper colony for small- and medium-polyped species, and three orfour sections per colony for large-polyped species. Maximum oo-cyte diameter was obtained by measuring the largest oocyte ob-served in the histological sections. We counted the number ofgonads per polyp, and oocytes per gonad in both live tissue prep-arations and in serial histological sections. The mean number ofoocytes per gonad was calculated from serial and cross sectionsthrough the entire polyp. In the present paper, the term ``fecundity''refers to the number of oocytes per polyp.

Results

Patterns of sexual development

Most of the coral species studied in the present research(20 out of 23) were hermaphroditic (Table 1). In thethree gonochoric corals (Goniopora savignyi, Porites lu-tea and Pavona varians), sexuality was generally deter-mined at the colony level. In one sampled colony ofP. lutea, a few male polyps were observed in a pre-dominantly female colony. The main mode of repro-duction (21 out of 23 species) was broadcasting ofgametes (Table 1). The species Seriatopora caliendrumand Alveopora daedalea brood and release planulae [inaddition to Stylophora pistillata studied previously byRinkevich and Loya (1979a, b) and also included inTable 1].

Gonad arrangement and morphology varied accord-ing to family (Table 1). Members of the Pocilloporidaeexhibited a unique gonad arrangement. The gonads werefound on stalks arising from the column wall in the bodycavity (Table 1; Fig. 1A). Ova and spermaries co-oc-curred within the same polyp, but were segregated be-

692

tween di�erent pairs of mesenteries (Fig. 1B). In theother six families, gonads were located within the me-senteries (Table 1). In the Acroporidae, Faviidae,Mussidae and Oculinidae, the gonads ®lled the entiremesenteries. Variation between these families occurredin the degree of segregation of testes and ovaries. Threelevels of segregation were observed.

Faviids and the mussid Cynarina lacrymalis dis-played testes and oocytes intermingled throughout themesentery without a distinct region for each gonad(Fig. 1C, D, E). In the mussid Lobophyllia hemprichi,male and female gonads were arranged separatelywithin the same mesentery. In the family Acroporidae,male and female gonads were segregated on di�erentmesenteries in each polyp, and occupied one row in eachmesentery.

In the Poritidae and Agariciidae, gonads created lo-calized enlargements on only part of each mesentery(Fig. 2A). In the family Poritidae, gonads were arrangedin local enlargements which protruded out, and ap-peared as swellings or bulbs on the mesenteries(Fig. 2B). These structures occurred in both gonochoricand hermaphroditic species, and both male and femalegonads exhibited a similar bulbous appearance (Fig. 2C,D). In contrast, the gonads of Pavona varians (Agarici-idae) appeared as small thickenings within the mesen-

teries, which did not protrude out from the mesenterywall (Fig. 2E, F).

In broadcasting species, oogenesis was accompaniedby color change in the developing oocytes. For example,oocytes of Favia favus ®rst appeared green (with oocytediameter of 40 to 200 lm), then became greenish-red,and ®nally dark red (with a diameter of 360 lm) 2 weeksprior to spawning. Oocytes of Acropora eurystomachanged from pink to white, while in Galaxea fascicul-aris the color changed from white to pink (Table 1).These color changes enabled determination of the mat-uration stage of a coral in situ by examination of femalegonad color in gravid colonies. In brooding species, eggcolor did not change throughout oogenesis, apart fromthe progressive increase in the brownish color due to theincrease in zooxanthella numbers during maturation andembryogenesis.

Broadcasting species had annual gametogenic cycleslasting approximately 6 to 7 months. The onset ofoogenesis preceded spermatogenesis by 2 to 4 months(Fig. 3). Nevertheless, during the last 2 months of de-velopment, oocytes and testes matured synchronouslywithin colonies. Monthly changes in oocyte diameterconformed to either linear or exponential growthcurves (Fig. 4). The linear curves indicate that mostoocyte production occurred during the ®rst month of

Table 1 Reproductive patternsof 24 scleractinian coral speciesat Eilat, Red Sea (G go-nochoric; H hermaphroditic;B broadcasting gametes; Pbrooding planulae; C gonads inbody cavity; S gonads in me-senteries)

a Results from studies ofRinkevich and Loya (1979a, b)

Family,species

Sexuality Mode ofreproduction

Gonadarrangement

Oocyte color duringmaturation

PocilloporidaeStylophora pistillataa H P C Brown onlySeriatopora caliendrum H P C Brown onlyPocillopora verrucosa H B C White to cream

AcroporidaeAcropora eurystoma H B S Pink to whiteAcropora hemprichi H B S Pink to whiteAcropora humilis H B S Pink to whiteAcropora hyacinthus H B S Red to pinkAcropora scandens H B S Pink to whiteAcropora variabilis H B S White to pinkAstreopora myriophthalma H B S Red to purpleMontipora erythraea H B S Brown to cream

FaviidaeCyphastrea microphthalma H B S Green to brownFavia favus H B S Green to redFavites pentagona H B S Pink to redGoniastrea retiformis H B S Pink to redPlatygyra lamellina H B S Green to red

PoritidaeAlveopora daedalea H P S Brown onlyGoniopora savignyi G B S White to brownPorites lutea G B S Green to lilac

MussidaeAcanthastrea echinata H B S Green to orangeCynarina lacrymalis H B S Brown to orangeLobophyllia hemprichi H B S Orange to pink

OculinidaeGalaxea fascicularis H B S White to pink

AgariciidaePavona varians G B S White to cream

693

oogenesis. These oocytes then gradually increased insize until maturation. Linear growth curves were seenin oocytes of Montipora erythraea, Astreoporamyriophthalma, Favia favus and Galaxea fascicularis. Incontrast, the exponential growth curves represent con-tinuous production of oocytes over the ®rst 3 months,

during which all the oocytes are produced with verylittle increase in size. These oocytes then undergo arapid acceleration of growth rate. Exponential growthcurves were recorded in oocytes of Acropora eurystoma,Acropora humilis, Goniastrea retiformis and Platygyralamellina.

694

The gametogenic cycles in brooding species weredi�erent from those in broadcasting species in their timescale and degree of synchronization within the colonyand population. Gonads of the brooders Seriatoporacaliendrum and Alveopora daedalea matured in 2 to

Fig. 1 Histological sections of mature gonads of hermatypic corals inthe Red Sea. A Female gonad attached on a stalk (arrow) inPocillopora verrucosa. B Separated ovaries and testes in the polyp ofPocillopora verrucosa. Intermingled gonads within the mesenteries ofC Goniastrea retiformis, D Cyphastrea microphthalma and E Cynarinalacrymalis. F Longitudinal arrangement of spermatogonia andoogonia in the polyp of Acropora eurystoma. G Cross sectionillustrating spermaries and oocytes in di�erent mesenteries of A.eurystoma (Sp spermary; Oc oocyte; N oocyte nucleus; S mesentery)

Fig. 2 Anatomy of gonads in some gonochoric species in the Red Sea.ACross sectionof a femalepolypofPorites luteaandBmagni®cationofan oocyte in the gonadofPorites lutea.COvaries andD spermatogoniaof Goniopora savignyi. E Ovaries and F spermatogonia of Pavonavarians. (Sp spermary; Oc oocyte; N oocyte nucleus; Smesentery)

b

695

3 months (Shlesinger and Loya 1985). Synchronizationin the development of male and female gonads occurredonly in polyps which were in a similar reproductivephase, in terms of initiating oocytes, maturation beforefertilization, and stages during embryogenesis (Shle-singer and Loya, unpublished data). Colonies of thesebrooding species, and even di�erent polyps within agiven colony, did not release planulae synchronously.

Fecundity versus polyp size

Among broadcasting corals, the number of ovaries perpolyp was consistent within the small-polyped familiesAcroporidae (4 ovaries polyp)1) and Poritidae (12 ova-ries polyp)1). Additionally, the number of oocytes perpolyp and per gonad was also similar within these twofamilies. In contrast, the two families with large polyps(Faviidae, Mussidae) exhibited a wider variation in theabove gonad characteristics. The brooding species Sty-lophora pistillata, Seriatopora caliendrum and Alveoporadaedalea developed only one or two ovaries per polyp,with one to three oocytes per ovary. Thus, broodersproduce only one to six eggs per polyp, while broad-casters may release tens to thousands of eggs per polyp(Table 2).

At the family level, the number of ovaries and oocytesper polyp in corals in the Red Sea is similar to elsewhere(Table 3). A constant number of ovaries per broad-casting polyp is found in two additional small-polypedfamilies: the Pocilloporidae (6) and Agariciidae (12)(Tables 2, 3). Reported values of oocyte diameters de-pend on whether measurements were made on live tissueor histological sections. In broadcasting species from

Eilat, polyp diameter was signi®cantly correlated withthe number of oocytes produced per polyp (Kendell'scoe�cient of rank correlation, P < 0.02). Broadcastersin small-polyped families (Pocilloporidae, Acroporidae,Poritidae) produced a fairly low number of eggs perpolyp (20 to 75). In contrast, families with medium-sizedpolyps (Faviidae) had moderate egg production (105 to700 polyp)1), and large-polyped families (Mussidae,Oculinidae) exhibited very high egg production (1000 to3550 polyp)1, Table 2). No signi®cant correlation wasfound between polyp diameter and oocyte diameter(P > 0.05) or between oocyte diameter and mode ofreproduction (P > 0.05).

Discussion and conclusions

This study presents long-term (10 years) data on thereproductive characteristics of 24 scleractinian coralspecies in the Gulf of Eilat, Red Sea. The corals studiedcomprise approximately 70 to 80% of the coral cover inthe Gulf of Eilat and therefore play an important role inthis ecosystem. By monitoring all coral species at thesame site, we were able to determine whether the coralreproductive characteristics exhibited phylogenetic,geographic patterns, or both. The present work providesimportant information concerning approximate spawn-ing times of 11 coral species (Fig. 3) in addition to the 12species previously reported by Shlesinger and Loya(1985). These results further support our contention thatRed Sea corals exhibit temporal reproductive isolation(Shlesinger and Loya 1985).

Most of the species in the Red Sea reproduce duringthe summer months (Fig. 3; Shlesinger and Loya 1985).

Fig. 3 Reproductive patternsof 11 broadcasting species ofscleractinian corals at Eilat in1986. The last two species occurcommonly at 15 to 30 m depthand are rare in shallow water.The period of gonaddevelopment is represented by asolid line with indications of theonset of oogenesis ($) andspermatogenesis (#). Furthersynchronized development ofboth gonads is indicated by theline following the $# symbol.Approximate spawning time,determined by disappearanceof oocytes in the histologicalsections, is indicated by thesymbol

696

Two species in this study, Cynarina lacrymalis and Fa-vites pentagona, started oogenesis in late summer withgamete release in the spring. Stylophora pistillata is theonly other coral species from the Red Sea reported tohave such a gametic cycle (Rinkevich and Loya 1979a).Most of the studied species (20 out of 23) were her-maphroditic (Table 1). In the three gonochoric species,sexuality was determined at the colony level. Her-maphroditism or gonochorism followed family or genuslines. Most species within a given family exhibited thesame gonad morphology such as gonad arrangement,structure and position in each polyp. In addition to thenumber of gonads in each spawning polyp, the number

of oocytes per spawning gonad were consistent withineach family (Table 2). Therefore, these characteristicsappear to be phylogenetic in origin, as suggested byHarrison (1985).

In contrast, the rate of oocyte growth varied evenwithin the same family (Fig. 4). Although the end resultwas the same, some species exhibited linear oocytegrowth while others had exponential growth. These twostrategies of oocyte growth varied in their initial ener-getic allocation. In species that exhibit linear oocytegrowth, oocytes are synchronously formed during the®rst month of oogenesis and then grow at a constantrate. In species that exhibit exponential oocyte growth,oocytes are produced during a prolonged period ofabout 3 months. Then, energy is channeled into in-creasing oocyte diameter until maturity.

Harrison and Wallace (1990) mentioned that oocytesare obviously colored some weeks prior to spawning.The present report is the ®rst to indicate that oocytes ofbroadcasting coral species exhibit color changethroughout maturation (Table 1). This feature can beapplied as a ®eld tool to predict in situ, the approximatespawning time. No such color changes were observed inoocytes of brooding species.

The di�erent modes of reproduction, brooding versusbroadcasting, did not follow taxonomic lines (Table 1).Similar ®ndings were reported by Harrison (1985). Themajority of the coral species in the present study (21 outof 23) broadcast gametes (Table 1). Only two speciesAlveopora daedalea and Seriatopora caliendrum broodand release planulae. The only other brooding speciesreported from the Red Sea is Stylophora pistillata(Rinkevich and Loya 1979a, b). For the Red Sea corals,the reproductive mode re¯ects variation in timing dur-ing gonad development. Oogenesis is long in broad-casting species (5 to 7 months) compared to a shortertime period (<3 months) in brooders. Broadcasters re-lease gametes in a short (1 to 4 d year)1) period withcomplete synchronization in the population. Broodersrelease planulae over a number of months and there isonly partial synchronization within the population.Furthermore, broadcasting species produce many (>10)oocytes per polyp and many (>4 ) gonads per polyp(Table 2). Conversely, brooders produce few (1 to 4)oocytes per polyp and have few (1 or 2) gonads perpolyp (Table 2). In general, brooding species producefew, large eggs, which allows them to provide supple-mentary nutrition for each planula. Thus, di�erent re-productive modes may have arisen in response toenergetic constraints or environmental pressures.

In attempting to understand brooding and broadcastspawning in corals, several studies have concluded thatreproductive mode and fecundity are linked to oocytediameter and polyp size. Chia (1976) proposed thatbroadcasting organisms tend to have smaller eggs andhigher fecundity than related species that exhibit pa-rental care (brooding) and release large larvae. For RedSea corals, oocyte size had no relation to the mode ofreproduction and varied widely. The brooding coral

Fig. 4 Monthly increases in oocyte diameter during the oogenesis ofselected scleractinians in the Red Sea. Left: oocytes of speciesexhibiting linear growth curves; right: exhibiting exponential growth.Each month, for 5 to 6 months, 100 oocytes were measured for eachspecies

697

Alveopora daedalea and broadcasting species of the ge-nus Acropora had the largest oocytes (450 and 420 lm indiameter, respectively). Brooding species of Pocillopor-idae and Poritidae developed larger oocytes than thebroadcasting species in these families (Table 2). Oocytediameter appears to vary widely within individual spe-cies at di�erent locations (Stimson 1976; Harriott 1983a)or within genera (Wallace 1985; Kojis 1986), indicatingthat more extensive comparative studies are needed. Thewide distribution of oocyte sizes within species may bedue to either real di�erences between sites, or to diver-gence in technical procedures (i.e. measurements byhistological sections versus direct tissue measurements,see Wallace 1985). Comparison of oocyte diametersbetween geographical regions would be improved bystandardizing histological procedures and measure-ments.

Rinkevich and Loya (1979a) suggested that broodingand broadcasting species were separated according topolyp size, egg diameter and the anatomical sites ofgonads in the polyp. Szmant (1986) found that in Car-

ibbean corals large colony size and short annualspawning correlated with brooding species, while smallcolony size, multiple planulating cycles per year andoccupation of unstable habitats correlated with brood-ing species. Van Moorsel (1983) suggested that thenumber of sexual products may be a�ected by polypstructure. Wallace (1985) found in Acropora species thatoocyte diameter is inversely related to polyp fecundity.The present study demonstrates that among scleractini-ans in the Gulf of Eilat, the number of oocytes (i.e.polyp fecundity) is positively correlated to polyp diam-eter (Table 2). Nevertheless, the relationship betweenegg number, egg size and polyp size is complex, and isprobably linked to the energetics of reproduction in eachspecies. Furthermore, recent research demonstrates thatmode of reproduction may vary in some species. Forexample, Pocillopora damicornis (Ward 1992) and Go-niastrea aspera (Sakai 1997) exhibit broadcast spawningas well as brooding. At present there seems to be nouniversal relationship between coral size, polyp size and/or egg size and mode of reproduction.

Table 2 Quantitative characteristics of female gonads in Red Seacorals measured from histological sections. Oocyte diameter re-presents the largest measurements in cross sections. Estimation of

number of oocytes per polyp is derived from the average number ofoocytes per gonad multiplied by the number of ovaries per polyp.From each species, 30 colonies were sampled

Family,species

Polypdiameter(mm)

Oocytediameter(lm)

No. ofovariespolyp)1

No. ofoocytesgonad)1

Mean � SDoocytesgonad)1

Estimatedoocytespolyp)1

PocilloporidaeStylophora pistillataa, b 0.75 230 1±2 1±2 1.0 1±2Seriatopora caliendruma 0.75 240 1±2 1 1.0 1±2Pocillopora verrucosa 0.75 130 6 8±12 9.3 � 1.3 56

AcroporidaeAcropora eurystoma 0.75 420 4 5±8 5.2 � 1.5 21Acropora hemprichi 0.8±1.0 395 4 5±9 6.6 � 1.7 27Acropora humilis 0.7±1.0 395 4 5±8 6.6 � 1.8 27Acropora hyacinthus 0.75 375 4 4±6 5.1 � 0.8 20Acropora scandens 0.75 375 4 4±6 4.9 � 1.2 20Acropora variabilis 0.8±1.0 420 4 4±8 5.7 � 1.6 23Astreopora myriophthalma 2.5±3.0 395 4 8±15 10.7 � 2.2 43Montipora erythraea 0.75 290 4 4±10 6.7 � 2.0 27

FaviidaeCyphastrea microphthalma 1.3±1.5 290 12 8±10 8.6 � 0.8 105Favia favus 5.2±6.5 395 18 20±60 38.5 � 11.8 700Favites pentagona 5.5±6.5 240 18 10±30 19.7 � 6.9 355Goniastrea retiformis 3.0±4.0 315 18 20±45 33.1 � 7.0 600Platygyra lamellina 4.5±5.5 370 18 12±60 41.4 � 14.9 750

PoritidaeAlveopora daedaleaa 0.8±1.2 450 1±2 1±3 2.2 � 1.3 4Goniopora savignyi 2.0±2.3 265 12 5±8 6.3 � 1.9 75Porites lutea 1.25 160 12 4±5 4.9 � 1.6 60

MussidaeAcanthastrea echinata 10.0±12.0 300 24 20±62 46.6 � 8.8 1000Cynarina lacrymalis 16.0±20.0 265 60 45±120 79.7 � 21.3 3550Lobophyllia hemprichi 16.0±20.0 290 48 35±65 51.8 � 10.1 2450

OculinidaeGalaxea fascicularis 5.0±7.0 395 36 48±230 60.4 � 21.9 2160

AgariciidaePavona varians 0.6±0.8 50 12 1±2 1.6 � 0.8 18

a Brood planulae, unlike all other species which broadcast gametesbResults from studies of Rinkevich and Loya (1979a,b)

698

Based on the evolution of sexual reproduction in seaanemones (Chia 1976), we suggest that oviparity(broadcasting) is the ancestral mode of reproduction inHexacorallia, and viviparity (brooding) is a derived trait.Evidently, most stony corals retain the ancestral mode of

reproduction, i.e. broadcasting of gametes. The advan-tage of this reproductive mode is long distance dispersaldue to the long planktonic larval period (Harrison andWallace 1990; Richmond and Hunter 1990).

Table 3 Reproductive modes and fecundity in hermatypic corals.The literature survey is based on available data of gonad para-meters in scleractinians. Sources: A Rinkevich and Loya (1979a); BHarriott (1983b); C Stimson (1976); D Fadlallah (1985); E Szmant(1986); F Kojis (1986); G Wallace (1985); H Wyers (1985); I Har-

riott (1983a); J Szmant-Froehlich et al. (1985); K Babcock (1984); LKojis and Quinn (1981); M Glynn et al. (1994); N Glynn et al.(1996); O Szmant-Froehlich et al. (1980); P Fadlallah (1982). Allreferences used ®xed histological sections (P brooding planulae; Bbroadcasting species)

Family,species

Mode ofreproduction

Oocytediameter(lm)

No. ofovariespolyp)1

No. ofoocytesgonad)1

No. ofoocytespolyp)1

Source

PocilloporidaeStylophora pistillata P 230 1±2 1±2 APocillopora damicornis P 100 1±2 1 BPocillopora damicornis P 30 6 3±4 18±24 CPocillopora meandrina 30 6 18±20 108±120 CPocillopora verrucosa B 150 6 60±90 D

AcroporidaeAcropora cervicornis B 300 4±6 6±8 EAcropora cuneata P 325 4 1±3 FAcropora ¯orida B 584 4 6±14 GAcropora granulosa B 534 4 6±23 GAcropora hyacinthus B 622 4 4±11 GAcropora longicyathus B 591 4 6±18 GAcropora loripes B 560 4 5±22 GAcropora nobilis B 571 4 2±12 GAcropora palifera P 325 4 1±3 FAcropora sarmentosa B 652 4 6±14 GAcropora valida B 633 4 3±11 GMontipora sp. 280 4 3±4 12±16 C

FaviidaeCyphastrea ocellina P 24 CDiploria strigosa 330±370 10 HDiploria strigosa B 400 22±32 8±10 EFavia favus 360 1200 IFavia fragum P 250 9±22 1±2 JGoniastrea aspera B 350 1±13 KGoniastrea favulus B 510 1±13 LMontastraea annularis B 300 12 5±14 EMontastraea cavernosa B 350 24 12 E

PoritidaePorites astreoides P 50 6±10 2 EPorites australiensis 150 12 4±8 72 IPorites compressa 120 12 3±4 36±48 CPorites lobata 120 12 3±4 36±48 CPorites lobata 140±170 60±75 MPorites lutea 180 12 4±8 72 I

MussidaeLobophyllia corymbosa 600 2800 IMycetophyllia ferox P 300 12±24 2±4 E

SiderastreidaeSiderastrea siderea B 600 22±32 8±10 E

AgariciidaeGardineroseris planulata 117 26±34 NPavona explanata 12 20 240 CPavona gigantea B 110 18±229 N

RhizangiidaeAstrangia danae B 100±130 24 3000±4000 OAstrangia lajollaensis B 150 6000 P

DendrophylliidaeDendrogyra cylindrus B 300 10±30 E

699

Extreme physical disturbances which are character-istic of reef zones in shallow water may in¯uence theevolution of reproductive modes (Stimson 1978; Szmant1986). Species with large and massive colonies maysurvive these physical disturbances by increasing energyfor growth, and decreasing that for reproduction, whilestill broadcasting gametes. Species which have lowcompetitive abilities or require narrow and speci®cconditions for settlement and recruitment may haveshifted to a di�erent mode of reproduction, i.e. brood-ing.

Szmant (1986) suggested that brooding, with the re-sulting higher recruitment rates, appears to be anadaptation evolved by small-size coral species that su�erfrequent and early colony mortality. Brooding mayprovide a competitive edge during recruitment for somer-strategist species, such as Stylophora pistillata (Loya1976). Brooding results in more rapid settlement andmetamorphosis to the primary polyp, in comparisonwith the slower rate of metamorphosis of primarypolyps in broadcasting species (Shlesinger and Loya1991). Thus brooding would also bene®t species withspecialized environmental conditions. For example,colonies of both brooders studied in the present re-search, Seriatopora caliendrum and Alveopora daedalea,aggregate only in cavities and areas that are protectedfrom currents and wave action (Y.S., personal obser-vations).

Current information on brooding versus broadcast-ing corals illustrates the large structural and develop-mental di�erences taking place during reproductionbetween these two pathways. Fundamental questionsfrom an ecological point of view, however, remain.How do the reproductive characteristics of di�erentspecies relate to their evolutionary history? Which re-productive features have diverged according to eco-logical pressures and which were stimulated byenergetic processes within the colony? Finally, whichreproductive features are constrained by phylogenetichistory? We have found that in the 24 coral speciesstudied in the Red Sea certain reproductive traits, suchas gonad anatomy and sexuality (gonochorism andhermaphroditism) appear to be constrained by phylo-genetic origin. Furthermore, we suggest that in sclera-ctinian corals, broadcasting appears to be the ancestralmode, while brooding may have arisen in certain spe-cies in response to their individual ecological and en-vironmental requirements.

Acknowledgements We thank B. Coloroni for her histological help,and E. Kramarsky-Winter, D. Goulet, P. Harrison andB. Rinkevich who critically read this manuscript. We thankY. Benayahu, Z. Dar, L. Fienstein and A. Sha®r for their time anddiving company. This work was conducted at the InteruniversityInstitute of Eilat, Israel. We thank the former director, I. Paperna,the present director, A. Baranes, and the sta� for their hospitalityand assistance during the work. This research was supported by theIsrael Science Foundation (Y.L.) and the Nature Reserve Au-thority of Israel (Y.S.).

References

Babcock RC (1984) Reproduction and distribution of two speciesof Goniastrea (Scleractinia) from the Great Barrier Reef prov-ince. Coral Reefs 2: 187±204

Bermas NA, Alino PM, Atrigenio MP, Uychiaoco A (1992) Ob-servations on the reproduction of scleractinian and soft coralsin the Philippines. In: Richmond RH (ed) Proc 7th int coralReef Symp. Vol. 1. University of Guam, Mangilao, pp 443±447

Chia FS (1976) Sea anemone reproduction: patterns and adaptiveradiation. In: Mackie GO (ed) Coelenterate ecology and be-havior. Plenum Press, New York, pp 261±270

Dai CF, Soong K, Fan TY (1992) Sexual reproduction of corals innorthern and southern Taiwan. In: Richmond RH (ed) Proc 7thint coral Reef Symp. Vol. 1. University of Guam, Mangilao,pp 448±455

Fadlallah YH (1982) Reproduction ecology of the coral Astrangialajoleris in a kelp forest habitat. Oecologia 55: 379±388

Fadlallah YH (1985) Reproduction in the coral Pocillopora verru-cosa on the reefs adjacent to the industrial city of Yanbu (RedSea, Saudi Arabia). In: Gabrie C et al. (eds) Proc 5th int CoralReef Congr. Vol. 4. Antenne Museum EPHE, Moorea, FrenchPolynesia, pp 313±318

Gittings SR, Boland GS, Deslarzes KJP, Combs CL, Holland BS,Bright TJ (1992) Mass spawning and reproductive viability ofreef corals at the East Flower Garden Bank, northwest Gulf ofMexico. Bull mar Sci 51: 420±428

Glynn PW, Colley SB, Eakin CM, Smith DB, Corte s J, GassmanNJ, Guzma n HM, Del Rosario JB, Feingold JS (1994) Reefcoral reproduction in the eastern Paci®c: Costa Rica, Panama ,and the Gala pagos Islands (Ecuador). I. Poritidae. Mar Biol118: 191±208

Glynn PW, Colley SB, Gassman NJ, Black K, Corte s J, Mate JL(1996) Reef coral reproduction in the eastern Paci®c: CostaRica, Panama , and the Gala pagos Islands (Ecuador). III.Agariciidae (Pavona gigantea and Gardineroseris planulata).Mar Biol 125: 579±601

Glynn PW, Gassman NJ, Eakin CM, Corte s J, Smith DB, Guzma nHM (1991) Reef coral reproduction in the eastern Paci®c: CostaRica, Panama , and the Gala pagos Islands (Ecuador). I.Pocilloporidae. Mar Biol 109: 355±368

Harriott VJ (1983a) Reproductive ecology of four scleractinianspecies at Lizard Island, Great Barrier Reef. Coral Reefs 2: 9±18

Harriott VJ (1983b) Reproductive seasonality, settlement and postsettlement mortality of Pocillopora damicornis (Linnaeus), atLizard Island, Great Barrier Reef. Coral Reefs 2: 151±157

Harrison PL (1985) Sexual characteristics of scleractinian corals:systematic and evolutionary implications. In: Gabrie C et al.(eds) Proc 5th int Coral Reef Congr. Vol. 4. Antenne MuseumEPHE, Moorea, French Polynesia, pp 337±342

Harrison PL, Wallace CC (1990) Reproduction, dispersal and re-cruitment of scleractinian corals. In: Dubinsky Z (ed) Ecosys-tems of the world. 25. Coral reefs. Elsevier, Amsterdam,pp 133±207

Johnson KG (1992) Synchronous planulation of Manicina areolata(Scleractinia) with lunar periodicity. Mar Ecol Prog Ser 87:265±273

Knowlton N, Mate JL, Guzma n HM, Rowan R, Jara J (1997)Direct evidence for reproductive isolation among the threespecies of the Montastrea annularis complex in Central America(Panama and Honduras). Mar Biol 127: 705±711

Kojis BL (1986) Sexual reproductive in Acropora (Isopora) species(Coelenterata; Scleractinia). I. A. cuneata and A. palifera onHeron Island reef, Great Barrier Reef. Mar Biol 91: 291±309

Kojis BL, Quinn NJ (1981) Aspects of sexual reproduction andlarval development in the shallow water hermatypic coral Go-niastrea australensis (Edwards and Haime, 1857). Bull mar Sci31: 558±573

Kojis BL, Quinn NJ (1982a) Reproductive strategies in species ofPorites (Scleractinia). In: Gomez ED et al. (eds) Proc 4th int

700

coral Reef Symp. Vol. 2. Marine Sciences Center, University ofthe Phillipines, Manila, pp 145±151

Kojis BL, Quinn NJ (1982b) Reproductive ecology of two faviidcorals (Coelenterata: Scleractinia).Mar Ecol Prog Ser 8: 251±255

Kramarsky-Winter E, Fine M, Loya Y (1997) Coral polyp expul-sion. Nature, Lond 387: p 137

Kramarsky-Winter E, Loya Y (1996) Regeneration versus buddingin fungiid corals: a trade-o�. Mar Ecol Prog Ser 134: 179±185

Loya Y (1972) Community structure and species diversity of her-matypic corals at Eilat, Red Sea. Mar Biol 13: 100±123

Loya Y (1976) The Red Sea coral Stylophora pistillata is an rstrategist. Nature, Lond 259: 478±480

Richmond RH, Hunter CL (1990) Reproduction and recruitmentof corals: comparisons among the Caribbean, the tropical Pa-ci®c, and the Red Sea. Mar Ecol Prog Ser 60: 185±203

Rinkevich B, Loya Y (1979a) The reproduction of the Red Seacoral Stylophora pistillata. I. Gonads and planulae. Mar EcolProg Ser 1: 133±144

Rinkevich B, Loya Y (1979b) The reproduction of the Red Seacoral Stylophora pistillata. II. Synchronization in breeding andseasonality in planulae shedding. Mar Ecol Prog Ser 1: 145±152

Rinkevich B, Loya Y (1987) Variability in the pattern of sexualreproduction of the coral Stylophora pistillata at Eilat, Red Sea:a long-term study. Biol Bull mar biol Lab, Woods Hole 173:335±344

Sakai K (1997) Gametogenesis, spawning, and planula brooding bythe reef coral Goniastrea aspera (Scleractinia) in Okinawa, Ja-pan. Mar Ecol Prog Ser 151: 67±72

Shimoike K, Hayashibara T, Kimura T, Omori M (1992) Obser-vations of split spawning in Acropora spp. at Akajima Island,Okinawa. In: Richmond RH (ed) Proc 7th int coral Reef Symp.Vol. 1. University of Guam, Mangilao, pp 484±488

Shlesinger Y, Loya Y (1985) Coral community reproductive pat-terns: Red Sea versus the Great Barrier Reef. Science 228: 1333±1335

Shlesinger Y, Loya Y (1991) Larval development and survivorshipin the corals Favia favus and Platygyra lamellina. Hydrobiol-ogia 216/217: 101±108

Stanton FG (1992) Spatio-temporal patterns of spawning in thecoral, Montipora verrucosa in Hawaii. In: Richmond RH (ed)Proc 7th int coral Reef Symp. Vol. 1. University of Guam,Mangilao, pp 489±493

Stimson JS (1976) Reproduction of some common Hawaiian reefcorals. In: Mackie GO (ed) Coelenterate ecology and behavior.Plenum Press, New York, pp 271±279

Stimson JS (1978) Mode and timing of reproduction in somecommon hermatypic coral of Hawaii and Enewetak. Mar Biol48: 173±184

Stobart B, Babcock RC, Willis BL (1992) Biannual spawning ofthree species of scleractinian coral from the Great Barrier Reef.In: Richmond RH (ed) Proc 7th int coral reef Symp. Vol. 1.University of Guam, Mangilao, pp 494±499

Szmant AM (1986) Reproductive ecology of Caribbean reef corals.Coral Reefs 5: 43±54

Szmant AM (1991) Sexual reproduction by the Caribbean reefcorals Montastrea annularis and M. cavernosa. Mar Ecol ProgSer 74: 13±25

Szmant-Froelich A, Reutter M, Riggs L (1985) Sexual reproduc-tion of Favia fragum (Esper): lunar patterns of gametogenesis,embryogenesis and planulation in Puerto Rico. Bull mar Sci 37:880±892

Szmant-Froelich A, Yevich P, Pilson MEQ (1980) Gametogenesisand early development of the temperate coral Astrangia danae(Anthozoa: Scleractinia). Biol Bull mar biol Lab, Woods Hole158: 257±269

Tanner JE (1996) Seasonality and lunar periodicity in the repro-duction of pocilloporid corals. Coral Reefs 15: 59±66

Van Moorsel GWNM (1983) Reproductive strategies in two closelyrelated corals (Agaricia, Scleractinia). Mar Ecol Prog Ser 13:273±283

Van Veghel MLJ (1993) Multiple species spawning on Curacaoreefs. Bull mar Sci 52: 1017±1021

Van Veghel MLJ (1994) Reproductive characteristics of the poly-morphic Caribbean reef building coral Montastrea annularis. I.Gametogenesis and spawning behavior. Mar Ecol Prog Ser 109:209±219

Van Veghel MLJ, Bak RPM (1994) Reproductive characteristics ofthe polymorphic Caribbean reef building coral Montastrea an-nularis. III. Reproduction in damaged and regenerating colo-nies. Mar Ecol Prog Ser 109: 229±233

Van Veghel MLJ, Kahmann MEH (1994) Reproductive charac-teristics of the polymorphic Caribbean reef building coralMontastrea annularis. II. Fecundity and colony structure. MarEcol Prog Ser 109: 221±227

Van Woesik R (1995) Coral communities at high latitude are notpseudopopulations: evidence of spawning at 32°N, Japan.Coral Reefs 14: 119±120

Wallace CC (1985) Reproduction, recruitment and fragmentationin nine sympatric species of the coral genus Acropora. Mar Biol8: 217±233

Ward S (1992) Evidence for broadcast spawning as well asbrooding in the scleractinian coral Pocillopora damicornis. MarBiol 112: 641±646

Ward S (1995a) Two patterns of energy allocation for growth, re-production and lipid storage in the scleractinian coral Pocillo-pora damicornis. Coral Reefs 14: 87±90

Ward S (1995b) The e�ect of damage on the growth, repro-duction and storage of lipids in the scleractinian coral Po-cillopora damicornis (Linnaeus). J exp mar Biol Ecol 187:193±206

Wyers SC (1985) Sexual reproduction of the coral Diploria strigosa(Scleractinia, Faviidae) in Bermuda. In: Gabrie C et al. (eds)Proc 5th int Coral Reef Congr. Vol. 1. Antenne MuseumEPHE, Moorea, French Polynesia, pp 301±305

701