Ó Springer-Verlag 2000 Marine Biology (2000) 136: 1045±1056 C. Ramofa®a á S. C. Battaglene á J. D. Bell á M. Byrne Reproductive biology of the commercial sea cucumber Holothuria fuscogilva in the Solomon Islands Received: 28 June 1999 / Accepted: 8 March 2000 Abstract Reproduction of Holothuria fuscogilva (Selenka, 1867) in the Solomon Islands was investigated over a 4 yr period (1994 to 1998) by macroscopic and microscopic examination of the gonad tubules, the gonad index (GI) method, histological examination of gametogenesis, and spawning-induction trials. The gonad consisted of numerous tubules that dominated the coelom of gravid specimens. New tubules appeared in March, and grew in size and extent of branching until they reached their maximum size and maturity in August. Spawning occurred from August to October, with the majority of gametes released during October, although it was only partial in many individuals. After spawning, the tubules appeared wrinkled and resorbed into the gonad basis. A ®ve-stage gonad maturity scale based on the macroscopic appearance of the gonad tubules corresponded with discrete stages of gametogenesis identi®ed by histology. Gametogenesis was initiated in mid-March, with oogenesis and spermatogenesis occurring in parallel, followed by the growing stage (May to July) which was marked by active gamete development. Successful induction of spawning during the breeding period corroborated the GI and histological data. The uniform growth of gonad tubules indicated that H. fuscogilva in the Solomon Islands does not conform to the progressive tubule recruitment model Communicated by G. F. Humphrey, Sydney C. Ramofa®a (&) á M. Byrne Department of Anatomy and Histology, F13, University of Sydney, Sydney, New South Wales 2006, Australia Fax: 0061 (0)2 9351 6556 e-mail: cramo@anatomy.usyd.edu.au C. Ramofa®a á S. C. Battaglene á J. D. Bell International Center for Living Aquatic Resources Management (ICLARM), Coastal Aquaculture Centre, P.O. Box 438, Honiara, Solomon Islands described for other holothurians. An important application of this study is that the appearance of gonad tubules, removed by biopsy, can be used to determine the gonad condition of wild adults or captive broodstock. Introduction Aspidochirote sea cucumbers are a conspicuous component of the macrobenthos of tropical marine environments. The importance of these large, depositfeeding holothurians in benthic processes is well documented (Yingst 1982; Coulon and Jangoux 1993; Uthicke 1999). In addition to their ecological importance, aspidochirotes support numerous artisanal ®sheries for beche-de-mer throughout the Indo-Paci®c (Conand and Byrne 1993; Preston 1993; Conand 1997). Beche-de-mer is derived by processing the body wall of the sea cucumbers, and is exported mainly to China and Singapore. Global export trends indicate that wild stocks of aspidochirote sea cucumbers are currently over-exploited (Conand and Byrne 1993; Conand 1997). In the Indo-Paci®c region, where the sea cucumber ®shery bene®ts coastal communities in developing countries, emphasis is now being given to sustainable use of holothurian resources. Possible management measures include minimum size limits, closed seasons, bag limits and restrictions on the use of SCUBA for harvesting (Preston 1993; Conand 1997). However, the artisanal nature of the ®shery makes implementation of such measures dicult (Conand 1997; Battaglene and Bell 1999). Consequently, release of juvenile sea cucumbers reared in hatcheries is being assessed as an alternative method to restore and enhance wild stocks (Preston 1994; Munro and Bell 1997; Battagelene and Bell 1999), and recent studies indicate that some tropical species are particularly suitable for stock restoration and enhancement programmes (Ramofa®a et al. 1997; Battaglene 1999; Battaglene and Bell 1999; Battaglene et al. 1999). However, the success of such programmes 1046 depends on a thorough understanding of the biology and ecology of these animals, including knowledge of their reproductive cycle. Holothuria fuscogilva, commonly known as the ``white teat®sh'', is the most prized of the commercially important aspidochirote species (Holland 1994). It occurs throughout the Indo-Paci®c and has a patchy distribution in seagrass beds, on reef slopes, and in lagoons at depths of 3 to 40 m (Conand 1981, 1993a; Reichenbach 1999). Previous studies report that the annual reproductive cycle H. fuscogilva is variable, with spawning in December and January in New Caledonia, and from December to March during the northeast monsoon season in the Maldives (Conand 1981, 1993a; Reichenbach 1999). As part of a research project to assess the potential for releasing cultured juveniles to manage stocks, we determined the reproductive cycle of Holothuria fuscogilva in the Solomon Islands. Reproduction in this species was examined using the gonad index (GI) method, by histological examination of the gonads, and by induction of spawning. The macroscopic appearance of the gonads was examined as a rapid means of assessing reproductive maturity. The ®ndings reported here extend the knowledge of reproduction in tropical aspidochirotes, and provide important information for the culture of H. fuscogilva. Materials and methods Reproduction of Holothuria fuscogilva (Selenka, 1867) was investigated in Marau Sound (Fig. 1), Guadalcanal, Solomon Islands Fig. 1 Four collection sites (d) of Holothuria fuscogilva in Marau Sound in 1994 to 1997 (1: histological examination and spawning trials) and one site in Tulagi in 1997 and 1998 (2: spawning trials) (Shaded areas reefs; insert map of Central and Western Solomon Islands) (09°50¢S; 160°49¢E) over a 4 yr period. Specimens were collected at depths of 25 to 30 m from the lagoon ¯oor using SCUBA. As H. fuscogilva display no external sexual dimorphism, an attempt was made to collect at least 20 individuals each month to ensure that both females and males were obtained. Specimens were collected from February 1994 to February 1997; in May 1997; from July to November 1997; and in March 1998. The samples comprised individuals from a broad size range (1000 to 3000 g drained weight). We were unable to reduce this size range because of dif®culty in collecting specimens (density approximated 1/1000 m2). Each sea cucumber was dissected in the ®eld, and drained and weighed as described by Conand (1981) and Sewell and Bergquist (1990). The gonad was ®xed in 7% bu€ered formalin. Gonad wet weight was determined, and its maturity was assessed according to a ®ve-stage maturity scale (see Table 1). This scale was based on macro- and microscopic features including: tubule length and diameter, tubule branching, gonad colour, and presence of gametes and phagocytes. For each gonad, the length and diameter of 15 tubules selected at random were measured to the nearest millimetre. Smears from each tubule were examined microscopically for the presence of gametes and phagocytes. A gonad index (GI) was used to document the changes associated with gonad development of Holothuria fuscogilva. Although the e€ectiveness of the GI method is reduced for samples with a broad size range (Gonor 1972; Grant and Tyler 1983), it proved useful for H. fuscogilva because of its strong and consistent seasonal GI pattern over the 4 yr study. GI was expressed as a ratio of preserved wet gonad weight to wet drained body weight ´ 100, and was determined for all females and males in each monthly sample. Mean monthly GI was calculated for each sex. A balanced, twoway, ®xed-e€ects model ANOVA was used to assess variation in GI maxima (n = 9) between sexes and years (SeptemberpGI  data for 1994, 1995 and 1996). For analysis, the data were x-transformed to achieve homogeneity of variance (assessed using Cochran's test). Histology of gametogenesis was determined for gonad samples collected between May 1996 and November 1997 and in March 1998. Gonads preserved in formalin were rinsed in tap water and stored in 70% ethanol. For histology, tubules were dehydrated, embedded in paran, sectioned (6 lm thick), and stained with 1047 Table 1 Holothuria fuscogilva. Five maturity stages in reproductive cycle, based on morphology of gonad tubules, and corresponding gametogenic stages identi®ed by histology Maturity stage, Sex Gonad Tubule wt (g) Length Diam. (mm) (mm) Branching Condition Colour Histological stagea I Indeterminate 1±5 5±20 <0.30 0±1 Gametes not evident White Recovery Female Male 20±50 20±60 0.5±1.0 0.5±1.0 1±3 1±3 Growing oocytes (30±80 lm) Sperm developing White Growing Oocytes visible through thin tubule Translucent wall (140±170 lm) Tubules packed with sperm Creamy white Mature Tubules appeared Tubules reduced, relict oocytes brown present, empty lumen visible. Phagocytes present, small oocytes observed Unspawned tubules densely packed with sperm. Spawned tubules empty and reduced Partly spawned II 10±40 10±30 III Female 50±150 50±120 1±2.5 1±4 Male 50±100 50±180 1±2.0 1±4 IV Female 50±150 50±120 1±2.0 1±4 Male 50±100 50±180 1±2.0 1±4 Female 10±50 10±30 <1.00 1±3 Male 10±40 10±50 <0.80 1±4 V a Relict oocytes may be present, but tubules shrunken and wrinkled and reduced in size Shrunken and wrinkled tubules, relict sperm present, brown spots occasionally developed. White, or brown Spent Gonad basis with brown colouration See ``Results ± Histology'' for further details haemotoxylin and eosin (H/E). Five gametogeneic stages were de®ned: recovery, growing, mature, partly-spawned and spent, similar to the stages used in other studies of holothurian reproduction (Tanaka 1958; Sewell 1992). Some gonad sections were stained by the periodic acid±Schi€ (PAS) method, which stains echinoderm yolk and haemal ¯uid. Characterisation of each gametogenic stage was based on development and staining properties of the germinal epithelium. For ovaries, we evaluated oocyte eosinophilia and PAS staining response. Diameters of 10 oocytes from three ovaries from each of the di€erent stages of vitellogenesis were measured. Histology was used to determine whether the ®ve-stage maturity scale based on macroscopic appearance of the tubules represented meaningful di€erences in the gametogenic state of gonads. Spawning induction trials were conducted during the 1996, 1997 and 1998 breeding seasons. In 1996, specimens were collected from Marau Sound and in 1997 and 1998 specimens were collected from Tulagi (9°6¢S; 160°9¢E) (Fig. 1). In 1996 and 1997, sea cucumbers were induced to spawn by heat shock, i.e. raising water temperature by 2 to 3 C° above ambient (Ramofa®a et al. 1995). Specimens collected in 1998 were induced to spawn by the addition of the dried alga Schizochytrium sp. to holding tanks (Battaglene 1999). The proportions of individuals induced to spawn were recorded. Data on annual variation in day length were obtained from the Ministry of Meteorology, Solomon Islands, and data on water temperature (3 m depth) were obtained from giant-clam grow-out farms 3 km west of the Marau collection site for 1995 and 1996. Results A total of 724 Holothuria fuscogilva were examined from Marau Sound. These consisted of 319 females, 371 males, 11 specimens of indeterminate sex, and 23 individuals lacking a gonad. Most of the indeterminate individuals were encountered in April, and individuals lacking gonads were collected in February and April. The mean drained weight of females, males, individuals of indeterminate sex and specimens with no gonad were 1849.35 g (SE = 17.97), 1834.65 g (SE = 24.61), 1591.36 g (SE = 103.76) and 1408.61 g (SE = 107.89), respectively. During the breeding season, all H. fuscogilva examined (1000 to 3000 g drained weight) had gonads, indicating that the specimens lacking gonads in February to April would have had gonads during the previous breeding season (August to October, see subsection ``Gonad index''). The sex ratio of the gonochoric H. fuscogilva did not di€er from unity (v2 = 3.90; P > 0.05). Gonad morphology The gonad of Holothuria fuscogilva was a single structure consisting of numerous branched tubules arising from the gonad basis attached to the anterior body wall (Fig 2A). The gonoduct opened externally at the gonopore, dorsally above the mouth. During gonad development, branched tubules extended into the perivisceral cavity and dominated the cavity when gravid. The ®ve stages of gonad development based on tubule size and appearance are detailed in Table 1. 1048 Gonad growth in Holothuria fuscogilva involved formation of new tubules arising from the gonad basis with subsequent increase in tubule length and diameter (Table 1). In the initial stage of gonad growth (Stage I), the tubules had a mean length of 17.0 mm (SE = 0.4, n = 150) and, in general, were unbranched. Sex could not be determined at this stage because gametes were not evident in gonad smears. As gonad growth progressed (Stage II), females and males could be identi®ed by the presence of developing eggs and sperm (Fig. 2C, B, respectively). As the gonads approached maturity, the sex of specimens could be also be determined by gonad colour. Growing testes appeared creamy white, and tubules had a uniform appearance. When mature (Stage III), tubules were packed with spermatozoa and a variable number of tubules with an irregular beaded appearance were occasionally seen (Fig. 2B). These beaded tubules were observed prior to spawning, so their irregular appearance was not due to partial release of sperm. However, variable beaded tubules may persist after spawning. Mature ovaries were mustard in colour, and individual tubules had transparent thin tubule walls through which oocytes were evident (Fig. 2C, D). Tubule length was a good indicator of reproductive Fig. 2 Holothuria fuscogilva. Gonad anatomy. A Gonad basis (GB) and tubule attachment in partly-spawned male; B mature branched testis tubule with beaded appearance; C mature ovary tubule with oocytes clearly visible through gonad wall; D mature oocytes released from gravid tubule; E spawned (St) and unspawned (Ut) tubules of partly-spawned female; F spent ovarian tubules with wrinkled and shrunken appearance (Scale bars in A, B, E = 235 lm; in C, D = 897 lm; in F = 200 lm) maturity (Fig. 3), with the longest tubules present at the mature stage (females: x = 78.61 mm long, SE = 2.80, n = 112; males: x = 87.48 mm long, SE = 3.4 n = 102). Through the spawning season, the simultaneous presence of both spawned and unspawned tubules (Stage IV) indicated that partial spawning was characteristic of Holothuria fuscogilva (Fig. 2E). Examination of gonad smears revealed the presence of abundant phagocytes and oocyte debris in unspawned tubules, indicating that the gametes could be resorbed. Spent gonad tubules (Stage V) were wrinkled and greatly reduced in size (Fig. 2F). They occasionally developed a brown colour because of the presence brown bodies in the connective tissue of the gonad wall and in the lumen. Gonad index The mean monthly GI for female and male Holothuria fuscogilva displayed a distinct seasonal pattern (Fig. 4). In both sexes, the GI displayed synchronous gonad development, marked by a gradual increase in mean GI beginning in May or June, and a maximum in August, 1049 Fig. 4 Holothuria fuscogilva. Monthly variation in GI of females (d) and males (s) from Solomon islands over 4 yr (Vertical bars standard errors of mean) stained with PAS. The histological stages of oogenesis are as follows: Fig. 3 Holothuria fuscogilva. Monthly variation in tubule length and tubule diameter of females (d) and males (s) over 4 yr (Vertical bars standard errors of mean) September or October. The subsequent sharp decline in the GI indicated that most gamete release was complete by November (Fig. 4). During the spawning season, the GIs for male H. fuscogilva were consistently lower than those for females (Fig. 4). Over 4 yr, the maximum GI of females and males ranged from 3.69 to 4.59 and 2.70 to 2.84, respectively. The two-way ANOVA revealed that this di€erence between sexes was signi®cant in all years of the study (F = 5.30; P < 0.05). It also indicated that there was no signi®cant variation in GI among years (F = 1.50; P > 0.05). Histology Histology revealed that the ®ve gonad maturity stages designated by gonad tubule size and appearance correlated with the ®ve stages of gametogenic development. A description of the histological features of each gametogenic stage is detailed below. Females Based on their staining responses to PAS and H/E, the developing oocytes were categorized as pre(x = 13 lm, SE = 0.8, n = 30), early (x = 40.1 lm, SE = 0.7, n = 30) mid (x = 60.1 lm, SE = 1.0, n = 30)- or late-vitellogenic (x = 71.1 lm, SE = 1.26, n = 30). Previtellogenic and vitellogenic oocytes maintained their position along the germinal epithelium (Fig. 5A±D). Haemal ¯uid was not seen in sections Stage I: recovery. Ovaries in the recovery stage (Fig. 5A) had thick walls. Previtellogenic oocytes were basophilic and PAS-negative. In some sections, early vitellogenic oocytes were observed along the germinal epithelium and were still strongly basophilic, although the presence of yolk was evident among the scattered PAS+ granules. Relict oocytes from the previous reproductive season and phagocytes were occasionally seen. Stage II: growing. The growing stage (Fig. 5B) was characterised by active vitellogenesis. Early and midvitellogenic oocytes were abundant. These oocytes had a distinct germinal vesicle which persisted until spawning. Vitellogenic oocytes were surrounded by follicle cells throughout development. The gonad wall was greatly reduced in thickness as oogenesis progressed towards maturity. Stage III: mature. Mature ovaries were densely packed with PAS+ and eosinophilic late-vitellogenic oocytes, and the gonad wall was thin (Fig. 5C). The oocytes remained within their follicle, and the germinal vesicle started to take up an eccentric position towards the basal protuberance (Fig. 5D). The protuberance was PAS-negative and contained ®lamentous strands. Stage IV: partly spawned. Not all ovarian tubules released gametes during spawning. Partly spawned ovaries contained both spawned and unspawned tubules (Fig. 5E, F). Although phagocytes were present in spawned and unspawned tubules, there were more of these cells in the latter (Fig. 5F). Spawned tubules had a reduced diameter and a wrinkled appearance. The ovary wall gradually increased in thickness. In some specimens, gametogenesis was re-initiated and previtellogenic and early vitellogenic oocytes lined the germinal epithelium. Relict oocytes and debris were occasionally observed in the lumen (Fig. 5F, G, H). Stage V: spent. Spent ovaries were wrinkled and shrunken, with relict oocytes occasionally present in the 1050 Fig. 5 Holothuria fuscogilva. Oogenesis. A, B Recovering and early growing ovaries with previtellogenic (pv) and early (ev) vitellogenic oocytes. C, D Mature ovary with oocytes enclosed within their follicle (f) and germinal vesicle located eccentrically towards protuberance (p) which attached to germinal epithelium. E, F Partly-spawned ovary; many unspawned ova persisted during this stage; phagocytes (ph) were especially abundant in unspawned tubules (F). G, H Spent ovaries; intensive shrinkage of tubules occurred and a few relict oocytes persisted; in some sections, sex could be identi®ed only by presence of one or two oocytes (H) (Scale bars in A, D = 84 lm; in B, C, F = 185 lm; in E, G, H = 320 lm) lumen (Fig. 5G, H). Phagocytes were evident, and the gonad wall was thick. present along the germinal epithelium, and the gonad wall was at its maximum thickness (Fig. 6A). The lumen was empty. Males Stage II: growing. A striking feature of growing testes was the numerous infolds of the germinal epithelium, which extended for some distance into the lumen (Fig. 6B, C). These infolds were lined by a dense layer of Stage I: recovery. Testes in recovery stage could only be identi®ed by histology. A layer of spermatogonia was 1051 Fig. 6 Holothuria fuscogilva. Spermatogenesis. A Recovering testis with early spermatocytes lining germinal epithelium. B, C Growing testes with infolds of germinal epithelium (arrowed) and sperm developing spermatocyte columns (sc); spermatozoa began to ®ll lumen as growth progressed. D, E Mature testes; in early mature stage (D) spermatocytes persist along gonad wall, but are absent from fully matured testes (E). F Partly-spawned testis with spermatozoa in lumen. G, H Spent testes with residual spermatozoa or empty lumena. (Scale bars in A, C, D = 84 lm; in F, H = 185 lm; in B, F = 320 lm; in E = 741 lm) spermatocytes organized in short columns. In late growing-stage testes, the germinal infolds were reduced and spermatozoa were abundant in the lumen. The thickness of the gonad wall was greatly reduced. Stage III: mature. In mature testes, the infolds of the germinal epithelium were reduced or absent and the lumen was packed with spermatozoa (Fig. 6D, E). A few spermatocytes could still be present along the germinal epithelium (Fig. 6D). The gonad wall was at its minimal thickness. Stage IV: partly spawned. As for the ovary, partly spawned testes contained tubules that had, and those that had not, spawned. Spawning activity was indicated by a di€use arrangement of spermatozoa in the lumen. There was also a reduction in the size of the tubules, which were wrinkled and shrunken in appearance. Dense aggregations of spermatozoa and phagocytes were present in the lumen (Fig. 6F). Spermatocytes along the germinal epithelium indicated the onset of gamete oogenesis in some tubules. Some testes had a distinct gap between this spermatocyte layer and the spermatozoa in the lumen. Stage V: spent. Spent tubules were shrunken, and generally had an empty lumen except for a few relict 1052 Fig. 7 Holothuria fuscogilva. Gametogenic cycle of females (a) and males (b), collected from May 1996 to November 1998. Histograms showing percentage of individuals with gonads in each of ®ve maturation stages. (n sample size for each month; samples containing <5 females or males are not included) spermatozoa (Fig. 6G, H). Spermatogonia were scattered along the germinal epithelium. The gonad wall was thick, and the germinal epithelium once again took on its convoluted appearance in some specimens. Reproductive cycle The percentage of individuals at each gametogenic stage in monthly samples collected over 2 yr is shown in Fig. 7. By May of each year, most gonads contained new cohorts of developing gametes. Examination of gonads from March 1998 (n = 18) revealed that both spent (39%) and recovery (61%) gametogenic stages were present, indicating that re-initiation of gametogenesis 1053 occurred in March/April. This approximates the time when day becomes shorter than night (Fig. 8). In both sexes, gametogenesis occurred in parallel (Fig. 7), resulting in a similar increase in GI (Fig. 4). The rapid increase in GI from May to August coincided with active vitellogenesis and spermatogenesis, with maturity reached in August (Figs. 4, 7). Histology revealed that partial gamete release was underway by August (Fig. 7). Partial spawning was also characteristic of the specimens collected in September and October. The marked decrease in GI between September/October and November, and the presence of spent gonads in November indicated that most spawning activity occurred in September or October. Individuals with spent gonads and low GI were found from November to March/April (Figs. 4, 7). Fig. 8 Holothuria fuscogilva. Annual day-length cycle for Solomon Islands Induction of spawning Over 3 yr, spawning of Holothuria fuscogilva was induced successfully in August, September and October, particularly in males (Fig. 9). In 1996 and 1997, heatshock elicited gamete release in 10% of individuals (n = 134). Addition of the dried alga Schizochytrium sp. elicited gamete release in 36% of individuals (n = 47). Both methods were most successful in inducing spawning in September and October. The breeding period of Holothuria fuscogilva coincided with an increase in day length, which begins in August (Fig. 8). There was no clear relationship, however, between spawning and water temperature. Temperature varied annually between 28 and 31 °C. Discussion The reproductive cycle of Holothuria fuscogilva in the Solomon Islands was similar to that in New Caledonia (Conand 1981, 1993a) and the Maldives (Reichenbach 1999) in that spawning occurred in summer. However, the onset of spawning di€ered, starting in August in the Solomon Islands, November in New Caledonia, and December in the Maldives. In addition, mature individuals were present for 3 to 5 mo in the Solomon Islands, whereas they occurred throughout the year in the Maldives, with the greatest numbers recorded from August to December (Reichenbach 1999). The seasonal reproduction of Holothuria fuscogilva is typical of that of many tropical holothurians (Conand 1981, 1982, 1993a, b; Harriot 1985; Hopper et al. 1998; Reichenbach 1999), and contradicts the widespread assumption of year-round spawning of marine invertebrates in the tropics (Giese and Kanatani 1987; but see Ong Che and Gomez (1985), Ong Che (1990) and Chao et al. (1995) for species that have prolonged reproduction). Water temperature (Tanaka 1958; Conand 1981), photoperiod (Conand 1982; Cameron and Fankboner Fig. 9 Holothuria fuscogilva. Percentage of females and males induced to spawn from August to October. Sample size (n) combines data from 1996 to 1998 1986), water velocity (Engstrom 1980), salinity (Krishnaswamy and Krisnan 1967), a combination of water temperature and photoperiod (Costelloe 1985) and phytoplankton blooms (Hamel et al. 1993) have all been implicated as factors controlling gametogenesis and/or spawning in sea cucumbers. The onset of gametogenesis by H. fuscogilva in Marau coincided with the in¯ection point in March when the light period becomes shorter than the dark period, indicating that photoperiod may entrain gonad development, as documented for sea urchins (Pearse et al. 1986; Byrne et al. 1998; Walker and Lesser 1998). Spawning may be entrained to a lunar cue, with the crepuscular period de®ning the time of gamete release, as reported for a variety of echinoderms (Babcock et al. 1992; Byrne et al. 1998). Tanaka (1958) suggested that temperature may control spawning in holothurians, and Conand (1981, 1982, 1993a, b) demonstrated a consistent relationship between temperature and spawning in some tropical species. In the Solomon Islands there is no clear seasonal changes in sea temperature during the breeding season of Holothuria fuscogilva. Spawning was, however, induced by heatshock, indicating that H. fuscogilva may release gametes in situ in response to short-term heat stress. Although heat-shock is often used to induce spawning in holothurians, the mechanism by which this stimulus promotes 1054 oocyte maturation, ovulation and spawning is not known (McEuen 1987; Ramofa®a et al. 1995). Induction of spawning in H. fuscogilva in response to introduction of the dried alga, Schizochytrium sp., to holding tanks supports the supposition that phytoplankton exudates may in¯uence spawning in the wild. Spawning in response to phytoplankton has been reported for other sea cucumbers and sea urchins (Cameron and Fankboner 1986; Hamel et al. 1993; Starr et al. 1990). Histology revealed that many individuals initiated gamete release before the GI peaked, and that many partially spawned gonads contained numerous phagocytes. This demonstrates that changes in GI alone cannot be used to estimate the timing and duration of gamete release in Holothuria fuscogilva. On the other hand, there was good correlation between the maturity stages based on tubule size and appearance, and those based on histology. The fact that tubule morphology can be used to detect release of gametes is of great bene®t to breeding programmes. In particular, it means that the biopsy techniques described by Yanigasawa (1998) and Reichenbach (1999), in which tubules can be removed through a dorsal incision in the body wall near the gonopore, can be used to assess the condition of broodstock. This assessment method will be especially important in remote locations or when the sea cucumber are broodstock that cannot be sacri®ced. Gametogenesis in Holothuria fuscogilva was similar to that documented for other holothurians (Smiley et al. 1991; Sewell et al. 1997). Initiation of gametogenesis in the March/April period, and subsequent gametogenic maturation and gamete release occurred in parallel for both sexes. Like several other sea cucumber species (Sewell 1992; Hamel et al. 1993), spawning in H. fuscogilva is partial and it is not known if unspawned tubules eventually release their gametes. However, as seen for Psolus fabricii, the abundance of phagocytes in unspawned tubules suggested strong phagocytic activity, with potential sequestration of material for storage and eventual use in gametogenesis (Hamel et al. 1993). Reinitiation of gametogenesis in spawned and unspawned tubules resulted in the presence of overlapping generations of immature and relict oocytes, as documented for other holothurians (Sewell et al. 1997). Despite this gametogenic renewal, all the gonad tubules in H. fuscogilva were resorbed to the basis after cessation of breeding, and thus any immature oocytes present in the gonads were also resorbed. The rationale underlying reinitiation of gametogenesis immediately prior to tubule resorption is not clear, but has been reported for several holothurians (Choe 1963; Cameron and Fankboner 1986; Sewell 1992; Hamel et al. 1993). This phenomenon shows that the tubule recruitment model proposed by Smiley (1988), based on Parastichopus californicus, is not applicable to Holothuria fuscogilva. Unlike P. californicus, in which tubules are recruited progressively, all gonad tubules in H. fuscogilva were at a similar stage of development. Furthermore, H. fuscogilva di€ered from the tubule recruitment model in the presence of overlapping cohorts of oocytes in spawned tubules. In their review of gonad structure in the Holothuroidea, Sewell et al. (1997) noted that no tropical aspidochirote species is known to conform to the model. A feature of testis development in Holothuria fuscogilva, the presence of numerous longitudinal folds in the germinal epithelium, appears typical of holothurian spermatogenesis (review: Smiley et al. 1991). Two functions have been suggested for these folds; to increase the surface area for proliferation of spermatogonia (Cameron and Fankboner 1986), and to provide a reservoir of nutrients in the haemal ¯uid to support spermatogenesis (Smiley et al. 1991). The occurrence of Holothuria fuscogilva lacking gonads in the February to April samples shows that complete resorption of the gonad occasionally occurs at the end of breeding. This result was not a result of sampling individuals below the size at ®rst maturity. This phenomenon appears to be rapid, short-lived and variable among the population. As noted for H. fuscogilva in the Solomon Islands, individuals with indeterminate gonads also occur in April in New Caledonia (Conand 1981). Annual gonad resorption has also been reported for Stichopus mollis (Sewell 1992) and S. japonicus (Choe 1963). In S. mollis, total resorption occurs in populations at a lower latitude and higher water temperature, whilst populations at higher latitude undergo incomplete resorption and maintain a resting phase. These contrasting populations subsequently redeveloped their gonads in time to spawn in the next season. Our data di€ered from those of Sewell in that individuals lacking gonads or with indeterminate gonads were present in the population at the same time. We suggest that total gonad resorption may occur annually in H. fuscogilva, followed by rapid gonadal regeneration. This however, is asynchronous among individuals. A weekly sampling programme during the February to April period is required in order to verify this suggestion. In conclusion, we have shown that the breeding season of Holothuria fuscogilva in the Solomon Islands occurs from August to October, and that gonad tubule morphology provides a relatively rapid and reliable method of assessing reproductive condition without sacri®cing valuable and scarce animals. Biopsy of gonad tubules should prove to be particularly e€ective in assessing the availability of gametes from captive broodstock used in breeding programmes to restore and enhance wild ®sheries. Acknowledgements We thank the people of Marau for permission to collect specimens, special thanks to M. Ukaria and the late D. Adilamo who assisted us to collect specimens in Marau Sound. We are also grateful to sta€ at the ICLARM Coastal Aquaculture Centre, especially M. Sau, H. Rota, F. Lasi, H. Tafea and A. Hart. Financial support was provided by the Paci®c Science Foundation, the Australian Centre for International Agricultural Research,an AusAID Postgraduate Scholarship to CR and an Australian Research Council grant to MB. ICLARM Contribution No. 1545. 1055 References Babcock R, Mundy C, Keesing J, Oliver J (1992) Predictable and unpredictable spawning events: in situ behavioural data from free-spawning coral reef invertebrates. Invert Reprod Dev 22: 213±228 Battaglene SC (1999) Culture of tropical sea cucumbers for the purpose of stock restoration and enhancement. In: Baine M (ed) The conservation sea cucumbers in Malaysia: their taxonomy, ecology and trade. Heriot-Watt University, Stromness, Orkney Islands, Scotland, pp 11±25 (Proc int Conf Dep Agric, Kuala Lumpur, Malaysia, Feb. 1999) Battaglene SC, Bell JD (1999) Potential of the tropical Indo-Paci®c sea cucumber, Holothuria scabra, for stock enhancement. In: Howell BR, Moksness E, Svasand T (eds) Proceedings of the First International Symposium on Stock Enhancement and Sea Ranching. Fishing News Books, Blackwell Science, Oxford, pp 478±490 Battaglene SC, Seymour EJ, Ramofa®a C (1999) Survival and growth of cultured juvenile sea cucumbers, Holothuria scabra. Aquaculture, Amsterdam 178: 293±322 Byrne M, Andrew NL, Worthington DG (1998) Reproduction in the diadematoid sea urchin Centrostephanus rodgersii in contrasting habitats along the coast of New South Wales, Australia. Mar Biol 132: 305±318 Cameron JL, Fankboner PV (1986) Reproductive biology of the commercial sea cucumber Parastichopus californicus (Stimpson) (Echinodermata: Holothuridea). I. Reproductive periodicity and spawning behaviour. Can J Zool 64: 168±175 Chao S-M, Chen C-P, Alexander PS (1995) Reproductive cycles of tropical sea cucumbers (Echinoderma: Holothuridea) in southern Taiwan. Mar Biol 122: 289±295 Choe S (1963) Namako no kenkyu [Biology of the Japanese common sea cucumber Stichopus japanicus Selenka]. Pusan National University, pusan, Japan [in Japanese with English summary] Conand C (1981) Sexual reproduction of three commercially important holothurian species (Echinodermata) from the lagoon of New Caledonia. Bull mar Sci 31: 523±543 Conand C (1982) Reproductive cycle and biometric relations in a population of Actinopyga echinites (Echinodermata: Holothuroidea) from the lagoon of New Caledonia, western tropical Paci®c. In: Lawrence JM (ed) International Echinoderm Conference. Balkema, Rotterdam pp 437±442 Conand C (1993a) Reproductive biology of the holothurians from the major communities of the New Caledonian Lagoon. Mar biol 116: 439±450 Conand C (1993b) Ecology and reproductive biology of Stichopus variegatus an indo-paci®c coral reef sea cucumber (Echinodermata: Holothuroidea). Bull mar Sci 52: 970±981 Conand C (1997) Are holothurian ®sheries for export sustainable? Proc 8th int coral Reef Symp 2: 2021±2026 [Lessios HA, Macintyre I (eds) Smithsonian Tropical Research Institute, PanamaÂ] Conand C, Byrne M (1993) A review of recent developments in the world sea cucumber ®sheries. Mar Fish Rev 55: 1±13 Costelloe J (1985) The annual reproductive cycle of the holothurian Aslia lefevrei (Dendrochirota: Echinodermata). Mar Biol 88: 155±165 Coulon P, Jangoux M (1993) Feeding rate and sediment reworking by the holothuroid Holothuria tubulosa (Echinodermata) in the Mediterranean seagrass bed o€ Ischia Island, Italy. Mar Ecol Prog Ser 92: 201±204 Engstrom NA (1980) Reproductive cycles of Holothuria (Halodeimea) ¯oridana, H. (H.) mexicana and their hybrids in Southern Florida, USA. Invert Reprod Dev 2: 237±244 Giese AC, Kanatani H (1987) Maturation and spawning. In: Giese AC, Pearse JS, Pearse VB (eds) Reproduction of marine invertebrates: general aspects ± seeking unity in diversity. Boxwood Press, Paci®c Grove, California, pp 251±329 Gonor JJ (1972) Gonad growth in the sea urchin Stronylocentrotus purpuratus (Stimpson) (Echinodermata: Echinoidea) and the assumptions of the gonad index methods. J exp mar Biol Ecol 10: 89±103 Grant A, Tyler PA (1983) The analysis of data in studies of invertebrate reproduction. II. The analysis of oocyte size frequency data, and comparision of di€erent types of data. Invert Reprod Dev 6: 271±283 Hamel J-F, Himmelman JH, Dufresne L (1993) Gametogenesis and spawning of the sea cucumber Psolus fabricii (Duben and Koren). Biol Bull mar biol Lab, Woods Hole 184: 125±143 Harriot VJ (1985) Reproductive biology of three congeneric sea cucumber species Holothuria atra, H. impatiens, and H. edulis, at Heron Reef, Great Barrier Reef. Mar Freshwat Res 36: 51±57 Holland A (1994) The status of global beche-de-mer ®sheries with special reference to Solomon Islands and the potentials of holothurian culture. MSc thesis. University of NewCastle upon Tyne, Durham, England Hopper DR, Hunter CL, Richmond RH (1998) Sexual reproduction of the tropical sea cucumber, Actinopyga mauritiana (Echinodermata: Holothuroidea) in Guam. Bull mar Sci 63: 1±9 Krishnaswamy S, Krishnan S (1967) A report on the reproductive cycle of the holothurian Holothuria scabra Jaeger. Curr Sci 36: 155±156 McEuen FS (1987) Phylum Echinodermata, class Holothuroidea In: Strathmann MF (ed) Reproduction and development of marine invertebrates of the northern Paci®c coast. University of Washington Press, Seattle pp 574±596 Munro JL, Bell JD (1997) Enhancement of marine ®sheries resources. Rev Fish Sci 5: 185±222 Ong Che RG (1990) Reproductive cycle of Holothuria leucospilata Brandt (Echinodermata: Holothuroidea) in Hong Kong and the role of body tissues in reproduction. Asian mar Biol 7: 115±132 Ong Che RG, Gomez ED (1985) Reproductive periodicity of Holothuria scabra Jaeger at Calatagan, Batangas, Philippines. Asian mar Biol 2: 21±30 Pearse JS, Pearse VB, Davis KK (1986) Photoperiodic regulation of gametogenesis and growth in the sea urchin Strongylocentrotus purpuratus. J exp Zool 237: 107±118 Preston GL (1993) Beche-de-mer In: Wright A, Hill L (eds) Nearshore marine resources of the South Paci®c. Institute of Paci®c Studies, Suva, Forum Fisheries Agency, Honiara, International Centre for Ocean Development, Honiara, Canada pp 119±128 Preston GL (1994) Enhancement of the Paci®c islands invertebrate ®sheries by restocking. Proc 44th int Center living aquat Resour Mgmt (ICLARM) Conf 44: 88±91. [Munro JL, Munro PE (eds)] Ramofa®a C, Foyle TP, Bell JD (1997) Growth of juvenile Actinopyga mauritiana (Holothuroidea) in captivity. Aquaculture, Amsterdam 152: 119 ± 128 Ramofa®a C, Gervis M, Bell JD (1995) Spawning and early larval rearing of Holothuria atra. SPC Inf Bull Beche-de-mer 7: 2±5 Reichenbach N (1999) Ecology and ®shery biology of Holothuria fuscogilva (Echinodermata: Holothuroidea) in the Maldives. Bull mar Sci 64: 103±113 Sewell MA (1992) Reproduction of the temperate aspidochirote Stichopus mollis (Echinodermata: Holothuroidea) in New Zealand. Ophelia 35: 103±121 Sewell MA, Bergquist PR (1990) Variability in the reproductive cycle of Stichopus mollis (Echinodermata: Holothuroidea). Invert Reprod Dev 17: 1±7 Sewell MA, Tyler PA, Young CM, Conand C (1997) Ovarian development in the Class Holothuroidea: a reassessment of the `tubule recruitment model'. Biol Bull mar biol Lab, Woods Hole 192: 17±26 Smiley S (1988) The dynamics of oogenesis and the annual ovarian cycle of Stichopus californicus (Echinodermata: Holothuroidea). Biol Bull mar biol Lab, Woods Hole 175: 79±93 Smiley S, McEuen F-S, Cha€ee C, Krishan S (1991) Echinodermata: Holothuroidea. In: Giese AC, Pearse JS, Pearse VB (eds) Reproduction of marine invertebrates. Vol VI. Echinoderms 1056 and lophophorates. Boxwood Press, Paci®c grove, California, pp 663±750 Starr M, Himmelman JH, Therriault J-C (1990) Direct coupling of marine invertebrate spawning with phytoplankton blooms. Science, NY 247: 1071±1074 Tanaka Y (1958) Seasonal changes in the gonad of Stichopus japonicus. Bull Fac Fish Hokkaido Univ 9: 29±36 Uthicke S (1999) Sediment bioturbation and impact of feeding activity of Holothuria (Halodeima) atra and Stichopus chloronotus, two sediment feeding holothurians, at Lizard Island, Great Barrier Reef. Bul mar Sci 64: 124±141 Walker CW, Lesser MP (1998) Manipulation of food and photoperiod promotes out-of-season gametogenesis in the green sea urchin, Strongylocentrotus droebachiensis: implications for aquaculture. Mar Biol 132: 663±676 Yanagisawa T (1998) Aspects of the biology and culture of the sea cucumber: In: De Silva SS (ed) Tropical mariculture. Academic Press, London, pp 292±308 Yingst JY (1982) Factors in¯uencing rates of sediment ingestion by Parastichopus parvinensis (Clark), an epibenthic deposit-feeding holothurian. Estuar cstl, Shelf Sci 14: 119± 134