Chapter 6. Sexual Reproduction
Most asteroids possess 10 gonads, two in each ray with gonoducts opening in the interradii. In some coral-reef genera, such as Linckia and Nardoa, the gonads are arranged serially with numerous gonoducts existing along the length of the arm. When an individual is ready to spawn, the gonads can occupy the whole length of a ray. In most species, the sexes are separate, but hermaphroditism has been reported in Asterina burtoni (Achituv, 1972; Achituv and Malik, 1985; Achituv and Sher, 1991). In all coral-reef species studied previously, fertilisation is external with gametes being released directly into the water. An off-reef species, Euretaster insignis, belongs to the family Pterasteridae, other members of which are known to brood their young within the supra-dorsal membrane (McClary and Mladenov, 1989; McClary, 1990). The spawning of one individual might trigger other individuals to spawn, thus increasing the chance of fertilisation (Okaji, 1991). Alternately, synchronous spawning might be triggered extrinsically (see Yamaguchi and Lucas, 1984; Minchin, 1987).
The sexual reproductive cycle has been studied in some of the commoner coral-reef species e.g. Asterina burtoni, (by Achituv, 1972; James, 1972; Achituv and Malik, 1985), Linckia laevigata (by Yamaguchi, 1977 a), and Ophidiaster granifer (by Yamaguchi and Lucas, 1984). The type of larval development has also been studied in several species e.g. Astropecten polyacanthus (by Oguro et al., 1975), Acanthaster planci (by e.g. Henderson and Lucas, 1971), Gomophia egyptiaca (by Yamaguchi, 1974), Leiaster leachi (by Komatsu, 1973), Ophidiaster granifer, Ophidiaster robillardi and Ophidiaster squamous (by Yamaguchi and Lucas, 1984). The known forms of reproduction along with the type of larval development of most Guam species are tabled by Yamaguchi (1975 b).
The use of the hormone 1-methyl adenine to produce final maturation and subsequent release of gametes in asteroids is well documented (Kanatani, 1969, 1973). Yamaguchi (1977 a) described the injection of the hormone into the coelomic cavity of Linckia laevigata to assess the stage of development of the gonads. This procedure was also used by Yamaguchi and Lucas (1984). The presence and strength of response to treatment have been shown to depend upon the stage of maturation of the gametocytes (Kuborta et al., 1977).
The time required before a response is produced, following treatment, depends upon the proximity of the natural breeding season (Kanatani, 1969). A delayed response in the genus Echinaster was described previously by Turner (1976). Additionally, it is possible that not all individuals within a population are at exactly the same stage of gamete development at any time (see Pearse, 1968).
The reproductive analysis of each of the species entailed the injection of 1-methyl adenine into the arms of a sample of the population several times during each year. The chemical was obtained as anhydrous powder in 10 mg tubes from Sigma Chemicals. The anhydrous powder was kept frozen following delivery. The concentration of 1-methyl adenine used for injection was 0.0001 M. dissolved in sea-water. A working solution was prepared freshly for each sampling period and stored in a refrigerator at 4°C. It was necessary to warm the sea-water temporarily to 40°C to facilitate the dissolution of the chemical as the working solution was being prepared. Once the chemical was dissolved (requiring about five to ten minutes with stirring) it was immediately placed in the refrigerator. The solution was not warmed again before injection into the starfish.
The quantity injected per individual depended on the mean size of the species tested. It ranged from one millilitre per individual in small species such as Ophidiaster granifer to five millilitres per individual in large species such as Linckia laevigata. In large species, the hormone was injected into three of the arms of each animal while it was temporally removed from the aquarium. In smaller species the injection was administered aborally into one interradius. Following injection, each animal was returned to its aquarium along with other conspecifics which had also been injected. The test animals were then observed for several hours and any release of gametes through the gonopores was recorded. To maintain visibility within the test aquaria, specimens were removed and placed in a larger tank once they commenced spawning. Spawning always continued following the transfer. Gametes that were fertilised in aquaria were never released in the field.
The procedure used by Kanatani (1969) required the extraction of gonad for in vitro treatment with 1-methyl adenine. Other methods of determining reproductive periodicity, such as gonad index or histological examination require the test individuals to be killed, and the gonads removed. Many species of coral-reef asteroid occur in low abundance and the regular killing of test individuals would have required the use of much smaller sample sizes to ensure that the population was not reduced by periodic testing. In the present study the convenience of an in vivo treatment, that allowed the rapid testing of a large number of individuals of each species, was considered to outweigh the limitations imposed by the lack of detailed histological information.
Possible variability in response within samples required sample sizes in the vicinity of 20 to 30 individuals to ensure statistical significance of the different spawning frequencies, at different times of the year. Histological study would have provided direct qualitative evidence of gametogenesis. The dichotomous (presence / absence) spawning data obtained in the present study required larger sample sizes to demonstrate any periodicity conclusively. The G Test was used to establish that the observed response in the breeding season was significantly different from the null (low all year round) response. While in some cases, the “expecteds” were as low as two, the G statistic appeared sufficiently high to indicate a significant spawning response.
Release of gametes required no longer than 3 hours after injection except in Echinaster luzonicus. However, for every species that was studied, the response time varied throughout the year. This time ranged from just under three hours, two months before the breeding season, to as little as 15 minutes at the peak of the season. At this peak, if the water temperature within aquaria was allowed to rise above that of the reef flat (as it would on a very hot day), spontaneous spawning was observed in species of Linckia and Nardoa. Spawning in the field was not observed during this study. The result with 1-methyl adenine was always reduced if the test animals had undergone previous spawning.
Tables 6.1a to 6.8a list the spawning response to injection with 1-methyl adenine and Figures 6.1 to 6.8 graph the annual spawning pattern of each of the common species over the study period. Tables 6.1b to 6.8b show the results of the G test, comparing the spawning response in four seasons. Table 6.9 lists the reproductive strategies of each species.
Providing that the water did not become too cloudy, it was always possible to determine the sex of the individuals by the type of gamete released. The size, number and development of eggs was not studied in detail, but varied among species depending on the type of larvae produced.
The results of observations, relating to 1-methyl adenine injections, on the sexual reproductive cycles of the commoner asteroid species at Heron Reef are outlined below.
Females of Fromia elegans release from 100 to 200 very large (approximately 2.0 mm diameter) eggs from gonopores, two of which are located in each interradius. The eggs are bright red in colour and are of neutral buoyancy. Egg release may take up to three hours following injection but can occur after only 30 minutes. The males spawn within one hour and sperm are released through gonopores which are located slightly higher in each interradius than in the female. The fertilised eggs undergo lecithotrophic development. The peak of sexual activity occurred in early summer (November, December).
Females of Linckia guildingii release large numbers of small (approximately 0.1 mm diameter), colourless and negatively buoyant eggs through gonopores located serially along the arms. At no time was the sexually active proportion of the population very high. The peak in sexual activity that was apparent occurred in mid summer (December).
The reproductive cycle of Linckia laevigata, at Guam, was studied by Yamaguchi (1977 a). Females release very large numbers of small (approximately 0.1 mm diameter), colourless and negatively buoyant eggs through gonopores located serially along the arms. At Heron Reef, L. laevigata showed a positive response to treatment (approximately 1 million eggs shed) from mid-winter to mid-summer. Spontaneous spawning occurred only in the summer months (November, December).
Females of Nardoa novaecaledoniae release approximately 1000, large (approximately 1.0 mm diameter), orange and positively buoyant eggs from gonopores arranged serially in each arm. The eggs undergo lecithotrophic development. The peak of sexual activity occurred in late summer (January) but specimens dissected in mid-winter (July) showed extensive gonad development. No response to 1-methyl adenine injection could be produced at this time. It would appear that this species can spawn and undergo complete gametogenesis within six months, but spawning was observed only once a year.
Females of Nardoa pauciforis produce eggs which appear very similar to that of Nardoa novaecaledoniae. They are the two common species of this genus on the Great Barrier Reef and are both very similar as adults. They are distinguished by the compression of the distal plates of the arms in Nardoa novaecaledoniae. The peak of sexual activity occurred one month earlier (December) in Nardoa pauciforis than in Nardoa novaecaledoniae, indicating a degree of reproductive isolation in these species at the southern end of the Great Barrier Reef.
Ophidiaster granifer produces eggs which undergo parthenogenetic development (Yamaguchi and Lucas, 1984) and in this study only females (total of 7 individuals) were observed to spawn. Small numbers (20-60 per female) of large (0.6 mm diameter), neutrally buoyant, bright red eggs underwent at least initial development despite no obvious sperm having been released in the water. The only spawning activity in this species was observed in early summer (November, December).
Disasterina abnormalis does not appear to be abundant on the Great Barrier Reef other than in the Capricorn Group at its southern end. At Heron Island this species is abundant behind an extensive rubble bank on the northern side of the reef. The eggs of this species are small (approximately 0.1 mm diameter), colourless and sticky. They sank to the bottom of the aquarium and adhered to the glass, from which they were hard to dislodge. The type of development is unknown. The peak of sexual activity occurred in late spring (October).
Echinaster luzonicus liberates small numbers (20-100 per female) of positively buoyant and approximately 1.0 mm in diameter red eggs during late summer (February). Specimens with fully developed arms were selected for the injection of hormone as arm regeneration following autotomy, which is common in this species, may be at the expense of gonad development.
The remaining species either showed no response to 1-methyl adenine or were not sampled in sufficient numbers to establish reproductive periodicity. The following responses to treatment occurred:
Iconaster longimanus no response July (N=1)
Culcita novaeguineae one female February (N=4) no response October (N=4)
Asteropsis carinifera no response July (N=1)
Gomophia egyptiaca two males and one female December (N=3) no response February (N=1), May (N=1), July (N=2)
Linckia multifora no response throughout study (N=120)
Nardoa rosea one male February (N=1) no response July (N=2), November (N=4)
Ophidiaster armatus one male July (N=2) no response February (N=1), October (N=1), December (N=1)
Ophidiaster confertus no response July (N=1)
Ophidiaster robillardi no response May (N=7), June (N=1), October (N=4), November (N=1), December (N=3)
Tamaria megaloplax no response July (N=1)
Asterina burtoni no response throughout study (N=59)
Echinaster stereosomus no response July (N=2)
Table 6.9 The primary type of reproduction (REPRO) and the type of larval development (DEVEL) where known are shown for all asteroid species recorded from Heron Reef. – = not known; PLANK = Planktotrophic; LECITH = Lecithotrophic A = Achituv (1972); B = Barker (1977); Y = Yamaguchi (1975); * = this study
SPECIES REPRO. DEVEL. SOURCE Astropecten polyacanthus SEXUAL PLANK Y Iconaster longimanus - - Culcita novaeguineae SEXUAL PLANK Y Acanthaster planci SEXUAL PLANK Y Asteropsis carinifera SEXUAL PLANK Y Dactylosaster cylindricus - - Fromia elegans SEXUAL LECITH * Fromia milleporella - - Gomophia egyptiaca SEXUAL LECITH Y Linckia guildingii ASEXUAL PLANK * Linckia laevigata SEXUAL PLANK Y Linckia multifora ASEXUAL PLANK Y Nardoa novaecaledoniae SEXUAL LECITH * Nardoa pauciforis SEXUAL LECITH * Nardoa rosea SEXUAL LECITH * Neoferdina cumingi - - Ophidiaster armatus - - Ophidiaster confertus - - Ophidiaster granifer SEXUAL LECITH Y Ophidiaster lioderma - - Ophidiaster robillardi ASEXUAL - Ophidiaster watsoni - - Anseropoda rosacea - - Asterina anomala ASEXUAL - A Asterina burtoni SEXUAL - A Disasterina abnormalis SEXUAL - * Disasterina leptalacantha - - Tegulaster emburyi - - Mithrodia clavigera SEXUAL PLANK Y Echinaster luzonicus ASEXUAL LECITH * Coscinasterias calamaria ASEXUAL PLANK B
The species of coral-reef asteroids studied at Heron Island showed differences in the length of their breeding season and this may reflect on their colonisation ability (Mileikovsky, 1971). The length of the breeding season within a species might vary with latitude and the further the population is from the equator, the shorter may be the season for summer breeders. However, except for Linckia laevigata, the species studied at Heron Reef generally showed a one to two month breeding season. In two species, Linckia multifora and Asterina burtoni, no sexual activity was observed throughout the study. This might result from lower than required water temperature at Heron Reef for most of the year (see Mladenov et al., 1986). In this regard, Mortensen (1937) was able to obtain eggs from Linckia multifora in the Red Sea where the water temperature is higher than at Heron Reef. It might also be correlated with an increased emphasis on asexual reproduction in Linckia multifora once a reef has been colonised by a few sexually reproduced individuals. Although Asterina burtoni was not observed to undergo either sexual or asexual reproduction, the distinction between A. burtoni and the small fissiparous A. anomala is unclear. It is possible that A. anomala is an asexually reproducing form of A. burtoni.
Nardoa novaecaledoniae and N. pauciforis possess arms swollen with gametes for much of the year but still have only a limited breeding season. At Heron Reef, at the southern end of the Great Barrier Reef, Nardoa pauciforis is reproductively mature earlier in the summer than is Nardoa novaecaledoniae and its eggs are released in November or early December. At this time Nardoa novaecaledoniae is not capable of releasing eggs and sperm. Although both species look similar they appear to have limited interbreeding, at least over part of their geographic range. In general, temperature seems to be an important factor in gametogenesis, but the final spawning trigger is dependent on lunar/tidal cycles in many species (Pearse, 1970, 1975; Yamaguchi and Lucas, 1984).
In coral-reef asteroids the range in fecundity is extremely large. Fromia elegans, Gomophia egyptiaca, all species of Nardoa, Ophidiaster granifer and Echinaster luzonicus produce large, buoyant, highly pigmented and yolky eggs. While larval development was not studied in detail, the initial phases of lecithotrophic development were observed in these species. The large reserves of yolk should ensure that the resulting larvae need not feed while in the plankton. The number of eggs produced with this development was not studied in detail but appeared to be relatively small (that is, less than 1000 and sometimes much fewer per individual). The eggs are buoyant, opaque and 0.6 to 2.0 mm in diameter. For any species, the egg size, in combination with number of eggs liberated, is an index of reproductive effort. The energetic fecundity in relation to body size of different species, might represent qualitatively different reproductive strategies.
Planktotrophic larvae are produced by Astropecten polyacanthus, Choriaster granulatus, Culcita novaeguineae, Acanthaster planci, Asteropsis carinifera, all species of Linckia, Mithrodia clavigera, Leiaster leachi and Coscinasterias calamaria (Yamaguchi, 1975; Barker, 1977). With this type of larval development, many (up to 1 million), small (0.1 to 0.2 mm), non-yolky, transparent eggs are produced. These eggs appeared less buoyant than eggs that contain large yolk reserves. This may influence dispersion.
Larvae of species of Nardoa and other genera which undergo lecithotrophic development may be less likely to die of starvation in the plankton compared with those of species that undergo planktotrophic development and have an obligate larval feeding stage before settlement (see e.g. Thorson, 1950, 1966; Vance, 1973; Barker, 1977; Strathmann and Vedder, 1977; McEdwards and Janies, 1993). Lecithotrophic larvae have shorter development times, but the planktonic stage can be extended if suitable settlement sites are not available (Strathmann, 1978; Yamaguchi, 1974; Yamaguchi and Lucas, 1984). However, these larvae cannot remain in the plankton for as long as larvae with planktotrophic development. Their dispersal ability and genetic exchange is lower (Scheltema, 1968, 1971; Nishida and Lucas, 1988; Nash et al., 1988; Mladenov and Emson, 1990; Benzie and Stoddart, 1992 a,b). Yamaguchi (1975 b) has commented on the low abundance of lecithotrophic species on oceanic atolls. However, such species are well represented at Heron Reef, a situation that might result from the proximity of adjacent reefs, which would allow short lived larvae from one reef to settle on nearby reefs ( see Fisk and Harriott, 1990). Any species may have difficulty colonising over distances greater than its larval dispersal capacity and lecithotrophic species might suffer local extinction following large scale destruction or alteration of coral reef habitat.
Despite the high sexual reproductive effort displayed by most of the large-bodied species, there is little evidence of high recruitment of starfish at Heron Reef. Loosanoff (1964) observed periodic high recruitment during a 25 year study of the temperate species Asterias forbesi. Periodic high recruitment has also been observed in the coral reef species Acanthaster planci.
It is possible that many eggs are never fertilised when adult populations exist at low densities, such as at Heron Reef. Many fertilised eggs or subsequent larvae would also die from predation or starvation in the plankton (Jackson and Strathmann, 1981; Olsen, 1987). The availability of suitable settling substrate or post-settlement benthic predation might also limit the recruitment of juveniles.
It might be expected that planktonic regulation would be less constant than benthic regulation because of the relative unpredictability of small scale water circulation and the extremely patchy distribution of planktonic predators compared with a possibly more regular and species-specific mortality caused by benthic predators. The post-settlement survival of small juvenile Acanthaster planci was examined by Keesing and Halford (1992) and Keesing and Cartwright (1993) who found a difference in survivorship between caged specimens compared with uncaged specimens. For the less common species at Heron Reef, it is possible that many of their eggs are not fertilised. Adult numbers will be further regulated by a combination of either larval starvation or larval and juvenile mortality.
When the energy content of a liberated egg is considered, two different reproductive strategies are apparent. Sexual recruitment can follow either planktotrophic or lecithotrophic larval development (Hendler, 1975; Yamaguchi, 1977 b; Lessios, 1990; McEdwards and Chia, 1991). Because many small eggs can represent the same investment of energy as a few large eggs, the energetic fecundity per unit body weight can be similar in both strategies, despite the difference in numerical fecundity.
It has been suggested (Vance, 1973; Yamaguchi, 1973 a, 1973 b, 1977 b) that lecithotrophic development is an adaptation to high predation or starvation of larva. With this development it is possible the length of larval life can be shorter and hence larval survival should be favoured. On oceanic atolls, species with lecithotrophic development are never abundant and this could result from their poor dispersal ability (Yamaguchi, 1975 b). On Heron Reef, and possibly the Great Barrier Reef in general, where many reefs exist in relatively close proximity, lecithotrophic genera such as Nardoa, Fromia and Echinaster appear to be more abundant than they are on atolls. However, planktotrophic development is favoured where high dispersal is required or when planktonic predation is low and planktonic food is predictable (Menge, 1975; Mileikovsky, 1971; Vance, 1973; Yamaguchi, 1977 b). Many species with this type of larval development occur on oceanic atolls but they occur also on the reefs of the Great Barrier Reef.
Disasterina abnormalis liberated small sticky eggs which sank and adhered to the substrate. The type of development was not studied, but the low dispersion capacity of its eggs might explain its apparently limited distribution. This species was abundant locally. In all other sexually reproducing species, juveniles were uncommon and there was no evidence of either periodic high recruitment or mortality.
The extent of larval dispersion is an important factor in our understanding of the community dynamics within a reef or reef system. If high between-reef larval dispersal occurs, then the adaptive significance of the dispersal phase is the location of spatially and temporally isolated patches in the survival mosaic of each species. Alternately, if larvae recruit primarily into the parent reef population, as a result of circular water movement patterns (Atkinson, Smith and Stroup, 1982; Dight et al., 1990 a, b; Black and Moran, 1991; Black, 1993 but see Wolanski, 1993), then the adaptive significance of the dispersal phase is the avoidance of planktonic or benthic predation in shallow water.
The development time of one month, possessed by many asteroid larvae with planktotrophic development (Yamaguchi, 1977 b; Williams and Benzie, 1993), allows potentially high dispersal, and this development time can be extended further if no suitable settlement site is available. Larvae with lecithotrophic development are capable also of extending the length of the pelagic phase (Yamaguchi, 1974; Yamaguchi and Lucas, 1984). The larvae of coral-reef starfish generally require a solid substrate to complete their development, and a coralline algal substrate has been observed as the chosen settling surface for many species (Yamaguchi, 1973 b; Johnson et al., 1991). More complex species specific signals, located by sensitive chemosensory receptors might ensure settlement in habitats which are conducive to survival of post-settlement stages (Burkenroad, 1957; Morse, 1984). Yamaguchi (1977 c) showed that some juvenile starfish have exponential growth during the period following settlement and proposed that juveniles are subject to high mortality during this period. The juveniles transform to adult morphology at a certain size and before this size is attained may look quite different from adults (e.g. Culcita novaeguineae illustrated by Clark, 1921).
The phenomenon of aggregation (Ormond et al., 1973), parthenogenetic development (Yamaguchi and Lucas, 1984), hermaphroditism (Achituv, 1972) or asexual reproduction (Rideout, 1978) may be correlated with survival at low population densities. In low density, spatially dispersed populations of starfish, there is a low probability of locating a conspecific of the opposite sex at breeding time.