What do starfish do and where are they?


Chapter 3. Habitat

3.1 Introduction

The spatial distribution of coral-reef starfish has been studied at several different scales. The physical, biological and historical parameters that explain the distribution of these species on the global scale (Clark and Rowe, 1971) will not directly explain the spatial pattern of an assemblage on a single reef. The small-scale distribution of Linckia laevigata on the fringing reef at Guam, and some factors which determine the abundance of this species in different habitats, were described by Strong (1975). The distribution and movements of Linckia laevigata at Lizard Island were examined by Thompson and Thompson (1982). Laxton (1974) suggested that Linckia laevigata may alter its distribution following outbreaks of Acanthaster planci. The distribution of an assemblage of starfish on a reef that is not known to have undergone outbreaks of Acanthaster planci has not been studied previously.

Feeding of starfish was extensively reviewed by Sloan (1980) and Jangoux (1982) and there have been many detailed examinations of the diets, competitive interactions and niche separation of colder water species (Blankley, 1984; Menge, 1972 a, 1972 b, 1981; Menge and Menge, 1974). The ecology of the tropical omnivorous, Atlantic species Oreaster reticulatus has been extensively studied by Scheibling (1980, 1981 a, b, 1982). The known diets and habitat preferences of the Indo-West Pacific coral-reef species have been tabled by Yamaguchi (1975 b). The contribution of these aspects of niche specialisation to the co-existence of many asteroid species in the coral reef ecosystem is poorly understood.

The general correlation between food supply and growth in asteroids has been discussed (Mead, 1900; Wolda, 1970; Paine, 1976). Species such as Acanthaster planci and Culcita novaeguineae are known to prey on hard corals (Endean, 1969; Yamaguchi, 1975 b; Glynn and Krupp, 1986). The response of Acanthaster planci to different prey species has been studied (Ormond, Hanscomb and Beach, 1976). Asterina anomala, Asterina burtoni, Ophidiaster granifer and Gomophia egyptiaca are known to feed on sponges and ascidians (Thomassin, 1976; Yamaguchi, 1975 b) while Astropecten polyacanthus and other members of its genus are known to prey on molluscs (Christensen, 1970; Ribi and Jost, 1978; Jost, 1979). Coscinasterias calamaria along with most members of the Order Forcipulatida, which are primarily inhabitants of temperate waters, are known to also prey on molluscs (Sloan, 1980; Jangoux, 1982).

However, the vast majority of coral-reef starfish are thought to be general detritovores and to feed primarily on the epibenthic felt (Thomassin, 1976) which is widely distributed throughout the reef environment. The possibility of ciliary nutrition (filter feeding) in some asteroids was raised by Gemmill (1915).

With the exception of what were once considered “primitive” genera such as Astropecten and Luidia (see Blake, 1987), starfish generally feed by everting their stomach over the substrate and digestion is external (Blake, 1990). However, the forcipulate species Heliaster helianthus is known to possess a flexible feeding habit involving both intra-oral and extra-oral feeding (Tokeshi, 1991). When Acanthaster planci and Culcita novaeguineae feed on hard coral they evert their stomachs and leave white feeding scars where the living tissue has been digested off the skeleton (Yamaguchi, 1975 b). Predominantly epibenthic feeders such as members of the genera Linckia, Nardoa and Ophidiaster leave no such feeding mark to indicate the position of their everted stomach when they are removed from the substrate while feeding (Yamaguchi, 1975 b).

While there have been extensive studies of the micro-habitat requirements of some coral-reef animals, for example gastropods (see e.g. Kohn and Leviten, 1976; Leviten and Kohn, 1980; Reichelt, 1982) there has been little work done on this aspect of coral-reef asteroid ecology. The role of habitat complexity in determining the population densities of species of gastropod within the coral reef ecosystem was discussed by Kohn (1968). However, the temporal scale of many community studies is often insufficient to determine the stability, or otherwise, of the observed community structure.

The degree to which species specialise in their use of habitat and other resources is a major component of the complexity of any assemblage (Klopfer, 1959; Klopfer and MacArthur, 1960, 1961; MacArthur and Levins, 1964, 1967; May and MacArthur, 1972. Reichelt (1982) regarded the availability of refuges from predation, desiccation and turbulence as determining factors in the spatial pattern of many intertidal species of gastropod. Kohn (1968) suggested that habitat complexity and the resultant spatial heterogeneity may directly determine the diversity of species assemblages.

Other authors have stated that the species assemblages that they examined were predominantly random. Guilds of species which use their habitat in a similar fashion were proposed by Sale (1976, 1977). In determining the spatial distribution of organisms, many authors propose the use of null (or neutral) models to prevent random events from being misinterpreted as meaningful biological pattern (Connor and Simberloff, 1979; McGuinness, 1984). However, caution over the misuse of inappropriate null models, which incorrectly reject real pattern, has been suggested by other authors (Dunbar, 1980; Quinn and Dunham, 1983; Roughgarden, 1983; Gilpin and Diamond, 1982).

3.2 Methods

At the completion of each traverse, all specimens were identified, counted and measured. The estimated area of the traverse was also recorded. While traverses were not stratified with respect to habitat, the zone of maximum density (primary habitat) was quite apparent for some species. Where a species occurred in more than one zone, the zones of less-frequent occurrence were referred to as secondary habitats. The distinction between primary and secondary habitat was much clearer in some species than in others. Some species were found so rarely that their range of habitat is unknown. In these cases, the zone in which they were located is regarded as the primary habitat.

25 specimens of Linckia laevigata were tagged with small plastic clothing tags which were inserted through the body wall of one arm, near its base, in such a way that a number inscribed on each tag was visible on close examination of the specimen. These starfish were then released at a site that provided no refuge other than under boulders. All boulders within a radius of 30 meters were overturned during attempts made to relocate them at 24 hour intervals. Twenty-five specimens of each of five additional species, Linckia guildingii, Linckia multifora, Nardoa novaecaledoniae, Nardoa pauciforis and Echinaster luzonicus, were released following similar tagging. Other methods of tagging, such as stains (Loosanoff, 1937; Feder, 1955; Vernon, 1937), were not successful because of low intensity of staining.

The feeding of starfish at Heron Reef was not examined in detail. When starfish specimens were collected, their stomachs were often found everted over the substrate. The approximate size of the stomach was noted along with the type of food material on which they appeared to be feeding. Any mark that their feeding activity may have left on the surface of the substrate was recorded.

3.3 Results

Culcita novaeguineae did not occur commonly in any habitat that was sampled by the shallow-water traverses. This species appeared to be more abundant in coral pools adjacent to the lagoon than on either the reef crest or reef flat at the western end of the reef. Linckia guildingii, Linckia laevigata, Nardoa novaecaledoniae and Nardoa pauciforis appeared to be slightly more abundant in one zone (primary habitat) than another (secondary habitat), but this difference was not clear, and was not quantified. All the remaining species occurred with either much greater densities in their primary habitats compared with their densities in secondary habitats, or were found in only one major zone.

The species that are regarded as exposed sometimes were found under boulders or coral rubble but no regular pattern of concealment was apparent in these species. The cryptic species varied in their type of refuge, which ranged from small rubble to large boulders.

The primary (1′) habitat, secondary (2′) habitat and general pattern of concealment of each species are listed in Table 3.1. These data are not quantitative but represent a general impression of the overall distribution pattern for each species. While some species are extremely restricted in their spatial distribution, others are widespread and it was not possible to estimate the abundance of each species, in each zone, in each sampling period when specimens were primarily required for size-frequency and reproductive analysis.

The general location of each species at Heron Reef and the known diet of each of the species (after Yamaguchi, 1975 b) are listed in Table 3.2. While diet was not studied in detail, observations made during this study confirm those of Yamaguchi that most species of starfish feed on epibenthic felt.

Table 3.1 The species of Asteroidea, primary habitat, secondary habitat, and their habit (excluding those species that occur predominantly in the off-reef floor zone).

SPECIES                                      1′ HABITAT       2′ HABITAT        HABIT

Astropecten polyacanthus  floor         flat           exposed
Iconaster longimanus      slope         floor          exposed
Culcita novaeguineae      flat          crest, slope   exposed
Acanthaster planci        slope         flat           both
Asteropsis carinifera     crest                        cryptic
Dactylosaster cylindricus crest                        cryptic
Fromia elegans            slope         crest          exposed
Fromia milleporella       crest                        exposed
Gomophia egyptiaca        slope         crest          both
Linckia guildingii        flat          crest          exposed
Linckia laevigata         flat          crest          exposed
Linckia multifora         crest         slope          both
Nardoa novaecaledoniae    crest         flat           exposed
Nardoa pauciforis         flat          crest          exposed
Nardoa rosea              floor         flat           exposed
Neoferdina cumingi        slope         crest          both
Ophidiaster armatus       floor         crest          exposed
Ophidiaster confertus     crest                        cryptic
Ophidiaster granifer      crest                        cryptic
Ophidiaster lioderma      crest                        cryptic
Ophidiaster robillardi    crest                        cryptic
Ophidiaster watsoni	  crest			       cryptic
Anseropoda rosacea        flat                         cryptic
Asterina anomala          crest                        cryptic
Asterina burtoni          crest                        cryptic
Disasterina abnormalis    crest                        cryptic
Disasterina leptalacantha flat          crest          cryptic
Tegulaster emburyi        crest                        cryptic
Mithrodia clavigera       crest                        exposed
Echinaster luzonicus      crest         flat, slope    both
Coscinasterias calamaria  crest                        cryptic

Table 3.2

Diet and location of each species (excluding those species that occur predominantly in the off-reef floor zone). In some species the diet is unknown. asc = ascidian

SPECIES                       DIET         LOCATION
Astropecten polyacanthus      mollusc      fine sand 0-30 m
Iconaster longimanus          felt         rubble 20-30 m
Culcita novaeguineae          coral, felt  sand, rubble 0-25 m
Acanthaster planci            coral        live coral 0-25 m
Asteropsis carinifera          -           sand under rock
Dactylosaster cylindricus      -           rock under rock
Fromia elegans                felt         rubble 0-25 m
Fromia milleporella            -           rubble
Gomophia egyptiaca            sponge, asc  sand, rubble 0-20 m
Linckia guildingii            felt         sand, rubble
Linckia laevigata             felt         sand, rubble
Linckia multifora             felt         attached under rock
Nardoa novaecaledoniae        felt         sand, rubble
Nardoa pauciforis             felt         sand, rubble
Nardoa rosea                   -           sand, rubble 0-30 m
Neoferdina cumingi             -           sand, rubble 0-20 m
Ophidiaster armatus            -           sand, rubble 0-30 m
Ophidiaster confertus          -           rock under rock
Ophidiaster granifer          sponge, asc  attached under rock
Ophidiaster lioderma           -           rubble under rock
Ophidiaster robillardi         -           attached under rock
Ophidiaster watsoni		 -	   under boulder
Anseropoda rosacea		 -	   sand
Asterina anomala              sponge, asc  attached under rock
Asterina burtoni              sponge, asc  attached under rock
Disasterina abnormalis        felt         attached under rock
Disasterina leptalacantha     felt         attached under rock
Tegulaster emburyi             -           attached under rock
Mithrodia clavigera            -           rubble
Echinaster luzonicus          felt         sand, rubble 0-25 m
Coscinasterias calamaria      mollusc      attached under rock

In all species, excepting Acanthaster planci, the stomach was small, usually about the same area as the disc. At no time was the ingestion of large food material observed in any species but Astropecten polyacanthus and Coscinasterias calamaria were occasionally observed feeding on very small gastropods which were partially inside the mouth. In all other species the feeding was entirely extra-oral.

When the stomach was everted, the oral spines were oriented away from the mouth and both digestion and absorption occurred outside the body. At the completion of feeding, the stomach was withdrawn through the mouth and the oral spines were reoriented such that they occluded the mouth opening. All species commenced retraction of the stomach when removed from the substrate but the delay, before retraction was complete, varied from a few seconds in the case of epibenthic felt feeders with small stomachs to a few minutes in the case of Acanthaster planci. In all species, the oral spines could not reorient into the non-feeding (defensive) positions until the stomach was fully retracted inside the mouth.

The only two species which left feeding scars on the substrate were Culcita novaeguineae and Acanthaster planci. Both of these species feed on hard coral. Culcita novaeguineae was observed also feeding on bryozoan colonies at the base of the reef slope at a depth of 20 meters. All the other species had not altered the appearance of the substrate on which they had been feeding. Fromia elegans, Linckia guildingii, Linckia laevigata, Linckia multifora, Nardoa novaecaledoniae, Nardoa pauciforis, Disasterina abnormalis, Disasterina leptalacantha and Echinaster luzonicus were only observed to evert their stomach over substrate of either sand or coral-rock that was covered with a fine layer of organic material (epibenthic felt). Ophidiaster granifer and Asterina burtoni were observed with their stomachs everted over both solitary and colonial ascidians, but these two species also everted their stomachs over the epibenthic felt.

The results of movement studies of Linckia laevigata showed that this species is capable of moving at least 30 meters in a 24 hour period. Only 12 out of 25 specimens, which were released on coral rubble substrate at the outer, northern reef flat were able to be relocated 24 hours later. Eight specimens had not moved, three specimens had moved one meter, one specimen had moved 15 meters, but 13 specimens had moved a distance greater than 30 meters, and were not relocated. After a further 24 hours only two tagged specimens and no untagged specimens could be relocated in the vicinity of the point of release. Of the 25 specimens of six other species that were released following tagging, no tagged specimens were observed when attempts were made to relocate them after an interval of two months. It is possible that the small plastic tags were lost from the arms of these specimens.

3.4 Discussion

It can be seen from Table 3.1 that 16 of the 31 asteroid species recorded from Heron Reef were found in more than one of the major coral reef zones during this study. Echinaster luzonicus was found in all zones. The less common species were not encountered often enough for the data to show their complete distribution. With the exception of Culcita novaeguineae, Linckia guildingii, Linckia laevigata, Nardoa novaecaledoniae, Nardoa pauciforis and Echinaster luzonicus, the asteroids at or in the vicinity of Heron Reef can be divided into reef flat, reef crest, reef slope and off-reef floor species.

Species that occurred with their highest abundance on the reef flat are Culcita novaeguineae, Linckia guildingii, Linckia laevigata, Nardoa pauciforis and Disasterina leptalacantha. All species, excepting D. leptalacantha, are a large size when fully grown (large-bodied) and lie fully exposed in the daytime. However, they might complete their early development under boulders on the reef crest (Yamaguchi 1973 a, b). Disasterina leptalacantha is a small cryptic species found occasionally on the reef crest, or more often, under slabs of beachrock on the innermost part of the reef flat.

Linckia multifora, Nardoa novaecaledoniae and Echinaster luzonicus occurred with highest abundance on the reef crest. Species which were found exclusively on the reef crest are Asteropsis carinifera, Dactylosaster cylindricus, Fromia milleporella, Ophidiaster confertus, Ophidiaster granifer, Ophidiaster lioderma, Ophidiaster robillardi, Asterina anomala, Asterina burtoni, Disasterina abnormalis, Tegulaster emburyi, Mithrodia clavigera and Coscinasterias calamaria.

Predominantly reef slope species are Iconaster longimanus, Fromia elegans, Gomophia egyptiaca, Neoferdina cumingi and Acanthaster planci. Their distribution seems more closely allied to living coral than is that of the species which inhabit the reef flat or reef crest.

The off-reef floor is not a true coral reef environment although numerous solitary corals and alcyonarians occur. The asteroid fauna of this environment is made up chiefly of the widespread sub-littoral species Astropecten polyacanthus, Pentaceraster regulus and Euretaster insignis in the deeper water (30-40 meters). Nardoa rosea, Ophidiaster armatus, Tamaria megaloplax, Echinaster stereosomus and large individuals of Linckia multifora and Echinaster luzonicus are found in the shallower water (20-30 meters) near the base of the reef slope.

Of the 25 starfish species found on Heron Reef itself, 17 were located only in intertidal regions and an additional three were found predominantly in intertidal regions. The other five species have been found intertidally, but occur predominantly in subtidal habitats.

Some cryptic reef crest species such as Asteropsis carinifera, Linckia multifora, Disasterina abnormalis, Echinaster luzonicus and Coscinasterias calamaria occur under rocks which possess a sparse development of epibiota. This would appear to be a short lived micro-habitat. Other cryptic reef crest species such as Gomophia egyptiaca, Ophidiaster granifer, Ophidiaster robillardi, Asterina anomala and Asterina burtoni are more frequently associated with encrustations of sponges and ascidians (Yamaguchi, 1975 b). Ophidiaster confertus was found either under or attached to the side of large boulders on the reef crest. Other species such as Dactylosaster cylindricus, Neoferdina cumingi, Ophidiaster lioderma and Tegulaster emburyi were not located in sufficient numbers to establish any pattern of occurrence. The reef crest species, that are regarded as cryptic during the day, did not move into exposed locations at night and species could remain cryptic during periods of activity. The reef crest, interstitial environment which is composed of highly fragmented coral, continually percolated by sea water, might provide refuges from desiccation and predation for some species but all attempts to locate asteroids in this micro-habitat were unsuccessful.

The epibenthic felt which covers large areas of intertidal coral reef habitat is composed primarily of protozoans, algae and bacteria (Thomassin, 1976). At Heron Reef and on other reefs, this appears to be a substantial resource. It would be of interest to know if competitive exclusion of one or more species can occur when population densities are higher than those recorded at Heron Reef. The role of species specific enzymes in the digestion of different organisms composing the epibenthic felt has not been investigated to date. Each species might possesses enzymes which facilitate the efficient exploitation of a different component of the epibenthic felt. A biochemical study of the gastric mucosa of each asteroid species would be needed to establish resource partitioning on this microscopic scale.

The feeding and movements of Acanthaster planci was studied by Keesing and Lucas (1992). This species has been shown to possess an enzyme which can efficiently digest the wax ester, Cetyl Palmitate, which is stored in the soft tissues of hard corals (Benson et al., 1975; Brahimi-Horn, 1989). Consequently, this asteroid must be regarded as a highly specialised predator of scleractinians. It is possible that other highly specialised enzymes occur in coral-reef asteroids. To exploit epibenthic felt efficiently, a scavenger would utilise mechanical or bacterial breakdown of algal cell walls and in other taxa this would be accomplished within a gut and the freed inter-cellular material would be subsequently assimilated by the organism. The epibenthic feeding asteroids evert their stomach over the felt and digestion occurs externally. These asteroids rely on enzymes to complete the digestion process but there are few known eucaryote enzymes which are capable of chemically digesting cellulose walls of algae. Diatoms, which are common in the epibenthic felt, have siliceous walls and the chemical digestion of this material, by any organism, would seem impossible. It is apparent, however, that many species of asteroid coexist on this same resource and further work is needed on possible dietary (enzyme) specialisation.

Predation on molluscs has been recorded in Astropecten polyacanthus and Coscinasterias calamaria. The paucity of mollusc-feeding coral-reef asteroids on Heron Reef is noticeable, compared with their high abundance in temperate waters (see Menge, 1975; Kwon and Cho, 1986; Nojima et al., 1986). Individuals of Coscinasterias calamaria found in the reef crest habitats were only small (R <20 mm), as were the gastropod prey on which they were observed feeding. the mobility of starfish might be insufficient to allow large-scale foraging on a coral reef as well as sufficient aggregation for successful reproduction.

Blake (1983) suggested that the general body plan of molluscivorous starfish leaves them vulnerable to predators. The observed delay in oral spine relocation following the commencement of stomach retraction, together with the defensive nature of these spines, suggests that the oral region of a starfish may be especially vulnerable to predatory attack particularly during and immediately following starfish feeding. The absence of adult specimens of Coscinasterias calamaria and other molluscivorous species of starfish may be related to the dangers inherent in this type of feeding.

In summary, the coral-reef starfish that occurred at Heron Reef showed some inter-specific variation with respect to diet, but many species appeared to feed on the same food (epibenthic felt). These species also showed some inter-specific variation with respect to habitat, but in every coral reef zone, some species sought no refuge and occurred with exposed habits. Of the species that occur predominantly on coral reefs, clear examples of niche (dietary or microhabitat) specialisation are known only for Culcita novaeguineae and the predominantly subtidal species Acanthaster planci. Clear examples of competitive interactions were not observed during this study.



Introduction 1985 – What not to say


Introduction to initial thesis of 1985.

The primary hindrance to any creative synthesis is the failure to recognize the causes of dissension in science. In some cases a critic may doubt the integrity of either the scientist or the data. In these cases further repetition of experiments may alleviate skepticism. In other cases, there may be a philosophical disagreement which has its basis in differing beliefs held by scientists. These preconceptions, which all scientists possess, can severely limit the process of conciliation. Dissension, however caused, is not easily resolved, and often leads to polarization within science. The failure of science to recognize and acknowledge explicitly, the validity of differing viewpoints based on the same data demonstrates confusion over these two causes of dissension.

Additionally, our aversion to implausible conjecture has limited the diversity of models that can be open to empiric testing. It must be remembered that, when seen through the eyes of a different culture, a model held to be true universally might appear implausible. An example is our astonishment at the complexity of Micronesian fishhooks which have been individually crafted to capture specific types of fish (Johannes, 1981). The pattern of relative planetary movements in our solar system was simplified greatly when the geo-centric model was discarded, in favor of a model in which the planets moved in orbits about the sun (Whitlam, 1975). This and other conceptual leaps have occurred when cultural limitations were waived temporarily, and the consequently greater insight justified the cultural changes that followed. The phrase, “that is only conjecture”, seems to imply that many scientists regard conjecture as superfluous in the day to day running of science. In the pursuit of scientific rigor, we often overlook this necessary component of synthesis and find it increasingly difficult to model our thoughts in a manner that enables others to see a more distant horizon by standing on our shoulders.

Finally, the scientific process might not progress beyond the accumulation of facts. Often it appears impossible, from the data, to do more than simplify nature’s variability by statistics, to categorize its variety by description or to model its behavior mathematically. These techniques should be tools to further understanding of the interrelationships between the elements under study. To many workers the apparently stochastic nature of many phenomena prohibits a more detailed analysis.

In recent years there have been many studies in coral reef ecology, biology and biogeography. Often, they are related, one to another, by a common principle or factor (e.g. population outbreaks of Acanthaster planci, plans for reef management or theoretical questions about the causes of diversity). Whenever several independent researchers study similar phenomena or ask similar questions, the possibility of dissension arises. The ensuing debates, and even hostility, can appear as a failure in communication between the scientists concerned. More often the observed behavior results from philosophic disagreement taking its natural course. Philosophic changes in both science and society are often inseparable; they rarely occur, figuratively speaking, without bloodshed.

Philosophic clashes are likely to occur whenever the capacity to acquire data, strategic to the opposing viewpoints, approaches a limit imposed by logistics or technology. This is certainly the case with much of the ecological debate over population outbreaks of Acanthaster planci. The failure to recognize the symptoms of this fundamental disagreement can lead to much wasted effort, funds and, most of all, time. Often, we need to determine which observations are required to swing general support from one model to another. However, the competing models may have reached the same level of logistic untestability and there may be no logistically possible observation or series of experiments which could distinguish between the validities of either model.

There has been much debate about the relative roles of disturbance, stability and niche specialization, as factors contributing to the co- existence of the large number of species in some communities (e.g. Brown, 1981; Connell, 1978; Levins, 1963; MacArthur, 1955, 1969; May, 1972, 1974; Paine, 1966). By contrast, there has been little consideration of the possibility that these factors and associated models may be of secondary importance to community order or structure itself. The process of community succession follows a path of increasing complexity towards an hypothesized relatively static, climax community, the composition of which is determined by prevailing environmental as well as historical parameters (e.g. Dunbar, 1972). Throughout this process, as early (rapid) colonists are excluded by species that are competitively superior, the composition of the community changes. The number of species present in the community increases to a maximum at some stage of succession, prior to the climax and subsequently decreases as a result of the exclusion of inferior competitors (Connell, 1978). The extent to which disturbance, by creating spatial and temporal patches of early succession, acts to prevent the monopolization of available resources by a small number of superior competitors has been discussed extensively (e.g. Dayton, 1971; Levin and Paine, 1974). Other authors have either proposed or implied that high diversity communities are at equilibrium, and that species coexistence is mediated by the complex processes of interdependence and specialization that have evolved, in a physically benign environment, over long periods (e.g. Fischer, 1960; Sanders, 1969; and reviews by Goodman, 1975; Osman and Whitlatch, 1978; Pianka, 1966). Additionally, Jacobs (1974) and Peters (1976) have pointed out that the correspondence between stability in species composition and stage of community succession is tautological because a successional climax is defined in terms of its temporal stability.

The progressive increase in diversity, biomass, complexity and structure, resulting from the succession process, has been the focus of much discussion (e.g. Dunbar, 1972; Margalef, 1963; Odum,1969). In some instances (e.g. Sale, 1984), the co-existence of numerous species can be explained without resorting to complex, pattern oriented, models that require numerous assumptions which are testable only by prolonged, rigorous and exacting field observation. The almost universally accepted null hypothesis of “chance” does very little to enlighten biological scientists who want to understand or observe any existing inter-relationships between the species they study; that hypothesis, however, has gained a reverence totally unbecoming a statement that claims to say nothing at all (Dunbar, 1980; Roughgarden, 1983).

Any function which may be played by community order or structure has, in the past, been so secondary to the aims or objectives of the “experimental approach”, as to be uninteresting or considered logistically untestable (i.e. impractical or too difficult). However, there is a fundamental difference between a model being logistically untestable and it being logically untestable (for discussion compare Connell, 1980; Dunbar, 1980; Kuhn, 1970; Popper, 1983; Quinn and Dunham, 1983; Roughgarden, 1983; Simberloff, 1982). I do not propose that all the myths laid to rest by “Occam’s Razor” be resurrected and considered as reasonable explanations of available evidence. However, I do suggest that, in the field of coral reef ecology, our attempts to simplify the system under study have produced models that bear little relationship to reality. The rigid adherence to the least complex and ramifying hypothesis, has made it difficult to see beyond the generally accepted view of nature based on probability theory and chance

While there have been many taxonomic and biogeographic works dealing with the coral reef asteroid community (Clark, 1921, 1938, 1946; Clark and Rowe, 1971; Marsh, 1974; Marsh, 1974, 1976, 1977; Yamaguchi, 1975b), the ecological requirements of asteroid species occurring within the Indo-West Pacific region have not been studied extensively. It is known that many species occur on coral reefs throughout the region (Clark and Rowe, 1971), while others possess a more restricted distribution. Several asteroid species are known from only a few specimens and are considered to be rare (Clark, 1921; Yamaguchi, 1975b). The habitat requirements of coral reef asteroid species, and the ecological roles of rare as well as of more common species are not understood. It is not known whether rarity is a survival strategy, an abundance limit imposed by predators or a failure in competitive ability of a species on its path to extinction. These questions have not been answered for this or any other taxonomic group within the highly diverse and complex ecosystem of the coral reef.

Competition and other ecological models and corollaries draw their scientific context often, by analogy, from the corresponding pattern of interaction observed within contemporary human society. The influence of one’s cultural background in the initial perception and subsequent acceptance of the ecological generality of these analogies is overlooked often. A model is an abstraction only, but in common with all scientific models, socially analogous models can be raised, by consensus, to the status of paradigms, such that, observations which contradict the model are considered either inaccurate or implausible. Assume that the population size of some organism is limited by the level of juvenile recruitment in such a way that the density of adults is never sufficiently high for one individual to interact significantly with another (e.g. Dale, 1978; Doherty, 1982). If these assumptions were true but unknown, the interactions between adults and their ecological significance could be modeled incorrectly using competition theory. Observations, which are categorized within a severely limited body of theory, cannot be regarded as empiric support for any hypothesis, as biased observations can provide support for any model. It is possible that many organisms live presently at adult population densities, which are sufficiently low to preclude both, inter and intra-specific competition. In such species, the adult population density may be limited always, at some previous stage of the life cycle, and the adult populations may be free from density dependent interactions.

A range of reproductive strategies is found in coral reef asteroids. Sexual recruitment can follow either planktotrophic or lecithotrophic larval development (Yamaguchi, 1977b). The occurrence of parthenogenetic development (Yamaguchi and Lucas, 1984), hermaphroditism (Achituv, 1972) or asexual reproduction (Rideout, 1978) may be correlated with survival at low population density and the consequential low probability of locating an opposite sexed conspecific at breeding time. Within coral reef asteroids, asexual reproduction has been observed in Linckia guildingii, Linckia multifora, Ophidiaster robillardi, Echinaster luzonicus and Asterina anomala (Emson and Wilkie, 1980). This provides evidence that, in some species under certain conditions, genetic variability and potential dispersal are less important to the maintenance of population numbers, than is continuity of recruitment.

The larvae of coral reef asteroids 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, 1973b). More complex species-specific optima, located by sensitive chemo-sensory receptors might ensure settlement in habitats which are conducive to survival of post-settlement stages (Morse, 1984). Yamaguchi (1977c) showed that some juvenile asteroids 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 may look quite different from adults (e.g. Culcita novaeguineae illustrated by Clark, 1921).

A general paucity of information about juveniles characterizes available data on population structures of large bodied, coral-reef asteroids (Yamaguchi, 1973a). It is possible that populations are maintained either by continual low recruitment or occasional high recruitment, each coupled with iteroparity. The juveniles are cryptic and their apparent absence or rarity indicates that reproductive success is either constantly low, sporadic or both. Sporadic success may depend on factors such as availability of planktonic food, level of planktonic predation or mortality of settled larvae. These may average out over the life span of the adult resulting in stability of adult numbers. Population increases of coral reef asteroid species have been well documented for Acanthaster planci, and apparent population increases of Linckia laevigata following Acanthaster outbreaks have been described (Laxton, 1974).

The differing requirements for growth and successful recruitment of juveniles, within the coral reef asteroid community, will have resulted in diverse life history strategies. A conceptual dichotomy exists in our perception of the life history of all organisms, and is referred to as r- versus K- strategy (Pianka, 1972; Stearns, 1976). These different survival characteristics are thought to have evolved in response to specific types of environments (Hairston, Tinkle and Wilbur, 1970; Murphy, 1968; Wilbur, Tinkle and Collins, 1974). The spectrum of existing life history attributes, apparent in any community study (see e.g. Menge, 1975; Vance, 1973), represents many points on a continuum between the conceptually ideal r- strategists and K- strategists.

Goodman (1974) proposed that if a population’s size is limited mainly by competition then natural selection will result in an increased competitive ability (K-selection) and, in populations which are not resource limited, selection will result in an increased reproductive rate (r-selection). The reproductive effort (energy used for reproduction compared with the energy used for non-reproductive purposes) and age specific mortality schedule are an indication of the type of selection which has occurred during the evolution of a species (Pianka, 1972).

The longevity of a species is determined by the relative probability of juvenile and adult survivorship. In the simplest case, if the probability of a sexually mature organism’s survival from one reproductive season to the next is greater than the probability of one of the offspring reaching sexual maturity, then the species will exhibit iteroparity (see Goodman, 1974; Murphy, 1968). The weighting of selective attributes is arbitrary (e.g. niche specialization, number and size of eggs, longevity, possession of toxin), as the absolute ends of the r-K continuum do not exist in reality. Unpredictable environmental factors (e.g. perils of larval life and enormous potential dispersion) can result in a high numerical fecundity, and consequentially, most marine benthic invertebrates have a high energy cost associated with reproduction (Mileikovsky, 1971).

The dispersal stage of a population spreads the risk of local extinction in space and time (den Boer, 1968; Scheltema, 1971; Strathmann, 1974). The early stages of succession survive by being able to colonize regions quickly following disturbance. The resultant spatial and temporal variation in population size seems to characterize the typical r- strategists. Their populations are stable only when viewed on a larger scale. The spatial and temporal scale at which a species must be viewed for its numbers to be stable is an indication of its position on the r-K continuum.

The life history strategy, of each species, will be viewed in this context and a variety of strategies should be observed within the coral reef asteroid community. Early succession species would be expected to have large fluctuations while late succession species should have smaller ones. Stable, climax communities should be characterized by small fluctuations of their component species. The apparent stability of any biological system is dependent on the scale of observation (see Bradbury and Reichelt, 1982; Sale, 1984; Weiss, 1969). At the organismic scale, there would be neither temporal nor spatial abundance variation if an individual exactly replaced itself, without dispersing, then died. At the population scale, a level of numerical stability, consistent with a model of community equilibrium and climax, could be achieved if dispersion, larval survival and settlement phenomena did not result in greatly differing adult numbers from one year to the next.

Since the late 1950’s, coral reefs of the Indo-West Pacific region have experienced population outbreaks of the corallivorous asteroid Acanthaster planci (e.g. Bligh and Bligh, 1971; Branham, 1973; Chesher, 1969; Endean and Chesher, 1973; Endean and Stablum, 1975; Goreau, 1963; Heydorn, 1972; Kenchington, 1976; Marsh and Tsuda, 1973; Pearson, 1972). The resultant loss of hard coral cover on some reefs of the Great Barrier Reef was studied during the period of outbreak, and subsequently, so that both the short and long term effects of this predator would be known (Endean and Stablum, 1973; Pearson, 1981). The role of this predator in the elevation or lowering of coral species diversity on the Great Barrier Reef has not been studied adequately. It is apparent that some reefs become reinfested with Acanthaster planci about 15 years following the initial infestation (Cameron and Endean, 1982). It would appear, that when the quantity (not necessarily diversity) of a reef’s hard coral cover has regrown, the asteroid can recruit again in high numbers.

Although the Acanthaster planci population outbreak phenomenon has puzzled scientists for a quarter of a century, and although many explanatory hypotheses and models have been proposed (e.g. Birkeland, 1982; Endean, 1969; Flanigan and Lamberts, 1981; Randall, 1972; Sale, Potts and Frankel, 1976), there remains disagreement about the causes of the phenomenon. Additionally, there is disagreement about the need for reef management strategies that might mitigate the widespread effects of this coral predator. The extent of present population outbreaks and the possibility of past outbreaks (prior to 1960) have not been studied in sufficient detail to allow critical evaluation of either the problem itself, or the risks associated with incorrect management. We do not know what factors allow high recruitment of this asteroid on some reefs when, on other reefs, it maintains a low population density. The natural life expectancy, larval dispersal and adult migration of this asteroid, while central to an understanding of the phenomenon, are not understood sufficiently (Moore, 1978). The role of natural predators in maintaining high diversity and the possible survival strategy of rarity in the coral reef community have not been studied adequately.

Comparative data on other coral reef asteroids might contribute usefully to an understanding, or at least, enlarge our perspective of the Acanthaster planci outbreak phenomenon. With this broad aim in mind, the present study focuses on the community ecology of asteroid species at Heron Island. The population dynamics of other coral reef asteroids might show patterns of high recruitment similar to those of Acanthaster planci. The study of the abundance, longevity, population density, diet and reproduction of other coral reef asteroid species will allow comparison with Acanthaster planci as well as provide information on the mechanisms that maintain diversity within this coral reef community.


Species Present at Heron Reef


Chapter 2. Species Present

2.1 Introduction

An initial study of the echinoderms of the Great Barrier Reef was conducted by H.L.Clark during a visit to the Murray Islands in 1913 (Clark, 1921). This work was followed by that of A.A.Livingstone during the Great Barrier Reef Expedition (Livingstone, 1932) and that of Gibbs, Clark and Clark (1976). Two monographs dealing with the Australian echinoderm fauna were compiled by H.L.Clark (Clark, 1938; 1946). Extensive biogeographical studies of Queensland echinoderms were undertaken by Endean (1953; 1956; 1957; 1961; 1965) and many of the records therein relate to Heron Island asteroids.

In the Indian Ocean, a detailed account of the echinoderm species present in West Australian waters was provided by Marsh (1976). Elsewhere in the Indian Ocean, the asteroid (starfish) fauna has been studied at Mozambique (Jangoux, 1972 a; Walenkamp, 1990), South Africa (Thandar, 1989), Somalia (Tortonese, 1980), the Gulf of Suez (James and Pearce, 1969), the Red Sea (Clark, 1967 a; Tortonese, 1960, 1977, 1979), the Arabian Gulf (Price, 1981), the Iranian Gulf (Mortensen, 1940), India (Koehler, 1910; James, 1973), the Andaman and Nicobar Islands (Julka and Das, 1978) and the Maldive Islands (Clark and Spencer-Davis, 1966; Jangoux and Aziz, 1985).

In the Pacific Ocean, the starfish fauna has been studied in China (Liao, 1980), Hong Kong (Clark, 1982), Taiwan (Chao and Chang, 1989), the Philippines (Fisher, 1919; Domantay, 1972; De Celis, 1980), the Ryukyu Islands (Hayashi, 1938 a), the Ogasawara Islands (Hayashi, 1938 b), the Caroline Islands (Hayashi, 1938 c; Grosenbaugh, 1981; Marsh, 1977; Oguro, 1984), the Mariana Islands (Yamaguchi, 1975 b; Kerr et al., 1992), the Marshall Islands (Clark, 1952), Indonesia (Guille and Jangoux, 1978; Jangoux, 1978), New Caledonia (Jangoux, 1984), Tonga (Clark, 1931), South East Polynesia (Marsh, 1974), Hawaii (Fisher, 1906; Ely, 1942) and the general North Pacific region (Fisher, 1911, 1925). The geographical distribution of the shallow water species was reviewed by Clark and Rowe (1971).

There have been many taxonomic revisions within the Asteroidea. The works of Baker and Marsh (1974), Blake (1979; 1980; 1981; 1983; 1990), Jangoux (1972 b; 1980), Pope and Rowe (1977), Rowe (1977) and Marsh (1991) have included species of coral-reef starfish. All previous revisions were summarised in the specific descriptions and keys to the asteroid species provided by Clark and Rowe (1971).

2.2 Methods

Specimens of several species of starfish were required primarily for size-frequency and reproductive analysis. Sampling methods were chosen so as to ensure that the sample sizes were sufficient to allow statistical analysis of size-frequency and reproductive data in a reasonable number of species. Starfish were collected by means of quadrats, general searches and on traverses that were conducted primarily at the western end of Heron Reef (Figure 1). On Heron Reef, traverses ran between the cay and the reef crest (0.5 to 2 kilometres apart) and also between two points both on the reef crest (0.5 to 6 kilometres apart). Because the primary purpose of sampling was the collection of size-frequency and reproductive data, the traverses were not stratified with respect to habitat. Traverses were neither systematic nor random and most traverses included both reef flat and reef crest zones. All exposed starfish within a four meter width were collected for the length of the traverse. In addition to the collection of exposed starfish, a selection of large and small, dead coral slabs were overturned and cryptic specimens located beneath these slabs were collected. The lagoon and its adjacent coral pools were not sampled by traverse because of the difficulty in traversing this habitat.

All traverses were conducted within two hours of low tide, during the period of spring tides (full or new moon). When the water over the reef flat and reef crest was any deeper than this or under adverse weather conditions it was difficult to locate smaller starfish. Specimens that were required for reproductive studies were collected during general searches at these times but these specimens were not included in either the abundance or size-frequency data because this would have been biased towards the more visible species and individuals.

In total, 72 overlapping, intertidal traverses were conducted during the period from May 1978 to December 1982. The total area sampled by these traverses was approximately 120 hectares (1.2 square kilometres) which is equivalent to about five percent of the shallow-water, reef area of Heron Reef. The mean traverse length was just over four kilometres.

Cryptic species were also sampled using metre square quadrats in particular areas where previous traverse sampling had shown that starfish abundance was relatively high. These samples provided data for starfish present on a very small area of the reef crest. These quadrat samples cannot be regarded as random and they are not typical of the reef crest in general. The reef crest zone is extremely variable and spatial heterogeneity (patchiness) appeared to be highly dependent on the scale of sampling. These quadrat samples were undertaken to obtain estimates of the starfish density in these localised patches.

Subtidal specimens of starfish were collected on the reef slope and off-reef floor by the use of SCUBA. These subtidal samples were not used to determine subtidal starfish density because limitations in underwater visibility would have resulted in the underestimation of all starfish abundances. Detailed quadrat sampling would not have been directly comparable with intertidal traverse data and such sampling was not considered appropriate given the logistical constraints of extensive sampling using SCUBA. The off-reef floor was only rarely sampled and the species that occur in this habitat may be much more abundant than is apparent from the results obtained.

All starfish were identified, measured and placed along with conspecifics in glass aquaria at the Heron Island Research Station. Specimens were identified by reference to Clark and Rowe (1971). Specimens were also compared with their original descriptions where necessary. An examination of the specimens with a stereoscopic microscope was sufficient to distinguish all species. Juvenile identification was possible in all cases by reference to Clark (1921), Yamaguchi and Lucas (1984) or Yamaguchi (1973 a, 1973 b, 1974, 1977 a).

All individuals not required for taxonomic study were released in habitats similar to those where they were found. Specimens of all species studied were photographed live and some were preserved in alcohol. These are housed in the Department of Zoology, University of Queensland.

Throughout this thesis, unless some ecological parameter is given higher priority temporarily, the sequence in which species appear in tables is determined by their systematic position. The families are sequenced according to Blake (1979, 1980, 1981, 1987, 1990) and the classification of Clark and Rowe (1971). The genera are sequenced alphabetically within families.

2.3 Results

The species listed in Table 2.1 have either been recorded previously from Heron Reef, or were found in the present study and represent new records for the locality (marked with “*”). The species included are all those that occur either on the reef top (reef flat and reef crest) or on the reef slope extending to a depth of approximately 30 metres. At Heron Reef this is approximately the depth where the substrate of predominantly live coral or coral rubble changes to the finer sediments of the off-reef floor. Coral-reef species that do not appear to occur on the off-reef floor are marked “+”.

Table 2.1 Asteroid species recorded from Heron Reef.

Astropecten polyacanthus Muller and Troschel,1842

Iconaster longimanus (Mobius, 1859) *

Culcita novaeguineae Muller and Troschel,1842 +

Acanthaster planci (Linnaeus,1758) +

Asteropsis carinifera (Lamarck,1816) *+

Dactylosaster cylindricus (Lamarck,1816) *+
Fromia elegans Clark,1921 *+
Fromia milleporella (Lamarck,1816) +
Gomophia egyptiaca Gray,1840 +
Linckia guildingii Gray,1840 +
Linckia laevigata (Linnaeus,1758) +
Linckia multifora (Lamarck,1816) *+
Nardoa novaecaledoniae (Perrier,1875) +
Nardoa pauciforis (von Martens,1866) +
Nardoa rosea Clark,1921
Neoferdina cumingi (Gray,1840) +
Ophidiaster armatus Koehler,1910 *
Ophidiaster confertus Clark,1916
Ophidiaster granifer Lutken,1871 +
Ophidiaster lioderma Clark,1921 *+
Ophidiaster robillardi de Loriel,1885 *+
Ophidiaster watsoni (Livingstone,1936) +
Tamaria megaloplax (Bell,1884) *

Anseropoda rosacea (Lamarck,1816)
Asterina anomala Clark,1921 *+
Asterina burtoni Gray,1840 +
Disasterina abnormalis Perrier,1876 *+
Disasterina leptalacantha (Clark,1916) +
Tegulaster emburyi Livingstone,1933 *+

Mithrodia clavigera (Lamarck,1816) *+

Echinaster luzonicus (Gray,1840) +
Echinaster stereosomus Fisher,1913 *

Coscinasterias calamaria (Gray,1840) *

* new record for Heron Reef + coral-reef species

In addition to the preceding species, Anthenea aspera, Stellaster equestris, Metrodira subulata and Acanthaster brevispinus were recorded from the area by Bennett (1958). These species were dredged from a depth of 45 meters east of Wistari Reef and were not directly associated with any coral-reef habitat. Halityle regularis was recorded by Baker and Marsh (1974) and Andora popei was recorded by Rowe (1977) from the off-reef floor near Heron Reef. Pentaceraster regulus and Euretaster insignis were also observed on the off-reef floor.

The following brief notes relate to the species of starfish that have been located at Heron Reef (on the reef flat, reef crest or reef slope) either in this study or by previous workers.

Family Astropectinidae
Astropecten polyacanthus Muller and Troschel,1842

This species of starfish is not restricted to coral reefs, but occurs also in sandy areas along the east coast of the Australian mainland. It was not common but specimens were found during this study in the deeper waters of the off-reef floor. It is recognised by the many conspicuous sharp spines along the body margin. The tube feet do not possess suckers at their tips. It has been found on a sandy spit at Heron Island Reef by Endean (1965).

Family Goniasteridae
Iconaster longimanus (Mobius,1859)

This orange and white patterned starfish is immediately recognised by its long tapering arms. It was not common at either Heron or the adjacent Wistari Reef, but specimens were located during this study in about 20 metres of water on the deeper parts of the reef slope. They were usually associated with coral rubble. Some specimens that were collected had recently lost one arm.

Family Oreasteridae
Culcita novaeguineae Muller and Troschel,1842

The juveniles of this species (R less than 70 mm) look quite different from adults. This starfish is most commonly encountered on the reef flat although it occurs also on the reef crest. Its greatest abundance may be at the base of the reef slope or in the deeper coral pools adjacent to the lagoon. This large and conspicuous species was not common on the traverses at Heron Reef during the period of this study.

Family Acanthasteridae
Acanthaster planci (Linnaeus,1758)

This well known species was uncommon at Heron and the adjacent Wistari Reef during the period of the present study. Only five subtidal adults and one juvenile specimen were encountered. Endean (1961) recorded a single specimen from a pool near the reef crest at Heron Reef.

Family Asteropseidae
Asteropsis carinifera (Lamarck,1816)

This species is not common at the southern end of the Great Barrier Reef. It has been recorded as common at Mer in the Murray Islands (Clark, 1921). During this study, three specimens were encountered on the reef crest at Heron Reef.

Family Ophidiasteridae
Dactylosaster cylindricus (Lamarck,1816)

During this study, a single specimen was located on the reef crest at Heron Reef. Few specimens of this species have been found on the Great Barrier Reef or elsewhere throughout its range. This species can be distinguished from others in this family by the presence of only a few small granules in the centre of each plate of the body. The remaining granules are concealed by a skin-like membrane. It occurs on the rocky reefs off southern Queensland more frequently than it does at Heron Reef.

Fromia milleporella (Lamarck,1816)

One specimen was found on the reef crest. Endean (1956) found two specimens under boulders on the reef crest.

Fromia elegans Clark,1921

At Heron Reef, this starfish is relatively common in the reef slope zone. Most specimens have five even arms, but specimens with four and six arms were not uncommon. This species was found also on the reef crest lying exposed in small pools, and on the sand at the base of the reef slope in 20 metres of water.

Gomophia egyptiaca Gray,1840

At Heron Reef, the only individuals encountered, during this study, were coloured purple and brown with pink tips to the tubercles which cover the aboral surface of the body. Specimens of this species were usually found either concealed under boulders on the reef crest or crawling amongst dead coral rubble on the reef slope. This species is not common at Heron Reef. Of the small number of intertidal specimens collected, two were found in close proximity. Endean (1965) found only two specimens on the reef flat at Heron Reef.

Linckia guildingii Gray,1840

The Grey Linckia, while not as common as L. laevigata or L. multifora, is encountered frequently on reefs of the Great Barrier Reef. The grey coloration conceals the animal when crawling over dead coral clumps which are covered by filamentous algae, but the animal is conspicuous when on coral sand. Although the adult starfish is uniform grey in colour, juveniles are mottled white, grey and purple, and do not lose this appearance until a size of about 80 mm arm radius is attained.

Linckia laevigata (Linnaeus,1758)

The Blue Linckia inhabits intertidal reef areas throughout the Indo-West Pacific region. It attains a large size (arm radius 180 mm), is brightly coloured, and is usually found lying unconcealed on or near coral clumps in the reef flat. It can be found also on the reef crest, lying either exposed on the algal rim or partially hidden under coral boulders in the rubble zone. There is very little colour variation within this species on the Great Barrier Reef. The most frequent number of arms is five although arm number ranges from three to seven. The extremes are rare.

Linckia multifora (Lamarck,1816)

This species is usually found with one or more arms missing, these having been autotomised. The maximum size that this animal attains at Heron Reef is about 100 mm arm radius, but most specimens are approximately one-third this size. Sometimes the starfish will be found crawling in the open across the reef crest but more often it will be found under boulders. The specimens which occur under boulders are usually smaller and lighter in colour and do not have the brown coloration which is found in those that have adopted an exposed existence. The most common number of arms is six but the number varies between three and eight. It is unusual to find a specimen with all arms of equal length.

Occasionally specimens are found that do not belong clearly to either Linckia laevigata or Linckia multifora. These specimens are blue in colour but have pointed arms and show evidence of recent autotomous reproduction. There is a small row of granules between the furrow spines. One blue comet form has been found during this study. Because of their general morphology, these specimens have not been regarded as Linckia laevigata, but as colour variations of Linckia multifora.

Nardoa novaecaledoniae (Perrier,1875)

The two common species of Nardoa appear quite similar in overall appearance and differ in the arrangement of the plates which cover the arms. In Nardoa novaecaledoniae these plates are abruptly reduced in size in the outer one-third of each arm.

Nardoa pauciforis (von Martens,1866)

This starfish is slightly less common than the previous species but is not hard to find on Heron Reef. It occurs more commonly on the reef flat than on the reef crest but it can be overlooked in this habitat as both N. pauciforis and N. novaecaledoniae blend well with the background of living and dead coral. The animals are most conspicuous when crawling over the sand between coral clumps. The average individual size of this species is slightly larger than that of N. novaecaledoniae. Also, the arms are usually longer relative to the body than in N. novaecaledoniae. A diagnostic feature of N. pauciforis is the absence of an abrupt change in the size of the plates towards the outer one-third of the arms.

Nardoa rosea Clark,1921

This species is more frequently encountered in the deeper parts of the reef slope (20 meters) than on the top of the reef at Heron Reef, but is not common in any of these habitats. It is a beautiful starfish with an average size of 90 mm arm radius.

Neoferdina cumingi (Gray,1840)

This starfish is not encountered often on the reef top at Heron Reef, but is occasionally seen when diving on the reef slope. There is great variation in the number and pattern of the red spots which are conspicuous along the arms.

Ophidiaster armatus Koehler,1910

All members of the genus Ophidiaster possess four rows of papular areas on both sides of every arm, a total of eight rows per arm. Papulae are the respiratory organs and occur in groups of between five and twenty, each appearing as a small, transparent projection through the outside body wall. The extent to which each papula is extended is dependent greatly on the water conditions.

O. armatus is readily recognisable by its dark coloration, tapering arms and by the coarse feel of the animal due to the very rough granulation of its skin. This species is found in low numbers, mainly at the base of the reef slope at Heron Reef.

Ophidiaster granifer Lutken,1871
This species possesses the tapering arms and uneven granulation of the previous species, but it is easily distinguished by its general coloration, smaller size (25 mm) and shorter arms relative to the diameter of the disc. Specimens of this species are usually encountered under boulders on the reef crest where they occur with moderate abundance. They are always cryptic in their habits.

Ophidiaster lioderma Clark,1921

This moderate sized starfish (R=100 mm) is very rare indeed having been found on two known occasions only, in two localities which are far apart on the Great Barrier Reef. The original specimen was discovered by H.L.Clark when he visited the northern end of the reef and was based at Murray Island in Torres Strait at the turn of the century. During this study, a further specimen was located on the reef crest at Heron Reef and is now housed in the West Australian Museum.

This species is a medium-brown in colour and can be readily identified by the skin covered body which possesses microscopic granulation. The only other member of this family which has a covering of thick skin is Leiaster leachi but this species has no surface granulation whatsoever and is brightly coloured.

Ophidiaster confertus Clark,1916

Four specimens of this species were located on the reef crest at Heron Reef. This species which grows up to 160 mm arm radius occurs more commonly on the New South Wales coast than at the southern end of the Great Barrier Reef (Clark, 1946).

Ophidiaster robillardi de Loriel,1885

This species occurs in moderate abundance in patches at Heron Island. The extreme patchiness of the distribution and abundance of this species is attributable to low dispersion associated with asexual reproduction. The average size of specimens is 35 mm arm radius and about ten percent of the specimens encountered were comet forms resulting from autotomous reproduction.

Ophidiaster watsoni Livingstone,1936

Gomophia egyptiaca and Ophidiaster watsoni are very similar and may be conspecific. Endean (1956) found one specimen of O. watsoni under a boulder on the reef edge at Heron Island.

Tamaria megaloplax (Bell,1884)

This species is found in the deeper waters, on sand near the base of the reef slope at Heron Reef. It occurs much more commonly on rocky reefs in south-east Queensland, than it does on the Great Barrier Reef. The average size of specimens found in south-east Queensland is about 100 mm arm radius. The specimens show considerable variation in the degree of roundness of the plates on the arms. This genus is characterised by having only three parallel rows of papular groups on both side of each arm, unlike Ophidiaster which has four, and Hacelia which has five rows.

Family Asterinidae
Anseropoda rosacea (Lamarck, 1816)

A single specimen of this species was found on sand in a reef-crest pool at Heron Reef by Endean (1956).

Asterina anomala Clark,1921

This small starfish is usually hard to find as the maximum size of individuals found on the Great Barrier Reef is about 5 mm arm radius. The bright coloration is of little help in finding this species as the boulders under which it occurs are encrusted usually with other brightly coloured invertebrates such as sponges and ascidians. This species is probably much more common than it appears to be but its small size makes sampling extremely difficult.

The usual number of arms in this species is seven. Half of these are normally regenerating as this species reproduces by binary fission. In this process, the animal divides into two and both sides regenerate the missing arms. If the regeneration has not proceeded very far then three or four adjacent ambulacral grooves will not extend much beyond the mouth.

Asterina burtoni Gray,1840

The taxonomic positions of this and of the preceding species are not clear. While the coloration of Asterina burtoni is quite variable, ranging from grey, through green to red or purple, it does not exhibit the multi-coloured pattern possessed by the previous species. A. burtoni does not reproduce by fission at Heron Reef and consequently most specimens have five arms of equal length. The average size of specimens is 13 mm arm radius.

Disasterina abnormalis Perrier,1876

This species has been recorded at a few localities along the Great Barrier Reef, and also in Indonesia as well as in the South Pacific. When alive, the animal is covered by a relatively thick skin which conceals the underlying plates. Many of these plates bear some very short rounded spines but it is not possible to discern the diagnostic characters of this species unless the specimen is preserved and then dried.

At Heron Reef at the southern end of The Great Barrier Reef, this is the most abundant starfish found on the top of the reef. It lives amongst the broken coral rubble on the innermost portion of the reef crest. The average size of specimens is 15 mm arm radius, but this size varies with periods of growth and with recruitment of juveniles to the population.

Disasterina leptalacantha (Clark,1916)

This close relative of the preceding species grows to the same size, but is known only from the Capricorn Group at the southern end of the Great Barrier Reef. The difference between these two species is unmistakable as Disasterina leptalacantha possesses very long, extremely thin spines along the body margin, but in life these may be folded upwards against the side of the body and are overlooked easily. The coloration of this species is different from that of the previous one and the arms are also slightly longer. The reason for the apparent limited distribution of this species is unknown.

The main habitat of this species is amongst the broken slabs of beachrock at low tide level. It is not common but specimens will be found either adhering to the underside of the rocks or amongst the sand immediately under the rocks.

Tegulaster emburyi Livingstone,1933

During this study, one specimen of this species was located on the reef crest at Heron Reef. The only other known specimen of this species was found at North-West Island, also in the Capricorn Group. Both specimens were found under a dead coral boulder in the reef crest zone. This species is exceedingly rare and may also be highly restricted in its geographic range. Both known specimens were just under 20 mm in arm radius.

Family Mithrodiidae
Mithrodia clavigera (Lamarck,1816)

During this study, one specimen was located at Heron Reef, but it did not occur within the intertidal traverses. It was located on the reef crest in December 1984. The species has been found elsewhere in the South Pacific but is uncommon.

Family Echinasteridae
Echinaster luzonicus (Gray,1840)

This starfish ranges from almost black, through red, to speckled orange and black in coloration. Specimens with all arms of equal length are not common as this species reproduces by means of autotomy, and comet forms will be found along with the adults in most habitats. The habitat in which this species is most abundant is under coral boulders on the reef crest. However, specimens may be found in most other intertidal habitats as well as on the reef slope and extending down to the boundary with the off-reef floor. The specimens which are found sub-tidally are larger usually than those found intertidally. The average size of specimens varies from one reef zone to another, but on the reef crest it is about 47 mm arm radius. However, its size is dependent on the amount of autotomy which has occurred recently. The species can grow to about 90 mm. Some of the specimens encountered at the edge of the off-reef floor possess epiphytic ctenophores crawling over the arms of the starfish.

Echinaster stereosomus Fisher,1913

At Heron Reef, this species is found near the base of the reef slope. It occurs on the rocky reefs off southern Queensland more frequently than it does at Heron Reef.

Family Asteriidae
Coscinasterias calamaria (Gray,1840)

This is primarily a southern species (Clark, 1946). Barrier Reef specimens are small, up to 30 mm arm radius, compared with the much larger individuals found on the mainland coast. This species is capable of asexual reproduction by binary fission. At Heron Reef, C. calamaria maintains small patches of moderate abundance by asexual reproduction. Indeed, it appears unlikely that specimens grow sufficiently large to become sexually mature at Heron Reef.

2.4 Discussion

The following species represent new records for Heron Reef:

Iconaster longimanus, Asteropsis carinifera, Dactylosaster cylindricus, Fromia elegans, Linckia multifora, Ophidiaster armatus, Ophidiaster lioderma, Ophidiaster robillardi, Tamaria megaloplax, Asterina anomala, Disasterina abnormalis, Tegulaster emburyi, Mithrodia clavigera, Echinaster stereosomus and Coscinasterias calamaria.

This study has provided the most southerly records from Great Barrier Reef waters of Iconaster longimanus, Asteropsis carinifera, Dactylosaster cylindricus, Fromia elegans, Linckia multifora, Ophidiaster armatus, Ophidiaster lioderma, Ophidiaster robillardi, Tamaria megaloplax, Asterina anomala, Disasterina abnormalis, Mithrodia clavigera and Echinaster stereosomus.

Single specimens of both Ophidiaster lioderma and Tegulaster emburyi were recorded at Heron Reef during this study and these represent the only known specimens of these species apart from their holotypes. Additionally, this study has provided the first record of the predominantly temperate species, Coscinasterias calamaria on the Great Barrier Reef. Euretaster insignis, which has not been recorded in the vicinity of a reef of the Great Barrier Reef, was found on the off-reef floor between Heron and Wistari Reefs.

Ophidiaster watsoni and Anseropoda rosacea were recorded from Heron Reef by Endean (1957) but were not located during this study. The taxonomic position of the former species is unclear. Anseropoda rosacea either is very uncommon at present, or primarily inhabits the sandy bottom of the lagoon which was not sampled extensively. Ophidiaster hemprichi and Ophidiaster lorioli occur at Heron Reef (Marsh pers. com.), but were not located during this study. Halityle regularis and Andora popei have been recorded from the off-reef floor in the vicinity of Heron Reef, by Baker and Marsh (1974) and Rowe (1977) respectively. These species were not located during this study as the off-reef floor was not sampled as intensively as were the shallow-water zones.

Because of its southerly position on the Great Barrier Reef, some predominantly sub-tropical asteroid species (e.g. Ophidiaster confertus and Coscinasterias calamaria) occur at Heron Reef but appear to not occur further north on the Great Barrier Reef. Additionally, some predominantly mainland species (Endean, 1957) occur either on, or in close proximity to, reefs of the Great Barrier Reef. The biogeographical study of Endean (1957) has shown that a distinction must be made between the asteroid species which occur predominantly on coral reefs of the Great Barrier Reef and those which occur elsewhere in Queensland waters. The results of the present study are in accord with this view. Clearly, there is a coral-reef asteroid fauna exemplified by that of Heron Reef, which is different from that of off-reef waters. However, as noted by Endean (1957), some species which occur on reefs of the Great Barrier Reef are not exclusively coral-reef species. For example, species such as Archaster typicus, Protoreaster nodosus, Ophidiaster confertus, Tamaria megaloplax, Asterina nuda, Patiriella pseudoexigua, Anseropoda rosacea and Coscinasterias calamaria occur predominantly in habitats other than those provided by coral reefs.

It seems likely, because of the extremely southern position of Heron Reef and other reefs in the Capricorn and Bunker Group, that many of the predominantly coral-reef species do not occur there with the same abundance as they do further north where physical conditions such as low water temperature on the reef flat in winter may be less extreme. Additionally, the relative isolation of this group of islands and reefs from the rest of the Great Barrier Reef might influence the abundance of those species with a low capacity for larval dispersal. However, these factors do not appear to affect the abundance of species that are common throughout the Great Barrier Reef. At higher latitudes, such as that of Heron Reef, the factors just mentioned might increase the abundance range between the most common and the rarest species. This would be reflected in the extent of sampling that would be required to locate most of the species that occur in the locality.

When current asteroid species lists for Heron Island and other reefs of the Capricorn Group are compared with those of recent studies of the North Pacific coral-reef Asteroidea (Yamaguchi, 1975 b; Marsh, 1977) it is apparent that some of the species that occur at Guam or Palau (e.g. Archaster typicus, Celerina heffernani, Fromia indica, Fromia monilis, Nardoa tuberculata, Nardoa tumulosa, Neoferdina offreti, Asterina corallicola and Echinaster callosus), have not been recorded from the Capricorn Group. On the other hand, some of the species that have been recorded from Heron Island and other reefs of the Capricorn Group (e.g. Tosia queenslandensis, Iconaster longimanus, Fromia elegans, Nardoa pauciforis, Nardoa rosea, Neoferdina cumingi, Ophidiaster armatus, Ophidiaster lioderma, Disasterina abnormalis, Disasterina leptalacantha and Tegulaster emburyi), have not been recorded from either Guam or Palau.

Future investigations may reveal that some of the above similar but geographically separated coral-reef species (e.g. Fromia indica and Fromia elegans) are conspecific. However, future investigations may also confirm the restricted distributions of some of the species mentioned above.

Most of the coral-reef asteroids found on the Great Barrier Reef, including Heron Reef and other reefs of the Capricorn Group, have strong affinities with coral-reef asteroids of the Western Pacific region as noted by Endean (1957). However, a few species appear endemic to the reefs of the Capricorn Group or are essentially sub-tropical species that have extended their ranges to include the southernmost reefs of the Great Barrier Reef.



Why study starfish at Heron Reef?


Chapter 1. General Introduction

Coral reefs seem to defy many of the paradigms which characterise less complex biological communities. While there is general agreement that the biota of coral reefs exhibit high species diversity, some authors have characterised coral reef assemblages by selecting species with high population densities (Sale, 1974; 1976; 1977; 1984; Sale and Dybdahl, 1975; Connell, 1978). Other authors have included rarer species (Kohn, 1959; 1968; Den Boer, 1971; Grassle, 1973) and Endean and Cameron (1990 a) have emphasised the importance of the role of these rarer species and stated that rarity is virtually ignored in most ecological models of the coral reef ecosystem. They suggest that our understanding of coral-reef ecology is influenced strongly by the constraints of many of the analytical tools being used in reef studies. As a result they believe that most analyses have dealt primarily with species that are sufficiently numerous to provide statistically satisfactory numbers of records and that most studies have excluded rare species which, in fact comprise the majority of coral-reef species.

The complexity of coral reef ecosystems is not surprising given the great length of time that these ecosystems have been in existence. While the shallow water distribution of coral reefs has varied with the alternation of glacial and interglacial periods (Hays, Imbrie and Shackleton, 1976), in their broad biological form, coral reefs have existed since the Precambrian and reefs similar to present reefs have existed for around 50 million years (Newell, 1972). While stating that there is no general rule for coral-reef organisms, Endean and Cameron (1990 a) have suggested that the attribute of persistence possessed by most of the rarer species characterises the majority of coral-reef species and is responsible for both structuring and perpetuating this ecosystem. They regard the coral reef ecosystem as being an ordered and predictable system. However, other authors (Sale, 1977; 1991; Connell, 1978) have different views.

Sale (1991) regards reef fish communities as open non-equilibrial systems with living space determined in a random manner. Connell (1978) regards intermediate levels of disturbance as essential to the maintenance of diversity in this and other highly diverse and complex ecosystems. There has been much discussion of the meaning of stability (MacArthur, 1955; Dunbar, 1960; Leigh, 1965; May, 1972; Jacobs, 1974; Margalef, 1974; Goodman, 1975; Peters, 1976; Pimm, 1984).

Endean and Cameron (1990 a) have put forward the hypothesis that complex, high diversity assemblages of coral-reef animals are characterised by a preponderance of rare but long-lived species that they have termed persisters. These persistent species exhibit low recruitment, low adult mortality and relative constancy of adult population numbers and population structure. They occur in association with opportunist species that have high recruitment, a high adult mortality and varying adult population numbers and population structure. While individuals belonging to opportunist species are more abundantly represented than those belonging to persistent species, Endean and Cameron believe that the majority of species in the coral reef ecosystem are persistent species. This hypothesis has not been tested in the field.

As no general consensus relating to the organisation of coral reefs has been reached in the literature, the persister / opportunist distinction is examined in this thesis, rather than a deep analysis of the opposing views relating to stability. Events that are stochastic and unpredictable at one spatial or temporal scale may be predictable at another. In addition, the stability or otherwise of any system may be determined, amongst other things, by the particular set of species that is chosen to characterise the system.

The starfish fauna of coral reefs can be distinguished from the starfish fauna of surrounding waters (Endean, 1953; 1965) and coral-reef starfish may be regarded as an ecological entity. During studies of Queensland echinoderms, Endean (1953; 1957; 1961; 1965) found 18 species of starfish on Heron Reef. Although reference was made to the habitat, general abundance and biogeography of each of the species, no detailed study of the Heron Reef starfish assemblage was made. This study will compare a number of ecological parameters in several species of starfish occurring on this coral reef. The population stability of the less abundantly represented, persistent species will be contrasted with that of the more abundantly represented opportunistic species. For the purposes of this study, the population stability of each species refers to the constancy of its population size structure over time.

Clark and Rowe (1971) and Yamaguchi (1975 b) reviewed the geographic distribution of many coral-reef starfish. It is clear that specimens of some species are frequently encountered and appear to be relatively common while others are known from very few specimens and appear to be extremely rare. The ecological requirements of coral-reef starfish, as well as the role of both rare and common species, are not understood and it is not known whether rarity is a survival strategy, an abundance limit imposed by predators or a failure in competitive ability of a species on its path to extinction. These problems have not been addressed for asteroids or any other taxonomic group within the highly diverse and complex coral reef ecosystem.

It has been suggested that longevity may characterise species of predictable environments (Frank, 1968; Grassle, 1973) or species with unpredictable pre-reproductive survival (Ebert, 1982; Goodman, 1974; Murphy, 1968). Several authors (Frank, 1969; Grassle, 1973; Ebert, 1982) have found many coral-reef animals to be long lived and Endean and Cameron (1990 a) regard the long-term persistence of individuals at given sites as an ordering phenomenon in the coral reef ecosystem. Little information is available on the longevity of coral-reef starfish. Ebert (1983), Kenchington (1976), Cameron and Endean (1982) and Endean and Cameron (1990 b), believe that Acanthaster planci is a long lived species, but Lucas (1984) suggested individual senescence in this species at an age of approximately five years. Stump and Lucas (1990) reported a linear growth pattern in aboral spine ossicles of this species which supported this suggestion, however the maximum age of this species has now been re-evaluated to at least 12-15 years (R.Stump, Ph.D. thesis). Yamaguchi and Lucas (1984) demonstrated a short lived population structure in the small and cryptic starfish Ophidiaster granifer, but little is known of the longevity of other species of coral-reef starfish.

The severe effects of Acanthaster planci predation are well documented (Chesher, 1969 a,b; Endean, 1969) and the change in coral population structure following an A. planci population outbreak was reported by Cameron, Endean and Devantier (1991). Moran (1986) has compiled a bibliography on the Acanthaster planci population outbreak phenomenon. Research on temperate starfish species that undergo population outbreaks has been reviewed by Loosanoff (1961).

Little is known of the other coral-reef starfish species, and the reproductive patterns, population stability and diversity of starfish assemblages on reefs that have not carried population outbreaks of Acanthaster planci are poorly understood. Heron Reef is such a reef. It is a Marine National Park and is situated near the southern end of the Great Barrier Reef.

It should be appreciated that the number of species recorded in any study is determined by both the spatial and temporal scales of sampling as well as by the distribution and composition of the species in the assemblage (species richness or diversity). To allow some degree of standardisation for collection effort, the rate at which the number of species in a sample increases with area of the sample (the species-area relationship) and the range of abundances within this assemblage (species relative abundance) have been chosen as a more representative measure of species richness than the total number of species. The species-area relationship, relative species abundance, and constancy of numbers and mean size (population stability) of coral-reef starfish are unknown elsewhere.

The fact that this study was undertaken on a reef that appeared to have low starfish abundance (and was not known to have carried an Acanthaster outbreak) generally precluded small-scale analyses that were dependent on high population densities. Large-scale traverse sampling is analogous to manta tows that have been used to monitor populations of Acanthaster. This scale of sampling is useful to establish a general pattern of starfish abundance, but is not capable of providing detailed data on either microhabitat partitioning or small-scale abundance. It should provide a basis for future comparison with data from Heron and other reefs.

Sexual reproductive patterns have been studied in some of the coral-reef starfish species known to occur throughout the Indo-West Pacific. Most of these studies have been conducted on reefs that are known to have carried population outbreaks of Acanthaster planci (Yamaguchi, 1973 a,b; 1974; 1975 b; 1977 a; Yamaguchi and Lucas, 1984). In addition to providing reproductive data for these species from a reef that does not undergo such population outbreaks, this study will examine the reproductive patterns of previously unstudied species. When the timing and extent of sexual reproduction along with the type of larval development exhibited by the various species studied are correlated, inferences can be drawn regarding the reproductive effort and dispersal capacity of each species involved. Endean and Cameron (1990 a) have suggested that opportunists and persisters are basically different with respect to their rates of recruitment, and a pattern should emerge when data on reproduction and population structure for a number of coral-reef starfish species are compared.

Several species of coral-reef starfish are known to exhibit asexual reproduction. The extent of asexual reproduction in the population maintenance of each species is an indication of the adaptive significance of this low-dispersal reproductive strategy. Many authors have commented on the role that may be played by this form of reproduction (Rideout, 1978; Yamaguchi, 1975 b; Ottesen and Lucas, 1982; Yamaguchi and Lucas, 1984) and Endean and Cameron (1990 a) have suggested that this mode of reproduction may assist species to withstand disturbance. Most species of starfish cannot reproduce asexually but are still capable of great powers of regeneration. Missing limbs in species that do not reproduce asexually may indicate sub-lethal predation.

Recruitment, migration and mortality ultimately determine the spatial and temporal distributions of the starfish populations in the Heron Reef assemblage. There is a distinction to be drawn between reproduction and recruitment as well as between predation and mortality. Recruitment is a process that is complete only when an offspring reaches maturity and reproduces itself. Similarly, predation may only be sub-lethal and autotomised limbs may be regenerated or may become asexual recruits. Starfish mortality occurs only when all fragments of a starfish have died. For logistical reasons, it was decided not to examine potential predators in this study. Likewise, a detailed examination of larval settlement processes was not undertaken. Migration of starfish is poorly understood as there is considerable difficulty in relocating tagged specimens particularly in autotomous species.

However, the interaction between the major determinants of population size mentioned above will influence the size-frequency distributions of each species. These distributions will be compared over time at Heron Reef and with size-frequency data from other localities. Mean individual size will vary with periods of recruitment and mortality, and size-frequency distributions that are constant over a study period of several years will suggest stability within the age structure. Alternately, such a finding could reflect the apparently static nature of a long-lived species when observed on a comparatively short time scale, even one of several years. However, study of the latter alternative could not be pursued beyond the time frame of this study which embraced five years.

Knowledge of the spatial pattern, fecundity and population dynamics of each of the coral-reef starfish species represented is essential to an understanding of the stability or otherwise of the populations of species comprising the coral-reef asteroid assemblages of Indo-West Pacific reefs. This knowledge is also essential to an understanding of outbreak phenomena, such as population outbreaks of Acanthaster planci. The obtaining of comparative distribution and reproductive data on many starfish species from Heron Reef will clarify the factors that influence diversity and stability within this assemblage.

With these broad aims in mind, this study focused on Heron Reef and sought answers to the following questions:

What starfish species are present at Heron Reef?
What is the spatial pattern for each species?
What is the population structure of each species?
What is the reproductive mode of each species?
Is the mean individual size stable for each species?
How is abundance distributed within this assemblage?

Thesis to be defended

In this study of the shallow-water asteroid assemblage of Heron Island reef, an Indo-West Pacific coral reef that has not been known to carry an outbreak of Acanthaster planci and hence can be regarded as a reef that has not been subject to a major disturbance at least in the immediate past, the thesis to be defended is:

1. The asteroid assemblage is comprised of numerous persistent species and a smaller number of opportunistic species.

2. The persistent species are relatively uncommon (rare) and possess relatively stable population densities and population size structures and have low rates of recruitment.

3. The opportunistic species exhibit localised high density, significant population fluctuations and are characterised by high recruitment (either sexual or asexual).

Site of study

Heron Reef (23° 27′ S, 151° 57′ E) lies in the Capricorn Group which is towards the southern end of the Great Barrier Reef. It is a lagoonal platform reef with a vegetated cay at its western end (Figure 1). The cay supports a tourist resort and research station. Heron Reef has been zoned as Marine Park A within the Capricornia Section of the Great Barrier Reef Marine Park, and prior to this was protected, from over-collection, by a regulation of Queensland State Fisheries. The western end of the reef is easily accessible from the cay but access to the eastern end requires the use of a small boat.

The major habitat zones used in the present study are described in detail by Jell and Flood (1978).

These zones are: 1. Reef flat (with lagoon)

2. Reef crest or reef rim

3. Reef slope

4. Off-reef floor

At the western end of Heron Reef, where studies were made, the reef flat is the sub-tidal habitat nearest to the cay. It is chiefly comprised of dead and living coral clumps which vary in size from a few centimetres in diameter to dead coral boulders or living micro-atolls with diameters of several metres. The dead coral clumps can, at certain times of the year, be obscured by a prolific growth of algae. The chief physical parameter that separates the reef flat from the lagoon is the water depth at low water spring tides. The water depth can vary from less than half a metre at the western end of the reef where sedimentation is great to more than a metre at its transition into lagoon east of the cay. The lagoon is up to six meters in depth at Heron Reef and has scattered coral outcrops which may reach the surface. It is regarded as an extension of the reef flat for the purposes of this study. At the innermost part of the reef flat (adjacent to the cay) a series of strata composed of cemented sand and coral fragments occurs. The strata are called beachrock.

The reef crest is the outer region of intertidal coral growth and is shallower than the previous zone. It is the most turbulent of all coral-reef zones being exposed to direct wave action at all stages of the tide. It has little fine sediment other than that which is trapped within the algal turf and which has accumulated under boulders. Living coral growth is usually low in profile and the general substrate is comprised of cemented reef rock strewn with broken coralline material. This material ranges in size from single coral fragments which are a few centimetres in diameter, to large boulders that are greater than two meters in diameter.

The reef slope is subtidal and supports extensive coral growth to a depth of approximately 20 meters. The coral growth tapers off to almost negligible coral cover at a depth of approximately 30 meters where the slope merges with the off-reef floor. This transition may be sudden on some reefs which possess almost vertical reef slopes, but at Heron Reef the transition is gradual. This zone is less physically controlled than are the previous zones. After periods of severe swell there may be areas of broken coral colonies but generally, as depth increases, the direct effect of wave action decreases. The substrate is of poorly sorted sediments as well as living and dead coral colonies, together with their epibiota.

The off-reef floor between Heron Reef and the adjacent reefs is over 40 meters deep and in places supports a well developed fauna of alcyonarians and solitary hard corals along with their associated epibiota. The off-reef floor is the deepest of the reef zones and provides habitats that are clearly different from the shallow water habitats provided by the other three zones. The sediment found on the off-reef floor is varied and its composition is dependent on currents as well as on surge effects during heavy wave action.


What’s an outbreak?


Chapter 4. Population Density

Summary: Figure 4.14 / 4.15 (above) is a composite graph that shows most inter-tidal species of starfish at Heron Reef are rare. It graphs the population distributions of the six relatively abundant species, namely Echinaster luzonicus, Disasterina abnormalis, Asterina burtoni, Nardoa novaecaledoniae, Linckia multifora and Linckia laevigata. It also  graphs the population distributions of the six less abundant species, namely Asterina anomala, Ophidiaster granifer, Nardoa pauciforis, Linckia guildingii, Culcita novaeguineae and Fromia elegans. The abundances of the 12 remaining species that occurred on traverses were very low and were not analyzed.

4.1 Introduction

It is well known that the scale of observation is critical for the determination of the spatial distribution pattern of a species. Differing scales of analysis can produce apparently differing results even with the same data. The properties or parameters that emerge from studies of communities can be dependent on which scale of organisation, space or time is chosen (Bradbury and Reichelt, 1982).

While the abundances of the various species of starfish will be partially determined by the small-scale distribution of scattered resources, the overall spatial distribution of each species will be a composite pattern influenced by food, refuge and predator abundance as well as aggregation behaviour (Patton et al., 1991; Stevenson, 1992; Iwasaki, 1993). Each of these factors can vary at a number of scales.

For each species within this assemblage, population aggregation may vary either spatially (from one location to another) or temporally (over time at any one location). If there is an equal probability of locating a species at every point within its spatial distribution, then individuals of that species are distributed at random. However, if the geographical range of a species or its abundance variation within that range is attributed to either physical or biological parameters, then non-randomness of the spatial distribution of that species is directly implied.

If the scale of observation is such that individuals of a particular species would be expected to be distributed randomly throughout habitats, which themselves are distributed randomly in space, then the expected distribution of individuals in space will be clumped, not random. If low density populations of starfish are not expected to be distributed randomly, then density estimates can seriously underestimate the standard error of the mean. Failure to determine the degree of positive skewness in the density distribution results in poor repeatability. Population density estimates of non-random species are credible only when the extent of the positive tail of this distribution has been determined adequately.

4.2 Methods

Specimens were collected primarily for size-frequency and reproductive analysis. For logistical reasons, it was not possible to estimate the density of each species, in each zone, in each sampling period. For each species, the density on each traverse was calculated by dividing the number of individuals by the estimated area of the traverse. The mean density of each species was then calculated by taking the arithmetic mean of the 72 traverse densities. It is represented as the average number of individuals found per hectare.

Because starfish are not distributed randomly, the total number of individuals of each species divided by the total area is not equal to the mean of the individual traverse densities. The standard deviation of density was calculated from the 72 traverse densities and represents the overall variation in density across all the traverses.

Because of the nature of traverse sampling, the density of most species is only approximate. Exposed species are reasonably well estimated but the cryptic species are greatly underestimated in their abundance because not all coral rocks and boulders in each traverse were overturned. Although the undersurface of rocks was examined closely, the nature of the substrate would make detection of the smaller species less reliable than the detection of larger species. When specimens were located within the sediment under rocks, individuals that were buried deeply within the rubble or sediment under these rocks would not have been found.

It should be noted that, in addition to patchiness, the number of individuals of each species recorded on different traverses varied because of variation in the size of traverse. The total number of each species also varied between sampling periods as a result of variation in the number of traverses undertaken in each sampling period.

Disasterina abnormalis was sampled in detail because it occurred in one region at a high density. This was the only species that could be sampled in this manner and this species appeared to occur at this density in only one region. The mean individual density per square meter, over a number of contiguous quadrats, and a Chi-square value (with Yates’ correction) of the inter-quadrat variation was calculated for Disasterina abnormalis. Twenty (metre square) quadrats were laid at Site 1 in April 1980 and again in July 1980. Forty (metre square) quadrats were laid at Site 2 in April 1980. All specimens occurring within the quadrats were counted and measured. It is to be noted that this was a region of northern reef crest where the density of this particular species was known from traverse data to be high.

4.3 Results

The data presented in Table 4.1 show the densities of all the intertidal asteroid species that occurred within traverses during this study at Heron Reef. For all species, the standard deviation was greater than the mean density. This, together with Figures 4.2 to 4.13, indicates variations in density that are greater than the expected Poisson variation. The results of quadrat density analysis of Disasterina abnormalis are shown in Table 4.2. The variation was analysed using chi square and individuals were clumped at the metre square scale.

Figure 4.1a graphs the linear relation between the total number of individuals and the total sample area. Figure 4.1b graphs the number of species in each of five (log) average density ranges. This illustrates how the average density of starfish species is distributed within this assemblage.

Figures 4.2 to 4.13 graph the population distribution of each of the common species, over the 72 traverses. Each graph displays the number of traverses on which a species occurred at a particular density. The density axis has been logged to facilitate the display of an extremely wide range of density.

Figures 4.14 and 4.15 are composite graphs of the population distributions of these species. Figure 4.14 graphs the population distributions of the six relatively abundant species, namely Echinaster luzonicus, Disasterina abnormalis, Asterina burtoni, Nardoa novaecaledoniae, Linckia multifora and Linckia laevigata. Figure 4.15 graphs the population distributions of the six less abundant species, namely Asterina anomala, Ophidiaster granifer, Nardoa pauciforis, Linckia guildingii, Culcita novaeguineae and Fromia elegans. The abundances of the 12 remaining species that occurred on traverses were very low and were not analysed.

Table 4.1

The density of each species that occurred on intertidal traverses expressed as mean density (number per hectare), standard deviation (S.D.) and number (N) of individuals.

SPECIES                                               MEAN DENSITY           S.D.                N

Culcita novaeguineae                                       0.14                0.48                15
Asteropsis carinifera                                        0.02               0.12                  3
Dactylosaster cylindricus                                0.01               0.12                  1
Fromia elegans                                                   0.10               0.42                16
Fromia milleporella                                          0.002            0.01                  1
Gomophia egyptiaca                                         0.12               0.57                  6
Linckia guildingii                                               1.27               2.63               116
Linckia laevigata                                               4.01                4.87              509
Linckia multifora                                              7.51               17.30              522
Nardoa novaecaledoniae                                3.19                 3.72              326
Nardoa pauciforis                                             1.60                 1.77              187
Nardoa rosea                                                     0.002              0.01                   1
Ophidiaster armatus                                       0.02                0.11                   4
Ophidiaster confertus                                    0.03                 0.15                   4
Ophidiaster granifer                                       1.56                 2.67                116
Ophidiaster lioderma                                    0.02                 0.15                     1
Ophidiaster robillardi                                   0.58                  2.58                  24
Asterina anomala                                           0.23                 0.58                   17
Asterina burtoni                                             3.27                 6.99                208
Disasterina abnormalis                                5.68               10.04                500
Disasterina leptalacantha                           0.23                 1.31                      7
Tegulaster emburyi                                       0.01                0.07                      1
Echinaster luzonicus                                  16.16               24.67               1402
Coscinasterias calamaria                            0.11                  0.65                      7

Table 4.2

Density and patchiness of Disasterina abnormalis.

The variation in number of individuals within adjacent square metre quadrats at two study sites and two sampling periods. DENSITY (the number of individuals per square metre), CHI-SQUARE (calculated from the inter-quadrat variation), PROB (the probability of this variation being random) and the NUMBER of individuals in the sample are tabled.

PERIOD                           DENSITY              CHI-SQUARE               PROB.      NUMBER

APRIL 1980 SITE 1                8.4                    58 (d.f.=24)               <.001                  161
APRIL 1980 SITE 2               0.7                                                             n/s                    29
JULY 1980 SITE 1                  8.9                     11 (d.f.=11)                    n/s                    98

4.4 Discussion

The traverse data do not allow for a statistically valid comparison of density among different sites or different sampling periods. Four species of starfish appeared to demonstrate changes in density during the study period. Two of these species were capable of asexual reproduction and these species demonstrated periods of autotomy followed by periods of growth. These species were Linckia multifora and Echinaster luzonicus. Asexual reproduction following a sexual recruitment was suggested by Ottesen and Lucas (1982) and Yamaguchi and Lucas (1984) as the reason for the greatly different abundances of all asexually reproducing species at different places on the same reef.

The other two species that showed a large change in abundance were Disasterina abnormalis and Asterina burtoni. At the commencement of the sampling program, the density of Asterina burtoni appeared to be about half that of Disasterina abnormalis under boulders on the reef crest. The abundance change in Asterina burtoni could not be analysed accurately because Asterina burtoni did not occur in great abundance in any known habitat. As a result, it was not possible to sample its density using meter square quadrats. A temporal variation in the abundance of A. burtoni was recorded by Price (1981) in the Arabian Gulf. Disasterina abnormalis had periodic high recruitment with resultant changes in both its abundance and size-frequency distribution. For Linckia multifora, Echinaster luzonicus and Disasterina abnormalis, the changes in mean individual size results from periods of high recruitment which are discussed in the chapter on Population Stability. The range in abundance of Acanthaster planci in both outbreaking and non-outbreaking populations was investigated by Moran and De’ath (1992 a).

The results of quadrat sampling in an area of reef crest north of Heron cay where the density of Disasterina abnormalis was known to be high (Table 4.2, Site 1) showed an average density of 8.4 individuals per square metre. This region was the innermost part of the reef crest and was sheltered partially from heavy wave action by a bank of rubble which extended for about one kilometre. One hundred metres west of this rubble bank (Table 4.2, Site 2), in an otherwise similar region of the inner reef crest, the density of this species was less than one individual per square metre.

The number of individuals of Disasterina abnormalis per square metre showed too much variation in the April 1980 (Site 1) sample for the individuals to be randomly distributed at the time of sampling. However, in the July 1980 (Site 1) sample, the individuals were not significantly clumped. It is of interest that the density per square metre of this species did not differ significantly between these two sampling periods. The only difference was in the degree of aggregation. Antonelli and Kazarinoff (1988) regarded the degree of aggregation of A. planci as an important factor in the modelling of population regulation by predators.

The quadrat samples produced only a minute subset of the known number of species, because such a small area was sampled. It was not feasible to sample extensively by quadrat as the patchy distribution of all these species required a large-scale estimate of spatial density variation.

Environmental heterogeneity might account for the observed clumping of individuals that were found primarily under boulders or rubble, but does not explain the variation in abundance of exposed asteroids on traverses which crossed what appeared to be similar habitat. The effect of variation in physical parameters, such as depth of water, amount of siltation of substrate, and strength of wave action is unknown, and factors such as these might account for some of the observed differences in abundance.

It can be seen from Figures 4.2 to 4.13 that individuals of each of the species were usually either absent or reasonably well represented on traverses. Individuals of the more abundant species did not occur at higher densities on every traverse, but were located on more of the traverses and occurred more often at moderate and higher densities. Individuals of the less abundant species were absent from most of the traverses and occurred less often at the moderate densities and never at higher densities. The possibility of variation in the abundance of species from reef to reef also exists. This would be more noticeable if reefs maintain semi-closed circulations. The numbers of one species may gradually build up by local recruitment if larvae recruit to the parent population.

Some of the rarer species of coral-reef starfish are known only from their holotype or perhaps one or two paratypes and appear to exist at population densities which defy our normal understanding of population dynamics and reproductive strategies. It is not clear how these species survive and which, if any, ecological requirements or constraints limit their distribution or abundance. It is not known whether these species are rare because their necessary ecological requirements are met at only a small number of points or whether their rarity is a result of intense predation.

Recruitment involving survival to reproduction must occur at some points within the distribution of each species unless we are observing the process of extinction. Considering both the number of species involved and the fact that species such as Tosia queenslandensis, Ophidiaster lioderma and Tegulaster emburyi are considered rare throughout their geographical range, rare species must demonstrate physical or behavioural attributes which are adaptations to existence in low density populations. Levins and Culver (1971) suggested that specialised rare species might play a key role in ecosystem modulation and they raised the possibility of specialised predation or competition among rare species.

Any assumption of a species abundance indicating its successfulness or adaptive nature should be questioned and the concept of species adapted to live in sparse populations offered as partial explanation of the high diversity in many ecosystems. The influence of specific predation can result in the rarity of a species and adaptations to this might represent a viable survival strategy (Connell, 1970). Spawning aggregations, extended gamete survival and high gamete specificity, hermaphroditism, parthenogenesis and asexual reproduction are all ways of ensuring continuity of offspring in rare species. Certain very specialised species might occur only at a certain resource optimum and their populations will be limited to the number of these sites of optimum habitat.

It is not known to what extent population fluctuations are normal on coral reefs. In the larger species of starfish at Heron Reef, the overall impression was that population fluctuations were low compared with the fluctuations that are known to occur on other reefs and in temperate ecosystems.

Because the abundance data resulting from the traverse samples was biased towards large, exposed individuals, it would be unwise to use this traverse data for a direct density comparison over repeated sampling periods. The non-randomness of the spatial distributions of these starfish populations, as evidenced by the large range in local density that was recorded in the populations of many of the species, further limits the validity of such a density comparison. For this reason, the variation in mean individual size was considered a more appropriate measure of change in the population structure of the species.