Chester Clinic

 

Ambl-ku.jp

Coral Reefs (2011) 30:227–235 Assessing feeding electivity in Acanthaster planci: a nullmodel analysis M. Tokeshi • J. R. P. Daud Received: 8 February 2010 / Accepted: 26 October 2010 / Published online: 21 November 2010Ó Springer-Verlag 2010 Feeding electivity was investigated in a non- Acanthaster. The present study emphasizes the merit of outbreaking population of the crown-of-thorns starfish testing the observed patterns, using null models for a rigor- Acanthaster planci (L.) from North Sulawesi, Indonesia. A ous assessment of feeding preferences.
null model-based approach was used to assess the feedingpattern of Acanthaster in relation to the availability of coral Electivity index  Crown-of-thorns starfish  prey in the field. Of a total of 70 species of corals recorded Faviidae  Prey selection  Corallivore  Sulawesi as prey, massive species, particularly of Faviidae, tended tobe more frequently consumed than would be expectedunder the assumption of random feeding by A. planci.
Branched and encrusting/laminar forms of corals thatoccurred in relatively exposed sites were apparently not Among many organisms associated with coral reefs, the preferred, pointing to the importance of non-acroporan crown-of-thorns starfish Acanthaster planci (L.) is one of massive species of corals in cryptic habitats as prey for the most important predators of corals with its wide dis- A. planci. The null model-based electivity index Z intro- tribution in the Indo-Pacific and destructive effects upon duced here directly measured the deviation from random coral reefs. A large body of information has been accu- feeding, while two common indices (Ivlev's and Vander- mulated on its general biology and ecology (Moran ; ploeg and Scavia's) only partially reflected such deviations Birkeland and Lucas ), including the aspects of pre- (hence, prey selection cannot be accurately demonstrated dation. Particular attention has been drawn to its ability to by these). Electivity values (Z) for poritid species and feed on scleractinian corals and to cause massive mortality Acropora palifera, the most common Acropora species in among them; a widely held view is that the species pref- the study site, were significantly negative, indicating erentially feeds on tabular and branched corals of the genus apparent avoidance of them by Acanthaster. Our results Acropora (Keesing and Lucas De'ath and Moran indicate that accessibility to different coral species and the ; Pratchett On the other hand, while much choice/avoidance of certain species are the important ele- attention has been drawn to its population outbreaks and ments of feeding in non-outbreaking populations of consequent, often dramatic, decimation of coral reefs Acanthaster inhabiting spatially variable reef environments.
(Done ; Yamaguchi ; Birkeland and Lucas ; A similar consideration may apply to the feeding patterns of Berumen and Pratchett Pratchett ; Pratchett other corallivores that possess superior/inferior mobility to et al. Kenyon and Aeby there is remarkablylittle detailed information on its predatory habit in differentreef sites during the periods in which the species occurs at Communicated by Environment Editor Prof. Mark Hay relatively low- to intermediate-densities. It has been sug-gested that the starfish's feeding behavior may change with M. Tokeshi (&)  J. R. P. Daud their densities, with food selectivity being high at low AMBL-Kyushu University, Amakusa, Kumamoto 863-2507, densities and declining with increasing densities and scar- city of corals toward the end of an outbreak (Birkeland and Coral Reefs (2011) 30:227–235 Lucas ). For a comprehensive understanding of the fisheries and the local diving industry, there have been ecology of this important predator, more information is heavy concentrations of A. planci since 2003 ([20 indi- needed on its feeding patterns, particularly under non- viduals ha-1 in 2006–2009) that have caused extensive reef damages, alongside anthropogenic influences (due to One of the difficulties in deciphering the feeding pat- people using reefs in various manners; fishing, shellfish terns of versatile predators such as Acanthaster in open collection, trampling, abrasion by snorkelers, etc.). The habitats concerns the assessment of preferential/non-pre- outbreaking populations of Acanthaster gradually moved ferential feeding. Even if laboratory/controlled field from east to west within the southern reefs of Bunaken, and experiments indicate a possibility of selective feeding on coral devastation was widespread in the south-central reef some prey categories, data from natural field conditions flats by November 2009; however, the reefs on the eastern cannot readily be interpreted as showing the occurrence of coast of the island, with distinct topography and reef preferential/non-preferential feeding, unless the observed morphology, were free from Acanthaster outbreaks until patterns are demonstrably different from patterns expected the end of 2009.
by chance. In the case of Acanthster feeding, however, Feeding habits of A. planci were observed in the shallow there has been no study to date that has rigorously tested reef habitats (about 4 km long by 80–150 m wide) on the the departures of field-observed data from random or by- eastern side of Bunaken Island, where reef flats gradually chance patterns. Further, while electivity indices have gave way to reef slopes offshore without forming apparent sometimes been used to interpret data on feeding patterns, reef edges or crests. Shallow reefs between about 1–7 m of the relationship between index values and the degree of depth, where scleractinian corals were most abundant, were deviation from randomness has rarely been checked or systematically searched using SCUBA, taking note of the clarified. It is, therefore, important to recognize that dif- occurrence of corals with feeding scars (dead patches of ferent electivity indices may vary in their indications of coral, 0.2–0.5 m in longest dimension) and Acanthaster feeding preference and avoidance patterns.
individuals in contiguous quadrats (10 by 5 m). In areas The present study employs a null model approach for with low density of feeding scars and Acanthaster, 50 explicitly assessing deviations from random patterns of by 10 m quadrats were also employed, depending on the feeding, while taking into account the availability of dif- local topography. A. planci individuals we observed in our ferent prey, in order to elucidate the feeding patterns of study site were all similar in size (in the range of Acanthaster under non-outbreak conditions. While null 20–30 cm diameter) and hence believed to be of the same model-based analyses have been adopted in a variety of cohort. The relative abundances of coral species were studies on community structure and interspecific patterns separately assessed using a total of 27 transects, each 10-m (Gotelli and Graves Tokeshi ; Arakaki and long, noting the shapes and sizes of colonies traversed by a Tokeshi ), there is a paucity of such analyses on the transect line.
feeding ecology of marine predators.
For comparing coral growth forms, we have adopted a simple system of classification with 5 forms only: massive,encrusting, laminar, branched, and tabular. While many Materials and methods corals change forms as they grow and in different micro-habitats, we designated a most representative form observed for each coral species in the state they wereconsumed by A. planci. Growth forms of ‘free-living' Data on the feeding of Acanthaster were collected on the fungiid species were categorized as ‘laminar', because their eastern side of Bunaken Island (124°470E, 1°370N) in North exposed body surfaces were roughly in one plane.
Sulawesi, eastern Indonesia, in 2006–2008. According toour continuous observation of coral assemblages since Null models and a new electivity index 2000, the occurrence of A. planci on the eastern reefs ofBunaken Island has been low to intermediate with no Feeding data were analyzed by comparing the observed evidence of past outbreaks or reef devastation (Acanthaster frequencies of feeding on different coral species with density of 0.7–6.5 individuals ha-1 in 2005–2009, which is values derived from null models in which starfishes were well below the suggested threshold levels of outbreak- assumed to forage randomly on available corals. For cal- culating the expected frequencies of feeding under the Reef; e.g., C15 individuals ha-1, Moran and De'ath assumption of no positive/negative feeding choices, corals ; [10 ha-1, Keesing and Lucas In contrast, on consumed were selected randomly among all species partially enclosed reefs of the southern side of the island, depending on the availability of each species for the same which have been heavily exploited by both subsistence number of feeding incidences (feeding scars) as observed, Coral Reefs (2011) 30:227–235 and the process was replicated 10,000 times. The avail- ability of different coral species was expressed as propor-tional abundance values in two different ways: model 1 Feeding electivity in Acanthaster used total estimated cover (area) of colonies, while model 2used total perimeter lengths. Thus, these models assumed A total of 46 coral genera recorded from the study site on that Acanthaster's choice of coral prey was dependent the east coast of Bunaken divided into two equal-numbered upon their relative abundances expressed as either one- or groups in terms of feeding preferences by Acanthaster, one two-dimensional measures of colony size. This analysis with positive values of electivity (Z) and the other with was conducted separately for two different taxonomic negative values (Fig. Genera of the family Faviidae resolutions: the genera-based classification and the species- dominated the group with positive Z values, with Favites, based classification of coral prey. Using these models, Goniastrea, and Montastraea showing the three highest electivity (Z) of a particular coral species or genus by electivity values and seven out of ten highest ranking Acanthaster was calculated as where g0 (g1) is the proportion of simulations with thecalculated values of feeding frequency smaller (larger) thanthe observed value (0 B g0, g1 B 1.0 and 0 B g0 ?g1 B 1.0). Z ranges from 1.0, the maximum (positive)electivity, to -1.0, the minimum (negative) electivity,while values around zero indicate random feeding based onavailability. This index has an advantage of showing thetendencies of preference and avoidance as positive andnegative values and enables the comparisons of uncon-sumed potential prey alongside the assessment of thoseprey actually consumed. Another advantage is that indexvalues can closely reflect their statistical significance.
The positive/negative electivity values were judged to besignificant (i.e., significant preference or avoidance of aparticular species), if max (g0, g1) [ 0.975, correspondingto P 0.05 in a two-tailed test (max (g0, g1) [ 0.995 forP 0.01 and max (g0, g1) [ 0.9995 for P 0.001).
The performance of this index was compared with two other indices of electivity, Ivlev's ((Ei) and Van-derploeg and Scavia's ) (Wi), using the same feedingdata. The indices are Ei ¼ ðri  piÞ=ðri þ piÞ Wi ¼ ðX  n1Þ=ðX þ n1Þ where X ¼ ðri=piÞ=Rðri=piÞ where ri and pi are the proportions of prey/food i in the dietand in the environment, respectively, and n is the numberof kinds of food. Another index by Jacobs (Di = (ri - pi)/(ri ? pi - 2ri pi), was also tested, but itsvalues were very close to Ivlev's (as expected from itsformulation) and, therefore, not included in the results.
While these indices are supposed to give a quantitativemeasure of deviation from random feeding, no formalexamination of such deviation has been conducted. These Fig. 1 Electivity (Z) values of different coral genera as prey toAcanthaster planci in the shallow reef habitats on the east coast of indices were each plotted against Z to see how well the Bunaken, North Sulawesi. The mean of Z values derived from models latter's data structure (i.e., degree of departure from ran- 1 and 2 is given for each genus. A significant departure from domness) is reflected in E or W.
randomness is indicated by ***P 0.001, **P 0.01, *P 0.05 Coral Reefs (2011) 30:227–235 genera belonging to this family. In contrast, the group with Table 1 List of coral species preyed upon by Acanthaster planci in negative Z values included diverse genera belonging to east Bunaken, North Sulawesi, Indonesia different families, with Acropora and Porites showing strongly negative electivity values. Of those genera with relatively weak electivity values (-0.5 Z 0.5), nega-tive values occurred more frequently than positive ones, indicating a fairly sharp discrepancy between preferred and Goniastrea minuta non-preferred prey groups.
Goniastrea edwardsi From a total of 154 feeding scars observed, 70 species of corals were identified as being consumed by Acanthaster Favia lizardensis (Table ). Electivity values were variable among different coral species, ranging from 1 to -1, but the majority of species (61 spp.) yielded positive values with only nine species demonstrating negative ones. The values from two Favia helianthoides null models were highly correlated (P 0.001) and yiel- ded similar results. Of all species, 18 species had positive Z values that were judged to be significantly large on the basis of model 1, of which 11 species belonged to Faviidae.
Favites stylifera While species of Faviidae showed predominantly positive electivity values, Acroporidae showed a mixed pattern with Leptastrea bottae both positive and negative electivity values among differ- Oulophyllia bennetae ent species. Of nine species that showed negative electiv- Oulophyllia crispa ity, values for Acropora palifera and Porites cylindrica were significant (P 0.01 and P 0.001, respectively) on Platygyra ryukyuensis the basis of model 1, indicating a clear tendency of Platygyra sinensis avoidance of these species by Acanthaster.
Platygyra verweyi In terms of coral morphology, massive forms were more predominant (34 spp.; Table Fig. ) than branched ones Montastrea colemani (18 spp.) among consumed corals, while encrusting and Montastrea valenciennesi laminar forms were also frequently consumed (18 spp.).
Cyphastrea ocellina Among massive corals, Goniastrea spp. and Favites abdita Cyphastrea agassizi were common and preferred prey (Z & 1, P 0.01), andspecies with large polyps such as Symphillia and Oulo- phyllia (Fig. f) were also consumed. Other common Acanthastrea faviaformis prey were encrusting/laminar forms of Montipora, includ- Lobophyllia hemprichii ing M. danae and M. grisea. It was also notable that a species of Millepora, non-scleractinian corals generally reported to be free from Acanthaster predation (Endean Coeloseris mayerii ; Glynn ; Lewis but see Keesing ), was Pachiseris gemmae observed to fall prey (Fig. d). Among branched forms, Hydnophora rigida and Seriatopora histrix were frequently Physogyra lichtensteini consumed (Fig. h), with significantly high levels of feeding preference (Z & 1, P 0.01).
Echinopora lamellosa Performance of electivity indices Plotting of Ivlev's E and Vanderploeg and Scavia's W against Z (Fig. ) shows that E and W indices do not Montipora confusa necessarily reflect deviation from random feeding. For relatively large deviation from randomness (e.g., Z [ 0.8), Montipora efflorescens both E and W tend to show increasing scatter (variation along y axis), indicating that the degree of deviation is not Coral Reefs (2011) 30:227–235 Table 1 continued the minimal values (-1) of these indices do not indicatedepartures from random feeding.
Montipora informis Montipora foliosa Evaluation of feeding electivity Montipora porites Montipora hispida The present study points to the importance of massive Montipora turtlensis corals, particularly the species of Faviidae, as food for a non-outbreaking population of A. planci in North Sulawesi.
Acropora solitaryensis The data clearly demonstrated that faviid species tended to Acropora brueggemanni be more preferred by A. planci predators than the species of Acropora latistella other families. In particular, Z values for faviids were Acropora cytherea significantly higher than those for acroporid species Acropora palifera (Mann–Whitney U test, P 0.001). At the same time, the Acropora kimbeensis study emphasizes the importance of rigorously assessing the preferential/non-preferential feeding of different coral Astreopora myriophthalma taxa by A. planci. In this context, it is worth noting that the widely recognized consumption of tabular and branched Pocillopora damicornis Acropora species by A. planci is not necessarily a fixed or Pocillopora meandrina universal behavioral pattern under different reef conditions.
Pocillopora eydouxi In order to evaluate the feeding tendencies of A. planci, Pocillopora verrucosa the present study adopted an index of electivity that mea- Stylophora pistillata sured the degree of departure from a presumed random Seriatopora hystrix mode of feeding. This is based on the idea that an observed pattern of feeding needs to be assessed against the back- Porites cylindrica ground of non-selective feeding, and this can only be Alveopora marionensis accomplished through simulation of random feeding pro- cesses. Just as feeding of a particular prey item may be due Merulina scabricula either to a chance encounter or to preferential foraging, no Merulina ampliata consumption of a particular prey category may also be due Hydnophora grandis to either chance or active avoidance. Thus, it is desirable Hydnophora rigida that an index of electivity reflects the degree to which an Hydnophora microconus observed situation departs from a random situation. While this is considered important for a rigorous evaluation of non-random patterns of feeding, no study has examined the performance of conventional indices of electivity in terms Heliofungia actiniformis of deviations from randomness. The analysis in the present study suggests that use of two common indices of electivity Millepora sp.
is problematic as their values only partially mirror thedegree of deviation from random feeding. The problem is Growth forms are indicated as: m massive, e encrusting, l laminar, most acute for the assessment of significant feeding pref- b branched, t tabular. Electivity (Z) values based on two null modelsare shown (* 0.05, ** 0.01, *** 0.001) erences, as the values from both Ivlev's index E andVanderploeg and Scavia's W are quite variable for rela-tively high values of Z ([0.8). In other words, both of these accurately expressed by these values. In the case of coral indices tend to underestimate the cases of significant species with significantly large departures of positive Z preferences; for example, feeding on Faviidae species values, corresponding values of E varied from 0.470 to cannot reliably be assessed by these. At the same time, 0.977 and W from -0.198 to 0.908. On the other hand, both these indices tend to indicate incorrectly negative food indices correlated well with Z under situations not signifi- selection where in fact there is no significant departure cantly departing from randomness (within an intermediate from randomness. Indeed, an example of somewhat com- range of Z, -0.5 Z 0.5). The plot clearly shows that plex results of prey selection in corallivorous gastropods

Coral Reefs (2011) 30:227–235 Fig. 2 Acanthaster plancifeeding on different species ofnon-acroporan corals inBunaken, eastern Indonesia.
a Goniastrea edwardsi, notealso untouched Acroporapalifera in the right front;b Leptastrea bottae; c Faviastelligera; d Millepora sp.;e Oulophyllia crispa; f Detailsof Oulophyllia crispa withmesenterial filaments extendedto resist in vain the attack ofA. planci; g Leptoria phrygia;h Hydnophora rigida(consumed Seriatopora hystrixalso visible in the foregroundand background; note alsountouched Acropora palifera inthe left back) (Morton and Blackmore ) may be partly due to the use affect their availability to A. planci predators. In this of Jacobs's index (a slight modification of E, see Materials regard, given the complex three-dimensional structures of and Methods).
reef environment, the two-dimensional measure of colony In evaluating feeding electivity, as the encounter rate size (model 1) is probably more relevant than the one- would vary with the availability of different prey under dimensional measure (model 2), though largely similar random feeding, some assumptions must be made regard- results were obtained by both measures.
ing prey encounter. However, in the absence of precise An analysis of feeding patterns of A. planci from the information on how A. planci detects prey in natural reef Great Barrier Reef (De'ath and Moran revealed that environments, some uncertainties always remain as to species of Acropora were most preferred, followed by which characteristics of coral colonies would most strongly those of Montipora, while massive corals particularly of Coral Reefs (2011) 30:227–235 particularly low value of Z (=-0.9977) despite being the most common species of Acropora in the shallow reefhabitats (Fig. a, h), suggesting that Acropora species arenot uniformly preferred by the starfish.
Variable feeding in Acanthaster and other corallivores While factors such as ‘innate' preferences and previousfeeding experiences either as juvenile or as adult mayinfluence prey selection (Collins Ormond et al. ;Pratchett accessibility to particular coral coloniesmust also constitute an important, sometimes overriding,factor. In our study site, acroporids in exposed, fast-currenthabitats and table-formed species standing on a thin centralstalk may be far less accessible prey than massive corals of various sizes in a close vicinity of a starfish's refugia.
There are also possible deterrence effects of coral symbi-onts such as toxic gobies and crustaceans that occur inassociation with different species of Acropora (Lassig; Glynn , Pratchett ). Accessibilityalso appears to be an important factor in the case of solitaryfungiid species that are often found in microhabitats noteasily accessible for A. planci, such as small crevices oramong other corals, rubble substrates; indeed, individualsof fungiids in accessible open locations were seen to bereadily consumed by the starfish.
The present study has clearly demonstrated that Acanthaster in the natural environment is capable ofpreying upon a diversity of coral species, particularly Fig. 3 Relationship between electivity index Z and a Ivlev's index massive ones. Previous studies on the Great Barrier Reef E and b Vanderploeg and Scavia's index W also drew attention to the predation of massive corals byA. planci and their slow recovery after A. planci outbreaks the genus Porites were not favored. Our result also con- (Endean et al. ; Cameron et al. ). One possibility firms that massive species of Porites such as P. lobata and regarding the starfish's predation behavior is that at low- to P. lutea, which were very common in the study site intermediate-densities crown-of-thorns individuals may (*65% occurrence in all quadrats), were not consumed. It feed on corals that are close to their shelters, most typically is also worth noting that branched species of Porites such the underside of a coral rock. Starfishes are likely to as P. nigrescens and P. attenuata were not preyed upon by encounter massive and encrusting corals when they crawl Acanthaster despite their commonness (combined occur- out of such hiding places in most shallow reef areas of rence *53%); the most abundant of branched Porites, eastern Bunaken. Feeding on a variety of non-acroporan P. cylindrica, showed the minimal value of Z (=-1.0).
corals has also been reported for Acanthaster populations Therefore, irrespective of growth forms, poritid species in in the subtropical waters of Okinawa, southern Japan eastern Bunaken are generally not prone to consumption by Acanthaster. While some studies reported Acanthaster Apart from A. planci, various reef-associated taxa have feeding on Porites (Goreau ; Barnes et al. Glynn been known to demonstrate a range of obligate and facul- ; Done we rarely found such feeding. Several tative corallivory and feeding electivity (Cumming ; traits of Porites may explain low preference by Acanth- Cole et al. ; Rotjan and Lewis ). However, except aster in our study; these include low energetic values, in cases where feeding is apparently restricted to a single or release of repellent substances (Brauer et al. and the few target species (e.g., a butterflyfish Chaetodon trifas- presence of symbiont species that deter the attack by sea- cialis feeding almost exclusively on Acropora hyacinthus, stars (DeVantier et al. DeVantier and Endean ; Pratchett ), it is not straightforward to recognize Pratchett Another interesting case concerns Acro- feeding preferences/avoidances in the absence of a null pora palifera, a species with stout branches, that had a model type test on field-derived data. The present study Coral Reefs (2011) 30:227–235 draws attention to the importance of assessing the depar- Collins ARS (1975) Biochemical investigations of two responses tures of observed feeding patterns from randomly expected involved in the feeding behaviour of Acanthaster planci (L.). 3.
Food preferences. J Exp Mar Biol Ecol 17:87–94 patterns. In this regard, as mobility and life-time move- Cumming RL (1999) Predation on reef-building corals: multiscale ments vary among corallivore taxa (fishes, molluscs, etc.), variation in three corallivorous gastropods, Drupella spp. Coral the availability/accessibility of different coral species to a Reefs 18:147–157 particular corallivore needs to be carefully assessed. For De'ath G, Moran PJ (1998) Factors affecting the behaviour of crown- of-thorns starfish (Acanthaster planci L.) on the Great Barrier example, while highly mobile corallivorous fishes (e.g., Reef. 2: Feeding preferences. J Exp Mar Biol Ecol 220:107–126 Chaetodontidae) may have access to nearly all coral col- DeVantier LM, Endean E (1988) The scallop Pedum spondyloideum onies within a patch of reef, molluscan predators with more mitigates the effects of Acanthaster planci predation on the host restricted mobility (e.g., Coralliophila, Hayes ; Oren coral Porites: host defence facilitated by exaptation. Mar EcolProg Ser 47:293–301 et al. Drupella, Cumming ; Morton and DeVantier LM, Reichelt RE, Bradbury RH (1986) Does Spirobran- Blackmore ) may only have limited access to different chus giganteus protect host Porites from predation by Acanth- coral species on relatively small spatial scales. It remains to aster planci: predator pressure as a mechanism of coevolution.
be seen to what extent different cases of corallivory (e.g., Mar Ecol Prog Ser 32:307–310 Done TJ (1985) Effects of two Acanthaster outbreaks on coral in Drupella, Morton and Blackmore can be recog- community structure—the meaning of devastation. Proc 5th Int.
nized as significant departures from random feeding.
Coral Reef Symp 5:315–320 Predation patterns in reef environments apparently vary Endean R (1973) Population explosions of Acanthaster planci and depending on many factors including the local distributions associated destruction of hermatypic corals in the Indo-Pacificregion. In: Jones OA, Endean R (eds) Biology and geology of of different coral species. Given the complex variation in coral reefs. vol II: Biology 1. Academic Press, New York, species composition and local topography of different reef sites, more research is needed to gain a better under- Endean R, Cameron AM, DeVantier LM (1989) Acanthaster planci standing of the feeding behavior of corallivores such as A.
predation on massive coral: the myth of rapid recovery ofdevastated reefs. Proc 6th Int Coral Reef Symp 2: 143–148 planci under different environmental conditions in a range Glynn PW (1973) Aspects of the ecology of coral reefs in the western of spatio-temporal contexts.
Atlantic region. In: Jones OA, Endean R (eds) Biology andgeology of coral reefs. vol II: Biology 1. Academic Press, New Thanks are due to S. Arakaki, M. Yoko-o, and York, pp 271–324 other members of the Amakusa Biological Laboratory for various Glynn PW (1974) Acanthaster: effect on coral reef growth in Panama.
forms of help. This research was supported by the Kyushu University Science 180:504–506 P & P program, a Japanese Government postgraduate studentship (to Glynn PW (1980) Defence by symbiotic Crustacea of host corals JRPD), the 21st century COE and the GCOE programs (at Kyushu elicited by chemical cues from predator. Oecologia 47:287–290 University) of the Japan Ministry of Education, Culture, Sports, Glynn PW (1987) Some ecological consequences of coral-crustacean Science, and Technology, and the grants-in-aid for scientific research guard mutualism in the Indian and Pacific Oceans. Symbiosis from the Japan Society for the Promotion of Science.
Goreau TF (1964) On the predation of coral by the spiny starfish Acanthaster planci (L.) in the southern Red Sea. Sea Fish ResStn Haifa Bull 35:23–26 Gotelli NJ, Graves GR (1996) Null models in ecology. Smithsonian Institution Press, Washington Arakaki S, Tokeshi M (2010) Analysis of spatial niche structure in Hayes JA (1990) Prey preference in a Caribbean corallivore, coexisting tidepool fishes: null models based on multi-scale Coralliophila abbreviata (Lamarck) (Gastropoda, Coralliophil- experiments. J Anim Ecol 79. doi: idae). Bull Mar Sci 47:557–560 Ivlev VS (1961) Experimental ecology of the feeding of fishes. Yale Barnes DJ, Brauer RW, Jordan RW (1970) Locomotory response of University Press, New Haven, Connecticut Acanthaster planci to various species of coral. Nature 228: Jacobs J (1974) Quantitative measurement of food selection. Oeco- logia 14:413–417 Berumen ML, Pratchett MS (2006) Recovery without resilience: Keesing JK (1992) Influence of persistent sub-infestation density persistent disturbance and long-term shifts in the structure of fish Acanthaster planci (L.) and high density Echinometra mathaei and coral communities at Tiahura Reef, Moorea. Coral Reefs (de Blainville) populations on coral reef community structure in Okinawa, Japan. Proc 7th Int Coral Reef Symp 2:769–779 Birkeland C, Lucas JS (1990) Acanthaster planci: a major manage- Keesing JK, Lucas JS (1992) Field measurement of feeding and ment problem of coral reefs. CRC press, Boca Raton movement rates of the crown-of-thorns starfish Acanthaster Brauer RW, Jordan MR, Barnes DJ (1970) Triggering of the stomach planci (L.). J Exp Mar Biol Ecol 156:89–104 eversion reflex of Acanthaster planci by coral extracts. Nature Kenyon J, Aeby GS (2009) Localized outbreak and feeding prefer- ences of the crown-of-thorns seastar Acanthaster planci (Echi- Cameron AM, Endean R, DeVantier LM (1991) The effects of nodermata. Asteroidea) on reefs off Oahu, Hawaii. Bull Mar Sci Acanthaster planci predation on populations of two species of massive coral. Hydrobiologia 216(217):257–262 Lassig BR (1977) Communication and coexistence in a coral Cole AJ, Pratchett MS, Jones GP (2008) Diversity and functional community. Mar Biol 42:85–92 importance of coral-feeding fishes on tropical coral reefs. Fish Lewis JB (1989) The ecology of Millepora: a review. Coral Reefs Coral Reefs (2011) 30:227–235 Moran PJ (1986) The Acanthaster phenomenon. Oceanogr Mar Biol Pratchett MS (2005b) Dietary overlap among coral-feeding butter- Annu Rev 24:379–480 flyfishes (Chaetodontidae) at Lizard Island, northern Great Moran PJ, De'ath G (1992) Estimates of the abundance of the crown- Barrier Reef. Mar Biol 148:373–382 of-thorns starfish Acanthaster planci in outbreaking and non- Pratchett MS (2007) Feeding preferences of Acanthaster planci (L.) outbreaking populations on reefs within the Great Barrier Reef.
under controlled conditions of food availability. Pac Sci Mar Biol 113:509–515 Morton B, Blackmore G (2009) Seasonal variations in the density of Pratchett MS, Schenk TJ, Baine M, Syms C, Baird AH (2009) and corallivory by Drupella rugosa and Cronia margariticola Selective coral mortality associated with outbreaks of Acanth- (Caenogastropoda: Muricidae) from the coastal waters of Hong aster planci L. in Bootless Bay, Papua New Guinea. Mar Kong: ‘plagues' or ‘aggregations'? J Mar Biol Assoc UK Environ Res 67:230–236 Rotjan RD, Lewis SM (2008) Impact of coral predators on tropical Oren U, Brickner I, Loya Y (1998) Prudent sessile feeding by the reefs. Mar Ecol Prog Ser 367:73–91 corallivore snail, Coralliophila violacea on coral energy sinks.
Tokeshi M (1999) Species coexistence: ecological and evolutionary Proc R Soc Lond B 265:2043–2050 perspectives. Blackwell Science, Oxford Ormond RFG, Hanscomb NJ, Beach DH (1976) Food selection in the Vanderploeg HA, Scavia D (1979) Calculation and use of selectivity crown-of-thorns starfish Acanthaster planci (L.). Mar Behav coefficients of feeding: zooplankton grazing. Ecol Model Physiol 4:93–105 Pratchett MS (2001) Influence of coral symbionts on feeding Yamaguchi M (1986) Acanthaster planci infestations of reefs and preferences of crown-of-thorns starfish. Mar Ecol Prog Ser coral assemblages in Japan: a retrospective analysis of control efforts. Coral Reefs 5:23–30 Pratchett MS (2005a) Dynamics of an outbreak population of Acanthaster planci at Lizard Island, northern Great Barrier Reef(1995–1999). Coral Reefs 24:453–462

Source: http://ambl-ku.jp/tokeshi/tokeshiDaud2011.pdf