FISHERY BULLETIN: VOL. 71. NO. 4 



Behavior 



Table 3 summarizes the 56 observations made 

 during the 48 h of the Pink Cliff study. In 38 

 instances (68% ), the specimen was feeding, that 

 is, it was immobile, hunched or puffed up (Figure 

 4a), with stomach everted and covering the 

 prey. Feeding was observed in 31 of the 43 

 daytime observations (72% ) and 7 of the 13 night 

 observations (54%). On 12 occasions, 6 at night 

 and 6 during the day, the asteroids were "rest- 

 ing," i.e., they were immobile and the stomach 

 was within the gastric cavity. Six times the ob- 

 served specimen was "traveling," moving over 

 the substrate on tube feet, one slightly lifted 

 arm leading (Figure 5). Traveling was only 

 observed during the day, but the evidence is 

 insufficient for postulating a diurnal rhythm. 



Feeding 



Analysis of the 38 feeding observations indi- 

 cates a definite prey preference. Poiites com- 

 prised 79% of the food (29 observations, plus 

 one of simultaneous feeding on Porites and 

 coralline algae). PociUopora was fed on two 

 times (5% ) and coralline algae two times (5% ). 

 On four occasions (11%) it was impossible to 

 determine the prey. 



Specimen no. 8, which fed simultaneously on 

 coral and alga, was also the sole predator on 

 PociUopora. Unusual behavior was also exhibit- 

 ed by no. 6 which, on three occasions, displayed 

 the singular habit of feeding on two or three 

 Porites colonies at once. 



Movement 



The sequence of behavior^ observed suggests 

 a general pattern of movement. The individual 

 remains for one or more hours over a small 

 Porites enci'ustation, and then moves off, usual- 

 ly to another corallum which may be several 

 meters distant. A series of feeding stops is 

 followed by a resting period, which may also 

 last for several hours, or by several rest stops 

 interspersed with short moves. It should be 

 noted that, even though specimen no. 3, 5, 6, 

 and 8 were at rest in consecutive observations 



separated by intervals of 3-13 h, in each case the 

 individual had moved between the observation 

 periods. 



The specimen which reacted abnormally to 

 the 1970 tagging operation provides some evi- 

 dence for the rate at which A. eUisii is capable 

 of moving: when returned to the substrate, it 

 immediately embarked on a continuous journey, 

 covering a distance of approximately 10 m in a 

 little over an hour. More refined observations 

 are necessary for estimation of rate of move- 

 ment, but a rough idea can perhaps be gained 

 from the Pink Cliff study. Dividing the total 

 elapsed time between midpoints of the first and 

 last observation periods (33 h) by the summed 

 distances moved (123 m) gives an average 

 ground coverage of 0.373 m/h. A higher figure 

 is derived from considering the two cases (be- 

 tween period 2-3 and 5-6) in which the inter- 

 vening intervals were approximately equal 

 (about 31/2 h). Distances covered between 

 those periods ranged from to 7.0 m and 

 averaged 2.1 m, giving a rate of 0.6 m/h. Fur- 

 thermore, average rate in the rock pile and 

 coral gardens, where coral are most abundant, 

 was 0.246 m/h, approximately half of that on 

 the relatively coral-poor platform (0.464 m/h), 

 suggesting correlation with density of the corals. 



ACANTHASTER ELLlSll 

 AS PREDATOR 



Crucial to an evaluation of A. ellisii as a 

 potentially destructive coral predator is its 

 foraging behavior throughout its range. North 

 of the island complex, small crustose colonies of 

 Porites are the only positively identified prey of 

 A. ellisii. In the island complex, which falls in 

 the middle of its known range, Porites apparent- 

 ly remains the major food item, with gorgonians, 

 algae, and PociUopora attaining minor im- 

 portance. (It might be mentioned here that feed- 

 ing by A. ellisii on PociUopora was inferred, 

 rather than actually observed, by Dana and 

 Wolfson (1970).) The few observations which 

 have been made farther to the south suggest 

 that PociUopora is more frequently fed on than 

 Porites. 



This apparent shift in prey preference may 



938 



