BARHAM, GOWDY, and WOLFSON: ACANTHASTER IN GULF OF CALIFORNIA 



reflect the distribution of corals in the Gulf of 

 California. From an ecological standpoint, the 

 ranges for coral genera given in the most re- 

 cent treatment of the subject (Squires, 1959) 

 can be misleading. For example, Squires shows 

 the northern limit for extant Pocillopora species 

 as a line extending across the Gulf from about 

 the level of Santa Rosalia to Guaymas. The 

 line is based on a single specimen of Pocillopora 

 robiista (= elegans) collected by an Allan Han- 

 cock Expedition at Isla San Marcos (Durham 

 and Barnard. 1952). John Garth of the Hancock 

 Foundation informs us that the specimen was 

 dredged in 20 fathoms south of Isla San Marcos 

 (lat. 27°09'05"N; long. 112°04'25"W) and 

 appears to have been taken alive. In our experi- 

 ence, the only common coral north of the 

 island complex is Pontes, and this Isla San 

 Marcos specimen must represent an isolated 

 outpost of Pocillopora. Squires' Table 1, sum- 

 marizing the diminution of the hermatypic coral 

 fauna in the northern regions of the Gulf, gives 

 a more accurate picture of the situation as we 

 have seen it. It seems obvious, then, that if 

 A. ellisii is primarily an obligate coral feeder, 

 its major prey through the northern part of its 

 range must be Porites. 



The varying growth forms of the corals may 

 also enter into the prey preference of A. ellisii. 

 In gross morphology, Porites is highly plastic, 

 showing phenotypic gradation from encrust- 

 ing colonies in the northern part of its range to 

 columnar, branched, or nodular forms in the 

 south. However, the influence of latitude is can- 

 celled in protected situations in the lower mid- 

 Gulf, such as Bahia Coyote and the inner bay 

 at Puerto Escondido, and Porites tends to 

 assume the dome-shaped heads and large nodular 

 form typical of the southern type. We have 

 never seen A. ellisii in these quiet-water en- 

 vironments. 



It is noteworthy that we have never seen A. 

 ellisii on large or ramose corals at any locality. 

 The Porites being preyed on in all of our obser- 

 vations were so small that the sea stars appeared 

 to be sitting on a coral-free substrate; only 

 when the animal was lifted could the coral be 

 seen and identified. Furthermore, our observa- 

 tions of predation on Pocillopora all involved 

 feeding at substrate level. In a crude field trial, 



a traveling specimen immediately began to feed 

 on a broken branch that we had placed in its 

 path; and the three observations at Pink Cliff 

 station (all by specimen no. 8) involved one 

 instance of feeding on the underside of a low 

 branch and two on broken branches. However, 

 at the end of the study, no. 8 fed on a ramose 

 Pocillopora when dropped directly on top of 

 it, and Faulkner (pers. comm.) reports that 

 south of La Paz, where Pocillopora assumes a 

 densely packed, reeflike formation, A. ellisii 

 feeds on top of the colonies. It thus appears 

 that some factor or factors may prevent the 

 asteroid from mounting upright, solitary corals. 



Weber and Woodhead (1970) reported that 

 fish and crabs living among the branches of 

 Stylopora, a close relative of Pocillopora, de- 

 fend that coral from attack by A. pla)ici. Pocil- 

 lopora's symbionts (Figure 3) perhaps serve 

 a similar function. 



The experimental work of Barnes, Brauer, and 

 Jordan (1970) and Brauer, Jordan, and Barnes 

 (1970) may explain the marked preference 

 of A. ellisii for discrete, low coralla. Working 

 with A. plaiici, these investigators demonstrated 

 that, while the tube feet are sensitive to nema- 

 tocysts, the stomachs respond positively to 

 nematocyst-containing tissue. From these 

 experiments, and from the observation that the 

 sea star uses its arms and spines as much as 

 possible when climbing coral heads, they infer 

 that the humped feeding position of A. pla)ici 

 may serve to keep the tube feet free of the prey's 

 stinging polyps. If A. ellisii responds to nemato- 

 cysts in the same ways, small, isolated, crustose 

 coralla that can be easily straddled and then 

 covered by the stomach would constitute more 

 acceptable prey and would be selected over 

 larger growths. 



Examination of small Porites fed upon by 

 A. ellisii suggests that the entire colony is not 

 necessarily destroyed; white patches, which 

 indicate that the zooxanthellae have been di- 

 gested, are left mainly on regions that protrude 

 above the general level of the substrate. We 

 consistently observed, however, that the entire 

 corallum appears to be covered with a mucous 

 film, and suggest that the extent of the damage 

 inflicted on small Porites by A. ellisii bears 

 further investigation. 



939 



