availability, less water flow, or larval scarcity as depth 

 increases. 



INTRODUCTION 



Benthic assemblages inhabiting rock surfaces often take many 

 years to develop on initially cleared rock or other hard surfaces 

 (Osman, 1977, 1978; Sebens, 1985, 1986; Paine and Levin, 1981). 

 The time scale of recolonization depends on the availability of 

 propagules of the component species, their advection toward the 

 substratum, and on their interactions with species already 

 inhabiting the surface. Most studies of epifaunal encrusting 

 community development have been conducted in shallow water 

 temperate habitats (Karlson, 1978; O'Connor, et al. 1980; Kay and 

 Keough, 1981; Sebens, 1985, 1986) or on the undersides of coral 

 plates in cryptic reef habitats (Jackson, 1977a, b, 1979; Buss and 

 Jackson, 1979; Buss, 1980). Colonization of artificial or 

 natural surfaces in deeper water, and the type of assemblages 

 that develop, are virtually unknown. 



Experiments in early community development often employ 

 artificial plates constructed of plastic, rock, asbestos or other 

 suitable material, frequently roughened to enhance successful 

 larval settlement and adhesion (Sutherland, 1975; Sutherland and 

 Karlson, 1971; Schoener, 1974; Schoener et al . , 1979; Shin, 1981; 

 Harris and Irons, 1982). Such plates are easy to monitor because 

 of their flat surfaces and identical sizes. They are, however, 

 usually separated from the substratum and thus from certain 

 benthic predators which could influence community processes. On 

 the other hand, they are also separated from adjacent sessile 

 species which could colonize plates by lateral growth. Suspended 

 plates thus sample recruitment (minus mortality) of species with 

 larvae in the water column just above the bottom. Alternatively, 

 other studies have used scraped natural surfaces in contact with 

 the adjacent substratum and its occupants (Dayton, 1971, 1975; 

 Kay and Keough, 1981; Sebens, 1985, 1986; Witman 1987). This 

 method has the advantage of allowing all normal community effects 

 to influence settlement, recruitment, and subsequent growth and 

 survival of colonists. However, it is then difficult to 

 distinguish colonization from the water column from lateral 

 growth or asexual proliferation by adjacent encrusting species. 

 Also, scraped areas often retain some fragments of colonies or 

 algal thalli within small crevices, and thus there is another 

 source or recolonization. It has been argued that this method 

 mimics naturally occurring patches more closely than do any of 

 the artificial substratum methods. Finally, at least one study 

 conducted in the rocky intertidal zone followed naturally created 

 patches of a wide size range for several years, quantifying the 

 changes in patch size and occupants (Paine and Levin 1981). 



In the present study, we used a number of different 

 experimental techniques to explore early community development at 

 four depths (30, 50, 65, 80 m) along a transect down the slope of 

 Ammen Rock Pinnacle, on Cashes Ledge in the center of the Gulf of 



46 



