794 



Fishery Bulletin 92(4). 1994 



m but did not affect identification and counts, since 

 it exceeded 2 m (see Methods section). 



Species composition differed over five different 

 habitat types (Table 3). Nearly three times as many 

 species were identified from ledge habitats (63) than 



tS LEDGE 



(S3 LIVE-BOTTOM 



□ SAND 







SUMMER 1963 FALL 1983 SPRING 1986 SUMMER 1986 



Figure 3 



Number of species observed at Gray's Reef National Ma- 

 rine Sanctuary by survey and habitat. 



S3 LEDGE 



ESI LM-BOTTOM 



□ SAND 



1 





SUMMER 1985 



PAI_L 1985 



SPRING 1996 



SUMMER 1986 



Figure 4 



Density offish (estimated means) at Gray's Reef National 

 Marine Sanctuary by survey and habitat. Number offish/ 

 m over sand for fall 1985, spring 1986, and summer 1986 

 was 0.02. 



from sand habitats (22), and over one-third as many 

 species were seen on ledge as were seen on either 

 dense (46) or moderate (46) live-bottom. More spe- 

 cies were recorded over dense and moderate live-bot- 

 tom than over sparse live-bottom (33). Mean rela- 

 tive abundances also were related to habitat; high- 

 est values were found for ledge habitat, progres- 

 sively declining values for the three live-bottom 

 habitats, and lowest values for sand. On average, 

 abundances over ledges exceeded those over sand 

 bottoms by a factor of 50. 



Cluster analyses for each of the four surveys 

 (Fig. 5) indicate clear distinctions in community 

 composition among habitats. Generally, sites over 

 ledges and dense live-bottom areas were classi- 

 fied similarly and were distinct from sites found 

 over sparse live-bottom and sand. Classification 

 of moderate live-bottom sites was more variable. 

 Many species were present in more than one habi- 

 tat (Tables 1 and 3), and few individual species 

 could be considered indicators of a single habitat 

 type. The following species were present at more 

 than half the respective habitat sites. Ledges were 

 characterized by black sea bass, 6 belted sandfish, 

 gag, scamp, sand perch, round scad, tomtate, 

 sheepshead, spottail pinfish, longspine porgy, 

 cubbyu, Atlantic spadefish, slippery dick, 

 doctorfish, and planehead filefish. Dense live-bot- 

 tom was characterized by black sea bass, belted 

 sandfish, tomtate, longspine porgy, and slippery 

 dick. Moderate live-bottom had black sea bass, 

 belted sandfish, round scad, longspine porgy, and 

 slippery dick. Sparse live-bottom had black sea 

 bass, round scad, longspine porgy, and slippery 

 dick. Sand habitats were relatively depauperate 

 but were best characterized by the presence of 

 pearly razorfish. 



Discussion 



Comparison of the fauna of GRNMS with that of 

 other reefs off the southeastern U.S. coast sug- 

 gests a high level of variability among reef com- 

 munities. The fish species composition at GRNMS 

 differs considerably from that of an intensely stud- 

 ied reef in 30 m of water, 44 km south of Beaufort 

 Inlet, North Carolina (Parker, 1990). Of 113 spe- 

 cies observed by divers at the two reefs, only 42 

 (37%) were common to both (Table 1). Twenty-eight 

 species were unique to GRNMS (Table 1) and 43 

 species were seen only at the North Carolina site. 

 Although more effort was expended at GRNMS 



Scientific names of fishes in this study are listed in Tahle 1. 



