475 



Abstract-Videotapes collected by the 

 research submersible Jafio in the Ben- 

 guela ecosystem during spring 1997 

 were analyzed to determine demersal 

 nekton assemblage composition, fish be- 

 havior, and microscale habitat asso- 

 ciation, and habitat .selection. Demersal 

 fish assemblage diversity was gener- 

 ally low. and their composition was a 

 function of substratum tyjje. Individual 

 species showed an often marked asso- 

 ciation with either rocky or soft sub- 

 strata (or both), and a strong or weak 

 selection for the associated biota. Soles 

 and dragonets actively selected areas of 

 bioactive soft substrata, to which they 

 are behaviorally and morphologically 

 adapted. False jacopever were associ- 

 ated with crevices in areas of high-re- 

 lief rocky substrata. Kingklip are pis- 

 civorous and were largely confined to 

 holes at the base of rocks and favored 

 areas without a conspicuous epifauna. 

 Juvenile hake and gobies avoided ex- 

 tremely rocky areas and were largely 

 indifferent to the presence of benthic 

 invertebrates — behavior that is con- 

 sistent with their planktonic diets. 

 These results represent the first direct 

 observations of demersal nekton in 

 the region and are important because 

 they allow better interpretations of the 

 results of trawl studies. 



Habitat use by demersal nekton 

 on the continental shelf in the 

 Benguela ecosystem, southern Africa. 



Mark J. Gibbons 



Zoology Department 



University of the Western Cape 



Private Bag X 17 



Beliville 7535, Republic of Soutfi Africa 



E-mail address: mgibbonsfa'uwcacza 



Andre J. J. Goosen 

 Patti A. Wickens 



De Beers Manne (Pty) Ltd. 



PO Box 87 



Cape Town 8000, Republic of South Africa 



Manuscript accepted 23 January 2002. 

 Fish. Bull. 100:475^90 (2002). 



The Benguela ecosystem is character- 

 ized by dynamic coastal upwelling (Shan- 

 non, 1985). It is among the most produc- 

 tive ecosystems in the world (Gushing, 

 1971) and supports several commer- 

 cially valuable pelagic and demersal 

 fisheries. Although we have a fairly 

 good understanding of the role that the 

 physical and biological environment 

 play in influencing pelagic fishes (see 

 references in Payne et al., 1987, 1992), 

 the same cannot be said for the demer- 

 sal species. 



Although phytoplankton biomass and 

 production in the Benguela ecosystem 

 are high (Brown et al., 1991), there is 

 a large mismatch with zooplankton 

 (Hutchings, 1992) that results in signifi- 

 cant sedimentation (Hutchings, 1992), 

 and periodic anoxia on the bottom 

 (Bailey and Rogers, 1997). Bottom cur- 

 rents are generally sluggish to the south 

 (Shannon, 1985), and the seafloor is 

 characterized by quartz sands, diatoma- 

 ceous ooze, and foraminiferal deposits 

 (Bailey and Rogers, 1997). Trawl stud- 

 ies have revealed that demersal fish di- 

 versity in the region is low and that as- 

 semblages can be subdivided into those 

 inhabiting neritic, shelf, and slope areas 

 (Macpherson and Roel, 1987; Roel, 1987; 

 Macpherson and Gordoa, 1992; Smale et 

 al., 1993). Correlations between demer- 

 sal assemblages and the biophysical 

 milieu (including landscape) have not 



been made owing to the general lack 

 of environmental data, especially at the 

 microscale. 



Submersibles provide a way to sample 

 and observe the biophysical environ- 

 ment of the sea floor. Although the 

 method precludes an appreciation of 

 the infauna, the technique allows for 

 the simultaneous collection of informa- 

 tion on the macroscopic benthic fauna 

 and fishes, as well as on the physical 

 features of the environment, across a 

 wide variety of substrata. Elsewhere 

 in the world, submersibles have been 

 used to describe the physical and bio- 

 logical environment of benthic commu- 

 nities (Sibuet et al. , 1988) to determine 

 habitat use by nekton (Parker and 

 Ross, 1986; Felley and Vecchione, 1995) 

 and in data collection for stock assess- 

 ment (Parker and Ross, 1986; Giguere 

 and Brulotte, 1994). 



During 1996, the research submers- 

 ible Jago was used to examine some 

 offshore areas near the mouth of the 

 Orange River. A total of 17 dives were 

 made and the archived videotapes were 

 subjected to semiquantitative biologi- 

 cal analysis (Gibbons et al., 2000). The 

 Jago was commissioned again (by De 

 Beers Marine [Pty] Ltd.) in 1997 to col- 

 lect a larger number of higher quality 

 videotapes from the same area. In this 

 article, we report on microscale habitat 

 association and habitat selection by 



