668 



Fishery Bulletin 88(4). 1990 



classes up to 5 m, and in 2-m classes beyond 5 m. Fish 

 were not recorded beyond 10 m from the transect. Both 

 divers had a good knowledge of the fish fauna and also 

 had a good training in visual censuses. The transects 

 were finished between 30 and 45 minutes before the 

 start of the trawling. 



For this type of data, many density estimators may 

 be chosen (Burnham et al. 1980). Among these esti- 

 mators, Kulbicki and Duflo (unpubl. data) show that the 

 most robust one for fish density (D) in similar condi- 

 tions is given by: 



where D = fish density (fish/ha), 

 p = number of species, 

 nj = number of fish of species i, 

 dj = average distance (m) of species i to the 



transect, and 

 L = length of the transect (200 m). 



Biomass B (kg/ha) was calculated in a similar fashion: 



B = (10/2L) 5! (Wi/dj) 



D = (10J/2L) 5! (n,/(l, 



1 = 1 



where W; = weight of species i (g). 



The weights of fish were estimated from length-weight 

 relationships (Wantiez and Kulbicki In press). 



