672 



Fishery Bulletin 88(4), 1990 



Table 5 



Comparison of density and hiomass estimates from transects 

 and trawls for main families and species, St. \'incent Bay, New 

 Caledonia. 



Ratio = density from transects /density from trawls. 

 r = correlation coefficient, 



* .significant at o<0.05, ** significant at o<0.01. 

 ' night station (n°6) not included. 



species are rather uniformaly distributed on the bot- 

 tom. Conversely, species such asLethrirmfi nematacan- 

 thus (r = 0.03) or Pnstotis jerdoni (r = 0.04) have a very 

 patchy distribution. A pratical consequence is that a 

 larger number of stations will be needed to get a good 

 stock assessment for the latter species than for the 

 former. 



Biomass estimates from transects are 9.1 times 

 larger than those from the trawls (Table 7), this dif- 

 ference being significant (F test for paired comparison 

 (Sokal and Rohlf 1981) at o = 0.05). Although density 

 estimates for all species were not correlated between 

 trawls and transects (r= -0.17), biomass estimates 

 show some correlation between the two methods, if the 

 night station n°6 is excluded (r = 0,72, significant at 

 a = 0.07). The absence of correlation between the two 



methods in the density estimates may be explained by 

 the large contribution of small fish such as Pomacen- 

 tridae to transect density estimates, whereas these 

 species escape the net. Conversely, biomass estimates 

 are almost unaffected by these species because of their 

 small weight. At the species level the correlations be- 

 tween the two methods for biomass estimates are of 

 the same magnitude as those observed for density 

 estimates. 



Discussion 



The comparison of two methods with such different 

 concepts requires some adjustments. Trawls swept 

 approximately 5600 m- and transects covered be- 

 tween 700 and 1000 m- (depending on species) for 

 each station. The trawls did fish over the transect lines, 

 but a perfect match would mean a trawl track with a 

 precision of ± 1 m which is untractable even in shallow 

 water. This is, in our opinion, the main source of differ- 

 ence in the number of species given by each method. 



