Anderson et al : Current velocity and catch efficiency in sampling settlement-stage larvae of coral-reef fishies 



411 



Indeed, there may be important consequences in sam- 

 pling reef fish larvae among multiple sites that vary in 

 current velocity. Current-dependent catch efficiency could 

 confound estimates of spatial variation in larval abun- 

 dance if those sites vary substantially in current velocity. 

 For example, deploying light traps to sample larval abun- 

 dance at several locations may result in higher abundances 

 at sites with low current. This result could indicate actual 

 larval distributions, such as accumulation of larvae on the 

 leeward side of an island, or it could simply reflect current- 

 dependent catch efficiency. In the latter case, estimates of 

 lan'al distribution and subsequent interpretations of such 

 distributions (e.g. larval retention) would be flawed. 



As with all sampling devices, Hght traps and channel nets 

 have both advantages and disadvantages for estimating 

 lar\'al abundance. Light traps may collect a larger propor- 

 tion of settlement-stage larvae for those fishes that exhibit 

 a photopositive response, but they catch fewer taxa and 



sample an unknown volume of water (Cheat et al., 1993) in 

 relation to channel nets. Our results indicate an important 

 factor in selecting an appropriate device for estimating 

 larval supply — namely, the hydrodynamic conditions under 

 which the device will be deployed. Use of one sampling de- 

 vice, either light traps or channel nets, would have resulted 

 in biased and potentially inaccurate relative estimates of 

 larval abundance between So. Bock and S.E. LSI. In more 

 recent studies, several methods have been used to estimate 

 the abundance of as many larval fishes as possible (e.g. Leis 

 et al., 1998), and in a comparison of light traps and towed 

 nets in sampling freshwater fishes, Gregory and Powles 

 ( 1988) concluded that both sampling devices should be used 

 to avoid bias in the collection of larvae. 



The use of both light traps and channel nets simultane- 

 ously can provide less biased estimates of spatial variation 

 in taxonomic composition and larval supply of coral-reef 

 fishes among sites that vary in current velocity. This com- 



