FISHERY BULLETIN: VOL. 80, NO. 1 



Table 5.— Ratios of night to day catches (number per square meter) of Nematoscelis megalops as a func- 

 tion of size for stations where both the MOCNESS 1 and the MOCNESS 10 were taken. <» indicates 

 only the night tow caught individuals in the given size class; indicates the opposite patterns. 



DISCUSSION 



From this application of the Barkley avoid- 

 ance theory, it appears that estimates of N. mega- 

 lops water column abundance could be substan- 

 tially underestimated by both nets, even at night. 

 Minimum probabilities of capture derived from 

 best fits to model expectations are 0.1 or less for 

 night catches and 0.01 or less for day catches. 

 However, the fact that we cannot demonstrate a 

 dependence of the ratio of night to day catches on 

 the size of individuals caught strongly suggests 

 the size dependent swimming speed assumption 

 required to apply the model is not valid for this 

 species, a result which is apparently supported 

 by Kils's (1979) data for Euphausia superba 

 escape swimming (tail swimming). Being unable 

 to make this assumption means that the field 

 population size-frequency distribution which 

 was observed is probably not materially affected 

 by avoidance. Undeniably some fraction of the N. 

 megalops population is avoiding the net systems, 

 and the problem is serious enough to merit an 

 effort to reduce this bias, i.e., to prevent the 

 avoidance from taking place. 



The usual strategies suggested to reduce net 

 avoidance, increasing net speed or net size, have 

 serious shortcomings in this case. Our evidence 

 strongly implies that N. megalops' response to 

 increased net size is to increase its reaction dis- 

 tance so that the catch rate remains relatively 

 constant. Barkley (1972) reached the same con- 

 clusion in a comparison of 1 m diameter net and 3 

 m IKMT (Isaacs-Kidd midwater trawl) catching 

 rates of the northern anchovy, Engraulis mor- 

 dax. It is possible that by going to still larger nets 

 (i.e., >10 m 2 mouth areas), a reduction in the bias 

 could be effected. However, larger nets would be 



impractical, if not impossible, to handle on most 

 oceanographic vessels. 



As Barkley (1964) has demonstrated, in- 

 creased net speed is not a feasible strategy for 

 avoidance reduction since increasing the towing 

 speed of a net requires a compensatory reduction 

 in net size. The practical limits to increasing the 

 tow speed are reached at 2 to 3 kn, because of un- 

 avoidably extreme wire angles and inordinate 

 amounts of wire required to fish even at moder- 

 ate depths (to 1,000 m). High speed tows gener- 

 ally result in damaged specimens, reducing their 

 value in studies requiring taxonomic identifica- 

 tion or in physiological and biochemical mea- 

 surements. Finally, as speed of net is increased, 

 the effects of escapement through the meshes is 

 enhanced. 



Another means of reducing avoidance, that of 

 camouflaging the net to reduce an animal's abil- 

 ity to detect its presence and thereby reducing 

 the avoidance reaction distance, has been dis- 

 cussed briefly by Clutter and Anraku (1968). 

 There is evidence that it may be an effective 

 strategy for species such as N. rae.ga/o/xs'(LeBras- 

 seur and McAllister, unpublished data cited by 

 Clutter and Anraku 1968). To use this approach, 

 one must first know what kind of a signal the ani- 

 mal is using to detect the oncoming net. Camou- 

 flaging the net can be accomplished by reducing 

 the signal until it becomes part of the back- 

 ground (omnidirectional noise). Alternatively, 

 the noise level could be increased until the signal 

 is no longer detectable. 



Signals emanating from a net and towing 

 cable include deformation of flow, near field 

 (displacement dominant) or far field (pressure 

 dominant) sound, and light (bioluminescence) 

 (Clutter and Anraku 1968). The importance of 



88 



