WIEBE ET AL.: AVOIDANCE OF TOWED NETS BY NEMATOSCELIS MEGALOPS 



these different signals obviously depends upon 

 the net structure and towing cable configuration 

 and upon the ability of N. megalops to sense the 

 various signals. Although there is no direct ex- 

 perimental information about N. megalops' sen- 

 sory capabilities or about the signals being gen- 

 erated by MOCNESS, it seems clear that the 

 primary avoidance stimulus involves day to 

 night variations in light. Nemaioscelis megalops 

 must use vision to detect the net and can better 

 avoid the net during the day than at night be- 

 cause during the day the net is better illumi- 

 nated. A fundamental link between the amount 

 of light present and the magnitude of the avoid- 

 ance is provided by our observation that as indi- 

 viduals live deeper in the water column under 

 substantially reduced daytime light levels, day/ 

 night differences in catch rates decline. 



But if we accept the results gained by the 

 application of Barkley's model which indicate 

 substantial avoidance takes place at night in the 

 absence of bright sunlight, then other factors 

 must also be important. We propose that bio- 

 luminescence is the principal signal and that 

 vision remains the principal means of detection. 



Three linesof evidence support the importance 

 of bioluminescence as an avoidance cue. First, in 

 an experiment conducted in the early 1960's, 

 Boden (1969) equipped an IKMT with light 

 meters so that he could monitor the amount of 

 light produced above, below, in front of, and in- 

 side the trawl as it was towed at night. Biolumi- 

 nescent light above the trawl was less than below 

 the trawl but both were considerably lower than 

 that ahead of or in the net. Light within the net 

 was so bright that it recorded off scale and indi- 

 vidual flashes were often too numerous to be re- 

 corded as such. Light ahead of the net was also 

 exceedingly bright. Boden (1969) speculated 

 that the light ahead of the net was caused by or- 

 ganisms flashing either in response to the light 

 within the net or to pressure or sound waves 

 propagating forward from the net. Second, 

 Neshyba's (1967) experiments with a submarine 

 photometer and strobe light showed that meso- 

 pelagic and epipelagic organisms could be 

 stimulated to produce significant amounts of bio- 

 luminescence (10 4 juW/cm 2 ) for a sustained pe- 

 riod by proper strobe light flashing. In the ab- 

 sence of artificial flashing, he observed a much 

 lower level of irregular flashing (10 8 -10 7 /zW/ 

 cm 2 ) similar to that reported by Kampa and 

 Boden (1956), Clarke and Backus (1964), and 

 Boden et al. (1965). Third, it is known that the 



eyes of euphausiids and decapod shrimps living 

 at midwater depths during the day (i.e., 200-600 

 m)are sensitive to light levels (10 7 to possibly 10 9 

 yuW/cm 2 , Clarke 1970) significantly lower than 

 that produced as a result of bioluminescence. 



These lines of evidence suggest that the light 

 generated by organisms when they come in di- 

 rect contact with the nets or encounter turbu- 

 lence caused by the net is used by individuals 

 ahead of the net to detect its presence and begin 

 an avoidance response. It seems likely that the 

 light ahead of the net observed by Boden (1969) 

 was caused by the same kind of response mech- 

 anism described by Neshyba(1967), i.e., flashing 

 in response to flashing. 



The tactic of reducing the visual contrast be- 

 tween a net and the surrounding water was 

 demonstrated by LeBrasseur and McAllister 

 (unpublished data cited by Clutter and Anraku 

 1968) to reduce the avoidance error for euphau- 

 siids both day and night. However, if biolumi- 

 nescence in and ahead of the net is an important 

 cue as we suspect it to be, then a more active 

 means of camouflaging the net is required. 



It is known from recent evidence (Warner et al. 

 1979) that decapod Crustacea living at the same 

 depth as N. megalops are easily "blinded" by even 

 moderate amounts of light. This suggests the 

 possibility of equipping the mouth of a net with a 

 "blinding" light system to be used to periodically 

 illuminate a region ahead of the net with enough 

 light to temporarily blind individuals in the net. 

 With the light out, individuals so affected by the 

 light pulse would be unable to see and, therefore, 

 to respond to the much lower light generated by 

 zooplankton being captured by the net. We pos- 

 tulate that individuals outside the zone of tem- 

 porary blindness may respond by electing a 

 startle response, but, because the volume illumi- 

 nated would be so large, their movement would 

 be random with respect to the volume to be fil- 

 tered by the net. Clearly, considerably more re- 

 search is required before this strategy could be 

 considered feasible. 



There are two precautionary notes that must 

 be made. First, in spite of avoidance error, verti- 

 cal distribution patterns obtained in sampling 

 this species with MOCNESS at different times 

 under different hydrographic regimes are repli- 

 cable (Fig. 3). That is, although avoidance error 

 is strongly affecting the numerical estimates, the 

 shape of the vertical distributions seem much 

 less affected. Thus, in spite of the avoidance, we 

 believe we are obtaining valuable ecological in- 



89 



