any particular zooplankton group. Further, the 

 mechanisms by which an oncoming net is de- 

 tected and avoided are not well known (Clutter 

 and Anraku 1968). 



Theoretical studies of avoidance (Barkley 

 1964, 1972; Clutter and Anraku 1968; Murphy 

 and Clutter 1972; Laval 1974) have drawn atten- 

 tion to important behavioral aspects of avoidance 

 and how these behavioral features are likely to 

 interact with net size and towing speed. Avoid- 

 ance theory provides a framework for the design 

 of avoidance field studies and the interpretation 

 of their results, as Barkley's (1972) examples 

 clearly demonstrated. We have applied avoid- 

 ance models to data on the euphausiid Nema- 

 toseelis megalops to determine its response to dif- 

 ferent net types under different conditions. Since 

 our data on this relatively abundant species 

 indicated substantial avoidance effects, we have 

 examined it in some detail. 



In studies of N. megalops vertical distributions 

 (Wiebe and Boyd 1978; Boyd et al. 1978) night 

 tows were consistently observed to produce 

 higher numerical density estimates than tows at 

 the same station during the day. These tows were 

 taken with a multiple net system with aim 2 

 mouth opening (MOCNESS, Wiebe et al. 1976). 

 Although horizontal patchiness may have con- 

 tributed in an unbiased way to the day/night dif- 

 ferences in catch rate, the overall comparison of 

 day and night tows strongly suggested greater 

 avoidance during the day. Unfortunately, there 

 were too few day/night pairs at a single station to 

 demonstrate this by pairwise comparison. 



During 1976 and 1977, as part of a multidisci- 

 plinary study of Gulf Stream cold core rings ( Lai 

 and Richardson 1977; Richardson 1980), more 

 complete observations of day/night vertical dis- 

 tributions were made at stations in Slope Water 

 and cold core rings. Tows were taken with aim 2 

 MOCNESS and with a 10 m 2 MOCNESS. As 

 will be demonstrated below, both net systems 

 catch N. megalops without apparent size dis- 

 crimination between the two. Both net systems 

 are avoided to a certain extent, based on day/ 

 night catch ratios, but some revealing differ- 

 ences are evident. Using data from both sizes of 

 nets, it is possible to apply Barkley's ( 1972) avoid- 

 ance model to obtain independent estimates of 

 the parameters of reaction distance and percent 

 capture within a certain animal size range, and 

 through a comparative analysis reach some ten- 

 tative conclusions regarding avoidance mech- 

 anisms in the species studied. 



FISHERY BULLETIN: VOL 80, NO. 1 



METHODS 



The combinations of 1 m 2 and 10 m 2 MOC- 

 NESS tows were taken at nine stations (Fig. 1), 

 with additional information about each tow in 

 Table 1. Stations 1-4 were sampled in April 1977 

 on RV Knorr cruise 65 and stations 5-9 were sam- 

 pled on Knorr 71. Station 1 was in cold core ring 

 "Bob," a 2-mo old ring; station 2 in ring "Al," a 

 6-mo old ring; station 6 in "Emerson," a rela- 

 tively old ring; stations 7 and 8 in "Franklin," a 

 middle-aged ring. The other stations were lo- 

 cated in the Slope Water. For some comparisons 

 data collected with the 1 m 2 MOCNESS on 

 earlier cruises will be used; information about 

 these tows (those noted in Table 2) is reported by 

 Ortner et al. (1978). Station positions for the re- 

 maining tows are given in Table 1. 



At the majority of stations day and night tows 

 were taken with both the 1 m 2 and 10 m 2 MOC- 

 NESSes. The 1 m 2 net was equipped with nine 

 nets of 333 ^ m nylon mesh netting dyed dark 

 blue and was fished obliquely so that eight strata 

 were sampled. Generally, the strata sampled 

 were 1,000-850, 850-700, 700-550, 550-400, 400- 

 300, 300-200, 200-100, 100-0 m; usually between 

 600 and 1,000 m 3 were filtered in each strata. 

 Occasionally in the Slope Water when sharp 

 hydrographic gradients were present, the sam- 

 pling intervals were modified to bracket the 

 physical discontinuities. 



The 10 m 2 MOCNESS is a scaled up version of 

 the 1 m 2 net system described by Wiebe et al. 

 (1976). On the tows reported here, it was 

 equipped with five nets of 3 mm nylon mesh 

 white netting. This net system was fished 

 obliquely with four of the five nets sampling 250 

 m intervals from 1,000 to the surface. Each net 

 filtered between 13,000 and 43,000 m 3 . As with 

 the 1 m 2 MOCNESS, the first net is fished while 

 lowering the system to the maximum depth of 

 the tow to provide uniform drag and to prevent 

 kiting of the system at the start of the stratified 

 oblique haul. Neither system has a bridle in front 

 of the nets. 



The nets of both systems are opened and closed 

 with commands sent via conducting cable from a 

 surface deck unit. For both systems on Knorr 65, 

 we used an underwater unit which measures and 

 telemeters to the surface depth temperature, 

 conductivity, angle of the net system from the 

 vertical, flow past the net, and information re- 

 garding the electrical and mechanical function 

 of the opening/closing mechanism. The trans- 



76 



