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



meter for the water column is less than the cor- 

 responding night catch. In every case, sampling 

 extended below the maximum depth of occur- 

 rence of the population and there is no evidence 

 that any individuals of the population migrated 

 vertically out of the depth zone sampled during 

 the day. Therefore, it is highly significant that all 

 of the day values were less than the respective 

 night ones (P<0.005). This result gains impor- 

 tance if we also consider 10 other day/night pairs 

 of 1 m 2 MOCNESS tows in which N. megalops 

 was collected on previous cruises (Chain 125, 

 Knorr 53, Knorr 62). For nine of these pairs, 

 moderately to dramatically higher catches in the 

 night tow were obtained (Table 2). The single ex- 

 ception to this pattern was a pair of Slope Water 

 tows taken near the continental shelf in the wake 

 region of a warm core ring (tows 41, 42). But 

 these two tows were displaced in space by several 

 miles, and the night tow was taken nearer the 

 warm core ring where a lower catch might have 

 been expected. 



Of the 18 day/night pairs of 1 m 2 MOCNESS 

 tows, 17 yield higher density estimates at night. 

 Patchiness in the distribution of N. megalops 

 contributed to variability to these estimates but 

 as an unbiased variance component, it does not 

 affect our expectation that one-half of the day 

 and one-half of the night tows in day/night pairs 

 should be the larger. Thus it is unlikely that 

 patchiness of this species is responsible for the 

 significantly higher night catches that we have 

 observed (P<0.001). We know of no other expla- 

 nation than avoidance to explain this result. 



There are only five pairs of 10 m 2 MOCNESS 

 observations of the vertical distribution of N. 

 megalops. For two of these, the integrated day 

 catch is larger than the corresponding night 

 catch and, therefore, night catches are not sig- 

 nificantly larger than day catches (P>0.05). This 

 result either means that there is no day/night dif- 

 ferential avoidance of the 10 m 2 net or that in the 

 face of other sources of error such as patchiness, 

 we have too few day/night pairs of observations 

 to observe the avoidance effect. If avoidance 

 were affecting only the smaller net then at least 

 we would expect that the 1 m 2 net day catches per 

 unit volume would be consistently smaller than 

 the corresponding 10 m 2 net day catches. We 

 might also expect that night catches with the 1 

 m 2 net would be smaller than the 10 m 2 net. 

 Neither comparison yields a significant result 

 (P>0.05; day MOCNESS 1 tows greater than 



day MOCNESS 10 tows in four out of seven com- 

 parisons; night MOCNESS 1 tows greater than 

 night MOCNESS 10 tows in three out of seven 

 comparisons). Thus within the limits of error, by 

 day or by night both net systems provide com- 

 parable estimates of the number of N. megalops 

 living in the water column at a given station. 



It is possible that the lack of differences in the 

 catching rates between the two nets is due to the 

 different mesh sizes. Small individuals might 

 have been caught more efficiently by the 1 m 2 net 

 while larger individuals could have avoided this 

 net better and conversely for the 10 m 2 net except 

 that small individuals would have been lost due 

 to escapement through the mesh. The size-fre- 

 quency data in Figures 4 and 5 do not support 

 this possibility. While there is considerable vari- 

 ability between net tow pairs, in terms of abso- 

 lute abundance, neither net system systemati- 

 cally catches large or small individual N. mega- 

 lops in the size range counted better than the 

 other. A similar observation can be made if com- 

 parisons are made on the relative abundances in 

 a given size class (Fig. 6). 



There is one other potentially significant trend 

 in the data that is important to note. The magni- 

 tude of the day/night avoidance does not appear 

 to be uniform with depth. For the 1 m 2 MOC- 

 NESS, largest differences between paired night 

 and day catches where both are positive occur 

 when the center of distribution of N. megalops is 

 above 300-400 m and minimum differences 

 occur at or below these depths (Table 2). Linear 

 regression of the ratio of night to day catch (N/D) 

 versus depth of the center of the distribution at 

 night (50% of occurrence with depth) is signifi- 

 cant at P = 0.1. There is a similar pattern in the 

 10 m 2 MOCNESS tows, although as mentioned 

 above, the day/night differences in catching 

 rates are considerably smaller. 



In summary, there is clear evidence for differ- 

 ential day/night avoidance of the 1 m 2 MOC- 

 NESS. Furthermore, there are no significant 

 differences in the size range of adolescent or 

 adult N. megalops caught by the 1 m 2 or 10 m 2 

 MOCNESS systems nor in either system's esti- 

 mates of its abundance in the water column at a 

 given station when day or night pairs are com- 

 pared. Although differences between pairs of 

 day/night catches for the 10 m 2 MOCNESS are 

 statistically not significant, the entire data set 

 when considered as a whole strongly suggests 

 that N. megalops is also avoiding the 10 m 2 net, 

 albeit to a lesser extent. 



81 



