VERTICAL DISTRIBUTION AND MIGRATION RECONSIDERED 277 



not sampling the krill by day with anything Uke the same success as it is at night, and this is probably 

 because (i) the daytime subsurface swarms present a difficult target, even for the NiooB, and 

 (2) because the krill (p. 275) are evidently well able to jump clear of a net of such small diameter. 

 Moreover, since avoidance is manifested principally as a daytime surface phenomenon, the very low 

 average catch-figure for the day N 70 B would seem to provide some further evidence that many of the 

 daytime swarms are gathered in not very badly illuminated conditions not very far below the surface. 



With nets set so close together on the same warp we might well think that wherever there should 

 be krill in their way both nets of a pair would sample it, each according to its catching capacity, 

 in other words that both would produce at least positive results. Yet there are many instances 

 where the NiooB gives positive, and its partner negative results, and vice versa. Thus at night 

 there are 35% cases where NiooB struck the krill and its pair (N70B) missed it. By day there 

 are 26% such cases. Again, at night there are 7% cases where N70B struck the krill and its pair 

 (NiooB) missed it, and there are 13% such cases by day. Combining night and day observations 

 and taking all instances where either net of a pair was positive and its partner negative we find that in 

 41 % of all cases where paired nets were used we struck the krill with one of them and missed it with 

 the other. In view of how very closely these nets were operating together this suggests very strongly 

 that the krill cannot be distributed broadcast in the sea, that they must in fact be in swarms, many of 

 them often of very small dimensions. Nor can they, it seems, be disposed in depth, in layers say 

 20, 30 or 40 m. thick, for if they were it would be virtually impossible for one net of a pair to sample 

 them and the other not. They must it would appear, both at the surface and below, be gathered 

 together in the shallow rafts that have so often been seen from the decks of vessels. 



There are a number of instances among the 148 paired samplings where, although the horizontal 

 net showed that there was in fact krill at the surface, neither net of an oblique pair took any krill at 

 all. Of the 80 night pairs 13 % were negative, and of the 68 day pairs as much as 32% were negative, 

 the higher day negative percentage being no doubt due partly to the difficult target presented by the 

 daytime subsurface swarms and partly to avoidance both at the surface and perhaps not far below. 



In 35 instances where both nets of a pair gave positive results at night the NiooB produced the 

 larger catch in 29, or 83 %, the N70B producing the larger catch (or the same catch as the N lOoB) in 

 six instances only. There are 19 cases where both nets gave positive results by day and in 89% of 

 them the N looB again produced the larger catch. It is clear from the length frequencies in our log- 

 books that these were instances where both nets of a pair were passing through the same swarm and 

 it is, therefore, only to be expected that the larger net in the vast majority of cases would sample it 

 with the greater success. 



It is of some interest to compare these results with those from paired oblique nets worked below 



100 m. There are no such pairs, 97 for night and 13 for day. The average gatherings for night and 



day are shown below. 



Paired oblique nets worked below 1 00 metres 



Average catch night Average catch day 



NiooB N70B NiooB N70B 



10 13 



The day results, it will be seen, are scarcely worth considering since the observations are so few. The 

 night figures, however, point to a distinct scarcity of deep subsurface swarms, 65 % of the hauls being 

 negative for both nets of a pair. In fact if we consider all pairs that took between them either no 

 animals or less than 10 the percentage of negative or negligible results for night and day combined is 

 as high as 94. There are only three instances involving the deep night gatherings where catches of 

 between 100 and 1000 were obtained, these combined gatherings accounting for 91 % of the total of 



