VERTICAL DISTRIBUTION AND MIGRATION RECONSIDERED 375 



rapid dodging movement can readily, it seems, spring clear of a smaller net with approximately half 

 the catching area, avoiding it evidently outright. 



Fig. 17, it will be recalled, shows our average horizontal gatherings for night and day in the sur- 

 face (o-io m.) layer and at 30 m. depth intervals below, the picture presented being based on our total 

 horizontal bathymetric data from the krill-rich regions of the circumpolar sea. It does not, however, 

 show the true relative density of the population at these levels because our observations, although 

 extremely abundant in the surface (0-5 m.) layer, do not cover even remotely so thoroughly every 

 5 m. interval below. As Figs. 61 and 62 so distinctly show, they are in fact so widely spaced in the 

 vertical plane that in many instances it seems we could have been missing the subsurface swarms 

 simply through not fishing at the precise levels where in their shallow rafts they might have chanced 

 to lie. If this be true a more accurate representation of the relative abundance of these animals at 

 and below the surface by night and by day is perhaps to be obtained by multiplying the average 

 gatherings from the surface by 10, the average gatherings between 10 and 50 m. by 40, the average 

 gatherings for every 50 m. interval below this by 50, and then, using a true depth scale, expressing 

 the results (Fig. 63) as area histograms. This treatment of the data, if it be legitimate, while again 

 distinctly suggesting a major gathering of the daytime swarms between 10 and 50 m., would seem to 

 suggest too that in so far as numbers go more than half the daytime subsurface population is concen- 

 trated at this relatively high level, the remainder below, more spread out, and in a declining abundance 

 that seems largely to peter out below 150 m. It will be seen, too, that the total daytime subsurface 

 population is approximately equal to the night population at the surface, strongly suggestive it would 

 seem that most of the massed night swarms desert the surface by day. 



As revealed here the night position is chiefly remarkable for the massive accumulation of the krill 

 at the surface then. It is remarkable too, however, for the almost equally large numbers that seem to 

 be gathered between 10 and 100 m., but particularly between 10 and 50, and I hesitate again to say 

 whether this is a real accumulation or whether it springs from contamination with the surface popula- 

 tion. It might be real or not, but I strongly suspect not (see pp. 160 and 270). 



On the whole I am inclined to the view that such inferences as may be drawn from this final pre- 

 sentation of the bathymetric data should be accepted with reserve. They involve the assumption 

 that the vast majority of the night swarms on the surface seek deeper levels by day. Yet we know that 

 some do not and that there is active avoidance of the day stern net by those that remain. It seems to 

 me that if the swarms do largely desert the surface by day and, as our observations suggest, gather 

 principally between 10 and 50 m., the massing of the patches there, and perhaps at slightly deeper 

 levels, ought to have presented a far surer target for our horizontal nets than all along it seems to have 

 done. As I have already said (p. 270) the feeding of the whales and the distribution of sighted patches 

 both distinctly suggest that daytime surface swarming is not an uncommon event. Thus although 

 Fig. 63 distinctly suggests a major vertical movement involving the sinking of virtually the entire 

 night-time surface population to deeper daytime levels, I do not take this as indisputable evidence 

 that all or virtually all the night swarms go down. The numbers of individuals on the surface at night 

 and below the surface by day it is true are approximately equal, but this could well have sprung from 

 the multiplication of the day subsurface averages by a somewhat arbitrary, and perhaps too high, 

 factor. The massive and distinctly unnatural night accumulation that seems to be gathered between 

 10 and 100 m., lohere the total population appears to be almost as large as it is on the surface itself, could 

 equally well have sprung from the same somewhat arbitrary treatment of the figures. 



Further information as to the habit, behaviour and vertical distribution of the krill in the sea comes 

 from the paired net hauls we made obliquely in the surface (loo-o m.) layer. These nets, a stramin 

 net (NiooB) and a silk net (N70B), were fished one close behind the other at the end of a warp 



31-2 



