VERTICAL DISTRIBUTION AND MIGRATION RECONSIDERED 269 



narrow stratum all along its horizontal path. The fact that we do, however, strike the swarms when 

 towing obliquely by day, sampling them often with conspicuous success, he thinks is strong evidence 

 that some of them must in fact be making daytime migrations to deeper levels. In other words, in 

 so far as the light factor is concerned, the subsurface swarms he suggests, in the darkness or gloom of 

 their daytime environment, are being sampled with much the same advantage, although not with 

 the same certainty, as are the swarms on the surface at night. He thinks, however, that such as do 

 migrate vertically by day may only go a short way down, perhaps no farther than 40 or 50 m., and 

 indeed the average daytime subsurface catch-figures for such moderate to substantial gatherings as 

 we have made between 5 and 100 m. (p. 160, Table 31) could well I agree be taken as evidence of 

 such a migration. It might in fact be that it is somewhere between 10 and 40 m. that most of the 

 migrating swarms, in their vertical scatter, are gathered during the daylight hours, the moderate, yet 

 distinct, peak in the average daytime subsurface concentration we have recorded between these levels 

 (p. 161, Fig. 17) suggesting that this is what is actually taking place. 



He believes too, as I do, that the smaller gatherings of the night oblique net vis a vis those of the 

 night net on the surface are simply due to the fact that whereas the oblique net passes for only a few 

 minutes through the now richly populated surface zone the horizontal net does so for half-an-hour. 

 It is equally clear to us both (see again Table 31) that such daytime migration as he envisages does 

 not involve the older swarms in any mass descent into the warm deep layer. 



The ratio of the average night catch of the oblique nets to the average night catch of the surface 

 nets, where both nets were used simultaneously, is i :4 (p. 282, Table 60), the corresponding ratio 

 for the same nets used in the same way in daylight being 7:1. He regards this as strong confirmatory 

 evidence that a vertical migration must in fact be taking place, believing that if the swarms were 

 confined to the surface by day they would avoid the oblique net there just as readily as they would, 

 and are known in fact to do, the day stern net on the surface. In other words the day oblique net 

 must be sampling the swarms, not in the well lit surface zone, but in the gloom or darkness of 

 deeper levels. He agrees of course that surface swarming is a daytime as well as a night phenomenon, 

 but, as he has only seen the daylight swarming very occasionally, during two seasons south, he believes 

 it to be a relatively rare event, suggesting that such behaviour may be similar to that of the copepod 

 Calanus finmarchicus which, while normally having a distinct vertical migration, sometimes (Gough, 

 1905) occurs in immense numbers on the very surface in brilliant sunlight. He points too to the 

 evidence (p. 160, Table 31) that, like Calanus, the krill do not in fact all come to the surface at night, 

 the occasional quite large subsurface gatherings obtained between 5 and 100 m. in darkness suggesting 

 that some of the swarms remain down. He calls attention finally to the fact that stern nets on the 

 surface in daytime do occasionally strike the larger congregating krill and sample them with great 

 success, the obvious inference being that in such instances at least they had not been very successfully 

 escaping. 



In an overall comment on these alternative yet not altogether conflicting views I would point out that 

 if Professor Hardy's vertical migration is in fact taking place, as it could well it seems be, I believe 

 like Mackintosh (p. 162) that it does not by any means involve the whole of the adult or half-grown 

 surface population, and would recall too what Hardy himself has said (p. 163) that the behaviour of 

 the krill that take part seems to be very erratic. I would say too, as Fraser (1936) and Hardy (1955) 

 have already found, that neither the late larval nor early adolescent swarms seem to be involved in 

 this movement to anything like the same extent, if indeed they are involved in it at all. 



There can be little doubt, I believe, that such swarms as probably do go down by day do not all 

 go to the same level, but become scattered about at random perhaps over a considerable vertical 

 horizon. Good evidence of this is provided by the relatively small size of the great majority of our 



