270 DISCOVERY REPORTS 



subsurface gatherings from between 5 and 100 m. (p. 160, Table 31) and of such other gatherings 

 as we have from deeper levels. These do suggest very strongly that the swarms by day rarely if ever 

 occupy the same deep horizontal stratum, for if they did, some, perhaps many, of the 353 subsurface 

 hauls we made between 5 and 250 m. would surely have hit it and sampled it with enormous success. 



On the surface at night (p. 159, Table 30) we strike the krill with conspicuous regularity, 74% of 

 hauls being positive, 54% yielding catches of 10-100 or more. Below the surface we see the position 

 in reverse, 73% of hauls being negative and only 10% with catches of 10-100 or more. The sub- 

 surface position it seems to me, both by day and by night (Table 31) can only be explained on two 

 alternative assumptions, either the subsurface swarms are scarce, or, if not exactly scarce, are so 

 isolated in space that both horizontal and oblique nets are offered only a feeble chance of striking 

 them. Both assumptions I believe in the long run may prove to be correct. 



If densely packed euphausians regularly produce the ' noisy ' echoes that Gushing and Richardson 

 (1956) seem to have detected in the North Sea, the listening technique recently advocated by Barnes 

 (1959) could, it seems, possibly help us to determine accurately where the subsurface swarms are by 

 day and whether they are abundant or otherwise. The undersea observation chamber described by 

 Inoue, Nishizawa and Fukuda (1957), a miniature underwater laboratory that can be operated down 

 to 200 m., is another instrument that might help to correct our imperfect understanding of the distribu- 

 tion and movements of the krill in the sea. It might, for instance, prove particularly useful for 

 observing at close quarters the extent, draught and behaviour of the subsurface swarms, and perhaps 

 provide accurate information as to how abundant they really are and as to how far down they extend, 



I would call attention, too, to the occasional large, or moderately large, gatherings, especially the 

 shallower ones, we make below the surface at night (Table 31). Are they in fact from swarms that 

 have migrated away from the surface by day but have delayed their return, or from swarms that went 

 down at night, or are they in fact gatherings actually made at the surface when shooting the large open 

 net, with its inevitable halt (p. 160), through the densely crowded surface zone? The relatively large 

 average night gatherings we record between 10 and 70 m. (p. 161, Fig. 17), somewhat larger it 

 will be seen than our corresponding gatherings from this horizon by day, lead me strongly to suspect 

 that much of this apparently substantial night subsurface concentration springs from contamination 

 with the crowded surface population. 



With reference to the occasional, but conspicuous success with which the stern surface net some- 

 times samples the daylight swarms I would recall that the only occasion when this happened, where 

 large krill (and really large gatherings) were involved, was at Station 150, where (p. 265) stern nets, 

 fishing far back in the wake, were it seems sampling closed-up swarms, which having shortly before 

 been split wide open by the passing vessel, could not in any case have been so successfully sampled by 

 the ordinary stern net which normally fishes much closer to the hull. Avoidance of the net, I would 

 emphasise again, both by day and by night, springs perhaps as much from the thrust of the hull upon 

 the swarms as from the actions of the krill themselves. These special nets, for instance, could equally 

 well, I think, have been sampling swarms that (see below) had sounded as the vessel came upon them, 

 but had returned to the surface far back in the wake, to the distinct advantage of the net that was 

 fishing there and the manifest disadvantage of the same net had it been fishing nearer the hull. 



I am not so sure, however, that daytime swarming at the surface is a relatively rare event. The pro- 

 fusion of swarms seen from 'WiUiam Scoresby ' in the Weddell Sea (p. 148), and by Gunther (p. 150) 

 and others when whale-marking, shows that at certain times and over certain wide areas it is exceed- 

 ingly common, while the distribution of total patchiness recorded by Discovery observers (p. 411, 

 Fig. 144) reveals it to be a phenomenon of widespread occurrence, to be seen at one time or another 

 almost anywhere in the East Wind-Weddell surface stream where the krill are manifestly so 



