274 DISCOVERY REPORTS 



there. Both avoidance and vertical migration it seems could be contributing, equally perhaps, to the 

 daytime surface scarcity we record. If however, avoidance is to be ruled out and the majority of the 

 swarms do in fact desert the surface by day, then it can again only be concluded that such as go down 

 become so widely scattered in the vertical plane that they present a very difficult target for our deep 

 horizontal nets. 



(3) The swarms that go down by day (represented, say, by our samples of 100 or more) do not 

 appear to go to any great depth, few in fact to below 100 m. The majority it seems might be going to 

 levels between 40 and 80 m., with more perhaps, if the average catch-figures are any indication, above 

 50 m. than below. The average day and night gatherings at South Georgia for all levels sampled by 

 the N70H and NiooH nets combined work out as tabulated below, 



Average catch 



Vertical horizon (tn.) Daylight Darkness 



0-5 40 1683 



5-50 184 297 



50-100 88 86 



100-200 39 8 



the daytime data distinctly suggesting a major massing of the subsurface swarms between 5 and 

 50 m., a phenomenon Fig. 17 it will be recalled also seems to stress with equal emphasis. 



(4) Although subsurface swarms (represented in this instance say by samples of 50 or over) seem 

 to be revealed by the South Georgia nets with much the same frequency by night as by day, I would 

 again call attention to the possibility that the night subsurface gatherings, especially the shallower 

 ones, being distinctly larger on the average than they are by day, may spring in part at least from 

 contamination with the densely populated surface zone. 



(5) There is no obvious rhythm in the pattern of the vertical movements that take place. If there 

 was we should expect perhaps to see a gradual descent to deeper levels after sunrise, perhaps a con- 

 spicuous deep massing during the height of the day, with a gradual return towards the surface towards 

 nightfall. None of this, however, appears to be happening, and it seems clear enough that subsurface 

 swarms are liable to be encountered at any level down to about 150 m. virtually at any hour of the day 

 or night. I would conclude, therefore, that the apparently erratic vertical movement of the krill to 

 which Hardy and Gunther (p. 163) originally called attention can hardly in fact be other than real. 



(6) The swarms that go down clearly do not all go to the same level, never it seems, neither by day 

 nor by night, occupying such a narrow horizontal stratum as they do when on the surface at night. 

 If they did we should sometimes it is certain, with so many deep horizontal net hauls, have hit this 

 critical layer and sampled it with enormous success, producing far larger subsurface gatherings than 

 the South Georgia nets have revealed. The haphazard levels at which our nets have it seems struck 

 the deeper swarms in these surveys can only it seems point to the same conclusion. 



(7) It will be seen that out of 51 daytime surface towings with the large loo-cm. diameter hori- 

 zontal net 39 are negative, eight produce very small or negligible catches, four sampling the older 

 swarms with moderate or substantial success. But the corresponding gatherings of the smaller jo-cm. 

 diameter net without exception are negative. This is a remarkable difference, providing very strong 

 additional evidence that in any assessment of the vertical distribution of these animals, and of their 

 vertical movements, active avoidance of the surface net by the older individuals in daylight is a factor 

 that must always be reckoned with. From these strikingly contrasting results, obtained in almost, but 

 not quite, the same water, and almost, but not quite, simultaneously, we can fairly in fact conclude 

 that throughout daylight in the surface zone such older and more active animals as may be there, 

 though not always able to avoid the stramin net with its relatively large mouth-opening, with their 



