328 DISCOVERY REPORTS 



avoidance during the daj^ than during the night (Holt and Tattersall, 1905; Frost, 1932; Gardiner, 

 1934). Southern and Gardiner (1932) found that when they arranged a number of species in the 

 order of the difference between day and night catches they were also in order of size. They concluded 

 that the larger animals were more successful in avoiding the net and that this was a factor of some 

 importance when considering the variation between day and night hauls. Hardy and Gunther (1935), 

 when considering the vertical migration of the larvae oi E.frigida Hansen, 191 1 , and Thysanoessa spp., 

 found that there was no significant difference between the relative populations in the upper 1000 m. 

 as shown by the nets by day and by night. The same seems to apply to adolescent and adult 

 E. triacantha. In Table 6 the same data as were used in preparing Fig. 9 C are given, except the mean 

 catches for each depth range are based upon the actual catches and not those corrected for a 250 haul. 

 The totals of these means can be taken as a measure of the total population, as shown by the N70V, 

 in each 4-hourly period. Although these totals show a variation they do not indicate any significant 

 increase in the amount of avoidance during the day. This does not mean that the animals do not 

 avoid the N 70 V, it just shows that there is no appreciable diurnal variation in the amount of avoidance, 

 and consequently reasonable comparisons can be made of day and night hauls. It should be pointed 



Table 6. The mean catch with the Njo V in each depth range, uncorrected for a 250 m. haul and divided 

 into ^-hourly periods. The totals represent the relative total population for each period. {These figures, 

 corrected for a 250 m. haul have been used in preparing Fig. gC. The numbers of hauls in each interval 

 are given in Table 5) 



Time (hr.) 



out that here we may be dealing with a special case and that this does not apply to all species. It has 

 been shown above that the larger E. triacantha, i.e. those that would be more successful in avoiding 

 the net, tend to inhabit deeper water where the increase in illumination during the day is not very 

 great, and that in the upper waters, where the diurnal variation in illumination might be expected to 

 have an effect, the animals are smaller and consequently less successful in avoiding the net. Mr Marr 

 (personal communication) has found that in the case of adult E. superba, which inhabits the surface 

 water and displays no diurnal migration, illumination has a very marked effect on the amount of 

 avoidance, so much so that daylight hauls with the N70V very rarely take any adult E. superba at all. 

 Both Hardy and Gunther (1935) and Mackintosh (1937) have shown that the horizontal distribution 

 of planktonic animals may be influenced by variations in their vertical distribution. Mackintosh 

 suggested that populations, at least of certain common species in the Antarctic, could maintain their 

 particular horizontal ranges by means of seasonal or diurnal vertical movements between water 

 masses movmg in different directions. To the south of the Antarctic convergence the directions of 

 movement of the water masses are fairly well known (for details see Deacon, 1937). North of about 

 65° S the Antarctic surface water, a layer of cold water 100-250 m. in thickness, is moving to the east 

 and north and below this lies the warm deep water which generally has a strong southerly component. 

 We have seen above that the diurnal vertical movements of E. triacantha are sufficient for the animals 

 to alternate between the two currents and it follows that, provided the relative speeds of the currents 



