104 DISCOVERY REPORTS 



considered together the peculiar nature of the larger of the two groups is evident. The 

 average growth curve on p. io8 below shows that the average length of the larger group 

 is in each instance much greater than that represented in the graph. Thus in September 

 the normal average is about 12 mm. as opposed to 18 mm. in John's larger group. In 

 October the normal average is about 13 mm. as opposed to 27 and 25 mm. in the 

 "small adults" from Sts. WS 279 and WS 288. Ruud's mean length diagram (1932, 

 p. 41, fig. 8) shows that in October the mean length of adult krill is about 50 mm., so 

 that our "small adults" fall somewhere between the normal adolescent average length 

 and fully adult average length. 



The origin of these anomalous forms is not clear, and what the factors are that in- 

 fluence their difference in size is open to speculation. It may be that they are derived 

 from batches of eggs laid very early in the summer season that have pushed ahead with 

 development, obtaining the full benefit of the summer season, whereas the majority of 

 the larvae encountered only obtain benefit from the latter part of the summer, having 

 been derived from eggs laid well on in that season. It may be, as John suggests, that 

 specially favourable environmental conditions, for instance a rich diatom flora, may have 

 been the lot of some of the larvae for a greater period than the remainder. Whatever be 

 the true explanation, the fact that forms occur having this peculiarity of length leads to 

 enquiry as to when they become sexually mature. If the growth rate in these eu- 

 phausians is similar to that in normal circumstances there is every possibility that they 

 reach a size at which they are sexually mature before the coming of the second winter. 

 If this does occur the complications to be encountered in unravelling the constitution 

 of the adult krill population are greatly increased. 



The occurrence of young E. superba differing widely in length frequency in different 

 localities at approximately similar times was described by John in a comparison of 

 larvae from the Bellingshausen Sea, from the Weddell Sea ice-edge and from stations 

 in open water west of South Georgia. The period covered was little more than a month 

 and the contrasted resuhs are referred to on p. 141 and shown in Fig. 61. He suggests 

 that better conditions for growth, such as the more abundant diatom flora, may account 

 for the larger size of some of the larvae as compared with the others. 



Apart from one instance, mentioned below, no attempt has been made in this 

 paper to correlate the occurrence of diatoms with the krill results. It is regretted that it 

 has not been possible to fulfil the promise implied in Hart's Phytoplankton Report 

 (1934, p. 11). The detailed investigation of this very important field of research will, it 

 is felt, clear up many of the problems connected with the growth and distribution of 

 krill. 



The correlation is inserted as it has a bearing on John's suggestion concerning varia- 

 tion in richness of the food supply. In Fig. 28 is shown the average length of adolescent 

 E. superba and the abundance of Thalia ssiosir a antarctica at the stations indicated. The 

 presence, but not the abundance, of Clwetoceros socialis is also represented in the figure ; 

 this small colonial form occurred in such great numbers that its abundance could not be 

 estimated satisfactorily. The stations from which the data were taken were all made in 



