io8 



DISCOVERY REPORTS 



lengths are derived, in two of the three half-months, from larvae taken near South 

 Georgia and in the third half-month from the South Shetland-South Sandwich area. 

 The low figure for the first half of March is caused by the great numbers of Meta- 

 nauplius and ist Calyptopis taken at stations just north of the South Shetlands in that 

 half-month. But this indefinite division into a north and south area breaks down alto- 

 gether in the second half of March and first half of April, where the contrasted results 



Fig. 30. Graph showing average growth rate of Euphausia superba during the first fifteen 



months of larval life. 



of the two periods are both derived from larvae from the Bransfield Strait and South 

 Sandwich Islands-Burdwood Bank Line. 



The growth curve can be divided into four parts: a low, gently increasing portion 

 from November to March, a steeper portion between March and June, a part having 

 little or no upward tendency between June and August, and a steeply ascending portion 

 from August onwards. 



The first part of the curve coincides with the observed period of spawning. The average 

 length of the larval krill population will be kept down by the constant addition of very 

 small developmental forms; it is a period of production and growth. The second part 

 extends from the end of the spawning period to the time when winter conditions become 

 eflFective. During this period the average length is not aflFected by influx of very young 



