MOULTING 257 



being disregarded. They occurred, as might be expected in gatherings of such short duration, in very 

 small numbers, singly or in twos or threes, at all depths from the surface down to looo m. The monthly 

 frequency of their occurrence between these levels (Figs. 58-60), based on samples from the East Wind 

 drift, the Weddell drift and the Bransfield Strait, seems to throw a certain light not only on the 

 probable depth at which these animals cast their skins, but also on the period during which the older 

 stages of the surface population are feeding and growing at their maximum intensity. 



Let us consider first the probable level or levels at which the moulting takes place. At first sight it 

 would appear from Fig. 58 that there is a marked increase in the frequency of the moult with 

 increasing depth suggesting that the krill at a time when they must be % 



assumed to be most vulnerable to their surface predators might seek the 

 comparative safety of deep water in order to shed their skins. Since, how- so 



ever, our observations at all levels (pp. 157-70) are directly at variance 

 with the existence of such a movement there must be some other explanation 

 of these apparently contradictory appearances. Obviously it must be that 2s- 



the skins, being dead things, are sinking, and that in consequence of the 

 variable duration of our vertical hauls they tend to appear with greater 



DEPTH 



25- 

 O 

 75 

 50 



75 



50- 

 25- 



75- 

 50- 

 25- 

 O 



-o- 



34 



^" 500- 



® 



frequency in the three deep 250 m. hauls than they do in the shorter ones °- 



above 250 m. For example, if there were one sinking skin located some- 

 where in the 250-0 m. layer it is obvious that a net fishing in only part of 

 this layer, say from 100 to 50 m., would have less chance of capturing it than Fig. 60. Moulting. Frequency 

 one fishing at the same time through the whole 250-0 m. water column. Thus of occurrence of cast skins 

 if the three uppermost net hauls, 250-100 m., 100-50 m. and 50-0 m., be between the surface and 1 000 m. 



, . 111-11 /T^. ^ in the Bransfield Strait, 



represented as one contmuous 250 m. haul, then it would appear (rig. 59) 



that the frequency of occurrence of casts in the four now equally distributed horizons between the surface 

 and 1000 m. is for all practical purposes the same. And from this it could be argued, the casts being 

 dead and sinking, that the moulting takes place somewhere near the surface, or at any rate above 

 250 m., where the main concentrations of the whale food have been shown to be. 



The vertical frequency of occurrence of casts in the Bransfield Strait for the months of November 

 and February combined is shown separately in Fig. 60. From this the same general inferences 

 can be drawn as from Fig. 59, namely, that the sinking skins are more or less equally abundant at 

 all depths covered by our observations and that they are probably shed at the surface. 



Turning now to the distribution of the casts in time it will be seen that on the whole there are 

 enough observations from September to June to indicate a period of intensive growth lasting from 

 November to March, marked by an abundance of cast skins in the plankton, followed by a period 

 of slackened growth, marked by the absence of cast skins from the plankton, the latter period 

 extending from April to June and probably, if we had observations to show it,^ covering July and 

 August as well. As Bargmann (1945) has noted, the periods November to March and April to June 

 correspond respectively although somewhat roughly with the times of the phytoplankton maximum 

 and post-maximal decrease. It is not surprising, therefore, that the former should be found to 

 correspond exactly with the period of rapid growth she records (p. 251, Fig. 54) between October 

 and March, when this species in its second year practically doubles its average length, and that the 

 latter no less exactly should correspond with the period of arrested development indicated by the 

 flattening out of the growth curves of her second-year euphausians between April and June. It is 

 interesting too to note that the period November to March coincides with the known protracted 

 spawning season of the krill. 



^ I did not examine our samples for July and August. 



29 DM 



