172 DISCOVERY REPORTS 



because they came from a sample fixed in formalin 27 years ago and were not taken directly from the 

 sea. Ponomareva (p. 145), using fresh Thysanoessa inermis, and estimating the wet weight from 

 specimens dried off on filter paper, gives 0-57 g. as the average wet weight in samples ranging from 

 15 to 26 mm. long, and using this figure, assuming it to be the same in the corresponding length 

 range in E. siiperba, the density of the krill would work out at 23-18 g-jm^, approximately 40 times the 

 equivalent weight of whale flesh in the East Wind-Weddell stream. 



My estimates for the absolute density may be wrong, but at least they cannot be wildly wrong in so 

 far as even the lowest result, 2-50 g./m.^, is still substantially higher than what it seems might be the 

 corresponding density for whales. It would obviously be wildly wrong if it were less. Furthermore, 

 Mackintosh and Brown's admittedly rough estimate for the Antarctic whale population has been 

 thought to be too low (see p. 145); and if it should prove to be larger than these authorities have 

 computed, and Vangstein (1956) seems to think it might be far larger, then the krill population that 

 supports it would have to be correspondingly large, which suggests that of my three estimates for 

 absolute density the two higher are the least likely to be wildly out. Even these, however, in the long 

 run may prove to be low, perhaps much too low. They may refer in fact only to the immediate vicinity 

 of the surface, perhaps only to the top 5 m. For as yet we have been able to form no clear picture 

 as to how many, how spaced and how large are such patches as probably exist below the surface, 

 beyond, although possibly (pp. 268-78) not very far beyond, the range of human vision. We must also 

 allow for the possibility that many patches may be more than i yard thick. Gunther, for instance 

 (p. 149), mentions 'a metre or two' and Matthews (p. 149) 2 fathoms. If there are in fact many such 

 patches then obviously our highest estimate for absolute density, even for the surface zone alone, is 

 much too low. 



Food 

 As both Barkley and Hart (p. 45) have shown, Fragilariopsis antarctica is the most frequently occur- 

 ring and abundant organism in the stomach contents of the krill. It would appear in fact from Barkley's 

 findings that this species constitutes their staple diet, the average preponderance of this over all other 

 organisms in the stomachs being as Si-i to 18-9%. Many other organisms or remains of other 

 organisms, are recorded by Barkley, and a list of these is given below. I take this opportunity to 

 publish at the same time a second list compiled from some field notes made by Dr Hart in 1937-8. 

 In gratefully acknowledging this valuable source of information, I would point out that Dr Hart 

 himself did not at the time consider his results of much value, since they were based on the examina- 

 tion of only forty stomachs. In view of Barkley's findings, however, they prove to be of great interest. 

 Neither list refers to any particular stomach, both showing the total organisms, other than Fragilari- 

 opsis antarctica, recorded in the combined stomach contents of the krill these workers examined. The 

 organisms considered by both Barkley and Hart as next of importance in the diet are shown in bold 



type. 



Broadly the two lists agree, both seeming to emphasise that small, spineless diatoms are the principal 

 constituents of the stomach contents. It is perhaps a little surprising that Barkley does not record 

 Thalassiosira, but possibly his ' Kleine Coscinodiscen ' would include this species. In summarising 

 his results Barkley states that the dietary of all sizes of the krill ranging from 10 mm. long, which 

 would include both Fifth and Sixth Furcilias, to 60 mm. long, is essentially the same. In his 

 notes Hart observes a tendency for small animals to feed more voraciously than large ones and for 

 adult males to feed more voraciously than gravid females.^ 



The occurrence of Cocconeis ceticola in three of Barkley's samples, one of them with 115 cells, is 

 of great interest. Hitherto this species had been known (Bennett, 1920; Hart, 1935; Peters, 



^ Bargmann (1945) states that the gravid females are unable to feed. 



