METHODS EMPLOYED 



IS 



volume: of sample in c.c. 



estimated diatom total 



in millions 



to the sea water. Having been allowed 24 hours to settle the superfluous liquid was 

 decanted off and the catches transferred to | lb. jars. Before these jars were available 

 glass tubes plugged with cotton-wool wrapped in tissue paper were used : a number of 

 these could be stored in a large jar. When very large catches were obtained, as in 

 November 1930, it was found necessary to retain only a small portion, for it is obviously 

 impracticable to preserve many successive catches of a litre or more of diatoms when 

 working a large number of close stations. 



For various reasons the volume of the settled sample does not give a trustworthy guide 

 to the number of organisms contained. The most important of these lies in the fact that 

 comparatively small numbers of spiny forms, 

 such as large Chaetocerids, or species forming 

 long robust chains, such as Rhizosolenia styli- 

 formis, occupy a large apparent volume, as com- 

 pared with that taken up by quite large numbers 

 of the more compact forms, such as Coscinodiscus 

 spp. Further, an apparently small sample may 

 contain millions of minute forms such as Nitz- 

 schia and Fragilaria. However, the volumes have 



been given at the top of the tables of analyses, as 



it has been found that where the proportions of 



the principal species remain fairly constant, they 



do show a fair agreement with the estimated 



numbers of organisms (see Fig. 3): also, in the 



case of very large samples, they afford some 



justification for the minute fractions examined. 



On the other hand, in the very rich and varied 



samples taken round South Georgia, it will at 



once be seen that the volumes do not possess any 



constant relation to the numbers of organisms 



(e.g. Tables X-XVI). 



The actual estimations were made by Hen- 

 sen's methods, as adopted by Prof. Hardy (in 



press). It was found that it was impossible to 



do this work at sea. Accordingly, as many 



samples as possible were worked up whenever the ship was in port, and the remainder 



while serving at the Marine Biological Station in South Georgia, and in England. 



Owing to the necessity of obtaining a preliminary idea of the nature of the phytoplankton, 



in some cases the practice was adopted while abroad of counting much smaller fractions 



of the larger samples than are justified by the limits of accuracy of the method, if a true 



quantitative analysis is aimed at. Two circumstances led to the retention of this rough 



comparative method, after consultation with Prof. Hardy and Dr W. H. Pearsall of 



Leeds. 



=T= 



376 378 379 380 381 3BE 383 384 385 386 387 388 



Fig. 3. Graph showing the close relation 

 between volume of sample and estimated 

 numbers of phytoplankton organisms, found 

 only when the quality of the plankton re- 

 mains fairly constant. 



