METHODS EMPLOYED 17 



more concrete expression of relative abundance than the "common", "very rare" and 

 so on, of earHer plankton workers. With the smaller samples some approach to the 

 actual numbers has probably been attained, as shown by Hardy. It may also be 

 claimed that the counts show the relative proportions of the leading forms at each 

 station with some accuracy. The personal error involved may be taken as the same all 

 through, as care has been taken to repeat some of the earlier analyses after a higher 

 degree of proficiency had been attained. 



The centrifuge method evolved by Gran permits more accurate estimations to be 

 made, but can only be employed with advantage when the dominant forms are known, 

 and it is possible to arrange the method of investigation accordingly. Gran (1932, 

 p. 343) has himself pointed out that the success attending the employment of this method 

 in the northern hemisphere has been largely due to the knowledge of the general dis- 

 tribution of the more important forms gained during earlier work with nets. 



Very little was known about the distribution of the leading forms in the south when 

 these observations were begun, but by the examination of a large number of net hauls 

 it has been possible to obtain a general view of the nature and extent of the principal 

 associations. If the centrifuge method had been adopted it would not have been possible 

 to make observations at so many different points during the same period of time, and 

 for this reason alone it appears that net hauls are still the most valuable method in 

 deahng with relatively unknown areas. At the few stations at which it was practicable 

 to use the centrifuge method, excellent results were obtained on Gran's lines, and it is 

 hoped to employ this method more frequently in the future. 



One definite advantage presented by net hauls lies in the capture of the rarer forms. 

 It is true that the numbers estimated are even more in error than those of the more 

 numerous species, but at least their presence is noted, whereas they would almost 

 certainly be missed by the centrifuge method. This in itself would be of small im- 

 portance, were it not for the fact that some of the rarer forms seem to be confined to 

 particular types of water, whose probable extent can thus in some measure be defined 

 by their occurrence. The presence of small numbers of littoral forms also often provides 

 a valuable clue as to the extent of inshore influences. 



In making the analyses given in this paper, all the forms were identified down to 

 species as far as possible, with the aid of the works already mentioned. There did not 

 seem to be any new forms among the more important species. Two categories of organ- 

 isms were less thoroughly dealt with : the diatoms belonging to the genus Coscinodisciis, 

 and the smaller and less numerous dinoflagellates. 



With regard to the genus Coscinodisciis, only one form, the large, easily recognized 

 C. botivet, was ever taken in large numbers at any distance from the land. On the occa- 

 sion when other members of the genus reached their highest relative abundance round 

 South Georgia, on the January-February 1930 survey, the combined numbers recorded 

 formed only 1-2 per cent of the total phytoplankton. As regards species, this genus is 

 extraordinarily richly represented, as will be seen from Van Heurck's list (1909, pp. 

 83-5), where over a hundred named species and varieties are recorded as occurring in 



