METHODS 267 



depends upon the actual manipulation, and many of the earlier counts have had to be 

 discarded. The method finally adopted, and which gave what seemed to be fairly 

 consistent results, was as follows: A small electric centrifuge carrying four 12-5 c.c. 

 tubes was employed and the samples centrifuged for 5 min. at 2000 r.p.m., the highest 

 speed at which the machine could be run at sea without excessive straining. Longer 

 periods of centrifuging did not lead to appreciable increase in the number of organisms 

 deposited. The supernatant liquid was very carefully drawn off with a special pipette 

 with a recurved tip, similar to the arrangement employed by Marshall (1933, p. 112). 

 We found that this gave very much more consistent results than pouring, as recom- 

 mended by Nielsen (1933). The liquid remaining in the tip of the tube (about 0-3 c.c.) 

 was then cautiously agitated with a straight pipette to remove the crust of organisms 

 adhering to the glass, and transferred to a cell on a large squared slide. Here it was 

 trapped under a no. i cover-glass of the largest rectangular size, and the organisms 

 counted under an ordinary microscope in the usual way, with the aid of a large me- 

 chanical stage. 



Recentrifuging of the supernatant liquid usually gave about lo'^'o of the original 

 count for most species, so to allow for this and loss in manipulation 12 c.c. were 

 reckoned as 10 c.c. in working out the results (cf. Gran, 1929, p. 6). 



Series of counts from o, 5, 10, 20, 50 and 100 m. were obtained from 119 stations, 

 apart from the earlier experimental efforts which had to be discarded. While the work 

 was in progress Nielsen's (1933) severe criticism of the method appeared, from which 

 it seemed that centrifuge results could not even be considered as roughly comparable 

 at different stations. It had long been known, of course, that the method did not approach 

 the ideal of an ' absolute ' estimation (Allen, 1919), and in view of the new unfavourable 

 evidence it seemed useless to persevere with it. Unfortunately, the alternative sedi- 

 mentation method advocated by Nielsen did not lend itself to our immediate purpose, 

 for reasons discussed in the next section of this paper. It is felt, however, that these 

 counts still provide a valuable clue to the probable type of depth distribution of the 

 phytoplankton as a whole, and some evidence regarding organisms which may be 

 missed by the nets. They have accordingly been considered briefly from these points 

 of view, though it is now evident that the full data are not worth publishing. 



These centrifuge counts strongly supported the impression gained from the experi- 

 mental work of Marshall and Orr (1928) that within the Antarctic zone production 

 would be limited to the upper 50 m. or so by the minimum light requirements of the 

 organisms. We had further evidence of this from experimental net hauls, which 

 prompted us to use the 50-0 m. Harvey net haul as our best indication of the relative 

 order of production throughout, though on rare occasions large quantities of diatoms 

 are to be found at lower levels. 



The presentation of the results is based on arithmetical means of the observations at 

 mean dates, in several regions or areas within the Antarctic zone. The areas have been 

 chosen according to the degree of uniformity of the conditions, both physical and 

 biological, observed within them, as described on pp. 278-80. It will be realized that 



