43§ 



DISCOVERY REPORTS 



hemisphere, would also follow naturally. During the month of November most of the 

 other southern diatom genera reach their great maxima in the northern part of the 

 Antarctic Zone, and to a lesser extent in the southern part of the sub-Antarctic Zone — 

 the true home of R. curvata — also. I believe therefore that the drop in the November 

 figures may have a real significance, these essentially summer forms tending to be kept 

 down by the vast numbers of other species present. The rise from mid-season onwards, 

 with a well-marked maximum in February, is beyond doubt. If the two largest catches 

 in this month are excluded, on the grounds of the exceptional conditions under which 

 they were obtained, the mean figure is reduced to 11,270 and still remains more than 

 twice as great as that for any other month. The slight difference between the December 

 and January figures, on the other hand, is probably not significant, being due to insuf- 

 ficiency of data for the latter month, when the ships have nearly always been working 

 too far south to capture the species. 



The slight increase in May is probably a real one ensuring a sufficiency of individuals 

 to maintain the stock through the unfavourable winter period when mortality is certainly 

 high. Moreover, the measurement data indicate a high degree of auxospore formation 

 during May, as will be seen from Table VIII. 



It should be borne in mind that the counts of this never very abundant species, small 

 as they are, are obtained from quite large fractions of the catch, and are very much more 

 accurate than ordinary estimates of the whole sample, where the numbers of some species 

 often run into millions. 



MEASUREMENT INVESTIGATION 

 The method of collection and presentation of the data on variation in size has already 

 been fully described (pp. 420-21). The seasonal variation in percentage frequency of the 

 different diameter classes is shown in Table VIII. The percentages of individuals of 

 diameter less than 60/x is shown to the left, and the percentage of greater diameter in 

 the right-hand column. 



Data not strictly comparable. 



