RHIZOSOLENIA CURVATA 42 i 



The figure arrived at was 62-7^.. In order to indicate the extent to which the individuals 

 tended to be larger or smaller than the mean in any given month, the percentages in the 

 several diameter classes above and below 60/x were therefore added together and entered 

 in two additional columns, the percentages below 60/x to the left, and those above 60/x 

 to the right of Table VIII. 



In the attempt to correlate size and temperature suitable samples have been grouped 

 together for each degree of temperature over the whole of the organism's temperature 

 range. As far as possible these samples were selected from among those taken in the 

 Scotia Sea and adjacent areas, during the months in which R. curvata appears to find its 

 seasonal optimum. As with the endeavour to follow the seasonal variation in size, how- 

 ever, it was necessary to bring in some data not strictly comparable with those forming 

 the main body of the evidence, in order to cover the lower extreme of the range, for 

 which spring observations only were available. At all the other temperatures ample 

 data were available from samples taken in late summer (Table IX). In all other respects 

 the data indicating the degree of correlation between size and temperature have been 

 treated in the same way as those indicating the seasonal variation in size. 



One other aspect of size variation was studied with a view to illustrating the applica- 

 tion of these measurement investigations to the use of R. curvata as an indicator species. 

 This was a correlation between size and distance from the Antarctic convergence. It is 

 obvious that this relation should be a broad reflection of the temperature effect, slightly 

 masked by the seasonal variation. For this purpose all the measurements obtained in 

 the initial examination of the samples were used. The samples were grouped according 

 to their distance from the convergence, and the percentage frequencies of the diameter 

 classes in each grouping were worked out and tabulated as before (Table X). The 

 average numbers of R. curvata per net haul in each of these distance groupings has also 

 been shown in the same table. 



I wish to express my thanks to Miss E. C. Humphries, who has prepared the text- 

 figures for publication, and to the staff of the cryptogamic and general sections of the 

 library of the British Museum (Natural History) whose ready assistance has been very 

 helpful. 



OBSERVATIONS IN THE SCOTIA SEA AND 

 SOUTH GEORGIA AREA 



The fact that R. curvata is occasionally to be found in small numbers round South 

 Georgia, which lies some 220 miles to the south of the Antarctic convergence, was first 

 established by Hardy (1935, p. 53). It was obvious even at that time that it was never 

 frequent there, but was relatively abundant in sub-Antarctic waters to the north and 

 west. Subsequent examination of more than 120 samples from the South Georgia area, 

 many of the results of which have already been published (Hart, 1934, Tables II-XVI), 

 proved that R. curvata was very rare and usually found only to the west of the island, 

 in numbers fewer than 1000 cells per net haul. Nevertheless, the undoubted presence 

 of more or less living individuals in this area, well south of the Antarctic convergence, 



