3i8 DISCOVERY REPORTS 



autumnal maximum is probably even more variable. Some of our earlier work, and 

 observations in the adjacent waters of the Scotia Sea, suggests that in some years it may 

 take place as late as April or even May, and that sometimes it is hardly apparent at all. 

 Wherever the seasonal cycle has been studied intensively in temperate — polar waters, 



Z5000 



20000- — 



15000 



10000 



5000- 



Julu August September October November December January FebruarLj March April Mai^ 



June 



Fig. 13. South Georgia area. Seasonal variations in plant pigments per m.', means of 50-0 m. hauls at 

 mean dates. Numbers of observations shown in brackets. Note scale is necessarily much smaller than that 

 used for oceanic regions and for the Scotia Sea. © = 1933-4. = i934~S- • = 193^7' 



Table 8 



