5 4 DISCOVERY REPORTS 



the far south, and well after the maximum concentration at the convergence. The 

 increase of phosphate must be due to regeneration, but as the salinity and temperature 

 results between o and 60 m. at Sts. 13 15 and 13 16 show, it was due to regeneration in 

 sub-Antarctic surface water which had passed over the Antarctic convergence. In the 

 south of the sub-Antarctic zone the main outburst of phytoplankton occurs earlier than 

 south of the convergence; regeneration will set in earlier too, so that the effect of a 

 withdrawal of phosphate at the ice-edge in February was masked in the region between 

 64I and 59 S by the presence in the 0-60 m. layer of sub-Antarctic surface water which 

 was showing a high degree of regeneration. If we assume that at approximately 66° S the 

 phosphate content in February 1934 was lower than in March of the same year and that 

 the mean integral average of the 0-100 m. layer in February was about 100 mg., then 

 a minimum withdrawal of about 20 mg. would occur in this layer during the phyto- 

 plankton season in this latitude. At the Antarctic convergence no estimate of the 

 consumption of phosphate can be given owing to the absence of observations unaffected 

 by sub-Antarctic water. 



SEASONAL VARIATION OF SILICATE IN THE 

 SURFACE LAYER IN 8o° W 



The mean integral averages of the 0-100 m. layer have been calculated and are plotted 

 in Fig. 17. North of about 63^° S only a very small variation in silicate content of this 

 layer was found in September, October, November and December. In 1934 the 

 September, October and November observations just south of the convergence were 

 made within a period of 64 days, during which time the increase in concentration of 

 phytoplankton in this position was very small. This lack of variation in the silicate content 

 of the surface layer during these months also reflects the paucity of the concentration 

 of phytoplankton. 



The silicate values in March 1934 were everywhere lower than in any other month, 

 and the curve in Fig. 17 for this month does not show the peaks which are such a 

 conspicuous feature of the phosphate curve. This is because the silicate content of 

 sub-Antarctic surface water is considerably lower than that of Antarctic surface water, 

 and the ratio of phosphate to silicate in the Antarctic water is much smaller than in sub- 

 Antarctic water. Hence the presence of sub-Antarctic water in the 0-60 m. layer at 

 Sts. 1 3 15 and 13 16, which is showing a high degree of regeneration of phosphate, will 

 not cause corresponding peaks in the silicate curve for these stations. 



In all months a fall of silicate occurs northwards from the ice-edge as far north as 

 about 62|°-63|° S, where a horizontal portion is found in all the curves. There is no 

 counterpart of this horizontal portion in the phosphate curves except in October ; the 

 existence of this horizontal part of the curves is not dependent on phytoplankton activity 

 alone, because it occurs both in September and October when the concentration of 

 phytoplankton is extremely small and in November and December when the amount 

 of phytoplankton is increasing. 



The curves show that there was a rapid decrease of silicate in the 0-100 m. layer 



