RHINCALANUS GIGAS 303 



July or August 1932, but in September in the South Pacific Ocean Rhincalamis had 

 practically disappeared from the catches. 



It seems, then, that R. gigos spent the summer months November 1932 to at least 

 February 1933 at the surface, mainly in the loo-o m. layer. The progressive reduction 

 of the catches during the winter months April-September 1932 in the West Wind Drift 

 current around the Antarctic Continent, together with the descent of the catches from 

 the upper to the lower nets in April, suggests that the species left the surface and 

 descended into depths below 250 m. during the winter. During September 1932 it was 

 out of range of the 250-100 m. net but came within its range in October 1932. 



The same process appears to have taken place in the previous season, 193 1-2, but the 

 movements seem to be less clearly defined. In November of that season the larger pro- 

 portion of the catches at most of the stations was in the lower nets. It is noticeable, 

 however, that stations immediately south of the Antarctic convergence in the Drake 

 Passage (Sts. 731, 733, 745) show a majority in the upper nets. During December in 

 the western Scotia Sea and in South Atlantic water north and west of South Georgia the 

 stations immediately south of the convergence (Sts. 755, 775 and 751, which was taken 

 almost on the convergence) again show a majority in the upper nets, while those farther 

 removed from the convergence (Sts. 753, 757, 759, and 774) show a majority in the 

 lower nets. At three stations also during these months — Sts. 726, 737 and 769, which 

 were taken in places where warm deep water wells upwards to the surface — a majority 

 was also found in the upper nets. St. 726 is situated near the coast of South America, 

 St. 737 near the continental shelf, and St. 769 on the ridge between the South Shet- 

 lands and the South Orkneys. At the few stations taken in the second half of 

 December 1932 and in January and February 1933 the majority of the catch was in 

 the upper nets. 



Thus it appears that the rise of Rhincalanus to the surface in the season 193 1-2 in 

 Antarctic water away from the convergence took place in early December, nearly a 

 month later in the year than in the following season 1932-3, when the rise to the surface 

 occurred at the beginning of November. Deacon (1936) has found that the hydrological 

 data show the season 193 1-2 to have been colder and " later " (with respect to the south- 

 ward movement of the isotherms) than the season 1932-3. The average temperature of 

 the surface 50 m. of water in the region round South Georgia between 52 and 56° S and 

 33-40° W was 1-50° C. in the middle of January 1932 and i-88° C. in the middle of 

 January 1933. In 193 1-2, however, the rise oi Rhincalanus to the surface appears to have 

 already taken place in November at stations near the convergence, so that it is possible 

 that the spring ascent to the surface may take place earlier near the convergence than 

 elsewhere. The precise conditions under which this ascent takes place are at present 

 unknown, but it is perhaps to be expected that it will occur earlier near the convergence 

 than elsewhere, since the warming effect of the approach of summer has been found 

 (Deacon, 1936) to become marked sooner in the region just south of the convergence. 

 It is permissible then to suggest that the spring ascent is connected with the attainment 

 of a certain temperature by the water at the surface, and, indeed, that the whole 



