424 



DISCOVERY REPORTS 



region, where the convergence lies more or less normal to the path of the prevailing 

 westerly gales instead of parallel to it, as elsewhere. Once within mixed water on the 

 Antarctic side of the convergence, the general north-easterly trend of the surface drift 

 would tend to sweep the species to positions opposite the northern loop of the S, 

 although their probable " point of entry" is the eastern extremity of the southern loop. 

 Only one of the Scotia Sea records of R. curvata lies beyond the extreme probable 

 limit of sub-Antarctic influence, which Deacon puts at approximately 150 miles. This 

 station, 658, was worked over the bank between the Shag Rocks and South Georgia, so 

 that the presence of the species there is probably due to the same factor which is thought 

 to account for its presence round the island of South Georgia itself, namely subsurface 



Fig. 1. Observations of Rhizosolenia curvata in the Scotia Sea. 



transport followed by vertical mixing. A few other more easterly Scotia Sea records lie 

 close to the Shag Rocks, and it is possible that these are also due to the same factor. 

 Comparison of Fig. 1 with the bathyorographical chart of the area published by Herd- 

 man (1934, pi. xlv) shows this possibility very clearly. 



One other feature of the records in the Scotia Sea which is worthy of special note is 

 the distribution of the richer hauls of R. curvata in the later part of the year. It can be 

 seen from Table II that out of the ten catches in which more than 10,000 individuals 

 were recorded, seven were obtained in February, two in March and one, the smallest, 

 in April. This is in broad agreement with the results of a study of the seasonal variation 

 in abundance described later, but it may be mentioned here that it seems to present an 

 interesting analogy to the seasonal behaviour of other members of the genus in corre- 

 sponding latitudes in both hemispheres. These also tend to show their maximum 



