428 



DISCOVERY REPORTS 



except at St. 883, where Deacon (1937) has already been able to demonstrate a direct 

 southward movement of surface water by hydrological criteria. Deacon also con- 

 sidered that conditions at St. 891, even farther within the normal limits of the Antarctic 

 Zone, were indicative of the operation of this same factor. This reversal of the normal 

 direction of the surface drift is believed by Deacon to be a fairly frequent occurrence 

 during the coldest months of the year. 



Table IV. Observations in the Southern Indian Ocean and south of Australia 



Two other points of interest arise from consideration of these results. Firstly, it can 

 be seen from Table IV that while the largest hauls were obtained from stations worked 

 in the sub-Antarctic surface water towards the upper limit of the temperature range of 

 the species, at the highest temperatures of all, at Sts. 849, 881 and 893, the falling off in 

 numbers is very distinct. This tallies perfectly with the results obtained from the 

 Southern Ocean as a whole, which show that quite close to the upper limit of its tem- 

 perature range, R. curvata exhibits a marked decrease in numbers and in the size of the 

 individuals. Secondly, there is a strong suggestion of a slight secondary autumnal 

 maximum in the numbers of this never very abundant species, furnished by the 

 relatively rich hauls at Sts. 882, 883 and 1368. This again has been found to hold for the 

 Southern Ocean as a whole, but it is not nearly so well marked when all the areas, each 

 with its own local variations in environment, are considered together. 



OBSERVATIONS SOUTH OF THE TASMAN SEA AND 

 IN THE SOUTHERN PACIFIC OCEAN 



Our records of R. curvata from this area are mostly from its eastern and western 

 extremes. In the central part of the Southern Pacific Ocean very few stations have been 

 worked in the sub-Antarctic Zone, and these at the least favourable time of the year, 

 when the species is known to show its minimum abundance elsewhere. In plotting the 



