RHIZOSOLENIA CURVATA 423 



of Bouvet Island, close to the Meteor bank, and south of Kerguelen on the eastern 

 margin of the submarine ridge connecting Kerguelen and Heard Island. Thus both 

 these records are from localities in which sharp changes in bottom relief may be expected 

 to complicate the circulation of the water layers. 



The results obtained in the Scotia Sea proper are given in Table II, and the positions 

 of the stations are plotted in Fig. 1 . 



Broadly speaking, the largest numbers of R. curvata lay on the sub-Antarctic side of 

 the convergence, e.g. Sts. WS 69, WS 519. Several catches of more than average 

 quantity were, however, obtained on the Antarctic side of the convergence. At some 

 of these the high-surface temperatures furnish good evidence of sub-Antarctic water 

 having been transported beyond its normal limit. Truly Antarctic surface water is rarely 

 warmed above 4-5° C, while the extreme limit is probably about 5 C. in February 

 close to the convergence. Yet Sts. 634 and 830, where the surface temperatures were 

 4-57 and 5-13° C. respectively, and where a considerable quantity of R. curvata was 

 present, lay well within the normal limits of the Antarctic Zone. Again, at St. 1335, where 

 another moderately rich haul was obtained, the temperature, 3 -17° C, was very high 

 for a station so far within the limits of the Antarctic Zone, if the time of year is taken into 

 consideration. This argument gains force from comparison with data given by Deacon 

 (1933, p. 199, fig. 14) for a station lying approximately the same distance within the 

 convergence, but farther eastwards, due north of South Georgia. 



Examination of the distribution chart in conjunction with the data tabulated above, 

 however, shows that so many records of R. curvata have been obtained on the Antarctic 

 side of the average position of the Antarctic convergence in the Scotia Sea, at quite low 

 temperatures, that they cannot all be explained simply on the grounds of actual move- 

 ment of unmixed, or slightly mixed sub-Antarctic water beyond its normal limit, even 

 when full allowance is made for seasonal variation in temperature. It is believed that 

 the presence of the species in these localities is to be explained by its ability to persist 

 for some time in mixed water, initially formed by direct transport but so cooled and 

 diluted that its mixed origin is no longer demonstrable by ordinary physical criteria. 

 We know that the species can remain alive for short periods round South Georgia, so 

 that it should be quite possible for it to continue for a considerable time in mixed water 

 in which the sub-Antarctic element has been almost completely dissipated, especially 

 during the warmest months of the year. In spring the very small numbers of the species 

 present at the stations where the lower temperatures were recorded strongly suggest 

 that the environment was unfavourable, while the measurement investigations described 

 in a later section of this paper (p. 438) supply good evidence that it can never persist 

 within the Antarctic Zone indefinitely. 



The distribution chart (Fig. 1) shows that nearly all the records of R. curvata on the 

 Antarctic side of the convergence in the Scotia Sea lie east, or north-east of the easterly 

 projection of sub-Antarctic water where the S -shaped bend in the convergence occurs. 

 The preponderance east of the upper, or more northerly loop is slight but fairly well 

 marked. Mixing across the convergence will obviously tend to be most frequent in this 



