434 DISCOVERY REPORTS 



it lay 65 miles north of the average position of the convergence, and 35 miles north of 

 its probable actual position at the time when the observations were made. There is thus 

 no doubt of the mainly sub-Antarctic composition of the surface water there. 



The differences noted between the "probable average" and probable actual positions 

 of the convergence show clearly the difficulties introduced by quite small variations in 

 the position of the convergence into comparisons of the type presented in this paper. 

 It is impossible to give more exact data than the probable average position plotted on the 

 distribution charts, except where very full series of observations normal to the isotherms 

 are available, as here. A similar anomaly is shown by the figures for St. 1475, where 

 small numbers of R. carvata were obtained. Here the very low surface temperature, 

 0-40° C, was sufficient to show, even in the eastern South Pacific, that the degree of 

 sub-Antarctic mixing, due to strong northerly winds prevalent at the time, was but 

 slight. Yet this station lay on the sub-Antarctic side of the average position of the con- 

 vergence. The actual position of the convergence at the time, however, lay 45 miles to 

 the north, and so the apparent anomaly explains itself. 



Apart from these indications of the probable sources of discrepancies in the more 

 widely scattered data from other areas, the broad significance of these records in longitude 

 8o° W is quite clear. They give strong support to the general theory of the distribution 

 of R. curvata which cannot be postulated so definitely for the less well-worked areas. 

 Thus all the heavier catches were in sub-Antarctic water, for on the only occasion upon 

 which a considerable haul was obtained south of the average position of the con- 

 vergence, it seemed highly probable that the convergence had actually shifted even 

 farther south (cf. the relatively high surface temperature at St. 13 15). 



In the other areas investigated all the way round the world, the most southerly 

 records of R. curvata lie well north of latitude 60 ° S. In the eastern South Pacific, in 

 longitude 8o° W and at Sts. 730 and 1466 for example, records south of 6o° S are com- 

 mon. Reference to Fig. 7 in the concluding section of this paper, where nearly all the 

 records have been plotted on a circumpolar chart, will show at once how closely this 

 great southward extension of the species is correlated with the southward sweep of the 

 convergence where it rounds Cape Horn. 



The observations in longitude 8o° W also show that in five out of six successive series 

 of observations the maximum numbers of R. curvata were recorded at the stations im- 

 mediately to the north of the convergence. This lends very strong support to the view 

 that it is in the colder portion of the sub-Antarctic Zone that the species finds its 

 optimum. The fact that the highest numbers of all have been obtained at somewhat 

 higher temperatures in the Scotia Sea and southern Indian Ocean does not detract from 

 the value of this evidence. The more northerly position of the convergence and the fact 

 that the sub-Antarctic Zone itself is much narrower in those regions leads to a crowding 

 together of the isotherms. Actually, therefore, these very rich stations (e.g. Sts. WS 69, 

 WS 520 and 882) are no farther from the convergence than the southerly stations in 

 8o° W, and it is still quite correct to regard the former as lying within the colder part of 

 the sub-Antarctic Zone. When it is added that the temperatures in the warmer part of 



