THE SUBTROPICAL CONVERGENCE 59 



winter and the 14-5° C. isotherm in summer belongs entirely to a movement of sub- 

 tropical water towards the south, and is not mixed with sub-Antarctic water. The mean 

 annual position of such a boundary determined from the monthly temperature charts of 

 Klaehn (191 1, pi. 35) is shown in Fig. 4 (p. 19). It does not extend south of 43-44° S, 

 although the Brazil current can, according to Klaehn, be traced to 49° S, and it can only 

 be concluded that in the last 5-6 of latitude the subtropical water of the current be- 

 comes more and more mixed with sub-Antarctic water. The monthly temperature charts 

 show that the relatively unmixed subtropical water penetrates 3-5 farther south in 

 summer than in winter, owing to the greater strength of the Brazil current in 

 summer (see Klaehn, pp. 657-60). 



Farther east the subtropical convergence recedes gradually northwards. Between 

 41 43' S, 42 21' W and 35 18' S, 19 01' W the convergence lies approximately along 

 the line of observations made at Sts. 72-8 (Station List, 1929). Sts. 243 and 245 in 

 38 48' S, 27 22' W, and 38 20' S, 22 18' W {ibid.) appear to lie just north of the con- 

 vergence. Observations made in January 1926 near Tristan da Cunha in 36 55' S, 

 12 12' W, and farther south in 39 25' S, 12 08' W (Sts. 4 and 7, ibid.), indicate that 

 the convergence was then just south of the island, but others, made between Sts. 78-83 

 and Sts. 246-9, in June 1926 and June 1927, show that the convergence was probably 

 north of the island. Observations made by the 'Challenger' (St. 133) and the 'Gauss' 

 (Drygalski, 1926, pp. 423 et seq.) point to the same conclusion. In November 1933, 

 however, the convergence was quite definitely 5-6 south of the island. 



The data collected so far show therefore that the position of the convergence fluctu- 

 ates over at least 6° of latitude. They give some indication that the movement is seasonal, 

 towards the south in summer, and towards the north in winter ; but the high latitude 

 of the convergence in November 1933— at the end of winter— suggests that it has other 

 large movements, probably dependent on corresponding meteorological changes. The 

 mean position of the convergence indicated by the rather scattered data is shown in 

 Fig. 4. 



In the eastern half of the Atlantic Ocean as far as 15° E the convergence lies between 

 35 and 38 S. It then bends sharply southwards as far as 40 S and there, owing to the 

 presence of a tongue of water which pushes its way towards the west from the Agulhas 

 current, it bends back in a salient towards the west before it continues its general trend 

 towards the east in about 44 S. 



The data available from this region suggest that the position of the convergence varies 

 over a range of at least 100 200 miles. The position shown in Fig. 4 is as nearly as 

 possible the average boundary between the two waters, but it is not unlikely that sub- 

 tropical water will frequently be encountered outside it. The tongue of water which 

 flows westwards from the Agulhas current into the Atlantic Ocean south of 40 S 

 appears to be particularly subject to variations. The temperature and salinity distribu- 

 tion in October 1930 showed that the subtropical water extended as far as St. 450 in 

 44 58' S, 12° 58' E, and in March 1933 it was found as far west as St. 1 165 (section 7) 

 in 41° 01' S, 9 34' E. In deciding the position of the convergence the observations 



8-2 



