3i6 DISCOVERY REPORTS 



chaetognath Sagitta gazellae was present in both Antarctic and Subantarctic waters the catches were 

 extremely small in a region loo miles north to 200 miles south of the convergence and he distinguished 

 separate races in the population to the north and south of it. It will be clear from Fig. 3 that the pattern 

 of distribution for E. triacantha is the reverse of that for S. gazellae and that the catches in fact reach 

 a maximum in the vicinity of the convergence. David suggests that it is likely that the sudden changes 

 in physical conditions, brought about by the fluctuations in the position of the Antarctic convergence, 

 are unfavourable to S. gazellae. In the case of E. triacantha, however, such conditions are evidently 

 favourable. 



-r 

 500 



■fOO 

 NORTH 



200 too A 



NAUTICAL 



100 200 

 MILES 



300 



400 500 



SOUTH 



600 



800 



Fig- 3- The number of adolescent and adult E. triacantha in a series of shallow oblique night hauls with the i m. net from 

 north-south lines of stations plotted according to the distance in nautical miles from the Antarctic convergence. The number 

 per 20 min. haul is expressed as a percentage of the total catch for each line. The position of the convergence has been taken 

 from Mackintosh (1946, Table 9). 



Seasonal variation 

 In Fig. 4 the distribution of adolescent and adult E. triacantha in relation to the Antarctic convergence 

 is shown for each season of the year. The mean catch per 20 min. haul has been plotted at intervals 

 of 50 miles north and south of the convergence. The crosses indicate the intervals for which no data 

 are available. In order to obtain sufficient observations lines of stations from several different regions 

 of the Southern Ocean have been used; the number taken from each sector in each season is shown 

 in Table i . The exact position of the convergence was indeterminate for a line of stations (23 1 1-2320) 

 along the Greenwich meridian in April and the mean position of 49° 50' S, based on nine other lines 

 along the same meridian, has been used as an approximation. This is unlikely to fall more than 

 20 miles from the actual position and as the observations are in 50-mile groups no significant error 



