l82 



DISCOVERY REPORTS 



Figs. 2-4 include the positions listed in Table 9 except No. 20 which was vague and unreliable. 

 Different positions have been joined up, wherever it seemed reasonable to do so. This seems to provide 

 the best basis for estimating the mean position of the convergence, and it gives an impression of the 

 extent to which its position is liable to vary. The majority of positions are in the Atlantic sector. Data 

 are still rather scarce elsewhere, and the points plotted in Fig. 4 are joined by lines drawn simply in 

 what appear to be the most probable positions. In Plate I the convergence is drawn to represent as 

 nearly as possible the mean of the lines in Figs. 2-4. Wherever possible the average latitude has been 

 calculated within narrow limits of longitude, and those records which are marked ' Good ' or ' V. good ' in 

 Table 9, were counted twice in the averages so that some bias should be given to the more reliable data. 



Fig. I. The Antarctic convergence in the Falkland Islands Sector, September to December 1934. Numbers refer to the serial 

 numbers in Table 9 (p. 205). Black dots denote well-defined positions, and rings approximate or uncertain positions. Pecked 

 lines show the ascertained position of the convergence (apart from unknown minor irregularities), and dotted lines show its 

 probable course where it is not checked. Continuous lines show the tracks of the 'Discovery 11' (see also Fig. o, p. 201). 



The mean line of the convergence, shown in Plate I, like most of the results set forth in this paper, 

 is an estimate which is open to adjustment in the light of any additional data which may be obtained 

 in the future. It is obvious from Figs. 2-4 that it is based on far better material between 80° W and 

 30° E than elsewhere. In the Falkland region the estimated position must be very near the true mean 

 position, but in the Pacific sector, between no and 160? W it may be far from accurate. 



Some comments are needed on the extent of variation from the mean. I have made a rough measure- 

 ment of the distance of each plotted point from the estimated mean position, and the resulting figures 

 given in Table i are perhaps worth noting. This table includes all the positions in Table 9 for which 

 a measurement could be made, and the possible extent of the displacement may not be so great as it 

 seems to suggest. Most of the larger deviations were at places where either the actual record of the 

 convergence was not very certain, or where the mean position and course of the convergence are based 

 on inadequate data. Furthermore, it is known that sub-Antarctic surface water is sometimes thrust 

 a considerable distance south of the normal position of the convergence, and the extensive loops in 

 the convergence (seen in Figs. 2, 3) show that there is an element of chance in the latitude at which 

 it is found at an isolated crossing. This last factor may indeed account for many instances of apparent 

 displacement from the mean position. Probably a displacement of 50 miles or so either way is not 

 uncommon, but it may be that the extreme displacement does not exceed about 100 miles. 



