THE ANTARCTIC CONVERGENCE 



189 



Table 4. The observed middle temperatures of the co?wergence gradient for month and latitude 



It is evident' that the average figures can be improved by some method of smoothmg, even though 

 the data are insufficient for an accurate estimate of the true averages for each month and latitude. 

 I have therefore adopted the rough and ready method of plotting the averages and drawmg curves 

 by eye to make as good a fit as possible. Figures showing the average for all latitudes in different 

 months, and for all months in different latitudes might be misleading, for there are insufficient winter 

 observations in the higher latitudes. In Fig. 6 A-D therefore the monthly averages are plotted separately 

 for each range of latitude and the curves are drawn both to represent the monthly trend of the points, 

 and to resemble each other, for it is assumed that the rate of change does not differ very much in 

 different latitudes. In Fig. 6 E the four curves are shown on the same scale and are marked with the 

 middle latitude of each range. Thus the second curve applies to all records between 50 and 54 S and 

 can be taken to represent 52° S. It will not be far wrong to say that those north of 50 represent 48 b 



and those south of 58° represent 60° S. 



Table 5. Provisional estimate of the average middle temperature of the convergence gradient 

 in different months and latitudes {figures smoothed) 



Table r is derived from Fig. 6, and it includes figures for interpolated latitudes and a slight extra- 

 poltion to 46 Id 6." S. The'steps leading up to this table have been explained ■" <i«f --'_;'/- 

 Le idea of the extent to which it can be rehed on. This can be judged from F,g. 6 A-D and the 



