^ THE CURRENTS OF THE ATLANTIC. 405 



may call northern, central, and southern Europe, respectively he 

 finds that the mortality in the last is by much the greatest, while 

 it is nearly equal in the two other divisions. In this comparison 

 M. Quetelet has considered merely the effect of latitude upon 

 mortality. But as we have already shown that places in the 

 same parallel of latitude often have widely different temperatures, 

 it seems plain that the comparison should be made differently, and 

 that we should group together for the purpose places under the 

 same, or nearly the same, isothermal line. When we compare in 

 this manner the death-rates of the several countries of Europe 

 with their mean temperatures, we find that the mortality is greatest 

 in Italy and Greece, where the temperature is highest ; and that 

 it decreases as the temperature decreases, down to a certain limit, 

 after which it appears to increase again. 



But this is obviously an inadequate mode of considering the 

 problem. It seems plain that, for animals as well as for plants, 

 the salubrity of a climate will depend on the extremes of tempera- 

 ture, much more than on the mean ; and that, of these extremes, 

 the most influential on life will be that most removed from the 

 normal temperature which is best adapted to the species. Accord- 

 ingly, we should expect the mortality to be influenced chiefly by 

 the summer temperature in hot climates, and by the uinter tempera- 

 ture in cold ones. Arid this I find to be the case in Europe, where 

 alone we have the data requisite for the comparison. 



The following are the principal conclusions which we seem to 

 be warranted in drawing : 



1. In the southern half of Europe the mortality depends upon 

 the temperature of summer, being greatest where that temperature 

 is greatest, and diminishing with it down to a certain limit. 



2. In the northern half of Europe, on the contrary, the mor- 

 tality depends on the temperature of winter, being greatest when 

 that is least. 



3. The boundary line between these two regions is not far from 

 the mean yearly isothermal of 50, which is accordingly the line 

 of least relative mortality. 



4. The mortality attains an absolute minimum in the British 

 islands, at the western extremity of this line, the annual range of 

 temperature being there least. 



These conclusions will be evident on the inspection of the fol- 

 lowing Table, in which the mortality is compared with the 



