314 



DR R. E. SCORE SBY- JACKSON 



logical agent at any given time, that at least the character of the weather imme- 

 diately preceding the period under examination should be ascertained, the influence 

 of sustained heat or cold, for example, as will be shown hereafter, being remark- 

 able. Nevertheless, when the period of comparison extends over an entire month, 

 I think we may safely believe that the fluctuations in the death-rate — in so far as 

 they are dependent upon weather at all — are dependent upon the weather of the 

 particular month in which the deaths are recorded. This is not uniformly so 

 Suppose, for example, that the fourth week of January were intensely cold, and 

 the subsequent month of February comparatively mild, it is quite possible that a 

 large number of deaths, caused by the cold of January, might fall to be registered 

 in February ; so that if either of the months were examined separately, an erro- 

 neous impression concerning the influence of temperature would result. If the 

 periods of observation had been daily or weekly, I would on no account have 

 separated them, because necessarily the ratio of deatlis of one day, or even of one 

 week, must often be, to a certain extent, modified by the weather of the previous 

 days or week; but, I repeat, where each subject of comparison extends over a 

 whole month, errors from such a cause as I have now explained must be very 

 trifling. It is, however, to obviate such errors that I have constructed Table B, 

 in which both the meteorological and necrological data are arranged appositely in 

 the order of the sequence of the months of the several years. This explanation is 

 also offered to objections which may be raised against the order of the months 

 in Table A.] I need not mention the years from which the several months are 

 taken ; that will be seen on reference to the larger table : — 



Months of Lowest Mean Temperature, 



„ Highest, „ 



Mean of the Six Years, . . 



January. 



February. 



35-5 280-8;340 

 39-6!243-4:405 



3306 



March. 



37-81256-7 



246-94302320 



April. 



41-3222-7 

 45-42230 



May. 



June. 



491 

 51-9 



225-4 

 193-5 



37-4 265-3 38-2 257-4 39-8|249-8 43-2 242-8 50-3 219-5 



52-41215-0 

 58-9j219-3 

 55-9208-6 



Months of Lowest Mean Temperature, 



„ Highest, „ 



Mean of the Six Years, . . . 



July. 



August. 



53-8il92-3 54-4192-4 

 59-0 181-3 60-0224-1 



September. 



50-2 

 56-1 



56-8 204-5 57-2;189-4 52-5 



182-5 

 210-8 

 187-7 



October. 



44-9|205.3 

 49-6208-0 

 47-2198-2 



November. 



37-1234-7 

 43-7231-3 



December. 



34-0 

 44-9 



263-4 

 215-7 



39-5237-138-8247-9 



If this table speak truly, it leads to the conclusion, that for every diminution 

 of mean temperature below 50° there is a corresponding increase of mortality ; 

 but that from mean temperatures above 50° a diminution is favourable to vitality, 

 at least if the temperature have been for any length of time above 50°. In 

 other words, mean temperature and mortality from all causes have an inverse 

 relationship below 50°, a direct relationship above 50°. But it must be borne in 



