ON THE INFLUENCE OF WEATHER UPON MORTALITY. 333 



the mortality,; but above 50° the relationship becomes direct, the death-rate in- 

 creasing with the temperature. The months during which the latter condition is 

 observable, are probably July and August ; but in Scotland the mean temperature 

 does not often rise so high as to render it a cause of alarm. 



2. Over the whole year the relationship between the monthly range of tem- 

 perature and the death-rate is inverse ; but during the months of July, August, 

 and September, it is direct. A similar relationship exists between the mean 

 daily range of temperature and the death-rate. 



3. It is probable that dry cold is more fatal than humid cold. 



4. Extremes of temperature are always fatal, but eminently so when long 

 sustained. 



B. Atmospheric Pressure. 



1. The results afforded by comparison of the relative height of the barometer 

 with the death-rate from all causes are conflicting, but there is probably a pre- 

 ponderance in favour of the supposition that the relationship between the two 

 is inverse in the colder, and direct in the warmer months. 



2. The relationship between the monthly range of the barometer and the 

 death-rate is direct. 



C. Drought and Humidity. 



The relationship existing between humidity (irrespective of temperature) and 

 mortality appears to be direct in the colder, and inverse in the warmer months. 



D. Winds. 



Winds blowing from a point between N.W. and S.E. (north about) attend a 

 high death-rate. Winds blowing from a point between S.E. and W. (south about) 

 occur more frequently during months in which the mortality from all causes 

 is low. 



If in the foregoing tables the diflPerence between the mortality with this and with 

 that kind of weather be often apparently trifling, it must be remembered that one or 

 two deaths, more or less, per 100,000 living, is a matter for serious consideration. 

 It must be remembered also, that in the mortality from all causes, there are many 

 compensating influences at work which tend to reduce and soften what avouIcI 

 otherwise appear as remarkably prominent features in the inquiry. As a single 

 example of this, it is plain that if deaths from diarrhoea were excluded, the death- 

 rate from all causes would be relatively higher in cold weather than it wow 

 appears to be, for deaths from diarrhoea being low in cold and high in warm 

 weather, tend to equalise the surface over the whole rate of mortality. 



VOL. XXIII. PART II. 4 Y 



