PART I.] Monthly Soil-temperature in Relation to Cholera-prevalence. 231 



be taken into account. This, however, is not all that is required to secure a determi- 

 nation of the relations borne by the water-level in one year to that in another ; for it is 

 evident that the variation in the amount of loss by evaporation must importantly modify 

 the effect of the addition by rainfall. Even if the data of temperature and atmospheric 

 humidity be taken into consideration along with those of rainfall, only very unsatis- 

 factory results are obtained, compared with those furnished by direct observation. That 

 this is the case is very distinctly shown by the preceding table (XXXIV), showing the 

 particulars of rainfall, temperature, and humidity, from 1870 to 1876, arranged, as far as 

 possible, in a way to facilitate their application to questions of water-level. 



(f) Soil-temperature. 

 TABLE XXXV. 



Comparison of average Monthly Soil-teTnperature (3 years) and Cholera-prevalence. 



The diagram, in addition to the lines indicating soil-temperature and cholera- 

 prevalence, contains a third line of the average atmospheric temperature, in order to 

 allow of ready comparison of the relations of the air above and within the soil in this 

 respect. 



Soil and Atmospheric Temperature , 

 curves [average of 3 years]. 



= Soil Temperature curve. 

 = Air Temperature curve. 



Total monthly Cholera Deaths [;i8 

 years]. 



Diagram 8.— Average Soil-temperature at 6 feet below the surface, and Cholera-prevalence. 



The data regarding temperature are those of the soil at a depth of six feet from 

 the surface, and therefore represent the conditions of a stratum towards the lower portion 



