1896.] on the Past, Present and Future Water Supply of London. 67 



Comparing the curves in tlie diagram, it is seen that, with a 

 few exceptions, a remarkably close relation is maintained between 

 them. 



The only exception of any importance to the rule that the number 

 of microbes varies directly with the flow of the river, occurring 

 during the thirty-two months through which these observations were 

 continued, happened in November 1892, when the flow increased 

 from 501 millions of gallons in October to 1845 millions in 



Fig. 14. 



November, whilst the microbes actually diminished in number from 

 2216 to 1868 per cc. Neither the sunshine nor the temperature 

 records of these two months, however, afford any explanation of this 

 anomalous result : for there was a good deal of sunshine in October 

 before the collection of the sample, and the temperature was higher ; 

 whilst in November no ray of sunshine reached the Thames during 

 the three days preceding the taking of the sample, and the tempera- 

 ture was nearly 4° C. lower than in the preceding month. I have 



f2 



