Report on the Bacteriology of Water. 379 



Quantitative Analysis of Thames Water in August, 1893 (Table Q). 



If we analyse these August numbers, we get as the average of 



14 two-days' plates, 1157; of 24 three-days' plates, we have 1530; of 



15 four-days' plates, we have 2208 ; of 8 five-days' plates, we get 

 3575 ; and of 8 six-days' plates, we have 3037, as the approximate 

 averages. Of course these numbers can only be regarded as approxi- 

 mations, but I think they are of use as indicating what might be 

 looked for if the necessary large series of observations could be made 

 over several years ; and they certainly serve to put us on our guard 

 against placing too much confidence in the gelatine method unless a 

 great number of observations are made, extending over a long period. 

 Of course the great difficulty to be contended against is that of 

 keeping the plates long enough ; if only one badly liquefying form is 

 present, it ruins the plates before the slowly developing ones germi- 

 nate out. 



Quantitative Analysis of Thames Water in October, 1893 (Table U). 



The following table gives the details of the analysis for October, 

 and again I have recorded all the points. The chief feature 01 

 interest is the much longer time during which the plates could be 

 cultivated, in spite of the prevalence of liquefying forms, where care 

 was taken to dilute sufficiently. 



VOL. LVI. 2 D 



