Report on the Bacteriology of Water. 345 



The two plates from flask F 4 (not exposed), examined at the same 

 time, gave 10,800 and 7,200 per c.c. in two days. Total, 18,000 ; 

 mean, 9000 per c.c. ; and, in three days, they showed 13,200 and an 

 uncountable number, owing to liquefaction. 



These results were ddcidedly mystifying at first, for they showed 

 apparently a stimulating effect of exposure to the light ; bat on going 

 further into the matter it seems more probable that what really 

 happens is, that (1) the sunlight was not powerful enough in blue- 

 violet rays to produce any appreciable inhibition in the time ; and (2) 

 the flask F 4, covered in tin foil, &3., did not become warmed so rapidly 

 as the other, and consequently still showed the retarding action of 

 the icing to which it had been subjected. 



That this explanation is right is borne out by the behaviour of the 

 plates taken at 4 P.M. i.e., after four hours' further exposure of the 

 flasks for the covered flask, although enormously increased in 

 bacteria, was still behind the exposed one. 



After two days' incubation, the two plates from F 3 (exposed) at 

 this period gave 18-4,800 and 165,000 respectively. Total, 349,800 ; 

 mean, 179,900 per c.c. ; and, after three days, 231,000 and 207,900. 

 Total, 438,900 ; mean, 219,450 per c.c. 



Whereas the two plates from the non-exposed flask gave respectively 

 19,500 and 12,000 per c.c. ; total, 31,500 ; mean, 15,750 per c.c., after 

 two days' incubation, and were not counted further. 



Considering the enormous expenditure of time and trouble involved 

 in making and counting these plates, we were somewhat discouraged 

 by these negative results, which are summarised in the accompany- 

 ing Table J (p. 346, &c.). 



XII. 



On August 24th two flasks of Thames water labelled F 5 and F 6 

 were exposed exactly as the last, but the weather was fine and we 

 had much bright jsun and blue sky, with rapidly moving white. 

 clouds. 



The first exposure was from 9.30 A.M. to 4 P.M., and the arrange- 

 ment as before. The temperature, as indicated by thermometers in 

 control flasks, rose occasionally over 35 C, but was usually not above 

 30 C, and somewhat higher in the covered flask than in the un- 

 covered one. 



Four plates of the water with which the flasks were charged were 

 made at the time of starting the experiment, and were incubated as 

 long as possible. In four days they gave us 1980, 1650, 1320, and 

 2970 colonies per c.c. Total of the four samples, 7920 ; average, 1980 

 colonies per c.c. 



After six and a half hours' exposure, of which about four hours was 

 brilliant sunshine as far as we could estimate, two plates from each 



