Report on the Bacteriology of Water. 341 



we note that 

 cation would normally occur. 



much greater if we note that during the period an enormous multipli- 



X. 



On August 15th two Erlenmeyer flasks were charged as before 

 with Thames water, and labelled A and C. A was exposed over 

 mirrors, and C wrapped up. We introduced the difference here of 

 having an additional mirror behind the flask, as well as that below. 



The day was bright, with plenty of sunshine all the time, and A 

 was exposed for 5| hours from 10.30 a.m. to 4 p.m. and then, 

 samples taken from both. 



Meanwhile, the average of seven samples taken at the commence- 

 ment gave 1644 as the number per 1 c.c. in the Thames water at 

 starting. 



Unfortunately, the temperature rose during the next twenty-four 

 hours sufficiently to soften the gelatine of the plates taken after the 

 first 5^ hours, so we could not count these. 



On August 16lh another bright, clear day the flask A was again 

 exposed for six hours, and C (wrapped up) beside it, both flasks 

 having stood all night (nearly 12 hours), at 18 C in a cupboard in 

 the laboratory. 



After this second exposure the plates gave for A about 6000 per 

 c.c., and for C over 174,000 per c.c. ; showing that the exposure to the 

 light had kept down the numbers in A, in spite of the interval of 

 twelve hours in a warm, dark cupboard, when, of course, the bacteria 

 npt killed off by the first day's exposure multiplied rapidly. 



Here, again, I was struck by the diminution of the liquefaction 

 on the plates from the flask A; it did not look like merely fewer 

 liquefying forms, but as if those that were present really liquefied 

 less rapidly and less efficiently than those from the flask not exposed 

 to light. 



XL 



On August 22nd two Erlenmeyer flasks, marked F 3 and F 4, were 

 charged to a depth of 1 in. with Thames water, properly collected, 

 &c. (see Table J). 



Flask F 3 was exposed to the sun with a mirror below ; F 4 stood, 

 covered, by its side. The exposure lasted from 11.30 A.M. to 4.30 P.M. 

 (being five hours), but only about one and a half to two hours at 

 most was good sunlight, the day being showery with snatches of blue 

 sky at intervals. 



Four samples taken at the beginning of the experiment gave 800, 

 1408, 748, and 1462 per c.c. as the numbers after two days' incubation. 

 Total of four plates = 4418 ; average = 1104 colonies per c.c. to start 

 with. 



