Report on the Bacteriology of Water. 337 



However, on comparing the action of a thick crystal of rock-salt 

 with that of water and alum, I was unable to detect any such marked 

 difference as would seem to follow if that conclusion were correct. 

 As will be seen more clearly later, when I come to discuss the results 

 obtained with the spectrum, the infra-red rays are themselves utterly 

 without perceptible effect.* 



VIII. 



The following series of experiments, designed to estimate the 

 degree of light action on water bacteria, was carried out, under my 

 direction and supervision, by Miss Hayward, of University College, 

 London, and I owe it to her to state that their successful carrying 

 out would have been impossible on account of the numerous plates to 

 be counted, in short periods, and involving very large numbers but 

 for her untiring industry and devotion to the work. 



Series I. 



About 150 c.c. of Thames water, collected at 10 A.M. on August 12, 

 were distributed equally in three Erlenmeyer flasks, properly steri- 

 lised, and the flasks labelled A, B,' and C ; and at 11 A.M. a 1-drop plate 

 was made from each flask. These plates, examined and counted on 

 August 14 at 11 A.M., gave respectively 1560, 1700, and 1080 colonies 

 per c.c. i.e., an average of 1446 colonies per c.c. developing in two 

 days. On keeping the plates another day, two of them gave 3705 

 and 1656 respectively, while the third was uncountable and liquefied, 

 the mean being 2080. 



We, therefore, assume that the water contained at the outset about 

 2700 bacteria per c.c., capable of developing in three days. 



The flasks were then placed as follows : A was suspended by the 

 neck so that it could be exposed to what sunshine there was, and at 

 the same time be illuminated from below by the light reflected from 

 a plane silvered mirror. 



B and C merely stood by the side of the stand supporting A, C 

 being covered with tin-foil and black paper, while B was exposed to 

 the light from above and at the sides. 



After standing thus from 12 noon to 4.30 P.M., the weather being 

 very cloudy with only occasional bursts of sunshine, fresh samples 

 were taken from each flask by means of sterilised pipettes, and new 

 plates made to see if any changes had occurred of note. 



Taking first the plate from A, after two days' incubation, we found 

 1599 colonies per c.c. ; and after a further twenty-four hours, 2760 

 per c.c., of which 12 per cent, were liquefying forms. The number 

 of living bacteria capable of developing in two to three days, there- 



* Since writing this the experiments with the spectrum have been published 

 separately (See ' Proc. Roy. Soc.,' 1894, vol. 54, p. 472, Abstract). 



