48 Report S.A.A. Advancement of Science. 



water issuing from the lower point is subject to contamination 

 (obvious or otherwise) from animal sources, its chlorine-content will 

 be variable and a wrong interpretation might therefore be placed on 

 an increased chlorine content of the lower water. An inconvenience 

 of the method is therefore that a series of determinations, extending 

 over some days, of the chlorine in the water, must be made before 

 the trial. On the whole, therefore, sodium chloride cannot be 

 regarded as suitable for the purpose in view. 



A lithium salt has been used for the purpose, and we might 

 expect that it would be safer, because more characteristic, in its 

 indications. I can see little objection to the use of lithium salts, 

 although I believe that fluorescein will usually be found better. It 

 is conceivable that the case might arise where coloration of the 

 water was prohibited, either for aesthetic reasons or to avoid arousing 

 suspicion, and in such instances, I think recourse might be had to a 

 lithium salt. In passing, it may be noticed that lithium chloride, 

 when present to the extent necessary for its easy recognition, is not 

 poisonous. Lithium chloride was used by Frankland, in determining 

 the source of the contamination of a Dublin water-supply, in an 

 examination, which has since become classical. 



Potassium, or sodium, permanganate has been suggested as a 

 material to use : but as most natural waters exert some reducing 

 power on permanganate, and as organic matter with which water 

 might come into contact in the underground passage, would reduce 

 the permanganate, this substance could only be of practical value 

 where the underground channel is very short and fairly free from 

 debris. 



Other materials, which impart a distinctive colour to the water, 

 have been used : I refer to the aniline dyes, and it is to one of 

 these — fluorescein — that I desire here to make more particular 

 reference. 



Before discussing this matter, reference may be made to one 

 other group of substances, which might be made use of in tracing 

 water connections, — organisms or bacteria, not normally present in 

 either of the two waters under investigation. The bacteria used 

 must be non-pathogenic, and for ready recognition a pigmenting 

 organism such as B. frodigiosus would be a convenience. If the 

 connection between the tw^o points were by a fairly clear channel, 

 this method might be successfully employed, but if the water is 

 passing through debris in the channel, this is liable to become an 

 efficient bacterial filter for the time beinf^ and thus prevent the 

 passage of the organisms. Furthermore, the possible effect of sedi- 

 mentation, and the life of the bacteria in the water, would require 

 to be taken into consideration, and finally, Busch {Centralblaft f. 

 Bakteriol. II., vol. xvi., 119, 1906) has shown that the recovery 

 of a bacillus, even in an open stream, at a point a mile or two 

 below that at which the organism was added, is not always a marter 

 of ease. In the tracing of underground waters, I do not think that 

 great reliance can be placed on bacteria. 



