Sterilisation Experiments on the Cambridge Water. 567 



quantities of chlorine, say from one part in a million of water to 

 one part in eight millions, may thus be pumped into the water 

 entering the chlorinating vessel. Here the two are left in contact 

 as they pass through the cylinder, the chlorine can then, if 

 necessary, be wholl}^ or partly neutralised by the bisulphite of 

 soda solution which is thrown into the water as it leaves the 

 cylinder. The water as it leaves the cylinder flows into (4) a 

 galvanised iron tank over a slotted weir so graduated that the 

 amount of water passing over it at any given time may be easily 

 measured. All samples of the treated water were collected as it 

 passed over this weir. With this apparatus, capable of treating 

 84,000 gallons per day of 12 hours, sufficient for a community 

 of from 2000 to 2500 people or, if running for 24 hours for 5000 

 people, a series of experiments were made. At first there was 

 some difficulty in obtaining a steady working of the pumps, then 

 certain difficulties arose in connection with the sterility of the 

 bisulphite solution and again with the balancing or neutralising 

 of the chlorine by the bisulphite owing to the marked instability 

 of the latter salt which seemed to become oxidised very rapidly 

 indeed. Ultimately these difficulties were overcome and out of 

 six experiments made between the 8th December 1909 and the 

 6th January 1910 samples varying from 110 to 150 c.c. of treated 

 water collected at the weir were found to contain not a single 

 Bacillus coli or any of its congeners. In all these experiments one 

 part of chlorine per million of water was used and, to get rid of 

 the taste of chlorine, it was always necessary to neutralise by 

 means of bisulphite of soda. When complete neutralisation was 

 obtained neither taste nor smell of chlorine remained. The amount 

 of chlorine was then halved and the amount of water tested for the 

 presence of coli increased. 800 c.c. of the treated water taken at 

 the weir contained not a single Bacillus coli or any organism 

 resembling it. Between the 10th January and the 2nd February 

 1910 fourteen further experiments were carried out, the amount 

 of chlorine per million of water being again halved, only one part 

 of available chlorine being added to four million parts of water. 

 Here, again, larger quantities of water, in no case less than half 

 a litre, were tested, and in every instance it was found that the 

 organisms of the Bacillus coli group had been destroyed. The 

 method of testing, though simple, was very efficient. A strong 

 solution of McConkey's bile salts glucose litmus medium was 

 prepared and a small quantity placed in a litre flask. This, 

 plugged with cotton-wool, was thoroughly sterilised and taken 

 out to the Fulbourn pumping station where the experiments were 

 carried out. At the weir the cotton-wool plug was removed and 

 half a litre of the treated water was allowed to flow directly into 

 the flask ; the lip of this flask had, of course, been carefully pro- 



