568 Prof. Woodhead, The results of 



tected from dust by a cotton- wool plug covered with tough paper 

 that would stand sterilisation by heat. Sometimes three or four 

 of these large flasks were used during a single day, whilst smaller 

 quantities of the treated water — 50 and 100 c.c. — were tested in a 

 similar fashion especially during the earlier experiments ; latterly, 

 however, when it was found that the sterilising process was so 

 complete, only the larger quantities were used. These flasks were 

 then incubated for forty-eight hours ; if the Bacillus coli was 

 present acid was formed, the litmus turning red and gas bubbles 

 making their appearance at the surface. Where there is any doubt 

 as to the formation of gas this can readily be determined by 

 plunging a warm platinum wire into the medium ; if there is no gas 

 no bubbles make their appearance, whilst if gas is being formed 

 a large number of small bubbles rise to tlie surface at the margin 

 of the fluid. This test, known as the " presumptive coli test " is 

 admirable where it is desired to prove a negative, but by itself it 

 is not sufficient to indicate the presence of the true Bacillus coli ; 

 for our purposes, however, it was all that was necessary, and as 

 we had an enormous number of samples to test we accepted the 

 absence of acid and gas as a proof of the absence not only of the 

 true Bacillus coli but of its congeners. Even where the Bacillus 

 coli had been completely eliminated a few colonies of spore-bearing 

 organisms sometimes made their appearance on agar- and gelatin- 

 plates; these however were present in exceedingly small numbers 

 and may be disregarded; they are not anthrax bacilli nor are they 

 Bacillus enteritidis sporogenes, an organism we have never found 

 in the Cambridge water. 



The above experiments seldom extended over more than four 

 or five hours each and were devoted mainly to getting the ap- 

 paratus " tuned " and although such comparatively good results 

 were obtained the experiments during this period were not looked 

 upon as being either crucial or final. Only on one occasion — 

 when the pumps were being regulated — was any sample found to 

 contain a coliform organism in 50 c.c. of treated water. 



In order to obtain prolonged tests under actual working con- 

 ditions a second series of experiments was carried out. This series 

 extended over 15 days, the pumping engine running from 5.30 a.m. 

 to 5.30 p.m. each day. During this period half-hourly observations 

 were taken and on certain days samples were sent to Professor 

 Frankland, Dr Thresh, Dr Houston and Dr Otto Hehner. Here 

 again on the first and second days of this series we were occupied 

 in regulating the pumps and testing the process of neutralisation 

 of the chlorine by the bisulphite of soda solution and the results 

 obtained were somewhat irregular. At certain periods complete 

 sterilisation of the water was attained whilst at others, especially 

 when the pumps were being regulated and we had an excess of 



