Report on the Bacteriology of Water, 



2(55 



not only small doses of the liqnid to be tested, e.g., 1 or 3 drops from 

 a pipette discharging 25 drops to the 1 c.c, bnt also diluted the 

 liquid with sterile distilled water, e.g., 1 : 4, or 1 : 9, &c, facts which 

 we bring out duly in the tables. 



To those critics who would remark on the dangers of the " dilu- 

 tion method" above referred to, I would reply in two ways: (1) it 

 is the only practicable method available for getting over the difficulty 

 of plates so densely crowded with colonies that no attempt at count- 

 ing (or estimating) is possible ; and (2) it certainly does not lead to 

 exaggeration of the numbers of colonies, but in the contrary direction, 

 and, therefore, the final numbers obtained are more likely to be below 

 than above the truth. 



As a matter of experience, I am, in fact, more and more assured 

 that the whole procedure of gelatine plate cultivation leads to under- 

 estimation rather than to over-estimation, and this is obviously a 

 fault on the better side of exactness, since we have to be content with 

 approximations. Nevertheless, great care has to be taken in all 

 stages of manipulation, and the more so because it is always neces- 

 sary to multiply out the final results. 



The results of the daily examination of the non-infected flasks show 

 the usual rise to a maximum, and then fall in the numbers of normal 

 aquatic organisms existing in the Thames water. I now pass to 

 the results obtained with weak and strong anthrax respectively, in 

 the crude Thames water, i.e., without filtering or sterilising in any 

 way, Tables C (I) and C (II). 



