ESTIMATION OF B. COLI 141 



occasionally reproduce so rapidly in the lower dilutions as to 

 prevent the growth of the coliform bacteria and so give a 

 negative gas test when a much higher dilution of the same 

 sample shows copious gas formation. 



The usual amounts of lactose and peptone employed in the 

 fermentation test are 1 per cent of each, but Chamot and Sher- 

 wood ^ have shown that a lactose content of 0.6 per cent pro- 

 duces equally satisfactory results as 1.0 per cent. Under 0.6 

 per cent the results were irregular and the total volume of gas 

 small, whilst quantities much exceeding 1.0 per cent retarded 

 the rate of gas formation. With normal acidities they found 

 that the total gas volume was proportional to the concentration 

 of the nitrogen whether present as peptone, beef extract or 

 infusion. With increasing amoimts of peptone the increase 

 in gas volume was rapid until 4.0 per cent was reached and when 

 both final volume and rate of production were considered, 

 it was found that a concentration of 3.0 to 4.0 per cent was the 

 optimum. Potassiimi chloride (0.6 per cent) hastened gas 

 formation and was found superior to phosphates and other 

 salts. The concentrations finally recommended were lactose 

 0.8 per cent, peptone 3 to 4 per cent, KCl 0.6 per cent, and the 

 reaction +1-0 per cent. With lactose bile the nitrogen content 

 should be sufficient with the addition of only 1.0 per cent of 

 peptone, but in other media the higher amount should be em- 

 ployed. For the concentration method the author uses ordinary 

 lactose broth or lactose bile salt broth in preference to lactose 

 bile on account of the irregularities often fovmd with lactose 

 bile and due to the variations in composition. 



The number of tubes to be employed in order to obtain 

 reasonably precise results is the second point for consideration. 

 It has been usual to use such dilutions of milk that the quan- 

 tities represent decimal fractions of 1 c.cm. and to endeavour 

 to obtain at least one positive and one negative result. Al- 

 though, in many instances, no attempt has been made to con- 

 vert such positive and negative findings into mathematical 

 expressions, others have attempted to do so by taking the 



