352 IVAN C. HALL AND LILLIAN J. ELLEFSON 



frequently due to sporulating organisms of little or no signifi- 

 cance from the sanitary viewpoint. These are mainly anaerobic 

 bacilli, but aerobic gas producing spore bearers also occur. 

 Both groups are inhibited by gentian violet in proportion to the 

 concentration, and, whereas B. coli is not inhibited by a much 

 greater concentration of this dye, its use for the selective inhibi- 

 tion of the former organism is urged. Tests were made at a 

 concentration of 1-100,000 and 1-20,000 in the presumptive 

 test and at 1-100,000 in a litmus lactose agar plating medium. 

 In all, 65 samples of water, from which B. coli had previously 

 been isolated, were tested; 54 were found to contain both sporu- 

 lating anaerobes and B. coli, 1 contained only B. coli and no 

 anaerobes, and 10 showed anaerobes only, B. coli having disap- 

 peared therefrom on standing. The tests showed that gentian 

 violet in the presumptive test not only does not interfere with 

 the isolation of B. coli but actually favors it according to the 

 statistical analysis of the results. In the 10 samples set down 

 as containing only anaerobes, presumptive tests were negative 

 in the dye test — positive in the standard test. A most striking 

 feature is the nearly complete inhibition of growth in the case 

 of heated samples tested in dye broth — these containing, with a 

 few exceptions, only sporulating organisms. 



Some positive tests were obtained in gentian violet lactose 

 broth from which B. coli could not be isolated; barring the 

 possibility that these were due to anaerobic Gram-negative 

 non-sporulating bacilli, it is believed they can be avoided com- 

 pletely by increasing the concentration of the dye. It may be 

 suggested, however, that if one part gentian violet in twenty 

 liters of lactose broth inhibits nearly 95 per cent of the spurious 

 presumptive tests, as inspection of the proportion (4 to 78) of 

 false tests obtained with and without gentian violet in the 

 heated samples shows, it would scarcely be worth while to 

 increase the concentration greatly for the extra 5 or 6 per cent. 

 Even if we analyze the figures in the least favorable manner, 

 i.e., to show the percentage of samples yielding B. coli from 

 positive presumptive tests in all our examinations we have the 

 following : 



