190 BAENETT COHEN 



The distilled water of the laboratory, from a gas-heated Stokes 

 automatic still, was used in all cases except where otherwise 

 mentioned. In experiments requiring water of exceptional purity 

 the latter was prepared by double distillation out of acid and 

 alkali in Pyrex glass, with precautions for the exclusion of car- 

 bon dioxide. 



The tap water used was Potomac River water which had pre- 

 sumably been treated with alum and filtered through sand. 

 Ficker (1898) has shown that an "oligodynamic" property is 

 readily acquired by water which is allowed to remain in contact 

 for any long time with the metallic fixtures in ordinary plumbing. 

 To prevent this occurrence in our water, the latter was allowed 

 to run freely for several hours before being used for experiment. 

 During filtration, a by-pass provided a constant flow of fresh 

 tap water under a head of one meter. 



The buffer solutions were prepared with the precautions de- 

 scribed by Clark (1920)* The only divergence made here was 

 to dilute these buffers to a concentration of M/500. The pH 

 values of these dilutions were then determined colorimetrically. 

 Tests were made to determine the minimal amount of buffer 

 necessary to maintain constant hydrogen ion concentrations 

 under the conditions of these experiments; and it was found that 

 the M/500 concentration of buffer answered this purpose most 

 satisfactorily. 



Tlie bottles in which the bacterial viability was studied were of 

 approximately 1000 cc. capacity, of ordinary glass and with 

 ground stoppers. Before use, they were thoroughly cleansed 

 with fresh chromic acid cleaning mixture, rinsed, steamed in an 

 Arnold sterilizer for several hours, and well rinsed with distilled 

 water. 



Except for the experiments with double-distilled water, in 

 which cases the containers received an internal coating of puri- 

 fied paraffin, no attempt was made to prevent possible solution 

 of the glass in the contained water. This procedure was followed 

 because in preliminary experiments a comparison of the \'iabiUty 

 curves from water in Jena glass containers and in the above 

 softer glass bottles showed no appreciable differences. A\Tien 



