534 H. J. CONN 



are averaged, there is very little choice between crystal violet, 

 methyl violet 6B, and gentian violet. A few more F grades 

 appear with crystal violet than with either of the others but not 

 enough to be of any significance. For our practical purposes 

 these results indicate that the Gram stain can be made with a 

 dye bearing any one of these three names, provided it comes 

 from a satisfactory source. 



Quite in contrast to the uniformity of these average results 

 is the variation in the reports from the individual investigators. 

 In spite of the standardized technic used in the work there is not 

 a single sample in regard to which entirely consistent results 

 have been made and practically every sample that has been 

 tested more than four or five times has been reported good or 

 excellent by some investigator and unsatisfactory by some other. 

 This suggests that either the personal equation is one that cannot 

 be eliminated entirely or else that some point in the technic 

 requires further standardization. The chief points which have 

 been mentioned so far as susceptible to further standardization 

 are: temperature at which the cultures are incubated, strength 

 of alcohol used for decolorization, length of time of decoloriza- 

 tion, and nature of the counter-stain used. It is very doubtful 

 whether variation could be avoided by controlling all these 

 factors but the committee hopes to plan a further investigation 

 along this Une simply to test out some of these points in regard 

 to the Gram technic. In a recent paper Burke (1922) has sug- 

 gested the importance of some of these factors, but it seems as 

 though a cooperative experiment might give more light on the 

 subject than any study made by a single observer. 



In looking up the results of the individual manufacturers of 

 these stains, it is to be noticed first that nearly every sample 

 which has ranked as high as G -|- or E — is one that has been 

 tested so few times in this work that the results are not especially 

 significant. For this reason it is regarded as impossible to de- 

 cide which is the best of the samples tested. It is perhaps of 

 some significance, however, to notice those few samples that 

 have ranked below G — in the general average. These samples 

 are crystal violet of Harmer, H. S. Laboratories, Leitz and 



