334 On Certain Bacteria from the Air of New York Citij. 



was not a very strong i)ink tint in the medium. Other ao[ar cul- 

 tures repeated the characters. Colonies on agar plates were 

 moderately" brittle, the surface ones large, flat, and marked with 

 coarse granules centrally, and a rim-like margin. Morphology 

 normal. In short, this form corresponded in all respects with the 

 culture from which it was apparently derived, except in the pe- 

 culiar alteration in the consistence of the growth mass. But the 

 difference is one that Avould be usualh- considered specific* 

 Now the usual precautions to avoid contamination were observed 

 here and it is rather unlikely that this culture was a contamina- 

 tion, especially considering the improbability that a contamina- 

 tion should reproduce all the morphological and biological charac- 

 ters so exactl}'. Moreover, I have seen no such species as this 

 would have to be considered. In spite of these considerations, 

 I am unable to make the assertion that we have not to do with a 

 contamination, but the following experiment occurred to me to 

 test the matter further. I have shown above that there is a ten- 

 denc_> for a variety to return to the parent form. Therefore, if 

 this be a variet3', some of the colonies should show a return to 

 normal B. lactis erythrogenes. If, however, it be a different 

 species accidentally introduced into the original gelatin plate, we 

 should not expect to see any such result. Some agar plates wer^ 

 consequently made and on the third dilution there resulted one 

 hundred and twenty-five large surface colonies, besides many deep 

 ones. Of these one hundred and twenty -five, three showed the 

 smooth surface and soft texture of B. lactis erythrogenes. The 

 others were all alike, granular centrall}-, with a rim-like margin, 

 and liable to crack in two under the needle. Cultures from the 

 former resulted in a normal growth of B. lactis erythrogenes^ the 

 others reproduced the wrinkly growth. Now there are but two 

 alternatives ; either this is a discontinuous variation, which occurs 

 in this species, and which I have thus happened upon in two in- 

 stances, possessing a certain tendency to return to the type, or 

 there is a species closely allied to B. lactis erythrogenes, but dif- 

 fering in texture of growth, which shows a marked tendenc}' to con- 

 taminate cultures of this species (! ;, but is not otherwise com- 

 mon. In my last experiments this " species " must have entered 



* Dr. Prudden has called my attention to the fact that a very similar differ- 

 ence in growth form exists between the Bocilliis fubirciilosis and the 7?. tubercu- 

 losis gaUinarum. 



