On Certain Bacteria from the Air of New York City. 337 



three different forms, differing onl^y in manner of growth and oc- 

 curring so sporadical!}', is much against tlie contamination 

 theory. Thus it becomes evident that this line of research is not 

 competent to positively exclude the possibility^ of contamination, 

 though it has rendered it a complicated explanation. It was 

 abandoned at this point. 



Now, on the other hand, in making plates of the soft form, the 

 difficulty of the cells possibly adhering in masses is reduced to a 

 minimum. A softer growth is scarcely to be imagined than that 

 of the normal B, lactis erythrogenes. In the hanging drop, the 

 cells are seen singly, or, rarely, in pairs or short chains. It is 

 therefore, highly improbable that these should form masses and 

 entangle a few of the Avrinkly form. The worst that might be ex- 

 pected would be that the soft and wrinkly forms should adhere in 

 approximately' equal numbers, and the resulting colonies could 

 not be mistaken for pure colonies of the soft form. Moreover, 

 whereas a slight growth of the soft form might escape detection 

 in the growth mass of the wrinkly form, it seems almost certain 

 that in the reverse condition the mixture would be easily de- 

 tected. Taking advantage of these conditions, the question was 

 approached from this side. An agar culture was made from a 

 soft colon}' occurring on the first " reversion plate " made from 

 the original wrinkl}- culture. It appeai'ed to be a pure culture of 

 normal B. lactis erythrogenes. A pair of third dilution plates 

 prepared from it exhibited about 140 colonels, all of the normal 

 soft form.* A second pair of plates exhibited 340 colonies, and 

 of these one had returned to the wrinkly form, and gave a typical 

 wrinkly growth on agar ; twelve exhibited a new variet}', not 

 previously met with, and the rest were of the normal soft form. 

 The new variety possessed a slighth' irregular surface, but was 

 soft to the needle. Its growth masses on agar were slightly 

 wrinkl}', with the outline fineh' marked just before the edge as if 

 milled, and altogether were intermediate between the wrinkly 

 form and the normal soft form, but nothing like a mixture of the 

 two. Thus it would seem to be established that these various 

 growths are discontinuous varieties of B. lactis erythrogenes^ 



* There occurred also ou these jjlates some colouies of Bacillui^ ramosus de- 

 rived from the sterilized water used iu making the plates. The water was 

 boiled just before using as an extra jjrecautiou, but the resistant spores of this 

 species were uot destroyed. 



