On Certain Bacteria from the Air of New York City. 335 



\\\\ nitrate culture, but not entirely overpowered the B. lactis 

 erythrogenes, so that both could appear in the plate in the pro- 

 portions described. 



To eliminate this last alternative, plates were again made, this 

 time from a growth taken from one of the rough surfaced colo- 

 nies just described. This culture has now been through the 

 usual process of purification, and ought not to be contaminated. 

 It exhibited the characteristic wrinkled surface of the form under 

 investigation. Now it is to be remarked that even if no rever- 

 sion were demonstrable, it would not prove that this form was 

 originalh' a contamination, for the original tendenc}' to reversion 

 being three out of one hundred and twent^'-five, I might have 

 practically' eliminated it by this selection. On the other hand, 

 if reversion occurred, it would be evidence in favor of this form 

 being a discontinuous variation. The plates were made with the 

 following result : One third-dilution plate gave about one hun- 

 dred and fifty surface colonies, all of the wrinkl}' form; a second 

 third-dilution plate gave about one hundred and fort}- colonies, 

 three of which were of the normal soft form. Here again the re- 

 version showed itself, but with a distinct tendencj' to diminution ; 

 three out of two hundred and ninety.* These wrinkly colonies 

 did not reproduce the form of colony described above ; they 

 were more folded and wrinkled with uneven edge ; but the growth 

 on agar tubes was not noticeably difterent. In a second "gen- 

 eration " agar culture made from one of these tube cultures when 

 it had become old, a narrow border of the soft growth appeared 

 around the edge, suggesting that either reversion had occurred 

 in the culture itself or that the latter was not pure. Conse- 

 quentlv plates were again prepared. The resulting colonies were 

 obtained only few on a plate (about fifty-five surface ones) so as 

 to give them plenty of room. Of these colonies none developed 

 the normal soft form, but when the}' had become old some ex- 

 hibited the bordering soft growth just as in the tube from Avhich 

 they were taken, but the border was present in varying degrees. 

 Four cultures were made from different appearing colonies as fol- 

 lows: 1, from a colony all wrinkly, resulting growth all of the 

 wrinkl}' form; 2, from a colony before definite characters had ap- 



* Notice the much less tendency to reversion in the discontinuous variation 

 than among the slight continuous ones. Is this a general character of these 

 classes of variation ? 



