New York Agricultural Experiment Station. 115 



The lack of even the wine color was taken as evidence that the 

 starch had not been attacked. This was the case with 80 per ct. 

 of the cultures. Of the remainder the starch was practically all 

 destroyed in 4 per ct. and 16 per ct. were classed as feeble. The 

 separation of the cases where starch is attacked into two classes is 

 unsatisfactory, since there was no available turning point between 

 those where the action was barely perceptible and where it was 

 practically complete. 



Among the observed reactions which were not enumerated in 

 Table I, the results with Gram stain are the most striking. With 

 the exception of a single culture, No. 28, all of the isolated cul- 

 tures were tested in this way and all but six gave positive re- 

 actions. The reaction of the single strain under No. 30 was 

 variable and that of the single strains representing No. 1 and 

 No. 2 was negative. One of the two strains under No. 7, one 

 of the six strains under No. 12 and one of the three strains under 

 No. 54 were also negative. 



This proportion of Gram positive strains, 96 per ct., is sur- 

 prisingly large and may be taken as one of the general character- 

 istics of the udder flora. 



Relative frequency of occurrence. — Some forms were observed 

 in only a series of samples from a single cow while others were 

 common in samples from various sources. In some cases the 

 bacteria in a given quarter were few and the flora mixed while in 

 others they were abundant and often consisted of practically a 

 single form. These conditions can perhaps be best illustrated 

 by tabulating the forms recognized in the case of the fifteen cows 

 from the Station herd. These data are given in Table X. 



L. B. 



H. 



15,39 

 39,23 



* The numbers given in this and the following tables are the culture num- 

 bers given in Table IX. 



