94 STORRS AGRICULTURAL EXPERIMENT STATION. 
all air bacteria the average was only 250 per cubic centimeter, 
in 20 case rising above 850. ‘These numbers agree well with 
those reported by Van Slyke at the Geneva Experiment Sta- 
tion (12th An. Rep. N. Y. Exp. Sta., p. 184, 1902). 
It will be seen that, instead of concluding that the milk ducts 
furnish bacteria by thousands per cubic centimeter of milk, our 
experiments indicate that the uncontaminated milk contains 
only small numbers. From our experiments the chief contam- 
ination of milk would seem to be external; from those of Har- 
rison and Cumming it would seem to be the milk ducts. 
Species of bacteria.—'The most striking difference, however, 
between the results that we have obtained and those of Harri- 
son and Cumming isin the species of bacteria. It should be 
stated that our studies have involved individual tests of about 
one hundred cows and several scores of experiments upon the 
mixed milk of a herd of thirty cows. Moreover, they have 
been carried out both at Middletown and at Storrs, places about 
thirty miles apart. The results, therefore, are not isolated 
and dependent upon a small amount of data, but have been 
confirmed by some hundreds of experiments. It may be fur- 
ther stated that these hundreds of experiments practically a// 
confirm each other, and whereas there are some variations in 
the percentage of lactic bacteria in milk freshly drawn, the 
numbers are always small and never approximate the conditions 
described in Harrison’s experiments. 
The most striking fact of our analyses compared with those 
of Harrison and Cumming is the small numbers of the lactic bac- 
teria. Whereas Harrison found 95 per cent. or so of the bac- 
teria belonging to the lactic types, we have found the numbers 
far smaller. The highest numbers were 70 per cent., and this 
only in two cases. Commonly there are less than 50 per cent., 
and in many cases below 30 per cent. 
When a closer study is made of these lactic organisms, we 
find that in practically no‘instance do they consist of the three 
types which Harrison finds. We occasionally find B. lactzs 
aerogenes, B. colt, and some other closely allied forms, but 
practically never either B. lactis acidi I. or IT, This result has — 

