552 
have under consideration that a larger proportion of dairies than even 
10 per cent must be classed as infected. 
To obtain further information regarding the intermittent distribu- 
tion of tuberculous milk by infected dairies.® samples of milk were 
bought from a dairy, from which several months previously a sample 
of milk had been obtained that was found to be infected with virulent 
tubercle bacilli, on 30 different days and injected into guinea pigs. 
Among the 30 samples the second, third, and eighth were found to 
contain tubercle bacilli; the remaining 27 were not infected. If we 
add the sample of milk which first showed the infected character of 
the dairy to the 30 later samples, we have 31 from one source among 
which 4, or about 13 per cent, were found to contain tubercle bacilli. 
It does not require much reasoning to conclude from this evidence 
that the chances for discovering an infected dairy by testing one 
sample of milk from it may be equal to only 13 per cent, and the 
chances that the one test will not reveal the infected character of a 
dairy may be nearly eight times as great as the chances that it will. 
I do not wish to create an exaggerated idea of the proportion of 
dairies that intermittently distribute tubercle bacilli in milk, because 
the facts are so grave that, without exaggeration, they are almost be- 
yond belief. It is well, however, to know the truth, and through 
knowing it. to be convinced that the milk of no dairy can be accepted 
as permanently free from tubercle bacilli unless it is obtained in a 
clean, wholesome environment from cows shown by the application of 
the tuberculin test to be free from tuberculosis. 
The available data regarding the frequency with which tubercle 
bacilli occur in butter and other dairy products than milk are very 
meager for the United States, but when we know that tubercle bacilli 
in milk are transferred to the cream, butter, cheese, etc., made from 
it. we can readily infer how commonly these products are infected. 
Relative to the infection of cream and butter the following para- 
graph from a report of the United States Secretary of Agriculture 
is very significant : 6 
The examination of sediment taken from cream separators of public cream- 
eries throughout the country has demonstrated the presence of tubercle bacilli 
in about one-fourth of the samples. 
In a recent publication of the United States Bureau of Animal 
Industry it was pointed out that both the tendency of tubercle bacilli 
to rise with cream and a comparison of European statistics relative 
to the frequency with which tubercle- bacilli have been detected, 
respectively, in milk and butter indicate that when tubercle bacilli 
a Unpublished work of the experiment station of tbe United States Bureau of 
Animal Industry. 
* Annual Report of the Secretary of Agriculture, Washington, D. C.. 1907, 
p. 30. 
