551 
contained tubercle bacilli.® The third investigation showed that 
or 5.5 per cent, of 36 samples of milk contained tubercle bacilli. 
The milk was obtained from 26 dairies, among which 2, or 7.7 per 
cent, were distributing infected milk.^ The fourth investigation 
showed that 17, or 16 per cent, of 107 samples of milk contained 
tubercle bacilli, and that among 8 samples of commercially pasteur- 
ized milk one was found that contained live tubercle bacilli.® 
The four investigations taken together show that among the 439 
samples of milk 36, or 8.2 per cent, were infected with live, virulent 
tubercle bacilli. 
The fact that one among eight commercially pasteurized samples of 
milk contained living tubercle bacilli is conclusive proof that some of 
the so-called “ pasteurization,” commercially practiced, is worse than 
useless, and has the evil tendency to quiet the mind regarding grave 
dangers it does not correct. 
It is a serious charge against the milk commonly sold by dairies to 
say that fully 1 sample among every 12 contains living, virulent 
tubercle bacilli, and yet this is the most favorable conclusion we can 
draw from four of the most recent and most reliable investigations 
with which the writer is acquainted. 
A further analysis of the two among the four milk investigations 
that give the number of dairies from which milk was tested proves 
that the conditions are worse than their superficial appearance indi- 
cates. These two investigations show that 17, or 6.5 per cent, of 259 
samples of milk obtained from 128 dairies were infected, and that the 
infected milk was sold by 13, or 10 per cent, of the dairies. The two 
investigations also show that the total number of samples of milk 
obtained from the 13 infected dairies is 31, of which 17 were infected 
and 14 were free from infection. Hence, the difference between the 
percentage of infected milk samples and the percentage of infected 
dairies can not be explained on the assumption that it is due to the 
more frequent duplication of tests with milk .from the noninfected 
than from the infected dairies ; it is shown on the face of the evidence 
that the difference of the two percentages is due to the fact that 
infected dairies distribute infected milk intermittently and not 
continuously. 
The intermittent distribution of infected milk by infected dairies is 
not only interesting because it may be related to the intermittent ex- 
pulsion of tubercle bacilli by cattle with their feces, but also because 
it justifies that we should draw the conclusion from the milk tests we 
® J. R. MoWer, same work as above, pp. 493-i95. 
* Unpublished work of the experiment station of the United States Bureau of 
Animal Industry. 
® Dr. Alfred Hess, of New York. Paper presented at the International Con- 
gress on Tuberculosis, Washington, D. C., 1908. 
