156 
per cent. Several conditions no doubt influence the proportion of 
persons affected, the most important of which must be the amount 
of infection in the milk. In the Palo Alto epidemic if was deter- 
mined that the milk became infected through the water used for 
washing the cans and also at times for diluting the milk. This water 
was obtained from a creek which received the drainage from several 
houses in which 'there were patients with typhoid fever. The water 
of the creek for two or three weeks must have been quite heavily 
charged with typhoid bacilli, so that probably the majority of the 
milk cans washed in this water received some of the organisms. 
It is easy to understand how a milk supply thus almost if not quite 
continuously infected for several weeks may cause the infection of a 
large proportion of the persons who use that milk; but when the 
infection is introduced into the milk at irregular intervals for a like 
period, as would be expected when the infection is conveyed on the 
hands or clothing of persons or by flies, etc., a very small proportion 
of the -consumers of the milk may become infected. 
The susceptibility of the people supplied with infected milk, of 
course, would affect the proportion. In a community where typhoid 
fever had been prevalent for years, and in which there would be a 
number of persons rendered relatively immune by previous infection, 
we would expect less susceptibility than in a community where the 
disease had never prevailed. 
That it takes susceptibility plus exposure to infection for the 
disease to occur was well shown by an instance in the course of a milk 
outbreak in the District of Columbia in the fall of 1906. In a 
children’s home having about 100 inmates 7 children came down with 
typhoid fever within a period of two or three days. The way in 
which the milk was delivered to and served at the institution made it 
practically impossible for the 7 children affected to have drunk milk 
from any one can, or one day’s delivery, from which at least 75 per 
cent of the children did not drink. Thus of 75 children almost cer- 
tainly drinking infected milk only 7 had the disease. It is conceiv- 
able that in such an instance the typhoid bacilli in the can or cans of 
infected milk were not uniformly distributed through the bulk of the 
milk, but were in clumps so that only one or two of the children 
drinking from a 5-gallon can of milk actually received any of the 
bacilli ; but such a view is contrary to what is known by actual test 
about the distribution of bacteria such as the B. typhosus in a liquid 
such as milk, so it is much more reasonable to conclude that all of the 
75 children received some of the bacilli and the escape of the majority 
was due entirely to lack of susceptibility at the time the organisms 
were ingested. 
