54 TYPHOID FEVER IN DISTRICT OF COLUMBIA. 
were in contact with cases outside of the District of Columbia, leav- 
ing 32 cases, or 4.28 per cent of the 866 cases, which were attributed 
to infection by direct contact. Among these 32 cases were 2 instances 
of colored servants contracting the disease from white employers, 
but no instances were determined in which white persons contracted 
the disease from colored servants. 
Fourteen cases gave a history of slight or occasional association 
with patients in the febrile stage. Of these, 5 were attributed to 
infection by milk, leaving 9 for consideration as suspected contact 
cases. In 21 cases indirect conveyance of the infection by persons 
or flies was strongly suggested. Twelve cases had free and intimate 
association with patients just convalescing from t}^hoid fever. One 
case developed the disease within thirty days after moving into a 
house which had been vacated by persons among whom there had 
been typhoid fever during the three months prior. 
It seems fair to estimate the chances of these 44 cases having become 
infected by contact at about 50 per cent; and so adding 22 of these 
to the 32 which were considered almost certainly infected by contact, 
there are 54 cases, or 6.23 per cent of the total, which may reason- 
ably be attributed to infection by contact. 
No satisfactory information could be obtained upon the question 
of direct or indirect contact with persons recently (three to four 
months) convalescent f rom the disease or wdth chronic bacillus-carriers. 
Therefore, it may be argued that more cases may have contracted 
their infection in this way; how many more is problematical. 
That the bulk of the infection in Washington is not spread by 
contact is indicated by the fact that the number of cases diminished 
through the latter half of August and through September, while all 
the conditions for the spread of the disease by this means seemed 
favorable; and, further, by the fact that the disease prevails to a 
much greater extent in Washington than in some other American 
cities in which the conditions for the spread of the disease by con- 
tact appear to be equally favorable. On the other hand, arguments 
may be adduced indicating that contacts play a larger role in the 
problem, but on account of our meager data we prefer to leave 
the subject for further study, which is now going on in the 
Hygienic Laboratory. 
The difficulty of determining just how many cases should be attrib- 
uted to infection by contact will, we believe, become evident to any- 
one studying the epidemiology of typhoid fever in a large city. To 
estimate in figures the actual number infected by contact among the 
cases which occurred in the District of Columbia during the period 
in which we were studying the situation would be mere speculation, 
and in attributing a given number of cases to infection by contact, 
milk, or any other one factor, we have taken into consideration all 
