9 
5. The inoculation of guinea pigs shows that in the course of non- 
miliary tuberculosis, even when it lias assumed a rapid form, febrile 
or hemorrhagic, the bacillemia does not exist. The only case which we 
know of (Ettlinger) perhaps may be explained by the existence of a 
short premiliary stage terminating abruptly in this observation by a 
hyperacute peritonitis. 
6. The results by the use of the inoscopic method of Jousset and 
the method of homogenization are not to be admitted without great 
reserve. These methods, in order to preserve their value, ought to be 
followed with a perfect asepsis very difficult to realize. 
7. The most grave causes of error which are attached to the 
methods recently proposed for the direct finding of tubercle bacilli 
in the blood reside in the fact of the existence in our surroundings : 
a. Of acid-resistant bacilli, well studied by Koch, Petri, Moeller, etc. 
b. Of common saprophytic bacteria, which have acquired in the 
organic media where they are accidentally found, an acid-resistant 
power more or less accentuated. 
TECHNIQUE EMPLOYED IN THE COLLECTION AND EXAMINATION OF 
THE BLOOD. 
In order to endeavor to find out why the results of the older 
observers differed from those of Rosenberger, I undertook a study of 
tubercle bacilli in the blood. 
My studies have been made on 48 cases of human tuberculosis and 
13 guinea pigs and 8 rabblits experimentally infected. The study 
consisted of the examination of smears, cultures, and the inoculation 
of blood into fresh guinea pigs. With a few exceptions, which are 
noted in the protocols, the following procedure was employed: 
In the human cases the blood was drawn from the vein at the bend 
of the elbow. Usually 10 or 20 c. c. of blood were taken and mixed 
with sufficient sodium citrate solution to give approximately a 1 per 
cent solution of the citrate, and thoroughly shaken. As soon as the 
blood was received at the laboratory it was divided into two lots. 
One lot was at once centrifuged for one hour at high speed. Two 
smears were made from the sediment, using in every case drawn-out 
pipettes that had never been used. Five drops of the sediment were 
then planted in a fermentation tube of lactose-bouillon and 5 drops 
on glycerin-potato; 2.5 c. c. of the sediment was next injected into 
the peritoneal cavity of a young guinea pig. 
The cultures were placed in the incubator at 37° C. ; the fermenta- 
tion tube was examined for growth in twenty-four, forty-eight, and 
seventy-two hours, and discarded after that time. The potato cul- 
tures were sealed with paraffin and kept under observation for two 
months. The guinea pigs were weighed at intervals of a week. 
