474 BACILLI IN CHRONIC INFECTIOUS DISEASES. 
resisting power against heat than the bacilli themselves, would there- 
fore simply be an admission that soma bacteriologists had made a 
mistaken inference based upon incomplete data. In view of the 
facts stated we can simply repeat what was said at the outset, viz., 
the question as to spore formation has not been definitely deter- 
mined. 
The tubercle bacillus is a strict parasite, and its biological char- 
acters are such that it could scarcely find natural conditions, outside 
of the bodies of living animals, favorable for its multiplication. It 
therefore does not grow as a saprophyte under ordinary circum- 
stances. But it has been noted by Roux and Nocard that when it 
has been cultivated for a time in artificial media containing glycerin 
it may grow in a plain bouillon of veal or chicken, in which media it 
fails to develop when introduced directly from a culture originating 
from the body of an infected animal. This would indicate the pos- 
sibility of its acquiring the ability to grow as a saprophyte ; and we 
can scarcely doubt that at some time in the past it was a true sapro- 
phyte. The experiments of Nuttall indicate that the bacillus may 
multiply, under favorable temperature conditions, in tuberculous 
sputum outside of the body. And it is extremely probable that mul- 
tiplication occurs in the muco-purulent secretion which accumulates 
in pulmonary cavities in phthisical patients. In these cavities its de- 
velopment may, in a certain sense, be regarded as saprophytic, as it 
feeds upon non-living organic material. 
Koch first succeeded in cultivating this bacillus upon coagulated 
blood serum, prepared as directed in Section VIII., Part First, of the 
present volume. Roux and Nocard have since shown (1888) that it 
grows very well on nutrient agar to which glycerin has been added 
(six to eight per cent), and also in veal broth containing five per cent 
of glycerin. It is difficult to obtain pure cultures from tuberculous 
sputum, on account of the presence of other bacteria which grow 
much more rapidly and take full possession of the medium before the 
tubercle bacillus has had time to form visible colonies. For this rea- 
son it is best to first inoculate a guinea-pig with the tuberculous spu- 
tum and to obtain cultures from it after tuberculous infection has 
fully developed. The inoculated animals usually die at the end of 
three or four weeks. It is best to kill one which gives evidence of 
being tuberculous, and to remove one or more nodules from the 
lungs through an opening made in the chest walls. The greatest 
care will be required to prevent contamination by other common 
microérganisms. The instruments used must be sterilized by heat, 
and the skin over the anterior thoracic wall carefully turned back ; 
then, after again sterilizing knives and scissors, cut an opening into 
the chest cavity, draw out the root of the lung, and take up with 
