jan. 2,1917 Diagnosis of Tuberculosis by Complement Fixation 
15 
antigenic and the anticomplementary properties was of such short range 
that the reliability of the tests with this antigen had to be questioned. 
The same experience was obtained with an antigen prepared from de¬ 
fatted organisms. The only advantage of this antigen as compared with 
the preceding one was noted to be in the shorter time in which the reac¬ 
tion took place, and in some instances it appeared more reliable. The 
precipitation method of Stimson (35) was also tried, but without much 
success. It was the intention also to employ the combined Besredka- 
Stimson method, as advocated by Craig (14), but this was discarded 
when a communication was received from Dr. Craig in which he stated 
that the antigen devised by him might be improved. 
An antigen prepared from an unheated liquid medium and Berkefeld- 
filtered proved to be of no value. The same in combination with the 
bacteria gave results similar to the antigen prepared from tubercle bacil¬ 
lary emulsion and from defatted organisms; and when reinforced with 
lipoids, the reactions obtained were much more distinct. At the same 
time, however, many nonspecific reactions were obtained. According to 
Calmette (10), the serum of a tuberculous individual contains more 
lecithin than that of a healthy person. Whether this observation has any 
bearing on the more definite reaction obtained with antigens containing a 
lipoid substance or whether the more distinct reactions are due to changes 
of lipoidic bodies present in the serum can not be stated. 
In using Besredka’s antigen for comparative tests it was desirable to 
establish whether the nonspecific reaction reported by Bronfenbrenner 
(7.) and Stimson (35) could be observed in tests conducted with serums 
of cattle. In these instances the presence of syphilis, which was assumed 
to be the cause of some atypical reactions, was excluded. Nevertheless 
similar disturbing factors were observed in the tests with bovine serums. 
It has been stated that some of the nonspecific reactions might have been 
due to the possibility that the serums originated from animals which had 
been previously tested with tuberculin. There were, however, non¬ 
specific reactions obtained in animals belonging to a herd in which the 
tuberculin test had not been applied for at least one year, the animals of 
this herd having been proved free from tuberculosis for years by periodical 
testing. Table IX serves as a comparison as to the results obtained with 
Besredka’s antigen and the one prepared by the writers from tuberculin 
precipitate and defatted organisms. 
The number of cases tested and the results recorded in Table IX 
would not appear to justify any definite claims as to the comparative 
value of Besredka’s antigen and the one prepared by the writers. How¬ 
ever, in the course of these experiments numerous samples of serums 
were repeatedly tested with various antigens in order to establish their 
relative accuracy, and, based especially on these results, a considerable 
advantage appeared to be in favor of the antigen prepared from a bacillary 
emulsion and tuberculin precipitate. 
67904°—17 - 2 
