172 PROTEIN POISONS 



medium be removed. This is too powerful a poison to be 

 injected repeatedly even in small doses into the animal 

 body. 



One of us has for the past two years used solutions of 

 the cell residue in the treatment of tuberculosis in man. 

 The most suitable preparation is a 1 per cent, solution 

 filtered through porcelain. The cell residue in weighed 

 quantity is placed in a bottle with the proper volume of a 

 0.5 per cent, solution of carbolic acid, and the bottle is 

 carried on a mechanical shaker for twenty-four hours, 

 after which the content is passed through a porcelain filter. 

 Such a solution will keep indefinitely. We have used this 

 solution sufficiently to justify the following statements: 

 (1) It is of no value in advanced cases of pulmonary tuber- 

 culosis. (2) It may prove harmful even in initial cases if 

 the dose be too large or if small doses be too frequently 

 repeated. (3) When properly used in initial cases or in 

 localized tuberculosis, its action is apparently prompt and 

 specific. If the tubercle bacilli wholly disappear from the 

 sputum, as they may, the injections should be repeated 

 at intervals of from two to four weeks for some months. 

 We .wish it clearly understood that in well-established cases 

 of pulmonary tuberculosis no benefit from this treatment 

 can be expected. We believe that in initial cases this pre- 

 paration is preferable to any form of tuberculin. 



Toxophor Group. White and Avery 1 have reported an 

 interesting research on the split products of the cellular 

 substance of the tubercle bacillus, especially of the toxo- 

 phor group. They used a strain virulent to guinea-pigs. 

 This was grown on glycerin broth cultures for six weeks, 

 and the cellular substance was washed with alcohol and 

 ether, ground in a ball mill and split up by our method. 

 The toxophor obtained by them agreed with that which 

 we have prepared. It is a yellowish-brown powder of 

 characteristic pungent odor, readily soluble in alcohol. Its 

 aqueous solutions are faintly turbid, and give the biuret, 



1 Jour. Med. Research, 1912, xxvi, 317. 



