SARA E. BRANHAM 715 



Most of the evidence concerning the antigenic activity of substances that are 

 not proteins has been obtained from bacteria. Nicolle' reported an antigen in several 

 members of the colon-typhoid group which produced agglutinins in the blood of 

 injected animals, and which was soluble in absolute alcohol and in ether. Perlzweig 

 and Steflfen^ obtained an actively immunizing antigen which survived digestion with 

 trypsin, was non-lipoidal, and gave Millon, xanthoproteic, and ninhydrin reactions. 

 They believed that this antigen probably adhered to the protein molecule, but was 

 itself not a protein. The apparently protein-free lipoid fraction obtained from the 

 tubercle bacillus by Dienes and Schoenheit^ is one of a number of similar lipoidal 

 materials that have been reported. The experiments of Ferry and Fisher'' have 

 indicated that the antigenic elements of many bacteria may be more or less loosely 

 bound to the cells. They obtained antisera of high titre by using saline washings of 

 the bacterial cells, whereas they found the washed cells themselves relatively low in 

 antigenic properties. The most active washings gave some protein reactions, but 

 pneumococcus washings containing as little as 0.0298 per cent nitrogen were good 

 antigens. 



There have been other reports of antibody production in animals with non- 

 protein antigens. In many of these instances a negative biuret test has been the 

 criterion for judging the protein-free state of the material used. In others, materials 

 were not considered protein free if they gave any of the typical amino-acid reactions. 

 The biuret test is not very sensitive. Even in the absence of all interfering substances 

 it can detect only about one part of protein in ten thousand parts of water. It is 

 quite conceivable that an amount of protein of antigenic significance may be unde- 

 tected if such a criterion be employed. 



Apparently very little work has been done on the amount of protein necessary 

 to stimulate antibody formation. Wells^ succeeded in producing fatal sensitization 

 of guinea pigs with single doses of crystallized egg albumin as small as o.oooooi gm. 

 Osborne, Mendel, and Harris^ reported that, although 0.004 g"^- of their purified 

 ricin proved fatal to medium-sized rabbits, lesser amounts conferred an immunity to 

 this dose. Seibert^ demonstrated agglutinins in the serum of rabbits immunized with 

 an amount of concentrated pyrogenic water which probably contained 0.000062- 

 0.000118 gm. protein. In our own work^ we have succeeded in demonstrating agglu- 

 tinins, precipitins, and complement-fixing antibodies in the sera from rabbits that 

 received an amount of synthetic medium-culture filtrate from Bacterium enteritidis 

 which we estimated to contain at most 0.000003-0.000004 gm. of protein. It may 



'Nicolle, C: Ann. deVInst. Pasteur, 12, 161. 1898. 



^ Perlzweig, W. A., and Steffen, G. I.: J. Exper. Med., 38, 163. 1923; Proc. Soc. E.xper. Biol., &' 

 Med., 20, 378. 1922-23. 



^Dienes, L., and Schoenheit, E. W.: Am. Rev. Tiiberc, 8, 73. 1923. 



*• Ferry, N. S., and Fisher, L. W.: Brit. J. Exper. Path., s, 185 and 205. 1924; /. Lab. 6* Clin. 

 Med., 10, 817. 1925. 



s Wells, H. G.: The Chemical Aspects of Immunity, p. 196. 1925. 



'Osborne, T. B., Mendel, L. B., and Harris, I. F.: Am. J. Physiol., 14, 259. 1905. 



' Seibert, F. B.: ibid., 71, 621. 1925. 



* Branham, S. E., and Humphreys, E. M.: /. Infect. Dis., 40, 516. 1927. 



