4 Journal of Agricultural Research voi. vm, no. i 
These contradictory results stimulated the investigators to further 
work. Dieterlen (17), continuing the experiments of Christian and 
Rosenblat (11), found that normal animal serums do not contain anti¬ 
bodies but that spontaneously or artifically infected animal serums 
possess them. Szabdky stated that rabbits always possess antibodies 
against tuberculin, but that they disappear whenever the animal becomes 
tuberculous. In guinea pigs such antibodies are invariably absent. 
Lowenstein (26), employing old and new tuberculin as antigen, found anti¬ 
bodies present in a high percentage of tuberculous individuals in advanced 
stages of the disease. In cases of recent infections they ^ere almost 
invariably absent. In cases treated with tuberculin they usually dis¬ 
appear for a short time and reappear about the eighth day after the 
injection and subsequently increase in amount. Ruppel (33) noted a 
marked rise in the quantity of antibodies in the serums of bovines 
inoculated with human tubercle bacilli and treated with tuberculin. 
Emery (18) employed sterile bacillary emulsion as antigen, using the 
complement present in the serum to be tested, and human red blood cor¬ 
puscles, also a special technique by varying the time during which the 
ingredients remained together at 38° C. He obtained 82 per cent of 
positive reactions in 56 cases of tuberculosis and 17.6 per cent positive 
reactions in 34 nontuberculous cases. 
Porter (32) investigated the sera of cattle and obtained complement 
fixation in normal as well as in tuberculous cases, the reaction being 
more characteristic in the advanced cases than in the early stage of the 
disease. 
Wyschelewsky (43), also working with cattle, confirmed Porter's 
results. Both observers concluded, however, that this method is not 
suitable for the differentiation of latent cases from th'e more progressive 
ones. 
Deilmann (16) compared the various extracts from the tubercle bacilli 
such as proteins, fatty acid, and tuberculo-nastin individually and also 
each in turn with tuberculin for their antigenic properties, and concluded 
that all the extracts iact as antigens, but none are as suitable in this respect 
as an emulsion of the entire bacillus or as tuberculin. 
Calmette and Massol (10) and also Letulle (24) used as antigens watery 
extracts of bovine tubercle bacilli and bacilli macerated in peptone. 
They followed a more minute technique by adding varying doses of com¬ 
plement and succeeded in demonstrating antibodies in a large percentage 
of tuberculous cases. They consider that antibodies are not invariably 
present, because in many cases a substance is present in the serum which 
inhibits the reaction. They further obtained the most marked reactions 
in cases of a slow progressive character, while in acute cases the results 
were more often negative. 
Hammer (22), in using a mixture of tuberculous syno'vial membrane 
and tuberculin as antigen, obtained out of 96 cases, 50 positive and 46 
