The streptococcal antifibrinolysin test of Tillett and Garner (1) had promise of being a practical serological method for the diagnosis of a recent hemolytic streptococcal infection. The test is simple to perform, inexpensive, and uniform in its results. In the experience of Tillett and his associates (2,3), the test was generally positive in diseases caused by the hemolytic streptococcus, and negative in non-streptococcal infections. Subsequent observations by other investigators, however, indicated that the antifibrinolysin test is non-specific. Myers, Keefer and Holmes (4) obtained strongly positive tests in 2 of 6 patients with gonococcal arthritis, and in 2 cases of bacterial endocarditis due to the streptococcus viridans. Waaler (5,6) found the test to be positive in 3 of 4 patients with streptococcus viridans bacteremia, and in 6 patients with non-streptococcal pneumonia. He concluded that the presence of a positive test is of little diagnostic value.The antifibrinolysin test has been used as a laboratory aid in our pediatric clinic for over 3 years. Our early experience with the test corroborated the findings of Tillett and associates. Although in very young infants variable results were obtained, we found the test generally to be positive in children recovering from a hemolytic streptococcal infection, and negative in other types of infectious disease. Furcolow and Fousek (7) of this clinic reported in 1936 the results of 84 tests on 70 patients.-The positive tests numbered 37, and 36 of these were from definite cases of streptococcal infection; the exception was a patient with staphylococcal bacteremia. Of the negative tests none was from cases of hemolytic streptococcal disease. As our series increased, we occasionally encountered positive tests in nonstreptococcal infections without an associated bacteremia but we were inclined to view the positive results in these cases as an indication of a recent mild streptococcal infection. In many of the 1 Presented in part at the 1938 meeting of The Society for Pediatric Research. children there was a history of a recent infection which could have been caused by the hemolytic streptococcus. We did not feel that the rather frequent association of positive tests with a bacteremia interfered with the practical value of the method so long as we were aware that such an exception existed. Such cases were not common and the presence of organisms in the blood stream was easily demonstrated by culture. It did not seem fair to regard the test as non-specific on this basis since in the presence of a bacteremia it might be expected that all of the body's methods of defense might be brought into play.The specificity of the test in pediatric patients first became open to our suspicion during the pneumococcal pneumonia season of 1936 to 1937. Almost all these patients with pneumonia showed strongly positive antifibrinolysin tests at the time of admission to the hospital. In the large majority of these cases, there was no history of recent illness, and no clinical or bacte...