The comparative values of the direct, the indirect complement-fixation and the agar-gel immunodiffusion tests were assessed for the diagnosis of equine infectious anemia. Antibodies were detected on the agar-gel immunodiffusion test as early as 18 days post-inoculation in the serums of experimentally infected horses and were readily detectable in all the subsequent bleedings. Complement-fixing antibodies, demonstrable by the direct method, were detected commencing about the same time. However, these were not long-lasting and were replaced by the non-complement-fixing antibodies demonstrable by the indirect method; although both types of antibodies could be detected in some sera at the same time. In a herd of 55 horses, 28 were positive on the agar-gel immunodiffusion test, and among these 28 horses, 24 of them reacted on either the direct or indirect complement-fixation test or both. Thirteen horses that were negative on the three tests at the first sampling, reacted on the agar-gel immunodiffusion test 43 days later. Ten of these positive animals had direct type of complement-fixing antibodies; only one had the indirect; and two of them were negative on both tests. It appeared that the AGI test was a more reliable technique than either the direct or indirect complement-fixation tests, particularly when dealing with serums which contained small amounts of antibody. The sequential appearance of the two different types of complement-fixing activity might be used to determine the evolution of the disease on a herd basis.