Horse anti-human thoracic duct lymphocyte globulin (ATDLG) has been used successfully for the treatment of severe aplastic anemia, although not all lots have comparable efficacy. We have characterized the antibody specificities contained in one lot of Swiss ATDLG found to provide a response rate of 69% and another lot that provided only a 31% response rate. Antibody specificities were analyzed quantitatively by competitive inhibition assays with the use of a panel of fluorescein-conjugated murine monoclonal antibodies that recognize T cell antigens, common leukocyte antigens, and "la-like" antigens. Although there was wide variation in the amounts of individual antibody specificities within each lot, the effective lot of ATDLG contained an average of 2 1/2 times as much of each antibody specificity as the less effective lot. There were only two antibody specificities that differed remarkably from this pattern; and these deviations did not appear sufficient to account for the variation in ATDLG efficacy. It is possible that antibody specificities other than those tested were responsible for therapeutic efficacy. Alternatively, the data suggest that it might be possible to achieve improved results for the treatment of severe aplastic anemia with higher doses of less effective lots of ATDLG.