A study of leukocyte antibodies is presented using (1) the sera of rabbits immunized with human leukocytes, and (2) the sera of three patients screened for the presence of such antibodies from among 36 patients with hematologic disease, 31 of whom (including the 3 studied in detail) had received multiple transfusions. The following technics are described and were employed: Leukoagglutination, leukoprecipitation including tube and agar-plate methods, agglutination of antigen-coated tanned and untanned sheep erythrocytes, the effect of antisera upon phagocytosis of heat-killed staphylococci by leukocytes, and upon ameboid motility of leukocytes. The leukoagglutinin test gives reliable clearcut results providing that appropriate controls are included and certain criteria adhered to, in order to facilitate the recognition of clumping due to other factors than true antigen-antibody union. No leukoprecipitins were detected in human sera with the technics used in this study. Immune rabbit sera, on the other hand, gave two reaction-lines in agar media, when set up against leukocyte extract. Immune rabbit sera reacted strongly with antigen-coated tanned sheep red blood cells. Human sera did not so react. One of the three selected human sera reacted with antigen-coated untanned erythrocytes, suggesting the presence of a polysaccharide antigen extractable from human leukocytes and capable of stimulating antibody formation in the human. Immune rabbit sera, and other human sera, did not react in this test. A suggestive but perhaps not a conclusive effect upon phagocytosis of bacteria by leukocytes exposed to human leukocyte antibody for 1 hour could be demonstrated. By means of ameboid motility studies, a cytotoxic effect of the human antisera upon human leukocytes could be demonstrated after 18 hours of incubation, but not after 3 hours. This was interpreted as evidence of a delayed reaction. Certain cardinal points from a clinical and theoretical standpoint with regard to the genesis of leukocyte antibodies in man are briefly reviewed. A possible analogy between leukocyte antibody formation and the homograft reaction is discussed. It is suggested that the rarity of leukocyte iso-antibody formation following transfusion is related to the fact that the intravenous pathway may be a poor route of immunization for these antigens.