Abstract In a case of primary atypical pneumonia (blood grounp O), incomplete antibodies of the cold agglutinin type with a very wide thermal spectrum (35 to 37 C.) led to a severe immunopancytopenia in the form of severe hemolytic anemia, agranulocytosis, and thrombocytopenia. It is obvious that hemolysis was not induced by cold agglutinins which were active 1:1024 at 4 C. and only 1:32 at 22 C. The indirect cold antibodies which were probably responsible for the hemolysis which occurred were thermolabile (1:16 at 22 C., 1:8 at 37 C.) and could be demonstrated only in the antiglobulin plasma test. During the acute phase, too, a hemolysin of the Dacie type active for normal erythrocytes, but even more active for trypsinized cells, at 37 C. and at a pH of 6.8 and 7.6 could be demonstrated. The severe hemolysis could be stopped by ACTH treatment but the titer of the complete and incomplete cold antibodies showed no distinct change and the direct and indirect Coombs test remained positive for a long time. However, there was a distinct diminution of the thermal amplitude of the incomplete antibodies. A severe agranulocytosis (140 granulocytes) developed which could be improved each time by cortisone. However, permanent improvement resulted only after the disappearance of the incomplete antibodies one year later. A leukopenic factor could be demonstrated by transfusion of 300 cc. blood to a recipient of the same blood group producing a pronounced fall in the granulocytes of from 4200 to 910. Control transfusions from a normal inidividual to the same recipient showed no change and the same is true for the transfusion experiment performed one year later from the now cured patient to the same recipient. Leukocyte agglutination tests were negative, but it is presumed that agglutinins of the incomplete type may have been present. Opsonins could not be demonstrated. The thrombocytopenia never descended below 50,000. The bone marrow showed a distinct shift to the left, i.e. a predominance of unripe forms. On the basis of our previous experimental work (amidopyrine agranulocytosis) it is believed that the changes in the marrow are brought about by a depletion and exhaustion of the bone marrow due to the enormously increased peripheral destruction of the granulocytes by agglutination rather than to an actual inhibition of the marrow. The relation of this immunoleukopenia in our previous findings to the occurrence of leukocyte agglutination in agranulocytosis of allergic origin is discussed. Agranulocytosis and leukopenia may be produced on an immunologic basis in many other cases and thus some form of agranulocytosis may be related to the mechanism of erythrocyte destruction in acquired hemolytic anemia and to the agglutination of thrombocytes ins essential thrombocytopenia on an immunologic basis.