In the studies of complement fixation described in this paper, the antigens were prepared from (a) normal monkey red cells, (b) parasitized red cells of monkeys dying with Plasmodium knowlesi infection, (c) the spleens of monkeys dying with Plasmodium knowlesi infection; the sera came from (a) normal human beings, (b) patients with syphilis, (c) patients with paresis who were receiving malaria therapy with Plasmodium knowlesi, Plasmodium vivax, or Plasmodium falciparum, and (d) patients with malaria alone. The malarial antigens gave negative complement fixation reactions with 70 to 80 per cent of the luetic and normal sera and weak or doubtful reactions with the remaining 20 to 30 per cent. With the exception of one antigen prepared from spleen, there was no evidence that the malarial antigens were more reactive with Wassermann-positive than with Wassermann-negative sera. Some human sera give weak complement fixation with antigens prepared from normal monkey erythrocytes, and the percentage of these positive reactions is slightly higher with malarial sera than with normal or luetic sera. The most sensitive and specific malarial antigen was prepared from dried parasitized red cells by extraction with saline, freezing, and thawing. This P. knowlesi antigen gives strong complement fixation with malarial sera from human beings infected with P. knowlesi, P. vivax, or P. falciparum. The titer of complement-fixing antibodies reaches a maximum about 1 month after the beginning of the acute infection. At this time all of the P. knowlesi sera tested were positive. After 4 months the reaction diminishes rapidly in titer but may remain positive for 12 months or longer. With P. knowlesi infections in man, the complement fixation reaction remains positive for some time after the infection has apparently disappeared as judged by daily smears and inoculation of monkeys with the blood. The complement fixation reaction in malaria is group-specific rather than species-specific. Sera from patients infected with P. vivax or P. falciparum react in the same way with the P. knowlesi antigen as the homologous sera. Absorption of malarial human sera with normal monkey erythrocytes does not remove the immune bodies which fix complement with malarial antigens.