1. 1. Investigations have been made on the activity of normal human serum, and of baboon plasma in killing non-pathogenic trypanosomes such as Trypanosoma equiperdum. Baboon plasma is 5 times more active than human serum but otherwise its activity is qualitatively the same. 2. 2. The activity varies according to whether it is measured in vitro or in vivo (i.e. in mice). It is postulated that there are 2 closely related substances, A which acts in vitro and B which acts in vivo. Usually both substances are present but in suitable circumstances they can be obtained separately. Most human and baboon sera contain much B, but some contain very little A. Usually there is 20–40 times as much B as A. 3. 3. In vitro, there is a lag period of about 3 4 hour and then trypanocidal action is completed in less than 6 hours. In vivo the action is a little slower. 4. 4. Substance A is suppressed if the human serum is injected intravenously into rats and mice but B is made more active. The substances could not be removed from serum by absorption with large numbers of trypanosomes. 5. 5. Human serum and baboon plasma were fractionated by ion-exchange chromatography and by gel filtration using Sephadex G-150 and Sepharose 6B. The active fractions were further examined by Laurell immunoelectrophoresis. These techniques indicated that the active substances are α 2 macroglobulins with molecular weight approximately 300,000. B has a slightly greater molecular weight and is slightly more acidic than A. It is postulated that both substances possess the same protein nucleus, but that they differ slightly in the configuration and/or composition. Some sera, which were initially inactive in vitro, yielded highly active fractions after gel filtration; this activation might be due to conversion of B into A or to removal of inhibitors which were masking A. 6. 6. When trypanosomes are exposed to serum there is a latent period of half an hour, after which the trypanosomes become distorted until they are almost globular. As shown by the electron microscope, there is progressive disintegration of the cytoplasm and large clear spaces appear beneath the pellicle. The other structures of the trypanosome persist unchanged for a long period, but ultimately the trypanosome disintegrates. 7. 7. As shown above the trypanocidal substances are α 2 macroglobulins and they are probably synthesized in the liver (not in the spleen and lymph glands). Accordingly their activity is quite different from the standard immunological processes. Their action might be due to either a coincidental stimulation of the serum protein by the surface of the trypanosome, or it might be a biological mechanism to protect men and baboons from many species of African trypanosomes, or that it might be part of a properdin system against infections in general. 8. 8. The trypanocidal action of human serum has been used by Rickman and Robson (1970) in the Blood Incubation Infectivity Test to differentiate T. rhodesiense from T. brucei. Improved methods for conducting and interpreting this test will be described in a second communication.