The ability of human eosinophils (EOS) to take up and kill intracellular (amastigote, AMA) forms of Trypanosoma cruzi was investigated as a part of our efforts to define the role of inflammatory cells, typically found in acute chagasic lesions, in clearance and destruction of the parasite. In experiments in which purified EOS and AMA were mixed and allowed to interact for 1 hr, EOS uptake of the parasites was demonstrable by electron microscopy and was confirmed by measurements of incorporation of [3H]AMA. The degree of parasite uptake increased with the AMA to EOS ratio, and was detectable as early as 20 min after the initial contact. After 1 hr, a time when minimal AMA damage was noted, 40% of the EOS were found to contain parasites. Damage inflicted on the AMA was frequently seen 2 hr after removal of free AMA; largely destroyed parasites were a common occurrence after 4 hr. The electron microscopic observations were confirmed by kinetic measurements of the release of 3H-labeled materials by EOS that had ingested radiolabeled AMA. The AMA were systematically found within the EOS vacuoles and were never seen free in the cytoplasm. Fusion of the EOS granules with AMA-containing vacuoles was frequently observed and the presence of EOS granule major basic protein (MBP) in the vacuole content was demonstrated immunocytochemically. MBP was around the AMA and bound to parasite structures, including the subpellicular microtubules. Heparin, a polyanion previously shown to abrogate MBP-mediated destruction of blood forms of T. cruzi, inhibited intracellular killing of AMA by EOS. Two other polyanions, chondroitin sulfate and dextran sulfate, had similar effects. In an EOS-free system, the presence of MBP resulted in AMA lysis; this reaction was inhibited by anti-MBP antibody. Taken together, these results favor the notion that EOS contribute to the elimination of T. cruzi AMA from infected tissues through phagocytosis, leading to an intravacuolar lytic process in which at least MBP plays a role.