Glutathione export from uninfected human erythrocytes was compared with that from cells infected with the malaria parasite Plasmodium falciparum using two separate methods that distinguish between oxidized (GSSG) and reduced (GSH) glutathione. One involved enzymatic recycling with or without thiol-masking; the other involved rapid derivatization followed by HPLC. Glutathione efflux from uninfected erythrocytes under physiological conditions occurred predominantly as GSH. On exposure of the cells to oxidative challenge, efflux of GSSG exceeded that of GSH. Efflux of both species was blocked by MK571, an inhibitor of mammalian multidrug-resistance proteins. Glutathione efflux from parasitized erythrocytes was substantially greater than that from uninfected erythrocytes. Under physiological conditions, the exported species was GSH, whereas under energy-depleted conditions, GSSG efflux occurred. Glutathione export from parasitized cells was inhibited partially by MK571 and more so by furosemide, an inhibitor of the 'new permeability pathways' induced by the parasite in the host erythrocyte membrane. Efflux from isolated parasites occurred as GSH. On exposure to oxidative challenge, this GSH efflux decreased, but no GSSG export was detected. These results are consistent with the view that the parasite supplies its host erythrocyte with GSH, much of which is exported from the infected cell via parasite-induced pathways.