Mouse macrophages purified by elutriation from thioglycollate-induced peritoneal exudate cells were labelled with indium-111-oxine and injected intravenously into mice. A substantial amount of unbound radioactivity remained in the circulation, suggesting that the radionuclide was not stably bound to the cells. Culture experiments with radiolabelled cells showed that indium-111 was released in the medium. Another cell marker, PKH-95, an iodine-125-labelled aliphatic compound insertable into the cell membrane, bound more stably than indium-111. Five minutes after injection of 1 2 5 I-PKH-95-labelled macrophages, about 98% of the cells were in a non-circulating pool. It was checked that PKH-95 labelling did not compromise the viability and functions of the macrophages and that autologous erythrocytes and blood mononuclear cells labelled with PKH-95 remained in the circulation after i.v. injection. One hour after injection, 1 2 5 I-PKH-95-labelled macrophages were distributed mainly in lung (36%), liver (19%) and spleen (5%). Subsequently, radioactivity decreased in the lung while increasing in liver, spleen and in an artificially induced footpad inflammation. The radioactivity accumulation in the inflammation persisted at least for 7 days. It represented a small proportion of radioactivity injected (0.2%) but was trapped very specifically in the inflammation. This raised the hypothesis that macrophages of the non-circulating pool could be released in the circulation and recruited into the inflammation with slow kinetics.