Intravenous injection of trypsin in the rat induces early lung leucostasis and emphysema of delayed onset. This report confirms that this emphysema is not rat-specific and that the leucostasis is not related to the presence of contaminating endotoxin in the trypsin. In hamsters (n = 37), leucostasis did not occur when they were injected with heat-treated trypsin, but numerous granulocytes were sequestered in the vessels of hamsters receiving a fresh solution of trypsin. In these hamsters, the number of granulocytes harvested by lavage increased significantly (1.87 x 10(6) per ml, P < 0.001) compared with hamsters injected with either heat-denatured trypsin (0.89) or saline (0.86), or compared with controls (0.86). Emphysema was inconstantly observed in hamsters 6 or 12 weeks after injection with trypsin for 1 h. It was frequently (17/20) present and always (20/20) well developed (intercept + 180 per cent) in the 2-h perfused hamsters whose lungs were abnormally heterogeneous (index + 100 per cent) relative to the seven controls and to the nine saline-injected hamsters. The efficiency of trypsin in triggering emphysema (percentage of hamsters having abnormal values of intercept) was dependent on the time of perfusion. This form of experimental emphysema is thus considered to be due to an endotoxin-independent leucostasis.