Transient plugging of microcapillaries by leucocytes is a possible reason for the occurrence of acute hypoxia in tumours. We compared the abilities of nicotinamide at 1000 micrograms ml-1 and 150 micrograms ml-1 and pentoxifylline at 300 micrograms ml-1 to increase the filterability of normal and artificially activated human leucocytes through 8 microns pores, as a model for the capillary bed. Using a St George's filtrometer, filterability of treated leucocyte suspensions was compared with control for three to six sequential 60 microliters samples, normalising control values to unity. Pentoxifylline at 300 micrograms ml-1 halved the ratio of treated to control value to 0.47 +/- 0.13 (2 s.e.), P = 0.001 (i.e. an increase in filterability), and nicotinamide at 1000 micrograms ml-1 reduced it to 0.69 +/- 0.22, P = 0.04, but the clinically achievable 150 micrograms ml-1 was ineffective (0.82 +/- 0.25, P = 0.24). Filterability of artificially activated leucocytes was reduced (3.9 +/- 1.20) but was restored to control values of unity by 1000 micrograms ml-1 nicotinamide and 300 micrograms ml-1 pentoxifylline and partially restored by 150 micrograms ml-1 nicotinamide (1.2 mM), which was isoeffective with 100 micrograms ml-1 pentoxifylline (0.37 mM). Pentoxifylline is therefore more effective on a molar basis and was shown to affect both polymorphonuclear leucocytes and lymphocytes, while nicotinamide only affects lymphocytes. The data are consistent with the hypothesis that both agents modify acute hypoxia by increasing leucocyte filterability.