EDTA-anticoagulated blood samples from 19 HIV-1-positive subjects and 13 healthy laboratory worker controls were analysed for three lymphocyte subpopulations (CD3, CD4, CD8 T cells) (lysed whole blood method, Becton Dickinson FACScan flow cytometer) at 0, 24, 48, 72 and 96 h after venesection, having been stored at either 4 degrees C, 12 degrees C, 16 degrees C or 21 degrees C. In samples stored at 4 degrees C and 12 degrees C there was a significant fall in both %CD3 and %CD 4, and a significant rise in %CD8. At 16 degrees C the %CD8 remained stable, while there were marginal rises in %CD3 and %CD4. At 21 degrees C, the %CD8 again remained stable, while %CD3 and %CD4 rose significantly with time. These trends were independent of HIV-1 status. At each temperature studied, the rates of change of lymphocyte subpopulations were independent of each other. These results suggest that a temperature range of 14-16 degrees C may be optimal for sample storage prior to measurement of T cell subsets. They emphasise the importance of strict control on conditions if samples are to be kept for any length of time before analysis.