A modification of the MT-4 cell plaque assay for human immunodeficiency virus (HIV) is described, which gave reproducible results with all 4 HIV-1 strains and the two HIV-2 strains that were used. The main feature of this new method is the use of a tetrazolium (MTT) staining procedure. The number of plaques read after 4-6 days was essentially the same as the number of infectious units derived from the 50% cell culture infective dose (CCID50) in MT-4 suspension cultures. For a selected group of antiviral compounds the 50% plaque-inhibitory doses were comparable with the 50% inhibitory doses (ID50) in suspension cultures. In the plaque assay HIV-1 (HTLV-IIIB) and HIV-2 (LAV-2ROD) were equally susceptible to azidothymidine (AZT), and the same was true for didehydrodideoxythymidine (D4T). For these compounds it was irrelevant whether they were already present during the initial HIV adsorption phase or added immediately thereafter. Pentosan polysulfate proved about 20-fold more inhibitory to HIV-2 than HIV-1. There was a 5-fold increase in activity if present during the virus adsorption stage.