The sensitivities of six commercial combined anti-HIV-1/anti-HIV-2 enzyme immunoassays (EIAs) were evaluated, one assay (ELAVIA) based on whole virion antigen, two assays (Abbott, Wellcozyme) based on antigens expressed from recombinant DNA, and three assays (Biochrom, IAF Biochem and Pharmacia) based on synthetic peptides as antigens. All the kits investigated performed well on a panel of 47 routine anti-HIV-1-positive specimens, but on series of anti-HIV-1-seroconversion specimens from seven plasmapheresis donors, two of the peptide assays, Biochrom and Pharmacia, performed less well than the other assays. On a panel of anti-HIV-2-positive specimens, all the assays except Biochrom detected all 33 positive sera, though the reactions of some of them in the Abbott assay were relatively weak. In deciding whether to introduce a combined assay in place of an anti-HIV-1 assay, cost, specificity, the availability of confirmatory tests and the prevalence of HIV-2 in the locality, as well as sensitivity, should be considered.