THE striking similarity between the first two human immunodeficiency virus type 1 (HIV-1) isolates Lai/LAV (formerly LAV, isolated at the Pasteur Institute1,2) and Lai/IIIB (formerly HTLV-IIIB, reported to be isolated from a pooled culture at the Laboratory of Tumor Cell Biology (LTCB) of the National Cancer Institute3,4) provoked considerable controversy in light of the high level of variability found among subsequent HIV-1 isolates5. In November 1990, the Office of Scientific Integrity at the National Institutes of Health commissioned our group to analyse archival samples established at the Pasteur Institute and LTCB between 1983 and 1985. Retrospective analyses6,7 have shown that contamination of a culture derived from patient BRU by one from patient LAI was responsible for the provenance of HIV-1 Lai/LAV; the contaminated culture (M2T-/B) was sent to LTCB in September 19836,7. Our goals were to determine which HIV-1 variants were present in the samples and the sequence diversity among HIV-1 isolates from the earliest stages of the AIDS epidemic. We examined archival specimens and report here the detection of six novel HIV-1 sequences in the cultures used to establish the pool: none is closely related to HIV-1 Lai/IIIB. A sample derived from patient LAI contained variants of both HIV-1 Lai/IIIB and HIV-1 Lai/LAV, and a sequence identical to a variant of HIV-1 Lai/IIIB was detected in the contaminated M2T-/B culture. We conclude that the pool, and probably another LTCB culture, MoV, were contaminated between October 1983 and early 1984 by variants of HIV-1 Lai from the M2T-/B culture. Therefore, the origin of the HIV-1 Lai/IIIB isolate also was patient LAI.