In 1983 Luc Montagnier and his colleagues claimed to have discovered a novel retrovirus presently known as human immunodeficiency virus (HIV). By 1984 HIV was almost universally accepted to be the cause of AIDS. However, 20 years later, HIV cannot account for the phenomena for which the retroviral hypothesis was proposed, namely, Kaposi's sarcoma, decrease in T4 lymphocytes and thus the opportunistic infections in AIDS patients which were assumed to be the direct results of this decrease. Agents other than HIV to which patients belonging to the AIDS risk groups are exposed cause decrease in T4 cells. Neither have the main predictions of the HIV hypothesis been fulfilled. HIV seropositivity in the developed countries still remains restricted to the original high risk groups, no HIV vaccine exists, and no successful animal model. has been developed. In this communication, we critically analyse the evidence which in 1983 was claimed to prove the existence of HIV. The phenomena which Montagnier and his colleagues considered proof for the existence of HIV are detection of reverse transcriptase activity; the presence of retrovirus-like particles in the culture; immunological reactivity between proteins from the culture supernatant which, in sucrose density gradients, banded at the density of 1.16 g/ml ("purified virus") and antibodies in a patient's (BRU) serum. Reverse transcriptase activity can be found in viruses other than retroviruses and in all normal cells. Reverse transcription can be brought about not only by the enzyme reverse transcriptase but also by normal, cellular DNA polymerases. Retrovirus-like particles are ubiquitous in cultures not infected with retroviruses, especially in conditions employed by Montagnier et at. From the reaction between proteins in the "purified virus" and antibodies in the patient serum Montagnier concluded that the proteins were HIV proteins and the antibodies were HIV antibodies. Since all antibodies are polyspecific, from such a reaction it is not possible to define the origin of even one reactant let atone both. Even if this were possible, since Montagnier's "purified virus" did not contain particles with the "morphology typical of retroviruses", the proteins cannot be retroviral. We conclude that, these phenomena are non-specific to retroviruses and thus cannot be considered proof for the existence of a unique retrovirus HIV. (C) 2004 Elsevier Ltd. All rights reserved.