The overall aim of this project was to investigate HIV-1 genetic diversity in South Afri ca and to characterise the immune response in mice to a South African subtype C gp120. To investigate the relationship between subtype and mode of transmission, samples were collected from individuals infected by heterosexual and male homosexual transmission from patients attending local HIV/AIDS clinics in Cape Town (n=49) and Bloemfontein (n=4). Isolates were subtyped using heteroduplex mobility assay (HMA) based on the V3-V5 region of the env geneusing reference plasmids (2 B, 2 C and 1 D) representative of local subtypes. HMA identified four env subtypes: A, B, C and D. Subtype B viruses were found in 92.9% (26/28) of the male homosexual/bisexual group and subtype C viruses in 77.2% (17 /22) of the heterosexual group. Subtype B viruses were also identified in two heterosexual patients, one patient infected by blood transfusion and in two patients with. unknown mode of transmission. Subtype D viruses were found in one male homosexual patient and one heterosexual patient and a husband and wife couple were infected with subtype A viruses. A significant association between subtype and mode of transmission (p=<0.0001) was identified, confirming two independent epidemics. To determine the subtype distribution of HIV within urban heterosexual populations throughout South Africa, samples were collected from women attending antenatal clinics in Johannesburg (n=34), Pretoria (n=S) and Durban (n=20). Samples from Bloemfontein (n=24) were taken from individuals attending an HIV/AIDS clinic. All eighty-three samples were subtyped by HMA in the env region as before. The predominant subtype circulating within the urban heterosexual population throughout South Africa was identified as subtype C (92.8%) although subtype B was also detected (7.2%). It may thus be beneficial if a HIV vaccine for South Africa is based on a subtype C model. In addition, a rapid method for identification of HIV-1 gag subtypes was developed based on restriction fragment length polymorphism (RFLP) analysis of 400bp (p17) or 650bp (p17 and 5' p24) long PCR fragments. This strategy was appl i ed to eighty-six samples (Cape Town n=47, Johannesburg n=20, Bloemfontein n=17 and Durban n=2) previously subtyped by either sequence analysis of the gag p17 region (n=31), heteroduplex mobility assay (HMA) based on the env gene (n=76), or both (n=21). RFLP analysis identified two subtype A, twenty-five subtype B, fifty-eight subtype C and one subtype D isolates. There were no discrepancies between RFLP and sequence gag subtypes, demonstrating the reliability of this method and no discordance between gag RFLP subtypes and env HMA subtypes, indicating no recombinant viruses in the genomic regions analysed.