T hree laboratory workers have been infected with the IIIB strain of HIV; their antibody response to HIV has been studied in serial serum specimens. Because the infecting virus is known, the fine specificity of the antibody response was studied on the homologous strain of HIV. Anti-p17, anti-p24, anti-gp160, CD4/gp120 blocking and neutralizing antibodies developed in parallel. Epitope mapping of the anti-gp160 response indicated several regions that consistently induced an antibody response. Serum contained antibody which reacted with V3-specific peptides corresponding to the very tip of the loop and crossreactivity was seen with V3 loop peptides from other sequence divergent strains of HIV. Antibody to the V1 loop was produced at levels comparable with that seen for the V3-loop. Anti-V1 neutralized HIV with a titration curve equivalent to an anti-V3 monoclonal antibody. Because the infecting virus is known and serial reisolates have been obtained, we explored the relationship between production of antibody to a given epitope and mutation in the virus. The data suggest that an association exists, but do not clearly indicate that antibody drives the selection for mutant viruses. The findings presented here provide a fine specificity analysis of the evolution of the antibody response to HIV in greater detail than has previously been performed.