To determine the relationship between antibodies to herpes simplex virus (HSV) types 1 and 2 and diagnosis of orolabial and genital herpes, a cross-sectional survey was done among 869 sexually transmitted disease clinic attenders and 1594 blood donors in London. Among clinic attenders, the prevalence of HSV-1 infection was 59.5% and that of HSV-2 infection was 22.7%, and among blood donors the prevalence was 44.6% and 7.6%, respectively. The sensitivity and specificity of a diagnosis of oral herpes for the presence of HSV-1 antibody was almost identical in the 2 groups (clinic attenders: sensitivity, 33.1%, and specificity, 91.4%; blood donors: sensitivity, 32.3%, and specificity, 94.3%). A diagnosis of genital herpes was less sensitive for antibody for HSV-2 among donors than among clinic attenders (P < .001); however, the specificity was similar in the 2 populations (clinic attenders: sensitivity, 32.1%, and specificity, 96.6%; blood donors: sensitivity, 17.5%, and specificity, 99.5%). False-positive clinical histories were also relatively common (clinic attenders, 12%; donors, 6%). The sensitivity of the diagnosis of genital herpes would be improved if accurate serologic assays for detection of HSV type-specific antibodies were more widely available.