Type-specific serologic assays for herpes simplex virus (HSV) types 1 and 2 based on glycoprotein G-1 (gG-1) (HSV-1) and gG-2 (HSV-2) discriminate between antibodies against HSV-1 and HSV-2. We previously developed a Western blot assay using gG-1 and gG-2 expressed in baculovirus, performed extensive validation studies, and determined that it was both sensitive and specific for type-specific detection of HSV antibody. Here we report that, among a cohort of Thai military recruits, the serostatus of some individuals changed from positive to negative over time (6.6% among those ever positive for HSV-1, and 14.9% among those ever positive for HSV-2). We tested a subset of these specimens in three other gG-based assays: an enzyme-linked immunosorbent assay, an immunoblot strip assay, and a Western blot assay. Positive-to-negative shifts occurred in every assay; the frequency of the shifts ranged from 6. 1% to 21.2% of the specimen sets tested. There was only limited agreement among the assays concerning which individuals lost reactivity. This inaccuracy, exhibited by all of the assay protocols, was not predicted by validation studies employing specimens from cross-sectional studies and was most pronounced in HSV-2 testing. This argues for the inclusion of serial blood specimens in serologic assay validation procedures.