An in situ enzyme-linked immunosorbent assay (ELISA), performed directly on fixed varicella-zoster virus-infected monolayers, was used to quantitate viral antigen, and by color reduction it was used to evaluate the activity of drugs against varicella-zoster virus. Color production in the ELISA (optical density) was directly related to the dose of input virus. Antigen representing 5 to 10 plaques could be detected 3 days after infection. The ELISA was specific and reproducible, as shown by absorption and repeat experiments, respectively. A color-reduction test by ELISA was compared with the conventional plaque-reduction assay for its ability to measure the antiviral activity of acyclovir, bromovinyldeoxyuridine, trifluorothymidine, and vidarabine against four strains of varicella-zoster virus. In all cases but one the 50% inhibitory doses were lower when measured by ELISA than by the plaque-reduction assay. This in situ ELISA color reduction method had the following advantages over the conventional plaque-reduction assay: (i) the endpoint was an objective measurement; (ii) there was less well-to-well variation; (iii) the assay was sensitive to changes in plaque size as well as plaque number; (iv) it was less labor intensive.