A sensitive in vivo system was studied for use in evaluating relatively small quantities of antiviral compounds. Swiss mice were injected intracerebrally with 10 LD(50) of type 1 herpes simplex virus (HSV) or vaccinia virus. Test compounds were injected intracerebrally into mice at a standard time interval after virus inoculation. Significant increases in the number of survivors in drug-treated, infected animals as compared to saline-treated virus controls was the criterion for evaluation of antiviral activity. 9-beta-d-Arabinofuranosyladenine (Ara-A) was used to determine an optimal time of therapy for this system. Significant survival increases occurred when the drug was injected 3, 6, 12, and 48 hr after HSV inoculation, with maximal effect occurring after the 6-hr treatment interval. Determination of brain virus titer of 6-hr Ara-A- and saline-treated mice at various intervals indicated that treatment with the drug delayed development of detectable levels of HSV by 1 day. 1-beta-d-Arabinofuranosylcytosine (Ara-C), trifluorothymidine (TFT), and 5-iodo-2'-deoxyuridine (IDU) were also evaluated with the use of the 6-hr treatment time against HSV and vaccinia virus. Ara-A and TFT increased the number of survivors among mice infected with either virus, whereas Ara-C increased the survivors among HSV-infected, but not vaccinia virus-infected, mice. IDU was inactive against both viruses.