Laboratory studies were undertaken to differentiate between rubella virus isolates of 'wild' and vaccine origin. Groups of 4 to 5 rabbits (total of 114) were inoculated subcutaneously and intradermally with two consecutive doses of one of seventeen rubella virus strains at a 4-week interval. Sera were collected from each rabbit biweekly for a period of 8 weeks and once 14 weeks after the first inoculation, and were tested for rubella virus hemagglutination-inhibiting (HI) antibodies. It was found that rabbits inoculated with the Cendehill vaccine virus showed no antibody response to two doses of virus, while one dose of each of two 'wild' virus strains elicited high antibody titers in all but one rabbit inoculated. These antibody titers increased following the second virus dose and remained at a high level for 14 weeks. In contrast, rabbits inoculated with 8 virus strains isolated from Cendehill vaccinees, exhibited no immune response to the first virus dose and showed transient, low antibody titers to the second dose. This moderate reversion of the rabbit immunogenic marker of the Cendehlll virus in the human host, was not sufficient to render the 'one dose--30 day' test unsuitable for the characterization of rubella virus isolates.