The Big Blue lacI transgenic rodent assay, which uses the lambda LIZ/lacI gene as the target for mutation, provides a convenient short-term assay for the study of mutation in vivo [Kohler et al. (1991): Proc Natl Acad Sci USA 88:7958-7962; Provost et al. (1993): Mutat Res 288:133-149). However, the interpretation of data from transgenic animal assays is sometimes complicated by mutants that appear as sectored mutant lambda plaques. These mutants can form a significant fraction of the mutant plaques [Hayward et al. (1995): Carcinogenesis 16:2429-2433]. Thus, in order to accurately determine in vivo mutant frequencies and mutational specificities, it is necessary to score sectored plaques and partition them from the rest of the data. In this study, the specificity of mutation in sectored plaques recovered from untreated and UVB-treated Big Blue mouse skin was analyzed and compared to mutations recovered from lambda LIZ/lacI grown on the Escherichia coli host. The mutational spectra of sectored plaques from untreated and UVB-treated mice were remarkably similar to each other and resembled those recovered from the lambda LIZ/lacI phage plated directly on E. coli. Both the sectored mutants and those recovered in lambda LIZ/lacI phage differed from the spectra of spontaneous mutants in E. coli and in Big Blue mouse skin. While sectored mutants from UVB-treated mouse skin and lambda LIZ/lacI mutants were also different from spontaneous mutants recovered from Big Blue liver, these was little difference between sectored mutants from untreated mouse skin and spontaneous liver mutants (P = 0.07). The mutational spectra of sectored plaques is thus largely consistent with their origin as spontaneous mutations arising in vitro during growth of the lambda LIZ/lacI shuttle vector DNA on the E. coli host, although the potential contribution from lesions in mouse DNA being expressed ex vivo in the E. coli host cannot be excluded.