Multicellular spheroids were grown from mixtures of rat brain tumor cells sensitive (9L) and resistant (R3) to 1,3-bis(2-chloroethyl)-1-nitrosourea. Percentages of each cell subpopulation in these spheroids were estimated with the sister chromatid exchange assay and were found to be approximately the same as those used to initiate spheroids. Spheroids grown from 9L cells alone had a higher growth rate than spheroids grown from R3 cells alone. However, the growth rate of mixed-cell spheroids was essentially the same as that of pure 9L spheroids and was independent of the percentages of R3 cells in mixed-cell spheroids. The sensitivity of 9L cells in mixed-cell spheroids treated with 1,3-bis(2-chloroethyl)-1-nitrosourea, estimated by changes in the number of sister chromatid exchanges per metaphase induced by treatment, decreased as the percentage of R3 cells increased. These effects are probably the result of an interaction between the two cell subpopulations held in three-dimensional contact, a situation similar to that in tumors in situ. The results suggest why one cell subpopulation of tumors does not become dominant during growth and indicate that interactions between cell subpopulations can influence the sensitivity of one subpopulation to 1,3-bis(2-chloroethyl)-1-nitrosourea.