The rate of spontaneous mutation resulting in electrophoretic variants per cell generation in a human lymphoblastoid cell line, on the basis of experiments described in this paper, is found to be 7.2 x 10-8 per locus. A review of similar data on electrophoretic variants resulting from spontaneous mutation in the human germ line leads to an estimate of 3.3 x 10-8 per locus per cell generation. It is argued that the similariy of these two estimates, despite an average cell generation time of 18.5 hr for the cultured somatic cells but about 26 days in the germ line, suggests that spontaneous mutation involving nucleotide substitutions is much more dependent on cell generation than on time. This finding permits the inference that environmental (exogenous) variables make a relatively small contribution to the rate of this type of human germinal spontaneous mutation. While in vitro somatic-cell mutation rates, such as derived in this study, provide a basis for modeling the contribution of nucleotide substitutions in multihit/clonal theories of carcinogenesis, it is also argued that the complex of events involved in carcinogenesis, including chromosomal rearrangements and mitotic recombination, could have very different individual probabilities. Estimates for the rates of these other types of mutation are needed to provide a better understanding of the manner in which multiple mutations accumulate in malignant cells. review published data on the rate with which mutation results in electromorphs in the human germ line and convert these data to a rate per cell division, to render the findings comparable to the somatic cell studies. The data on mutation rates are still minimal for our purposes, but numbers of this type are hard won, and we consider the methodological points to be developed-which should guide future studies-to be as important as the present data.The somatic and germinal rates which are derived are in surprising agreement. On the other hand, the errors to be attached to the two estimates at this time are such that the two estimates might differ by a factor of 3, the germ-line rates being lower. The hypothesis will be advanced that the degree of agreement between these two estimates suggests that extraneous perturbations (radiation, chemicals) play relatively little role in spontaneous germ-line mutation rates involving single nucleotides. The somatic rates observed provide a basis for modeling the contribution of nucleotide substitutions to multihit carcinogenesis. However, we will argue that great caution must be exercised in applying the observed in vitro somatic-cell spontaneous mutation rates to theories of oncogenesis. These rates may not be applicable to cancer-related genes or to genetic lesions other than nucleotide substitutions.The question of the contribution of "exogenous" factors to contemporary "spontaneous" germ-line mutation rates in humans is one of long-standing interest. In this paper, we will pursue this question through a comparison of the rate of spontaneous mutation in a cultu...