Mendelian and non-Mendelian heredity: a reappraisal

Abstract The first report of a non-Mendelian gene (Correns 1908) appeared just eight years after the rediscovery of Mendel’s laws and, in this sense, Mendelian and non-Mendelian heredity have both been available for analysis and comparison for almost sixty years. Yet it is only now that non-Mendelian genes are becoming accessible to the kind of intensive investigation required to reveal their chemical identity, their mode of replication and transmission, and their role in cellular heredity. In the classical studies of non-Mendelian heredity in plants by Correns, von Wettstein, Baur, Renner, Michaelis, Oehlkers, and later by Rhoades [for reviews see Caspari (1948), Rhoades (1955), Jinks (1964)] the existence of hereditary factors showing a maternal rather than Mendelian pattern of inheritance was clearly demonstrated. Further progress, however, was severely impeded by two kinds of experimental difficulty: (1) The stability of non-Mendelian genes toward conventional mutagens made it very difficult to acquire mutants to study in genetically characterized, inbred lines. The attempt to explore natural variability of non-Mendelian genes by inter-varietal and inter-species crosses led for the most part to hopelessly complex results involving multigenic interactions (e. g. Michaelis 1958). (2) The typical pattern of maternal inheritance, useful in the initial discrimination between Mendelian and non-Mendelian genes, was subsequently a great obstacle in the analysis of segregation and recombination. The few instances of male transmission of non-Mendelian genes through the pollen provided some of the clearest insights into the workings of the system, but here too the mutations available for analysis were too few and the experimental procedures too complex.