DEMEREC1-3 applied the term `selfers' to ``auxotrophic mutants of bacteria in which transduction with homologous phage gives rise to significantly larger numbers of wild-type variants than occur by spontaneous reversion in uninfected control bacteria''. Lissouba et al.4 found that some mutants affecting ascospore pigmentation in Ascobolus immersus produce wild-type `recombinants' in homo-allelic crosses, whereas numerous other whitespore mutants when selfed never give wild-type revertants. However, this phenomenon has not been investigated in detail. We have examined one such mutant, `186', more precisely, because it produced the `recombinants' in relatively high frequency. Homo-allelic crosses of this mutant were carried out and tetrad analysis applied. It was characteristic that wild-type spores were found mainly in asci with 6 : 2 segregation, namely, 6 spores `186' and 2 wild. In the case of normal back mutation before meiosis, asci with 4 mutant and 4 wild spores should be obtained. Among 127,964 tetrads examined, 38 of the 6 : 2 and 3 of the 4 : 4 asci were found (frequencies 0.000297 and 0.000023, respectively). Asci with segregation 4 : 4 in homologous cross may represent the normal rate of spontaneous back-mutations that correspond to Demerec's control experiments, because only one set of genes is present during mitosis. In turn, the 6 : 2 tetrads can originate only during meiosis. In meiosis two sets of genes are present in the nucleus, so that the situation is analogous to the transduction experiments with `homologous phage'. Assuming that the 4 : 4 asci result from back-mutations in mitosis and the 6 : 2 asci from back mutations in meiosis, it is obvious that the frequency of this phenomenon is about 10-fold higher in meiosis than in mitosis. Presumably this difference is even greater if we take into account the clonal effect which works only on back-mutations in mitosis, increasing the number of wild-type nuclei that could enter meiosis.