Twenty-four reciprocal crosses between two green and six chimera cultivars, containing respectively normal green and mutant white plastids in their germ layers, were analysed for their segregation patterns of green, variegated and white embryos, and for their fertility data.

Analysis of variance of fertility data showed that W × G crosses were generally less fertile than the reciprocal G × W crosses, but this was not correlated with selection against any particular class of embryo.

Analysis of segregation patterns showed that they were essentially the same with a constant female and varying male parent, but change the female and the pattern could be radically altered. The six different mutant females with both constant green males could be arranged in a sequence in which white plastids were increasingly successful, and white and variegated embryos increasingly frequent. A similar sequence was demonstrated with three green females and two constant white males. It was concluded that the major control of plastid inheritance was determined by the female nuclear and plastid genotypes, with the male having only a minor, modifying influence even when male plastids were more successfully transmitted than female ones.

The very low frequencies of variegated embryos in some crosses led to a rejection of the classical hypothesis of sorting-out from mixed cells, and its replacement by the hypothesis that pure green or pure white embryos arise by the replication of only one plastid at a time, whereas variegated embryos arise by replication of both plastids.