The thesis is composed of two parts that are related by the theme of genetic markers. The first part involves the application of genetic markers to investigate the mating system, population genetic structure, and evolutionary relationship of the three Chinese larch taxa: Larix gmelinii, L. olgensis and L. principis-rupprechtii. Seventeen populations of the Larix taxa listed above were analysed using eight polymorphic allozyme markers. Results indicated that mating system was variable among taxa and among natural populations within taxa. Population differentiation of each taxa was very small, showing that less than 2% of total genetic variation occurred among populations. Spatial distribution of genetic variation of L. gmelinii was random, but a weak pattern of isolation by distance was detected in L. olgensis. The genetic relationship among the three taxa elucidated by allozyme markers indicated that the genetic distances were very low between them. Based on morphological traits and the results obtained by allozyme and cpDNA sequence markers, it is reasonable to consider L. olgensis and L. principis-rupprechtii to be two varieties of L. gmelinii rather than two separate Larix species. In the second part of this thesis, theories of plant population genetic structure were developed to incorporate biparentally, paternally, and maternally inherited genes into a variety of models. Population differentiation for each of the three plant genomes was formulated in the island, stepping stone and isolation by distance models of population structure. The results showed that maternally inherited organelle genes maintain larger differentiation than paternally inherited organelle genes, which in turn maintain larger differentiation than biparentally inherited nuclear genes. In the stepping-stone model, differences in genetic correlation with distance among the differently inherited genomes were conditional on the values of long and short distance migration for pollen and seeds. The relative contribution to migration of seed and pollen flow can be estimated in terms of gene frequency data or DNA sequence data.