Although dispersal in the giant panda (Ailuropoda melanoleuca) is a demographic mechanism which can potentially counteract the negative effect of habitat fragmentation, little is known about dispersal in this species because of difficulties in observing individuals. Using data from faecal microsatellite genotyping, we compared the spatial distribution of giant pandas in two populations and the proximity of relatives in one key population to infer their dispersal pattern. We conclude that giant pandas exhibit female-biased dispersal because: (i) vAIc (variance of assignment index) for females was significantly larger than for males, suggesting that females comprise both 'local' and 'foreign' genotypes; (ii) the average spatial distance of related female dyads was significantly larger than that of males; (iii) larger r (relatedness), F(ST) (genetic variance among populations) and mAIc (mean of assignment index) values were found in males using the software FSTAT, although the differences were not significant; (iv) males set up territories neighbouring to their birth place; (v) significant population structure using microsatellites with a concomitant lack of mitochondrial structure was found in a previous study, possibly indicating more extensive female dispersal; and (vi) female-biased dispersal was strongly supported by evidence from concomitant ecological studies. Considering previous ecological data and life-history characteristics of the giant panda, female-biased dispersal is most likely to be due to competition for birth dens among females, inbreeding avoidance and enhancing inclusive fitness among related males.

Taxus yunnanensis, which is an endangered tree that is considered valuable because it contains the effective natural anticancer metabolite taxol and heteropolysaccharides, has long suffered from severe habitat fragmentation. In this study, the levels of genetic diversity in two populations of 136 individuals were analyzed based on eleven polymorphic microsatellite loci. Our results suggested that these two populations were characterized by low genetic diversity (NE = 2.303/2.557; HO = 0.168/0.142; HE = 0.453/0.517), a population bottleneck, a low effective population size (Ne = 7/9), a high level of inbreeding (FIS = 0.596/0.702), and a weak, but significant spatial genetic structure (Sp = 0.001, b = −0.001*). Habitat fragmentation, seed shadow overlap and limited seed and pollen dispersal and potential selfing may have contributed to the observed gene tic structure. The results of the present study will enable development of practical conservation measures to effectively conserve the valuable genetic resources of this endangered plant.