Data file description This data file contains all the genotypic and geographic information analyzed in our article. This data file was written in an Excel working sheet, including individual id, subpopulation name, genotype data and coordinates for each sampled individual tree. Each row represents the dataset of a single individual and each column represents a characteristic (e.g. “subpopulation”) or an allele in a locus (i.e. the genotype of a locus was described by two columns). This data file was created following the data form required by GenAlEx (Peakall & Smouse 2006) and can be easily transferred into the data forms of other genetic softwares used in our research article (e.g. SPAGeDi and FSTAT). Note that the effective population densities (De) used in the inference of effective gene dispersal ranges (σ) are given as following: 0.002613 in subpopulation “HY Post-fragmentation”; 0.001880 in subpopulation “LS Pre-fragmentation”; 0.002633 in subpopulation “LS Post-fragmentation”; 0.004368 in subpopulation “XS Pre-fragmentation”; 0.003456 in subpopulation “XS Post-fragmentation”. The estimations of De in population HY and LS approximately equal to D/4, where D is the census population density in each sample plot. Values of De in population XS were slightly higher than D/4 since the local density of seed bearing trees in this population was higher than other two populations according to the field observation in continuous two years, i.e. 2009 and 2010.