Spatial genetic structure in populations ofQuercus mongolicavar.grosseserrata(Fagaceae) from southern Korea

Abstract

Multilocus, putative allozyme genotypes were mapped and sampled from two local populations of Quercus mongolica Fischer ex Turcz var. grosseserrata (Bl.) Rehder & Wilson (Fagaceae) (each area is 100 m × 100 m, one with Sasa cover (N = 62) versus a second without it (N = 384)) occurring in undisturbed forests near Nogodan, Mount Jiri in southern Korea. Ripley's L-statistics and spatial autocorrelation analysis (a coancestry coefficient, fij) were used to test the prediction that because of low seedling establishment in a population with dense Sasa cover, there would be no spatial aggregation or hyperdispersion of individual trees and little evidence of fine-scale genetic structure in the population. As predicted, the Sasa-covered population showed no evidence of significant aggregation of individuals (P < 0.01) up to an interplant distance of 50 m and a random distribution of putative genotypes in the population. By contrast, the L-statistics conducted in the Sasa-free population indicated significant aggregation of individuals at interplant distances extending from 4 to 50 m. Spatial autocorrelation analysis revealed small but significant (P < 0.01), positive, fine-scale genetic structure extending from 10 to 30 m. A very similar result was obtained from 100 replicates each consisting of 62 trees in the Sasa-free populations by applying rarefaction and bootstrapping. These findings support the hypothesis that ground vegetation such as Sasa spp. has an impact on fine-scale genetic structure. The weak spatial genetic structure found in the Sasa-free population may primarily be due to limited acorn dispersal coupled with overlapping seed shadows and (or) secondary acorn dispersal by rodents.Key words: allozymes, Fagaceae, ground cover, Quercus mongolica var. grosseserrata, Sasa spp., spatial genetic structure.