Abstract: We analyzed the spatial distribution pattern and correlation of the top four dominant tree species in a 2 hm² karst secondary forest plot of Tianlong Mountain in central Guizhou, using pairwise correlation function g(r) combined with a completely random model (CSR). The results showed that the diameter structure of trees followed an inverted J-shape, and that more trees belonged to diameter class Ⅴ (≥10 cm) driven by the dominant trees of Lithocarpus confinis and Platycarya longipes. L. confinis presented an inverted J-shaped distribution, and the population could renew very well and was in the primary growth stage. The abundance of P. longipes and Itea yunnanensis increased gradually with increasing diameter class. The density of grown and large trees was far more than the young and small individuals, which indicated poor population regeneration, and the population was in the middle and late growth stages. The top dominant tree species, except L. confinis, showed clustering distribution at large scale, which was decreased gradually with scale and trended to distribute randomly. The pattern was particularly prominent in the diameter class for young trees. Different diameter classes of different tree species presented diffe-rent spatial distribution patterns which influenced by many factors. In terms of interspecific associations, the four dominant tree species showed negative or no associations. The higher importance value of tree species, the lower the degree of association with other dominant species. The two negative correlation tree species had the lowest degree of correlation at small scale. With the increase of spatial scale, the degree of negative correlation decreased gradually, and tended to be no correlation.

Key words: karst secondary forest, spatial distribution, interspecific association