Soil development and plant N limitation dataset

The study site is located at Maocun ecological experimental station (25°12′30′′ N, 110°33′45′′E), 30 km to the southeast of Guilin City, Guangxi Zhuang Autonomous Region in China. The area is a typical karst landscape formed from upper and middle Devonian limestone. We selected two types of study sites where one had progressed to the woodland stage and the other remained in the shrubland stage. These two types of study sites had been growing crops before the 1960, and the natural vegetation recovery (mountain closure for natural regeneration) occurred from 1960s after cropland abandonment. Six independent plots (20 m × 20 m) with at least a 50-m distance were selected, three as woodlands and three as shrublands. For each of the plots, a nearby farmland has been selected as control site, in order show the pre-succession soil condition. Within each designated plot, soil samples were carefully collected using 5-points sampling method from the surface layer (0 ~ 10 cm) and mixed thoroughly to form one composite sample. Soil development indexes include particle-size distribution (percentage of clay content), clay mineral content (total Al2O3 and free Al2O3 (extractable Al2O3)), siliceous content (SiO2), iron oxide mineralogy (Fed (dithionite-extractable iron oxide)/Fet ((total iron oxide)), Sa (Sa = SiO2/Al2O3), and Saf (Saf = SiO2/(Al2O3 + Fe2O3). Community-level plant foliar N, P and δ15N data were also analyzed and included in this dataset.