1. Afforestations are globally increasing to produce timber and pulp wood, and to enhance carbon sequestration aiming at the mitigation of CO2 emissions. Large areas have been and will be afforested, particularly in China, in order to compensate for the negative impacts of former clear-cuttings and to improve ecosystem services such as carbon sequestration or ecosystem nutrient retention, attributable to minimized soil erosion or leaching losses. We wanted to know whether monocultures or species mixtures of native species in afforestations promote nitrogen acquisition and retention. 2. To test the effect of species richness on system N retention and tree sapling N uptake we conducted a 15N tracer experiment in a young tree plantation. To this end, saplings of four abundant early successional tree species have been planted in monocultures, in two- and four-species mixtures and as single trees.

To test for the effect of competitor presence, saplings have been planted either as single tree or 16 saplings in an array of four by four. The 16 saplings have been used to establish monocultures, two-species combinations and four-species combinations. The four monocultures of each species, all six possible two-species combinations, and one four-species combination made a total of eleven species combinations.In the species combinations, each species was represented by the same number of individuals in both, the peripheral rows (i.e. 12 individuals) as well as in the center (i.e. four individuals). To avoid edge effects, all analyses were performed using the four central individuals. All treatments were replicated four times, once in each of the four blocks. The total number of plots was 88 (11 species combinations x four blocks + 11 single saplings x four species – spread over four blocks). All treatments were randomly assigned to plots within blocks. The experiment was run from March 2009 until September 2010 when sapling harvest took place. Pulse labelling with 15NH415NO3 (98% 15N) was performed in August and in September 2009. Tracer application at 10 cm soil depth took place at eight points in a grid of 7 cm spacings surrounding single trees and each of the four central individuals. Each sapling received an amount of 0.01314 g of 15NH415NO3 per application dissolved in water. 15N addition was carried out by means of an Eppendorf Repeater (4780, Eppendorf AG, Hamburg, Germany) extended by an injection needle with four openings.Young, fully developed leaves have been sampled 6 days before and after each application as well as in April and September 2010. Leaves have been dried in a drying oven (24h, 70°C). The species specific allocation of biomass to foliage was estimated using allometric equations on the basis of the ground basal diameter (measured at the time of leaf sampling, R² of allometries ≥ 0.76; (Trogisch 2012)). The species mean leaf biomass per individual and plot has been used for the estimation of the N pool. Fine roots (diameter < 2 mm) have been excavated in September 2010, washed, dried (48h, 60°C) and weighed. Species mean root biomass per individual and plot has been calculated for the estimation of the N pool. Fine root samples were pooled per species and plot for 15N analyses. We restricted biomass sampling to plant organs of high N content. Five soil samples were taken by means of a soil corer (10 cm depth) before and after the first 15N application, after the second application, as well as in April and September 2010. Soil samples were pooled per plot and dried (24h, 70°C). The soil pool was calculated as the volume of soil below the four central individuals (30 × 30 × 20 cm soil depth) and multiplied with bulk density (1.09 g cm-3 and 1.64 g cm-3 for 0-10 cm and 10-20 cm depth respectively).

The experiment was set up in the northeast part of Jiangxi Province, P.R. China (N 29° 06.29 E 117° 55.28).

Pulse labelling with 15NH415NO3 (98% 15N) was performed in August and in September 2009. Leaves and roots have been harvested in September 2010

Four tree species. Schima superba, Castanea henrii, Quercus serrata, Elaeocarpus decipiens

15N contents from references and samples were calculated as:δ15N (‰) = [(Rsample/Rstandard)-1]1000, where Rsample and Rstandard are the ratios of 15N to 14N of the sample and the standard, respectively. The standard is conventionally set as atmospheric N2 (Rstandard = 0.0036765; (Coplen et al. 1992). 15N enrichment is the per mille isotope enrichment of a sample (δ15Nsample) vs. a reference (δ15Nref; for leaves and soil: samples taken before the first application, for roots: samples from untreated saplings; (Fry 2006):15N enrichment (‰) = [(δ15Nsample - δ15Nref) / (δ15Nref + 1000)] * 1000.15N tracer recovery (at.%) in leaves and fine roots of saplings and of soil was calculated as:15Nrec = [mpool*(at.%15Npool – at.%15Nref)/(at.%15Ntracer – at.%15Nref)] where mpool is the amount of N in the biomass of leaves/fine roots per sapling (g), at.%15Npool, at.%15Nref and at.%15Ntracer is the at.%15N of labelled leaves/fine roots/soil samples, the unlabelled reference and the added tracer, respectively (Fry 2006). 15N tracer recoveries in percent (% 15Nrec) represent masses of 15N tracer recovered as percent of total 15N tracer masses applied per sapling. In addition to the recovery of leaves, fine roots and soil, N retention was calculated as the total recovery (sum of plot soil, leaf and fine root recovery as these components take up most of N) for the plots with competitors. To calculate the plot recovery of leaves and fine roots, we used the biomass of leaves and roots of all four central individuals and the respective species specific enrichment values. The relative leaf, root and soil recovery in N retention was calculated as percentages of total recovery.