While diversity-invasibility relationships have often been addressed in experimental grassland communities, experimental approaches in forests are less common, especially in species-rich subtropical regions. Forest ecosystems feature higher complexity than grassland ecosystems, since they are composed by different vertical layers covering a high level of horizontal heterogeneity. In the context of a large-scale tree diversity experiment in subtropical China (BEF-China) with diversity levels from 1 to 24, we established a seed addition experiment with groups of native and exotic herbaceous species. In this early-successional system, seed mixtures for both groups were created with high and low functional diversity, sown into subplots along the tree diversity gradient of and harvested after one vegetation period.

The seed addition experiment was established on Site A for analyzing the impact of tree diversity on herb layer invasibility and finished with harvest in October 2010. The seed addition experiment was conducted within 32 plots representing the full tree diversity gradient at Site A. These selected 32 plots included the tree diversity levels one (n = 16), two (n = 8), four (n = 4) eight (n = 2), 16 (n =1) and 24 (n = 1). Within the northern five rows of each plot, six subplots (2 m × 2 m) containing four trees, respectively, were randomly selected and established in March 2010. Thus, the seed addition was based on a total of 191 (32 × 6) subplots, each experiencing a neighbourhood tree diversity ranging from one to four (1, 2, 4). Within each subplot, seeds were added to a core area of 1 m² in the centre and surrounded by four target tree individuals, whereas the adjacent area served as a buffer zone The six subplots per plot were assigned to different seed addition treatments. Seed addition was applied by using five different seed mixtures, which were composed by six different species each. The species are annuals, bi-annuals or perennials and represent typical species of ruderal, disturbed sites like clear-cut forest areas. Two seed mixtures were composed by native species (N) or exotic species (E) only. Within each of the groups of native and exotic species, one seed mixture was composed by functionally dissimilar species (N high and E high) and a second mixture by functionally similar species (N low and E low). A fifth seed mixture was composed by three exotic and three native species (EN) and a sixth subplot per plot received no seed addition and served as an internal control. In April 2010, each seed addition subplot received a total of 2400 seeds per m², i.e. 400 seeds per each species were sown into the subplot’s core area. A full vegetation survey including cover, maximum height and mean height of all naturally occurring species in the herb layer was conducted in April, May, August and September 2010.

Main experiment, Site A, no shrub VIPs 29°08'-29°11'N 117°90'-117°93'E

Initial seed addition was performed in April 2010

The following herb species were used for seed addition: Ageratum conyzoides, Amaranthus spinosus, Bidens frondosa, Chenopodium ambrosioides, Conyza canadensis, Erigeron annuus, Gynura crepidioides, Oenothera biennis, Solidago canadensis, Artemisia lavandulaefolia, Artemisia anomala, Bidens biternata, Chenopodium album, Corchoropsis tomentosa, Lactuca formosana, Rumex japonicus, Siegesbeckia glabrescens, Solanum nigrum

For each subplot, the coverage of each herb layer species is given