, 2009). The

resulting reforested landscape is highly fragmented, with a mosaic of different forest and scrub types, farmland and settlements ( Ma and Fu, 2000). The area surrounding the BFERS is dominated by secondary Q.wutaishanica woodland, while stands of the native birch species Betula platyphylla (Sukaczev) and B. dahurica (Pall.) have become established, especially at higher elevations. Natural regeneration has also led to the establishment of a mixed forest Dolutegravir of broadleaved and conifer species, while non-extractive pine (P.tabulaeformis) and larch (Larix principis-rupprechtii (Mayr.)) plantations cover significant areas. P. tabulaeformis is a popular plantation species naturally co-occurring with Q. wutaishanica at elevations of 1200–2000 m, whereas L. principis-rupprechtii grows naturally at elevations between 1610 and 2445 m in northern AZD6244 nmr China ( Zhang et al., 2009), although larch monocultures are commonly encountered at lower altitudes.

We selected study sites in the five dominant forest types: larch, pine, mixed, oak and birch forest. These all harbour a well-developed and diverse understory of subdominant trees, shrubs and herbs. All study sites were located on steep slopes of 15–39° between 1165 m and 1410 m, with larch and birch forest sites located on north-exposed slopes in accordance with their general distribution, while sites representing the other forest types varied in their exposition. Following exploration of forest type boundaries on the ground, four plots were selected in each forest type to survey vegetation and sample ground beetles. Plots were positioned at least 50 m away from each other to ensure sample independence (Digweed et al., 1995). A distance of at least 15 m was kept to any path or open space to minimise edge

effects. This was deemed sufficient since carabids do not respond strongly to edge effects in forest landscapes (Heliölä et al., 2001). Plots were located Nintedanib (BIBF 1120) in areas that appeared representative of the overall forest structure, and plot locations were recorded using GPS. In the centre of each plot location, two pitfall traps were set two metres apart, giving a total of eight traps per forest type. Plots were necessarily grouped relatively closely together due to the small patch size of each forest type and the need to avoid transitional zones. Plot locations were selected to provide distinct results in relation to the specific carabid assemblages supported by each forest type. Sampling occurred over ten weeks between July and August 2011 and over thirteen weeks between June and September 2012, to coincide with peaks in carabid activity reported from the same area (Yu et al., 2006). Plastic cups with a diameter of 7.5 cm and a depth of 10.2 cm were used as pitfall traps, protected by a metal roof positioned ∼6 cm above the cups.