This research is done to quantifiy individual tree growth and mortality within and among different species in the comparative study plots. Local neighbourhood interactions are analysed by the means of competition indices. Questions are: (a) does diversity reduce competition? (b) does competition change with successional stage? The data might provide valuable information to be compared with data from the main experiment. We chose 20 target trees each of the two most abundant evergreen species (Castanopsis eyrei FAGACEAE and Schima superba THEACEAE ) and of the two most abundant decidious species (Quercus serrata and Castanea henryi FAGACEAE). Competitors were defined as all trees with a dbh> 10 cm in middle and old successional forests and >3 cm in early successional forests and within a radius that is 1/2 of the height of the target tree. The positions of the competitors in relation to the respective target tree were evaluated (distance and azimuth). Following parameters were measured of all trees: total height, heigth of bifurcatin point, dbh and the crown width in 8 directions.

Study design We decided on an individual-based approach to elucidate our hypotheses. We selected four tree species of high abundance (Yu et al., 2001): S. superba Gardn. et Champ., C. eyrei (Champ. ex Benth.) Hutch., Q. serrata var. brevipetiolata and C. henryi (Skan) Rehd. et Wils. called target species. The target species belong to different functional groups, since S. superba and C. eyrei are evergreen and Q. serrata and C. henryi are deciduous. Data sampling was conducted during summer and autumn 2008 on the 27CSPs Twenty target trees per species were chosen randomly within the plots from all individuals complying with the following criteria: (i) single stemmed; (ii) diameter at breast height (dbh, 1.3m above ground) >10cm (intermediate and old permanent plots) or dbh >3cm (in young permanent plots); (iii) crown position in the upper canopy layer; (iv) each target species could only be selected once per plot. Thus, the 20 target trees of each species were spread over 20 of the 27 permanent plots with the exception of C. henryi (10 trees in 10 plots) for which only 10 individuals fulfilling all the criteria were found. Local biotic conditions (local species richness, local functional diversity, competition) were assessed by recording the position, size and identity of the neighbours of each target tree. Neighbours were defined as all individuals whose top height (corrected for the slope position (z-axis)) cut the hull of a reversed cone with an opening angle of 70◦ and positioned with its tip at the foot of the target tree (following the method of Biging and Dobbertin, 1992). They also had to fulfil the minimum dbh criterion (criterion (ii) above). Each target tree, together with its local neighbours formed a group.Each target tree had a mean number of 10.9 (±6.0 SD) neighbours, resulting in a total number of 837 surveyed individuals.

CSPs 29°08'-29°17'N 118°02'-118°11'E Measurements and calculated variables Crown radii in the eight subcardinal directions were determined by means of a crown mirror. The crown projection area (hereafter crown area) was calculated using the formula for a polygon. In cases of extraordinary crown displacement – the crown projection did not include the stem base – the distances to the proximal and distal edge of the crown were measured in all possible directions (if this was only possible for one direction, four crown radii were measured as follows: the distances to the proximal and distal edge of the crown were determined and, starting at the centre of this crown diameter, on the axis perpendicular to it. In this case, crown area was approximated as a quadrangle). Relative crown displacement (rd) was considered to be the distance of the centre of gravity of the crown area from the stem base, divided by the mean crown radius (Longuetaud et al., 2008). In cases where the centre of gravity perfectly matches the stem base, rd is equal to 0. Main tree stems may deviate from vertical stature in two different ways: stem inclination and stem bendiness, with the former involving straight but leaning stems and the latter bent or curved stems (Schamp et al., 2007). In the forest interior of Gutianshan NNR bent growth forms are confined to a relatively small number of trees. To obtain the degree of stem inclination, we first measured the height of the bifurcation point (height at which the lowest living crown branch of the tree branches off, excluding epicormics or springs), and then the horizontal distance of the bifurcation point from the stem in the direction of the slope. The angle of inclination was calculated as tan alpha of these two distances. If the stem was inclined towards the slope, stem inclination was considered to be negative. This method is similar to that used by Matsuzaki et al. (2006), where stem inclination was defined by referring to a straight line between the stem base and the top of the tree. However, in our study we used the bifurcation point as a reference point in order to separate the effect of crown asymmetry from that of stem inclination.The dbh (measured with a diameter measurement tape) and total height were recorded for all trees. The relative position of the neighbours to the target tree was measured as the horizontal distance from stem base to stem base. All height and distance measurements were conducted using a Forester Vertex Hypsometer (Haglöf, Sweden). All neighbours were determined to species. Local species richness (i.e. the species richness of the group) was estimated using the rarefaction method (Hurlbert’s formula, 1971) of the vegan package in R (subsample size: two trees). We applied the same method to obtain ameasure for local functional diversity. For this purpose we classified the species into seven functional groups. Since our hypotheses are based on growth characteristics of crown and stem, the trees were allocated to the functional groups according to adult tree height (canopy tree vs. sub-canopy tree) (Poorter et al., 2006), leaf longevity (evergreen vs. deciduous) (Deng et al., 2008) and leaf morphology (simple leaves, compound leaves, needles). We decided to establish an additional functional group for Fabaceae, because of their well known ability to fix nitrogen and the resulting differential preconditions for growth.To characterise the competitive power of the target trees competition indices have been calculated. Upper canopy heightwasdefined as themeanheight of the highest 20% trees in the respective group.For each tree group, data for slope inclination and aspect were collected and transformed after Beers et al. (1966).

Target tree species are: Schima superba, Quercus serrata, Castanopsis eyrei, Castanea henryi

very steep terrain, very hot, multilayer canopy -> difficult to see the crowns of upper canopy individuals, long sampling period

We tested speciesspecific differences of crown area, relative crown displacement and stem inclination using a mixed effects model with species as fixed effect and plot as random effect. We refined our models by testing for main effects of abiotic and biotic environmental variables in order to verify their effects on crown and stem parameters. Biotic variables included species richness, functional diversity, competition index, the dbh of the target tree and mean upper canopy height, while the abiotic variables were slope inclination, slope aspect and soil depth. Target species identity was also included as a categorical predictor variable. Similar to the previous models, plot was included as random effect.