From February 2011 till November 2012 the upper 15 cm of soil layer was sampled every 2–3 weeks, except for the winter when the sampling intensity was decreased. During 2011, 20 samples were collected at every sampling campaign for each genotype. During see more 2012, the number of samples was different at each sampling date, following the expected intrinsic variability of the Fr biomass based on the experience of the previous growing season (i.e. 2011). At each sampling campaign in 2011 and in 2012, half of the samples were collected in the narrow and half in the wide inter-rows, randomly distributed

over the planted area within the former pasture. Fine roots were picked from each sample by hand while: (a) separating weed roots (Wr) from poplar roots, (b) sorting poplar roots in dead and living roots, and (c) sorting Fr in two diameter classes: 0–1 mm and 1–2 mm for independent Fr productivity and mortality calculations of each diameter class (see below for more details). Poplar roots were sorted from find more Wr based on morphological characteristics. Poplar roots showed a brown color and a dense ramification pattern, while Wr had a lighter color and less ramification. The sorting of dead (necromass) and living Fr was based on the darker color and the poorer cohesion between the cortex and the periderm of the dead roots (Janssens et al., 1999). After washing, fine roots were oven dried at 70 °C for 1–4 days to

determine the dry root mass. Fr mass of one core sample picked for x min (i.e. 5–20 min) was converted into total Fr mass in the

sample (i.e. after 60 min picking duration) using Richard’s equation (as explained Vitamin B12 in detail by Berhongaray et al. (2013b)) and expressed in g DM m−2. Subsamples of dried Fr were ground for further C and N-analyses. More details on Fr collection and data processing can be found in Berhongaray et al., 2013a and Berhongaray et al., 2013b. The aboveground woody biomass was calculated for both genotypes from previously published data for the first rotation (Verlinden et al., 2013b) and from new measurements for the second rotation. A detailed inventory of stem and shoot diameter (D) distribution and of mortality was carried out for each genotype at the end of each rotation in December 2011 and December 2013. The number of shoots per tree was counted, stem and shoot diameter at 22 cm above the soil was measured for one entire row per monoclonal block and the number of missing trees was counted. Based on the stem diameter distribution of the plantation, ten trees of each genotype were selected for destructive harvest, covering the widest possible range of number of shoots and of stem and shoot diameter. Stem and shoot diameter at 22 cm was measured on the selected trees with a digital caliper (model CD-15DC, Mitutoyo Corporation, Japan, 0.01 mm precision), before the tree was harvested.