3 cm2 for Skado. LAImax also showed a high variation among the genotypes, with a slightly higher variation in GS2 (COV of 32–35%) as compared to GS1 (COV of 23–32%). As a result of the very fast growth of all genotypes during the second year after establishment, values of LAImax doubled or tripled from GS1 to GS2. Genotypic means (±standard deviation) of LAImax of both growing seasons are shown in Table 3. LAD showed a somewhat higher variance (COV 31–41%). Minor genotypic and annual differences were expressed in SLA, with a mean value of 12.69 m2 kg−1. For both tree height and stem diameter, and hence also total biomass, variation

among genotypes increased in GS2 as compared to GS1. Biomass production had a COV twice as large (38%) as stem diameter (15%) and tree height (19%). The mean biomass production over both growing

learn more seasons ranged from 1.52 Mg ha−1 yr−1 for Brandaris to 7.22 Mg ha−1 yr−1 for Hees (Table 2 and Table 3). Differences in bud set and bud flush dates in GS2 were limited. Except for the T × M genotypes, which had both the earliest start and the latest end of GS2, all other genotypes had their bud flush as well as their bud set within maximal two weeks separated from each other. Sensitivity NVP-BKM120 to rust also showed a rather high variation (COV’s between 23 and 46) among the 12 genotypes (Table 2 and Table 3), confirming the importance of this selection criterion in most breeding and selection programmes of poplar. An overview of the results of the correlation analysis between biomass

production and the different leaf characteristics, the growth traits, the phenological parameters and rust sensitivity is given in Table 4. The mean biomass production was strongly positively correlated with stem diameter these and height growth, with LAImax (see also Fig. 1) and LAD, and also with SLA. Negative correlations were found with the degree of rust infection (Fig. 1). Similar correlations as for biomass production were found for diameter growth. Height growth on the other hand, was neither correlated with LAI nor with LAD, and only weakly with rust infection. Tree height was significantly correlated with the individual leaf area as well as with the timing of bud set in GS2. Phenological dates were poorly related to other parameters. The few significant correlations with phenological dates showed that the later bud set, the higher the biomass production and the RUE; this was also explained by the lower rust infection. LAImax and LAD of GS2 were negatively correlated with the rust infection during GS1. In GS1, the number of stems grown from a cutting was inversely proportional to the height reached after the first (establishment) year and also to the individual leaf area. The individual leaf area on its turn was negatively correlated with the nitrogen concentration in the leaf (Fig. 2). With regard to the wood characteristics, few correlations with other traits were observed.