Resistance to leaf rust in Populus has been shown to be under strong genetic control ( Rajora et al., 1994 and Dunlap and Stettler, 1998). Since leaf rust resistance is often

strongly correlated at different tree ages, early selection for this trait appears feasible ( Rajora et al., 1994). The results of this study confirmed that biomass production decreased with increasing rust infection ( Fig. 1) in line with previous reports ( Royle and Ostry, 1996, Steenackers et al., 1996 and Dunlap and Stettler, 1998). The infection of rust was more severe and started earlier in the year in GS1 than in GS2, and had therefore a larger impact on biomass growth. In GS1 the rust infection on Robusta caused a sudden decrease in LAI, a black coloration of leaves and leaf fall after week 35 (

Broeckx et al., 2012a). As expected, genotype Robusta was most susceptible to rust among all the Cilengitide genotypes. Robusta is the oldest of the genotypes ( Table 1), and is known for poor rust resistance ( Centrum voor Genetische Bronnen, 2013 and Steenackers et al., 1990) and slow growth ( Barigah et al., 1994, Ceulemans et al., 1996 and Meiresonne, 2006). P. deltoides species are frequently used for (back)crossing to breed rust-resistant genotypes ( Steenackers et al., 1990 and Steenackers, 2010). Genotype Grimminge, which is a back-cross of (P. trichocarpa × P. deltoides) with a P. deltoides maternal parent, showed an intermediate rust infection of all genotypes of this study.

Besides Robusta, PD0325901 clinical trial also the P. Interleukin-2 receptor nigra genotypes showed a rust score in the higher range ( Fig. 1; Table 3); this probably also explained their low productivity. Nowadays, breeding and selection strategies in Flanders aim at partial resistance and tolerance to rust rather than complete resistance due to newly arising, more virulent pathotypes that caused the breakdown of rust resistance between 1980 and 2000 ( De Cuyper, 2008). Concerning the wood characteristics, only poor genotypic variation was observed (Table 2). Similar to previous findings (Benetka et al., 2002), very small differences (COV <1%) in the HHV between the six genotypes were found, with a mean value of 19.45 MJ kg−1. This is well within the range reported in a review study on biomass quality of poplar which also concluded the small variation in HHV that exists among poplar species (Kenney et al., 1990). Therefore, selection for this trait presumably only brings about little genetic improvement. Wood quality has the lowest priority among the selection criteria for breeding and selection programs for poplar cultivars in Flanders, in particular with regard to SRC cultivation (Steenackers et al., 1990). Despite poor variation in wood density as well, a significant negative correlation with biomass production was found (Fig. 1). The lower yielding genotypes (e.g.