Growth, stability, and genotype x environment (GxE) interaction were investigated for 69 clones after five years at six sites in Minnesota. Fifty-three clones were Populus deltoides x Populus nigra (DxN) crosses, nine were P. deltoides x P. maximowiczi, ten other crosses. Most clones were previously screened for growth and disease resistance in Minnesota. Five-year diameter (DBH) and basal area (BA) at 1.38 m averaged 93.5 mm and 72.11 cm<
sup>
2<
/sup>
, respectively, over the six sites. DBH site means varied from 109.0 to 79.4 mm. The fastest-growing clone BA was 64% and 49% larger than the mean of the two commercial standards and the mean of the population, respectively. Site, clone, and clone x site effects were highly significant in the ANOVA. The variance component for clone was over twice that of clone x site (GxE), indicating a relatively small reduction in genetic gain due to GxE. Clonal rank did not change between sites. GxE interaction was dominated by relative performance differences of clones on the different sites. Twenty-six percent of clones were stable (little change in growth between sites), 74% unstable. Stability coefficients of the unstable clones varied over a 99% range, indicating the population had high and variable phenotypic plasticity. Only 15% of clones were both stable and fast growing. Seven putatively superior clones, all DxN, were selected for future testing under near-commercial conditions. The results, if representative of other inter-specific Populus populations, indicate it will be difficult, and probably impractical, to reduce GxE with standard quantitative genetics methods in hybrid poplar tree improvement programs without sacrificing productivity gains.