Variation in non-target traits in genetically modified hybrid aspens does not exceed natural variation.

Genetically modified hybrid aspens (Populus tremula L. x P. tremuloides Michx.), selected for increased growth under controlled conditions, have been grown in highly replicated field trials to evaluate how the target trait (growth) translated to natural conditions. Moreover, the variation was compared among genotypes of ecologically important non-target traits: number of shoots, bud set, pathogen infection, amount of insect herbivory, composition of the insect herbivore community and flower bud induction. This variation was compared with the variation in a population of randomly selected natural accessions of P. tremula grown in common garden trials, to estimate how the "unintended variation" present in transgenic trees, which in the future may be commercialized, compares with natural variation. The natural variation in the traits was found to be typically significantly greater. The data suggest that when authorities evaluate the potential risks associated with a field experiment or commercial introduction of transgenic trees, risk evaluation should focus on target traits and that unintentional variation in non-target traits is of less concern.