To date, few studies have focused on the effects of genetically modified trees (GM trees) on the environment. One concern with GM trees is that they may have unanticipated effects on non-target organisms, i.e. effects on organisms that are not direct targets of the genetically modified trait. The main objective of this thesis was to study potential non-target effects from the interaction between GM trees and natural enemies, including phytopathogens and herbivorous insects. To study this I used a system consisting of GM trees featuring changes in growth-related characteristics, and naturally occurring enemies. The GM trees used were the aspen hybrids Populus tremula x tremuloides: one unmodified wild type clone T89 (control) and transgenic lines with altered expression of gibberellin (GA 20-oxidase), sucrose (SPS) or pectin (PME); and Populus tremula x alba: one unmodified wild type clone INRA 717-1-B4 (control) and lines modified to suppress the activity of the enzymes in the lignin biosynthetic pathway, i.e. CAD, COMT, CCR or CCoAOMT. The natural enemies used were the parasitic phytopathogens Melampsora pinitorqua, M. populnea and Venturia tremulae, and the herbivorous leaf-beetle Phratora vitellinae. To address this question inoculation experiments, feeding preference experiments, analyses of secondary chemistry and field inventories were performed. The results of the studies showed that the GM trees significantly affected the interaction with the natural enemies, both in the laboratory as well as in the field. For instance, both M. pinitorqua and V. tremulae showed an altered disease incidence on the GM trees of P. tremula x tremuloides compared to the unmodified wild type T89, where all tested transgenic lines exhibited altered susceptibility to the pathogens. However, there were also differences in aggressiveness to the aspens depending on pathogen population. The results from the field inventory showed that lines within all tested transgenic construct, COMT, CAD, CCoAOMT and CCR of P. tremula x alba differed significantly from the wild type INRA 717-1-B4 in susceptibility to M. populnea. In addition, the susceptibility to the rust also differed significantly between lines carrying the same transgenic constructs. Furthermore, we found that overexpression of SPS in P. tremula x tremuloides, unintentionally induced changes in plant secondary chemistry, where the GM-line SPS33A exhibited the largest deviation from the wild type T89 in contents of plant phenolics and nitrogen, and that these changes coincide with a concurrent decrease in herbivory by P. vitellinae on this line. I argue that the altered interactions are the result of physiological changes in the trees. They can originate from direct effects i.e. altered expression of the modified trait, indirect effects of the genetic modification process e.g. pleiotropy, or effects from the transformation process e.g. position effects, to which the tested natural enemies respond. The result stresses the importance of further research on the causes and mechanisms responsible for the altered interaction between GM trees and non-target organisms, as well as evaluating the potential environmental effects of cultivation of GM trees in the field. Such research will require collaboration between researchers from different disciplines, such as plant ecology and physiology, functional genomics, proteomics and metabolomics.