Local vessel wall shear stress is considered to be important for vessel growth. This study is a theoretical investigation of how this mechanism contributes to the structure of a vascular network. The analyses and simulations were performed on vascular networks of increasing complexity, ranging from single-vessel resistance to large hexagonal networks. These networks were perfused by constant-flow sources, constant-pressure sources, or pressure sources with internal resistances. The mathematical foundation of the local endothelial shear stress and vessel wall adaptation was as follows: delta d/delta t = K*(tau-tau desired)*d, where d is vessel diameter, tau desired is desired shear stress, and K is a growth factor. Single vessels and networks with vessels in series developed stable optimal diameters when perfused at constant flow or with a constant-pressure source with internal resistance. However, when constant-pressure perfusion was applied, these vessels developed ever-increasing diameters or completely regressed. In networks with two vessels in parallel, only one; vessel attained an optimal diameter and the other regressed, irrespective of the nature of the perfusion source. Finally, large hexagonal networks regressed to a single vessel when perfused with a pressure source with internal resistance. The behavior was independent of variation in parameters, although the adaptation rate and the diameter of the final vessel were altered. Similar conclusions hold for models of vascular trees. We conclude that the effect of shear stress on vascular diameter alone does not lead to stable network structures, and additional factor(s) must be present.