The conventional picture of what happens during a stress relaxation usually involves imagining the response of a single dislocation to a steadily decreasing stress. The velocity of this dislocation decreases with decreasing stress in such a way that we can measure the stress dependence of the dislocation velocity. Analysis of the data from a different viewpoint enables us to calculate the apparent activation volume for the motion of the dislocation under the assumption of thermally activated glie. Conventional thinking about stress relaxation, however, does not consider the eventual fate of this dislocation. If the stress relaxes to a low enough level, it is clear that the dislocation must stop. This is consistent with the idea that we can determine the stress dependence of the dislocation velocity from relaxation data only for those cases where the dislocation's velocity is allowed to approach zero asymptotically, in short, for those cases where the dislocation never stops. This conflict poses a dilemma for the experimentalist. In real crystals, however, obstacles impede the dislocation's progress so that those dislocations which are stopped at a given stress will probably never resume motion under the influence of the steadily declining stress present during relaxation. Thus one couldmore » envision stress relaxation as a process of exhaustion of mobile dislocations, rather than a process of decreasing dislocation velocity. Clearly both points of view have merit and in reality both mechanisms contribute to the phenomena.« less