the disaster they do not thereby, it is true, serve the purpose of wish-fulfilment, the hallucinatory conjuring up of which has, under the domination of the pleasure-principle, become the function of dreams. But we may assume that they thereby subserve another purpose, which must be fulfilled before the pleasure-principle can begin its sway. These dreams are attempts at restoring control of the stimuli by developing apprehension, the pretermission of which caused the traumatic neurosis. They thus afford us an insight into a function of the psychic apparatus, which without contradicting the pleasure-principle is nevertheless independent of it, and appears to be of earlier origin than the aim of attaining pleasure and avoiding ‘pain’. This is therefore the moment to concede for the first time an exception to the principle that the dream is a wish-fulfilment. Anxiety dreams are no such exception, as I have repeatedly and in detail shown; nor are the ‘punishment dreams’, for they merely put in the place of the interdicted wish-fulfilment the punishment appropriate to it, and are thus the wish-fulfilment of the sense of guilt reacting on the contemned impulse. But the dreams mentioned above of patients suffering from traumatic neuroses do not permit of classification under the category of wish-fulfilment, nor do the dreams occurring during psycho-analysis that bring back the recollection of the psychic traumata of childhood. They obey rather the repetition-compulsion, which in analysis, it is true, is supported by the (not unconscious) wish to conjure up again what has been forgotten and repressed. Thus the function of the dream, viz. to do away with the motives leading to interruption of sleep by presenting wish-fulfilments of the disturbing excitations, would not be its original one; the dream could secure control of this function only after the whole psychic life had accepted the domination of the pleasure-principle. If there is a ‘beyond the pleasure-principle’ it is logical to admit a prehistoric past also for the wish-fulfilling tendency of the dream, though to do so is no contradiction of its later function. Now, when this tendency is once broken through, there arises the further question: are such dreams, which in the interests of the psychical binding of traumatic impressions follow the repetition-compulsion, not possible apart from analysis? The answer is certainly in the affirmative.