That in some cases there is what we may call a deliberate subconscious selection in the imagery presented to consciousness in dreams, there can be no doubt. But mental imagery is deeper and more elemental than any of the higher psychic functions even when exerted subconsciously. Just as the immense procession of continuous and totally unfamiliar imagery which is evoked by the action of mescal on the visual centres has no more connection with the subject's volition or desires than the procession of the starry skies, so likewise, we seem bound to admit, it may be in the case of a succession of separate images in dreams. It is nearly always possible to find a link of connection between any two images chosen at random, and the link is often a real subconscious link, but not necessarily so. Discontinuous images may arise, it seems probable, from a psychic basis deeper than choice, their appearance being determined by their own dynamic condition at the moment. We must, as Baron Mourre[15] not quite happily puts it, take into account 'the physiological state of ideas.' If we hold to the belief that dreaming is based on a fundamental and elementary tendency to the formation of continuous or discontinuous images, which may or may not be controlled by psychic emotions or impulses, we shall be delivered from many hazardous speculations. When we thus start with the recognition of a more or less spontaneous procession of images as the elemental stuff of dreams, one of the first problems we encounter is the relation of attention to that imagery. What is the degree and the nature of the attention we exert in dreams?