Ocular vision is the perception of material objects, in accordance with optical laws, from a definite point in space. Our review of hallucinations has already removed two of these limitations. If I see a hallucinatory figure--and figures seen in _dreams_ come under this category--I see something which is not a material object, and I see it in a manner not determined by optical laws. A dream-figure may indeed seem to _conform_ to optical laws; but that will be the result of self-suggestion, or of organised memories, and will vary according to the dreamer's visualising power. While a portrait-painter may see a face in dream which he can paint from memory when he wakes, the ordinary man's dream-percept will be vague, shifting, and unrememberable. Similarly, if I see a subjective hallucinatory figure "out in the room," its aspect is not _determined_ by optical laws (it may even seem to stand _behind_ the observer, or otherwise _outside_ his visual field), but it will more or less _conform_--by my mere self-suggestion, if by nothing else--to optical laws; and, moreover, it will still seem to be seen from a fixed point in space, namely, from the stationary observer's eyes or brain. All this seems fairly plain, so long as we are admittedly dealing with hallucinatory figures whose origin must be in the percipient's own mind. But so soon as we come to quasi-percepts which we believe to exist or to originate somewhere outside the percipient's mind, our difficulties come thick and fast. If there be some external origin for our inward vision (which thereby becomes _veridical_) we must not any longer assume that all veridical inward vision starts or is exercised from the same point. If it gets hold of _facts_ (veridical impressions or pictures, not mere subjective fancies), we cannot be sure _a priori_ whether it somehow goes to find the facts, or the facts come to find it. Again, we cannot any longer take for granted that it will be cognisant only of phantasmal or immaterial percepts. If it can get at phantasmal percepts outside the organism, may it not get at _material_ percepts also? May it not see distant houses, as well as the images of distant souls? Hazardous as these speculations may seem, they nevertheless represent an attempt to get our notions of supersensory things as near down to our notions of sensory things as we fairly can. Whatever may be our ultimate conception of an ideal world, we must not for the present attempt to start from any standpoint too far removed from the temporal and spatial existence which alone we know.