The birth of a dream is then no mystery. It resembles the birth of all our perceptions. The mechanism of the dream is the same, in general, as that of normal perception. When we perceive a real object, what we actually see--the sensible matter of our perception--is very little in comparison with what our memory adds to it. When you read a book, when you look through your newspaper, do you suppose that all the printed letters really come into your consciousness? In that case the whole day would hardly be long enough for you to read a paper. The truth is that you see in each word and even in each member of a phrase only some letters or even some characteristic marks, just enough to permit you to divine the rest. All of the rest, that you think you see, you really give yourself as an hallucination. There are numerous and decisive experiments which leave no doubt on this point. I will cite only those of Goldscheider and Mueller. These experimenters wrote or printed some formulas in common use, "Positively no admission;" "Preface to the fourth edition," etc. But they took care to write the words incorrectly, changing and, above all, omitting letters. These sentences were exposed in a darkened room. The person who served as the subject of the experiment was placed before them and did not know, of course, what had been written. Then the inscription was illuminated by the electric light for a very short time, too short for the observer to be able to perceive really all the letters. They began by determining experimentally the time necessary for seeing one letter of the alphabet. It was then easy to arrange it so that the observer could not perceive more than eight or ten letters, for example, of the thirty or forty letters composing the formula. Usually, however, he read the entire phrase without difficulty. But that is not for us the most instructive point of this experiment.