The result of such researches as have been made may be said, briefly, to show that human testimony has much less value than is normally assigned to it and, in particular, much less value than it is held to possess in a Court of Law. The experimental results obtained in this field are, indeed, often startling. It is hardly too much to say that one’s first impulse, on becoming acquainted with the results hitherto reached, is to fall back on a general and dismayed scepticism regarding the value of human testimony to anything whatever. But a closer examination of the results show us that this attitude is unwarranted, and reinforces the common-sense assumption that the value of human testimony is a matter of degree, varying from complete worthlessness to a very fair presumption that the event occurred as stated. The investigation is useful chiefly in showing us what factors influence this value. It is convenient to separate out these factors according to the scheme recently employed by Dr. Edmond Locard, in his analysis of police records over a number of years. The statements made by a witness repose, in the first place, on sensations which he has experienced. It might be thought too obvious to be worth mentioning that we require the witness who heard a sound, for instance, to have reasonably good hearing, and yet there are many cases where simple preliminary considerations of this kind are not taken into account. Professor Zöllner’s famous book _Transcendental Physics_, for instance, alleged marvels that occurred in the presence of Slade, the medium; and these alleged marvels, of great influence in spreading a belief in spiritualism, were witnessed to by four professors, Zöllner, Fechner, Scheibner and Weber. But a member of the Seybert Commission, Mr. George S. Fullerton, as a result of personal interviews, found that two of these professors, Fechner and Scheibner, were partially blind at the time. Their sensations, therefore, in this respect, were untrustworthy. But defects of this kind may usually be determined and this factor conditioning the witnesses’ credibility allowed for. Where a witness makes appeal to a sensation which may be checked the check should always be imposed. Thus Dr. Locard gives an instance where a witness stated that an event occurred in a mill at a certain hour. How did he know the hour? By hearing a clock strike at the time the event occurred. A test was made, and it was found that the noise of the mill made the striking of the clock quite inaudible. The witness then remembered that he did not hear the clock strike until he had left the mill. Similarly, witnesses have testified that they saw a man leave a doorway, their post of observation being one from which the doorway could not be seen. Sensations may often be checked, however, and, to a careful inquirer, they need not be a grave source of error. But the next stage is concerned with the witness’s perceptions. Of his