We find in it none of the familiar features of myth or of untrained fancy; the reports have not given wings to a quadruped, or horns and hoofs to a carnivor; they contradict nothing that is known. Can we fairly suppose that this complete agreement, alike in what they contain and in what they do not contain, is the accidental result of a hundred disconnected mistakes?

It is most instructive, in this connection, to compare first-hand (and the better sort of second-hand) narratives with others. I have already spoken of the greater general sobriety of the first-hand evidence. I may now add that the suspiciously startling details which often characterise the more remote narratives are precisely of the sort which the telepathic hypothesis could by no possibility be made to cover. To wet the pillow or leave the door open would be quite an ordinary breach of manners in the popular "ghost," or the second-hand apparition of doubtful authority. I have mentioned the real dripping letter conveyed by the phantasmal midshipman. I may further recall the scar reported to have been left on the lady's wrist by the touch of the well-known "Beresford" apparition; and the wounds alleged to have been produced on the bodies of absent witches, by blows and sword thrusts directed to their "astral" appearances. No marvels in the least resembling these find any place in our first-hand records; yet why should they not, if those records are fundamentally untrustworthy? The existence of such features in other narratives sufficiently shows how wide is the possible range of incidents, in stories where the ordinary limitations of communication between human beings are alleged to have been transcended.

Of this wide field, the hypothesis of the action of mind on mind, which we are endeavouring to develop, covers only a single well-defined portion. By what fatality, if error is widely at work in the case of our first-hand evidence, do its results always fall inside and not outside this very limited area? If our witnesses are assumed to sit loose to the facts which they have known, why should they bring their accounts into rigid (though purely accidental) conformity with a theory which they have not known?

§ 19. What I have here indicated is the general impression produced by the evidence in our own minds. In our view, the reality of telepathy (even apart from a consideration of the experimental evidence) may be not unreasonably taken as proved. Having formed this view, we are bound to state it; but we expressly refrain from putting it forward dogmatically, and from saying that to reject it would argue want of candour or intelligence. We hold that, in such a matter, it is idle to attempt to define the line of complete proof; and the proof given - if it be one - is far from being of an eclatant or overwhelming sort. To those who do not realise the strength of the a priori presumption against it, it may easily look more overwhelming than it is. To others, again, it may appear that, on the hypothesis that the faculty has acted as widely as we have supposed, the highest evidential standard ought to have been reached in a larger number of cases. To us it rather seems that the evidence that we find is just about what might have been expected.

We see nothing in the mere existence of telepathy that would tend to make reserved people mention strange experiences, or to make careless or busy people keep conscientious diaries - or generally that would lead the persons immediately connected with a telepathic case, in which their emotions may be deeply involved, to act with a single eye to producing a clinching piece of evidence for the future benefit of critical psychological inquirers. It would, of course, be useless for us to urge that evidence which falls short of the best is still as good as can be expected, unless we were able to present a certain nucleus of fairly conclusive cases, and this we think we can do. But if the proof is held to demand more cases of the highest evidential quality, we must trust to time for them. The ideal collection would, of course, be one where every independent instance should be so evidentially complete that it must be either (1) telepathic, or (2) a purely accidental coincidence of a most striking kind, or (3) the result of a fraudulent conspiracy to deceive, in which several persons of good character and reputation have taken part.

In our view, this point has been reached in a sufficient number of the examples here given to exclude the second and third of these alternatives; but these examples constitute only a very small minority compared with the mass of cases which are merely confirmatory - strongly confirmatory, as we think, but still confirmatory only and not crucial. And the collection so far falls short of the ideal.

In saying, then, that telepathy may not unreasonably be taken as proved, I do not wish for a moment to imply that the proof which we give is the one which we should eventually desire to see given. To no reader, we think, will the various imperfections and weak spots of our case be more patent than they have been to ourselves. Some of these are beyond remedy - as the absence of contemporary documents. Others may possibly be remedied at a later stage - for instance, the suppression of names.1 It has been impossible to bring home to all our informants that where a person refuses to a phenomenon, belonging to a certain class, the direct testimony which' he would give, if needful, to any other sort of personal experience, the world is sure to take the view that he lacks that complete assurance of the reality of the experience which alone can make his evidence worthy of serious attention. This is not always just; since the reason why he suppresses his name may be, not that he doubts the truth of his evidence, but that he regards the truth in this particular department of Nature as something disgraceful or uncanny; or it may be mere fear of ridicule, or a shrinking from any form of publicity. But meanwhile the defect must not be extenuated.