That seems an extraordinary statement. Generalisations based on a moderate number of instances, without an adequate link of theory, do not indeed belong to highly organised and deductive science, truly; neither did meteors a century ago; and surely there are or have been facts in biology, in geology, in meteorology, and especially in the reports of geographical explorers, which could equally be disclaimed as unscientific, if tried by this singularly severe test. A votary of a deductive science may not have a very high opinion of the Baconian method' of investigation, yet surely the objection to induction is here expressed too strongly.

Nor is ability to reproduce and display a recorded phenomenon a necessary condition essential to acceptance of testimony concerning it, else must the historic appearance of certain new stars be regarded as incredible, and themselves be ousted from the scientific domain.

In another part of his essay Professor Newcomb draws an interesting and instructive parallel, or contrast rather, between the present outcome of two discoveries of Sir William Crookes's. Crookes discovered cathode rays; and behold every laboratory in the world was at work, and presently X-rays and radium resulted. About the same time Crookes also discovered, or at least published some observations upon, certain unexplained physical movements and materialisations; but in this direction, Professor Newcomb implies, nothing has securely been established at all. Very few have even tried to repeat the experiments.

The question he wishes to raise is, why this difference?

An answer is given by Professor Newcomb himself, though I should word it differently. He says these latter things do not belong to science. I say they have been observed before scientific men were ready for them. The general public, if not led by physicists, would have seen no whit more sense or meaning in the cathode rays, than the men of science were able to see in the unexplained physical movements.

But whereas for the first class of discovery every laboratory in the world was already equipped, for the second research there exist, or at any rate did exist, no fitted laboratories at all. The first discovery might have been made by any one of a hundred professors. The second observation depended for success on the presence and the willingness, the leisure and complacency, of a subject with extraordinary and exceptional faculties. The canons of evidence in this subject, moreover, are far from established; while constant precaution has to be taken against fraud.

Discoveries of the first class belong to the domain to which all men of science, and indeed the general public, have become by habit inured. Observations of the second class belong to a new and mistrusted region, full of danger, and strewn with the bones of former explorers.

There was a time when a not dissimilar assertion could be made of the first class of observation likewise.

Roger Bacon investigated things belonging to the first of the two enumerated classes, but he suffered for his temerity, and his discoveries underwent the fate of practical extinction. The world was not ready; laboratories did not exist; open-minded men were few and far between. Ordinary people might have repeated some of his observations, had they chosen, but it would have been useless if they had; they would have been obliged to forsake them and flee. It was safer to regard the ill-understood results as magical and diabolic, and to torment and ridicule the unfortunate pioneer; ridicule which, by the way, has survived, in witty fashion, even down to the latest Oxford pageant, when the greatest experimental philosopher in the history of that university was exhibited as a showman with a genuinely comic penny-in-the-slot machine.

In modern times pioneers are treated in more friendly fashion, they are pitied rather than abused, and unless they are impatient or impulsive they may well rest content with the reception accorded to their occasional utterances.

They can afford to be patient; time is on their side. And if it should really turn out that they are self-deceived, if it be really only a will-of-the-wisp that they are pursuing, then nothing ought to give them greater satisfaction than to have the futility of their quest pointed out, and to have their feet once more set upon the solid macadamised road of orthodox science.

Let us now enter upon Professor Newcomb's criticism more in detail. The two phenomena specially selected for criticism are:

(1) Thought transference, or telepathy of an experimental and controlled kind between persons generally in the same room, or at a comparatively short distance from one another.

(2) Phantasms of the dying; which, as he well knows, we endeavour, as far as may be, to explain by unconscious and spontaneous telepathy from one person to another across a considerable distance. For the least strained assumption is that the dying person unconsciously transmits an impression, or acts as telepathic agent, just before he dies; and that is why we commonly speak of these death-wraiths as phantasms of the living.

Our own position with regard to the two groups is as follows:

In the experimental cases of telepathy the difficulty is to be quite certain that all known processes of sense have been excluded; and this is often the only difficulty, since in those experiments which can be regarded as successful, the hypothesis of chance connection is quite preposterous. It is quite clear that the connection is due to some cause; the only possible question is whether that cause or connexion is telepathic, whether, in fact, all normal means of communication have been excluded with absolute security. This can perhaps only be shown conclusively by increasing the distance between the two experi

menters to several miles, which has been done successfully in some instances. So far for the experimental cases.

For the spontaneous cases, however, the opposite difficulty holds. When the agent is in Australia and the percipient in England, no one can suppose that the causal connexion between event and phenomenon lies in hyperaesthesia of the ordinary channels of sense. The main point in dealing with these cases, therefore, is to ascertain whether there is any causal connexion at all; that is to say, any connexion beyond the possibilities of chance.

All this is explicitly stated in our Proceedings, vol. x. pp. 27, 28, and Professor Newcomb's objections fall under the same heads, which I will consider separately.

To group 1, that is to experimental telepathy, our critic opposes the contention that the more thoroughly you take precaution against collusion and mal-observation, the less notable is the result obtained. I am not prepared to admit that, but it is a straightforward question of fact, which some study of our records might answer, but which renewed experiment will answer better.

6

Unfortunately he also goes on to say that we have kept no record of non-successes: the probability of success cannot be stated because we have no record of the failures, the number of which defies estimation.' But with that I really must join issue. It is a thoughtless slander which should not have been perpetrated. We are quite aware of the necessity of recording failures as well as successes. We should indeed be in an infantile stage of the investigation if we were blind to the possibilities of chance-coincidence, and if we only recorded a few successes obtained out of many thousands of experimental trials! In every series of telepathic experiments that we have ever published, the number of failures has always been recorded, and has invariably been taken into account in any deduction. Some of our investigators have even taken the trouble to see what sort of a result would be obtained by chance alone-drawing out pictures in pairs, from a set of 2000 diagrams, for instance, and seeing what, if any, correspondence ever exists between the components of any single pair. The series is recorded in the Proceedings of the Society for Psychical Research, vol. vi. pp. 398–405.

Another singularly mistaken statement follows: 'nothing bearing on [experimental telepathy is found in its recently published Proceedings.'

But in Proceedings, part 54, published in October 1907, are to be found the experiments of Miss Miles and Miss Ramsden, who have carried out experimental thought-transference over some considerable distances. And another report fills the latest part of the Proceedings, namely a detailed account of sittings with Mrs. Piper, where the whole subject of cross-correspondences is developed, which are definitely experimental. And if telepathy is not the explanation of these,

as I myself am inclined to think it is not, though clearly the record does bear upon the subject,' then something still more surprising and far-fetched will have to be postulated.

I assert, therefore, much more strongly than Professor Newcomb can deny, that direct experiment has established the possibility of an immediate kind of thought-transference between individuals.

He considers it extremely unlikely that such a faculty as thoughttransference should exist. But of that I really cannot judge: all we can say is that it is not very usual, in a form sufficiently developed to lend itself to experiment.

Its unusualness itself gives him another argument against the possible existence of the telepathic faculty. When we go to sleep in London, he says, we are surrounded by millions of other intelligences, some of them in a state of emotional excitement. How is it that we do not become aware of all these thoughts? How is it that we can keep our own ideas secret at all?

Well, it is a definite question-possibly susceptible of an answer; like the somewhat similar question, How it is that, with sensitive ears and a noisy larynx close to each other in the same head, we do not deafen ourselves by our own speech? In wireless telegraphy the difficulty is a real one: the receiver has to be thrown out of action and short-circuited whenever the sender adjacent to it is in operation, and the receiving human operator must be isolated from stray noises. Whereas in ordinary speech we all know that we can carry on conversation in a crowded hall, and with both communicators talking at once sometimes. Protection has been provided for in the structure of the head.

But reverting to the cognate case in telepathy. We must admit that, however it happens, it is an undoubted fact that the faculty of isolation, the power of secreting and isolating thoughts, exists, and is absurdly familiar to the human race. So much so, that to suggest any leakage of thought from one individual to another excites incredulity. A sceptic is nearly always on strong and popular ground; prejudice is always on his side. Clearly most people are opaque to telepathic impulses, and are presumably retentive of their own thoughts. It is only the few here and there who are found to be leaky; or, more likely, it is only the few here and there who can make any use of the leakage.

The fact could hardly be otherwise, as things are; for if telepathic communication had been common, instead of exceptional, humanity would have been aware of it from time immemorial, and it would have been incorporated as one of the root-experiences of the race. It is not in the least more unlikely a priori than is the power we possess of communicating with each other by vibrations of the air and by marks on paper. The tacit assumption underlying Professor Newcomb's objection is that every faculty possessed, or initially

possessed or residually possessed, by the human race must be common and familiarly known. But that assumption is clearly gratuitous.

We will pass to group 2-the case of phantasms, visions, premonitions and such like. First, he says that tales of these are often untrue or exaggerated. I agree tales of them often are; and rigid inquiry is necessary to secure a trustworthy record. Inquiry and collection of documentary evidence is a troublesome process, but that trouble has had to be taken; and in the book, Phantasms of the Living, as well as in Mr. Myers's book and our Proceedings, a considerable number of substantially true narratives are embodied. Here and there one has been admitted which was found not to stand subsequent test. Such lapses have been exceedingly few-not more than four in number, I believe-but they did occur; humanum est errare, even among the leaders of the Society for Psychical Research. Of these broken-down cases the Hornby' case, which Professor Newcomb quotes, stands out strongly; for both its assertion and its denial were made exceedingly public.

But I notice a singular phenomenon. Now that it has failed it is emphasised as having been a case of unique value. Professor Newcomb says of it: I only recall a single case in which the correctness of a telepathic narrative was tested by independent and conclusive authority.'

Why this emphasis? Surely not because it is a discredited case? Might some of the established cases be regarded as equally weighty and well evidenced, if only they had happened to break down? That is, if only the evidence for them had happened to turn out weak!

I ask this, not in a spirit of mockery, but of wonder. I have noticed the same tendency so often, and am never able to explain it in a polite and conciliatory manner, as I would wish.

The first objection of Professor Newcomb to the veridical nature of any hallucination amounts, then, to this: that all such correspondence between appearance and reality is of an imaginary character, that visions are seldom recorded at the time, and that they grow more wonderful in the memory. If the stories were dissected down to their bare bones, he thinks, they would evaporate in common-place.

Very well, that is one definite objection which has to be faced. On the strength of our record I meet it with a direct negative; and so it becomes a matter upon which to go to the jury.

Some objection is directed against the antiquity of some of our records. It is true that at first we had to deal with an accumulated mass of evidence, with the result that in Phantasms of the Living a few cases are published as much as twenty years old at date of publication (1886). The cases now reported to us are chiefly recent ones, and we rarely, if ever, publish any more than four or five years old.