It does not occur to him that this account of the matter is in itself harder to accept than the fact of a subjective auditory hallucination. To realise this would require a certain amount of definite psychological knowledge. Things are sufficiently explained to him if they seem to cohere in an evident way. Or if he is sensible that his version of the matter introduces or suggests a decided element of the marvellous, still the marvel is of a sort which is a legitimate subject of human speculation, and with which it is interesting to have been in personal contact. And not only has his reason thus followed the line of least resistance; his memory has also been relieved by the unity which he has given to its contents. It has now got a single and well-compacted story to carry, instead of two disconnected items. It has, so to speak, exchanged two silver pieces, of different coinages and doubtful ratio, for a single familiar florin.

The above is no mere fancy sketch; it represents what is really not unlikely to occur. When we were just now considering how far an honest and intelligent witness is likely to imagine afterwards that a passing impression which at the time was vague and unrecognised had really been distinct and recognised, it will be remembered that such a perversion seemed decidedly unlikely - that we saw no ground for assuming that an error of that type had entered into anything like a majority of the cases where we have no conclusive evidence that it has not entered. But with the dates it is otherwise. We have received several illustrations of the liability of even first-hand witnesses to make times exactly coalesce without due proof of their having done so, or even in spite of proof that they did not do so. Having by a reasoning process of a vague kind come to the conclusion that the two events were simultaneous, they will be apt to note any items of facts or inference which tell in this direction, and not such as may tell in the other. An informant sometimes by his very accuracy reveals the attitude of mind which might easily produce inaccuracy in other cases.

He will tell us that all that was proved was that the death fell in the same month as the impression; but that it is "borne in on him" that it was at the same hour. A good many people upon whom such a conviction is "borne in" will treat that as if it were itself the evidence required. One sort of case in which the tendency in question has been specially evident is that where the death has taken place at a great distance from the percipient. The instinct of artistic perfection overshoots the mark, when a ship's log in the Indian Ocean shows that death took place at a quarter-past 3, and a clock on an English mantelpiece reveals that that is the very minute of the apparition. Telepathy, like electricity, may " annihilate space"; but it will never make the time of day at two different longitudes the same. This particular error would not, it is true, completely vitiate the case from our point of view, since the 12 hours' interval would not have been exceeded; but pro tanto it, of course, diminishes the credit of the witness.

§ 12. Let us now examine the two dates separately, and see where the danger more particularly lies, and what tests and safeguards can be adopted. And first as to the date of the event that has befallen the agent. As we have seen, it is almost always first from the percipient's side that we hear of this event; and to him the knowledge of it came as a piece of news, sometimes by word of mouth, sometimes in a letter or telegram, occasionally in some printed form. In very many cases the date would, of course, be part of the news. Now, if his own experience was impressive enough to have caused him real anxiety or curiosity, and if his recollection is clear that the news came almost immediately afterwards - say within a couple of days - and that the time of the two events was there and then compared, and found to coincide, the coincidence will then rest on something better, at any rate, than the mere memory of a date. It will depend on the memory that a certain unusual and probably painful state of mind received remarkable justification, and that this justification in turn produced another state of mind which was also of an unusual type.

If there was really no such synchronism as is represented, then not only the abstract fact of correspondence, but a distinct and interesting piece of mental experience must have been fictitiously imagined. Now, it may be said, I think, as a rule, that a fictitious imagination of this sort needs some little time to grow up; that it is decidedly improbable that any case which is definitely recorded very soon after the event will have suffered this degree of misrepresentation. But a few years will give the imagination time to play very strange tricks. We have had one very notable proof of this, in a case where a curiously detailed vision of a dead man, which (so far as we can ascertain) must have followed the actual death by at least three months, was represented to us, after an interval of ten years, by the person who had seen it - a witness of undoubted integrity - as having occurred on the very night of the death. We may be right in regarding so complete a lapse of memory on the part of an intelligent witness as exceptional; but we should certainly not be justified in assuming that it is exceptional; and no case of anything like that degree of remoteness can be relied on, without some evidence beyond the percipient's mere present recollection that the event which befell the agent took place at the time mentioned.

The evidence may be of various sorts. If the exact date of the percipient's experience can be proved, then it is often possible to fix the other date as the same, by letters, diaries, or obituaries, or by the verbal testimony of some independent witness. If no such evidence is accessible, or even if the exact date of the percipient's experience is forgotten, it may still be possible to obtain corroboration of the coincidence from someone who was immediately cognisant of the percipient's experience, and who had independent means of ascertaining the further fact and of noting the connection at the time. But the absence of a written record of either event is, of course, a decidedly weak point.