But to pick out arbitrarily some point that favours clairvoyance, and at the same time to pass over unsuitable elements, cannot give a correct picture of the probability that the clairvoyant was accurate in his prediction. It almost invariably happens that prophecies to which but one meaning can be ascribed are only published when the event prophesied has occurred, a circumstance that does not exactly show that much credence is to be placed in previsional clairvoyance. Nevertheless, a few cases have been published beforehand, though certainly with marvellously bad results. Only, spiritists are much too fond of suppressing their failures; for example, I have been unable to find that any further mention has been made of the prophecy published in 1899 that Dreyfus would die in 1904. There can be no sense in calculating the probability of success in cases of clairvoyance unless the failures are included in the calculation.

7. Coincidence. E.g., a command given in thought may be obeyed, because by chance, or for some other reason, experimenter and subject think of the same thing. In telepathy the first order thought of is nearly always that the right arm should be raised. This source of error is both great and interesting. It has lately been carefully examined by a member of the American branch of the Society for Psychical Research, C. S. Minot Thus, it has been discovered that every one prefers certain figures, etc., which recur strikingly often, even when the choice is left open. Now, when in a telepathic experiment A. thinks of a number which is to be divined by B. without the latter making use of any known method of perception, it would be necessary to discover if they prefer the same figures, if they have the same " number habit" This must also be weighed in experiments with cards, in which it appears to me the ace of hearts is very often chosen. It is evident that great care must be exercised in drawing conclusions, and that the study of mysterious phenomena leads to the recognition of important laws.

We have already come across coincidence as a source of error in the supposed appearance of dying persons to their friends who are far away. What Parish considered as coming under the fourth source of error was really a special case of coincidence. Coincidence and fictitious interpretation together very frequently lead investigators into error, because the greater the number of interpretations that can be given to a manifestation - i.e., in clairvoyance - the more likely will the case be one of coincidence.

It sometimes happens that the probability of coincidence is so great that the manifestation is utterly valueless. In the case of the medium, Mrs. Piper, it is mentioned that Hyslop was asked by the ghost of his father at one of the seances: "Do you remember the penknife with which I used to clean my nails?" Hyslop said, "No"; and the spirit replied, "You must remember the little knife with a black handle, that I first of all carried in my waistcoat-pocket and then in my coat-pocket ?" Hyslop then wrote to some of his relatives and learned that his father had possessed a black penknife with which he cleaned his nails. " Only it appeared that Hyslop senior did not carry the knife either in his waistcoat-pocket or in his coat-pocket, but in his trouser-pocket." In the preface to the German edition of the book in which Mrs. Piper's feats are described, Schrenck-Notzing remarks with regard to the above and other cases: - "These examples are quite satisfactory as far as minor details are concerned." I must acknowledge that this example does not satisfy me, and that I am as little satisfied with the other examples with regard to either major or minor details. In any case, numbers of people carry a black-handled knife with which they clean their nails.

Hyslop's experiment failed in one important point; his father did not carry the knife in his waistcoat-pocket or coat-pocket, but in his trouser-pocket. This point is certainly mentioned, but only in a way that would make the error appear to be of no importance. I do not quite understand what things of this nature are intended to prove. At all events, as far as the knife was concerned, the probability of chance success was so great, that the whole episode cannot be said to have proved anything.

If any good is to come from the investigation of the phenomena of occultism, the way in which the minutes of occultistic seances are kept is of the greatest importance. I consider a systematically kept protocol a necessity, but I no longer think such documents as convincing as I used to do. A superficial glance at the published protocols is often convincing, and spiritists are very much astonished when doubts are expressed as to the accuracy of the reports.

Unfortunately, it is generally found to be quite right to doubt whether these reports are a faithful reproduction of what actually occurred. A few years ago, I was able, in an essay entitled Physician and Judge, to throw some light on the untrustworthiness of the official records kept of the proceedings in magistrates' courts, about which I gave the following explanation: - "These records are not calculated to give a correct view of how such cases are conducted. If a protocol is to have any claim to accuracy it ought not to be a mere resume of the proceedings, but should be an accurate transcription of the shorthand notes of every question and answer. It ought to be possible to gather from the report whether the witness hesitated or vacillated, whether a question was first of all answered in the negative and whether an affirmative answer was only given after the question had been frequently repeated. In such a case the affirmative answer ought not to be quoted by itself. It is only when the protocol records that the witness hesitated or was at first silent, that any weight can be laid on such a document." The above concerned judicial records. But I hold the opinion that we must demand as much of the records of occultistic investigations.