volume, are in themselves sufficiently numerous, as a simple calculation will show, to preclude explanation by chance, if no serious error has vitiated the records.
When Miss Campbell and Miss Despard—to take an illustration from the preceding chapter—are occupied, the one in present sensation, the other in imagination with the same scene, the conditions, as said, can be effectively controlled. Further, the experimenters have some experience in recording their observations: the time of the experiment is of their own choosing, so that they are not taken unawares: the records are practically contemporaneous with the events; each is made before any knowledge of the other's experience is forthcoming. Lastly, both parties are necessarily concerned to be as accurate as possible in describing their own side of the experience, since any fanciful embellishment may impair the accuracy of the correspondence. But when, to take the strongest case, a man sees the vision of a friend at the time of his death, we have no such safeguards to ensure the accuracy of the record. The vision finds him unprepared and often unable to appreciate its significance. Even when the impression produced is such as to induce the percipient to make a note of the circumstance or to write a letter about it before the correspondence with the death is known, it is but rarely, as the following narratives will show, that the contemporary record is preserved. When no note is made, and we have to depend entirely on
