Brown dreamed nothing at all, but he gets excited when he hears the bad news at home; he thinks, by false memory, that he has a recollection of it, he says to his wife, 'My dear, didn't I tell you, last night, I had dreamed all this?' and his equally excited wife replies, 'True, my Brown, you did, and I said it was only one of your dreams.' And both now believe that the dream occurred. This is very plausible, is it not? only science would not say anything about it if the dream had _not_ been fulfilled--if Brown had remarked, 'Egad, my dear, seeing that horse reminds me that I was dreaming last night of driving in a dog-cart.' For then Brown was not excited. None of this exquisite reasoning as to dreams applies to waking hallucinations, reported before the alleged coincidence, unless we accept a collective hallucination of memory in seer or seers, and also in the persons to whom their story was told. But, it is obvious, memory is apt to become mythopoeic, so far as to exaggerate closeness of coincidence, and to add romantic details. We do not need Herr Parish to tell us _that_; we meet the circumstance in all narratives from memory, whatever the topic, even in Herr Parish's own writings. We must admit that the public, in ghostly, as in all narratives on all topics, is given to 'fanciful addenda.' Therefore, as Herr Parish justly remarks, we should 'maintain a very sceptical attitude to all accounts' of veridical hallucinations. 'Not that we should dismiss them as old wives' fables--an all too common method--or even doubt the narrator's good faith.' We should treat them like tales of big fish that get away; sometimes there is good corroborative evidence that they really were big fish, sometimes not. We shall return to these false memories. Was there a coincidence at all in the Society's cases printed in the Census? Herr Parish thinks three of the selected twenty-six cases very dubious. In one case is a _possible_ margin of four days, another (wrongly numbered by the way) does not occur at all among the twenty-six. In the third, Herr Parish is wrong in his statement.[4] This is a lovely example of the sceptical slipshod, and, accompanied by the miscitation of the second case, shows that inexactitude is not all on the side of the seers. However the case is not very good, the two percipients fancying that the date of the event was less remote than it really was. Unluckily Herr Parish only criticises these three cases, how accurately we have remarked. He had no room for more.