But, as at different times they guessed all the numbers between 90 and 100, we believe that they were not aware that our series stopped at 90, in which case their chance of being right in a single guess was 1 to 90. No one will suppose, therefore, that 117 complete successes in 664 guesses was the result of chance. Good days and bad days alike are included in the numbers given, though, as will be seen in the sequel, on some days no success at all was obtained. It was clear that the power of divining the numbers was exceedingly variable, but whether the difference was in the agent or the percipient, or on what circumstances it depended, we have so far been unable to discover.

Eight persons, at least, besides Mr. Smith, tried to act as agents, but either failed to hypnotise the percipients or to transfer any impression. Nor did others succeed in transferring impressions when the hypnotic state had been induced by Mr. Smith. Mr. Smith himself did not succeed except when the percipients were hypnotised.

[The account, written at the time, of the first day's experiments with each of the four percipients - Mr. W., Mr. T., Mr. P., and Miss B. - is next given. The numbers used were all those from 10 to 90 inclusive; each number was drawn at random from a bag, and placed in a little box which Mr. Smith held in his hand with its back towards the subject, who sat facing him with closed eyes. It was ascertained by experiment that the subject would have had to move his head several feet to see the number. He was only told that he was to see numbers of two figures.

All the experiments are given in order in a tabulated form, including in each case the number drawn, the number guessed, and the conditions. The following is a summary of the trials with P. and T.: -

Of the 139 trials "with Mr. Smith in another room," there were 52 in which the two rooms were only separated by a curtain; 3 of the complete successes were obtained under these conditions. In the other 87 trials Mr. Smith (accompanied by one of the other experimenters) was in the passage outside, with the door closed, or in a room below. With T. 79 trials were made with Mr. Smith in another room, but the success obtained was not beyond what might be expected by chance.

The nature of the percipients' impressions and other cognate points are fully discussed in the paper. A detailed discussion is also given as to how far the conditions were adequate to prevent the impressions reaching the percipients through any of the ordinary channels of sense. The question of the possibility of unconscious whispering or counting by the agent is fully dealt with, and an analysis is given of all the wrong guesses, in order to test whether the mistakes are such as could have been produced by this means. The conclusion is reached that the mistakes could not be thus accounted for.1

The question of the bearing of "number-habits " on these experiments is also discussed. On this subject, see the "Note on Number-habits," in 630 A.

1 Messrs. Hansen and Lehniann, in a pamphlet entitled Ueber unwillkürliche Flüstern ... (in Wundt's Philosophische Studien, vol. xi. part 4) endeavoured to show that the success of these experiments was due to unconscious whispering, on the ground that the mistakes made in them were similar to the mistakes that occurred in their own experiments in unconscious whispering. Professor Sidgwick, however, showed (Proceedings S.P.R., vol. xii. pp. 298-315), by analysing all the experiments, that the number of correspondences in mistakes in the two sets were not more than might have been produced by chance, and that, therefore, no argument could be founded on them.

The experiments were continued at intervals during 1890 and 1891 by Mrs. Sidgwick, assisted by Miss Alice Johnson, Mr. G. A. Smith being the hypnotiser, and the principal subjects being P., T., and Miss B. The experiments were all made at Brighton, some in Mr. Smith's rooms, and some in Mrs. Sidgwick's lodgings.

Three different kinds of experiments were made: (1) in transferring numbers of two digits; (2) in transferring ideas other than numbers; (3) in producing local anaesthesia and rigidity by mental suggestion.1

The first kind of experiments afforded important confirmation of the earlier series, in that success was obtained with agent and percipient in different rooms, no one in the same room with the percipient knowing what the numbers to be guessed were. A full description, with a plan, is given of Mr. Smith's rooms (one of the "arches" on the beach at Brighton), in which most of the experiments were made. The arch contained two floors, the upper entirely occupied by a sitting-room, and the lower consisting of a little lobby and two small rooms. The first conditions of the evening's experiments with Miss B. are described as follows: -

The persons present were Miss B., Professor Barrett, Mr. Smith, Mrs. Sidgwick, Miss Johnson, and T.; but T., with whom we have been experimenting, was hypnotised, and not attending to what was going on. Mr. Smith hypnotised Miss B. and tried three numbers while in the room with her and speaking to her. There was some success with these, and then Mr. Smith and Miss Johnson went downstairs to the lobby. Miss Johnson drew a number at random out of a bag containing all the numbers from 10 to 90, and handed it to Mr. Smith, who then tried to communicate it mentally to Miss B. A little bell or gong was rung upstairs by Mrs. Sidgwick as soon as a guess had been made, and again when the party upstairs were ready to begin another. As the bell was rung Mrs. Sidgwick, who had previously been put en rapport with Miss B., said, "Therés another, Miss B.," or words to that effect; and Miss B. began to look out for an impression, which usually came to her in a visual form. Mrs. Sidgwick noted down her guesses and remarks, while Miss Johnson noted the numbers that she drew from the bag, and at intervals we compared notes.... It is hardly necessary to say that, as Mrs. Sidgwick was herself wholly ignorant of the number, it was impossible for her remarks to give any indications to Miss B.