In the following cases, where the exact nature of the impression received was not apparently consciously classified by the percipient, it may be presumed to have been either of a visual or an auditory nature. M. Charles Richet (_Revue Philosophique_, Dec. 1884, "La suggestion mentale et le calcul des probabilités") conducted a series of experiments in guessing the suits of cards drawn at random from a pack. 2927 trials were made: ten persons besides M. Richet himself--who acted sometimes as agent and sometimes as percipient--taking part in the experiments. In the 2927 trials the suit was correctly named 789 times, the most probable number of correct guesses being 732. A similar series of trials was conducted, on Edmund Gurney's initiative, by some members of the S.P.R. and others. There were 17 series, containing 17,653 trials, and 4760 successes; the theoretically probable number, on the assumption that the results were due to chance, being 4413. The probability for some cause other than chance deduced from this result is .999,999,98, which represents perhaps a higher degree of probability than the inhabitants of this hemisphere are justified in attaching to the belief that the ensuing night will be followed by another day.[18] In a similar series of experiments carried out under the direction of the American S.P.R. the proportion of successes was little higher than the theoretically probable number.[19] But in the absence of details as to the conditions under which the experiments were made, no unfavourable inference can fairly be drawn from these results. At any rate some very remarkable results were obtained later, in a series of trials made on the lines laid down by the committee of the American Society. The agent in this case was Mrs. J. F. Brown, the percipient Nellie Gallagher, "a domestic lately come from the county of Northumberland, in New Brunswick." The experiments appear to have been carried out with great care, and the results are recorded and analysed at length (_Proc. Am. S.P.R._, pp. 322-349). 3000 trials were made in guessing the numbers from 0 to 9 or from 1 to 10 inclusive. The order of the digits in each set of 100 trials was determined by drawing lots. The agent sat at one side of a table, the percipient at the other side. At first the percipient sat facing the agent, but after about 1000 trials had been made her back was turned to the table--and this position was continued to the end. The paper containing the numbers to be guessed was placed in the agent's lap, out of sight of the percipient. There was no mirror in the room. In the result the digits were correctly named 584 times, or nearly twice the probable number, 300. The proportion of the successes steadily increased, from 175 in the first batch of 1000 trials, to 190 in the second, and 219 in the third batch. No. 3.--By DR. OCHOROWICZ.