APPENDIX 459 



in this case quite impossible to exclude. Monsieur Claparede, cool 

 observer as he in general is, was greatly impressed at the outset ot 

 his experiments with Zarif by the fact that the horse spontaneously 

 rapped out in his code Schlprrd. Allowing for the resemblance 

 in German of Sch to C and for the absence of vowels in the 

 horse's notation, this is equivalent to the name Claparcde. 

 Commenting on it, Krall at first said that Zarif had never heard 

 the name, but afterwards corrected himself and said that Claparede's 

 coming had been mentioned before the horse that morning. The 

 legitimate inference is not that the horse wanted to show that he 

 knew Claparede's name, but that when his spontaneous rappings 

 began Herr Krall tended involuntarily to check them successively 

 at the point at which they made an intelligible letter, and as the 

 shaping of these letters into a name proceeded, the process would 

 become more certain. 



In fact, whatever may be said against the horses as arithmeticians 

 may be said with greater force against them as grammarians. It 

 is on their handling of arithmetical problems that they must stand 

 or fall, and on this point we may conclude negatively that, at 

 any rate, they do not do arithmetic after the manner of the 

 arithmeticians. But if the horses have not all the powers claimed 

 for them, are we to conclude that they have none at all beyond a 

 preternatural gift of responding to signs which no human observer 

 can detect ? In advancing this theory we are in some danger of 

 falling as deep into the bog of the miraculous on the one side as 

 the opposite party do upon the other. It is quite useless to bring 

 in telepathy to our aid, for telepathy would not explain the cases 

 in which the experimenter himself does not know the problem. 

 We have seen that genuine cases of success under these conditions 

 are very few, and our principal witness, Dr. Haenel, is not wholly 

 satisfactory, for he is apparently an enthusiast, and in his enthusiasm 

 omits some details which we should very much wish to have. 

 Let us, however, for the moment take his evidence, assuming its 

 soundness, and consider what powers it implies. It will be seen 

 that the experiments with Muhamed fall into three divisions. 

 The first consist in simple additions and very simple multiplica- 

 tions, which were for the most part solved accurately. The 

 second consist of a series of problems described as similar, in which 

 the horse failed and which, most unfortunately, he does not 

 specify. The third consists of the extraction of the fourth root. 

 Now this last problem was handed to Dr. Haenel by Krall 

 himself, and it is not an unfair inference that it is one in which 

 the horse might be expected to succeed. If that is so, it was 

 because the horse had memorised it, and we have seen evidence of 

 memorising, partial or complete, in some of the errors already 

 referred to. At any rate, without either calculation or memorising, 



