THE SEPTEMBER-COMMISSION 259 



at any of the experiments, was asked to put the questions 

 to the horse. Mr. von Osten went into the house, accom- 

 panied by a member of the commission. And again, Mr. 

 Schillings would go out before the second part of the 

 test, without having met Mr. von Osten. 



Five tests were made in this way. They were not 

 attended by such amazing success as were the preceding 

 ones, but nevertheless the results were surprising. The 

 horse nearly always repeated the number itself, instead 

 of performing the operation required. Since, however, 

 Mr. Schillings, owing to a misunderstanding, had, in the 

 first two cases, said to the horse : " You are to repeat 

 this number for Mr. von Osten ", the errors might appear 

 to be a result of this request. 



At the final discussion, the result of which was the 

 unanimous declaration which was given for publication, 

 not only the data obtained during these two sessions, but 

 also the earlier experiences of some of the members of 

 the commission were taken into consideration. None of 

 the tests witnessed could be referred to chance or to the 

 use of tricks. Count zu Castell pointed out that in the 

 course of eight days he had elicited forty correct 

 responses from the horse, among them some in regard 

 to which he himself had been momentarily in error. 

 Other members recalled the many instances in previous 

 exhibitions, during which both Mr. Schillings and Mr. 

 von Osten were absent, when questions were put to the 

 horse by others. The commission also had access to a 

 detailed account written by Professor Stumpf on Mr. 

 von Osten's method of instruction, based on the ex- 

 planations and demonstrations which Mr. von Osten 

 had himself given. As a result of these considerations 

 the commission felt under obligations to give public 



