320 NATURE STUDIES. 



series of 382 trials, the average number of successes 

 at the first attempt by an ordinary guesser would be 

 7J. Of our trials, 127 were successes on the first 

 attempt, 56 on the second, 19 on the third, making 

 202 in all. On most of the occasions of failure, 180 

 in number, second trials were made ; but in some 

 cases the guesser professed inability, and declined to 

 make more than one, and in others we allowed three ; 

 no trial beyond the third was ever allowed. During 

 the last day or two of trial, after it had occurred to 

 us to notice the point, we found that of the failures 

 to guess a card at the first trial, those wrong both in 

 suit and number were a small minority. Our most 

 striking piece of success, when the thing selected was 

 divulged to none of the family, was five cards running, 

 named correctly on a first trial ; the odds against this 

 happening once in our series were considerably over 

 a million to 1. We had altogether a good many 

 similar batches, the two longest ruus being eight 

 consecutive successes, once with cards and once with 

 names; where the adverse odds in the former case 

 were over 142 millions to one, and in the latter some- 

 thing incalculably greater. If we add to these results 

 others obtained on previous visits, it seems not too 

 much to say that the hypothesis of mere coincidence is 

 practically excluded." 



" The exceptional nature of this inquiry/' proceeds 

 Professor Barrett, " goes far to invalidate arguments 

 founded on character and demeanour; and, on this 

 head, we will only state our conviction that any candid 



