3i8 NATURE STUDIES. 



' Michael/ and then almost immediately ' Davitt/ 

 To avoid any association of ideas, we then chose 

 imaginary names, made up by ourselves at the 

 moment, as ' Samuel Morris/ ' John Thomas Parker/ 

 1 Phoebe Wilson/ The names were given correctly 

 in toto } at the first trial in five cases out of ten. Three 

 cases were complete failures, and in two the names 

 given bore a strong resemblance to those selected by 

 us, ' Jacob Williams/ being given as ' Jacob Wild/ 

 and ' Emily Walker/ as ' Enry Walker/ It was now 

 getting late, and both we and the younger children 

 were very tired : and four concluding attempts to 

 guess the name of a town in England were all failures, 

 though one of us had previously obtained remarkable 

 success in this very experiment." 



It appears to me that the failures in these and 

 other cases yet to be cited, are as important a part of 

 the evidence in favour of mind-reading or mind- 

 ruling, as the successes. For they tend to show that 

 there was no general system of deception by which 

 the members of the family who had been present 

 when the names were selected informed the children 

 by signals previously agreed upon. However, as 

 it will be obvious that there can be no absolute 

 certainty on this point in cases in which any mem- 

 bers of the family knew what was selected, we 

 proceed to consider cases in which only the com- 

 mittee of investigation knew the words or things 

 chosen. 



In the cases last considered, the explanation may be 



