THE EVOLUTION OF INTELLIGENCE 109 



laboratory conditions of control, or at least under conditions 

 of a somewhat artificial character. Evidence of the first type 

 tends to be anecdotal in character and, in the nature of the 

 case, is rarely able to present any trustworthy account of the 

 life history of the individual creature whose behavior is under 

 discussion, or indeed of the immediate antecedents of the par- 

 ticular episode under consideration. It would be at once 

 unjust and ungracious to deny substantial value to much of 

 the information emanating from such sources. Indeed, with 

 the case of the instincts nearly all of our knowledge derives 

 from this source. But on the other hand, evidence of this 

 character may be subjected to the most searching scrutiny, for 

 to it attaches not only more than the usual dangers of inaccu- 

 rate observation, but also the all but inevitable tendency to 

 anthropomorphic interpretation. This is the happy hunting 

 ground of the nature faker and of the manufacturers of 

 Arabian Nights tales and Baron Miinchhausen legends of 

 animal life and behavior. None but a hopeless ingrate could 

 recall Mowgli and his friends without a sentiment of deep and 

 lasting gratitude; but, if taken as more than literary romance, 

 the accounts of Kaa, of Shere Khan and the Banderlog, to 

 mention no other instances, can only result in profound mis- 

 apprehension. 



On the other hand, the laboratory school suffers from the 

 charge of subjecting animals to abnormal conditions, under 

 which their reactions are distorted and in which it is impos- 

 sible to expect normal expressions of their intrinsic intelligence. 

 As regards the larger undomesticated animals, there is doubt- 

 less some force in this type of criticism, and indeed when it 

 comes to certain forms of procedure, to be mentioned in an 

 instant, it may be alleged with some show of proof, that the 

 very conditions under which the observations are made tend 

 to defeat their own purposes. For example, it is a not un- 

 common procedure in animal experimentation to create in- 



