THE INTELLIGENCE OF MAMMALS 235 
of course have been the result of the dog’s studying the 
hinges, latch and general make-up of the gate, and conclud- 
ing that if the latch were raised the gate would be free to 
swing open. Such a course would be a very natural one 
for a human being, but few would consider that a dog 
would be likely to follow it. The dog might, however, 
learn to open the gate by watching someone do it and then 
imitating him. In this case the dog might be thought to 
conclude that “since a man lifted the latch and went out, 
therefore, I can lift the latch and go out.” As a matter of 
fact the dog learned to make his escape in neither of these 
ways. His method of learning the trick which was watched 
from the beginning was as follows: Being placed in the 
yard from which he was anxious to escape Toby poked his 
head between the bars of the fence in various places and by 
chance placed it under the latch and raised it, when the 
gate swung open and he scampered out on the street. The 
method pursued was one of trial and error. The fortunate 
movement which effected the dog’s liberation was associated 
with the perception of the latch, but the association was not 
perfect at first. Only after ten or twelve experiences, ‘in 
which the number of times the dog poked his nose through 
wrong places gradually diminished, did he learn to go 
directly to the right place, and raise the latch. 
The experiments of Thorndike have convinced him that 
the intelligence of animals is limited to the type that we 
have just considered. He holds that animals do not draw 
inferences and that, barring the apes, it is doubtful if they 
have ideas. Thorndike’s experiments were among the 
first systematic attempts to get at the nature and limitation 
of animal intelligence by means of experimental methods. 
Through his pronounced spirit of iconoclasm toward anec- 
dotal psychology and anthropomorphism Thorndike was 
