The investigation by Hamilton (1911) of reactive tendencies in P. rhesus and irus yielded preeminently important data concerning complex behavior.  For the ingenious quadruple-choice method devised by this observer showed that mature monkeys exhibit fairly adequate types of response.  As Hamilton’s interest centered in behavior, he did not discuss ideation, but this does not prevent the comparison of his data with those of the present report, and the agreement of his findings with my own is obvious.

My work contrasts sharply with that briefly mentioned above in that I applied systematically and over a period of several months an experimental method suited to reveal problem solving ability.  Previously, the so-called problem or puzzle-box method had been used as a means of testing for the presence of ideas.  For this I substituted the multiple-choice method.  One of the chief advantages of this new method is the possibility of obtaining curves of learning for the solution or attempted solution of relational problems of varying difficultness.  I am confident that these curves of learning will prove far more valuable than such data as are yielded by the puzzle-box method.

The Pithecus monkeys, which I studied intensively, yielded relatively abundant evidences of ideation, but with Thorndike I must agree that of “free ideas” there is scanty evidence; or rather, I should prefer to say, that although ideas seem to be in play frequently, they are rather concrete and definitely attached than “free.”  Neither in my sustained multiple-choice experiments nor from my supplementary tests did I obtain convincing indications of reasoning.  What Hobhouse has called articulate ideas, I believe to appear infrequently in these animals.  But on the whole, I believe that the general conclusions of previous experimental observers have done no injustice to the ideational ability of monkeys.  It is clearly important, however, that we always should take into account the species of animal observed, for unquestionably there are extreme differences in mental development among the monkeys.

As I view my results in the light of their relations to earlier work, I am strongly impressed with the importance of the use of improved methods for the study of complex behavior.  The delayed reaction method of Hunter, the quadruple-choice method of Hamilton, and my multiple-choice method offer new and promising approaches to forms of activity which thus far have been only superficially observed.

The ability exhibited by Skirrl to try a method out and then to abandon it suddenly is characteristic of animals high in intelligence.  Most of the problems which I presented to my animals would be rated as difficult by psychologists, for as a rule they involved definite relations and demanded on the part of the subject both perception of a particular relation and the ability to remember or re-present it on occasion.