INTEGRATION OF MOVEMENTS IN LEARNING IN THE RAT 445 



an antithetical division of movements is a needless one. In a 

 functionally developed rat, every movement made, results from 

 an effective adjustment to the outside of the problem box, 

 inside of the hood which covers the box, or from an effective 

 adjustment leading to the raising of the latch. The neck reflex 

 movement which raises the latch does not, accordingly, appear 

 because of the production of a number of movements, for every 

 movement is effectively performed at some configurated part 

 of the problem box. When the functional condition of the rat 

 is not the best, then ineffective and effective adjustments to 

 configurated parts of the problem box are in evidence. These 

 different adjustments can not be regarded as antithetical sorts, 

 for both are not produced in all rats. Effective adjustment to 

 the problem box, or solving it, may or may not take place. 

 The conditions which determine ineffective adjustment can not 

 be adequately expressed by saying that the "successful" move- 

 ment did not appear, and that it appears only in a few cases. 



Present theories of learning can not account for the divergent 

 results obtained in the first trial. The most difficult thing for 

 the adherents to a definite theory of learning to explain, is the 

 number of times the problem box remains unsolved. In such 

 instances repeated touching of the underside of the latch evi- 

 dently has no definite effect on the nervous system, neither 

 increasing the so-called permeability of the synapses, nor select- 

 ing neural arcs. This fact seems to show conclusively that 

 definite sensory excitations are not so essential to the solving of 

 the problem in the first trial, as the working concept of " trial 

 and error" holds to be necessary to produce the " successful" 

 movement. 



In succeeding trials, when the problem is solved, an associa- 

 tion of the senses does not account for the results obtained. It 

 would be difficult to account for the fluctuations in the unequal 

 antagonistic action of the rat's limbs conditioning imperfect 

 responses by the assumption that definite sensory excitations 

 produce these fluctuations. The imperfect responses seem not 

 to be conditioned by definite sensory excitations, but by changes 

 in the nervous system which produce fluctuations in the rat's 



