132 K. S. LASHLEY AND S. I. FRANZ 



of the cortex is inadequate. From experiments 29, 31, and 33 

 in which animals were overtrained it seems that an amount of 

 overtraining equal to three or four times that required for learn- 

 ing is insufficient to reduce the habit to lower brain levels. The 

 difference in the complexity of the maze and inclined-plane box 

 habits and the apparent failure of the latter to be reduced to 

 subcortical levels within the limits of these experiments suggest 

 that the amount of practice necessary for the assumption of the 

 cortical functions by subcortical ganglia is proportional to the 

 complexity of the habit. Until further evidence is accumulated 

 this can not be considered more than a probable assumption. 



The data presented in section IV shows that complete destruc- 

 tion of the frontal pole results in the loss of the plane-box habit, 

 whereas the loss of the temporal and parietal regions is without 

 effect upon the habit. This indicates that conduction path- 

 ways involved in the performance of the habit pass through the 

 frontal pole, but an attempt to localize these more accurately 

 brings to light a complexity of function which has not before 

 been suggested. Partial destruction of the frontal pole did not 

 always result in the loss of the habit. An attempt to find a 

 correlation between the part destroyed and the retention or loss 

 of the habit revealed the further fact that no single part of the 

 frontal pole escaped destruction in all the animals which retained 

 the habit (fig. 14). It seems then that for retention some part 

 of the frontal pole must be preserved but no particular part 

 seems necessary. This is, perhaps, what might be expected if 

 we abandon the purely diagrammatic concept of the reflex cortical 

 arc as a single chain of neurones and consider that for the per- 

 formance of even a simple movement a number of such arcs are 

 required. With a vast number involved in complex habitual 

 acts, it is very improbable that all would be projected on a re- 

 stricted area of the cortex; the experimental results suggest 

 rather that in their cortical relations these arcs are widely dis- 

 tributed over the frontal pole so that a partial lesion results in 

 only the partial destruction of the arcs involved in the perform- 

 ance of any simple act. 



The experiments of series V have shown that the rat may form 



