78 HARRY MILES JOHNSON 



too far from the free end to open the box with one effort. The 

 blind dogs at the first trial passed around the box, stopping to 

 scratch at the southeast end, and then trying the southeast 

 and northeast corners before attacking the latch at the south- 

 west. Both learned to stop at the southwest corner during the 

 first series of ten trials: Dog i at the third and Dog 2 at the 

 eighth trial. 



SUMMARY 



The above experiments have shown that it is impracticable 

 to attack the " Molyneux problem" by using dogs rendered 

 temporarily blind by this operation. 



It is evident from the records that vision is not necessary to 

 enable the dog to make quite complicated adjustments. The 

 behavior of my animals, particularly that of Dog 5 in the con- 

 trol tests indicates that the dog may make little use of vision. 

 His reaction to the rope and to the disoriented box strengthens 

 the belief that the normal as well as the blind dogs, in ordinary 

 situations rely greatly on kinaesthetic and muscular sense- 

 processes in making their adjustments. This accords with the 

 data accumulated by Watson, Richardson and Vincent in the rat. 



The rate and methods of learning in blind and normal dogs 

 shows sui-prisingly little difference. 



Dogs given only ten trials a day required fewer trials for 

 learning, on the average, than those given twenty trials. This 

 fact suggests the question, what is the optimal number of daily 

 trials to educe perfect habits with the least effort? 



