The Outline of Science 



to waltz round and round in circles without sufficient cause and 

 to trip sideways towards its dormitory instead of proceeding in 

 the orthodox head-on fashion. But this freak is a very educable 

 creature, as Professor Yerkes has shown. In a careful way he 

 confronted his mouse-pupil with alternative pathways marked by 

 different degrees of illumination, or by different colours. If the 

 mouse chose compartment A, it found a clear passage direct to 

 its nest; if it chose compartment B, it was punished by a mild 

 electric shock and it had to take a roundabout road home. Need- 

 less to say, the A compartment was sometimes to the right hand, 

 sometimes to the left, else mere position would have been a guide. 

 The experiments showed that the dancing mice learn to discrimi- 

 nate the right path from the wrong, and similar results have been 

 got from other mammals, such as rats and squirrels. There is no 

 proof of learning by ideas, but there is proof of learning by 

 experience. And the same must be true in wild life. 



Many mammals, such as cats and rats, learn how to manipu- 

 late puzzle-boxes and how to get at the treasure at the heart of a 

 Hampton Court maze. Some of the puzzle-boxes, with a reward 

 of food inside, are quite difficult, for the various bolts and bars 

 have to be dealt with in a particular order, and yet many mammals 

 master the problem. What is plain is that they gradually elimi- 

 nate useless movements, that they make fewer and fewer mistakes, 

 that they eventually succeed, and that they register the solution 

 within themselves so that it remains with them for a time. It 

 looks a little like the behaviour of a man who learns a game of 

 skill without thinking. It is a learning by experience, not by 

 ideas or reflection. Thus it is very difficult to suppose that a rat 

 or a cat could form any idea or even picture of the Hampton 

 Court maze which they nevertheless master. 



Learning Tricks 



Given sufficient inducement many of the cleverer mammals 

 will learn to do very sensible things, and no one is wise enough to 



