EVOLUTION looi 



rushing down the stair or climbing down the wire-netting, that there 

 could be no doubt as to their associating bell and food. The females 

 learned a little more quickly, showing first signs of recognition after 

 twenty-four meals, and learning definitely after thirty-seven lessons. 

 What a contrast to Pavlov's three hundred! It need hardly be said 

 that the mice did not see their teacher at meal-times. 



A hundred and forty meals were given, but after the fortieth there 

 was no appreciable difference in the manner or the average rate of 

 the response. But there were marked occasional variations, apparently 

 connected with appetite, and some individuals seemed more alert 

 than others. These experiments began with twelve males and eleven 

 females, and were continued for about two and a half months, when 

 the numbers had been reduced by disease and accidents to seven 

 and eight respectively. The bell-ringing then ceased, and the mice 

 were paired, yielding eventually a first filial generation of seven 

 males and 14 females. In due time these were ready for lessons 

 exactly similar to those their parents had received. Miss Dean found 

 that the first hints of recognition appeared in both sexes at the 

 twenty-sixth meal ; and that all the males had learned after forty-four 

 meals, and all the females after forty-six, rather longer than before ! 

 Thus the net result was negative; the rate of learning had not 

 changed appreciably. But no one will attach great importance to 

 experiments extending over no more than two generations. 



This brings us to the experiments on white rats, begun in 1920 

 by MacDougall, the well-known psychologist, at that time in 

 Harvard. He does not claim to have established the validity of the 

 Lamarckian principle — i.e. the transmissibility of representative 

 results of acquired modifications, but he thinks that his facts point 

 towards this conclusion. His first notion was to train the rats to run 

 in a hollow wheel like a squirrel-cage, which included an electrified 

 section whose mild shock could be evaded by a jump. The rats soon 

 learned to run the wheel smoothly and to leap prettily across the 

 electrified section. But in the following generations there was any- 

 thing but improvement. A tendency developed to dash up the 

 sidewall instead of leaping. Though the bolder spirits would still leap, 

 there seemed to be a growing timidity or hereditary phobia, which 

 rather spoilt the experiment. 



The next experiment was made with an ingeniously contrived 

 tank, so adjusted that a rat placed in the water could find its way 

 out by either of two routes — one gangway (b) electrified and illu- 

 mined, the other {a) with free egress, but not lighted up. The question 

 was how long the rats would take to learn to avoid the less desirable 

 way out, and whether the number of trials would diminish in the 

 course of generations of training. Sometimes the bright electrified 

 gangway was alternated in position with the dim safe gangway, and 

 untrained rats were of course used as controls. The general result, 



