342 INTRODUCTION TO EVOLUTION 



individual during its lifetime inherited by its offspring? Our everyday ob- 

 servations would lead us to answer in the negative. It is fortunate, for 

 example, that mutilations are not inherited. A man who loses a finger need 

 not fear to father a family lest his children be born with fewer than the 

 usual number of digits. Weismann, the German critic of Lamarckism, re- 

 moved the tails from mice for many generations. Of course the mice in the 

 last generation were born with as long tails as the mice of the first genera- 

 tion had had at birth. Such evidences against Lamarckism have been 

 criticized on the ground that the mutilation is something done to an animal, 

 something in which the animal does not participate actively. Developments 

 produced by the activity of the individual do not seem to offer more con- 

 vincing evidence in support of Lamarckism, however. If our hypothetical 

 college sprinter continues training after he leaves college and marries will 

 his sons be born with more highly developed leg muscles than they would 

 otherwise have had? We know that they will not. Is the son of a concert 

 pianist bom with more skill in his fingers than would have been his if his 

 father had been a lawyer? While we know that aptitude for music is in- 

 herited, we are equally certain that the son will have to begin with simple 

 finger exercises just as his father did before him, despite the years of train- 

 ing the father received before the son's birth. 



In this connection we may well consider a series of experiments which 

 have been interpreted as demonstrating inheritance of effect of training. 

 McDougall (1938) trained white rats in performance of a simple problem: 

 escape from a tank of water following a certain route. The trained rats were 

 mated and offspring were raised. The offspring, in turn, were taught the 

 problem; they then became the parents of a third generation. And so on for 

 forty-five generations. McDougall found marked and progressive decrease 

 in number of errors made in learning the problem as generation followed 

 generation. Taken at its face value, this finding would seem to indicate that 

 offspring were really profiting from training given their parents, that an 

 acquired character (training) was being inherited. 



This experiment has been repeated by Agar and his colleagues (1954), 

 whose final report records results for fifty generations, covering a period of 

 twenty years. They started with a single pair of albino rats of the Wistar 

 strain. The offspring from this pair were divided into two groups. Members 

 of one group were trained and then used as parents of a second generation. 

 Members of the other group were not trained but were used as parents to 

 start a control line running parallel with the trained line. In each generation 

 some of the rats in the trained line were trained, and then mated to produce 



