62 Life and Death, Heredity and Evolution 



is to pick out races that already have these peculiarities. 

 This is unquestionably the explanation of the effects of 

 selection in by far the greater number of experiments or 

 observations where it is found to have an effect. 



Such results were long interpreted as showing actual steps 

 in evolution; by selection changes in hereditary characters 

 were thought to be produced, and new hereditary characters 

 obtained. On the basis of such interpretations the rate of 

 change through selection, the rate of evolution, was 

 measured. 



But what really occurs in all such cases is a gradual 

 picking out of existing races with certain characteristics 

 (in our example, with numerous spines), and discarding the 

 rest. So far as the selection is based on these differences 

 between preexisting stocks, no evolutionary change has been 

 produced or measured. 



But can selection do nothing more than this? Are there 

 no actual changes in hereditary characters? Are these 

 diverse races really permanent in their hereditary char- 

 acters? Or do changes gradually occur even within such 

 stocks? How do these diverse races happen to exist? Can 

 several diverse races be produced from a single one? 



On this question a great deal of work has been done in 

 the unicellular forms ; we shall examine the results. 



One of the animals most studied from this point of view is 

 Paramecium. If we take a wild set of these animals, it is 

 easy, as we have seen, to bring about differences by selection, 

 for all we have to do is to pick out such as we please of the 

 different races that exist. But suppose we take but a single 

 race ; suppose we begin with a single individual and get our 

 entire population from this one parent. Shall we then be 

 able by selection to bring about hereditary differences? In 

 other words, do any hereditary changes occur in such a 

 single race? This question is evidently the fundamental one 



