560 THE CAUSES OF EVOLUTION 



fluctuations are not inherited, they may be weeded out by natural 

 selection and thus produce no results. Selection of mutations, on 

 the other hand, will result in the establishment of new types. The 

 matter is more difficult to study where the sexes are in separate 

 individuals and cross fertihzation must occur, as in most animals, 

 although many cases of this sort have been investigated in the 

 fruit-fly. Analysis can be more easily accompHshed in a self- 

 fertilizing species, Hke the bean plants studied by the Danish botan- 

 ist, Johannsen, or in a protozoan that reproduces asexually, like 

 the paramoecia, studied by Jennings (Fig. 308). 



Present Status of the Mutation Theory. — The Mutation 

 Theory, therefore, does not supplant the Theory of Natural 

 Selection. It shows the material upon which selection must work. 

 It is a s'tudy of the variation and heredity which are taken for 

 granted in the scheme of natural selection. In a concrete instance, 

 one might suppose selection to weed out, in each succeeding genera- 

 tion, the fluctuations that rendered individuals less fit to survive. 

 Under these circumstances no evolutionary change would be 

 effected, because such variations would not be inherited and the 

 individuals exhibiting them would give a similar population in 

 each generation, no matter how rigorous the selection. When, 

 however, a mutation occurred that was sufficiently advantageous, 

 selection would act upon the individuals exhibiting it, and upon 

 their descendants, and might estabHsh a new type. The older 

 type might be obhterated if it were at sufficient disadvantage in 

 competition with the new one; or the old and new types might 

 persist side by side, as often seems to be the case. In this manner 

 the appearance of mutations and their selection might bring about 

 evolution by a series of steps which Darwin would have considered 

 too large to be possible. We have seen that Darwin recognized the 

 existence of the extreme variations which, even in his day, were 

 known as mutations and called " sports " by breeders, but he did 

 not suppose them to be sufficiently frequent to be important for 

 evolution. 



In further support of this relationship between mutations and 

 selection, it may be said that many domesticated breeds seem to 

 have arisen from mutations that appeared fully formed, and the 

 descendants of which were selected with considerable inbreeding 

 until the type was established. The origin of the Ancon sheep, 

 a short-legged breed descended from a single ram that appeared 



