256 The Mechanism of Evolution in Leptinotaesa 



huge majority of cases corresponds to the upper limits of the natural " vari- 

 ation " of the species, does not transcend the natural variability, although a 

 fairly stable race is often capable of being created at this upper limit or at any 

 other point by this process. 



Of late years it has become the custom to interpret any such result on the 

 assumption that the race obtained is an isolated genotype or biotype. Is this 

 the truth of the matter ? It is the experience of so many observers that these 

 selected races revert to the parental standard when selection has ceased that 

 there must be some foundation in truth, and it must represent some result of 

 natural operations. Isolated genotypes or biotypes are asserted not to change, 

 to be difficult of modification, and Jennings says no one has thus far been able to 

 modify one of these isolated groups into another. It is not likely that intercross- 

 ing with other biotypes is solely responsible for this regression, and why should 

 the "selected" group so constantly move back to the mediocre of the entire 

 population ? Why this submergence of an improved, " efficient " (?) race by the 

 mediocre. 



In the population of L. muUitwniata I have shown that biotypes of the prono- 

 tal color-pattern are present, have been separated, and the distinguishing charac- 

 ters of these have been given. In the general population I can easily create a 

 race at any of the extremes of the natural variability by the use of quantitative 

 accumulation. The relation of this to the biotype question is not as simple as 

 one might suppose and will be considered later. In the isolated biotype, what 

 can quantitative accumulation accomplish ? 



A good test of this can be made by starting with biotype 4 and trying to 

 accumulate pigment and produce an end-result like biotype 1. Three sources 

 of error must be rigorously eliminated : (1) it must be certain that the biotype 

 4 is pure, contains only biotype 4; (2) the culture must be kept under uniform 

 conditions to prevent any action of external agents upon the pattern; (3) the 

 choice must be made upon the amount of pigment, and no attention paid to the 

 fluctuations in the pattern type. The problem is, can accumulation of pig- 

 ment or any other character produce directly or indirectly a change of type? 

 I have made many tests of this kind, of which a good example is the following, 

 wherein the attempt was made to change biotype 4 into biotype 1. 



Start was made upon a pure extracted biotype 4 stock derived from parents 

 which were obtained in nature at Chapultepec, and which in Fj gave only bio- 

 types 4, 5, and 6. In Fj pure-breeding examples (homozygous) of biotype 4 

 were found and these were carefully pedigreed through Fj, F^, F^, and F,, and 

 from the latter generation the basis of this culture came, while the stock was 

 continued until F^, when it was accidentally lost. It did not show any diver- 

 gent individuals or tendency to be transmuted iato any other type. 



Three males and three females of normal average aspect were mated, and gave 

 a progeny, Fj, of 143, distributed as shown in figure 42. From these, 4 males 

 and 4 females of the most extreme class were mated and gave a progeny of 79 

 individuals, with the distribution shown in figure 42, of which the upper range 

 was above the normal range of the biotype. In F^, F,, and F4 the same method 

 was followed of mating from the extreme upper class, and this rapidly caused 

 the movement of the population towards a condition of extreme pigmentation, 

 imtil in F^ the upper limit had reached a state just below that of biotype 1. 



