appearance; of ne:w characters. 271 



In as far as we can judge from these experiments, while the selection of 

 the inheritable variations when continued for ten or twelve generations is able 

 to give well-defined races which persist as long as selection is practiced, it 

 is not able to establish these races so that they can exist alone. The results 

 from the experiments fall in line with the general experience of breeders of 

 domesticated plants and animals. In these beetles, as De Vries also points 

 out in plants, selection is able to produce races by isolation only (selection) 

 of extreme heritable individual variations. 



I am aware that against this view the selectionists will at once say that far 

 more generations are needed, which I grant is possibly true ; but how many ? 

 I have shown that geographical races of this beetle living in places where 

 there have been the same environmental and local selective tendencies for at 

 least 400 or 500 generations have not yet become permanently m.odified in 

 form in such locations. It is true that a large part of geographical variation 

 is purely somatic, but not all ; hence, if local selective influences are so power- 

 ful, it would seem that a series of several hundred generations were ample 

 time for selection to begin to show its action. In experimental breeding 

 selection works rapidly, and in nature why should it not act with equal rapid- 

 ity — that is, where it is operative at all ? In any event, until there is evidence 

 to the contrary it will not be unfair to hold that pedigree cultures of ten to 

 fifteen generations do show fairly well how potent local or artificial selection 

 really is as a means of modification in these beetles. We must not confuse 

 local or artificial selection (isolation) with the specific selection characteristic 

 of the entire species. The former is productive of divergence, the latter of 

 conservation, and my experiments are in reality a contest between the two 

 with the conservative one victor in the end. 



EXPERIMDNTAI, PEDIGREE BREEDING OF NEW CHARACTERS AND SPECIES. 



There have arisen in my experiments, and I have found in nature, varia- 

 tions of L. decemlineata which dififered from the parent stocks in one or 

 many characters and stood beyond the limits of fluctuating variations. These 

 sports — saltatory variations, discontinuous variations, or "mutations," what- 

 ever we may call them — often bred true, and were able in many instances, but 

 not in all, to hand on their variations in full intensity. Whenever found 

 these interesting variations have been seized upon and made the basis of 

 instructive experimental cultures. 



It has been shown that in nature three well-defined species of Leptinotarsa — 

 melanothorax , ruhicimda, and angustovittata — arise through rapid transfor- 

 mation, and that mbicunda is possibly becoming established as a permanent 

 species. In treating of variations I have figured under the head of extreme 

 variations some of those observed to have arisen from decemlineata. 



20— T 



