278 PRODUCTION OF RACES AND SPECIES IN LEPTINOTARSA. 



become established. My original material in this series of cultures came 

 from potato fields widely separated at Clifton, Ohio, so that had I not gath- 

 ered and cared for the material nothing would have come of it. I doubt if 

 it is possible in the light of what we know of the natural history of decem- 

 lineata for the species pallida to become established as an independent species. 

 L. deccmlineata, although the parent of many of these variations, gives them 

 in nature so rarely that the collection of immense numbers is necessary to 

 discover even single specimens. L. deccmlineata differs widely in this 

 respect from many plants, as also from its tropical relatives. 



In the selection experiments with melanism and albinism a single male of 

 the mutant pallida appeared in the sixth generation. As far as could be deter- 

 mined this specimen was just like those from nature. It was crossed with two 

 female deccmlineata of modal color. The offspring of this cross, when reared 

 under normal conditions of existence, gave a hybrid progeny in which deccm- 

 lineata was dominant. These were bred inter se, and gave: (A) L. decem- 

 lincata, 29 male and 33 female (62), and (B) L. pallida, 11 male and 8 

 female (19), without any intermediates. From these cultures were again 

 made, and from (A) I obtained of (A'), L. decemlineata, 62 male and 71 

 female (133), (B'), L. pallida, 54 male and 67 female (121) ; and from (B) 

 I obtained L. pallida, 46 male and 44 female (90), and no trace of other con- 

 ditions. (B) was now reared through six lineal generations, and gave pure 

 pallida only. The (A) series was also reared through six generations, giv- 

 ing in each generation the two forms, pallida and deccmlineata, the pallida 

 breeding true to type where isolated. I did not try in this series the experi- 

 ment of rearing pallida along with decemlineata in a state of nature, owing to 

 lack of room, but I can see no reason why it should have behaved differently. 



The general behavior of this series I have put in the form of a diagram 

 (text-fig. 20), which shows the history of each generation in the experiment. 



In this series, as in the previous set, decemlineata is the dominant member 

 of the cross in the hybrid generation, but in the succeeding generation we get 

 a Mendelian separation, the two forms appearing in each subsequent genera- 

 tion. The chief point of interest to us is the further demonstration of the 

 fixity of type in pallida, no matter whether we obtain it from nature or from 

 material in captivity. This point, which is an important one, is for pallida 

 sufficiently demonstrated by the two sets of cultures described. Thus far we 

 have considered only one extreme variation, but before attempting to arrive 

 at any final decision we must examine the cultures made from other variations 

 from decemlineata. 



One not quite as common as pallida, and differing from it almost diametric- 

 ally, is melanicum. Although one of the commonest of the variations of this 

 class arising from decemlineata, it is one of the most difficult to propagate. 

 Out of the 3L specimens found in nature, 20 were males, and these were 



