262 PRODUCTION OF RACDS AND SPIJCIIiS IN L^PTINOTARSA. 



shown that color variations produced by environmental conditions during the 

 larval and pupal periods are not inherited at all. 



In another part of this experiment a single pair of beetles gave off- 

 spring of two sorts — one with the extreme conditions of the parents, the other 

 with the conditions of the first set. The parents of this series were from 

 the same stock as that which gave entire negative results, and were indistin- 

 guishable from several other like pairs that were not able to transmit their 

 peculiarity to the next generation. There is, therefore, a difference in the 

 variations experimented with, one being inheritable, the other not. The rela- 

 tionship and behavior of these two series I have represented on plate 25. In 

 this plate, in which the curves of the selected parents and offspring are 

 brought together, it is clearly demonstrated in the series A that there is no 

 accumulation of the character through selection. The series B, however, 

 which arises from a single pair indistinguishable from the rest of the parents 

 of generation 5, shows two distinct groups of offspring. These were isolated 

 and reared in several succeeding generations. One of these groups (5') was 

 always strongly like the parents, and showed little or no tendency to revert to 

 the racial standard as did the individuals of B" . In the pair that were the 

 parents of series B, one or both had the selected variation of transmissible 

 quality, and passed it on to the following generation, but to part only. It 

 would appear, therefore, either that one parent had the character selected 

 transmissible and the other not, and that part of the offspring received the 

 transmitted character in inheritable form and part did not, or it is possible 

 that both parents had the character latent in some germ cells and not in others, 

 and that in reproduction there was a union of like germs. I believe that the 

 first proposition is the correct one ; but which parent possessed the character 

 in transmissible form I do not know in this case, and for this series of experi- 

 ments this is of little importance. 



A glance at plate 25 shows that the series of cultures B' are very different 

 fromi A or B'^ in that in B' selection was able to keep offspring up to the 

 parental mode, and to raise the mode of the race. In the others, however, 

 i. e., B^' and A, this was not possible, nor was I able to maintain the same, 

 nor approximately the same, modal standard for parents and offspring. 

 In B' , however, I was not able to carry the selective development of the race 

 beyond the usual range of variations of the species ; that is, within the num- 

 ber of generations experimented upon in B' , and the same result was obtained 

 in six other identical series. The limits of variability of the species seem to 

 be limits beyond which artificial selective processes are unable to carry the 

 modification of the race. On the other hand, by selection we can rapidly 

 create races up to the natural limit of variation of the parent form, and main- 

 tain them there with ease by the same process, and I can see no reason why 

 artificial selection should not be able to carry the species beyond its natural 



