II.] CONCEPTION OF "ACQUIRED CHARACTER." 77 



notion, while less clearly defined, was essentially the 

 same, and he collected a mass of evidence to show that 

 such characters are transmissible ; and he even went 

 further than Lamarck, and attempted to show that 

 mutilations may be hereditary. Weismann's early defi- 

 nition of acquired characters is plain enough. Such 

 characters, that is, the somatogenic, "not only include 

 the effects of mutilation, but the changes which follow 

 from increased or diminished performance of function, 

 and those which are directly due to nutrition and any of 

 the other external influences which act upon the body." 

 Standing fairly and squarely upon this definition, it is 

 easy enough to disprove it — that is, to show that some 

 characters thus acquired are hereditary. But the mo- 

 ment proofs are advanced, the definition is contracted, 

 and the Neo- Darwinians declare that the given character 

 was potentially present in the germ and was not prima- 

 rily superinduced by the external conditions — a position 

 which, while it allows of no proof, can neither be over- 

 thrown. A cow lost her left horn by suppuration, and 

 two of her calves had rudimentary left horns ; but Weis - 

 mann immediately says, " The loss of a cow's horn may 

 have arisen from a congenital malformation . ' ' Certainly ! 

 And it may not; and the presumption is that it did not. 

 A soldier loses his left eye by inflammation, and two of 

 his sons have defective left eyes. Now, "the soldier," 

 says Weismann, "did not lose his left eye because it 

 was injured, but because it was predisposed to become 

 diseased from the beginning, and readily became in- 

 flamed after a slight injury!" This gratuitous manner 

 of explaining away the recorded instances of the sup- 

 posed transmission of mutilations, and the like, is com- 

 mon with the Neo -Darwinians, but it must always create 



