58 DISCONTINUOUS VARIATION. 



abnormalities if they possess them, but this prepotency 

 is quite limited in its range. Darwin in his " Variation 

 of Animals and Plants under Domestication " * men- 

 tions an instance of transmission of supernumerary 

 digits through five generations, whilst in other cases 

 they have reappeared after an interval of even three 

 generations. " But," says Darwin, " we must not over- 

 estimate the force of inheritance. Dr. Struthers as- 

 serts that cases of non-inheritance and of the first ap- 

 pearance of additional digits in unaffected families are 

 much more frequent than cases of inheritance." 



Unless much stronger evidence than that hitherto ad- 

 vanced be obtained, it therefore follows that, according 

 to the known laws of inheritance, suddenly occurring 

 variations, unless artificially selected, must inevitably 

 be swamped by intercrossing and disappear. Suppos- 

 ing, on the other hand, any such variation is artificially 

 isolated, and bred in and in with its own offspring, then 

 it may be possible to establish a distinct race, bearing in 

 undiminished degree all the abnormal characteristics of 

 the original variety. For instance, Darwin thus 

 records the origin of the ancon sheep :f "In 1791 a 

 ram-lamb was born in Massachusetts, having short 

 crooked legs and a long back, like a turnspit dog. 

 From this one lamb the otter or ancon semi-monstrous 

 breed was raised ; as these sheep could not leap over the 

 fences, it was thought that they would be valuable. 

 The sheep are remarkable for transmitting their char- 

 acter so truly that Colonel Humphreys never heard of 

 ' but one questionable case ' of an ancon ram and ewe 



*Vol. i. p. 457, Ed. ii. 

 \Ibid., vol. i. p. 104. 



