358 THE INFLUENCE OF LIVING SUKKODNDINGS. 



of new characters in addition to sucli as are already present in 

 the parents ; but these too are rendered to a certain degree fluc- 

 tuating, so tliat any exact, or even approximate, repetition of them 

 in the offspring is put out of the question. The rigid constancy 

 of the parental characters maintained by interbreeding is so 

 completely interrupted by hybridisation that the organisms 

 thereby exposed to the means of selection acting on the young 

 animals are to a certain extent perfectly new. Certainly, we as 

 yet know very little concerning the mode in which hybridisation 

 affects any other characters than the colouring ; the comparative 

 diihculty of breeding such hybrids, and the short series of years 

 that have elapsed since more attention has been paid to such 

 cases than formerly, leave no room for surprise that available 

 material is so extremely scanty in this direction. But this 

 cannot alter the results that have been attained so far; for if 

 after longer investigation we should be brought to the conclu- 

 sion that in the animals at our command for expeiiment the 

 colouring of the skin, or of its covering, still seems especially 

 adapted to exhibit the effects of hybridisation, while other organs 

 as the skeleton — for instance — are not affected at all, or in a very 

 insignificant degree, neither of the principles above laid down 

 would be disproved, but merely restricted in their application. 



If we now compare and contrast inbreedurg and hybridisa- 

 tion — cross-breeding — it is well known that the very essence of 

 the former process is the union of very closely allied individuals 

 of the same race or species, while that of the latter, on the con- 

 trary, is the union of individuals very distinct from each other. 

 And since this contrast, so far as we learn from the numerous 

 experiments now within our knowledge, points to one and the 

 same cause as that by which the difference in the results of 

 the fertile union is determined in each case, v,e may deduce 

 another general principle which ^xill lead us to still wider con- 

 clusions : namely, that the more remote the systematic affinity 

 is of two animals that unite to pioduce young, the greater is the 

 probability that, together with a pei-fectly undetermined mix- 

 ture of the parental characters, new characters may arise which 

 do not occur in either i^aient. The cause of the disturbance 

 thus arising in the constancy of the specific character is the act 



