SystciiuUic and Sc.vual Relationship. 595 



apply the principle in those numerous cases in whicli they 

 fail. W'kh regard to the former point, it should he noted 

 that there are numerotis natural hybrids which cannot as 

 yet be made artificially, as for instance Kibes Cordunia- 

 nuni ; in other words that many possible hybridizations 

 do not succeed within the narrow limits of an experiment. 

 The impossibility of a successful cross can therefore 

 hardly ever be proved experimentally. W'itli regard t<> 

 the second point it suffices to cite the fact that in the 

 great majority of genera no specific hybrids exist at all ; 

 and that therefore here the delimitation of genera accor- 

 ding to this principle, would fail entirely. 



We come now to the species. Kolreuter expressed 

 the view that crosses wdthin the limits of these are fertile 

 and give fertile offspring; wdiereas crosses between si)e- 

 cies would either show a lessened fertility or at least 

 would prodtice infertile hybrids. Gartner and most of 

 the more recent investigators have subscribed to this 

 view, except that they regard diminished fertility, ratlier 

 than the actual absence of it, as the index of the bound- 

 aries of the species.^ 



But even to these generalizations the exceptions are 

 so numerous that unanimity in their ap])lication has not 

 yet been reached. The parallel between sexual affinity 

 and systematic relationship holds good in general, but 

 fails only too often in particular cases.- Naudix re- 

 garded these deviations from the rule as exceptions,-' and 

 Abbado and several other investigators have claimed the 

 determining cause of these exceptions in individual cases 

 to be the task of hybridological researches.-* 



'Gartner, loc. cit., pp. 163-164, 578-579. etc. 



"See MuRBECK, Botaniska Noiiscr, 1901. p. 214- 



'Ch. Naudin. UHyhridite dans les z'c^rtaux. 1869, p. MS- 



* Abbado, L'ihridismo nei vegctaVi, 1898, p. 48. 



