Systcjiiafic and Sexual Relationship. 597 



species of Diplaciis {Miumlns) do not seem to l)c luur- 

 phologically more remote from one anotlier than Tro- 

 paeolwn uiajiis and Tr. inimis, Nicotiaiia latissima and 

 A^. Marylandica, N. nistica and iV. Texana or Pisinn sa- 

 tiviiiii and P. arvcnsc. Nevertheless the results iA cross- 

 ing" in the former cases exhibit all the characters of hy- 

 brids, but in the latter those of mongrels." Focke sum- 

 marizes his conclusions on this subject in the thesis that 

 ''systematically probable" crosses often miscarry, whilst 

 improbable ones sometimes succeed {loe. cit., p. 457). 



Even regarded from this point of view the difference 

 betw^een IMendelian and uni-sexual crosses offers itself 

 as a criterion for distinguishing betw^een species and 

 varieties. 



But sexual afifinitv does not ahvavs irive reliable in- 

 dications. In the first place Gartner frecjuently found 

 that fertility, as measured by the number of seeds ripen- 

 ing in a capsule, is different in reciprocal crosses. This 

 shows that it is not merely determined by the degree of 

 relationship, but obviously by some other causes besides. 

 The best known of these is the length of the style; and 

 the recent investigations of Burck on the concentration 

 and stimulating properties of the fluid secreted by the 

 stigma have thrown much valuable light on this subject.^ 

 In extreme cases one of the crosses succeeds well, whilst 

 the other does not at all, as for instance Mirabilis Jala pa 

 X longiflova, Geuin urhanum X nrale, Sophronitis X 

 Cattleya, and so forth. In the second place some crosses 

 do not succeed in spite of a very close apparent relation- 

 ship, as for instance betw^een Anagallis arreiisis and eoe- 

 rulea (Gartner). 



^W. Burck. Over de bezve^iu^ der sfempcis by M'unulus en To- 

 renia. Sitzunp^sber. d. Kon. Xkad. d. Wet., Amsterdam. 1901, and 

 in previous articles. 



