Distinction Between Species and Varieties. 
It may, however, also happen that in forms which 
differ from one another in two points only, one of these 
would give a bi-sexual union, whilst the other would lead 
to a uni-sexual cross. Such is the often cited instance of 
L\clinis I'cspertina X dinrna. One of the characters 
A'oulcl follow MENDEL'S law in crosses, whilst the other 
would tend to produce a constant intermediate form. 1 
In this case, according to the former character, one parent 
would be related to the other as a derivate variety, but ac- 
cording to the latter as a homonomous elementary species. 
Exactly the same, though to a greater extent, must 
be true of tri-polyhybrids. The points of difference be- 
tween their parents can be all uni-sexual or all bi-sexual, 
or some of them uni-sexual and others bi-sexual. In 
the first case the parents are to be considered as ele- 
mentary species, in the second as varieties; in the third, 
however, the principle affords no decision. 
It is just this case which appears to be the commonest 
in nature. In experiments in hybridization, we must, if 
we wish at all to elucidate the laws to which the results 
conform, confine our attention to certain points of differ- 
ence and leave all the rest out of consideration as of 
subordinate importance. MENDEL did this in his experi- 
ments with peas, and the same has to be done in crossing 
Maize, the races of which do not differ from one another 
exclusively in varietal characters ; so also in Lychnis ves- 
pertina and dinrna and in many other cases. 
It would take us too long to continue this discussion 
further, and to accumulate examples; my meaning will, 
I believe, be sufficiently clear. It may be expressed in 
*I call to mind the Oenothera Pohh'ana (O. Iatay,brevistylis} , one 
of my crosses, the /afa-character of which behaves as a mutational 
character in a cross ; whilst the shortness of the stamen behaves as 
a Mendelian character. 
