8 HEREDITY, AND THE ORIGIN OF SPECIES. 
stable combination of the characters of the parental forms 
in such manner as to give rise to a new type unlike either 
of the parents, variously intermediate and constant to the 
new type in succeeding generations. More than a thou- 
sand such fixed hybrids, or hybrid species are known, some 
of which have been formed anew experimentally and are 
thus beyond doubt. It is to be seen therefore that hybridi- 
zation has played, and is playing, no small part in the com- 
position of the flora of the earth, and that it must be con- 
sidered as an active, and not unimportant factor in the 
evolution of plants. 
It must be admitted at once that such fixed hybrids are 
exceptions to accepted principles of dominance, and of the 
purity of the germ-cells, as the fixed hybrids depart en- 
tirely from the Mendelian procedure. As a further compli- 
cation we have in some instances a hybrid progeny com- 
prising several fixed types, each constant to itself. The 
only possible explanation of such phenomena in the simpler 
case calls for a fusion of the germ-cells and a double 
longitudinal division of chromosomes. 
Such splittings are qualitatively unsymmetrical, since 
the fixed hybrids are variously intermediate between the 
parents, sometimes being goneoclinic to the one furnishing 
the pollen, and in other cases to the one bearing the egg- 
cell. When it comes to the composition of a hybrid pro- 
geny which comprises several fixed types in the first gene- 
ration, breeding true in succeeding generations, we can 
only look to the researches of Wilson and others upon the 
quantitative and qualitative inequality of the chromosomes 
to furnish us the proper clue. It is only fair to say that no 
investigation of such cases has yet been made. Now as to 
a third category in which the hybrid is polytypic in the 
first generation, and some or all of the types may split in 
the second generation into forms, which may or not be 
identical with the grandparents, we evidently have a com- 
