48 BOTANICAL GAZETTE [JANUARY 
other parent, a guess which might be fortified to some extent, perhaps, 
by the facts of distribution. In certain meristic qualities, such as 
the rate and total amount of growth, the hybrid may even exceed that 
of either parent, or be less than either. 
The mistaken impression prevails that hybrids bear defective 
pollen, but this is generally found to be the case only when the cross 
is unbalanced in many characters, particularly those appertaining 
to the reproductive functions. The infertility of many animal hybrids 
has also strengthened the assumption that plant hybrids share this 
defect. Many hybrids are quite as prolific in mature seeds as either 
parent, while in some that are offered by the seedsman and nursery- 
man they are claimed to excel in this respect. 
In any case where the comparative anatomical method is used, 
care must be taken to make observations upon material from similar 
stages of development. This is forcibly impressed upon one after 
following the growth of a hybrid which in the seedling stages shows 
a predominance of the qualities and anatomical characters of one 
parent, a different arrangement in the adult shoot, and a still different 
balance in the flower and fruit. 
The third method of study of a supposed hybrid is one which 
involves pure cultures of its progeny for one or two generations. 
If it should be a fixed hybrid no results will be secured which will 
be of value in the solution of the problem, since, so far as any facts 
offered by such cultures are concerned, the plant behaves as any 
other species. Presumably most of the species of suspected hybrid 
origin are of this character, but some of them undoubtedly will be 
found to be constantly re-formed and to offer alternative inheritance, 
and hence this test should be applied whenever practicable. 
If the supposed parents differ in but one or a few characters, 
and the hybrid shows alternative inheritance, the solution of the 
main question lies near at hand. It is not such simple questions as 
this however that we are usually called upon to solve. The re 
difficulties lie in the hybrids with the component qualities in stable 
combination, making a fixed hybrid, and with the forms which exhibit 
an interlocked combination of the ancestral characters in the first 
generation which resolve into the possible combinations of the dis- 
similar characters in the second, by which an enormous range of indi- 
