300 BROOKLYN BOTANIC GARDEN MEMOIRS 
are not able to survive successfully they are found to the exclusion of 
one or both of these. This last conclusion upsets the conventional 
assumption that hybrids can only exist where their originating species 
occur side by side. It is clear from the general results of the highly 
important systematic and geographical investigations of Kerner that 
new species may appear as the result of spontaneous hybridization. 
The more recent evidence supplied by the investigations of Brainerd 
upon the violets and certain Rosaceae point equally positively in the 
same direction. This author has made it clear that a number of 
recognized species of Viola and Rubus are in reality hybrids in their 
origin. A particularly interesting result reached by Dr. Brainerd is 
that these hybrid species may become absolutely fixed in spite of their 
mode of origin and respond not only to recognized systematic but also 
to genetical criteria for species. 
It is too often assumed at the present time that the best criteria 
of species are physiological. On this basis the capacity to breed true 
in cultures and to produce offspring which comply with the tests of 
genetical analysis is regarded as of paramount importance. Since 
many known hybrids comply equally with recognized species with 
these tests it has become clear that what a plant does in cultures can- 
not be accepted as an infallible evidence of its antecedents. Where 
physiological criteria fail, we turn to the more constant ones furnished 
by morphological characters. It has been recognized for nearly a 
century that sterility is often a marked feature of hybrids, especially 
when they result from the crossing of somewhat incompatible species. 
The causes of incompatibility are apparently unknown as often 
species more different in their external characteristics and more 
widely separated in geographical range can be crossed with greater 
success than those nearly related on the evidence of external features 
and geographic coincidence. For example the horse and the zebra 
produce fertile hybrids, while the horse and the ass, when crossed, give 
rise ordinarily to infertile mules. Similarly our common canoe, yellow 
and black birches, which often grow side by side without hybridizing 
all apparently cross with a considerable degree of readiness with the 
more isolated shrubby birch of swamps, Betula pumila, according to 
the investigations of Jack and Rosendahl in this country. 
Hybrids may present in the case of plants a number of interesting 
morphological characteristics. For example we frequently find a 
high degree of imperfection in their gametic cells, male and female, 
with the emphasis of sterility nearly always on the male. This feature 
is often so marked that it is impossible to fertilize a hybrid with its 
own pollen, even when the ovules present a considerable degree of 
fertility. The morphological imperfection in pollen grains can obvi- 
