484 BOTANICAL GAZETTE [DECEMBER 
species E. latifolium) the pollen showed a strong tendency to 
abortion in varying degrees. Conditions similar to those were 
observed also in E. Jatifolium, and more generally in the subgenus 
Lysimachion, where there was quite habitual tendency to abortion 
of the microspores. This situation is due, without doubt, to the 
more constant presence of species between which crosses are possible. 
In contrast to this condition, there is that presented by the 
2 essentially monotypic species of Gongylocar pus, the geographically 
limited Epilobium angustifolium, and the strictly monotypic 
Zauschneria californica. In these plants there appears no evidence 
of pollen degeneracy, which (in view of the general principle of 
sterility in hybrids) is precisely the condition one would expect. 
It is, nevertheless, just as significant that these species which are 
unable to cross should present uniformly well developed pollen 
grains. This generally fertile state is in striking contrast to a 
habitual degenerate condition where there is frequent interfertili- 
zation. Thus it is apparent that reduced fecundity is valuable 
evidence of hybrid derivation, while a uniform development of the 
microspores is, other things being equal, as equally indicative of 
uncontaminated parentage. 
From these studies of the Onagraceae in relation to geographical 
distribution, it is evident that impotent microspores are generally 
present in plants growing sufficiently near closely allied species to 
be the result of intercrossing. Opposed to this type of spore 
formation, there is a habitually perfect condition of the pollen 
chosen from plants growing in more isolated localities, and from 
monotypic species where interbreeding is manifestly impossible. 
Thus it is evident, from the morphological standpoint, that the 
Onagraceae intercross quite freely, and consequently there must 
be present many natural hybrids which greatly complicate genetical 
studies of species in this family. 
Conclusions 
1. It is apparent that abortive pollen, in a large though varying 
degree, is an indication of hybrid derivation. 
2. There are many recognized hybrid segregates which have 
remained constant and true to type. 
