230 — : Rhodora [DECEMBER 
formed by such irregular division are aborted. Thus there is a se- 
lective elimination of potential gametes. Further, many of the pollen 
grains, and possibly the egg cells, are not functional. Again, an 
extremely large percentage — often the majority — of the zygotes 
formed are not viable. This will be apparent to anyone who takes 
the trouble to make germination tests of Oenothera seeds. Even 
after discarding the many seeds that casual examination shows to be 
worthless, it is seldom that over sixty percent of the remainder pro- 
duce mature plants, and the germination may drop as low as five per- 
cent. Now, it is quite likely that the only useful laws of heredity 
will be those which like the laws of physies and chemistry are mathe- 
matical descriptions of cycles of events from which predictions of 
what must occur under like circumstances may be made. Is it strange 
then, that many biologists are cautious when asked to accept as a basis 
for such descriptions, breeding results from plants like the Oenotheras 
where only a small portion of the facts can be known owing to the 
immense number of potential plants lost through the abortion of both 
zygotes and gametes? It is like asking a chemist to accept theories 
as to the structural formulae of organic compounds upon which only 
determinations of nitrogen and oxygen have been made. 
These facts are boldly disregarded by the author in his chapters 
on “Hybridisation and Hereditary Behavior” and “The Relation 
between Hybridisation and Mutation." "The Oenotheras, he says, 
have four main types of hereditary behavior: ‘ (1) mutation crosses, 
(2) Mendelian splitting, (3) blending and modification of characters, 
and (4) twin hybrids." Much of the discussion under these heads is a 
severe arraignment of Mendelism, but the author’s contempt for the 
Mendelian theory of heredity is not that bred of familiarity. If the 
reviewer has not misconceived matters, the author’s idea of Mendelian 
segregation is enticingly simple. If when two organisms are crossed, 
The F, generation is uniform and the Fs generation comprises two 
types in the ratio of three to one, the inheritance is Mendelian. No 
circumstance whatever may modify the definiteness of these phenom- 
ena under pain of their disqualification as examples of Mendelian 
inheritance. 
It would be rash to assert that Mendelism even in the broadened 
sense with which the word is used today, covers all types of inheri- 
tance. Correns’ and Baur’s experiments on chromatophores and 
Goldschmidt's work on the gypsy moth indicate the possibility of 
inheritance through the cytoplasm, while Mendelian data parallel 
chromosome distribution. 
On the other hand, it is even more bold to assert that inheritance in 
Oenothera is definitely nonmendelian. 'They cannot be placed in 
either category with certainty, but it ought to be emphasized that no 
single fact discovered by those who have made pedigree cultures of 
the group, precludes a Mendelian interpretation. Gates’ arguments 
against Mendelian interpretation of heredity in the evening primroses 
