THE EXPERIMENTAL STUDY OF GENETIC RELATIONSHIPS I47 
its behavior cannot be held to throw light upon the behavior of wild 
plants. It has now been shown that wild species show absolutely 
parallel phenomena. 
It has also been said that the non-Mendelian behavior of Oenothera 
hybrids shows that the genus is so exceptional in its genetic behavior 
that it provides no basis for generalization. In reply we may say that 
the Mendelian school have in general confined their crosses to varieties 
of cultivated plants. If they had ventured into the field of inter- 
specific hybridization they would have found plenty of parallels to the 
behavior of Oenothera. As a case in point we may cite the experience 
of W. Neilson Jones, who obtained matroclinic reciprocal hybrids 
between species of Digitalis. He did not think the results quite com- 
parable to those of Oenothera for the reason that the Oenothera hybrids 
reported up to that time had been patroclinic. Now, however, this 
argument is removed by the discovery of matroclinic hybrids in Oeno- 
thera. The enormous literature of orchid hybridization contains 
frequent allusions to unlike reciprocal hybrids. In this largest family 
of the monocotyledons may be found numerous examples of both 
patroclinic and matroclinic hybrids. Many of them may prove to be 
cases of parthenogenesis, but the situation demands a much more 
thorough study than has yet been given it. The so-called "false 
hybrids" of Fragaria, as well as other Rosaceae, should be carefully 
investigated both cytologically and genetically. This much is sure; 
it is not yet time to speak of the universality of Mendelian phenomena, 
or of the exceptional nature of Oenothera. I am inclined to believe 
that such groups as the Orchidaceae may even provide parallels for the 
mutability of Oenothera. For example, Miss Pace has recently 
studied the cytology of Spiranthes cernua and S. gracilis from material 
collected near Chicago. Nine other species are interpolated between 
these two by Oakes Ames, in his Monograph of the American Species of 
Spiranthes. Yet they differ from one another in somewhat the same 
way that Oenothera Lamarckiana differs from Oe. gigas. Miss Pace 
finds 15 and 30 chromosomes in Spiranthes gracilis, as the reduced and 
unreduced numbers, but 30 and 60 in 5. cernua. The latter may well 
be, as Miss Pace intimates, a tetraploid form of the former. Here is a 
fertile field for the experimentalist. 
Jeffrey has lately argued that all plants which have any defective 
pollen grains are in a state of genetical impurity, and that any con- 
clusions drawn from their genetical behavior, in connection with the 
