GENETICS: CLAUSEN AND GOODSPEED 243 
functional ovules, the number of which is apparently constant within 
rather narrow limits. Assuming that segregation and recombination 
take place normally and in accordance with the chromosome view of 
heredity, these functional ovules represent the Tahacum and syhestris 
extremes of a recombination series, the vast majority of the members of 
which fail to function because they are built up from incompatible 
elements derived from both systems. The evidence for such a con- 
stitution of the functional ovules is furnished by the results of back- 
crosses of the hybrid with the parents. When the Fi hybrid is crossed 
back with syhestris, the progeny consists of abnormal, sterile individuals 
and a few typical syhestris individuals which are completely fertile and 
breed true. On the other hand when the Tahacum parent is crossed 
back onto the Fi hybrid the progeny consists of Tahacum forms some of 
which are completely fertile and others of which are sterile like the Fi 
hybrids. The hereditary phenomena, therefore, displayed by these Fi 
species hybrids confirm the conception that they represent a contrast 
between reaction systems, the elements of which display a considerable 
degree of mutual incompatibility. It follows, then, that the type of 
behavior displayed by species hybrids may be considered as dependent 
upon the degree of incompatibility of the elements of the reaction systems 
therein involved. Sterility in such cases is merely a logical consequence 
of this same incompatibility, and the degree of sterility may be regarded 
as an expression of its extent. 
The adoption and application of such a reaction system concep- 
tion to hereditary phenomena has far reaching consequences. When, 
for example, this conception is applied to the Oenothera phenomena, it 
at once follows that the widespread occurrence of partial sterility, the 
significance of which has never been definitely determined, must be of 
primary importance in the formulation of any consistent explanation 
of the hereditary phenomena displayed in Oenothera. Until it is possible 
to define clearly the exact significance of this partial sterility, it is 
obviously useless to attempt to apply any rigid Mendelian analysis. 
Moreover, the Oenothera phenomena belong to several different cate- 
gories, three of which, at least, may be clearly distinguished. In the 
first place, there are some strict factor mutations in Oenothera, such as 
ruhricalyx. These mutations depend upon a particular change in some 
locus in the hereditary mechanism, and they display normal Mendelian 
behavior when tested with the forms from which they were derived. 
The extensive observations of Morgan and his associates on mutation 
in Drosophila, to say nothing of other well authenticated cases in both 
plants and animals, seem to establish the vaHdity and nature of this 
