466 
AMERICAN JOURNAL OF BOTANY 
[Vol. io, 
The impure species, although hybrid in its germinal constitution, breeds 
true because only such gametes unite and give progeny as will reproduce 
the heterozygous constitution of the parent plant. Other types of gametes 
to be expected from the segregation divisions of meiosis either do not 
develop or fail to function for one of various possible reasons, or, if they do 
unite, the zygote either can not develop at all or it produces an embryo or 
seedling which can not mature. In short, there are breakdowns at one or 
more of various critical points in the life history, and thus some form of 
sterility eliminates the development of all or nearly all groups of segregates 
possible to the hybrid in question, and only such combinations of gametes 
are effective as will give the genotype of the parent. 
This concept of the pure-breeding hybrid is not a fancy. We have 
excellent evidence that impure species are common in the genus Oenothera 
and that Oenothera Lamarckiana is one of them. Certain lines of Drosophila 
are known to be impure, and we owe to studies of Muller on such material 
the theory of balanced lethals which offers the best notion of a mechanism 
in heredity responsible for the generally true breeding of an impure species 
and for the appearance of occasional variants which some geneticists call 
mutants but which are really segregates from the heterozygous stock. 
The theory of balanced lethals postulates the presence of two different 
lethals, for example x and y, the first in one chromosome and the second in 
the other chromosome of a synaptic pair. The organism is therefore 
heterozygous for each lethal. The theory also assumes that each lethal is 
effective only in double dose. The reduction divisions in such material 
will give two classes of sperms and two classes of eggs, each class distin¬ 
guished by the presence of one of the two lethals. Thus there will be 
sperms x and y and eggs x and y and the chance mating of these will give 
zygotes in the following proportions ixx : 2 xy : lyy. Zygotes xx and yy, 
because they have lethals in double dose cannot develop progeny, but the 
zygotes with the heterozygous combination xy will live and reproduce the 
impure or hybrid parent type. Thus an impure species or race will breed 
true and maintain a constant state of hybridism unless the relative position 
of the lethals is changed by a crossover or unless a lethal becomes ineffective 
through a mutation. A crossover makes possible a class of zygotes free 
from both lethals, because sperms and eggs would be of the two classes 
xy and oo and the zygotes would be in the proportions ixxyy : 2xy : ioo. 
Through the class of segregates free from both lethals recessive characters 
would appear if genes responsible for their suppression were removed by 
way of the class homozygous for both lethals. The appearance of such 
recessives will simulate mutations although in reality they are manifesta¬ 
tions of a process of segregation. 
There is not time to carry farther a discussion of the bearing of the 
theory of lethals on the facts of hybrid sterility. It is not probable that 
conditions in Oenothera and Drosophila, two groups which have received 
