Nov., 1923] 
DAVIS — POLLEN- AND SEED-STERILITY 
465 
somes, and it must be confessed that we cannot show any reasons why they 
may not be there even though their presence adds mightily to the responsi¬ 
bilities of these heavily worked cell structures. We cannot say that slow¬ 
ness of germination on the part of a pollen grain may not be due to the 
character of substance or lack of substance in some region of a chromosome 
even though the pollen tube grows through activity of the ectoplast. On the 
other hand it may be that some lethals are present outside of nuclear 
structures. Breeding studies should be able to separate cytoplasmic 
lethals, if such exist, from those that are associated with chromosomes, 
since the two would be expected to follow different systems in inheritance. 
We have treated sterility as though it depended in the last analysis on 
factors internal to the organism which may be postulated as genes and which 
express themselves through inheritance. With respect to sterility in 
hybrids, its association with germ-plasm organization is generally evident. 
It is a fair question to ask whether sterility ever has a basis other than that of 
germ-plasm organization. The temptation is sometimes great to lay the 
responsibilities of sterility to causes outside the organism and thus to 
separate a sort of physiological sterility from that which has a genetical 
relation. There are, for example, forms of sterility brought out by disease 
or by physiological conditions harmful to the plant. Supposing such 
sterility to be only partial, as is frequently the case, we should not of course 
expect the inheritance of this acquired sterility for the same reasons that 
we do not expect the inheritance of an acquired morphological character. 
So far as I am aware there are no reasons to regard sterility as other than a 
characteristic dependent upon germ-plasm organization and dependent in 
the same sense as are morphological characters. 
As stated at the beginning of this paper, hybrid material is generally 
expected to exhibit in some degree sterility of various sorts. Hybrid 
material is also generally expected to show its heterozygous nature in 
breeding by throwing a varied assortment of forms in its progeny. There 
is coming to be recognized, however, a type of hybrid that reproduces itself 
perfectly, throwing at most only occasional variants. Such hybrids satisfy 
fully our concept of a species as a kind of animal or plant which breeds 
true. They are impure species because their germ plasm in the diploid 
condition carries different sets of genes affecting characters other than 
those of sex. The pure species in contrast has a germ plasm carrying two 
similar sets of genes each contributed by one of the parents and each with 
the same genetical constitution except when genes responsible for sex and 
sex-linked characters are concerned. 
Because the impure or hybrid species is of particular interest in relation 
to the problems of hybrid sterility I must discuss certain phases of this 
subject although they have been treated with some fullness in my earlier 
paper “Species, pure and impure.” 2 
2 Davis, B. M. Species, pure and impure. Sci., n. ser. 55: 108-114. 1922. 
