2 BULLETIN 869, U. S. DEPARTMENT OE AGRICULTURE. 
of most of the modern taxonomists as well. The groups of Schuebler 
(22) 1 , Serfage (23), Heuze (11, 12), Voss (25), Koernicke (13, 14, 15, 
16, 17), Atterberg (2, 3, 4), and Beaven (5) involved variations in 
density. In 1918 Harlan (10) offered an arrangement which elim- 
inated the question of density from the major groups. It was re- 
tained as a minor distinction only, because of the volume of the liter- 
ature in which it had been used. Its complete elimination would 
have left too little connection between the author's scheme and the 
previous usage. ] 
In classifying barleys, density is an obvious and attractive char- 
acter. When confined to type forms the separations are ideal, but, 
as with many things in taxonomy, its perfection depends on limited 
material. The more material that is assembled the more the sub- 
divisions of density have to be increased. Linnaeus (18) used the 
name Eordeum disticlion to designate the lax 2-rowed and E. zeo- 
criton to designate the very dense 2-rowed forms. Schuebler divided 
E. disticlion into erectum and nutans. Eriksson (S) used genuinum 
and patens to designate lax and dense subdivisions of erectum. 
Linnseus recognized liexastichum and vulgar e as the dense and lax 
groups of 6-rowed barleys. Koernicke divided liexastichum into 
pyramidatum and parallelum and recognized oraciiyurum and macro- 
terium of Alefeld (1) as dense and lax subdivisions of pyramidatum. 
The finer the groups were made, the more confusing became the dis- 
tinctions. The confusion indicated that, while there might be some 
genetic distinctions, from a taxonomic standpoint there was no clear 
separation. i Q 
In the mode of inheritance the situation is also complicated. As 
a size character, the accounts are quite favorable as to its constancy, 
and some varieties are traceable for centuries by this character alone. 
In recent times Blaringhem (7), possibly following the lead of the 
Svalof station, made quite elaborate studies of barley density in 
France. Harlan (9) found density to be quite a stable character. 
Regarding the mode of inheritance, the studies, however, are largely 
unsatisfactory. The taxonomic papers contain no comprehensive 
measurement of density. Many of the inheritance papers are equally 
inadequate. In many instances fertility and density are treated 
together, as by Von Tschermak (24). Density has been regarded 
as recessive by Blaringhem (7) and as dominant by Von Tschermak. 
The only paper which is directly concerned with the method of 
study used in this article is that of Biffen (6), who obtained results 
closely parallel to those presented herein. In three crosses to which he 
paid particular attention, Biffen found the F x generation to be slightly 
more dense than the lax parent, although the numbers of individuals 
in F x were small. The F 2 generation consisted in each case of plants 
1 The serial numbers in parentheses refer to "Literature cited," at the end of this bulletin. 
