6 BULLETIN 1386, U. S. DEPARTMENT OF AGRICULTURE 
The descriptions which follow this key include, in addition to the 
points specified, characterizations of the axis and branches regarding 
form, surface configuration, and pubescence; of the glumes, lemma, 
and palea in some detail regarding form, size, and pubescence; and 
of the caryopsis in regard to form, size, color, and markings. It is 
evident from Schumann's statements in reference to source of mate- 
rial that he worked mainly with dried specimens. 
According to Stapf's classification (16, pp. 104—154), all his cul- 
tivated types, a subgroup in which the mature spikelets persist, belong 
to a group in which the branches of the first order are divided and 
the racemes are -both terminal and lateral, this being opposed to a 
second group in which the primary branches are simple and the 
racemes terminal, and to a third in which these branches are usually 
solitary, but divided at the base. His main groups of the cultivated 
forms are established on the texture of the glumes, and the subdivi- 
sions of these groups are based on characters of both the panicle 
itself and the individual spikelets. His key and descriptions include 
characterizations regarding form and density of the panicle; ar- 
rangement, direction, size, and pubescence of the branches; number 
of joints in the racemes and form and pubescence of the rachises; the 
form, size, color, nervation, and pubescence of the glumes of the 
fertile spikelet; and some points in regard to the lemmas, palea?, 
caryopses, and the sterile spikelets. 
Sorgo varieties in the United States have been described by Wray 
(18), Pech (1.2). Collier (£), and Ball (0), but Wray's descriptions 
of types introduced by him from Africa under the supposed African 
names were not made from the botanical viewpoint, and they give 
little clue as to which, if any, of the present-day varieties correspond 
to them. 
Pech (12) presented descriptions of nine varieties and a key for 
their identification, employing such characters of the panicle and the 
seed as the form, size, and habit of the branches of the panicle, the 
relative length, color, and form of the glumes and grain, and the 
vestiture of the glumes. But his descriptions were meager, and 
none of the varieties were 4 known by the names now used; hence it 
is impossible to identify any of them with certainty as varieties now 
in cultivation, although each type is represented by a small sketch 
of both panicle and seed. A reproduction of his key follows. 
Artificial key to the varieties of sorgo, oy Pech 
Ripe seed longer than the glumes Liberian. 
Ripe seed equaling or snorter than the flumes. 
Glumes equaling the length of the seed. 
Olumes closed, hiding the seed Red Imphee. 
Glumes open, showing the seed. 
Glumes greenish white or ash color White Imphee. 
Glumes hlack or purplish black. 
Branches of the panicle thin ; panicle long, 
widely spreading True Chinese sorgo. 
Branches of the panicle compact. Panicle short, 
erect, more or less appressed to the axis. 
Glumes mostly downy Early Sorgo. 
Glumes mostly smooth Oomseana. 
Glumes longer than the seed Black Imphee. 
Ball (3) points out that one of these varieties, the Chinese, appears 
to be similar to some of our strains of Amber. Pech's Red Imphee 
may have been the Honey or the one known in some localities as 
