PANICLE CHARACTERS OF SORGO 25 
Color of the caryopsis. — In general, colors in the caryopsis are 
lighter than in the glumes. In one variety the caryopses are white, 
and the differences range through tints and shades of yellow, brown, 
and reddish brown to chocolate. In some varieties a red element is 
pronounced, though in none is the color truly red. Where exposed, 
the shade is usually somewhat darker than where protected from 
the light by the glumes, and generally the statements made con- 
cerning the glumes in regard to inconstancy of color likewise apply 
to the ca^opsis. 
RELATIVE LENGTH OF THE CARYOPSIS AND THE GLUMES 
A number of writers of ke}^s and descriptions have made use of 
the extent to which the glumes open and the caryopses are exserted. 
Among these are Arduino (7), Hackel (£, pp. 499-520), Stapf (61). 
Pech (12), Collier (5), and Ball (2). Both of these characters are 
dependent on variation in the comparative size of caryopses and 
glumes. The opening of the glumes and the protrusion of the 
caryopsis are both due to increase in the length of the latter beyond 
the length of the glumes. The car} T opsis is not, however, relatively 
longer in varieties in which it is exserted as compared with other 
kinds, but, generally speaking, the glumes are relatively shorter. 
Considered from a phylogenetic viewpoint, therefore, it may be 
assumed that varieties with exserted caryopses originated from a 
type with covered grains by the shortening of the glumes along 
with other parts of the panicles. 
In none of our varieties are cases to be found in which the pro- 
truding caryopsis is associated with the effuse type of panicle, as 
was indicated by Schumann (14-) regarding East African varieties. 
Since the caryopsis is inserted at the summit of a short axis, there 
are cases in which it is shorter than the glumes, yet protrudes some- 
what beyond them. Four conditions of caryopsis length, as com- 
pared with length of the glumes, therefore, really exist: (1) The 
caryopsis shorter than and not extending as far as the apices of the 
glumes; (2) the caryopsis shorter than the glumes but extending to 
their apices; (3) the caryopsis shorter than the glumes but extend- 
ing somewhat beyond their apices; and (4) the caryopsis actually 
longer than the glumes. The protruding caryopsis is well shown in 
the Sumac variety, which is represented in Plate II, C. The 
amount exserted differs somewhat with the variety. It is impos- 
sible, however ,to indicate the extent, and on\j three conditions are 
therefore of value for actual use. 
THE STERILE SPIKELET 
Although the sterile spikelets do not truly disarticulate from their 
pedicels or the pedicels at their bases from the joints of the rachises 
of the racemes, in many varieties the pedicels become fragile with the 
maturing of the panicle, so that they rupture easily, and the spikelets 
then fall off freely, usually enough remaining, however, to make it 
possible to determine accurately their form, nervation, pubescence, 
and other characters. In certain varieties, too, the pedicels remain 
tough and tenacious, so that in the main they continue attached to 
the rachis and retain the spikelets, this character being somewhat 
influenced by the degree of maturity of the panicle. 
