been recognized previously. Bews (1929) has pointed out 
that, in general, the spreading type of panicle is the most 
primitive and ‘‘Starting from this, the main evolutionary 
trend has been towards contraction, condensation [or 
compression], reduction and, as a result, increased pro- 
tection.’’ Arber (1984) has added that ‘‘With the con- 
sequent compression [from confinement during youth], 
we may, in some degree, associate the reduction within 
the inflorercence [or within the floral phytomer], which 
is SO conspicuous in the grasses. ”’ 
These evolutionary trends are apparent in a graded 
series of species starting with the Andropogoneae and 
terminating with the American Maydeae as follows: 
The spreading panicle of grasses such as Mrianthus have 
contracted to compact spikes as in 7'r7psacum along with 
various reductions: in sex development (perfect to uni- 
sexual); in numbers of spikelets (paired to single); in 
glume length (long to short); and in degree of pubes- 
cence (hirsute to glabrous). The last named feature may 
be, in part, related to an increase in lignification and a 
decrease in cell size, as suggested by preliminary studies 
in maize. Coincident with these reductions, the rachis- 
segments enlarge in diameter and an adnate prophyll 
develops on the inner surface of the concave segment. 
The wings of the prophyl! protrude laterally so that they 
clasp the outer glume of the enclosed spikelet and, there- 
by, complete the structure of this protective device. This 
adnate prophyll constitutes a distinguishing feature be- 
tween the caryopsis-bearing inflorescences of the Andro- 
pogoneae and the American Maydeae. In the staminate 
inflorescence of maize and teosinte, and in other non- 
cupulate panicles, the prophyll appears to be modified 
as a small axillary pulvinus. Cupules are weakly devel- 
oped in the staminate rachis of T’ripsacum. 
The Hordeae contain another independent series lead- 
[ 236 ] 
