3 
precise determination. Two other sheets, however, in Vanderyst’s 
collections (nos. 2661 and 3832) from the Kwango and Stanley Pool Dis- 
tricts of the Belgian Congo belong clearly to a distinct species of the 
genus, so that Diheteropogon is so far known from at least three species. 
The nearest approach to Diheteropogon may be found in Andropogon 
filifolius, and it is, in fact, a moot question whether it should not be 
transferred to Andropogon. Nees, led no doubt by a superficial 
resemblance with Heteropogon, actually described it as Heteropogon 
filifolius. It lacks, however, the extreme differentiation into homo- 
gamous and heterogamous pairs of spikelets which is so conspicuous 
a feature in Diheteropogon, although it possesses generally one or 
two homogamous pairs at the base of the peduncled or of both 
racemes, a character present in a less pronounced degree in most 
species of the section Piestiwm.—O. Svar. 
Fic. 1, part of fertile portion of raceme with the pedicelled spikelets removed, 
rtile spikelet with adjacent joint and pedicel; 3, dia- 
0 ; 7, valvul 
spikelet) ; 9, 1 ; 1 flattened out ; 11, same as in natura 
pikelet) ; 9, lower glume ; 10, upper glume Ha poet been 1A oT at ds bok 
tate ; e of u 
by error), valvule of same (9-14 of pedicelled spikelet). All enlarged, except 
Jig. 1, which is of natural size. 
