353 



©vtgmal Mxtxtlt^. 



PHYLLOEACIIIS, A NEW GENUS OF GRAMIXE.J^ 

 FEOM WESTERN TROPICAL AFRICA. 



By Henry Trimen, I\I.B., F.L.S. 



(Tab. 205). 



The grass here described is a Dative of the district of Pimgo 

 Andongo, in Angola, where the only specimens I have seen were 

 collected by Dr. Welwitsch, in March, 1857. Among the numerous 

 interesting S23ecies, many undescribed, of this family contained in 

 his herbarium, no one is more remarkable than the subject of this 

 notice, which, both in habit and in structure of the inflorescence, 

 appears to differ from all known genera to a very unusual extent. 

 I propose to name it PJu/UoracJiis, a descriptive title suf&ciently 

 self-exx3lanatory. 



Before giving a formal definition of this genus I may caU 

 attention to the peculiarities of the floral arrangement upon which 

 that definition is founded, and to some other particulars ; and I 

 cannot but regret that the specimens before me are not sufficient to 

 do this with completeness, since their flowering season being over 

 when gathered a large number of the spikelets have fallen, and 

 stamens are wanting in those which remain, which, however, 

 present us with the nearly-ripe grain. 



At first sight the long inflorescence might be supposed to be 

 partially enclosed in a leaf-sheath, and it is not until the edges of 

 the veined leafy organ are raised that this latter is seen to be the 

 rachis itself, greatly dilated, somewhat after the manner of the 

 Ceresia section of Paspahim, and bearing on its thickened central 

 rib the branches of the inflorescence. This consists of a number 

 (12-20) of contracted brandies or compound spikelets, sessile and 

 ai-ticulated on either side of the inner surface of the central rib of 

 the rachis ; they are not closely placed, but overlap one another, and 

 when detached, which readily happens, they leave a small oval scar. 

 The parts present in each are rather puzzling, and their i3recise 

 relationships not very clear, but it appears that each compound 

 spikelet must be regarded as consisting of an abbreviated secondary 

 rachis and three spikelets, of which the lowest one is alone fertile 

 and 1 -flowered, the upper two (borne upon the very broad, flattened 

 truncate secondary rachis, and pressed closely against the side of 

 the lowest spikelet) being much smaller, and consisting merely of 

 empty glumes. The lowest (fertile) spikelet has four glumes, if we 

 are to give the lowest minute and subulate bract that name ; '" the 



♦ General Munro, who has with his usual kimluess examined, at my request, 

 the structure of a compound spikelet of Phyllorachis, and agrees generally with 



N. s. VOL. 8. [December, 1879.] 2 z 



