280 



lis inferior/' are in my view, the relics of a whole rudimentary 

 spikelet. You will often have observed in several Routeloua:^ (the 

 genus to wliich Pentarrhaphis is very nearly allied) that the rha- 

 chilla of the spikelet is headed by a group of bristles (in B. imtlti- 

 seta, for example, by many) which are reduced glumes. Now in 

 Pentarrhaphis I believe the group of bristles at the base represents 

 a whole rudimentary spikelet. The number of these setae seems 

 to be somewhat variable, as is often the case with rudimentary 

 organs. As you have fuller materials for investigation I invite 

 you to make a closer study of it, and should you gain the same 

 opinion as myself regarding it, I propose for the species the name 

 Poitarrhaphis gcmiuata. Hack. & Scribn. 



" I believe that Polyscliistis, Prcsl. (in Rcl. Ha^nk. i. p. 294) is 

 quite the same genus, and his P, pauperciila seems even to be 

 very nearly allied to Pentarrhaphis gcniiiiata. But I can only 

 judge of this from Presl's figure, not from authentic specimens, 

 which I sought for in vain in the Vienna Herbarium. Presl says 

 that Haenke collected it in Luzone, but he very often confounded 

 plants from Mexico with those from the Philippines, and vice versa!' 



After making a careful investigation of the spikclets and the 

 bristles or seta: accompanying them, I communicated my conclu- 

 sions, detailed below, to Prof Hackel, and in reply he says: '* I 

 fully agree with you in the signification that you give to the 

 various bristles that accompany the base of the twin spikclets. 

 As to the number of five bristles, described by Kunth, I venture 

 the view that one of them was the outer glume, the rest forming 

 two pairs, each representing a sterile spikelet. I found also in 

 your specimen one example of a single spikelet [in the cluster] 

 and I think it was accompanied by five or six bristles." 



I found several such examples as here referred to, in my 

 specimens, always the lowermost clusters in the racemes. Fig. B, 

 Plate CVIII illustrates one of these. At the base of this spikelet on 

 the upper or outer side, is the second empty glume, a, while on 

 the lower or inner side are several strongly pilose seta^ (three of 

 which are undistingulshable from one another when dry, the fourth 

 being evidently the second spikelet very imperfectly developed. 

 From the position of these set.ne they might easily be mistaken 

 for a quadripartite lower glume, such as Presl describes for Poly- 





