231 



scJdstis, ■' gluma infcriora quadripartita, flora brevior, laciniis 

 lineari-subulatis, pilosis.'* Supposing one of these bristles to have 

 been bifid, as at c, plate CVIT, only more deeply so, there would 

 then exist a striking resemblance to the lower glume of Pcntar- 

 rhapJiis, as described by Kunth : ''gluma inferior ex aristis 

 quinque basi subconnatis composita." 



In the clusters or fascicles with two well developed spikelets 

 these spring from a very much shortened, hirsute axis so that 

 they appear to stand opposite each other, the backs of their lowest 

 glumes coming together, see figure A, plate CVIII. Between 

 the spikelets are three setae or awn-like bristles, one of which is 

 often more or less deeply two-parted. These bristles, shown 

 more clearly in figure C, are sub-equal in length, similarly pilose 

 and projecting in the same line of growth with the spikelets, and 

 appear to be, at least when examined dry, all of the same nature. 

 If moistened, however, and then carefully examined under a lense, 

 two of them will be found to have upon each side below a narrow 

 membranous margin (see figs. b\ b\ plate CVII); this character, 

 together with that of their position, determines them to be the 

 lower empty glumes of the two spikelets. The third seta is either 

 simple or bifid and originates at a point a little to one side or in 

 advance of the lower empty glumes. It may be said to occupy 

 the centre of the fascicle of which it forms a part, representing 

 the rudiment of a third spikelet. That it is not a continuation of 

 the partial rhachis, but really the representative of a third spike- 

 let, is determined, I think, by the following characters; (i) its 

 direction of growth in the same line with the fully developed 

 spikelets; (2) its springing from the same very hirsute and 

 condensed axis of these spikelets, which is (to me an important 

 point in this connection) articulated with the short pedicel from 

 the main rhachis; (3) its being pilose like the empty glumes, the 

 extended midribs or awns of which it most closely resembles. In 

 those Bouteloua^ with which lam familiar, wherein the rhachis of 

 the spikes is produced beyond the spikes, the primary direction 

 of this rhachis is continued, and although it may be reduced to a 

 sharp pohit or even become notched or bifid as in Bouteloiia di- 

 sticha, Bcnth. (^Polyodon distichum, HBK.) there is no question 

 as to its true character. 



