ME. A. G. MORE ON FESTUCA AMBIGUA. 191 



The three closely allied species may be thus contrasted : pre- 

 mising that the term Vulpia is here adopted for that section of 

 the Linnean Festuca which is distinguished by an annual root, 

 and by having awns longer than the pale. 



# Lower pale ciliated. 



1. Vulpia ciliata, Link. Panicle narrow, upright, its rachis downy. 

 Glumes widely membranous, pointed, very unequal, 3-5 to 1 ; the 

 larger glume reaching less than half-way up the included floret. 

 Awn of lower florets 1^ time as long as the pale. 



** Lower pale not ciliated, scabrous. 



2. Vulpia ambigua. Panicle narrow, close, upright, its lowest branch 

 reaching more than one-third up. Glumes widely membranous, sub- 

 acute or eroded, very unequal, 3-6 to 1 ; the larger glume reaching 

 one-third up the included floret. Awn twice as long as the pale. 



3. Vulpia pseudo-myurus, Soy. -Will. Panicle elongated, very 

 narrow, nodding, interrupted below, often partly sheathed ; its lowest 

 branch reaching one-quarter up the panicle. Glumes membranous- 

 edged, acuminate, unequal, 3 to 1 ; the larger glume reaching half- 

 way up the included floret. Awn nearly twice as long as the pale. 



In all these characters Vulpia ambigua is still more distinct 

 from V. Broteri, Boiss and Eeut, to which I was once inclined to 

 refer it. V. Broteri has a short, upright, oblong panicle, distant 

 from the sheath, and with its lowest branch reaching nearly half- 

 way up. Its glumes are far less unequal, and more pointed than 

 in V. ambigua : the larger glume reaches more than two-thirds up 

 the included floret, and the awns are more than twice as long as 

 the pale. Indeed, in the series, V. Broteri would take its place 

 between V. pseudo-myurus and V. sciuroides. 



There is still less fear of confounding V. sciuroides with any of 

 the above. V. sciuroides has much stouter spikelets, a short, up- 

 right, oblong panicle far removed from the sheath. Its glumes are 

 as 2 to 1, acuminate, with the dorsal nerve reaching to the top ; 

 larger glume nearly as long as the included floret, awn about 

 equalling the pale ; lower branches (when present) reaching nearly 

 half-way up the panicle. 



It is to be noticed that, in computing the proportions of the 

 two glumes to each other and to the included floret, those of the 

 terminal spikelet are not to be trusted, being always less unequal 

 than the others. In several instances I have found a well-developed 



p2 



