Petrie. — 0)1 Neiv Native Plants. 405 



of some of the geneva merged by Beiitham and Hooker in 

 Deschampsia clearly imply that a\Ynless species are included. 

 This is a point of great hnportauce, for in D. tenella, D. nova- 

 zclandice, and usually in D. imdlla, the absence of a distinct 

 dorsal awn is the only cardinal character in which these 

 species depart from the normal type of Dcscliampsia. In some, 

 it is true, the outer glumes are markedly unequal, but in 

 others they are not more unequal than is usual in the genus. 

 lu all other respects they are all true Descliampsias and have 

 little in common with Triodia ; and the Triodia (?) antarctica 

 of Bentham and Hooker agrees with the endemic New Zea- 

 land forms in every generically important feature. The chief 

 characters in wdiicli they differ from Triodia are the uniformly 

 membranous texture of the glumes and their few^ and faint 

 nerves, the small number of flowers in each spikelet (nearly 

 always 2) ; the uniformly stipitate upper flowers ; the more 

 numerous teeth or lobes of the flowering glume ; the sub- 

 median nerves of the hyaline palea ; the absence of any imper- 

 fect terminal flower ; the rounded-oblong (not plano-convex) 

 grain ; and the peculiar, broad, long, subulate, scarious ligule. 

 In addition to these points of diil'erence common to all the 

 si^ecies here described, D. chapmani has a distinct dorsal awn, 

 a character wholly foreign to Triodia. A similar dorsal awn 

 occurs frequently in the flowers of an allied form of which I 

 have specimens from Mr. T. Kirk, F.L.S., under the MS. name 

 of Triodia purpurea, and it occurs occasionally in the flowers 

 of D. pusilla, while in D. tenella it is hardly doubtful that the 

 nerve-like ridge on the back of the flowering glume is an awn- 

 like structure adnate to the glume. The remarkably uniform 

 character of the ligule is a point of some importance, the more 

 so as it is quite unlike that of Triodia, which is usually, if not 

 invariably, represented by a band of hairs. The New Zealand 

 species of Triodia, and also the British and Australian ones, 

 all agree in this. Indeed, the two native species of Triodia- 

 present few points of close alliance with the series of Dcscliamp- 

 sias noticed in this paper, while their differences are obvious 

 and striking. The result of this discussion seems to be that 

 the character of the genus Descliampsia, as given in the 

 "Genera Plantarum," needs to be amended, so as to include 

 the new species \\o\y brought to light from New Zealand and 

 its outlying islands. 



11. Lobelia linnceoides, sp. nov. 



This is Pratia (?) linnceoides, Hook, f., described on page 

 172 of Hooker's "Handbook of the New Zealand Flora." 

 When it was described the fruit was uuknow-n, and, as the 

 habit and foliage present a close resemblance to the indigenous 

 species of Pratia, it was provisionally ranked in that genus. 



