Plate 218.— IMPERATA CHEESEMANIl. 



Family GRAMINBiE.] [Genus IMPERATA, Cyr. 



Imperata Cheesemanii, Hack, in Trans. N.Z. Inst, xxxv (1903), 378 ; Cheesein. Man. N.Z. Ft. 843. 



This liandsome grass was first discovered by myself in August, 1887, on Raoul 

 or Sunday Island, the chief island of the Kermadec Group. At the time of my 

 visit it was abundant, especially on clif?s and steep slopes on the northern side of 

 the island. Mr. R. B. Oliver, who spent the greater portion of 1908 on Sunday 

 Island for the purpose of examining its vegetation, also found it not uncommon. 

 He records it from talus slopes on the cliffs and on the slopes leading down to the 

 central crater, from sand-dunes (where it grows intermixed with Ifomcea hiloba), 

 and from the gravel-flat in Denham Bay. I suspect, however, that with Imperata, 

 as with several other Kermadec Islands plants, the great increase in the number 

 of goats has reduced the area of distribution of the plant since the time of my visit. 



In my account of the " Flora of the Kermadec Islands " (Trans. N.Z. Inst. 

 XX (1888), 175) I referred the species to the widely distributed /. arundinacea, Cvr. 

 This, however, was a mistake, as was first pointed out by Professor Hackel, who 

 described it under its present name. He remarks that it is much nearer to 

 Z. exaltata, Brong., the typical form of which ranges from the Philippine Islands and 

 the Malay Archipelago to the New Hebrides, and which in a slightly different form 

 (var. caudata) stretches from Mexico and the West Indies to Argentina. Our plant 

 is much smaller in all its parts. 



The genus Imperata is widely distributed in the tropical regions of both 

 hemispheres, but the number of species probably does not exceed six or seven. 

 /. arundinacea, which is the best-known member of the genus, has been collected 

 near Kaitaia by the late Mr. R. H. Matthews, but there is a suspicion that 

 it is naturalized only. 



Plate 218. Imperata Cheesemanii, drawn from specimens collected on seaclifis on Sunday, or 

 Raoul, Island. Fig. 1, tip of leaf (enlarged) ; 2, ligule (enlarged) ; 3, two spikelets (x 8) ; 4, outer 

 glume ; 5, same seen from the front ; 6, second glume ; 7, the same seen from the front ; 8, third glume ; 

 9, palea ; 10, flovsering-glume (all enlarged) ; 11, ovary, style, and stamen removed from the spikelet 

 (xlO). 



