194 Transactions. 



This variety is easily distinguished from the extremely common var. 

 pusilla Bitter by its larger bright-green leaves, which have never brown 

 basal leaflets, and the pale but clear green calyx-segments and spines, 

 whereas the small basal leaflets of the var. pusilla are more or less deeply 

 stained with brown, and the spines are pale and stained here and there 

 with light-red. It might be thought that the differences in the colour of 

 the leaves was due to the effect of light, but the two varieties grow side 

 by side with their characters unchanged. 



Bitter makes his var. pusilla a subspecies of A. Sanguisorbae equivalent 

 in rank to his subspecies oiovae-zelandiae (= A. novae-zelandiae T. Kirk), 

 but it seems to me simpler to treat the last as an aggregate species and 

 to reduce all Bitter's other New Zealand endemic subspecies of A. Sanguis- 

 orhae to the rank of varieties, though this step must eventually lead to the 

 establishment of sub varieties. 



3. Celmisia Monroi Hook. f. 



Celmisia Monroi was first described by Sir Joseph Hooker in the Hand- 

 book, p. 133, from specimens collected by Monro on the Upton Downs, 

 Awatere ; by Haast near Mount Cook (Hopkins River) and elsewhere 

 in the Alps of Canterbury ; and by W. T. L. Travers in some part of the 

 latter. There was no " type " for the species, since it was not only de- 

 scribed from a number of individuals, but also the Awatere specimens 

 possessed glabrous achenes, while in those collected by Haast they were 

 hispidulous. The species under consideration has been imperfectly under- 

 stood by most authors, &c., and, according to Cheeseman, confounded with 

 small forms of C. coriacea Hook, f., a not surprising mistake. T. Kirk, 

 in the " Students' Flora," p. 288, enlarged Hooker's conception of the 

 species by including plants from the neighbourhood of Whangarei and the 

 Bay of Islands (Auckland). Cheeseman, in the Manual, p. 313, reduced 

 the compass of the species to that proposed by Hooker, and placed the 

 Whangarei plant under C. Adamsii T. Kirk as var. rugulosa Cheesem. 

 Recently (Trans. N.Z. Inst., vol. 44 (1912), p. 182) Petrie has removed 

 the Mount Cook plant from the species, describing it as C. Boweana, 

 thus limiting G. Monroi to the plant of the Awatere. 



C Monroi in its restricted significance, and as examined in the light 

 of fairly abundant material from different stations, does not altogether 

 match the description in the Manual. The leaves are not infrequently 

 much longer (21 in.) than therein described, and broader (1| in.), though 

 about f in. is perhaps the average breadth. The leaf-sheaths are not 

 always short, but may be of great length — e.g., a lamina 10|- in. long may 

 possess a sheath 9 in. long. The texture of the leaf is frequently thinner 

 than has hitherto been described, and more or less flaccid leaves are not 

 unknown. 



Possibly there are two distinct races represented amongst my material. 

 The first possesses leaves varying from linear-lanceolate to broadly lanceo- 

 late, and the second, of which I have only one specimen, is a quite different- 

 looking plant with rather short (9 in. long) linear or almost linear rigid 

 leaves, ^| in. broad, furnished with a short not densely woolly sheath. 



The " race " of C. Monroi with the more or less lanceolate leaves is 

 evidently closely related to C. coriacea, and some taxonomists may prefer 

 to reduce it to a variety of that aggregate species : but it can be distin- 

 guished by its narrower, thinner, and sometimes far shorter leaves, with 

 the lamina never rounded at the base but tapering gradually into the 



