58 Transactions. 



This well-marked variety is distinguished at a glance by the broadly 

 ■oblong-spathulate leaves with wide petioles shorter than the lamina and. 

 rounded, never acute, apex. 



Whether this variety is confined to Mount Egmont and the Pouakai 

 Range I do not know, but it is apparently the only form of the species in 

 that area. 



(c.) C. glandulosa Hook. f. var. y longiscapa Cockayne var. nov. 



Folia oblongo-lanceolata vel obovato-lanceolata, cum petiolo + 5-5 cm. 

 longa, glabra, apice acuta ; petioli laminam longitudine aequantes vel 

 superantes. Scapus + 20 cm. longus, gracilis, strictus. 



South Island : Fiord Botanical District — Upper Clinton Valley, on old 

 moraine. L. C. 



At first sight this variety looks so distinct from the " type " (var. vera) 

 that one feels inclined to separate it as a distinct species. But, except for 

 its larger longer-petioled leaves and much longer scape, its characters are 

 virtually the same as the plant of the Volcanic Plateau. It bears a good 

 deal of resemblance to Celmisia glabrescens Petrie of Stewart Island, but 

 the latter is separated by its still longer leaves, which are thinly tomentose 

 beneath and which lack the strong network of veins. 



*» 



18. Celmisia longifolia Cass. var. gracilenta (Hook, f.) T. Kirk, form 



with branched scape. 



In the garden at Dunedin of the late Mr. H. J. Matthews there were 

 at least two forms of Celmisia coriacea Hook. f. with branched scapes. To 

 these names were given by T. Kirk {Students'' Flora, p. 288), although he 

 considered that the branching was due to cultivation. Cheeseman supports 

 Kirk in this opinion, stating {Manual of the New Zealand Flora, p. 311), 

 " In cultivation it [C. coriacea'] varies still more largely, and often produces 

 branched scapes, a peculiarity quite unknown in any Celmisia in the wild 

 state, so far as my observations go." 



Some years ago, however, before the above statement was published, 

 I collected a number of specimens on the Port Hills, Banks Peninsula, from 

 a wild plant of Celmisia longifolia var. gracilenta with branched scapes. 

 Possibly the occurrence cf such branching may be commoner than hitherto 

 suspected. Nor do I see any reason for assuming that the branching in 

 Matthews's plants was the result of cultivation, especially as they were 

 grown under the purely natural conditions his garden supplied. Further, 

 according to Cheeseman (I.e.), Matthews stated that var. ensata T. Kirk, 

 another branched form, was collected by himself from a wild branched 

 plant growing near Lake Harris, Cheeseman remarking, however, that 

 " I have only seen cultivated specimens." 



19. Leptospermum scoparium Forst. var. incanum Cockayne var. nov. 



Folia lanceolata vel lineari-lanceolata, circ. 8 mm. longa, subtus praecipue 

 juventute 4; pilis sericeis albidis obsita ; flores magni petalis roseis leve 

 tinctis. 



North Island : North Auckland Botanical District — Common, especially 

 in the northern part of the district, in many places forming thickets. L. C. 



This well-marked variety is distinguished at a glance from any other 

 forms of the species with which it may be associated by its young branchlets 

 .and leaves hoary with silky whitish hairs which persist for a considerable 

 .time on older leaves especially near the base, its rather large flowers with 



