Systematic Botamj.] SUBANTARCTIC ISLANDS OF NEW ZEALAND. 465 



41. Deyeuxia Forstcri, Kunth, var. mi- No doubt a recent variation from D. Forsteri, var. Lyallii. 



cratkera, Hack. 



42. Deschampsia (jracillima, T. Kirk . . This seems to be nearest to the New Zealand species of the 



genus. 



43. ,, fcnicillata, T. Kirk . . I have seen only two very imperfect specimens of this, and 



cannot express any definite opinion as to its relationships. 



44. Poa Tennantiana, Petrie . . Very close to Poa joliosa, Hook, f., of which it is probably a 



recent derivative form, and belonging to the same group as 

 the Kerguelen P. Cookii, Hook, f., and the Fuegian P. flabel- 

 lata, Hook. f. 



45. „ antipoda, Petrie . . . . The nearest ally of this is given by Mr. Petrie as the northern 



P. trivialis. 



46. ,, litorosa, Cheesem. . . . . A well-marked species, not closely resembling any other. 



47. ,, ramosissima, Hook. f. . . The nearest relative of this appears to be the Kermadec 



Island P. polyphylla, Hack. 



48. ,, Hamihoni, T. Kirk. . . . I have seen no specimens of this, and Mr. Kirk's description is 



not sufficiently precise to allow its relationships to be made 

 out. 



49. ,, aucklandica, Petrie. . . . Appears to be a very isolated form, of doubtful affinity. 



50. ,, incrassata, Petrie . . . . The New Zealand P. exigua is probably the nearest ally to this. 



51. ,, hreviglumis, Hook. f. . . Most probably a mere variety of the New Zealand P. imbecilla, 



Forst., from which it differs only in the more minute outer 

 glumes. 



52. Atropis antipoda, Petrie. . . . A near ally of the New Zealand A. stricta, Hack. 



53. Festuca contracta, T. Kirk . . Very imperfectly known, but clearly distinct from any New 



Zealand plant. 



A study of the " tabular view " will show that only a small proportion of the 

 endemic species occurs in all the islands. As a rule, they have a more limited dis- 

 tribution, and there are several which do not extend beyond the confines of a single 

 island. For instance, two species are confined to the Snares, and are found in no 

 other part of the group ; six to the Auckland Islands ; four to Campbell Island ; 

 four to Antipodes Island ; and three to Macquarie Island. This must be regarded 

 as an important fact, proving a considerable amount of independent development in 

 the vegetation of the separate islands, and pointing strongly towards their probable 

 lengthened isolation. 



A careful scrutiny of the endemic species proves that they can be readily divided 

 into two distinct groups. First, a group consisting of species not closely related to 

 any other — some of them, in fact, so distinct as to make it difficult to assign their 

 true relationships. These must be looked upon as ancient and long-isolated types, 

 relics of what was once a numerous and widely spread flora, regarding the origin 

 and extent of which we have no precise knowledge, and the history of which is alto- 

 gether a matter of conjecture. In this category I should include Stilbocarpa (two 

 species), Pleurophyllum (three species), two species of Celmisia, Senecio antipodus, 

 Ligusticum latifolium and L. antipodum, Ahrotanella rosulata and A. spathulata, 

 Cotula lanata, Gentiana (four species), Veronica Benthami. Colobanthus muscoides, 

 Ranunculus subscaposus and R. aucMandicus, Plantago aucklandica, and a few others, 

 including six or eight species of grasses. Speaking approximately, I should place 

 about thirty-two species in this group. The second group, containing the remaining 

 twenty-one species, would include Ranunculus pinguis, Azorella reniformis, Veronica 

 odora, Coprosma ciliata, Ligusticum acutifolium, Olearia Lyallii, Myosotis capitata, &c., 

 all very closely allied to species at present living in New Zealand, and which conse- 

 31— S. 



