368 Transactions. / 



(forma) , Drapetes Dieffenbachii, and Anisotome aromatica. Euphratiia zelandica, 

 Forsfera tenella, and Oreomyrrhis andicola also occur, but not so abundantly. 

 This formation is found on all the tops from Mount Sinclair to Castle 

 Rock, but for some reason does not occur on French Peak (2,600 ft.) to 

 Saddle Mount (2,700 ft.). It reappears on Mount Bossu (2,500 ft.) and 

 the neighbouring Carew's Peak (2,600 ft.), and traces of it reappear on 

 Brasenose (2,500 ft.) behind Akaroa Township. Other species of the 

 subalpine pasture are Danthonia Raoidii var. rigida, Aciphijlla Colensoi, 

 Raoulia glabra, and R. subsericea. Danthonia Cunninghamii and Gunnera 

 monoica are also sometimes to be found. This little subalpine fiorula 

 is, of course, as much isolated as if it stood on an island oif the coast, 

 and it is somewhat difficult to account for its presence here. Either 

 it has been brought by birds or winds, or else it is a remnant of a 

 vegetation that in glacial times reached to the sea-coast and extended 

 widely over the country. We have not the data at present to solve the 

 problem, and it can only be considered in connection with the general 

 geological evidence and with the distribution of alpine species throughout 

 New Zealand. It is, of course, to be expected that in this latitude wind- 

 swept mountain-tops should carry some subalpine species above 2,500 ft. 

 Such species are found near Dunedin as low as 1,500 ft., and in Stewart 

 Island at sea-level. In some cases, however {e.g., View Hill, French 

 Peak), the forests were able to maintain themselves on hilltops up to 

 2,500 ft. This may be due to edaphic conditions, for these peaks are 

 perhaps less rocky than those which are bare. 



Vertical Distribution. 



We have now considered briefly the general distribution of the indi- 

 genous vegetation on the peninsula. By way of brief recapitulation, it may 

 be pointed out that in walking from the edge of Lake Ellesmere to the 

 summit of Mount Herbert one would pass through practically all the plant 

 associations described. Starting with the salt marsh and salt meadow, 

 these follow in succession : the coastal scrub, a heath, a totara - black-pine 

 forest, a Podocarpus Hallii - Libocedrus forest, and a narrow Ijelt of sub- 

 alpine scrub, and finally the subalpine pasture. The same series would 

 ])e passed through in ascending from the beach at Pigeon Bay to the 

 summit of Mount Sinclair, or from the shore of Lake Forsyth to the top 

 of Mount Fitzgerald. 



Endemism. 



The area is so isolated that it might be expected to show a certain 

 amount of endemism. This is most readily illustrated by the three 

 species Celmisia Mackaid, Veronica Lavaudiana, and Senecio saxifragoides, 

 Celmisia Machaui has been rejjorted from Mount Fyffe by H. B. Kirk, 

 but the report has not been confirmed ; and Veronica Lavaudiana has been 

 recorded by Lyall, W. T. \j. Travers, and J. B. Armstrong from the river-beds 

 of the Canterbury Plains. Now, F. Lavaudiana is a true cliasmophyte, and 

 [ should be very much surprised to see it growing in the shingle-bed of 

 one of our Canterbury rivers. It might perhaps be expected in the moun- 

 tain gorges of these rivers, but I have not seen it there, and much doubt 

 its occurrence elsewhere than on Banks Peninsula. Travers records it 

 from the Ashley, but though I have been at many points of the Ashley 

 river-system I have seen nothing resembling V . Lavaudiana there. A form 

 of V. Raoidii certainly occurs at White Rock in small quantity ; and 



