Ecological Botani/.] SUBANTARCTIC ISLANDS OP NEW ZEALAND. 211 



near camp, but the only plant collected was unfortunately lost. Probably these rare 

 plants, and others not specified, were once much more common, and denote a more 

 extensive flora in tJie past, decrease in land-surface having caused extinction of 

 species.* 



(/3.) TiiK Plant Formations. 



(i.) Dunes. 



Dunes occur only on the north coast of Enderby Island. These I had not time 

 to visit, and can only reproduce here briefly the main points regarding their vegeta- 

 tion from my former paper (-'Botanical Excursion," p. 236). 



The most iiiifortant ecological feature is a negative one — the absence of s fecial 

 sand -hi 11(1 i III/ plants, the common sand-sedge of New Zealand, Scirpus frondosus, 

 being wanting, although common in Stewart Island, and even in the Chatliams 

 (Cockayne, " Plant-covering of Chatham Island," p. 261). The fact is — though, of 

 course, this does not explain the absence of the above plant — that the wet climate 

 is quite suflacient to keep the dunes stable, and plants unadajjted to a loose and 

 easily moved substratum can exist quite well. Thus there is in some places a close 

 low growth of Pratia arenaria,'f its green creeping stems rooting at the nodes. Of 

 a similar habit are Lagenophora pumila, but its stems are beneath the sand, and the 

 less-spreading and closer-growing Epilohium confertifoUum. With these is a moss 

 of dense habit. The close growth is chiefly on the sheltered side of the dune gullies, 

 the patches of vegetation in exposed places being far apart. Ranunculus acaulis and 

 Crassula rnoschata are also common plants. 



On some parts of the dunes, particularly wliere they are drifting inland, Riimex 

 neglectus is the dominant jjlant. Elsewhere in the New Zealand region it is a plant 

 of gravelly beaches or coastal moors. Its stout far-creeping stems, however, fit it 

 admirably for spreading in sand, and where, through what is most probably the 

 action of cattle, the dunes have been disturbed, and are spreading inland, a pure 

 formation of the Ruine.r. absent in the jnimitive vegetation, has been produced. J 



(ii.) Voastal-rock Formations. 



The rocks may be more or less flat and raised but little above the water-surface, 

 or there may be clii?s varying much in height. 



On flat rocks green cushions of Colohanthus muscoides dominate, and in crevices 

 are lines of Crassula rnoschata and small tufts of Scirpus aucklandictis, both of which 

 also frequently grow upon the Colohanthus cushions. Generally there are a few 

 stunted plants of Blechnum durum and Asplenium obtusatum, depauperated Cotula 

 plumosa and C. lanata, and frequently abundant, though not everywhere, are the 

 dark-green glossy flat rosettes of the endemic coastal Plantago. 



On stony beaches the vegetation is scanty, and consists of a few patches of 

 Crassula rnoschata and Scirpus aucklandicus, except where a grassy or Pleurophyllum 



* This matter is gone into at some length in niy rcrently pubHshed report on Stewart Island. 



t The specific, name is quite misleading, since in Chathani Island it grows in abundance on all 

 kinds of soils, and in both wet and dry stations. 



i Undoubtedly the Rumex can grow upwards as it is buried, and tliis gives it, ready-made, a saiid- 

 liinding adaptation. 



