Ecoloijical Bolaiuj.] SUBANTARCTIC ISLANDS OF NEW ZEALAND. 201 



foetidissima, also at times a forest-tree, is still more striking, since in the wind-swept 

 o^ien it becomes a prostrate plant, creeping on the ground beneath the shelter of the 

 tussocks. So, too, on the Auckland Islands do Coprosma ■pnrviflom and the stiff- 

 stemmed, tomentose-leaved Cassinia Vauvilliersii, both xerophytes, form mats upon 

 the ground. Finally, on Antipodes Island, small plants, 20 cm. tall and less, of 

 the divaricatingly branched xerophyte Coprosma cuneata are common.* 



A somewhat similar case to the above is that of Ranunculus pimjuis. On the 

 Auckland Islands this plant varies so greatly according to station that it was only 

 by finding a series of intermediate forms that I could believe there were not two 

 distinct species in question. Even Hooker described two varieties as well as the 

 type. But if we consider a systematic variety as an hereditary entity — and any 

 other definition would frequently lead to the making of dozens of varieties — then 

 these distinctions of Hooker's are incorrect. Where there is perfect shelter, as in 

 small open spaces in the subalpine scrub, R. pinguis may have a far-spreading and 

 branching rhizome, a 1- or 2-flowered scape more than 28 cm. tall, furnished with 

 long-petioled bracts, and leaves 6 cm. by 7 cm., with petioles 23 cm. or more in 

 length ; but on the open wet peaty or stony ground this same species may consist of 

 a small rosette of leaves each 1 cm. by 9 mm., with petiole 5 mm., and a virtually 

 sessile flower in its centre. 



The cushion form, so characteristic of the subantarctic islands, seems at first 

 thought one of extreme stability, and yet in certain instances it is little more 

 '"fixed" than the tree form cited above. Thus, Professor C. Chilton brought back 

 from Camjibell Island a cushion of PJiyUacJme davigera, which he put in the moist 

 chamber attached to his laboratory. Within a few weeks merely, a remarkable 

 change came about : the shoots lengthened, the internodes became longer, and a 

 plant of quite a loose habit and an altogether different appearance from the normal 

 was produced. And yet this moist-air form must be considered just as much a 

 specific form as the dense cushion of the subantarctic bog, which latter is evidently 

 little more fixed than is the wind-shorn Metrosuhros lucida or the mat form of the 

 mesophytic forest-shrub Coprosma foetidissima. 



HfimenophjjUum rmdtifidtim is a case similar to the last. This excessive 

 hygrophyte, when growing on subalpine cliffs or wet peaty ground, has its fronds 

 archuig, and their segments all closely curled together ; nor does it in those stations 

 at any time assume any other habit. Yet when brought into a moist chamber, as 

 I have already shownt, it puts forth new fronds of the typical open form, resembling 

 exactly those of the interior of any New Zealand rain forest. 



A rather different case to any of the above is that of Poa litorosa. Where this 

 grass occurs in wet j^eaty meadows or swamps it builds ujj a tall trunk of its dead 

 roots, leaves, and stems ; but on rock-faces near the shore the grass grows in abund- 

 ance, but the trunk-building habit is wanting, and the leaves droop down the rock 

 like long coarse locks of loose hair. Poa foliosa behaves in a similar manner, both 

 on cliffs and on flat ground, in either of which stations trunkless plants may 

 occur (fig. 7). 



* So far as the Antipodes plant goes, the stimulus derived from its environment seems deep- 

 seated, since a plant in the experiment plot at Canterbury College, planted by me in 1903, at the 

 present time (August, 1909) has hardly increased in size, though in robust health. 



f Cockayne, " 15utanical Excursion," p. 267. 



