344 Tronsaciions. 



as sufficient room is allowed eastwards, southwards, and perhaps westwards, 

 for the plants driven out of the naountains to find a haven of refuge. On 

 the contrary, if it is assumed that an ice-sheet has covered the southern 

 end of Greater New Zealand, leaving no land easticards or westwards of the 

 present subantarctic islands, then, for reasons already given by Cockayne 

 (1909, p. 7), the present distribution of species is inexplicable. 



There then remains only to be considered the repcopling of the glaciated 

 area from the east during the retreat of the glaciers and the synchronous 

 advance of the plants. Let us consider, therefore, the state of affairs at the 

 time the terminal faces of the Canterbury glaciers, and of the Rangitata- 

 Rakaia in particular, extended just on to the Canterbury Plain, taking 

 into consideration at the same time the effect of elevation, the generally 

 accepted cause of our glacial period (Hutton. 1900 ; Haast, 1879 ; Park, 

 1910) upon the climate. 



Although the glaciers may have extended not merely on to the plain, 

 but even far beyond the foothills, it by no means follows that the eastern 

 ranges (Mount Somers, Mount Hutt, &c.) were altogether ice-clad, or even 

 snow-covered. On the contrary, increase in altitude of the actual Divide 

 would lead to a drier climate on the eastern ranges than the present one, 

 and to a much diminished snowfall, so that it is quite possible, even at 

 the greatest assum.ed extent of the glaciers (see Park 1909, 1910), these 

 •eastern mountains still preserved more or less of their primitive plant 

 covering. But the increase of drought would bring about a semi-desert 

 climate on the east, and, as Diels and Cockayne have shown, the forest- 

 plants would either perish or be modified (Diels, 1896 ; Cockayne, 1901). 

 So it is assumed that various xeromorphic gro-\\i:h-forms have arisen — 

 e.g., the divaricating, the flat-stemmed leafless, &c. On the mountains the 

 drought-conditions Avould be still more severe, and then may have originated 

 the Baoulia cushions of the rocks and the highly differentiated shingle- 

 slip plants. 



Whether the above speculations are approximately true or the con- 

 trary, the first ground available in quantity for plant-colonization would 

 be the rocky slopes above the ice-line of the valley glaciers, such as are to 

 be seen near any of the large eastern glaciers at the present time. The 

 populating bv plants of such ground would be very similar to the evolution 

 of fell-field, treated of below, while the final soil would be formed by slow 

 degrees from the easily disintegrated rock (greywacke, &c.) together with 

 humus from the plants, the plant covering favouring the formation of a 

 true soil. 



As the ice retreated, the smoothed lower slopes of the hills, the roches 

 montonees, the moraines, and the valley-floors, much of which would be 

 river-bed, Avould be by degrees exposed. Soil of a different character 

 to that mentioned above would finally accumulate, and would consist of 

 the wind-blown rock-flour from the river-beds, of the subglacial material 

 (boulder-clay, &c.) which would be left on the rock, &c., as the ice melted, 

 and, in special positions, of extensive deposits of lateral moraine. That 

 the surface soil of the valleys and lower slopes is principally wind-trans- 

 ported {i.e., loess), though beneath there may be clay, is evidenced by 

 what is going on to-do y during any dry wind, when clouds of dust rise from 

 the river-beds, but such loess may not infrequently be mixed with other 

 soil. 



Clay, formed in situ from the underlying rock, does not appear to be a 

 specially important constituent of the soil. For instance, on certain roches 



