ridges. (See Plate 3, fig. a, and compare Healy and Willett, 1938, p. 13.) The western margin of the block is not nearly so continuous as the steep slightly recessed eastward slopes of the Bryneira Range on the further side of the relatively low country occupied by the ultrabasic rocks, but is deeply trenched by numerous valleys extending back into cirques, some of which contain small glaciers. The shouldered spurs and ridges above the 5000–5500 level lead back to a highland area, on which lie nearly twenty square miles of glaciers, some of which descend with a comparatively gentle slope from the peaks and arětes, before discharging into steeply sloping hanging or valley-glaciers. The Ice Plateau discharging eastward by the Andy Glacier affords a notable example of these. The topography repeats on a broader scale the features of Skippers Range, the higher parts of which were moulded by the headward erosion of glaciers on the portion of a matureland surface remaining above the heads of youthful gorges, which were rapidly cutting back into the recently elevated Fiordland region just before glacial conditions became established. That the greater extension of ice should be on the eastern or lee-side rather than the western side of the main divide accords with the general experience in other lands. (e.g. Gilbert, 1904.) Attention was first called to it by McKerrow (1863) when discussing the sudden floods among these mountains. “Change of temperature is the necessary and immediate cause, but … a flood may occur without any perceptible increase of temperature for the wind by transporting the snow to a lower altitude occasions the same effect as a rise of temperature … The snow-line on the north-west (windy) side of the mountain range was higher than on the south-east, the sheltered side, showing the wind is a very decided cause in producing the effects now under consideration.” In addition to this the familiar contrast between the heating effects of the morning and afternoon sun may be recalled (cf. Garwood, 1910). It was among the peaks, arětes and glaciers of this portion of the region herein discussed, that the bulk of the exploratory work of one of us (Holloway) was carried out. East of this block the Dart Valley is cut to a depth of nearly 6000 feet below the summit-level, and probably follows a shatter-zone, which continued southwards, determined the western side of Lake Wakatipu, while northward it is followed by the Beansburn. Above the junction with this stream, the Dart follows a north-north-east to south-south-west direction. Up to the junction with the Brideburn, the valley has a generally alluviated floor and a fall that averages about 20 feet per mile, but above this it narrows and rises nearly 500 feet within two miles before opening out into Cattle Flat. East of the Lower and Middle Dart Valley is the Forbes Range, the topography of which has been described by Sims (1931), Henderson (1937), Healy and Willett (1938). It consists of foliated quartz-albite-sericite schists, often chloritic or rich in epidote, possessing a general dip towards the west at angles which near the summit of Mount Earnslaw, may decrease to about 20 degrees. The mountain rises a thousand feet above the general summit-level, and though it