C—l

108

or scattered in hopeless confusion in the bush along the banks. I think the last 6or 8 chains in the bush, before the top of the cataract is reached, was about the hardest piece of travelling I have experienced, owing to the difficulty of finding a way through and over these boulders. In places one would be nearly out into good going when a gap 50ft. deep and 20ft. broad on three sides of what turned out to be a boulder would necessitate a long detour back, to begin finding another route. It formed a maze of the first order. Opposite Regina Creek is an island (Crusoe's Island), and on the west side of the river a knoll (Queen's Knoll) of some 70ft. to 100 ft. high, both formed of large erratic boulders. This looked like the remnant of an old terminal moraine, and on following the probable line towards Mount McGloin I picked up several smaller heaps of erractics in the bush, proving that a terminal moraine of considerable size once existed here. On the hills where bush can find a hold round Cassell's Flat rata is the prevailing timber, but on the flat and lower terraces ribbon-wood, stunted totaras, cedars, &c, prevail, with a few open patches of grass. Here the river flows in deep smooth reaches, with short rapids between them, and is easily fordable in two places below the inflow of the Twain, and in several places between that point and the gorge of the Karangaroa, which shows clear water after the winter snow has melted—that is, after Christmas. At the upper end of Cassell's Flat the valley gradually becomes narrower and the river more rapid, until the whole floor of the valley rises suddenly to Camp 4, the river descending from the upper valley through a fine gorge and over two large cataracts. Above Camp 4 there are several stretches of open gravel beaches with rocky bluffs on the right bank, while the left bank is rockbound the whole distance from Cassell's Flat to within a mile of Troyte River. At this point the valley-bottom again becomes filled with glacial deposits in the shape of large boulders and small moranic drift, and the banks slope gently back to a height of some two or three hundred feet to the foot of the high precipices and terraces of bare ice-worn rock, with patches of stunted and " vagrant " vegetation with "no visible means of support." After passing Troyte River the valley becomes rather narrower, the hills gradually drawing together till the saddle is reached, but the river descends through a few small flats, which give one short spells of easy travelling. Stan-ding- on the Karangaroa Saddle the whole valley to the great cataracts can be seen at a glance, and forms a very interesting picture. On the right the hills slope back from the river at an average angle of 35°, showing in some places for the first three miles shattered bluffs and rock-faces at the top of the ridge with spurs starting out from the face of a precipice, and spreading towards the river in graceful slopes, while beyond the hills assume a different aspect, owing to the covering of soil and glacial deposit having slipped bodily into the river, notably at Lily and Coleridge Creeks, leaving bare rock-slopes, at an angle varying from 32° to 36°, and rising from the river's edge for some 2,000 ft. or 3,000 ft., interrupted here and there by low precipitous bluffs. Points A and AA were fixed on such a slope, and it was only by taking off my boots that I could go up the rockslope ; without boots it was easy travelling. The dog I had was also happy as long as we were moving, but when I stopped to take bearings it was very amusing to watch him trying to sit down, first facing up hill and then down, but the angle was too great for him to find a comfortable seat. On the immediate left while standing on the saddle is a precipitous range of hard gneiss rock, a narrow " knife-edge " ridge, dropping in precipices and terraces into the river, or rather to the slopes of debris at its foot. Above Christmas Flat (as I have named the basin at the head of the valley) there are three very distinct glacier terraces, but further down the valley I was unable to pick them up again with any distinctness. The Troyte River cuts through a fine unpassable gorge and drains Mount Fettes and Townshend, with one small glacier on the former, The spur which forms the left bank of this gorge is of the same hard rock, and is cut up into terraces and hummocks to an extraordinary extent, not unlike the terraced vineyards of Italy and the East are often represented in pictures. From Camp 5 down to the top of the great cataracts this side is rock-bound, the river running for a few chains on the top of an almost flat surface of rock, and then cutting the narrow and picturesque gorge, previously mentioned, through the formation. Below this it opens out on the right side into beaches and small flats; while the terrace of solid rock on the left side gradually gets higher as the river-bed falls, until at the mouth of the big gorge below Camp 4 it is fully 200 ft. in height. From this terrace the hill slopes back fairly gently, interrupted by two or three terraces of various heights, until it reaches the foot of towering cliffs of very hard rock. Whenever a creek descends from this range there is an open course of some 20 yards in breadth where the bare smooth rock is exposed, and it is then seen that the surface soil on which the bush grows is only a foot or two deep from the river right back to the foot of the precipices. On coming up the river the rata bush, which prevails in the valley, came to an abrupt end at the top of the cataracts, and the rock-slope and terrace was overgrown with good mountain-birch bush, continuing till Camp 5 was reached, when it stopped as suddenly as it had begun, and its place was taken by stunted rata and mountain vegetation. On the right bank, however, only a very few isolated birch-trees were seen growing amongst the ratas and other bush. But up at Christmas Flat I saw a patch of about a dozen large birches towering above the low alpine scrub. This terrace or rock-floor extending from Camp 5 to the cataract is, I believe, the old valley-bottom, which the glacial deposits left by the retreating ice and debris coming down in the ordinary course of things from the hillsides has covered from the saddle down to Camp 5, after which the river, meeting with less debris, has been gradually cutting down the floor, leaving the terrace above described, until it came to the rapid descent into Cassell's Flat, where it ha.s cut the deep rockbound gorge in which the two cataracts are found. The actual cataracts are due to the large erratic boulders left behind by the glacier in its retreat forming a bar across the valley, and, the hills being too steep to hold them, they have fallen and accumulated at the bottom of the valley and gradually collected in the gorge as the water has cut away the ground under-