Canterbury and Westland. 389 



it had not escaped the notice of the earliest settlers, by whom it was 

 generally known as " The Eailroad," and a number of theories were 

 propounded to account for its formation. One of them, believing that 

 a set of faults was running here parallel to each other, was sure that 

 there was a coal-field below ; another, by no means wanting in 

 intelligence, was certain that it was made by the hand of man, as an 

 artificial cover for an army. It was also described to me as a river- 

 bed, partly uplifted or sunk across its course, and only after repeated 

 careful examination was I able to assign to it a purely glacier origin. 

 This glacier channel, as before observed, ascends four river terraces, 

 crossing them diagonally as they gradually approach the lower gorge. 

 Here, where the river has cut through a bar of quartziferous porphyry, 

 they all unite, the river having cut almost straight down for six to 

 700 feet, leaviug nearly perpendicular walls on both sides. If I can 

 trust to some Aneroid observations made at both extremities, the 

 frontal moraine lies about 20 feet above the starting point on the 

 banks of the river. Here, not far from the starting point, a small 

 but well marked frontal moraine crosses the glacier channel (section 

 No. 8). At (B) it crosses a deep channel, in which a branch of the 

 Rakaia was once flowing ; the lateral remains are here very low, being 

 scarcely five feet above the floor. The glacier then ascended the slope 

 of the first terrace, having about 30deg. inclination, and possessing an 

 altitude of about 20 feet. On this slope the lateral moraine is well 

 developed on both sides. From the summit of the terrace the glacier 

 channel continues in the same direction, with well defined banks on 

 both sides, rising about 18 feet above its bed, but standing only six to 

 eight feet above the fluviatile deposits of which the terrace is formed. 

 The glacier has thus been able to lay its bed to that extent lower — 

 some 10 to 12 feet. After half a mile, another small frontal moraine 

 crosses the channel at (C). Having followed the first terrace for about 

 a mile, the second terrace, about 20 feet high, is reached, with a small 

 lagoon at its foot (D). After ascending also this terrace the lateral 

 moraine gradually gets lower, now being only ten feet high, and 

 another small terminal moraine (E) has been thrown down. Two 

 more terraces are afterwards ascended, 10 to 12 feet high. Having 

 reached the highest level of the old river bed, the channel then 

 ascends gradually towards the roclw moutonnee lying in front of it. 

 The ground assumes now a more hummocky surface, and the lateral 

 moraines are less marked. Numerous erratic blocks lie about, and 

 several small channels for the outlet of water are well defined. At 

 two spots only, before the glacier channel reached the highest level of 



