It was in 1886 that the first preliminaries in the actual construction of this final link in the coast-to-coast railway was made. The surveys showed the feasibility of the scheme, though it was pointed out that to climb over the Andes would entail work of a peculiar character, and that the cost would be tremendous. The critical point was the negotiation of the summit itself, for the mountain pass is at an altitude of 12,796 feet. Though commenced in 1886, the scheme experienced many changes of fortune which hindered construction time after time. Financial and labour troubles were the two most retarding factors. By 1891 only 57½ miles were open to traffic; four years later only witnessed the passing of the 90th milestone. Such slow progress was deplorable in comparison with the building of the Oroya and Mollendo railways. Then came a delay of four years, but in 1899 work was resumed and was pushed forward to completion. On the Chilian side, owing to similar troubles, construction was possible only in spurts, and even when the financial details were adjusted satisfactorily, the scarcity of labour remained a thorny problem. The surveys showed that the most practical route westwards from Mendoza was by following the course of the river of that name right into the mountain range. In this manner extensive blasting and heavy cutting could be avoided, except where the mountain-sides dropped abruptly into the river, and then these would have to be tunnelled. The constructional engineers followed this location, but only to run full-tilt into another difficulty which had not been foreseen. The Mendoza is a South American replica of China’s ill-fated Hoang-ho. In the low season its placid waters roll leisurely to the ocean, but when it is swollen by the melting snows it tears along with fiendish velocity. As its banks are composed only of the soft alluvium brought down from the mountains, the foaming waters do not find this a very difficult obstacle, and accordingly carry it away in tremendous quantities. As a result, the river is for ever changing its channel. To the railway engineers such eccentricities proved serious factors. They realised speedily that here was a situation peculiarly exasperating, for long lengths of track were swept away bodily time after time. It never could be anticipated where the turbulent water would break its bounds next. A stretch of permanent way, left safe and sound in the morning, sometimes was wiped out of existence before nightfall. All that could be seen of the work possibly was the rails dipping into the water on one bank and reappearing on the other, the intervening section describing a graceful festoon in the depths of the muddy torrent. At times the waters were more freakish. They would burst upon the track with such violence as to wrench the metals apart; then only the jagged, twisted ends jutting mournfully into the air on either side of the new river channel were the sole remnants of the track.