In a far-away corner of the same country, Upper Burma, may be seen another interesting example of the bridge-builder’s craft, carried out under particularly exacting conditions in a forbidding country. This is the Gokteik Viaduct, which carries the metre-gauge single track of the Burma Railway Company across the gorge of the same name. This structure was completed by the Pennsylvania Steel Company, of Steelton, Pennsylvania, and the award of the contract was criticised severely in Great Britain. But the Government wanted the valley spanned in the shortest possible space of time and at a moderate price. When the tenders invited from all parts of the world were opened, it was found that the British firms had been outclassed by their American rivals in both these essential factors. The location of the railway across this gorge was beset with peculiar difficulties. The question of the approach was trying to decide to the best advantage, and in fact so many surveys were made that one of the American engineers remarked “that he could not see the side of the cliff for survey pegs.” [Illustration: THE TRAINING-BUND UNDER CONSTRUCTION BY NATIVE LABOUR. AT THE EXTREME RIGHT A PIER IS BEING BUILT FOR THE BRIDGE] [Illustration: GENERAL VIEW OF THE PIER-BUILDING OPERATIONS FOR THE CURZON BRIDGE ACROSS THE GANGES AT ALLAHABAD The bridge is 3000 feet long. Temporary line at right.] The gorge itself is also somewhat strange; in fact, it is a curious wonder of Nature. The Chungzoune River flows through the rift, but out of sight, its course being through a natural tunnel, into which it disappears suddenly at a depth of 500 feet. When the line was first surveyed it was in accordance with a low viaduct, the approach thereto being over a suggested section of rack railway working on the Abt system, with grades of 1 in 12½. This was subsequently abandoned, and the surveyor was called upon to find a fresh location so as to eliminate the rack railway, and to give grades not exceeding 1 in 25, so as to permit the line to be worked by adhesion. This decision raised the height of the towers by 70 feet and increased its length to 1,350 feet. Even this did not meet with approval, for after prolonged deliberation a third location was demanded, to give an easier line yet. In this last survey the gradients were flattened to 1 in 40, with an attendant increase in the height of the structure as well as of its length. It was found impossible to improve upon the viaduct itself, so further surveys were carried out to improve the approaches, reducing their length and introducing curves at either end of the viaduct.