observed that reference had been made in the Paper to the question of inherent grain-size, and he would like to ask whether the Author would recommend the London, Midland and Scottish Railway to have some rails made of fine-grain steel, even if they were more expensive, in order to ascertain their wear. He did not think that enough broken rails occurred in Great Britain to justify spending much more per ton for the steel. In that respect British railways were very fortunate in comparison with others, as was shown by the statistics of broken rails given in the Proceedings of the International Railway Congress.The Author had given some very interesting data about the recording of rail failures, and Mr. Wallace had looked up some data of the London, Midland and Scottish Railway, which were admittedly, as he thought the Author would say, merely engineers' data, having been taken by semi-skilled labour instead of by skilled metallurgists. In 349 L.M.S. rails which were taken out before complete fracture took place, there were 502 defects (a defect which went through the web and the head being counted twice) 47 per cent. being in the head, 45 per cent. in the web, and 8 per cent. in the foot.With regard to the position of the fracture in the rail-length, in 667 L.M.S. rails, with a total of 780 defects (a defect extending over two locations being counted twice), 29 per cent. of the defects occurred in the running-on fishplate area, 26 per cent. between that area and the second sleeper, 35 per cent. in the middle of the rail (the middle being from the second sleeper at the running-on end to the last but one on the running-off end), 4 per cent. in the running-off fishplate area, and 6 per cent. between that area and the next sleeper.In those 667 rails the fractures were attributed to the following causes : corrosion fatigue and work fatigue, 30 per cent. ; transverse fissure, 6 per cent. ; inclusions, 21 per cent. ; wheel burn, 8 per cent. ; unallocated, 35 per cent.He had discussed with the Author several times the question of having a rail-failures chart, and he quite agreed .with all that the Author had stated about the necessity of collecting statistical data t o enable the problems which arose to be dealt with intelligently, but he foresaw difficulty in getting failures inspected promptly. It was expensive and troublesome to cut rails and send the pieces to the District Engineer's office, but, on the other hand, to send somebody to the spot from the
29Engineer's staff would involve excessive travelling time for people who were already extremely busy, if not overworked.With regard to the number of fcactures that occurred in cold weather, many years ago, when he was an assistant on a small railway in Ireland, he had kept a record of rail failures on that line and had prepared a graph showing the failures in each quarter of the year. After he had done that for about 12 years, he found that the first quarter of the year led easily, followed by the fourth quarter, whilst the two warm weather quarters were considerably ...
