208 KEPOKT — 1873. 



Except in cases of accident, when an engine or train leaves the line, rails of the 

 weight now used in permanent way are never known to be hent bj' the passnge of 

 trains, but brittle rails will break. 



The weight on the driving-wheel of a large engine is about 8 tons ; the amount 

 of vertical fall in passing along the line is necessarily very small ; and we know by 

 experience that this large weight with this small fall is suiEcient to break inferior 

 rails, while it leaves the good ones unbent and uninj ured. 



What we require of the test by impact is that it should be so arranged as to do 

 what the engines do, detect the brittle rails without destroying the good ones ; 

 whereas, as now applied, it destroys the 1 or 2 per cent, of the rails submitted to 

 the test, however good they may be, while it gives no information whatever 

 regarding the remaining 98 or 99 per cent, of the quantity. 



Another test for toughness or ductility which is very useful is the extension of 

 the metal beyond the limit of elasticity. 



In testing his fluid-compressed steel. Sir Joseph Whitworth employs this test 

 upon a piece of the mefnl inches in length. For a length of 2 inches at each 

 end a screw is cut for the purpose of enabling the hydraulic apparatus to bring the 

 strain to bear on the sample. The remaining 2 inches between the screwed portions 

 is accurately turned down until the sectional area is exactly ^ an inch. 



The sample is now subjected to strain, and the recorded extension occasioned by 

 the strain at the moment of rupture is treated as percentage or proportion of the 

 2 inches between the screws, and is described as the percentufie of ductility. 



Eut it is obvious the measure of ductility so obtained has reference to the par- 

 ticular length and dimensions of the specimen, and would be altogether varied if a 

 long bar were tested instead of a short one. 



There is, however, another evidence of ductility which, within certain limits, is 

 independent of length — that is, the diminution of sectional area which takes place 

 at the point of rupture ; and the ratio Avhich the original sectional area of the bar 

 bears to the sectional area of the fractured end appears to aflbrd a more definite 

 measure of ductility. 



Thus in the experiments of Mr. Kirkaldy, previously referred to, it appears that 

 in bars 50 inches long and 1'382 inch diameter, the sectional area of the fractured 

 end was in some cases less than five tenths of the original section. 



In the bars broken by the Committee, which were 14 feet long and 1| inch in 

 diameter, it was in the best samples under six tenths, while the best qualities of 

 wrought iron similarly treated showed a ratio of about five tenths. 



It is to be observed that such a degree of ductility as is presented by these 

 samples is not needed in engineering structures, the wrought iron frequently used, 

 and I may say generally used, for these purposes being of much less ductility. 



Without, however, attempting to say what description of test may be found the 

 best for ascertaining the property of ductility, it may be observed that what is 

 required for this test is a definite basis to act upon, and that the samples should be 

 so made as to render the test cheap, expeditious, and easy of application. 



The next requirement is that when a piece of metal has been tested, and its 

 qualities of strength and toughness ascertained, there should be some means of 

 denoting its quality in an authentic manner". 



To a certain extent this is already done in iron by the mark of the maker ; but 

 something more than this is necessary to fulfil the required conditions in steel. 



W^hat is termed steel, is iron with a small proportion of carbon in it. These two 

 ingredients are necessai-y to constitute steel ; and there may or may not be present 

 in very small quantities graphite, silicon, manganese, sulphur, and phosphorus. 



In connexion with the experiments made by the Committee, 14 of the samples 

 were tested by Mr. E. Richards, of the Barrow Steel Works, 5 of which were kindly 

 repeated by Dr. Odling. 



Although there are some discrepancies in the results which we cannot accoimt 

 for, yet some of the characteristics are brought out clearly. 



It appears that manganese maj' be present to the extent of four tenths per cent, 

 without injur}' either to the strength or ductility, but sulphur and phosphorus, 

 except in extremely small quantities, are fatal to ductility. 



in the samples tried by the Committee and Mr. Kirkaldy, the quantity of carbon 



