November i6, 191 i] 



NATURE 



93 



Reverting to experimental work in this country, early 

 in the last century George Rennie made investigations of 

 the resistance of structural materials, and Peter Barlow 

 made experiments on the strength of timber at the Dock- 

 yard and at the Arsenal. This led to his association with 

 engineers ; and he assisted Telford in calculations for the 

 Menai Suspension Bridge. His " Essay on the Strength 

 of Timber," in 1817, when developed in later editions, may 

 be regarded as the first general treatise in English on the 

 -:rength of materials. 



Down to about the end of the first quarter of the last 

 Lcntury most of the knowledge of strength of materials 

 was due to the work of Continental engineers and 

 physicists. Then an advance was made here. Thomas 

 Tredgold in 1820 published " A Treatise on the Principles 

 of Carpentry." This dealt scientifically and practically 

 with all the then known data of resistance. His book on 

 the steam engine, dealing specially with questions of 

 strength, was published in 1827, and republished down to 

 1850. It has been rather a fashion amongst elasticians to 

 ignore or depreciate the work of Tredgold. But he had a 

 practical insight into what was important and what was 

 negligible in engineering problems greater than that of 

 writers with more ample mathematical resources, and he 

 rendered essential services to the engineers of his time. 



A little later Prof. Eaton Hodgkinson began the re- 

 searches on strength of materials which give him a fore- 

 most place among careful experimenters. He is credited 

 with the discovery of permanent set and of the position 

 of the neutral axis. His paper of 1830 on iron beams, 

 and that of 1840 on columns, were very valuable contribu- 

 tions to practical science. 



There is not time to trace further the history of the 

 science of strength of materials. Experimenters and 

 laboratories in the last fifty years have increased 

 enormously, and theoretical investigations have been 

 pushed to great lengths. But I wish to take the oppor- 

 tunity of paying a tribute to one who seems to me the 

 prince of observers in this branch of science. I mean the 

 late Prof. Bauschinger, of Munich, professor of mechanics 

 and graphic statics at the Technical High School at 

 Munich for twenty-five years. There he established a 

 laboratory, rather more for research than instruction, 

 where engineering experiments were carried out with a 

 thoroughness and delicate accuracy never previously 

 equalled. In 1868 he published the result of indicator 

 observations. But his special field of work was that of 

 tpsts of materials. He created the first public laboratorv 

 for this purpose. He first applied Gauss's method of 

 reading by reflection in instruments for measuring the 

 deformation of bodies when strained. 



Amongst many researches remarkable for their extent 

 it is only possible just to mention one on cements and 

 cement and lime mortars, demonstrating, amongst other 

 things, the importance of fine grinding, and one on the 

 building stones of Germany. His researches on timber 

 first indicated the precautions necessary for securing com- 

 parable results, especially the law of variation of strength 

 with moisture. He carried out many researches on cast 

 iron, wrought iron, and steel, especially some with refer- 

 ence to the variation of the position of the elastic limit 

 under different conditions of straining. Perhaps one of 

 his most important achievements was the foundation of 

 the International Association for Testing Materials. 



Testing Materials for Quality. 

 Down to the middle of the last century the only gener- 

 ally used tests of the quality of iron and steel used in 

 construction were bend-tests, and in certain cases shock- 

 tests. Such other researches as were made were directed 

 to a different object — either to determine the constants in 

 formulae on which engineers relied or to prove the 

 sufficiency of complete structural members. It was the 

 mtroduction of Bessemer steel, and cases of unexpected 

 failure of steel structures, which forced on engineers the 

 necessity of syst<-matic tests of quality. An important 

 series of tests carried out by the late Mr. D. Kirkaldy in 

 i860 led to the adoption of a tension-test as the usual test 

 of quality. Its special merit is that exact figures can be 

 specified for elastic limit, resistance to fracture, and 

 elongation. 



NO. 2194, VOL. 88] 



With the introduction of definite tests, the importance of 

 accurate and reasonable specifications became urgent. 

 Recently the work of the Engineering Standards Com- 

 mittee has done very much to guide the engineer in 

 securing trustworthy material without imposing conditions 

 too irksome or costly on the producer. Hence it has 

 come about that the testing engineer has been created, 

 having functions partly as an investigator of the properties 

 of materials, partly as adviser of manufacturers, and 

 partly as inspector of material. 



Application of the Science of Strength of Materials to 

 Practice. 



The general object of the accumulation of experimental 

 and theoretical knowledge of strength of materials has 

 been to determine the minimum amount of material and 

 the best disposition of it in machines and structures to 

 secure safety. Putting it another way, by theory the 

 stress conditions due to any given straining action can be 

 calculated ; then it has to be determined what is the 

 greatest permissible working stress, and in what way does 

 it depend on such physical properties of the material as 

 can be ascertained by testing. 



The oldest, and still the most common, method of pro- 

 ceeding is to reduce the straining actions to simple tension, 

 thrust or shear, by calculations based on the assumption 

 of Hooke's law, and to provide material enough to limit 

 the intensit}' of these stresses to a fraction of the breaking 

 strength of test-pieces similarly strained. The ratio of 

 the breaking strength to the working strength is termed 

 the factor of safety. In by far the largest number of 

 cases with which an engineer has to deal, the breaking 

 strength of a structure cannot be calculated, for Hooke's 

 law ceases to be true for stresses much below the breaking 

 stresses. Hence the engineer sometimes proceeds one step 

 further. He tests a scale model of a structure to break- 

 ing, and from this deduces the breaking load of the full- 

 size structure by the law of similarity. 



In the case of complex structures, the determination of 

 the exact maximum stresses in which are beyond the 

 resources of mathematical analysis, the value of model 

 experiments is unquestioned. 



Now it is easy to show that the ratio of the breaking 

 stress to the working stress, or the breaking load of a 

 girder to the working load, is not a real factor of safety : 

 the point at which danger occurs in different cases is not 

 a fixed fraction of either the real or the calculated break- 

 ing load. It has even been contended, with some plausi- 

 bility, that the fashion of dividing the breaking stress by a 

 factor to find the working stress is a barbarous method. 

 The factor must be varied for different conditions, and 

 can onlv be fixed empirically. 



The ductile materials chiefly used in construction yield 

 or suffer a large deformation at about half to two-thirds of 

 the breaking stress. The deformations, if the yield stress 

 is exceeded, would be, at least in very many cases, far 

 too large to be tolerated in either machines or structures. 

 If the vield stresses are taken as the limits of safety, then 

 the real factor of safety is only about half the ratio of the 

 breaking to the working stress. Further, the ratio of the 

 yield stress to the breaking stress is not a very constant 

 ratio. 



Shall we, then, drop the breaking strength and adopt 

 the vield-point as the measure of the constructive value 

 of a material? Manv constructive materials, such as 

 stone, timber, and cast' iron, have no yield-point. Besides, 

 even for ductile rolled material, such as mild steel, the 

 point ordinarily determined as the yield-point is not a very 

 fixed point for a given material. It depends a little on 

 the rate of loading. Yielding really begins at the elastic 

 limit, a point below the yield-point, spreading along the 

 bar and becoming general over the bar at the yield-point. 



Some writers assume that the elastic limit as determined 

 in a tension test is the real measure of the constructive 

 strength of a material : but I am not sure that this is not. 

 when rigidly examined, the most ambiguous and elusive 

 of all the measures proposed. In what sense, then, is the 

 elastic limit found in a tension-test to be understood? 



No doubt there are cases where the primitive tensile 

 elastic limit does fix a superior limit to the stress consistent 

 with safety. But it cannot be taken generally as an exact 



