374 KEPORT— 1867. 



The ultimate elongations are unaccountably variable, and seem much below 

 what might have been expected ; even the greatest elongation, •1437, given 

 in the Table, is below the average for iron bars, whilst the least elongation, 

 •0037, produced by a strain of 38^ tons per square inch, is only about the 

 60th part of this average. 



Compression. — ^Table III. 



Thirty-two of the bars supported each a pressure of 100-7 tons per square 

 inch of section without undergoing any sensibble fracture, whilst twenty- 

 three bars were more or less fractured with this pressure. 



The mean value of the compression per unit of length, given in col. 6 of the 

 Table, taken for 24 of the best specimens, is -372 ; whilst the mean taken for 

 the remaining specimens is '232, giving a general mean deflection of "302. 



The work, u, expended in crushing the material in short columns is re- 

 markably large. The mean value of u, given in col. 7, taken for 26 of the 

 best specimens, is 41300 ; whilst the mean taken for the remaining speci- 

 mens is 25400, giving a general mean value of 33400. 



If 6000 be taken as the value of m, in the case of tensile strain, then the 

 work expended in rupturing the material by compression wUl be 5| times 

 the work expended in rupturing the material by extension. 



Tensile and compressive Resistances compared. 



Taking the mean tensile resistance to rupture at 47^7 tons per square 

 inch, it follows that their resistance to compression is more than double 



(2-1 times) their resistance to extension : thus -2pv7^=2-l. Hence it fol- 

 lows that the most economic form of a steel bar imdergoing transverse strain 

 would be a bar with double flanchcs, having the area of the bottom flauch 

 aboiit double that of the top flanch. 



This conclusion is borne out by the results of experiments on transverse 

 strain, where S^, the strain per square inch of the material at the elastic 

 limit, =6C=6x 6-83 tons=40-98, or 41 tons nearly; but the mean break- 

 ing strain per square inch by extension =47"7 tons, clearly indicating that 

 the compressive resistance in the former case was considerably .in excess of 

 the tensile resistance. 



It is important in every experiment on the strength of materials, which 

 enters so largely into constructive art, that we should be thoroughly ac- 

 quainted with the properties of the material of which the striicture is com- 

 posed, and that its resistance in all the different forms of strain should be 

 clearly and distinctly ascertained. In the foregoing experiments we have 

 determined the resisting powers of the different specimens to bending, ten- 

 sion, and compression ; but we have omitted that of torsion, or twisting, 

 until we have an opportunity of doing so upon the same identical bars. 

 These I hope to accomplish in a separate commiuiication, and also to give 

 some further results on an enlarged scale, calculated to coniinn what has 

 already been done, and to ascertain some additional facts in regard to the 

 changes now in progress in the manufacture of Bessemer steel. 



