252 Mr. Mallet on the Physical Conditions 



of an error, be abstracted from the average deduced from the remainder ; and 

 it seems wholly disproved by the experiments of Mr. Edwin Clarke (Britan. 

 Bridge, vol. i. p. 377), and by those of Navier (" Applic. de la Mecanique," 

 t. i. p. 30). The former gentleman, Mr. E. Clarke, whose experiments are by 

 much the most important we possess, inasmuch as he alone has attended to the 

 relative extension of the iron in either direction, found that bars cut longitudi- 

 nally and transversely from the same plate of fine fibrous iron of excellent 

 quality, were broken by strains per square inch of section of — 



Tons. Tons. 



In the direction of the fibre, . . 19-66 to 20-2 

 Across the fibre, 16-93 to 167 



— and that the ultimate extension of the plate in the line ofthejihre was double 

 as great as transverse to it. The latter, from the mean of ten experiments, found 

 the ultimate strength in the line of the lamination and fibre, to that transverse 

 to the same, in the ratio of 40-8 : 36-4 ; the iron being of a stiffness, that it began 

 to extend sensibly under from ^ to | the strain of rupture. 



The explanation offered by Mr. Fairbairn, that the difference may be owing 

 to better modes of " piling the rough bars," i. e., crossing them, before rolling, 

 cannot afiect the question. The principles here enunciated, upon which the 

 final direction of the fibre depends, as well as the facts known to every iron- 

 master who rolls boiler-plate, assure us, that no matter how the rough bars are 

 crossed or piled, the fibre of the rolled plate, if of well-manufactured iron, is 

 uniformly in the direction of lamination. And, were it otherwise, Mr. Fair- 

 bairn's experiments would be wholly inconclusive, as having been made on 

 iron, confessedly not having a distinctly longitudinal fibre, and, therefore, unfit 

 for the proposed inquiry. 



212. Taking the means of Mr. E. Clarke's experiments, then, at 20 tons 

 longitudinal, and 17 tons transverse, the value of the coeSicient TV in each case 

 will be — 



In the line of fibre, = 234-84 

 Across the fibre, = 30-47 



— taking the total extension = -0016 in the first, and half that in the second. 

 We find, therefore, that the elastic range ofuirought-iron, of any given gtmlity, 



