312 



Sir W. Thomson on the Fracture of Brittle 



[Feb. 25, 



sensitiveness of the galvanometer was exactly the same in all the observa- 

 tions, still it was probably not greatly different. 



Observations of the heat of the stars, if strictly comparable, might be of 

 value, in connexion with the spectra of their light, to help us to determine 

 the condition of the matter from which the light was emitted in different 

 stars. 



I hope at a future time to resume this inquiry with a larger telescope, 

 and to obtain some approximate value of the quantity of heat received at the 

 earth from the brighter stars. 



IV. " On the Fracture of Brittle and Viscous Solids by f Shearing/ 33 

 By Sir W. Thomson, F.B.S. Received January 2, 1889. 



On recently visiting Mr. Kirkaldy's testing works, the Grove, South- 

 wark, I was much struck with the appearances presented by some speci- 

 mens of iron and steel round bars which had been broken by torsion. 

 Some of them were broken right across, as nearly as may be in a plane 

 perpendicular to the axis of the bar. On examining these I perceived 

 that they had all yielded through a great degree to distortion before having 

 broken. I therefore looked for bars of hardened steel which had been 

 tested similarly, and found many beautiful specimens in Mr. Kirkaldy's 

 museum. These, without exception, showed complicated surfaces of frac- 

 ture, which were such as to demonstrate, as part of the whole effect in each 

 case, a spiral fissure round the circumference of the cylinder at an angle 

 of about 45° to the length. This is just what is to be expected when we 

 consider that if A B D C (fig. J) represent an infinitesimal square on the 

 surface of a round bar with its sides A C and B D parallel to the axis of the 

 cylinder, before torsion, and ABD'C the figure into which this square 

 becomes distorted just before rupture, the diagonal A D has become elon- 

 gated to the length A D', and the diagonal B C has become contracted to 

 the length B C, and that therefore there must be maximum tension every- 



Fig, 1. Fig. 2. 



where, across the spiral of which B C is an infinitely short portion. But 

 the specimens are remarkable as showing in softer or more viscous solids 

 a tendency to break parallel to the surfaces of "shearing" AB, CD, 

 rather than in surfaces inclined to these at an angle of 45°. Through the 

 kindness of Mr. Kirkaldy, his specimens of both kinds are now exhibited 



