Wii'es of varying Elasticity, 99 



the elastic properties of certain modifications in the nature of 

 the wire. Thus one definite length of the wire might be 

 heated to any required temperature, while adjacent portions 

 were by cold water or ice kept at a low temperature, then 

 (24) or (34) would indicate the effect either on the modulus 

 of elasticity or on the rigidity, allowance being made for 

 direct expansion by heat. So, again, observations made after 

 part of a wire exposed for some time to a definite high 

 temperature had cooled, would indicate the consequent modi- 

 fication of the elastic structure. Or, again, an electric current 

 might be run through a definite portion of the wire, while 

 the rest was exposed to no such influence, any heating-effect 

 being avoided by surrounding the whole wire with cold water 

 if necessary. By clamping the wire at different points, select 

 portions might be subjected to severe strain, while the rest 

 was free from strain ; then removing this strain and observing 

 the values of w or v at different points in the wire, now 

 exposed to one and the same force of tension or torsion, the 

 modification by strain of the elastic properties might be 

 deduced. Other applications will suggest themselves. 



The advantage of this method is that in every case the 

 same wire is dealt with and the uniformity of its original 

 structure may previously have been tested. Further, there 

 is no comparison of results deduced from experiments in 

 which any variation of the mode of application of the forces 

 of tension or of torsion may have occurred. In two sets of 

 experiments, even with the same wire, there is always risk 

 of error through variation in the suddenness with which the 

 stresses are applied. Again, if, as usual, the whole of the wire 

 experimented on has been subjected to certain modifications, 

 it can seldom, if ever, be restored exactly to its original 

 condition, so that one experimental comparison cannot be 

 repeated with accuracy, and in considering the effect of 

 other physical changes a continually increasing modification 

 in structure is introduced ; whereas in the single wire locally 

 and definitely modified, a comparison may be instituted 

 between any given portion and an unmodified portion sub- 

 jected to the same stress, neglecting or allowing for the 

 weight of the intermediate portion in the case of tension. In 

 many cases also it would be much easier to modify only a 

 comparatively short portion of a long wire; and from the 

 preceding results the difference between the observations 

 made, on a given length of the wire, before and after the 

 modification, would accurately disclose the change of state in 

 the small modified portion. 



The results obtained from the consideration of the torsional 



H2 



