



328 Instrument for measuring the expansion of Solid Bodies. 



* 

 * 



the atmosphere, this source of inaccuracy, if metals of high fusing- 

 , points be selected, becomes almost infinitely small, and would scarce- 

 ly be appreciated, even by the rigid verification above proposed. 



The second point is, to provide the means of measuring accurate- 

 * ]y the variations in length of bodies placed for experiment between 

 the fixed points. . , 



This may be done by means of a micrometer screw with a grad- 

 uated head and vernier attached : or in a better way by the bar of 

 experiment C, which abuts firmly against the more exhaustible bar 

 of the instrument A acting; at its free end against the screw d. This 

 , screw is made with a very oblique thread so that it will thrust out 

 with a small force and turn at the same time on its axis. By means 

 of the rotation of this screw, the expansion of the bar C between 

 the two invariable points MN may be measured. 



N M 



The rotation of the screw d may be measured by a system of 

 wheel work like that of a watch, or, (as this is liable to some inac- 

 curacies from complication) by means of a telescope mounted on 

 the axis of the screw and perpendicular to it, and ranging over a 

 graduated arc with a vernier, at a convenient distance. In this way 

 a minute of a degree equal to ^rio o °f a rotation, and perhaps a 

 second of a degree equal ta TTV |» 7V of a rotation of the screw 

 might be measured. The limits of the delicacy of this determina- 

 tion will depend on the smallness of the screw which must be strong 

 enough for the purpose named, and on the distance of the arc at 

 which minute divisions can best be seen by the telescope. This 

 method seems to be susceptible of much greater accuracy and more 

 rigid verification than any one hitherto employed. In the practical 

 use of the instrument there are only two corrections to be made and 

 they are of the nature of constants which are to be applied in all 

 the experiments except when the experimental bar is of the same 

 metal as the screw. They are, 1st, the expansion of the screw d 

 from o to o', and 2d, the correction for the expansion of a steel plate 

 of known thickness between the screw and end of the bar C. This 

 steel plate is acted against by the point of the screw, because this 

 point would otherwise indent the bar and the screw be thrust out a 

 less distance than the actual expansion. 





