326 



for measuring the expansion of 



of no means having been contrived to insure uniformity in the indi- 

 cations of this instrument, and there being no means of testing its 

 accuracy to the extent that many scientific researches demand ; and 

 as corrections for changes of temperature enter into all the investiga- 

 tions mentioned above, uniformity in the results, can be obtained 

 only when temperatures, and the changes of volume resulting from 

 varied temperatures can be accurately measured. 



In all accurate measurements of lineal expansion in sol- 

 ids, the first object is, to have two points which shall remain 

 invariably equidistant at all temperatures within the range 

 of experiment : and the second, is, to provide the means 

 of accurately measuring the variations in the length of the 

 body under examination, when it is placed between these 

 fixed points. 



1 I propose to accomplish the first of these objects, by 

 making use of two bars of different metals, whose lengths 

 are inversely proportional to their expansibilities, on the 

 principle of the compensation pendulum; that is, if both 

 the bars be equally heated, the shorter bar shall expand 

 exactly as much in length as the longer, and the distance 

 a b in the annexed diagram shall be equal, at one temper- 

 ature, to a' V at another. 



To verify the accuracy of this equal expansion, I pro- 

 pose to use the combined bars fastened together at (c) as 

 a balance beam ; the masses being so adjusted, as to throw 

 the center of gravity in the vertical plane passing through 

 (a), and perpendicular to the axis of the beam. A delicate 

 knife edge is attached to the end of the short bar at (a), 

 like that of a balance, and it rests on polished cylinders of 

 glass, transversely to their axes. The friction is thus ren- 

 dered null, by the contact of the knife edge and the sup- 



porting cylinders being reduced almost to a mathematical 

 point. After the center of gravity shall have been brought indefi- 

 nitely near to the knife edges, and below them by the ordinary ad- 

 justments of a balance, so as to render it as delicate a balance as 

 the inertia of the mass will allow, the combined mass is ready for 



ality 



If it be now subjected 



to varied temperatures, (being exposed to each a sufficient time for 

 its mass to be uniformly heated,) and it remains in equilibrio, the 



