Day — Residual Viscosity on Thermal Expansion. 343 



which, while they are attained by comparatively simple means, 

 yet add greatly to the value of the results obtained by its use; 

 notably, the way in which a base of constant length for a reason- 

 able length of time is secured, and, also, the construction of the 

 parts which suppout the movable levers, so as to accommodate 

 the use of the apparatus to bars of any length. 



A bar of the solid substance whose expansion is to be 

 observed is submerged in the bath, it, which is supported by 

 standards above a solid wooden base, pq. The temperature of 

 the bath is raised by means of a multiple gas-jet suspended 

 beneath it. 



The ends of the bar actuate contact levers of a convenient 

 pattern, s and x, which are free to move about horizontal axes. 

 Two rods, projecting from these levers in the line of their axes, 

 support the adjustable mirrors, u and v. The angular motion 

 of the levers, which measures the expansion of the bar, is 

 determined in the usual way by means of these mirrors and 

 the two telescopes, h and ^, with their attached scales. Any 

 lost motion of the levers is prevented by spiral springs, which, 

 having their ends attached respectively to the tixed parts of 

 the apparatus supporting the levers and to arms on the levers 

 themselves, keep them pressed iirmly aganist the ends of the 

 bar. Thumb-screws serve to regulate the tension of the springs. 



In observing expansions, the work is much facilitated, if, 

 instead of having both levers movable, one lever, with its mir- 

 ror, is maintained in a fixed position by its spring, while all of 

 the motion is exhibited by the other. In this case, the fixity 

 of the non-movable mirror is assured by means of its correspond- 

 ing telescope."^ 



Another arrangement of the optical parts of the apparatus is 

 shown in tig. 2. The telescope h is removed and a scale, h\ 

 fixed in its place. Both of the mirrors have a possible rotation 

 in the direction of the arrows. The reflected scale image 

 reaches the eye at the telescope, k'^ over the path indicated by 

 the dotted lines and arrow heads. Since the successive deflec- 

 tions of the scale-image by the two mirrors are of the same 

 sign, the deflection observed at the telescope will correctly 

 measure the sum of all possible motions of the levers due to 

 either expansion or contraction of the bar. It is clear that the 

 lengths of the contact levers will have to be proportional to the 

 corresponding scale distances, that of the stationary mirror, u', 

 being the longer. 



A base of constant length is secured by the following simple 

 means: A bar of \^" galvanized iron pipe, aa^ is supported in 

 a horizontal position directly above the bath upon uprights, hh, 



* I am aware that similar observations with two telescopes were made on a 

 certain occasion by Prof. W. Hallock. 



