produced in Compound Bars cf Mstat by Change oj Temperature. 63 



every one alike Increafed while the lines joining the lower extremities continue ftrait and unal- 

 tered in length, it may eafily beihewn that the right line will become regularly polygonal: and 

 if the number of divifions were indefinitely great, this polygonal line would become circular. 

 The Heel and brafs, in our experiments, are referable to this theorem ; for, while they 

 continue flrait, an indefinite number of perpendiculars may be conceived to be drawn 

 from the furfacc of the fleel to the furface of the brafs, at right angles to the plane of 

 contact. Change of temperature will render one bar longer than the other ; the perpen- 

 diculars will diverge, and the right line will become a circle. 



The (horter the perpendiculars, that is to fay, the thinner the bars, the greater will be 

 the divergency from a given difference of length, and confequently the fmallerthe circle j 

 that is to fay, the diameters will be nearly in proportion to the thicknefies. Acsordingly 

 it is found that very thin bars of this conftruftion are much more bended by the fame 

 change of temperature than fuch as are thicker. If a narrow piece of the thin tinned iron 

 called flieet-tin, about eight or ten inches long, be rubbed bright, and wetted with a fo- 

 lution of fal ammoniac, and upon this there be applied an equal tliin flip of fheet brafs, alfo 

 brightened, and wetted with the fame folutlon, and the faces be prefled firmly together by 

 confining them between two thick ftrait bars of metal, bound together with wire (while 

 in the vice), the whole apparatus may be then gently heated, till the coating of the tin is 

 fufed, and folders the faces together. This period is known by placing a fmall piece of tin 

 upon the outfide, which will flow when the heat is fufficiently elevated. When the 

 compound bar thus obtained is bended up in the figure of a pair of tea-tongs, the change 

 of pofition of its extremities is very perceptible, and is fuch that (the iron being outfide) 

 it not only opens vifibly when held to the fire or touched by the hand, but even when 

 breathed upon. 



The pradlical advantage of this conftru£lion is, that the range of motion greatly exceeds 

 that of the mere expanfion or contrailion. The difadvantage to be fufpeiSed is, that the 

 properties of the combination may change, fo that the original motion produced from a 

 known alteration of temperature may hot take place after feme years heating and cooling. 



When a thin bar of brafs is attached to another of Heel of nearly the fame tenacity and 

 flexibility, and both become expanded by heat, the form of a circular arc will be alTumed, 

 of which the radius may be difcovered by admeafurement ; and the difference in length of 

 the two bars afcertained by an eafy calculation. Or converfely, if their expanfion be known 

 at any given temperature, it will be eafy to determine th? curve. In order to afcertain 

 from dirc£l experiment, without reference to pyrometrical tables, the proportion between 

 the lengthening of the brafs beyond the ftcel, and the fpace moved through by the ex- 

 tremity of a bar of this kind, I had recourfc to experiment. A bar of plate (lecl, 0.033 

 inch thick, 0.55 inch wide, and 6.4 inches long, was riveted to a perfeftly fimilar and equal 

 bar of plate brafs by fmall iron rivets, as in Figure i, Plate V. It was then pinned at 

 one end againfl a flrait bar of copper, which it touclied througlvits wliole length, from 

 the pin to the otiier end that was left at liberty. The dillance from the pin to the free 

 end was 6.35 inches. Immediately oppofitc the lafl-mcntioned end a pin was fixed upright 

 in the copper, very near the end, but not touching it. On the pin was flided a fprinjj 

 clip, with a Qiort tail as at C in Figure 2, where AB rcprcfents the compound bar and 

 l)li the pin. The brafs fide of the bar was next the copper. When the temperature was 40" 



of 



