Ciiycalar^ produced in Compound MftaUic Bars by Heat and Cold. 6 1 



airf%vcred, tliat, from the neceflary circulation of heated air, through its difference of denfity, . 

 and^the fpeed with which metals condufl heat, it is not probable that any appreciable dif- 

 fjerence of effe£l can arife from the pendulum itfelf becoming heated or cooled more fpeedily 

 than any fixed part of the apparatus which may be applied to obviate its change of figure;, 

 and ftill more it is remarked, that the fmall variations would correcl each other in their 

 return. Thcfe arguments appear to poficfs confiderable pra£tical force in favour of fixed- 

 compenfations, which in fome iixilances are fimpler and cheaper than fuch as mave with; 

 the pendulum. 



The ordinary principle of thefe combinations is neverthelefs fuch as to imply the ufc of- 

 a lever. If vi'e fuppofe an inflexible bar of brafs, (landing upright upon a fteady fupport, 

 to afford tlie fulcrum to an horizontal lever extending each way ; and that a vertical br.r of. 

 fteel be attached to one extremity of the lever, and faftened beneath to the fame fupport, it is 

 certain that the other extremity of the lever will afcend by heat, and defeend by cold, ac- 

 cordingly as the fulcrum rifes and frills through a greater fpacc than the extremity of the. 

 fleel bar attached to the oppofite end of the lever. If the pendulum, therefore, be fuf- 

 pended by a flexible piece of fpring to that end of the lever which is fartheft from the fleel 

 bar, and the fpring be llightly confined by a notch in a metallic cock, as ufual, the afcent 

 and defcent of the whole pendulum, by the operation of this contrivance, will counteratt 

 die defcent and afcent of the ball by heat and cold. In order that it may accurately do 

 thio, it is requifite that tlie proportion of tlie arms of the lever fliould be variable by ad- 

 j.uftment. 



The direG oppofition of the expanfion of metals is certainly the mofl fimple and efFeftual; 

 and that by levers is evidently lefs fo, though capable of being magnified at pleafure. Tliis 

 direct aftion has been applied in watches to carry a curb or piece of metal which ferved to 

 lengthen or {hortenthe efTedlive part of the pendulum-fpring, accordingly as the difference 

 of temperature, by altering the diameter of the balance, might render it more or lefs eal'y to. 

 be moved with a determinate velocity. But I have not heard of any fimple contrivance of 

 this nature having been applied to the balance itfelf. It might not be dilhcult to confl;ru£l a 

 balance upon the principle of the gridiron pendulum ; but the difficulties of adjuftment for- 

 temperature, pofition, and rate would either prove exceedingly great, or call for expedients, 

 hitherto unthought of, to fimplify the proceffes. A very curious contrivance is at prefent 

 ufcd in our bed portable time-pieces, which diifei^ very much from thofe before defcribed. 

 1 do not know that it has ever been fcientifically treated, nor even who was the artift that 

 firil applied it to watch-work. One artift in a loofe uncertain manner informed me, that 

 he underftood the contrivance to belong to M. Berthoud at Paris; another pofitively allured 

 me, that the well-known Pinchbeck in Cockfpur-ffireet exhibited a metallic folid thermometer 

 on this principle many years ago. Cumming appears to have known nothing of it in 1766, 

 when he publiiiicd his hlcments ; neither do I find it in the French Encyclopcdie pub- 

 lilhed in 1779. I fliould be happy todo jufticc to the inventors by the aflTiilancc of com- 

 munications from corrcfpondcnts. The principle of this contrivance is, fimply, that if the- 

 faces of two ftrait b.irs of metal, differing in their cxpanfibility, be faffened together by 

 riveting or foldciing, they will continue ftrait only fo long as the original temperature 

 remains ; but if the temperature be augmented, the moft cxpanfible metal will become the 

 loDgt-r, and a flexure will take place, fo as to produce convexity on that fide and concavity. 



