282 INVENTIOX OF CHRO NOMETERS, &C. 



ball containing the spirits of wine, and Is a little bent down- 

 ^vards ; the mercury is in the lower part, and the vertical 

 branch of the tube, A K, is open at the upper end. These two 

 thermometers are placed opposite one another, the axis of the 

 balance being in the same plane with the centrical lines of the 

 tubes ; and the thermometers and balance are solidly attached 

 together, and form a well poised and steady regulator. At 

 the middle temperature of the atmosphere, the quicksilver 

 stands at K A ? K ; but, when an increase of heat occurs, the 

 alcohol, by its expansion, forces the mercury from the exterior 

 branch, ^K^, towards A K, and a portion of the mass of the 

 regulator contracts by approaching the centre. On the con- 

 trary, if the variation consists of an additional degree of cold, 

 the mercury moves towards the exterior branch, and the weight 

 towards the circumference of the balance becomes greater. 

 Thus, if the thermometers are well adjusted, the effects of all 

 the changes of temperature in the balance will be compensa- 

 ted, and the regulator will act with the same uniformity as 

 if its figure were not liable to such alterations. 

 Peter le Roy P. le Roy invented and executed this method of compcn- 



!,*1"°*.^"°^^ sation before he became acquainted with Harrison's contri- 

 «r Harrison's ^ 



expansion cm-b vance of the compound metallic thermometer; and he avows,* 



when he made 



his thermome> * Mcmoire sur la meilleure Maniere dc mesurer Ic Temps en 



ttical balance : jyjer, &c. p. 56 and 57.— This being an intercstinii point, I shall 

 subioin a literal copy of the passage of P. le Roy's memoir, in 

 which that author states his manner of making the compensation 

 balance. 



qiiand j' ai chcrch6 h compenser reflect du chaud et du 



froid p.ir des lames de cuivre et d' acier, rivees ensemble, comme 

 M. Harrison, j' ai tente, non de changer la longueur du spiral, 

 mais dc faire approcher ou eloigner par ce moycn, du centre du 

 balancier, une partie considerable de sa circonferance. Pour 

 ccla, j' ai employe un balancier ffig. 2) compose de deux demi-cer- 

 cles formes chacun d' une lame de cuivre et d' une d' acier, reunes 

 comme dan's le thcrmomctre dc M. Harrison. 



L' etfect rtpondoit asscz a mcs intentions, j' ai memc observe au 

 moyen de 1' index i, conduit par des semblables lames //(fig.3), que 

 pur le froid c-t le chaud, le raouvement dc ces lames suivoit asstz 

 txactement la marche du thermometre : il en resultoit une comr 

 pen-niion de !a chaleur et du froid, dont ou pouvoit augmenter oa 

 diminuer 1' ett'cct a volonte, en mettant plus ou moins denmsse auK 



exlrcniites des ces dcml-cerclcs. 



that 



