THERMOMETER. 



fucfc coniparifon. It is by this we Ihould know the heat or 

 cold of another feafon, of another year, another ch'mate, 

 &c. and what is the greateft degree of heat or cold in 

 which men and other animals can fubfift. 



M. de Reaumur contrived a new thermometer for the pur- 

 pofe ; wherein the incorivnionces above recited are pro- 

 pofed to be remedied. He took a large ball and tube, 

 and knowing the content of the ball as well as that of the 

 tube in every part, he graduated the tube, fo that the fpace 

 from one diviiion to another might contain a thoufandth 

 part of the liquor, which liquor would contain one thoufand 

 parts when it Hood at the freezing point : then putting the 

 ball of his thermometer, and part of the tube, into boiling 

 water, he obferved whether it rofe eighty divifions ; and if 

 it exceeded thefe, he changed his liquor, and by adding 

 water lowered it, fo that on the next trial from the freezin.g 

 point to the point of boiling water, it Ihould only rife 

 eighty divifions : but if the liquor, being too low, fell fhort 

 of eighty divifions, he raifed it by adding reftified fplrit to 

 it. The liquor thus prepared fuited his purpofe, and ferved 

 for making a thermometer of any fize, whxjfe fcale would 

 agree with his ftandard. Such liquor, or fpirits, being 

 about the ftrength of common brandy, may eafily be had 

 any where, or made of a proper degree of denfity by raifing 

 or lowering it. 



The abbe Nollet made many excellent thermometers upon 

 M. de Reaumur's principle. Dr. Marline, however, ex- 

 prefles his apprehenfions that thermometers of this kind are 

 conftrufted on principles, that will by no means be found fo 

 accurate as were to be widied and expefted. The balls, 

 or bulbs, being large, as three or four inches in diameter, 

 are neither heated nor cooled foon enough to fliew the 

 variations in the heat of bodies, and in the weather. Small 

 bulbs and fmall tubes, he fays, are much more convenient, 

 and may be conftrufled with fufBcient accuracy. Though 

 it muft be allowed, that Reaumur, by his excellent fcale, 

 and by depriving the Ipirit of wine of its air, and expelling 

 the air by means of heat from the ball and tube of his 

 thermometer, has brought it to as great a degree of perfec- 

 tion as it may poifibly admit ; yet it is liable to fome of the 

 inconveniences of J'pirit thermometers, and much inferior to 

 the mercurial thermometers. Thermometers of this kind, 

 and thofe of mercury, do not agree in indicating the fame 

 degrees of extreme cold ; for when the mercury has ilood 

 at 22° below o, the fpirit indicated only \%P \ and when the 

 mercury ftood at 28° or 37° below o, the fpirit refted at 25° 

 or 29°. See the defcription of Reaumur's thermometer at 

 large in Mem. de I'Acad. R. des Scienc. an. 1730, p. 645. 

 Hift. p. 15. Ibid. an. 1731, p. 354. Hift. p. 7. 



In 1740, M. Micheh du Crefl conftrufted a fpirit thermo- 

 meter, to which he annexes four fcales befides its own, w'z. 

 that of the old thermometer in the Obfervatory at Paris, 

 Reaumur's, de I'Ifle's, and Fahrenheit's. See Fixed Points 

 of Thermometers. 



Thermometer, Mercurial. It is a circumftance of prin- 

 cipal importance in the conftrufliofl of tliermometers, to 

 procure a fluid that meafures equal variations of heat by 

 correfponding equal variations in its own bulk or volume : 

 and the fluid which poffefTcs this eflential requifite in the 

 moft perfedl degree is mercury : the variations in its bulk 

 approaching nearer to a proportion with the correfpond- 

 ing variations of its heat than any other fluid. This general 

 propofition M. de Luc has very elaborately evinced, by 

 fhewing that the condenfations of fluids, which increafe in 

 bulk when they freeze, are not proportional to the dimi- 

 nutions of heat ; and that the dilatations of fluids, which are 



eafily converted into vapour by heat, are not proportional 

 to the augmentations of heat : whereas the bulk of mercury 

 is not enlarged when it freezes, and it refifts evaporation 

 more than every other liquid that has been ufed in the con- 

 ftruclion of thermometers. Befides, it is of all liquids the 

 moft eafily purged of its air. It is alfo the moft proper for 

 meafuring very confiderable variations of heat ; for, if a fcale 

 be graduated with o at the point of melting ice, and 80 at 

 that of boiling water, mercury well purged of its air will 

 indicate feven times this difference of iieat, or 561 degrees 

 in fuch a fcale ; as it will condenfe without freezing to — 261 

 of this fcale, and expand witliout boiling to 300 of the 

 fame fcale. Mercury is alfo more fenfible than any other 

 fluid, air excepted, and conforms more readily to the feveral 

 variations of heat. Moreover, as mercury is an homogene- 

 ous fluid, it will in every thermometer exhibit the fame dila- 

 tation or condenfation by the fame variations of heat. The 

 expanfion of mercury is fcarcely lefs regular than that of 

 folids, which probably approaches the neareil to the fteps of 

 the natural fcale, though not without fome inequality ; and 

 therefore a portion of mercury inclofed in a bulb of glafs, 

 having a fine tube connefted with it, forms a thermometer 

 the moft convenient and probably the moft accurate of any, 

 for common ufe ; the degrees correfponding very nearly 

 with thofe of the natural fcale, although, according to the 

 moft accurate experiments, they appeal- to indicate, towards 

 the middle of the common fcale of Fahrenheit, a temperature 

 two or three degrees too low. There is an inequality ot 

 the fame kind, but ftill greater, in the degrees of the fpirit 

 thermometer ; and this inftrument has alfo the diladvantage 

 of being liable to burft in a heat below that of boiling water; 

 neverthelefs, it is well calculated for the meafurement of 

 very low temperatures, fince pure alcohol has never yet been 

 frozen, while mercury has been reduced to a iolid by the 

 cold of Siberia and of Hudfon's Bay ; but both mercury 

 and linfeed oil fupport a heat of between 500° and 600", 

 without ebullition. 



In order to render thermometers uniform and comparable, 

 it is defirable that mercui-y, fo excellently adapted for this 

 purpofe, (liould be the only fluid ufed in the conftrudlion of 

 them, more efpecially as a thennometer with mercury may 

 be more eafily conftrufted than any others. De Luc's 

 Recherches, &c. vol. i. part ii. cap. 2. paflim. 



Dr. Halley, though apprifed only of fome of the remark- 

 able properties of mercia'y above recited, feems to have 

 been the firft who fuggefted the apphcation of this fluid 

 to the conftruftion of thermometers. Phil. Tranf. Abr. 

 vol. li. p. 34. 



Boerhaave (Chem. i. p. 720.) fays, thefe mercurial ther- 

 mometers were firft contrived by Olaus Roemer ; but the 

 claims of Fahrenheit of Amfterdam, who gave an account 

 of his invention to the Royal Society in 1 724 (Phil. Tranf. 

 N° 381, or Abr. vol. vii. p. 49.), have been generally 

 allowed. And though Prins and others, in England, 

 HoOand, France, and other countries, have made this inftru- 

 ment as well as Fahrenheit, moft of the mercurial thermo- 

 meters are graduated according to his fcale, and are called 

 Fahrenheit's thermometers. Thefe are made of different 

 lengths, and with fome variation in the form of the bulb, 

 according to the purpofes for which they are defigned. 

 Intteadofthe ball, ufed in the fpirit thermometer, a cone or 

 cylinder is annexed to the tube, which may be eafily enlarged 

 or diminiflied, and made of fuch a magnitude, that its 

 capacity may have a certain and known proportion to that 

 of the tube ; and by this means feveral thermometers may 

 be eonftrufted to the fame fcale : befides, the heat more 



eafily 



