TIIERMOMETEK. 



r It, or. iJ" it bf I'.ard, tliat it mould be kept boiling at lead 

 lA, minutes befurc it is ufed ; yet the committee recom- 

 mend, for the adjuftment of thcrmoneters mtended for nice 

 experiments, to employ rain or diftilled water, and to per- 

 form the operation in (leai.i. , , r 



It is obierved, that though the boiling point be placed lo 

 much higher on fomc of the thcrmomelcrs now made than 

 on other?, vet this does not produce any confiderable error 

 in the obfe'rvations of the weather, at leaft in this climate ; 

 L>T an error of ih" in the pofition of the boiling point, will 

 make au error oii'ly of iialf a degree in the pofition of 92", 

 and of not more tiian a quarter of a degree in the point 



of 62°. • ,. . f 



It is only in nice experiments, or in trying the heat ot 

 hot liquors, that this error in the boiling point can be of 

 much fignification. 



In adjufting the freezing, as well as the boibng point, the 

 quickfilver in the tube ought to be kept of the lame heat as 

 that in the ball. When the freezing point is placed at a 

 confiderable diftance from the ball, the pounded ice (hould 

 be piled to fuch a height above the ball, that th» error 

 which can arife from the quickfilver in the remaining pait of 

 the tube, not being heated equally with that in the ball, 

 fhall be very fmall, or the obferved point muft be conefted 

 on that account, according to the following table. 



The correftion in this table is expreffed in 1 000th parts 

 of the diftance between the freezing point and the furface 

 of the ice : e. g. if the freezing point ftands feven inches 

 above the furface of the ice, and the heat of the room is 6z^, 

 the point of 32° (hould be placed 7 x .00261, or .018 of 

 an inch lower than the obferved point. A diagonal fcale 

 will facilitate this correftion. 



The committee obferve, that in trying the heat of liquors, 

 care (hould be taken that the quickfilver in the tube of the 

 thermometer be heated to the fame degree as that in the 

 ball ; or if this cannot be done conveniently, the obferved 

 heat (hould be correfted on that account ; for the manner of 

 doing which, and a table calculated for this purpofe, wc 

 muft refer to their excellent report in the Phil. Tranf. 

 vol. Ixvii. part ii. art. 37. 



Several experiments made by governor Hutchins, at 

 Albany Fort, in Hudfon's Bay, in 1782, in purfuancc of a 

 method fuggefted by Dr. Black and Mr. Cavendifli, and 

 for wliich he obtained fir Godfrey Copley's medal from the 

 Royal Society in 1783, have not only confirmed the ob- 

 fervations before made, relative to the folid ftate into which 

 quickfilver can be brought by cold, its metalline fplendour 

 and polilh when fmooth, its roughnefs and cryftallization 

 where the furface was unconfined, its malleabihty, foftnefs, 

 and dull found when ftruck ; but have alfo clearly demon- 

 (Irated, that its point of congelation is no lower than — 40°, 

 or rather - 39°, of Fahrenheit's fcale ; that it will bear, 

 however, to be cooled a few degrees below that point, to 

 which it jumps up again on beginning to congeal ; and that 

 its rapid defcent in a thermometer, through many hundred 

 of degrees, when it has once paffed the above-mentioned 



limit, proceeds merely from ite great contraction in the aft 

 of freezing. See Phil. Tranl. vol. Ixxiii. pari ii. art. 

 •20, 20, 21. 



Tur.RMOMETERS, Obfcrvat'wns en the Conjliiidion of. It 

 is abfolutely iieceflary that thofe who would derive any ad- 

 vantage from thefe inftruments, (hould agree in ufing the 

 fame liquor, and in determining, according to the fame 

 method, the two fundamental points. If they agree in thefc 

 fixed points, it is of no great importance whether they di- 

 vide the interval between them into a greater or lefs number 

 of equal parts. The fcale of Fahrenheit, in which the fun- 

 damental interval between 2 1 2'^, the point of boiling water, 

 and 32°, that of melting ice, is divided into 180 parts, 

 (houid be retained in the northern countries, where Fahren- 

 heit's thei-mometer is ul'ed : and the fcale, in which the fun- 

 damental interval is divided into 80 parts, will ferve for 

 thofe countries where the thermometer of M. de Reaumur 

 is adopted. But no inconvenience is to be apprehended 

 from varying the fcale for particular ufes, provided care be 

 taken to fignify into what number of parts the fundamental 

 interval is divided, and the point where o is placed. 



With regard to the choice of tubes, it is moft defirable to 

 have them exaftly cylindric through their whole length. See 

 Mercurial Thermometer. 



The capillary tubei are preferable to others, becaufe they 

 require lefs bulbs, and they are alfo lefs brittle, and more 

 fenfible. Thofe of the mod convenient fize for common 

 experiments are fuch as have their internal diameter about 

 the fourth of a line : and thofe made of thin glafs are better 

 than others, as the rife and fall of the mercury may be 

 more diftinftly perceived. The length of nine inches will 

 ferve for all common cccafions ; but for particular purpofes, 

 the length both of the tubes and of the divifions fliould bs 

 adapted to the ufes for which they are defigned. 



In determining the beft fize of the balls or bulbs, it has 

 been ufual to compare new tubes with fuch thermometer? 

 as are well proportioned. But M. Durand has propofcd a 

 formula for finding the proportion which the balls ought to 

 bear to their refpedtive tubes. With this view he expreffes 

 the length of the tube, meafured in diameters of itfelf, by a ; 

 the whole capacity of the ball and tube by c ; the capacity 

 of the fundamental interval, exprelfed in the iame parts with 

 the whole capacity, ty d ; the number of degrees of the 

 fundamental interval by m ; the number of other degrees 

 which the fcale is to contain, befides thofe of the fundamental 

 interval both above and below it, by n ; and the diameter of 

 the ball meafured in diameters of the tube by b : and b = 



{/' 



a X 



d X in + n 



For two cylinders having 



equal bafes being as their heights, m : n :: d : — , which is 



m 



the capacity of that part of the tube which exceeds the 

 fundamental interval, to which adding d, that inter- 



val, we have the total capacity of the tube 

 d n + d 



dn 



+ d, or 



Subtrafting this from c, we (hall have the 



m 



. , d n -Ii- d m ^ m — dm ^ d n 



capacity 01 the ball :::= i- — =■ . 



mm 



If this quantity be divided by the capacity of the tube, 

 the quotient will ihew how often the capacity of the 

 ball contains that of the tube ; and this quotient is = 



.• in 



