Tho defects inseparable from all the above thermometers are, that the dilatations of the air are not proportional to the increments of heat ; that the length of the column of spirit or mercury varies with the temperature of the atmosphere ; also, that the air which is in con tact with the surface of the spirit in the open vessel, in the first kind of instrument, or with the top of the column of the spirit or mercury in the others, exerts more or loss pressure according to its density ; and thus the indications afforded by the thermometer are rendered erroneous, or require corrections which it is difficult to apply. Tho air-thermometer proposed by Dubuat, and of which the following is a brief description, possesses some advantages above those which have been mentioned ; but not being portable, it has never been employed.
It consists of a column of mercury in a tube, like that of a baro meter, hermetically sealed at the upper end, and bent below so as to form a short branch inclined at about 40' to the straight part of the tube ; this branch terminating with a hollow ball. The mercury occupies the straight part of the tube to the height of about 294 inches above the bend ; and at this bend it terminates without entering into the ball, which, by the construction, is a little above the bend. The part of the tube which is above the column of mercury is free from air, and when the bend is plunged in boiling water tho tube is to be in a slightly inclined position, so that a vertical line may pass through the two extremities of the mercurial column ' • then, upon the ball becoming cool, and the elasticity of the air in it being diminished, the weight of the mercury will cause it to descend in the long branch and rise in the other. The mercury is to be prevented from entering the ball by making the tube decline farther from the vertical position, so that the lower extremity of the mercury may remain in the vertical line before mentioned ; and the temperature of the air is to be deter mined by the height of the top of the column of mercury above a horizontal lino passing through the lower extremity, that is, by the cosine of the declination of the tube from the vertical. Since the air in the ball preserves constantly the same volume, the elasticity com municated to it by the heat of the atmosphere, or by the fluid in which the instrument is plunged, is always in equilibrio with the pressure of the column of mercury, which is the force acting against it, and is proportional to the vertical height of that column.
About the middle of the 17th century the members of the Accadernia del CNIMP1110 caused thermometers to be constructed in which, instead of air, alcohol or spirit of wine was employed. The fluid was intro
duced, as at present, into a glass tube terminating at bottom in a hollow ball, from which tho air had been expelled by heat. The oppo site extremity of the tube was then hermetically sealed, and a scale of equal parts was applied to the stem, by means of small beads of coloured enamel, for the purpose of expressing the temperature of the atmo sphere, or of the liquid which was to be examined. Alcohol dilates and contracts considerably with the variations of temperature to which it may be subject, though not in so great a degree as air. It is also capable of measuring very low temperatures; but as it is brought to a boiling state much sooner than water, it cannot be :employed to ascertain a high degree of heat. Spirit-thermometers were introduced into this country by Mr. Boyle, and they are still used for low tem peratures.
Sir Isaac Newton, being dissatisfied with the smallness of the range of spirit-thermometers, employed linseed-oil in tubes for the purpose of measuring degrees of temperature. This liquid has nearly the same amount of expansibility by increments of heat as alcohol; and it is capable of bearing a considerable amount of heat and of cold without either boiling or freezing ; but from its viscidity it adheres so much to the interior side of the tube as to render accurate observations quite impossible, and on this account it has not since Newton's time been employed for thermometers.
The thermometer which is now in general use is a slender tube of glass terminating in a ball containing mercury, the air having been expelled and tho tube afterwards hermetically sealed. The idea of employing this fluid for the purpose of measuring degrees of heat by its expansion is supposed to have first occurred to Dr. Halley ; and the reason why it was not employed by that philosopher appears to have been that the range of its expansion is much less than that of alcohol. According to llocrhaave (' Elements Chemise; 1732), the honour of having been the first to recommend a mercurial thermometer is to be ascribed to Romer, the discoverer of the motion of light, who is said to have invented it in 1709; but it was not till the year 1724 that such a thermometer was known in this country. In that year an account of a mercurial thermometer which had been invented by Fahrenheit, of Amsterdam, in 1720, was read before the Hoyal Society, and was published in the Philosophical Transactions' (vol. xxxiii.).