The only other air thermometer we intend to no tice is that of Amontons, who made the indicial fluid a column of mercury twenty-eight French inches long, so that the included air was subjected to the pressure of two atmospheres; by this me thod he was enabled to measure high temperatures, such as that of boiling water, without a scale of such great length as the dilatation of air under the common pressure would have required. It was, however, subject to the same great defect as the Sanctorian thermometer, and was besides very un wieldy and liable to accident. In the seventeeth century a modification of the air thermometer was proposed by Van Helmont and by Sturmius, which has recently been revived under the names of Ther moscope and Differential thermometer; as this, however, does not belong to the simple form of the instrument, we shall give some account of it in another part of this article.
The defects of the air thermometer having been duly appreciated by the Florentine .qcadcmia del Cunento, that enterprising body published, in the first volume of their transactions,' a description of a new thermometer, in which spirit of wine was used as the expanding substance, which, as it might be hermetically sealed up in a glass tube or bulb, was free from any defect arising from press ure, as well as the possibility of any loss of fluid by evaporation. This instrument was constructed much in the same way as at present, the spirit be ing dilated till it filled the whole tube, when it was quickly sealed, and on cooling, the fluid retired, leaving nearly a vacuum above it. The great defect of the Florentine weather glass, as it was commonly called, was the want of any fixed scale of gradua tion, on which account no instrument except those graduated by the original one of the academy, could be comparable with any other, the only direc tion being that the cold of ice and snow should make it stand at 20 deg. and the greatest summer heats at Florence, at 80 degrees.
The spirit thermometer was faulty in several other respects, yet it cannot but be thought fortu nate that this fluid, which is esteemed the second best for filling thermometers, should have been so early thought of. The Florentine academicians sometimes bent the tube of their instrument into a tornous form, but this does not appear to be the first attempt of the kind, for we observe at fol. 220 of Sanctorius's Commentaries on Avicenna, a curi ous figure of a similar contrivance.
From the invention of the spirit thermometer may be dated the epoch of the application of truly scientific effects to the art of thermometry. Boyle having received one of the Florentine instruments in England, set himself to effect some radical im provements upon a contrivance which it was ob vious was capable of shedding new light and preci sion over every branch of natural inquiry. He
therefore proposed, as a fixed point for graduation, the thawing of oil of aniseeds, which he preferred to the freezing of water, (another point he also mentioned,) as being more easily obtained in all seasons of the year, and because he had great doubts as to the constancy of the point of freezing of different kinds of water. Indeed this doubt served for long after to delay the adoption of the fixed points at present employed. Derham, Ilal ley, INIusschenbroek, and other philosophers, till far on in the last century, believed that water froze at different temperatures in different latitudes, which Martine,l- only ninety years since, cites some experiments of his own to disprove. Besides this, Boyle falls into another mistake in not taking two fixed points, but endeavouring to compute the ab solute expansion of the spirits, and graduating the scale to ten-thousandths, or some other fixed num / her of parts of expansion, a proposal which, though extremely philosophical in the abstract, is too difficult of execution to be successfully adopt ed in practice. A similar plan was proposed by Hooke4 The acute Halley afterwards turned his attention to this subject of growing importance. He first proposed the temperature of deep pits for a stand ard, naturally enough suggested by the constancy of the thermometer observed by De La Rive and others, in the caverns of the Observatory at Paris. But the difficulty of finding such situations, inde pendently of the variation of their temperature in different latitudes, was enough to set aside this pro posal. Halley afterwards devised the excellent standard derived from boiling liquids, though he did not appreciate its real merit. He tried the point of vaporization of water, mercury, and spirit of wine, and unfortunately preferred the latter; a point very uncertain, as depending greatly upon the strength of the fluid. The boiling point of water, though Halley very much overlooked it, he found to be very steady in its indication.' Nor is it easi ly affected by minute accidental admixtures. The principal source of error is that the vaporization of fluids is affected by the state of atmosphere at the time, as to pressure, the temperature being highest when the pressure is greatest, and the reverse. By subsequent examination, this source of error has become merely the object of an equation table which will be given in the course of this article.