As the instrument is entirely founded on this principle, that heat or fire expands all bodies, as cold condenses them, there was a necessity of employing a fluid easy to be dilated. A quantity of it is seated in one part in the bulb. This being expanded by heat, is pushed forward into a fine tube, or capillary cylinder, so small, that the motion of the fluid in it is speedy and perceptible. Some thermometers have been constructed with their reservoir composed of a larger cylinder; but in general, at present, they are made globular. The smaller the bulb is, the sooner it is heated through, and the finer the tube, the greater will be the length of it, and the more distinct the degrees. It is scarcely possible that any glass cylinder, so very small, should be perfectly regular; the quicksilver, during the expansion, passing through some parts of the tube wider than others, the degrees will be shorter in the first case, and longer in the latter. If the divisions, therefore, are made equal between the boiling and freezing points, a thermometer, whose cylinder is irregular, cannot be true. To rectify this inconveniency, the ingenious Mr. Bird, of London, puts into the tube about the length of an inch of mercury; and measuring, with a pair of compasses, the true extent of this body of quicksilver in one place, he moves it from one end to the other, carefully observing where it increases or diminishes in length, thereby ascertaining the parts, and how much the degrees are to be varied. By this contrivance, his thermometers are perfectly accurate, and exceed all that were ever made before. I shall not trouble my reader with numerous calculations that have been made, to express the quantity of particles of the liquor contained in the bulb, in order to determine how much it is dilated. This, Dr. Martine seems to think a more curious than useful enquiry. It is sufficient, for our purpose, to know how the best thermometers ought to be constructed: they who have leisure and inclination, may be agreeably entertained by the author last cited. By observing the rise of the mercury in the thermometer, during any given time, as, for instance, during the time of the day, we ascertain the degree and value of the heat of every part of the day, from whence may be fixed the medium of the whole time, or any part thereof. By repeated experiments, it appears, the medium heat of most days is usually indicated at eight o’clock in the morning, if the instrument is placed in the shade, in a northern situation, and out of the reach of any accidental heat. Though water is not so readily affected as air by heat and cold, yet, as all bodies long exposed in the same place, become of the same degree of heat with the air itself, no great error can arise from estimating water, in general, to be of the same heat as the air, at eight o’clock in the morning, in the shade.