In the older thermometers, the scale was arbitrary, so that no comparable readings could be taken by means of different instruments. In the finest modern instruments, also, the scale is usually quite arbitrary, being, in fact, engraved on the tube during the process of calibration above described. But then, by careful observation, certain definite temperatures are measured in terms of this arbitrary scale, so that the value of a degree and the position of some definite zero-point are determined for it, and the result engraved on the tube. These numbers enable us, by an easy calculation, to reduce the observed reading of the fine instrument to its equivalent in some of the standard scales.
At present, we assume, what is very nearly true for mercury at least, that equal increments of bulk correspond to equal increments of temperature. All, then, that is necessary is to fix two definite temperatures, and assign their positions on our scale. Water being one of the most common bodies in nature, and being everywhere easily obtainable in a state of great purity, is usually employed; and its freezing and boiling points are taken as the definite points. The temperature of freezing water or of melting ice is almost absolutely fixed, for (see HEAT) pressure alters it only very slightly. It is otherwise with the boiling-point of pure water, for this is considerably raised by increase of pressure; so much so, in fact, that if the barometer be not attended to, an error of several degrees is possible. Hence we must define the particular pressure, usually 30 in., at which the boiling-point is to be determined. The thermometer, constructed (so far) as above described is to have its bulb, and nearly the whole of the portion of the stem which contains liquid, immersed in pounded ice, from which the melted portion is freely trickling; and when the level of the spirit has become stationary, its position, the is marked on the tube. Similarly, the baiometer standing at 30 in., the is inclosed in the steam immediately above the surface of water freely boiling. We thus obtain the It only remains that we decide by what numbers these points shall be indicated, because (on account of the nearly uniform expansion of mer cury) then the remaining divisions can be at once filled in by dividing the interval between them into equal parts, or, if necessary, allowing for a alight taper in the tube. The only scales which require mention are those of Fahrenheit, Reaumur, and Celsius. Of these, the first is commonly used in Britain, the second in Germany, and the third in Trance; but this last, under the name of the centigrade scale, is almost exclusively used by scientific men of all nations. The relations of these scales will be easily understood
by means of the following figure: In the Fahrenheit scale, the freezing-point is 32°, and the boiling-point 212°, so that the space between these is divided into 212 — 32, or 180, equal parts or degrees. In the others, the freezing-point is the zero, but the boiling-point is 80° and 100° respectively. It is of course perfectly easy to reduce from one of these scales to another. Thus—What is the cen tigrade reading for 77° Fahr. (See the dotted line in the figure)? The numbers in Fahren heit's scale are all too great by 32, because 32°, and not 0°, stands for the freezing-point. Subtract this from 77, and we have 45. Hence the required number of centigrade degrees must bear the same ratio to the 100 from freezing to boiling in that scale that the 45 bears to the between the same limits in Fahrenheit's. The requisite number is therefore 45 180 100 = 25° C. In words—To convert Fahrenheit to centigrade, subtract 32, and multi 10 5 ply by —180, or Vice vers6—To pass from centigrade to Fahrenheit, multiply, by 50 32. Thus the Fahrenheit value of 50 C. 5 + 32 = 122, as in the figure. Of course the similar processes with Reaumur's scale present no difficulty.
It is supposed that Fahrenheit fixed his zero at the point of greatest cold that he had observed, possibly in Iceland, more probably by means of a freezing mixture, such as snow and salt, or sal-ammoniac. It is much to be desired that the centigrade scale alone were employed.
A mercurial thermometer ceases to be of use for temperatures only a little above the freezing point of mercury; but it has a wide range upward, as mercury does not boil till about 600° C. On the other hand, a spirit-thermometer, though of little use beyond about 50° or 60° C., as alcohol boils at 70° C., is useful for any degree of cold yet pro duced, as alcohol has never yet been frozen. When extreme sensitiveness is required, ether being considerably more expansible than alcohol, is sometimes employed; as by Thomson in detecting the effect of pressure on the freezing-point of water. Water, again, would be about the very worst substance with which a thermometer could be filled; for not to speak of its expanding in the act of freezing, and therefore necessarily bursting the instrument, if it were ever allowed to reach the freezing-point, its scale would read partly backward and partly forward; for as ice-cold water is gradually heated up to 4° C., it contracts, and begins to expand again after that limit has been passed.