294 THERMOMETERS. [135. ingly instruments termed thermometers (from Oɛpuòs, hot, uiтpov, a measure) have been constructed for this purpose. FIG. 116. Air, from the delicacy of its indications, and the regularity of its expansion, would seem to be the material best fitted to measure changes of temperature, and indeed it was the substance first tried. The air thermometer consisted simply of a bulb of glass, with a narrow elongated stem dipping into some liquid (fig. 116); as the bulb became heated the air expanded, and depressed the liquid in the stem as it cooled the air contracted and the liquid rose; a scale attached to the stem gave the amount of the expansion or contraction. But the size of the instrument and the extreme delicacy of its indications limited the range through which it could be employed, and impaired its utility. It was also soon observed that differences of atmospheric pressure, entirely independent of temperature, caused an alteration in the volume of the air: this may be seen by introducing such an instrument under the receiver of an air-pump, and moving the handle; very slight motion causes a great alteration in the height of the column of liquid in the thermometer.* This diffi ciable differences are observed. The following table contains a summary of the results of these experiments: For ordinary purposes, sufficient accuracy is, however, attained by assuming the expansion of gases and vapours by heat, between o° C. and 100°, at of the volume at o° C.; this is about 94 for each degree of Fahrenheit, or for each degree Centigrade, upon the volume occupied by the gas at 32° F. or o° C. *The expansion of air, however, when measured with due precaution, gives a more accurate means of estimating temperature than any other mode hitherto devised, but the application of air to this purpose requires elaborate and special contrivances. 136.] GRADUATION OF THERMOMETERS. FIG. 117. 295 culty was obviated by a modification of the instrument in which atmospheric pressure was altogether excluded; two bulbs were blown at the extremities of a stout tube, which was bent twice at right angles, and a liquid was included in the stem (fig. 117). The instrument in this form, however, would not indicate general changes of temperature, but only differences between the temperature of the two bulbs: if one were warmer than the other, the air expanded in that bulb and drove the liquid in the stem towards the other bulb: it was hence termed the differential thermometer or thermoscope, and formed, in a limited number of cases, an instrument of great sensitiveness and utility. (136) Graduation of the Thermometer.The next great improvement in the ther mometer was made by the Florentine academicians: they substituted the expansion of a liquid for that of air, employing spirit of wine for this purpose, and they divided the stem of the instrument, arbitrarily as before, by means of small dots of enamel, placed at equal distances upon the tube. As the scales of different instruments were not divided upon any uniform principle, the results which they furnished did not admit of direct comparison. This evil was, however, remedied by Newton, who applied Hook's observation, that the melting point of ice always occurs at a fixed temperature; and finding that the boiling point of water under certain standard circumstances is equally constant, he proposed these as fixed points, between which the scale should be divided into a certain number of equal parts; the scale being continued above and below, with similar divisions, as far as might be necessary. Unfortunately, this interval has in different countries been differently subdivided. In England, Fahrenheit's division into 180° has hitherto been principally employed; the zero, or o°, upon this scale being 32 of these degrees below the freezing point of water. In France, and generally on the Continent, the Centigrade division introduced by Celsius prevails; the interval between the freezing and the boiling points being subdivided into 100°, the degrees being counted upwards and downwards from the freezing point, which is reckoned o°, the lower temperatures being indicated by the prefix of the negative algebraic sigu -. THERMOMETRIC SCALES. [136. 80°, Réaumur, whose division is still used in Germany and Russia, divided the same interval into making his zero at the freezing point of water. The conversion of degrees upon one scale into those of another is easily effected by the use of the following formulæ : Fahrenheit to Centigrade, 9 (F°—32)=C°. C°+32=F°. R° +32=Fo. 100 Fahrenheit to Réaumur, Centigrade to Réaumur, Réaumur to Centigrade, (F-32)=R°. Co=R°. R°=C°. The annexed cut (fig. 118) exhibits the three scales placed side by side, and shows the corre50sponding values through a considerable range of temperature.* 5040 .300 The employment of spirit of wine as the expansible liquid for measuring temperature, is attended with advantages where low temperatures are concerned, inasmuch as spirit of wine has never yet been solidified by cold. But owing to the low temperature at which the spirit boils, it cannot be applied for high ranges of temperature; the heat of boiling water would be sufficient to burst the thermometer, in consequence of the generation of the vapour of alcohol within the instrument. For most purposes mercury is a more convenient thermometric to liquid; it embraces a considerable range of temperature, freezing at 38°.8 C., or 37°9 below o° F., and not boiling under 350° C., or 662° F. It has also the advantage of not adhering to the sides of the tubes. (137) Tests of a Good Thermometer.-A good mercurial thermometer should answer to the following tests when its bulb and stem to the top of the mercurial column are immersed in melting ice, the top of the mercury should indicate exactly o C.; when suspended with its bulb and scale immersed in the steam of water boiling in a metallic vessel, as represented in * A table of the degrees of the Centigrade scale, with their comparative values on that of Fahrenheit, will be found in the Appendix to this volume. 138.] MINIMUM AND MAXIMUM THERMOMETERS. 297 fig. 119 (the barometer standing at 760mm.), the mercury should remain stationary at 100° C. When the instrument is inverted, The the mercury should fill the tube, and The expansion of which a thermometer thus takes cognizance is not the entire expansion of the mercury, but the difference between the expansion of the mercury and of the glass; both FIG. 119. expand, but the mercury expands the more of the two, and the .column of metal therefore rises in the stem of the instrument. If a thermometer be graduated immediately after it has been sealed, it is liable to undergo a slight alteration in the fixed points of the scale, owing to the gradual contraction of the bulb, which does not attain its permanent dimensions until after a lapse of several months. From this circumstance, the freezing point may be elevated from to a degree C.; and thus the graduations throughout the scale would indicate a temperature higher than the true one by the amount of the error. In some thermometers, even after long use, the bulb, as Despretz and Pierre have shown, after exposure to a heat not exceeding that of boiling water, does not, on cooling, contract at once to its proper dimensions; and thus a temporary displacement of the graduation and depression of the freezing point below the mark, owing to a temporary dilatation of the bulb, is caused every time such instruments are heated to 100° C. (138) Different Forms of Thermometer.-The variety of cirunder which thermometers are used, necessarily demands a considerable variety in their form. It is desirable, cumstances * In accurate observations, when the column of mercury in the stem of the instrument is exposed to a temperature different from that of the bulb, a correction must be made on this account. If the bulb of the thermometer be placed in the steam of boiling water while the stem projects into the atmosphere, and is of the temperature say of 15° C., an error exceeding 1°.5 C. might easily be made. Wherever it is practicable the whole column of mercury should be raised to the same temperature as the bulb. 298 MINIMUM AND MAXIMUM THERMOMETERS. [138. for delicate experiments, to reduce the mass of the instrument, in order to diminish the amount of heat required to raise its temperature to that of the bodies with which it is brought into contact; but where minute subdivisions of a degree require notice, it is better to employ a thermometer with a large bulb and a fine bore. A useful form of the instrument is the selfregistering maximum and minimum thermometer.—The common maximum thermometer, or that of Rutherford, consists of a mercurial thermometer, with a horizontal stem, in the bore of which a small piece of steel wire is included above the mercury. As the mercury expands it pushes the steel before it, and when the mercury contracts, and recedes towards the bulb, the wire does not follow it. The minimum temperature is observed by means of a spirit thermometer, arranged like the mercurial one, but the index consists of a small piece of enamel, sunk below the surface of the liquid column. As the spirit descends, it carries the index with it by capillarity, as soon as the upper extremity of the index reaches the surface of the liquid; but the liquid, on expanding, readily passes by the enamel, and leaves it at the lowest point to which the column had retreated. Six's thermometer is constructed upon a somewhat similar principle, but it is less portable, and more liable to derangement. A simpler form of maximum thermometer has been constructed by Negretti and Zambra; it is merely an ordinary thermometer, placed horizontally, with a contraction in the tube, just above the bulb, so that it allows the mercury to pass when it expands, but, owing to the narrowing, the metal does not recede when the temperature falls. It therefore indicates the highest temperature attained since the last observation. In the maximum thermometer of Phillips the index is merely a small portion of the mercurial column itself, which is detached from the rest, and separated from it by a minute bubble of air. In both the forms last mentioned the true place of the mercurial column is restored by placing the instrument vertically, and giving it a slight swinging motion. In deep-sea soundings in which the temperature was determined by self-registering thermometers, it was found that the pressure acting on the exterior of the bulb produced an elevation of the mercurial column, and recorded temperatures several degrees too high. This has been prevented by blowing another bulb round the thermometer bulb, and nearly filling this with alcohol, the air being removed from the space at the top of the bulb by boiling the spirit before sealing. Any pressure on the |