the liquid in B, expanding, lowers the mercury level at A and raises that at C. There is on each side a scale, but the one at A is numbered downward. The temperature at any time can be read on either scale. Above the mercury, at both A and C, there are placed on the inside of the tube small iron indexes i and i'. Each index is so adjusted that it can slide up and down the inside of the tube, but with enough friction so that its weight will not move it. When the mercury at C rises, it pushes the index i' up the tube; but when the temperature falls, and the mercury at C also falls, the index will be left standing in the tube at the highest point reached by the mercury. Hence this index will give the highest, or maximum, temperature. On the other side the mercury will also push the index up, but only while the temperature is falling. The index on the left side will therefore give the lowest, or minimum, temperature. The alcohol, or whatever liquid is used, wets these indexes and flows past them, exerting inappreciable forces on them. As is well known, mercury does not wet iron, and because of its very large surface tension it can exert considerable force on the indexes. To set the instrument the indexes are jarred down, or pulled down by a small magnet.

187. The clinical thermometer. The clinical thermometer is a maximum thermometer designed to indicate the highest temperature to which it is exposed. Just above the bulb is a narrow constriction through which the mercury flows with difficulty. When the mercury in the bulb is expanding, mercury is forced through; but when the temperature falls, the mercury does not flow back through this constriction, but stands in the tube just as it was at the highest temperature. To set the thermometer the mercury is jarred back through the constriction to the bulb by one or two vigorous shakes.

188. Resistance thermometers. The electrical resistance of a coil of wire changes when the temperature of the wire changes. The apparatus needed to measure the electrical resistance of a wire is portable, and the process is not difficult. The change in electrical resistance thus affords a relatively simple method of measuring changes in temperature. It is especially useful for the measurement of temperature in inaccessible places; for example, a coil of wire has been buried in a silo, with the connecting wires extending to the outside. With suitable apparatus the temperature of the fermenting silage is obtained whenever desired.

Resistance thermometers can be used through a wide range of temperature-from very low temperatures up to about 1000° C., or 1800° F. (bright-red heat). For both high and low temperatures the best thermometer coils are made of platinum wire.

189. Gas thermometers. When air or any other gas is heated, it tends to expand; but if prevented from expanding, its pressure will increase. Either the increase in pressure or the increase in the volume of a gas can be used for indicating changes in temperature. On account of the relatively simple laws governing the behavior of gases, gas thermometers are more accurate than any other type. They are not so easy to use, however; hence it is now a common practice, where a high degree of accuracy is desired, to calibrate some other type of thermometer by the more accurate gas thermometer and then to use the calibrated one.

Through a great range of temperatures, from the lowest that are measured up to nearly 1500° C. (white heat), the gas thermometer is the standard thermometer. For very low temperatures helium gas at reduced pressure is used; for high temperatures nitrogen is used. Current practice favors that type of gas thermometer in which the volume of the gas is kept constant and the temperature changes are measured by changes in pressure of the gas.

190. Thermoelectric thermometers. In Fig. 138, ABC represents an iron wire and ADC a copper one. If the junction A of the two wires is heated, and the other junction, C, is kept cold, an electric current will flow around the circuit in the direction indicated by the arrows. This is an example of a general rule: When

Aso

Iron
B

D
Copper
FIG. 138

C

ever two wires of different materials are connected together at both ends, an electric current will be developed if the two junctions are not at the same temperature. When the necessary electrical measurements (which need not be described here) can be made, this effect can be applied to the measurement of the difference of temperature of the two junctions.

The magnitude of the current and its direction of flow depend on the materials used. For ordinary temperatures copper and an alloy called constantan (trade name, "advance") are perhaps the best. For high temperatures wires that will not oxidize or readily melt are used. A thermoelectric couple composed of platinum and a platinum-rhodium alloy can be used up to nearly 1600° С.

191. Measurement of high temperatures. In many industries certain processes must be carried on at high temperatures, and it is often an important matter to know when the temperature is correct, for a change of 20° C. may alter the product. Hence a great deal of attention has been paid to the measurement of high temperatures. As has already been stated, the resistance thermometer can be used up to a bright-red heat, and the thermoelectric couple still higher. But the most convenient pyrometers (as high-temperature thermometers are called) compare the light or heat radiated from the hot body with that radiated by some other substance. For example, in one type the electric current through an incandescent lamp is changed until the filament of the lamp is of the same brightness as that of the hot body in the furnace, where the temperature is desired. When the electric current through the lamp has been measured, the corresponding temperature can be found from a chart.

192. Standard temperatures. The two fixed temperatures, the melting point of ice and the boiling point of water, are convenient for ordinary ranges of temperatures; but for both low-temperature and high-temperature measurements it is convenient to have a number of known temperatures to use in calibrating, or testing, the instrument one is using. Such low temperatures as the boiling points of liquid hydrogen and oxygen have been carefully determined for this purpose. The melting of many pure metals occurs at very definite temperatures, and these are known. The Bureau of Standards at Washington is prepared to furnish samples of chemically pure metals the melting points of which are as follows:

Tin.

Zinc

.

231.88° С.

419.44° С.

Other temperatures at which natural processes, such as melting or boiling, take place are accurately known. These temperatures which are known to a high degree of precision are called standard temperatures. It is now relatively simple to calibrate any type of thermometer through any desired range by the aid of these standard temperatures.

PROBLEMS

1. Reduce 60° F. to C.; 70° F. to C.; 20° F. to C.

2. Reduce 30° C. to F.; 20° C. to F.; 4° C. to F.

3. Reduce -10° C. to F.; -40° C. to F.; -273° C. to F.

4. Reduce 0° F. to C.; -40° F. to C.; 50° F. to C.

5. Reduce 20° C. to K.; 68° F. to K.; 14° F. to K.

CHAPTER XVII

EXPANSION

Expansion and contraction produced by changes in temperature, 193. Differential expansion, 194. Law of linear expansion, 195. Special cases of expansion, 196. Compensated pendulums, 197. Compensated balance wheels, 198. Volume expansion, 199. Expansion of liquids; relative expansion, 200. Measurement of absolute expansion of a liquid, 201. Expansion of water; maximum density, 202. Forces involved in expansion, 203. Expansion of gases, 204. Expansion at constant pressure, 205. Changes in pressure at constant volume, 206. Summary of gas laws, 207. General gas law, 208. Examples, 209. Limit of application of gas laws, 210. The gas thermometer, 211. Summary, 212.

193. Expansion and contraction produced by changes in temperature. It is not difficult to show that the length of a metal rod increases when the temperature is raised. A simple method can be understood by the aid of Fig. 139.

One end of a metal rod is firmly clamped, and the other is allowed to rest on a small sewing-needle which is lying on a smooth surface, such as glass. When the rod is heated, it will expand, causing the needle to roll. A light index fas

FIG. 139

tened to the needle will turn as the needle rolls, thus making the motion visible. If the diameter of the needle and the angle through which it turns are both measured, it is not difficult to measure the amount of the elongation.

Another simple method is to send an electric current through a stretched iron wire two or three meters long. If the current is large enough to heat the wire quite hot, one can plainly see the sag of the wire.

A rise of temperature produces expansion, but a lowering of temperature causes contraction. This is plainly seen in either one