The coefficient of expansion of carbon dioxid at a constant pressure of 760 millimeters of mercury must be considered as still somewhat uncertain, though the value given in the table appears to be the best now attainable. The ((natural zero" of the constant-pressure ther mometer lies in about the same general region as the natural zero of the constant-volume in strument. The temperature of melting ice, as referred to the '
In the platinum-resistance thermometer, tem perature is defined as proportional to the elec trical resistance of a coil of pure, annealed platinum wire. The "thermometer" itself con sists of a coil of the wire, wound upon a sheet or strip of mica, and placed in one of the arms of a Wheatstone's bridge, so that its resistance may be accurately determined. It is usual to denote a temperature as defined by the plati num-resistance thermometer by the symbol "pt" ("platinum temperature"). We have, therefore, pr- CR, where R is the observed resistance of the coil at the temperature denoted by pt and C is a constant whose value is to be determined. If Ro and pto and R. and prow are the respective resistances and platinum-resistance tempera .tures at the freezing and boiling points of water, then we have, precisely as in the case of the constant-volume gas thermometer, pt----* 100 R.), as the platinum-resistance temperature, as reckoned from the ((natural zero* of the platinum-resistance therrnotneter. The "reduced platinum temperature,* obtained by subtracting pt. from the temperature pt as here computed, is the one that is commonly used, however; and Callendar and Harker and Chappuis have shown that the reduced plati num-temperature can be expressed in tertn of the reduced gas thermometer scale by means of a simple equation of the form: T T pi= T + A. — --11 100 100 A being a constant whose value is to be deter mined experimentally. Callendar and Griffiths. for the purpose of determining. A, recommend that the resistance of the platinum coil of the thermometer be observed at the temperature of boiling sulphur; the "reduced temperature" of this boiling point being, according to their ex periments with the constant-pressure air ther mometer, 444.53° C (Eumorfopoulos states that the boiling point of sulphur on this scale is between 443.58° and 443.62° C. See 'Proceed ings of the Royal Society,' 1908 A, 81, p. 339.
Compare, also, Callendar and Moss, in the same publication, 1909 A, 83, p. 106). The platinum resistance thermometer gives great promise of being a highly valuable instrument in the future. Indeed, it is so already• but it does not yet appear to be capable of 'deterntining the abso lute values of temperatures closer than to 0.01° C. It may be used as a differential ther mcnneter, however, so as to give results of a far higher order of accuracy. For this pur pose two similar coils or strips of platinum are used, these being placed in two of the arms of a Wheatstone's bndge, so that the smallest de parture from equality in their resistances can be observed. (See RESISTANC.E, ELF.cnucAt). Langley's bolometer is an instrument of this sort. It is used to explore the solar spectrum, and consists of two strips of platinum foil, which are placed across the spectrum to be examined, with their edges toward the source of the light. The two strips are placed in the two arms of a sensitive Wheatstone's bridge, and so long as both the strips are exposed to radiation of the satne intensity, the balance of the bridge is preserved. When one of the strips coincides with a Fraunhofer line, how ever, while the other is still exposed to the full radiative power of the source of light, the balance is destroyed, and the existence of the line is thereby demonstrated, even though the line be in the infra-red, where it is not visible to the eye.
Thermo-electric couples have been used to a considerable extent for the measurement of temperature, and Regnault experimented with them somewhat, but showed that they are dis tinctly inferior in accuracy to the other known methods of determining temperature. At ex ceedingly low temperatures, however, they are often of great value. Wroblewski, for example, made use of thermo-couples Quite extensively for temperature measurements in his researches on the critical points of the gases which are liquefiable only at extremely low temperatures. The platinum resistance thermometer is more generally favored, however, for this purpose; though it cannot be used for temperatures too close to the absolute zero on account of the anomalous and sudden changes of resistance that occur in that region. (See RFSISTANCE, Eutcrium). At these extremely low tempera tures the helium thermometer is still useful, however.
Consult Guillaume, 'Thermometrie de Pre cision,' and Preston, 'Theory of Heat.' See, also, the numerous scientific papers of Kamer lingh-Onnes relating to low-temperature re search.