THERMOMETER, ELECTRICAL 21. The convenience of the mercurial thermometer lies in the fact that it is complete in itself, and can be read without subsidiary appliances beyond a magnifying glass. Its weakness lies in the very limited range of each single instrument, and in the troublesome and often uncertain corrections which must be applied to its readings in all work of precision. Electrical thermometers have the disadvantage of requiring auxiliary apparatus, such as galvanometers and resistances, the use of which involves some electrical training. But they far surpass the mercurial thermometer in point of range, delicacy and adaptability, and can be applied to many investigations in which ordinary thermometers are quite useless.
There are two kinds of electrical thermometers, which depend on different effects of heat on the electrical properties of metals: (i) The Thermocouple, or Thermopile, which depends on the production of a thermoelectric force when the junctions of different metals in an electric circuit are at different temperatures; and (2) the Electrical Resistance Thermometer, the action of which depends on the fact that the resistance of a pure metal to the passage of an electric current increases very considerably when the temperature is raised. The theory of the thermocouple is discussed in the article THERMOELECTRICITY, as it possesses many points of interest, and has been studied by many skilful experimentalists. The electrical resistance thermometer is of more recent origin; but although the theory has been less fully developed, the practice of the method bids fair to surpass all others in the variety and accuracy of its applications. In order to secure the widest possible range and the greatest constancy, in either variety of electrical thermometer, advantage is taken of the great stability and infusibility characteristic of the metals of the platinum group. Other metals are occasionally used in work at low temperatures with thermocouples for the sake of obtaining a larger electromotive force, but the substitution is attended with loss of constancy and uncertainty of reduction, unless the range is greatly restricted.
22. Applications of the Thermocouple. The principal uses of the thermocouple in thermometry are for measuring high temperatures, and for measuring small differences of temperature, more particularly when the temperature is required to be measured at a point, or in a very small space. The electromotive force of the couple depends only on the temperature at the plane of junction of the two metals, which can be very exactly located. A typical instance of a measurement to which the thermocouple is peculiarly suited, is the determination of the cyclical variations of temperature at accurately measured depths from one-tenth to one-hundredth of an inch in the metal of the cylinder of a heat engine, the interior surface of which is exposed to cyclical variations of temperature in the working of the engine. 1 The exact depth of the plane of junction can be measured without difficulty to the thousandth of an inch. The insertion of the wire makes the least possible disturbance of the continuity of the metal. There is no lag, as the thermometer itself is part of the metal. The instantaneous value of the temperature at any particular point of the stroke can be measured separately by setting a periodic contact to close the circuit of the galvanometer at the desired point. A further advantage is gained by measuring only the difference of temperature between two junctions of a thermocouple at different depths, instead of the whole interval from some fixed point. None of these advantages could be secured by the use of any ordinary thermometer; some depend on the fact that the method is electrical, but some are peculiar to the thermocouple, and could not be otherwise attained.
On the other hand, the thermocouple is not well suited for thermometry of precision on account of the smallness of the electromotive force, which is of the order of ten microvolts only per degree for the most constant couples. By the use of very delicate galvanometers it is possible to read to the hundredth or even in special cases to the thousandth of a degree on this small difference, but unfortunately it is not possible to eliminate accidental thermal effects in other parts of the circuit due to small differences of temperature and material. These accidental effects seldom amount to less than one or two microvolts even in the best work, and limit the accuracy attainable in temperature measurement to about the tenth of a degree with a single platinum thermocouple. This limit can be surpassed Hall, Trans. Amer. Inst. Elect. Eng. 1891, vol. viii. p. 226; Callendar and Nicolson, Proc. Inst. C. E. vol. cxxxi. p. I.
by using couples of greater thermoelectric power and less permanence, or by using a pile or series of couples, but in either case it is doubtful whether the advantage gained in power is not balanced by loss of simplicity and constancy. A method of avoiding these effects, which the writer has found to be of great use in delicate thermoelectric researches, is to make the whole circuit, including all the terminals and even the slidewire itself, of pure copper. Platinoid, german silver, constantan and other alloys most commonly used for resistances and slidewires, are particularly to be avoided, on account of their great thermoelectric power when connected to copper. Manganin and platinum-silver are the least objectionable, but the improvement effected by substituting copper is very marked. It is clear that this objection to the use of the couple does not apply so strongly to high temperatures, because the electromotive force of the couple itself is greater, and the accuracy attainable is limited by other considerations.
23. The Resistance Thermometer. In practice the resistance thermometer is almost invariably made of platinum, since there is very seldom any advantage to be gained by the substitution of baser metals. The instrument is for this reason often referred to simply as the " platinum thermometer." It is important that the platinum should be pure, both for the sake of uniformity and also because the change of electrical resistance with temperature is greatly diminished by impurities. The observation of the fundamental coefficient, which is -00390 (or rather larger than the coefficient of expansion of a gas) for the purest metal hitherto obtained, is one of the most delicate tests of the purity of the metal. In addition to the constancy and infusibility of the metal, a special advantage which is secured by^the use of platinum is the close agreement of the thermodynamical scale with the platinum scale of temperature, as defined by the formula p/ = ioo(R-R )/(R 1 -R ), . . . (24)
in which the symbol pi stands for the temperature on the platinum scale centigrade, and R, Ri and RO are the observed resistances of the thermometer at the temperatures pt, 100 and o C. respectively. A platinum thermometer is generally arranged to read directly in degrees of temperature on the platinum scale, just as a mercury thermometer is graduated in degrees of the mercury scale. The reduction to the scale of the gas thermometer is most conveniently effected by the difference formula / pt=dt(t ioo)/io,ooo, . . . (25)
in which a is a constant, called the difference-coefficient, the value of which for pure platinum is about 1-50, but varies slightly for different specimens. This formula was first given by the writer as the result of a series of comparisons of different platinum wires with each other and with other metals, and also with an air thermometer over the range o to 625 C. The platinum wire in these comparisons was enclosed inside the bulb of the air thermometer itself, and disposed in such a manner as to be at the mean temperature of the bulb in case the temperature was not quite uniform throughout (Phil. Trans. A. 1887, p. 161). The formula was subsequently verified by C. T. Heycock and F. H. Neville (Jown. Chem. Soc. February 1895;, by the observation of a number of higher points up to the freezing-point of copper at 1082 C., which they showed to agree with the most probable mean of all the best determinations by various methods of gas thermometry. At still higher temperatures, beyond the present range of the gas thermometer, the writer has succeeded in obtaining presumptive evidence of the validity of the same formula by comparison with the scales of the expansion and the specific heat of platinum, which appear to follow similar laws (Phil. Mag. December 1899). If we assume that the coefficient of expansion of platinum, the coefficient of increase of resistance, and the specific heat are all three linear functions of the temperature, we obtain results which are in agreement within the limits of error of observation up to the fusing-point of platinum itself. The same formula has been independently verified by the comparison of platinum thermometers with the constantvolume nitrogen thermometer by Harker and Chappuis (Phil. Trans. A. 1900), working at the International Bureau at Sevres, over the range o to 650 C. It has also been shown to satisfy very closely the observations on the variation of electrical resistance of other metals over wide ranges of temperature. Although the theoretical explanation of the formula has not yet been given, owing to our ignorance of the true nature of electrical conduction and of the molecular constitution of metals, it may be regarded from an empirical point of view as being one of the most accurately established of all thermometric formulae. It will be observed that it also represents the simplest possible type of divergence from the thermodynamical scale.
24. Methods and Apparatus. The methods of electrical thermometry may be roughly classified under two heads as (i) deflection methods, in which the temperature is deduced from the observed deflection of a galvanometer; and (2) balance methods, in which the resistance or the electromotive force is balanced against a known adjustable resistance or potential difference. The former methods are most suitable for rough work and rapid reading, the latter for accurate measurements. In the practice of the deflection method it is customary to use a movable-coil galvanometer, the sensitiveness of which can be varied by varying the resistance in circuit, or by varying the stiffness of the suspension. The accuracy attainable is of the order of one-half of i per cent, on the deflection, and is limited by variations of resistance of the galvanometer, and by the imperfect elasticity of the suspension. In any case the scale of the galvanometer should be calibrated or tested for uniformity. In this kind of work the thermocouple has the advantage over the resistance thermometer in that the latter requires an auxiliary battery to supply the current; but in many cases this is no disadvantage, because it permits a greater latitude of adjustment, and makes it possible to obtain greater power than with the thermocouple.
In cases where it is desired to obtain greater accuracy without abandoning the quickness of reading which is the principal advantage of the deflection method, it is possible to combine the two methods by balancing part of the potential difference by means of a potentiometer and using the galvanometer for the small changes only. In cases where the greatest accuracy is required, a very sensitive galvanometer should be used, and the whole of the potential difference should be balanced as nearly as possible, leaving very little to depend on the deflection of the galvanometer. The degree of sensitiveness and accuracy obtainable depends primarily on the delicacy of the galvanometer, on the power available, and on the steadiness of the conditions of experiment. For thermometry of precision the resistance thermometer possesses three very great advantages over the thermocouple: (i) The power available, owing to the use of a battery, is much greater; (2) it is possible completely to eliminate the errors due to accidental thermal effects by reversing the battery; (3) the Wheatstone bridge method can be employed in place of the potentiometer, so that the constancy of the battery is immaterial, and it is not necessary to use a standard cell. The conditions to be satisfied in the attainment of the greatest possible accuracy in the measurement of temperature by this method differ somewhat from those which obtain in ordinary measurements of resistance, so that a special type of apparatus has been evolved for the purpose, a brief description of which will be given.
25. Compensated Bridge Apparatus. It is necessary that the thermometer should be connected to the measuring apparatus by wires or " leads " of considerable length, generally at least two or three metres, in order to avoid exposing the galvanometer and resistance box, or other delicate parts of the apparatus, to changes of temperature. It is also essential that the leads should not be too thick or heavy, for convenience in handling and to prevent conduction of heat along the stem of the thermometer. The resistance of that part of the leads which is exposed to variations of temperature necessarily changes, and would give rise to serious errors if it were not determined or compensated. The method now generally adopted in accurate work is to compensate the variations of resistance of the leads by an exactly similar pair of dummy leads called the " compensator " and connected as shown diagrammatically in fig. 6. The battery, consisting of a single cell, with a rheostat and reversing key in circuit, is connected to the terminals AB of the two equal resistance coils AG, GB, which form the ratio arms of the balance. These coils must be carefully tested for equality of temperature-coefficient, and placed in close proximity to each other so as to be always at the same temperature. If they are interwound on the same reel, they must be most carefully insulated from each other. In parallel circuit with the ratio coifs are connected the compensator CC' and the balancing resistances C'E, on one side of the bridge-wire EF, and the compensating resistances FP and the pyrometer and leads PRP' on the other side. The galvanometer is connected to the point G between the ratio coils, and to the sliding contact D on the bridge- wire. Since the ratio coils are always equal, equal changes of resistance on either side of D are eliminated, and do not affect the balance. Thus the changes of the pyrometer leads PP' are balanced by the equal changes of the compensator leads CC' on the other side. As a further refinement, which is of some importance in delicate work, the ends of the compensator leads are connected by a short piece of the same wire as the pyrometer coil. For instance, in observing the variations of temperature of the steam in the cylinder of a steam engine at different points of the stroke with a very delicate thermometer made of wire one-thousandth of an inch in diameter (Proc. Inst. C. ., vol. cxxxi. fig. 16, p. 23), the ends of the fine wire attached to the thick leads could not follow the rapid variations of temperature, and it was found necessary to adopt this, device to eliminate the end-effect. It is also useful in other case* to eliminate the effect of conduction along the leads in cooling intends of the fine wire coil. The balancing resistances C'E are made of some alloy such as manganin or platinum-silver, the resistance Ceilfc FIG. 6. Diagram of Compensated Bridge Method.
of which varies very little with change of temperature. Platinumsilver is probably the best material, as it can be perfectly annealed at a red heat without risk of burning, and is then extremely constant. Unless the box can be kept at an absolutely constant and uniform temperature, which is not impossible but often inconvenient, it is necessary to allow for the change of resistance of the balancing coils C'E due to change in the temperature of the box. The temperature of the coils cannot be accurately determined with a mercury thermometer unless they are immersed in oil, but even in that case it is necessary to know the temperature-coefficient of each individual coil. A more convenient and accurate method, which eliminates the correction automatically, is to compensate each individual coil of the balancing coils C'E by a corresponding compensating coil at FP on the other side of the bridge-wire. The compensating coils are made of platinum, also annealed at a red heat, and each is placed in the box in close proximity to the coil it is intended to compensate. Each balancing coil and its compensator are tested together at various temperatures between 10 and 30 C., and are adjusted until their difference remains constant for any small variation of temperature in the neighbourhood of 20 C. This method of compensation was applied by the writer in 1887, but has not been generally adopted on account of the labour involved in adjusting the coils. The absolute values of the resistances are immaterial, but it is necessary to know the relative values with the greatest possible accuracy. For this reason it is preferable to arrange the resistances in the binary scale, each resistance being equal to twice the next smaller resistance, or to the_ sum of all the smaller resistances, the two smallest resistances being made equal. This arrangement permits the greatest accuracy of comparison in the simplest manner with the fewest observations. The bridge-wire EF provides a continuous scale for reading small changes of resistance. Any change of resistance of the pyrometer coil necessitates the movement of the balance point D through an equivalent resistance along the bridgewire. The equivalent resistance of the bridge- wire per unit length of the attached scale is preferably adjusted, by means of a shunt shown in parallel with it in fig. 5, to be an exact submultiple of the smallest resistance coil. It is usual also to adjust the resistances of the thermometers so that their fundamental intervals are convenient multiples of this unit, generally either 100, 200, 500, or 1000, so that the bridge- wire may read directly in degrees of temperature on the platinum scale. It is easy to get a scale of 10 cms. or more to the degree, and it is not difficult with a suitable galvanometer to read to the ten-thousandth part of a degree. The length of the bridge- wire itself need not be more than 20 or 30 cms., as the balancing resistances enable the scale to be indefinitely extended. Thus the instrument possesses the great advantage over the mercury thermometer that the most open scale may be easily secured without unwieldy length, and without restricting the range of each thermometer.
26. Errors and Corrections. It is most instructive to consider the errors and corrections involved in platinum thermometry on the same lines as those on which the corresponding errors of the mercury thermometer have already been treated.
I. The changes of zero of the mercury thermometer arise chiefly from the small expansibility of mercury combined with the imperfect elasticity of the containing tube. In platinum thermometry the containing tube has nothing to do with the reading, and the effect of any possible strain of the fine wire of the coil is minimized by its small dimensions and by the large temperaturecoefficient of the increase of resistance, which is more than twenty times greater than the coefficient of apparent expansion of mercury in glass. It is not surprising, therefore, that the changes of zero of a platinum thermometer should be practically negligible, provided that the wire is not strained and contaminated with impurities. It is probable that with ordinary care the changes of zero due to exposure to any given limits of temperature are in all cases less than the limit of accuracy of observation, due to other causes at the extreme limit of the range considered.
II. The fundamental interval of each thermometer must be determined as usual by observations in ice and steam, and a correction must be applied by the method already described in the case of the mercury thermometer. The difference of the temperature of the steam from 100 C. should be determined on the platinum scale by the approximate formula dpti = -985*, = -0362(11-760) oooo2o(H-76o) 2 . (26)
III. Pressure Correction. The effect of change of pressure on a platinum thermometer of the ordinary tube form is of course nothing, as the wire itself is not exposed to the pressure. Even if the wire is naked and directly exposed to large changes of pressure, the change of reading is almost inappreciable. Similarly there is no source of error analogous to the effects of capillarity, which are so troublesome with delicate mercury thermometers.
IV. Stem 'Exposure. The reading of a platinum thermometer with compensated leads depends only on the temperature of the coil of wire forming the bulb, and not on the temperature of the stem, provided that the immersion is sufficient to avoid errors due to conduction or convection along the stem. It is desirable that the top of the bulb should be immersed to a depth equal to from three to ten times the diameter of the tube, according to the accuracy required.
V. Scale Correction.- The reduction to the thermodynamical scale may be effected, within the limits of probable error of the most accurate measurements at present available, by the very simple difference formula (25) already given, over the whole range from 100 C. to +1100 C. This is in striking contrast with the mercury thermometer, which requires a cubic formula to cover the range o to 200 C. with equal accuracy. The value of the constant d in the formula varies but little, provided that the wire be fairly pure and the thermometers properly constructed. In order to determine its value in any special case, it is best to take an observation at the boiling-point of sulphur (S.B.P.) for temperatures above o C., or at that of oxygen for temperatures below o C. down to 200 C. It appears probable that there is a point of inflection in the curve of resistance-variation of platinum and some other metals in the neighbourhood of -200 C., and that the formula does not apply accurately below this point. It has become the custom to assume the boiling-point of sulphur (S.B.P.) under normal pressure to be 444-53 C., as determined by Callendar and Griffiths, using a constant-pressure air thermometer, and to take the rate of change of temperature with pressure as '082 per mm. from Regnault's observations. According to experiments made at Kew Observatory with platinum thermometers (Chree, Proc. R. S., 1900), the rate of change is somewhat larger than that given by Regnault's formula, namely, about '090 per mm., and it appears desirable to determine this constant with greater accuracy. The difference between the above formulae reaches a tenth of a degree if the barometer differs by 12 mm. from 760 mm. The uncertainty in the absolute boiling-point of sulphur, however, is probably somewhat greater than one-tenth of a degree, on account of the uncertainty of the expansion correction of the gas thermometer (Phil. Mag., December 1899). The thermodynamical correction of the gas thermometer, which amounts to half a degree at this point, is also to some extent uncertain, on account of the extrapolation. Provided, however, that some exact value of the S.B.P. is chosen for reference, for the reduction of observations with platinum thermometers, the results so reduced will be strictly comparable, and can be corrected at any subsequent time when the value of the S.B.P. is more accurately determined. The boiling-point of oxygen may be taken as -182-5 C. with sufficient approximation for a similar purpose.
VI. Calibration Correction. The calibration of the resistance box and the bridge-wire corresponds to the calibration of the stem of the mercury thermometer, but the process is much simpler for several reasons. It is more easy to obtain a uniform wire than a uniform tube. The scale of the wire is much more open, it corresponds to a very small part of the whole scale, and the process of calibration is easier. One box when calibrated will serve for any number of thermometers of different ranges and scales, and covers the whole range of temperature (see CALIBRATION).
27. Electrical Precautions. The platinum thermometer is so far superior to the mercury thermometer in all the points above enumerated that, if there were no other difficulties, no one would ever use a mercury thermometer for work of precision. In using a platinum thermometer, however, some electrical training is essential to obtain the best results. The manipulation and adjustment of a delicate galvanometer present formidable difficulties to the non-electrical observer. Bad contacts, faulty connexions, and defective insulation, are not likely to trouble the practised electrician, but present endless possibilities of error to the tyro. A useful discussion of these and similar details is given in the paper by Chree already referred to. Bad insulation of the pyrometer and connexions can easily be detected, in the compensated instrument already described, by disconnecting one of the C leads from the battery and one of the P leads from the bridge-wire. Under these conditions the galvanometer should not deflect tf the insulation is perfect. Defective insulation is most likely to be due to damp in the thermometer at low temperatures. This source of error is best removed by drying and hermetically sealing the thermometers. Trouble from bad contacts generally arises from the use of plugs for the resistance coils. If plugs are used, they must be specially designed so as not to disturb each other, and must be well fitted and kept very clean. Mercury cups with large copper terminals, well amalgamated, as used with standard resistance coils, are probably the simplest and most satisfactory method of changing connexions. Accidental thermoelectric effects in the circuit are a possible source of error, as with the thermocouple, but they are always very small if the thermometer is properly constructed, and are relatively unimportant owing to the large K.M.F. available. In any case they may be completely eliminated by reversing the battery. The he.ating effect of the current through the thermometer is often negligible, but should be measured and allowed for in accurate work. With a current of 01 ampere the rise of temperature should not exceed ijj or i;f of a degree. With a delicate galvanometer it is possible to read to the tenthousandth of a degree with a current of only 002 ampere, in which case the heating effect is generally less than rs^s of a degree. It can be very easily measured in any case by changing from one cell to two, thus doubling the current in the thermometer, and quadrupling the heating effect. The correction is then applied by subtracting onethird of the difference between the readings with one and two cells from the reading with one cell. The correction is always very small, if a reasonably sensitive galvanometer is used, and is frequently negligible, especially in differential work, which is one of the most fruitful applications of the platinum thermometer.
28. Construction of Thermometers. One of the chief advantages of the platinum thermometer for research work is the endless variety of forms in which it may be made, to suit the particular exigencies of each individual experiment. It is peculiarly suited for observing the average temperature throughout a length or space, which is so often required in physical experiments. For this purpose the wire may be disposed in a straight length, or in a spiral, along the space in question. Again, in observing the temperature of a gas, the naked wire, on account of its small mass and extremely low radiative power, is far superior to any mercury thermometer. The commonest form FIG. 7. of platinum thermometer (fig. 7), and the most platinum suitable for general purposes, contains a coil B from Thermometer. ^ in. to 2 in. long, wound on a cross of thin mica, and enclosed in a tube, about 4 to J in. in diameter, of glass or porcelain according to the temperature for which it is required. The pyrometer leads and the compensator leads are insulated and kept in place by passing through mica disks fitting the tube, which serve also to prevent convection currents up and down the tube. The protecting tube of glass or porcelain is fitted with a wooden head A carrying four insulated terminals, PP, CC, to which the pyrometer and compensator leads are respectively connected, and which serve to connect the instrument to the measuring apparatus. For work of the highest precision these terminals are often omitted, and the leads are directly soldered to a flexible cable in order to avoid possible errors from thermoelectric effects and changes of resistance of the screw terminals. For temperatures above 500 C. the protecting tube must be of porcelain, and the leads of platinum throughout that part of the tube which is exposed to high temperatures. For lower temperatures a tube of hard glass and leads of copper or silver may be employed, but it is better in any case to make the lower part of the leads of platinum in order to diminish the conduction of heat along the stem. For laboratory work a tube 30 or 40 cm. in length usually suffices, but for large furnaces the length of the protecting tube is often 5 to 10 ft. In the latter case it is usual to protect the porcelain tubes with an external steel tube, which may be removed for delicate measurements. _ 29. Special Forms of Thermometer. In the measurement of linear expansion it is a great advantage to employ a thermometer with the bulb or sensitive portion equal in length to the bar or column under test, so as to obtain the mean temperature of the whole length. In measuring the linear expansion of a standard metre or yard, a fine platinum wire enclosed in a glass capillary, or otherwise insulated, is employed, its length being equal to that of the bar. The same method has been applied by Callendar (Phil. Trans. A, 1887) and Bedford (Phil. Mag., 1898) to the expansion of glass and porcelain at high temperatures, employing a fine wire supported along the axis of the tube under test. An equivalent method, applied to the expansion of silica by Callendar, is to enclose a rod ' of the material inside a platinum tube which is heated by an electric current. This is a very rapid and convenient process, since the mean temperature of the rod must be equal to that of the enclosing tube. Any temperature up to the melting-point of platinum is readily obtained, and easily regulated. The temperature may be obtained by observing either the resistance of the platinum tube or its linear expansion. Either method may also be employed in J. Joly's meldometer, which consists of an electrically heated strip for observing the melting-points of minerals or other substances in small fragments. In observing the temperature of a long column of mercury, as in the method of equilibrating columns for determining the absolute expansion of mercury, a platinum thermometer with a bulb equal in length to the column may similarly be employed with advantage. The application is here particularly important because it is practically impossible to ensure perfect uniformity of temperature in a vertical column, 6 ft. or more in length, at high temperatures.
30. Sensitive Thermometers. Where quickness of reading is essential, the mercury thermometer, or the tube form of electric thermometer, is unsuitable. In cases where the thermometer has to be immersed in a conducting liquid or solution, the fine wire forming the bulb may be insulated by enclosing it in a coiled glass capillary. This method! has been employed by Callendar and Barnes and by Jaeger, but the instrument is necessarily fragile, and requires careful handling. For non-conducting liquids or gases the bare wire may be employed with great advantage. This is particularly important in the case of gases owing to the extreme sensitiveness thus obtained and the almost complete immunity from radiation error at moderate temperatures. Thermometers constructed in the form of a flat grid of bare wire mounted on a mica and ebonite frame have been employed by H. Brown (Proc. R. S., 1905, B 76, p. 124) for observing the temperature of leaves and of air currents to which they were exposed. They have also been employed for observing the air-temperature for meteorological purposes in Egypt and Spain with very satisfactory results (Proc. R. S., 1905, A 77, p. 7). The fine wire, owing to its small size and bright metallic surface, very rapidly acquires the temperature of the air, and is very little affected by radiation from surrounding objects, which is one of the chief difficulties in the employment of mercurial thermometers for the observation of the temperature of the air.
For the observation of rapidly varying temperatures, such as those occurring in the cylinder of a gas- or steam-engine, an electrical thermometer with very fine wire, of the order of -ooi in. diameter is practically the only instrument available. The temperature at any particular moment may be obtained by setting a mechanical contact-maker to close the circuit at the desired point. The sensitive part of the thermometer consists simply of a loop of fine wire from half an inch to an inch long, connected by suitable leads to the measuring apparatus as employed by Burstall (Phil. Mag., October 1895) in the gas-engine, and Callendar and Nicolson (Proc. Inst. C. E., 1898) in the steam-engine. The explosion temperatures cannot be satisfactorily measured in a gas-engine in this manner, because the radiation error at high temperatures is excessive unless the wire is very fine, in which case it is very soon melted even with weak mixtures. Callendar and Dalby accordingly devised a mechanical valve (Proc. R. S., A 80, p. 57) for exposing the thermometer only during the admission and compression strokes, and have deduced the actual explosion temperatures from the indicator diagram. B. Hopkinson (Proc. R. S., A 77, p. 387) succeeded in following the course of an explosion in a closed vessel by means of a similar thermometer connected to a galvanometer xxvi. 27 of short period giving a continuous record on a moving photographic film. When the flame reached the wire the temperature rose 1200 C. in about A of a second, which illustrates the order of sensitiveness attainable with a fine wire of this size. O. R. Lummer and E. Pringsheim, in their measurements of the ratio of the specific heats of gases by observing the fall of temperature due to sudden expansion, employed a very thin strip of foil with the object of securing greater sensitiveness. This was a somewhat doubtful expedient, oecause such a strip is extremely fragile and liable to be injured by air currents, ana because the sensitiveness is not as a matter of fact appreciably improved, whereas the radiation error is increased in direct proportion to the surface exposed. One of the principal sources of error in employing a short loop of fine wire for observing rapidly varying temperatures is that the ends of the loop close to the thick leads are affected by conduction of heat to or from the leads, and cannot follow the rapid variations of temperature. This error may be readily avoided by the method, first employed by Callendar and Nicolson, of connecting the compensating leads with a short length of the same fine wire. The end effect is then eliminated by observing the difference of resistance between two loops of different lengths. Thermocouples of very fine wire have also been employed for similar measurements, but they are more difficult to make than the simple loop of one wire, and the sensitiveness attainable is much less, owing to the small E.M.F. of a single thermocouple.
31. Radiation Thermoscopes. For measuring the intensity of radiation, some form of thermometer with a blackened bulb or sensitive area is employed. It is assumed that the rise of temperature of the thermometer is approximately proportional to the intensity of the radiation according to Newton's law of cooling (see HEAT) for small differences of temperature. A mercury maximum thermometer with a small blackened bulb is still very generally employed in meteorological observatories for registering the maximum solar radiation. But the indications are Rable to error and very difficult to interpret, and an instrument of this type is not sufficiently sensitive or quick in action for weak sources of radiation. Sir John Leslie employed an air thermoscope, similar to that of Galileo (HEAT, fig. l), with a blackened bulb. This has the advantage of a small capacity for heat, and is still employed in various forms for demonstration purposes, but is not sufficiently sensitive for accurate work. Electrical thermometers are now generally employed on account of their superior sensitiveness, and also on account of the greater facility of adaptation for the requirements of each particular experiment. The most familiar instrument is M. Mellpni's thermopile, which is built up of a number of small bars of antimony and bismuth, or other alloys of high thermoelectric power, arranged in the form of a cube with alternate junctions on opposite faces. When connected to a galvanometer of suitable resistance, this arrangement gives a high degree of sensitiveness on account of the multiplication of couples, but owing to the large mass of metal involved in its construction it takes a considerable time to acquire a steady state. This defect has been remedied in the radiomicrometer of C. V. Boys (Phil. Trans., 1888, 1 80 A, p. 159) by employing a single junction attached to a small disk of very thin copper. The free ends of the minute bars forming the couple are connected to a loop of thin copper wire suspended by a fine quartz fibre between the poles of a magnet. This arrangement forms a very delicate galvanometer and gives the maximum sensitiveness attainable with a single couple, since all unnecessary connecting wires are avoided. It is incomparably quicker and more dead-beat in action than the ordinary thermopile, but has the disadvantage that it must be set up permanently on a steady support and the radiation brought to it in a horizontal direction. An instrument of similar delicacy is the radiometer, the action of which depends on the repulsive effect of the residual gas in a nearly perfect vacuum on a delicately balanced vane suspended by a fine fibre. An instrument of this type was first constructed by Sir William Crookes (see RADIOMETER); the instrument was applied to radiation measurements, and its sensitiveness greatly improved by E. F. Nichols. It requires a very steady mounting, like the radiomicrometer, but has the additional defect that the radiation must be introduced through a window, which may give rise to selective absorption. Other varieties of thermopile, in which the sensitive parts are constructed, as in Boys' radiomicrometer, so as to have a very small capacity, but are connected like the ordinary pile to a separate galvanometer, have been employed by Lord Rosse for observations of lunar heat and by W. H. Julius and Callendar for the solar corona.
In cases where the radiation can be concentrated on a very small area, such as the receiving disk of the radiomicrometer, the thermoelectric method is probably the most sensitive. But if there is no restriction as to the area of the receiving surface, considerable advantage may be gained in convenience of manipulation, without loss of sensitiveness, by the electric resistance method. An instrument of this type was first employed by S. P. Langley (Proc. Amer. Acad., 1881, 16, p. 342) under the name of the bolometer, by which it has since been known. The sensitive surface is made in the form of a blackened grid of thin metallic foil, generally platinum coated with platinum black, connected in one of the arms of a Wheatstone bridge. The rise of temperature of the grid when exposed to radiation is measured by its increase of resistance in the usual manner. In order to compensate for changes of temperature of the surrounding air the balancing resistance is made of a precisely similar grid, placed in close proximity to the first but screened from radiation. The foil should be as thin as possible consistent with strength, in order to secure the maximum sensitiveness. For spectroscopic work a single strip or linear bolometer is employed. For absolute measurements, where it is necessary to absorb the whole radiation admitted through a given area, two grids are placed with the strips of one behind the interspaces of the other.
32. Absolute Measurement of Radiation. In many cases the object is not to secure the maximum degree of sensitiveness, but an absolute measurement of the intensity of the radiant energy, in calories per square centimetre per minute, or other suitable units. For this purpose some form of radiation thermometer is generally employed, but the method of procedure is modified. The earlier methods as exemplified in C. S. M. Pouillet's pyrheliometer, or L. I. G. Violle's actinometer, consisted in observing the rate of rise of temperature of a small calorimeter, or thermometer of known thermal capacity, when exposed to a given area of the radiation to be measured. To secure greater sensitiveness A. P. P. Crova substituted a copper disk with an attached thermocouple for the calorimetric thermometer. The method is very simple and direct, but has the disadvantage that the correction for external loss of heat is somewhat uncertain and difficult to apply, since the conditions are unsteady and the observation depends on rate of change of temperature. For this reason static methods, depending on the steady temperature finally attained, in which the rate of loss of heat is directly determined by an electric compensation method, have come more prominently into favour in recent years. In K. J. Angstrom's pyrheliometer (ActU Soc. Upsala, 1893) two similar blackened strips of equal area and resistance are fixed side by side in a suitable case in such a manner that either may be exposed to the radiation to be measured while the other is simultaneously heated by an electric current. Attached to the backs of the strips, but insulated from them by thin paper, are the two junctions of a thermocouple which indicates when the temperatures are equal. When this condition is secured the intensity of the radiation is equal to the rate of generation of heat per unit area by the electric current, which is deduced, from a knowledge of the resistance and area of the strip, by observing the current required to balance the radiation. The instrument is very quick and sensitive in action, and the method avoids any assumption with regard to the rate of loss of heat, except that it is the same for the two similar strips at the same temperature. The accuracy of the method is limited chiefly by the measurement of the resistance and width of the strips, and by the difficulty of securing exact similarity and permanence in the attachment of the junctions of the thermocouple. Small differences in this respect may be eliminated by interchanging the strips, but there remain outstanding differences between different instruments of the same make which often exceed 5 per cent.
An electric method proposed by F. Kurlbaum (Wied. Ann., 1898, 65, p. 748) consists in observing the rise of temperature produced by radiation in a bolometer grid, then cutting off the radiation and observing the increase of current required to produce the same rise of temperature. There is no difficulty in this case in measuring the area exposed or the resistance of the bolometer, and no uncertainty can arise as to the temperature of the strip, because the heated strip itself serves as its own thermometer. The current is easily deduced from a knowledge of the resistances and the E.M.F. of the battery. The chief source of uncertainty mentioned by Kurlbaum lies in possible differences between the effects of radiation and current-heating near the ends of the strips, the area so affected representing a large proportion of the whole area. In Angstrom's method this is not so important because the temperature indicated by the couple is that near the middle of the strip. In the case of the bolometer this end-effect may be compensated, as explained by Callendar (Proc. R. S., 1907, 77 A, p. 7), in the same manner as for sensitive thermometers, by employing two similar bolometers with strips of different lengths.^ An important defect of all the methods so far considered is that the measurement depends on the coefficient of absorption of the black with which the receiving surface is coated. _ The error is probably small, of the order of I or 2 per cent., but is difficult to determine accurately, and varies to some extent with the quality of the radiation. The absorptive power is generally less for rays of great wave-length than for visible rays. If we assume that the loss of heat by conduction and convection is independent of the nature of the surface the defect in question may be avoided by the following method. Two bolometer strips, one bright and the other black, but otherwise exactly similar, are simultaneously exposed to the radiation to be measured, and are traversed by the same electric current. The black strip will be more heated by the radiation than the bright, but the rise of temperature of the bright strip due to the current will be greater than that of the black strip because its emissive power is lower. If the current is adjusted until the temperatures of the two strips are equal the losses by convection ana conduction will be equal, and also the rate of generation of heat by the current in each strip. The rise of temperature must therefore be such that each strip loses as much heat by radiation to the surrounding case as it gains from the incident radiation to be measured. Assuming Kirchhoff's law, the ratio of the emissive to the absorptive power is the same for all bodies at the same temperature, and is equal to the emissive power of a perfectly black body. The rise of temperature of each strip, when balance is attained, will be the same as that of a perfectly black strip under the same conditions of exposure. The electric current in this method serves to eliminate losses by convection and conduction, and the result is obtained in terms of the observed rise of temperature and the radiation constant for a black body. The method works well for a source at 100 C. ; but, for a high temperature source, a correction is required because the absorptive powers of the strips may differ appreciably from their emissive powers.
Another electric compensation method of special interest is the method of the " Peltier cross." A small disk of copper is supported by two thermoelectric couples forming a cross. One of the couples serves to measure the rise of temperature, while the other is traversed by an electric current, which may be employed to compensate the radiation by the heat absorption due to the Peltier effect. The advantage of this method is that the Peltier effect is easily determined from an observation of the thermoelectric power (see THERMOELECTRICITY) in absolute measure, and that it is Eroportional to the first power of the current. Loss or gain of eat by conduction from the supporting wires, and changes of temperature in the surrounding case, are readily compensated by mounting two similar disks side by side. Small differences between the disks are eliminated by exposing them to radiation alternately, with reversal of the current, so that the irradiated disk is cooled or the other disk heated by the Peltier effect. The current is adjusted in each case so that the temperatures of the disks are equal, as indicated by the second couple connecting the disks. The method is about equal in sensitiveness to that of Angstrom, but it is easier to secure conditions of exact similarity and to measure the quantities involved in the absolute determination, namely, the area of the hole through which the radiation is admitted, and the coefficient of the Peltier effect. The uncertainty due to imperfect blackness of the disks may be eliminated by using cups in place of disks; and the sensitiveness and range may be increased by using thermopiles in place of single couples.
33. Optical or Radiation Pyrometers. Since the intensity of radiation increases very rapidly with the temperature of the source of radiation, instruments for measuring radiation may be applied for measuring temperature, assuming that the laws connecting radiation and temperature are known. The advantage of this method is that the measurement may be made from a distance without exposing any part of the measuring apparatus to the destructive action of high temperatures. Apart from the difficulty of calibrating the measuring apparatus to give temperature in terms of radiation, the chief source of uncertainty in the application of the method is the emissive power of the source of radiation. The methods principally employed may be divided into two classes: (i) Radiation methods, depending on the measurement of the radiant energy by means of a radiometer, thermocouple or bolometer; (2) optical or photometric methods, depending on the colour or luminous intensity of the radiation as compared with a suitable standard.
Of the radiation methods the simplest in theory and practice depends on observing the total intensity of radiation, which varies as the fourth power of the absolute temperature according to the Stefan-Boltzmann law (see HEAT) for a perfectly black body or full radiator. In applying this method it is very necessary to allow for the emissive power of the source, in case this does not radiate as a black body. Thus the emissive power of polished platinum at 1000 Abs. is only 10 per cent., and that of black iron oxide about 40 per cent, of that of a black body; and the percentage varies differently for different bodies with change of temperature, and also for the same body according to the part of the spectrum used , for the measurement. Owing to the rapid increase of radiation with temperature the error due to departure from black body radiation is not so serious as might be imagined at first sight. If the temperature of a polished platinum strip -at 1500 C. were estimated by the radiation formula, assuming the constant for a perfectly black body, the error for red light would be about 125. for green about 100, and for blue about 75. Such errors may be corrected when the emissive power of the source at various temperatures is known from previous experiments, but it is preferable to observe, whenever possible, the radiation from the interior of a uniformly heated enclosure which approximates very closely to that of a black body (see HEAT).
Radiation pyrometers of this type are generally^ calibrated by the method of sighting on the interior of an electric furnace containing a thermocouple or gas-thermometer by which the temperature is measured. The gas-thermometer has been employed for verifying the law of radiation up to 1500 C., but the difficulties of obtaining accurate results with the gas-thermometer increase so rapidly above 1200 C. that it is questionable whether any advan tage is gained by using it beyond this point. The law of radiation has been so closely verified by observations at lower temperatures that the uncertainty involved in applying it at higher temperatures in the case of a black body is probably less than the uncertainty of the gas-thermometer measurements, and much less than the uncertainty of extrapolating an empirical formula for a thermo couple. Thus L. F. C. Holborn and W. Wien (Wied. Ann., 1805 (6), by extrapolating their thermoelectric formula, found the value I5 8 7 C. for the melting-point of palladium, whereas Violle founc 1500 C. by the calorimetric method, and Callendar and Eumorfopoulos (Phil. Mag., 1899, 48) found 1540 and 1550 C. by the methods of the expansion and the change of resistance of platinum respectively. By a later thermoelectric extrapolation Holborn anc Henning (Berlin Akad., 1905, 12, p. 311) found 1535 C. for the melting-point of palladium, and 1710 C. for that of platinum, values which were strikingly confirmed by J. A. Harker at the National Physical Laboratory, and by Waidner and Burgess at the Bureau of Standards, U.S.A. Holborn and Valentiner employin an optical method (Ann. Phys., 1907, 22, p. i) found 1582 C. anu 1789 C. for palladium and platinum respectively. There can be little doubt that the extrapolation of the parabolic formula for the thermocouple at these temperatures is quite untrustworthy (see THERMOELECTRICITY) and that the values given by the electrical resistance method, or by the laws of radiation, are more likely to be correct. Assuming that the total radiation varies as the fourth power of the absolute temperature, a radiation pyrometer can be calibrated by a single observation at a known temperature, such as the melting-point of gold, 1062 C. if a black body is employed as the source; and its indications will probably be accurate at higher temperatures under a similar restriction. If the pyrometer is sighted on the interior of a furnace through a small observation FIG. 8. Fery's Mirror Pyrometer (Camb. Scient. Inst. Co.). For temperatures from 500 C. to 1 100 C hole it will indicate the temperature of the furnace correctly, provided that the temperature is uniform. But it must be remembered that this condition does. not generally exist in large furnaces. Suppose, for instance, that it is required to find the temperature of the molten metal on the hearth of a furnace viewed through a thick layer of furnace gases, which are probably at a much higher temperature. It is evident that the radiation from the intervening flame may be much greater than that from the metal, and may introduce serious errors. The same objection applies with greater force to optical pyrometers, as the luminous radiation from gases may be of a highly selective character. If, on the other hand, it is required to observe the temperature of metal in a ladle before casting, the surface of the metal must be cleared of scum, and it is necessary to know the emissive power of the metal or oxide exposed.
For scientific measurements of temperature by the radiation method, the thermopile, or bolometer, or radiomicrometer. previously calibrated by exposure to a black body at a known temperature, is directly exposed at a known distance to a known area of the source of radiation. The required result may then be deduced in terms of the area and the distance. The use of extraneous optical appliances is avoided as far as possible on account of selective absorption. .For practical purposes, in order to avoid troublesome calculations and measurements, an optical arrangement is employed, either lens or mirror, in order to form an image of the source on the receiving surface. Fig. 8 illustrates Fery's mirror pyrometer, in which a mirror M, focused by the pinion P, forms an image of the source on a disk, supported by wires of constantan and copper forming a thermocouple, connected by the brass strips D and R to the terminals b, b'. The observation hole in the wall of the furnace is sighted through the eyepiece O, and is made to overlap the disk slightly. The rise of temperature of the junction js assumed to be proportional to the intensity of radiation, and is indicated by the deflexion of a delicate galvanometer connected to the terminals b, b'. A lens may be substituted for the mirror at high temperatures, but it is necessary to allow for the selective absorption of the lens, and to a less extent for that of the mirror, by a special calibration of the scale.
Assuming Wien's laws for the distribution of energy in the spectrum (see HEAT), the temperature of a black body may also be measured by observing (i) the wave-length corresponding to maximum intensity in the normal spectrum, which varies inversely as the absolute temperature, or (2) the maximum intensity itself, which varies as the fifth power of the absolute temperature, or (3) . the intensity of radiation corresponding to some particular radiation or colour, which varies as an exponential function, the exact form of which is somewhat uncertain. Methods (i) and (2) require elaborate apparatus and arc impracticable except for purposes of scientific research. The exact application of method (3) is almost equally difficult, and is less certain in its results, but for optical purposes this method may be realized with a fair degree of approximation by the use of coloured glasses, and forms the basis in theory of the most trustworthy optical pyrometers.
34. Optical or Photometric Pyrometers. The change of colour of a heated body from red to white with rise of temperature, and the great increase of intrinsic brilliancy which accompanies the change, are among the most familiar methods of estimating high temperatures. For many processes eye estimation suffices, But a much greater degree of accuracy may be secured by the employment of suitable photometers. In Mesure and Nouel's pyrometric telescope, the estimation of temperature depends on observing the rotation of a quartz polarimeter required to reduce the colour of the radiation to a standard tint. It has the advantage of requiring no auxiliary ^apparatus, but, owing to the lack of a standard of comparison, its indications are not very precise. In the majority of photometric pyrometers, a standard of comparison for the intensity of the light, either an amyl-acetate or gasoline lamp, or an electric glow-lamp, is employed. The optical pyrometer of H. L. Le Chatelier (Comptes Rendus, 1892, 114, p. 214) was one of the earliest, and has served as a model for subsequent inventors. The standard of comparison is an amyl-acetate lamp, the flame of which is adjusted in the usual manner and viewed in the same field as the image of the source. The two halves of the field are adjusted to equality of brightness by means of a cat's eye diaphragm and absorption glasses, and are viewed through a red glass, giving nearly monochromatic radiation in order to avoid the difficulty of comparing lights of different colours. Assuming Wien's law, the logarithm of the intensity of monochromatic radiation for a black body is a linear function of the reciprocal of the absolute temperature, and the instrument can be graduated by observing two temperatures; but it is generally graduated at several points by comparison with temperatures observed by means of a thermocouple.
The Wanner Pyrometer (Phys. Zeits., 1902, p. 112) is a modification of Konig s spectrophotometer, in which the two halves of the field, corresponding to the source and the standard of comparison, are illuminated with monochromatic red light polarized in planes at right angles to each other. The two halves may be equalized by rotating the analyzer, the circle of which is graduated to read in degrees of temperature. The instrument has a somewhat restricted range of maximum sensitiveness, and cannot be used below 900 C. owing to the great loss of light in the complicated optical system. It cannot be sighted directly on the object since no image is formed as in the Le Chatelier or Fery instruments, but the methods of securing monochromatic light by a direct vision spectroscope, and of adjusting the fields to equality by rotating the analyser, are capable of great precision, and lead to simple theoretical formulae for the ratio of the intensities in terms of Wien's law.
The Fery Absorption Pyrometer (Journ. Phys., 1904, p. 32) differs From Le Chatelier s only in minor details, such as the replacement of the cat's eye diaphragm by a pair of absorbing glass wedges. The principles of its action and the method of calibration are the same. The pyrometers of Morse, and of L. F. C. Holborn and F. Kurlbaum depend on the employment of a glow lamp filament as standard of comparison, the current through which is adjusted :o make the intrinsic brilliancy of the filament equal to that of :he source. When this adjustment is made the filament becomes invisible against the image of the source as background, and the temperature of the source may be determined from an observation if the current required. Each lamp requires a separate calibration, Dut the lamps remain fairly constant provided that they are not overheated. To avoid this, the source is screened by absorption jlasses (which also require calibration) in observing high tempera:ures. Except at low temperatures the comparison is effected by placing a red glass before the eyepiece. At low temperatures a ipecial advantage of the glow-lamp as a standard of comparison s that it matches the source in colour as well as in brightness, so that the instrument is very sensitive. At high temperatures the red glass serves chiefly to mitigate the glare.
35. Registering and Recording Thermometers. The term registerng thermometer is usually applied to an instrument with an index which requires setting, and when set will indicate the maximum or minimum temperature occurring, or will register the temperature at particular time or place. A recording instrument is one constructed to give a continuous record of the temperature, and requires a revolving drum or some equivalent clockwork mechanism for recording the time. The most familiar types of registering thermometers are modifications of the common liquid-in-glass thermometer.
John Rutherford's maximum, invented before 1790, was an ordinary mercurial thermometer placed horizontally; the column pushed before it a small steed index which was left at the highest point reached and was drawn down again to the liquid by a magnet when the instrument had to be reset. It is little used now. Negretti and Zambra's maximum has a constriction in the tube near the bulb, past which the mercury easily expands but cannot return when the temperature falls, since the column breaks at the narrowed point when the fluid in the bulb begins to contract. The instrument is set for a fresh observation by shaking the detached portion of the column back down the tube. The clinical thermometers used by physicians are instruments of this type, and are made with a very open scale to read only in the neighbourhood of the normal temperature of the human body. In the Phillips or Walferdin maximum a portion of the mercury is separated from the rest by a minute bubble of air. It is placed horizontally and as the temperature rises the detached portion of the column is pushed forward but is not withdrawn when the main column retreats towards the bulb in cooling. It is set for a new observation by bringing it into a vertical position and tapping it slightly. By reducing the length of the index and the bore of the stem this thermometer may be made suitable for use in any position without altering its register.
The minimum thermometer in most common use is that of Rutherford, invented in 1790. It is a spirit thermometer, preferably filled with amyl alcohol, to reduce risk of distillation, in the column of which a small porcelain index is included. The instrument is hung horizontally, and, as the temperature falls, the index is drawn back through surface tension by the end of the column. When the temperature rises the liquid flows past the index, which is left at the lowest point attained. To prepare the instrument for a fresh observation it is inverted, when the index falls back against the end of the column. James Six's combined maximum and minimum thermometer (Phil. Trans., 1782) consists of a U-tube, the bend of which is filled with mercury. One leg contains spirit above the mercury and terminates in a bulb also full of spirit. The other leg also contains a column of spirit above the mercury, but terminates in a bulb containing air and vapour of spirit mixed. With increase of temperature the spirit in the full bulb expands; the mercury in consequence is pushed round the bend and rises to a greater or less extent in the other leg, carrying before it a steel index which thus marks the maximum temperature. With cold the spirit in the full bulb contracts, and: the mercury moves back carrying with it a second index which marks the minimum temperature. The instrument is set by drawing down the two indices upon the two ends of the mercury column by means of a magnet.
With a mercury thermometer a continuous record of temperature can only be obtained by the aid of photography, a method which has been in use for many years at some first-class observatories, but which cannot be generally employed on account of the expense and the elaborate nature of the apparatus required. The commonest type of recording thermometer works on the principle of the Bourdon pressure-gauge. The bulb consists of a curved metallic tube filled with liquid, the expansion of which with rise of temperature tends to straighten the tube. The movements are recorded on a revolving drum by a pen carried at the end of a light lever attached to the bulb. This form of instrument is widely employed for rough work, but it has a very limited range and is unsuitable for accurate work on account of want of sensitiveness and of great liability to change of zero, owing to imperfect elasticity of the metal tube. For accurate work, especially at high temperatures, electrical thermometers possess many advantages, and are often the only instruments available. They are comparatively free from change of zero over long periods, and the thermometer or pyrometer itself may be placed in a furnace or elsewhere at a considerable distance from the recording apparatus. The principal types are the thermocouple and the platinum resistance thermometer already described, which may be employed for recording purposes, without altering the thermometer itself, by connexion to a suitable recording mechanism. The methods in use for recording the indications of electrical thermometers may be classified as in 24 under the two headings of (i) deflexion methods and (2) balance methods. Deflexion methods, in which the deflexion of the galvanometer is recorded, are more suitable for rough work, and balance methods for accurate measurements. The most delicate and most generally applicable method of recording the deflexions of a mirror galvanometer is by photographing the movements of the spot of light on a moving film. Almost any required scale or degree of sensitiveness may be obtained in this manner, but the record cannot be inspected at any time without removal and development. Since the forces actuating the needle of the galvanometer are very small, it is out of the question to attach a pen or marking point directly to the end of the pointer for recording a continuous trace on a revolving drum, because the errors due to friction with the recording sheet would be excessive. This difficulty has been avoided in many electrical instruments by depressing the pointer so as to mark the paper only at regular intervals of a minute or so, leaving it completely free for the greater part of the time. The record thus obtained is discontinuous, but is sufficient for many purposes. For accurate measurement, or for obtaining an open scale over a particular range of temperature, it is necessary to employ some form of balance method as already explained in 24.
36. Electric Recorder, Balance Method. The application of the electric balance, potentiometer or Wheatstone-bridge for recording changes of resistance or electromotive force has been effected by employing a galvanometer of the movable coil type as a relay. The deflexion of the galvanometer to right or left, according as the resistance or E.M.F. increases or diminishes, is made to actuate one or other of a pair of motors for moving the contact point' on the bridge wire and the recording pen on the drum in the corresponding direction. A continuous record free from friction error is thus obtained, since the galvanometer does not actuate the pen directly. With an electrical resistance thermometer it is possible in this way to obtain continuous pen-and-ink records on a scale of an inch or more to the degree, reading to -01 C. and practically free from zero error over any desired range from 200 to +1500 C. With a thermocouple, employing the potentiometer method, the same apparatus can be used with advantage, but it is not possible to obtain so open a scale on account of the smallness of the thermoelectromotive force available.
The attainment of sufficient delicacy in the relay mechanism turns on the employment of a rotating or vibrating contact in combination with a moving coil galvanometer of the siphonrecorder type. This was first successfully effected by Callendar (Trans. R. S. Canad., 1897) for records of radiation and temperature, and has since been applied to submarine telegraphy by S. G. Brown and by A. Muirhead. The mechanism of Callendar's electrical recorder, as arranged for temperature measurements, is described and illustrated in Engineering, May 26, 1899, and in a treatise on Pyrometry by Le Chatelier and Boudouard. Electrical recording instruments of both types are now coming into extensive use for industrial purposes in the measurement of furnace temperatures, etc., for which they are particularly suitable, because the recording apparatus can be placed at any distance from the furnaces which may be considered most convenient, and can be connected to any one of a set of furnaces in succession whenever it is desired to obtain a record.
AUTHORITIES. There is no special work on the subject of thermometry in English, but most of the principles and methods are described in text-books on heat, of which Preston's Theory of Heat may be specially mentioned. For recent advances in thermometry the reader should consult the original papers, the most important of which have been cited. The greater part of the recent work on the subject will be found in the publications of the Bureau International des Poids et Mesures de Sevres (Paris), of the Reichsanstalt (Berlin), of the Bureau of Standards, U.S.A. (Washington), and of the National Physical Laboratory (London). (H. L. C.)
Note - this article incorporates content from Encyclopaedia Britannica, Eleventh Edition, (1910-1911)