No. 9. AN ELECTRIC PYROMETER. DR C. W. SIEMENS selected as the subject for his Bakerian Lecture, 1871, the effect of temperature on the resistance of metals, and a short notice of his results is given in the Proceedings of the Royal Society, April 1871. His experiments ranged between freezing point and 1000° C., and gave rise to an empirical1 formula, which differed from that suggested by Dr Matthiessen, but satisfied the results of the latter observer obtained solely at lower temperatures. Siemens' formula is R=a √t+ Bt+y, where R is the resistance of the coil at absolute temperature t, and a, B, y are constants, which vary for different metals. An account of the pyrometer as generally in use is given with diagrams in the Brit. Assoc. Reports for 1876. The resistance is measured differentially with a voltameter, and the result is easily obtained by means of a slide rule suitable for rapid and practical determinations of temperature. A number of observations were made with an instrument supplied by Messrs Siemens, Bros., which in practice is exceedingly convenient, but in consequence of its large dimensions and the massiveness of the iron case in which it is enclosed it was not found possible to determine the constants of the instrument, and in consequence the following experiment was arranged with a view of obtaining the values of the constants for platinum. 1 The form in which the formula is cast was suggested by a consideration of the circumstances upon which the variation of resistance is likely to depend. To determine the temperature of a small glass cylinder in a Bunsen flame by means of an electric pyrometer. A fine spiral wire of platinum was enclosed in a cylinder of white German glass about three inches in length. This plan was adopted because the action of the flame upon the resistance of the wire when unprotected is not yet ascertained; some previous determinations made it seem likely that this influence was very considerable. The ends of the platinum wire were passed through a screen, which secured stability, and were then connected with a Wheatstone's Bridge in order to measure the resistance. The coil was first inserted in a Bunsen flame sufficiently large to envelope the whole, and was screened from draughts of air. The glass bent under its own weight when redhot, but the resistance became fairly constant after about a quarter of an hour, and a maximum resistance was measured. The coil was next inserted in melting ice and also in steam at temperature 100°6 C. Two other determinations of the resistances at higher temperatures, 154° and 202°.5, were obtained with the coil in a bath of paraffin; the former was taken so as to furnish a test equation, and was not so accurately observed as that at the higher temperature. It has been observed1 that wires exposed to high temperatures are liable to a variation in resistance after cooling, either of increase or decrease. In order to determine any change of this kind, the platinum spiral was again inserted in steam at temperature 100° C., an interval of three weeks having elapsed since its insertion in the Bunsen flame, and an increase of onehundredth of an ohm was found, or a change of one in a thousand. 2 Temperature Determination.-The same mercury thermometer was used throughout the experiments and finally compared with an air pyrometer, and with Jolly's air thermometer. The most trustworthy result gives the reading of the mercury thermometer as 2°.3 in excess of the air thermometer at 199° C., and the temperatures 154°, 202°-5 must therefore be corrected to 152° and 200°2 C. This result agrees satisfactorily with that 1 Brit. Assoc. Reports, 1876. 2 See p. 11. obtained by calibrating the mercury thermometer and applying Regnault's corrections approximately'. The reading of a mercury thermometer exceeds that of an air thermometer at 200 °C. by an amount varying from one to five degrees, depending on the kind of glass used in manufacture. In this case the calibration error amounted to 5° in excess at 208° C., and the mercury thermometer probably reads 2° or 3° higher than the air thermometer at this temperature. Thus, if both were accurately calibrated, the corrected value of 202°5 is from 199°.5 to 200°5; and the experiment with Jolly's air thermometer gave 200°.3 C. Values of the Constants.-The values of a, B, y as found by Dr Siemens are 0·03937, 0·002164, and — 0·2413, where it is to be observed that y is negative. The values obtained by these experiments are 0023, 001456, and 5905, where a is negative and y is positive. Although these are entirely different from those obtained by Siemens, they satisfy the test equation within the limits of the combined errors of observation and variation of resistance, namely, one-hundredth of an ohm. It was afterwards found that a small hole was fused in the glass cylinder by the Bunsen flame, and that paraffin had entered and surrounded the wire, but apparently without vitiating the results. A possible source of slight error is the fact that redhot glass ceases to be a non-conductor and a correction may be necessary for the resistance in the Bunsen flame. The values of the constants without applying corrections for the mercury thermometer are a='004807, B001257, y='5274, which agree more nearly with those given by Siemens than the values finally obtained. The uncorrected temperature obtained is 986° C., whilst the corrected value is 941° C.3. This result might be compared 1 Mémoires de l'Académie, Tom. xxi. 2 Glass conducts as an electrolyte at a temperature below that of fusion. Encycl. Brit., Art. Electrolysis. Buff. Ann. de Chimie (3) XLII. p. 125, 1854, gives results for the resistance between the interior and exterior of a tube of glass. Its effects in this case would be inappreciable. 3 [This result is very much too high. Siemens' constants give the value about with those obtained by means of Favre and Silbermann's calorimeter1. Although the difference in the results is large, it must be remembered that the experiments with the calorimeter were to determine the temperature of a small piece of metal after being cooled by a fall of a foot through the air, and the loss of heat from a small piece of metal must be very great under such circumstances. Results. Temp. Cent............. 0° 100°.6 152° 200°.3 t° 1.15 1.23 2.28 The values of the constants are determined from the first, third, and fourth columns, and when supplied in the equation formed from the data of the third, give the resistance as 1.16 ohms. Thus the difference is within the limits of the observation errors. The final equation is 2.28-0023 √t+001456t+5905, and the value of the temperature of the glass in the Bunsen flame is 941° C. A. S. EVE. 470° C. for the temperature, which is probably not far from right. As a matter of fact the curve representing the variation of resistance of platinum with variation of temperature is very nearly a straight line, and the results obtained by assuming Siemens' formula and determining the constants by observations at comparatively low temperatures may be very wide of the truth in consequence of small experimental errors. This subject has been lately taken up and worked out in the laboratory by Mr H. L. Callendar, of Trinity College. W. N. S.] 1 See p. 57. No. 10. LAPLACE AND LAVOISIER'S ICE-CALORIMETER USED FOR MEASURING THE HEAT GENERATED IN AN INCANDESCENT ELECTRIC LAMP. BLACK, who discovered and investigated the phenomena of latent heat and specific heat, introduced also the two most important methods for measuring quantities of heat. The effects of heat on bodies, on which these methods are based, In applying the second method he determined the amount of heat lost by a known mass of the body examined in cooling through a certain range of temperature, or, in changing its state, by the amount of ice converted by it into water at the same temperature as the ice. For this purpose he used a clear transparent block of ice in which a hole had been made and which could be covered over with another flat piece of ice. The body to be examined was then introduced into the cavity, which had previously been carefully dried, and left there till it had come into thermal equilibrium with the ice. No communication of heat from outside could take place, as the only effect produced by the heat of the surrounding atmosphere is to melt some of the ice on the outer surface of the block. Laplace and Lavoisier made use of the same method for measuring quantities of heat, but they somewhat modified the working of it. A description of the apparatus employed by them and of the experimental details will be found in Oeuvres de Lavoisier, Vol. II. p. 283, or in Mémoires de l'Académie des Sciences, 1780. This modification of Black's original process does away with the necessity of procuring a perfectly clear block of ice of suitable size, which is not always easy, but on the other hand some sources of error are introduced, from which the first is free. |