INVESTIGATIONS ON LIGHT AND HEAT, MADE AND PUBLISHED WHOLLY OR IN PART WITH XXVII. CONTRIBUTIONS FROM THE PHYSICAL LABORATORY OF HARVARD UNIVERSITY. A STANDARD OF LIGHT. BY JOHN Trowbridge. Communicated May 26, 1885. THE discussions in the Paris Conference of 1881-84 upon the subject of a standard of light, which resulted in the adoption of the light emitted by a surface of platinum at the point of solidification, seemed to assort ill with the discussions which led to a reaffirmation of the value of the C. G. S. system of absolute physical units, and a recognition of the relations between work and heat, and electrical energy. The solidification point of platinum may be a fixed point in nature; but it has not been shown how this fixed point can be connected with that great web of physical measurements which has been woven by Weber, Helmholtz, Thomson, Maxwell, and other physicists. It is true that during the discussions of the Conference reference was made to a proposition of Schwendler, that the light emitted by a strip of platinum rendered incandescent by a known electrical current should be taken as a standard. This proposition, however, received little support; and the Conference finally adopted the light emitted by solidifying platinum as a standard. It seems highly desirable that any standard of light which may be adopted should be connected with the present system of absolute measurements. The suggestion of Schwendler, therefore, seems to merit more attention than it has received. The suggestion of employing the light from a strip of platinum rendered incandescent by an electrical current is really due to Dr. John W. Draper, of New York, who in 1847 enunciated it as follows: "A surface of platinum of standard dimensions raised to a standard temperature by a voltaic current will always emit a constant light. A strip of that metal one inch long andth of an inch wide, connected with a lever by which its expansion might be measured, would yield at 2,000° a light suitable for most purposes." * It has been urged against this standard that different specimens of platinum will emit different amounts of light with the same difference of potential; and that it would be difficult to carry out a measurement of the light and the strength of the current all at the same instant. With a view to obtaining a knowledge of the practical difficulties in this measurement, I interposed a fine platinum wire between the poles of a battery, and endeavored to measure the light emitted, together with the difference of potential at the extremities of the wire and the amount of current which passed through a tangent galvanometer. The difficulties, however, in using a fine platinum wire with a moderate battery power were great. The wire would fuse before the measurements could be satisfactorily made. I then employed a strip of platinum foil 5 mm. wide, about 5 cm. long, and about .02 mm. in thickness. This was placed in a shunt circuit of a small gram machine in order that if the strip should fuse the Dynamo machine might not race. With the proper speed and a suitable adjustment of resistances, the light from this platinum strip could be maintained very constant. The strip was placed in a long Ritchie photometer box, which was provided with two mirrors inclined according to the plan of Ritchie. One half of the photometer disk was illuminated by the incandescent strip, and the other half by a sperm candle. The electrical current was measured by a tangent galvanometer of which the reduction factor was .44 C. G. S. system. The difference of potential at the ends of the strip was measured by a Thomson quadrant electrometer, the deflections of which were compared with that of a Daniell cell, the electromotive force of which was approximately 1.09. A Thomson voltmeter was also used. The indications of this instrument agreed with those of the electrometer. The following table gives the deflections of the instrument. One Daniell cell gave a deflection with the electrometer of 1.3 centimeters. The resistance of the platinum strip when cold was .2 of an ohm. It will be seen from the above results that the current varied approximately from 8 to 6 webers, with an electromotive force of from 3.8 to 2.6 volts, while the resistance varied from .47 to .44 of an ohm, the resistance when cold being .2 of an ohm. The range of the indications of the electrical instruments was comparatively small, while the light varied enormously. It is evident that the chief difficulty of this method is in measuring a strong current with accuracy: for an increase in the current represented by a fraction of a degree of the tangent galvanometer will result in a very large increase in the light from the incandescent strip. I next endeavored to ascertain if a thermal junction enclosed in an Edison incandescent lamp, at the centre of the carbon loop, would be sensitive to changes in the heat radiation of the lamp. It is evident that, if this were the case, the carbon loop might be raised to the same point of incandescence in successive times, assuming that the thermal junction at this point of incandescence receives the same amount of radiant energy. Mr. Edison kindly provided me with a lamp in which one thermal junction of an alloy of iridium platinum and platinum was inserted at the centre of carbon loops. The other junction was placed in ice and water. The thermo-electric force of this combination, however, was extremely feeble. The difficulty of inserting wires of other metals into glass prevented me from carrying this idea further. Instead of the thermal junction, a small loop of extremely fine platinum wire was placed at the centre of a carbon loop in an Edison lamp. This fine wire constituted a bolometer strip and made one branch of a Wheatstone's bridge, it being my intention to place a similar strip in another branch of the bridge, thus making a bolometer. The lamp was placed in a photometer box, and its light was compared with that of a candle as it was raised from a red glow to a light of fifteencandle power. At the same time the resistance of the fine platinum wire was measured by a Wheatstone's bridge. The following table gives the results. This method seems to be quite sensitive. The change in resistance is large when estimated by the number of ohms necessary to restore a balance to the bridge. It was noticed that at a certain point a comparatively small increase in heat radiations was accompanied by a large change in the amount of light emitted. This phenomenon had been noticed early by Dr. J. W. Draper. One Leclanché cell with five ohms in the circuit beside the resistance of the strip was sufficient to raise the latter to a red heat, and precautions were then necessary to prevent a change of resistance from the heating effect of the battery employed with the Wheatstone's bridge. Being desirous of ascertaining whether the resistance of the platinum wire changed after it had been heated to a red heat and had been allowed to cool, I arranged the resistance of the battery circuit outside the bridge, so that the wire could be raised to a red heat, and then, having quickly weakened the battery circuit, remeasured the resistance of the strip. No difference could be perceived in the resistance of the strip. This illustrated the fact discovered by Professor Langley, that thin strips of metal arranged as bolometer strips give up heat very quickly. The results of this experiment led me to think that a bolometer strip of definite surface could be placed at a fixed distance from a carbon loop of definite dimensions inside an exhausted glass vessel. The amount of radiation which the bolometer strip receives could be calculated; and we might base our standard of light upon the point of incandescence which would give a definite radiation at a fixed distance. We could not distinguish by this method the energy produced by rays of different refrangibility. It seems desirable, however, to substitute for the uncertain estimation of colored lights by the eye an instrument which will measure the energy produced by the radiating source at a certain distance. Within certain limits I found that the bolometer strip would indicate an increase or decrease of the amount of radiant energy received while the difference in color of the incandescent lamp made the observer at the photometer entirely uncertain of his measure ments. Owing to the difficulty of obtaining the proper apparatus for the prosecution of the study of this method, I then studied the question of the practicability of employing a thermopile to measure the amount of radiation from an incandescent strip of platinum at a fixed distance. Within a long photometer box was placed a thin brass vessel containing water. Steam was passed by means of a rubber hose into the water of this vessel which was thus maintained at a constant temperature of about 94° C. The outside of the vessel was about 92° C. VOL. XX. (N. S. XII.) 32 This was ascertained by making the side of the vessel constitute one metal of a thermal junction. Between this vessel and the platinum strip, which was made incandescent by a current of from 8 to 9 webers, was placed a thermopile. The face of the thermopile was thus exposed to the radiation from a given amount of heated surface at a constant temperature, while the other was exposed to the radiation of a given surface of platinum. The faces of the thermopile were provided with the customary cones, and a series of diaphragms of thick card-board extended between the radiating surface of the vessel containing the heated water and the platinum strip. The thermopile was connected with a short coil galvanometer, and was moved until the galvanometer needle came to zero. This arrangement was extremely sensitive,a movement of a centimeter in the position of the faces of the pile being sufficient to drive the spot of light from the galvanometer mirror off the scale, corresponding to a movement of nearly fifty centimeter scale divisions. There is no difficulty in effecting a balance as quickly as an ordinary photometric measurement is made. While one observer compares a candle or other source of light with the light from an incandescent strip of platinum, another could make the measurements with the thermopile, and could obtain the amount of energy radiated by the incandescent strip in terms of the constant source of heat. It is necessary to reverse the faces of the thermopile, or to place a second constant source of heat on the same side upon which the incandescent strip is placed. The following table indicates the character of the results. The reduction factor of the galvanometer was .44 in the C. G. S. system. When the photometric indications were the same, the thermopile indicated a large change in the amount of heat received. Thus the heat indications within the range in which the experiments were taken were far more sensitive than the photometric indications. |