RADIATION
647
T
minals PP attached thereto, at a sufficient distance from, the ends to avoid the question of heat conduction to the electrodes. These potential wires, which were from 0-003 to 0-025 mm. in diameter, accurately defined the length of the central part of the receiver which was utilised in the measurements. By
-t=0-5mm.
Rear View of Receiver
7mm./
•BO
S R S
End Section through A-A
Tftfl. 6. '
exposing the whole length of tho receiver to radiation, conduction losses from its end do not enter the problem.
The effect of tho presence of the potential terminals was determined (20) (by using a third terminal P') and found negligible, viz. 0-3 per cent. The method of finding tho difference of potential between two terminals attached to the receiver |
seems more certain than to find the difference of potential between two heavy electrodes as used by Ger-lach (31).
The operation of this type of radiometer, with a hemispherical mirror placed in front of tho receiver (20, 24, 38), may introduce errors when the receiver is heated by electrical means. Hence, after conferring with specialists in geometrical optics who were in agreement with tho writer's opinion that a reflecting enclosure is likely to introduce errors, the receiver was used without the hemispherical mirror. Subsequently this mirror was employed in a separate experiment to determine the diffuse reflection (24) from lampblack and platinum black, and finally in determining the loss by reflection from some of tho receivers actually used in the observations. Hence, the loss of radiation incident upon the receivers was probably accounted for as accurately as it would have been by employing a " blackening " device in front of the receiver.
In order to test the question of the accuracy of the corrections used for eliminating the loss by reflection, a series of observations was made on one receiver. In this test the slits
in front of the receiver, and all other conditions remained unchanged. The only variation was in smoking the platinum black receiver with a sperm candle after making the first set of observations. The reflecting power of platinum black is taken to be 1-7 per cent (39) and that of lamp-black 1-2 per cent. The respective determinations, after correcting for energy lost by reflection, were cr=5-822 and a — 5-814. They differ by only about 0-1 per cent, which is very satisfactory, and shows that tho reflection factors were well determined.
Tho assembled apparatus is shown in Fig. C, in which A and B are the water-cooled diaphragms, S the shutter (see Fig. 3), F the radiometer, and D the telescoping diaphragmed tubes enclosing tho optical path from which the moisture was removed by means of phosphorus pontoxido P. Subsequent tests showed that in this outfit the effect of atmospheric absorption was loss than 0-1 per cent (20), which conclusion is substantiated by recent measurements made by Gerlaoh (33).
Tho thermopile was connected with an unusually well shielded iron-clad Thomson galvanometer of special design (38, 41), which served merely as a null instrument to indicate the rise in temperature of the receiver when exposed to radiation and when heated electrically.
The method of observation consisted in exposing tho
R
D
_ 79 — i i--------„ J
I*. I--------12.s- - *|
U— 15-5 and 30am: -
Horizontal Section
PlQ. 0.
receiver to radiation and noting the galvanometer deflection ; then in heating the strip electrically to give, within 1 per cent, the same deflection. Tho measurements of tho electric power put into the atrip wore made with the same potentiometer used in measuring the temperature of tho radiator. No difficulty was experienced in determining temperature equilibrium on heating the receiver by radiation and by electrical means.
The exact amount of energy necessary to cause the same deflection as that produced by absorbing radiant energy is obtained by multiplying the observed energy input by the ratio of the galvanometer deflections. This gives the " constant" of each receiver. In order to reduce all tho measurements to a commone strip, thus eliminating any irregular heating ctrased by inequalities in the thickness of the receiver, which was blackened oleotrolytioally with platinum black. The manganin strip is heated electrically to the same temperature as that attained by the strip when exposed to the radiation from a black body, which was heated by afcoam. From a knowledge of the resistance of the strip and tho electric current tho energy input, and henco tho value of tho radiation constant, is determined.