TRANSACTIONS OF SECTION A. 331 
thin covering of fine-grained lampblack, a copper globe, from which the radiation 
takes place, is suspended. The copper globe is 5 centimétres in diameter, while 
the diameter of the envelope is 10 centimétres. In the experiments carried out at 
low temperatures up to the present the surface of the hanging globe has been either 
coated with a very fine covering of lampblack, or else silvered electrolytically 
and polished to the highest degree attainable. The polish is very much better 
than even the mirrored surface obtained by silversmiths, and, as formerly shown, 
the degree of polish makes a very great ditference to the result. 
The enclosing copper envelope is connected to a pair of five fall Sprengel pumps, 
and the following results have been obtained at pressures} not greater than 0-1 m. 
or 0:2 m. (one-tenth or two-tenths of a millionth of an atmosphere). 
When the vacuum is complete the enclosure is cooled down by the application 
around it of a bath of liquid air, contained in a Dewar double-walled vessel. 
Under these circumstances the suspended globe, which at the beginning is at the 
general temperature of the laboratory, cools by radiation. The cooling process is 
carried on for some hours; and during the earlier part of this time the difference 
of temperatures of the cooling globe and the enclosure is read off by means of a 
pair of thermo-electric junctions, one of which is at the centre of the globe * and 
the other immersed in the liquid-air bath. The readings are taken by means of a 
suitable moving coil galvanometer, at equal intervals of time, given by a chrono- 
meter which beats half-seconds.° 
When the temperature of the globe has been reduced nearly to the temperature 
of the surroundings—that is, nearly to the temperature of the liquid-air bath, a 
process which, with the high vacuum mentioned. occupies, as has been said, many 
hours—the conditions are reversed. The liquid air surrounding the outer envelope 
is replaced by hot water, or hot oil, and the hanging globe begins to receive heat, 
by radiation, from the surrounding walls. The reading of the galvanometer, con- 
nected with the thermo-electric couple, taken at equal intervals of time, gives 
os more the difference of temperature between the globe and the surrounding 
envelope. 
From the readings thus taken the rate of cooling, or the rate of heating, of the 
globe is calculated ; and, finally, the absolute loss (or gain) of heat per unit of 
cooling surface, a unit difference of temperatures of cooling surface and surround- 
ings, per unit of time, is obtained. 
The method of standardising the thermo-junction need not be discussed in this 
abstract, although it is not devoid of interest. It was explained fully in the paper. 
Neither is it necessary to go at length into details of the calculations. It will be 
sufficient here to give the results of the experiments, remarking at the same time 
that the work is by no means concluded. 
In a paper communicated to the Royal Society in 1893, and published in the 
‘Philosophical Transactions,’ the results are given of experiments carried out in 
an exactly similar way to that described above. The same copper globes were 
used as the cooling bodies, and the same spherical shell as the envelope. The only 
practical difference between those experiments and these now described is that the 
latter were carried out at the lowest temperatures the author could reach, and the 
former at the highest. The latter are near the temperature of liquid air, and the 
earlier at some three or four hundred degrees higher in the scale. The results of 
the two series of experiments are thus exactly comparable. 
1 It is difficult to maintain these extremely low pressures. The application of 
the intense cold of the liquid air to the copper shells, which constitute the enclosure, 
causes them to warp, and develops frequently an excessively minute leak, which causes 
a very slight rise of pressure, This curious phenomenon needs investigation. 
2 Many important details of similar experiments are given in a paper communi- 
cated some years ago to the Royal Society’s Philosophical Transactions. 
8 It has been shown in previous papers that the difference of temperatures (under 
present circumstances) between the centre of the cooling globe and the surface is so 
small that it may be neglected, and, this being the case, it is obviously better to put 
the junction at the centre of the globe than at any other point of it. 
