216 On Determinations of the Absorption of Heat by Liquids. 



a better method of experiment, finds the absorption by a corre- 

 sponding thickness of liquid under 20 per cent.* 



Summing up the results of this inquiry, the main points are 

 as follows. : — When the more diathermic liquids are introduced 

 between two plane parallel plates of rock-salt separated by a very 

 small interval, the rays from an artificial source are found to be 

 more freely transmitted than when air intervenes between the 

 plates. For a space '02 inch wide the increased transmission 

 amounts with bichloride of carbon to about 12 per cent., with 

 bisulphide of carbon to 9 per cent., and with chloroform to 4*5 

 per cent. This effect disappears and less heat is transmitted (1) 

 when the transcalency of the liquid diminishes; e. g. the same 

 thickness of sulphuric ether intercepts 30 per cent, of the heat 

 previously passing through the empty cell; (2) when the distance 

 between the plates is increased beyond, say, ^V of an inch in the 

 case of bisulphide, and T 2 o of an inch in the case of bichloride of 

 carbon. The increased transmission by these liquids reappears, 

 however, in thicker layers when plane parallel glass plates are 

 substituted for rock-salt, and continues, apparently indeed aug- 

 menting, as the depth of the cell increases, so far as the experi- 

 ments were carried. Bisulphide of carbon, poured into a cell 

 with glass sides 1*2 inch apart, increases the heat falling on the 

 pile 6 per cent., and bichloride of carbon a still larger amount. 

 Altering the temperature or nature of the source, the size of the 

 aperture in a screen behind the cell, or the position of the cell, 

 makes no material change in these results. But altering the 

 character, or augmenting the thickness, of the walls of the cell 

 has considerable influence. For example, if the cell-walls be of 

 glass, increasing their thickness from one- to three-tenths of an 

 inch raises the heat falling on the thermoscope 6 per cent, when 

 equal depths of the selfsame liquid are poured into the cell. 

 Again, by merely changing the parallel sides of the same cell 

 from rock-salt to precisely similar plates of glass, the very same 

 liquid can be shown to intercept a certain quantity of the heat 

 falling on the thermo-pile in the one case, and to augment that 

 quantity in the other — the difference amounting to upwards of 

 10 per cent, of the total radiation through the empty cell. 



* Dr. Tyndall has pointed out (Phil. Trans. 1864, p. 328.; Phil. Mag. 

 vol. xxviii. p. 43i)) how seriously the glass cells used by Melloni must have 

 influenced the results of his experiments. Melloni, indeed, saw this him- 

 self, but endeavoured to prove, by trying two of his liquids between rock- 

 salt plates, that the error vanished in the thickness of liquid he employed ; 

 (La Tliermochrose, pp. 1(13 and 19!)). The liquids chosen were colza oil 

 and water — the worst he could have selected; for had Melloni made the 

 Comparison with more diathermic liquids, he would have discovered his 

 error. 



