, Determinations of the Absorption of Heat by Liquids. 215 



rock-salt cell the heat-rays would not be sifted in any sensible 

 degree, and hence the absorption by the deep layers of liquid 

 which were employed would be sufficient entirely to mask the 

 opposite effect observed with a glass cell. I regret that when 

 making these experiments it did not occur to me to try how far 

 the foregoing assumption is true, by causing the beam to pass 

 through a plate of glass before it entered the rock-salt cell ; 

 and now I have not the opportunity of making this crucial test. 

 But the following experiment, which strikingly shows the influ- 

 ence of sifting, may be taken in support of the view here stated. 

 The divergent rays direct from the source of heat used through- 

 out these experiments were caused to pass through an empty 

 glass cell \ an inch deep. Falling on the pile, they gave a de- 

 flection represented by 100. Filling the cell with bisulphide of 

 carbon augmented this number to 110, or an increased heat of 

 10 per cent. The cell was emptied, replaced, and a second cell 

 3 inches deep, already filled with bisulphide of carbon, was inter- 

 posed in the path of the rays from the source. The beam had thus 

 to traverse 3 inches of bisulphide of carbon before it reached the 

 empty cell, so that on passing through an additional \ inch of the 

 same liquid, but little or no further absorption would take place; 

 and the action of the liquid in concentrating the rays should thus 

 be more pronounced. Representing by 100 the radiation through 

 the 3-inch stratum of liquid and the empty J-inch cell, filling 

 the latter with bisulphide of carbon raised this figure to 120, or 

 an increased heat of 20 per cent. Compared with the former 

 observation, this experiment shows that even in a glass cell at 

 least 10 per cent, of the total radiation is absorbed by a layer 

 '4 an inch thick of bisulphide of carbon ; much more likely, 

 therefore, is it that in a rock-salt cell the absorption by deep 

 layers of liquid should disguise the causes tending to augment 

 the transmission. 



Finally, it may be asked if pouring the bisulphide of carbon 

 into a glass cell increases the heat reaching the pile, how is it 

 that Melloni, who, as is well known, used glass cells in his experi- 

 ments on the diathermancy of liquids, does not mention the fact ? 

 This is accounted for, I believe, by Melloni' s method of experi- 

 ment : he did not determine the transmission through his empty 

 glass cell first, and then through the liquid added to the cell ; but 

 noting only the free radiation to his pile, he then introduced the 

 cell and liquid at once. Consequently he assigned an absorption 

 to some liquids which probably was wholly due to the glass walls 

 of his cell. Hence it is that whilst Melloni states a layer of bi- 

 sulphide of carbon "36 inch deep absorbs 37 per cent, of the heat 

 from an argand lamp, Dr. Tyndall, who used a rock-salt cell and 



