210 Mr. W. F. Barrett on Sources of Error in 



filling the cell with bisulphide of carbon increased this number to 



106. 



The experiment was twice repeated with the same result ; and 

 the effect was not uncertain or transitory, but a steady perma- 

 nent augmentation. 



I next tried a series of cells with parallel glass ends of varying 

 distances asunder, ranging from 01 of an inch to 3 inches, bi- 

 sulphide of carbon being used in each case. The source and the 

 disposition of the apparatus were the same throughout, and the 

 result of all the experiments showed that increasing the depth of 

 the cell did not, as was the case with rock-salt cells, reverse the 

 effect, but rather augmented the increased transmission — the 

 transmission through an empty glass cell 3 inches deep being- 

 raised 9 per cent, by filling it with bisulphide of carbon. 



Returning to rock-salt plates, for the sake of comparison a 

 cell 2 inches deep was tried. As anticipated, instead of increased 

 transmission, considerable absorption was now observed when the 

 bisulphide of carbon was poured into the cell. x\nother similar 

 rock-salt cell, 0*8 of an inch deep, was also tried, with the same 

 result. In fine, with cells having rock-salt ends, after a thick- 

 ness of 0*2 of an inch had been attained, absorption invariably 

 was found on the introduction of bisulphide and bichloride of 

 carbon ; but with cells having glass ends, augmented transmission 

 was always observed with both liquids, the bichloride of carbon 

 having always a slight advantage. 



There can therefore be no doubt about the fact that the ab- 

 sorption of the more diathermic liquids is entirely disguised 

 when glass cells and a source of rays of the character here used 

 are employed in their examination. Moreover the effect was 

 found to remain the same when the nature of the source was 

 altered, or its temperature raised or lowered, or when the cell 

 was moved nearer to or further from the source of heat : nor was 

 the result changed by causing the diaphragm in the screen (D, 

 fig. 1) to be larger or smaller than the diameter of the aperture 

 in the cell. The same result was also obtained whenever the 

 experiments were made. For example, I here collate a number 

 of determinations, some made at an interval of twelve months 

 apart. The radiation through an empty glass cell 1*2 inch deep 

 being taken as 100, the numbers 



106-2, 106-9, 104-6, 106, 106, 



or a mean of 6 per cent., expressed the increased temperature 

 noticed when the cell was filled with bisulphide of carbon. Ne- 

 vertheless the result is just as certain that when the walls of the 

 cell are of rock-salt this increase is not shown ; numerous experi- 



