Determinations of the Absolution of Heat by Liquids. 213 



cell, whilst, further, bichloride and bisulphide of carbon are two 

 of the most diathermic liquids known. The index of the bichlo- 

 ride is nearer that of glass and rock-salt than the index of the 

 bisulphide, just as might be expected. Melloni has stated that 

 7'7 per cent, of the incident rays are reflected from a single 

 polished plate of rock-salt*. Although this figure is probably 

 rather too high, still there can be no doubt that any moderately 

 transcalent liquid, when poured into a cell whose walls toere suffi- 

 ciently near together, would, by its optical density, increase the 

 transmission more than it would intercept it. And this is ex- 

 actly what has been found in the experiments described in the 

 early part of this paper As, moreover, this abolition of reflec- 

 tion of the inner surfaces of the cell will take place at all thick- 

 nesses of the liquid layer, the absorption hitherto attributed to 

 the majority of liquids must be too low on account of this oppo- 

 sing effect — an effect as yet disregarded, but here shown to be 

 very sensible. 



The foregoing explanation, however, is quite incompetent to 

 explain the second group of facts — the augmented heat observed 

 when certain liquids were enclosed in vessels with glass ends. 

 For here the augmentation continued until the thickness of the 

 liquid was increased to the limits of experiment ; and in this 

 case the loss from absorption by the liquid must have been greater 

 than the gain, even if the interior reflection had been entirely 

 abolished. 



Here the cause is, I believe, mainly due to a well-known effect 

 of plane surfaces on the refraction of divergent rays. This will 

 be clear if we consider the case represented in fig. 2. Let S be 

 a source of rays diverging 

 till they meet the cell C, 

 and let P be the surface of 

 the thermo-pile. If, now, 

 the vessel C be empty, 

 the rays, after passing 

 through its sides,, will 

 move in the direction of 

 the dotted lines. Now let this vessel be filled with a liquid whose 

 density is approximately the same as the walls of the cell, the rays 

 will then take a course indicated by the continuous lines. In the 

 latter case the angle of the rays emerging from the cell will be 

 more acute. The distance of the divergent point S from the pile P 

 has thus virtually been diminished, and, in consequence, many 

 rays which, when the cell was empty, escaped the pile, now fall 

 on its surface. To obtain this action it is not, of course, neces- 

 sary that the refractive power of the liquid should be the same 

 * La Thtrmochrose, p. 1.95. 



Fi<?. 2. 



