SiO M. MELLONI ON THE POLARIZATION OF HEAT. 



coloured glasses to act in this case like the green and opake-black glasses 

 of the preceding table, and vice versa. 



Moreover, the white, red, orange, yellow, blue, indigo, and violet 

 glasses all produce very nearly the same alteration in the value of the 

 index of polarization of the green tourmalines : and we have just ob- 

 served that these glasses have a very marked effect on the pale yellow 

 tourmalines. This effect, varying in different systems, but constant in 

 each particular system through the whole series of plates, is perfectly 

 analogous to the uniformity of the ratios which, notwithstanding the 

 changes of intensity, are observed to exist between the quantities of 

 heat transmitted by the same glasses (coloured and uncoloured) when 

 successively exposed to the rays emerging from the several kinds of 

 screens*. We are thus brought back once more to one of the fore- 

 going conclusions, namely, that the colouring substances have no power 

 of elective absorption with respect to the rays of the calorific stream 

 which passes through the glass. 



Before we conclude our observations on the effect produced by the 

 screens in the index of polarization of the tourmalines we shall make 

 some remarks on the effects produced by the variation of thickness in 

 the interposed substance and by the solution of salts in water. 



On inspecting the second table it is easy to see that the influence of 

 each substance in augmenting or diminishing the index of polarization 

 is more powerful in proportion to the greater thickness of the substance. 

 Thus, water reduced to a layer of 0'"'"*74 in thickness causes the index 

 of the pale yellow tourmalines to rise from 22 to 36, while a layer of 

 gmra raises it to 67. On the other hand, a plate of dark glass 0"""*81 

 in thickness, which causes the same index to descend from 22 to 4, 

 would reduce it even to 1"5 if its thickness were about 2"™. All this 

 accords with the experiments of successive transmission, which show 

 that the calorific stream emerging from a given substance beco»ies more 

 simple, or, if the expression be preferred, more purified in proportion 

 to the greater thickness of the substance through which it has passed. 

 We find this to be the case with white light also in penetrating coloured 

 media. 



Tartaric acid, rock salt, and alum dissolved to saturation in water 

 make no sensible change in its action on the index of polarization of 

 the tourmalines. We have already shown, in a preceding memoir, that 

 alum and rock salt (which, of all perfectly diaphanous and colourless sub- 

 stances, are those that possess the maximum and the minimum of diather- 

 mancy) do not by being dissolved in water affect the diathermanous power 

 of this liquid*. The sensible equality of the action of these solutions 

 and that of pure water on the natural index of polarization of tlie tour- 



* Aim. lie Chlm. et de Phijs., toni. Iv. p. 5.5. 



