WHICH PRODUCE LIGHT AND RADIANT HEAT. 391 



calorific and luminous constitution of the solar spectrum by the inter- 

 position of diaphanous substances,) if, instead of water, a plate of glass 

 be employed, the same variations will take place, though on a more 

 extended scale ; in other words, the last obscure limit of the normal 

 spectrum and the maximum of temperature approaches the most re- 

 fracted part more nearly than it does when the medium is a layer of 

 water of equal thickness*. In all cases, the ratios of luminous intensity 

 between the several parts of the spectrum remain invariable, because of 

 the perfect transparence of the media traversed by the solar rays. 



But if the plate of uncoloured glass be removed and coloured glass 

 substituted for it, the luminous spectrum will be entirely altered. If, 

 for example, a blue cobalt glass be employed, not only does the orange 

 disappear, but a great part of the green, a little of the blue, and the 

 middle of the red band, so that the spectrum then looks like a series of 

 zones, more or less broad and luminous, mixed with obscure bands. A 

 finely violet-coloured glass usually effaces the orange and the yellow, 

 and leaves but the red on one side and the blue and indigo on the 

 other. In fine, a red glass, as it intercepts the other rays almost en- 

 tirely, may be said to afford a passage to none but rays of its own 

 colour. 



Now, in examining the distribution of the heat of the obscure and 

 the luminous bands, so capriciously coupled together in these several mo- 

 difications of the spectrum, we find the calorific energy more or less di- 

 minished according to the nature of the glass interposed ; but the maxi- 

 mum always remains nearly in the same position, and the temperatures 



• I have shown in my first Memoir on the transmission of heat through solid 

 and liquid bodies how we may account for the different positions taken by the 

 maximum of temperature in the solar spectra produced by prisms of different 

 substances. 



The above-mentioned experiments decisively prove that the position of this 

 maximum must depend not only on the matter, but also on the mean thick- 

 ness of the prism. To convince ourselves of this, we have only to take a large 

 hollow prism, filled with water, and partly cover one of its lateral faces with an 

 opake plate laid in the direction of its length, so as to leave the side situated 

 towards the refringent angle perfectly free. Upon measuring the temperatures 

 of the different zones of the spectrum, we shall see that the maximum of heat 

 which, when the prism is entirely free, is found in the yellow, now that one 

 face is partially covered by the plate, approaches the last red limit and that 

 more closely in proportion as the portion which remains uncovered in the di- 

 rection of the edge is smaller. These variations are reproduced with more 

 or less energy by employing for the construction of the prisms sohd diaphanous 

 bodies or liquids different from water ; but there are no variations when we 

 employ rock salt. Hence it is clear that this substance, which transmits all the 

 calorific radiations of terrestrial sources with the same intensity, transmits solar 

 heat also without producing any change in the relative intensities of the dif- 

 ferent rays. It is for this reason that I have thought it advisable to make use 

 of a rock-salt prism in the dispersion of solar heat, and afterwards to consider 

 the alterations produced in the relative intensities of the refracted rays by the 

 interposition of transparent bodies. 



