Intelligence and Miscellaneous Articles. 315 



first trials a flint-glass prism and a glass concave mirror silvered 

 at its surface instead of an achromatic lens. The employment 

 of a network of lines engraved on the metal instead of the prism 

 would permit the influence of any elective absorption to be elimi- 

 nated. 



I have made numerous determinations of thermal curves of the 

 solar spectrum on exceptionally fine days, at different periods in 

 the years 1877 and 1878. These curves differ in the ratio of their 

 respective ordinates, but especially in the position of the thermal 

 maximum, as has been shown by Melloni. These curves were ren- 

 dered comparable with one another by bringing them to the scale of 

 the wave-lengths, and reducing, by means of the dispersion-curve of 

 the prism, the intensities to those which would correspond to the 

 theoretic case of the normal spectrum — that is, of constant dis- 

 persion. 



The following are, for the luminous part of these spectra, the 

 means of a number of concordant observations made under excel- 

 lent atmospheric conditions : — 



millim. milliiu. millim. millim. millim. millim. 



Waye-lengths 0-000676 0-000605 0-000560 0-0(30523 0'000486 0-000459 



Intensities 1000 820 760 070 540 460 



I have represented by 1000 the thermal intensity which corresponds 

 to a red radiation of wave-length 0*000676 millim. ; the intensities 

 measured in the ultra-red cannot find a place in this Table, the cor- 

 responding wave-lengths not being accurately known. 



Now here are the ratios of the luminous intensities of the same 

 radiations of the spectra of the following sources, compared with 

 the light of the sun : — 



millim. millim. millim. millim. millim. millim. 



Wave-lengths 0-000676 0-000605 0-000560 0-000523 0"000486 0-000459 



Electric fight 1000 707 597 506 307 228 



Drummond light... 1000 573 490 299 168 73 



Moderator lamp .. . 1000 442 296 166 80 27 



The electric light was from 60 large Bunsen elements, Foucault 

 regulator, with M. Carre's carbons in the focus of a metallic con- 

 cave mirror; the Drummond light, oxygen and illuminating-gas 

 thrown upon lime ; the moderator lamp, fed with colza-oil. I 

 measured the ratio of the intensity of each of the radiations of these 

 spectra, corresponding to the wave-lengths of the preceding Table, 

 to the intensity of the same radiation in the solar spectrum, repre- 

 senting these latter by the value of their thermal intensities, and 

 always representing by 1000 the intensity corresponding to the 

 wave-length 676. 



1'or luminous radiations which have undergone no weakening by- 

 previous transmission, there would be proportionality between the 

 thermal and luminous intensities of one and the same radiation, 

 whatever its origin, as MM- Jamin and Masson have demonstrated ; 

 but the experiments of M. Desains* have shown that, in the con- 

 * Compter Rendus, t. lxvii. p. 297. 



