10 S. P. Langley — Observations on Invisible Heat-Spectra. 



beat waves that penetrate our atmosphere.* To show this 

 more clearly we have drawn the solar spectrum given by the 

 rock-salt prism in its true position (though not in its true 

 amount) relatively to that of the heat curves cited. The max- 

 imum of the latter lies in every case, it will be seen, far below 

 the very lowest part of the solar invisible heat. 



(2) In spite of the compression of the infra-red by the prism, 

 these heat curves extend almost indefinitely in the direction of 

 the smaller indices, so far that we can, in fact, represent only a 

 part of this extent in our plate. The measures already cited 

 in case of the curve for the Leslie cube at 178°, for instance, 

 show very measurable heat at a deviation of 33°, which corre- 

 sponds to an index of refraction of 1/4511, while the smallest 

 index given in the plate is 1*49. 



(3) An increase of temperature increases every ordinate, but 

 not in like proportions, ordi nates corresponding to the heat in 

 the more refrangible parts always growing more rapidly than 

 those for less refrangible heat. 



(4) As a necessary consequence of this, follows the (inde- 

 pendently observed) fact of the progressive movement of the 

 maximum ordinate toward the more refrangible end, as the 

 temperature rises. 



(5) These prismatic curves are not symmetrical, the greater 

 portion of the area in every case lying below the maximum, 

 i. e., toward the greater wave-length, and the descent being 

 always most abrupt on the more refrangible side. 



As the heat spectra from surfaces at the temperature of boil- 

 ing water or melting ice are those to which the chief interest 

 attaches, in connection with the temperature of the soil, and as 

 these are not well shown on the same scale of ordinates with 

 that of the red-hot copper, we give an independent representa- 

 tion of these two in Plate 1, but upon the wave-length, not the 

 prismatic scale. Their maxima of heat fall at points. in the 

 normal spectrum which (as we explain later) are only approxi- 

 mately determinate on this scale, but which are probably at 

 least as low as the points (/) corresponding to the boiling water 

 maximum, and (g) corresponding to the position of the maxi- 

 mum ordinate in the spectrum of ice, at the melting point, or 

 lower. No attempt is made in this plate to represent the rela- 

 tive amounts of heat in the solar and Leslie cube curves, but 

 only their positions on the wave-length scale ; and here also it 

 will be understood that the latter curves really extend far 

 further to the right than the limits of the plate admit of show- 

 ing them. 



* The distinctive character of these radiations is also well shown by the fact 

 we have found that a thick film of lampblack, which is nearly as impervious to 

 the dark solar heat as to light, transmits more than 50 per cent of the rays in 

 question. 



