S. P. Langley — Observations on Invisible Heat-Sjoectra. 5 



trum by means of a prism or grating, and to determine the 

 indices of refraction of its prominent parts, and by inference, 

 their wave-lengths. We have now been engaged on this re- 

 search at Allegheny at intervals for two years, a time which 

 will not appear extravagant to one acquainted with its extreme 

 difficulties. Not to dwell on these in detail, I will mention, 

 only, that the grating can not well be used on account of its 

 overlapping spectra, if for no other reason, and that the most 

 transparent glass, which we have found to be comparatively 

 diathermanous to dark solar heat, turns out to be. almost 

 absolutely athermanous to the heat from a surface at the tem- 

 perature of boiling water. 



Grlass being useless here, almost the only material of which 

 we can form our prism is rock-salt, and we must have not only 

 an entire train of lenses (both colli mating and observing) of 

 salt, as well as the prism, but the pieces must be of exceptional 

 size, purity and perfection of figure, to contend with these 

 special difficulties, and they must be maintained in condition, 

 in spite of the incessant deterioration of this substance. Finally, 

 as we wish to determine wave-lengths, these measures must be 

 very accurate, and the prism be capable of giving fixed points 

 of reference like the visible Fraunhofer lines. The prism we 

 are now using was made by Mr. J. A. Brashear of Pittsburgh, 

 and when freshly polished, gives a spectrum not only filled with 

 hundreds of Fraunhofer lines, but which shows distinctly the 

 nickel lines between the D's, and is probably the finest one 

 ever produced from this material. 



Such measures on the collective heat of black bodies as those 

 of Melloni and Tyndall have been made on large radiating sur- 

 faces, like those of the Leslie cube, but in order to form a spec- 

 trum, of this as of any other source, we must, of course, take 

 only such a limited fraction of the side of the cube as is repre- 

 sented by a narrow spectroscope slit ; so that both from its 

 minute amount and feeble intensity (even if we can pass it 

 through a prism, to form a spectrum), it is absolutely inappreci- 

 able, in anything like homogeneous portions, to the most deli- 

 cate thermopile, and difficult of attack even by the bolometer. 



We have employed, as radiating surfaces, Leslie cubes cov- 

 ered with lampblack and filled with boiling water or aniline, 

 the former givifg a radiating surface of temperature of 100° C, 

 the latter one of 178° 0. and also cubes filled with freezing 

 mixtures, with the latter of which Mr. F. W. Very, of this ob- 

 servatory, conducted in the cold days of last March one series 

 of measures in which the radiator was the bolometer itself, at 

 a temperature of — 2° 0. and the source radiated to, a vessel 

 filled with a mixture of salt and snow at a temperature of 

 — 20° C, thus determining the distribution of energy in the spec- 



