156 S. P. Langley on the Selective Absorption of Solar Energy. 



screen is interposed between the bolometer and the grating ; 

 for the temperature of the screen itself, as it is replaced or 

 withdrawn, will certainly affect such measurements as these. 

 Through the whole course of the experiment the bolometer is 

 uninterruptedly exposed to radiations from the grating, 

 whether reflected by it or emanating from its own substance. 

 The interruption of the solar radiation is effected at the other 

 end of the train, 5 metres beyond the grating itself. In the 

 gratings employed one of the second spectra is very feeble, or 

 almost lacking. The rays of the second spectrum are neces- 

 sarily superposed on those of double the wave-length in the 

 first ; and as all evidence of solar radiation in the most sensi- 

 tive apparatus at the sea-level dies out near \=0 M, 3 in the 

 ultra-violet, it follows that we can measure down in the first 

 spectrum as far as \= (F'6, or in fact further, without any fear 

 whatever of our results being affected by the underlying 

 second spectrum, even if that were a strong one. Underlying 

 (f'l, which is near the limit of the red in the first spectrum, 

 is O^^ in the second, where heat is practically still negligible. 

 We have therefore (knowing the amount of heat in the second 

 spectrum at 0^*5, and knowing that our ultimate point of 

 measurement at 1^*0 in the first spectrum overlies (F'5 in the 

 second) the means of asserting with confidence that no con- 

 siderable error can be introduced from this cause, after an 

 allowance has been made here for the minute effect of this 

 second spectrum. An allowance is also made to reduce the 

 effect to that which would have been observed with a grating 

 so coarsely ruled as to cause no considerable deviation from 

 the slit of any portion of the spectrum measured. The bolo- 

 meter (in a constant position relative to the concave mirror 

 such that the optical axis of the latter bisected the angle be- 

 tween its central thread and the centre of the grating) was 

 moved, together with the mirror, by a tangent-screw in arc, 

 so that the spectrum appeared to traverse its face. 



The actual angular deviation of any ray under examination 

 was obtained from a divided circle on which the arm carrying 

 both mirror and bolometer moved. A particular description 

 is not given, as the whole apparatus was replaced by a more 

 perfect one later. That actually used will be intelligible by 

 the sketch, fig. 1, where S is the slit, Gr the grating, M the 

 concave mirror, B the bolometer, and C the divided circle. 



The light came from the silvered mirror of a heliostat, 

 passing through the slit at a distance of about 5 metres from 

 the grating, which was bolted immovably above the centre of 

 the circle of a massive dividing-engine, with the grating's 



