and Unmeasured Wave-lengths. 39 



below the range of vision, but of their wave-lengths nearly 

 nothing has, till lately, been ascertained, partly for want of 

 sufficiently delicate heat-recognizing apparatus, and still 

 more from the fact that it is difficult to use the grating here, 

 owing to the overlapping spectra, and to the consequent neces- 

 sity we have till lately been under of separating these rays 

 only by the prism, which gives no measure of their wave- 

 lengths. Physicists have accordingly attempted to find 

 these, by observing what deviations correspond to known 

 wave-lengths in the visible portion, and by trying to deter- 

 mine from theoretical considerations what relations should 

 obtain in the infra-red ; but the various formulas by which 

 these supposed relations have been expressed have not till 

 lately been tested. The difficulty has been partly overcome 

 in the last few years by the application of the linear bolo- 

 meter to the spectrum formed by the concave gratings with 

 which Prof. Rowland has furnished us ; the deviations of the 

 heat-rays having first been observed and the principal lines 

 of the infra-red region mapped by the joint use of the 

 bolometer and a flint-glass prism, in 1881. It will be 

 remembered that one of the best known formulae on which 

 physicists till lately relied for determining the relations of 

 wave-lengths to deviations was Cauchy's ; that this set an 

 absolute limit to the wave-length which any prism could 

 under any circumstances discriminate, and that this supposed 

 extreme wave-length was somewhere between 10,000 and 



o 



15,000 on Angstrom's scale. Besides this theoretical limit, 

 it was supposed that glass absorbed dark heat to such an 

 extent that the longer solar heat-waves would be stopped in 

 the substance of the prism, even were there no other obstacle. 



In 1881, however, we found at Allegheny by actual trial 

 that heat-waves, whose wave-length was far in excess of the 

 theoretical limit, passed through a flint-glass prism, so that it 

 was ascertained both that this supposed limit did not exist 

 and that common glass was nearly diathermanous to all the 

 dark heat which comes to us from the sun. By means of a 

 glass prism and the bolometer, we were thus able to pursue our 

 researches and map the infra-red or invisible solar spectrum 

 to a point where it actually came to an end. What the wave- 

 length of this point was we could not tell, for it lay entirely 

 outside of what theory had till then pronounced possible. 



Next, using the grating, we have at Allegheny determined 

 the wave-lengths of most of the newly discovered solar-heat 

 region, by direct observation, and shown that it extended to 

 the unanticipated length of 2'7 //, (e, PI. IV. fig. 1) (i. e. 27,000 



o 



on Angstrom's scale). I cite these facts, which have already 



