514 Prof. S. P. Langley on the Invisible 



prismatic, not the normal^ scale. They are identifiable with 

 the portion extending from X to Y on the latter. 



Observations made during the winter indicate that the band 

 at 2^*64 (figs. 3 a, 3 b) is, with a high sun, largely filled up, 

 especially on the less refrangible side. At noon a subordinate 

 maximum has been found within the low sun limits of this 

 band at 2^*94, and a second one at 2 K 80 frequently accom- 

 panies it, producing subordinate minima at 2' A, 89 and 3 M *02. 



As the absorption increases, with a sinking sun, these sub- 

 ordinate maxima disappear to a very great extent, that at 2^*80 

 being the first to vanish as well as the quickest to grow, so 

 that at noon, on a cold day, it not only surpasses the maximum 

 at 2^94, but even begins to approach that at 3^20, while, 

 when the sun's altitude is less than 10°, the nearly uniform 

 part of the band extends from 2 K 45 to 3^*15 without a break. 

 Minor cold bands at 3^*37 and 3^*69 are suspected. The evi- 

 dence for their existence may be seen by inspecting the high 

 sun and low sun curves given in fig. 4. From 4 M, to 4 M, 5 

 we have another region of almost complete absorption, followed 

 by a maximum at 4?' 6, beyond which lies the longest break 

 of all, stretching from 5^ to ll/\ 



The solar heat throughout the region from 5^ to ll/* is 

 probably far greater than it here appears, and if the parts 

 struck out by our atmosphere w 7 ere restored it would probably 

 be found that a not wholly inconsiderable portion of the sun's 

 heat lies in this region ; for it must be borne in mind that 

 even the maxima are doubtless in some degree affected by a 

 linear absorption, so that, because a part of the spectrum 

 appears to be almost entirely transmitted, we cannot infer that 

 it necessarily is so, or that it has not, after passing through 

 the upper strata of the air, already parted with a considerable 

 portion of its energy. 



While the position of a line or band caused by light from 

 the centre of the sun is, in the upper spectrum, unchangeable, 

 in this extreme low spectrum (if we could photograph it, like 

 the upper, on cold and on hot days, with high sun and with 

 low) the absorption would be seen to increase not symme- 

 trically with the centre of the band, but more on one side 

 than the other, so as to considerably modify the position of 

 greatest absorption. This seems to be the explanation of a 

 curious fact which could not have been anticipated in advance 

 of observation ; that is, that the centres of several of these 

 bands and lines are under some conditions found to be shifted 

 to a recognizable extent, and hence their w r ave-lengths are, 

 within certain limits, variable. 



It seems as though the absorption, ivhich we see exercised 



