Any investigations on the subject of atmospheric absorption are of such importance in the study of meteorology, that we have deemed it advisable to present a preliminary notice of certain results obtained by us, without waiting to present a more detailed account which will be communicated at a future date. From 1874, when one of us commenced photographing the spectrum in the above region, till more than a year ago, the extremely various manners in which the absorptions took place caused considerable perplexity as to their origin, and it was only after we had completed our paper on the absorption of certain liquids, that a clue to the phenomena was apparently found. Since that time we have carefully watched the spectrum in relation to atmospheric moisture, and we think that more than a year’s observations in London, when taken in connexion with a month’s work, at an altitude of 8,500 feet on the Riffel, justify the conclusions we now lay before the Society. A study of the map of the infra-red region of the solar spectrum, and more especially a new and much more complete one, which is being prepared for presentation to the Royal Society by one of us, shows that the spectrum in this part is traversed by absorption lines of varying intensity. Besides these linear absorptions, photographs taken on days of different atmospheric conditions, show banded absorptions superposed over them. These latter are step by step absorptions increasing in intensity as they approach the limit of the spectrum at the least refrangible end. In the annexed diagram, fig. 4 shows the general appearance of this region up to λ 10,000 on a fairly dry day : the banded absorption is small, taking, place principally between λ 9420 and λ 9800 : a trace of absorption is also visible between λ 8330 and λ 9420. On a cold day, with a north-easterly wind blowing, and also at a high altitude on a dry day, these absorptions nearly if not quite disappear. If we examine photographs taken when the air is nearly saturated with moisture (in some form or another) we have a spectrum like fig. 1. Except with very prolonged exposure no trace of a spectrum below λ 8330 can be photographed. Fig. 2 shows the absorption bands, where there is a difference of about 3° between the wet and dry bulb, the latter standing at about 50°. It will be noticed that the spectrum extends to the limit of about λ 9420, when total absorption steps in and blocks out the rest of the spectrum. Fig. 3 shows the spectrum where the difference between the wet and the dry bulb is about 6°. Figs. 5 and 6 show the absorption of thicknesses of 1 foot and 3 inches of water respectively, where the source of light gives a continuous spectrum ; ⅛ inch water merely shows the absorption bands below 9420. It will be seen that there is an accurate coincidence between these “ water bands” and the absorption bands seen in the solar spectrum, and hence we cannot but assume that there is a connexion one with the other. In fact, on a dry day it is only necessary to place varying thicknesses of water before the slit of the spectroscope and to photograph the solar spectrum through them, in order to reproduce the phenomena observed on days in which there is more or less moisture present in the atmosphere. It is quite easy to deduce the moisture present in atmosphere at certain temperatures by a study of the photographs. There does appear a difference, however, in the intensity of the banded absorptions in hot weather and in cold about up to 50°. In the former they are less marked when the degree of saturation and the length of atmosphere traversed are the same as in the latter.
VII. The influence of water in the atmosphere on the solar spectrum and solar temperature
