92 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1917. 



and from 13.0 to 16.0^, respectively. In these bands its absorption is nearly 

 total from 4.0 to 4.8/t and from 14.0 to 15.6/* even when carbonic-acid gas is 

 present in much less quantities than the atmosphere contains. * * * In the 

 absence of water vapor the total absorption possible by carbonic-acid gas would 

 be 14 per cent. In all the lower regions of the atmosphere, however, water 

 vapor is present in such quantities as almost completely to extinguish the radi- 

 ation of the earth's surface in these two special regions. * * * It therefore 

 does not appear possible that the presence or absence, or increase or decrease, 

 of the carbonic-acid contents of the air are likely to appreciably influence the 

 temperature of the earth's surface. 



It seems certain, in view of what has been said that the earth's solid and 

 liquid surfaces, and the lower parts of the atmosphere, contribute directly 

 almost nothing to the amount of radiation which the earth as a planet sends 

 to space. The earth's surface and the lower atmosphere, of course, exchange 

 radiation together, and by this process and by convection the heat of these 

 regions ascends toward space. But convection grows less and less as the air 

 becomes rarer, and must at length cease to be an appreciable factor. It is the 

 water vapor and carbonic-acid gas far above the earth's surface, where the 

 absorption of the rays by the water vapor and carbonic-acid gas lying still 

 higher becomes small, that form the true radiating surface of the earth con- 

 sidered as a planet. * * * With the scanty material at hand, and in con- 

 sideration of the distribution of water vapor in the free air, it seems safe to 

 put the effective position of the radiating surface at fully 4,000 meters above 

 sea level * * * at a probable mean temperature of 263° absolute centi- 

 grade or — 10° centigrade. 



Some writers have misinterpreted these remarks and understood 

 us as supposing that there' is a special layer at 4,000 meters elevation 

 above sea level which prevents radiation escaping from below and 

 whose own radiation passes unhindered to space. Our meaning was 

 quite different. Every layer from sea level to the limit of the atmos- 

 phere contributes something to the total radiation output of the earth. 

 But, because of the great absorption of superposed water vapor and 

 clouds, the lower solid and liquid and atmospheric layers contribute 

 little, while because of their dryness the higher atmospheric layers 

 contribute little. Roughly estimating the various factors, we con- 

 cluded that the center of activity of the radiation of the earth as a 

 planet could be set at about 4,000 meters elevation. 



How far are these conclusions now to be altered ? As to the effect 

 of cloudiness, not at all. As to water vapor Mr. Fowle finds the 

 following results on the percentages of absorption of rays from a 

 perfect radiator at the earth's mean temperature in atmospheric 

 columns containing besides carbon-dioxide sufficient to produce maxi- 

 mum absorption, water vapor which if precipitated would produce 

 certain depths of liquid water: 



Ppt. water cm 0. 003 0. 03 0. 3 3. 



Absorption 49 57 66 75 



In order to apply these data I give figures for the average quantities 

 of terrestrial water vapor which, according to Hann, exist in vertical 



