F. W. Very — Sky Radiation and Isothermal Layer. 375 



explanation limits local convection to a superficial layer about 

 1500 meters deep within which the ordinary fair-weather cumuli 

 originate. Temperature variations, both diurnal and annual, 

 continue, however, above this point. The range of temperature 

 at 8000 meters is over 50 per cent of that at the surface, and 

 variations of 18° C. on dates only a few days apart are found 

 at an altitude of 17,000 meters. 



Irregularities in the distribution of the chief absorbents of 

 radiation in the atmosphere, which are aqueous vapor and its 

 products, the hydrols, are certainly responsible for most of the 

 atmospheric thermal variations, and if the term "convection" be 

 extended to include not only the process of thermal exchange 

 between the earth's surface and the first cumulus level (which 

 is the sense in which Clayton employs the word in the passage 

 cited), but also the wider cyclonic and planetary circulation, we 

 may accept Blair's statement, understanding by this a convec- 

 tional process by which warm, moist air is carried upward from 

 the earth's surface, or by secondary convection from local cloud 

 layers, or on a larger scale in a planetary circulation where 

 moist air, starting from the earth's surface in the tropics, travels 

 polewards, gaining both easting and altitude as the latitude 

 increases. 



Convection, in the sense of an overturning of gravitational 

 equilibrium, has practically ceased in the upper layer of temper- 

 ature-inversion ; but persistence of momentum carries some part 

 of the jDoleward planetary circulation into this elevated region, 

 thereby continually replenishing it with sufficient moisture to 

 serve as a potent absorbent of the solar rays within the limits of 

 the numerous and broad aqueous bands of the infra-red spec- 

 trum, and at times bringing an unusual accession of vapor. 

 Here is the main source of occasional exceptional heat in the 

 upper inversion layer. Nevertheless, the constancy of this 

 layer throughout the globe differentiates it immediately from 

 an effect of the ever-varying winds, and the sudden increase in 

 the second differential of the thermal energy transferred from 

 layer to layer on entering this region, marks the process by 

 which the isothermal layer originates as a radiant one. Here, 

 and here only, can there be any direct radiation of the atmos- 

 phere to space. The radiant process does indeed exist at all 

 levels in the atmosphere, but in the lower air it is an indirect 

 successive alternation of radiation and absorption of radiation 

 by innumerable steps — a mode of communication more rapid 

 than that by penetration through local molecular motion, but 

 still one which is slow compared with the velocity of radiation.* 



* Compare the account of this step-by-step process of atmospheric radiation 

 which is given in my work on "Atmospheric Badiation," pages 114-115 and 

 124. 



