3C4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 51 



If therefore the ascent of the air suddenly becomes a descent, 

 then, at least in the first moments, re-evaporation must take place, 

 that is to say, the processes attending the immediately preceding 

 moments will be exactly repeated in the reverse order. 



If all precipitation were carried up with the ascending current of 

 air, as it is in the case of clouds from which no rain has as yet fallen, 

 then in general by reversing the movement the air would arrive 

 at its starting point in the same condition in 'which it had left the 

 earth. If, on the other hand, water has actually fallen from the 

 cloud, then by reversing the movement the air enters sooner into the 

 dry stage and its warming follows a very different law from that of 

 its cooling. 



The process is therefore reversible in its smallest details but not 

 in the larger nor as a whole, and it is precisely to this peculiarity 

 that, as is well known, we owe the Foehn phenomena, the differences 

 of the weather in the areas of high pressure and low pressure, the 

 peculiarities of the windward side and leeward side of mountain 

 ranges, etc. 



We meet analogous conditions in the warming and cooling of the 

 atmosphere in contact with a terrestrial surface that is subject to 

 insolation and radiation. 



A limit to the warming of the lowest layer of air is soon set by the 

 occurrence of unstable equilibrium, but, on the other hand, its cool- 

 ing can proceed as long as the radiation continues and so long as no 

 rapid renewal of the air is produced either by the drainage away of 

 the cooled air or by the wind. It is well known that the very low 

 temperatures that are observed in valleys are thus produced and 

 especially on tundras in winter and to a less extent also in other 

 seasons during very clear calm nights. The same is true of the so- 

 called inversions of temperature that were first observed in moun- 

 tainous regions and which for a long time were supposed to be 

 principally confined to such regions. 



Scientific balloon voyages have shown that these inversions occur 

 regularly at times of overpowering radiation and gentle atmospheric 

 motion. 



Moreover, balloon voyages furnish us far more complete pictures 

 of the temperature inversions than do simultaneous observations at 

 a summit station and at a neighboring valley station, in which latter 

 case such an inversion can pass entirely unnoticed when the location 

 of the (stratum of) highest temperature lies only slightly above the 

 level of the lower station, so that the temperature of the upper 

 station is still lower than that of the lower station. Thus, for exam- 



