THEORETICAL DISCUSSION. 267 



Imagine that the air at about 5,000 metres altitude is saturated vfith. moisture and has 

 a temperature of — 60°C. its condition will then be represented on the diagram by the point 

 A. If this air is now carried downwards without receiving or losing any heat to the groimd, 

 its temperature and pressure at the ground will be represented by the point B. If while it 

 is at the ground it receives heat its temperature will change, but not its pressure. If its 

 temperature rises 10° its condition will be represented by the point C. Now let it be carried 

 upwards without receiving or losuig any further heat and its condition at each height will be 

 represented by the line CD. At D it meets its saturation line and a further increase in 

 height will cause precipitation. It will be noticed that D is higher than A. Let another 

 mass of the same air from the same layer be lowered from A to B and then lose heat by 

 radiation, or any other method, until its temperature is reduced 10°C. Its condition will 

 then be represented by E, now let it rise adiabatically along the line E F it will then be 

 saturated at F, i.e., it will be saturated at a lower height than that from which it started. 

 In these two cases the point of saturation has been raised and lowered in consequence of 

 adding and subtracting heat respectively. This relationship is true generally ; the formal 

 proof is very easy, but need not be given here ; it is sufficient to state that if air is 

 saturated at any height in the atmosphere and subsequently receives heat it must rise to a 

 greater height before it becomes saturated again and conversely if heat is abstracted from it, 

 it will become saturated at a lower altitude than it had originally.* This is true no matter 

 how or where the heat is added or subtracted. We at once see why it is that as a general 

 rule an anticyclone in temperate arid tropical regions produces dry weather. The air enters 

 the anticyclone in the upper atmosphere and descends towards the ground which has been 

 made hot owing to the bright sunshine due to the absence of clouds. In consequence the air 

 in such an anticyclone on the whole receives more heat from the hot gromid than it loses 

 by radiation, it would therefore have to rise to a greater height than that where it entered 

 the anticyclone before precipitation occurs. This is practically impossible and therefore pre- 

 cipitation is of exceedingly rare occurrence in the anticyclones with which we are the most 

 famihar. 



In the Antarctic, however, the conditions are reversed. It is only during a very short 

 period of the year and then only during parts of each day that the air receives more heat 

 than it radiates. This is shown by the tendency to form ' temperature inversions ' near the 

 ground which are especially well marked during the wintet. 



We will now consider an actual case. The curve J H K in iigin-e 83 represents the 

 actual temperature found at different heights on August 17th, 1911. The temperatiu-e gradient 

 actually found between 1 and 2 kilometres, represented by the part of the curve H K, 

 has been assumed to be representative of the gradient throughout the upper atmosphere ; 

 this part of the curve has been continued as a straight line to the top of the diagram. 

 Now suppose that the air enters the Antarctic as a saturated current at about 6 kilometres 

 altitude. Its temperature and pressure will then be represented by the point G approximately 

 and G P will be its saturation line. Now a mass of air which has just entered the Antarctic 

 under these conditions is radiating heat and in consequence becoming denser, it therefore 

 descends. If it descends rapidly it warms up owing to adiabatic compression, its temperature 

 at each height being given by a line thi'ough G parallel to the adiabatic lines. It wall be 

 seen fi-om the diagram, however, that this would cause its temperature to be higher than that 

 of the surrounding air the temperatm'e of which is given by the line G H. Its higher 



* To be quite correct pressure instead of height ought to have been used as the criterion in this statement. 

 But our whole discussion is simplified by assuming that the relationship between pressure and height remains 

 constant, so that the two terms are synonymous. Also it is easier to visualise height than pressure and the 

 use of the height instead of the pressure tends to a clearer exposition. 



