SUNLIGHT AND ITS MEASUREMENT 161 



aqueous lines between 0.535m and 0.650 n produce a depression 

 not shown in this graph but whose effect may be noted by the 

 decreased transmission shown in the upper graphs of figure 3. 

 It will, of course, be evident that the absorption in the water 

 vapor bands will increase and decrease with changing atmos- 

 pheric contents of water so that the areas of the corresponding 

 depressions in the curve will change likewise. 



The effect, of water vapor upon the incoming radiation is, of 

 course, dependent upon the total amount of vapor through 

 which the beam passes. This amount of vapor is usually stated 

 as the number of centimeters of ''precipitable" water, i.e., the 

 depth of water that, when evaporated into a vertical column 

 of air of the same cross section, would produce the amount of 

 vapor that would actually be encountered between the observing 

 station and the limit of the atmosphere in a cylinder of the same 

 cross section. There is, however, no simple relation by which 

 vapor pressure determinations at the level of the observing 

 station can be converted into equivalent values for depths of 

 precipitable water because air currents produce unknown changes 

 in the moisture content of the air above the level at which the 

 determinations are made. 



Empirical expressions have been formulated by Harm 15 and 

 by Humphreys 16 to express the mean or average relation be- 

 tween these two quantities. 17 At elevations above sea level 

 allowances must be made for altitude. 18 



15 Harm, J. von, Lehrbuch der Aletereologie. 3te Aufl. 224-226. Leipzig, 

 1915. 



16 Humphreys, W. J., The amount and vertical distribution of water vapor 

 on clear days. Bull. Mt. Weather Obs. 4: 121-128. 1911. 



17 If the depth of precipitable water vertically above a station at sea level is 

 denoted by Q and the prevailing vapor pressure in centimeters of mercury at 

 the same sea level station by eo, the usual form of the expression connecting them 

 is Qo — Ke , in which K is an empirical constant. Hann found K = 2.3, while 

 Humphreys takes it as 2.0 for "clear days." 



18 For a discussion of this question and a more accurate method for determining 

 precipitable water see Fowle, F. E., The spectroscopic determination of water 

 vapor. Astrophys. J. 35: 149-162. 1912. 



Idem ibid. Ann. Astrophys. Obs. Smithsonian Inst. 3: 171—181. 1913. Ibid. 

 The determination of the aqueous vapor above Mount Wilson. Astrophys. 

 J. 37: 359-372. 1913. 



Idem. ibid. Ann. Astrophys. Obs. Smithsonian Inst. 3: 182-193. 1913. 



