DISTRIBUTION OF ATMOSPHERIC PRESSURE VON BEZOLD 369 



many charts contain no portion of the equator, but should speak 

 of a degree of the meridian, or a degree of latitude, since on every 

 chart a portion of the meridian appears or can be easily drawn. 

 The degree measured on the meridian, therefore, under all circum- 

 stances corresponds closely to a length of 1 1 1 kilometers. 



Therefore by comparing any distance on any chart with a degree 

 of latitude we can express it in fractions or multiples of in, in 

 kilometers. 



Since the formulae (i) collectively contain certain lengths (or 

 distances on the earth's surface) therefore they are specially 

 convenient for studies based on the synoptic weather map. On 

 the other hand, they suffer from the defect that in contrast with 

 formula (2) they contain two variables, i. e., G and p or / and p, or 

 strictly speaking three, since p itself depends on pressure and tem- 

 perature. 



In the ordinary discussions we consider only one variable G, since 

 p is assumed to be constant. 



But this is only a crude approximation for in fact 



/? 273 /? T 



p = Po 760' 27ST~t° Tp = Po J T 

 where for simplicity /? = 760, T = 273, T = 273 + / or the abso- 

 lute temperature, and p — 1.293, tne value that p assumes corre- 

 sponding to the normal pressure /?„ and the temperature T . 



If we substitute this value in the formulas (1) and recall that 



13.6 ^ ot 13.596 p._ V9m 



i°o -* l°o ■* 



is simply the gas constant for dry air occurring in the law of Mariotte- 

 Gay-Lussac and which is ordinarily represented by the letter R, 

 then we have 



r-iiiui-p R s (la) 



dB T 



r = TilJ R s M 



JB T 



r = T^8 (ir) 



These formulae contain altogether three independent variables, 



i. e.,/?, Tand G, — or/?, Tand — , — or/?, Tand / — instead of the one 



dl 



independent variable that is implied in the ordinary approximate 



consideration of this subject. 



