264 Bigelow — Radiation in the Earth? s Atmosphere. 



The value of Kj/K is a wide variable, 3*5 to 5*2 at least, and 

 Humphreys' arguments for 1/b = 2 are not valid. There is 

 no evidence that the planetary radiation described in his paper 

 can be verified. It can be shown that his theorem of the 

 relation between the " thickness of the water layer that would 

 result from a condensation of the water vapor in the atmos- 

 phere above any given level . . . .may be expressed 

 approximately by d = 2w" is not generally true. 



Having derived values of log c and a for different conditions 

 in the atmosphere, as represented by, (1) twelve balloon ascen- 

 sions ; (2) the northern hemisphere generally ; (3) cyclones and 

 anti-cyclones ; (4) the diurnal convection at Cordoba, which 

 will be published in Bulletin No. 3, Oficina Meteorologica 

 Argentina, we can often avoid the long computation of the full 

 series of formulas by computing K 10 approximately from T 10 , 



by (11). 



Table 10. 

 Log (K 10 ) = log c + a log T 10 



z 



logK^o 



log(K I0 ) 



Alog(K 10 ) 



a 



log c 



log(K, )o 



6000 



4-779 



4-780 



-o-ooi 



3-820 



— 5-560 



4-781 



5000 



4-827 



4-789 



+ 0-038 



3-822 



— 5-564 



4-826 



4000 



4-871 



4-848 



+ 0-023 



3 825 



-5-571 



4-869 



3000 



4-905 



4-886 



+ 0-019 



3-827 



— 5-575 



4-906 



2500 



4924 



4-921 



+ 0-003 



3-828 



— 5-578 



4-923 



2000 



4-946 



4-955 



-0-009 



3 831 



— 5-585 



4946 



1500 



4-968 



4-987 



-0-019 



3-834 



— 5-592 



4-970 



1000 



4-986 



4-971 



+ 015 



3-837 



— 5-598 



4-984 



From the data in the last two columns of Table 1, log c and 

 a, compute log (K 10 ) in column 3, Table 10, and compare with 

 log K 10 in the second column derived from Table 1, as com- 

 puted through the entire series of formulas. The differences 

 A log (K 10 ) may be compensated by changing the value of a in 

 Table 1 at the following rate, for A log (K 10 ) = +0*0005 we have 

 A a = +0*001. Correcting a of Table 1, as in column 5, Table 

 10, and taking the corresponding log c from Table 3, we 

 find by formula (11) the log (K 10 ) thus adjusted in the last 

 column, which conforms with log K 10 in the second column. 

 The required changes in a are a few units in the third decimal, 

 and thus the entire series of formulas is checked. 



* Mt. Weather Bulletin, vol. 4, part 3, x, xi, xii. 



