EartKs Nonadiabatic Atmosphere. 523 



The ascension of April 27, 1909, is given separately in order 

 to show more fully the scope of the computing, in Table 1, 

 and the summary in Table 2 is arranged to bring out the 

 relations in the Middle Latitudes and in the Tropics. The 

 data for Victoria Nyanza are taken from the Meteorol. Zeit., 

 Dec, 1910, which is a mean of several ascensions, especially 

 Aug. 30 and Sept. 5, 1909. Above the 17,000 meter level the 

 temperature of Aug. 30, lower than —75° C, should be veri- 

 fied for several reasons before accepting them as characteristic 

 of the strata at the elevation 17 to 19 kil. over the equator. 



The observed temperature at the given height is the basis of 

 the reduction, the value of n being found by taking the 

 adiabatic gradient for the difference of elevation and dividing 

 it by the observed difference. It varies from the adiabatic 

 system at all strata, and it may become very large, having 

 positive or negative values in nearly isothermal layers. The 

 value of n is assumed to be uniform in the stratum from which 

 it is computed by the top and bottom temperatures. The 

 check P = pRT is complete throughout. The pressure being 

 reduced to meters of mercury agrees closely with that observed. 

 The gas coefficient and the specific heat are variables charac- 

 teristic of nonadiabatic atmospheres. Since the adiabatic 

 specific heat is C^ a =993 # 58 it is seen that (Cp a —Cp lt ) (T a — T ) 

 becomes a very large quantity, especially in the upper strata, 

 and it is equal to i(q\—q\) + (Q i — Q a ) the kinetic energy of 

 circulation and of heat radiation between the two levels. The 

 second check equation is closely fulfilled, but the small 

 differences indicate that the data of the temperatures are not 

 perfectly adjusted to the elevation. It should be especially 

 noted that the equation in common use by meteorologists, 



is unbalanced, lacking the large heat term (Q, — Q„), and on that 

 account the ordinary dynamic data are never adjustable with- 

 out the compensation for the loss of heat by radiation. It is 

 seen that the entropy increases with the height if the tempera- 

 ture steadily diminishes, but changes from negative to positive 

 values in cases of an inversion, as in the stratum 500 to 1000 

 meters. The third check equation (Q, — Q ) = (Wj — W„) + 

 (U t — U ) is confirmed, and the escaping heat depends upon an 

 excess in the loss of the inner energy over the external work. 

 The loss of the inner energy is nearly constant per 1000 

 meters, U 1 — U =7152, but the work diminishes with the alti- 

 tude and the decrease of the density, so that Q 1 — Q , the loss 

 of heat, increases with the elevation, whose rate of change will 

 be further noted. The radiating function, 



