SHORT MEMOIRS ON METEOROLOGICAL SUBJECTS. 



observations); for the first intervals, from 1 to 8,000 feet, I united each 

 tbree into one mean value, and thus obtained the following: 



The reduction of the observed values of the vapor-tension to the 

 nearest intervals occurring in the table was made by means of interpo- 

 lation after the values of G had been computed from every observation. 

 These numbers show an increase of the value of G with the altitude. 

 The method of least squares now gives us the expression 



C = 20251 + 0.11334 7i in English feet, 

 or 



C= 0172 + 0.11334 /t in meters, 



where h is to be expressed in units of either English feet or meters 

 respectively. The mean value of G as computed by this formula is 

 6,517 meters. Since small changes in G have but slight influence upon 



the value of -^, we can also try the effect of considering C as a constant 



i^o 



equal to 6,517 meters. 

 The following table shows that in fact this is perfectly allowable : 



Altitudes (in thousands of English feet) ... 1 4 



Observed^ 0.87 .64 



Computed jp ( C increasing) 0. 88 .64 



Computed 2> (C constant) 0.90 .65 



Whence it appears that the olserved values of vapor-tension at different 

 altitudes are represented with almost perfect accuracy hj the formula 



—h —h 



p = po lO'isn = 2^0 e 2^*^o. 



We must for the present leave undecided the question whether the 

 constant 6517 varies with the temperature, as may be suspected from 

 the slower diminution of the vapor-tension among the Himalayas, since 

 the observations are too few to determine such a coefficient with any 

 security. 



From the expression jnst deduced empirically for the diminution of 

 vapor-tension with altitude, as compared with the formula previously 

 given, which would obtain for an independent atmosphere of aqueous 

 vapor, we conclude that the ratio of the weight of the aqueous vapor 

 actually present in the atmosphere, as compared with the weight result- 

 ing from the Dalton hypothesis, is as the ratio of the values of G in both 

 formulae, or as 2830 to 12829, or as 0.22 to 1. Whence, the iceight of the 

 vapor present in then-hole atmosphere above the place of observation, according 

 to Dalton^s hypothesis, is 4.5 times greater than it is in reality. 



8 12 16 20 24 28 



.42 .27 .18 .13 — — 



.42 .28 .19 .13 .09 .06 



,42 .27 .18 .12 .08 .05 



