in the Air upon the Temperature of the Ground. 241 



And it seems as if periodic alterations with the time of 

 observation occurred in his series. On what circumstance 

 these alterations with the time depend one can only make 

 vague conjectures : probably the clearness of the sky may 

 have altered within a long period of observation, although this 

 could not be detected by the eye. In order to eliminate this 

 irregular variation, I have divided the observations into four 

 groups, for which the mean quantities of carbonic acid (K) 

 and of water- vapour (W) were 1*21 and 0*36, 2 # 21 and 0'86, 

 1-33 and 1-18, and 2*22 and 234 respectively. With the 

 help of the mean values of the heat-radiation for every group 

 of rays in these four groups of observations, I have roughly 

 calculated the absorption coefficients (x and y) for both gases, 

 and by means of these reduced the value for each observation 

 to the value that it would have possessed if K and W had 

 been 1*5 and 0*88 respectively. The 21 values for the 

 different rays were then summed up, so that I obtained the 

 total heat- radiation for every series of observations, reduced to 

 K = 1'5 and W = 0*88. If the materials of observation were 

 very regular, the figures for this total radiation should not 

 differ very much from one another. In fact, one sees that 

 observations that are made at nearly the same time give also 

 nearly equal values, but if the observations were made at very 

 different times, the values differ also generally very much. 

 For the following periods I have found the corresponding 

 mean values of the total radiation : — 



Period. 

 1885. Feb. 21-June 24 



1885. July 29-1886. Feb. 16. 



1886. Sept. 13-Sept. 18 



1886. Oct. 11-Nov. 8 



1887. Jan. 8-Feb. 9 



In order to reduce the figures of Langley to comparability 

 with one another, I have applied the reduction factors 

 tabulated above to the observations made in the respective 

 periods. I have convinced myself that by this mode of 

 working no systematic error is introduced into the following 

 calculations. 



After this had been done, I rearranged the figures of 

 Langley 's groups according to the values of K and W 

 in the following table. (For further details see my original 

 memoir,) 



Mean 



Reduction 



value. 



factor. 



4850 



1-3 



6344 



1-00 



2748 



2-31 



5535 



1-15 



3725 



1-70 



