﻿fowle] solar radiation measures 401 



Actinometric measures, indeed, in common with the bolometric 

 process, depend on the estimation of the transmission of the earth's 

 atmosphere and are at a great disadvantage as compared with the 

 process of observation of the absorption of the solar envelope orig- 

 inally devised by Mr. Langle) where, as seems very possible from 

 the solar studies now being carried on here, the changes in the effec- 

 tive emission of the sun itself are determined. In the latter process 

 less than three minutes of constant conditions in our atmosphere 

 are required to make an observation of the apparent absorption of 

 the solar envelope. The actinometric and holographic processes, on 

 the other hand, demand two or three hour's constant condition of 

 atmospheric transmissibility, which occurs comparatively seldom in 

 such a locality as Washington. 



The observations discussed in this paper were made with two 

 aetinometers. One was of the Crova pattern, a glass and mercury 

 thermometer, the blackened bulb of which was situated in the centei 

 of a metallic sphere, polished nickel without, blackened within, and 

 having an opening allowing a beam of sunlight of slightly less 

 diameter than the bulb to fall on the latter. The other actinometer 

 was home-made and consisted of a thin flat cylindrical copper recep- 

 tacle filled with mercury, blackened over the exposed end with 

 platinum black, and the rest nickeled and surrounded by a wooden 

 sphere, bright within, having the beam similarly limited by dia- 

 phragms. A mercury thermometer inserted in the copper vessel 

 served to indicate the rise in temperature. Observations are made 

 in the same manner with both aetinometers. By cutting the sun- 

 light off from the aperture, a cooling correction is determined for 

 two minutes ; the instrument is then exposed to the solar radiation 

 for two minutes, and a final cooling correction for two minutes com- 

 pletes the observation. The radiation value is obtained from the 

 rate of rise during exposure, corrected by the mean of the cooling 

 corrections determined before and after the exposure to the sun. 

 Values of the solar constant have been computed by Bouguer's 

 formula from actinometric measures alone, to see how far short 

 they fall from the spectro-bolometric determinations from the same 

 data. 



In this discussion of actinometric measures the observations will ac- 

 cordingly first be treated as if they followed the formula E x = E ( ,a 7 "> 

 which is better adapted to the purpose in the linear form 



log E x = log E -\- m x log a 



where log E x and m x are employed as variables. Thus in order to 



