820 S. A. Hill — Solar Thermometer Observations at AUahahad. [No. 4, 



From the monthly means of the excess temperatures and of these 

 elements, as well as the proportionate figures for dust haze, we may- 

 compute the transmission co-efficients of the several constituents of the 

 atmosphere by a modification of Pouillet's formula. For, putting R to 

 stand for the observed radiation (as given in Table I.) and S for the solar 

 " constant," we have R = S. a^' pf^y^', where a and P are the trans- 

 mission co-efficients of air and vapour, each of one inch pressure, y the 

 transmission co-efficient for dust haze of one-tenth of the density ob- 

 served in May, c the atmospheric thickness traversed by the rays at 

 noonday compared with the vertical depth of the atmosphere, and h, /, 

 and d the pressures of air and vapour and the proportion of dust 

 respectively. Remembering that, for the obliquities we have to deal 

 with, e is as nearly as possible equal to the secant of the zenith distance, 

 we may for purposes of calculation put the equations into the form :— 

 log R = log S + & sec. zloga+f sec. z log*/? + d sec. z log y, and 

 the equations for the nine months are more than sufficient to determine 

 all the unknown quantities. 



The observations of Mr. Hennessey at Mussoorie and Dehra in 1869 

 and 1879, however, probably give more accurate values for the trans- 

 mission CO- efficients of air and water vapour than could be determined 

 in this way, and I therefore adopt them now. These values {see P. R. S., 

 No. 219, 1882, page 435) are .— 



a = -998555 

 P = -72783 



Inserting these in the equations for the nine months, we find y = '99449 

 and S = 82°-29. 



The value of y shows that in May, when the dust haze is a maximum 

 and the sun's rays nearly vertical, the absorption by dust does not much 

 exceed 5 per cent. This result is probably only brought about by the 

 circumstance that the reflection from dust particles in all directions round 

 the thermometer bulb is nearly equal to the absorption by those particles 

 which lie in the direct path of the solar beam, bat it suffices to show that 

 any error of moderate extent in estimating the proportionate quantity of 

 dust for a given month or year will have little effect on the computed value 

 of the solar constant. The quantity of water vapour in the air can be 

 determined with a considerable degree of exactness, and, when its 

 pressure at the place of observation is about an inch, it absorbs 27 per 

 cent, of the incident radiation. The figures in Table I. can therefore 

 now be corrected for atmospheric absorption as were those in my first 

 paper. The resulting values of the solar '' constant," corrected also to 

 the earth's mean distance from the sun, are the following*— 



