318 !^. A. Hill — Sular Thermometer Observations at AUahahad. [No. 4, 



Table I. — Excess Temperattire of the Sun Thermometer on Clear 

 Bays above the Air Temperature at Noon. 



Year. 



Jan. 



Feb. 



Mar. 



April. 



May. 



June. 



Oct. 



Nov. 



Dec. 



Mean. 



1876 



66-i 



65-7 



63-9 



60-6 



56-5 



57-3 



60-3 



64-6 



64-6 



620 



1877 



630 



66-7 



63-6 



62-4 



60-0 



57-4 



^ 59-8 



630 



63-4 



621 



1878 



65-9 



643 



65-2 



64-5 



62-3 



57-8 



59 6 



64-7 



64-6 



63-2 



1879 



64-2 



63-9 



63-7 



62-6 



61-8 



61-0 



611 



64-9 



63-5 



631 



1880 



630 



63-8 



60-3 



59-9 



62-0 



61-5 



58-3 



62-5 



60-6 



61-3 



1881 



62-3 



60-9 



61-9 



60-4 



60-3 



630 



54-8 



61-8 



59-9 



60-6 



1882 



60-5 



61-5 



60-1 



611 



60-6 



59-1 



54-6 



62-2 



58-2 



59-8 



1883 



60-9 



61-5 



610 



58-4 



57-6 



57-1 



59-2 



59-8 



58-2 



59-3 



1884. 



58-3 



60-5 



59-3 



611 



58-9 



55-9 



64-1 



65-3 



63-5 



60-8 



1885 



63-2 



66-2 



630 



62-2 



610 



58-1 



61-6 



64-3 



63-6 



626 



Mean 



62-7 



63-5 



62-2 



61-3 



60-1 



58-8 



59-4 



63-2 



62-0 



61-5 



In this table, as in those already published, clear days are understood 

 to be those on which the cloud proportion at 10 A. m. and 4 p. M. 

 does not exceed one-fifth of the expanse. Such days are so uncommon 

 in July, August, and September that these three months could not be 

 included. The column of annual means indicates clearly a cyclical 

 variation in the ten years, the maximum occurring in 1878-79 and the 

 minimum falling about the end of 1883 ; hence, if such a variation 

 appears in the computed values of the so-called solar " constant " given 

 below, it cannot be set aside as a merely arithmetical result, since it is 

 founded upon a similar variation in the data obtained by direct obser- 

 vation. The monthly means at the foot show a maximum in February 

 and another in November, with a faintly marked minimum in the cold 

 weather and another more distinctly marked in June. They should, 

 however, be multiplied by the square of the earth's mean radius vector 

 for each month to correct for the varying distance of the sun, before they 

 can be properly compared, a point which was overlooked in my paper 

 of 1883. When thus corrected they stand as follows : — 



Jan. Feb. March. April. May. June. Oct. Nov. Dec. 



60-7° 62-0° 61-7° 61-8° 61-5° 60-7° 58-9° 61-8° 60-0° 



If the absorptive power of the atmosphere were constant, we should 

 expect to find the maxim-um result in June, when the sun rises highest 

 in the heavens, and the minimum in December, when the incidence is 

 most oblique and the thickness of the atmosphere to be traversed by 

 the rays is greatest. The actual variation which gives maxima in 

 February or April and in November and minima in December and October 

 must be due to the modifying influence of a variability of absorption. 



