1886.] S. A. Hill — Solar Thermometer Observations at Allahahad. 317 



the sun-spot period was suggested. Last year also, I published a paper 

 ill this Journa],* in which a very similar conclusion was arrived at, by 

 a totally distinct method, from the observations made on certain days 

 at Lucknow. In the first paper, the monthly mean excess of the solar 

 thermometer on clear days above the air temperature at noon was com- 

 pared with the barometric pressure and the tension of vapour, as well as 

 with a number representing a proportionate estimate of the amount of dust 

 and haze in the month, and then, by combining the observations of the 

 several months by means of Pouillet's formula (which makes allowancQ 

 for the varying thickness of the atmosphere traversed by rays from the 

 sun at different altitudes), transmission co- efficients for the various 

 constituents of the atmosphere were computed. These co- efficients then 

 served to correct the original observations to what would have been 

 their probable value if there had been no atmospheric absorption, and it 

 was in the annual means of such corrected observations that the varia- 

 tion in the sun-spot period appeared most distinctly. 



In discussing the Lucknow observations, those taken with a non- 

 registering thermometer at different hours of the same day were com- 

 pared, and the " solar constant " was computed by Pouillet's formula. 

 The resulting mean values, though founded on at most only four days* 

 observations in each month, gave a very similar variation to that deduced 

 from the Allahabad results, the maximum falling in 1878, which was 

 the minimum year of the last sun-spot cycle. 



The solar thermometer brought into use at Allahabad in 1876, a 

 large one by Negretti and Zambra, continued to be used under nearly 

 identical conditions until this year, when it was removed to the new 

 observatory. It had been my intention to keep the exposure of the in- 

 strument as nearly constant as possible for ten or eleven years, when if any 

 appearance of a cyclical variation were found in the observations, it 

 might with much probability be assigned to real changes occurring in 

 the sun. During my absence from India in 1883-84, however, the 

 position of the instrument was changed, and the readings of it were 

 permanently increased (apparently by reflection from a small mango 

 tree) to the extent of about 2 degrees as empirically determined. In 

 Table I., which gives a summary of the observations for ten years, the 

 vitiated observations thus empirically corrected are printed in sloping 

 figures. 



* YoL LIY, Part II. 



