abbot] 



STUDIES OF SOLAR CONSTANT OF RADIATION 



8i 



peated comparisons of the two instruments and by comparisons of 

 the pyrheliometer with another type that they give proportional 

 results under widely differing conditions of wind and temperature, 

 so that I have no question of the relative accuracy of the actinometric 

 data employed in computing values of the solar constant within two 

 percent. There is, on the other hand, room for question as to the 

 absolute magnitudes of the values given, for these depend on the 

 constants and the theory of the mercury pyrheliometer. Steps are 

 being taken to get. further checks on this matter, and in a later pub- 

 lication it is expected to recompute the data in accord with later 

 information. For the present then, the values in the following table 

 are to be held as relatively accurate and consistent among them- 

 selves, but subject later to correction by a common multiplying 

 factor. 



Table II. 

 Values of the Solar Constant of Radiation. From Bolographic Studies. 



The holographs used in the computations extend from wave- 

 length 0.37/i to wave-length 2.5/y. with the exception of those of 

 October, 1902, which reached only to a wave-length of 0.48// in the 

 violet. For these latter holographs a correction of about twelve per- 

 cent was applied, founded on the later work, and thus the results for 

 October, 1902, are entitled to slightly less weight on this account. 

 All the areas have been extrapolated for the radiations lying outside 

 at both ends of the region o.yj n to 2.5 /i, but the corrections so applied 

 amount to less than one percent altogether. Their magnitude was 



