lO SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 72 



the silvered surface of the rear mh-ror. Thus 3.95 is regarded as 

 more nearly correct for ordinary w^ork with beams w^hich (unlike 

 direct run-rays) subtend large angles. 



The second comparison was made at Mount Harqua Hala, Arizona, 

 by Dr. Abbot, November 10, 1920, using secondary pyrheliometers 

 S. I. No. 32 and A. P. O. No. 9. 



First determination — Melikeron normal to sun's rays. 



Calories by Pyr. No. 32 and No. 9= 1.531 (mean value). 

 (Current)^ of Melikeron = .347 (mean value). 



1. 531 



— T^- = 4.41 = constant of Melikeron No. 2. 



Second determination — Melikeron at 7° angle to sun's rays (cos. 



7° = -992). 



Calories by Pyr. No. 32 and No. 9= 1.538 (mean value). 

 (Current)^ of Melikeron— .378 



^^^^-t||^^^ = 4.04 = constant of Melikeron No. 2. 



This result shows for normally incident rays a considerably greater 

 absorption and scattering by the rear mirror than was the case in 

 the comparison of August 30. From the deteriorated appearance of 

 the silvered mirror on November 10 this was quite to be expected for 

 sun rays, but the deterioration was probably quite negligible for 

 earth rays. The best constant, then, of Melikeron No. 2 without glass 

 hemisphere is the mean of 3.95 and 4.04, or 4.00 which is now the 

 adopted value, applicable for all rays not at strictly normal incidence. 



At Mount Wilson and Mount Harqua Hala numerous comparisons 

 on the night sky were made between Melikeron No. 2 and Pyrgeo- 

 meter No. 22. The two instruments, leveled, were mounted at the 

 same height and within less than 6 inches of each other. Exactly 

 similar bright tin-box shutters were used on both instruments. Using 

 the above adopted constant of Melikeron No. 2, a value of the con- 

 stant of Pyrgeometer No. 22 was obtained each time. The results are 

 summarized in table HI. 



The mean of all, under these varying conditions of air temperature 

 and water-vapor content is 9.72. There is perhaps some evidence in 

 these values that the constant of Pyrgeometer No. 22 is a function of 

 both air temperature and water-vapor content. But further compari- 

 sons under a wider range of air conditions are needed to confirm it. 



To illustrate this indication, values are given in the table computed 

 by the formula : 



Constant = 1 1 .50 — 3. 1 2p — i .47 ( t — 60° ) ^'^. 



