NO. 5 PYRHELIOMETRV ALDRICII AND ABBOT 3 



1. An eyepiece of improved design was used to read the silver- 

 disk pyrheliometer. 



2. The rate of the seconds pendulum was carefully adjusted to 

 indicate exact i -second intervals. 



3. A high-sensitivity D'Arsonval galvanometer was used. The 

 total deflection for uncompensated solar heating was 44 cm. as com- 

 pared with 10 cm. in the previous work. 



4. All current measurements were made with a direct-reading 

 potentiometer, using a 3-ohm standard resistance and a certified 

 standard cell. Currents with this arrangement were read tcj i part 

 in 5,000. 



Comparisons were made on 2 days, August 26 and 27. Excellent 

 skies prevailed on both days. In all the comparisons, C. G. Abbot 

 read the silver-disk pyrheliometer and operated the shutter of the 

 standard pyrheliometer. L. B. Aldrich made the galvanometer and 

 current measurements. 



Two silver-disk instruments, S.I. Xo. 5bi8 and S.I. No. 79 were 

 used. They had been carried by hand, one by each of the authors, 

 from Washington, D. C, to Mount Wilson. The adopted constant 

 of S.I. No. 5bi8 (Smithsonian scale of 1913), as stated in our previ- 

 ous papers, is .3715. That of S.I. No. 79, as determined from 32 

 comparisons against substandard A, P.O. No. Sbis in November and 

 December, 1946, is .3736. 



The results of our comparisons are summarized in table i. With 

 S.I. No. 5bi8, the mean of 18 comparisons against standard No. 5 

 gives .3626 as the constant of S.I. No, 5bis. Thus the ratio of 

 Smithsonian revised scale of 191 3 to the scale of standard No. 5 is 



•■J^ ^ = 1.0245. The mean of 15 comparisons between S.I. No. 79 



and Standard No. 5 gives .3650 as the constant of S.I. No. 79, 

 and the ratio of the scale of 1913 to that of Standard No. 5 is 



•3/3 =1.0235. It is interesting to note that the average deviation 



of individual comparisons is only one-half of one percent, and the 

 probable error of the means one-tenth of one percent. 



The mean ratio for all 33 comparisons is 1.0240. 



In 1932, 37 comparisons gave a mean ratio of 1.0248. In 1934, 

 42 values gave 1.0237. Thus the 1932, 1934, and 1947 means agree 

 within I part in 1,000. We conclude that the scale of Smithsonian 

 revised pyrheliometry of 191 3 is very nearly 2.4 percent too high. 

 Our silver-disk instruments have remained unchanged. 



