6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 95 



The reader will notice that l)y the method of reduction adopted, 

 whereby for every determination the same instrument is read both 

 warm and cool by the same observer, and the results thus obtained 

 are subtracted, personal equation is eliminated and it is not necessary 

 to know the reduction factors applicable to the instruments comparecl. 

 The result K = 0.00109 ^-^ identical with that found by Abbot and 

 Aldrich 25 years ago. and indicates no desirable correction to the 

 published preferred value, K = 0.0011. 



5. Although our coefficient as heretofore published appears to be 

 applicable to the measurements of the 1 1 silver-disk pyrheliometers 

 referred to in this paper, some one may raise the question as to 

 whether it is applicable to the instrument S. I. 12 at Potsdam which 

 was employed by Fuessner. Inasmuch as all of the pyrheliometers 

 we have prepared and sent out from the Smithsonian Institution are 

 made of similar materials, and with their receivers of identical sizes, 

 it is highly improbable that the Potsdam pyrheliometer is exceptional. 



But there is a further check on this question. In October 193 1 the 

 writer carried silver-disk pyrheliometer S. I. 5bis to Potsdam, and at 

 2 separate times involving a considerable range of temperatures, 

 comparisons were made between it and the Potsdam instrument S. 1. 

 12. The results were uncommonly harmonious, as both instruments 

 were reduced with the usual value of the temperature coefficient K. 

 Hence we may believe that as concerns temperature these two instru- 

 ments, S. I. 5i,is and S. I. 12. are alike. But in 1932 and 1934, silver- 

 disk pyrheliometer S. I. 5bis was carefully compared with the new 

 water-flow double-chamber standard pyrheliometer No. 5 on Moimt 

 Wilson. These comparisons involved a range of temperature from 

 26° to 46° C. We were unable to detect any differences in results 

 showing the influence of this large range in temperatures. Had the 

 coefficient K been omitted, as recommended by Fuessner, a progressive 

 range of 2.2 percent would have appeared. Hence we may be sure 

 that S. I. 5bis is no exception in temperature behavior to the 11 others 

 referred to in this paper, and the Potsdam comparison indicates that 

 the instrument S. I. 12 used in Fuessner's investigations behaved 

 similarly. 



Conclusion. — I do not think it devolves on me to suggest what con- 

 ditions of instruments or observation may have led Fuessner to the 

 false conclusion he has published that the temperature coefficients 

 should be omitted from reductions of measurements with the silver- 

 disk pyrheliometers. I feel sure that if he had reflected that such a 

 change would destroy all of the conclusions on solar variability and 

 its consequen.ces which have come from the work of the Astrojibysical 



