SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 60 



July 28, 1910. It is possible that A. P.O. 9 was inadvertently not 

 fully exposed at this time. We shall adopt : 



A-P^8 = x.0068. A P O 9 = lJ0426m 



A.P.O. 8 bi 



A.P.O. 8 bis 



A.P.O. IV 

 A.P.O. 8 bis 



= 0.7398. 



From Stan^rd2 and AJ>Q 8 we find Standard 2 Xli(X)68=0 ^ 



A.P.O. 8 A.P.O. 8bis A.P.O. 8bis 



From Standard 3 and A J> 0.8 we find Standard 3 =0,3765X1.0068=0.3791. 

 A.P.O. 8 A.P.O. 8bis A.P.O. 8bis 



From Standardj and A^OIV we fi d Standard 3 =Q .5x40x0.7398=0.3809. 

 A.P.O. IV A.P.O. 8bis A.P.O. 8bis 



also 0.5094X0.7398=0:3768. 



From S 4^^4 and A ROo^ find Standard 4 6l8x 6=a3772 , 



A.P.0. 9 A.P.O. 8bis A.P.O. 8bis 



Besides these values we have given two direct comparisons of ---- ^ _ • 



A.P.O. 8bis 



In combining the results to obtain the best value of the constant of 

 Secondary Pyrheliometer A. P. O. 8 b ; s , we have been guided by the 

 view that a completely independent set-up of apparatus is a more 

 weighty condition than is a small probable error. This amounts to 

 saying that we have considered constant errors of more importance 

 than accidental ones. But admitting this, we have also taken some 

 notice of the number of observations made on different occasions, and 

 of their accordance. These considerations have led us to the follow- 

 ing values for Constant of Secondary Pyrheliometer A. P. O. 8bi s : 



The constants of the silver disk pyrheliometers sent out by the 

 Smithsonian Institution to various observers are derived by compari- 

 sons of those silver disk pyrheliometers with one or the other of the 

 pyrheliometers named above. The following table gives the results 



