NO. 2^ CORRECTIONS TO SILVER-DISK I'^R 11 ELIO.M E TRY Al'.mrr 3 



2. We have now two first-rate solar stations, Montezuma and St. 

 Katherine, which he in o]:)posite hemispheres. Summer at the one 

 comes during the winter at the other. ^\y colleague. L. B. Aldrich, 

 has collected for me groups of solar-constant values of two kinds 

 (table I ) . In the one kind, group A, it was winter at Montezuma and 

 summer at St. Katherine. In the other kind, group B, it was summer 

 at Montezuma and winter at St. Katherine. We shall now see what 

 the actual mean values of the solar constant were found to be at the 

 two stations and what they would have been found to be if the silver- 

 disk pyrheliometer correction factors had been omitted. The solar- 

 constant values given are all to be understood as prefixed with 1.9. 

 Thus for 42 read 1.942. They are not absolutely final values, as we 

 are still engaged in removing small sources of error before final 

 publication, but they are the best we now have. 



From these typical good days it is seen that with an average temper- 

 ature difference between group A and group B of — 18°7, the actual 

 mean difiference of solar-constant values was —0.004 calorie, or 2/10 

 percent. Had the pyrheliometer corrections been omitted it would 

 have been +0.037 caloric, or 2 ])ercent. a discrei^ancy 10 times as 

 large. 



3. The correction factors K and K' to the silver-disk pyrhelionietry 

 were not guessed at. As regards K. they were the results of observa- 

 tion, as regards K', of well-accepted principles of exact thermometry. 



As for the factor K : 



In the years 1910 and 191 1, L. B. Aldrich and the writer compared 

 silver-disk pyrheliometers S. I. i, S. I. 2. S. I. 3, S. I. 4, A. P. O. 8. 

 and A. P. O. IV on several different days with temperatures ranging 

 from 11° C. to 43° 5 C. The instruments were repeatedly interchanged 

 as regards temperature. I will omit details, as I shall give below a 

 later comparison made in the same way. 



A general summary of all of these experiments of the years 19x0 

 and 191 1 is as follows : 



Tai'.i.k 2 



Coniiiai-isons of .,. .Vssumed 



pyrheliometers \niiilier \ alue ot K weight 



S. I. I with A. P. (). S 



S. I. 2 '• 



S. I. 3 '• •• ■• TO -0.00104 3» 



S.I. 4 ■ 



A. P. O. IV with A. P. O. 8 



1/2' 



" The higher weight given .S. I. .? with .\. P. O. 8, and the lower weight given S. I. 4 witl 

 A. P. O. 8, depended iirincipally on the wide temperature differences with the former pair 

 and the narrow temperature differences which happened to have prevailed with the latter pair 



