NO. 19 THE SILVER DISK PYRHELIOMETER — ABBOT 9 



Several questions have arisen regarding the accuracy of the silver 

 disk pyrheliometer : 



1. As to the effect of variations of the light of the sky, it might 

 seem that since the pyrheliometer is exposed to about 80° of solid 

 angle, of which the sun occupies only about o.°2, the sky light might 

 be quite considerable. To test this question a screen which limited 

 the solid angle to 5° was fixed to one instrument, and another in- 

 strument with the usual arrangements was compared with it at 

 Washington. No alteration of the relative readings due to the use 

 of the screen could be found on a very clear day. On another day, 

 less clear, a change of relative readings of about 0.5 per cent was 

 found. On a very poor day the effect may reach i or even 2 per 

 cent. On Mount Wilson the sky is so clear that its effect would be 

 negligible. 



2. As regards the inclination of the instrument, experiments were 

 made by affixing a mirror to one pyrheliometer, so that the sun at 

 about 45° from the zenith could be observed in two positions, verti- 

 cal and horizontal, of the instrument, without shifting the mirror 

 with respect to the pyrheliometer. Thus equipped, the pyrhelio- 

 meter was compared with a second, read in the usual manner. No 

 alteration of the relative readings could be found depending on the 

 inclination. 



3. As regards atmospheric pressure, a metal box was fitted to 

 one pyrheliometer and a glass plate fixed over the tube for admit- 

 ting sun rays, so that the air could be exhausted from the interior 

 of the pyrheliometer. Comparisons were made at atmospheric 

 pressure, and at about 1/760 of atmospheric pressure, with another 

 instrument read as usual. No changes of the relative readings de- 

 pending on air pressure could be discerned. 



4. As regards age of the blackened surface of the silver disk, we 

 have not very long-continued records. But a copper disk instru- 

 ment has been employed on Mount Wilson since the spring of 1906, 

 and has been many times compared with other copper disk instru- 

 m.ents there. It has been several times cleaned and re-blackened. 

 There is no evidence that changes as great as i per cent have ever 

 occurred due to defects in the blackening. The instrument appears 

 now to have the same relative readings compared with another in- 

 strument at Washington which has been unchanged and seldom 

 used since 1906, that subsisted between them five years ago. An- 

 other copper disk instrument was loaned in 1907 to the Weather 

 Bureau. Its blackening has been unchanged, and it now (1911) 



