32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



by a diaphragm to about 1 mm. from the edges. On this part of its 

 length the black strip will be heated but will not radiate, and the edges 

 will therefore be slightly above the temperature of the surroundings. 

 As I have made a detailed study of these edge-effects in the case of the 

 pyrheliometer-, 1 where I found that they affected the result only to 

 about 1 per cent, I will not dwell upon them here. In the case of the 

 pyrgeometer, the influence will result only in an unsteadiness of the 

 zero, due to convection currents. The two mentioned effects will 

 probably affect the result to not more than about ±2 per cent, even 

 under unfavorable conditions. 



Much larger are the accidental errors in the measurements of 

 the humidity. The ventilated psychrometer, used in these measure- 

 ments, has been subjected to several investigations and critical dis- 

 cussions and it is therefore unnecessary to go into details. It will 

 be enough to state that the results are probably correct to within 

 5 per cent for temperatures above zero, and to within about 10 per 

 cent for temperatures below o°. 



1 Met. Zeit., 8, 1914, p. 369. 



