4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 95 



As for the factor K': 



When exposure to the sun drives mercury into the capillary of the 

 thermometer, the length of the column of newly appearing mercury 

 depends on the temperature prevailing in the thermometer stem. The 

 coefficient of linear expansion of mercury in a capillary of the glass 

 used in the thermometer may be taken as 0.00014 per degree Centi- 

 grade. The rise of temperature of the pyrheliometer is therefore 

 corrected to what it would have been at a stem temperature of 

 20° C, namely, by the corrected temperature rise multiplied by 

 0.00014(20° — T). 



4. The indirect and direct evidence I have thus far brought forward 

 seems to me more weighty than that which moved Fuessner to his 

 conclusion. For his observations involved the peculiarities of several 

 other types of pyrheliometers and were made in a high latitude at 

 rather low sun. Yet since, if sustained, Fuessner's conclusion invali- 

 dates nearly the whole program of the Smithsonian Astrophysical 

 Observatory for the past 25 years, I have felt it to be necessary to 

 redetermine the factor K from new observations. For this purpose 

 I directed Messrs. Zodtner and Greeley, both highly skilled observers 

 of long experience, now operating our station at Table Mountain, 

 California, to make numerous direct comparisons of silver-disk pyr- 

 heliometers A. P. O. 10 and A. P. O. 12, using pendulum beats to 

 fix the times of observing. Their program is as follows : 



Observing in a warm room with opened window with one pyr- 

 heliometer, and immediately outside in winter's cold with the other, a 

 series of absolutely simultaneous comparisons is made. Time signals 

 for both observers were given by a pendulum. Then, without remov- 

 ing the instruments, they are allowed to return to nearly the temper- 

 atures of their respective surroundings, and a second comparison is 

 made with exchanged observers. Then, on a second day, or later in 

 the same day, with the instruments exchanged, so that that which was 

 warm has now become cold, and that which was cold has now become 

 warm, the same program is repeated. In this way the instruments and 

 the observers were exchanged again and again, until satisfactory 

 results had been obtained. These observations were made in good 

 sky conditions with fairly high sun at the latitude 34° North. 



As the matter is of importance to support the accuracy of the last 

 25 years of work of the Astrophysical Observatory, I give the results 

 in some detail. In the following table the readings marked R represent 

 the corrected rise of temperature of the pyrheliometer after proper 

 corrections are applied for irregularity of the bore of the stem and 

 for reducing the temperature of the stem to 20° C. 



