74 ANNUAL EEPORT SMITHSONIAN INSTITUTION, 1910. 



the possibility of solar variability wliich was the result to be determined by the 

 expedition. This inquiry has, indeed, been the primary one in all the subse- 

 quent work, but not to the exclusion of attempts to fix the value of the solar 

 constant itself. 



There were at that time two principal and seemingly formidable difficulties 

 hindering the determination of the solar constant of radiation. First, there 

 was no instrument capable of absorbing fully and adapted for measuring com- 

 pletely the energy received at the earth's surface, excepting, perhaps, the little- 

 known and rarely used instrument invented by W. A. Michelsou, of Russia, about 

 1S94. Second, there was grave doubt if a true estimate of the loss of radiation 

 in traversing the air could be made. Langley has somewhere described the first 

 obstacle as " formidable," the second as " perhaps insurmountable." 



As stated in previous reports, much attention was given from 1903 onward to 

 devising a standard pyrheliometer, and thus establishing the absolute scale of 

 radiation measurements. A considerable degree of success seemed to be attained 

 in 190G, but the results obtained in that year were found, by comparison with 

 instruments of the United States Weather Bureau, to differ so much from the 

 generally adopted scale of Angstrom that further work, involving finally the 

 construction of two additional water-flow pyrheliometers, was done. The last 

 of these instruments, and by far the most perfect of them all, was completed 

 and tried at Mount Wilson in October, 1909. A fairly close agreement seemed 

 to hold between it and its immediate predecessor, but when the electrical con- 

 stants of both instruments were determined with extreme care in Februai-y, at 

 AVashington, by Mr. Aldrich, the gap widened. A source of error, till then 

 little regarded, was reconsidered, and painstaking comparisons of pyrheliometers 

 were carried through at Washington by Messrs. Aldrich, Abbot, and Fowle. 

 These were finished in June, 1910, and the two standard pyrheliometers were 

 found to agree together well within the probable error of the highly accurate 

 experiments. Not only so, but each instrument was found to take up and 

 measure between 99 and 100 per cent of such various quantities of electrically 

 introduced heat as were used as tests. Finally these definite measurements 

 indicated that while the results published at page 46 in volume 2 of the Annals, 

 made with standard pyrheliometer No. 1, are 4 or 5 per cent above the true 

 scale, yet when all the experiments made with that instrument, at Washington 

 as well as Mount Wilson, are collected their mean result is almost in exact 

 agreement with the results obtained in 1910 with standard pyrheliometers 

 Nos. 2 and 3. 



It may now be accepted that the absolute scale of radiation is established 

 within three parts in 1,000, and that we may express all our measurements of 

 solar radiation made since 1902 with this degree of accuracy in absolute calories 

 per square centimeter per minute. 



Three secondary pyrheliometers, the cost of whose construction after my 

 designs has been defrayed from the Hodgkins Fund, have been standardized and 

 sent to Russia, France, and Italy. Two others have been sold by the Institution 

 to the United States Agricultural Department. Thus steps are being taken to 

 diffuse the standard scale of pyrheliometry. The new scale is about 5.2 per cent 

 above that of new Angstrom pyrheliometers. 



The second obstacle mentioned above seems now less serious than the first. 

 It was found in 1905 and 1906 that practically identical values of the solar 

 constant resulted from good series of spectro-bolometric observations of the 

 same day taken at Washington (sea level) and Mount Wilson (6,000 feet eleva- 

 tion). But in August, 1909, Mr. Abbot ascended Mount Whitney (14,500 feet) 

 with a complete spectro-bolometric outfit, and, notwithstanding many days of 



