112 ANNUAL KEPOET SMITHSONIAN INSTITUTION, 19 2 8 



material and personnel has occupied much time of the director and 

 staff in Washington. 



Several years having gone by since the station at Table Mountain 

 began its regular work, enough data had accumulated to justify a 

 statistical study over the whole period, to detect any systematic 

 errors. Minute systematic errors in the uncorrected results are in- 

 evitable. We are attempting to determine the intensity of the sun's 

 energy not only as it is received at the observatory but also as it was 

 in free space outside the atmosphere. Humidity and dust produce 

 effects which it is impossible to ascertain precisely on any given indi- 

 vidual day by any method. Hence only by comparing the average 

 run of the results over a term of years with the average run of atmos- 

 pheric conditions during the same interval can these not quite negli- 

 gible residual systematic errors be determined and allowed for. 

 Such a study of the Table Mountain work has been in progress. 

 When completed there were revealed certain discordances between 

 Table Mountain and Montezuma which, though small, demanded still 

 further study. 



As so often has happened in the history of science, this study by 

 my colleague, Mr. Fowle, of a perplexing discordance has brought a 

 new discovery of some importance. It is that the ozone existing in 

 the atmosphere at a level of 30 to 50 kilometers (18 to 30 miles), and 

 which is formed from the atmospheric oxygen by the action of 

 invisible ultra-violet sun rays, is variable in amount over Table 

 Mountain, though nearl}' constant in amount over Montezuma. The 

 discrepancy in the final results of radiation work between the two 

 stations appears to be due mainly to this variability of atmospheric 

 ozone at Table Mountain. Regular observations of ozone are now 

 in progress there in cooperation with Doctor Dobson, of Oxford, 

 England. 



The tedious but necessary computations and statistical comparisons 

 involved in the work of systematizing and correcting the preliminary 

 results of the observations, onlj^ part of which is indicated in the 

 discussion above, have employed Mr. Fowle and two computers con- 

 tinually during the year. 



(&) Appai'atus — Under the direction of the writer and his col- 

 league, Mr. Aldrich, the instrument maker, Mr. Kramer, has con- 

 tinued to make apparatus for radiation investigations. One instru- 

 ment upon which much attention has been lavished is a new form of 

 pyrheliometer to measure more accurately and conveniently the sun's 

 radiation. So accurate and stable is the silver-disk pyrheliometer 

 which we have employed for nearly 20 years, and of which over 50 

 copies have been furnished by the Smithsonian at cost to other insti- 

 tutions at home and abroad, that it is hard to prepare a new instru- 

 ment superior to it. Yet there are two or three slight sources of 



