EEPOET OF THE SECRETARY. 75 



and large enough to produce considerable changes of the earth's mean 

 temperature. 



During the past year the work of the Observatory has been chiefly directed 

 toward testing the supposed variability of the sun and increasing our knowl- 

 edge of it. As I have said in my last report, this investigation has three main 

 branches as follows : 



First. The determination of the intensity and variation of the total solar 

 radiation reaching the outer limit of our atmosphere. Second, the examination 

 of the distribution of solar radiation over the sun"s disk for the purpose of 

 detecting changes of absorption in the solar envelope. Third, the reduction 

 of temperature measurements from numerous meteorological stations to note 

 departures of the temperature of the earth from its mean. 



General view of the results thus far ohtained since ■January. 1902. 



Before reporting fully this year's work I give in the accompanying chart, 

 Plate V, a general view of the results of the last two and a half years 

 touching this question. Three series of observations are represented in the 

 chart, namely : 



A. Average departures from mean temperature for 80 stations distributed 

 over the North Temperate Zone, represented l)y the full line at the top of the 

 diagram. 



B. The transmission of the solar envelope for radiations of wave length 

 0.50/u, a (green), iis computed from spectrobolometric observations of the sun's 

 disk at Washington, represented by the line at the middle of the diagram. 



C. The solar radiation outside the earth's atmosphere as computed from 

 spectrobolometric observations at Washington, represented by the lower line of 

 the diagram. 



As the observations of type B and earlier ones of type C are frequently 

 separated by long intervals of time, the lines connecting the points should not 

 be interpreted as necessarily indicating the intervening values of the quantity 

 observed. 



Observations of the solar constant taken in Washington are seldom of a satis- 

 factory character, owing to the scarcity of days when uniform transparency of 

 the atmosphere persists for the several hours required for the determinations. 

 Accordingly only a few of the observations of type C are entitled to great 

 weight, and these are designated in the diagram by the letter G to denote it. 

 Observations of good character^ but less satisfactory than these first, are desig- 

 nated by M, denoting medium weight. A still less perfect class, to which, un- 

 fortunately, most of the observations of 1904 and 1905 belong, is designated by 

 the letter L, signifying of little weight. Some observations still less satisfactory 

 than these I have omitted from the diagram as deserving no weight at all. 



Considering now the diagram it will be seen that early in the year 1903 thei-e 

 was a long-continued period of abnormally high terrestrial temperature, and 

 this was followed by a still longer period of abnormally low temperature. Cor- 

 responding well with this state of affairs are the observations of total solar 

 radiation, and these observations are fortunately almost all of great, or at least 

 medium, weight. In September, 1903, a single observation of the transmission 

 of the solar envelope was made, and this, compared with others of more recent 

 date, is low, indicating that diminished transparency of the solar envelope was 

 the probable reason for the small amount of i-adiation and low temperature 

 observed at the same time. 



In the year 1904 there was no long-continued period of abnormal temperature 



a Results of a similar trend are obtained for all wave lengths, and this wave 

 length is chosen merely in illustration. 



