312 ABBOT, FOWLE, ALDRICH : VARIATION OF THE SUN 



We were somewhat unlucky in oui' expeditions. In 1911 a box 

 containing the bolometer and other necessary parts was delayed 

 one month in reaching Algeria, so that a long period of good 

 weather in August was lost. Also the months of September, 

 October, and November, 1911, proved less favorable than usual 

 at Mount Wilson and less favorable than had been hoped at 

 Bassour. Thus the number of days in 1911 in which good obser- 

 vations were secured in both places was rather small. In the 

 year 1912, although the sky was generally cloudless, the eruption 

 of the volcano of Mount Katmai in Alaska of June 6 and 7 so 

 filled the sky with haze, both at Mount Wilson and at Bassour, 

 that a great many days of July and August were rendered unsuit- 

 able for comparison between the two stations. Thus it occurred 

 that of 75 days in which observations were secured at both sta- 

 tions in the years 1911 and 1912, only 48 were found good enough 

 for satisfactory comparisons of the solar constant values obtained. 



For the purpose in view namely, to show whether the apparent 

 fluctuation of solar radiation is due to something outside the 

 earth, it is immaterial whether the days of observat'on are con- 

 secutive or not. It is only required to know whether, if high 

 values are found at Bassour, high values will occur on the same 

 day at Mount Wilson, and, if low values are found at Bassour, 

 low values will be found on Mount Wilson. It matters not whether 

 the days in question be found in one year or another, provided 

 that they be numerous enough to exclude the probability that 

 an agreement, if obtained, is owing wholly to chance. 



The accompanying illustration gives the results of all the days 

 found suitable for comparison between Bassour and Mount Wil- 

 son. Ordinates are solar constant values as obtained at Mount 

 Wilson, abscissae are solar constant values as obtained at Bassour. 

 Circles represent the results of days of the year 1912, and crosses 

 represent the results of days of the year 1911. If the solar radia- 

 tion had varied, and all determinations of it had been free from 

 error, the points must all have lain upon the straight line inclined 

 at 45 degrees to the axis. As it is impossible that results shall 

 be entirely free from error, we must expect that the points repre- 

 senting individual days will be well represented by the 45 degree 



