ABBOT: SOLAR RADIATION MEASUREMENTS 227 



son, California. This station is usually occupied from May to 

 November. Its main investigation comprises spectro-bolometric 

 determinations of the so-called "solar constant of radiation" after 

 the general method of Langley. This term designates the amount 

 of heat per square centimeter per minute which would be pro- 

 duced by completely absorbing the sun's radiation outside our 

 atmosphere at the earth's mean solar distance. 



Early results indicated that this quantity is not really a con- 

 stant, but varies over a range of several per cent, both from 

 year to year and in short irregular periods of days or weeks. 

 Confirmation of these results has been secured in many ways, 

 so that now there remains, I think, but one possible explanation 

 of the phenomenon other than that the sun itself varies in its 

 emission by several per cent from time to time. This alterna- 

 tive possibility is that atmospheric changes occur simultaneously 

 over the whole earth which lead to variable erroneous determina- 

 tions of the so-called "solar constant," and that the errors thus 

 produced are nearly equal and introduce apparent variations in 

 the same sense, however far apart the two simultaneously ob- 

 serving stations may be. I believe it is easier to admit that 

 the sun itself is variable as supposed. Other irregularly vari- 

 able stars are numerous. There is no reason why the sun, too, 

 may not be variable. 



It is true that recently Dr. Guthnick of the Berlin-Babels- 

 berg Observatory has made a good many photo-electric measure- 

 ments of the relative brightness of the planets Jupiter and Sat- 

 urn compared to reference stars. His results thus far have not 

 confirmed the variability we have found. But they have not 

 disproved it. His observations have, heretofore, seldom been 

 taken on coincident days with ours. This inconvenience we 

 expect will be remedied for the year 1920. Guthnick's results 

 show a range of several per cent. This may be in part really 

 solar. If no solar variation was suspected one would attribute 

 it to experimental error. If it should prove that Guthnick in 

 the future finds variations in planetary brightness similar to our 

 solar changes but not coincident in time, we must recall that 

 the planets generally lie in different directions. If the solar 



