138 PROCEEDINGS OF THE AMERICAN ACADEMY 



being 31°, 100 ohms were added to the resistance of the circuit, and 

 when the sun's rays were thrown directly into the thermograph a 

 deflection of 114.6 scale divisions was obtained. Without the 100 

 ohms of additional resistance, the deflection would have been : 



100 X 114.6 _..,_. 1 V • • 

 • — „ ^..^. — = 04/00 scale divisions. 

 0.209o 



That same evening the moon was observed at an altitude of 30° 15'. 

 the deflection being 124.59 scale divisions. From its augmented semi- 

 diameter, 14' 52".7, and the same constants.as in the previous example, 

 we find the deflection for the uncondensed moonbeam to be 0.2945 

 of a scale division. 



The ratio of 54700 to 0.2945 is 186,300. The results obtained by 

 the two methods are here tabulated. 



Ratio of Sun Radiation to Moon Radiation. 

 First Method. Second Method. 



March 30, 1888 . . . 181,000 October 18, 1888 . . . 187,200 



April 2.5, 1888 .... 180,900 October 21, 1888 . . . 186,300 



October 18, 1888 . . . 167,400 



Mean of the five results . . . . . 184,.560 



The results given by the two methods do not show the wide dis- 

 agreement so noticeable in the comparative light of the two bodies, as 

 found by different methods. This fact is strong evidence in favor of 

 the accuracy of the mean result. Jo reality, the problem of the total 

 comparative radiation of the sun and moon is in many respects simpley 

 than the problem of the comparative light intensity of the two bodies. 

 In the former case, there is freedom from all personal bias ; the meas- 

 urements are all given by the indications of the instruments directly, 

 and, in general, the chances for error seem to be fewer. 



Lord Rosse found the ratio of solar to lunar radiation to be 80,000 

 to 1. Langley obtained the ratio 96,509 to 1. Provided the moon 

 were a flat disk, reflecting perfectly all of the sun's rays that fall upon 

 it, we could not receive more than one 97,000th part of the solar heat 

 from such a disk, as Zolner has shown. The close agreement of his 

 result with this number Langley considers to be largely a matter of 

 chance, or, rather, of constant errors, tending in an unknown degree 

 to increase the observed values. 



It would be a priori quite improbable that we should, under the 

 circumstances of reflection from the lunar surface and subsequent 



