BEPOET OP THE SECEETAEY. 87 



tion of radiation over the sun's disk. Two improvements were intro- 

 duced in the apparatus. 



The coelostat used in solar constant work was provided with stel- 

 lite mirrors in place of silver on glass, so that now the optical train 

 of the spectrobolometer for solar constant work contains exclusively 

 nontarnishable mirrors. This allows us to compare as never before 

 the distribution of radiation in the solar spectrum from day to day. 

 We hope now to determine surely whether the variations of solar 

 radiations are uniform over the whole spectrum or predominate in 

 certain wave lengths. 



A specially designed vacuum bolometer like the one employed at 

 Hump Mountain and in Chile and wholly sealed in glass so as not 

 ever to require attention to renew the vacuum has been substituted. 

 This bolometer was constructed to exact specifications as to length, 

 breadth, and thickness of strips in accordance with completely 

 worked-out unpublished theory of the bolometer. We are sure that 

 it is the last word on the subject as regards adaptability to our in- 

 vestigation. All expected results are obtained in its actual use. 



The sky was more cloudy than usual on Mount Wilson both in 1917 

 and 1918, during the time of our expeditions. In the winter, in 

 November, December, and January of 1917-18, the Carnegie Institu- 

 tion observers reported unusually good weather for the season. 



Mr. Aldrich carried on in 1917 some investigations to determine 

 whether long- wave rays, not transmissible by rock salt (that is, ex- 

 ceeding 20 microns in wave length) , occur in the solar beam at the 

 earth's surface. The experiments indicated that they do not. He 

 also investigated the transmissibility of the atmosphere for rays of 

 more than 20 microns in wave length by means of the pyranometer 

 with and without a rock-salt cover. The experiments showed that 

 even toward the zenith these rays are wholly cut off by the lower- 

 layers of the atmosphere. Hence we may conclude from both the 

 solar and nocturnal observations that our atmosphere is opaque to 

 rays exceeding 20 microns in wave length, such as are emitted in 

 recognizable quantities by bodies at terrestrial or solar temperatures. 

 This result is in harmony with Mr. Aldrich's laboratory experiments 

 above mentioned. 



In June, 1918, experiments were begun at Mount Wilson to deter- 

 mine the distribution of solar radiation along that diameter of the 

 solar image which is at right angles to the east and west diameter 

 investigated in our usual tower telescope work. A special apparatus 

 was arranged to slowly rotate the second mirror of the coelostat, and 

 thus produce a regular drift of the solar image along any desired 

 diameter. Preliminary results obtained show that the differences, if 

 any, between the distribution of radiation along different solar di- 

 ameters do not amount to 1 per cent. 



