82 REPOKT OF THE SECRETAEY. 



to know in advance how to design and con.sti'nct a l)olonieter to give the best 

 resnlt under stated conditions. 



In accordance with the results thus reached, a vacuum bolometer was con- 

 structed for the observation of stellar heat which is several times as sensi- 

 tive for the purpose as the best bolometer hitherto prepared here. 



Cwlostat. — A new coelostat with two 15-inch mirrors was obtained for the 

 proposed expedition to a high-altitude station. The general design was pre- 

 pared at the Observatory, and the construction was by Mr. M. E. Kahler, of 

 Washington, excepting some work done on the clock at the Observatory shop. 

 A photograph of the apparatus as in use at Moinit ^^'ilson is here shown 

 (Plate YII). 



Several unusual features may be noted. The rotating mirror is shown 

 mounted on a carriage which may travel on ways either east and west or 

 north and south, so as to be adjusted for the position of the sun at any 

 time in the year. The carriage is of the same height as the lower base of 

 the support of the second mirror, and the two mirrors may be interchanged 

 if desired, so that the rotating one can be fixed in one place and the east- 

 and-west and north-and-south motions can be made with the other. The 

 driving clock is fastened to the support of the polar axis and designed to run 

 equally well in any position, so that the adjustment of the axis for different 

 latitudes can be made without altering the driving mechanism. In order to 

 conform to the motion of the carriage, the clock is driven by springs instead of 

 weights. At your suggestion there was introduced a driving spring, which 

 is itself kept wound to a nearly constant tension by two larger springs. Thus 

 the driving force is luiiforin though the two larger springs run down. The rate 

 of their unwinding is governed by an escapement driven from the governor 

 train. The governor itself is of the centrifugal type, but has springs instead 

 of gravity as the governing force, and will thus run in any ijosition. The 

 accuracy of the clock is ample, so that the reflected beam from the cielostat 

 frequently remains constant -in direction within one minute of arc for a half 

 hour. 



Sfantldrd pi/rJiclionicter. — Mention was made in my last year's report of a 

 new form of pyrheliometer then under construction. This instrument has 

 been completed and forms a part of the equipment of the INIount Wilson expe- 

 dition. It appears to justify all the hopes that had been connected with it. 

 As stated last year, it receives solar rays in a hollow blackened chamber of the 

 shape of a test tube, from wiiich little radiation can escape by reflection, or 

 heat by convection, owing to the deei)ness of the chamber and to its numerous 

 blackened diaphragms. A current of water circulates around this chamber 

 and takes up the heat absorbed on its walls. Platinum resistance wires serve 

 to determine the temperature of the water before and after its passage around 

 the chamber. Thus the sun's rays entering a known aperture produce a 

 measurable rise of teujperature in a known amount of water. A certain check 

 is had on the accuracy of the measurement, for a coil of wire is introduced 

 within the rear end of the chamber, and in this coil heat may be produced 

 electrically at a known rate. This heat warms the air and indirectly reaches 

 the walls of the chamber, and may l>e measured as if it were produced by 

 radiation. 



At nine recent trials made on several different days, with the instrument in 

 different positions and with different rates of flow of the water current, the 

 heat " found " ran between 97 per cent and 103 per cent of the heat intro- 

 duced. The mean of nine trials gave 100.4 per cent " found." If, then, the 

 electrically supplied heat is thus closely measurable, much more should that 



