210 CARNEGIE INSTITUTION OF WASHINGTON. 



second with its apex at 268° right ascension, +28° decUnation, and a 

 group velocity of 17 km. Mr. Raymond and Mr. Wilson have been 

 associated with the Director in this investigation. 



DIURNAL TERM IN CLOCK-RATE. 



In the last report of the Department, Year Book 1920, attention 

 was called to a diurnal term in the right ascensions determined at 

 Albany, Greenwich, Cape of Good Hope, and Pulkova. Mr. Varnum 

 has continued his investigation of this phenomenon, employing various 

 tests to determine the nature of the disturbing element causing the 

 phenomenon. The fact that clocks of different makes, those running 

 free, and those under barometric and temperature control, exhibit 

 the same phenomenon seemed sufficient to discredit the idea that the 

 irregularities lay in the clock itself. It did seem very reasonable that 

 light passing through miles of an atmosphere which is subject to chang- 

 ing conditions might appreciably affect the observations. The greatest 

 ca.'e was exerted to free the observations from all known sources of 

 error. This was more readily effected because the observations had 

 been planned in a manner to thoroughly discuss any peculiarities which 

 might be outstanding in them. The residual errors of separate series 

 of observations were then tested on the assumption that changing 

 conditions of the atmosphere produced a varied refractional effect. 



It is reasonable to suppose that the heating and cooling effects of 

 the atmosphere due to the position of the sun might produce effects 

 in the atmosphere which would introduce a diurnal term in the ob- 

 servations. In order to inspire confidence such a term should improve 

 the observational data. It was found that the introduction of such 

 a term in the right-ascensions greatly reduced the systematic correc- 

 tions in right-ascension depending upon right-ascension and declina- 

 tion. A like improvement was also noted in the effect east minus 

 west and in upper minus lower. In addition, it was noted that series 

 of observations which appeared to be inharmonious proved to be those 

 series where the refractional effect is most pronounced, so that appar- 

 ently discordant results are harmonized. If the phenomenon is a 

 natural one due to variation in refraction, then it should affect the 

 zenith distances as well as the right-ascensions. 



The work upon this phase of the problem has barely started, but for 

 the few series treated it has yielded interesting results. It has appar- 

 ently demonstrated that the observations taken at any particular 

 time are subject to a systematic shift. WTien the refractional term is 

 introduced, the individual star positions, with very few exceptions, are 

 brought into greater accordance. Considerable work is yet to be 

 undertaken upon the investigation, but it seems eminently worth 

 while, as the results promise to correct the positions of the poles and 

 the equator; thus, to a large extent, observations taken north and south 



