202 CAKNEGIE INSTITUTION OF WASHINGTON. 



the preliminary reduction, only those series have been used where the 

 clock correction could be determined from groups of stars 12 hours 

 apart and where the clock-rate could be derived from successive 

 transits of the same stars. Thus errors in right-ascension depending 

 upon right-ascension and any existing diurnal effect are practically 

 eliminated, as are also the assumed positions of the clock stars. 



The series of observations used extend over periods of 24 to 130 

 hours, with a group of clock stars every 12 hours. In a large majority 

 of cases there are also successive transits of a circumpolar star to effect 

 a determination of the azimuth independent of the assumed right- 

 ascension. This should insure a reasonable degree of freedom from 

 systematic error and error in the assumed place. For the preliminary 

 discussion, 347 series containing 4,652 observations were available. 



After correcting the resultant clock corrections for a term in right- 

 ascension dependent upon declination, they were collected in the order 

 of right-ascension, and a preliminary value of the term in right- 

 ascension depending upon right-ascension was determined. Applying 

 this term, the observations were then collected in order of the local 

 mean time of observation and a preliminary diurnal term was found. 

 After correcting for the diurnal term, a preliminary term east minus 

 west was formed. Three approximations were employed to deduce 

 the final results. 



The final terms in right-ascension dependent upon right-ascension 

 are: 



Albany -0!008 sin R. A. +0?012 cos R. A. (corrn. to P. G. C.) 



San Luis -O.OOS sin R. A. +0.009 cos R. A. (corrn. to P. G. C.) 



The diurnal term is: 



-|-0!0114 sin M. T. -0^0073 cos M. T. (corrn. to obsd. Transits) 



In order to test the reality of the diurnal term, a rough treatment 

 was made of material obtained from the following catalogues : Green- 

 wich Obsei-vations for 1907-8; Pulkova Series II, Vol. 20, 1894-6; 

 Cape Fundamental Catalogue for 1900. 



For Greenwich and Pulkova observed clock corrections were given. 

 Those series were selected where there was a fair distribution of obser- 

 vations throughout 24 hours. The resultant diurnal terms are listed 

 below. 



Albany +0.0114 sin M. T. -0.0073 cos M. T. 



Greenwich - .0058 - .0129 



Cape - .0054 - .0069 



Pulkova - .0033 - .0078 



For Albany there is a well-marked maximum at 8 hours, for Green- 

 wich at 13 to 14 hours, for Cape at 14 hours, and for Pulkova at 13 to 

 14 hours. This peculiarity suggested an effect depending upon longi- 

 tude. The material was then arragned according to Greenwich mean 

 time and further treated as a cosine term. The tabulated results give 



