126 TUCKER— EFFECT OF DIURNAL VARIATION 



19 13 — 003 



20 15 — 019 



21 14 _^ 003 



22 16 — 008 

 ^7> 18 + 005 



016 



022 



on 



013 



008 

 Average 13 + 0.006 + o.oio 



From our observations extending over a period of a quarter of a 

 century the mean excess of the hourly rate at night over the average 

 hourly rate during one day is 0^.006. This corresponds to a variation 

 in the daily rate of an amplitude of approximately 0.3 of a second. 

 The difference between observed and interpolated clock corrections 

 would be a maximum for an interval of six hours, and would amount 

 to over o^03. Double this difference would occur between the ob- 

 servations near sunset and sunrise, and an observed difference of 

 0^.06 has been found between clock corrections at those epochs of 

 the day. These numerical results are still subject to revision, as more 

 precise values are to be anticipated from our current series of 

 observations. 



In fundamental right-ascension observations differences of this 

 size should occur, but the alternate observations of groups of stars, 

 twelve hours apart, has smoothed out this effect in our adopted 

 systems. 



It is not often necessary to carry the daily rate forward more 

 than two hours except in fundamental work, and the difference be- 

 tween daily and hourly rates would rarely introduce an error exceed- 

 ing 0^0 1. 



It has been our custom generally to adopt the hourly rate derived 

 during the period of observation, when that period is of sufficient 

 length, in reducing transit observations. The results thus obtained 

 conform to the adopted right-ascension system, with its errors in- 

 cluded. The actual performance of the clock has been of secondary 

 importance in deriving the right ascensions. 



* [one star] w Hcrculis, mistake in P.M. 



