450 



SCIENCE 



[N. S. Vol. XXV. No. 638 



ing the performance of his clock, I have 

 used the adopted rates as given by Auwers 

 in his rereduction of Bradley. The month- 

 ly means of the rates from July, 1758, to 

 July, 1759, were taken and the difference 

 of each rate from its monthly mean. Then 

 the mean of these differences without re- 

 gard to sign was taken for each month. 



result. However, there are two well-known 

 clocks which should be mentioned in this 

 connection, and in conclusion I will give 

 some hitherto unpublished data concerning 

 the clock with which I have been working 

 during the past three years. 



Probably no clock has had its rate more 

 thoroughly discussed than Hohwii No. 17, 



VABIATION OF DAILY BATE OF THKEE GEEENWICH SIDEREAL CLOCKS 



The rates of two other clocks of the 

 Greenwich Observatory were likewise dis- 

 cussed, the standard clock for the year 

 1850 and that for 1900, the adopted daily 

 rates as published in the annual volumes 

 being used. The first of these was kept in 

 the observing room and thereby subjected 

 to large variations of temperature, while 

 the second, made in 1871 by E. Dent and 

 Company, was fixed to the north wall of 

 the magnetic basement, as in this apart- 

 ment the temperature is kept nearly uni- 

 form. The pendulum of this latter clock 

 is provided with barometric as well as 

 thermometric compensation. 



I sought to make a similar comparison 

 for the clocks of the other large observa- 

 tories, but soon found that the information 

 concerning the performance of the various 

 clocks was given in such a form that it was 

 in almost every case either impossible or 

 extremely laborious to secure the desired 



the standard clock of the observatory at 

 Leiden. 



It was set up in the transit room in 1861 

 and in December, 1898, was removed to 

 the large hall of the observatory, where, 

 enclosed in two wooden cases, it was placed 

 in a niche cut in the pier of the ten-inch 

 refractor. To further guard against sud- 

 den changes of temperature the niche is 

 closed by a glass door. At the meeting of 

 the Royal Academy of Sciences at Amster- 

 dam, held September 27, 1902, Dr. E. F. 

 van de Sande Bakhuyzen submitted the fol- 

 lowing formula as the best representation 

 of the daily rate of the clock : 



Dailyrate = — 0M73+0'.0157(h-760"""') 



— 0».0253 ( t-10° ) + 0".00074 ( t-10° ) ' 

 + 0'.0465 cos 27r(T-May 3)/365 



and gave the result of a comparison of the 

 observed daily rates 1899-1902, the average 

 interval of time for each rate being six 



