G Proceedings of the Royal Society of Victoria. 



clock pendulum were in every case selected for observation ; this 

 edge would be the apparent right edge to an observer on the 

 opposite side of the rooro, and consequently was the same as that 

 observed in Melbourne. The card disc on the clock pendulum was 

 that used in Melbourne, and the method of observing was the same 

 as that adopted in Melbourne, Kew and Greenwich. Four sets of 

 swings, two for each face, were effected with each pendulum, the 

 two sets for any one face being taken on the same day. The 

 discordance between the results for opposite faces with the Pendu- 

 lum No. 11* was very marked; but there is no doubt that this pen- 

 dulum is slightly bent, and very little question that its knife-edges 

 are not accurately perpendicular to the pendulum bar. I noticed 

 both these defects on the first arrival of the pendulums from 

 England. Fortunately, so long as they are constant they do not 

 afiect the accuracy of differential observations. 



Clock Rates. 



The Shelton clock was compared directly with the Siderial 

 clock of the Observatory at the beginning and end of each day's 

 work, in order to determine the difference of their rates. The 

 comparison was effected by means of a tape chronograph of 

 Morse's pattern, constructed by Messrs. Siemens Bros., which 

 worked very uniformly. The chronograph spaces were measured 

 off by means of a divided lens, the halves of which were mounted 

 on brass sliding pieces carrying scales ; this instrument being 

 used in much the same way as the heliometer. The scales were 

 graduated in inches and tenths, and hundredths were estimated. 



The error of the Siderial clock was determined by Transit 

 observations. Unfortunately the nights were so cloudy for most 

 of the time that star observations could not always be obtained, 

 and sun transits had perforce to be resorted to ; this cannot, 

 however, have affected the results to any serious extent, as on 

 those occasions when both sun and stars were observed the 

 difference of the deduced rates was never more than one or two 

 hundredths of a second. Both clocks behaved well ; their rates 

 are given in Table II. 



» AlluOed to by Baracclii, I.e., p. 1G6. 



