2:^0 r rowed lugs of the Royal Societi/ of Victoria. 



With the pendulums is su])])]ied :i " dummy pendulum," 

 of identical material and dimensions, into which two holes 

 are sunk for the reception of the bulbs of a pair of 

 thermometei-s. The dummy is placed in the same vacuum 

 chamber with the pendulum when vibrating, and close to it. 

 Their temperatures may accordingly be assumed as identical, 

 and the temperature of the dummy, as given by the 

 thermometers, can be employed for deteriiiining the temper- 

 ature corrections to be applied to the pendulum. Tlie 

 corrections to be applied to the thermometers have been 

 determined at Kew. 



(b) The clock employed for the (jbservation of coincidences 

 is a siderial clock, made by Shelton, and was used by Sabine 

 on his exjiedition in 1822. Its mean daily rate is very 

 constant, but it is subject to rather considerable horar}- 

 fluctuations of rate. The clock has an arrangement which 

 allows of its being re-wound without loss of driving power 

 during the winding. 



(c) The vacuum apparatus consists of a cylinder of sheet 

 copper, half closed at the top by a thick brass plate for 

 supporting the agate planes, and closed in above this by 

 a glass bell, ground to lit the brass plate ; it is closed at the 

 bottom by a metal hemisphere. It has one glass window 

 about lialf-way up, through which the thermometers are 

 read, and four others in the plane of the tail-piece of the 

 pendulum. Through one pair the coincidences are observed, 

 the other pair allowing a side view of the tail-piece, which 

 is necessary for determining the amplitude of its vibration. 

 To the sill of the back window is attached a bi-ass plate 

 bearing two scales at right angles to each other etched on 

 ground glass, and with well blackened divisions for 

 measuiing this amplitude. 



The cylinder is supported by three large levelling screws 

 on a heavy iron girder, which is itself bolted to a very 

 massive timber framework fastened together with iron bolts 

 and clam])s. The massiveness of the cylinder and fram«^ 

 render it quite impossible for the oscillations of the 

 pendulum to be communicated to the supports. 



The starting and stopping of the pendulum is perf )rnied 

 by means of levers worked from outside the cylinder by 

 metal rods passing through stuffing boxes, and cases filled 

 with oil to pi-event leakage of air. 



