ON THE MEASUBEMENT OF THE LUNAR DISTURBANCE OF GRAVITY. 103 



that space should be made of a heat-conducting material, which should 

 itself form the support for the pendulum. The whole instrument, 

 including the basement, was to be immersed in water, and the basement 

 itself was to be carefully detached from contact with the building in 

 which it stands. By these means we hoped to damp out the short oscil- 

 lations due to local tremors, but to allow the longer oscillations free to 

 take place ; but above all we desired that changes of temperature in the 

 instrument should take place with great slowness, and should be, as far 

 as possible, equal all round. 



We removed the pavement from the centre of the room, and had a 

 circular hole, about 3 feet 6 inches in diameter, excavated in the ' made 

 earth,' until we got down to the undistni-bed gravel, at a depth of about 

 2 feet 6 inches. 



We obtained a large cylindrical stone 2 feet 4 inches in diameter and 

 2 feet G inches in height, weighing about three-quarters of a ton. This 

 we had intended to place on the earth in the hole, so that its upper sur- 

 face should stand flush with the pavement of the room. But the excava- 

 tion had been carried down a little too deep, and therefore an ordinary 

 flat paving stone was placed on the earth, with a thin bedding of cement 

 underneath it. The cylindrical block was placed to stand upon the 

 paving stone, with a very thin bedding of lime and water between the 

 two stones. The svirface of the stone was then flush with the floor. We 

 do not think that any sacrifice of stability has been made by this course. 



An annular trench or ditch a little less than a foot across is left round 

 the stone. We have lately had the bottom of the ditch cemented, and 

 the vertical sides lined with brickwork, which is kept clear of any con- 

 tact with the paA^ement of the room. On the S. side the ditch is a little 

 wider, and this permits us to stand in it conveniently. The bricked 

 ditch is watertight, and has a small overflow pipe into the drains. The 

 water in the ditch stands slightly higher than the flat top of the cylin- 

 drical stone, and thus the whole basement may be kept immersed in 

 water, and it is, presumably, at a very uniform temperature all round. 



Before describing the instrument itself we will explain the remaining 

 precautions for equalisation of temperature. 



On the flat top of the stone stands a large barrel or tub, 5 feet G 

 inches high and 1 foot 10 inches in diameter, open at both ends. The 

 diameter of the stone is about 2 inches greater than the outside measure 

 of the diameter of the tub, and the tub thus nearly covers the whole of 

 the stone. The tub is well payed with pitch inside, and stands on two 

 felt rings soaked in tar. Five large iron weights, weighing altogether 

 nearly three-quarters of a ton, are hooked on to the upper edge of the 

 tub, in order to make the joint between the tub and the stone watertight. 

 Near the bottom is a plate-glass window ; when it is in position, the 

 window faces to the S. Tliis tub is filled with water and the instru.ment 

 stands immersed therein. 



We had at first much trouble from the leakage of the tub, and we 

 have to thank Mr. Gordon, the assistant at the Laboratory, for his ready 

 help in overcoming this difficulty, as well as others which were per- 

 petually reciirring. The mounting of the tub was one of the last things 

 done before the instrument was ready for observation, and we must now 

 return to the description of the instrument itself. 



We used the same pendulum-bob as before, but we had its shape 

 altered so that the ends both above and below were conical surfaces, 



