ON TBB MEASUEEMBNI OF THE LUNAE DISTURBANCE OF GRAVITY. 97 



to have a flat bottom to the pendnlum. This mistake we have remedied 

 in the final experiment described in the present paper. 



The damping effect of the water on the oscillations of the pendulum 

 and of the mirror was very great, and although the incessant dance of 

 the light contmued, it was of much smaller amplitude, and comparatively 

 large oscillations of the pendulum, caused by giving the piers a push, 

 died out after two or three swings. A very slight push on the stone piers 

 displaced the mean position of the light, but jumping and stamping on 

 the pavement of the room produced no perceptible effect. If, however, 

 one of us stood on the bare earth in the ditch behind, or before the 

 massive stone pier, a very sensible deflection of the light was caused ; 

 this we now know was caused by an elastic depression of the earth, which 

 tilted the whole structure in one or the other direction. A pull of a few 

 ounces, delivered horizontally on the centre of the lintel, produced a clear 

 deflection, and when the pull was 81bs., the deflection of the spot of hght 

 amounted to 45 cm. Wc then determined to make some rough systematic 

 experiments. 



The room Avas darkened by shutters over all the windows, and the 

 doors were kept closed. The paraffin lamp stood at three or four feet to 

 the S.E. of the easterly stone pier, but the light was screened from the 

 pier. 



We began our readings at 12 noon (March 15, 1880), and took eight 



between that time and 10.30 p.m. From 12 noon nntil 4 p.m. the lamp 



was left burning, but afterwards it was only lighted for about a minute to 



take each reading. At 12 the reailing was 595 m.m., and at 4 p.m. it 



was 936 m.m.' ; these readings, together with the intermediate ones, 



showed that the pendulum had been moving northwards with a nearly 



uniform velocity. After the lamp was put out, the pendulum moved 



southward, and by 10.30 p.m. was nearly in the same position as at noon. 



During the whole of the two following days and a part of the next we 



took a number of readings from 9 A.M. nntil 11 p.m. The observations 



when gi-aphically exhibited showed a fairly regular wave, the pendulum 



being at the maximum of its northern excursion between 5 and 7 p.m., 



and probably furthest south between the . same hours in the morning. 



But besides this wave motion, the mean position for the day travelled a 



good deal northward. We think that a part of this diurnal oscillation 



was due to the warping of the stone columns from changes of temperature. 



An increase of temperature on the south-east faces of the piers canned 



the lintel towards the north-west, and of this displacement we observed 



only the northerly component. The lamp produced a very rapid efi'ect, 



and the diurnal change lagged some two hours behind the change in the 



external air. The difference between the temperatures of the S.E. and 



N.W. faces of the pier must have been very slight indeed. At that time, 



and indeed until quite recently, we attributed the whole of this diurnal 



oscillation to the warping of the piers, but we now feel nearly certain that 



it was due in great measure to a real change in the horizon. 



We found that warming one of the legs of the iron tripod, even by 

 contact with the finger, produced a marked effect, and we concluded that 

 the mode of suspension was unsatisfactory. 



' I give the numbers as recorded in the note-book, but the readings would some- 

 times diifer by 2 or 3 m.m. within half-a-minute. The light always waves toand 

 fro in an uncertain sort of way, so that it is impossible to assign a mean position 

 witli any certainty. 



1881. H 



