ON EARTH TREMOR?. 301 



from 9.12 a.m. to 3.12 p.m. ; it was then turned down, and readings were 

 continued for six hours longer. The results are illustrated by the curve 

 in fig. 6. The effect of expansion is shown by the movement to the east 

 for a short time after the gas is lighted, and that of contraction by an 

 inci-ease in the rate of westerly deflection immediately after the gas is 

 turned down. The most important movement is that of more than a 

 second to the west while the gas was up, and to the east when it was- 

 turned down. There can be little doubt that it is due to the action of 

 convection currents in the paraffin oil, in which the mirror and frame are 

 immersed. 



An attempt was next made to determine the part of the instrument 

 whose expansion produced the first deflection. For this purpose a card- 

 board box, 12 inches long, 5J inches broad, and 4 inches deep, was filled 

 with hay. It was first placed so as to shield the upper part of the 

 instrument, leaving the mirror-box and levelling-screws uncovered, with 

 the result of preventing the first easterly movement. The gas was kept 

 alight for two hours (11.30 a.m. to 1.30 P.M.) ; it was then turned low 

 down, and readings were taken at intervals during the next three hours. 

 After one minute the image was seen to the left of the vertical cross- wire, 

 indicating a deflection to the west ; and this westei-ly movement continued 

 at a very nearly uniform rate during the whole time the gas was alight, 

 and also for about half an hour afterwards. Its velocity then rapidly 

 diminished, and after a quarter of an hour an easterly movement set in, 

 whose velocity soon became uniform and equal to that of the westerly 

 movement. The total westerly deflection of the mirror in. tlie first 2^ 

 hours was equivalent to a tilt of 1^ seconds. 



The box was then placed resting with its longer side on the wooden 

 cover of the drain-pipe. Strings were tied to each end and brought 

 round by hooks to the telescope-seat, so that the box could be moved 

 backwards and forwards in front of the mirror-box and levelling-screws. 

 These were shielded by the box from the gas-jet, which was kept alight 

 for two hours (12.40 to 2.40 p.m.). In order to take the readings, the 

 box was drawn to one side for a few seconds and then replaced. 

 Readings were also continued for several hours after the gas was turned 

 down. The resulting curve is similar to that in fig. 6. It follows, there- 

 fore, that the first easterly movement which takes place after the gas is 

 lighted is caused by the expansion of the brass-tube more than by that 

 of the levelling-screws. 



It is evident from these experiments that the natural movements of 

 the pendulum in its present foi-m and position may be sei'iously affected 

 by the slight changes of temperature which take place in ordinary cellars. 

 To guard against these, Mr. H. Darwin proposes to alter the mode of 

 suspension of the mirror. The two hooks by which it hangs on the silver 

 wire will be placed in a line at right angles to the plane of the mirror. 

 The instrument will also be protected by a case, and perhaps by immersion 

 in water. For photographic registration, it would be better to use a 

 glow electric lamp, or, better still, an induction-coil spark, instead of a 

 continuously burning gas-jet. 



A few experiments were made on the bending of the cellar-floor by 

 means of a heavy weight, but without much success. This was due 

 partly to the tilting so caused being nearly or quite masked in the daily 

 movement of the mirror, partly because to move the weight I was obliged 

 to enter the cellar, and frequently approach close to the instrument. The' 



