ON THE MEASUREMENT OF THE LUNAR DISTURBANCE OF GHAVITT. 113 



The principle employed is as follows : — There is a very stout vertical 

 stand, supported on three legs. At the top and bottom of the vertical 

 shaft are fixed two projections. Attached to each projection is a fine 

 straight steel clock spring ; the springs are parallel to the vertical shaft 

 of the stand, the one attached to the lower projection running upwards, 

 and that attached to the upper one running downwards. The springs 

 are of equal length, each being equal to half the distance between their 

 points of attachment on the projections. 



The springs terminate in a pair of rings, which stand exactly opposite 

 to one another, so that a rod may be thrust through both. 



A glass rod has a heavy weight attached to one end of it, and the 

 other end is thrust through the two rings. The rings are a little separated 

 from one another, and the glass rod stands out horizontally, with its 

 weight at the end, and is supjDorted by the tension of the two springs. 

 It is obvious that if the point of attachment of the upper spring were 

 vertically over that of the lower spring, and if the springs had no 

 torsional elasticity, then the glass rod would be in neutral equilibrium, 

 and would stand equally well in any azimuth. 



The springs being thin have but little torsional elasticity, and 

 Professor ZoUner arranges the instrument so that the one support is very 

 nearly over the other. In consequence of this the rod and weight have 

 but a small predilection for one azimuth more than another. The free 

 oscillations of the hox^zontal pendulum could thus be made extra- 

 ordinarily slow ; and even a complete period of one minute could be easily 

 attained. 



A very small horizontal foi'ce of course produces a large deflection of 

 the pendulum, and a small deflection of the force of gravitation with 

 reference to the instrument must produce a like result. He considers 

 that by this instrument ho could, in the first form of the instrument, 

 detect a displacement of the horizon through 0"'00035 ; in the second his 

 estimate is 0"-001. 



The observation was made by moans of a mirror attached to the 

 weight, and scale and telescope. 



The maximum change of level due to the moon's attraction is at St. 

 Petersburg 0"-0174., and from the sun 0"-0080 [C. A. T. Peters, 'Bull. 

 Acad. Imp. St. Petersbourg,' 1844, vol. 3. No. 14] ; and thus the insti-a- 

 ment was amply sensitive enough to detect the lunar and solar disturb- 

 ances of gravity.* 



Professor ZoUner found, as we have done, that the readings were never 

 the same for two successive instants. The passing of trains on the rail- 

 way at a mile distant produced oscillations of the equilibrium position. 



' We are of opinion that M. Zijllner has made a mistake in using at Leipsig 

 Peters' results for St. Petersbiarg. Besides this he considers the changes of tlie 

 vertical to be 0'''017i on eacli side of a mean position, and thus sa^'s the change is 

 0"'0348 altogether. Now a rough computation which I have made for Cambridge 

 shows that the maximum meridional horizontal component of gravitation, as due to 

 lunar attraction, is 4'12 x 10"'* of pure gravit}'. This force will produce a dellec- 

 tion of the plumb-line of 0"'00S."), and the total amplitude of meridional oscillation 

 will be 0"'0170. The maximum detiection of the plumb line occurs when the 

 moon's hour-angle is ± 45° and ± 135° at the place of observation. The change at 

 Cambridge when the moon is S.E. and N.W. ^s 0''-0l'16. The deflection of the plumb 

 line varies as the cosine of the latitude, and is therefore greater at Cambridge than at 

 St. Petersburg. Multiplying 021!} by sec 51-43' cos 60° we get -0174, and thus my 

 calculation agrees with Peters'. 



1881. ' I 



