ON THE MEASUREMENT OF THE LCNAH DlSTDEBANCE OF GEAVITY. 109 



Arraugements had also to bo nii>de for working the sensitisev, cen- 

 ti'aliser, and disturber from outside the I'oom. 



A scafibldiug was erected over the tub, but free of contact therewith, 

 and this supported a system of worm-wheels, tangent screws, and pulleys 

 by which the three requisite movements could be given. The junctures 

 with the sensitising and centralising rods were purposely made loose, 

 becaiise it was found at first that a slight shake to the scaffolding 

 disturbed the pendulum. 



The pulleys on the scaffolding are driven by cords which pass to the 

 observing window. 



On the window-sill we now have two telescopes, four pulleys, an 

 arrangement, with a scale attached, for raising and depressing the guide 

 weight, and a gas tap for governing the flame in the room. 



After the arrangemeiits which have been described were completed we 

 sensitised the instrument from outside the room. The arrangements 

 worked so admirably that we could produce a quite extraordinary degree 

 of sensitiveness by the alternate working of the sensitising and central- 

 ising wheels, without ever causing the image of the lamp to disappear 

 from the field of view. This is a great improvement on the old arrange- 

 ment with the stone gallows. 



We now found that if one of us was in the room and stood at about 

 16 feet to the S. of the instrument with his feet about a foot apart, and 

 slowly shifted his weight from one foot to the other, then a distinct 

 change was produced in the position of the mirror. This is the most 

 remarkable proof of sensitiveness which we have yet seen, for the instru- 

 ment can detect the difierence between the distortion of the soil caused 

 by a weight of 140 lbs. placed at 16 feet and at 17 feet. We have not as 

 yet taken any great pains to make the instrument as sensitive as possible, 

 and we have little doubt but that we might exceed the present degree of 

 delicacy, if it were desirable to do so. 



The sensitiveness now attained is, we think, only apparently greater 

 than it was with the stone gallows, and depends on the improved optical 

 arrangements, and the increase of steadiness due to the elimination of 

 changes of temperature in the support. 



From July 21 to July 25 we took a series of readings. There was 

 evidence of a distinct diurnal period with a maximum about noon, when 

 the pendulum stood furthest northwards ; in the experiment with the 

 stone gallows in 1880 the maximum northern excursion took place 

 between 5 and 7 p.m. 



The path of the pendulum was interrupted by many minor zigzags, 

 and it wonld sometimes reverse its motion for nearly an hour together. 

 During the first four days the mean position of the pendulum travelled 

 southward, and the image went oS" the scale three times, so that we had 

 to recentralise it. In the night between the 24th and 25th it took an 

 abrupt turn northward, and the reading was found in the morning of the 

 25th at nearly the opposite end of the scale. 



On the 25th the dance of the image was greater than we had seen it 

 at any time with the new instrument, so that we went into the room to 

 see whether the water had fallen in the tub and had left the top of the 

 copper tube exposed ; for on a previous occasion this had appeared to 

 produce much unsteadiness. There was, however, no change in the state 

 of affairs. A few days later the image was quite remarkable for its 

 steadiness. 



