ON THE MEAisURKMENT OF XHE LUNAll DISTUllBANCE Ok" GKAVITV. 107 



After tlie mirror was hung and the mirror-box screwed on, a brass cap 

 was fixed by screws on to the flange at the top of the copper tube. This 

 cap has a tube or chimney attached to it, the top of which rises five inches 

 above the top of the cap or lid from -which the pendulum hangs. From 

 this chimney emerges a rod attached to the screw by which the sensitising 

 apparatus is actuated, and also the silk by which the guide weight is 

 raised or depressed. 



The copper tube, with its appendages, was then filled with a boiled 

 mixture of filtered water and spirits of wine by means of a small tap in 

 the back of the mirror-box. The mixture was made by taking equal 

 volumes of the two fluids ; the boiling to which it was subjected will of 

 course have somewhat disturbed the proportions. Poiseuille has shown ' 

 that a mixture of spirits and water has much greater viscosity than either 

 pure spirits or pure water. When the mixture is by weight in the pro- 

 portion of about seven of water to nine of spirits, the viscosity is nearly 

 three times as great as that of pure spirits or of pure water. As the 

 specific gravity of spirits is about -8, it follows that the mixture is to 

 be made by taking equal volumes of the two fluids. It is on account of 

 this remarkable foot that we chose this mixture in which to suspend 

 the pendulum, and we observed that the unsteadiness of the mirror was 

 markedly less than when the fluid used was simply water. 



The level of the fluid stood in our tubular support quite up to the top 

 of the chimney, and thus the highest point of the pendulum itself was 

 5 inches below the surface. 



The tub was then let down over the instrument, and the weights 

 hooked on to its edge. The plate-glass window in the tub stood on the 

 S. opposite to the mirror-box. The tub was filled with water up to 

 nearly the top of the chimney, and the ditch round the stone basement 

 was also ultimatelv filled with water. The whole instrument thus stood 

 immersed from top to bottom in water. 



Even before the tub was filled we thought that we noticed a diminu- 

 tion of unsteadiness in the image of a slit reflected from the mirror. The 

 filling of the tub exercised quite a striking effect in the increase of 

 steadiness, and the water in the ditch again operated favoui-ably. 



We met with much difiiculty at first in preventing serious leakage of 

 the tub, and as it is still not absolutely watertight, we have arranged a 

 water-pipe to drip about once a minute into the tub. A small overflow 

 pipe from the tub to the ditch allows a very slow dripping to go into the 

 ditch, and thus both vessels are kept full to a constant level. We had 

 to take this course because we found that a rise of the water in the ditch 

 through half an inch produced a deflection of the pendulum. The ditch, 

 it must bo remembered, was a little broader on the S. side than else- 

 where. 



In Maj^, 1881, we took a series of observations with the light, slit and 

 scale. The scale was about 7 feet from the tub, and in order to read it 

 we found it convenient to kneel behind the scale on the ground. I was 

 one day watching the light for nearly ten minutes, and being tired with 

 kneeling on the pavement I supported part of my weight on my hands a 

 few inches in front of the scale. The place where my hands came was 

 on the bare earth fx-om which one of the paving stones had been removed. 

 I was surprised to find quite a large change in the reading. After 



' Pogycndmfs Ainialni, 1843, vol. r,S, p. 437. 



