638 MISCELLANEOUS PAPERS. 



from air distuibauces by means of a wooden box covered with clotli 

 and fastened directly to tbe pilhxr. The front of the box can be taken 

 oft", so as to remove or rephice the peudukim, while an opening, pro- 

 vided with a cover, in the top allows the weights to be moved at will. 

 On the sides in front of the mirrors, already described, are small square 

 windows closed with glass. Later this box was covered with a coat of 

 white lead. 



The telescope with which the oscillations were observed had an open- 

 ing of 8^'" and SO*^^'" focal distance. It was fastened on a brick pillar in 

 communicating passage. The scale utilized in the observations was 

 on glass; it was 50*='" in length, and was placed at a distance of 500'"' 

 from the mirror. The magnifying power of the telescope was such that 

 the tenth of a millimeter could be approximated, and after some prac- 

 tice the twentieth, a value which represented about a second of arc. 

 The graduation error in the scale was never greater than 0.01"'"'. 



THE ATTRACTINa MASSES. 



Until now balls of lead have been universally used. But on account 

 of the softness of lead it is almost impossible to secure a geometric 

 body homogeneous in its structure; besides, the continued suspension 

 would bring about, through the action of gravity, deformities. For 

 these reasons, notwithstanding its small specific gravity, the preference 

 was given to cast iron, and for practical convenience and the possibilitj^ 

 of a more exact figure the cylindrical shape was chosen. They were 

 cast under pressure, and upon a subsequent weighing and examination 

 for specific gravity they appeared practically identical. 



These masses, each of which weighed 325 kilograms, were so hung by 

 a wire rope over a system of pulleys that they always balanced one an- 

 other, and could be easily moved, one up, the other down. The length 

 of the rope was so adjusted that when one of the cylinders was directly 

 opposite the upper end of the pendulum the other would be on the oppo- 

 site side and directly on a line with the lower end. In order to obviate 

 any lateral motion and consequent shaking, each cylinder was provided 

 with guiding rollers which ran up tracks attached to uprights, firmly 

 held in their proper places. This and similar precautions were so well 

 taken that during the entire series of observations no jarring was ob- 

 served. 



ARRANGEMENT OF THE OBSERVATIONS. 



When the pendulum was put in motion, before the cylinders were put 

 in place, it was noticed if the amplitude was bisected by the vertical 

 through the point of support. If not, the screw in the upper end, al- 

 ready referred to, was made use of, aided, if necessary, by a change in 

 the position of the thin plates on the top of the rod. 



The position of equilibrium was next determined by placing, for in- 

 stance, the eastern weight near the upper end of the pendulum, observ- 

 ing a series of four consecutive elongations, then changing the weights, 

 observe another series of four elongations. Both series were, with few 



