206 Mr. T. Gray on Instruments for Measuring 



of B is slightly different, a light bracket being introduced for 

 the purpose of raising the ring R, and allowing room for the 

 registering-lever. The ring R is pivoted so that it can turn 

 round a vertical axis, thus rendering more definite the point 

 which remains steady. A stiff bar, b, fixed to the post P at 

 one end, carries the universal joint in which the writing-lever, 

 Z, turns. This universal joint consists of four very light bent 

 springs, fixed at the lower end to a round ring, and at the 

 upper end to a small disk through which the lever I passes, 

 and to which it is soldered. The four springs are arranged at 

 right angles to each other, and offer a slight elastic resistance 

 to the rotation of the lever Z, and consequently of B and A. 

 This resistance is made so small that the period of oscillation 

 of B is very long. The reason for using this peculiar form of 

 joint is to prevent a gradual displacement of R by the top of 

 I when the plate on which I writes is kept moving by clock- 

 work. The static friction, even with the precaution taken in 

 this machine, will probably be much greater than the pressure 

 of I; but a slight motion might take place during the shaking. 

 The ideas of calculating the positions of H and W and of 

 pivoting the ring R which are used in this paper, are taken 

 from Prof. Ewing's paper " On an Astatic Horizontal-Lever 

 Seismograph," communicated at last meeting to this Society, 

 or rather from a conversation which I had with him previous 

 to the reading of that paper. I have generally been in the 

 habit of finding approximately the steady part by experiment j 

 and in many cases this is almost the only satisfactory method*. 

 In using this instrument, the two gates A and B are ad- 

 justed so that their planes are accurately at right angles to 

 each other, each inclined at an angle of 45° to the front of the 

 post P. The gates are so made that when in this position the 

 lever I is vertical. At the lower end of the lever I a thin 

 sewing-needle, n, slides up and down through two small loops. 

 Under this needle the plate p, smoked to receive the record, 

 is placed. If it is desired to separate the different motions, p 

 must be moved by clockwork. 



V. On a Conical-Pendulum Seismograph, 



In a paper on " Steady Points for Seismographs," commu- 

 nicated to the Seismological Society of Japan in March 1881, 



* In the small instrument which I described before the Seismological 

 Society, the levers were so light and small compared with the ring R that 

 no determination of the exact proper positions of the axes were made. That 

 instrument promises to act very fairly as a seismograph ; but I expect the 

 instrument just described to act better, because of smaller friction, due to 

 great height, finer and more scientifically arranged hinges, and greater 

 length of arm. 



