April 12, 1888] 



NATURE 



571 



pendulum," first as a multiplying index, and secondly as a means 

 to check the motion of a large pendulum. In pendulum seismo- 

 graphs, which I have largely used in Japan (see Trans. Seis. Soc, 

 vol. iv. p. 91), a b was loaded with a brass ball, and thus the centre 

 of oscillation raised above b. The moment that a b exerted on 

 w was not, however, sufficient to prevent w from swinging, and 

 its movements were retarded and rendered "dead beat" by 

 frictional resistance directly applied to the surface of w, which 

 was a disk of lead suspended horizontally. During the last two 

 years I have had several seismographs constructed in which a b 

 was long ; and, as near to a as possible, a weight sufficiently 

 large to render w feebly stable was placed. This important 

 suggestion of loading ab originated with Mr. T. Gray. Later, 

 Mr. Gray drew attention to the necessity of rendering an ordinary 

 pendulum, for small displacements, absolutely astatic, and he 

 suggested various means by which this might be accomplished 

 (Trans. Seis. Soc, vol. iii. p. 145). 



In the same publication, vol. v. p. 89, Prof. Ewing, attack- 

 ing the same problem, described a duplex pendulum, a modified 

 form of which he described in vol. vi. p. 19. In vol. viii. p. ^i, 

 Prof. Sekiya described an improved form of Prof. Ewing's in- 

 strument (see also Nature, vol. xxxiv. p. 343^ In the duplex 

 pendulum seismograph an ordinary pendulum is rendered astatic 

 for small displacements by placing an inverted pendulum beneath 

 it, and so uniting the bobs of the two pendulums that any hori- 

 zontal motion is common to both, and the jointed system so 

 proportioned that neutral or feebly stable equilibrium is obtained. 

 Although these instruments are.forseismometrical work, theoretic- 

 ally good, in practice such of them as I have had, which are 

 the best to be obtained in this country, present many serious ob- 

 jections. Among these objections I may mention the following : 

 (i) the difficulty of adjustment ; (2) the difficulty of inserting 

 and removing smoked glass plates ; (3) the fact that the pointer 

 being cranked at its upper end it does not give so satisfactory a 

 record in directions at right angles to the plane of the crank as is 

 desired ; (4) their incapability of recording an earthquake of 

 greater amplitude than 5 mm. 



By introducing arrangements for adjustment, alteration in 

 the form of the recording index, &c., these instruments might 

 be improved. Possibly in the instrument recommended by 

 Prof. Ewing for use in Observatories (see Nature, vol. xxxiv. 

 p. 343), although it appears to be practically similar to those 

 I have in Tokio, the objections may not be so serious. 



The instrument of this class which I have in all respects found 

 the most satisfactory is, in its essential features, shown in Fig. 2, 



stiftened in the centre by a small transverse table which carries 

 the bar B. w is so suspended that it can be readily shifted 

 laterally or vertically. Below there is a small shaft which carries 

 the smoked plate. By means of a wedge this can be raised or 

 lowered, and the plate brought to any degree of contact with the 

 sliding pointer. This portion of the apparatus is so simple that 

 a record-receiving surface is instantly adjusted or removed by 

 the movement of a handle connected with the wedge. The in- 

 strument is an outcome of instruments which have preceded it, 

 and it may be regarded as a modification of an old type where 

 a b c has been prolonged upwards and the balance load placed 

 above a instead of being between a and b. Its chief recom- 

 mendations are; (i) its smallness ; (2) the simplicity and few- 

 ness of its parts ; (3) the ease with which it may be used ; (4) its 

 large range of motion ; (5) the accuracy of its diagrams. The 

 test for accuracy has been made by placing the instrument upon 

 a specially designed shaking table, the absolute movements of 

 which are recorded by a multiplying lever. 



Comparing the diagrams given by the machine with those 

 given by the table, it is found that for all small displacements, 

 whether in right lines or complicated curves, the diagrams, 

 20 or 30 mm. in length, are practically identical. For diagrams 

 50 mm. in their greatest dimensions, composed of a complication 

 of curves if anything greater in complexity than those yielded by 

 ordinary earthquakes, some differences occur, the extent of which 

 may be judged of by the accompanying diagram. Fig. 3. Figs. 4 



Fig. 3. 



and 5 are examples of the diagrams obtained for small displace- 

 ments. These diagrams are fair specimens, but have not been 

 selected as particularly good examples. The multiplication of 

 the table diagram, marked T, is 6 -3, while that of the seismograph, 

 marked S, is slightly over 6. 



Diagrams of the old type of seismograph with the weight 

 on a b have also compared favourably with the table motion. I 

 regret, however, to say that the diagrams given by one of Prof. 



Fig. 2. 



in which w represents a lead ring about 7 inches in diameter, with 

 a small tube, a, fixed in a plate at its centre. \v is supported by 

 three strings Or wires, s. The indicating pointer \^iv ab c, pro- 

 longed downwards, at the lower end there being a needle as a 

 writing-point sliding in a small tube, w a /» is a light steel rod with 

 a ball forming a universal joint on the tube at a, and a point, I, 

 pivoting in the fixed steel bar B. The stability of the system is 

 readily altered by raising or lowering the small weight w. For 



small displacements neutrality is obtained when ———. where 



p — ab,\^ the length of the main pendulum, and / the length of 

 the inverted pendulum. 

 The whole is carried on a tripod about 2 feet 3 inches high, 



S. 



Fig. 4. 



Fig. 5- 



Ewing's duplex pendulums, with the exception of their amplitude, 

 in no way resembled the table motion. This instrument was ad- 

 justed to have extremely feeble stability. With a second of 

 Prof. Ewing's instruments, which was adjusted by Prof. Sekiya s 

 assistant, who understood the machine, the distortion was not 

 so great, but the diagrams were complicated by the swinging of 

 the pendulum after the shaking had ceased. The pendulum m 

 this instance had a period of about two seconds, which was 

 much too short. John Milne. 



