6io 



NA TURE 



{Oct. 



purpose. This glass plane must be perfectly horizontal. 

 A circular disk of lead, g, is inclosed between glass planes, 

 d, and rests on three perfect spheres, /, which should be 

 preferably of glass or ivory. Rigidly attached at its centre 

 on the lower side is a conical spire, m, whose point 

 reaches just the level of the glass plane c, and has fixed 

 to it the silk thread n. From the centre of the upper side 



-Ball and plane seismograph fur indicating amplitude and : 

 all phases of a shock. 



projects the steel stylus o, bearing a fine platinum wire at 

 its extremity, which may be used with the double azimuth 

 circle of Fig. 3 in a direct manner in case of large earth- 

 quakes, or by a lever supported in gimbals, as also in 

 Fig. 3. Another glass plane, <•, with a large circular aper- 

 ture at its centre for the free movement of the azimuth 

 stylus o, has a wrought-iron trellis work backing h. This 

 reposes on the springs k, which are regulated by the 



Fig. 3. — Azimuth regist 





milled-head screws /, so that it is only pressing sufficiently 

 on the upper balls f to keep them in place. The upper 

 springs i are introduced to allow slight freedom of motion 

 to prevent breakage of the plates in almost vertical shocks 

 or from the expansion of the lead disk, balls, &c. The 

 silk thread is connected to the registering apparatus in 

 the same manner as in Fig. 1, the slight weight of which 

 will tend to draw back the rolling lead disk to its central 



position, and so prevent it shuffling out of its place, and 

 yet have almost no effect in modifying the register of the 

 absolute wave amplitude. 



In working over an earth-shaken district of small area, 

 such as that of Ischia, an error of observation of azimuth 

 of even a few degrees matters little in determining the 

 exact position of the epicentre. But on the contrary, in 

 large areas such as the Neapolitan earthquake of 1S57, 

 and to a far greater extent in widespread disturbances 

 such as the great Lisbon catastrophe, an error of a few 

 minutes of a degree is sufficient to produce great diver- 

 gence in the orientation of the azimuth and a consequent 

 incorrectness in the location of the epicentre. In most 

 seismographs so far employed, especially those of Italy, 

 no attempt has been made to divide the circle into eight 

 divisions, so that an error of nearly 45" could occur. 



Fig. 3 represents a separate apparatus, although it 

 would probably in practice be found preferable to replace 

 the pendulum by the rolling disk and balls as already 

 mentioned when describing Fig. 2, except that the contact 

 rings /; would then be inverted. A pendulum, a, with a 



Fig. 4. — Horizontal compo 



■e-path registe: 

 : earthquakes. 



length of suspension suitable to circumstances of observa- 

 tion, carries a steel spring wire, b, which slides in the 

 cylinder c, which, together with the light wire arm/, forms 

 a universal lever moving about the fulcrum at d, which 

 are gimbals. This lever should be so balanced that, if 

 placed in a horizontal position, the part above d should 

 counterpoise the part/ The lever carries a fine platinum 

 wire, g, which, when at rest, is the centre of the two con- 

 tact circles //, //, A". This part of the mechanism is in 

 connection with one pole of a battery. 



The contact circles A seen in detail, //, //, consist of a 

 suitable number of brass segments, /, which have a 

 V-shaped groove on their upper surface, and the edges, 

 both inner and outer, are bevelled off. Each one is insu- 

 lated from its fellow by the vulcanite plates m, which 

 project a little on the inner, upper, and outer sides, and 

 are sharpened to a knife edge. 



The registering apparatus consist of a number of long 

 soft-iron spring styles bolted to the column z', with their 



