ON THE FACTS AND THEORY OF EARTHQUAKE PHENOMENA. 93 
Reverting now to the time balls B,, B*, those which, being left behind, record 
the instant of the arrival of the shock at the instrument,—it has been stated 
that we have no occasion to determine their subsequent movements ; it 
may be well, however, to clear our notions generally as to what these will 
be. Rotation is almost instantly communicated to these balls by adhesion 
with the moving planes on which they rest. The block moves off horizon- 
tally (in the direction of the wave) from the ball, which rolls thus with a 
retarded motion up the inclined plane in a relatively opposite direction. The 
block attains its maximum velocity V, and, coming to rest, reverses the direc- 
tion of its own motion, and now follows back after the ball that it had left 
behind, which it may overtake, and strike, with a relative velocity equal to 
the sum of its own velocity and that of the ball, or to their difference, depend- 
ent upon the state of motion of the ball at the moment of impact. The 
impact calling forth elastic force from ball and block, the former will be 
thrown up along the inclined plane; but the extent of this movement, or 
whether it occur at all, will depend upon the dimensions and velocity of the 
wave itself (resolved into the line of movement on the inclined plane) and 
upon the elasticity, &c. of the ball and block. These we have no occasion 
to pursue further: the actual movements of these balls, B, and B?, how- 
ever, will be found recorded in time also, by their own pencil-tracers on the 
cylinder; but the only indication that concerns us, is the first instant of 
broken contact, as already explained. 
A single seismometric observatory, such as has been now described, set up 
within a given region of disturbance, is capable of giving the elements, neces- 
sary for the calculation of the position of the seismic focus, but without the 
power of controlling the accuracy of the results, except in so far as coinci- 
dent repetitions may confirm or refute them. But if three such seismome- 
tric observatories be set up within the region chosen, in positions that shall 
form the angles of a triangle with respect to each other, at moderate distances 
apart (from 15 to 30 miles), and these be all connected by galvanic wires, 
so that the whole of their records shall be made upon a single paper cylinder, 
moved by a single clock in one of the three observatories, we then have a 
further control, and an independent method of obtaining, both the hori- 
zontal component of direction, and the surface-velocity, from which, by 
methods yet to be stated, the depth of origin may be calculated without 
direct ascertainment of the vertical component in Z. The cylinder must in 
this case carry twelve pencil-tracers, four leading from each observatory. 
This leads us to the second and somewhat simpler form of seismometer 
proposed by me, and shown in figs. 4,5, 6 and 7 (of Plate XV.). In some re- 
spects, the principles of this instrument are the same as of that just described : 
like the former, it is a double instrument, each instrument having two move- 
able balls; but their action is different. Fig. 4 represents, in elevation, one 
of these instruments (let us suppose, that N. S.) as seen looking eastward, 
and the upper part of which is seen in plan in fig. 5. ss is the floor of the 
observatory within which the two similar instruments are placed. ¢¢is a 
shallow and flat-bottomed dish or basin of some feet in diameter, and about 
nine inches in depth, formed by a circular wooden curb or rim secured to 
the floor. 
___ In the centre of this, there stands up vertically a very stiff pillar or upright, 
rigidly secured into the floor, and which may be either of hard stone, hollow 
cast iron, or of hard wood, but best of the second. Its upper end is formed 
of wrought or cast iron in the form shown; and into it are secured the vertical 
supports of hardwood, s,s, which are placed with their parallel and vertical 
axes in the plane of the meridian or at right angles thereto, and are prepared, 
