420 EARTHQUAKE PHENOMENA. 



sucli that the set of cylinders when placed upright and equidistant 

 thereon shall have a space greater than the altitude between each. 

 Thus, if the cylinder of largest diameter have b = 0.5 of a foot, the 

 length of plank will, for a set of six, as in the figure, be about 12 feet. 

 These base planks being fixed level arid solid, the floor is to be levelled 

 up with dry sand to their upper surfaces, and the two sets of cylinders 

 adjusted to their places, one set running in an E. and W., the other 

 in a N. and S. cUrection, so that in whatever direction the horizontal 

 component of shock may move, the overthrown cylinders of one or 

 the other set shall fall transversely to the lengths of the plank bases, 

 and lodging on the sand-bed, remain exactly in the position as to azimuth 

 in which they were overthrown. If now a shock of any horizontal velo- 

 city, capable of overthrowing some of the cylinders, but not all of 

 them, arrive, it will throw down at once all the narrower ones, and 

 up to a certain diameter of base. For example, suppose a N. and S. 

 shock of such velocity as to overthrow W 6, W 5, and W 4, leaving 

 W 3, W 2, and W 1 standing, then V will have been greater than the 

 velocity clue to the overthrow of W 4, and less than that due to the 

 overthrow of W 3, and, within those limits, may be found from the 

 preceding equation. The cylinders here overthrown, W 6, W 5, and 

 W 4, will be found with their axes lying N. and S., at rest upon the 

 sand-bed. The cylinders N 6, N 5, and N 4 will be also overthrown; 

 but in this case they will fall in the line of their own plank basis, and 

 may roll, and so give no indication as to direction of shock in azimuth. 

 Hence the necessity for two sets of cylinders. One set, however, will 

 be sufficient, if space enough be provided between the cylinders, and 

 each be placed upon a cylindrical and separate basis of a diameter 

 equal to its own, and in height equal to the depth of the sand-bed.. 



This form of instrument, then, is capable of giving approximate 

 determinations of — 



1. The velocity of the horizontal component of shock, neglecting 

 the vertical component, which may be done where the angle of emer- 

 gence is not great. 



2. The surface direction in azimuth of the shock, or direction of 

 horizontal component of the seismic wave. 



3. Its absolute direction of primary movement, viz: the direction of 

 translation of the wave, which always coincides with the direction of mole- 

 cular displacement of the wave itself in the first half of its complete phase — 

 e. g., if a shock in N.S. azimuth throw the cylinders to the southward, 

 then the wave has traversed from S. to N. 



4. The exact time of the transit of the shock at the instrument may 

 be also indicated, if either the narrowest cylinders, N. 6 and W. 6, 

 (which by hypothesis must be always overthrown,) be connected with 

 a house-clock in the way about to be described, so as to stop it at the 

 moment of overthrow, or, still better, if a separate cylinder of even 

 less stability be appropriated to this purpose. 



Three such sets of instruments at distant stations may of course be 

 easily connected by galvanic wires, so as to give the transit time at 

 each accurately, and hence the transit rate. 



Three or more distant observers, with chronometers, may of course j+ 



