ON THE TRANSIT-VELOCITY OF EARTHQUAKE WAVES. 213 



through the rheostat (R), by conducting wires with the coil of the electro- 

 magnet (a), and terminating at the e+ pole at the mercury-cup (n), which 

 was in connexion with the other or e— pole of the batterj'. 



The rheostat was adjusted so tiiat the resistance equalled that of the 

 conducting wires along the telegraphic poles between C and D, E (fig. 1, 

 PI. IV.). In this state of things, when the lever (m) of the chronograph 

 was pressed down, the disk (f) instantly commenced rotating; but directly 

 afterwards the electro-magnet (a), whose current was established by the first 

 movement, attracted the lever armature (e) through the disk, and the latter 

 was arrested by being gripped between the pole of the magnet and the 

 armature. The arc of the circumference of the disk then, at the centre 

 of the magnec-pole (i.e. with 6 inches radius), that was intercepted between 

 the marked spot (p) whence it started and that at which it was arrested, 

 became a measure of the time lost or elapsed between starting the chrono- 

 graph at the observer's station and making contact at the firing-battery in 

 the actual experiments. 



The arc thus intercepted was converted into time, from the descent of the 



jj 

 weight (W), by the common formula l=2L-, s being given and equal to yVth 



4 



the length in feet of the arc described by the circumference of the disk before 

 being ari'ested ; and this was capable of being controlled by measuring by the 

 chronograph itself the actual time of a given number of successive revolu- 

 tions, and parts of revolutions, of the disk. The total number of complete 

 revolutions made being taken by reckoning the coils wound off the barrel 

 upon a mean of ten experiments with this apparatus, the delay at the contact- 

 maker appeared to be no more than 0"'0l43, which converted into distance, 

 at the greatest transit-rate observed, gives a correction of 17'3 feet per second, 

 and at that of the least of 12*S feet per second, both additive. 



It may be remarked that the small error due to inei'tia, &c. in this apparatus 

 tends nearly to correct itself, the extremely small time lost at starting of the 

 disk being very nearly equalled by its tendency to be carried a little too 

 far by the velocity impressed. The whole inertia also of the disk, barrel, Sic. 

 was extremely small in proportion to the moving weight W. 



Another correction requiring to be attended to in these experiments was 

 the time of hung-Jire in the charge of the mine, that is to say, the time 

 required for the burning of such a portion of the whole charge of powder 

 as should be sufficient to rupture the rock around, and so start off from 

 the focus the Avave-impulse perceived in the seismoscope — in other words, 

 the time lost between the instant of first ignition of the powder, viewed 

 as simultaneous with that of making contact at the firing-battery B, and 

 the starting of the wave of impulse to be measured. 



In my ibrmer experiments at Killiney Bay, it will be recollected that it 

 was in my power to determine this experimentally and rigidly, the moderate 

 charges of powder there employed admitting of this, and that I found it 

 amount for 25 lbs. of powder to 0""050513, or to about ^-^\\x of a second. 

 Such is, in fact, the time that the full charge of a 68-pounder takes to burn. 

 But in the present case direct experiment was impossible, and the value for 

 this correction can only be approximately obtained by observing the time that 

 elapsed in some instances between the moment of making contact at B, and 

 the first great visible movement of rock at the iace of the heading. This 

 observation I made in three instances, noting the time by a delicately made 

 chronoscope, by M. Robert, Rue du Coq, Paris. The results gave 0"'05, 

 0"'04', and 0"-8 for the time of hang-fire respectively, noting from the first 

 visible movement of rock at the face of the heading. This would give a mear 



