People often ask how quickly electricity travels, as if when we sent a telegraph signal along a wire a little bullet, so to speak, of electricity were shot along the wire like the carriers of the pneumatic tubes in the big drapers' shops. That is quite a misconception, for in reality the circuit of wire is more like a pipe full of electricity, and when we set a current flowing what we do is to set the whole of that electricity moving at once. If we think of a circular tube full of water with a pump at one spot in the circuit, we see that as soon as the water begins to move anywhere it moves everywhere. Moreover, if it stops at one point it stops simultaneously at every other point. While practically this is the case it is theoretically not quite so, for the inertia of the water when it is suddenly started or stopped no doubt causes a slight distortion of the tube itself resulting in a very slight (quite imperceptible) retardation of the movement of the water. Electricity also has a property comparable to the inertia which we are familiar with in the objects around us, and there is also a property in every conductor which to a certain extent resembles the elasticity of the water-pipe, whereby it may for a moment be bulged out. In a short wire, however (up to a mile or so), particularly if the flow and return parts of the circuit be twisted together, this electrical inertia practically vanishes and consequently we may say that for all practical purposes the current starts or stops, as the case may be, at precisely the same moment in every part of the circuit. That fact is of great value when, as in the case we are now discussing, we want to compare very exactly two events occurring very near together as to time but far apart as to place. [Illustration: BOMB-THROWERS AT WORK. Many kinds of bombs are used. One has a metal head and a handle about a foot long, with a streamer to ensure correct flight; another form resembles a brush when it is flying through the air; and a third, known as "the egg," is oval in form.] We need to compare the time when the shell leaves the gun with the time when it passes another point, say, one hundred yards away, and then again another point, say one hundred yards further on still. Supposing, then, a velocity of 3,000 feet per second, the time interval between the first point and the second and between the second and third will be somewhere about a tenth of a second. So we shall need a timepiece of some sort which will not only measure a tenth of a second, but will measure for us a very small _difference_ between two periods, each of which is only about a tenth of a second and which will be very nearly alike. That represents a degree of accuracy exceeding even what the astronomers, those princes of measurers, are accustomed to.