This exceedingly delicate timepiece is found in a falling weight. So long as the thing is so heavy that the air resistance is negligible, we can calculate with the greatest nicety how long a weight has taken to fall through a given distance. Near the muzzle of the gun there is set up a frame upon which are stretched a number of wires so close together that a shell cannot get past without breaking at least one of them. These wires are connected together so as to form one, and through them there flows a current of electricity the action of which, through an electro-magnet in the instrument house, holds up a long lead weight. At some distance away, say one hundred yards, there is a similar frame also electrically connected to an electro-magnet in the same instrument house. This second magnet, when energized by current from the frame, holds back a sharp point which, under the action of a spring, tends to press forward and scratch the lead weight. The third frame is likewise connected to a third magnet controlling a point similar to the other. To commence with, current flows through all three frames so that all three magnets are energized. The gun is then fired and immediately the shell breaks a wire in the first frame, cutting off the current from the first magnet and allowing the weight to fall. Meanwhile, the shell reaches the second frame, breaking a wire there, with the result that the second magnet loses its power, lets go the point which it has been holding back and permits it to make a light scratch upon the falling weight. This action is followed almost immediately by a similar action on the part of the third magnet, resulting in a second scratch on the lead weight. The position of these two scratches on the weight and their distance apart gives a very accurate indication of the time taken by the shell to pass from the first screen to the second and from the second to the third. From those times it is possible to calculate the initial velocity of the shell and the speed at which it will move in any part of its course. Indeed, with those two times as data, it is possible to work out all that it is necessary to know about the behaviour of the shell. This is rendered practicable by the fact that the moment the wire is cut the magnet lets go, no matter what the distance of the screen from the instrument may be. But for the instantaneous action of the current, allowance of some sort would have to be made for the fact that one screen is farther than another and the whole problem would be made much more complicated.