Figs. 3 and 4 give the details of the screens. Fig. 3 represents a piece of board 1 inch thick and 6 or 7 inches wide, and rather longer than the width of the screen to be formed. Transverse grooves are cut at equal distances, something less than the diameter of the shot, as shewn in the diagram. Staples of hard brass spring-wire (No. 14 or 15), are fixed with their prongs in the continuation of the grooves. Pieces of sheet copper A are provided, having two elliptical holes, the distance of whose centres equals the distance of the grooves. The pieces of copper A are used to connect each wire steeple, as C, with its neighbour on each side. Thus, Fig. 4, a, c, e, g, &c., represent these copper connections put in their places and holding down the wire spring, which, when free, are in contact with the tops of the holes; but, when properly weighted, they rest on the lower edge of the holes. Thus the copper c forms a connexion between the staples b and d; the copper e joins d and f, and so on. A galvanic stream will therefore take the following course, whether the springs be weighted or unweighted: copper a, brass b, copper c, brass d, copper e, brass f, copper g, &c. The current will only be interrupted when one or more threads have been cut and the corresponding spring is flying from the bottom to the top of its hole. About th of a second is required for the complete registration of such an interruption, the spring traversing about half an inch. The shelf B is placed for the weights to rest against, partly to prevent them from being carried forward by the shot, but chiefly to prevent the untwisting of the threads which support the weights. The weights used were about 2 lbs. each, and the strength of the sewing cotton for supporting them was equal to a stress of about 3 lbs., which was sufficient to withstand a tolerably strong wind. As the weights were equal the threads were kept equally stretched. The arrangement of the screens for an experiment is shown in Fig. 5. The wires for conveying the galvanic current are, like the common telegraph wire, carried on posts. abc is a continuous piece of wire; but there are interruptions between e and h, between i and 7, between m and p, &c., in Fig. 5. a The order to make the galvanic current circulate through the screens. course of the galvanic current is + a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, P, q, r, s, t. The ends a, t, are connected with the instrument and battery. The shot, being fired through the screens, in passing cuts one or more threads at each screen, so that corresponding to the instant at which the shot passes each screen there is an interruption of the galvanic current, and a simultaneous record on the paper. When the cylinder is filled with spirals, that is after five or six shots, it is transferred to the instrument, Fig. 6, where a is a circle divided into 300 equal parts, and the division is carried to 3000 by the help of a vernier. A small T-square, having a fine edge at b, moves along a brass straight-edge L, adjusted so as to be parallel to the axis of the cylinder. The mark 6 is carefully placed opposite each record on the paper by means of a tangent screw (not shewn in the figure), and the vernier is read. Fig. 6. The clock goes on breaking the galvanic circuit every swing of the pendulum, whether the marker m be in contact with the paper or notconsequently, whatever be the loss of time in the action of the marker, we may fairly suppose it to be constant. But it has been objected that the current having been circulating through the screens for several seconds, or even minutes, without interruption before the shot is fired, the records at the first and the following screens are not made under the same conditions. A careful inspection of the results actually obtained shewed that this objection had no foundation to rest upon for that particular experiment; still it seemed desirable to get rid of the objectionable arrangement lest an irregularity might make its appearance in some important experiment. The first screen was placed 120 feet from the gun, the second at 240 feet, and so on. One preparatory break would be obtained by placing a screen at the mouth of the gun; but in addition to this it appears well to introduce the ordinary self-acting contact breaker, made to beat as near as may be the intervals of time to be measured. The raising of a spring lever interrupts the main current of galvanism through the screens. The insertion of a pin to keep up the lever, re-opens a passage for the screen galvanic current through the contact breaker; this may be made also to ring a bell in the instrument room, to give notice that all things are ready for the experiment. The fly-wheel is then put in motion, the signal to fire is given; the pulling of the lanyard withdraws at once the pin, restores the main current, and fires the gun. The construction of the chronograph was commenced in August, 1864; it was ready for trial in June, 1865. It received its first partial trial before the Committee on Gun Cotton in July, 1865, in conjunction with Major Navez's Electro Ballistic Pendulum. The instruments gave a nearly constant difference of 20 f.s. in velocities of about 1500 f.s., and indicated a much greater consistency in the results than in those obtained by M. Melsens when engaged in comparing the two instruments of Major Navez and M. Boulengé. The chronograph remained at the proof butts from July to November, 1865, when it was taken down to Plumstead Marshes and placed in a splinter-proof, where it remained about a fortnight. Its powers to withstand damp and dust were well tested in this manner. Eighteen shots were fired in all. Two of them were fired by mistake when the cylinder was stationary. One shot carried away a screen, another cut the conducting wire at the second screen, &c. I am able, however, to give a good account of eleven out of the eighteen shots fired on November 23, November 29, and December 12, 1865. Before proceeding to give an account of the experimental trial of the chronograph, it is necessary to explain at some length the mode of dealing with the clock and screen records, because the absence of any provision to secure uniformity of rotation of the cylinder has appeared to many a great defect. The fact that the variation of velocity of a pendulum, falling from rest through a considerable angle, used in so many instruments, is enormously greater, seems to have been entirely overlooked. The reason for not attempting to secure uniform rotation in the present case was this: the thing has been often attempted, and, so far as I know, never satisfactorily accomplished, when such small intervals of time as the one-hundredth or one-thousandth of a second were of importance. It is sufficient to refer to the chronograph constructed by M. Breguet for Colonel Konstantinoff. M. Gloesener appears to be satisfied with the approach to uniformity of rotation made by that instrument, and gives four experiments in support of his favourable opinion. It will be sufficient to copy the second experiment, which gives the times occupied in making five revolutions in succession: 2"-05 1"-95 1′′.90 1-80 1"-60 1"-90 1′′.90 1"-80 giving a mean time of revolution of 0"-362. As this result was very far from satisfactory to me, and I had no hope of succeeding so well as M. Breguet, it was quite plain that the difficulty of making the cylinder revolve uniformly must be avoided in the construction of a good chronograph for experiments on gunnery. As already explained, in most, if not in all instruments used in Observatories for recording observations by the aid of electricity, a marker connected with the clock records the seconds. If this were done, it is plain that the length of spiral described in successive seconds would give the rate at which the cylinder was moving. The introduction of a heavy fly-wheel would prevent sudden variations of velocity. The axis was placed in a vertical position, in order to reduce the friction as much as possible, and to avoid errors that might be caused by one side of the fly-wheel being heavier than the other. All unnecessary resistances, all disturbing causes must be removed. It may ! Traité général des applications de l'électricité, &c., I, 374. then be assumed that the cylinder will lose its angular velocity according to some law, which law can be determined from the intervals between five or six successive records of the beats of the clock. The weight of the marking apparatus and electro magnets does, in some measure, act as a driver of the fly-wheel and cylinder, and so tends to keep up the velocity; but it was not introduced for that purpose, and its effect is slight. The gradual loss of angular velocity of the cylinder gives a little trouble in the calculation when extreme accuracy is desired, but creates not the slightest difficulty to the mathematician. The subject now must be divided into two parts; (1) where a thoroughly trustworthy velocity of a projectile is required, and (2) where the utmost attainable precision is desired, with a view to determine the exact amount and law of the resistance of the air to projectiles of various forms and sizes. In the first case, the intervals between at least three successive beats of the clock should be measured, so as to extend beyond the screen marks on both sides. Three screens are the least that can be used with safety, which may be placed 60 or 70 feet apart, or at greater equal distances, if possible. By taking three records from the clock, and three from the shot, we have a complete check. If the records be found to be consistent, and if A, B, C, represent the beats of the clock, and a the position of the three screen marks, then the length AB must be taken to represent a second, and the A B C a с distance from the first to the third mark at a would represent, very nearly on the same scale, the time occupied by the shot in passing from the first to the third screen-the use of the middle screen being merely to test the reliability of the screen records. The only calculation is comprised in the proportionvelocity in fs: distance between first and third screen :: AB: distance between first and third record at a. If the three screen records fall at b near B, then (AC) must be taken as the representative of a second; or if they fall at c, then length of BC must be used to represent a second. The velocity so determined would be thoroughly trustworthy, and sufficiently accurate for all practical purposes. In the second case, where the utmost attainable accuracy is desired, it becomes necessary to have more numerous equidistant screens spread over a greater distance. The screens are placed at equal distances because that facilitates calculation, and enables us to judge whether the records are consistent. This principle should be carefully observed in all experiments. Thus, if it be required to find how far the initial velocity depends upon the charge of powder, vary the charge by equal quantities. If several rounds are to be fired with each particular charge, begin with the highest or lowest, and fire one round of each charge up to the lowest or highest, and so on over again till the experiment be completed. If it be desired to find range tables corresponding to various elevations of the gun, fire one shot at 0°, one at 1°, one at 20, and so on up to the greatest elevation, and repeat till the experiment is completed. If now the mean range for each degree of elevation be taken, these ranges will follow some kind of law, and after small adjustments have |