that they could not be recovered after the explosion. Here are the fragments of a shell of the same size exploded by 765 grains of gunpowder. The difference between the size and number of the fragments in the two instances is very striking. In the case of the fulminate of mercury the explosive effect is exerted almost instantaneously in all directions, and the shell is therefore shattered into a very large number of fragments, the force of the explosion being almost entirely spent on the bursting of the shell; while in the case of gunpowder, the explosion being comparatively gradual in its nature, the force developed is only partly spent upon the fracture of the shell, and is still in course of development when this result is produced; hence, not only are the resulting fragments much fewer and larger, but a considerable projectile force is exerted upon them after their production, and they are consequently scattered to a much greater distance than those produced by the employment of fulminate of mercury.

When gunpowder is confined by means of a shot or shell in the barrel of a gun, the explosion of the first particles has the effect of overcoming the inertia of the projectile, and the action proceeding gradually, as compared, for instance, to that of the fulminate of mercury, the shot is projected with comparatively small strain upon the gun; but in employing the fulminate of mercury, in a corresponding experiment, it would be found that the enormous force, which reaches its maximum suddenly, would, almost simultaneously with the first movement of the shell, also discover the weakest parts of the gun; in all probability, therefore, the cannon would be burst, while we should not project our shot or shell to any great distance. In quarrying, where it is desired to detach large masses of rock or stone, without producing any great destructive effect, we employ a very slow-burning powder, the explosion of which exerts a force like that obtained with a wedge; if we were to employ fulminate of mercury in this instance, the particles of rock in the immediate vicinity of the charge would become shattered by the sudden and violent explosion, and the desired result would not be produced.

The action of gunpowder, gradual though it appears to be when compared with the action of a fulminate, may, in particular conditions, and under certain circumstances, be much too rapid. Recent investigations of the effects of gunpowder have shown that the power we possess of modifying its action so as to render it more gradual is exceedingly important. As an illustration of this, I may state that in long guns, and in cannon of large calibre, the charge of the powder used for the projection of the shot has been clearly shown to be completely ignited before the projectile is moved to any great distance along the bore of the gun; hence, we find that whenever explosions do occur in guns, in consequence, for instance, of inferior casting or metal, or an excessive charge of powder, the fracture of the gun is almost always confined to the part reaching from the trunnions to the breech. The American Dahlgren gun, of which this is a model, exhibits this great thickness at the breech end; this form has been adopted to enable the weapon to resist the comparatively enormous strain exerted on that part by the heavy charges employed. Where cast-iron cannon are still used, it will always be especially necessary, if we employ

a rapidly-burning powder, to make the gun comparatively very thick from the breech towards the trunnions; and the production of strong cast-iron guns of this form is attended with very considerable difficulties; but if we use a slower powder, we can employ a cannon of more uniform thickness, as the strain exerted by the exploding powder is distributed much more uniformly throughout the greater part, if not the whole, of the length of the gun. Again, in rifled guns, in which, in consequence of the accurate fit of the projectile, the friction between it and the bore of the gun is very great, in some of which, also, as in Sir William Armstrong's gun, the projectile has to be expanded by the explosion into the grooves of the cannon, a gradually progressive action of the explosion is obviously of very great importance. In mortars and very short guns, on the other hand, where we have a very small space for the projection of the shell, it is necessary to employ a very rapidly-burning powder. It has been constantly observed that, in firing mortars with the description of powder used for cannon, a portion of the charge has been thrown out unexploded.

The explosive action of gunpowder may, however, be easily regulated in a variety of ways. We may do so, for example, by altering the composition of the powder. By increasing the proportion of the charcoal beyond that indicated by theory as the smallest quantity which will combine with the whole of the oxygen in the saltpetre, we decrease the rapidity of burning of the powder, simply because we give the saltpetre a larger amount of carbon to oxidize, which it will do less energetically; the oxidation will, therefore, take place less rapidly, and less heat will be developed. Then, again, we can modify the rate of combustion of the powder by altering the method of manufacture, - that is, by rendering the mixture of ingredients more or less perfect. And, lastly, we may modify the action of powder—and this is really the correct way of going to work without interfering at all with the proportions of ingredients as indicated by theory, or with the intimacy of mixture essential to their perfect action, by simply modifying the state of division of the material; that is to say, by employing various sizes of grains of gunpowder, and also by submitting it to different degrees of compression. A comparison of the rate of burning of two or three different samples of powder of the same composition, but varying in size of grain, will show you that this modification in the rapidity of action of gunpowder may be effected with very great ease and nicety without interfering in any way with the ultimate amount of force exerted by the powder. Large grains, or rather lumps, have been tried, and, by employing them judiciously, it is found that they propel shot to an equal distance, and with even greater uniformity, than ordinary cannon powder. By combining the application of uniform and accurately regulated pressure with modifications in the composition of gunpowder, and by confining the material within a case or receptacle, so that, when ignited, it can only burn in one direction, the valuable arrangements known as fuses and time-fuses are obtained, by which charges of powder in shells may be ignited at any period, within certain limits, determined upon previously to the loading of the cannon. By mechanical arrangements, which regulate the amount of com

pressed powder to be burnt before the flame reaches the charge in a shell, the time of explosion may be adjusted with great nicety, subject, however, to slight variations which depend, as Dr. Frankland has recently shown, upon fluctuations in the density of the atmosphere. The production of rockets, signals, and numerous pyrotechnic arrangements, is based upon the principles applied in these fuses.

Although the gradual action of gunpowder is, as we have seen, of the greatest importance in most applications, there are certain instances in which more rapidly combustible substances, or more rapidly explosive bodies, undoubtedly might be employed with advantage. This is particularly the case, for example, in mining operations, where it is mainly desired to produce great destructive effects by the explosion. This circumstance has frequently led to the trial, and even occasionally to the use, for a brief period, in actual practice of bodies more readily and rapidly explosive than gunpowder. Only one explosive compound, gun-cotton, has been put to the practical test, but trials have been made of a variety of mixtures, in all of which chlorate of potash is employed instead of saltpetre, in admixture with very oxidizable substances. The preparation of any of these substances upon a large scale has, however, always been sooner or later attended by disastrous results, which in many instances have not been simply due to carelessness. Examples of these mixtures are Callow's mining powder, containing orpiment, or sulphide of arsenic, mixed with chlorate of potash, and German or white gunpowder, which consists of prussiate of potash, chlorate of potash and

sugar.

We have some of these mixtures which are so prone to change as to be ignited instantaneously by contact with powerful chemical agents, such as acids, for example. Chlorate of potash is most readily acted upon by sulphuric acid, which not only decomposes the salt, but also transforms the chloric acid into very unstable compounds, and the heat resulting from the chemical changes thus established by the acid in a small portion of the mixture of chlorate of potash with an oxidizable body, such as sugar or sulphide of antimony, is sufficient to ignite it, and thus the whole is almost instantaneously exploded. Again, friction will ignite some of these mixtures very readily, as by rubbing together in a mortar a few grains of chlorate of potassa and roll sulphur. Even in the manufacture and employment of comparatively so safe an agent as gunpowder, which may be subjected without ignition to tolerably powerful friction or percussion, and to the direct application of any temperature below that which suffices to ignite sulphur (about 550° Fahr.), the neglect of strict precautions for excluding the possibility of a particle of the powder being subjected to sudden and powerful friction, may, and frequently does, lead to accidental explosions. The occasional accidents in gunpowder manufactories are generally enveloped in mystery, in consequence of their fearfully destructive effects; in all cases, however, where it has been possible to trace the causes of such explosions, they have been found in the wilful or accidental neglect of simple precautionary measures, indispensable to the positive safety of the works and operators.

The more highly explosive mixtures, and some few explosive com

pounds, though inapplicable as substitutes for gunpowder on account of their great sensitiveness to the effects of heat, have, in consequence of this very quality, received important applications in numerous ingenious contrivances for effecting the ignition of gunpowder. Well-known instances of such applications are, the employment of fulminate of mercury in percussion caps; of a mixture of chlorate of potassa and sulphide of antimony in arrangements for firing cannon by percussion and by friction, and for exploding shells by percussion or concussion; and of the same mixture exploded at will by being brought into contact with a drop of strong sulphuric acid, for the ignition of submarine mines or of signals.

Other mixtures, combining a high degree of explosiveness with power of conducting electricity, have been successfully applied to the simultaneous ignition of numerous charges of gunpowder by electricity of high tension; by means of one of them, recently discovered, many mines may be simultaneously discharged, even by the employment of small magneto-electric machines; the necessity for the employment of voltaic arrangements in mining operations being thus entirely dispensed with.

One of the most highly explosive mixtures at present known, consisting of chlorate of potassa and amorphous phosphorus, has been most ingeniously applied by Sir William Armstrong to the ignition of his time-fuses, and to the production of concussion and percussion fuses, remarkable for the great ease with which they are exploded. The above mixture may be ignited by the application of a gentle heat, or by submission to moderate pressure: if it is made up into a hard mass by mixture with a little shellac varnish, the friction resulting from the rapid insertion of a pin's` point into the material suffices to ignite it, even when it is well covered with varnish. Thus, in Armstrong's time-fuse, which, when fixed in its place in the head of the shell, cannot, like ordinary fuses employed in smooth-bore guns, be ignited by the flame of the exploding charge of powder (as the shell accurately fits the bore of the gun), the fuse composition is inflamed immediately upon the firing of the gun, in the following manner: A small quantity of the phosphorus mixture is deposited at the bottom of a cylindrical cavity in the centre of the fuse, and over it is fixed a small plug of metal, with a pin's point projecting from its lower end. This plug is held in its place by a pin of soft metal, which, by reason of the vis inertia of the plug, is broken when the gun is fired, and the pin then instantly pierces the pellet of detonating mixture, which by its ignition sets into action the time-fuse. The distance between the pin's point and the phosphorus mixture, before the explosion, is only one-tenth of an inch. This arrangement exemplifies in a striking manner the delicacy of action which may be obtained by a judicious combination of simple mechanical arrangements and highly explosive materials.

The variety of work accomplished by the explosion of a charge of powder in an Armstrong gun loaded with a shell no less than five distinct and important operations being thereby effected before the shell leaves the gun- affords a most interesting illustration of the progress made in the application of explosives, and of the compara

tively great control which may be exercised over the operations of those destructive agents.

REPAIRING OF THE GREAT EASTERN.

In a voyage from Liverpool to New York, in October, 1862, the steamship Great Eastern struck upon a rock off the coast of Long Island, damaging and crushing in the iron plates on the port side for a distance of some sixteen feet. Owing to the huge dimensions of the vessel, and her inability to enter any existing dry-dock, the problem of repairing a damage of such an extent, situated some twenty-five feet below the water-line, was one not only of great engineering interest, but also one of great difficulty. Before detailing the manner in which this was accomplished, it is desirable to recall briefly the peculiarities of the vessel's construction.

The hull is formed of two distinct vessels as it were, one inside of the other. These skins are stayed to each other by a number of webs or partitions, that divide the vessel transversely into thirty-four spaces; they run the whole length from stem to stern. The webs are further crossed at right angles by thirteen separations which constitute a system of water-tight cells, each of which is entirely independent of the other, access being had to each cell through manholes, provided with plates, that open into them. It must be borne in mind also that there are, inside of the ship proper, two upright iron bulkheads that divide the hull into three long rooms. Now the man-hole plates previously mentioned communicate with each other from the upper series of cells in the ship's broadside down to the foot of the bulkhead before mentioned. There they stop. The arrangement on the other side is of course similar. The inner room has two man-hole plates on the inner skin, which allow access to the cellular divisions situated beneath it. These are connected through one another by the same plan as the others. In brief, the Great Eastern is a ship built up of a series of rectangular pipes, independent of each other, yet capable of being connected together.

Let us now return to the subject of the disaster. The fracture was entirely through the outer plating of the ship, and extended over three of the longitudinal cells. To close up the sides by any other means than with new plates was simply impossible, and these had to be put on while the vessel was in the water at her anchorage. The stubborn broken plates with their ragged edges afforded not the slightest hint that could be seized upon to accomplish the work short of much time and labor. Preliminary consultations resulted in deciding the authorities to adopt the expedient of a dam which should inclose the point of rupture on all sides, and which, by means of pumps, could be freed from water and rendered habitable while the operations were in progress.

A coffer-dam was built of heavy oak timber, semi-circular in form, and planked outside four inches thick. It was ascertained that thirtytwo tons of iron would be required to sink the scow, and it was forthwith partially submerged, while two chutes, hereafter mentioned, were affixed. Previously, however, two heavy chains had been attached to each side of it, in such a manner that the cable, fastened on to the