One who gives the matter any consideration must be struck by the fact that the effects of high explosives are greatly exaggerated, not only in the popular mind, but almost universally.
The idea that such public buildings as the Capitol at Washington, and the Houses of Parliament at London, can be destroyed by the explosive carried on the person of an anarchist, can hardly gain the credence of an intelligent man, although such is the apparent belief of many newspaper writers, but the equally unfounded idea that a battleship can be destroyed by a high explosive shell meets with common acceptance. I have heard officers of our own navy argue against the building of large armored vessels on the ground that the coming use of high explosive shell will render them worthless, and the Navy Department is deluged with letters from all over the country, and from abroad, proposing schemes for firing high explosive shell, the writers invariably claiming that a single shell on their system will destroy an enemy's ship. Moreover, this opinion is apparently held by numerous writers of authority. For example, Mr. H. W. Wilson, in his interesting book, "Ironclads in Action," says, referring to the dynamite cruiser Vesuvius: "The perfecting of the pneumatic gun would be the death-knell of the battleship, and it is hard to see what protection could be devised against its bolts." The same writer speaks disparagingly of the use of above-water torpedoes on cruisers, considering them a greater danger to the vessel carrying them than to her opponent, and that this view is quite generally held in our navy is shown by the outspoken statements of commanding officers and their requests, in some instances, to the Department, for permission to have the torpedoes removed from the vessels under their command.
Under these circumstances I have thought that it might be of interest to present to the readers of the Institute the conclusions which I have formed in regard to the value of high explosives in naval warfare, though I fear that these opinions are too contrary to those commonly held to command ready acceptance.
In the first place, it is well to try to form a clear idea of just what an explosion is, and, for our practical purposes, we may properly say, that an explosion is the sudden conversion of a solid or liquid substance into gases of high temperature. But there are different orders of explosion, depending entirely upon the rapidity of this conversion. If you set fire to a mass of gunpowder in the open air it explodes, but the process of its conversion into gases is so slow that before it is completely consumed the gases have expanded and occupy a much larger volume than did the original powder. If, on the contrary, you detonate a block of guncotton by means of fulminate of mercury, you thereby instantaneously convert ·the guncotton into gases which, having no time to expand during their formation, occupy only the same volume the original solid did. In each case the process is the same, but the difference in the time element produces a vast difference in effects. The explosion of low order, such as that of gunpowder, producing an effect like a push, where the explosion of a high order, such as the detonation of guncotton, produces an effect like a blow on the immediately surrounding objects. Now what we mean by a high explosive is an explosive which can, by proper means, be detonated, or instantaneously, before it has time to expand, converted into gases, and a very important question to consider is, how do the effects of such high explosives, when detonated, compare with the effects of gunpowder?
First, however, let us see what other differences exist between high explosives and gunpowder, besides the peculiar capacity of the former to be detonated, and, also, what differences there are between the principal high explosives themselves which have led to the unanimous selection of guncotton for naval uses.
In what follows the term density of charge refers to the ratio between the weight of an explosive and the volume of the space in which it is contained. If an explosive occupies a space which an equal weight of water would just fill, it is said to have a density of charge of unity, or since 27.73 cubic inches of water weigh one pound, we say that the density of charge is one when the explosive occupies a space of 27.73 cubic inches per pound. If there be 4 x 27.73 cubic inches, or 110.92 cubic inches, to each pound of explosive, we say the density of charge is ¼, etc.
Now, gunpowder burned in a closed chamber at a density of charge of 4/10 gives a pressure of 10 tons per square inch, while guncotton burned at the same density, by reason of the much greater volume of gases produced, gives 39 tons per square inch, and at a density of charge of unity, while gunpowder gives 42 tons per square inch, guncotton, if contained in an unyielding chamber, would give over 400 tons per square inch. But it has been proven experimentally that the products of detonation do not differ sensibly from those of simple combustion in a closed vessel with high density of charge. Moreover, experiments have shown that the velocity of propagation of detonation is about woo meters per second for nitroglycerine, 3000 to 4000 meters for dynamite, and 5000 to 7000 meters per second for guncotton. Imagine, then, a cubic foot of guncotton, having a density equal to that of water, detonated by a fulminate capsule placed on its top. In 1/20,000 part of a second from the beginning of the resulting conversion of the cubic foot of solid into gas, it will have all been converted into gas, which, before it expands, exerts a pressure of over 400 tons per square inch in all directions. Such an enormous pressure must evidently produce the deformation, rupture, and pulverization of materials placed in contact with the explosive. No containing vessel could possibly be made to resist guncotton detonated in its own volume, unless the weight of the explosive were so small that the surface in contact with it conducted off its heat rapidly enough to prevent full development of its explosive force. Of course, as its volume is diminished, the surface of the chamber would have a greater and greater ratio to the weight of explosive it contained, and consequently would more and more decrease the force of the explosion by the absorption of its heat, but the effect of the detonation of any reasonable quantity of guncotton in a closed chamber of its own volume, must inevitably shatter and pulverize the· containing walls. With gunpowder, on the other hand, there is no difficulty in constructing containing vessels which it cannot rupture, as instanced by our armor-piercing shell, which have to be loaded with a more powerful explosive than powder in order to be burst.
The difference, then, between detonated guncotton and exploded gunpowder is not only in the instantaneous conversion into gas of the first and the comparatively slow conversion of the latter, but also in the greater volume of gas given off by the former. One pound of gunpowder produces 7740 cubic inches of gases at atmospheric pressure and 0° C., while one pound of guncotton produces 23,810 cubic inches of gases, or three times as much, under the same conditions, about 14 cubic feet against about 4 ½.
Comparing guncotton with nitroglycerine, next, we find the latter somewhat the more powerful, giving off a less volume of gases, but at a higher temperature, and also, on account of its higher density, being capable of confinement in smaller space than an equal amount of guncotton.
Guncotton, however, is much preferable for naval use to nitroglycerine, dynamite, or any other of the high explosives, on account of its stability and safety in the wet state, and the fact that it cannot be detonated except by fulminate of mercury. Most high explosives, and especially nitroglycerine compounds, can be detonated by shock or heat; guncotton cannot be. It can be set on fire; though, when wet, this is almost impossible, but it only burns freely unless strongly confined. If confined and set on fire, of course the pressure rises until the surrounding walls yield, and if they are strong, an explosion is the result, but numerous experiments have shown that for the detonation of guncotton nothing but fulminate of mercury will serve. When wet, guncotton of good quality will keep indefinitely, with absolutely no change in characteristics. In this state, or when frozen, it can be detonated by the use of a primer of dry guncotton which is itself detonated by fulminate. No other means of detonating wet or frozen guncotton than by this combination of dry guncotton and fulminate has yet been found. Moreover, when wet, guncotton is almost completely inert.
That it is possible to explode wet guncotton under peculiar circumstances we have abundant evidence, but it is only under special conditions that this can occur, and as these are well set forth and explained in the Report of Her British Majesty's Inspector of Explosives in 1881, I shall quote from them as follows:
"It may at first sight appear remarkable that an explosion of wet tonite should be possible. But apart from the well-known fact that guncotton (of which tonite is only a variety) is frequently employed in the wet state, the initiative explosion being established by a detonator and small primer of dry guncotton, we have the evidence of a not inconsiderable number of accidental explosions of this class of material in a wet state as evidence that such accidents are possible.
"In 1878 no less than eight explosions of wet tonite during pressing occurred at these works; in 1879, there was another similar explosion, and another in 1880, and there have been four (exclusive of the present one) during the present year.
"Then we have had two accidents at the Government Factory at Waltham Abbey-one in 1873, and one in 1877; one at the Royal Laboratory, Woolwich, in 1875; and the ignition of a wet guncotton slab while sawing at the same place in 1877. There were also at one time not infrequent accidents of this sort at Stowmarket; and from some documents which have been furnished to me through the courtesy of M. Maurouard, Inspector General of 'Poudres et Salpetres' in France, it appears that three explosions have occurred during the pressing of wet guncotton at the factory at Moulin Blanc, viz.: on the 4th September, 1878; 11th July, 1879; and 8th September, 1879.
"All these accidents may no doubt be referred to the same primary cause, although brought about perhaps in different ways in each instance. On this subject, I will quote the explanation which we offer in our Annual Report for 1878, in special connection with the eight wet tonite accidents of that year -
"'This cause may be stated to be that some small portion of the guncotton is, during the process of pressing or sawing, exposed to the heating, and, therefore, drying influence of friction or pressure. If the friction or pressure be of such a character, or be so maintained upon any portion of the guncotton, as to dry the particles upon which it is exerted sufficiently to render them inflammable, the continuing friction or pressure will, as the next step, establish the actual inflammation of the particles. In the case of the guncotton which took fire in the process of sawing, it is reasonable to suppose that some dry or nearly dry guncotton sawdust had accumulated under the bed of the saw, or in the immediate neighborhood of the point where the inflammation became first established, and this, in its turn, caught fire and produced the accident. In the cases of accidents in the pressing of wet guncotton, whether in hydraulic or steam presses, or in the charging of a shell, if we once assume an ignition of a particle, the resulting explosion can be easily accounted for. The heat generated by the burning of the ignited particles would gradually dry the contiguous portions, arid the inflammation would extend with progressive energy until the pressure of the accumulating gases overpowered the resistance of the press, or shell or piston, and (in the absence of some safety-valve or easier method of escape) determined the violent disruption of the containing case. As to how the undue friction becomes first established, that is a point which would have to be determined for each case separately. It may arise from two metal surfaces with only particles of guncotton between them bearing upon one another, in consequence of some defect in the original construction of the machine or tool, or from failure in their application; or, it may be due to the presence of some grit or other rigid substance upon one of the metal surfaces with only a small portion of guncotton interposed. But however the accident may originate in each particular instance, it is, we believe, beyond a question that the rationale of such accidents is to be found in the explanation which we have given above. And it is important to call particular attention to this class of accident and nature of risk, because there is too common a tendency to assume that so long as substances of this sort are wet, nay, that even if they have been at any time wetted, they remain harmless until they have been formally dried, and thus precautions which would certainly not be neglected with the dry explosive are deemed unnecessary so long as the material is in what is assumed to be a damp condition, while arrangements which may be necessary for preventing any portion of the inflammable substance from resuming its inflammable condition during the operations of manufacture are overlooked or disregarded.'
"It will be observed that this explanation involves no suggestion of a detonative explosion having occurred, it simply assumes inflammation first of one particle and then in very rapid succession of other particles, and when such inflammation becomes established under pressure and the gases are unable to escape, an explosion is not only a natural but a necessary consequence. Where, as in the case of the slab of wet guncotton which ignited during sawing, the gases resulting from this inflammation are free to escape, the result is confined to ignition only and does not proceed to explosion.''
The foregoing quotation is important because of the very common belief, especially among those accustomed to the handling of guncotton, that it will stand anything if wet. It is true that wet guncotton can only be exploded by the combined use of dry guncotton and fulminate; it cannot be set on fire even, while wet-but it must never be forgotten that it is possible to dry and then explode guncotton by a single action. For example, wet guncotton will stand almost any pressure slowly applied, but apply a very high pressure very suddenly, and the resulting heat may first dry and then explode it. It is also important to point out that there is a great difference in guncottons with different percentages. of moisture. For instance, guncotton containing 15 per cent of water or less can be detonated by fulminate of mercury alone, but if it contains 18 per cent or over, it requires a dry primer as well as the fulminate. Again, perfectly dry guncotton will be detonated with certainty by 2.83 grains of fulminate in close contact with it, but if only atmospherically dried, it takes 5 grains for certain detonation.
Other interesting facts about guncot.ton are the following:
A dry primer detonated in close contact with wet guncotton will detonate any quantity of the latter, however large and however wet. Saturation, instead of lessening, seems to increase the effect of detonated guncotton. No increase of violence can be obtained by increasing the quantity of fulminate or of dry guncotton. A mass of wet guncotton will have just as great an effect when exploded by a 3-ounce dry primer and a 3-grain fulminate igniter, as if 10 pounds of dry cotton and 100 grains of fulminate were used to initiate the detonation. Close contact, however, is essential when such small quantities are used; the 3 grains of fulminate must not be separated from the dry cotton by more than the very thinnest film of metal, and the dry cotton must have the same close relation to the wet guncotton. No confinement of either wet or dry guncotton is needed for detonation, nor does confinement add in the least to the violence of their explosion when detonated.
Guncotton then, on account of its good qualities, having been adopted by all nations as the high explosive for naval warfare, we have to consider to what uses it is put and whether it can advantageously be used in other ways.
Undoubtedly, in its present use for torpedoes and for destroying obstacles of all sorts, it is much superior to gunpowder. As a submarine mine it will do the work of twice or three times its weight of gunpowder, and by means of two or three small blocks of it, heavy chains, steel rails, booms or other similar obstructions, can be cut in two, where they would be uninjured by large powder charges. But there is one use for high explosives in naval warfare which is not yet developed, but upon the development and towards the introduction of which into general use thousands of people, mostly cranks, but still many intelligent and sensible people, are laboring. I refer to the use of these explosives as the bursting charge of shells, and it is especially to the consideration of whether or not such shell would be more effective than the projectiles in present use, and, if so, how much more effective, that I shall now ask your attention.
What has already been said has shown the first and fundamental difference between the effects of detonated guncotton and of exploded gunpowder in ordinary iron or steel shell filled with them. In the first case, the shell is much more completely shattered and its small pieces are propelled with greater velocity than in the second case. It is this action which makes the shell or bomb containing high explosive so destructive when detonated in a crowd of people; there are so many more pieces than would be given by a powder-exploded shell that more people are killed or wounded.
But granting the fact that the high explosive produces a much more destructive effect upon materials in close contact with it, we have to consider how its action upon its less immediate surroundings compares with that of powder. The following experiment, made at the N. P. G., illustrates this point. Two similar wooden chambers were built, each 10 feet square and 8 feet high, with 6-inch timber posts at ,corners, sunk 2,0 feet in the ground; other 5-inch posts at the middle of the sides, also sunk 2½ feet in the ground; cross-timbers at top, bottom and middle of these posts and secured to them by 6-inch iron spikes, and sides and top from 2 inches to 3 inches thick securely nailed to the framework. In one of these a 4-inch forged steel shell filled with 2¾ pounds of shell powder was burst, and in the other a similar shell filled with 1.9 pounds of guncotton. In each case the shell was placed on a block of oak 1 foot square and 2 feet 7 inches long, in the center of the floor of the chamber; a square yard of cotton sheeting was hung on the wall 5 feet away, and an old sheepskin sponge head was placed 6 feet above and directly over the shell. The result of the powder explosion was as follows: The oak block had a small dent where the base of the shell stood; the cotton sheeting took fire; part of the wool was blown off the sponge head but not set on fire; the chamber itself was badly wrecked; its rear wall was forced outward bodily and pierced with five holes; the three sides and door were blown out and badly splintered, some pieces being blown 25 feet away; the heavy door was blown 10 feet away and fell on its face, pierced in two places; the sides were badly splintered and broken up, many planks showing marks of small pieces of shell; the front half of the heavy roof was lifted and fell on top of the rear half, one plank being lodged in a tree nearby; the largest piece of the shell found weighed 4 pounds; some pieces struck the roof, and had the building stood it was evident that the dispersion of fragments would have been found to cover a large area in all directions.
The result of the detonation of the guncotton shell was as follows: The oak block was split; nothing was set on fire; the chamber was but little damaged; there was one hole in the roof; the back was blown outward bodily, as in the powder explosion, and had 17 small holes in it; one side had a piece 4 x 7 feet blown out of it, pierced by 4 holes, and its remaining area had 14 holes in it; the front was pierced by 19 holes; a small panel in the door was either blown out or carried away by a fragment; the left side was pierced by 12 holes.
Thus we see illustrated the peculiar difference in effect of the two explosives-the guncotton produces much more violent local effects; breaks the shell into more numerous fragments, and projects them with greater velocity; but its racking effect is much less than that of the powder; the energy of the high explosive is much the greater; but it expends itself largely in rending and shattering materials in close contact with itself.
There would not, then, seem to be any special advantage to be gained by the use of a high explosive as the bursting charge of common shell of small caliber. In order to detonate such shell, fulminate igniters are absolutely necessary, and if wet gun cotton is used, a dry primer is also necessary. It is evident, on the face of the thing, that of a shell containing a high explosive coupled with a detonator of five or more grains of fulminate of mercury is a much more dangerous article to handle and fire than is an ordinary powder shell, and to justify the use of the former in place of the latter, considerably greater destructive effect must be demonstrated. But this does not exist under ordinary circumstances of naval warfare. The high explosive shell of small caliber, if fitted with an ordinary detonator, will explode harmlessly outside a ship's plating, or, at the best, after partly penetrating it, in which latter case a great part of the fragments would be harmless. If, on the other hand, fitted with a delayed action detonator, which is perfectly feasible, and made strong enough to penetrate moderately thick plating, the capacity of the shell is thereby reduced and the small quantity of high explosive which it contains will be unlikely to cause any greater damage than would result from the same shell filled with powder; if it burst in a group of men, it will probably injure more, but its general effect will be less. It is important to understand in this connection, that if a shell is not strong enough, or has not energy enough to get through plating, the fact that it detonates on impact will not add, but rather will detract, from its effect. For example, there were fired several years ago at the N. P. G. a number of 6-inch guncotton shells, and one of them against a 7-inch armor-plate. The plate, which was of nickel-steel, was set up on edge and supported by timber struts from the rear, being very much less strongly held than it would have been on a ship's side. The shell, weighing 78 pounds and containing 12 pounds of guncotton, including the dry primer, struck about 2 feet from bottom and 3 feet from edge, where the plate was 6 inches thick, with a velocity of 1225 f. s., and detonated on impact, being broken into innumerable small fragments. The plate itself was neither knocked down, cracked, nor injured in any way, except for a few scratches and a slight depression made by the point of the shell. In fact, I was satisfied at the ·time that an empty cast-iron shell with same striking energy would have done fully as much damage, and probably more,
We may conclude, then, at least as far as shell of small caliber are concerned, that to be of any real value in naval warfare, they must be strong enough to go through the plating they are likely to be fired at, and that, if so, their destructive effect would be but very slightly, if at all, increased by loading them with a high explosive, instead of with ordinary black powder.
When, however, we extend our inquiry to the field of large caliber shell, and especially such shell as are designed to carry very large quantities of high explosives, we are met by the fact that this field is but ill-explored. Almost nothing is known of the effects of such shell, and it has been hitherto almost wholly from the study of the results of accidental explosions that we have been able to judge them. Now there are two theories, so to speak, upon which the use of high explosives, as the bursting charges of large shell, may be advocated. You may say: although it is desirable, viewing the extent to which ships are covered with armor, to make all large-caliber shell of sufficiently thick walls to at least penetrate the thin armor, 4-inch, 5-inch or 6-inch, of an opponent, still the effect of such shell, loaded with a high explosive, will be vastly greater than if loaded with powder. Or, second, you may say: let us not try to pierce the armor of an opponent, but rather let us throw against him such huge masses of high explosive that he will be destroyed by their detonation; his sides blown .in and his crew killed by the concussion. The second of these is, of course, the theory usually held. To combat the first theory we must show that to substitute 20, 30 or 40 pounds of high explosive with a detonator, for about an equal weight of gunpowder, will not cause any marked increase in the destructive effect of the explosion of a large-caliber, thick-walled shell. To combat the second theory, we must show that the effects of the detonation of 200, 300 or even 400 pounds of a high explosive on the outside of a ship, would be by no means so terrific as is usually imagined.
Let us see, first, what past experience can tell us of the effects of the explosion of moderate quantities of high explosives.
On February 26, 1884, an infernal machine, containing 21 pounds of dynamite, about the equivalent of the same weight of dry guncotton, was detonated in the baggage-room of Victoria Station, at London. This station was a wooden structure, with light wooden partitions, and was very much damaged, especially in the vicinity of the place of explosion, but the local effect is well illustrated by the fact that the brick wall of the London, Chatham and Dover Station adjoining was uninjured. This wall, 18 inches thick, was separated 2 feet 4 inches from the light wooden side of the Victoria Station, and the explosive was about 3 feet inside the latter station, so that it was only from 5 to 6 feet from the brick wall. Similar infernal machines had been placed at the same time in the Charing Cross, the Paddington, and the Ludgate Hill stations, but they all failed to explode.
On March 15, I883, an infernal machine containing, probably, 28 pounds of dynamite, was exploded at the Government Offices in Whitehall, London. The explosive had been placed on the sill of a window, inside the stone balustrade of a room near the center of the basement floor of the building, which is 4 stories high and of about 280 feet front. Practically all the windows over this whole front were broken by the concussion, but the serious damage was but slight. The official report says: "But the structural effects of any importance were, it will be observed, of an essentially local character, having, in fact, been limited to the room and balustrade where the explosion occurred; to some slight cracking of the masonry above and on either side, and to the injury to the floor of the room immediately above. If the intention was 'to blow up' the Government Offices, the result fell almost ludicrously short of the conception."
In I884 Commander Folger made a series of experiments at the N. P. G. to determine the effects of high explosives against the sides or decks of ships. The following are extracts from his report on the subject. First, as regards the explosion of dynamite against vertical plating:
"In this experiment two targets were used. The first consisted of nine 1-inch wrought-iron plates secured by nine 1-inch bolts to 20 inches of live-oak backing, which last was well braced in the usual manner.
"This target did not by any means represent in its capacity for resistance the side of a modern armor-clad, the plating being laminated, and the separate sheets greatly weakened by numerous perforations. They had been manufactured as flooring plates of a monitor's turret, the holes serving for purposes of ventilation. The bolting, also, was defective, being insufficient in quality, in weight, and of poor material.
"Two hundred and sixty-five pounds of dynamite were exploded in close proximity and in contact with this target in ten successive' charges, varying in weight from 5 pounds to 75 pounds.
"The injury done the target was hardly noticeable after 70 pounds had been exploded at the same point in six successive charges, and was limited to an inappreciable indentation of the plating, the rupture of one bolt at a defective weld, and a slight springing back of the structure.
"With the so-pound charge the indent deepened to 3 inches with a diameter of 2 feet 6 inches. Five bolts were broken, and the plates bending forward were slightly separated at the edges.
"With the charge of 75 pounds (in contact, as were all charges above 20 pounds), the indent was deepened to 7½ inches, including the spring outward of plate edges, with a diameter of 3 feet. Four bolts retained fair holding power. The structure had sprung back 2 inches, and had recovered itself in the elastic earth. The middle upright of the backing was somewhat shattered. The outer plate was cracked a distance of 2 feet at the indent in the lines of the perforations.
"While the target was considered as not having sustained vital injury, its weak construction, in particular as regards bolting, was apparent, and its condition unsuitable for the prosecution of the experiments with still heavier charges. The plating was therefore removed, the injured backing timber replaced, and the bracing strengthened by a second series of support.
"As plating, eleven 7/8-inch sheets, which having been manufactured for the sides of an armored vessel, possessed a slight crown, were secured to the backing, using fourteen 1.5-inch bolts of wrought iron. The plates were likewise somewhat weakened by old bolt-holes, though less than those of the previous target. The completed structure, however, though still deficient in strength, represented more nearly the conditions sought for.
"A charge of 100 pounds of dynamite, suspended as in all previous cases of contact so that its weight assisted the bearing pressure, and with outside ignition-85 grains of fulminate in two primers-was exploded against the target with the following results:
"Depth of indent, 2½ inches; diameter, 2 feet. The edges of the plates at the ends were sprung forward about 2 inches, stripping the beveled bolt-heads through the outer plate. The backing structure had sprung back about 2 inches, and had recovered itself. None of the plates were cracked.
"In all of the previous experiments the bed logs and earth in front of the target were shattered and torn, a hole in the latter of a depth of perhaps 3 feet being made at the explosion of the charge of 100 pounds. It being apparent that the angle between the ship's side and the water surface offered conditions favorable to the development of more violent action of the gases, a structure representing the water surface-a 1-inch plate upon logs-was built up to the level of the center of the target, and a charge of 75 pounds was pressed firmly into the angle, and exploded, again using two primers containing 85 grains of fulminate,
"The effects of this charge were notably more decided than those of the preceding one of greater weight, the indent being deepened to 3.0 inches, and the backing and bracing considerably racked and strained. The injuries, however, could not be considered as vital, and the target was left standing without repairs of any description for further experiments. The water surface structure was, of course, demolished, the iron plate being blown to fragments and the logs almost pulverized.
"The conclusion to be drawn from the results obtained is, that a modern armor-clad will not receive material injury by the explosion in superficial contact with iron over water plating of very large charges of dynamite."
Second, as regards the explosion of guncotton on horizontal plating:
"In constructing the target for this experiment, an endeavor was made to represent the strength of 2 inches of steel suitably supported.
"Three perforated wrought-iron plates, similar to those already described in the dynamite experiments, were joined firmly together, using eight 1-inch bolts. The three-fold plate thus formed was bolted to heavy timbers distant from each other about 4½ feet. Two 1-inch bolts were placed at each end, passing nearly through the timber. To eliminate, as far as might be, the effects of the perforations, sheets of ¼-inch iron were inserted on each side of the middle plate, while a third was placed upon the target to receive the charge. There was thus a total strength (nearly) of 3¾ inches of iron.
"Twenty-five pounds of wet guncotton were exploded upon this target, using for ignition (placed above) 20 ounces of dry cotton with 85 grains of fulminate.
"A hole, 7 by 8 inches, about one-half the area of the base of the charge, was blown through all plates with the effects of the explosion reaching a depth of about 18 inches in the rather friable earth beneath. The bend of the plates extended to the timbers, without starting the bolts, with a depth of about 70 inches."
Recently the Ordnance Department of the army made the following experiments at Sandy Hook: A 12-inch steel shell, containing 78 pounds of emmensite, was suspended vertically with its point in contact with a horizontal 3-inch steel plate resting upon timbers about 2 feet apart on the ground. Sandbags were piled around so as to enclose shell and plates. The detonation of this shell produced practically no effect upon the plate, but the sideways blast was very violent, throwing the sand-bags in all directions and to considerable distances.
A similar shell loaded with smokeless rifle powder, suspended in the same way, and exploded by simple ignition, produced fully as much effect as the emmensite shell, and in neither case were the results, except as regards greater fragmentation of the shell itself, any greater than would have resulted from a black powder charge.
All of these examples lead to the same conclusion-that the effects of the detonation of moderate quantities of high explosives are very great as far as immediate surroundings are concerned, but are comparatively slight otherwise. In the cases of the infernal machines and of Capt. Folger's experiments, the high explosive charges were not confined in strong-walled shell, and, consequently, their energies were not expended in shattering such shell, and they did more damage to surrounding objects than would otherwise have been the case. The army test of I2-inch shell, first loaded with emmensite and detonated, and then loaded with powder and exploded by ignition, gave a more direct comparison than we have elsewhere, and this surely does not indicate any greatly increased destructive effect from the high explosive. The truth is, that people underestimate the violence of gunpowder shell explosions, and exaggerate the others.
What the effects produced by bursting a 13-inch forged steel shell with gunpowder are, may be judged from the following test made at the N. P . G. in May, 1895:
A closed triangular chamber 70 feet high was made with three armor-plates as the sides, 18 inches, 18 inches and 12 inches thick respectively, and the two former weighing, with their backings, about 43 tons each. The two perpendicular sides of this triangle were 15 and 70 feet long respectively, and the short side was formed by the 12-inch plate, which was securely backed and supported by heavy braces and struts from behind. The shell, loaded with 53 pounds of ordinary shell powder, was placed on its side on a 1½-inch plate about 4 feet square on the ground in the middle of the triangular space. The top of the enclosed space was covered by three 3-inch plates, each 100 x 5 feet, placed side by side. On explosion, the shell burst into many pieces, sixty being recovered, amounting in weight to three-quarters of the total. One of the 43-ton sides of the chamber was moved bodily to the left, its outer end 8 feet and its inner end 3 feet; the other 43-ton side was moved bodily one foot to the right; the three covering plates, which weighed 3 tons each, were thrown into the air, one landing 6 feet away on top of the adjoining butt, a second falling back outside, and the third falling back inside the space enclosed by the vertical armor-plates. The 1½-inch plate on which the shell rested was dished all over its surface, the greatest amount being 8 inches, and a piece 16 x 6 inches was sheared off its side. The heavy armor-plates were dented in several places by fragments of the shell, but otherwise uninjured. Considering the foregoing, I think we may fairly reach the same conclusions in regard to a large-caliber shell of ordinary capacity as in the case of the smaller shell, namely, that to be effective they must be made strong enough to get through the ship's side before they break up or are burst, and that if so made the charges of high explosive which they will contain, amounting to about the same weight as if of black powder, while breaking them into more pieces, will yet not have a sufficiently superior destructive effect, if any more so, to justify their use instead of shell loaded with black powder. If the shell breaks up on the plating, the detonation of its charge will be comparatively harmless, and, since a very large part of the area of modern battleships is covered with armor, this renders it desirable to make the walls of the 13-inch, our largest shell, so thick as to reduce its capacity to about so pounds of powder. This shell will go through 6-inch armor whole and, consequently, will be effective over almost the entire area of an opposing ship, but to load it with guncotton and a detonator would, in my opinion, not render it sufficiently more destructive than when loaded with black powder to justify incurring the additional risk to our own ships. The premature detonation of such a shell loaded with guncotton would undoubtedly disable the gun and almost certainly put a turret out of action, while the premature explosion of such a shell loaded with powder would have no injurious effects whatever. On the other hand, the only gain from the use of the guncotton bursting charge would be a somewhat greater number of pieces from the burst shell with a correspondingly increased probability of killing and wounding men exposed to its pieces. Whether loaded with powder or with guncotton, the shell would be harmless if it struck armor which it could not penetrate, and in case it did penetrate the plating and burst inside the ship, the structural damage done by the high explosive would not be greater but rather Jess than that done by the powder.
It should be clearly understood that this is a question of expediency only; undoubtedly we can load our ordinary shell with guncotton and fire them with reasonable safety from our ordinary guns with the moderate velocities, and we don't do it merely because we don't think it desirable, not because we are waiting for someone to invent a way to do it. The trouble is that the disadvantages outweigh the advantages, The only advantage is greater fragmentation of the shell when it bursts, while, on the other side, are the disadvantages of greater cost, greater danger in using, and greater necessity for care to prevent accident or deterioration. The powder shell, after being loaded and fused, is perfectly safe and needs no looking after, but if you use guncotton you must be always on the lookout to see that the main charge is kept wet, and that the detonator is kept perfectly dry; then you must, for safety and ease of examination, keep your dry primers and your fulminate separate, so that each shell must be prepared just before firing; then the detonating fuse, with the delay-action arrangements necessary to prevent harmless explosion before penetration, is not only expensive, but much more liable not to function properly than is the simple percussion fuse of the powder shell. Finally, as regards danger, while we say it is reasonably safe to fire moderate charges of wet guncotton with properly arranged detonators, yet we must not forget that any shell may explode prematurely. Now, if the powder shell explodes in the gun, it does nothing more than perhaps scratch the bore, but if the guncotton shell detonates in the gun, it will probably destroy it Suppose, for example, an obstruction in the bore, which checks the shell up a little in its motion, this is exactly as if it struck the enemy's side and, therefore; if the detonator functions properly you at once have a premature detonation.
It seems evident that to justify incurring these disadvantages there must be shown a far greater gain in efficiency than I have been able to find exists, and, therefore, I am forced to conclude that the use of detonative charges of high explosives in ordinary shell is undesirable. For armor-piercing shell, which cannot be burst by powder, it is of course desirable to use an explosive that will break them up after their passage through armor, but the capacity of such shell is so small that any delay action detonator would leave no room for explosive, and so we must be content not to detonate, but simply ignite their charges by the ordinary fuse.
One more word about the safety of firing ordinary shell loaded with high explosives, and I am done with them. The springs and air-cushions and other devices for lessening the shock of firing on the explosive charge are utterly useless; indeed, they are worse than useless, for in many cases they would really increase the danger of premature explosion; the only precaution to take is not to have too long a column of the explosive; it is its own inertia, setting it back and compressing its rear layers, that is the danger, and the greater the weight borne on the rear layer the more likely it is to explode. Cushioning can do no good for the reason that the maximum pressure does not come onto the explosive until the shell has moved several calibers, before which time any practicable spring would have been compressed to its limit, so that just when the cushioning is most needed it is not there.
All the foregoing must be understood as applying to shell of ordinary capacity. We have yet to consider the advantages and disadvantages of using high explosives in special shell of very large capacity.
In a paper which I had the honor of reading before the War College some time since, on the subject of probable future developments of ordnance, after stating an opinion similar to the foregoing, I said in conclusion: "In other words, the Only possibility with the high-explosive shell is to use as great a weight of explosive as can be safely thrown, to provide means for its detonation on impact, and to trust to the effects of such a detonation outside the ship." "How destructive these effects would be we do not know, but for my own part I believe they are greatly overestimated."
In a succeeding paper I shall describe some experiments which have been made on this subject, and endeavor to show that the use of high-explosive shell, even with the largest admissible bursting charges, is not desirable.
II
In my preceding paper I endeavored to show that the use of a high explosive as the bursting charge of shell of ordinary dimensions and capacity was undesirable, as not affording sufficiently greater destructive power to justify the greater risk incurred in its use. My conclusions were as follows:
(1) The effect of the explosion of a shell of ordinary dimensions and capacity outside a ship, by reason of its being unable to penetrate the plating which it strikes, is practically nil, whether it be loaded with powder or with a high explosive.
(2) Consequently it is very desirable to make shell sufficiently strong, which means thick-walled, to penetrate the plating of an opposing vessel without breaking, so as to burst inside her.
(3) Such thick-walled shell, when loaded with a high explosive, are more dangerous to handle and fire, and are very little, if any, more destructive in their effects than if loaded with gunpowder.
(4) Consequently, if anything is to be really gained by the use of high explosive shell, it must be by means of shell of very large capacity, or what have been aptly called "torpedo shell."
In this paper, then, I shall consider the popular theory that overwhelmingly destructive effects will be produced by the detonation of very large charges of high explosive against, or even near, a ship.
Much less is known of the field of serious damage which would result from the detonation of 300 or more pounds of guncotton than of that which smaller charges would cause. Experiments have shown that the effects of say so pounds of guncotton are extremely local, but perhaps the explosion of 500 pounds would be a very different matter, and it was on account of this lack of knowledge founded upon actual experience that the Bureau of Ordnance undertook the experiments with the Gathmann shell, which I shall now refer to.
But first let us examine the results of a great accidental explosion of which we have detailed information, and see how far-reaching its effects were. I refer to the famous Stowmarket explosion, which took place at the guncotton works of that name in England, on August 11, 1871.
On this occasion, 13½ tons of dry guncotton exploded, and as there had been a few years before an explosion of about 51 tons of black powder at Erith, the official report compares the results of the two explosions and points out the comparatively local effect of the guncotton.
The 13½ tons of guncotton which exploded at Stowmarket was stored in three small wooden magazines, side by side, 14 feet apart, and only separated from each other by brick walls. The noise of the explosion was heard 30 miles and the shock was felt 7 miles. Some windows were broken 4 miles away, and some window-frames and sashes up to about one mile away. Slight material damage was done at distances from one-quarter to three-quarters of a mile away, but, as a general rule, not beyond 450 to 500 yards. Wooden cottages of very light construction were seriously injured at a distance of from 300 to 400 yards from the seat of the explosion, and the brick factory buildings, at distances varying from 20 to so yards, were practically destroyed. In the fall of the factory buildings a number of people were killed and wounded, the total of casualties being 24 killed and so injured. There is, however, no record of death or injury to any person from the effect of concussion merely. It is stated in the official report on the explosion, that a man standing on the railway, 76 yards from the magazines which exploded, was blown over a hedge and stripped of all his clothes, but was uninjured. The crater of this explosion was 3S yards by 22 yards long, 9 to 10 feet deep. The cause of the explosion was almost conclusively proved to be the presence of acid in the guncotton, probably due to careless manufacture, though perhaps to malice. A secondary explosion of several hundredweight of dry guncotton took place while an attempt was being made to remove it from the ruins of a store-house in which it had been kept and which had been overthrown and set on fire by the first or main explosion. This produced a crater 7 2/3 x 8 yards by 3 to 4 feet deep. Terrible as the results of this tremendous explosion appear, I cannot doubt that its effects upon such a structure as that of a modern steel ship would have been insignificant except at very close range. The forces which will overthrow light wooden or brick walls are insignificant compared with those which a steel ship will withstand. Of course, people in the open, or at open ports, and near enough to be struck by the rush of heated gases produced by an explosion, will be either burned to death or killed by the dynamic effect of the blow, but one must be very close to be in danger of this. Far-reaching destruction can only result, if at all, from the concussive effect of the air-wave which is propagated from the seat of the explosion. Take, for instance, the detonation of 300 pounds of guncotton, which would occupy a spherical space of about one foot radius. When the gases, rushing outward in all directions, occupy a space of 20 feet radius, they will have expanded 8ooo times, and if confined in such a space, would exert a pressure of only one atmosphere. Of course, the dynamic energy of the gases at 20 feet from the center of such an explosion is still great, and they would do more damage than the mere static pressure indicates, but the illustration shows how rapidly they lose their force as they expand. But independent of and prior to any direct effect of the expanding gases, the enormous pressure caused by their almost instantaneous production gives an impulse to the air, something like a sound-wave, which transmits the initial pressure in all directions. It is this wave of intense compressure, followed by expansion, which breaks windows, etc., even miles away from an explosion. Unfortunately, but little is really known in regard to the intensity of the effects of this air-wave, Within the rather limited field of the direct action of the gases of explosion, it is not possible to draw any exact line between their mechanical effects and those of the air-wave, and although we know that all really distant effects are due to the air-wave, almost nothing is known of how serious such effects are as regards human life.
Granting that at 10 feet or even a less distance from the explosion of 300 pounds of guncotton, the direct action of the gases would be withstood by the side of a steel cruiser, what would be the effect upon people behind that side? If the ports were open, the inrush of heated gases might be sufficient at so short a distance to destroy life; but supposing that there were no opening in the side; could the shock injure people behind it? It is almost inconceivable to me that injury should be done to people completely sheltered by rigid plating as long as such plating is not broken through or down, but I have found that it is not an uncommon opinion that the shock, or concussive effect of an explosion, can be transmitted through an unyielding obstacle and injure people behind it. This was one of the objections urged a few years ago to the Holland submarine boat project. It was said that the explosion of its own torpedo, even at such a distance as not to injure its hull, would yet disable its occupants. Experiments made at the torpedo station, however, completely disproved this. A Lay torpedo shell, 28 feet long by 3 feet diameter, built of about 1/8-inch iron plating and containing animals, was suspended 15 feet below the surface, and mines, consisting of 81 pounds of dry guncotton, also submerged 15 feet, were detonated at various distances varying from 305 feet to 80 feet. The last explosion, at 80 feet, bent in the plating of the boat at several places, but in no case were the animals injured, and it was apparent that a further increase in the severity of the shock would simply result in breaking through the plating and so killing the animals by drowning.
From our experiences up to this point, then, we could safely conclude that to do serious damage to a ship, or to her crew, by the detonation of a high explosive in the air outside the ship's plating, it would be necessary to use a sufficiently large weight of the explosive to actually blow in the plating, and that such moderate weights of guncotton as had been tried, up to 100 pounds, had shown but trifling effects of this sort. It remained to be determined, then, by actual test, how serious would be the effects of the largest charges of guncotton which would be likely to be used in shell, and when Mr. Louis Gathmann, of Chicago, presented to the Department his very ingenious plan for firing extraordinarily large charges of guncotton from powder guns, it was decided to give his ideas a test and at the same time to find out whether or not such shell would have the destructive effect generally supposed.
Mr. Gathmann's plan was to construct his shell with a very strong head and very thin walls; the rotating band to be at the bourrelet, or junction of head and body of shell; the outside diameter of body to be considerably less than the diameter of bore of gun, so that the powder gases would surround and press upon the shell; and the base of the shell to be closed only by a gas-tight plunger or piston. By this construction, the shell being filled with wet guncotton, the powder pressure would be transmitted by the movable base-plug and the wet guncotton to the inner walls of the shell, exactly balancing the powder pressure on the outside of the shell walls, and, consequently, theoretically, the shell walls could be made as thin as one pleased without risk of their being either burst or collapsed. Thus, instead of having to put 90 per cent of the total weight of the projectile into the steel body in order to have it strong enough to stand being fired from a gun with high velocity, Mr. Gathmann proposed to put only 60 per cent or less, and to have 40 per cent in the form of explosive.
There were just two weak points in the foregoing theory: 1. The powder pressure in a large gun is by no means uniform throughout the chamber; on the contrary, from the moment the charge is ignited the gases are rushing to and fro, producing excesses of pressure now here and now there, and, consequently, a very thin-walled shell might easily be broken up in the gun, notwithstanding the arrangement for producing equal pressures inside and outside its walls. 2. The entire propelling pressure on the base of the shell was transmitted through the wet guncotton, and, consequently, if wet guncotton could be exploded by such a pressure as was used in the gun, the Gathmann shell would be a failure.
The first difficulty was overcome by making the shell of very strong material-nickel-steel forgings-and their successful operation, as far as this point was concerned, was demonstrated by firing two of them filled with water.
As to the second point, as we knew wet guncotton would stand almost any slowly-applied pressure, and as the charge of the shell was to be of saturated guncotton put in copper cases, and with all interstices filled with water so as to completely fill the shell cavity, we felt reasonably sure that the danger of premature explosion would not be great. As the sequel proved, however, this source of danger was underestimated.
The gun, improvised for the test of the Gathmann shell, was a 13-inch tube forging, bored, rifled and chambered like a 12-inch gun, and having a 12-inch breech mechanism fitted in a heavy steel forging screwed over its rear end. It was calculated to stand an internal pressure of about 10 tons per square inch within its elastic limit, and would undoubtedly have stood very much more without bursting.
The shell contained from 307 to 320 pounds of guncotton each, and were fitted with detonating primers consisting of a fulminate igniter, half a pound of dry guncotton, and about three pounds of wet guncotton contained in a steel primer case with walls one-half inch thick and screwed into the nose of the shell. In the first attempts to explode one of these shells the necessity for close contact between the primer and the main charge was well illustrated. In the primer case a strong diaphragm had been fitted separating the one-half pound of dry from the three pounds of slightly wet guncotton, and the one-half inch steel walls of the primer case separated the entire detonating arrangement from the main charge of the shell. The result of the first attempt to explode one of the shells by electricity, using 35 grains of fulminate in the center of the one-half pound of dry guncotton was as follows: There was a very mild explosion, scarcely audible at the bomb-proof nearby; the head of the shell was ruptured, a number of pieces being found in the vicinity; the forward copper case containing wet guncotton was torn and its contents scattered all about; and the main body of the shell containing the remaining four copper cases of wet guncotton was pushed some ten or twelve feet from its original position. Evidently the one-half pound dry primer had detonated, but had merely set on fire the three pounds of wet guncotton, separated from it by the diaphragm, and the combined force had nearly ruptured the front end of the shell without exploding or setting on fire the main charge of wet guncotton.
A second attempt was then made with a shell having the diaphragm separating the wet and dry parts of the detonating primer removed. The results were practically the same as before, the head of shell ruptured and the main body, with its charge, thrown a few feet to the rear, sliding along the ground.
To prove that these results were entirely due to the separation of detonator and main charge by the walls of the primer case, and not to any lack of strength or sensitiveness in the guncotton, one of the copper cases containing 71 pounds of wet guncotton, and taken out of the shell which had just failed to function, was exploded on a live oak block 14 x 16 x 15 inches long, by means of a one-pound dry primer and a service igniter of 35 grains of fulminate placed on top of it. The result was perfect detonation, and the complete disappearance of the oak block, not a recognizable portion of it being found. The results other than this, however, were almost nil.
The next experiment made was with a shell having the walls of its primer case perforated by ten holes, each about 1½ inches in diameter, so as to allow a more direct action of the primer upon the main charge of wet guncotton. The dry guncotton in the primer was also increased to 10 ounces instead of 8 as before, and the primer case was fitted with a cover of tin sheeting to keep the water from entering from the shell body and drowning the dry primer. The igniter, as before, was 35 grains of fulminate of mercury, exploded by electricity from a bombproof. The shell was placed on its side on the beach, between two screens constructed as follows: Three uprights 8 x 12 inches were let into the ground about 3 feet each and supported from the rear by braces 10 feet long and 6 inches square. To these uprights were secured, each by six 1½-inch bolts, steel plates about 13 feet long by 7 feet wide and 5/8 of an inch thick. These screens were so feet apart, facing each other, and the shell was 15 feet from one and 35 feet from the other. On firing there was complete detonation. A hole- in the ground 14 x 9 feet and 10 feet deep was the principal result. One small fragment of the shell had been driven through each of the screens, and other fragments had struck them, making only slight dents. The plates were not bent, displaced or injured in any other way. Only a few fragments of the shell itself were found and these were small, showing how completely it had been destroyed. A very slightly-built wooden Iatrine standing on the sea-wall, 93 feet from the shell, and with nothing interposing between them, was only injured by the explosion to the extent of having two or three of its boards displaced by being pulled off from the nails holding them to the framework.
The trifling effects of this detonation of a shell containing 320 pounds of guncotton was, I need hardly say, a complete surprise to most of those present. That a light screen, in no way approaching in strength the side of a steel cruiser, and only 15 feet away from the explosion, should be uninjured was rather a damper to the enthusiasm of people who prophesied the destruction of a ship by a single one of such shells. It was said, however, that even granting the experiment to show only extremely local effects from the explosion, yet these effects were tremendous upon objects in contact with the explosive, and as the idea was to explode the shell by impact against an enemy's ship, it by no means followed that such an explosion would be comparatively harmless. It was still claimed that no armor-plate, however strong and well backed, could withstand the explosion of such a shell in actual contact with it. Accordingly it was determined to make one further experiment to test this point.
The plate used was the ballistic test-plate which had represented the Kearsarge's side armor, and was of nickel-steel, face hardened, backed with 12-inch oak and two ½-inch skin plates, weighing 28 tons in all. It was 7½ feet high by 16 feet long, and of a thickness tapering from 16½ inches at top to 9½ inches at bottom. This plate was placed against an ordinary target structure, but not tied to it in any way, and other armorplates were placed at right angles to each end of it to confine fragments. The shell, containing 307 pounds of wet guncotton and a primer of 8 ounces dry, 2 pounds nearly dry, and 35 grains fulminate, was suspended by ropes so that it lay horizontally against the plate and along its middle line. It also had a light wood blocking support. The detonation was perfect so far as could be judged by the character of the report and results, but only comparatively trifling damage was done. The surface of the plate was scarred and roughened as if it had been partly fused, for a length of about 5 feet and a width of 8 inches abreast the suspended position of the shell, and there was some very slight flaking from the surface near the old shot holes. A hole was made in the ground along the front of the plate about 2½ feet deep, and partly as a result of this, and partly, perhaps, by rebound from its supports, the plate was swung away from its original position, one end leaving the target structure about 2 feet Two 1-inch steel plates on the ground below the projectile were broken up and their pieces driven into the ground and dispersed about the vicinity. The side of the shell itself next the plate was flattened out and burned into a ragged sheet 5 feet long by 8 inches wide, practically fitting into the fused portion of the plate, but detached and falling to the ground. The remainder of the shell was dispersed in very small fragments.
The armor-plates at the sides were practically unchanged in position, and the effect upon surrounding objects, such as a lot of oak blocking on the ground 40 or 50 feet away, was very much less than would have been that of the blast of a 12-inch or 13-inch gun fired with its muzzle at the point of explosion of the guncotton shell.
To observe the effect upon animal life, four chickens had been placed in different positions. One on the ground behind the center of the plate and one foot from its backing. This chicken was entirely uninjured, and upon being released showed no effects of the explosion, thus proving conclusively that people behind an armor-plate need have no fear of high-explosive shell.
One on the ground directly in front of the center of the plate and 43 feet from it. This chicken was killed by a flying fragment, which, of course, indicates nothing of importance. Its feathers were not blown off, except where it was struck, and it appeared to have but that one injury.
A third chicken was placed in one corner of the cellulose compartment of a coffer-dam, practically in front of the plate and 50 feet from it. This coffer-dam, consisting of an armor-plate with a thin steel compartment for cellulose behind it, had been fired through and its back, which faced the explosion, had been torn wide open, so that the chicken, in one corner, had much the position which a man would have if he stood well outboard and close to the side of a ship's port. This chicken was entirely uninjured.
The fate of the fourth chicken is rather mysterious. It was placed in another coffer-dam similar to the one just mentioned, but much nearer the explosion, about 25 feet, and having no perforation in it, so that it could only be affected by concussion coming from above down into the compartment. This chicken was found to be dead, but with no external marks of injury, and what killed it is, as I have said, not at all certain. Possibly it was internally injured by the concussion from the air-wave propagated in all directions and reaching it from above, but more probably it was killed by being in contact with the very thin plating of the coffer-dam and being struck by that, which of course must have yielded somewhat.
This experiment would seem to completely dispose of the theory that a high-explosive shell of very large capacity will blow in the side of an armored vessel if exploded against it. But there is one more objection which has been raised to this conclusion. It is said that the motion of the projectile will greatly increase the effect of its explosion on striking. The idea is that the energy of the projectile will act, as it were, as a tamping to the explosive, the gases of which will tend to continue onward and so will act in that direction with greater force than if they came from a stationary explosive. This is undoubtedly true to a certain extent, but the increase in effect can be only trifling. Suppose, for example, that at the instant of explosion the projectile were advancing at the rate of 1500 f. s., then the gases of explosives would expand in the outward direction with 1500 f. s. greater velocity than they would to the rear; but what is 1500 f. s. more or less when we consider the velocity with which the gases rush from the explosive center under the influence of the pressure of hundreds of tons per square inch produced by a detonation. No, this tamping effect of its striking velocity can add nothing material to the force of a high-explosive shell. We can safely conclude that such a shell, even if of the largest practicable capacity, exploding against the armored side of a vessel would be practically harmless. We can also safely conclude that such a shell, exploding in the air, even very near to an unarmored vessel, would do no serious structural damage to her. In other words, the effective sphere of the torpedo shell is far more limited than is usually supposed; it is superior in its effect to an ordinary powder shell only if it explodes inside a ship or under water close to her side. We have then to decide whether the greater effects of the large capacity high-explosive shell within the above limits are sufficient to justify incurring the greater dangers of their use.
It must be remembered that in the very nature of things that which is dangerous to an enemy is dangerous to ourselves. The powder which propels our shell against the enemy may burst our own guns or explode in handling; the fuse which should act on impact with the opposing ship may act prematurely to our own destruction, etc. The best that we can do is to devise and apply safety devices to render our weapons innocuous until the moment when we wish them to act upon an enemy, but there is always the possibility of premature action with injury to ourselves instead of our opponents. The real question, then, as to the use of any weapon is, how do the chances of its damaging an enemy compare with the chances of its damaging us, and how would the results of its successful use compare with those of its accidental premature use? Take, for example, the use of the Whitehead torpedo on cruisers, in regard to which there is much difference of opinion; if the chance of an accidental explosion before firing is very small and the results of such an explosion would not be very serious, and if, at the same time, a successful hit with the torpedo would put an enemy out of action and there is a fair chance of making such a hit, then the use of the torpedo is not only justified but demanded. But if, on the other hand, an accident of very possible occurrence may prematurely explode your torpedo, and if this explosion would put your own ship out of action, then the use of the torpedo would not be advisable. As regards the torpedo, I am strongly of opinion that the first statement of the case is correct; nothing but a direct blow upon the detonator could explode the torpedo, and the chance of this happening is almost infinitesimal. Moreover, even if it happened, and the torpedo were detonated while in its launching tube, I am satisfied that the damage done would be confined to the torpedo-room. The largest torpedo, the long 18-inch Whitehead, only carries 220 pounds of wet guncotton, and its explosion in a torpedo compartment of ordinary size (16 x 64 x 8 feet= 8192 cubic feet on Kearsarge and Kentucky), while it would of course kill everyone in that compartment, would not seriously damage the ship nor, in my opinion, do any damage at all outside of the compartment. It would be much better to have the explosion take place above the water-line than below it, for it might tear a good-sized hole in the ship's side around an underwater tube, while if exploded when being used through an open port above water, the damage to the side would be of no importance, and, moreover, in the latter case there would be a free escape of the gases of explosion. On the other hand, the chance of hitting with your torpedo an enemy attempting to ram you would be very good, and the effect of hitting him would probably be to disable him by at least filling several of his compartments with water and perhaps sinking him.
The use of the above-water torpedo on unarmored ships is, then, in my opinion, very desirable, but when it comes to the use of torpedo shell, it seems to me that the best of the argument is on the other side. We have seen that the chance of the torpedo shell doing very great damage to an enemy, even if it hits him, is comparatively small. All battleships nowadays are well covered with armor against which such shell would be harmless, as they would explode outside it; only in case they struck where they could penetrate would they do great damage. Moreover, the reduced velocity necessary for the safe firing of torpedo shell considerably reduces their chance of hitting at all. Can we afford to not only reduce the chance of hitting the target but also to greatly reduce the area of target against which a hit will count, for the sake of somewhat greater effects from a successful hit? If a successful hit meant destruction to the enemy, we might, but not otherwise, I think. Then, again, the premature explosion of the torpedo shell would do us the maximum harm that such a shell could do. The above-water torpedo, if exploded prematurely, explodes in the air in a large compartment containing nothing of vital importance, while if it hits. it explodes against the underwater side of the enemy where it will do its maximum of damage. The high-explosive shell, on the other hand, if exploded prematurely, is just in the position with regard to our own ship that we hope it to reach in an enemy's ship before exploding; if successfully fired and exploded, it is more likely to burst harmlessly outside the enemy than to penetrate and burst within her. As to the actual chance of such a premature explosion of a torpedo shell, I still hold exactly the same opinion as when I said some time ago, in a paper read at the War College on " Probable Developments in Ordnance," that we could safely fire about 175 pounds of wet guncotton with a velocity of about 1600 f. s. from our present 12-inch gun. I did not mean that there would be absolute safety with such a shell, or as much safety as with a powder shell, because any shell may explode prematurely, and the pre mature explosion of the guncotton shell would be far more serious than that of the ordinary shell. I do mean, though, that the danger of firing torpedo shell is not too great, provided their use is demanded by other considerations. It is the fact that I do not believe their effectiveness is sufficiently greater than that of powder shell that causes me to conclude that we should not use them; it is not that I believe their use, in a proper way, impracticable. One must be satisfied with low pressure in the gun and, consequently, low velocity of the projectile, and one must not sacrifice safety to the desire to throw too large a weight of explosive. It was going too far which caused the failure of the Gathmann shell. If we had been satisfied to fire it with, say, 4 tons pressure in the chamber instead of 8, it would probably not have exploded prematurely. I don't advo cate, and never did, such an arrangement as the Gathmann shell. I believe .if we use high explosive shell at all they should be like any other shell, only thinner-walled; but, nevertheless, the Gathmann shell was based upon a correct theory and would no doubt function properly in actual practice as long as the pressure on the wet guncotton was kept low enough. The premature explosion of. the one shell that was fired was undoubtedly due, in my opinion, to the setting on fire, either by pressure alone, or by the powder gases getting into the shell body, of the wet guncotton charge. There was no detonation, only an explosion of low order. The rear end of the gun was broken into large pieces, and the rifling remained quite perfect everywhere, whereas a detonation would have smashed the lands quite flat in wake of the shell. The pressure gauges all registered 28 tons, and I am satisfied it was nothing but the excessive pressure produced by the combustion of 394 pounds of powder and 300 pounds of guncotton which burst the gun; the gases could not get out of the muzzle fast enough and so the pressure rose till the gun walls gave way, The effects of the explosion can be seen from the photographs, and it is very surprising to see how the guns and carriages mounted on one side of the gun which had burst escaped injury. Whether such an explosion would have burst a service 12-inch gun cannot be absolutely determined, but I am very strongly of the opinion that it would not have. The detonation of such a shell in the bore would have burst any gun, but with free escape for the gases at the muzzle it is not likely that the simple combustion of the guncotton would produce a sufficient pressure to actually burst a modern gun.
Finally, then, the matter stands thus:
(1) The high explosive shell, whether it contains an ounce or 300 pounds of explosive, must penetrate the side of a ship before bursting in order to do serious damage; however large its charge, it will burst harmlessly against armor.
(2) This makes it necessary to arm our battleships with high-power guns of large caliber, firing armor-piercing projectiles, as the only means of getting through the thick armor of opposing battleships and thus reaching their vital parts.
(3) We can fire thin-walled shell loaded with guncotton from these same high-power guns by considerably reducing the velocity, but there is much more danger to ourselves with such shell than with powder shell; the chance of hitting with them, on account of their reduced velocity, is much less, and they would be effective only against that comparatively small and unimportant part of a modern battleship which is unprotected by armor.
(4) We can use guncotton, or some other high explosive, instead of powder in thick-walled common shell, such as will pass unbroken through the armor, but its use is more dangerous, and its effect but slightly, if at all, superior to that of powder under the same conditions.
Therefore, in the present state of my knowledge, I am forced to conclude that the advantages of using high explosives in shell are outweighed by the disadvantages, and that unless some new light is thrown upon the subject in the future, we should continue to use powder for bursting charges, and not high explosives.