plates, and to cast them in chill moulds. "These iron moulds," it is said, "chill the surface of the plate and give it great hardness and elasticity. When sufficient metal to form the plate has been poured in, molten metal is run round the mould itself. By these means, whilst the surface of the plate is chilled by contact with the mould, the interior is prevented from cooling too rapidly, and the particles of iron have time to assume their natural positions, with the result that the interior and inner side of the plate retain the high degree of tensile strength which belongs to the iron used." Wrought-iron plates are partly penetrated by shots which cannot get through them, and thus would be liable to be destroyed by a continuous fire from guns of comparatively small calibre. The advocates of chilled cast-iron armour believe that it would be absolutely impregnable, as regards guns of less calibre than those capable of penetrating it. It is, however, admitted that a thicker plate would be required than in the case of wrought-iron to resist complete penetration by any given gun, and for this reason alone it appears altogether improbable that it will ever be used for ships' armour plates. It is true it would cost much less than wrought-iron; but the great consideration in ironclads is weight, and the armour which is really the cheapest in the end is that which, weight for weight, is the most effective. This is due to the fact that greater weight of armour necessitates larger displacement, and this requires a larger area of armour, more powerful engines, &c., which again further increase displacement, and so every additional ton of armour tells much more in the total expense of the structure than in its own prime cost. Herr Grüson's cast-iron armour was tested by shot from some of the Krupp guns, and although it was reported to be a success, the German Government do not appear to think very much of it, seeing that they have used ordinary rolled plates, even on land defences constructed since the trials. Cast-iron armour was also tried at Spezzia, in the notable series of experiments made by the Italian Government, in November, 1876, but the results were not at all satisfactory.

The proposal to adopt cast-iron as the material for armour plates, although it led to no immediate result, served to direct attention

to the question of the defects of wrought-iron, and also to the consideration of means for avoiding them, without at the same time sacrificing the well-known advantages of the old material. The adoption of mild steel in place of puddled iron for the frames and skin plating of war vessels has also served to bring about trials of new materials for armour. These trials are described at length in an able paper recently read before the Institution of Mechanical Engineers, by Captain C. O. Browne, R.A., of Woolwich, and entitled, "On the Construction of Armour to resist Shot and Shell." Captain Browne's paper gives an admirable resumé of the present state of the armour question, and he describes the more important experiments in England upon wrought-iron armour of the last few years, as well as the Spezzia experiments upon steel armour plates, and some more recent trials of composite iron and steel armour. First he compares the action of different kinds of guns and projectiles upon armour. The first large guns which were used in his country, and those which were for a long time after favoured in America, were designed to throw a large shot with low velocity. Their action upon armour was to break it up, and this method was successful with the inferior plates at first tried. We remember seeing the effect of a large gun known as the Horsfall gun, at Shoeburyness, upon a target representing the side of the Warrior; a breach was made in the target, but it was a smash such as suggested a waste of force. As better wrought-iron armour was produced of greater thickness it was seen that the best method was to pierce or punch rather than smash the plate, and for this purpose the present style of heavy gun throwing a comparatively small shot with great volocity has obtained. The first projectiles, and those best adapted for smashing, were spherical. Flat-headed cylindrical shot were then tried, and they are most effective merely for punching a hole in a plate, but a cylindrical shot with an ogival head has been found most effective in penetrating the ordinary armour and backing. The flat-headed shot punches a hole through the plate and cuts out a disc from it which meets with great resistance in the wood backing. The pointed shot on the contrary punches a hole, but instead of tearing out a disc cleaves the plate and passes through it, and then through the wood backing. This projectile

finds out especially the peculiar weakness of the sandwich system, and Captain Inglis states on this point as the result of experiment that the resistances of armour 7 inches thick made up by one, two, or three plates, varies respectively as 100, 96, and 89; and that a single.plate 17 inches thick is about equal to three 6 inch plates separated by 5 inch layers of teak. If such be the case, the English Inflexible with two 12-inch armour plates must be superior to the Italian Dandolo with one plate 21 inches thick; but there is a farther consideration, it is exceedingly difficult to make reliable armour plates of the greater thickness, unless they be narrow, and one result of the Spezzia experiments was to prove the great weakness of narrow plates, for every one of the thick plates broke across when fired at by the 100-ton gun.

In comparing the effect of guns upon plates a formula is used deduced from practical experience, and by which the efficiency of each gun is expressed in terms of the thickness of wrought-iron plate it is capable of penetrating. The resistance offered by armour to penetration was for some time considered as varying with the square of the thickness; it has been determined within the last few years that the resistance varies with the 1 ths power of the thickness, and thus the gain of extra thick plates is less than was formerly supposed. Using this rule, and ascertaining also the striking velocity of shot from a given gun at a given range, the weight of shot also being known, the amount of penetration of a plate of average quality may be made a matter of calculation, and plates may be matched against guns. If a gun not capable of piercing through a given plate, it will obviously penetrate it to a less depth than it would be equal to if the plate were only just a match for it, and this being taken into account the results of successive experiments come very near the predicted result. It is interesting to notice the details of some of the experiments as showing exactly what may be expected of armour as regards the effect of guns already carried on board ships.

The 38-ton gun, as at first tried in 1876, was fired with a charge of 130 lbs. of powder and a projectile weighing 812 lbs. By the formula used by the Director of Artillery its calculated power

of penetration would be 19 inches; it did actually pass through a sandwich target having three plates each 6 inches thick. Afterwards the 38-ton gun was altered so as to take a larger charge of powder; its calculated penetrating power was then 21 inches, and it was opposed to a target having a total of 26 inches of iron; it penetrated this to a depth of 20 inches. The gun which recently burst on board the Thunderer was a 38-ton gun, and there are no larger guns than this afloat at present. It is therefore of some interest to know that its limit of penetration under the most favourable circumstances is 21 inches. The Inflexible is the only ship in the English Navy which has armour exceeding this thickness, and the Dandolo and Duilio are the only war ships in the possession of any foreign power which come up to it. Thinner armour of course is of some use against the 38-ton gun when the destructive shell is substituted for the shot, which merely penetrates.

Similar results were obtained by the tests of the 80-ton gun, which has a projectile weighing 1,700 lbs., and was in the first experiments fired with a charge of 370 lbs., which gives a calculated penetrating power of 28 inches. It was tried against a sandwich target consisting of four 8-inch plates, and actually penetrated 25 inches. The 80-ton gun was also chambered to take a larger charge of powder, when its shot would, by the formula, go through a solid plate 30 inches thick. It was tried against a target containing 32 inches, which was more than a match for it, and actually penetrated as far as 27 inches. Experiments were also made with shells both from the 38 and the 80-ton guns, but the experiments were made upon unbacked plates, which circumstance detracts much from their value. On this point Captain Browne remarks:-" It would be of great practical value to our fleet if it were known exactly what thickness of armour was sufficient to cause even the largest shell to explode, before it had power to fire the backing and wood of the ship; because if a vessel were rendered secure against being set on fire, she might stand a great quantity of shot fragments and langridge passing through her side. Probably all our newest ironclads, say those carrying over 12 inches of armour, are in this condition of com

parative safety, perhaps many more; in any case this would seem to be a subject well deserving further experiment."

The most effective weapon against wrought-iron armour is, as we have said, a comparatively small shot at high velocity, and the most effective projectile one which is best adapted for punching. The cast-iron armour to which we have referred is of an exactly opposite character, difficult to punch, easy to smash. Captain Browne dismisses it from consideration as a material for ship's armour. He says it has been used in much greater thickness than wrought iron, and, for a time, shot produces no apparent effect, but after continued fire it cracks, and soon after crumbles to pieces.

The information which we have as to the efficiency of steel armour, is partly derived from the trials made with the 100-ton gun, by the Italian Government, at Spezzia. A detailed account of these experiments was given by our Italian contemporary, Rivista Marittima, in the number for December, 1876. Wroughtiron armour plates of English manufacture, 22 inches thick, were used on some of the targets, and on others were steel plates of the same thickness, manufactured by Schneider and Co., of Creusot, in France. We are not informed of the precise character of the steel, but we should suppose it was hard steel, and it appears to have been manufactured by the Bessemer process, and afterwards by hammering. The result of the experiments was that the shot from the 100-ton gun went completely through the iron plates and the backing behind them, showing that there was much greater penetrating power than was used up by the plate, but it only just penetrated the steel plate. All the plates, however, were broken. In some other experiments with much smaller guns, the iron plate was partially penetrated, but little injured, while the steel plates showed signs of cracking, which clearly prove that though superior to the iron as regards one very heavy blow, they were much inferior in powers of resistance to the fire of comparatively small guns.

There are also some important results of more recent English trials of new kinds of armour, but we think they are anything but conclusive. The first series were carried out at Shoeburyness, the