error to limit criticism to the longitudinal extent of the belt armor in modern ships, and to exclude consideration of the vertical extent of the armor above and below the load-line. Apart from any discussion of the question from the artillerist’s point of view, or any attempt to determine the probability or otherwise of the wholesale destruction of the unarmored portions of modern battle ships by shell-fire from large guns, or by the projectiles from rapid-firing and machine guns, it is perfectly obvious to any one who will examine into the matter that the risk of damage to the light superstructures situated above the belt must be greater than the corresponding risk of damage to the narrow strips of side area exposed at the unarmored ends of the _Admiral_ class, between the level of the belt-deck and the water-line. “Sir Spencer Robinson, after his inspection of the models shown him at the Admiralty, recognizes the fact that in the French belted ships (of which the _Amiral Duperré_ is an example), if the light sides above the belt-deck are destroyed or very seriously riddled in action, the ship would be capsized in a very moderate sea-way. He further emphasizes the statement that the ships of the _Admiral_ class in the English navy, if similarly treated, would also capsize under the same conditions, and he appears to be surprised at the admission having been made. The fact is that there has never been any assertion that the _Admiral_ class would be safe against capsizing independently of assistance given to the armor-belted portions by the unarmored structure situated above. On the contrary, from the first, in the design of these ships, it was recognized that their stability, in the sense of their power to resist being capsized, if inclined to even moderate angles of inclination, was not guaranteed by the armor-belts. In this respect they were in identically the same position as all other armored ships with shallow water-line belts and isolated armored batteries placed high above water. “What has been said respecting the _Admiral_ class is this: If the unarmored ends above the protective deck were completely thrown open to the sea, then the initial stability (that is to say, the stiffness of the ships or their power to resist small inclinations from the upright) would still be guaranteed by the central armored portions. So fully did we appreciate the fact that the life of the ship in action (as determined by her power to resist large inclinations) depends greatly upon the assistance given by the unarmored superstructures to the armor-belted parts, that we were careful to make the structural arrangements of the superstructures above the belts such that they could bear a very considerable amount of riddling and damage from shot and shell without ceasing to contribute in the most important degree to the buoyancy and stability.