The bare enumeration of Rumford's published papers would occupy considerable space, but many of them have more to do with philanthropy and domestic economy than with physics. We have seen that, when guest of Lord George Germaine, he was engaged in experiments on gunpowder. The experiments were made in the usual manner by firing bullets into a ballistic pendulum, and recording the swing of the pendulum. Thompson suggested a modification of the ballistic pendulum, attaching the gun-barrel to the pendulum, and observing the recoil, and making allowance for the recoil due to the discharge from the gun of the products of combustion of the powder, the excess enabled the velocity of the bullet to be calculated. Afterwards he made experiments on the maximum pressure produced by the explosion of powder, and pointed out that the value of powder in ordnance does not depend simply on the whole amount of gas produced, but also on the rapidity of combustion. While superintending the arsenal at Munich, Rumford exploded small charges of powder in a specially constructed receiver, which was closed by a plug of well-greased leather, and on this was placed a hemisphere of steel pressed down by a 24-pounder brass cannon weighing 8081 pounds. He found that the weight of the gun was lifted by the explosion of quantities of powder varying from twelve to fifteen grains, and hence concluded that, if the products of combustion of the powder were confined to the space actually occupied by the solid powder, the initial pressure would exceed twenty thousand atmospheres. Rumford's calculation of the pressure, based upon the bursting of a barrel, which he had previously constructed, is not satisfactory, inasmuch as he takes no account of the fact that the inner portions of the metal would give way long before the outer layers exerted anything like their maximum tension. When a hollow vessel with thick walls, such as a gun-barrel or shell, is burst by gaseous pressure from within, the inner layers of material are stretched to their breaking tension before they receive much support from the outer layers; a rift is thus made in the interior, into which the gas enters, and the surface on which the gas presses being thus increased, the rift deepens till the fracture is complete. In order to gain the full strength due to the material employed, every portion of that material should be stretched simultaneously to the extent of its maximum safe load. This principle was first practically adopted by Sir W. G. Armstrong, who, by building up the breech of the gun with cylinders shrunk on, and so arranged that the tension increased towards the exterior, availed himself of nearly the whole strength of the metal employed to resist the explosion. Had Rumford's barrel been constructed on this principle, he would have obtained a much more satisfactory result.