Whether or not the lock which has been described in the preceding paragraph was the origin of the spring padlock, constructed entirely of metal, may perhaps be doubtful; but it is evident that the principle of its construction was the same. In both systems the bolt was secured by the end of a spring catch. It is only necessary to transfer the fixed end of the spring from the door to the bolt, and the notch from the bolt to the door, to make it resemble the spring catch of the Roman padlock about to be described. The Roman iron padlock and key represented in figs. 21C to 22C, Plate V., which is put together from specimens in my collection obtained partly from Jordan Hill, near Weymouth, and other sources, consisted of a square box, having a bar, _d_, on the top, and parallel to it, which was attached to one end of the box by means of a curved portion. The bolt _a_ was provided with two perpendicular bars, _b b_, at the end of which were rings, _c c_, which slipped on to the parallel bar _d_. At the end of the bolt were two or more catch springs, _e_, put on like the barbs of an arrow. These, being placed into the hole of the tube _f_, at the same time that the rings were slipped along the bar, collapsed and sprung open again, after having passed the opening, thereby fixing the bolt in the tube. To open the lock, a pin or key, _g_, having a return at the end, in which was a slit made to fit the springs, was pressed in at the opposite end, so as to close up the springs, after which the bolt could be drawn out of the box. This action is better shown in the succeeding examples of modern spring locks of the same kind. The case of a similar padlock to the above was found with Roman remains at Irchester, near Wellingborough, Northampton, by the Rev. R. BAKER, in 1878, and is figured in the Associated Architectural Society's Reports, vol. xv., plate iv., 1879.