in Wales.[11] It has a flat handle and appears to be adapted to be pressed downwards as if for opening a latch. Figs. 35B, 36B, Plate IV., are nearly similar ones, and were discovered in the Roman Villa at Hartlip, in Kent.[12]

Figs. 37B and 38B, Plate IV., are from drawings taken by me in the Musée de Saint Germain, and were found at St. Pierre-en-Chastre, Oise; others are figured in M. Liger's 'La Ferronnerie.'[13] Fig. 39B, Plate IV., is in the British Museum, and was found within the entrenchments at Spettisbury, near Blandford; it was presented to the Museum by Mr. J. Y. Akerman. Figs. 40B and 41B, Plate IV., are two found by me in pits in the interior of Mount Caburn Camp, near Lewes.[14] Fig. 41B is of large size, 8 inches in length, and sickle-shaped. All the objects discovered in this camp proved it to be of the late Celtic period; the tin coins found associated with these remains, the bone combs, pottery, and other objects belong to an age anterior to the Roman conquest. Fig. 42B, Plate IV., is a similar one found by Mr. Park Harrison in similar pits in the neighbouring camp of Cissbury,[15] in Sussex, which has been shown to have been occupied by people of the same age as Mount Caburn, viz.: the late Celtic period. It will be seen that some of these keys, all of which are of iron, have a small return or pin at the end, which is adapted to fit into a hole, and in the Cissbury specimen this end is flattened, as if to enable it to fit an aperture of special dimensions.

But for whatever purpose these crooked keys were used, whether as latch-keys, as keys for single-tumbler pins, or as hooks to pull back a plain iron or wooden bolt, the large size of some of them, especially that from Caburn, fig. 41B, and sickle shape, corresponds with remarkable accuracy to the description of a Greek key given by Eustathius, and quoted in Parkhurst's 'Hebrew Lexicon.' He says that they were "in the shape of a sickle, and that not being easily carried in the hand on account of their inconvenient form they were carried on the shoulder, as we see our reapers carry on their shoulders at this day their sickles, joined and tied together." Callimachus, in his hymn to Ceres, says that the goddess, having assumed the form of Nicippe, her priests carried a key, κατωμαδιος, that is, fit to be borne on the shoulder.[16] This also explains, I presume, the passage in Isaiah, "and the key of the House of David will I lay upon his shoulder; so he shall open, and none shall shut; and he shall shut, and none shall open."[17] It will be seen that the specimen found by me in Mount Caburn corresponds exactly with the description given in the above quotations, the curved portion of the key being 7¼ inches in diameter, a bundle of them tied together would exactly fit the shoulder, as represented in fig. 43B, Plate IV. As we know from the researches of Mr. Evans and others that imitations of the coins of Greece spread throughout Gaul and Britain, some of which, of very debased form and cast in tin, were found in the camp at Caburn in association with the sickle-shaped keys, and others have been found in connection with relics of the same period elsewhere, there is no inherent improbability in the supposition that the keys may have followed a like route.[18] Should further discoveries tend to confirm this connection, it would be a remarkable testimony to the value of archæological investigation if the well-known passage in the 'Odyssey' about the key of Penelope were to find its definite interpretation on the shores of Sussex.[19]

We must now return to fig. 2, Plate I., in order to trace the third class, C, of locks and padlocks fastening with a spring catch. It seems probable that fixed locks may have preceded hanging ones, although, on the other hand, the want of some contrivance for securing property must have been felt in connection with saddle-bags, panniers, and other appliances of nomadic life, and in a condition of society which preceded the use of fixed abodes. Be this as it may, it seems possible to trace the employment of spring locks by means of survivals from the common door-bolt.

The origin of the spring padlock, in the present state of my knowledge on the subject, is doubtful. The sequence which I here assume is only tentative, and it is probable that connecting links with more primitive contrivances may be supplied hereafter. The defect of the common bolt, as I have already shown, was its insecurity as an outside fastening; in fact it afforded no security at all, and to remedy this defect and make it inaccessible, except by means of a key, several different contrivances appear from the first to have suggested themselves; amongst others, one of the simplest was adopted in connection with the Scandinavian bolt, a specimen of which, probably a modern survival of an ancient form, was exhibited in the Scandinavian Section of the Exhibition of 1867, and is figured in M. Liger's work.[20] We must suppose the handle in fig. 2, Plate I., and its neck connecting it with the bolt, to be removed, leaving only the slit in the door along which the neck of the handle slid, and that a similar slit was made in the bolt also. The key, which was of iron, was T-shaped; it was inserted from the outside through the slit in the door, and in the bolt, with the arms of the T in a horizontal plane; it then received a quarter turn so that the arms of the T were brought into a vertical plane, and it was then pulled back, when the returns of the T were made to fit into two holes provided for them on either side of the slit in the bolt, on the inside, figs. 1C and 2C, Plate IV. By this means the key obtained a grip of the bolt, and it was only necessary to press it to one side in