BYZANTINE BICEPHALIC SIGNET OF THE SAME METAL. 307 used ill favour of the Christian origin of my bronze key- ring'^ confirm the latter inference. The kindly words upon it arc such as may be found both on pagan and Christian rings and ling-stoncs of that period, for the most part, however, conveyed in the Greek rather than in the Latin langunge. In support of the former suggestion I have figured another bronze key-ring from my own collection, found at Rome ; devoid of ornamentation, but corresponding in form, and probably of about the same period as its more noble contemporary. This was undoubtedly made for service as a key. In the Waterton Collection, at the South Kensington 3Iuseum (No. 551, '71), is one with a circular lift, pierced with seven holes, and attached to the hoop by a neck, as is the case on mine. Our information on the origin, uses, and import of these key-rings, which are found abundantly on various Roman sites, and in great variety of form, is still very imperfect. Learned writers ^ have connected them with various quota- tions from the classic authors, having reference to the habit of securing the casket or the wine cupboard of ancient days, but in so doing they have frequently superadded confusion to our imperfect knowledge ; the method of securing by the impression of the signet ring on clay or wax, attached to the door by a cord or otherwise, being sometimes con- founded with the rings in question, which were presumably formed for turning locks or lifting latches. That they were for both these uses is assured by an e.xamination of various examples in collections, and by a reference to the figures given by Licetus,* Montfaucon, and other writers. The lateral wards attached to a pipe projecting from the ring- formed handle can be for no other use than turning on a central pin, and driving the bolt ; wdiile on other varieties the projecting tongue could onl}' be used for lifting a latch, on the plan of the so-called " French latch-lock." One - Arch. Journal, vol. x.xviii, p. 290. Georgius Longiis and Francisco de Corte ' See Gorlaei, Dactylioth. 42, 205-209. have badly reproduced. One of them
- Licetus de Anulis, I6ir>, figures has the figure of a rudder between two
seven varieties, of which one is a latch. ears of corn engiaved on the bezel. His Six are with pipes ; of these Nos. 2, 5 and No. 8 is precisely similar to one in my (J are, I Bunpect, the original cut*, which collection.
