as in the Reeth find. Other specimens of these smaller discs may be seen in Plate LXXIV., Figs. 1, 3, 8 and 10.

No doubt the hanging ornaments attached to phalerae varied considerably in shape. On the horse of Bassus (Fig. 42, page 296) long floating streamers take the place of the more rigid leaf-ornaments we have been describing, and for these also there were probably employed metal terminals much like the two figured in Plate LXXII., Figs. 8 and 10, and some of those shown in Plate LXXVI., Figs. 4 to 10, 12, 13, 15, 16, though they may equally well have served for the ends of girdles. In the remarkable finds from the mosses of Schleswig, especially that from Thorsbjerg, we have interesting examples of the richness of early harness. Probably these contain objects that are not purely Roman, and they are perhaps a little later in date than the earliest occupation of Newstead. At the same time the Roman influence is strongly marked. It is instructive to note that the Thorsbjerg find included a number of hanging terminals which are identified as belonging to harness, and which clearly belong to the same class as those of which we have been speaking (Plate LXXII. and Plate LXXVI.).[1]

Terminals of the same pattern as those in Plate LXXII. were found at Haltern.[2] In these, too, the surface of the upper portion of the mounting was plated with silver, a method of decoration less common in the Antonine period. Of the examples brought together in Plate LXXVI., Fig. 8 from the Principia and Fig. 7 from the Praetentura are very similar. The longer is plated with some white metal, and inlaid with a design in niello. Fig. 9, which is of bronze like the preceding pieces, is leaf-shaped. In the case of Fig. 6 a portion of the leather to which the mounting was originally attached still survives. It is, of course, possible that some of the metal endings may have come from the long fringed portion of military girdles. Such are shown hanging down in front in many of the monuments.

Strap Mountings

The Thorsbjerg find, which has already proved helpful, further illustrates the use of such loops as we find in Plate LXXII. These are made of comparatively thin metal, except for the actual loops, which are stronger. It is evident that they were fixed to the ends of straps, for in some of them the rivets are still in position. A similar article was employed at Thorsbjerg as part of a chain-bridle, its purpose being to serve as the joint

1 Engelhardt, Denmark in the Early Iron Age, plate 15, figs. 38 to 48.

2 Mitteilungen der Altertums-Kommission fur Westfalen, Heft iii. Taf. xv. 4.