^{page 179 note a} Report for 1894 of Northumberland Excavation Committee, Archaeologia Æliana, vol. xvii. pp. xxii.-xxxii. Sec also Mr. Haverfield's notes, post.

^{page 180 note a} See below, p. 20.

^{page 180 note b} See below, p. 15.

^{page 182 note a} One probably found in the Perugia district is engraved in Montelius, Spännen från Bronsåldern, p. 186, fig. 188. But this earlier type is certainly rare in Italy, though a later variety with the foot no longer open but forming a single plate is occasionally found in N. Italy. I have seen examples from Liguria.

^{page 182 note b} See Tischler, Gurina, 26.

^{page 183 note a} See Evans, A. J., “On a Late-Celtic Urn-Field at Aylesford, Kent, etc.” Archaeologia, lii. 382, 383Google Scholar.

^{page 184 note a} In the British Museum.

^{page 184 note b} Excavations in Cranbourne Chase, vol. ii. pi. xcvii. 9–12.

^{page 184 note c} Vol. xxxi. (1845) pp. 279 et seqq. “On some Roman Vestigia recently found at Kirkby Thore, in Westmoreland.”

In Sir John Evans's Collection.

^{page 185 note a} ArchÆological Journal, viii. (1851) 35et seqq. The fibula is engraved on p. 39. For the inscription on the ring referring to the DeÆ Matres, see C. I. L. vii. 1299, and Haverfield, Romano-British Inscriptions, iii. 158 (in Archœological Journal, 1. 303), where the objects are said to have been probably found at Backworth.Google Scholar

^{page 185 note b} Less conspicuous examples of the same class are to be seen in the British Museum, from London (Roach Smith) and Farley Heath, Surrey. But in the south of England this type is exceptional and exotic.

^{page 189 note a} In varieties of this fibula, however, its form is lozenge-shaped.

^{page 189 note b} From Sir John Evans's collection. Numerous examples from the Grallo-Roman station of Châtelet, between St. Dizier and Joinville, are engraved in Grivaud de Vincennes, Arts et Métiers des Anciens, pl. 40.

^{page 188 note a} Thus in a homogeneous deposit found at Hochbübel near Meran, fibulÆ of this type are associated with others of the Certosa type, the date of which is fixed by the associated Greek pottery in the Bologna graves to the early part of the fifth century, B.C. Double disked brooches such as those described below were found in a bucket at Nonsberg near Dercolo, in a broken condition, associated with now Certosa fibulÆ, a proof in this case of anteriority of date. The evidence of the graves of Idria di Bača, near Gorizia, is also conclusive as to the very early date of the brooches with a single disk on the retroflected end. In Grave 14 one such occurred with a bronze vessel, very similar to that from the cremation grave (No. 500) of the great Hallstatt cemetery. The early date of the retroflected fibulÆ in south-east Europe has not been sufficiently taken count of by archÆologists.

^{page 189 note a} S. Reinach, Art. Fibula in Darenberg et Saglio, Diet. des Antiquités. A specimen from this site is in the Museum of St. Germain.

^{page 189 note b} Bianchetti, I Sepolcreti di Ornavasso (Turin, 1895), p. 32, pi. x. 18

^{page 189 note c} H. Lugon, Anzeiger für Schweizerischer Altertümskunde, 1892, No. 2, cited by Bianchetti, 1.a.

^{page 189 note d} Lindenschmit, Die Alterthümer etc., band iii. heft ix. taf. 3. Lindenschmit commits himself to the statement that this fibula is common in Rhenish graves of the third century or even later; but this is more than doubtful.

^{page 189 note e} In a letter to Mr. Haverfield, who kindly asked his opinion for me

^{page 191 note a} See especially Smith, J. Alexander, “Notes on a Bronze Late-Celtic Armlet,” in the Proceedings of the Society of Antiquaries of Scotland, xv. 316seqqGoogle Scholar. Examples of these are cited from Stanhope, Peeblesshire (op. cit. xv. 318, fig. 1); Alvah, Banffshire (op. cit. vi. pl. iii. and xv. 328, fig. 8, and 329, fig. 9); Castle Newe, Aberdeenshire (op. cit. xv. 330, fig. 10); Drumside, Belhelvie, Aberdeen-shire (op. cit. xv. 333, fig. 14 and 334, fig. 15); Bunrannocli (?), Perthshire (op. cit. xv. 336, fig. 16 and 337, fig. 17); Pittkelloney, Muthill, Aberdeenshire (op. cit. xv. 340); Kinghorn, Seafield, Fifeshire (op. cit. xv. 342); one example has been found near Newry, County Down, Ireland (op. cit. xv. 362. fig. 31).

^{page 191 note b} This illustration has been kindly lent by the Society of Antiquaries of Scotland.

^{page 191 note c} Archaeologia, xiv. 90.

^{page 191 note d} York volume of the Archaeological Institute, p. 10 and pl. 24. The thin plated bronze buckle of this find is suggestive of a somewhat late date. It approaches the types of the Danish Moss finds. e.g. Thorsbjerg.

^{page 191 note e} Lindenschmit, Die Alterthümer, etc., band ii. heft vi. taf. 2, fig. 2.

^{page 192 note a} Proc. Soc. Ant. Scot., xv. 316 seqq.

^{page 192 note b} Ibid. 351, fig. 28.

^{page 192 note c} Wilson, , Prehistoric Annals of Scotland, ii. 146, fig. 131Google Scholar.

^{page 192 note d} This illustration has been kindly lent by the Society of Antiquaries of Scotland.

^{page 192 note e} Proc. Soc. Ant. Scot., xv. 345, fig. 22, and xxv. 505, fig. 30.

^{page 193 note a} Hampel, Pannonische Costumbilder, Ungar. Revue, 1881. 147–153. This form of fibula also extends to Noricum. Tischler, Gurina, 26 note.

^{page 194 note a} Tischler, Gunrina, 27.

^{page 194 note b} Pannonians appear in one of the Cumberland forts just south of the Wall (C. I. L. vii. 417, Ephem. vii. 978), and we have the tombstones of individual Pannonians at Littlechesters (C. I. L. vii. 723), and probably at Housesteads (ib. 692). (F. H.)

^{page 195 note a} The excavations of 1895 seem to show that the masonry of the Wall was bonded into the masonry of the fortress wall. In other words, there are indications that the Wall and this fortress were built at one and the same time.

^{page 195 note b} The areas of the fortresses, according to Bruce, are as follows: Birdoswald and Bowness, 5½; acres; Chesters, 5¼; Housesteads and Benwell, 5; Halton Chesters, 4¼; Wallsend, Carrawburgh, Carvoran, 3½; Rutchester, Chesterholm, Great Chesters, 3¼; Burgh, 3; Castlesteads, 2¾; Stanwix, 2½; the area of Newcastle is unknown. East of Chesters (inclusive) the north faces of the fortresses project beyond, the Wall; west of it (from Carrawburgh inclusive), the north faces of the fortresses coincide with the Wall, or else as at Chesterholm, Carvoran, and Castlesteads the fortress is south of the Wall.

^{page 197 note a} Bruce, Handbook, p. 101.

^{page 197 note b} Bruce, Wall, 155.

^{page 197 note c} See der Obergermanisch-raetische Limes, No. 44, Kastell Murrhardt, pp. 9–10.

^{page 197 note d} The western gateway, excavated in 1895, was found to have been wholly walled up; the masonry indicated that one-half had been closed first and the other at a later period.

^{page 198 note a} Hall, Rome, Archaeologia Æliana, xvi. 443Google Scholar.

^{page 198 note b} See Müller, Albert, Philologus, xl. (1881) 248 foll.Google Scholar; Lindenschmit, Tracht und Bewaffnung des römischen Heeres, pl. xii. Presumably it was also used by auxiliaries, as on the Wall.

^{page 198 note c} It will be found in full, Archaeologia Æliana, xvii. pp. xxx–xxxiGoogle Scholar

^{page 198 note d} Excavated in 1892–3 by the Cumberland and Westmorland Architectural and Archaeological Society; see their Transactions, xiii. 375 et seqq.