Late Roman Belt-fittings from Canterbury

LATE-ROMAN BELT-FITTINGS FROM CANTERBURY B.M. AGER This note describes and discusses six late Roman and one specifically late Romano-British belt-fitting from recent excavations in Canterbury.1 Descriptions Strap-end of copper alloy plate, the attachment end broken off (Fig. l,a); leaf-shaped with an elongated, forked tip. It is bordered by punched U-shapes standing on contour lines and two more of these punches are placed base to base in the middle of the front. There is incised feathering at the tip, the edges of which are bevelled underneath, and a median groove with two punched dots in it towards the beginning of the tip. Length: 78 mm.; width: 14 mm. Found in the late fourth-century lower 'dark earth' layer overlying the Roman levels at Stour Street in 1986; 681/501. Strap-end of copper alloy (Fig. 1,b), composed of two dart-shaped leaves of plain sheet joined by a single bronze rivet at the butt end, which has a scalloped edge and two lateral lobes. Length: 25 mm.; width: 9 mm. Found in the 'dark earth' layer on the Marlowe IV site in 1980; 443/1583. Strap-end of plain copper alloy sheet with split end for insertion of the strap (Fig. 1,c); damaged and tip missing. There are two round holes 1 I should like to thank Mrs. P. Garrard, of the Canterbury Archaeological Trust, for inviting me to publish these fittings and for her kind assistance in providing me with copies of the drawings by M. Duncan, M. Ellam, R. Mair and G. Hulse, which are reproduced here by generous permission in advance of site reports. 25 B.M. AGER C e Fig. 1. Late-Roman Belt-fittings from Canterbury (Scale: 1/1). 26 a LATE-ROMAN BELT-FITIINGS FROM CANTERBURY barely above the base of th,e split which can only be decorative; the holes for the attachment rivets are at the butt end and both are damaged. Length: 31 mm.; width: 17 mm. Found in a pit dated c. 850-1050 on the Marlowe IV site in 1980; 70A/243. Strap-end of copper alloy (Fig. 1,d) of broad, pointed oval shape, broken across the base of the split end, the start of which is just visible from one side. The front is decorated with a central, punched ring-and-dot motif, with four smaller ones equally spaced around it. Length: 25 mm.; width 17 mm. Found in a pit dated c. 1400-1550 on the Marlowe Theatre site in 1982; 55/274. Buckle of copper alloy (Fig. l,e) with oval loop, around the hinge-bar of which is doubled the bevel-edged belt-plate of oval sheet. Of the original three rivets, one copper one remains in the front and one bronze one in the back-plate. The front is decorated with a ring-anddot motif next to the slot for the tongue and with single punched circles around the rivet holes. Length: 55mm. (approx.); height: 33 mm. Found in the primary layer of c. 350-375 of a Roman building on the Marlowe IIB site in 1979; B 266G/1983. Buckle belt-plate (Fig. 1,f) of the same type as above, but half missing. Two ring-and-dots remain and one bronze attachment rivet. Length: 23 mm.; height: 22 mm. Found in the late Roman third- to fourth-century demolition layer at Stour Street in 1986; 248/244. Buckle of copper alloy (Fig. 1,g) with damaged rectangular plate doubled over the hinge-bar of a trapezoid loop with roughly bevelled edges. A slot is cut in the plate for the tongue, which is missing. Length: 52mm. (approx.); height: 33 mm. Found in a layer dated c. 400-450 on the Marlowe IIB site in 1979; B318/1843. Discussion More than twenty examples of fork-tipped strap-ends are now known from Britain. 2 They belong principally to buckles of Hawkes and 2 More recent and additional finds to those noted by S.C. Hawkes, 'A late Roman buckle from Tripontium', Transactions of the Birmi11gham and Warwickshire Archaeological Society, Ixxxv (1972), 145-59,. note 10, and G. Clarke, The Roman Cemetery at Lankhills; Pre-Roman and Roman Winchester, II (Oxford, 1979. Winchester Studies 3), 281-2, include the pieces from Chichester, West Sussex in A. Down, Chichester Excavations, III (Chichester, 1978), Fig. 10.37, no. 94; Lydney Park, Glos. in R.E.M. and T.V. Wheeler, Report on the Excavation of the Prehistoric, Roman, and 27 B.M. AGER Dunning's horse-headed type IB,3 which appear to have been manufactured during the last three decades of the fourth century and were being worn out by the early fifth.4 One of these buckles, typologically early and nearly new in condition, from Tripontium, Warks., was found in the upper level of a fourth-century ditch and was probably lost in the third quarter of that century. 5 They are an entirely insular late-Roman type, although the strar,-ends are based on the amphora-shaped form of Simpson's type 'b'.6 The strap-ends have been dubbed the 'Tortworth' type by Clarke,7 after an example from Tortworth, now in Avon.8 post-Roman Site in Lydney Park, Gloucestershire (Reports of the Research Committee of the Society of Antiquaries of London, ix, 1932), Fig. 19, 82; Cirencester, Glos. in J. Ward, The Roman Era in Britain (London, 1911), Fig. 70, i; Winterton and Wintringham, Lines. published by K.A. Leahy, 'Late Roman and early Germanic metalwork from Lincolnshire', in (Eds.) N. Field and A. White, A Prospect of Lincolnshire (Lincoln, 1984), 23-32, Fig. 2, 17 and 19; and an unpublished example from excavation at the County Hospital, Dorchester, Dorset recorded by C. Sparey Green, 'A late Roman Buckle from Dorchester, Dorset', Britannia, xv (1984), 260-4, p. 262. A strap-end from Canterbury itself, listed by Mrs. Hawkes but unpublished at the time, has since appeared in S.S. Frere and S. Stow, Excavations in the St. George's Street and Burgate Street Areas (The Archaeology of Canterbury, VII, Maidstone, 1983), Fig. 61,6. 3 S.C. Hawkes and G.C. Dunning, 'Soldiers and Settlers in Britain, fourth to fifth Century', Med. Arch., v (1961), 1-70. 4 S.C. Hawkes, 'Some recent Finds of late Roman Buckles', Britannia, v (1974), 386-93, p. 387. See also Clarke op. cit. in note 2, 287. 5 Hawkes op. cit. in note 2. 6 C.J. Simpson, 'Belt-buckles and Strap-ends of the later Roman Empire: a preliminary Survey of several new Groups', Britannia, vii (1976), 192-223. 7 Op. cit. in note 2, 281. 8 Clarke is fundamentally mistaken in saying that the Richborough example is ornamented with a Quoit Brooch Style animal and for including the Wor Barrow strap-end (not 'nail-cleaner'; cf. the Colchester hinged strap-end noted below) amongst early Anglo-Saxon Quoit Brooch Style objects. Neither of these pieces has been included in the latest catalogue of items decorated in this style listed by V.I. Evison, The Fifth-Century Invasions South of the Thames (London, 1965), 62; on the contrary, they are given as examples of the transference of a flat, pictorial style across the Channel to late Roman Britain before the introduction of the style, ibid., 59. The resemblance noted by Clarke between the face with an absurd, triangular nose incised on the strap-end from Rockboume, Hants. and the terminal faces of the Quoit Brooch Style belt-set from Mucking, Essex, grave 117, appears to owe more simply to equally inept copying of a common Roman motif than to any closer affinity. It should rather be noted that the Christian symbolism apparent on some of these strap-ends, as observed by Hawkes, op. cit. in note 2, 151-8, is totally absent in the Quoit Brooch Style. As mentioned elsewhere, the choice and even the number of motifs employed in this style appear to be determined by pagan considerations; see my article, 'The smaller Variants of the Anglo-Saxon Quoit Brooch', in (Eds.), S.C. Hawkes et al. AngloSaxon Studies in Archaeology and History, iv (1985), 1-58, pp. 12, 15 and note SS. It is 28 LATE-ROMAN BELT-FITIINGS FROM CANTERBURY It was originally proposed by Mrs. Hawkes that these fittings were items of military equipment, issued to a kind of yeomanry force based on the towns of late-Roman Britain.9 But she has since suggested that, in view of their slightness and fragility especially, they seem more appropriate to non-official civilian dress, while it is probably significant too, in so far as regards fashion, that they are sometimes found in Anglo-Saxon female graves; even the wider, chip-􀁑arved fittings, generally accepted as being from cingula militiae, the belts worn by both civil and military officials from the reign of Diocletian onwards, were very occasionally worn by women on the Continent.10 This reassessment is supported by Leahy's observation11 that IB buckles (together with IA) are much more common on civil than on military sites. Furthermore, two fork-tipped strap-ends of a related, hinged type were found in a child's burial of the mid to later fourth century at Butt Road, Colchester.12 Also it seems that, contrary to an opinion sometimes expressed, belts did in fact form part of late Roman female costume. A woman is shown wearing a belt ornamented with propeller-sha􀁒ed stiffeners on a stele from Intercisa, possibly of the third century;1 another woman with a belt can be seen in the panel below. Another Roman woman, possibly the wife of the barbarian ieneral Stilicho, is portrayed on a late fourth-century ivory diptych1 wearing a belt, apparently studded with gems, over her dress, a fashion that is likely to have been imitated in other ranks of society. Thus, in itself the wearing of belts in the fourth and fifth centuries is neither specifically military nor even Germanic and male. The form of the belt and the contexts of finds are more reliable indicators and, in this light, type I fittings can no longer be assumed to be possible evidence for British military dispositions at this time, but seem more likely to be civilian. It is much to be hoped that fresh evidence will finally resolve this question. argued there that the style is Anglo-Saxon, though adapted largely from late Roman Continental models. Tortworth type strap-ends should be distinguished from so-called 'nail-cleaners' which also have forked tips, but are narrow-stemmed and may form part of a cosmetic set, e.g. the examples from London and other places, illustrated by J. Ward, op. cit. in note 2, Fig. 70c, d, h and possibly j-1. 9 Op. cit. in note 3, 41. 10 Op. cit. in note 4, 393. 11 Op. cit. in note 2, 23. 12 P. Crummy, Aspects of Anglo-Saxon and Norman Colchester (CBA Research Re􀁨rt no. 39, 1981), Fig. 15. 1 H. Bullinger, 'Spatantike Gurtelbeschlage: Typen, Herstellung, Trageweise und Datierung', Dissertationes Archaeologicae Gandenses, 12 (Bruges, 1969), Tafn. 67 and 69. 14 P. Dixon, Barbarian Europe (Oxford, 1916), ill. on p. 16. 29 B.M. AGER The other items of belt-equipment from Canterbury, however, probably would have been made for cingula militiae, and are comparable with examples that are commonly found at or near military sites in the frontier districts of the Roman Empire. But, since our finds are unassociated, it is impossible to be certain whether their last owners were soldiers (serving or retired) or civilians and, if the former, whether they came from regular provincial units, either billeted in the town or in transit, or from auxiliary barbarian forces, e.g. laeti.15 Such fittings are not indicative of any particular national origin.16 Their plainness does not preclude official rank, however, as simple forms are found associated with crossbow brooches, which are a sign of high status 1 e.g. at Lankhills, Winchester, Hants., graves 13, 23, 81 and others. 7 The pointed shape and pair of lobes at the neck of the small, dart-shaped strap-end from the Marlowe IV site connect it with a larger version from Wizernes, Pas-de-Calais, France, 18 probably from an inhumation grave which could also have contained a buckle datable to Bohme's phase I c. A.D. 350-400, or at least to the last third of the fourth century. 19 The perforated strap-end with split end, also from Marlowe IV, is a rather crude example of Simpson's amphora-shaped type 'b', with the C or heart-shaped cut-outs reduced to just two holes. It is clearly a residual find, but on the Continent this type was in use from c. A.D. 350 until c. A.D. 390, and in Britain possibly until c. A.D. 410, as at Lankhills.20 The pointed oval strap-end from Marlowe Theatre is residual, too; it has a close parallel, also with ring-and-dot decoration and broken at almost the same point, but with more rounded outline, in a provincial Roman example of Bohme's phase III, c. A.D. 400-450, from the cremation cemetery at Perlberg, Lower Saxony, Germany. 21 The two buckles with oval plates from Marlowe IIB and Stour St. belong to Simpson's group II, and the context of c. A.D. 350-375 for the 15 Further possibilities are considered by S. James, 'Britain and the late Roman Army' in (Eds.), T.F.C. Blagg and A.C. King Military and Civilian in Roman Britain (BAR Brit. Ser. 136, 1984), 161-86, pp. 167-8. 16 Simpson, op. cit. in note 6, 205. 17 Clarke, op. cit. in note 2, figs. 67, 68, 71. 18 H.W. Bohme, Germanische Grabfurule des 4. bis 5. Jahrhunderts zwischen unterer Elbe und Loire (Munich, 1974), 337-8, Taf. 147, 7-8). 19 On historical grounds the beginning of this phase is placed slightly later, during the last third of the fourth century, by G. Mildenburger, in his review of Bohme's work in Germania, 53 (1975), 256-63. 20 Clarke, op. ciJ. in note 2, 281. 2l Op. cit. in note 18, Taf. 33, 7. 30 LATE-ROMAN BELT-FITTINGS FROM CANTERBURY former is important for the dating of this type. Four examples were found at Lankhills, where they were in vogue between c. A.O. 350-410, which is broadly consistent with Continental dating to c. A.O. 340-400, though mostly before c. A.O. 370.22 Lastly, the buckle with trapezoid loop from Marlowe JIB belongs to Simpson's group IV. It is quite possible that it is contemporary with its context of c. A.D. 400-450 rather than residual; two examples from Lankhills are dated c. A.D. 350-390 by Clarke, who concludes that they were becoming increasingly common towards the end of the fourth century.23 All of the cingulum fittings from Canterbury could have been made at a local workshop, such as the one which is known to have existed at lckham, Kent,24 rather than at distant state factories in Pannonia or lllyricum. Intended as official issue, might they then have been purchased by the townsfolk at a late Roman equivalent of the army surplus store? More evidence from cemeteries should help us to answer the questions involved in the military/civil options and their further implications for late Romano-British history. 22 Clarke, op. cit. in note 2, 270-272. 23 Ibid., 272-3. 24 C. Young, 'The late Roman Water-mill at lckham, Kent, and the Saxon Shore' in (Ed.} A. Detsicas, Collectanea Historica. Essays in Memory of Stuart Rigold, Maidstone, 1981, 32-40. The damaged oval buckle-plate 248/244 from Stour Street (Fig. lF) was found associated with a metal-working hearth beside the Roman Watling Street and, while there is no certain evidence that this or other belt-fittings were being manufactured here, a slight possibility remains. 31