9
Some Recent Finds of Late Roman Buckles Author(s): Sonia Hawkes Source: Britannia, Vol. 5 (1974), pp. 386-393 Published by: Society for the Promotion of Roman Studies Stable URL: http://www.jstor.org/stable/525745 . Accessed: 24/06/2014 20:15 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. . Society for the Promotion of Roman Studies is collaborating with JSTOR to digitize, preserve and extend access to Britannia. http://www.jstor.org This content downloaded from 188.72.96.55 on Tue, 24 Jun 2014 20:15:54 PM All use subject to JSTOR Terms and Conditions

Some Recent Finds of Late Roman Buckles

Embed Size (px)

Citation preview

Page 1: Some Recent Finds of Late Roman Buckles

Some Recent Finds of Late Roman BucklesAuthor(s): Sonia HawkesSource: Britannia, Vol. 5 (1974), pp. 386-393Published by: Society for the Promotion of Roman StudiesStable URL: http://www.jstor.org/stable/525745 .

Accessed: 24/06/2014 20:15

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Society for the Promotion of Roman Studies is collaborating with JSTOR to digitize, preserve and extendaccess to Britannia.

http://www.jstor.org

This content downloaded from 188.72.96.55 on Tue, 24 Jun 2014 20:15:54 PMAll use subject to JSTOR Terms and Conditions

Page 2: Some Recent Finds of Late Roman Buckles

386 NOTES

essentially non-coin-using context. Equally puzzling is the agency of transmission of the coin, though the presence of St. Germanus in Britain at about the period of issue of the coin demonstrates one avenue of contact with the continent which was still open. It should be emphasized, however, that the VOT PUB coin is unlikely to have derived from Gaul since its circulation is restricted to Italy. On the whole it would seem that this single coin represents little more than a casual drift from the shrinking centre of the Empire to the fringes of an increasingly irrelevant former province.

Department of Archaeology, University of Durham

Some recent finds of Late Roman Buckles. Mrs. Sonia Hawkes writes: In March 1972 a decorated bronze buckle-plate was picked up from 'the surface of the Roman fort at Greta Bridge' in the North Riding of Yorkshire (Fio. 3, I).64 It is a fine example of the Type I which has been much written about in recent years,65 and which is considered to have been produced by Romano-British workshops during the second half of the fourth century. The fort and its vicus are at present ill explored66 and this plate, though a surface find, may be regarded as a precious indication of late occupation. It is a front-plate only, length

7"35 cm, width 1-6 cm, competently engraved with four

cross-hatched lozenges, which haye punched circlets at their angles, and bordered by simple angular punch work. Though now incomplete and slightly damaged, it shows little sign of original wear. Type I buckle-plates with stylistically similar geometric ornament have been found at nearby Catterick67 and at several sites further south.68 This is not the only new find of its type. From a late Roman pit on a site at Red Lodge, Ducklington, Oxon.,69 has come a battered fragment which, despite its slovenly and uncharacteristic zig-zag ornament, is likely to have been part of another Type I plate (FIG. 3, 2). Such plates, long and narrow, made of doubled sheet metal, were attached to small D-shaped buckles with confronted dolphin-heads (Type IA) which, on the best- made examples, generally had crests in the form of horse-heads (Type IB). To the already lengthy list of Type IB buckle-loops we may now add one in Stroud Museum,70 found in 197 on the surface of Site 15 in the field called Lower Chessalls at Kingscote, Gloucs.7 Here there was an extensive Roman settlement which had remained in occu- pation until at least the end of the fourth century. The buckle (FIG. 3, 3) must once have

closely resembled one from another and similar Gloucestershire site, at Wycomb7Z dolphins with buffer-jaws, circlet-eyes on well-moulded heads, a line of punched arcs at the base of each horse-head-but as well as plough damage, it shows much original wear, especially on the left side. A broken perforation through the remains of the right

64 Information from the finder by courtesy of Mrs. S. Thubron. The object, sent to Oxford in March 1973 for study and drawing (by Mrs. Marion Cox), has now been returned to the finder, Mr. E. W. Lloyd, by permission of whom it is published here.

65 S. C. Hawkes and G. C. Dunning, Medieval Archaeol. v (1961), 1-70 but esp. 21 iff. and Catalogue 41 ff-.; 43-4 Bericht der Ri6misch-Germanischen Kommission (1962-3), 155-231 but esp. 18 1 ff. and Catalogue 204 ff. S. C. Hawkes in A. C. C. Brodribb, A. R. Hands and D. R. Walker, Excavations at Shakenoak i (1968), 96-1oi, and Shakenoak iii (1972), 74-7; Trans. Birmingham Warwickshire Archaeol. Soc. lxxxv (I972), 145-59.

66 B. R. Hartley in R. M. Butler (ed.), Soldier and Civilian in Roman Yorkshire (1971), 58; cf. p. 413 below. 67 M. Pocock, Yorkshire Archaeol. J. xliii (197i), 187, pl. I; Hawkes, op. cit. (1972), 146, fig. 1,4-. 68 From Dorchester-on-Thames, Cirencester, Duston, Silchester and especially Popham, Hants., cf.

Hawkes and Dunning, op. cit. (1961), 47, 50, figs. I,I4 and 15, n-q. 69 Information from Mr. M. Aston of the Oxford City and County Museum. The site is being excavated

by Mr. G. Williams and the Witney Archaeological Group, by whose permission it is published here. The drawing (by Mr. M. Rouillard) is based on a Museum photograph and a sketch by Mr. Williams.

70 Reg. no. 71.421. It is published here by permission of the Museum's Committee. I am much indebted to the Curator, Mr. L. F. J. Walrond, for sending a detailed description and drawing on which that published (by M. Rouillard) is based.

71 B. N. Eagles and V. G. Swan, Trans. Bristol & Gloucestershire Archaeol. Soc. xci (1972), 6o ff. 72 Hawkes, op. cit. (note 65) (1972), 146, fig. I,3.

This content downloaded from 188.72.96.55 on Tue, 24 Jun 2014 20:15:54 PMAll use subject to JSTOR Terms and Conditions

Page 3: Some Recent Finds of Late Roman Buckles

NOTES 387 horse-head suggests that it may even have seen secondary use as a pendant. The fourth find, a Type IB buckle, complete with plate, is not figured. It was found very recently, together with two Saxon brooches, in an early fifth-century woman's grave at Mucking, Essex.3 This is one of a group of burials in the south-east part of Saxon cemetery II which had accompanying finds of late Roman and early Saxon character, including another dolphin buckle and a fixed-plate buckle with chip-carved attachment-plate. They are the earliest burials in this cemetery. The ornamental details on the Type IB buckle are at present obscured bycorrosion, but in general it seems to be as abraded by wear as the comparable buckle in the similarly dated Saxon woman's grave at Dor- chester-on-Thames.74 Of the cruder Type IA loops, two degenerate examples have recently been recognized in the material from the Saxon cemeteries of Long Wittenham, Oxon., and Kempston, Beds.7s

These finds bring the total of complete or fragmentary Type I buckles known to me to over 40. Of these three come from Roman forts, 12 from walled towns, one from a fortified road station, seven from rural settlements, six from villas, two from the graves of womenfolk of Saxon mercenaries in the late Roman forces, and seven (mostly of Type IA) from later Saxon graves. The dating evidence lacks any real precision, but the abraded condition of Type IB buckles from fifth-century contexts, and the occurrence of a typologically early specimen in pristine condition in a deposit of c. 370,76 suggest that they were being manufactured during the last three decades of the fourth century and being worn out by the early fifth. Possibly Type IA buckles continued to be made in the early fifth century. The overall distribution is thickest in the South and West Midlands and generally southerly, but the find from Greta Bridge reinforces the impor- tance of the small group, comprised otherwise of the Catterick and Stanwick buckles, apparently isolated in the north (FIG. 4).

Other finds of late Roman zoomorphic buckles, found or brought to my notice since the publication of the original catalogue, belong chiefly to my Type IIA, a mainly insular series nearly as numerous as Type I, which is characterized by loops in the form of confronted dolphins with involuted tails and cast hinged plates with arcade decoration in openwork.77 Complete specimens are rare and it is possible that some loops published in the original list, e.g. Nos. 5 and 7, may have had a different form of plate,78 but until there is a more numerous and representative corpus of this material generally it is best not to attempt a reclassification. The additions consist of two loops and four plates. The loops both come frorr Lincolnshire: one (FIG. 3, 4), width 5-7 cm, a fine example with realistic dolphins with notched crests and body scales, complete with a simple tongue, is a surface find from Dragonby by Scunthorpe, a rural site with evidence of late fourth-century occupation;T9 the other (FIG. 3, 5), width 4-2 cm, a more stylized piece with missing tongue, was found in 1968 at Lincoln just outside the western defences of the Lower Colonia, close to a projecting rectangular tower of fourth-century date.so There is an isolated plate, width nearly 4 cm, with three elongated and three circular openings, amongst the finds from the late fourth/early fifth-century occupation of the

73 I am much indebted to Mr. W. T. Jones for showing me this buckle, for sending me further details by letter, 20 January 1974, and allowing me to make use of them in this note. The grave number is 987; the associated brooches a small-long brooch and a brooch of 'Luton' type, better known as a Stiitzarrmfibel of Werner's type B, Bonner Jahrbficher clviii (1958), 376 ff.

74J. R. Kirk and E. T. Leeds, Oxoniensia xviii (1952-3), 63-76; Hawkes, op. cit. (1972), 150. 75 V. I. Evison, Antiq. Journ. xlviii (1968), 238, fig. 4a, e. 76 Hawkes, op. cit. (note 65) (1972), 150 f. 77Hawkes and Dunning, op. cit. (note 65) (1961), 21 if, 50 ff, map fig. 9; op. cit. (1962-3), i8i if, 212 if,

map fig. 9- 78 Giles Clarke, Antiq. Journ. 1 (I970), 292-8, fig. 4, no. 92. 79 Information from Jeffrey May, 2o September 1967. I am indebted to him for a photograph of the buckle. Cf.. also J. May, Antiq. Journ. 1 (1970), 222-45, for the site generally.

8oj. B. Whitwell, Roman Lincolnshire (1970), 137. I am indebted to him also for information by letter, 7 March 1969, and for copies of drawings of the Lincoln and Dragonby buckles.

This content downloaded from 188.72.96.55 on Tue, 24 Jun 2014 20:15:54 PMAll use subject to JSTOR Terms and Conditions

Page 4: Some Recent Finds of Late Roman Buckles

388 NOTES

e3

q 4

2

5

6

., 8

FIG. 3 Late Roman buckles (Nos. 1-7 (1), No. 8 (i)).

This content downloaded from 188.72.96.55 on Tue, 24 Jun 2014 20:15:54 PMAll use subject to JSTOR Terms and Conditions

Page 5: Some Recent Finds of Late Roman Buckles

NOTES 389 hillfort on Traprain Law in East Lothian.8, Another such plate, in broken condition but originally with four rectangular and eight circular openings (FIG. 3, 6) was found at Cirencester in 1956 'in a deposit dated by underlying coins of 345-8'.82z The late Roman fort at Richborough has yielded a fragment, and a burnt scrap from the Saxon cemetery at Droxford, Hants., may have been part of another.83

From Wye in Kent comes a plate similar to that from Cirencester but cast in one piece with a simple non-zoomorphic loop (FIG. 3, 7): it is thus a variant to be placed in Type IIB. Despite its broken tongue it is in unusually good condition and still attached to its plate are the flimsy strips of sheet bronze which were applied to the inside of the belt to reinforce the rivets. It is difficult to imagine how it could have survived in this state unless it were still attached to at least part of its belt when buried. The belt will have been a little over 4 cm broad. In close association with the buckle was a well- preserved iron knife, length originally just over I17 cm, of distinctive type with broad, short blade, which has a curved cutting edge and a markedly angled back, a long tang and two oval iron plates, thinly capped with bronze, which had supported the now decayed hilt and pommel (FIG. 3, 8). Parallels to this knife occur with buckles, other belt- fittings and cross-bow brooches, in the Lankhills cemetery at Winchester, with male burials of the mid fourth century which are regarded as both Germanic and military.84 Abroad they are to be found in the cemetery of the fort of the Litus Saxonicum at Ouden- burg in Belgium.8s One would have expected the Wye finds to have come from a grave too, but in fact they appear to have come from a coin-rich rubbish deposit on a settle- ment site, datable perhaps within the 370's.86 In view of the character of knife and buckle and their high state of preservation, the context seems distinctly odd and it is to be hoped that further excavation will determine the exact nature of the site.

Buckles of Type II known to me now number over 30. Of these two come from a fort of the Litus Saxonicum, one from a western port, 12 from walled towns, five from rural settlements, three from villas, two from a temple and five from Anglo-Saxon graves. They are distributed more evenly over the civil zone than the Type I buckles; the most northerly find (except that at Traprain) is in north Lincolnshire and there is not the same westerly concentration (FIG. 5). We need not go into the details of the insular dating evidence; it has been discussed before87 and tells us merely that Type II buckles in incomplete and broken state were being discarded as rubbish during the late fourth and fifth centuries. There is need, however, to reconsider when and under what circum- stances the type was first introduced into this country. Most if not all were made in British workshops and, to judge from their variety in quality and style, probably over a

81 My thanks to Mrs. E. Fowler for bringing it to my notice after omission from the list in 1961. It is published in Proc. Soc. Antiq. Scotland xlix (1914-5), 174, fig. 26,3, from Area B, Level 2 (upper).

82 1 am indebted to Dr. R. Reece for permission to publish the buckle and for information about the find context. It came from the garden of Querns Lane House, from a site on the north side of the road dividing Insulae III and IV and including part of that road. Although pottery and coins continued to the end of the fourth century, there was no evidence of any later, post-Roman, occupation. The site was described briefly in Trans. Bristol & Gloucester Archaeol. Soc. lxxv (1956), 203. 83 B. W. Cunliffe (ed.), Fifth Report on the Excavations of the Roman Fort at Richborough, Kent (1 968), pl. xxxv, no. 103. V. I. Evison, op. cit. (note 75), 238, fig. 4,i.

84Giles Clarke, op. cit. (note 78), 295, figs. 4-6; Martin Biddle, in Vor- und Friihformen der europdischen Stadt im Mittelalter (Gottingen, I973), 232 iff-

8sj. Mertens and L. van Impe, Het Laat-Romeins Gravefeld van Oudenburg (Archaeologia Belgica 135, 1971), pls. xvi, xxv,

LU, LXIII, XCII. For a similar knife from Treigne, cf. J. Dasnoy, Annales Soc. arche'ol.

Namur liii (1966), 222 ff., fig. 19. 86 Information via Miss Louise Millard, Canterbury Museum, from Mr. J. Bradshaw of the Ashford

Archaeological Society. The site is in Charity Field (TR 0494461o) and is described as a rubbish-filled hollow, completely sealed and one metre deep, containing 'a tightly packed deposit of charcoal, animal bones, pottery and small domestic items. Coins are plentiful, many barbarous and fragmentary, the identifiable ones ranging from 323 to 378'. Knife and buckle were found at the same depth and 40 cm apart. There is a brief note in Arch. Cantiana lxxxvii (1972), 233.

87Hawkes and Dunning, op. cit. (note 65) (1961), 25 f. 26

This content downloaded from 188.72.96.55 on Tue, 24 Jun 2014 20:15:54 PMAll use subject to JSTOR Terms and Conditions

Page 6: Some Recent Finds of Late Roman Buckles

390 NOTES

longish period of time. Unlike those of Type I, however, their plates can be paralleled abroad and it is likely that the model and inspiration came originally from northern Gaul.88 Several of these openwork plates on the continent come from graves and are known to have been mounted on belts, 4-5 cm broad, which were furnished also with amphora-shaped strap-ends and sets of propeller-shaped stiffeners. It seems probable that some at least of our Type II buckles had formed part of comparable belt-sets. Lack of grave-finds here has so far deprived us of direct evidence of associations, but it should be remembered that both amphora-shaped strap-ends and propeller-shaped stiffeners do occur here as isolated finds, notably in the Roman fort at Richborough.89 On the continent belts with such fittings were worn during the middle decades of the fourth century, and in two cases plates of the type that interest us were buried c. 350. Thus our Type II buckles may have come into use considerably earlier than previously allowed.

Writing in 196I about the recovery of so many of these and other types of late Roman belt-fittings from the towns and rural sites in the civil zone of Roman Britain, I suggested that they must be military gear, and posed and attempted to answer the question: 'Does the whole series reflect some otherwise unknown disposition of troops?' That my ideas of that time are still considered worth adopting is at once gratifying and alarming. Much new information has come to light since, especially from foreign publi- cations, and it is fully time for a fresh appraisal of the whole topic. That is beyond the scope of this note, but there are one or two points that must be made because they affect our understanding of the Type I and II buckles.

First then, the bronze-mounted belts described above, like the later and more

opulently decorated belts with fittings in chip-carving style, are cingula. They were worn as a badge of office in the fourth century and the splendour of the fittings must have varied according to rank. The cingulum was worn primarily by army officers, but since by this period civil servants ranked as soldiers it was not exclusively military.90 Thus our British buckles and belt-fittings, which surely belonged to some form of the cingulum however provincial in style and make and lowly in grade, may have been worn by civil as well as military officials. This might be thought to dispose of the hypothesis that the occurrence of belt-fittings indicates a military presence in town and country during the fourth century. But there are other considerations. Some of our finds of this type come from forts, many come from walled towns of which a number had had their defences modernized during the second half of the century to allow for the installation of ballistae,91 great engines for keeping attackers at a distance which must have needed professionals to operate them effectively. It has been suggested that there may have been city garrisons under the command of the vicarius.92 There is also a strong case, only now being can- vassed seriously, for the existence before 369 in southern Britain of a field army, billeted in towns.93 Stevens has also suggested that these comitatenses, largely barbarian by this

period, may, like later federates, have been given allotments of some proportion of Roman rural estates, especially in areas vulnerable to attack from outside enemies or local brigands. All such forces will have used official equipment of some sort, and at

88Hawkes and Dunning, op. cit. (note 65) (1961), 21 f., fig. 6 and (1962-3), 181, Taf. 60, 2-3, 61;

Dasnoy, op. cit. (note 85), 218 ff., fig. 17; H. Bullinger, Spdtantike Girtelbeschldge (Bruges, 1969), 67, Abb. 18, 56, Taf. ii, xxxi, xxxii, Falttafel A.

89J. P. Bushe-Fox, Second Report on the Excavation of the Roman Fort at Richborough (1928), pl. xxi, nos.

47 and 52, and Fourth Report (I949), pl. xxxvI, nos. 12-3, pl. LIII, 209; B. W. Cunliffe, op. cit. (note 83),

pl. xxxvii, no. 119. 90o A. H. M. Jones, The Later Roman Empire (1964), ii, 566. I am greatly indebted to Dr. John Morris

for information on this subject. 91 P. Corder, Arch. Journ. cxii (1955), 20-42. 92 S. S. Frere, Britannia, a History of Roman Britain (1967), 359. 93 C. E. Stevens, in Roman Frontier Studies 1967 (Tel Aviv, 1971), 35 f.; K. M. Martin, Latomus (1969),

408 ff.; S. S. Frere, in R. M. Butler, op. cit. (note 66), 18.

This content downloaded from 188.72.96.55 on Tue, 24 Jun 2014 20:15:54 PMAll use subject to JSTOR Terms and Conditions

Page 7: Some Recent Finds of Late Roman Buckles

NOTES 391

TYPE I LATE ROMANO-BRITISH BUCKLES

SITES BUCKLE-LOOPS PLATES

-I I\A IB

".

.

....Roman' Anglo-Saxon ) (1 ]

Undated a 0

Tf IPONTIUM ( 4M -WATER NEWTON

CRENCESTER RICHBOROUGH

\#SILCHTER' ?

---- - --2.: .5 .. - . -... - /

...

K( 1

50,

0 "0, .,d

o , Ko

b s'o o r5'oKms.

FIG. 4

This content downloaded from 188.72.96.55 on Tue, 24 Jun 2014 20:15:54 PMAll use subject to JSTOR Terms and Conditions

Page 8: Some Recent Finds of Late Roman Buckles

392 NOTES

SLATE ROMANO-BRITISH BUCKLES

....SITES TYPE IIA 11B

Roman A V .ool?

. . ..... - " . . . . .

: . ...........Anglo-Saxon

A V

- LEICESTEWR A

.C.T WATER NEWTON

?'A

\ST ALBANS

CAo E5 C E CRENCESTE

.. . ....:RICHBOROUGH

S ,AV,

. . - ;oo /so , ms.

FIG. 5

This content downloaded from 188.72.96.55 on Tue, 24 Jun 2014 20:15:54 PMAll use subject to JSTOR Terms and Conditions

Page 9: Some Recent Finds of Late Roman Buckles

NOTES 393 present the belt-fittings constitute the bulk of the evidence. Finally, though the bits and pieces of this equipment from settlement sites are equivocal, we are at last beginning to obtain evidence of a more positive kind from the cemeteries of our late Roman towns. The men buried at Winchester and Gloucester,94 with their status-proclaiming cross- bow brooches, cingula with buckles and strap-ends, and their knives, must, from the very fact that they were interred fully and officially clothed, surely have been pagan bar- barians, probably Germans, and thus soldiers rather than civilians. All in all, one-sided as this argument has been, there remains a case for saying that, while some of our belt- fittings may have been worn by civil officials, many, perhaps the majority, had been worn by the military. The date of the Winchester burials accords well with what Stevens hassaid about its being Constans in 342/3, rather than Count Theodosius, who remodelled the British sector of imperial defence and gave it a comitatus. It is in this context and at this date, I think, that we must envisage the introduction of our Type II buckles.

As to the British buckles of Type I, all previous discussion has assumed that they too had a military function. If this is correct, these small buckles, with their narrow plates and the slender strap-tags that match them in style,95 indicate an insular type of cingulum with a strap-like belt no more than 2 cm broad. But is it correct? The contrast in style and size between these and the apparently contemporary continental belts, up to 9 cm broad, with their heavy fitments in chip-carving style, is very marked indeed. Admittedly few of these latter reached Britain and those that did are confined mostly to its eastern side;96 admittedly there is no evidence that British workshops were making chip-carved belt-sets before the fifth century, and even then only in the south-east;97 admittedly, then, the Type I buckles and tags seem the only candidates for the official cingulum of the more westerly areas in the latter part of the fourth century. But the doubt remains. They are so very slight and fragile that they seem more appropriate to non-official civilian dress. It will not have escaped notice that Type I buckles from Saxon graves were all worn by women, two of whom belonged to groups serving in some military capacity under late Roman or early sub-Roman authorities. This female emphasis may be coincidental, of course. Belts were a feature of Germanic female dress, and these particular women may simply have selected what seemed most attractive in the late Roman repertory. On the continent, after all, there are even cases of women wearing the broad belt with chip-carved fittings.98 Yet, with German elements present in the population from the middle of the fourth century, it is not impossible that there were changes in the civilian dress of late Romano-British men and women.

In conclusion one can only say that many more finds are needed, from well-dated contexts on Roman sites and especially from graves other and earlier than those of Saxon federates' ladies, to clarify the chronological situation, to tell us more about the workshops engaged in the production of these buckles, and, above all, to inform us about the identity and status of their wearers.

35 Beaumont Street, Oxford

94Britannia iv (1973), 309. 9s S. C. Hawkes, op. cit. (note 65) (1972), 148 ff. 96 S. C. Hawkes and G. C. Dunning, op. cit. (note 2) (1961), map fig. 4 and (1962-3), map fig. 5. 97 V. I. Evison, op. cit. (note 75), 231-46, pl. LIII. 98J. Werner, Archaeol. Geographica i (1950), 59. For helpful remarks about the possible use of the

Type I buckles in female dress I am indebted to Dr. J. P. Wild, whose ideas, expressed in a letter, 12 December 1972, coincide very closely with my own though we are neither of us prepared to commit ourselves too positively in the present state of the evidence.

A Christian Monogram from Richborough, Kent. Dr. Kevin Greene writes: In 1968 W. H. C. Frend wrote that 'If (therefore) the impact of Christianity was to be felt anywhere in Britain in the fourth century one would expect it on the Wall and at other military sites, such as Richborough or Caernarvon'.99 Since then, P. D. C. Brown

99 In M. W. Barley and R. P. C. Hanson (eds.), Christianity in Britain, 300-700 (Leicester, 1968), 40.

This content downloaded from 188.72.96.55 on Tue, 24 Jun 2014 20:15:54 PMAll use subject to JSTOR Terms and Conditions