Hostname: page-component-7c8c6479df-995ml Total loading time: 0 Render date: 2024-03-29T09:57:56.747Z Has data issue: false hasContentIssue false

The London Amphitheatre: Excavations 1987–1996

Published online by Cambridge University Press:  09 November 2011

N.C.W. Bateman
Affiliation:
MoLAS, London

Extract

It is now ten years since the first discovery of the Roman amphitheatre in London, and under ordinary circumstances it might be expected that final publication of this important monument would now be imminent. However, what is perhaps not realized is that excavation was still taking place until late in 1996. It has been a massive project, generating more than 20,000 contexts and, although a preliminary assessment has been completed, it will clearly be some time before a final publication is possible. In view of the long time-scale of the project, therefore, it has been suggested that a provisional statement on the amphitheatre should be disseminated, partly, of course, with the intention of generating interest among, or receiving feedback from, other archaeologists and historians. To some extent it is also necessary to establish the limits of the project to counter some of the slightly misleading information and speculation that has already appeared in a number of other sources. It should, however, be remembered that the interpretations offered below are only provisional.

Type
Articles
Information
Britannia , Volume 28 , November 1997 , pp. 51 - 85
Copyright
Copyright © N.C.W. Bateman 1997. Exclusive Licence to Publish: The Society for the Promotion of Roman Studies

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 The site was directed by the author for MoLAS (Museum of London Archaeology Service) and has been generously funded by the Corporation of London. First notice of the discovery was given in Frere, S., ‘Roman Britain in 1987’, Britannia xix (1988).Google Scholar Three provisional summaries have in fact appeared: Bateman, N., ‘The discovery of Londinium's amphitheatre: excavations at the Old Art Gallery site 1987–1990’, London Arch. Vol. 6 No. 9 (1990), 232–41Google Scholar ; Bateman, N., ‘The London amphitheatre’, Current Arch. 137 (1994a), 164–71Google Scholar ; and Bateman, N., ‘Guildhall: beyond the amphitheatre’, London Arch. Vol. 7 No. 10 (1994b), 258–62.Google Scholar Annual excavation summaries have appeared in London Arch. ‘Annual Roundups’ for 1993, 1994, and 1995 (Greenwood, P. and Maloney, C., London Field-work and Publication Round-up, London Archaeologist, Vol. 7 No. 9 (1994)Google Scholar ; Vol. 7 No. 13 (1995); and Vol. 8 Supp. 1 (1996).

2 For example, in his otherwise excellent book Wiedemann suggests that the Guildhall amphitheatre measured 130 m by 110 m and was thus larger than El Djem, Carthage, and Lyon (T. Wiedemann, Emperors and Gladiators (1992), 45). His information is probably derived from D. Perring, Roman London (1991), 63. In fact, as will be apparent below (see pp. 69ff.), it is still unclear how large it was, though there seems very little likelihood of it exceeding 110 m by 88 m. Later, Wiedemann (p. 178) discusses evidence for the cult of Hercules at the London amphitheatre. In fact no such evidence has been identified.

3 e.g. J. Morris, Londinium (1982), 245; and S. Frere, Britannia (1967), 307.

4 W. F. Grimes, The Excavation of Roman and Mediaeval London (1968).

5 Excavations for the west wing of the Guildhall in the late 60s and early 70s, directly over what must have been the western entrance, failed to produce any evidence of an amphitheatre. In those remote days coverage was provided by just one archaeologist who concentrated his efforts in the area by the fort ditches (site directed by Peter Marsden, see Museum, Guildhall, ‘Archaeological finds in the City of London 1965–6’, Trans. London and Middlesex Arch. Soc. xxii (1) (1968), 410).Google Scholar In 1985 a site (GDH 85) was excavated near what is now known to be the outer wall line of the amphitheatre. See Steadman, K., DUA Archive Report: GDH 85 (1986 unpub.).Google Scholar

6 The masonry remains of the amphitheatre have been preserved in situ while the new building was constructed beneath, around, and above them. They will eventually form the centrepiece of a new museum designed around the results of the excavations.

7 In the main areas excavated in 1996, south of the Guildhall, there is only a single basement being constructed, and the conditions of Scheduled Monument consent prohibit excavation below a certain depth. This part of the arena and its covering silts are thus preserved for posterity.

8 See J.-C. Golvin, L'Amphithéâtre romain: Essai sur la théorisation de sa forme et de sesfonctions (1988), 357 and tab. 45. He also points out (p. 411) that most amphitheatres respect not the adjacent street grid but the local topography.

9 British amphitheatres which have clear evidence of an exclusively timber phase followed by partial or complete rebuilding in masonry include Chester (Thompson, F.H., ‘The excavation of the Roman amphitheatre at Chester’, Archaeologia cv (1976), 127239CrossRefGoogle Scholar ); Silchester (Fulford, M., The Silchester Amphitheatre: Excavations of 1979–85, Britannia Monograph 10 (1989))Google Scholar ; and Cirencester (Wacher, J., ‘Cirencester 1963. Fourth interim report’, Antiq. Journ. xliv (1964), 918).CrossRefGoogle Scholar Other timber amphitheatres such as Dorchester (Bradley, R., ‘Maumbury Rings, Dorchester: the excavations of 1908–1913’, Archaeologia cv (1976), 197)CrossRefGoogle Scholar were never rebuilt in masonry.

10 Over a thousand individual timbers, ranging from thin planks to beams c. 300 mm square, have been found and removed during the excavation. Over half of these were in good enough condition to sample for dendrochronology. However, apart from their value for dating, study of the timbers selected by the Romans, the joints used, and the tool-marks visible on them will provide valuable new insights into developments in Roman carpentry and timber management.

11 At Silchester the two uprights representing the gateposts of the south entrance in Timber Phase One were linked by a shallow trench which was presumed to have contained a threshold beam. See Fulford, op. cit. (note 9), 19-21.

12 Compare Silchester in section below pp. 61-2.

13 See below p. 62 for comparisons.

14 Inscription reads: ICLV three times, MIBL and MI[BL]. See Hassall, M.W.C. and Tomlin, R.S.O., ‘Roman Britain in 1994: Inscriptions. The City’, Britannia xxvi (1995), 382.Google Scholar

15 For discussion of the bank and the angle of seating at other amphitheatres in Britain see below p. 62.

16 The dimensions given here do not include a possible railing above the coping stones. Compare Chester where Thompson suggests c. 2.40 m of wall, not including further parapets, though Sunter suggests more still (Thompson, op. cit. (note 9), 233); Fulford reconstructs the height of the Silchester arena wall as 2.75 m without, and 3.75 m with, a parapet on top (Fulford, op. cit. (note 9), 48). Golvin gives an average of 2.63 m for the arena wall without balteus for amphitheatres, 2.72 m for semi-amphitheatres like Paris, and slightly lower, 2.15 m, for converted Eastern theatres. See Golvin, op. cit. (note 8), 316-17.

17 The paradigm for these might be thought to be the well-known amphitheatre depicted on Trajan's Column. See K. Lehmann-Hartleben, Die Trajanssäule (1926), esp. 139 and Taf. 46 for amphitheatre scene. See also below for comparisons, pp. 63-4.

18 Similar coping stones are also known from Chester (Thompson, op. cit. (note 9), 144 and 147); and Silchester (N. Sunter in Fulford, op. cit. (note 9), 173). At Caerleon fifteen were found: see Wheeler, R.E.M. and Wheeler, T.V., ‘The Roman amphitheatre at Caerleon’, Archaeologia lxxviii (1928), 111218, esp. 118–19).CrossRefGoogle Scholar

19 See Golvin, op. cit. (note 8), 314.

20 For comparison with other amphitheatres see below p. 63.

21 Feature [1567]. In the centre of the rectangular iron block was a pronounced depression with a diameter of c. 100 mm and a depth of c. 60 mm. The iron was set in the timber in such a way that the level at the base of the socket was virtually the same as that of the contemporary arena surface.

22 For discussion of British and other amphitheatres see below p. 63.

23 The stones were examined in situ on 27/1/94 by Dr T. Blagg who said that he knew of no British parallels for them; and thought the suggestion reasonable that they were part of a liftable trap (pers. comm.).

24 Provisional identification by Jane Sidell of MoLAS. Very few of the many bones recovered from the site have been identified or analysed yet; the skull of a large bovid, possibly a bull, was found in one of the perimeter drains. None of the faunal assemblages from the other amphitheatres in Britain have anything more exotic than horse, sheep, or cattle. See Fulford, op. cit. (note 9), 137-8; Wheeler and Wheeler, op. cit. (note 18), 214-15. Investigation of this important faunal group will surely be one of the most exciting aspects of future research.

25 For comparison with other British amphitheatres and discussion of carceres see below p. 64.

26 Similar systems have been identified at Caerleon. See G. Boon, Isca. The Roman Legionary Fortress at Caerleon, Mon. (1972), 93.

27 At sites to the north and east canalization may have had negative effects leading to cyclical patterns of erosion, silting-up, and renewed canalization, resulting in further erosion. See C. Maloney, The Upper Walbrook Valley in the Roman Period, CBA Res. Rep. 69 (1990), esp. 120.

28 For comparison with drainage systems in other British amphitheatres see below p. 65.

29 John Humphrey, in discussing the arenas of circuses, mentions the experiences of the makers of the film Ben Hur who had repeated problems with the surfaces used. Originally they had several layers of rock and rock lava. They noted the need for the surface to be ‘hard enough to hold the careening horses and chariots yet soft enough (i.e. with a sanded top) not to lame the horses.… (they) also discovered that the the track required a drainage system in case of rain … thus after a days trial the crushed yellow rock and much of the layer of lava below were removed, leaving only one and a half inches of crushed lava as the new uppermost layer…’. Humphrey also notes that ‘The archaeological evidence from at least two circuses (Caesaria and Carthage) suggests that the track beddings were either deliberately mortared or appeared so through the action of water … one should probably restore above the firm base a fairly thin sandy cover similar to the 1.5 inches of the movie’ (J.H. Humphrey, Roman Circuses (1986), 83). He also noted that, as in Greece, the surface was probably rolled at regular intervals (see also Miller, S.G., ‘Turns and lanes in the ancient stadium’, AJA lxxxiv (1980), 166).Google Scholar This seems to be entirely consistent with the evidence of the Guildhall site. Here the first ‘floor’ or ‘surface’ was of hard mortar (opus signinum) and the second ‘surface’ above it was of c. 5-10 cm of soft sand. There is every reason to think that the criteria for providing a sound surface in an arena where men and animals were to be fighting and charging around would be not dissimilar to those for the surfacing of the circus.

30 See below p. 65.

31 Fulford, op. cit. (note 9), 29, 31, 36 and fig. 82.

32 ibid., fig. 81 and esp. fig. 69.

33 Compare ibid., 27; and N. Sunter in Fulford, 163.

34 Fulford, op. cit. (note 9), figs 5 and 19-21.

35 ibid., 164 and figs 67-8.

36 At Carmarthen timber-revetted seating terraces were let into the natural and redeposited natural which formed the cavea bank (Little, J.H., ‘The Carmarthen amphitheatre’, Carmarthen Antiquary vii (1971), 5863, esp. 59-60).Google Scholar Something similar was identified at Cirencester (Brown, P.D.C. and McWhirr, A., ‘Cirencester, 1966’, Antiq. Journ. xlvii (1967), 185–97).CrossRefGoogle Scholar

37 Fulford, op. cit. (note 9), 13-14; also 163; cf. figs 69 and 70.

38 ibid., 173.

39 See Wheeler and Wheeler, op. cit. (note 18).

40 Boon, op. cit. (note 26), 94. See also Thompson, op. cit. (note 9), 231 and n. 5; and Fulford, op. cit. (note 9), 186.

41 Thompson, op. cit. (note 9), 142.

42 ibid., 236 and fig. 53.

43 Golvin, op. cit. (note 8), 293 table 32. Th e figure of 33.28 degrees is for the podium level. Th e average angle becomes steeper higher in the cavea, reaching c. 37.24 degrees in the summa cavea.

44 This depends on the position of the back of the cavea, which is discussed below p. 72. However note that if the cavea was only 16 m wide the rear elevation would still be 7.50 m.

45 White, G.M., ‘The Chichester amphitheatre: preliminary excavations’, Antiq. Journ. xvi (1936), 149–59;CrossRefGoogle Scholar esp. 151-2,154, and 156.

46 Wheeler and Wheeler, op. cit. (note 18), 118.

47 Thompson, op. cit. (note 9), 144.

48 J. Wacher, Cirencester Roman Amphitheatre, HMSO Leaflet (1981), 3.

49 Floca, O. and Vasiliev, V., ‘Amfiteatrul militar de la Micia’, Sargetia v (1968), 121–52, esp. 151-2.Google Scholar

50 See Golvin, op. cit. (note 8), 318-19.

51 Lehmann-Hartleben, op. cit. (note 17).

52 Compare Thompson, op. cit. (note 9), 163; and Sunter, fig. 53 in Thompson, 236.

53 However at Caenvent there appeared to be a small stretch of outer wall c. 25 ft from the arena wall. The absence of this wall on the north side has been taken to indicate that the amphitheatre was never finished, though alternative suggestions are surely possible. See Ashby, T., Hudd, A.E. and Martin, A.T., ‘Excavations at Caerwent, Monmouth-shire on the site of the Romano-British city of Venta Silurum, in the years 1901–1903’, Archaeologia lix(I) (1904), 105.Google Scholar Also note that there is some doubt about how high the back wall of the Chester amphitheatre was actually taken in masonry (see note 52 above).

54 Fulford, op. cit. (note 9), 13-14.

55 The amphitheatre was quite small, probably with seating for no more than a thousand. It was built under Hadrian and survived till the mid-third century. See Floca and Vasiliev, op. cit. (note 49), esp. figs 2 and 13.

56 See Heidenreich, K., ‘Das Amphiteater der Colonia Traiana bei Xanten, ein Versuch seiner Wiederherstellung’, Bonner Jahrbucher cxlv (1940), 3363Google Scholar , esp. Taf. 6 and 9.

57 See Thompson, op. cit. (note 9), 222–30 and figs 48–9. For timber structure on masonry footings see below note 104.

58 ibid., 170-3.

59 Boon, op. cit. (note 9), 97; also Wacher, op. cit. (note 48). Compare also the possible side-chamber in Silchester Timber Phase Two (Fulford, op. cit. (note 9), 169 and fig. 71), although there was no suggestion of wild animals in this ‘room’.

60 Golvin, op. cit. (note 8), table 41.5, 330.

61 On the reference to the Nemesis inscription: see Thompson, op. cit. (note 9), 166–70. For the close relation of the Nemesis cult to amphitheatres see M. Hornum, Nemesis, The Roman State and the Games, Religions in the Graeco-Roma n World Vol. 117 (1993).

62 Golvin, op. cit. (note 8), 323 and esp. table 38. On the entrance for Libitina see also R. Auget, Cruelty and Civilisation (1972), 55.

63 Wheeler and Wheeler, op. cit. (note 18), 115.

64 Fulford, op. cit. (note 9), 23-5.

65 Heidenreich, op. cit. (note 56), 40-1.

66 Thompson, op. cit. (note 9), 151 and 152.

67 Little, op. cit. (note 36), 59; and Fulford, op. cit. (note 9), 23-5 and 47-9.

68 Wheeler and Wheeler, op. cit. (note 18), 115.

69 White, op. cit. (note 45), 151-2.

70 Bradley, op. cit. (note 9), 38.

71 Little, op. cit. (note 36), 58.

72 Thompson, op. cit. (note 9), 150-1. A very few patches of stone cobbling survived over this sand in places — though whether they represent a complete resurfacing is unclear. Over these was a dark soil c.0.30-0.45 m thick which Thompson interprets as signifying ‘falling out of use’. Overlying this were more patchy remains of stone ‘slabbing’, possibly a cobbled surface, which must have been later than 273. However, it should be noted that the excavation strategy for 1967 involved deliberately machining-off of all but the lowest 2 ft of arena fill. This would mean that any later evidence would not have been seen. Note also that ‘stone slabbing’ is not really the kind of surface one might expect in an arena, especially as there is some evidence of hearths on it! When Thompson reports that there is clear evidence that the amphitheatre was ‘brought back into commission in the final quarter of the third century’ the question is ‘what sort of use’?

73 Fulford, op. cit. (note 9), esp. figs 38, 40, 42 which show an accumulation of up 0.75 m of successive sand and gravel surfaces. Professor Fulford has since confirmed that there was no clear alternation of ‘soft’ and ‘hard’ surfacings, though ‘more lightly metalled surfaces’ sealed periods of silting (pers. comm.).

74 See Bateman, op. cit. (note 1, 1994b).

75 Golvin, op. cit. (note 8), 409.

76 See p. 64 and note 62 above.

77 Steadman, op. cit. (note 5). See also p. 71 below.

78 The Roman masonry building which lies right across the width of King Street due south would have extended across any direct access road from the south. See Sites 45 and 48 in R. Merrifield, ‘Gazetteer’, The Roman City of London (1965).

79 ibid., Site 126.

80 Information on these and all other timbers taken from a provisional assessment by Ian Tyers (then of MoLAS Dendrochronology Lab.) of 29/4/94 (unpub.). These two timbers are Nos [16445] and [16839].

81 Damian Goodburn (MoLAS timber specialist) has pointed out that the timbers used as post-pads were all offcuts deriving probably from the process of in-situ trimming and the final fitting of timbers delivered to the work-site in a state of partial preparation. This is a phenomenon observed on several Roman sites in London. There are three specific reasons why such timbers are unlikely to have been waiting for use for more than a year. Firstly, there were few large splits in the end grain, which commonly occur in large oak in the first few months of drying after felling. Secondly, the sapwood was in good condition with no traces of decay or borer attack. Thirdly, large structural woodwork generally employed green timber in the Roman period.

82 Tyers, op. cit. (note 80).

83 In fact one as of Nero (64-6) has been identified so far from what appeared to be stream gravels [18152] preceding the construction of the amphitheatre. It should be noted that the site has been particularly rich in coins. Over 1,300 coins of all dates have been found, the bulk of them being from the Roman part of the sequence. The quantity of coins is directly attributable to the controlled use of on-site metal-detecting as a recovery technique. The very size of the coin collection has meant that it has so far only been possible to date coins from key contexts.

84 For instance, small groups of pottery found with the first surface [12861 et al.] and a constructional make-up [18018] of the rebuilt amphitheatre are dated after 120.

85 I am grateful for this and many other provisional pottery dates to Jo Groves of the Museum of London.

86 This represents the largest single deposit of Roman glass ever found in the North-Western provinces of the Empire. The next largest is from Saintes and similar dumps are known from Nijmegen and Lausanne. The glass at Guildhall is all cullet: nearly 100,000 fragments of up to a thousand individual vessels, as well as glass-blowing waste and window-glass. All of this was collected by the Romans for a re-working which never happened. I am indebted for all of the information about the glass deposit to John Shepherd of the Museum of London. It is well known that there was glass-working in the upper Walbrook valley nearby (See Maloney, op. cit. (note 27), 124; and J.D. Shepherd and M. Heyworth, ‘Le travail d u verre dans Londres romain (Londinium): un état de la question’, in D. Foy and G. Sennequier (eds), Ateliers de verhers de l'antiquité à la période pré-industrielle (1991), 13-22).

87 Ian Betts, MoLAS building material specialist (pers. comm.) See also p. 78.

88 The dendrochronological date is derived from two separate piles, [16079] and [16080], of the drain recut extending to the north. These are likely to be the latest of all such dates (Tyers, op. cit. (note 80)).

89 Five coins of 270–280 have been identified. Provisional identification of two coins of the House of Constantine (330–341) from the same layer [12033/12197] imply that it might be even later, bu t further work is necessary before this can be accepted.

90 It is worth noting that recent work inside the fort at the Shelley House site suggests that it may have been occupied by the military as late as c. 250 (pers. comm. D. Lakin, site supervisor). That the amphitheatre might, at an even later date, have been used as a market is, of course, still a possibility. It has been suggested that the amphitheatre at Cirencester, for example, could well have served as an extra-mural fairground or market-place in the fourth century. See Wacher, op. cit. (note 48).

91 Construction fill [22338]. Robin Symonds of MoLAS (pers. comm.) has commented that it is not currently possible to be more specific about the date of this.

92 From Fills [11825] and [11867]. Several mid-fourth-century coins of the House of Constantine; and two coins of the House of Valentinian 367–80.

93 For brief summaries see Bateman 1994a, op. cit. (note 1).

94 For example, though the arena wall at Caerleon seems to have been carefully laid out, the outer perimeter was not. The width of the bank supporting the seating varies form 12.3 to 13.6 m. On the evidence of inscriptions it would appear that gangs from at least seven centuries were involved in construction at Caerleon. There is also evidence from elsewhere that working-parties from different centuries may have been organised to work in competition rather than co-operation. See Evans, E., ‘Military architects and building design in Roman Britain’, Britannia xxv (1994), 143–64, esp. 152.CrossRefGoogle Scholar

95 See J.-C. Golvin, ‘Origine, fonction et forme de l'amphithéâtre romain’, in C. Domergue, C. Landes and J.-M. Pailler, Spectacula. I: Gladiateurs et amphithéâtres. Actes du colloque tenu à Toulouse et à Lattes le 26, 27, 28 et 29 mai 1987 (1990), 15-27.

96 In fact these curves have been created using an AutoCad facility and are thus true ellipses. As stated above, the real curvature of the arena wall would have been marginally different.

97 Marek Ziebart (MoLAS Principal Surveyor) writes: ‘Since there are mathematical models for the original setting out and construction of the amphitheatre walls, these can be tested by statistical analysis of the remains found and located so far. This would enable a number of models to be rejected, modified or confirmed, and hence the predictive modelling of the amphitheatre dimensions could be assigned statistical weighting and be used, with some degree of certainty, to delineate problem areas. The model would be tested using least squares analysis, which forms the basis of standard mathematical curve fitting. The work would be carried out using a commercially available software package such as Mathcad or Unistat’.

98 P. Marsden, Notes on City Excavations 1949–55 (1968, unpub.); see also Merrifield, op. cit. (note 78), Site 129. The location and alignment of this feature is only roughly fixed in relation to one of the buttresses of the mediaeval Guildhall overhead, a point which needs to be remembered in all subsequent discussion. The wall was left in situ.

99 To suggest that the east–west element was part of another wall lying behind the arena wall itself would leave only a very small space between them, not much more than a metre. This would be too small for a chamber and there is no evidence elsewhere for an internal corridor running behind the arena wall.

100 Note the western recess of the Silchester amphitheatre which, either by choice or bad design, lies to the north of the short axis. Fulford, op. cit. (note 9), figs 20 and 28.

101 See p. 66 above.

102 Note that the north return identified at the west end of this wall cannot be simply interpreted as the side of another entrance-way. A test-pit excavated some way to the north (see FIG. 3) has proved that the brickearth bank continued across this area. The wall may represent another chamber in the external body of the amphitheatre.

103 There are many amphitheatres which are constructed very differently on either side of the long axis. A good example is the early second-century amphitheatre at Trier, where the amphitheatre was built into a steep hillside and only the lower south side had a full masonry elevation, including a monumental entrance for the public (P. von Zabern, Trier: Kaiserresiden: undBischofssitz (1984), fig. 66 et al.). Although the amphitheatre in London was built on to a gentle hill rising to the north and west, the use made of a previous water-course meant that the southern cavea was actually founded higher than the eastern entrance.

104 Such a construction would seem to be implied at the theatre at Frejus for example. Here, a curious polygonal ring of small shallow foundations lay beyond the ends of the massive radial walls within and under the cavea. Modern interpretation now sees these as the footings for a timber gallery forming the outer ring of the cavea. See A. Grenier, Manual d'archéologie gatto-romaine 3.2: architecture: ludi et circenses (1958), 739-41. Grenier does acknowledge that such a construction was unusual.

105 Merrifield, op. cit. (note 78), Sites 130, 131 and 132; and Archaeologia lxiii, 314-15. Recordings of black mud’, ‘marshy hollows’, etc. in building work observations of 1671 (Wren, St Lawrence Jewry), 1908, and 1911. Merrifield suggests these represent a stream-bed which was certainly open in the Roman period and that it determined the otherwise anomalous NW–SE alignments of at least two Roman masonry buildings identified beside King Street to the south. One of these buildings had a tessellated pavement and may date from the early third century (ibid., Sites 132 and 151). This would suggest that the stream or drain was still open at that date.

106 The road is cited as Road 7 in Perring, D. and Roskams, S. with Allen, P., Early Development West of the Walbrook, CBA Res. Rep. 70 (1991), 114.Google Scholar See also Shepherd, J., ‘The pre-urban and Roman topography in the King Street and Cheapside areas of the City of London’, Trans. London and Middlesex Arch. Soc. xxxviii (1987), 1157, esp. 22.Google Scholar It was interpreted as an early-to-mid-second-century extension of the Flavian grid to the south.

107 I am indebted to Dave Bentley of MoLAS for this new information which has come about as a result of his work on establishing an Autocad database for the Archive of old DUA and MoLAS sites. He points out that the alignments of these roads should not be treated as at all certain where they extend much beyond the junction with the Cheapside road.

108 See for instance Golvin, op. cit. (note 95), fig. 5, which shows the variation between a true ellipse and an arena constructed of four true arcs. See comment in note 97 above regarding future proposed methodology.

109 It should be pointed out that even if the wall does represent part of the amphitheatre it is just as likely to have functioned as a footing for a timber superstructure as to have continued vertically as a masonry wall.

110 The ratio of 15:21 is about 70 per cent. Although the overhead structure at Guildhall would appear to have been constructed entirely in timber, a similar effect is achieved on the banks of amphitheatres such as Silchester. Both the Timber 2 and Stone 1 phases of the amphitheatre are reconstructed with timber seating over the front c. 70 per cent of the bank rising at a shallow angle, while the outer part of the bank falls at something approaching an angle of rest (Fulford, op. cit. (note 9), figs 70 and 73).

111 Note that although Golvin does not provide this figure, the average width of cavea for the 94 structures pleines amphitheatres in his tables 28 and 29 is c. 17 m, Golvin, op. cit. (note 8).

112 In discussing ways of calculating the number of spectators from available space, Bomgardner agrees with Golvin's estimation of space used for a single spectator as 0.28 m2 and that an allowance of c. 10 per cent of the total cavea space should be excluded from calculations for corridors, entrances etc. He disagrees however with Golvin's resultant formula of [2.5 × (cavea total surface area) = probable seating estimate], and suggests instead 2.25 as the constant. See Bomgardner, D.L., ‘A new era for amphitheatre studies’, JRA vi (1993), 375–90, esp. 386.Google Scholar

113 Golvin, op. cit. (note 8), 291 and n. 12. He actually provides an average figure of 1,450 m2 for his structures pleines types a and b, and 1,862 m2 for types c and d. From these he derives an average of 1,700 m2 for all non-monumental amphitheatres. He also points out (pp. 280-30) that this average figure for the arenas of semi-amphitheatres is very similar: it is the cavea size which differentiates the two types.

114 T. Darvill and C. Gerrard, Cirencester: Town and Landscape (1994), 79. Golvin expresses it differently: he suggests that the average area of the arena of all structures pleines was c. 35.5 per cent of the total area. See Golvin, op. cit. (note 8), 291 and n. 14. This translates to a factor of 1.82 for the relation of arena to cavea.

115 Thus the ratio of the 15 m small cavea (3035 m2) to the small arena (1886 m2) = 1.61; of the 21 m small cavea (4068 m2) to the small arena (1886 m2) = 2.16; of the 15m large cavea (3242 m2) to the large arena (2193 m2) = 1.48; of the 21 m large cavea (4943 m2) to the large arena (2193 m2) = 2.25.

116 Fulford, op. cit. (note 9), 177. As Bomgardner points out, in spite of the start made by Fulford and Golvin's recent work, there is still no satisfactory survey of the British amphitheatres (Bomgardner, op. cit. (note 112), 379).

117 Caerleon: Wheeler and Wheeler, op. cit. (note 18) and Boon, op. cit. (note 26). Chester: Thompson, op. cit. (note 9). Carmarthen: Little, op. cit. (note 36). Chichester: White, op. cit. (note 45). Cirencester: Brown and McWhirr, op. cit. (note 36); Wacher, op. cit. (note 48); and T. Darvill and C. Gerrard, Cirencester: Town and Landscape (1994). Dorchester: Bradley, op. cit. (note 9). Silchester: Fulford, op. cit. (note 9).

118 Tomen-y-Mur: Gresham, C.A., ‘The Roman fort at Tomen-y-Mur’, Arch. Camb. xciii (1938), 192211.Google ScholarRichborough: Cunliffe, B. (ed.), Fifth Report on the Excavation of the Roman Fort at Richborough, Kent, Report Res. Comm. Soc. Antiq. XXIII (1968).Google Scholar Caistor St Edmund: Maxwell, G.S. and Wilson, D.R., ‘Air reconaissance in Roman Britain 1977–84: III. Civilian sites’, Britannia xviii (1987), 148.CrossRefGoogle Scholar Frilford: Maxwell and Wilson, ibid.; and Hingley, R., ‘Recent discoveries of the Roman period at the Noah's Ark Inn, Frilford, South Oxfordshire’, Britannia xiii (1982), 305–9.CrossRefGoogle Scholar

119 Winterslow: Vatcher, F. de M., ‘The excavation of the Roman earthwork at Winterslow, Wilts’, Antiq. Journ. xliii (1963), 197213.CrossRefGoogle Scholar Charterhouse-on-Mendip: the report on Charterhouse specifically mentions that there was no sign of a ‘smoothed floor’ over the exposed natural. See Gray, H. St George, ‘Excavations at the “Amphitheatre”, Charterhouse-On-Mendip, 1909’, Proc. Somerset Arch, and Natur. Hist. Soc. lv (1910), 118–37, esp. 128.Google Scholar An alleged amphitheatre close to and similar to the Charterhouse structure was found just outside the Romano-British village of Woodcutts excavated by Pitt-Rivers in 1884-5 (ibid., 135).

120 References t o it appear in R.G. Collingwood and I. Richmond, The Archaeology of Roman Britain (1969), 118 and fig. 42e, but the ‘amphitheatre’ is not discussed anywhere else!

121 The structure was oval in shape though poorly set out, and only had one entrance, which would be unusual for an amphitheatre. The indications are that it was a late building and, moreover, was never finished (See Ashby et al., op. cit. (note 53), 87-124). Boon suggests that the amphitheatre may have been similar to the military amphitheatre at Micia in Dacia (Boon, op. cit. (note 26), 136 and n. 330) (see note 49 above for Micia). Caerwent was of course a civitas capital.

122 Newstead: a very clear elliptical shape revealed by aerial photography. See Keppie, L.J.F., ‘Roman Britain in 1993: Scotland; Newstead’, Britannia xxv (1994), 261.Google Scholar See also ‘Amphitheatre unearthed’ report in the The Times 21 October 1996. Catterick: see The Times, 9 Sept. 1995, and Moloney, C., ‘Catterick Race Course’, Current Arch. 148 (1996), 128–48.Google Scholar Lying just outside Roman Cataractonium, the banked oval enclosure measured c. 140 m along its longer axis, and the bank itself was some 40 m wide. Among other factors, however, the lack of any obvious entrance ways into the enclosure may cast some doubt on its interpretation as an amphitheatre.

123 There appear to be no real parallels for this structure. A sunken ‘arena’ was surrounded by a timber shuttering wall possibly as high as c. 2.40 m. The only reason for identifying it as a gyrus seems to have been that it was flat, circular, and enclosed. The term ‘gyrus’ for cavalry training-rings comes from Xenophon, who specifically mentions the need for soft sand, but no real surfaces were identified in the ‘arena’ of the Lunt at all. There would seem to be little reason to sink a gyrus into the ground. See Hobley, B., ‘Excavations at ‘The Lunt’ Roman military site, Baginton, Warwickshire 1968–71. Second interim report’, Trans. Birmingham and Warwickshire Arch. Soc. lxxxv (1973), 792.Google Scholar

124 Although Carmarthen is certainly near a fort, and is included by Le Roux among others in lists of amphitheatres originating with forts (see P. Le Roux, ‘L'amphitheatre et le soldat sous l'Empire romain’, in Domergue et al., op. cit. (note 95), 203–15, esp. 203), dating evidence seems to link the construction of the amphitheatre with the period when the fort was abandoned and the town became a civitas capital. See J.S. Wacher, The Towns of Roman Britain (1995), 392–3.

125 The existence of an amphitheatre at York has been suggested by the discovery in a Roman coffin of a fragment of bone which carries the inscription ‘domine victor vincasfelix’ (Lord Victor may you have a lucky win). The site of the amphitheatre just to the east of the fortress may be suggested by peculiarities of the later street plan. See P. Ottaway, Roman York (1993), 33-4 and fig. 1.

126 Wacher, op. cit. (note 124), 56 and 250.

127 James, H., ‘Excavations in Roman Carmarthen’, The Carmarthen Antiquary xxviii (1992), 536, esp. 13.Google Scholar

128 Thus Dr Bersu suggests in a brief appendix that a bank across the arena might imply conversion to a theatre. See Gresham, op. cit. (note 118).

129 Thus, even after the construction of Rome's first stone amphitheatre by Statilius Taurus, emperors from Augustus to Caligula continued to give gladiatorial combats in the Saepta (Cassius Dio Lix.10.1–5). Venationes were as frequently given in the Circus Maximus and at other sites in Rome (SHA Had. 19.2–9; Cassius Dio Liv.34.1–2). In the East, at Corinth for instance, Roman, as opposed to Greek, games took place in the amphitheatre, the theatre, and the odeon. In Ephesus gladiator games were held in the theatre; in Daphne at the stadium (Hornum, op. cit. (note 61), 53). There are many other examples.

130 Golvin carefully distinguishes ‘semi-amphitheatres’ like Lutece from ‘theatre-amphitheatres’ like Sanxay. At one level he sees them as two forms of ‘economy measure’ for communities which could not afford both a theatre and an amphitheatre. At another level he links both types to Gaulish rural cult sanctuaries and says that there is little to distinguish one type from the other in purely functional terms. He notes the smallness of the stage in both types, and suggests that something specifically different from normal theatrical shows must have been performed here, possibly cultic and possibly simultaneously with the munera in the arena. See Golvin, in Domergue, op. cit. (note 95), 15-27; also Golvin, op. cit. (note 8), 225-36. Hönle and Henze see two distinct traditions in the Gaulish structures (Mehrzwecktheater): one represented by buildings such as Lutèce, Senlis, Grand, and Lillebonne; and another by buildings such as Sanxay and Drevant (A. Hönle and A. Henze, Römische Amphitheater und Stadien (1981), 172-5).

131 Canterbury: Frere, S., ‘The Roman theatre at Canterbury’, Britannia i (1970), 83113CrossRefGoogle Scholar ; Verulamium: Frere, S., Verulamium Excavations: Vol. II, Report Res. Comm. Soc. Antiq. XLI (1983)Google Scholar ; Colchester: see Crummy, P., ‘The Roman theatre at Colchester’, Britannia xiii (1982), 299302CrossRefGoogle Scholar ; Gosbecks: see Dunnett, R., ‘The excavation of the Roman theatre at Gosbecks’, Britannia ii (1971), 2747.CrossRefGoogle Scholar Note should also be made of yet another possible theatre at Cirencester. Observations over the years of two concentric walls might represent a theatre very similar in size and shape to those of Verulamium and Canterbury. See Darvill and Gerrard, op. cit. (note 117), 62.

132 One of the few examples of civic euergetism known from Britannia, the inscription records the dedication of a new stage building by a local magistrate under Antoninus Pius (RIB 707). Recently however, the identification of Brough with Petuaria has been questioned. Brough, it seems, was more probably a military or naval base, and the civitas capital must be sought elsewhere. Wacher is undoubtedly correct when he points out that the existence of a theatre could suggest a rural sanctuary complex rather than a town (op. cit. (note 124), 394-7).

133 The building lasted from ‘before 100’ to the early third century. The theatre was then rebuilt ‘on a classical plan’ around 210-20; see Frere, op. cit. (note 131). Some have seen the transformation more simply as a ‘theatre’ replacing an ‘amphitheatre’ (D. Mackreth, ‘Roman public buildings’, in J. Schofield and R. Leech (eds), Urban Archaeology in Britain, CBA Res. Rep. 61 (1987), 133-46, esp. 137).

134 The theatre was probably not built before 90 and more likely in Hadrianic times. See Frere, op. cit. (note 131). The temple complex to the north, on the other side of the valley, survived until the third century and succeeded the late pre-Roman Iron Age ‘royal’ burial ground at the Folly Lane site (R. Niblett pers. comm.). Liversidge believed the theatre had some cultic function attached to the temple and pointed out that such festivals often included gladiatorial games. See J. Liversidge, Britain in the Roman Empire (1968), 369.

135 Wacher, op. cit. (note 124), 229.

136 Golvin, op. cit. (note 8), 232-4. He suggests that they probably did show munera and venationes in this class of building, though the shows must have been smaller and less important than those in the so-called ‘semi-amphitheatres’ or ‘proper’ amphitheatres. His explanation of this type of building as merely an extreme economic response to local needs is surely insufficient, and does not do justice to the wealth and status of Verulamium. Later (p. 236), however, he appears to contradict himself by implying that the reason Verulamium had a theatre-amphitheatre and not an amphitheatre, was that, unlike towns like Silchester, Dorchester, Caerwent, and Chichester, it was not a ‘new town’ but an important, and presumably pre-existing, cult or ‘sanctuary’ centre. This too appears deficient as an explanation.

137 M. Millett, The Romanization of Britain (1990), 87-8. That gladiatorial displays took place either at this theatre, or perhaps at an as yet undiscovered amphitheatre in Colchester, is perhaps suggested by the discovery of a gladiator's helmet at Hawkeden, not far from Colchester. See Painter, K.S., ‘A Roman bronze helmet from Hawkeden, Suffolk’, British Museum Quarterly xxxiii (1969), 121–9.CrossRefGoogle Scholar

138 The earliest timber phase, of possibly Hadrianic date, was a free-standing structure somewhat similar to the early structure at Chester. The later rebuild, probably in the mid-to late second century, was, interestingly, exactly the opposite of the London amphitheatre: that is to say that the cavea bank was revetted on the outside in masonry while the arena/orchestra was surrounded by a rather fragile timber structure. See Dunnett, op. cit. (note 131).

139 The Times 9 Sept. 1995. The structure appeared to be very late, probably after 337. Evidence for timber seating set over and into the natural curve of a valley head suggested an auditorium c. 30 m wide. There is evidence of continuity into the post-Roman period.

140 For the best discussion of euergetism see P. Veyne, Bread and Circuses (trans. 1990 from Le pain et le cirque 1976).

141 Tacitus, Agricola 21. Millett, op. cit. (note 137), 69, notes that this is in fact the only passage in ancient literature which appears to substantiate a ‘policy’ of urbanization, and as such that it is frequently given too much weight.

142 Blagg, T., ‘Architectural munificence in Britain: the evidence of inscriptions’, Britannia xxi (1990), 1331.CrossRefGoogle Scholar

143 Millett, op. cit. (note 137), esp. 82, 91, and 107. London and the coloniae were of course different in genesis and development to the civitas centres.

144 Betts, I.M., ‘Procuratorial tile stamps from London’, Britannia xxvi (1995), 207–29.CrossRefGoogle Scholar

145 Grimes, op. cit. (note 4), 17-40.

146 Perring, D. and Roskams, S. with Allen, P., Early Development West of the Walbrook, CBA Res. Rep. 70 (1991). 114.Google Scholar

147 Millett, op. cit. (note 137), 88.

148 Perring, op. cit. (note 2), 38–40. Interestingly he suggests that the earliest possible date for a fort on this site would seem to be provided by a coin of 71 found in a pit beneath the bank by the fort wall.

149 The road to the gate: Blair, I., ‘The finding of the Foster Lane glass’, Popular Archaeology v, no. 4 (1983), 23.Google Scholar Within the fort: a provisional assessment of work at Shelley House (Feb. to June 1996) indicates a phase of activity including kilns and burnt debris which was replaced by a phase with several Flavian timber buildings. These were rebuilt at least once before they were demolished to make way for masonry barrack buildings after c. 120. There was nothing specifically military about either the timber buildings or the finds associated with them, but then the same might be said of the later stone buildings and their associated finds (pers. comm., D. Lakin, site supervisor).

150 Millett, M., ‘Evaluating Roman London’, Arch. Journ. cli (1994), 427–34, esp. 434.Google Scholar

151 Millett, ibid., 433. Morris (op. cit. (note 3), 104) and Millett (ibid., 430) suggest that the eastern hill of Londinium was the location of a conventus civium romanorum, and the western hill a military enclave. The lack of earthen defences around London, and the relatively late construction of masonry walls around the city, have been seen as ‘a consequence of the city being a Roman community which had developed separately from the indigenous system’ (Millett, op. cit. (note 137), 140).

152 Frere suggests that the existence of numerous legionary gravestones indicates that London was the army headquarters but adds that the purpose of the fort may partly have been to accommodate an urban cohort (Frere, op. cit. (note 3), 125 and 196). Salway however disgrees with the idea of an urban cohort in London (P. Salway, Roman Britain (1981), 162). Estimates of the number of men housed in the stone fort vary. Morris suggests up to 1,500 (op. cit. (note 3), 166); Perring suggests 1,500-2,000 (op. cit. (note 2), 40). Compare this with figures of more than 6,000 provisionally suggested for the amphitheatre audience on p. 73 above.

153 A previous suggestion to that effect was based upon an erroneous assumption that the amphitheatre had a masonry outer wall. L. Ebbatson drew attention to the fact that none of the ‘civil’ amphitheatres in Britain have an outer masonry wall, whereas both the ‘military’ ones do (‘The London amphitheatre’, London Arch, vi (1) (1988), 22). Although the premise was incorrect, certainly as regards her assumption that London's amphitheatre had a continuous masonry outer wall, her conclusion that the amphitheatre was associated with the fort may still be partially valid.

154 Collingwood and Richmon d confuse the military amphitheatre with the ludus. Discussing certain British amphitheatres, they describe a ludus as ‘a structure closely resembling an amphitheatre. It differs from an amphitheatre not only in purpose but in proportions, for its arena is much larger than that of an amphitheatre in relation to the surrounding seating. It takes an amphitheatre's form because military weapon-instructors were originally gladiatorial trainers’ (Collingwood and Richmond, op. cit. (note 120), 117). Golvin, however, lists only 14 places where there were true ludi: i.e. gladiatorial training arenas and barracks: Capua, Pompeii, Cyzicus, Rome, Alexandria, Nikopolis, Praeneste, Pola, Verona, Lambaesis, Gemellae, Aquincum, Salone, Pergamum. He also notes that in most of them the arenas were very small, being no more than c. 200 m2 in area (Golvin, op. cit. (note 8), 152-3). That the military amphitheatre and the ludus were quite distinct seems to be confirmed by the fact that, for example, at Gemellae (Numidia) a ludus has been identified within the camp, while a slightly later and much larger amphitheatre has been identified outside the camp (Golvin, op. cit. (note 8), 155). It is surely possible that the so-called gyms within the camp at Baginton (see p. 75 and note 123 above) was actually a ludus!

155 ‘L'existence d'une architecture propre aux arènes militaires est loin d'être prouvée’ (Le Roux, op. cit. (note 124), 206).

156 Golvin, op. cit. (note 8), 154-6; Bomgardner, op. cit. (note 112), 381.

157 Golvin, op. cit. (note 8), 156 and note 461. He was not the first to do so. In discussion of the Caerleon amphitheatre Boon stated: ‘It has also been claimed that the military structure (termed ludus) is distinguished by plan from its civil counterpart, in that its arena is larger in relation to the seating. … This is questionable. The arena must always have been calculated with regard to the type and extent of the proceedings envisaged: but the seating is likely to have reflected very narrowly the potential total of spectators, of whatever kind’ (Boon, op. cit. (note 26), 99). In spite of this Bomgardner still has his doubts and suggests it may be unnecessarily pessimistic to believe there is no correlation with arena size (Bomgardner, op. cit. (note 112), note 45).

158 Bomgardner, op. cit. (note 112), fig. 5 and 388ff.

159 ibid., 382.

160 ‘Le soldat fut assurément un bâtisseur d'amphithéâtre’ (Le Roux, op. cit. (note 124), 205). At the Caerleon amphitheatre, for instance, plaques record the involvement of cohorts and centuries of legio II in its construction (Wheeler and Wheeler, op. cit. (note 18), 114). At Mesarfelta an inscription records the repair of the amphitheatre c. 180 by a cohort of legio VI Commagenorum. At Dura Europos an inscription records building of the amphitheatre in c. 216 by vexillations of IV Seythia and III Cyrenaica (Golvin (1988), op. cit. (note 8), 155).

161 Le Roux, op. cit. (note 124), 205-6: ‘il me semble peu probable qu'on ait soustrait l'armée a ses tâches impériales pour la mettre au service de communautes responsables de leurs programmes d'urbanisme et de leur gestion’. Note also Millett, following Blagg, suggests that ‘the use of military craftsmen, architects, surveyors and workmen on civil projects is difficult to substantiate’ (op. cit. (note 137), 72).

162 ‘Amphithéâtres établis en liaison étroite avec des camps et utilisés par les soldats’, Le Roux, op. cit. (note 124), 205. A list of amphitheatres that follow these criteria is surprisingly short. It includes Aquincum, Caerleon, Carmarthen, Carnuntum, Charterhouse-on-Mendip, Chester, Dura-Europos (Syria), Gemellae (Afr.), Lambaesis (Afr.), Mesarfelta (Afr.), Micia (Dacia), Noviomagus (Germ.), Porolissum (Dacia), Thevestis (Afr.), Tomen-y-Mur, Vetera-Birten (Germ.), and Vindonissa (Le Roux, op. cit. (note 124), 203-5 and note 26). All of which, with the exception of Tomen-y-Mur and Carmarthen, were of legionary not auxiliary units. Xanten should possibly be included because it was used by the soldiers of the nearby camp of Vetera. With Augst there is no proof that it was built by soldiers of the I Adiutrix and VII Gemina (ibid., 205).

163 Le Roux, op. cit. (note 124), 205. J. Liversidge, for instance, described Chester as a ‘military amphitheatre’, but then also went on to say that it probably had ‘room for both the Twentieth legion and the civilian population’ (op. cit. (note 134), 377).

164 Golvin, op. cit. (note 8), 156. Note that an inscription (CIL XIII.11253) found in the military amphitheatre at Carnuntum reserves seats for civil magistrates. It appears they occupied a tribunal directly opposite that of the legionary legate. See Kolendo, J., ‘La répartition des places aux spectacles et la stratification sociale dans l'Empire Romain’, Ktema vi (1981), 301–15, esp. 312.Google Scholar

165 Le Roux, op. cit. (note 124), 206. Thompson identifies the mistaken origin of the concept of amphitheatres as training ludi in Richmond's analysis of the Trajan's Column amphitheatres (op. cit. (note 9), 142-3).

166 Boon, op. cit. (note 26), 99.

167 Maurin, J., ‘Les Barbares aux arènes’, Ktema ix (1984), 103–11, esp. 105-6.Google Scholar Maurin suggested that the munera comprised not just a sequence of random individual combats but structured symbolic contests between barbarian ‘types’, which were defined by their divergence from an idealised Roman norm. The underlying basis of all gladiatorial combat was the opposition of armaments, of techniques, of defence and attack, of light and heavy armour. The two extremes, light and heavy, were polarized around the normative figure of the citizen legionary.

168 Especially in provincial capitals such as Lyon, Merida, and Tarragona. On the relation of provincial organizations to emperor worship see S.R.F. Price, Rituals and Power: the Roman Imperial Cult in Asia Minor (1984). On the basis of an inscription recorded by Roach Smith (but now lost) Merrifield suggested that London may well have become the provincial capital and the centre of the Imperial cult. The text reads ‘Num(ini) C[aes(aris) Aug(usti) ] Prov[incia] Brita[nnia]’ (R. Merrifield, London City of the Romans (1983), 83). Wilkes, however, casts doubt on this reading and interpretation (Wilkes, J.J., ‘Review’ in Britannia xii (1981), 412–16).CrossRefGoogle Scholar

169 Le Roux. op. cit. (note 124), 207. Webster said much the same thing twenty years ago: ‘What is also forgotten is that these shows were organised for special festivals and that they were preceded by ceremonies and sacrifices. It is more than likely that the legionary amphitheatres were primarily for this purpose’ (G. Webster, The Roman Imperial Army (1969), 201). Campbell points out that the army ‘formed the largest and most coherent single group of those expressing devotion to the emperor in this way’. The Feriale Duranum found at Dura-Europos associates 27 out of the 41 listed military festivals with the imperial cult (J.B. Campbell, The Emperor and the Roman Army (1984), 100 and note 68).

170 Interestingly, a earcerarius legionis — legionary prison-keeper — was granted a special seat at the civil amphitheatre of Aquincum, probably in recognition of his importance in maintaining a supply of prisoners and condemned men. See Kolendo, op. cit. (note 164), 313-14.

171 See Le Roux, op. cit. (note 124), 210.

172 Tacitus, Agricola 16.

173 Tacitus, Agricola 17. However, an interesting question also arises as to whether the date of planning the amphitheatre was significantly different from the date when construction started.

174 Cassius Dio LXV.15.2.

175 On Vespasian's use of the amphitheatre as a political tool see Hönle and Henze, op. cit. (note 130), 55. On the astute political development of the munera generally by all the Flavians see M. Clavel-Lévèque, L'empire en jeux (1984), esp. 167.

176 Tacitus, Ann. III.55.

177 See Wiedemann, op. cit. (note 2). 42.

178 Grenier, for instance, notes the importance of the Flavian period for amphitheatre building in Gaul (op. cit. (note 104), 691 et al.). On the Flavians need for continued military backing and legitimation note that Vespasian, Titus, and Domitian received more acclamations as imperator than any emperor other than Claudius, see Campbell, op. cit. (note 169), 124.

179 Salway, op. cit. (note 152), 157.

180 SHA Had. II.

181 Frere, op. cit. (note 3), 126. Other possible governors of Britain at the time include M. Appius Bradua — described only as holding the office under Hadrian, but who may have been appointed in the last year of Trajan, from 115-118, or later from 128-31.

182 Cassius Dio LXVIII.15.1; SHA Had. 14.11.

183 Cassius Dio LXIX.9.1-6.

184 SHA Had. 19.2-9.

185 Cassius Dio LXIX. 10.1.

186 SHA Had. 10.

187 Possibly the greatest recommendation of the munera as far as Cicero was concerned was its promotion of ancient Roman ‘virtus’ and manliness, and as a schooling against pain and death (Cicero, Tusc. 11.7.41). Pliny said that free citizens were made braver by the example, inspired to face wounds and look scornfully on death (Pliny, Panagyhc. 33.1).

188 Fishwick, D., ‘The Temple of the Three Gauls’, JRS lxii (1972), 4652.Google Scholar

189 The Chester timber amphitheatre was probably associated with legio II Adiutrix which was in Chester from the late 70s to c. 86. The masonry amphitheatre was associated with legio XX which arrived in 86/7. Thompson suggests a date of c. 90–120 (Thompson, op. cit. (note 9), 164 and 182).

190 Bradley sees the Dorchester amphitheatre as built by the military in the early Flavian period and falling out of use when they departed in the second century. (Bradley, op. cit. (note 9), 74-8). Fulford however suggests the possibility of a mid-second-century date for its construction (op. cit. (note 9), 192).

191 The Wheelers suggest c. 78–90 for the construction of the first amphitheatre, probably nearer the beginning. This was unusually early for a masonry amphitheatre but was seen by them as the first fruit of the consolidation of the western frontier under Agricola. They also conclude that the middle phase of reconstruction should be dated to c. 125, possibly associated with the return of the legion after 122 from building Hadrian's Wall. Boon disagrees with this date, suggesting that it is based on one worn coin of Hadrian, but in fact, there is more than one coin and two of them are specifically mentioned as being in mint condition (Wheeler and Wheeler, op. cit. (note 18), 148-9 and 152-4). The final period of renovation starts in 212-222. There is no evidence for activity later than 296 (ibid., 151).

192 Severus was well known for his enthusiasm for rebuilding public buildings (See SHA Sev. 23; and Cassius Dio LXXVII. 16.3). He was also unusually overt politically in giving particularly large munera and venationes as a celebration of the tenth anniversary of his own accession (Cassius Dio LXXVII.3-5).

193 Wacher, op. cit. (note 124), 391-3.

194 White, op. cit. (note 45), 157; also Wacher, op. cit. (note 124), 259 and 265.

195 Wacher, op. cit. (note 124), 261.

196 Fulford, op. cit. (note 9), 179; cf. also 9 and 14.

197 Wacher, op. cit. (note 124), 275-6.

198 More speculatively he suggests that Stone Phase Two, which should be before the 250s, may have been linked to the celebration of Rome's Millennium in 248 (Fulford, op. cit. (note 9), 55-6). The amphitheatre remained in use up to the fourth century (ibid., 56). It is interesting to note the evidence for a possible mid-third-century revival at the London amphitheatre (see Dating above), and especially the dendrochronological date of 243. The young Gordian III (238-44) was the grandson of Gordian I (238), who was known for his passion for munera (SHA Gord. Tres 3.5-8; and 4.4) and only the second emperor after Titus to strike a coin showing the Colosseum (Wiedemann, op. cit. (note 2), 18).

199 Millar, F., ‘Emperors, frontiers and foreign relations, 31 B.C. to a.d. 378’, Britannia xiii (1982), 123.CrossRefGoogle Scholar

200 There is only one other example of a town given Trajan's name at which there is an amphitheatre: Sarmizegetusa in Dacia, which received the name of Ulpia Traiana. Trajan's involvement in Dacia hardly needs pointing out.

201 See A.J. Brothers, ‘Buildings for entertainment’, in I.M. Barton, Roman Public Buildings (1989), 119.

202 See comment in note 178 and Grenier, op. cit. (note 104), 691f.