¹ Milne, G. and Milne, C., Current Arch, lxvi (1979), 198–204.Google Scholar

² This paper was compiled from the archive reports for Peninsular House, PEN 79, Areas B and C., and Pudding Lane, PDN 81, Areas C and F. The excavation also examined Roman, Saxon and post-Medieval buildings and associated activity to the north of these areas, and this material will be discussed in separate reports.

³ Battarbee, R., Diatom analysis of River Thames foreshore deposits…at Pudding Lane, London. Working Papers in Palaeoecology, No. 2 (University College London, 1982), 7.Google Scholar

⁴ Milne, G., Britannia xiii (1982), 271–6.CrossRefGoogle Scholar

⁵ A timber bridge pier at Mainz was larger at c. 13 m × 7 m, and an example from Laupen was smaller, at c. 5 × 5 m: see H. Cüappers, Die Trier Römerbrücken (1969), 185 abb. 156 and 198, abb. 169.… The piers of the Claudian timber bridge over the Rhine were c. 7 m × 4 m. See Fehr, H., Bonner Jahrbuch. clxxxi (1981), 287–301, fig. 2, andGoogle ScholarSchieferdecker, F. D., Bonner Jahrbuch. clxxxi (1981), 313–25, fig. 2.Google Scholar

⁶ A timber from Area B thought to be from the landing-stage could have been felled c. A.D. 50 to A.D. 60, according to the analysis of dendrochronological sample by Jennifer Hillam, University of Sheffield. The timber used in this and the other main waterfront structures was oak, identified by V. Straker, Museum of London.

⁷ Milne, op. cit. (note 4), fig. 3 (but scale is wrong), and pl. xxvib.

⁸ The post-and-plank revetments were superficially similar to those recorded to the east in 1974 on site marked 5 on fig. 1: see Jones, D. and Rhodes, M., Excavations at Billingsgate Buildings 1974, London Middx. Archaeol. Soc. Spec. Pap. No. 4 (1980).Google Scholar

⁹ All pottery dates in this report are provisional and were kindly supplied by Dr P. Tyers, Museum of London.

¹⁰ Dendrochronological analysis by Jennifer Hillam, University of Sheffield.

¹¹ However, nineteenth-century sewer excavations located a ragstone and tile wall c. 0·8 m wide with adjoining hypocaust in the middle of the Lane: Merrifield, R., The Roman City of London (1965), 286.Google Scholar

¹² G. E. Rickman, Roman Granaries and Store Buildings (1971), 1.

¹³ Horrea Galbana, the largest warehouse in Rome, covered 75,000 sq. m (see Rickman, op. cit. (note 12), 5) compared with c. 300 sq. m for Buildings 1 and 2 in London.

¹⁴ Rickman suggests that the different plan of some of the Portus buildings may be due to the ‘exigencies of space and the limitations which a harbour shape imposed on the buildings which surrounded it’, op. cit. (note 12), 131.

¹⁵ The early third-century building on the east mole was c. 100 m long and c. 13 m deep. See R. Bartoccini, Il porto romano di Leptis Magna (1960), 120; Rickman, op. cit. (note 12), 134–6, fig. 29.

¹⁶ See, for example, Manning, W. H., Saalburg Jahrbuch xxxii (1975), 105–29;Google ScholarGentry, A., Roman military stone-built granaries in Britain, BAR, 32 (1976); andGoogle ScholarMorris, P., Agricultural Buildings in Roman Britain, BAR, 70 (1979).Google Scholar

¹⁷ Rickman, op. cit. (note 12), Appendix 1, 297. For flooring in granaries, see also Black, E. W., Britannia xii (1981), 163–5.CrossRefGoogle Scholar

¹⁸ Rickman, op. cit. (note 12), 232. Ventilation-openings in the walls were almost invariably associated with raised-floors, whether of timber or tile and cement.

¹⁹ cf. the plan of the Trier horrea in Rickman, op. cit. (note 12), 265, fig. 65, and E. M. Wightman, Roman Trier and the Treveri (1970).

²⁰ A discrepancy which cannot wholly be accounted for by London's smaller population.

²¹ It is also possible that goods were sold directly from these buildings. The open-fronted form of Buildings 1 and 2 is similar to that of the north and east sides of the Horrea Agrippiana in Rome, which Rickman concludes ‘had a purpose that was at least partly that of retailing goods’, op. cit. (note 12), 97.

²² Perhaps near the Forum, some 300 m north of the harbour? The problem of establishment, subsequent ownership and management of horrea is discussed extensively in Rickman, op. cit. (note 12), chapter 5. From the first century onwards, the general pattern is of increasing Imperial intervention in ownership and control of warehouses. In the provinces, ‘… it is certain that the upkeep and control of the horrea… seem normally to have been the responsibility of the governor’, pp. 180–1. If Buildings 1 and 2 are to be interpreted as ‘temporary’ horrea (? transit sheds) operating in association with larger Imperial horrea, then they too should have been under the control of the governor.

²³ Of the c. 175 kilograms of pottery recovered from these waterfront deposits, c. 75 per cent were amphora sherds: this contrasts with domestic sites elsewhere in the City where the proportion is usually only 20–30 per cent. (Information from P. Tyers.) Numerous finds of barrel-staves and woodwaste interpreted as broken-up packing-cases also suggest mercantile activity in the immediate vicinity. (Information from B. Davies and C. Rochester.)

²⁴ For substantial evidence of a similar timber-faced quay c. 60 m long to the west of London Bridge see Miller, L., London Arch, iv (1982), 143–7 (Site 1 on FIG. I), and for the evidence of a quay on Regis House (Site 2 on FIG. I), see R. Merrifield, The Roman City of London (1965), 284.Google Scholar

²⁵ Not illustrated, but occupied similar area to Building 4 on FIG. 2d.

²⁶ The size of the column fragment suggests that it may have come from the courtyard of a private house, rather than a major public building. (Information from T. Blagg, University of Kent.)

²⁷ For report on the use of air-dried brick walling in London, see S. Roskams and D. Perring, The development of Roman London west of the Walbrook (in preparation), and also archive report on building materials from Pudding Lane site by Frances Pritchard.

²⁸ Samian pottery from these deposits has been provisionally dated to c. A.D. 120–30, but some of the coarse wares could be as late as 150.

²⁸ Pottery from this deliberately infilled drain has been dated to the later second century. The high proportion and range of finewares can be paralleled on the New Fresh Wharf excavations to the south (see note 30): warehouse clearance?

³⁰ The early third-century quay has been found at Seal House, New Fresh Wharf and, most recently, Billingsgate Lorry Park sites: numbers 7, 8 and 9 on FIG. I. See Miller, L. and Schofield, J., Excavations at New Fresh Wharf, Part I, Roman. (London Middx. Archaeol. Soc., Spec. Pap., forthcoming.) Dendrochronological analysis of timber from the quay suggests a felling date of c. AD. 210, which is compatible with the dates for the pottery from the infill deposits.Google Scholar

³¹ ‘T’-shaped foundations supporting a vaulted superstructure are known from the Porticus Aemilia in Rome and Trajan's harbour at Portus for example. See Rickman, op. cit. (note 12), 99, fig. 22; and p. 125, fig. 27a.

³² Dyson, T. and Schofield, J., London Middx. Arch. Soc. xxxii (1981), 39–44.Google Scholar

³³ Provision may have been made for temporary lockable shuttering of, for example, wooden slats as in the Horrea Agrippiana. Rickman, op. cit. (note 12), 91.

³⁴ The pottery from the late Roman contexts, including the backfill deposit of this drain, seems to represent a ‘standard domestic assemblage’; with a ‘reasonable proportion’ of finewares to coarsewares and no amphorae. (Information from Caroline Rochester.) cf. notes 23 and 29.

³⁵ The hexagonal Trajanic harbour at Portus provided c. 2000 m of quayside, which presumably functioned with the extensive riverside quays in Ostia and Rome itself. In London there is evidence for c. 70 m of quay east of Fish Street Hill, with at least 150 m to the west (see note 24), a total of c. 220 m. If the fragments of quay observed near St James Garlickhithe (Dyson and Schofield, op. cit. (note 32), 39, pl. 10) and south of Cannon Street station (Marsden, P., London Middx. Arch. Soc. xxvi (1975), 52–4) are part of a continuous development, a maximum length of c. 620 m for the first-century quay could be suggested, since there is some evidence that it did not extend west or east of these points.Google Scholar

³⁶ R. E. M. Wheeler, An inventory of the historical monuments in London, vol. III, Roman London (1928), 12.

³⁷ P. Marsden, A Roman ship from Blackfriars, London (1966).

³⁸ Marsden, P., Int. Journal Nautical Arch, iii (1974), 55–65, andCrossRefGoogle ScholarMarsden, P., London Middx. Arch. Soc. xxi (1965), 118–31. cf.Google ScholarCasson, L., JRS Iv (1965), 31–9, withGoogle ScholarMarsden, P. in Milne, G. and Hobley, B. (eds.), Waterfront archaeology in Britain and northern Europe, C.B.A. res. rep. 41, (1981), 10–12.Google Scholar

³⁹ See note 30.

⁴⁰ This change was also reflected on Area A, just north of Area B on the Peninsular House site (see FIG. I), where important evidence for a late Roman fishing industry (which included the native manufacture offish products) was recovered. This represents more evidence for the decline of imported luxury goods, such as Mediterranean garrum and hallec, and their replacement by local commodities in the third and fourth centuries.

⁴¹ See note 32. Further evidence of late Roman activity in the area comes, for example, from the town house and bath-building near Billingsgate (site 6 on FIG. I) (P. Marsden, Roman London (1980), 151–5; 180–6) as well as from the buildings and associated features on the terrace overlooking the waterfront on the Pudding Lane and Peninsular House sites, but is outside the scope of this report. In the west of the City, late Roman waterfront occupation was recorded at St. Peter's Hill site: see Britannia xiii (1982), 374.Google Scholar

⁴² See notes 12 and 16.

⁴³ Weeks, J. in McGrail, S. (ed.), Woodworking techniques before A.D. 1500, BAR Int. Series, 129 (1982), 157–68, surveys th e present state of knowledge of Romano-British carpentry, from which it was ‘possible to distinguish a series of Roman joints, and to acknowledge their lack of sophistication compared with the intricate joints of Norman church roofs’ (p. 157).Google Scholar

⁴⁴ An assessment of the scale and scope of the tidal harbour can suggest the type and size of vessels which might use it. In addition, the well-finished sections of dugout drain excavated on the quayside (PL. 2a) show that the necessary skills to construct a dugout boat existed in London. Barges of the Zwammerdam 2, 4 and 6 type may therefore have been working the Roman Thames. See Weerd, M. de in Taylor, Joan Du Plat and Cleere, H. (eds.), Roman Shipping and Trade, C.B.A. res. rep. 24 (1978), 15–21.Google Scholar

⁴⁵ cf. Hanson, W. S., Britannia viii (1978), 293–305, andCrossRefGoogle ScholarMeiggs, R., Memoirs of the American Academy in Rome xxxvi (1980), 185–95.CrossRefGoogle Scholar

⁴⁶ Medieval riverfront revetments on the Thames utilized smaller timbers, and show the greater economy and skill of the medieval carpenter. See Milne, G. in McGrail, S. (ed.), Medieval ships and harbours in northern Europe, BAR Int. Series, 66 (1979), 145–53.Google Scholar

⁴⁷ Willcox, G., London Middx. Arch. Soc. xxvi (1975), 285.Google Scholar