¹ ‘Gallo-Römische Strassensiedelungen und Kleinhausbauten.’ Bonner Jahrbücher, cxxviii (1922), 77 ff.Google Scholar

¹ Op. cit., 82.

³ Richmond, , Roman Britain² (London 1963), 77.Google ScholarRivet, in Wacher, J. S. (ed.) Civitas Capitals of Roman Britain (Leicester 1966), 109.Google Scholar

⁴ Dorchester: RIB, 235.

⁵ Irchester: RIB, 233. It is true that this strator's only link with Irchester may be the fact that he died there. On the other hand, the distribution of the staff of provincial governors on special duties in other provinces suggests that stationing of this man here is a distinct possibility.

⁶ Especially pp. 131–79.

⁷ Wacher, J. S. (ed.), Civitas Capitals of Roman Britain (Leicester, 1966).Google Scholar

⁸ Procs. Dorset N. H. & A. S., liv (1932), 87.Google Scholar

⁹ Arch.J., civ (1947), 62.Google Scholar

¹⁰ VCH, Wilts., i, 1,109 f.

¹¹ I exclude from this account the vici outside the forts of northern Britain, some of which clearly rose to the size, and perhaps the dignity, of small towns: Salway, , The Frontier People of Roman Britain (Cambridge 1965)Google Scholar

¹² Proc. Cambridge Antiq. Soc, lx (1967), 39 ff.Google Scholar

¹³ VCH, Essex, iii, 171 ff (Rivenhall); ibid., 133 f. (Gestingthorpe); JRS, xlviii (1958), 98; li (1961), 175 (Thistleton Dyer).Google Scholar

¹⁴ p. 85, fig. 5.

¹⁵ JRS, liv (1964), 164 f.Google Scholar

¹⁶ Canterbury: JRS, xliii (1953), 128: xliv (1954), 103. This was pointed out to me by Prof. Frere.Google Scholar

¹⁷ JRS xliii (1953), pl. xiv, 1.Google Scholar

¹⁸ Two more instances, out of many, may be noted. Brough (Notts.) has an acreage of some 6·75 within its innermost defensive circuit (undated). Outside this nucleus, building material and pottery extend over a further 18 acres, to north and south of the defences. Even allowing for the spreading of Roman material over the modern surface by agriculture, occupation was clearly by no means confined to the enclosed area: information from Mr. D. F. Mackreth and from personal observation. Irchester (Northants) measures nearly 20 acres within its defences, but a recent plan shows surface-finds covering more than 100 acres: Arch. J., cxxiv (1967), 66, fig. 1.Google Scholar

¹⁹ VCH, Oxon., i, 282, fig. 19. JRS, xliii (1953), 92.Google Scholar

²⁰ Trans. E. Herts. A.S., xiii (1950–1954), 98 ff, and fig. 1.Google Scholar

²¹ Wedlake, W. J., Excavations at Camerton (Camerton Excavation Club 1958), pl. iii.Google Scholar

²² JRS, xliii (1953), 91. St. Joseph, in Civitas Capitals of Roman Britain, 25 ff. and pl. iv.Google Scholar

²³ JRS, 1 (1960), 217 f.Google Scholar

²⁴ JRS, xliii (1953), 90 f.Google Scholar

²⁵ Arch. J., cxix (1962), 114ff., esp. fig. 1.Google Scholar

²⁶ JRS, li (1961), 132 f.Google Scholar

²⁷ JRS, xliii (1953), 92Google Scholar and pl. xiv. Frere, , Britannia (London 1967), pl. 32Google Scholar, 6. Collingwood, and Richmond, , The Archaeology of Roman Britain (London 1969), pl. x (b).Google Scholar

²⁸ Ibid, and Arch. J., cxxiv (1967), 65 ff. Mildenhall (Wilts.) has a more regular lay-out: Civitas Capitals of Roman Britain, 24.Google Scholar

²⁹ JRS, lV (1965), 87.Google Scholar

³⁰ Corder, P. (ed.), The Roman Town and Villa at Great Casterton (Rutland), Report III (Nottingham 1961), 33f.Google Scholar

³¹ It is worth noting that the later town defences enclose an area of settlement along Ermine Street, and not along the road which ran through the early fort (Ryhall Road). Thus, by the end of the second century, if no earlier, the core of the settlement lay about 400 yards from the site of the fort.

³² JRS, xliii (1953), 92: xlviii (1958), 98.Google Scholar

³³ Wedlake, op. cit. (n. 21), p1. iii.

³⁴ Corder, op. cit. (n. 30) 33 f.

³⁵ Trans. Thoroton Soc, lxxiii (1969), p. 56, fig. 23.Google Scholar

³⁶ Trans. Lichfield & S.Staffs A. & H.S., v (1963–1964), 15f.Google Scholar

³⁷ Trans. Thoroton Soc, lxxiii (1969), 63.Google Scholar

³⁸ Arch.J., cxx (1963), 1 ff.Google Scholar

³⁹ Corder, op. cit. n. 30. 35 ff. This interpretation as an aisled structure is that of the excavator, but it should be noted there is no direct evidence to demonstrate that the holes identified as those for aisle-posts really belonged to this building.

⁴⁰ JRS, li (1961), 171: lii (1962), 167.Google Scholar

⁴¹ Procs. Bristol Univ. Spelaeological Soc, xi (1967), 126 ff. The massive walls, and the position of the site, mark Gatcombe as an oddity. Is it a defended villa?Google Scholar

⁴² JRS, li (1961), 172: lv (1965), 209.Google Scholar

⁴³ VCH, Essex, iii, 72.

⁴⁴ JRS, xliii (1953), 90f.Google Scholar

⁴⁵ Trans. Birmingham A.S., lxxv (1959), 1ff.Google ScholarJRS, xlvi (1956), 130.Google Scholar Another corridor house stood close by: JRS, lviii (1968), 187.Google Scholar

⁴⁶ Wedlake, op. cit. (n. 21),49, fig. 13.

⁴⁷ JRS, xxxi (1941), 45, fig. 9.Google Scholar

⁴⁸ Trans. Bristol & Glos. A.S., lvii (1935), 234Google Scholar, referring to Witts, , Archaeological Handbook of Gloucestershire, (Cheltenham, 1883), 56.Google Scholar

⁴⁹ Sussex Arch. Colls., lxxxi (1940), 54 ff. has the fullest plan of this building.Google Scholar

⁵⁰ JRS, lv (1965), 220.Google Scholar Borough Farm: VCH Sussex iii (1935), 25.Google Scholar

⁵¹ Boon, , Roman Silchester (London 1957), 138 ff. and general plan.Google Scholar

⁵² Archaeologia, lxii, 1 ff. and 405 ff., and pl. lxiv.

⁵³ Antiquity, xxxviii (1964), 103 and figs. 1, 5 and 6.Google Scholar

⁵⁴ MPBW Guide, Wroxeter Roman City (prepared by G. Webster), Insula 15.

⁵⁵ Report of the Research Comm. of the Woolhope Club (1916). JRS, xliii (1953), 92 and pl. xiv.Google Scholar

⁵⁶ VCH, Oxon., i, 281.

⁵⁷ Report of the Research Comm. of the Woolhope Club (1916f.

⁵⁸ JRS, xlviii (1958), 139 f.Google Scholar

⁵⁹ Lewis, , Temples in Roman Britain (Cambridge 1966), 203, fig. 126.Google Scholar

⁶⁰ JRS, lvii (1967), 201 and fig. 20.Google Scholar

⁶¹ Wedlake, op. cit. (n. 21), 52, fig. 15. Lewis, op. cit. 96 and 192, fig. 98. Whatever its function, this building is oddly planned. It is to be noted that a neighbouring building, evidently a dwelling, also has a semicircular yard: Wedlake, op. cit., 51, fig. 14. A very interesting, but incompletely studied, group of buildings at Wycomb (Glos.) seems to have included a temple and possibly a small theatre: Lewis, op. cit. (n. 59), 194, fig. 109. The well-attested association of periodic fairs and markets with religious sites in the Celtic world may be recalled.

⁶² Trans. E. Herts A.S., xiii (1950–1954), 98 ff., and esp. fig. 2.Google Scholar

⁶³ VCH, Oxon., i, 281, summarizing work in the late nineteenth cent. by J. L. Myres and P. Manning. Excavation unpublished in full; records among the Manning MSS. in the Ashmolean Museum, Oxford.

⁶⁴ Zeitschrift für schweizerische Altertumskunde und Kunstgeschichte, iii (1941), 77:xxv (1968), 102.Google Scholar

⁶⁵ Lewis, op. cit. (n. 59) 184, figs. 68–69.

⁶⁶ Corder, op. cit. (n. 30), 49 f. A mansio has recently been identified at Godmanchester (Hunts.): Current Archaeology, Sept. 1969, 136.

⁶⁷ Trans. Birmingham A.S., lxxiv (1958), 12ff.Google Scholar

⁶⁸ Trans. Lichfield & S.Staffs. H. & A.S., viii (1966–1967), 11 and fig. 1. Dr. Webster (see preceding note) had already pointed out the general similarity to military bath-houses.Google Scholar

⁶⁹ VCH, Oxon., i, 284.

⁷⁰ Arch.J., civ (1947), 62.Google Scholar

⁷¹ Rivet, The Distribution of Roman Villas in Britain in Research Report no. 1 of the Council for British Arch., 32; idem in A. L. F. Rivet (ed). The Roman Villa in Britain (London 1969), 180 f.

⁷² Corder, op. cit. (n. 30), 59.

⁷³ JRS, lv (1965), 88.Google Scholar

⁷⁴ F. Oswald, The Last Days of Margidunum (1956), 4, and personal observation.

⁷⁵ Both sites are as yet unrecorded in full. One, a corridor-house, was found, and robbed of its stone, in the late nineteenth century (SK 722425). The other, represented by a very large scatter of building debris and pottery, was discovered in 1967 (SK 727418). Since the latter is only half a mile from the other, it may well not be a villa at all, although it includes at least one stone building. Alternatively, it may not be contemporary with the other.

⁷⁶ Reports and Papers of the Lincolnshire Arch. & Archit. Soc, vii (1957), 13. Other finds in the near vicinity indicate the former existence of stone buildings.Google Scholar

⁷⁷ East Midland Arch. Bull, ix (1966), 11 f.Google Scholar

⁷⁸ For instance, Ilchester (Som.), where a large courtyard house stood only half a mile from the walls: JRS, lviii (1968), 199Google Scholar. Godmanchester (Hunts.), with a villa half a mile away to the north: JRS, xlv (1955), 89Google Scholar; xlviii (1958), 98. Proc. Cambridge Antiq. Soc., lxi (1968), 19Google Scholar. Asthall (Oxon.), with one villa within a mile and possibly two: Oxoniensia, xx (1955), 29Google Scholar; VCH, Oxon., i, 330. Bourton Bridge (Glos.), where the settlement spreads for more than half a mile to the east of the Fosse Way, and includes the agricultural building earlier called the ‘Leadenwell villa’: Trans. Bristol & Glos. A.S., lv (1933), 377; lvi (1934), 99; lvii (1935), 234. Several other sites display this merging rather less clearly, and thus cannot yet be brought into the argument.Google Scholar

⁷⁹ Rivet, loc. cit. (n. 71).

⁸⁰ See note 13, p. 117 above.

⁸¹ And close to cities too: e.g. Wroxeter, : JRS, lv (1965), 87. Excavation has not yet demonstrated whether these are contemporary with the city or earlier.Google Scholar

⁸² JRS, lv (1965), 87.Google Scholar

⁸³ E.g. in the Welland valley: Thomas, C. (ed.) Rural Settlement in Roman Britain (CBA Research Report no. 7), 15Google Scholar ff. The same seems to be true of some of the Trent valley sites: e.g. Rampton (Notts.), East Midland Arch. Bull., ix (1966), 41 ff., and Shelford (Notts.), unpublished.Google Scholar

⁸⁴ E.g. Irchester, : Arch.J., cxxiv (1967), 70Google Scholar ff. Camerton: Wedlake, op. cit. (n. 21), 40 f. Dorchester-on-Thames: Arch. J., cxix (1962), 128.Google Scholar Thorpe (Notts.): JRS, liv (1964), 159Google Scholar. Ancaster: recent unpublished excavations by the Univ. of Nottingham. Old Sleaford: JRS, lii (1962), 167. The large number of Belgic coins from Braughing points to a pre-Roman origin.Google Scholar

⁸⁵ Possibly this view has come to general notice through the writings of geographers who were compelled to rely upon out-of-date archaeological information: e.g. H. C. Darby, ‘The Clearing of the English Woodlands’, Geography 1951, 71 ff., and Hoskins, W. G., The Making of the English Landscape (London 1955), 35Google Scholar ff. It is worth remembering that these two writers did not have the advantage of using the third edition of the O.S. Map of Roman Britain. On the progress and difficulties of clearance in the early Medieval period: Hoskins, in Poole, A. L. (ed.) Mediaeval England² (Oxford, 1958), 9 ff.Google Scholar

⁸⁶ Antiq. J., xliv (1964), 19Google Scholar ff. and Proc. Cambridge Antiq. Soc, lx (1967) 39Google Scholar ff. for the region about the Wash. RCHM (England), Dorset Inventory (forthcoming) for parts of the Wessex downland.

⁸⁷ In general, Richmond in Civitas Capitals of Roman Britain, 82 f.

⁸⁸ Hartley, B. R., Notes on the Roman Pottery Industry in the Nene Valley, (Peterborough Mus. Soc. 1960), 4 ff.; O.S. Map of Roman Britaini³, fig. 3.Google Scholar

⁸⁹ I am indebted to Dr. J. P. Wild for information about recent excavation and field-work in the Nene valley.

⁹⁰ JRS, liv (1964), 208 (Mancetter): li (1961), 173 (Hartshill). The floruit of these kilns was in the second and third centuries.Google Scholar

⁹¹ Trans. London & Middx. A.S., xvi (1951); xviiGoogle Scholar (1952–54). VCH, Middlesex, i (1969), 66.Google Scholar

⁹² Trans, Birmingham A.S., lxiv (1941–1942), 42 f., for refs. to Saxon salt-springs. It is not clear to what extent these were exploited commercially by the Romans.Google Scholar

⁹³ JRS, li (1961), 172: lv (1965), 209.Google Scholar

⁹⁴ Wedlake, op. cit. (n. 21), 82 ff.

⁹⁵ JRS, liv (1964), 165Google Scholar. Trans. Woolhope F.C., xxxviii (1965), 124 ff.Google Scholar

⁹⁶ Fieldhouse, May, and Wellstood, , A Romano-British Industrial Site near Tiddington, Stratford-on-Avon (Birmingham 1931).Google Scholar

⁹⁷ Trans. Woolhope F.C., xxxviii (1965), 124ff.Google Scholar

⁹⁸ A very small part of the settlement was excavated, and while that part was certainly concerned with metal-working, this does not indicate that the whole place necessarily was. The ‘industrial settlement’ boils down to one tile-kiln, sited in a rough stone enclosure or building, an iron-smelting furnace within another enclosure, and a small lead-smelting furnAcc. Relative dating is quite uncertain, and there is thus no good reason for supposing that these modest installations were in use simultaneously.

⁹⁹ But several others may prove on further examination to have been industrial centres. Kettering (Northants.) probably had some connexion with the nearby iron-stone deposits (VCH, Northants., i, 194). Ancaster presumably was a centre for quarrymen, and for craftsmen dependent upon the excellent Ancaster limestone. Norton (Yorks. E.R.) was a site where pottery-making rose to great importance in the late third and fourth centuries. This settlement is distinct from the vicus of the fort at Malton, which lies three miles away to the east (Clark, Kitson, Gazetteer of Roman Remains in East Yorkshire (Leeds, 1935), 113Google Scholar). Wilderspool (Lanes.) was certainly preoccupled with industry, but its history was peculiar, in that all the large-scale activity dates from the second cent., and most of it from the Antonine period. (T. May, Warrington's Roman Remains (1904), corrected in Northern History, iii (1968), 13–16).Google Scholar

¹⁰⁰ Rivet, in Civitas Capltals of Roman Britain, 110.

¹⁰¹ The Eifel: Germania, xxxiv (1956), 99–125.Google ScholarSchleswig, : Archaeologia Geographica, viii/ix (1959/1960), 7; x/xi (1961/3(1965)), 19.Google Scholar