¹ Rivet, A. L. F., ‘The British Section of the Antonine Itinerary’, Britannia i (1970), 34 f.CrossRefGoogle Scholar

² op. cit., 62.

³ ibid., 38–9. Confirmatory evidence for the length of the Roman mile in Britain is provided by milestone and milecastle spacings on Hadrian's Wall.

⁴ ibid., 38.

⁵ Collingwood, R. G. and Wright, R. P., he Roman Inscriptions of Britain, i (1965).Google Scholar

⁶ All measurements quoted in this paper are in Roman miles, taking 1,480 m. as the unit. Rivet's tables show very few plus errors in the Itinerary as a whole and where they do exist there is invariably uncertainty about the location of the place concerned, or about the accuracy or the route of the iter.

⁷ i.e. the outermost limit of a settlement area, including its cemeteries and extra-mural buildings in the case of a defended town.

⁸ That quite small forts did have territoria, albeit of unknown size, is indicated by RIB 1049, from Chester-le-Street, which refers specifically to the territorium. This fort only covered an area of c. 2-5 ha. (Britannia ii (1971), 251). The subject of territoria will be discussed later, p. 99.Google Scholar

⁹ The discrepancy has been previously explained by the tenuous suggestion that the distance was measured from Lincoln, which is 50 miles away, with at least four other towns in between. It is more logical to assume that a stone so close to Leicester and which states the distance as 2 miles a Ratis was in fact measured from there. See also Rivet, op. cit. (note 1), 38, n.17.

¹⁰ Caesaromagus is in the Moulsham district of Chelmsford and not at Widford, as Rivet states. This has been clear since 1849: VCH, Essex, iii (1963), 63Google Scholar; Essex Archaeol. and Hist, iv (1972), 3 f; P. J. Drury, ‘Chelmsford’, in W. J. Rodwell and R. T. Rowley (ed.), The ‘Small Towns’ of Roman Britain (1975), forthcoming.Google Scholar

¹¹ Few distances given in the Peutinger Table are correct as they stand. It is not difficult to appreciate how errors occurred in copying a worn and damaged manuscript and to correct them accordingly. The commonest fault is the loss of digits and letters (mainly from the beginning of words). Hence, the Kelvedon-Colchester distance may have originally agreed with iter IX and read viiii miles.

¹² There is no reason to question the initial entries of the iter, but there are problems further on.

¹³ Full consideration of this evidence will be published elsewhere, and it is only incidental to the subject under discussion. For Heybridge, see VCH Essex, iii (1963), 146; Rodwell and Rowley, op. cit. (note 10), and for recent excavations see HMSO, Archaeological Excavations 1972, 50.Google Scholar

¹⁴ The villa at Stanton Chare, north of Ixworth, would suit the Colchester distance very well. The inevitable correction for the journey to Caistor would still have to be made, this time by the addition of x, to form m.p. xxviii.

¹⁵ M. R. Hull, Roman Colchester (1958), pl. XLII.

¹⁶ Since the extent of the minor towns at Chelmsford and Kelvedon are well known, it is possible to measure precisely from their outermost limits.

¹⁷ At least two roads seem to have linked Heybridge and Colchester, but it makes no difference to the mileage in question.

¹⁸ The dykes, especially to the south of Colchester, are very incompletely known and cannot be discussed in detail here; there is good reason to believe that there are three principal systems: one cuts off the whole Colchester peninsula (Gryme's Dyke), while the other two formed roughly concentric defences to the south and west of the town, as indicated in FIG. 2. The evidence for the ‘middle’ system is substantial and much of it served as the boundary of the ancient Borough Liberty. The evidence for the ‘inner’ defence is less substantial.

¹⁹ EHR ccvi (1937), 197. Stevens suggested that the Duel Cross milestone, found 3 miles north of Aldborough, recorded the distance to York; he was using an earlier interpretation (not accepted in RIB) that XXC meant 20 miles to Colonia. There is, unfortunately, no evidence to show that York was ever known locally as Colonia (in the way that Colchester was); on the Castleford milestone it is referred to as Eb(uracum). For a reconsideration of this stone see p. 95 below.Google Scholar

²⁰ Rivet, op. cit. (note 1), 40.

²¹ I. D. Margary, Roman Roads in Britain (1967), 523.

²² Cf. J. S. Wacher, Excavations at Brough-on-Humber, 1958–61, (1969), 26.

²³ Professor Rivet has pointed out to me that there is no precedent in any of the Antonine Itineraries for returning a journey to a place through which it had already passed. This is an objection to the solution proposed, but since it is necessarily based on negative evidence, the. point cannot be pressed further. Indeed, it is perhaps worth noting that I. A. Richmond and O. G. S. Crawford in their assessment of the Ravenna Cosmography (Archaeologia xciii (1949), 12, 18) not only demonstrated that it was in part based on road itineraries (with some very close correspondences to the Antonine Itinerary), but also found it necessary to postulate several ‘returns’ to places of major importance.Google Scholar

²⁴ Rivet, op. cit. (note i), 40.

²⁵ It should, moreover, be noted that there appears to be a major cross-roads here, which could be the focus of a substantial settlement.

²⁶ Margary, op. cit. (note 21), 523.

²⁷ One final point remains to be discussed in connection with this iter: the distance from High Rochester to Corbridge is given as m.p. xx, when it is actually 25 miles. Rivet suggested the addition of v to correct the reading. This seems inevitable, but it upsets the itinerary total again, which ought then to read m.p. clxi. This may simply be a scribal confusion of x with v, which Rivet has shown to be the most common error of all (op. cit. (note 1), 62).

²⁸ But not so in itinera II, V and VIII, where it is simply Eburacum. Leug. is presumably a scribal misreading for Leg.; although all the British itinera have distances recorded in miles, many of the Continental distances are recorded in leugae, for which the abbreviation leug. is normally used.

²⁹ The legionary praetorium and the provincial governor's praetorium may have been one and the same at this period; cf. I. A. Richmond, in R.C.H.M., Eburacum (1962), p. xxxvii.

³⁰ That Delgovicia did lie on a route between Brough and York is confirmed by its appearance in one of the detectable ‘itineraries’ in the Ravenna Cosmography (note 23). Indeed the list there reads like a tour of military installations, running south from the Wall, to Lanchester, Binchester, Bowes, Catterick, York, Delgovicia and Brough-on-Humber. The absence of Aldborough (not known to be a military installation) is interesting and perhaps significant.

³¹ It is clear that the stone was an ‘old’ one at the time when the main inscription was cut (A.D. 268–70), as there are obvious traces of a primary text visible.

³² Other tombstones found inside the colonia walls, presumably in situations of reuse, include RIB 251, 265 and 268. The walls themselves contain tombstones RIB 250, 256 and 263.

³³ It is worth noting that Sapperton would fit very well if we assumed a scribal error and read m.p. xxv for m.p. xxx. Furthermore, this figure comes at the bottom of a suspicious list of repetitive mileages, from Villa Faustini onwards: m.p. xxxv, xvii, xxxv, xxv, xxxv, xxx. We have already shown that one of these must be wrong (see p. 80 and note 14). Moreover, if Camboricum is Hockwold-cum-Wilton and Durolipons is Cambridge, further problems are encountered with this run of figures, since Hockwold–Cambridge can never have been as little as 25 miles: it would have been at least 30. It is most unfortunate that this suspicious run of figures coincides with a series of place-names for which we have no independent evidence outside this iter, viz.: Villa Faustini, Camboricum, Durolipons and Causennae.

³⁴ It is not the first time that Saltersford has been suggested: J. B. Whitwell, Roman Lincolnshire (1970), 65 (with other references).

³⁵ If Bovium is at Holt (iter If), then another instance of deviation from the main route is inevitable, but again the case is not clear-cut. However, it was shown by Richmond and Crawford that branches from the main routes are in evidence in the Ravenna Cosmography: see note 23.

³⁶ Not alongside the Fosse Way, as in RIB.

³⁷ Collingwood and Wright, op. cit. (note 5), 699.

³⁸ Rivet, op. cit. (note I), 57.

³⁹ This section of the paper is based upon a copyright thesis submitted to London University in 1972: Roman London, its Roads and the Antonine Itinerary.

⁴⁰ Rivet, op. cit. (note I), 38.

⁴¹ The name is recorded in iter IX only; see Rodwell and Rowley, op. cit. (note 10).

⁴² It may be possible to fix a point on the Rochester–London milestone survey if RIB 2219 was the third stone from Rochester. It was found 3½ miles west of Durobrivae and could be used to support the suggestion that the town-zone extended for ½ mile along the road in that direction.

⁴³ This is very close to what Rivet suggested. It is not difficult to see how the scribe arrived at xviiii; he seems to have confused the preceding and succeeding entries of x and viiii respectively.

⁴⁴ There is a lack of material evidence and Rivet's argument for Crayford as Noviomagus is powerful. We cannot, of course, assume that Crecganford of A.D. 457 is to be equated precisely with the site of modern Crayford. Good archaeological evidence is needed before the matter can be pursued further: see G. Webster in Rodwell and Rowley, op. cit. (note 10).

⁴⁵ Since Rivet wrote his paper, much archaeological work has been done in Staines and the limits of Pontes better denned. I am indebted to Mrs. M. Rendell for information which she has kindly supplied.

⁴⁶ The interpretation proposed here differs substantially from that suggested by Rivet (op. cit. (note I), 49f), but does involve fewer changes to the text and has the advantage of adding up to the stated total.

⁴⁷ One possible objection might be raised over the Regnum-Clausentum distance, for which no correction has been allowed here. Rivet has already discussed this in detail (op. cit. (note I), 50) and concluded that Regno refers to the town of Chichester and not to the boundary of the territory of the Regni. Clausentum is given as 20 miles from Chichester and 10 miles from Winchester, which places it a mile north-west of Wickham (west of the river Meon), as noted by Rivet; there is no good reason to alter the figure in the iter to make Clausentum fit Bitterne. The only historical mention of Clausentum is in the Antonine Itinerary and since the meaning of the name remains obscure, no help can be derived from topography, etc. The identification of Clausentum with Bitterne has been accepted with little reservation by most recent writers and the Ordnance Survey, Map of Roman Britain (1956). For a discussion of the name see M. A. Cotton and P. W. Gathercole, Excavations at Clausentum, Southampton, 1951–54, (1958), 6; the primary source was not examined in detail and the Bitterne-Clausentum identity ‘established’ on no firmer evidence than the speculations of eighteenthand nineteenth-century antiquaries. Clearly, the itinerary evidence has simply been strained in the past, in order to fit an attractive site to the name. Potentially, Wickham is just as attractive, bearing in mind the fact that a settlement of vicus status is indicated by the name alone; cf. Med. Archaeol. xi (1967), 90; see also, B. Cunliffe, The Regni (1973), 72.Google Scholar

⁴⁸ E.g. VCH Essex, iii (1963), 24.Google Scholar

⁴⁹ K. Jackson, in Rivet, op. cit. (note 1), 73. Nor does it help to assume that the iter is incorrect in its mileage and place Durolitum at Romford, because it would then be only 13½ miles from the centre of London. Iter IX gives the Durolitum-Londinium distance as 15 miles, and this would also have to be altered to suit.

⁵⁰ VCH Essex, iii (1963), 88. See also note 41.Google Scholar

⁵¹ The archaeology of the Passingford Bridge area remains unexplored and much of the potential area for Durolitum is under parkland, which could be responsible for its concealment.

⁵² Margary, op. cit. (note 21), 57. Also discussed by the present writer (see note 39).

⁵³ It is perhaps no coincidence that an uninscribed stone stood by the roadside here until sometime in the last century; local antiquaries believed it to be a Roman milestone. It seems fairly certain that it can be identified with a well-known marker called the Ossulstone, to which reference is made in Medieval and later documents and maps. Oswulf's Stone was a sufficiently important feature of the Anglo-Saxon landscape to give its name to the Hundred in which London lies (the Ossulstone Hundred). For discussion, see: J. E. B. Gover, A. Mawer and F. M. Stenton, The Place-names of Middlesex (1942), 81; Trans. London & Middx. Archaeol. Soc. iv (1875), 62.Google Scholar

⁵⁴ The validity of its findspot is confirmed by the discovery of a second stone of later date close by (RIB 2266).

⁵⁵ Collingwood and Wright, op. cit. (note 5), 713.

⁵⁸ It is evident that the erection of a series of milestones around Aldborough was undertaken as a planned event in the reign of Decius (249–51) since the only three stones from the area all bear his name (RIB 2276–8). They are of buff gritstone and two were certainly cut by the same hand (RIB 2276–7).

⁵⁷ That the S was not merely the mason's signature, or other unimportant mark, is indicated by the size, prominence, symmetrical placing and enclosure, with XXC, between stops.

⁵⁸ That one milestone was not just a replacement for the other seems certain from their close temporal proximity: RIB 2235 was erected under Florian (A.D. 276) and RIB 2238 under Victorinus (268–70).

⁵⁹ Where the milestone RIB 2297 was found, and the probable base of another AA ³ viii, 141.

⁶⁰ For full discussion see, Birley, E., CW², liii (1953), 52fGoogle Scholar; and Graham, A. J., JRS lvi (1966), 96. The stone was found exactly 53 miles from the edge of Roman Carlisle.Google Scholar

⁶¹ JRS lv (1965), 224.Google Scholar

⁶² As concluded by Graham, op. cit. (note 60).

⁶³ RIB 2222, Respublica Belgarum; RIB 2250, Respublica Civitatis Dobunnorum; JRS lv (1965), 224, Respublica Carvetiorum.Google Scholar

⁶⁴ RIB 2240, Respublica Lindensis, and probably RIB 2241 also.

⁶⁵ It seems likely that many of the small towns were of vicus status, but the subject of non-military vici in Britain is a difficult one: see J. S. Johnson in Rodwell and Rowley, op. cit. (note 10). For an important paper which offers potential in pinpointing some sites which were certainly known as vici, see Med. Archaeol. xi (1967), 90; Rodwell in op. cit. (note 10).Google Scholar

⁶⁸ The explanation for part or all of this problem may be found in the name Sulloniacae (‘The Estate of Sullonius’); this could not have been a vicus and therefore would not have had a town-zone. Private estates are very rarely included in the Itinerary, although the roads must have been lined with them; presumably where a mileage was recorded it was to the centre of the estate. The only other certain example of a private estate being mentioned in the British itinera is Villa Faustini, but there are also several uncertain instances, such as Vagniacae and Bremetonnacum. This does not mean, of course, that all other places mentioned were of sufficiently high status to have had a town-zone.

⁶⁷ Rivet uses the term pomerium as equivalent to our ‘town-zone’. The areas outside towns which fall within the town-zones are too large to be included in the pomerium. In Classical usage this word seems to be reserved for a town's defences and a small area outside which was kept clear of buildings, etc.: Livy, i, 44.4. British town-zones demonstrably include buildings and cemeteries. I am grateful to Mr. J. S. Johnson for his assistance on the matter of pomeria.

⁶⁸ It is uncertain whether land-allotment by centuriation was still being practised under Claudius; if it was then it should certainly appear around Colchester, if nowhere else in Britain. There is, however, no convincing sign of centuriation in the road- or field-systems around Colchester, and Tacitus's reference to land-appropriation for the Colonia cannot be taken to imply the existence of such a system: Annals, xii, 32.5, where he says the colonia was founded in agros captivos. See also Annals xiv, 31.5. Rectilinear field-systems have been recorded in various parts of Britain, but none can be confidently assigned the term centuriation; cf. O. A. W. Dilke, The Roman Land Surveyors (1971), chapter 13.

⁶⁹ E.g. Chester in iter II, or Caerleon in iter XII.

⁷⁰ The problem of the growth of canabae and coloniae within a territorium legionis is discussed by Mócsy, ‘Das Territorium legionis und die Canaae in Pannonien’, Acta Arch Acad. Scient. Hungaricae iii (1953), 199 f.; and ‘Zu den Prata Legionis’, Studien zu den Militärgrenzen Roms (1967), Report of the Sixth Limes Congress, p. 211–14.Google Scholar

⁷¹ As suggested for Lincoln: Whitwell, op. cit., 26, 40; but he was thinking in terms of a territorium covering many square miles. It should not be forgotten that, although the existing milestones may be late, the road-survey itself must go back to the first century–in most cases to the Julio-Claudian period–and therefore to the moment in time when the legionary territoria were in being.

⁷² See also note 70. There is no evidence for the size of legionary territoria in Britain. Richmond assumed that the territorium of York would cover ‘many square miles’ (op. cit. (note 29), p. xxv); and H. von Petrikovits has suggested that legionary fortresses in Germany had small, inner territoria covering some 7,500 ha. and much larger, outer territoria of many square kilometres, upon which lay farms under legionary control (Das römische Rheinland, 1960, 63 f.). Perhaps, then, our measurements in Britain are mapping out the inner territorium legionis (although much smaller than its German counterpart), which was later transformed into the colonial territorium.

⁷³ S. S. Frere, Britannia (1967), 205. If the town-zones do represent diminished territoria around coloniae, an interesting problem is raised in relation to municipia, since they too could have territoria. Unfortunately, the evidence for municipia in Britain is far from conclusive: several have been suggested, but Verulamium alone is attested by documentary evidence. It is clear that Verulamium had but a modest town-zone, like other civitas-capitals, and a possible explanation for this may be sought in the careful handling of the Britons following the Boudiccan revolt: it may be a direct result of this that coloniae were only awarded restricted territoria, whilst municipia had none.

⁷⁴ Hence the perpetuation of the idea that ‘London Stone’ was the milliarium aureum for Britain: R. Merrifield, Roman London (1969), 95–7; or Stevens's attempt to make Durobrivae (Water Newton) into a tribal capital because of its associated milestones (op. cit. (note 19), 199); or the unquestioned acceptance of the Lincoln central milestone. On the question of the status of Water Newton, I. Hodder and M. Hassall have recently shown that it was a town of some importance which might have been a ‘late’ elevation to the rank of civitas—capital: Man (New Ser.) vi, (1971), 391 f. Even so, it is unlikely to affect the positioning of milestones since the original survey of the area must date to the Julio–Claudian period and it is highly improbable that whenever a town underwent a change of status all the milestones in the district were replaced, to accord with the new caput viae. For discussion of other milestone systems in the Empire see: Mommsen, Th. in Corpus lnscriptionum Latinarum, VIII (ii), 1881Google Scholar, 859; and Goodchild, R. G. in Berytus, ix (1949), 91 f.Google Scholar

⁷⁵ For example: there are minus errors, somewhat greater than any known in Britain, observable in mileages approaching Cologne, thus: Neuss-Cologne, m.p. xvi, is 23½ miles in reality (Cuntz, O., Itineraria Romana, i, 1929, 370.6); Bonn-Cologne, m.p. xi, is actually 17 miles (370.7); Julich–Cologne, m.p. xviii, is 28 miles (375.9, 376.1). Similar results obtained for Xanten: Kalkar Xanten, m.p. v, is really 9½ miles (370.2); but note the lack of a significant minus error between legionary fortresses, as in Britain, thus: Vetera–Neuss, m.p. xxxvi, is in fact 37 miles on the ground (370.4, 5). Hence, it would appear that town-zones could be similarly plotted around military coloniae in Germany.Google Scholar