¹ Strang, A., Ptolemy's Geography Reappraised, unpub. Ph.D. thesis, University of Nottingham (1994), 23Google Scholar for full list of Ptolemy's known work as the Geographia partly depends on some of his previous work.

² Dilke, O.A.W., ‘Cartography in ancient Europe and the Mediterranean’, in Harley, J.B. and Woodward, D. (eds), The History of Cartography I (1987), 268Google Scholar , apparently no maps accompanied the Greek manuscript of the Geographia.

³ Strang, op. cit. (note 1), 189-202. Comprehensive lists are given of some 296 codices and their repositories. Among some 102 Greek manuscripts there are possibly 32 with maps, codices with 26/27 maps are designated the ‘A’ Recension and later variants with 64/65 maps the ‘B’ Recension.

⁴ It is likely that most if not all of the Greek manuscripts emanated from Maximus Planudes’ original, which may have been codex Vaticanus graecus 177 which claims him as owner. Three ancient Geographia manuscripts are known in Arabic ( Karamustapha, A.T. in Harley, J.B. and Woodward, D. (eds), The History of Cartography II, 10Google Scholar ) but without translation and suitable research it is not known whether they could have been derived from a separate or earlier origin.

⁵ The earliest printed maps from Ptolemy's Geographia appear (in Latin) to have been produced in Bologna possibly as early as 1462. Subsequently, variants proliferated throughout Europe as the art of printing rapidly spread across the Continent.

⁶ The latest known printing of the Geographia is the reprint of E.L. Stevenson's Geography of Claudius Ptolemy (in English) in 1991. The maps of codex latinus VF.32, Naples, Cosmography, Maps from Ptolemy's Geography were recently (1990) reproduced by Magna Books, Leicester.

⁷ FIG. 1, is reproduced (with kind permission of the authors) from B. Jones and D. Mattingly, An Atlas of Roman Britain (1990), 19 , also see Ordnance Survey, Map of Roman Britain (4th edn, 1978), 15 and (3rdedn, 1956), 20.

⁸ William Camden, Britannia (1599 etc.), showed early interest in Ptolemy's Britain, and Mercator had published the Ptolemy map of 1578. It was John Horsley, Britannia Romana (1733), who first attempted to ‘rectify’ this map and to seriously analyse Ptolemy's place-names of Britain. W. Roy, Military Antiquities of the Romans in Britain (1793) , also considered Ptolemy's Geographia based on Mercator's map but based his place-name identities on the work of the discredited Richard of Cirencester (1747/8) so was badly misled. More recent work on Ptolemy's Britain is reviewed in Table 1 etc.

⁹ For comprehensive consideration of place-names derived from many sources including Ptolemy's Geographia refer to A.L.F. Rivet and C. Smith, The Place-Names of Roman Britain (1979).

¹⁰ Strang, op. cit. (note 1), includes reappraisal of the Roman Lower Danube region and Ptolemy's geographic methods.

¹¹ Strang, op. cit. (note 1), 59-61.

¹² See note 4, codex Vaticanus graecus 177 (thirteenth/fourteenth-century) is also the first to record, in a colophon, Agathodaimon's claim to be ‘a technician of Alexandria, who drew the whole world from the Geographia of Ptolemy’.

¹³ Eratosthenes (late third/early second century B.C.) had estimated the 360 degree earth's size to be represented by 700 stades/degree (Strang, op. cit. (note 1), 46, 185), whereas the actual size is 600 stades/degree (= 75 Rm./degree) (idem, 46). Unfortunately Ptolemy followed Poseidonios, Strabo and Marinus with a value of 500 stades/degree (=62.5 Rm./degree) making his world one sixth too small in size or effectively 31 per cent reduced in surface area. Ptolemy also adopted a 63°N latitude limit (Geographia 1.8) imposing a further restriction on available surface area.

¹⁴ Both of the first two maps in note 7 suffer from this defect which can encourage unrealistic distance measurements being made within distorted areas of the maps.

¹⁵ Ordnance Survey, Map of Roman Britain (3rd edn, 1956), modified.

¹⁶ Rivet and Smith, op. cit. (note 9), 150-80; Cuntz, O., Itineraha Romana I (1929); Astley, H.J. Dunkinfield, ‘Notes on the Ninth Iter of Antoninus’, Norfolk & Norwich Arch. Soc. xvii (n.d.), 1–30Google Scholar; Reed, N., ‘Pattern and purpose in the Antonine Itinerary’, American Journ. Phil, xcix (1978), 228–54CrossRefGoogle Scholar; Rivet, A.L.F. and Jackson, K., Britannia i (1970), 34–82CrossRefGoogle Scholar; Rodwell, W., Britannia vi (1975), 76–101CrossRefGoogle Scholar, and Jones and Mattingly, op. cit. (note 7), 23-9.

¹⁷ Rivet and Smith, op. cit. (note 9), 149-50 ; K. Miller, Die Peutingersche Tafel (1916).

¹⁸ Rivet and Smith, op. cit. (note 9), 216-25 ; O. Seeck, Notitia Dignitatum (1876) ; Hassall, M.W.C., ‘Britain in the Notitia’, in Goodburn, R. and Bartholomew, P. (eds), Aspects of the Notitia Dignitatum, BAR Int. Ser. 15 (1976), 103–17Google Scholar ; Ward, J. Hester, ‘The British sections of the Notitia Dignitatum’, PSAS iv (1973), 253–63Google Scholar ; Petrikovits, H. von, ‘More problems with the Notitia Dignitatum’, Britannia xi (1980), 423–7CrossRefGoogle Scholar ; Jones and Mattingly, op. cit. (note 7), 33-7.

¹⁹ Rivet and Smith, op. cit. (note 9), 185-215 ; M. Pinder and G. Parthey, Ravennatis Anonymi Cosmographia et Guidionis Geographica (1860) ; Richmond, I.A. and Crawford, O.G.S., ‘The British section of the Ravenna Cosmography’, Archaeologia xciii (1949), 1–50CrossRefGoogle Scholar ; Dillemann, L., Archaeologia cvi (1978), 61–73Google Scholar ; Jones and Mattingly, op. cit. (note 7), 29-33.

²⁰ Rivet and Smith, op. cit. (note 9), 73-4 ; C. Müller, Geographi Graeci Minores (1855).

²¹ See L. Bagrow, History of Cartography (1964, enlarged by R. Skleton), pl. X etc. Strang (op. cit. (note 1), 67; 264) indicates that this map, drawn to a rectilinear longitude-ratio of 0.666-0.669 (c. 41.7 Rm. per degree), is significant. Britain, when drawn separately from Europe, was restricted in configuration on Ptolemy's smaller world. Hence England might well need a ratio of this order for compatability whereas Scotland required a much smaller ratio for realism when projected in a conic or homeotheric world context.

²² Horsley, op. cit. (note 8), 361 : ‘If a degree of longitude in any part of Britain be according to Ptolemy, 40 miles (as some affirm) it must be in the south of England where the latitude is least. Nor must we here allow them the usual length of the English computed miles. A degree of latitude, or a degree of the great circle, seems to me, according to Ptolemy, to be near enough our usual reckoning 60 computed miles’ (i.e. probably c. 65 Rm. per degree latitude and c. 43.7 Rm. per degree longitude).

²³ Bradley, H., ‘Ptolemy's Geography of the British Isles’, Archaeologia xlviii (1885), 378–96.Google Scholar

²⁴ See Strang, op. cit. (note 1), 20 ; 63-4 for analysis of Ptolemy world map projections.

²⁵ T.G. Rylands, The Geography of Ptolemy Elucidated (1893). Small and full-sized worlds should each revolve 360 degrees in 24 hours (i.e. 15 degrees per hour). However to cover the same longitudinal distance on the small Ptolemy world it would need to rotate 18 degrees per hour. Rylands failed to reconcile these conflicting elements in his suggestion of faulty celestial observations.

²⁶ Tierney, J.J., ‘Ptolemy's map of Scotland’, JHS lxxix (1959), 132–48.CrossRefGoogle Scholar Tierney recognised that Ptolemy's coordinates were mainly based on land/sea measurements and directions. He discounted reliable astronomical observations and accepted that general location error could be up to 30 minutes of arc.

²⁷ Richmond, I.A., ‘Ptolemaic Scotland’, PSAS lvi (1922), 288–301.Google Scholar

²⁸ idem, 288.1 concur wholeheartedly with Richmond and believe that the vital clue has now been found.

²⁹ The ratio 0.275 gives a longitudinal scale of 17.2 Rm./degree (cf. ultimately 25.8 Rm. per degree) for Scotland which is an over-reduction but an improvement on the often accepted value of 34.4 Rm. per degree (ratio 0.55).

³⁰ Ordnance Survey, op. cit. (note 15), 20.

³¹ Rivet, A.L.F., ‘Some aspects of Ptolemy's geography of Britain’, in Chevalier, R. (ed.), Littérature gréco-romaine et géographie historique (1974), 56.Google Scholar This is the first time a linear scale was added to the severely distorted Ptolemy map of Britain.

³² Rivet and Smith, op. cit. (note 9), 112-14.

³³ Similarly, J.C. Mann's view, PSAS cxx (1990), 62Google Scholar that, ‘internally the measurements are correct, and Scotland in Ptolemy's map can be put right by a simple 90° turn to the left’, is not considered a serious comment.

³⁴ We must avoid the north of England as it has been twisted to be directly north of London and several grossly distorted or elongated regions (e.g. East Anglia, Land's End area) and some ‘places’ are obviously out of position.

³⁵ Strang, op. cit. (note 1), 76-8 , for selection of London as preferred, preliminary datum for map comparison.

³⁶ idem, 79-80, for the rather limited selection permissible.

³⁷ Initially the two grids (at 41.25 and 25 Rm. per degree) are matched at Ptolemy's 20° (London) longitude line but with later refinement at 21 ° with refined (final) scales of 41.67 and 25.8 Rm. per degree respectively. By inspection and trial testing it is possible to configure the grid relationship between England and Scotland.

³⁸ See Strang, op. cit. (note 1), 55-6 and 76-8, for convenient comparison between the Ptolemy world and the real world. Previous methods of ma p comparison/analysis are also discussed and the conclusion reached that for Ptolemy's highly distorted map, superimposition would be most effective using Ptolemy isopleths only, so that distances represented by the modern master ma p are indicated at their actual magnitude, in any direction. This is a novel approach and the reverse of applying modern isopleths to ancient maps (e.g. W.R. Tobler, ‘Medieval distortions; the projections of ancient maps’, Annals of the Association of American Geographers (1966), 365 ) which only emphasises the variability of scale of the ancient ma p and does not give true linear scale between isopleths.

³⁹ Strang, op. cit. (note 1), 76-8, 272.

⁴⁰ ‘Known’ places are chosen in England only, that are not obviously, drastically out of relative position or too close to the pivot point. For FIG. 6, Ituna (Solway/Eden), Vedra (R. Wear), Ganganorum prom. (Braich-y-Pwll), Maridunum (Carmarthen), and Tamarus fl. (R. Tamar) are for example shown connected to their real locations.

⁴¹ See Strang, op. cit. (note 1), Appendix 1, and later for definition and explanation of positional error associated with Ptolemy places.

⁴² Pytheas’ Thule is well argued by C.F.C. Hawkes, Pytheas; Europe and the Greek Explorers (1975), 35 and R. Selkirk, The Piecebridge Formula (1983), 173-4 , to identify with Iceland whereas the Thule of Agricola (Tacitus, Agricola 10.6) and of Ptolemy ( O.A.W. Dilke, Greek and Roman Maps (1985), 136 ) undoubtedly refers to Shetland.

⁴³ This would seem to support somewhat the basis, method, and assumptions of the analytical exercise.

⁴⁴ This sea crossing would have been an important (known) feature for early voyagers, unable to hug the coast (see Strabo, Geography IV.3.4, IV.5.1-5 and in Rivet and Smith, op. cit. (note 9), 91-2).

⁴⁵ C. Müller, Claudii Ptolemaei Geographia (1883).

⁴⁶ I suspect that Verubium prom. (Noss Head) and Virvedrum prom. (Duncansby Head) also followed, having previously moved I° east rather than ½° as shown.

⁴⁷ ‘This [Hibernia] lies above it, 30 miles from the tribe of the Silures by the shortest crossing’ (Pliny the Elder, Natural History IV. 102-4; written in the a.d. 70s, see Rivet and Smith, op. cit. (note 9), 80 and J.C. Mann and R.G. Penman, Lactor xi (1985), 14).

⁴⁸ The apparent deletion of Cape Wrath prevents its interposition between Orcades and Orcasprom.

⁴⁹ Strang, op. cit. (note 1), 91.

⁵⁰ idem, 275.

⁵¹ idem, 92-4; Appendix 7, Table A.

⁵² idem, 93, Appendix 8 and fig. M.

⁵³ During the analysis some bias was experienced, necessitating the original datum (London) being adjusted by 1/32° east and 1/16° north.

⁵⁴ The birectangular scale-match was optimised at 21°E rather than the initial choice of 20°E.

⁵⁵ Strang, op. cit. (note 1), 93-4, Table 10.

⁵⁶ idem, Appendix 8, Table B.

⁵⁷ idem, Appendix 9.

⁵⁸ idem, 152-5, this is particularly evident in Ptolemy's Map IX of Europe (lower Danube region of the Roman Empire).

⁵⁹ idem, 118-20, 293-4.

⁶⁰ Curle, A. ‘Description of the fortifications on Ruberslaw etc.’, PSAS xxxix (1904-1905), 225–6.Google Scholar

⁶¹ Rivet and Smith, op. cit. (note 9), 484.

⁶² J.K.S. St Joseph in B.R. Hartley and J.S. Wacher (eds), Rome and Her Northern Provinces (1983), 222-34.

⁶³ Rivet and Smith, op. cit. (note 9), 390.

⁶⁴ Maxwell, G.S., Ptolemy and the Map of Roman Britain, lecture at University of Leeds, 4 February 1994.Google Scholar

⁶⁵ Robertson, A.S., ‘Roman coins found in Scotland, 1971-82’, PSAS cxiii (1983), 421.Google Scholar

⁶⁶ Rivet and Smith, op. cit. (note 9), 316-17.

⁶⁷ idem, 301.

⁶⁸ Richmond, I.A., ‘A new Roman mountain road in Dumfriesshire and Roxburgshire’, PSAS lxxx–lxxxi (1946), 103–17.Google Scholar

⁶⁹ Graham, A., ‘The Roman road to Raeburnfoot’, PSAS lxxxii (1948), 231–4.Google Scholar

⁷⁰ Strang, op. cit. (note 1), 124-5, 295.

⁷¹ This would certainly not be the capital place, civitas of a tribe (e.g. this would have been Traprain Law for the Votadini and possibly Eildon Hill North for the Selgovae).

⁷² Rivet and Smith, op. cit. (note 9), 499.

⁷³ R.M. Ogilvie and I.A. Richmond, Cornelii Taciti de vita Agricolae (1967), 243-4.

⁷⁴ J.D. Leach and J.J. Wilkes, ‘The Roman military base at Carpow’, in Limes, Akten des XI Limeskongresses (1977), 47-61.

⁷⁵ Rivet and Smith, op. cit. (note 9), 373.

⁷⁶ idem, 465.

⁷⁷ Strang, op. cit. (note 1), Appendix 10.

⁷⁸ idem, 123 and Appendix 10 for relationship with Fl. Loxa, identified as the Findhorn and not the Lossie which has been the traditional preference. Also see Breeze, D.J., ‘Agricola in the Highlands?’, PSAS cxx (1990), 56–8Google Scholar; Macdonald, J. in Arch. Journ. (1891), 48, 361–95Google Scholar; Mann, J.C., Roman Northern Frontier Seminar, unpub. typescript, Dept. of Archaeology, University of Newcastle on Tyne (1970) 1, 14–17; and D.H. Sellars (ed.), Moray: Province and People (1993), 47-74Google Scholar.

⁷⁹ Derivation of the root Banna is given by Rivet and Smith, op. cit. (note 9), 261-2, and there seems to have been a tendency for meaning of ‘peak’ rather than ‘horn’ to be preferred. However, if we consider the pre-Hadrianic promontory site of Birdoswald (Banna, RIB 1905, Rudge Cup, Amiens patera, Ravenna Cosmography 107₂₈, and Hassall, op. cit. (note 18), 113), which was little more than a Roman signal station then (P. Howard, Birdoswald Fort on Hadrian's Wall (1969), 9, 21), it admirably suits the description of ‘horn or tongue’ from its situation within the sharp sweep of the Irthing river. Secondly if we now consider Horncastle (Bannovallum, Ravenna io6₅₅) described by Rivet and Smith, op. cit. (note 9), 265, as on the ‘spur’ at the junction of the river Bain and another, this fits a similar interpretation. Similarly, Whilton Lodge (Bannaventa, Antonine Itinerary 4705) is described by Dix, B. and Taylor, S., ‘Excavations at Bannaventa 1970-1’, Britannia xix (1988), 299CrossRefGoogle Scholar as ‘market on the spur’. Ptolemy's Bannatia seems to coincide with the Roman site at Cardean and G.S. Maxwell in The Romans in Scotland (1989), 109, describes the Cardean fort as ‘at the confluence of the Isla and the Dean’.

⁸⁰ Strang, op. cit. (note 1), 120; Appendix 10.

⁸¹ idem, 224, also see RIB 899 and Hassall, op. cit. (note 18), 111.

⁸² There is no justification for removing Ptolemy's Salinae to a Droitwich identity as in Rivet and Smith, op. cit. (note 9), 120, 451.

⁸³ G.H. Orpen, PRIA xxiv (1894), 115-28 and PRIA xxxii, C3 (1913), 41–57.Google Scholar

⁸⁴ Strang, op. cit. (note 1), Appendix 7.

⁸⁵ N. Higham and B. Jones, The Carvetii (1985).

⁸⁶ B. Hartley and L. Fitts, The Brigantes (1988).

⁸⁷ Rivet and Smith, op. cit. (note 9), 437.

⁸⁸ H. Ramm, The Parisi (1978).

⁸⁹ K. Branigan, The Catuvellauni (1987).

⁹⁰ M. Todd, The Coritani (1973).

⁹¹ Hassall, M.W.C. and Tomlin, R.S.O., ‘Thorpe in the Glebe. Lead sealing’, Britannia xiv (1993), 318.Google Scholar

⁹² Tomlin, R.S.O. ‘Non Coritani sed Corieltauvi’, Antiq. Journ. lxiii (1983), 353–5.CrossRefGoogle Scholar

⁹³ A. Detsicas, The Cantiaci (1983).

⁹⁴ B. Cunliffe, The Regni (1973).

⁹⁵ R. Dunnett, The Trinovantes (1975).

⁹⁶ G. Webster, The Cornovii (1975), 7.

⁹⁷ Rivet and Smith, op. cit. (note 9), 121.

⁹⁸ To be compatible with W.J. Watson's The History of the Celtic Place Names of Scotland (1926), 17 , which suggested relationship to Ca m Smeart, the Smertae would have to be encircled by the Lugi. Also it is probably more reasonable to have named a burial place in a n enemy's territory in any case.

⁹⁹ ibid., 21.

¹⁰⁰ RIB 2091, see also Woolliscroft, D.J., ‘The outpost system of Hadrian's Wall’, Brit. Archaeology vi (1988), 25.Google Scholar

¹⁰¹ Strang, op. cit. (note 1), 105-11.

¹⁰² idem, 110-17, 290-2.

¹⁰³ Richmond, I.A. and Steer, K.A., ‘Castellum Veluniate and civilians on a Roman frontier’, PSAS xc (1956), 1–6.Google Scholar

¹⁰⁴ Strang, op. cit. (note 1), 127-30, 297.

¹⁰⁵ Jarrett, M.G., ‘Legio II Augusta in Britain’, Arch. Camb. cxiii (1964), 52Google Scholar ; P.T. Bidwell, Roman Exeter; Fortress and Town (1980), 56 ; and R.J.A. Wilson, A Guide to the Roman Remains in Britain (1988), 190.

¹⁰⁶ A.R. Birley in R.M. Butler (ed.), Soldier and Civilian in Roman Yorkshire (1971), 82 and Wilson, op. cit. (note 105), 227.

¹⁰⁷ Military diploma July a.d. 112, D.J. Breeze and B. Dobson, Hadrian's Wall (3d edn, 1987), 62-4 and Greenstock, M.C. (ed.), Lactor iv (1971), 32.Google Scholar

¹⁰⁸ A. Strang, ‘Ptolemy's Geography of the Lower Danube region’, JRA (1998, forthcoming).

¹⁰⁹ Hind, J.G.F., ‘The Romano-British name for Corbridge’, Britannia v (1980), 165–71CrossRefGoogle Scholar; A.K. Bowman, Life and Letters on the Roman Frontier (1994), 22.