1. The ‘Note on Celtic origins’ on this page explains that, whatever later interpretations the name acquired, the ‘Celts’ (with a hard ‘c’) were not an ethnic group but the inhabitants of Iron Age Europe who shared certain cultural traits and who were variously blond, dark-haired, red-haired, curly-haired, tall or short, pale or swarthy, belligerent or pacific. A map of the greatest extent of Celtic lands (sometimes referred to as ‘Keltika’) is on this page. The same map depicts the multiple origins of the Celts according to a Celtic legend.
2. For precise details, see here.
3. ‘Gaulish’ was usually synonymous with ‘Celtic’ but was also applied – as it is today and in this book – to the inhabitants of the country known as Gaul (roughly coterminous with France). The Gauls spoke a Celtic language known as Gaulish, which was a branch of Continental Celtic.
4. ‘Oppidum’ is the word used by Julius Caesar to refer to the towns of the Gauls. It is now the usual term for the fortified settlements or hill forts of the Celts, which ranged in size from about fifteen to several hundred hectares (see here).
5. The places mentioned in this chapter are shown on the map on this page.
6. This is the Irish word; the ancient Gaulish equivalent has yet to be rediscovered.
7. See here and here, and ‘A Traveller’s Guide to Middle Earth’, here.
8. Since the angle changes with latitude, a local standard would have been chosen in order to produce a straight rather than a curved line. (See here.) Even on a short stretch, standardization would have been necessary since the exact point of sunrise or sunset cannot be observed directly: the sun’s light is refracted by the earth’s atmosphere, and the horizon is almost never flat and unobstructed. In 600 BC, the angle was 0.5° less than it is today. The angle in question is the solar azimuth angle (the number of degrees, measured clockwise from north, to the point on the horizon where the sun rises).
9. The course of the Via Domitia is shown in fig. 46.
10. The complete list, with coordinates, can be found at www.panmacmillan.com/theancientpaths.
11. The Greek and, probably, Gaulish, form was ‘Mediolanon’. The better-known Latin form, ‘Mediolanum’, is more commonly used. The plural is ‘Mediolana’. ‘Lanon’ is unrelated to the Welsh ‘llan’ and Breton ‘lann’ (an enclosed piece of land, especially the site of a monastery or church). The nearest Welsh equivalent is ‘llawn’ (‘full’ or ‘complete’).
12. An additional problem is that the names on maps can be just as deceptive as the maps themselves. Toponymy is like a dream in which familiar faces turn out to belong to strangers. The city of Milan owes its name to ‘Mediolanum’, but most of the other places called Milan have an entirely different origin. A village that was called ‘Moulins’ in the eighteenth century (now Molain) is mentioned in an eleventh-century document as ‘Villa Mediolanis’, but all the other ‘Moulins’ are mills. The etymological record shows that, of the thirty-nine ‘middle sanctuaries’ included by Vadé in his calculations, sixteen were never Mediolana. One of the key points in his system is Meilhan-sur-Garonne, which, though it looks like a name derived from ‘Mediolanum’, was probably the estate of a Roman citizen called Aemilianus.
13. http://confluence.org/
14. Nanterre, Nemours, Nîmes, Nempnett Thrubwell, etc. See also here and here and, for a longer list, www.panmacmillan.com/theancientpaths.
15. Cartographic triangulation consists of taking three points, measuring the angles of the resulting triangle, and the exact distance on the ground between two of the points. The lengths of the other two sides are provided by basic trigonometry. Another example of a Roman survey line (Watling Street) is shown in fig. 72.
16. See Chapter 8 and the map of Celtic sun-horses, here.
17. From Augusta Suessionum (Soissons) in the east, the Peutinger Map shows a logical set of distances between staging posts, but it ceases to be logical if Samarobriva is Amiens. From Rodium to Setucis, the distance is ten leagues (twenty-seven kilometres in a straight line). From Setucis to Samarobriva, the distance is also ten leagues (twenty-seven kilometres to the Parc Samara site, but only fifteen to Amiens).
18. A Metonic calendar covered a cycle of lunar phases, which lasts about nineteen years. It has been suggested that a Druid education lasted twenty years so that a complete set of astronomical observations of the Metonic cycle could be collected.
19. On rhumb lines or lines of constant bearing, see here and here.
20. The same presumption of illiteracy is often extended to Celtic society as a whole, though Diodorus Siculus reported that the Celts ‘cast letters to their relatives onto funeral pyres in the belief that the dead will be able to read them’. See also here.
21. Pythagoras taught in the sixth century BC. ‘Pythagorean’ is the interpretation of classical, non-Celtic writers, who saw similarities between Druidism and Hellenic philosophies, just as later writers perceived analogies with Brahmanism. These filiations are neither implausible nor provable.
22. The four Celtic festivals occur on the ‘cross-quarter days’, halfway between solstice and equinox: Samhain – winter solstice – Imbolc – vernal equinox – Beltane – summer solstice – Lughnasadh – autumn equinox.
23. This may have been a function of female Druids: Plutarch and Polyaenus attributed the power to stop battles to ‘Celtic women’.
24. On the location of this ‘consecrated place’ mentioned by Caesar, see here.
25. These and other lines are illustrated on the large map on this page.
26. If so, the line would be slightly closer to Châteaumeillant than Châteaumeillant is to Belvianes: since klimata or lines of latitude are based on length of day rather than distance on the ground, the lines become increasingly crowded as they approach the poles, which is one reason why the theory of equidistant Mediolana would never have revealed a pattern.
Except where stated, the assumed tolerance is less than ten seconds of day length (even measuring the length of day to the nearest minute would have been a remarkable achievement). The bearings are precisely accurate. Far less rigorous criteria are normally applied to the tangents of Roman roads and to the orientation of towns and temples.
27. Two further coincidences on this Aquitanian and Armorican line, apart from those mentioned later (here): 1. The point south of Alesia from which the line is projected lies just outside the village of Bezouce (Gard), which is thought to have been the capital of a tribe, the Budenicenses, whose existence is attested only by two inscriptions. 2. The point of intersection with the Alesia line lies a few hundred metres from the castle of La Rochefoucauld – a possible oppidum site and the home, in the tenth century AD, of a man said to be a son of the Fairy Mélusine, who was also present at Châteaumeillant. (But Mélusine was an exceptionally well-travelled fairy.)
28. See the note on portolan charts, here.
29. The date set by the Helvetii for their migration was around the time of the spring equinox, when the sun sets due west – the direction in which they intended to travel. Heading west from their assembly point, Geneva, they would indeed have reached the lands of the Santones, just to the north of La Rochelle.
30. Tarquinius Priscus was on the throne at the time when Massalia was founded (c. 600 BC), but the legend conflates different periods. The Etruscan city of Melpum, on the site of the future Mediolanum (Milan), was destroyed in 396 BC, and the Celts entered Rome in 387 BC. The Biturigan hegemony probably dates from the fourth century BC.
31. ‘Tricastinis redditi’: (lands) restored to the Tricastini.
32. For these and the following sites, see the large map on this page.
33. Les Courens (Beaumes-de-Venise), Saint-Christophe (Lafare), le Clairier (Malaucène).
34. Five other tribes can be added to the partial lists supplied by Livy. The Salassi from Aosta settled at Eporedia (Ivrea). The Insubres or Insubri are named as settlers by Polybius: the first part of ‘Insubri’ means ‘path’ or ‘direction’; the second suggests an offshoot of the Uberi who lived at the source of the Rhone and the Pillar of the Sun on the equinoctial line from Mediolanum Biturigum. The Mediomatrici gave their name to Mezzomerico. The Allobroges are mentioned in this connection by Geoffrey of Monmouth. The Veneti were said by Claudius to have invaded Rome with the Insubres.
35. ‘They profess to know the size and shape of the earth and the universe, the motion of the sky and the stars, and what the gods want’ (here).
36. The bearing from Metelen to Wolkersdorf is within 0.15° of the Gaulish standard (122.32°). The bearing from Wolkersdorf to Ruse, despite the great distance, is accurate to within one point of a 128-point compass (124.74°). Medieval portolan charts used a 32-point compass.
37. Within one point of a 128-point compass (124.62°). See here.
38. The name is recorded in AD 899 and probably means ‘boundary hill fort’. Its modern name is Ensérune.
39. Ariège,* Aube, Creuse, Drôme, Durance, Duria (Dora), Garonne,* Indre, Marne, Meuse,* Orne, Rhine, Rhone, Somme, Yonne. (The asterisk indicates the traditional rather than hydrographic source.)
40. Only one line survives of Varro’s Bellum Sequanicum: ‘Deinde ubi pellicuit dulcis levis unda saporis . . . ’ (‘Then, when the smooth swell of a sweet taste enticed . . . ’). This may be an allusion to the wine-inspired migrations of the Celts.
41. The unnamed place may have been somewhere in the Thuringian Forest (no. 3 in fig. 43) between the Elbe and the Rhine.
42. This was the so-called battle of the Sambre. Pierre Turquin showed conclusively in 1955 that the river ‘Sabis’ of Caesar’s text is not the Sambre but the Selle, and that the battle took place on the site of the village of Saulzoir (Nord). This is now corroborated by the Druidic system: Saulzoir lies precisely on the northern solstice line (fig. 50).
43. See the map of the Gaulish confederation on this page.
44. ‘Uxellodunum’ survives in Exoudun, Issudel, Issoudun and a few other place names. Most have probably vanished. The Uxelodunum (sic) on Hadrian’s Wall above Carlisle is now Stanwix (‘Stone Way’).
45. The Pictavi or Pictones of Lemonum (Poitiers).
46. The likeliest location of ‘Mediomanum’, which the Ravenna Cosmography (c. AD 700) places between Levobrinta (Forden Gaer) and Seguntio (Caernarfon), is Tomen y Mur, a Roman fort complex and legendary palace of Lleu (Lugh). The Cosmography also lists a ‘Mediobogdum’, once thought to have been the Roman fort at Hardknott Pass, but now identified with the fort near Kendal in the ‘middle of a bend’ in the river Kent, four kilometres west of the meridian. The ‘medio’ in this case may be purely geographical.
47. Rubers Law, Trimontium, Traprain Law and the Caterthuns, on which, see here.
48. The mid-point of the Antonine Wall survey line is Medionemeton. The site most likely to have performed the same function on Hadrian’s Wall is the halfway house of Vindolanda. The name is also recorded as Vindolana, which might be translated as ‘white church’ or Whitchurch.
49. Few of the other traditional candidates fit Tacitus’s description: Caer Caradoc near Church Stretton (a negligible river, and in Cornovian rather than Ordovican territory; the name ‘Caradoc’ is a corruption of ‘Cordokes’); Caer Caradoc near Clun (a negligible river); Llanymynech (where advance and retreat would have been relatively easy); the Herefordshire Beacon or British Camp (no river). The most likely sites, apart from Dinas Emrys, are Breidden Hill and Cefn Carnedd, both above the river Severn.
50. ‘Oxubii’ (‘People of the Ox’) was the name of a Mediterranean Gaulish tribe in the region of Fréjus.
51. On the transferral of Salisbury Cathedral from Sorviodunum (Old Sarum) to its current site in 1220, see here.
52. Ermine Street, which ran approximately north from London, shows no obviously significant bearing, whereas Ermin Street (north-west from Silchester) follows the British solstice bearing. Medieval scholars chose what they believed to be the more important ‘Ermyngestrete’, though in some versions of the legend (e.g. Robert of Gloucester’s thirteenth-century Chronicle), ‘Eningestret’ is paired with ‘Ikenildestrete’, as it is on the map above: ‘Lyne me clepeth eke thulke wey, he goth thorgh Glouceter, / And thorgh Circetre [Cirencester] euene also’.
53. Few of the other traditional candidates fit Tacitus’s description: Caer Caradoc near Church Stretton (a negligible river, and in Cornovian rather than Ordovican territory; the name ‘Caradoc’ is a corruption of ‘Cordokes’); Caer Caradoc near Clun (a negligible river); Llanymynech (where advance and retreat would have been relatively easy); the Herefordshire Beacon or British Camp (no river). The most likely sites, apart from Dinas Emrys, are Breidden Hill and Cefn Carnedd, both above the river Severn.
54. This was the name of a Romano-British chieftain. The later version, by Geoffrey of Monmouth, identifies the boy as the mathematician, astronomer, bard and prophet, Myrddin or Merlin.
55. The participle, ‘vastatum’ (‘ravaged’), may indicate a deliberate depopulation of tribal territories.
56. The battle may have been fought somewhere in the Grampian Mountains, but the name ‘Grampian’ is a false clue: it was first applied to the mountains of the southern Highlands by a sixteenth-century Scottish historian whose Latin text misspelled ‘Graupius’ with an ‘m’.
57. By an inscrutable coincidence, this line of latitude passes close to a small white house on the outskirts of Pitlochry at the foot of the Pass of Killiecrankie called Tigh na Geat (formerly, Taigh nan Teud, or ‘House of the Harpstring’). The house was once said to mark the southerly limit of the Lords of the Isles’ influence and the exact centre of Scotland.
58. Caereni, ‘Shepherds’; Caledones, ‘People of the Rugged Fortress’; Carnonacae, ‘Hornèd Ones’; Cornavii, ‘Sailors’; Damnonii, ‘People of the Lower World’ or ‘Keepers’ or ‘Magistrates’; etc. The tribes are listed in Ptolemy’s Geography (second century AD), probably from information supplied by Agricola’s officers.
59. ‘Pap’, meaning ‘nipple’ or ‘breast’, was a common name for a hill. ‘Maiden’ is incorrectly assumed to refer to the mountain’s fancied resemblance to a virgin’s breast.
60. Names derived from ‘briga’ (‘high place’) are common in the Celtic world, especially in Iberia. There is no known connection between this tribe and the Brigantes of Britain.
61. Deuteronomy, 32:10 (the phrase was usually quoted from the Vulgate). King James version: ‘In a desert land, and in the waste howling wilderness’.
62. Nemthor or Nemthur, which is probably Dumbarton (one of several possible birthplaces).
63. www.heritagegateway.org.uk/gateway/advanced_search.aspx
65. For example: www.movable-type.co.uk/scripts/latlong.html
66. In gpsvisualizer.com, the data should be preceded by the following header row: ‘name,latitude,longitude’.