In this section Ptolemy comments upon the practical pre–requisites required before embarking on map making, the sectional layout he has adopted for describing the habitable world, the sequence in which the co–ordinates are presented for a particular region and the method in which they are to be applied to the actual map surface itself. As it will be seen, this last direction is the source of much confusion.

§1 emphasises that the subject is map making and its practical implication in the representation of the known world in easily assimilated images, so that locations appear in true relationships to each other. In §2 Ptolemy indicates that the co–ordinates he lists for well known locations are trustworthy because they have been known for a long time. For other co–ordinates not supported by long time knowledge he suggests that they be regarded as relative to the nearest well-known location until proved otherwise. Thus Ptolemy is warning against the too ready acceptance of locations off the beaten track as being tentative and not to be relied upon. In the context of the rapidly expanding Roman Empire, particularly those parts acquired after Marinus wrote his commentaries, it would be best to treat them with caution. In §3, he notes that his presentation of the co–ordinates will show longitude before latitude against each location and that there will be space left alongside each for future corrections to be added.

Paragraphs §4–5 are most important if we are to comprehend how errors have been allowed to develop in the making of maps from Ptolemy's instructions. §4 appears to direct the actual insertion of the co–ordinates to be in a right–hand orientated manner, by which he apparently means that which would be adopted by a right–handed person. In §5, he therefore directs that the map projections should be completed, with their respective co–ordinates, starting from the north, for latitudes, and from the west, for longitudes. Presumably so that the writing hand shall not obscure what is being, and what has been, written. At first glance this is an innocuous direction aimed at pure convenience. If, however, there is a fundamental difference in the distance value of one degree of latitude and longitude, such as that between Posidonius and Eratosthenes, a too slavish adherence to this direction can have catastrophic consequences when new, unproven data is to be added. Consider the scope of the rectilinear graticules that Ptolemy directs us to create for Europe.

Map No. Country Longitude Latitude
2.SPAIN 1°W47°N

The scope of the individual co–ordinates given for each regional map is designed to cover the extremes of latitude and longitude covered by the region in question. Thus for the region of Hibernia and Albion, the longitudinal co–ordinate for the extreme west of Hibernia, that of the Southern Promontory, is given, in the commentaries, as 7° 40˘, while the longitudinal co–ordinate for the extreme east of Albion, that of the Varubium Promontory, is given in the commentaries, (incorrectly as it happens), as 31° 30˘. Similarly the northernmost latitude, that of Thule, is given as 63° 15˘ while the southernmost latitude, that of the Ocrium Promontory, is given as 51° 30˘. Thus Map 1 of Europe embraces a projection range of 7° to 32° longitude and 63° to 51° latitude.

Accepting that the well known locations, particularly for the map of Albion, are spatially correct and are all calculated on the same geodesic time base used by Marinus, then there can be no argument that their positions in the rectilinear graticule will also be correct. However, if subsequent to their original calculation, another set of locations were added based on a different geodesic time base, in the belief that the rectilinear graticule is on the same time base, then serious errors of location will occur. That this may have happened for all the locations north of a specific point in northern Britain, say somewhere within the territory of the Brigantes, must be a subject for serious consideration and will be dealt with later in the commentary.

§6–7 are concerned with the separation of the regions into Europe, Africa, Asia and beyond. §8–9 indicate the desirability of drawing in the boundaries of regions and provinces and noting unusual features, pointing out, at the same time, the opportunity afforded in devoting a whole map to one region or province complete with notes on the peoples therein. §10 allows that, should we wish, there is no reason why both the lines of longitude and latitude should not be drawn straight, providing the positioning of locations allows for this spatially. §11 declares that the instruction is now complete and the actual practice of map making can now begin.




This section of Book II begins a sequence listing all the principal geographical locations of Ptolemy's time and their navigational co–ordinates, which continues up to and including Book VII. It is quite evident that many, if not all, of the co–ordinates representing these locations are the result of calculation of distances rather than observational readings. That is to say the locations, having been plotted on a map surface using distance measurement and trigonometric calculation, are then read off as a series of co–ordinates. In normal circumstances, where the distance allocated to a degree of latitude and longitude is fixed and inviolate, this would be quite correct. However, where there might be an uncertainty as to which particular reading is obtained from which particular geodesic coefficient. Eratosthenes or Posidonius, then one is on very uncertain ground indeed. In this context then, all of these locations fall naturally into five categories and for each one there is a need to makes some assumptions before considering what Ptolemy has to say about them. Firstly there are those locations untouched by Roman occupation and colonization and therefore very much as Marinus might have found them. Hibernia is a case in point. Secondly those locations partly occupied and colonized by Rome for whom Ptolemy presents us with the base that Marinus provided upon which he admits to making his own annotations. Britannia is a case in point. Thirdly there are those longstanding provinces of Rome which would have contained mostly Romanized settlements built over older locations and with which Marinus would have been familiar. Africa, Spain Gaul and, of course, the hinterland of Italy itself are cases in point. Fourthly there are those locations who were initially occupied by the Roman army, who then subsequently withdrew; Greater Germania is a case in point. Fifthly there are those locations where Roman expansion occurred subsequent to Marinus and before Ptolemy; Scotland being a case in point. For the first category we should accept that the readings are reasonably homogenous and are a pure distillation from the sources of Marinus but mostly without means of verification. For the third category we should assume that all locations that pre–date Marinus have been unchanged by Ptolemy, unless specific mention is made in the text of Book I. (There are certain readings that Ptolemy takes exception to!). For the second and fourth categories we cannot accept, with certainty, any location as belonging to one time base or the other and all must be evaluated on an individual basis.

Further, before commencing the commentary on Ptolemy's given co–ordinates, whether of his own calculation or taken unchanged from the commentaries of Marinus, it is an appropriate point to consider them in the light of their modern day equivalents. The co–ordinates quoted in the Geographia are generally to the nearest five minutes and this argues a certain exactness of approach on the part of Ptolemy, or Marinus, since it is to the order of plus or minus an average of six Roman miles. However, as regards to his longitudinal co–ordinates, these cannot be directly compared with present day readings since these are taken as degrees east or west of the Greenwich Mean of 0° longitude. Ptolemy's readings are taken from a mean representing the farthest point west of the known, habitable world of his day. This is taken to be the location of the westernmost point of the Fortunate Isles and would have been located on Hierro Island in the Canary group. The present day longitude assigned to Hierro is a few minutes under 18° so that a simple reverse reading should give us a reading with which to make a direct comparison with Ptolemy. For instance, later in this text, Ptolemy gives a longitudinal reading of 20° 0˘ for Londinium whereas present day reading, with London almost exactly on the meridian of zero degrees, a reverse reading from Hierro would give a reading of just under 18°. This would give us a discrepancy of just under +11% between Ptolemy and ourselves.

In respect of his latitudinal co–ordinates, there is of course no difference in notation between Ptolemy and ourselves today, except that we use the suffix N (North) or S (South), to distinguish which hemisphere we are operating in, taking the equator as our base line in each case. However although Ptolemy's system numbers from the equator, for all intents and purposes he admits that most of his measurements are made relative to the parallel that passes through the island of Rhodes, which bisected the known, habitable world of his day. In the actual drawing of maps, however, he directs that measurement be made from the northernmost point of the known, habitable world of his day, which was Thule, to which he assigned the latitude 63° 15˘. Thule is, by a rather inconclusive argument, generally assigned to the present day Shetland Islands, although a description by both Pliny (Natural History IV, 102–104) and Solinus (Collectanea Rerun Memorabilium 22, 1–12), indicating that a frozen sea lies just to the north of Thule, might logically place it at the Faroe Islands. The mean point assigned to the latitude of Shetland today is 60° 30˘ while that of Faroe is 62° 0˘. Once again, taking Londinium for purposes of comparison, Ptolemy assigns a co–ordinate of 54° 0˘ while the present day co–ordinate is 51° 30˘. Bearing in mind Ptolemy's reverse computation directive, where distances are measured from the north and the co–ordinates computed accordingly, there would appear to be a net discrepancy, between Ptolemy and ourselves, of a little under +11%, if Thule is taken to be Faroe, or –3%, if Thule is taken as Shetland.

The geodesic time base of Eratosthenes provided a distance value for one degree, for both longitude and latitude, of about +11% on the value of that assigned today. While that of Posidonius provided a value that about –25%. As we progress through the list of Ptolemy's co–ordinates it will be a useful exercise to bear these equivalents very much in mind.

Section 2 itself is concerned with the island, as the Romans saw it, of Ireland which Ptolemy refers to by its ancient name of Hibernia rather than its Latin name, Britannia Parva. Hibernia was not a part of the Roman empire as such but was undoubtedly a place of trade to the Phoenician merchants and, presumably, to the Romans subsequently. However its geographical components were not altered by any Roman occupancy so that what Ptolemy reveals to us of its physical characteristics can be taken to be very much as it was known in the time of Marinus and for a long time previously. This allows us to examine an area of country untouched by any military incursion, with its accompanying environmental impact, between the time of Marinus and this work of Ptolemy.

With Hibernia, Ptolemy sets the pattern which he uses consistently throughout Books II to VII, when describing Roman provinces or non-Roman regions. Initially he describes the outline of the country concerned then fills in the internal detail with the names of ethnic occupants and their places of habitation, usually from north to south and in parallel corridors roughly one degree of longitude in width. The general shape of the country in question is therefore made immediately apparent at the outset and usually proceeds in an anti–clockwise manner from the north eastern extremity.

In the case of Hibernia Ptolemy describes this outline in §1–7 and in these paragraphs Ptolemy identifies the four cardinal extremities as,

NORTH WESTNORTHERN11° 0'61° 0'9°13'54°42
NORTH EASTROBOGDION16°30'61°30'11°4655°18'
SOUTH WESTSOUTHERN7°40'57°45'7°20'51°36'
SOUTH EASTIERON14°0'57°30'11°40'52°10'

thus confirming the familiar rhomboidal shape of Ireland. However, if the two rhomboids are drawn as diagrams, based on these cardinal points, it will be seen that, while the latitudinal distance, north to south, is reasonably comparable, in the longitudinal distance, east to west, Ptolemy's distance exceeds ours by about two thirds. It will be seen that while the two western co–ordinates are reasonably placed, for purposes of comparison, the two eastern co–ordinates are placed much too far to the east. The other noticeable feature is that while the latitudinal distances are comparable, Ptolemy's set of co–ordinates are all about six degrees north of their present day equivalents. This combination has the effect of placing the whole of Hibernia too far to the north and the east coast itself too far to the east. This may well be the reason that Ptolemy has associated the Outer and Inner Hebrides, the Isle of Man, Jura, Islay and the Mull of Kintyre with Ireland rather than with Britain. When we look at Ptolemy's description of the island of Albion in the next section, particularly the territory north of the line of Hadrian's Wall, further reasons for the anomaly will be discovered.

Paragraphs §1/2,4,6 and 8 describe the physical characteristics of the coasts of Hibernia, thus §1/2 describe the north coast from west to east and the locations described consist of three promontories and two river outflows. The two eastern of the three promontories can be reasonably identified with Rathlin rather than Fair Head and Malin Head on the north coast proper but the third promontory, lying 1° 30˘ to the west and 0° 40˘ to the south of Malin Head would bring us into the vicinity of Rossan Point on what is really the north–western coast of Donegal. The two river outflows could be logically assumed to be the major outlets of Lough Swilly and Lough Foyle, although the longitudinal co–ordinate of the second outflow would place it more in the vicinity of Ballycastle Bay.

§4 describes the west coast from north to south and the locations mentioned consist of two promontories, six river outflows and one town. If Rossan Point, as already mentioned, is taken to be the northern promontory, the associated river outflow might be that of Donegal Bay and the town itself in the Donegal/Bundoran area. The next five major outflows could be assigned, in sequence, to Westport and Galway bays, followed by the mouth of the Shannon river, Castlemaine Harbour and the Kenmare river mouth. The southern promontory could then be assigned to Dursey Head.

§6 describes the south coast from west to east and the locations mentioned consist of two promontories and two river outflows. The first, westerly located promontory being assigned to Dursey Head, the first river outflow would almost certainly be Cork Harbour since its longitudinal co–ordinate has the correct alignment with the Northern promontory. For the same reason the second river outflow can be reasonably assigned to Waterford Harbour since its longitudinal co–ordinate has a correct alignment with the Venicnium Promontory (Malin Head). The Ieron Promontory could then be assigned to Carnsore Point.

§8 describes the west coast from south to north and the locations mentioned consist of three promontories, five river outflows and two towns. The most southerly promontory having been assigned to Carnsore Point, the first river outflow cannot be Wexford Harbour, since the distance of the latitudinal co–ordinate that Ptolemy assigns must be in the vicinity of Dun Laoghaire, making the associated town almost certainly Dublin. The second river outflow could be assigned to the river Boyne while the third could be assigned to Belfast Lough and the associated town to Belfast itself. The second promontory could then be assigned to Black Head and the fourth river outflow to Larne Lough. The fifth river outflow could be assigned to Red bay and the final, northern promontory has already been assigned to either Rathlin or Fair Head

All of the suggested locations made for those noted in §1/2,4,6 and 8 are quite speculative; their only claim to any validity being a spatial relativity in accordance with Ptolemy's given co–ordinates. It is acknowledged that they are lacking in any other supporting evidence and that, in most cases, they conflict with the locations already assigned by historians. The only argument in their favour is that they agree with Ptolemy's narrative and assume that he has transmitted the Phoenician data correctly.

§3,5,7 and 9 are concerned with naming the ethnic groups or tribes that inhabit Hibernia and the associated location they occupy. As regards the north coast §3 indicates that the Vennicni occupy the north western corner and the Rhobogdi occupy the north eastern corner. §5 indicates that, as regards the west coast, the Erdini occupy the coast below the Vennicni, and below these are the Nagnatae. Then follow in succession the Autini, Gangani and below these latter are the Vellerbori. §7 indicates that, along the south coast there are three tribes; the Hiberni alongside the Vellerbori and above them the Usdiae and to towards the east, the Brigantes. §9 indicates the tribes associated with the east coast, of whom there are six. The Darini are the most northern, alongside the Rhobogni already mentioned. Below these are the Volunti followed closely by the Blani and the Cauci. Below these are the Manapi and the Coriundi, located above the previously mentioned Brigantes.

§11 lists the location of the dwelling places or towns and it is interesting to note that of the sixteen tribal groups noted there are only seven inland habitations and three coastal towns. The inland towns, and their possible tribal associations, being,

Regia 2Gangani

The three coastal towns, and their possible tribal associations, being,


Once again these assignments are quite speculative and are based on the spatial relationships arising from Ptolemy's co–ordinates

§11/12 are concerned with offshore islands that Ptolemy associates with Hibernia. However from the north–easterly co–ordinates given it is clear that all are more closely associated with the island of Albion than Hibernia. §11 is concerned with a group of five under the general heading of 'Ebuda' which would appear to be the Outer Hebrides, while further to the east another group, among which another island also called 'Ebuda', would appear to be the Inner Hebrides and finally 'Epidium', taken to be the Kintyre peninsula. §12 lists another group of islands not considered to be associated with one another, Monasida which could be Mull, two uninhabited islands, Edra and Limnos which could be Jura and Islay and lastly, Mona which would seem to the Isle of Man.

This completes Ptolemy's survey of Hibernia, a country with no reasonable identification with the Roman world. It must be accepted that the picture Ptolemy gives us is an ancient one, based on navigational and commercial data arising from Greek and Phoenician trading routes dating back centuries before Roman expansion began.




Section 3 is concerned with the island of Britain which Ptolemy refers to as Albion, rather than its Latin name, Britannia Magna. As in the case of Hibernia this is an anomaly that can only be explained if we accept that Ptolemy copied his description unchanged from the commentaries of Marinus. By the time of Ptolemy nomeclature throughout the empire had been fully romanized and this was in common use throughout. It is hard to accept that Ptolemy, commencing a new work of his own, would have used an ancient name for Britain, which had been a Roman province for almost one hundred years.

As with Hibernia, it would be best to establish some comparative spatial constants in relation to the overall size of the main island of Great Britain. The general shape is that of a scalene triangle whose apex is pushed over towards the to right hand side, thereby making an obtuse angle to the bottom left, as assumed by Eratosthenes and noted by Diodorus Siculus. Within this triangular shape, Stroma Island forms the northern apex, Lands End the western, obtuse angle and Margate the remaining eastern angle.

In modern maps this triangular mass of Great Britain lies between the co–ordinates, converted into the calibration of Ptolemy, of 12° 17˘ (5°43˘W) 50° 4˘ – Lands End, 19° 24˘ (1°24˘E) 51° 23˘ – Margate and 14° 52˘ (3°8˘W) 58° 40˘ – Stroma Island. In the list of Ptolemy, similar extreme points would be 11° 0˘ 52° 30˘ – Bolorium Promontory (considered to be Lands End]) and 22° 0˘ 54° 0˘ – Cantium Promontory (considered to be a point approximating to Margate]) and 31° 20˘ 60° 15˘ – Orcas promontory (considered to be a point approximating to Stroma island). However, in making any comparisons there is a difference in the equivalent miles per degree to take into consideration; the geodesic time base of Posidonius being about 20% less than today's actual figure while that of Eratosthenes is about 11% greater. Also, Ptolemy's basic grid numeration commences in the west, for longitude and for which, therefore, reverse computations with the Greenwich Mean of 0°, have to be made, while, for latitude, his grid numeration commences in the north.

Now it is quite evident that, while the shape of Ptolemy's Albion is generally that of the Great Britain of today, up to a certain latitude in northern Britain, something quite strange has happened to the basic triangular shape to the north of this point. The top third of the scalene triangle has been pushed towards the right, as though hinged at some point on that same right hand side, and the left hand side thereby greatly extended. This wild skew to the east has caused the overall shape to become an irregular pentagon. Where exactly this 'turning point' is located on the ground is not immediately obvious but it generates a cumulative effect which becomes most noticeable in Scotland, north of the line between the Forth/Clyde estuaries. In looking at Ptolemy's table of co–ordinates for the north of Britain therefore we should treat with suspicion any longitudinal co–ordinates that occur further to the east than 22° and of any undue compression in latitudinal degrees between 58° and 60°.

The first six paragraphs are concerned with the co–ordinates for the whole coastal outline of Britain and it is within these paragraphs that any fundamental errors and assumptions must be sought. These coastal outlines would have been surveyed from the sea and their co–ordinates recorded, but whether this was during one complete circumnavigation of Britannia or whether it was achieved piecemeal, as trade and military penetration moved northwards, is not known. However, Tacitus, in the Agricola (20–38), records that Agricola, having subdued the tribes in north Wales, advanced into the territory of the Brigantes and during the years AD. 81/4 and extended his campaign as far as the tribes of northern Scotland. His land advance was accompanied by a supporting fleet of ships and some of these vessels completed a circumnavigation of northern Britannia from the Clyde estuary on the west coast to the Forth estuary on the east coast. Thus Agricola's campaign traversed precisely that part of Britannia when this gross distortion in co–ordinates has occurred and must surely prompt the question as to whether his surveyors contributed wholly or partly to this fundamental error.

In section 3.1, the longitudinal co–ordinates given to the eleven locations marked as being on the north coast of Albion are the most suspect, since they are wildly at odds with the general lie of the land. In their existing locations, these have been assigned by modern historians to existing place names from the Mull of Galloway, on what is in fact the west coast, right around to Duncansby Head at the north east point of Scotland. This is based upon the theory that the north coast of Scotland would have reasonably been assumed by ancient navigators to include what is clearly the west coast.

If we allow ourselves to consider that the region north of the territory of the Brigantes may not have been surveyed until about AD83. under Agricola, then we would be justified in assuming that any geodesic survey carried out might well have been made by using the co–efficient of Posidonius. The basis for this assumption is that Posidonius worked more in relation to Republican Rome than Alexandria and his scientific findings would therefore have more relevance and acceptance there. On the other hand, since the remainder of the Britannia, south of this 'turning point', would have had a cartographic history of some length, bearing in mind the evidence of long held trading facilities, the co–ordinates in use might well have been based on the co–efficient of Eratosthenes. Of course this assumption must also rest upon the amount of data from Marinus that Ptolemy inserted unchanged as opposed to what was the result of his own knowledge. That Ptolemy opted for the former is, I think, borne out by the curious content of the map of southern Albion which reflects a situation that could only have existed at least one hundred years before Ptolemy's day.

If there was any such conflict in the use of a geodesic time base, it can be quite simply demonstrated that a 'turning' towards the east and a 'compression' towards the south would have certainly resulted for all territory exposed to the new geodesic time base. In view of the scope and extent of Agricola's expansion, this point could well have been as far south as Eboracum, Deva or even Lindum, the sort of area from where the expedition would have commenced.


Although it is not the function of this commentary to offer a working solution to the 'turning of Scotland' a case for consideration can be briefly demonstrated, on this initial set of co–ordinates. The argument pertains to the nature of the rectilinear graticules Ptolemy used for plotting purposes and the north west bias on which they are based. If a graticule is constructed to scale on plain paper to Ptolemy's instructions, numbering the degrees of longitude along the top, from 0°, representing the westernmost point he allows for the habitable world, and ascending at intervals scaled at the 500 stadia per degree, according to Posidonius. Likewise the degrees of latitude down the left hand side, descending from 63˘, also at scaled intervals of 500 stadia per degree. Another graticule should now be constructed in a similar fashion, this time on transparent paper but scaling the intervals of longitude and latitude at 700 stadia per degree, according to Eratosthenes. Now, on the first graticule, the eleven co–ordinates we have discussed in §1 should be entered as a single dot each. Now, if we lay the transparent copy precisely over this copy we can now read these original dots, through the transparency, as completely new co–ordinates. It will be seen that the co–ordinates will have shifted north and west and, as it were, reverted to the co–efficients of Eratosthenes and they now read thus:


NovantarumLong. 21° 00'.|. Lat. 61° 40'Long. 15° 00'.|. Lat. 62° 05'
RerigoniusLong. 20° 30'.|. Lat. 60° 45'Long. 14° 40'.|. Lat. 61° 25'
VindagoraLong. 21° 20'.|. Lat. 60° 30'Long. 15° 20'.|. Lat. 61° 15'
ClotaLong. 22° 15'.|. Lat. 59° 40'Long. 15° 55'.|. Lat. 60° 40
LemannoniusLong. 24° 00'.|. Lat. 60° 20'Long. 17° 05'.|. Lat. 61° 05'
EpidiumLong. 23° 00'.|. Lat. 60° 40' Long. 16° 25'.|. Lat. 61° 20'
LongusLong. 24° 00'.|. Lat. 60° 40' Long. 17° 05'.|. Lat. 61° 20
ItisLong. 27° 00'.|. Lat. 60° 40'Long. 19° 20'.|. Lat. 61° 20'
VolsaLong. 29° 00'.|. Lat. 60° 30'Long. 20° 40'.|. Lat. 61° 15'
NavarusLong. 30° 00'.|. Lat. 60° 30'Long. 21° 25'.|. Lat. 61° 15'
TarvendumLong. 31° 20'.|. Lat. 60° 15'Long. 22° 20'.|. Lat. 61° 00'

While the effect of this experiment will have been to move the co–efficients of longitude west and the co–efficients of latitude north, by a significant amount each, this does not really help resolve the problem. Certainly the longitudes have a progression eastwards that commences at an acceptable western value, but the latitudes are in opposition to any idea of a clockwise notation from Galloway round to Duncansby Head. This combination could, however, support a progression from Cape Wrath that would follow the north coast to Duncansby Head, then southwards into the Moray Firth then eastwards again to reach Dunnet Head, without any great conflict in geodesic values. However this would conflict with nomenclature already assigned and accepted by historians, particularly that of Clota for the Clyde river, as mentioned in Tacictus, Agricola 23,3: Bearing in mind the Celtic derivation of this name as being a river goddess, Watson CPNS 44, 71: too much reliance on Tacitus might well be misleading. While this small experiment has been of interest it is not intended to pursue its ramifications further at this stage. Sufficient to ask that it be borne in mind when considering the remainder of Britannia. At the same time a re–examination of the text of the Agricola of Tacitus is clearly indicated, particularly sections 22, 23 and 24, marking the third, fourth and fifth year of campaigning in northern Britain.

Paragraphs §2/3 are concerned with the co–ordinates for the west coast of Albion and, if the argument previously stated holds good, it seems certain that at some point along its length there occurs a point at which the geodesic time bases changes. On a cursory study of the sequence of the co–ordinates those between Moricambe Bay and Setiantorum Town & Harbour appear to be ideal candidates for this occurrence. Moricambe Bay is taken to be the inlet leading to the Flavian fort of Kilbride , rather than Morcambe Bay itself, and Setiantorum to be a harbour and town at the mouth of the Mersey. Whereas the latter would almost certainly be long known and used by Phoenician merchants, the former would fit quite closely to the location that Tacitus gives for Agricola's departure for the far north of Britannia. – 'Quinto expeditionum anno nave prima transgressus ignotas ad id tempus gentes crebis simul se properis proeliis domuit;' Agricola 24.1,3.

Leaping ahead to paragraphs §5/6, concerning the east coast of Albion, a similar point must also occur within that sequence also and at approximately the same latitude. In this respect Gabrantuicorum Bay (Good Harbour) might well be considered to be the northern limit to Phoenician trade and hence the Eratosthenes time base, while the group of locations clustered around the 58° parallel, and seeming to be in the general area of the Tyne estuary, might well be considered to be the commencement of new territorial advancement.

Paragraph §4 is concerned with the south coast of Albion, from the Lizard to Margate, and contains no problem areas save the identification of certain locations.

These six sections are vastly important and their understanding vital to the whole process of the further understanding of the remaining books of this work of Ptolemy. Clearly, something went catastrophically wrong with the surveying of northern Britain, for within an otherwise clear depiction of a province, surveyed co–ordinates were allowed to be recorded that were completely at odds with the general lie of the land. Within a relatively small area, at no great distance from long held, clearly established, base reference points, readings were taken and set down that were patently incorrect. That they were the result of surveys carried out during the campaign of Agricola must be given serious consideration, in which case they should be seen to have occurred where Roman expansion was taking place outside the limits of what had, up to that time, been the habitable world. Since this was a world that had long been traversed by trade routes that were constantly in use and were well established and recorded, it must give cause to wonder if this same set of circumstances occurred elsewhere in the Roman empire, where military expansion went beyond commercially established frontiers.

Paragraphs §7– 17 of Ptolemy's text deal with the disposition of the tribal entities north of a parallel that marks the northern limits of the territory of the Brigantes. Effectively this should be seen as being along a line linking the Solway/Tyne estuaries. In view of what was discussed in the opening paragraphs of this commentary, there is a distinct possibility that such dispositions may have to undergo some drastic revision. It was Ptolemy's habit to list the tribal centres of habitation, within a province, in sequence from the most northern to the most southern. In this context, the tribes of northern Britain, down to the Brigantes who, as it were, formed a coast–to–coast 'plug', observe the same registration. Thus, by Ptolemy's account, the Novantes are the most northern of this group and the most southerly are the Taecali. Since this would appear to conflict with long held opinions on tribal dispositions it is necessary to examine this in more detail.

If we allow Ptolemy's original co–ordinates to be revised, as above, the general areas of the tribal groupings, taking the mean values of the 'towns' they occupy, now become:–

. Novantes 13° 35˘ 61° 30˘

. Selgoves 13° 20˘ 60° 50˘

. Damnoni 14° 20˘ 60° 55˘

. Oadini/Otadini 14° 25˘ 60° 50˘

. Vecomagi 19° 25˘ 60° 30˘

. Venicomes 17° 10˘ 60° 0˘

. Taecali 18° 45˘ 59° 35˘

There are a number of tribes for whom Ptolemy gives no centres of habitation and no co–ordinates at all. They are general compass directions only. Thus,

. Epidi – East of the Damnoni

. Ceronei – East of the Epidi

. Creones – East of the Ceronei

. Cornonai – East of the Ceronei

. Carini – East of the Cornonai

. Cornavi – East of the Carini

. Caledonians – South of the Vecomagi

. Decantes – East of the Caledonians

. Lugi – East of the Decantes

. Smertae – North of the Lugi

It should be observed that the groupings of the centres of habitation lie in two distinct areas, near to the west coast (Novantes, Selgovae. Damnoni and Oadini/Otadini), or near to the east coast (Vecomagi, Venicomes and Taecali). Between them lie all the tribes for whom Ptolemy gives no co–ordinates at all. The implication seems to be that there was a considerable hinterland for which no precise knowledge was available, and this would accord generally with the mountainous topography of mid–Scotland.

If this is an acceptable argument then, notwithstanding deductions based on the distorted map of Scotland, a totally different view of the northern reaches of the province of Britain emerges; one that would seem far more logically based. What would seem to emerge is a reversion of Scotland to its proper elevation, showing a country in which, while the coastal areas are seen to be more developed, the wilder, more mountainous inland areas are still untouched. As such it reminds us of the description that Julius Caesar applied to southern Britain during his first expedition.

Paragraphs §16–30 are concerned with the tribal groupings from the Brigantes to the Dumnoni, once again in a strict north/south sequence. The co–ordinates given for the tribal locations in this sequence, particularly those of longitude are well within the parameters established for Britain in general. It is not proposed that they should be subjected to the same correction procedures as those for northern Britain since they appear to follow on quite logically to these amended northern co–ordinates. Apart from the identification of specific names, this southern list of Ptolemy contains few surprises in its content. However the omission of place names that should have been well established by his time causes us to speculate on the exact date attributable to this particular section of the map. Either it is basically a list drawn up well before the Claudian occupation, probably from commercial sources, to which Ptolemy has written in additional information or it has been subjected to a very peculiar form of corruption in copying. Indeed, it would seem that apart from late military amendments, it contains information of a purely commercial nature, of more use to the merchant than a disinterested traveller does. Thus ports and harbours, mines, salt deposits and agrarian areas are all present; details of urban facilities that might aid the overland traveller, only occasionally present. There are certain observations that can be made from these lists by comparing the relative positions of well-known locations. The towns of Lindum and Deva are useful reference points for such an exercise, since there can be little argument about their identities. To take Lindum and Deva and consider Ptolemy's co–ordinates:–

Lindum 18° 40˘ 56° 45˘

Deva 17° 30˘ 56° 30˘

If we consider them as being roughly on the same latitude then Ptolemy separates them, east to west, by a distance of 1° 10˘ of longitude. The geodesic time base of Posidonius allows 62˝ Roman miles per degree whereas that of Eratosthenes allows 87˝ Roman miles. 1° 10˘ according to Posidonius would give us about 73 Roman miles; according to Eratosthenes it would give us about 102 Roman miles. The actual measured distance on the Tabula Imperii Romani produced by the British Academy is about 103 Roman miles. Upon closer examination, this imaginary line from Lincoln to Chester would appear to be the northern limit within which Ptolemy's co–ordinates, and the distances between towns that can be deduced from them, bear a reasonable relationship to those of today. For instance, in respect of Eboracum (York), Ptolemy has a longitude of 20° 0˘ which places it on a longitudinal meridian which would be 1° 20˘ east of the one through Lincoln! A patently absurd situation. The same situation obtains for Caturactonium (Catterick) yet Vinnovium (Binchester) would appear to have the correct relative co–ordinates and distance. There are many such anomalies and A.L. Rivet & C. Smith, The Place–Names of Roman Britain. pp. 119/120: have some excellent tabular data in which this is most apparent. It is most noticeable that north of the Deva/Lincoln line, with the exception of Vinnovium, locations cease to have the same relativity with locations in the south yet have a common relativity with each other. Can it be that it also marks the limit of co–ordinates of Marinus and the beginning of the co–ordinates inserted by Ptolemy, for the sake of filling in the gaps in the commentaries of Marinus? Unfortunately, this is a rather simplistic approach and a more detailed analysis is required.

From the territory of the Brigantes southwards, Ptolemy lists thirty-eight towns or settlements associated with the various tribes. Of these only twenty–one can be identified from Roman sources, vide The Place Names of Roman Britain, Rivet/Smith – Atlas of the Roman World, Cornell/Matthews. Omission of certain Roman locations from the list, that should have been known to Ptolemy, pose problems in dating. These problems are brought more clearly into focus if we make a comparison of each of Ptolemy's co–ordinates with their equivalent today. Modern longitudes are reversed in keeping with Ptolemy's mean in the Fortunate Isles:

Long. Lat. Long. Lat.
Cataractorium20° 00'58°00'16°22'54°22'
Camulodunum (Yorks.)18°45'57°00'UnknownUnknown
Venta (Icenorum)20°30'55°20'19°45'52°39'
Maridunum (Moridunum)15°30'55°00'13°40'51°52'
Aquae Sulis17°20'53°40'15°38'51°23'
Venta (Belgarum)18°40'53°30'16°41'51°04'

In the above list there is an appreciable difference in the error margins between the co–ordinates north and south of the Deva/Lindum parallel. Cataractonium, Isurium, Eboracum and Petuaria show an average error factor of +21.5% in longitude. The remainder, with the exception of Corinium (+15%) and Isca Dumnoniorum (+22%), show an average error factor of +11% in longitude

Agricola's army, moving north from Wales would have passed through Deva and from there almost certainly on to Eboracum, Isurium and Cataractonium; places that may not have been of great importance to the Phoenician traders but which assumed major military significance on the advance northwards. The calculation of their exact locations would have been a routine matter for the army surveyors who, measuring distance and direction on the march from say Deva, would then have recorded the data and, by simple trigonometry, may even have located them, relative to Deva, on a map surface. Most probably it was not until Ptolemy that any attempt would have been made to set navigational co–ordinates for them and when he did so, Ptolemy assigned impossible longitudes to them. Could this have come about by using the Posidonius coefficient instead of that of Eratosthenes? By Ptolemy marking off a degree of longitude every five sevenths of the correct amount, thus increasing the longitudinal reading cumulatively. On a rectilinear graticule commencing at 7°0˘ longitude, as Ptolemy directs for Albion and Hibernia, a location with a true reading of 16°25˘ calculated with the wrong coefficient would give a false reading of 20°0˘. There would seem to be a strong case for considering that this hypothetical sequence of events did in fact take place. The latitudes given by Ptolemy do not seem to have been affected by the same factor and yet they do display the same grouping between north and south. The former show a reasonably constant error factor averaging +6%; the latter a similarly constant error factor averaging +4˝%. While no single cause would seem to explain this, the error must be the result of a combination of factors that must include the Posidonius/Eratosthenes factor but, as mentioned previously, would also appear to include Ptolemy's assumption on the exact location of Thule, his attribution to it of a latitudinal co–ordinate of 63°15˘ and his direction that latitudes be measured off southwards from this latitude

Moving north, and ultimately west, from these locations the army surveyors would have continued to measured the distance and direction marked, calculated the trigonometry and recorded the data. Ptolemy would have converted their data to degrees of longitude using the wrong coefficient unaware that, as the army's route moved further westwards from a falsely recorded base, there was a further compounding of the initial error which would ultimately produce longitudinal co–ordinates that would grossly distort and produce the 'turning of Scotland'. While the above argument is of course quite speculative, it almost certainly contains some indication of probable events. There is surely a weight of evidence that such events could have contributed to the misshapen outline that Ptolemy attributed to Britannia; it would indeed be a remarkable coincidence if the activities of Agricola, and their significant timing, were not also the catalyst.

However, what does seem to emerge quite clearly is that Ptolemy's co–ordinates have a rational basis and, with the constant error factors allowed for, can be relied upon. This point is most important in considering the remainder of the text and accepting that what Ptolemy says may be taken to have a validity that may not have been allowed in past considerations of his work.