The Date of the Gough Map

ABSTRACT

The date commonly given for the Gough map of Britain, about 1360, is, in the author's opinion, wrong. Arguments that have been offered to support such a dating are invalid. The best indication of the date of the map is the writing on it, which is essentially in a hand of about 1400, a dating endorsed by expert palaeographical opinion. Indeed, a few exceptional features of the handwriting may suggest a slightly later date. A few specific non-palaeographical features of the map confirm a date of production close to, or a little after, 1400. Comparison with other late medieval maps of large inland areas from any part of Europe shows how precocious or advanced the Gough Map is, even for the beginning of the fifteenth century. Arguments suggesting that the map had an earlier ‘prototype’, reflecting the affairs of King Edward I, are also found to be without merit.

La date communément avancée pour la carte de Grande Bretagne connue sous le nom de ‘Gough map’, aux environs de 1360, est fausse d'après l'avis de l'auteur. Les arguments mis en avant pour étayer une telle datation sont invalides. L'indicateur de datation le plus sûr, c'est le style d'écriture manuscrite qu'elle porte, qui remonte pour l'essentiel aux alentours de 1400, une datation confirmée par une expertise paléographique. En fait, un petit nombre de caractères exceptionnels de cette écriture manuscrite peut suggérer une date légèrement plus tardive. Un petit nombre de caractères non-paléographiques de la carte, très spécifiques, confirme une date de production très proche, ou de peu postérieure à 1400. La comparaison avec d'autres cartes tardives du Moyen Age représentant de larges aires territoriales de quelque partie de l'Europe montre combien la ‘Gough map’ est précoce ou avancée, même pour le début du XVe siècle. Les arguments suggérant que la carte est basée sur un prototype antérieur reflétant les affaires du roi Edouard 1^er sont considérés sans valeur.

Die Datierung um 1360, die normaler Weise für die Gough-Karte von Britannien angegeben wird, ist nach der Meinung des Autors falsch; die Gründe, die zu ihrer Untermauerung angeführt wurden, sind nicht stichhaltig. Der beste Hinweis für die zeitliche Einordnung ist die Schrift auf der Karte, die im Wesentlichen einer Hand um 1400 zugeordnet werden kann, eine Datierung, die durch paläographische Expertenmeinung bestätigt wird. Einige ungewöhnliche Elemente der Handschrift könnten sogar ein etwas späteres Datum nahelegen. Darüber hinaus bestätigen einige wenige spezielle, nicht-paläographische, Einzelheiten der Karte ein Herstellungsdatum um 1400 oder wenig später. Der Vergleich mit anderen spätmittelalterlichen Karten größerer Gebiete im Landesinneren beliebiger Regionen Europas zeigt, wie hoch entwickelt oder sogar richtungsweisend die Gough-Karte ist — selbst bei einer Datierung in das frühe fünfzehnte Jahrhundert. Argumente, die Karte basiere auf einem älteren Prototyp mit Hinweisen auf die Zeit König Eduard I., sind nach Ansicht des Autors nicht überzeugend.

La cronología con la que normalmente se fecha el mapa de Gough de Bretaña, hacia 1360, es, en opinión del autor, errónea. Los argumentos empleados en la defensa de esa cronología no son válidos. El mejor indicio de la fecha del mapa es el texto que aparece en él, realizado con una grafía de hacia 1400, fecha avalada por la opinión de un experto paleógrafo. De hecho, algunas características excepcionales de la escritura pueden sugerir una cronología algo más tardía. Algunas características del mapa muy específicas no paleográficas confirman una fecha de realización próxima, o algo posterior a 1400. La comparación con otros mapas bajomedievales de amplias áreas interiores de cualquier parte de Europa demuestra que el mapa de Gough es precoz, incluso para los inicios del siglo XV. Los argumentos que sugieren que el mapa tuvo un ‘prototipo’ temprano que reflejaba los asuntos del Rey Eduardo I se consideran carentes de interés.

KEYWORDS:

• Gough map
• medieval cartography
• early maps of England
• palaeographic dating
• Totius Britanniae
• Edward I
• Welsh castles

The manuscript map now in the Bodleian Library, Oxford, known as the Gough map after its eighteenth-century owner Richard Gough, occupies a special place in the corpus of surviving medieval maps from western Europe. It is large (55.3 × 116.4 cm), relatively well preserved, detailed and generally clear, with a wide range of pictorial signs and inscriptions in both red and black ink and with the sea and rivers tinted green. It depicts a recognizable England, Wales and Scotland, orientated to the east.

In several respects the Gough map qualifies as unique. It is not a mappamundi, but a geographical map of one part of the world. It shows an extraordinary number of rivers, settlements (generally correctly placed), and several networks of straight red lines linking some of the towns. No map that is comparable in its combination of area covered and detail of content is known to have been produced anywhere else in Europe before a date some years into the fifteenth century. The red lines, often mistaken for roads, raise various problems of interpretation. More fundamental are three other puzzles: the date of the map that has come down to us; the question of whether it was an original creation or a copy or re-working of a somewhat earlier map, now lost; and the question of whether its makers borrowed substantially from a conjectural much earlier ‘prototype’ map related to the wars of King Edward I. What follows is an attempt to answer these questions, or at least stimulate further discussion of them.

Familiar Arguments for a Date c.1360

Recent discussions of the Gough map follow various earlier writers in attributing the drawing or ‘creation’ of the map to a date soon after the middle of the fourteenth century. ‘The earliest possible date is 1355 . . . ’, writes Nick Millea; ‘the latest is 1366, when Sheppey (as it is called by the mapmaker) changed its name to Queenborough in honour of Queen Philippa, Edward III's wife’. ¹ In an earlier account of the map by E. J. Parsons, published in 1958, the title begins with The Map of Great Britain circa A.D. 1360. ² The most recent discussion of certain aspects of the map repeats the date ‘c.1360’, the authors asserting that ‘the surviving manuscript can be reasonably safely attributed to the middle of the fourteenth century’. ³ This dating is important in that it is likely to affect our understanding of the map. And it is almost certainly wrong.

It is not difficult, as will be seen, to accept that the Gough map was produced after 1355. What does need to be challenged is the traditional terminus ante quem of 1366. In fact there is no need to link the map with the re-naming of the town and castle of Sheppey at all. True, the name Shephey appears on the Isle of Sheppey off the coast of Kent, but it is patently the name of the island as a whole, corresponding with the names tenett (Thanet) and Foulnes for the neighbouring sea islands (Fig. 1). The map is assiduous in naming the major sea islands around the coast of England and Wales. The town, Queenborough, is not named, for the good reason that below the word Shephey most of the middle and lower part of the egg-shaped island is filled with an elaborate pictorial sign of a town with a castle, and no room remains for a word of nine or more letters to be fitted in without awkward cramping or spilling into the sea. The maker of the Gough map did not place town names, or parts of them, in the sea. It is unnecessary to try to explain here why this was so (the green tinting of the sea being only a partial reason); the fact is that this self-imposed rule was observed almost absolutely. ⁴ Whether the scribe could, with sufficient effort, have squeezed the name into the little segment available might be open to opinion, but to assume that he could and would have done so, and to base the dating of the map on that assumption, is quite unreasonable. ⁵

Fig. 1. Gough map, Bodleian Library, Oxford, Gough MS Gen. Top. 16. Manuscript. 55.3 × 116.4 cm. East at the top. Detail showing the Thames Estuary with parts of Essex (left) and Kent (right). The islands of Shephey (Sheppey, centre), Foulnes (Foulness, lower left), mereshey (Mersea, far left) and tenett (Thanet, top centre) are named. The cathedrals of Cantuar' (Canterbury) and Rowchester (Rochester) are prominent. Rochester Bridge over the Medway is circled. (Reproduced with permission from the Bodleian Library, Oxford.)

More specifically, the dating of the map to ‘circa 1360’, partly on the strength of its representation of Queenborough, is particularly ill chosen. Queenborough, as a substantial castle and town, did not exist in 1360. ⁶ The castle was commenced in 1361 and took shape only in the mid-1360s, probably being essentially complete soon after 1370. The attendant town, which was an important part of Edward III's project, seems to have been started (de novo) only in 1366 and was still far from complete in 1368. The map, however, emphasizes the town's domestic buildings in front of the castle, probably more so than with any other individually drawn castle in England, while correctly showing the town as unwalled. ⁷ What this means is that the representation of Queenborough was based on sound and particular knowledge that could have become available only after the new name had been adopted, which surely means that, whatever the reason for not writing the name, it was hardly chronological.

Another argument that has been offered for dating the map close to 1360 concerns the form of a rather crudely drawn ship placed in the sea northeast of the Orkney Islands (Fig. 2). Parsons, in his monograph of 1958, claimed that this was ‘a fourteenth-century warship similar to those used by Edward III, and shown on the Poole seal of 1325 and on the gold noble struck in 1344’, and offered this as grounds for dating the map. ⁸ The recent Bodleian publication repeats the observation, saying that ‘the ship is of a type built in the early fourteenth century’, and that this is ‘useful for dating the map’. ⁹ The ship, however, is of a type that remained in common use in the seas around Britain until well into the fifteenth century. Similar ships are seen in the Admiralty seals of Thomas Beaufort and John Holand datable to 1418–1426 and 1435–1442 respectively. ¹⁰ Indeed, the Gough-map ship could be said to show a perfect blend of the distinctive constructional features of the two ships portrayed on those seals. ¹¹

Fig. 2. Gough map. Detail to show the construction of the ship depicted off the east coast of the Orkney Islands. (Reproduced with permission from the Bodleian Library, Oxford.)

A third argument in favour of a dating close to the middle of the fourteenth century was advanced by the historian Sir Frank Stenton in the 1930s. As re-phrased by Parsons, the line of reasoning runs: ‘since Hessle, not Hull, is regarded as the northern landing-stage of the main ferry across the Humber, it would be unwise to place the date much later than 1350’. ¹² The map does not mark the routes of most ferries, and certainly not this one, and what Stenton seems to have been referring to is simply the fact that Hessle is included on the map. It is correctly placed, directly opposite Barton-upon-Humber, the main embarcation point on the south bank of the Humber, and it is hard to see why it should not be marked. The crossing from Barton to Hessle is about one-third of the distance from Barton to the medieval quayside of Hull and, in bad weather or in order to reach villages west of Hull, this route must have continued to be attractive. It certainly remained in use. Despite the steady growth of Hull, the Barton–Hessle ferry was still flourishing in the first half of the nineteenth century. ¹³ More generally, it is unwise to base any dating or other specific argument about the map on the inclusion or exclusion of any particular settlement. In the same part of the country as Hessle, for example, the unremarkable small towns of Great Langton and Seamer in Yorkshire, and Bitchfield in Lincolnshire, are depicted, while Selby is omitted.

A further element of the map that has been thought to suggest a date of compilation close to the middle of the fourteenth century is the name of one of only five towns identified on the vaguely drawn continental coastline shown along the eastern edge of the map. The name Sklus or Sluis has been seen as an allusion to Edward III's famous victory in the sea battle of Sluis in 1340, a relatively fresh memory in about 1360. The argument, however, is fragile. Sluis, as the outport or maritime gateway of Bruges, was one of the most important ports for shipping between England and the continent in the fourteenth and early fifteenth centuries and would have been worth marking for that reason alone.

Moreover, if any map dating from c.1360 really were concerned with celebrating Edward III's continental victories, it is unlikely that only a single engagement of 1340 would have been included. The Gough map can hardly be said to show an interest in English feats of arms when, supposedly in the very period of the battles of Crécy and Poitiers, there is not the slightest recognition of the claims of the King of England to be King of France, and no places recall the English king's triumphs even within a few miles of the coast. ¹⁴ At odds with the sustained involvement in France of Edward III and those close to him in the middle years of his reign, the map shows virtually nothing of that country. Despite ample room in the appropriate corner of the parchment, no idea is given of the northern French coastline, which was reasonably well known in England by the fourteenth century, leaving the Channel Islands apparently adrift in a wide ocean.

Interest in France and Flanders on the Gough map is confined to marking five of the best-known points of crossing, the most prominent of which is Calais. The representation of Calais as a handsome fortified town has been considered to help to date the map to soon after its acquisition by the English crown in 1347. The prominence given to the town, however, is no more than would have been justified at any time in the two centuries after that date. As ‘a small portion of England overseas’, easily accessible and well defended, it was the fulcrum of English commercial, diplomatic and military activity in northern Europe throughout at least the first of those centuries. ¹⁵ It could have been expected to have the same recognition on the map as any important English fortified town.

The Palaeographic Evidence

Overall, none of the features on the map, pictorial or toponymic, hitherto suggested gives a convincing indication of even an approximate date for its creation. Good palaeographic evidence for dating, however, does exist. There are, or originally were, some eight hundred and fifty words written on the map, of which about one third are still fully legible in generally available photographic reproductions, and roughly another third can be laboriously made out with the help of Parsons's transcriptions of fifty years ago. ¹⁶ On the whole, the words in red ink have best survived the scuffing and fading of the centuries. (Red ink was used for the names of the small number of districts or counties that are marked; for rivers, forests, mountains and lakes; for one or two notable features such as the Berkshire White Horse; for the few overseas sites; and for the occasional comment added to a particular location.) Unfortunately, these red-ink inscriptions are relatively difficult to date in an easily demonstrable way.

Anglicana Script

The place-names in red ink are written in more or less formal versions of the predominant cursive script used in England in the fourteenth century, known as anglicana. More exactly, they can be described as being in littera cursiva anglicana of the sub-type formata, tending in some cases to the most formal (or closest to non-cursive text hand) known as formata hybrida or ‘bastard’ anglicana, and even occasionally reaching the squared stiffness and formality of littera fere-textura. ¹⁷ Script of these categories, in the hands of most scribes, was liable to show little change in essential letter-forms from well before the mid-fourteenth century to well after the mid-fifteenth. The ductus, or detailed shaping of letters, may give some indication of date, but hardly allows the more confident dating that the use of alternative letter-forms from a different scribal tradition sometimes can. ¹⁸

Firmer dating, however, may be possible in the case of the names of individual settlements (towns, villages, castles and monasteries, to which can be added sea islands) that make up most of the script on the map. There are, or were, a little more than six hundred and fifty such names marked within Britain, all (with the exception of London and York) written in ink that was originally black or close to black and that now survives in various shades of grey-brown. More than two hundred and fifty of these names are still fully legible in the published photographic reproductions. About a further two hundred were wholly or mostly legible to Parsons in the 1950s (some under ultraviolet light) but can now be made out only with varying degrees of difficulty and incompleteness. Something approaching two hundred others, some already illegible to Parsons, have now in effect vanished from the most recent reproductions and, presumably, the original. What follows refers solely to the roughly two hundred and fifty names that can be clearly read. Whatever is or was true of the illegible or part-legible remainder could hardly make any difference to the arguments advanced here. ¹⁹

The town names, in black ink, are for the most part written in a slightly different form of cursiva anglicana script from that seen in the red ink. They are in a more fluent or truly cursive form, classifiable as littera cursiva anglicana media, sometimes verging on anglicana currens. They also draw to some extent on a different scribal tradition.

Evidence of Secretary Hand Influence

About 1370 a form of cursive script quite distinct from cursiva anglicana—one with several distinctive letter-forms and a novel ductus—was introduced into England; it is generally known as Secretary hand. In the 1370s, Secretary hand seems to have been used solely for documentary and archival purposes. ²⁰ By the mid-1380s it was beginning to be used for literary or other ‘book’ texts in Latin and French. ²¹ In these relatively early uses it normally appeared as a consistent, fully formed Secretary hand, written by scribes trained in that hand, and indeed from soon after 1410 we also begin to see a small number of literary texts in English written by scribes with a similar Secretary training. ²²

Quite different, and only showing itself gradually over several decades from about 1380 onwards, was the piecemeal borrowing of individual Secretary letter-forms into basically anglicana hands. The inclusion of the new letter-forms seems to have come about only haltingly. In English-language texts it was probably not until the third decade of the fifteenth century that the resultant ‘mixed’ hands were likely to contain as many Secretary features as anglicana ones. Of course, over the same period, many English scribes continued to write in purely anglicana cursive hands, their letter-forms being as difficult to date as those of the anglicana formata hands already described. However, those scribes whose anglicana did accept the influence of Secretary hand show a gradual or incremental process of adoption of the new hand's forms and features over the last few years of the fourteenth century and the first two or three decades of the fifteenth. This process, although admittedly haphazard, is one of which a great deal of evidence survives, and we have a rough idea of the chronology. It happens that the town names of the Gough map offer clear evidence of this gradual process of adoption.

The earliest Secretary letter-form to be widely adopted in basically anglicana hands—probably occasionally appearing in the 1380s and certainly quite often from c.1390 onwards—was the single-cell minuscule a (Fig. 3). This Secretary a is used on the Gough map fairly freely alongside the anglicana two-cell (or one-cell with tight overhead loop) form of a. One can see the difference of form in the neighbouring names of alton and alford (Alresford), basyngstok and bagshot, tame and walynford (all in Fig. 4), babreham (Babraham) and walden (Saffron Walden) (both in Fig. 5) and farne and the accompanying eland (Farne Island(s)), and haly and eland (Holy Island). Other examples of the Secretary form appear in Salesbery and Farnham (in Fig. 4), waltham (Waltham Abbey), Cantebrege (Cambridge), Croweland, walpole, Ramsey, Northamton (all in Fig. 5), and Schaftisbry. ²³ The scribe (or scribes) could evidently slip from one form to the other without thinking. This certainly suggests a date later than 1380 and most likely no earlier than 1390; anything more specific than that, however, it does not tell us.

Fig. 3. Selected letter-forms typical of English cursive hands c.1400–1425 as found on the Gough map. First two rows, minuscule a and minuscule w. Third row, normal Anglicana final s and the kidney-shaped form occasionally found in any cursive script after c.1380. Bottom row, three forms of minuscule r, each shown following a vowel: ‘descending’, the primary Anglicana form; ‘upright’, the primary Secretary form (variants not shown); ‘2-shaped’, a form used selectively in any type of hand.

Fig. 4. Gough map. Detail of central southern England, with the Thames and its tributaries left of centre and the Isle of Wight on the right. Oxford is shown, its name illegible, between two rivers bottom left. Below, selected place-names demonstrate differences of letter-forms. Top row (first three left to right): the cursiva anglicana minuscule a is seen in tame (Thame), basyngstok (Basingstoke) and alford (Alresford, Hampshire). Middle row: Secretary minuscule a is seen in walynford (Wallingford), alton (Alton, Hampshire) and Salesbery (Salisbury). The final examples in each row, witchirch (Whitchurch, Hampshire) and wynchester (Winchester), illustrate the Secretary letter w. In witchirch, notice also the use of 2-shaped letter r following letter i and the exceptional stretching of minuscule letter t. In the bottom row, the initial 2-shaped letter r of redyng (Reading) may well be an indicator of lateness of date. (Reproduced with permission from the Bodleian Library, Oxford.)

Fig. 5. Gough map. Detail showing the area between lenne (King's Lynn) on the left and london on the right. A ‘river’ is shown encircling Ely, to the left (see Fig. 8, below). Below, note the angularity of script in bedford and tortey (Turvey), and the stretching of minuscule t in the initial letter of tortey and in Northamton. The Secretary letter a is seen in Northamton, babreham (Babraham) and Cantebrege (Cambridge). The last example has exceptional angularity. The three forms of a in walden (Saffron Walden), babreham and Cantebrege respectively happen to illustrate clearly one aspect of the process of change taking place in early fifteenth-century cursive script in England. (Reproduced with permission from the Bodleian Library, Oxford.)

Fig. 8. A number of geographical errors point to a common source, direct or indirect, for the Gough map and the Totius Britanniae map. Details from the east-orientated Gough map have been rotated to match the south-orientated Totius Britanniae. The similarity is striking in all examples except in the case of the Isle of Axholme, the position of which in Totius Britanniae, in relation to the Trent and the Humber, is seriously distorted. (Reproduced with permisssion from the Bodleian Library, Oxford, and the British Library, London.)

The same can be said of the scribe's (or scribes') use of the Secretary ‘kidney-shaped’ or two-compartment final s. Just two examples are seen in the names that remain legible; these are peuins (Pevensey) and Lewes. This form appeared in basically cursiva anglicana script occasionally from c.1390, or perhaps a little earlier, although it was not widespread in such script until after 1400. How normal it would have seemed to the scribe or scribes of the Gough map is impossible to say, since so few legible names end in –s anywhere on the map.

We have much firmer and more striking evidence of date in the use of the Secretary form of the letter w, minuscule and possibly also capital. Throughout the fourteenth century the cursiva anglicana letter w had had a recognizable form, with its right-to-left approach sweeps or hooks at the top of its two strong downstrokes (sometimes replaced by an anti-clockwise swirl enveloping almost the whole letter), and its double back-loop against the second downstroke (see Fig. 3) The anglicana w occupied more space than any other letter; even when minuscule it could reach as high as the ascenders of other letters. ²⁴ Secretary w, in contrast, has an open, or splayed, down-up-down-up form, with the approach sweep or hairline to its first downstroke coming from the left, and the final up-stroke having a single hook or curl extending towards, but often not quite touching, the second downstroke. Apart from the first downstroke, Secretary w hardly rose above the headline.

One can, therefore, easily recognize the Secretary w when it appears, but what is surprising is how late it was in appearing. For obvious reasons the letter w does not occur in texts in Latin (except in proper names), and it is relatively infrequent in Anglo-Norman French. It is, however, a normal part of Middle English, and in those English-language texts of the early fifteenth century that contain various Secretary letter-forms, it might be expected that Secretary w would be among them. This seems never, or very rarely, to be the case. Even in Hoccleve's holograph texts of c.1422–1426, which at first glance seem to be wholly Secretary in letter-forms as well as in ductus, the letter w is always of an anglicana type. ²⁵ In fact it would be hard to identify any English-language manuscript text datable with confidence before 1430 that regularly uses the Secretary w, and even harder to identify one from before 1420 that uses Secretary w as its predominant form, or even uses it frequently. ²⁶

Yet the town names of the Gough map have several examples of Secretary w. They are outnumbered by examples of the anglicana form of the letter, but, as with minuscule a, this mixing of forms was only to be expected in vernacular manuscripts of the period. What is striking here is that the scribe, evidently without particular forethought, should slip into using Secretary w at all. Clear examples are seen in wynchester, witchirch (Hampshire) (both in Fig. 4), Lewes, yawhour (Yalding, in Fig. 1), wynchelsee, wynsour, wycombe, walden (see Fig. 5), nywmarkett, walsyngham, wircyster and newport (Gloucestershire) (the last two in Fig. 6). The name donemowe (Great Dunmow) has an exaggerated form of Secretary w with an idiosyncratic flourish.

Fig. 6. Gough map. Detail showing the lower Severn valley. On the left wircyster (Worcester) shows the use of Secretary w and of 2-shaped r following i. Secretary w is also seen in newport, located to the right of gloucester (Gloucester). Note the angularity of the initial letter of Bristowe (Bristol), a characteristic of the initial letters of several other town names on the map. (Reproduced with permission from the Bodleian Library, Oxford.)

Other letter-forms found in the names of towns on the Gough map may possibly offer further evidence of a post-1400 date. The use of the 2-shaped minuscule r immediately after any letter, rather than only after certain ones, seems not to be found, in basically anglicana hands, until well into the fifteenth century (see Fig. 3). Throughout the fourteenth century in England, a 2-shaped r had been used as one of the forms of minuscule r, but, it seems, only when following one of six letters. It was used routinely following o, frequently following b and p, less frequently following d, and occasionally following a or e. Otherwise, regardless of the type of hand, it was not used, or used only rarely indeed. From the late fourteenth century onward, at first only in some examples of Secretary hand, then progressively in both Secretary hand and in anglicana or ‘mixed’ hands, the 2-shaped r came to be used freely following any letter. In basically anglicana hands this change does not seem to have been at all advanced before the 1420s. It is, however, seen in several town names on the Gough map. ²⁷

Likewise the stretching of the stem of minuscule t further above the cross-stroke than it extends below it is a distinctly fifteenth-century mannerism, in fact only well-evidenced from the mid-century onwards. Until virtually the end of the fourteenth century, in all forms of script in England, the stem of minuscule t extended relatively little above the cross-stroke, and often not at all, while the foot of the stem curved firmly to the right, often as far as the cross-stroke above it. Gradually during the course of the fifteenth century, in cursive hands, minuscule t came to be often written with the cross-stroke relatively low and the stem above it reaching at least as high as an ascender, sometimes curving to the right, while the foot of the stem was reduced to little more than a twist of the quill. A little evidence of this is seen on the Gough map, most obviously in the name witchirch (in Fig. 4) where the upper stem is taller than a regular ascender and curves to the right. ²⁸ It cannot, however, be claimed that the use of these particular forms of either minuscule r or minuscule t is acknowledged as an indicator of date in existing palaeographical writing, and pending fuller discussion of the evidence than is possible here, the significance of these particular letter-forms can only be considered a suggestion.

The Ductus of the Script

Less contentiously, the ductus of the script on the Gough map tends to confirm the evidence of letter-forms. This is true of both the red-ink and the black-ink script, and of the anglicana letter-forms as well as the relatively few Secretary ones. Partly through the growing influence of Secretary hand, the roundedness of fourteenth-century cursiva anglicana hands tended to give way about the end of the fourteenth century and the beginning of the fifteenth to a greater angularity. The previous relative consistency in the pressure of quill on parchment was being replaced by a greater contrast of thick and thin strokes, which were now sometimes made with a broader tip, and by a clearer up-and-down rhythm. Rounded continuity of line tended to be interrupted: bows often became two or three flat or slightly ‘concave’ faces meeting at sharp angles, the so-called ‘broken strokes’, most obviously in the lobes of such letters as minuscule d and b. This, it should be said, had been anticipated to some extent (especially with letter d) in the third quarter of the fourteenth century. However, when the impression of angularity is more general and the use of broken strokes is seen in a variety of letters in a written text, a date no earlier than the end of the century is suggested.

The red-ink inscriptions on the map certainly show this angularity, even if inconsistently. Broken-stroke minuscule d, for example, is seen in the neighbouring Stranford, drowdaa, deuelyn and Mare occidentale. ²⁹ The word Stranford is conspicuously angular, its f virtually a straight line with a serif, and its second r (the 2-shaped r following o) a z-shape with top and bottom strokes curving firmly inward. This inward-curving z-like r appears quite often on the map, as in wallia borealis, yorkwold, Orkeney, Norfolk and dorset. The angular and distorted anglicana a of Stranford is seen elsewhere in, for example, Mare occidentale and holand. Initial capitals are in some cases distinctly angulated: the upper lobe of B in Boleyne, for example, is made up of straight lines, and the lower lobe of A in Ageland consists of a straight vertical linked by a hair-line to a notably inward-curving arc. It should be acknowledged, however, that the angularity of some red-ink names is best explained as a partial borrowing of the forms of littera textualis, as in the d of deuonia and dorset, the u of (albus) equus and most obviously the whole word Dacia.

There is no question of coming close to the forms of littera textualis in the more relaxed cursive of the black-ink town names, but they too have some evidence of Secretary-influenced angularity. Broken strokes are seen in the lobes of b and d in, for example, basyngstok (in Fig. 4), bedford (in Fig. 5), beuerley and dunstaple, and in the single cells of the Secretary a in yawhour, walpole, Cantebrege (in Fig. 5) and Schaftisbry, among others, besides various examples of letter o. ³⁰ The sprawling and spikey shaping of capitals in the initial letters of Bristowe (Bristol) (in Fig. 6), Reigate, Rumy (Romney) and Brumholm (Broomholm, Norfolk) hardly seems to belong to the fourteenth century, and the same is true of the slanting minims linked base-to-top with straight light strokes in such names as donemowe, tame, Rumy and wycombe. The minuscule h that looks much the same either way up, essentially two looped downstrokes linked by a straight line, as in huntington, hertford, heth (Hythe, Kent), hastynges and hauentre (Havant) would again be unusual before 1400.

The evidence derived from the ductus of the script on the Gough map may be limited on its own, but when it is taken together with the use of Secretary letter-forms, the overall assessment of date is not in doubt. In the words of the palaeographer Pamela Robinson, after studying details of the map, ‘a dating late fourteenth century or early fifteenth century is more appropriate for the handwriting than [one] before 1366’. ³¹ If a few particular letter-forms that have been noted here have pointed to a date somewhat later than 1400, this need not be problematic. The difference would only be a matter of a decade or two, and palaeographic dating often has to accept imprecision of this degree. It is possible in this case that an older scribe had picked up, in the first decades of the fifteenth century, a few of the fashionable letter-forms of the newer hand while still keeping the basic forms of his early training. Other explanations, however, are possible. The dating of the script is best left approximate.

Has the Handwriting Been Tampered with?

Any palaeographical discussion of the Gough map may be complicated, for some readers, by certain off-hand comments on the script made by Parsons in 1958. ‘Many names’, he wrote, ‘have been overwritten in a later hand . . . It is almost impossible to say when this overwriting was done, but in view of the skill Thomas Martin had with his pen, there is the possibility that it was his work’. ³² Martin, it should be explained, was the owner of the map, before Gough, in the mid-eighteenth century.

Parsons offered no evidence at this point in his text to identify these ‘many’ overwritten names. In his gazetteer, though, under the relevant counties (Kent, Sussex and Oxfordshire), he said that the names given to Mayfield and Lewes are each written twice with some variation, and that the name cheryng is also written twice. He also claimed that ‘the original forms of Heth, Rumy and Appeldre have been overwritten in another hand’, that ‘Yawhour has been written over Yaldy’, and that the more legible parts of the names chedyngfold and tetsworth (the rest being faint) have been ‘overwritten’ or ‘written in a later hand’. That is all he had to say as evidence of over-writing. He neither described the supposed differences between original and ‘later’ script, nor did he cite any additional names.

We can only guess what else Parsons might have been thinking of. A reasonable inference is that he had noticed that some town names, or parts of them, are clearly written in dark ink while neighbouring names, or the rest of the same word, are paler in varying degrees. Dark, he seems to have been suggesting, indicates re-inking. The obvious objection to such an idea is that the re-inking would be in precisely the area where it was least needed. It is generally agreed that when the map was in use it was scuffed or rubbed chiefly from the bottom or western edge, so that the loss of writing is most serious there, becoming less so as one goes upward and eastward. ³³ That, given the map's representation of the shape of the country, has meant that almost all the easily legible names of settlements lie east or southeast of a line from Shaftesbury to Bridlington.

The greater fading of the names in the western and northern two-thirds of Britain was presumably already serious by the eighteenth century, but it cannot have reached the point of illegibility, since most were still just about decipherable by Parsons in the 1950s. It is easy to imagine that these hundreds of faded but still legible names in the west and north of the map positively invited re-inking by anyone at that period prepared to do such a thing. Yet not one of these names shows any evidence of having been re-inked. All the darkest words or parts of words are found above the Shaftesbury to Bridlington line (when viewed with east at the top), precisely where probably all the names would have been fully legible in the eighteenth century. Would Martin really have mutilated his map so irrationally? ³⁴ And would anybody have done so in such an unsystematic manner, since even in this part of the map there is great variation of dark and pale (within individual names and from name to name), making them less rather than more comfortable to read?

Why, then, is there this variation in the ink? It becomes understandable if we consider the harshness of the map's treatment compared with that of most other examples of medieval writing on parchment. The map was, as far as we know, repeatedly opened or kept open and handled as a working map for perhaps a hundred and fifty years after its creation. In the twentieth century it was again, amazingly, permanently exposed to daylight by the institution responsible for its care. Under such conditions, its ink was tested beyond endurance. It may be assumed that the ink has survived best where it was thickest and that its original thickness would not have been constant. Place-names on such a map are not written in the continuous flow of a normal text; each is written individually and in no predictable order, possibly with pauses for the accompanying sign to be drawn. Moreover, two parts of the same word, as in tetsworth (in Fig. 4), could have been written with different dips of the pen, just as neighbouring names could have been written with the pen fuller or drier. Centuries of maltreatment and exposure to daylight will have shown up such differences in a way unlikely to happen with a map enclosed within a book or roll.

Ink deterioration in medieval manuscripts that have been exposed to daylight, frequent touching, damp and the smoke of open fires is often curious and unpredictable, as can be seen from two other English maps. Both survive only as fragments; the Duchy of Cornwall mappamundi comes from the late thirteenth century and the Aslake map, also a mappamundi, from the later fourteenth. ³⁵ They both show, the latter to an extreme degree, the arbitrary way in which different patches of originally presumably uniform script have survived. Particular words, or parts of words, have faded to illegibility while neighbouring words, or even isolated letters in the same word, have darkened, giving an illusion of overwriting. ³⁶ The Gough map may not have had the same brutal mishandling as these two fragments, and its contrasts of dark and light are generally less abrupt, but it is not at all surprising that its ink underwent essentially similar changes given its original use and rather careless modern treatment.

The strongest argument, however, against Parsons's speculation that some place-names had been over-written can be made only by looking at the Gough map as a whole and seeing the essential consistency of the style and forms of writing wherever it is legible. There is not a word, in red ink or in grey-brown, of which the overall impression and the letter-forms are incompatible with a date close to the beginning of the fifteenth century. Would an eighteenth-century botcher have slipped so perfectly into the ductus as well as the letter-forms of a particular, and quite short, late medieval phase? The same can be said of the few credible examples of overwriting or rewriting that Parsons gives, as already described. They too seem to belong, in their style and letter-forms, to the same period as the handwriting of the rest of the map. It is not difficult to imagine how this overwriting and rewriting came about. Eight of the nine names that he cites are found in one small area of the map, between Chichester and Canterbury. An original scribe, or someone close to him, could have opted to change the form of the names that are now Mayfield and Yalding (in each case substituting, as it happens, a more archaic form than that first used) and at the same time decided that five nearby names needed to be written more clearly. Not a single letter in the words involved is incongruous.

Further Palaeographic Considerations

If, then, it can be accepted that the map is not likely to have a stratum of eighteenth-century ink, there remain a few less extravagant possible challenges to a dating of close to 1400. It might, for example, be imagined that the Secretary letter-forms were only found, grouped together, in certain names, and that these could have been added or substituted some decades after wholly anglicana originals, which could accordingly be dated a little earlier. In fact this cannot be the case: the Secretary letter-forms are inextricably mixed with anglicana forms in the same words, and words with or without them are not recognizably different in ductus or aspect.

A final consideration in judging the date of the handwriting involves the identity of the scribe or scribes and his or their attitude to the words being written on the map. It is likely that the map was produced in London and possible that it was intended for the use of some branch of the king's government. Certainly the collaborative effort needed to produce it indicates an institutional rather than a private origin, in which case the writing could have been done by one or more of the clerks in royal service, who were among the earliest to adopt the new Secretary hand for documentary and archival work. In such a case—with a scribe viewing the place-names of the map as a matter of documentary record—the use of Secretary letter-forms, and perhaps other indications of ‘lateness’ of date, would indeed be compatible with a date earlier than the 1390s.

That line of argument, reasonable as it may seem in theory, is not borne out by the realities of the map. A self-confident scribe trained in Secretary hand would have used it consistently, whereas only about 5 per cent of the letters in the town names, and none of the script in red ink, use the Secretary letter-forms already described. Although the town names conform to that strange convention of late medieval English place-name records, the complete avoidance of the letters thorn or yogh, they do not seem to be executed with documentary discipline. They are written at various angles, sometimes on the curve, sometimes broken by the spire of the adjacent sign, often a little cramped and almost always subservient to the location of place sign, river or coastline.

This irregularity is understandable. If, as is generally assumed, the Gough map's function was primarily utilitarian, dedicated to the herculean task of showing where more than six hundred places are, it is only to be expected that elegance of form was subordinated to the conveyance of information. Similarly, the script, whether we think of it as being written by the compilers of the map themselves or as the work of a rather inartistic scribe, is likely to be the businesslike and unstudied hand of a literate man or men without special calligraphic training, neither ahead of scribal fashion nor behind it. Indications of date in the writing on the Gough map can, it seems, be judged as they would in any unexceptional book hand of the period.

Non-Palaeographic Evidence

The non-palaeographic evidence that suggests a date for the Gough map no earlier than about 1400 is of two types. On the one hand, certain specific details on the map seem unlikely to have been included before that date. On the other, the overall nature of the Gough map, when seen in the context of European map making in the late Middle Ages, gives an indication of how unlikely it is that such a map could have been produced at an earlier period.

Specific Evidence in the Map's Content

The use on the Gough map of the name Aberestwyth is a particularly telling piece of evidence. During the fourteenth century the principal part of the town now known as Aberystwyth was the castle and walled borough built for Edward I from 1277 onwards on the north bank of the Rheidol. This construction was contiguous with the ancient village of Llanbadarn Fawr, from which it took its original name. ‘In 1277’, writes J. Beverley Smith, ‘Edmund of Lancaster chose to build the new castle, not at Aberystwyth, but at Llanbadarn Fawr . . . [and] the documents consistently refer to the new castle and town as Llanbadarn Fawr. It was only later, about the time of the Glyndwr rebellion, that the name Aberystwyth came to be adopted in the records . . .’ ³⁷ (Glyndwr's rebellion occupied, roughly speaking, the first decade of the fifteenth century.) Between 1282 and 1399, the settlement was consistently referred to in the English Patent Rolls as Lampader, Lampadarn, Lampadervaur, Lampadervawe or something similar. ³⁸ Such English chronicles as mention it at this period seem always to refer to it as Llan Padarn, Lampader Vaur, Lampadernvaur, and the like, as do the Annales Cambriae presumed to come from the nearby monastery of Strata Florida. ³⁹

A different name, however, starts to be recorded from September 1402 when, for the first time, the Patent Rolls twice refer to the same settlement as Aberustuyth. A few entries with the older name follow, but in August 1410 we again have Aberustewith. In 1407, a remarkable document quoted in Thomas Walsingham's chronicle (from St Albans) makes clear how Glyndwr's lieutenants made strenuous efforts to give the castle, which they were occupying, and the attached town the name Aberustwith (variously spelled). ⁴⁰ In fact there had long been a small village of that name on the river Ystwyth, just over a mile south of Llanbadarn Fawr, and it had evidently became a matter of Welsh pride to impose its name on the larger and more important English creation on the Rheidol. In time the new name came to be generally accepted, even by the English, but it is unbelievable that an English cartographer, with or without some association with the English court, would have used any name on a map except Llanbadarn Fawr (or a variant) between the borough's foundation in 1277 and the end of the fourteenth century.

A second indication of date is the marking of the bridge over the Medway between Rochester and Strood, in Kent (see Fig. 1). This is almost the only bridge indicated on the map in the whole of the English midlands and south (although many other important bridges existed in the area, notably over the Thames, Severn and Trent), and some reason must have led to the exceptional attention. ⁴¹ This is not difficult to find. For most of the fourteenth century the ancient wooden bridge at the Medway crossing had proved unsatisfactory, repeatedly needing urgent repairs. In 1381, it was largely destroyed by ice and flood. After intensive fund raising a fine new stone bridge, pontem sumptuosissimum, was completed provisionally in 1391 and fully about 1398. ⁴²

What is remarkable about the rebuilding of the Medway bridge was the scale of public attention it attracted. Different aspects of the destruction and replacement of the bridge are noted in at least four surviving contemporary chronicles, in each case independently of the others. ⁴³ Provision for the new bridge is recorded in the Parliament Rolls for 1391 and 1397, and it is mentioned in the royal Patent Rolls at least twenty-seven times between May 1381 and November 1401. ⁴⁴ No other English bridge of the period, except perhaps London Bridge, seems to have received more than a fraction of this attention. It seems likely that the newly rebuilt bridge was depicted on the map.

The Precocity of the Gough Map

It is not always appreciated just how strikingly original the Gough map must have been in its day. Those who had the opportunity of seeing a map anywhere in western Europe in the fourteenth century were relatively few and limited mainly to monastic and ecclesiastical circles or to court and government. Sophisticated constructs such as portolan charts would have remained in the hands of navigators and their employers or, if lavishly decorated, of patrons. There was little call, given the political fragmentation of much of continental Europe and the almost total eclipse of Ptolemy's Geography, for the creation of extensive regional maps before the fifteenth century. Only the significance of Palestine for Christians prompted the occasional map of places mentioned in the Bible and the holy sites for which the crusaders were fighting and which pilgrims subsequently sought out.

There seems, in short, no geographical map from before the fifteenth century (as opposed to a theologically based mappamundi) that covers, as the Gough map does, so extensive an inland area as that of England and Wales (some 150,000 square kilometres) let alone with a comparable wealth of reliable topographical detail. More astonishingly, when the positioning of the settlements is measured by sophisticated modern cartographic analysis, as in the recent work at Queen's University, Belfast, the map apparently shows a ‘high degree of positional accuracy’ and ‘remarkably “modern” cartographic attributes’. ⁴⁵ On all counts, the overall finesse of the Gough map's geography reflects a remarkable accomplishment.

A further measure of the precocity of the Gough map is seen in its enduring influence. It, or copies of it, continued to be used as the model at home and abroad for later maps of Britain until well into the sixteenth century. Four maps from before about 1550 have been identified as owing a great deal to the Gough map: the manuscript map Angliae figura (late 1530s); the map the merchant Thomas Butler sketched into his private notebook about 1550; Sebastian Münster's woodcut map of England in his edition of Ptolemy's Geography, printed at Basle in 1540; and George Lily's copperplate Britanniae insulae, printed in Rome in 1546. ⁴⁶ In fact these late derivatives are notably less detailed and accurately delineated than the Gough map, produced up to a century and a half earlier.

Much of the data incorporated into the Gough map must have been accumulated over a long period. Stenton drew attention to a number of distances between towns on the map that are identical to those already recorded in Robert of Nottingham's account of a journey that he made in 1324–1325, measured in leucae or leagues, evidently the same unit of distance as used by the compilers of the Gough map. ⁴⁷ According to Stenton, the distances covered by Robert on three different days, and the cumulative distances covered in two or three days in four other cases, match those marked between the same towns on the map. From these Stenton concluded that the figures given on the Gough map are ‘genuine estimates of distance, such as were current among contemporary travellers . . . [and that] were tending to become stereotyped already in the fourteenth century’. ⁴⁸

Itineraries (lists of towns with distances between them) remained the standard way-finding aid from Classical to early modern times, some written down and more no doubt unwritten. ⁴⁹ They were also used in the compilation of maps, as is clearly demonstrated in Matthew Paris's map of the British Isles (1250s), where the vertical list of place-names from London to the River Tyne is particularly striking. ⁵⁰ It might be expected that, given the ubiquity of this cartographic resource, detailed maps of areas akin to that of the Gough map in western Europe would have been compiled in similar manner in the fourteenth, and at the turn into the fifteenth, century, but so far nothing comparable to the Gough map has come to light for these dates. Even in Italy, particularly advanced in cartographic skills at this period, maps comparable to the Gough map in their coverage and detail seem to have appeared a little later. ⁵¹

Whatever use was made of itineraries and other sources of travel information, the making of the Gough map needed something more. A systematic attempt must have been made to gather data through field ‘surveys’ and observers dispatched to measure distances and directions themselves or to make local enquiries. Attempts to add to the few already known astronomically defined co-ordinates for selected towns may also have been made. ⁵² To assemble so much information and then to transform it into a single image implies a high degree of administrative organization as well as remarkable cartographic confidence.

Is the Gough Map Original or a Copy?

So far, we have discussed the map as we have it. A different dimension, however, is introduced if the Gough map is considered to be a copy rather than the original. If so, was it a copy or a variant of a map made a few years earlier and now lost? Or was it dependent on a much earlier ‘original’, again conjectural, dating from the reign of Edward I (1272–1307).

The suggestion that multiple copies of the Gough map could have been made is not a startling one. As Peter Barber has remarked in relation to the Gough map, ‘it would have made sense to make several copies of a map which had cost so much in time, money and intellectual vigour to produce’. ⁵³ Nor is it any less credible that, out of several exemplars, only one should have survived. The question is whether the map before us looks like an original or a copy. The fact that it appears unfinished—inasmuch as it fails to mark routes and distances from, for example, London into Sussex, Kent and Essex—hardly tilts the argument either way. It is hard to believe that if these particular lines were present on the original map, a scribe who copied so much detail elsewhere would not have taken the trouble to copy them also. It is even harder to imagine these important routes being omitted from the original compilation.

Grounds can be found for regarding the extant map as a copy. For example, a degree of slovenliness in the placing of several towns mars the map's overall accuracy. Given the immense stock of topographical knowledge that the compilers must have acquired, would they really have imagined that Ipswich, a major port, was landlocked half way between two estuaries, that Burton upon Trent lay several miles from that river, that Piercebridge also lay far from its bridge across the river Tees, that Hedon lay north of Hornsea, or that Knowsley was nearer to Clitheroe than to Liverpool? Would they have omitted Eltham, Bath or Kenilworth, all places of prime importance at the time? These seem more like the oversights of an occasionally bored copyist than a compiler's ignorance. The argument, however, is far from conclusive.

In fact the whole idea of a dichotomy between an original and its copies may be misconceived. Instead, it may have been that when the Gough map was produced, there already existed not a finished or authoritative version, but a draft or number of drafts on which the makers of the Gough map proceeded to build. This is indirectly suggested by the survival of a map that is a near contemporary of the Gough map, the so-called Totius Britanniae map of Britain found in the codex British Library MS Harley 1808 (Fig. 7 and Appendix). ⁵⁴ The map, which is south-orientated, can probably be dated c.1400–1430. It is not only far smaller than the Gough map, but also a different type of map, marking only about one-sixth as many places as are found on the Gough map. It is for the most part geographically inaccurate. ⁵⁵ In fact, Totius Britanniae is better thought of as a pleasing illustration of part of the codex in which it is found rather than as a practical or informed map.

Fig. 7. Totius Britanniae Tabula Chorographica. British Library, MS Harley 1808, fol. 9v. Manuscript. 23 × 15 cm. South at the top. A relatively large number of places can be associated with the city of York, itself marked here by an exaggeratedly large pictorial sign. It is likely that the walled city represented in bird's-eye view on folio 45v in the same manuscript is also York, and the two illustrations may be linked. For the geographical errors common to the Totius Britanniae map and the Gough map, see Figure 8. (Reproduced with permission from the British Library, London.)

Totius Britanniae does, however, contain a number of distinctive features that give the impression that it could have been derived from the Gough map. Most notable of these are the marking of Plynlimon as a lake instead of a mountain, the circling of both the Isle of Ely and the Isle of Axholme with clearly defined but imaginary rivers, and the exaggeratedly prominent meanders of the Wear near Durham, all of which appear on the Gough Map. But this can hardly be a case of direct borrowing. Certain details on Totius Britanniae contradict the Gough map: instead of the Gough map's form of the name Aberystwyth the ‘older’ name lampader vawe is used, Baath is included, and the Latin form of Christchurch is given. A map produced as casually and sketchily as Totius Britanniae was hardly likely to have emended a more detailed and authoritative model in such ways. More significantly, the shared features listed above are among the Gough map's more obvious inaccuracies (Fig. 8). ⁵⁶ It is improbable that a borrower who took so little of the map's reliable information should have picked out several of its errors. It is much more likely that both maps were, in these particulars, indebted to some earlier map.

Indeed, it is easy to believe that a number of maps of Britain were being produced, perhaps by persons close to the royal household, about the beginning of the fifteenth century. ⁵⁷ Whether a final authoritative ‘master map’ derived from one or some of these maps ever existed is impossible to say. Certainly it is hard to see the Gough map itself as such a definitive version, or to know where to place it in this hypothetical group of related maps. Abundantly rich in reliable information but with some surprising inaccuracies, agreeably decorated (even with touches of gold) but by no means a de luxe production, evidently fit to be put on display but still unfinished, the Gough map's contradictions can be explained only by evidence now lost. Like many surviving written texts of the period, its affiliations can probably never be satisfactorily known. ⁵⁸

A Supposed Original Promoting Edward I

There remains the question of the claimed descent of the Gough map from a much earlier map, one ‘conceived’ or ‘drawn up’ as early as 1280. ⁵⁹ The chief grounds for this suggestion are said to be ‘some strange emphases’ in its content that associate it closely with Edward I's campaigns in Wales (chiefly in the years 1276–1277, 1282–1284 and 1294–1295), together with some supposed reflection of his engagement with Scotland (chiefly from 1291 to his death in 1307). ⁶⁰ In addition, what is said to be an emphasis on Arthurian myth has been taken as evidence of a wish to glorify Edward and promote his claim to rule the whole of Britain. ⁶¹ It is hard, however, to recognize these three ‘emphases’ and their claimed implications.

Edward I, Wales and the Map

In the thirty years following King Edward's first military successes in Wales in 1277 one of the principal means that he chose to enforce his control or ‘settlement’ of the principality was the building of a series of remarkable castles, probably the most impressive programme of military architecture seen in Europe in the Middle Ages. One of the characteristics of the Gough map as a whole is its hierarchy of castles and fortified towns. Individual castle signs range from the largest and most elaborate (as at Windsor), through the merely large and quite elaborate (such as those for Peel, Jedburgh and Painscastle), to the medium-sized (as at Peak Castle, Somerton (near Lincoln) and Wark), and down to the small but still clearly individualized (as at Hadleigh, in Essex, and Arundel), and the smallest, marked schematically (as at Tutbury and Clun). The signs for towns fortified by a castle likewise vary. They range from elaborate vignettes (London, York and Norwich), through those in which the castle (represented by a tower) is accompanied either by buildings and a town wall (as at Newcastle upon Tyne, Nottingham and Ludlow) or by buildings without a town wall (as at Lancaster, Newark and Lewes), down to the innumerable examples marked only by a tower and a single domestic building (as at Clitheroe, Newcastle-under-Lyme, Hertford and Helmsley).

Following the campaign of 1276–1277, Edward built or rebuilt four principal castles, Flint, Rhuddlan, Llanbadarn Fawr and Builth. None of these, with the possible exception of Builth, is portrayed on the map as having a castle, or even as a walled town, although each castle had a carefully planned fortified borough joined to it (Fig. 9). The building of defensible towns or ‘plantation boroughs’ stocked with English burgesses was a vital part of Edward's plans for north and central Wales. The Gough map shows almost no awareness of this despite having the appropriate signs in use elsewhere.

Fig. 9. Wales on the Gough map. The map gives no hint of the architectural, strategic and military-historical importance of the great castles and planned towns that resulted from Edward I's first campaign in Wales (1276–1277), namely Flint, Rhuddlan, Llanbadarn Fawr (Aberystwyth) and Builth (the last either not shown or mislabelled). The castles and boroughs that followed his second main campaign (1282–1283), Conwy, Caernarfon, Harlech and the expanded Criccieth, are shown as relatively low in the settlement hierarchy or without any indication of their military significance. Edward's last great castle, Beaumaris, is significantly misplaced. The much older castles of Painscastle and Radnor/Presteigne, of little importance to Edward, are shown as grander than any that he built, and the small and probably decayed castle of Llanfihangel-y-Creuddyn is given unaccountable prominence. (Reproduced with permission from the Bodleian Library, Oxford.)

The four castles just named were not only strongholds but also key locations in Edward's continuing Welsh campaigns. Flint was on occasion his base and mustering point, notably in 1277. Rhuddlan was his most important base in Wales in the early campaigns, especially in 1282–1283 when he spent many weeks continuously in residence. There, on different occasions, he collected armies, received embassies, took the oath of allegiance of the rebel leader Llewelyn, and in 1284 issued his defining Statute of Rhuddlan for the government of Wales. Yet castle and borough, in both these places, are simply marked on the Gough map as the lowest category of ordinary non-military town or village.

Llanbadarn Castle, augmented with one of the first and most successful of Edward's walled boroughs, was also notable for the events that followed its construction. Seized by rebels in 1282, regained, then besieged and relieved in 1294–1295, it was particularly associated with Edward's brother Edmund and became the centre of control for northern Ceredigion. ⁶² But Llanbadarn, like Flint and Rhuddlan, is marked as the lowest category of unfortified settlement.

Builth was of exceptional importance to Edward I. Not only was it a particularly massive castle with attached borough, but it also controlled a ‘land’, or miniature county, described by F. M. Powicke as ‘a royal march in front of marchlands, of peculiar importance’. ⁶³ Repeatedly attacked or used as a rallying point, Builth was the approximate site of Llewelyn's final defeat and death in 1282. Yet, according to Parsons and those following him, it does not appear on the Gough map at all. This may not be quite true, for this part of the map (Radnorshire and the adjacent counties) is inaccurate and probably has some mislabelling. The place sign just south of the river Wye, its name legible to Parsons as ‘Radnor’, could have been intended for Builth. ⁶⁴ If so, it is being shown as the lowest level of town with castle, with only a partial wall.

After Edward's second campaign, in 1282–1283, he began another castle-building phase, notably at Conwy, Harlech and Caernarfon. Not only are these castles among the most impressive in Christendom, but all were associated with significant events in the years immediately following their construction. Little hint of their importance is given on the Gough map, however. Conwy is depicted as a walled town with a castle, but only of the smallest category; Harlech is given no castle at all, but is identified as an unfortified small settlement like any other, as is Criccieth, where the castle was much strengthened after its capture from the Welsh. Caernarfon is shown as a town with a castle, but only of the smallest size and without a town wall.

A little more recognition was accorded to Beaumaris, the last and most elaborate of Edward's major castles, built after the campaign of 1294–1295 on the southeastern corner of Anglesey, strategically sited to look south over the Menai Straits to the mainland. On the Gough map, though, although marked as a castle, the small-to-medium sized sign bears no relation to the size or contemporary significance of Beaumaris. Moreover, it is incorrectly positioned on the northern shore of the island, facing the Isle of Man.

Some castles of politico-military importance in Edward's Welsh campaigns, such as Montgomery, Chirk and Bere, are not included on the Gough map. In contrast, other castles that had been important only in an earlier period are portrayed with strangely inflated pre-eminence. Painscastle (near Hay-on-Wye) is shown as far grander than any castle in north Wales, as is the castle that could be either Radnor or Presteigne. At the same time lanmihangel, evidently what remained of a small twelfth-century castle at Llanfihangel-y-Creuddyn, a few miles inland from Llanbadarn, is shown as if at least equal to Beaumaris. ⁶⁵ Particularly puzzling is the castle, identified by Parsons as Gwytherin, with a sign just like that of Beaumaris. If a castle ever existed in the village of that name, it played no known part in any of Edward I's campaigns. ⁶⁶

Another unidentifiable castle was noted by Parsons in the south of Denbighshire. This seems to be labelled engan, but, again, whatever may have been intended, such a place has no known connection with Edward I's campaigns. It should be noted that the relative positions of Rhuddlan, Ruthin, Denbigh and St Asaph (easy to picture) are grossly inaccurate, and that the castle misread by Parsons as hawarden seems to have been wrongly labelled on the map in the first place. ⁶⁷ Altogether, in their treatment of much of north and central Wales, the Gough map's compilers seem to have been floundering to a degree not seen in almost any part of England. We are led to the conclusion that, wherever their information for this area came from, it could not have been from consistently reliable observation on the ground or from any recognizable recollection of what had mattered to Edward I.

The Routes in Wales

Proponents of the idea of an early ‘prototype’ for the Gough map have made much of the fact that the map marks out a route, with the usual red lines, along the north coast and down much of the west coast of Wales, as if this reflected particular knowledge of Edward I's campaigns. In fact this was, as it still is, one of the two or three most important routes from England into and through Wales. It followed the coast because that is where, for centuries prior to Edward's campaigns and after them, the majority of substantial settlements of north and west-central Wales have been found. It represents, roughly speaking, the road taken by Gerald of Wales (Giraldus Cambrensis) in 1188 when he travelled from Llanbadarn to Chester at surprising speed. ⁶⁸

Edward I's settlement of Wales improved the condition and increased the traffic along the pre-existing track, and, having been improved, it proceeded to serve the various needs of a coastal strip that, at least eastwards from Anglesey, soon became closely linked with Chester. ⁶⁹ As it happens, the military and political importance of the route did not end with Edward I. In October 1400 Henry IV took a small army along it from Chester to Caernarfon and back, while in the previous year Richard II had made his last journey before his captivity along most of the coastal route marked on the map, from south of Llanbadarn, through Harlech and probably Caernarfon, to Conwy, from where he was taken along the coast to Chester. Altogether, this route was well known and well used. Flint, Rhuddlan and Harlech are shown on it because they were inevitable staging posts for travellers.

The Gough map marks only one other route within Wales with a red line. This too represents an ancient line of movement between Hereford, on the Wye, a tributary of the Severn and gateway to the Severn valley, and the southwestern extremity of Wales by way of Brecon and Carmarthen. Apart from the alternative approach to Brecon (up the Usk from the south coast), the Hereford route offered the easiest east–west passage across central Wales. It was used by invading English rulers at various times. Henry IV took this route in September 1403, when he travelled, with a small army as well as his usual baggage train, to Carmarthen in just five days and back in four. ⁷⁰ More than a century earlier, in June and July 1287, Edward I's lieutenants had executed a number of marches and counter-marches with small armies in the general area of Hereford, Brecon, Llywel and the upper Tywi. ⁷¹ But it would be wrong to associate the route primarily with military use. Southwestern Wales, with the important cathedral town of St David's, was one of the most Anglicized parts of the principality. The Hereford–Brecon–Carmarthen route from the western English midlands afforded a manageable approach to it for travellers of all sorts. ⁷²

It has been claimed that Edward I used the Hereford–Carmarthen route in 1295, but that is not the case. ⁷³ He did travel, in June of that year, from Whitechurch, south of Cardigan, with two unidentified stopping places, to the villages of Llywel and Llangadog, where he spent four days in local movements, before going south to Ystradgynlais and Merthyr, and then north, through Brecon, Builth and Radnor, to Welshpool. ⁷⁴ The villages of Llywel and Llangadog do happen to be on the Hereford–Carmarthen route and are marked on the Gough map, but the relationship between Edward's movements and the Hereford–Carmarthen route is obviously limited. ⁷⁵

Edward I, Scotland and the Map

There is no need to attach any political or propagandist significance in the fact that Scotland is included on the Gough map. It would have come as naturally to the map's makers to represent the whole island of Britain as it did to Matthew Paris in the 1250s and to early sixteenth-century English mapmakers. Nor is it surprising that a border was not marked. Even where local opinion was agreed on where the dividing line lay on the ground, a boundary on the map would have offended English sensitivities. Both Edward I and his son Edward II (r. 1307–1327) had claimed and fought for sovereignty over Scotland, and although by the end of 1314 they had lost what they had gained, English kings continued to believe firmly in their right to some sort of overlordship. Indeed Henry IV attempted to enforce this militarily as late as 1400. Despite this ending in inglorious retreat from Edinburgh, for the rest of his reign his tortured relations with Scotland and its regents were based on the assumption that he had a right to influence and interfere in the affairs of the Scots. ⁷⁶

On the Gough map the most reliably located Scottish towns are confined for the most part to English-speaking areas, particularly along the east coast as far north as Montrose, and to some of the Gaelic-speaking western Lowlands. The scene of Edward I's most important victory in Scotland, Falkirk, is not indicated. Some of the towns that are marked represented high points in Edward I's more successful campaigns, but these are not differentiated from towns lacking such associations. Given the freedom with which non-topographical inscriptions are scattered over various parts of Scotland, it would have been easy to point out where Edward's ambitions had seemed to be fulfilled, but no such information is given. For example, an inscription tells us that the abbey church of Arbroath was founded in honour of St Thomas Becket; nothing, though, reminds us of the formal abdication of the Scottish king John Balliol (1296), which took place in nearby Brechin and Montrose.

Twelve comitatus or earldoms are marked in Scotland, eight in the Highlands. The claim that the omission of the earldom of Douglas (to which could be added those of the Scottish March and of Angus), and the inclusion of the earldom of Buchan have significance for the dating of the map is unconvincing. ⁷⁷ The earldoms were clearly not plotted on the basis of freshly gathered topographical information. Even making allowance for the inaccurate representation of the Highlands, the positions of several of the seats of the earldoms cannot have been much more than guesswork, while all twelve titles could have been taken from a secondary source, not necessarily up to date. It is easy to think of reasons why the earldoms of Douglas and March might have been deliberately omitted at certain points in the period when the Gough map was apparently being produced. ⁷⁸ The inclusion of the earldom of Buchan has no bearing on the date of the map. ⁷⁹

The Map's Treatment of Arthurian Myth

It has been claimed that the makers of the Gough map were trying to advance the cause of Edward I by emphasizing on the map certain associations with the myths of King Arthur and thus the idea that Edward was the ‘new Arthur’, king of all Britain. In order to assess how interested in things Arthurian the mapmakers were, it is necessary to appreciate just how saturated with Arthurian myth English literary culture was in the fourteenth and fifteenth centuries. It is not sufficient simply to count the many romances describing the events of Arthur's life, and the far greater number that use his court as their point of reference in telling of the adventures of Arthur's knights and associates. It is above all to be seen in what passed as serious history. The legendary history of Britain that, as popularly understood, reached its climax in the story of Arthur and that had been invented, as far as we know, by Geoffrey of Monmouth in the 1130s with his Historia regum Britanniae [History of the Kings of Britain], had promptly become accepted by many literate people as historical fact. In the words of Richard Barber, ‘Some fifty other chroniclers writing in Latin used it in the years up to 1420 to greater or lesser extent . . .’ ⁸⁰ To this number could be added about a dozen writing in English or Anglo-Norman French over the same period. And all the chronicles placed important events of Arthur's reign in actual, or occasionally fictitious, locations in various parts of Britain.

For a fourteenth- or early fifteenth-century mapmaker the temptation to mark the many places associated with Arthur on a map of Britain must have been great. We might have expected to find something Arthurian displayed at Winchester, Carlisle or Bath—an allusion to the Round Table, perhaps—but nothing of the sort is found at these places on the map, and Bath is not even marked. There could have been a reflection of the shameless exploitation of supposed Arthurian connections by the monks of Glastonbury; instead the map simply marks Glastonbury as a town with a fairly large church, like so many other places. The makers of the map could have suggested a site for the Last Battle in, or near, Cornwall or indicated one of the notable sites of Arthur's activities on the continent, such as Mont Saint Michel; again there is nothing of the sort. Above all, they could have singled out Caerleon, the chief location of Arthur's court and a town associated more than any other with his pomp and pageantry. On the Gough map, however, Caerleon is shown as an insignificant place, a settlement of the lowest level in the map's hierarchy, lacking even the customary dab of red ink. In short, every easy opportunity to remind the Gough map's readers of familiar Arthurian landmarks has been missed.

A few items on the map have nevertheless been claimed to show Arthurian preoccupations, three of them near or fairly near Carlisle. The area around Carlisle is one of those parts of the Gough map that have been convincingly claimed to have a greater concentration of information than would have been expected, and this, surely, has to be remembered when considering why three secluded places, namely Inglewood Forest, the Tarn Wathelan and Pendragon Castle, have been marked. The first two have a small place in late medieval English literature; the last has none. Nor do the three have a significant place in the actual story of Arthur. The three surviving romances (together with a ballad) that refer to Inglewood Forest are only superficially Arthurian, telling self-contained stories that could be given almost any romantic setting. ⁸¹ Pendragon Castle is a tiny fortress between Kirkby Steven and Hawes, of no known Arthurian associations beyond its borrowing of the surname of Arthur's father. Why it should be marked on the map is probably a matter of local knowledge now lost.

Tintagel, which is shown on the map, certainly did have an important role in the story of Arthur as well as being a prominent settlement in late medieval Cornwall. It can be conceded that its inclusion may owe something to the appeal of the romances in which it appears. ⁸² But to see that inclusion as having an ideological or political message is straining the evidence. Likewise to see the small note near Dartmouth, Devon, stating that ‘Brutus and the Trojans landed here’, as an endorsement of the political ambitions of Edward I is to veer towards modern fantasy. It would have seemed a reasonable embellishment for makers or users of the map steeped in the History of the Kings of Britain. Like many early maps in which empty spaces seemed to ask to be filled with mythological figures or exotic creatures, the Gough map has its notices that Welsh soothsayers were to be found on Bardsey and wolves abounded in a particular part of Scotland. And even if the allusion to Brutus had been intended to convey a hint that Britain, the island of Brutus, was a single unit under the rule of the English king, such a note would not have been specific to any one king. ⁸³ Much more likely it was simply a reassuring reminder that the community using the map had an illustrious heritage from the ancients.

This consideration of the claims of an early ‘prototype’ for the Gough map can be summarized simply. Rather than showing an informed and admiring interest in Edward I's activities and pretensions, the map's makers were surprisingly ignorant about his campaigns in Wales, showed little interest in his successes in Scotland, and, for Englishmen of their day, seem to have been distinctly uninterested in Arthurian myth.

A New Date for the Gough Map

This article does not suggest that we can give a precise or narrow date to the Gough map. The palaeographic evidence, which of course relates only to the document as we have it, gives us no more than an approximate date, probably a decade or two after 1400 rather than any time before it. It is common when dating late medieval manuscripts by the characteristics of the script to do so in terms of quarter centuries, and the script of the Gough map can reasonably, in the absence of further information, be put in the first quarter of the fifteenth century.

The date suggested by the content of the map has to be equally approximate. The matter is complicated by the fact that the map as we have it could be either a copy of a lost original or just one, and not necessarily the earliest, of a family of related maps, the rest of which are lost. Even so, such a family, assuming its members shared essentially the same content as the Gough map, is unlikely to have been produced before the beginning of the fifteenth century. It could be conceded that the use of the name Aberystwyth and the marking of the bridge over the Medway, both indicating a date close to or after 1400, were changes effected in the process of copying or re-working. Nevertheless, the overall maturity of the Gough map and the confidence and efficiency with which it (or its exemplar) was produced, cannot easily be ascribed to any time much before the turn of the century. No attempt is made here to attribute the creation of the Gough map to any particular event or activity in the government of England in the early fifteenth century, but it would be no surprise were further research to place the content of the map more firmly in, or close to, the reign of Henry IV (1399–1413).

Notes

Notes and References

1. Nick Millea, The Gough Map: the Earliest Road Map of Great Britain? (Oxford, The Bodleian Library, 2007), 13.

2. E. J. S. Parsons, The Map of Great Britain circa A.D. 1360 Known as the Gough Map (Oxford, The Bodleian Library, 1958).

3. Keith D. Lilley, Christopher D. Lloyd and Bruce M.S.Campbell, ‘Mapping the realm: a new look at the Gough Map of Britain (c.1360)’, Imago Mundi 61:1 (2009): 1–28, at 2.

4. The only exceptions to this assertion in names that are at all legible seem to be that the last letter of both mereshey (Mersea) and Motland (Looe Island) slip into the sea.

5. The struggle that the scribe had to fit the name of nearby Rochester into the space available around the sign, breaking it into two lines, might have deterred him from trying to insert a word like quenburgh (its minimum length) into the space available for it: the two lines of Rowch/ester have greater vertical spread than the space would have allowed.

6. On the well-documented question of the building of the castle and town of Queenborough, see H. M. Colvin and R. A. Brown, The History of the King's Works, volumes 1 and 2, The Middle Ages (London, Her Majesty's Stationery Office, 1963), 2: 793–804, particularly 796–98, 800 and 802–3.

7. Ibid., 803: ‘the town . . . does not appear to have been in any way fortified’. Considering its location and intended role, the mapmakers might well have guessed, without first-hand information, that it would be fortified. The prominence of the domestic buildings in the Sheppey town sign is made more noticeable by the fact that the ‘twin’ castle of Hadleigh, on the opposite side of the Thames, is not shown as having any domestic buildings.

8. Parsons, Map of Great Britain (see note 2), 2.

9. Millea, Gough Map (see note 1), 43.

10. See Gillian Hutchinson, Medieval Ships and Shipping (London, Leicester University Press, 1994), 28.

11. On the seals one ship is a ‘hulk’ and the other a ‘hulk hybrid’. In a hulk the strakes (side planking) curve up to the sheerline before reaching the stem and stern posts; in a cog, they cluster at each end post. In a hybrid the clustering occurs at just one end. The ship on the Gough map is essentially a hulk, but with partial clustering of the strakes at the prow. This does not prove that it was drawn as late as the dates of the two seals, but it does suggest that it might have been drawn only a little earlier. As far as we know, during the later fourteenth century and in the early fifteenth hulks became more common and cogs less the norm in English waters.

12. Parsons, Map of Great Britain (see note 2), 2. See also F. M. Stenton, ‘The road system of medieval England’, Economic History Review 7:1 (1936): 1–21 at 7.

13. Alun D'Orley, The Humber Ferries (Knaresborough, Nidd Valley Narrow Gauge Railways, 1968), 48. D'Orley notes that the crossing had a paddle-steamer and ‘was much used by cattle-drovers, for it was on the route along which they brought their beasts south from the Vale of York’ (ibid., 34).

14. Some indication of a claim of sovereignty could easily have been added. Contemporary portolan charts, for example, commonly marked territories with banners or shields of arms to show who held or claimed sovereignty over them. Edward III's courtiers and loyal subjects would certainly have taken his claims over much of France for granted: after the Treaty of Brétigny (1360), the fleurs de lis continued to feature on his shield. Even with the coast of France shown as it is on the map, places inland of such importance to Edward as Guines (an actual possession), Crécy and Rouen could easily have been included. It may be relevant that after Edward III's death, throughout the reign of Richard II (1377–1399) and through most of Henry IV's (1399–1413), the claim to the French throne and ideas of military glory in France were of relatively little interest.

15. For the role of Calais in English affairs in the early fifteenth century, see E. F. Jacob, The Oxford History of England. The Fifteenth Century 1399–1485 (Oxford, Oxford University Press, 1961), 106, from which the quotation is taken. Calais held the Staple or monopoly for exporting wool and other commodities to the continent, intermittently up to 1392 and continuously thereafter.

16. The total includes all words on the map, except for about seventy repetitions of the formula fl for flumen. On the question of the legibility of the script in modern reproductions, see note 23 below.

17. A brief account of the varieties of cursive anglicana script between the 13th and 15th centuries is found in Jane Roberts, Guide to Scripts Used in English Writings up to 1500 (London, The British Library, 2005), 161. For ‘text’ or textualis / textura hands, see 140–41. The terminology used in what follows here is generally that of Julian Brown.

18. The term ‘ductus’ is not easily defined and is not used consistently by all palaeographers. In this article it describes a scribe's particular shaping of letters, personal or following changing fashion, which is only incidental to the essential structure of a letter-form. Whereas different letter-forms are usually easily recognized and described, variations of ductus are more subtle and tend to be described impressionistically. The ductus of a hand is usually understood to be something shared by a succession of words, and not just characteristic of a single word; when shared widely enough, and assessed in sufficiently general terms, it merges into the ‘aspect’ or ‘style’ of the hand.

19. The signs used to mark the towns and other settlements have survived far better than the names, not only in their dabs of red ink but also in their black ink outlines. Even where the name has mostly or wholly disappeared the sign is often still clear. This raises questions that can hardly be answered here. Parsons in 1958 claimed that the signs of certain towns had been re-inked, but even with the towns that he named in this respect there seems little if any evidence of this. The darkest-inked sign on the map seems to be that of St Helier on Jersey, but even that shows no duplication of lines such as would surely have occurred at some point if a botcher chose to re-ink it.

20. For an account of the introduction and development of Secretary cursive script in England, see M. B. Parkes, English Cursive Book Hands 1250–1500 (Oxford, Oxford University Press, 1969), xix–xx. Examples of its early use are reproduced in plates 9 and 10. See also Roberts, Guide to Scripts (note 17), 211.

21. Our understanding of the early use of Secretary hand in literary or libraria texts in England may be distorted by our knowledge of one precocious example, a treatise in Latin (Rolle's De Emendatione Vitae) in British Library MS Additional 34,763 (item 2). This is indeed in an essentially Secretary hand, dated by scripsit 1384. The date has to be accepted, with the reservation that we do not seem to have any other non-documentary or non-archival manuscript text from England in a Secretary hand that can be confidently dated within the next twenty years.

22. Probably the earliest such text to survive is the copy of Hoccleve's The Regement of Princes found in British Library MS Arundel 38, datable 1411 or 1412; see Andrew G. Watson, Catalogue of Dated and Datable Manuscripts c.700–1600 in the Department of Manuscripts, The British Library, 2 vols. (London, The British Library, 1979), 1: 88 and 2: plate 329. The ductus is thoroughly Secretary, the letter-forms about 80 per cent so.

23. Place-names are here transcribed with abbreviations silently expanded. Parsons's 1958 transcriptions were always consulted, although he does mistranscribe a few names in slight respects, including Schaftisbry among the words just cited. My readings have been taken in the first place from the two-sheet facsimile issued by the Bodleian Library in conjunction with Parsons's 1958 monograph, and the Library's 1996 poster-form reprinting of the same facsimile. The initial readings have been corroborated by use of a high resolution scan of the map, the excellent detail of which unfortunately emphasizes the deterioration of the script of the map since the 1950s; in some respects the clarity of some readings in the 1958 photocopies may never be regained. The photocopies reproduce the script of the map rather more clearly than most of the facsimiles in Millea, Gough Map (see note 1). The reproduction of the map placed online by the team at Queen's University, Belfast (http://143.117.30.60/website/Gough Map/viewer. htm) is of little help in showing the script of the place-names. No doubt partly, again, because of the continuing deterioration of the original, the black-ink script in this reproduction is too pale and fragmented to be legible without foreknowledge in the case of many words that are fully legible in the earlier photocopies.

24. A rather different form of w is found in many textualis rotunda or fere-textura manuscript texts in England over the same period. This does not have the anti-clockwise sweeps of the cursive letter, but a slight and stiff approach stroke from the left on the first down-stroke. It cannot, however, be mistaken for the Secretary w, being an essentially non-cursive form and having the double back-loop against the second down-stroke characteristic of textualis as well as anglicana cursiva hands.

25. See Thomas Hoccleve, A Facsimile of the Autograph Verse Manuscripts, ed. J. A. Burrow and A. I. Doyle (Oxford, Oxford University Press, 2002), passim. The same is true of Hoccleve's earlier surviving autograph texts.

26. It might seem that the earliest known English non-archival text using a clear Secretary w is the copy of Nassington's Mirror of Life reproduced in Watson, Catalogue of Dated and Datable Manuscripts (see note 22), 2: plate 351, under the date ‘1418’. However, Watson acknowledges that the dating is uncertain (ibid., 1: 156), and it was not accepted by the British Library cataloguers who considered that the writing was ‘probably of the middle of XV cent’: see G. F. Warner and J. P. Gilson, Catalogue of Western Manuscripts in the Old Royal and King's Collections, 4 vols. (London, The British Museum, 1921), 2: 241.

27. For example, the 2-shaped r is used after i in wircyster (Worcester, in Fig. 6) and witchirch (Whitchurch, Hampshire, in Fig. 4), after t and u in petreburgh (Peterborough) after g in grenested (East Grinstead), after a in ware and after e in bery (Bury St Edmunds), berkwey (Barkway), hertford, Tilberi (Tilbury), cheryng (Charing), derbi (Derby) and elsewhere. It is particularly striking that it is used as an initial in redyng (Reading, in Fig. 4).

28. See also tetford and, as less extreme examples, abyngton (Abingdon) (in Fig. 4), caxton, Northamton, tortey (Turvey) (in Fig. 5), and brydlyngton (Bridlington).

29. Other examples are seen in Arderne (Arden), lyndese (Lindsey), yorkwold (Yorkshire Wolds) and kendale.

30. The sharp angles or ‘points of breaking’, where the strokes that make up these lobes and cells meet, are only rarely drawn out into the ‘horns’ that are a feature of some Secretary-influenced cursive hands from the second decade of the 15th century onwards. The clearest example of this is seen in the a of Cantebrege (in Fig. 5).

31. I am most grateful to Pamela Robinson of the University of London for permission to quote from private correspondence. Her assessment was made after consultation with other expert palaeographers who saw sample facsimiles.

32. Parsons, Map of Great Britain (see note 2), 3.

33. Ibid., 3: ‘Users leaning over the map would rub this part with their sleeves’. This assumption is corroborated by recent publication of 15th-century evidence from an Oxford college of the fixing of a map to a mensa (normally a ‘table’), which implies a fairly horizontal structure that one could lean over: Rodney Thomson, ‘Medieval maps at Merton College, Oxford’, Imago Mundi 61:1 (2009): 84–90, at 88, no. 7, s.a. 1461: ‘pro mensa deseruiente mappam . . . ad disponendam mensam pro mappa . . . ad figendam pellem ad mensam . . .’ The nail-holes around the edge of the Gough map show that it was at some time fixed to a frame. The pallor of the western and northern parts of the map could also be ascribed to uneven exposure to light over a long period.

34. There seems to be no reason in any case to suspect Thomas Martin, a fairly scholarly antiquarian, of wanting to practise his scribal skills on an ancient document that he obviously valued. There is even less reason to suspect the presumed previous owner, Peter le Neve, or the known following owner, Richard Gough, of doing the same thing.

35. A good reproduction of the Duchy of Cornwall mappamundi fragment is found in The Hereford World Map: Medieval World Maps and Their Context, ed. P. D. A. Harvey (London, The British Library, 2006), 20. A particularly useful monochrome reproduction of a substantial part of the Aslake mappamundi fragment is found in P. D. A. Harvey, Medieval Maps (London, The British Library, 1991), 36. See also Peter M. Barber and Michelle P. Brown, ‘The Aslake world map’, Imago Mundi 44 (1992): 22–44. All the script in the monochrome reproduction just cited is transcribed at 36–38 in that article.

36. In the Duchy of Cornwall fragment, see, for examples of freakish darkening of occasional letters, the first line (Orientalis plaga) and the ninth (meridiana pars) of the spandrel of marginal text. Examples in the Aslake fragment are too obvious and numerous to need suggesting.

37. J. Beverley Smith, ‘The foundation of the borough’, in Aberystwyth 1277–1977, ed. Ieuan Gwynedd Jones (Llandysul, Gomer Press, 1977), 14–27, at 18.

38. The Calendar of Patent Rolls (London, 1901—) is a long and incomplete series of volumes containing the record of royal grants and administrative decisions as issued in letters patent. Variously edited, it is arranged under the reigns of monarchs, those here concerned running from Edward I to Henry IV.

39. Annales Cambriae, ed. John Williams ab Ithel, Rolls Series (London, 1860), 105, 106, 108 and 109. See also, for example, Thomae Walsingham . . . Historia Anglicana, ed. H. T. Riley, Rolls Series (London, 1863), 18 (Rex . . . in Occidentali Wallia apud Lampader Vaur . . . castrum construxit insigne), and 21.

40. See The St. Albans Chronicle 1406–1420, ed. V. H. Galbraith (Oxford, Oxford University Press, 1937), 22–27. The indenture in question describes Prince Henry, having besieged the castle for more than a year without success, parleying with the rebels as equals and obviously prepared to swallow his pride on the question of nomenclature.

41. A bridge is marked near Camelford in Cornwall, probably at Wadebridge.

42. R. H. Britnell, ‘Rochester Bridge 1381–1530’, in Traffic and Politics: The Construction and Management of Rochester Bridge, AD 43–1993, ed. Nigel Yates and James M. Gibson (Woodbridge, Boydell Press, 1994), 43–106, particularly 43–50. For background information, see Colin Flight, The Earliest Recorded Bridge at Rochester (Oxford, Tempus Reparatum, 1997).

43. The Westminster Chronicle 1381–1394, ed. L. C. Hector and Barbara F. Harvey (Oxford, Oxford University Press, 1982), 1–2; Eulogium Historiarum vel Temporis, ed. F. S. Hayden, 3 vols., Rolls Series (London, 1863), 3: 351 and 367; The Chronicle of Adam Usk 1377–1421, ed. C. Given-Wilson (Oxford, Oxford University Press, 1997), 36–37; and Galbraith, St. Albans Chronicle (see note 40), 22. The enthusiastic description of the new bridge just quoted is from this last chronicle. The author, Thomas Walsingham, virtually repeated his reference to the bridge (though not the sumptuosissimum) in his Ypodigma Neustriae. Two of these chronicles mention the bridge in reference to the death in 1407 of Sir Robert Knolles, one of its two main benefactors.

44. Rotuli Parliamentorum, Petitiones & Placita in Parliamento, 7 vols. (London, 1767–1832), 3: 289–90 and 354. Under the 1397 entry: ore le dit Pount est fait de novel de pere & en un meillour & pluis seure lieu q'il estoit devant (Now the said bridge is made anew in stone and in a better and more secure place than it was before).

45. Positional accuracy is indicated by different regression procedures. See Lilley and Lloyd, ‘Mapping the realm’ (note 3), 18 and 19.

46. Respectively, Peter M. Barber, King Henry's Map of the British Isles: BL Cotton MS Augustus I.i.9 (London, The Folio Society, 2009), passim; Daniel Birkholz, ‘The Gough map revisited: Thomas Butler's The Mape off Ynglonnd, c.1547–1554’, Imago Mundi 58:1 (2006), 23–47; Peter M. Barber, ‘The British Isles’, in The Mercator Atlas of Europe. Facsimile of the Maps by Gerardus Mercator contained in the Atlas of Europe, c. 1570–1572, ed. Marcel Watelet (Pleasant Hill, Oregon, Walking Tree Press, 1998), 43–77, and Peter M. Barber, ‘Mapmaking in England, c.1470–1650’, in The History of Cartography, Vol. 3, Cartography in the European Renaissance, ed. David Woodward; 2 vols. (Chicago, University of Chicago Press, 2007), Bk 2: 1589–669, esp. 1620–1622. For the last three maps, see also Millea, Gough Map (note 1), 50–57.

47. Stenton, ‘Road system of medieval England’ (see note 12), 13–14.

48. Ibid., 14.

49. For some medieval examples, see Catherine Delano-Smith and Roger J. P. Kain, English Maps: A History (London, The British Library, 1999), 148–53.

50. For reproductions of two of the four versions, see Harvey, The Hereford World Map (note 35), 74–75. See also P. D. A. Harvey, ‘Matthew Paris's maps of Britain’, in Thirteenth Century England IV. Proceedings of the Newcastle upon Tyne Conference 1991, ed. P. R. Coss and S. D. Lloyd (Woodbridge, Boydell Press, 1992), 109–21.

51. Four large and detailed maps of Italy dating from the first half of the fifteenth century (and three from later in the century) are discussed by Marica Milanesi, ‘Antico e moderno nella cartografia umanistica: le grandi carte d'Italia nel Quatrocento’, Geographia Antiqua 16–17 (2007–2008): 153–76 and plates. In relation to the two earliest maps (Archivio di Stato, Florence, and British Library MS, Cotton Roll xiii.44), Milanesi notes that these show Italy as a single unit, from the Alps to Calabria, despite a millennium of political fragmentation, and that the primary objective appears to have been to ‘supply the correct location of modern inhabited places’ (p. 161). While these maps are later than the Gough map, it is clear that we are only just beginning to learn about the production and use of large-area topographical maps in the early 15th century. I am indebted to Peter Barber for this information and comment.

52. See Lilley and Lloyd, ‘Mapping the realm’ (note 3), 15–17. On the availability of co-ordinates for latitude, and more rarely, longitude, by the early 15th century, see Delano-Smith and Kain, English Maps (note 49), 43, and related n. 108.

53. Barber, King Henry's Map (see note 46), 30.

54. Totius Britanniae is reproduced in full in colour in Barber, King Henry's Map (see note 46), 31, and in black and white in Delano-Smith and Kain, English Maps (see note 49), 21.

55. For example, Northampton and Daventry are placed near to the Lancashire coast, roughly north of Conwy and Caernarfon. Within a London–Huntingdon–Worcester–London triangle, which on the Gough Map would enclose about 25 towns, the Harley map has only one, named Stafford.

56. Certain areas on the Gough map, even within England, cannot have been mapped from first-hand observation and for these its makers must have relied on secondary sources, with some misinformation. The topography of the fenlands around the isles of Ely and Axholme would have been little known to outsiders and anyway difficult to map. Their knowledge of the mountainous parts of Wales was evidently slight.

57. I am grateful to Peter Barber for valuable discussion and suggestions on this topic.

58. Comparison may be made with the text of Chaucer's Troilus and Criseyde. Despite the survival of 16 manuscript copies spread in date over a century from the time of its composition, the mid-1380s, the textual tradition and the relative authority of the different versions are still in various respects unknowable.

59. Millea, Gough Map (see note 1), 13. See also Daniel Birkholz, The King's Two Maps (New York and London, Routledge, 2004), 67.

60. Millea, Gough Map (see note 1), 13.

61. Millea, Gough Map (see note 1), 14–15 and 41–43; and Birkholz, King's Two Maps (see note 59), 94–98 and 113–48, particularly 115–17, 121 and 134–41.

62. Beverley Smith, ‘The foundation of the borough’ (see note 37), 18.

63. F. M. Powicke, The Oxford History of England. The Thirteenth Century (Oxford, Oxford University Press, 1953), 411.

64. In that case, the large unnamed castle sign immediately to the northeast of the river name would represent Radnor, not Presteigne, and the small settlement between Builth (as here assumed) and Llandovery, its name ‘obliterated’ according to Parsons, would be Llanwyrtyd. Almost all the names on this part of the map seem now to be much less legible than they were to Parsons in 1958.

65. This is not, pace Parsons, Llanfihangel Castell Gwalter, but the settlement still prominent enough in the 16th century to be marked by Saxton as Llanyhangle ycrothen.

66. Parsons's reading is supported by the corresponding wetheryn on the Angliae figura map; see Barber, King Henry's Map (note 46).

67. From its location, the place should be Caergwrle (Hope), but the name on the map almost certainly ends with the letters werk, perhaps suggesting Basingwerk, the location of a small 12th-century castle as well as the well-known monastery, but which is in fact on the coast.

68. Gerald of Wales: the Journey through Wales and the Description of Wales, trans. and ed. Lewis Thorpe (London, Penguin Books, 1978), 32 and 35–36. In just seven or eight days, Gerald and his aged archiepiscopal companion not only covered about a hundred and sixty kilometres of the route, but spent a day on Anglesey, and held a number of church services. They rode from Bangor to Rhuddlan, about forty-eight kilometres, in a single day.

69. The link with Chester and its region is reflected in arrangements made for the maintenance of the castles in north Wales, which were supervised from Chester: H. J. Hewitt, Cheshire under the Three Edwards (Chester, Cheshire Community Council, 1967), 50; and J. T. Driver, Cheshire in the Later Middle Ages (Chester, Cheshire Community Council, 1971), 55–56. Slate, lead and cattle were among the goods regularly sent from Wales into Cheshire. All ports along the north Wales coast, as far as Beaumaris, also served the Anglo-Irish traffic.

70. Henry IV travelled from Hereford to Hay, Talgarth and the upper Usk, stopping for a day on the way out at Defynnog near Llywel. The return journey was presumably along the same route and without the rest day: J. E. Lloyd, Owen Glendower: Owen Glyn Dwr (Oxford, Oxford University Press, 1931), 73–75. This was, in the words of a more recent historian of the rebellion, ‘the familiar route from Hereford to Brecon and thence to Carmarthen’: R. R. Davies, The Revolt of Owain Glyn Dwr (Oxford and New York, Oxford University Press, 1995), 238. Henry had already used the route from Hereford at least as far as the upper Tywi, in October 1401: Ian Mortimer, The Fears of Henry IV (London, Vintage Books, 2008), 242, especially n.41.

71. See J. E. Morris, The Welsh Wars of Edward I (Oxford, Oxford University Press, 1901), 208–12.

72. The mid-14th century chronicler Ranulph Higden claimed that one of the four principal roads (regales vias) of Britain (which he called Rikenildstrete) ran for its first part from St David's to Worcester. It may be inferred, from his text, that this road did not go round by the south coast of Wales, to which he allocated another of his four ‘royal’ roads: Polychronicon Ranulphi Higden, Monachi Cestrensis, ed. Churchill Babington and J. R. Lumby, 9 vols., Rolls Series (London, 1865–1886), 2: 46.

73. Millea, Gough Map (see note 1), 14.

74. Henry Gough, Itinerary of King Edward the First throughout His Reign, 2 vols. (Paisley, Alexander Gardner, 1900), 2: 130.

75. The direct route between Llywel and Llangadog, which is shown on the Gough map and which has been claimed by Millea (Gough Map (see note 1), 14) to reflect particular knowledge of a campaign of Edward I, was a well-attested track between what is now Trecastle, about a mile from modern Llywel, and Llangadog. (Stenton, ‘ Road system of medieval England’ (see note 12), 9, n.3, points out that Trecastle is in the parish of Llywel; the name Trecastle does not seem to have been widely used before the 16th century.) Eighteenth- and early 19th-century maps show a substantial road on this line: see, for example, John Rocque's map of ‘Carmarthenshire with Glamorganshire’ of 1753, and the map of south Wales published by Henry Teesdale in 1829, both reproduced in John Booth, Antique Maps of Wales (Westbury, Cambridge House Books, 1978), 81 and 107.

76. The persistence of the English kings’ assumption of their right to the overlordship of Scotland is reflected in the Angliae figura map (late 1530s) on which the colouring of Scotland (as of Ireland) and the closed crown over the title cartouche ‘implictly symbolised Henry VIII's “imperial” sovereignty over the whole of the British Isles’: Barber, King Henry's Map (see note 46), 48.

77. Millea, Gough Map (see note 1), 14 and accompanying n.24.

78. Douglas was always a particular thorn in Henry IV's side, and George Dunbar, Earl of March, after a long and complicated involvement in Henry's dealings with Scotland, turned traitor to him before the end of the reign.

79. Although the earldom of Buchan had lapsed in the early fourteenth century, it was revived twice, first about 1382 for Alexander Stewart (‘The Wolf of Badenoch’) and then again in 1406 for his kinsman John.

80. Richard Barber, King Arthur in Legend and History (London, Sphere Books, 1973; Ipswich, Boydell Press, 1974), 47–48.

81. Helaine Newstead, ‘Arthurian legends’, in A Manual of the Writings in Middle English 1050–1500, ed. J. Burke Severs and Albert E. Hartung (New Haven, Connecticut Academy of Arts and Sciences, 1967–, in progress), 1:38–79, esp. 61–66. All three of these romances seem to have been written after 1400; we probably have no literary record of Inglewood Forest or the Tarn Wathelan before that date.

82. These romances, it should be said, were as likely to be about Tristan and Isolde as Uther and Igerne.

83. In Henry IV's attempt in 1401 to assert his overlordship of Scotland, ‘the English recited the time-honoured rigmarole beginning with Brutus the Trojan’: Ranald Nicholson, Scotland, the Later Middle Ages (Edinburgh, Oliver and Boyd, 1973), 219–20.

84. The first chronicle was written in a predominantly Secretary hand in which there is the occasional Secretary minuscule g capped with the distinctive horizontal bar. The second chronicle is in an angular anglicana hand, impossible to date narrowly, but proved by its reference to the death of Henry IV to be no earlier than 1413.

85. The physical arrangement of the codex is difficult to discover without unbinding.

86. The page is reproduced in black and white in Delano-Smith and Kain, English Maps (see note 49), 180.

87. See Stenton, ‘Road system of medieval England’ (note 12), 13 and 20.

88. Ibid.

89. The towns are not located accurately on the map, but that simply reflects the style of the map as a whole.

90. Parsons's gazetteer, in his Map of Great Britain (see note 2) has a note, under Cardiganshire, referring to ‘a drawing of a range of mountains, scarcely visible’, southeast of Plynlimon, but he makes no mention of a name. No particular mountain exists in that vicinity, although the surviving district name Elan might be related.

APPENDIX. A Note on Totius Britanniae

The map of Britain in British Library MS Harley 1808, at folio 9v, was given the name Totius Britanniae Tabula Chorographica by the cataloguers of the Harley collection. The map measures approximately 23 by 15 centimetres and occupies a full page. It is preceded and followed by two incomplete chronicles of England, each seemingly unique. The one before the map, Anni a Natiuitate Christi, finishes at 1122, and the one that follows it, Cronica facta sub compendio, continues to roughly 1413. Both of these are in hands that seem to be datable c.1410–1430. ⁸⁴ The map on folio 9v is not, however, recognizably linked with either of these items. It was drawn on the last of three pages that were apparently originally unused, probably at the end of a gathering. ⁸⁵ There is, moreover, no reason to associate the map with the third item in the codex, beginning with an illuminated page at folio 30v, which rehearses a little of the Brutus myth or legendary history of Britain.

If the map has any connection with anything else in the codex, it would be with the full page illustration on folio 45v and a written text on the following thirteen leaves (folios 46r–58v). The illumination is a bird's-eye view of a circular walled city, thought to be York, enclosing a scatter of buildings and vignettes assumed to have been associated with events in the cathedral's history. ⁸⁶ The text contains an account of the history and privileges of the archbishopric of York. The script could be considered an anglicana hand with strong Secretary influence or, probably more convincingly, a ‘mixed’ hand that is rather more Secretary than anglicana. It is unlikely to have been written before the 1420s.

This script is unlike that of the inscriptions on the map of Britain on folio 9v, which rules out any simple and obvious link between them. However, the map and the city view could have been drawn by the same person. The distinctive dark green of the sea on the map is also found in some parts of the background of the city view, and the pink touching on most of the towns on the map is like that on the towers of the city's walls.

More significantly, the content of the map shows a preoccupation with York. Its sign for the city is far more elaborate than that for any other place, even London, and one of the map's few moderately competent representations of topography is the sequence of towns from Huntingdon to York by way of Wansford, Stamford, Grantham and Doncaster, a frequently used route to the city. The inclusion of Burton upon Stather, Lincolnshire, while omitting half the major towns of England, may seem surprising until it is realized that Burton upon Stather was a point of embarcation for travel by river to York. ⁸⁷ A related point of disembarcation en route to York was Howden, which is also on the map. ⁸⁸ The marking of eight towns close to the Humber (Grimsby, Barton-upon-Humber, Brigg, (formerly Glanfordbrigg), Beverley, Hull and Hornsea, as well as Burton upon Stather and Howden), a far greater concentration of towns than in any other equivalent area covered by the map, could reflect the strong ties of Beverley with York. ⁸⁹ If it were the case that the map was drawn on a previously blank page in association with the York text, then it could well be of the same date as the hand of that text, which seems to be no earlier than c.1420.

The script on the map itself neither confirms nor precludes such a dating. Of necessity, the writing is much smaller than elsewhere in the codex. It is generally tight and upright, a constriction that could mean that the scribe was not writing fluently in his normal hand, as for continuous text, and in writing more carefully he might have been less likely to allow himself innovations of style. The hand is essentially anglicana but with a scattering of examples of Secretary a (as in Eboracus [sic], Barton, dauentre), but this alone points to no more precise a date than some time in, roughly, the half-century after 1380. However, the distinctive form of capital B seen in Beaulieu, Bristow, Baath, Brendwod, Barton, Beuerlay and elsewhere—in which the upright is preceded by a 2-shaped flourish—is characteristic of the years after 1400 rather than before. The script of Totius Britanniae thus confirms that the familiar dating of the map as ‘c.1400’ is not unreasonable, provided it is not taken too strictly, so that the suggested alignment of date with the apparently later folios 46r–58v is by no means impossible.

It was suggested above not that Totius Britanniae was derived directly from the extant Gough map, but that both maps must have been indebted in one way or another to a common source (see Fig. 8). Totius Britanniae was unlikely to have been copied directly from the Gough map since it contains features absent from the latter, as already noted (such as the use of the name Lampader vawe and the inclusion of Bath, neither particularly remarkable). More significant is the presence of the name mons Ellenyth, with appropriate sign, to the southeast of the Plynlimon ‘lake’, which does not seem to be marked on the Gough map. ⁹⁰ And then there is a small town, apparently named Gosford or Gofford (?Copford) where one would expect Ipswich or Colchester.

Most curiously, in labelling Carmarthen as kermerðin the Totius Britanniae map seems to be using the archaic English letter eth. The letter eth went out of use early in the fourteenth century, and one can only guess that the map's author was relying on some map or document relating to Wales from a much earlier period. Nevertheless, he evidently had no great knowledge of Wales, for which he marked no more than fifteen places, with St David's and Haverfordwest both placed, remarkably, on the north coast.