This suite of webpages was prompted by a remark of Paul Harvey about portolan chart colour. "The convention deserves detailed investigation. Its origin may be functional or decorative - or neither" (2009, p.52).

The C&SA was made possible by the scans accompanying Ramon Pujades's Les cartes portolanes (2007) and his important achievement in assigning the undated works to their appropriate author (or at least place of production) and the correct period. No serious portolan study can be done in the future without access to this work. Please urge your library to acquire it, via the Institut Cartogràfic de Catalunya.

The 2007 Pujades DVD cannot be manipulated or downloaded. Since it was impossible to carry out side-by-side comparisons, relying instead on drawings and memory, there will inevitably be some inaccuracies in my work, for example in the designation of ambiguous colours.

The concentration in these pages on my occasional points of difference with Pujades should not obscure the fact that I agree with him on most issues.

Rather than just describing what can be seen on, at most, selected charts, the C&SA justifies its conclusions, about what is unique, rare or expected, on the basis of its large body of systematically analysed data, drawn from almost all surviving works of the period up to 1469 and some beyond.

Although there are few prior assertions underlying these pages, one is that portolan charts are a major, if not the primary, source for their own history.

The C&SA considered features most of which were ignored in the past as being trivial or too small to notice. A good example is the thousands of small or even minute offshore details. [See 'Small hydrographical details'.]

Some of the many questions asked have not yet been adequately answered. Sometimes the analysis throws up observations that are not yet understood. [See Some areas for possible future research.]

Sometimes what was found was unexpected. [See Expectations contradicted by the C&SA findings.]

The colouring of islands and estuaries forms a fundamental part of the portolan chart construction process. Colours distinguish visually the coastline from offshore features, and one island from another. No work after the Carte Pisane and Cortona chart is without coloured islands, right up into the 18th century. [See Notes: Introduction.]

It was already evident, from a study of toponymy and coastal outlines, that significant development took place in the first 30 years of the 14th century. The C&SA data back that up - indeed confirming that most long-lasting conventions had appeared by 1330 - even if few of them would seem to us to represent progress.

Just as comparative analysis of toponymy has revealed patterns of interconnections between chartmakers which cannot be seen from an individual chart, so the codification of island colour and shapes, as well as certain key conventions, can lead to the identification of previously unsuspected visual, chromatic and constructional 'signatures'. [See Notes: Introduction.]

Most chartmakers were no more than extremely careful copyists. They were trained to imitate, not innovate and, apart perhaps from place-name changes, few made any positive contribution to portolan chart development.

Some features required active collaboration between two or three specialists, or forethought if a single person was involved, for example in the coordination of a plethora of small hydrographical details in the Aegean, in the handling of 'name labels', or the instances where a name's initial letter was placed in a separate box on a nearby island.

The way that an armorial device might be applied to Corsica, Sardinia and Sicily demonstrates how aesthetic considerations could overrule practical ones, since the device sometimes obliterated part of the toponymy. In a similar way, the occasional gold leaf covering those and other large islands might entirely smother the traditional coastal place-names (although the Catalan Atlas was careful to avoid that by placing them in the sea). [See The presence of an armorial device within the separate countries of the British Isles, as well as within Corsica, Sardinia and Sicily (a Microsoft Word document).]

Close examination of different types of features confirms the care taken to reproduce precisely and consistently the smallest hydrographical detail. Even when drawn freehand, the outlines usually remained true to their traditional 'signature shape'.

Colour choices can be a guide to chartmaking lineage. However, there was greater consistency over island shape and the use of standard conventions.

To us, the minor details documented in the C&SA are no more than trivial parts of the background, indeed some are barely visible. But it is clear that the copyists did not make such distinctions. The hierarchy of value and importance that historians have imposed on portolan chart content is ahistorical. Every reproduced feature on the chart was the result of a conscious process or an act of memory, and each, whether large or small, was placed in the new chart with the same care. [See Conservatism and workshop practice.]

The treatment of the Aegean islands has wider implications for our understanding of portolan chart function. Their placement is far too erratic to have been of any use to a helmsman trying to make his way through or between them. But the charts offer a geographical catalogue of the islands, with the approximate position and name of each. For a visiting ship this would be a valuable aide-mémoire, even if a local pilot or past experience was needed to thread a safe way through the labyrinth. [See A deeper understanding of the function of portolan charts.]

Portolan charts were certainly used at sea, but how? It is likely when one or more islands would have been in view at any one time, that sailors would use the visual association of the easily memorised 'signature shapes', as a convenient way to monitor their progress and plan their route. This applies also to the continental coastlines, which were understood by both producer and user to comprise a sequence of correctly positioned headlands, river mouths and settlements, not a continuously surveyed outline. Whatever navigational use might have been made of the portolan charts for longer voyages, within sight of land - rather than being an instrument for way-finding - it would have served as a means of recognising the ship's position in relation to a visible feature, or of sequencing a route by means of its place-names. [See A deeper understanding of the function of portolan charts.]

The 2007 Pujades scans from the Cornaro Atlas (a Venetian collection in the British Library dating from 1489 or a little later) feature copies of several 15th-century chartmakers. The C&SA could not be applied here, because the atlas's copyist was not interested in a chartmaker's visual and chromatic signatures. However, his faithful copying of toponymy and hydrographical details gives us access to an extra eight works by named 15th-century chartmakers, with parts of three others. [See A note on the Cornaro Atlas.]

The C&SA findings, which considered over 100 features relating to 51 geographical entities, are set out on 13 detailed tables and several comparative ones. [See the Colour & Shape Analysis Menu page.]

Neither the Carte Pisane nor Cortona chart have island and estuary colouring, which is first seen on either the work of Vesconte (1311) or on the oldest surviving (anonymous) Genoese chart in the Riccardiana Library, Florence. Since both those use fewer colours than later works and no earlier instances have been noticed on other medieval maps, it seems likely that this convention, to last well into the 18th century, was introduced by the portolan chartmakers of the early 14th century. [See Earliest Italian works and also How the colour was used and applied .]

There are some instances of great variation, both within and between the centres of production; in other cases there are strong regional or universal associations with a specific colour. [See Consistency conclusions.]

Even if the level of overall consistency varies considerably between chartmakers, there are still clear patterns running through the colour choices of the least consistent. In other words, all chartmakers were aware of the detail of island colouring and all made some conscious efforts to maintain some consistency. Most also took great care when applying the paint.

Colour choices were not determined by the availability of pigment, nor was red used to indicate importance.

No chartmaker (of those for whom we have at least three works) established a chromatic signature for the entire chart and faithfully observed it thereafter, until the middle of the 15th century.

Since charts used four or more different colours, the observed patterns could not be explained by chance.

The association of a particular colour with a well-known Mediterranean island was logical, even if the Catalans and Venetians chose a different colour. In the same way, it is not surprising that inconsistent colouring was used for the imaginary Atlantic islands or the generalised island shapes used to denote the deltas.

Those who showed strong consistency, particularly Roselli and Benincasa, may have worked from a colour pattern. Others perhaps memorised those cases of strong colour association and improvised for the rest. [See Colour Consistency Tables (a Microsoft Word document).]

Some - Vesconte, Roselli and Benincasa for example - seem to have realised the advantage of consistent colouring conventions after the time of their early productions.

Benincasa's work demonstrates that he achieved almost complete consistency, since the occasional differences can be viewed as intentional minor variations, in some cases motivated by a desire for added variety. Indeed, we can learn more from the exceptions, since those show a mind at work rather than just slavish copying. [See Consistency: Benincasa and also Colour Consistency Tables (a Microsoft Word document).]

Benincasa's consistency and distinctive colour conventions allow the confident identification of the few unsigned works that can be attributed to him, or to his successors who continued his style until the mid-16th century. [See Colour and shape on the productions of Benincasa's successors and imitators (a Microsoft Word document).]

Some conventions were widely observed, e.g. the red for Gozo; others were peculiar to the different places of production, particularly Majorca and Venice. The near-contemporary Roselli and Benincasa mark a partial merging of different regional conventions.

Island colouring has previously unsuspected diagnostic value, in terms of determining authorship, place of production and lineage.

The continental coastlines would have been copied by some direct means from the workshop pattern, which would almost certainly have included the largest islands. However, the smaller islands may well have been copied freehand, while the smallest islets were supplied in red with the colourist's brush. The medium-sized islands that form the subject of this analysis might have been either traced or drawn freehand.

Not surprisingly, islands in the Atlantic that were based on myth or uncertain reports, are presented in rudimentary form: circles, crescents, rectangles, trefoils, etc. What seems not to have been commented on before is that most medium-sized islands in the Mediterranean were also given imaginary forms. [See Island shapes as a mnemonic device.]

Examination of the Aegean islands shows that almost all, down to the smallest islet, were accorded a clearly unrealistic, but repeated, 'signature shape', akin to a logo. Where most true island shapes are unmemorably irregular, the symmetrical geometric outline given to each served as a visual mnemonic device - speeding up production, since they could be readily memorised by the draftsman, while simultaneously making it easy for the user to find their 'virtual' way through the archipelago's confusing reality. It was Vesconte who started to introduce these imaginary outlines, which are still discernible in some charts of the late 17th century, although the most extreme fabrication, relating to Limnos with its 'lollipops', is first seen on the 1330 Dalorto/Dulceti chart. The Ionian island of Zakynthos is another good example of a mnemonic shape. [See Explanatory notes and wider implications (Islands - under Zakynthos, Limnos, smaller Aegean Islands).]

That these formalised islands were often given consistent colouring must have reinforced the mnemonic purpose.

In a few cases the mnemonic, 'signature' island outlines co-existed with attempts, often by Venetian practitioners, to achieve a realistic form. However, most chartmakers persevered for long afterwards with the obviously artificial shapes, based on geometric designs or the natural world. In other words, those continued to serve their purpose.

Two islands, the British Isle of Man and Paxos in the Ionian chain, were given the form of a cross. Catalan chartmakers abandoned a realistic outline for the I. of Man in favour of the symbolic cross form, thus demonstrating an active preference for the mnemonic form.[See Island shapes as a mnemonic device, also Islands with a simple geometric form and The features chosen to illustrate the varied portrayal of islands and estuaries on portolan charts up to 1469, Nos 3 & 31.]

For an as yet unexplained reason, the Catalan chartmakers introduced 'lollipop' extensions to the south and east coasts of their own island, Majorca and one out of the north coast of Minorca. [See Majorca (and the next entry for Minorca).]

The imaginary outlines were treated with the same respect by the copyists as those we would consider 'realistic'.

The main river deltas, Rhône, Danube, Dnipro and Nile, were conveyed by means of simple island shapes. Their number, arrangement and colouring often help to point to authorship. [See River deltas.]

Changes in the portrayal of the Danube delta help to confirm the dating of the British Library's late Vesconte atlas. [See Attributions: Vesconte.]

'Name labels', an uncoloured strip left for the island name, which are found by 1330 and were a development out of island colouring, seem to have been another innovation of portolan chartmakers. They were applied to certain islands in the Aegean (to save space), to Majorca (as an alternative to the Catalan stripes), and less frequently to the mythical islands of the north Atlantic, the Ionian Islands, and the island shapes in the Dnipro delta. These necessitated cooperation between the colourist and the scribe, and in the few cases where the edges of such labels were actually inked in, perhaps the draftsman as well - though those would probably represent three stages of production by a single individual in most cases. The use of multiple name labels links three anonymous Venetian atlases (Pujades A 29, 30 & 31). [See Name labels.]

The number and alignment of the red and gold stripes over Majorca followed established conventions and forms part of a chartmaker's signature. [See Majorca's stripes.]

Other diagnostic features, i.e. those that were not universal, can be found in the crosses on Lanzarote, Rhodes and Chios, a white centre given to Tenerife (inferno), and the elaboration of one or more of the mythical circular islands into the form of a button. [See Other colour conventions.]

What appears to us to be a trivial point, not previously commented on, is the unexplained convention that led Dalorto/Dulceti to detach the initial letter of both Aigues Mortes and Damiata (in the Nile delta). Nevertheless, like so many of the very small chart details, this was also the result of a conscious decision by those who perpetuated that custom. It must have formed part of apprenticeship training and helps to distinguish the work of individuals and centres. [See Detached initial letters.]

A similar convention involving the first letter of pinea, shown near Aigues Mortes, proved to link some early-15th-century Venetian works. [See Attributions: Cesanis.]

Attributions

Workshops

The pioneering work by Ramon Pujades in his 2007 Les cartes portolanes in identifying, for unsigned work, their most likely place of production, authorship and approximate dating, underpins my own investigations. Nearly all of his judgements are substantiated - in the case of the Ziroldi group down to their precise dating.

By confirming which anonymous charts are not by any of those whose signed work has survived, it has proved possible significantly to increase the number of unknown practitioners. It is the contention of this analysis that the work of perhaps 25 unnamed chartmakers survives in the form of 30 unattributed works, where Pujades had referred to 'ten or so' individuals. [See Tables of signed, attributed, anonymous and workshop productions (a Microsoft Word document - Table C).]

The main difference between my own findings and Pujades's assertions relates to the existence and nature of supposed workshops (ateliers), and specifically that 'by the end of the fourteenth century there were ateliers specialising in the serial production of low-cost nautical charts' (2007, p.498b). The C&SA found no direct evidence for those, in the sense of operations involving several fully-trained practitioners working together, apart from confirming that there seems to have been some kind of Vallseca atelier. [See What is meant by an atelier? and also Tables of signed, attributed, anonymous and workshop productions (a Microsoft Word document - Table B).]

Instead, it is asserted here that any chartmaker who employed apprentices or assistants would produce charts that were at least partly the result of collaboration. Those unsigned works that appear to be indistinguishable from their signed equivalents have therefore been attributed to the master himself, rather than to a hypothetical atelier. This applied to works assigned by Pujades to supposed ateliers run by Vesconte, Dulceti, Rafel Soler, Ziroldi, Roselli and Benincasa. [See How would the productions of an atelier be different from those of the master himself?]

There is no evidence that all chartmakers had apprentices or even assistants. This could hardly have applied generally to the Venetians, a number of whom were seamen. It should therefore not be assumed that no practitioner worked alone. [See What is meant by an atelier?]

It has always been assumed that the signature on the 1448 Bianco chart should be taken literally, i.e. that he had carried his equipment and patterns with him and drawn it when his galley was in London. This is unlikely and evidence from the chart indicates that it was probably drawn earlier in Venice, with just the signature added in London. [See Andrea Bianco's "London" chart of 1448.]

It is argued that several different workshop models (or patterns) may well have been needed. Besides a complete pattern for 'pouncing' the coastline's nodal points, there could have been others, covering the chart in sections. Those would have been used for transferring the detailed coastal configurations between the headlands, the toponymy and perhaps the colour. [See Stages in the construction of a chart under 'Patterns'.]

In the case of the most prolific 15th-century chartmaker, Grazioso Benincasa, Pujades's claim that he had 'the support of a well organised atelier employing permanent collaborators' (2007, p.497a) is challenged. [See What is meant by an atelier?]

A few of the attributed works are visually or constructionally different from those by the chartmaker to whose workshop Pujades assigns them. They have been treated here as the work of hitherto unknown practitioners, perhaps unauthorised copyists. Insofar as their toponymy reproduces that of the known chartmaker, that would indicate the source of the model used, not the authorship of the work in question. [See How would the productions of an atelier be different from those of the master himself?]

Pujades interprets the 1433 Vallseca contract as evidence of 'serialized, nearly mass production of navigational charts' (2009, p.328). A counter-argument is presented that this was an exceptional arrangement concerning an untypical chartmaker. [See The Vallseca contract of 1433.]

The C&SA's data on the large number of Venetian productions in the 15th century, about half of them anonymous, suggests different working methods might have operated in that city, conceivably along some kind of co-operative lines, but probably not state-sponsored. [See Venetian practice.]

The cheaper charts, that must have formed the vast bulk of what has now been lost, required almost all the time-consuming construction processes involved in producing a de luxe version. There is no evidence of the content or quality being reduced. [See Stages in the construction of a chart and What is the difference between a plain and a de luxe chart?]

We know of no apprentice records involving unrelated Italians, though such indenturing must have occurred. Several Catalan instances are documented but the C&SA findings were not able to conclusively confirm other suggested links. [See Traces of master/apprentice relationships.]

Why are some reliably attributed charts not signed? In three cases this is evidently because the signature was trimmed off along with the chart's neck. For others, some reasons are offered, but several cases remain unexplained. [See To sign or not to sign?]

It is likely that some unsigned surviving works are copies of others, presumably unauthorised. The atlas in Parma, reproducing one of Benincasa's, is a good example. [See Copies and imitations and Colour and shape on the productions of Benincasa's successors and imitators (a Microsoft Word document).]

---

Individual attributions

Pujades has rescued from '15th-century' obscurity the Genoese chart in the Ricciardiana Library in Florence, dating it to the first quarter of the 14th century. It is here argued that this work is indeed very early and may possibly pre-date the 1311 Vesconte chart. [See Genoese first quarter 14th century.]

The C&SA results for the supposed Cresques atelier show a surprising lack of uniformity. [See Cresques workshop.]

The C&SA substantiates the attribution of four unsigned atlases to Ziroldi. They also confirm Pujades's suggested chronological sequence. While Ziroldi was inconsistent in his use of colour, his profile shows an unexplained pattern in the delta colouring in which the four middle works are noticeably different from the two earlier and the three later. [See Ziroldi and also 15th century Italian charts: Part 2 (a Microsoft Word document).]

On the supposed Vallseca atelier, the C&SA results reveal considerable variation between the productions of what appear to be several different people. This seems to indicate lack of close oversight by the person in charge. [See Vallseca.]

The shadowy Catalan chartmaker Rafel Soler, for whom Pujades has been able to identify two works, seems to have provided Benincasa with at least two of his 'signature' conventions. [See Soler.]

One of the three charts attributed to Roselli was almost certainly an autograph production from which the signature was trimmed off (Modena); however, the C&SA findings repudiate the claimed connection of the other two (in Paris and Yale). [See Roselli.]

The attribution to Roselli of the Catalan Estense world map, actually an extended portolan chart, is confirmed and its dating narrowed to 1462-4. [See 'Catalan Estense world map'.]

---

Benincasa

See various references above, and the Conclusions to 'The style and content of Grazioso Benincasa's charts: imitation, innovation and repetition'.

Benincasa was the first to show (in 1468) the west African coast down to Liberia, along with the Cape Verde Islands. However, a toponymic analysis shows that Petrus Roselli, in 1464, had - at least partially - pre-empted Benincasa. This seems not to have been noted before. That Benincasa did not always extend his atlas coverage to the final point of the Portuguese discoveries indicates that he sometimes put logistical convenience ahead of cartographic completeness. [See West African coast]

Top of page

---

Toponymy

New Names

All 1,800 recurring names along the coastline between France and Morocco have been listed and analysed. These include most of the toponyms on a standard portolan chart. [For the full name list see Table of 'Significant Names' (a Microsoft Word document); and for an extended commentary see Innovative Names.]

Perhaps 350 rare or irregular names found up to 1469 may have been excluded. [See 'Rare and irregular names'.]

The baseline for this study is provided by roughly 1000 'Foundation Names', those found initially on Vesconte's earliest works of 1311-1313. 'Significant Names' - those added subsequently - total around 75% of that original selection, and those discarded at some point before 1600 nearly a half. [See Innovative Names: 'General'.]

Over three quarters of the names first found on Vesconte's earliest productions were still there 300 years later. [See Innovative Names: 'General' and the following section on Vesconte.]

The way that Vescontian innovations were added to the corpus of Foundation Names can be seen in The addition of 'Significant Names' to the 31 sections of coastline (tables) (a Microsoft Word document).

'Precursor Names', of which at least 112 have been identified, were those added after the earliest Vesconte treatment (by him or by others) despite having already appeared in two 13th-century portolani or on two charts considered to predate 1311. [See 'Toponymic transmission before 1311 ['Precursor names']'.]

New names can be seen on the work of all the chartmakers who signed and dated their productions up to 1440. [See Innovative Names: 'Introduction'.]

The importance of Dalorto/Dulceti, the Pizzigani brothers and Francesco Beccari is clearly reinforced. What was surprising was the major contribution made by the Oliva family in the second half of the 16th century. [See The addition of 'Significant Names' to the 31 sections of coastline (tables) (a Microsoft Word document) and also 'Oliva'.]

The route by which a new name passed to chartmakers in another centre was erratic, accidental, piecemeal and sometimes delayed by a century or more. The practitioners of 15th-century Palma, Majorca, when they belatedly repeated certain names that were, by then, a standard feature of Italian work, did so, not on a single occasion, but in at least seven small selections. [See 'Toponymic transmission after 1313' and Table E in Second appearance of northern Adriatic names on Italian and Catalan charts respectively (up to 1469) (a Microsoft Word document).]

Not only was there often a delay before a name started to appear on the charts but the second appearance (at least on surviving charts) might come decades later. It was also not uncommon for names to disappear and then reappear perhaps a century later. All of which warns against loading too much weight on the inclusion/omission of a single name without corroboration from its specific history. [See Rare names, toponymic gaps and reintroductions (a Microsoft Word document).]

It has been assumed that, since the charts were provided for trading voyages, toponymic changes would mirror that. However, no bias has been found reflecting the known patterns of trading voyages or the location of the trading entrepôts set up by the Venetians, Genoese or Catalans. Instead, the key seems to have been propinquity, or just chance. The Venetians were toponymically more active in the Adriatic, and the Catalans and Genoese in the sea they shared. [See 'Geographical distribution', and the following section on 'Trading routes'.]

Given differences over time it is hard to generalise about the areas of greater or lesser change. Overall, the Adriatic shows the most toponymic dynamism and parts of the eastern Mediterranean and North Africa the least. Against that, the west coast of Asia Minor and Morocco display more new recurring names than the south coast of Italy. And the west coast of France was, comparatively, the most static of all the 31 sections in the 14th and yet the most dynamic in the first half of the 15th. [See Tables 4 & 5 in The addition of 'Significant Names' to the 31 sections of coastline (tables) (a Microsoft Word document).]

Sometimes a cluster of added names points to a single [unknown] informant. [See 'Trading routes', and the next section on 'Beccari'.]

Perhaps half of the introduced names referred to natural features, which, by definition, were not new. But the same applies to settlements, whose inclusion usually reflected growing relevance rather than creation. [See 'The meaning of names'.]

Names, abbreviated even down to their initial letter, are occasionally found on Portuguese charts for places on Spain's west coast - an indication that, however international the chart market might have been in the Mediterranean, such charts were intended for Portuguese use alone. [See 'Portuguese charts'.]

Top of page

---

Abandoned names

For the evidence behind the following statements see Abandoned portolan chart names and, for the details, The disappearance of 'Significant Names' from the 31 sections of coastline (tables) and The disappearance of 'Significant Names' from the seven main regions(charts & tables) (both Microsoft Word documents).

Over 450 names were removed from the charts before 1600, having first formed part of the recurring pattern of portolan chart toponymy. As a measure of chart dynamism, that represents a little under half of the names found on the earliest works of Pietro Vesconte (1311-13). Nor was it just a case of abandoning 'old' names; roughly half of those rejected had themselves been added to the original selection.

The second half of the 16th century was ruthless in discarding names that were no longer needed. No less than 60% of the names that disappear from the charts over those three centuries vanished in the period 1550-1600.

However, that figure is distorted by the large number of archaic names retained by Benincasa's successors, Freducci and Sideri, up to 1570, long after others had abandoned them [see Note 3 in the general table].

The pace of removal quickened markedly in the 1580s, and the provisional finding that many more disappeared in or close to the first decade of the 17th century suggests that the period 1580-1610 may have been one of mass extinction.

Of the seven broad geographical regions, the Atlantic coasts and those of the Adriatic and Morea saw the largest volume of toponymic disposal. At the other end of the scale, and at about one third of the rate, the long stretch from the southern coast of Asia Minor round to Morocco was the most static.

---

Red Names

For conclusions relating to the comprehensive survey of Red names, carried out during 2013, see here.

Top of page

---

Most of the summary conclusions listed above point to the four main essays (or their accompanying tables):

Wider Implications of the 'Colour & Shape Analysis'

Introductory Notes on Workshops

The style and content of Grazioso Benincasa's charts: imitation, innovation and repetition

Innovative Portolan Chart Names

---

Overview of these pages   |  Portolan Charts Main Menu  |  Colour & Shape Analysis Menu   |  Toponymy Menu

---

HOME (main menu)  |  Tony Campbell articles  |  Portolan charts menu  |  Index   |   Sitemap   |   What this site is about   |   What's New

---