The Medieval and Early Modern Nautical Chart:
Birth, Evolution and Use
MEDEA-CHART
Host Institution: FCIÊNCIAS.ID – Associação para a Investigação e Desenvolvimento de Ciências. Faculdade de Ciências, Universidade de Lisboa – FCUL (Faculty of Sciences, University of Lisbon – FCUL)
Financing institution: European Research Council (ERC-2016-STG), MEDEA-CHART 714033
Principal Investigator: Joaquim Alves Gaspar, Faculty of Sciences, University of Lisbon
ABSTRACT
Of all the technical and scientific developments that made possible the early modern maritime expansion, the
nautical chart is perhaps the least studied and understood. This fact is very surprising as it was through those
charts that the newly discovered world was first shown to the amazed eyes of the European nations. Although the History of Cartography is a well-established academic discipline and old charts have been examined for many years, their detailed technical study is still in its infancy. What is the origin of the preMercator nautical chart, how charts evolved technically over time and how they were used at sea are all critical questions that remain to be answered. I intend to approach these challenges in a truly interdisciplinary
way, by using innovative and powerful tools as a complement to the traditional methods of historical research: analytical cartometric methods, numerical modelling and the examination of the manuscripts through
special lighting. By applying these tools to a large sample of charts of various periods and origins, I aim to
unveil hidden graphic content related to their construction and use, to characterize their main geometric features, to establish meaningful connections with contemporary navigational methods and exploration missions, and to numerically simulate their construction by taking into account the explanations given in the
textual sources. The effectiveness of those techniques has already been demonstrated in my previous studies,
such as in the solution of an historical enigma which had been alive for more than a century: the construction
of the Mercator projection, in 1569. Now, I propose to handle a broader and more complex set of questions,
which has eluded the historians of cartography for even a longer period. The clarification of these issues will
have a ground-breaking impact, not only in the strict field of the History of Cartography, but also in the context of the intellectual history at large.
1
EXTENDED SYNOPSIS
1. Outline & methodology
Emerging suddenly at the end of the thirteen century and exhibiting an unprecedented accuracy – when compared with the contemporary erudite cartography –, the portolan chart is among the most intriguing medieval
artefacts. Its advent has been considered a major turning point, not only in the History of Cartography (Cortesão, 1969, 215-16) but in the history of civilization itself (Magnaghi, 1939, 330). However little is known
about the genesis of these amazing documents, which has been the object of hundreds of studies since Nordenskjold (1897) produced his classical essay on the history of charts. From the time the first portolan charts
of the Mediterranean were drawn to the full adoption of the Mercator projection (late eighteenth century),
there was a span of some five hundred years, during which the world was explored and depicted in nautical
charts with increasing accuracy and detail. Paradoxically, and despite their irreplaceable role in the European
expansion, little is known about how these charts were constructed and how they evolved technically over
time. Two episodes are especially obscure: the birth of the portolan chart, at some uncertain date of the thirteenth century; and the development of the latitude chart, following the introduction of astronomical navigation, near the end of the fifteenth century. The historical importance of these developments greatly transcends the practical use of charts at sea and the conveying of geographical knowledge. In fact the practise of
navigation and the making of nautical charts were key elements in the development of the modern concept of
physical space, where the spherical shape of the Earth gradually acquired an increasing practical relevance.
The process was made possible by the participation of pilots and cartographers who, unlike its predecessors
in the Antiquity, were driven by practical necessity rather than intellectual curiosity. While some early modern university-educated cosmographers and mathematicians did comment on the geometry and use of contemporary charts, they were hardly ever involved directly in their conception and making. This fact partially
explains the scarcity of textual sources and makes the genesis and evolution of the nautical chart a puzzling
and fascinating historical subject. In fact, a clear perspective of the progress in cartographic and navigational
techniques greatly transcends the strict scope of the History of Cartography, as it may shed relevant light into
broader aspects of the intellectual history, such as the role played by medieval and early-modern artisans in
the complex process we conventionally designate by scientific revolution (see Leitão, 2013, 11-39; Sánchez,
2010; Sandman, 2004).
The fact that the earliest extant portolan charts appeared almost fully-formed, combined with the old myth
of the ‘dark ages’ and the absence of contemporary sources explaining how they were made, contributed to
the emergence of alternative hypotheses – some of them bizarre and none founded on historical evidence –,
postulating an earlier origin (Phoenician, Greek, Roman, Byzantine, Arabic, etc.). In the words of Tony
Campbell (1987, 380-90), whose chapter of the History of Cartography contains a summary of the various
conflicting theories, the question of the origin ‘is perhaps the most intractable’. This was the view of the
leading researcher some thirty years ago, just when the first effective attempts were made to apply numerical
techniques to the study of old charts. It is noteworthy that the present opinion of the traditional historians of
Cartography about this issue does not appear to have changed much despite the progress made in the last
years using such techniques. One key factor may
explain their apparent difficulty in acknowledging
the outcome, if not the relevance, of a technical
approach to the study of the old charts: their imperfect knowledge of relevant navigational and cartographical concepts, especially those with a mathematical component.
It is my intention to demonstrate that portolan
charts were indeed medieval creations and that
they were constructed on the basis of information
collected by mariners, with the primary purpose of
supporting navigation.
The application of cartometric techniques to the Figure 1 - Grid of meridians and parallels in Jorge de Aguiar
study of old maps was proposed a long time ago. portolan chart of 1492 (reproduced from Gaspar, 2010, 94).
At the origin of one of the presently used analytical tools is the pioneering work of the geographer Waldo
Tobler (1966), who first proposed a method to determine the mesh of meridians and parallels implicit in a
2
non-georeferenced map, and suggested that such mesh could be used to identify the underlying map projection. However only in the last decades of the twentieth century were those techniques made attractive to researchers by the growing accessibility of computer power and the development of user-friendly applications.
Two early works are to be mentioned, both focused on the origin of the medieval portolan chart: the monograph of Jonathan Lanman (1987), who investigated a possible match between the navigational information
found in medieval portolani and the geometry of contemporary charts, and the dissertation of Scott Loomer
(1987), who assessed the geometry of a number of portolan charts, in order to investigate their evolution and
identify the underlying projection. A more sophisticated approach was pioneered twenty years later, in my
two articles The Myth of the Square Chart (Gaspar, 2007) and Unveiling the mystery of portolan charts
(Gaspar, 2008), where the analytical and modelling tools later formally described in my PhD dissertation
were tested (Gaspar, 2010). In these works a meaningful connection between the geometry of the old nautical charts and the underlying navigational methods, under the influence of magnetic declination, was established for the first time.
Concerning the latitude chart, which was likely developed near the end of the fifteenth century following
the introduction of astronomical navigation, no written sources have survived describing its genesis. Thus the
research on the subject can only resort to the few extant charts of the transitional period, complemented by
the descriptions given in some later sources of the sixteenth century. Although there is an extensive literature
on the Portuguese cartography of the time, dominated by Cortesão & Mota’s Portugaliae Monumenta Cartographica (1966), no technical analysis was ever made of those charts – aside from the one in my own works
–, whose study was almost exclusively focused on dating, authorship, geographical content and historical
context. In my previous work, I have analysed the geometry of the five earliest charts of Portuguese origin,
and was able to draw relevant historical conclusions about how they were made and how they relate with the
navigational methods and the contemporary voyages of exploration (Gaspar, 2010; 2012). Now it is time to
extend this methodology to a much larger scope and over a broader period, in order to establish their genesis and technical evolution over time, under a new light.
1.1. Cartometric analysis
Two main analytical techniques are central to the present research proposal: the determination of the geographical grid implicit in the old charts (which do not usually depict meridians or parallels); and the assessment of planimetric accuracy. With the first technique one aims at assessing the main geometric features of
the charts, mainly by visual inspection. For example (Fig. 1), the counter-clockwise tilt of the geographical
grid implicit in Jorge de Aguiar’s (1492) representation of the Mediterranean clearly reveals the effect of an
eastward magnetic declination on the navigational data used in its construction. With the second technique
(assessment of planimetric accuracy), which is entirely novel in the field, it is possible to extract detailed
information about the cartographic standards of each chart (e.g. the adopted length of the degree of latitude),
as well as about the methods and sources used in its construction. For example, by representing graphically
the distribution of the latitude errors along the coasts of Europe and Africa, as depicted in the Cantino planisphere (1502), it is possible to identify the methods used to represent the various regions and, in some cases,
to determine the origin and date of the
information (Fig. 2). In the present
case the relatively high accuracy of the
latitudes in the northern Atlantic coast
of Africa (from 0º to 30º N), when
compared with other regions, has permitted the establishment of a historical
connection to the astronomical survey
ordered by king John II of Portugal
c.1485, according to Columbus’s testimony (Gaspar, 2010, 152); likewise,
the distribution of the latitude errors in
Figure 2 - Distribution of latitude errors in the Cantino planisphere along
the southernmost stretch of the African the coast of Africa (reproduced from Gaspar, 2010, 151).
Atlantic coast (blue circles, on the left)
suggests that the origin of the information is the exploration voyage of Bartolomeu Dias, made in 1487-88
(Idem, 155). In order to understand how the old charts were constructed, one has first to characterize them
geometrically, and only then, seek to relate the results of the analysis with what is known about the cartographic and navigational methods of the time.
3
1.2. Numerical modelling
A second method which complements the analytical techniques described above is the numerical simulation
of the geometry of the old charts, based on the concept of empirical map projection. The idea was first proposed by Tobler (1977), who used a sample of distances between places in the Mediterranean in order to
simulate the geometry of portolan charts. Central
to his proposal was the implicit assumption that
nautical charts were constructed by transferring to
the plane the distances estimated by the pilots at
sea. In my previous work, I have generalized Tobler’s approach to both distances and rhumb-line
directions, under the effect of magnetic declination (Gaspar, 2010, 73-84). To the effectiveness
of my method much contributed the excellent
observations of magnetic declination made by
João de Castro in the sixteenth century (Castro,
1538), complemented with the outputs of a modFigure 3 – Output of simulation model EMP. Compare with
ern geomagnetic model (Korte & Constable,
Fig. 1.
2005). This was the first time that the geometry of
old nautical charts was simulated numerically, taking into account the navigational methods described in the
textual sources and considering the effect of magnetic declination. Fig. 3 shows the output of my EMP (Empirical Map projections) model, for the Mediterranean basin, to be compared with Aguiar’s chart in Fig. 1.
The comparison demonstrates that the chart’s main geometric features are well explained by the use of uncorrected magnetic courses and estimated distances, plotted on the plane with a constant scale, as if the Earth
were flat. An important detail to note is that the best matches were obtained when attributing a larger weight
(80%) to the magnetic courses over the estimated distances. This fact is consistent with the navigational
practise of the time, when the pilots (correctly) trusted more the courses indicated by the compass than the
distances estimated on the basis of more or less subjective factors. Being able to reproduce the main geometric features of the charts using this methodology is a giant step towards actually understanding how they
were made and evolved over time.
1.3. Special lighting
One of the least understood processes pertaining to the navigational practice in the Middle Ages and early
modern period is the use of charts. Very few visible marks survive (the graphite pencil was unknown and ink
was not used on the expensive parchment), the textual sources are scarce and they only usually explain the
basic operations. We know from the contemporary descriptions that a pair of dry point compasses was used
to measure distances and directions on the chart. Before the advent of astronomical navigation, the problem
of finding the ship’s position at sea was solved by dead reckoning, that is, on the basis of the course steered
and the distance travelled or, when close to land, using bearings measured with the marine compass. Thus all
we may expect to find are pricking marks, and perhaps, some faint traces made by the compass’s points on
the parchment – probably not on the richly decorated charts, which were not carried aboard, but more likely
on the simpler ones. A recent examination of a chart of the sixteenth century has demonstrated how visual
inspection of the parchment through tangential lighting may reveal many of those marks clearly enough.
An additional difficulty is the poor condition of the old charts caused by its long exposure to light and
dirt, very often affecting legibility. An appropriate way to circumvent the difficulty is, as is routinely done in
the study of old parchment or papyri manuscripts, to use multispectral imaging. This method was never applied to the study of old maps until very recently, with the analysis made by Chet Van Duzer of the YaleMartellus world map of c. 1492, which produced excellent results (Duzer, 2015).
2. Objectives & challenges
The main purpose of the present project is to solve a series of critical questions which have eluded the historians of cartography for a very long time, pertaining to the birth, evolution and use of pre-Mercator nautical
charts.
The most important questions that I intend to handle are: i) when, where and how were the earliest portolan charts constructed? How were portolan charts updated? ii) How and when were the first latitude charts
developed, following the introduction of astronomical navigation? iii) How did the latitude chart evolve
technically during the sixteenth century? iv) How did cosmographers, cartographers and pilots deal with the
known geometric inconsistencies of the early modern charts, caused by magnetic declination and the implicit
4
assumption of a flat Earth? v) How were nautical charts used at sea? These are among the most relevant open
questions in the History of Cartography and Navigation. What I propose is systematically apply the methodology described next to a large and carefully chosen set of charts of both periods, in order to produce groundbreaking answers to those questions.
The research will focus on two different, yet closely related, cartographic objects and periods: the portolan chart of the Mediterranean and Western Europe (from c.1250 on); and the latitude chart of the Atlantic
(from c.1500 on). In the following paragraphs an overview of the research methodology is presented.
2.1. Study of charts
The first step will be to choose the charts to be analysed. Most of them will be studied using the high resolution digital copies available in libraries and museums; others will have to be inspected on-site. Owing to the
large number of extant charts and the limited man power, the choice of those that will be subjected to analysis is a non-trivial decision that will be made taking into account date, origin, geographical coverage, historical relevance and availability, as explained in Part B2. Two subsets are to be formed: one of portolan charts
(e.g. those not incorporating observed latitudes); and another of latitude charts. Both subsets will include,
whenever possible, all the earliest exemplars of their kind, as well as those showing signs of innovation or
technical improvement.
Each chart will be subjected to a sequential set of analytical procedures: i) survey of literature and reassessment of authorship and dating, when unknown or problematic; ii) survey of cartographic content and
place names; iii) georeferencing and qualitative assessment of geographical grid (as illustrated in Fig. 1); iv)
quantitative assessment of planimetric accuracy.
Numerical simulations of the geometry of the charts will then be made using the methodology described
above. For the model EMP to be run the following input data are necessary: i) the maritime tracks representative of the routes used in the construction of the charts (known or assumed); ii) the spatial distribution
of the magnetic declination in the region, at the time the information was collected (as given by contemporary data or produced by modern geomagnetic models). Several runs of the model will be made for different
values of the relevant parameters: maritime tracks; spatial distribution of magnetic declination (as a function
of date and source); cartographic models adopted to represent the various regions (e.g. portolan or latitudetype), etc. In this experimental phase of the research, the various parameters of the model will be fine-tuned
in order to simulate the geometry of a given chart or group of charts, always in the light of the construction
methods described in the sources.
At this stage I expect to be able to fully characterize the geometry of the charts under analysis and its evolution over time, as well as to provide answers to some questions related to the details of the construction: for
example, the possibility of a piecewise combination of various regional representations (portolan charts) or
the possible connections with specific voyages of exploration (latitude charts). In the case of portolan charts,
the comparison between the results obtained in the cartometric analysis with the outputs of the EMP model
for different spatial distributions of magnetic declination (which vary with time) will make possible to produce a better estimate for the date of the earliest charts. In general, a good match between the geometric
features of the charts under analysis and the model outputs, corroborated by sound historical facts and reasoning, will validate the assumptions made about the construction method, origin of the information, dating,
etc.
Much is to be expected of the use of special lighting in the examination of the charts, despite the novelty
of the approach. The easiest technique – visual inspection under tangential light – is to be applied routinely
to all originals chosen for analysis. The aim is to detect marks of use and construction (wax stains, pricking
marks and traces), to be interpreted in the light of the navigational and cartographic methods of the time.
When deemed necessary, the manuscripts will be examined under ultra-violet lighting, in order to unveil
unreadable text and graphic content. The use of multispectral imaging will be reserved to carefully chosen
cases, whenever it is considered that the expected results are of vital importance to the project’s goals, thus
justifying the high cost of the method.
2.2. Study of textual sources
A number of textual sources describing the medieval and early modern cartographic and navigational methods will be re-visited and re-interpreted in the light of the project’s goals. Important examples are the manuscript of c.1200 known as the Liber de existencia riveriarum, containing the oldest known portolano and the
earliest reference to chart making (see Gautier-Dalché, 1995; Pujades, 2007), Benedetto Cotrugli’s De navigatione, c. 1465 (see Falchetta, 2009), the earliest known navigation treatise, some earlier Italian portolani
(such as Lo Compasso da Navigare, thirteenth century) as well a large number of treatises of the sixteenth
5
century discussing the chart’s geometry or describing how they were constructed and used (Pedro Nunes,
Martín Enciso, Alonso de Chaves, Martín Cortés, William Bourne and many others). Although the focus of
this research is on the charts, a systematic survey of all relevant texts is crucial to a complete clarification of
the topic.
2.3. Dissemination of results / outreach
The results of the project will be routinely disclosed through publications and conferences/lectures. Article
and paper proposals will be submitted to top level journals and forums if the fields of History of Cartography, History of Science and History of Navigation. An average number of one article and two paper
presentations per year per member is expected, from the second year on. A collective book will be prepared
and proposed to an international editor, aiming to complement the History of Cartography (The University
of Chicago Press) series with the results of the project. Eight international meetings dedicated to the project’s
subjects will be organized. During the last one an unconventional public exhibition is planned, in which the
geometry, construction and use of old nautical charts will be illustrated using digital techniques.
3. The research team
The research team will be composed by six members, two of them resident members of my research centre
(CIUHCT): Joaquim Alves Gaspar (the PI) and Henrique Leitão (Senior Researcher, SR). Owing to an uncommon combination of skills in the fields of History of Cartography, mathematical cartography, marine
navigation and hydrographic surveying, complemented by considerable experience in leadership, mentoring
and project management, I consider that I have the ideal profile to lead this research. Furthermore the SR and
I form a solid multidisciplinary core to the project: while my main expertise focuses on the technical aspects
related to navigation and cartography, he is a leading historian of early-modern science in Europe.
The remaining four members of the research team will be selected through international calls, two as
postdoctoral fellows and two as PhD students. All members will be trained in the specific tasks of the research, including the use of the cartometric and modelling techniques. The choice of the four external members is a critical step for the success of the project. I am confident that the care I intend to put into the preparation of the call and the evaluation process, the prestige of the ERC grants and the challenge posed by the
revolutionary approach of this research will guarantee a sufficiently large number of excellent candidates.
4. Relevance of the proposal
In previous works I have developed and applied innovative analytical and modelling techniques to the study
of a number of old nautical charts. The results of my work, which have been presented in the leading international journals and forums, have shown emphatically how effective and powerful those techniques are. In
one of the recent studies, I have solved an historical enigma which had been alive for more than a century:
the construction of the Mercator projection, in 1569 (Gaspar & Leitão, 2013; Leitão & Gaspar, 2014).
Now I propose to handle a broader and more complex set of questions which have eluded the historians of
cartography for even a longer period: the origin and evolution of the pre-Mercator nautical chart. By approaching these issues with such a novel and powerful methodology I aim to shed considerable light on the
subject and solve some long standing historical puzzles. Moreover I intend to demonstrate the effectiveness
and potential of my tools for solving other challenging problems, which will contribute to their widespread
use and change the discipline in a profound way.
The impact of this research is anticipated to greatly transcend the strict field of the History of Cartography, owing to its broader implications in the History of Science and the intellectual history at large. A good
example of such implications is the role played by early-modern pilots, cartographers and cosmographers in
the development of the modern concept of physical space and in the complex process that culminated in
what we designate conventionally by Scientific Revolution.
It is also important to stress an implicit, yet vital, goal of my proposal: to make a significant contribution
to the training of a new generation of historians of Cartography, prepared to apply interdisciplinary techniques to the study of old maps. In the last years I was very successful, through my individual work, in
demonstrating to the international community the effectiveness of the methods I have pioneered. Now the
time is right to disseminate this know-how and begin the slow process of making it canonical to the study of
old maps and charts. The expectation of such realization, which can only be nurtured by the support of an
ERC grant, is a very strong driving force behind my proposal.
Being able to combine effectively the interdisciplinary methods proposed in this project, in order to
produce results whose historical relevance greatly transcends the scope of its subject matter, will be an
eloquent way of demonstrating that the reality described by Charles Percy Snow in 1958 is beginning to
change, after all.
6
List of references
1. CAMPBELL, Tony (1987), ‘Portolan Charts from the Late Thirteenth Century to 1500’, in Woodward &
Harley (ed.), The History of Cartography, Volume One. Chicago & London: the University of Chicago
Press, 371-463.
2. CORTESÃO, Armando & MOTA (1987), Teixeira, Portugaliae Monumenta Cartographica, 6 vols,
facsimiled edition. Lisboa: Imprensa Nacional-Casa da Moeda [1960].
3. CORTESÃO, Armando (1969), History of Portuguese Cartography, Vol. I. Lisboa: Junta de Investigações do Ultramar.
4. CASTRO, João de (1958), ‘Roteiro de Lisboa a Goa’, in Cortesão & Albuquerque, Obras Completas de
D. João de Castro. Coimbra: Academia Internacional de Cultura Portuguesa, Vol. II: 1-169 [1538].
5. DUZER, Chet Van (2015), “Let there be multispectral light”: imaging the world map by Henricus Martellus at Yale, c. 1491, unpublished paper given in the International Conference for the History of Cartography, Antwerp.
6. FALCHETA, Piero (2009), ‘Il Trattato de Navigatione di Benedetto Cotrugli (1464-1465). Edizione
Commentata Del Ms. Schoenberg 47 Con Il Testo Del Ms. 557 di Yale’, Studi Veneziani, 15-333.
7. GASPAR, Joaquim Alves (2007), ‘The myth of the square chart’, e-Perimetron, 2, 2: 66-79.
8. GASPAR, Joaquim Alves (2008), ‘Dead reckoning and magnetic declination: unveiling the mystery of
portolan charts’, e-Perimetron, 3, 4: 191-203.
9. GASPAR, Joaquim Alves (2010), From the Portolan Chart of the Mediterranean to the Latitude Chart
of the Atlantic: Cartometric Analysis and Modeling. Universidade Nova de Lisboa: unpublished PhD
dissertation.
10. GASPAR, Joaquim Alves (2012), 'Blunders, Errors and Entanglements: Scrutinizing the Cantino Planisphere with a Cartometric Eye'. Imago Mundi, 64, 2: 181-200.
11. GASPAR, Joaquim Alves & LEITÃO, Henrique (2013), ‘Squaring the Circle. How Mercator Constructed His projection in 1569’, Imago Mundi, 66, 1: 1-24.
12. GAUTIER-DALCHÉ, Patrick (1995), Carte marine et portulan au XIIe siècle. Le Liber de Existencia
Riverierarum et Forma Maris Nostri Mediterranei (Pise, circa 1200). Rome: École Française de Rome.
13. KORTE, M. & CONSTABLE, Catherine (2005), ‘Continuous geomagnetic field models for the past 7
millennia: 2. CALS7K’, Geochemistry, Geophysics, Geosystems, 6, 1.
14. LANMAN, Jonathan (1987), On the origin of Portolan Charts. The Hermon Dunlap Smith Centre for
the History of Cartography. Occasional Publication No 2.
15. LEITÃO, Henrique (2013), ‘Um Mundo Novo e uma Nova Ciência’, in LEITÃO, Henrique, 360º Ciência Descoberta (catalogue of the exhibition of the same name, Lisbon, 2013). Lisbon: Fundação Calouste Gulbenkian, 15-39.
16. LEITÃO, Henrique & GASPAR, Joaquim Alves (2014), ‘Globes, Rhumb tables, and the Pre-History of
the Mercator Projection’, Imago Mundi, 66, 2: 180-195.
17. LOOMER, Scott (1987), A Cartometric analysis of Portolan Charts: a Search for Methodology. PhD
dissertation. Madison: University of Wisconsin.
18. MAGNAGHI, Alberto, ‘Nautiche carte’ (1929-39), in Encyclopedia italiana di scienze, lettere ed arti.
Rome: Istituto Giovanni Trecanni, 24: 323-31.
19. NORDENSKIÖLD, Adolf (1897), Periplus: an Essay of the Early History of Charts and SailingDirections, Trans. Francis Bather. Stockolm.
20. PUJADES. Ramón (2007), Les Cartes Portolanes. La representació medieval d’una mar solcada. Barcelona: Institut Cartogràfic de Catalunya.
21. SÁNCHEZ, Antonio (2010), ‘La voz de los artesanos en el renascimiento científico: Cosmógrafos y
cartógrafos en el prelúdio de la “Nueva Filosofia Natural” ’, in Arbor. Ciencia, Pensamiento y Cultura,
743: 449-460.
22. SANDMAN, Alison (2004), ‘An Apologia for the Pilot’s Charts: Politics, Projections and Pilot’s Reports in Early Modern Spain’, Imago Mundi, 56, 1: 7-22.
23. SNOW, Charles Percy (2001). The Two Cultures. London: Cambridge University Press [1959].
24. TOBLER, Waldo (1966), ‘Medieval Distortions: the Projections of Ancient Maps’, Annals of the Association of American Geographers, 56, 2: 351-60.
25. TOBLER, Waldo (1977), ‘Numerical Approaches to Map Projections’, in Kretshmer, E. (ed.), Studies in
Theoretical Cartography. Viena: Deuticke, 51-64.
7
PROJECT DESCRIPTION
1. State of the art & objectives
The main purpose of the present project is to solve a series of crucial questions which have eluded the historians of cartography for a very long time, pertaining to the birth, technical evolution and use of nautical
charts during the Middle Ages and early modern period. This goal is to be accomplished using innovative
interdisciplinary techniques – cartometric analysis, numerical modelling and the examination of the manuscript charts under special lighting – which will complement the traditional methods of historical research.
With the introduction of such techniques I intend, not only to solve long-standing problems, but also to
change the discipline in a profound way, by opening new and exciting opportunities for research. The most
important historical issues that I intend to clarify are:







When and where were the first portolan charts of the Mediterranean produced? Although there is presently some consensus for a medieval Italian origin, the supporting evidence is scarce and alternative explanations, occasionally bizarre, continue to find their way into the specialized literature.
How were the first portolan charts of the Mediterranean constructed? Although the few surviving textual
sources, as well as the geometry of the actual charts, indicate that they were drawn using navigational
data collected by pilots at sea, such evidence needs to be reinforced by a systematic and closer look into
the contemporary texts and, most especially, into the geometric features of the charts. Whether they are
piecewise constructions of various regional representations, as some suggest, or they were conceived as
single cartographic units, is uncertain.
How were portolan charts updated with new geographical information? The answer to this question is
particularly relevant for the earliest stages of chart evolution, when the representations of the Mediterranean, Black Sea and western African coast improved and consolidated. Being able to demonstrate that
the updating process was identical to the one used to construct the earliest charts will have a vital importance for the solution of the origin problem.
How and when were the first latitude charts constructed, following the introduction of astronomical navigation? Although a few early Portuguese charts are extant from the transitional period (late fifteenth and
early sixteenth centuries), no textual sources survive describing the process, which was most likely conducted by pilots and cartographers.
How did the latitude charts evolve technically? We know that, in the earliest exemplars, not all regions
were represented according to the latitudes of the places and there were marked scale differences among
regions. By systematically examining the geometry of a sample of charts distributed over a long period I
intend to clarify how cartographic accuracy improved over time.
How did cosmographers, cartographers and pilots deal with the geometric inconsistencies of the early
modern charts, caused by not considering the curvature of the Earth – as a result of the limitations imposed by the navigational methods – and by the effect of a space and time-varying magnetic declination?
I intend to clarify this question by combining the results of cartometric analysis with the information collected in contemporary texts: navigation treatises and rutters.
How were nautical charts used by the pilots at sea? No medieval sources are extant explaining the use of
portolan charts in the Mediterranean. Concerning the early modern period, only the basic operations are
usually documented in the sources. This issue is to be clarified by a systematic survey and analysis of the
textual sources, complemented by the physical examination of the charts, looking for marks of use.
These are among the most important open questions in the fields of the History of Cartography and History
of Navigation. The fact that no satisfactory solutions have been found so far is explained not only by the
scarcity of textual sources, but mostly by the difficulties posed by the complexity of the problems, requiring
considerable expertise in unrelated specialized fields, such as medieval and early-modern history, mathematical cartography and marine navigation. The impact of this research is expected to greatly transcend the
scope of History of Cartography expanding to related areas such as the History of Science and the intellectual history. Moreover it is my intention to contribute to the revitalization of the discipline through the training
of a new generation of historians of Cartography, prepared to apply consistently those interdisciplinary techniques to the study of old maps and charts.
a0. Introit
Some three hundred years separate the development of the medieval portolan chart, in some uncertain date of
the thirteenth century, from the construction of the Mercator projection, in 1569. Moreover, owing to its
8
incompatibility with the navigational methods of the time, only near the end of the eighteenth century – after
the longitude problem was solved and the spatial distribution of magnetic declination was known – was the
old latitude chart definitely replaced by the novel projection. For five centuries, an apparently naïve method
of cartographic representation, in which compass courses and estimated distances were transferred directly to
the plane with a constant scale, not taking into account the sphericity of the earth, served marine navigation
and played a fundamental role in European discovery and expansion. Paradoxically little is known about how
and when the earliest nautical charts were created, and how they evolved technically over time. Two episodes are particularly obscure: the birth of the Mediterranean portolan chart, which is considered by most
historians to have occurred during the thirteenth century, following the introduction of the marine compass in
the Mediterranean; and the development of the latitude chart (or ‘plane chart’), following the introduction of
astronomical navigation, during the second half of the fifteenth century. A particular circumstance appears to
have contributed to our fragmentary knowledge about the subject, making the evolution of the pre-Mercator
nautical chart an even more puzzling historical subject, which is the fact that such evolution was mostly conducted by artisans: pilots and cartographers. While mathematically-trained scholars may have participated in
the conception of the first portolan charts (the subject is still controversial, see Pujades, 2007, 515-18) and
university-educated people did comment on the geometry and use of charts in the early modern period (for
example, the mathematician Pedro Nunes, in 1537), there is no evidence that any of these scholars were directly involved in the making of the actual charts before the middle of the sixteenth century. The situation
only began to change with the creation of the positions of cosmógrafo-mor (in Portugal) and piloto-mayor (in
Spain), a royal decision which triggered an unprecedented collaboration between two very different cultural
and social groups (Leitão, 2006; 2007).
The research on pre-Mercator nautical cartography is usually split into two separate subjects and periods:
the portolan chart of the Mediterranean (c.1250 on) and the latitude chart of the Atlantic (c.1500 on). However it should be stressed that the two cartographic models co-existed in all nautical charts of the sixteenth
century, where the Caribbean Sea (up to c.1525) and the Mediterranean (up to c. 1600) continued to be represented according to the old portolan-type model.
In the following paragraphs a more detailed review is made of relevant literature on the three research
subjects: the portolan chart, the latitude chart and the use of charts at sea. Whenever appropriate, highlights
are made in bold type, indicating my intention to address particular issues. At the end of each part a short
summary is made of the most relevant points in the present state of knowledge (What we know), followed by
a list of research objectives (What I propose). These objectives are to be pursued through a detailed work
program, which is described in the next section (Methodology).
a.1. The portolan chart
There exists a vast and rich bibliography on the history of the medieval portolan charts, dating from the
eighteenth century to the present day, which addresses the fundamental questions concerning when, where,
who, why and how the earliest charts were drawn. The contemporary literature is dominated by two major
scholarly works: Tony Campbell’s Portolan Charts from the Late Thirteen Century to 1500 (1987), where a
review is made on the various conflicting theories pertaining to the genesis and construction of the first
charts; and Ramón Pujades’s Les Cartes Portolanes (2007), where a detailed analysis is presented about the
historical context in which the medieval nautical cartography was likely born. A great care is put by Campbell in his approach to the birth and construction of portolan charts, a care that is emphatically expressed by
his statement: “among the research problems connected with the portolan charts, the question of their origin
is perhaps the most intractable”. Although a larger emphasis of his historiographic review is clearly put on
the medieval origin theory, Campbell expresses many doubts about related issues, such as the use of the
magnetic compass and the adoption of a map projection in the construction. That is not the case of Pujades,
who considers that the earliest nautical charts could not have been produced before the beginning of the thirteenth century, when some mathematical developments took place in Europe (Idem, p. 515). A rediscovered
manuscript portolano (that is, a pilot’s rutter) from c.1200, the Liber de existencia riveriarum, referring to a
nautical chart which once accompanied the text (Gautier-Dalché, 1995), represents for him a strong confirmation of the close relationship between the sailing directions and the charts, as well as of the involvement
of seamen in their making. However, owing to the difficulties associated with the interpretation of a corrupted text written in Medieval Latin, Pujades and Gautier-Dalché disagree on whether the rutter contained in the
manuscript was made from an existing chart or the other way around (Gautier-Dalché, 2012). The text will
be re-visited and carefully examined in order to clarify this issue. Concerning where the first charts were
made, Pujades points to a specific coastal region between Pisa and Genoa, and to a milieu of ecclesiastic
scholars in contact with the world of maritime trade and navigation, where the first prototypes may have
9
been prepared during the first quarter of the thirteenth century (Pujades, 2007, 515). As to the how they were
made, Pujades clearly supports the mainstream hypothesis that charts were drawn using navigational information collected by the pilots and that the marine compass played a fundamental role in the process. This is
an important a-priori hypothesis of this project, to be confirmed by carefully comparing the geometric
features of the charts, as determined by cartometric analysis, with the results of numerical simulations.
Based on the regional scale differences exhibited by portolan charts, several authors have suggested that
they are a piecewise construction of local representations. While Campbell takes note of this hypothesis
without expressing his opinion (Campbell, 1987, 387-88), Pujades is clearly against it. For him, those differences result from the errors introduced by the expansion of the area covered – initially only the Mediterranean basin – to the Black Sea and western Europe (Pujades, 2007, 511). However, this interpretation is hardly
compatible with the fact that scale variations also exist among Mediterranean sub-basins, as noted by others
(for example, Loomer, 1987, 159-165). This subject will be clarified by first identifying the different
regions on the basis of scale variations, and then simulating numerically how regional representations
may have been assembled to produce a chart of the whole region.
The counter-clockwise tilt of the Mediterranean basin in all portolan charts up to c.1600 (see Fig, 1 in
Part B1) was noted by many authors, who produced different estimates of the rotation angle, ranging from 6
to 10 degrees (Campbell, 1987, 384-85). However few ventured into establishing a causal connection between the tilt and magnetic declination in the region because no reliable paleomagnetic data were available at
the time those studies were made. That was also the case of Campbell, who suggested that a better
knowledge of both the distortions of the charts and the value of magnetic declination during the thirteen century in the Mediterranean would permit to ‘place a firmer date on the initial compilation of the portolan
charts’ (Idem, 384-85). In this respect an important point was made by Lanman (1987, 23-32), who compiled
the average skewing of a sample of charts from c.1300 to 1662, and represented the values graphically as a
function of time. The results show that the average skewing of the charts remained approximately constant
up to about 1570, when it dropped swiftly to near zero, in the beginning of the seventeenth century. Lanman
concluded that such changes were the reflex of the secular variation of the magnetic declination in the region. However, the paleomagnetic data available to Lanman was too poor to support any firm conclusions on
this matter. In a previous work (Gaspar, 2008), I used his chart data together with the values of magnetic
declination produced by a modern geomagnetic model and was able to show that magnetic declination did
vary significantly between 1200 and 1600, while the average skewing of portolan charts was kept more or
less constant up to about 1570. Figure 1 illustrates the tilt of the charts (open circles) together with the variation of magnetic declination in the western and central Mediterranean, as given by the model of Korte &
Constable (2005). This result demonstrates that charts with a tilt close to 9 degrees continued to be used in
the Mediterranean well after magnetic declination has decreased to much smaller values. That is to be expected since new charts were routinely copied from the existing prototypes and the phenomenon of magnetic
declination was unknown at the time. Any differences between the directions indicated by the compass and
true North were considered to be the result of poorly magnetized needles, and was usually corrected by the
compass maker. This operation
was strongly disapproved by
the Portuguese pilot João de
Lisboa, in the beginning of the
sixteenth century, who specifically referred to the marine
compasses made in Flanders
and Genoa (Albuquerque,
1982, 14). Notice that such
kind of manipulation, if made
by pilots, would have eliminated most of the discrepancies
between the directions indicated by the charts and those given by the compasses, thus ex- Figure 1 – Secular variation of magnetic declination in four places of the central and western Mediterranean, in degrees (colored lines), together with the average tilt of some
plaining why it took so long to portolan charts from c. 1285 to 1600 (open circles). The large grey circle near the top left
be fixed (Gaspar, 2008, 197). corner marks the part of the graph where the average value of magnetic declination
Therefore the average skewing matches the average tilt of the oldest charts, suggesting that the earliest prototypes may
of the oldest charts illustrated have been produced between 1200 and 1250 (adapted from Gaspar, 2010, 98).
10
in Fig. 1 – a little more than 9 degrees – may reflect the average value of magnetic declination at the time the
earliest charts were produced. By intersecting the horizontal nine degree’s line with the lines representing
the secular variation of magnetic declination in the western and central Mediterranean, we find that those
prototypes may have been produced between 1200 and 1250. This type of analysis will be pursued with a
much larger number of charts and in a more systematic way, by replacing the ad-hoc skewing angles
measured by Lanman by a series of values determined in different regions, and then comparing the
results with the local values of magnetic declination given by a geomagnetic model. This way I expect
to improve, and make more robust, the estimate made above for the date of the earliest charts.
The possibility of a precise match between the courses and distances registered in the portolani and the
geometry of early portolan charts, hypothetically based on them, was investigated by Lanman (1987). This is
the earliest study where a systematic geometric analysis of a sample of charts was made. Although the theory
of Lanman – who tried to reproduce the coastlines of two charts by transferring to the plane the courses and
distances between adjacent ports, as registered in two portolani – is clearly an oversimplification, his research was a very important step in the right direction. A more powerful approach was proposed some twenty years later, in two of my articles (Gaspar, 2007; 2008), when innovative cartometric analysis and numerical modelling methods were tested for the first time. In these works a meaningful connection between the
geometry of the old nautical charts and the underlying navigational methods, under the influence of magnetic
declination, was first established. Two novel techniques were tested: the assessment of the geometric features of the charts by examining their implicit grid of meridians and parallels; and the numerical simulation
of their geometry, on the basis on the contemporary navigational methods and taking into account the effect
of magnetic declination. In these works I have concluded that the gross geometric features of the portolan
charts are well explained by the medieval cartographic methods, which consisted in transferring directly to
the plane of the chart the uncorrected magnetic directions and estimated distances observed at sea. In this
project I propose to pursue the same line of research by extending the analysis to a much larger sample of charts of different periods, in order to assess their technical evolution and confirm the a-priory
assumptions on the construction methods.
Another pioneering work in which cartometric techniques were used is the one of Loomer (1987), who
made an extensive analysis of 26 charts, from 1339 to 1508, trying to assess technical evolution and identify
the underlying map projection. The author found a high degree of correlation between the geometry of the
charts, and both the equirectangular and the Mercator projection, and concluded that the charts were likely
constructed using loxodromic course bearings, rather than distances. However he emphatically stressed that
no map projection was consciously employed in the construction, which was likely based in some kind of
triangulation. Loomer further concluded that the methodology for chart construction was fixed early in its
history, with no apparent improvement or degradation in cartographic accuracy over time, and that charts
were likely developed from a series of regional representations tied together (Idem, p. 168-9). An unexpected
finding of Loomer is that the tilt of the Mediterranean axis on the charts could not be the result of magnetic
declination. However his conclusion was negatively affected by the lack of reliable information on the distribution of magnetic declination, not available at the time the study was made. In this project I will use the
spatial distribution of magnetic declination in ancient times given by modern geomagnetic models,
such as the one described by Korte & Constable (2005), based on paleomagnetic data, and the one by Jackson et al. (2000), based on historical observations.
More recently Roel Nicolai (2014) completed a study of a series of portolan charts and concluded: first,
that the charts were made by assembling some five or six regional representations; second, that those representations were based on data collected by means of accurate geodetic surveys and were drawn using the
Mercator projection; and third, that the medieval navigational methods cannot explain the high level of planimetric accuracy of the charts. His proposed explanation is that the medieval portolan charts are copies of
older prototypes, probably of byzantine origin. However the author does not present any piece of historical
evidence for such claim and does not describe the geodetic methods allegedly used to perform the surveys.
Furthermore he fails to explain how the improvements made in the portolan charts, most especially during
the fourteen and fifteenth centuries, could be possible using the same unsatisfactory methods. Two reviews
of Roel’s work, refuting his theories, were made recently by Tony Campbell and myself (Gaspar, 2015c, 2024; Campbell, 2015, 25-27).
The careful and somehow pessimistic view of Tony Campbell made thirty years ago, about the historiographic difficulties associated with the origin problem, was justified by the lack of sufficiently clear historical evidence supporting any of the theories of the time. However the situation has evolved significantly since
then, as a result of three important developments: i) the rediscovery of the manuscript Liber de existencia
11
riveriarum, in which a clear connection is made between charts and portolani, and the involvement of seamen is explicitly mentioned (Gautier-Dalché, 1996); a better knowledge of the spatial distribution of magnetic declination in the area, which made possible to associate the tilt of the Mediterranean in all charts up to
1600 with its average value in the region, during the thirteen century (Gaspar, 2008); and the use of analytical and modelling technics in the study of the charts, which has already produced important results concerning the close relationship between their geometric features and the navigational data used in their construction (Gaspar, 2008; 2010).
Of the various conflicting theories reviewed by Campbell, the one asserting that the first charts were created during the Middle Ages and that no system of projection was explicitly used in its construction –which
was based on navigational data collected by pilots at sea – clearly prevails today. This interpretation is compatible with the geometry of the actual charts and corroborated by the counter-clockwise tilt of the Mediterranean basin, which reflects the average value of the magnetic declination at the time the earliest prototypes
were likely produced. Still outlier theories continue to find their way into the specialized literature, such as
the recent one by Nicolai (2014), mentioned above. This is a sign that our state of knowledge about fundamental questions pertaining to the origin and method of construction of portolan charts is still fragile and
fragmentary, despite the recent developments.
What we know:
- Three undated and unsigned portolan charts – the earliest known – are extant from about 1300: the Carte
Pisane (Bibliothèque nationale de France), the Cortona chart (Biblioteca dell'Accademia Etrusca di Cortona) and the Lucca chart (Archivio di Stato di Lucca). Their accuracy and detail suggest that chart’s
construction in the Mediterranean must have initiated at an earlier date, at least some decades before;
- Portolan charts could not have been created before the beginning of the thirteenth century, when some
mathematical developments took place (for example, the introduction of Arab numerals);
- A Latin manuscript of the beginning of the thirteen century is extant, the Liber de existencia riveriarum,
referring to a nautical chart and how information provided by ship’s officers and their portolani was used
in its construction;
- All portolan charts before c.1600 represent the Mediterranean with a counter clockwise tilt of about 8º to
10º. This feature is consistent with what we know about the spatial distribution of the magnetic declination in the region during the thirteenth century;
- The main geometric features of all portolan charts are invariant, which is an indication that the methodology for chart’s construction was fixed early in their history;
- Scale variations among Mediterranean sub-basins suggest that portolan charts are piecewise constructions
of regional representations;
- Improvements in cartographic accuracy and detail took place in the beginning of the fourteenth century,
in the representation of the Mediterranean, Black Sea and western Europe, and during the fifteenth century, in the representation of the western coast of Africa. After that, there is no evidence of any significant
improvement or degradation in charts’ accuracy over time;
- Numerous references to nautical charts were found in the inventories of ships, together with references to
marine compasses, sandglasses and dividers. This demonstrates that charts were usually carried on board.
In most navigational treatises, including Cotrugli’s De Navigatione (1465), references are made about the
use of charts on board.
What I propose:
- Specific objective 1.1: confirm (or disprove) that portolan charts were created to be used in navigation,
using navigational data collected by pilots;
- Specific objective 1.2: produce detailed evidence that the geometry of portolan charts reflects the spatial
distribution of magnetic declination in the region, when the data for the earliest charts were collected;
- Specific objective 1.3: determine more precisely when the very first prototype charts were produced;
- Specific objective 1.4: clarify whether the earliest prototypes were or not constructed by assembling regional representations;
- Specific objective 1.5: clarify how charts were updated and improved on the basis of new geographic
information.
a.2. The latitude chart
The introduction of astronomical methods, during the second half of the fifteenth century, is consensually
considered as a major technical breakthrough in the history of navigation. In my opinion its historical impact
greatly transcends that specific field. Being able to determine the position of a ship at sea by observing the
12
sky involves a dramatic change of paradigm: from one implicitly assuming a flat Earth, where all directions
and distances measured on its curved surface can be transferred directly to the plane with minimal errors; to
another, where the sphericity of our planet is deeply embedded in the very method used to find the ship’s
position and represent it on a chart. To my knowledge, this is the earliest known instance of a long intellectual process during which the shape of the Earth – which was known since antiquity – gradually acquired an
increasing practical relevance, not only in the technical fields of navigation and cartography but also in everyday life. In 1486 Christopher Columbus proposed to the Spanish monarchs Fernando and Isabella to reach
the Indies by sailing west, arguing that the voyage was not only possible but also shorter than the route followed by the Portuguese. Central to his arguments was a certain assumption about the size of the Earth,
which he considered to be much smaller than the model adopted by his contemporaries. This is the first historical episode where the shape and size of our planet acquired a vital political and strategic relevance. Another important one was Magellan-Elcano circumnavigation of 1519-22, which was prepared and executed
with the explicit purpose of demonstrating that some spicy islands in the other side of the world (the Moluccas) were located in the Spanish hemisphere, according to the terms of the treatise of Tordesillas (1494).
However the impact of this process onto the actual charts was long and painful. The history of nautical
cartography between 1500 and 1750 reflects a hard struggle between the intent of fully incorporating the
shape and size of our planet into the charts, championed by the cosmographers, and the practical impossibility of doing so because of the incompatibilities with the contemporary navigational methods, as claimed by
pilots. Some conflicts are known between the two groups, both in Portugal and Spain, with the cosmographers complaining about the ignorance and stubbornness of the pilots, who systematically refused to apply
their alleged improvements to charts (Sandman, 2001, 2004; Sánchez, 2013, 238-61). Incidentally the reason
was most often on the side of pilots, who were the sole responsible for the effectiveness and safety of navigation (Gaspar, 2014). Thus the crisis to which Randles (1988) referred to in his well-known article ‘From the
Mediterranean portulan chart to the marine world chart of the great discoveries: the crisis in cartography in
the sixteenth century’ had nothing to do with the deficiencies of cartography and all to do with those of navigation. Only after it became possible to determine longitude at sea and the spatial distribution of magnetic
declination was known – well into the eighteenth century – was it possible to fully adopt the 200 years-old
Mercator projection and abandon the old cartographic models for good (Monmonier, 2004, 79). It is my
intention to investigate this process, by focusing on the technical cartographic alterations that were
made (or just attempted) with the purpose of better reflecting the shape and size of the Earth into nautical charts.
The Cantino planisphere, drawn by an anonymous Portuguese cartographer in 1502, is the earliest extant
latitude chart, that is, a nautical chart incorporating astronomically-observed latitudes (for extensive studies
of the Cantino planisphere see Harrisse, 1883, 1892; Leite, 1923; Roukema, 1963; Cortesão & Mota, 1987;
Gaspar, 2010; 2012). Its model rapidly spread over Europe and became a de facto standard for all charts and
nautical planispheres produced in the sixteenth century and beyond. Little is known about the technical developments which preceded and followed the making of the Cantino planisphere, to which I have called the
“silent cartographic revolution” (Gaspar, 2013a, 70).
Despite the quality and thoroughness of the existing bibliography on the Iberian early-modern cartography, which includes such monumental achievements as Portugaliae Monumenta Cartographica (Cortesão
& Mota, 1987) and History of Portuguese Cartography (Cortesão, 1969), no significant contribution was
made by any of these works to our state of knowledge about the technical evolution of charts. That is because they were mainly focused on cartographic coverage, geographical content, authorship, dating and historical context, with little concern for chart construction and evolution. To the lack of technical studies much
contributed the well-known misconception championed by Cortesão and others, which I have called the
‘myth of the square’ (Gaspar, 2007), postulating that the nautical charts made by the Portuguese from the end
of the fifteenth century on were constructed according to the principles of the equirectangular projection
centred on the Equator, the so-called ‘square chart’ or platte carré. This mistake, which may have its origin
in a wrong interpretation of Pedro Nunes’s ‘Treatise in Defense of the Nautical Chart’ of 1537 (Nunes, 2004,
120-41), propagated to the present day and still contaminates the specialized literature (see, for example,
Snyder, 1993, 6-8; 2007, 274-278 and Monmonier, 2004, 28-29).
Aside from my own contributions (Gaspar, 2012; 2013; 2014; 2015a, 2015b; 2015d), a number of works
were published in the last decades discussing generic matters related to the genesis and evolution of early
modern Iberian charts (Mota 1973; 1977, Verlinden 1979, Marques 1987, Randles 1988, Cerezo 1994), or
addressing technical aspects related to their geometry and construction (Gernez, 1952; Cerezo 1992-94, Am-
13
aral 1995; Silió 1995; Robles 2010). However most of them focus on specific charts and/or features, and
none approaches chart construction and evolution in a systematic way.
What we know:
- Astronomical navigation was introduced in the second half of the fifteenth century by the Portuguese;
- No chart has survived of the fifteenth century, Portuguese or otherwise, where places were represented
according to their latitudes. The earliest extant latitude chart is the Cantino planisphere (1502); the earliest extant chart with a graphical scale of latitudes is likely the one by Pedro Reinel (c.1504);
- In all latitude charts of the beginning of the sixteenth century, coexist two cartographic models: the portolan chart model, used to represent the Mediterranean and the Caribbean Sea; and the latitude chart model,
used to represented Africa, Brazil and the Indian Ocean;
- Some charts are extant, Iberian and otherwise, with multiple scales of latitude. Such expedient was used
to accommodate in the same chart navigational information of latitudes, magnetic courses and distances;
- The latitude chart was definitely replaced by the Mercator projection only in the second half of the eighteenth century, after the longitude problem was solved.
What I propose:
- Specific objective 2.1: determine more precisely when and how the first latitude charts were developed;
- Specific objective 2.2: clarify how the traditional portolan-type charts of the fifteenth century could have
been used to support astronomical navigation, before the latitude chart was developed;
- Specific objective 2.3: clarify how latitude charts evolved technically from the beginning of the sixteenth
century on; in particular, how the old portolan-type cartographic representations of Europe, Africa and
Americas were progressively corrected on the basis of latitude observations;
- Specific objective 2.4: clarify if and how the knowledge of the secular variation of magnetic declination,
and its influence on the navigational accuracy of the charts, affected their construction;
- Specific objective 2.5: clarify how the geometric inconsistencies of the early modern charts, caused by
magnetic declination and the implicit assumption of a flat Earth, was dealt with by cosmographers, cartographers and pilots;
- Specific objective 2.6: clarify how the shape and size of the Earth became relevant for nautical cartography and how it was gradually reflected on nautical charts.
a.3. Use of charts at sea
The practice of navigation in Europe, during the Middle Ages and Renaissance, is addressed in various
scholarly studies, most of them focused on the Mediterranean (Gautier-Dalché, 1996; Kelley, 1999; Falcheta,
2008). However in few of these studies is the practical use of nautical charts in navigation discussed, and
when it is, their approach rarely goes beyond the transcription of the contemporary sources. Such is, for example, the case of Eva Taylor (1956), David Waters (1958) and, more recently, Eric Ash, (2007). In this last
work the author concludes that, because of the lack of physical evidence, ‘we may never be able to reconstruct the use of nautical charts at sea’.
The earliest known navigation manual explicitly referring to the use of charts is Benedetto Cotrugli’s De
navigatione, of 1465 (Falcheta, 2009). Here it is described how to find the course and measure the distance
between two places using two pairs of dry point compasses, how to estimate the position of the ship on the
basis of the course steered and the distance travelled, and how to determine the distance to land using two
bearings determined with the marine compass. An interesting detail of this text is the reference made to the
use of wax drops to mark the position of the ship on the chart. Traces of these drops will be looked for
when examining manuscript charts in search of marks of use. No other source dealing with the use of the
chart is extant before the sixteenth century, when numerous Iberian navigational treatises were written. A
point to note is that the descriptions made in those treatises about the use of the chart for solving various
navigational problems are identical to the one in Cotrugli’s manual. Examples of important titles are Martín
Fernández Enciso, Suma de Geografia (1519), Pedro Nunes, Tratado en defensam da carta de marear
(1537), Martín Cortez de Albacar, Breve compendio de la sphera y del arte de navegar (1551), Fernando
Oliveira, Ars Nautica (1570) and Francisco da Costa, Tratado da Hidrografia e Arte de Navegar (1599). To
these sources one should add a very large number of Iberian rutters and logbooks, most of them unedited.
What we know:
Nautical charts have been used to support navigation since the Middle Ages, both inside and outside the
Mediterranean;
14
-
-
Nautical charts were used, as in the present day, to support two complementary processes: the planning
of navigation, consisting in the retrieval of the necessary information pertaining to the passage: distances, dangers, durations, etc.; and the execution of navigation, consisting in determining and controlling
the position of the ship at sea;
Nautical charts and pilot’s rutters were complementary tools in the planning and execution of navigation.
What I propose:
Specific objective 3.1: clarify how portolan charts were used to support navigation, both in the planning
and execution phases;
Specific objective 3.2: clarify how the information contained in the rutters complemented the one of the
charts, both in the planning and execution phases;
Specific objective 3.3: clarify how nautical charts were used to support navigation, after the astronomical methods were introduced. In particular, clarify the use of graphical versus analytical methods in determining the position of the ship.
2. Methodology
In the present section the research program is described, in the form of a structured list of tasks, intended to
achieve the specific objectives articulated above. When appropriate, the analytical and modelling tools briefly introduced in part B1 are explained in more detail. The research tasks are organized in two groups: Study
of charts and Study of textual sources. The supporting tasks (e.g. software development) and the expected
outcome of the research (e.g. publications, meetings and exhibitions) are described in separate sub-sections:
Supporting tasks and Dissemination of results/outreach.
b.1. Study of charts
Most of the charts will be examined using the high resolution digital copies provided by libraries and museums; others will have to be inspected on site, in order to detect marks not perceptible on a photograph, or
when special lighting is considered necessary. All members of the research team will be involved in this
group of tasks.
Task 1.1 – Choice of charts
Two lists of charts to be studies will be created: one of portolan charts (here understood as those manuscript
charts which do not incorporate observed latitudes); and another of latitude charts (here understood as those
charts incorporating astronomically-observed latitudes, even if only locally). The choice of the charts that
will be subjected to analysis in each list will be decided on the basis of the following criteria: (1) Date and
historical relevance: both lists will include, whenever possible, all the earliest exemplars of its kind as well as
those charts showing relevant technical innovations or improvements; Origin: for the first list, charts of different origins will be chosen: Italian and Majorcan workshops as well as Portuguese or Arabic ones, when
appropriate. Concerning the second list, Iberian charts are of special relevance owing to the fact that they
were the first to incorporate observed latitudes and served as models for the contemporary European cartography. Charts produced in other European countries will also to studied whenever they show evidence of
cartographic innovation or reflect the content of lost Iberian prototypes, such as some Italian planispheres of
the sixteenth century; Geographical coverage: for the first list, the analysis will cover not only the representations of the Mediterranean and Black Sea but also of Western Europe and the western coast of Africa. These
regions are relevant because improvements were made in their representations during the fourteenth and
fifteenth centuries; Availability: a large number of the portolan charts of the Mediterranean have already
been digitized and can be studied through the available digital copies, which is not the case of most latitude
charts. Whenever it is considered necessary charts will be examined on-site and/or their images digitized. An
exhaustive list of charts to be used as a reference is in Extra Annex 1.
Task 1.2 – Survey of literature; assessment of authorship and dating
A quick survey of the secondary literature addressing each chart will be made, in order to retrieve relevant
information and review authorship and dating, which is often unknown or problematic.
Task 1.3 – Survey of place names
Lists of places of known geographical coordinates, which have been identified both in the old charts and in a
modern representation, will be made. Their positions will serve as control points in the determination of the
geographical grids, the assessment of planimetric accuracy and the comparisons between charts. This is a
very time consuming task, owing to the large number of toponyms depicted in the charts and the difficulties
15
associated with their identification. In order to facilitate the process a database of names and coordinates will
be created (see subsection b.3).
Task 1.4 – Georeferencing
Most of the cartometric techniques used in this research require that a previous correspondence between the
chart under analysis and the physical world is first established. This is accomplished by georeferencing,
which consists in assigning a geographic coordinate system to the chart under analysis, on the basis of a
sample of control points of known latitudes and longitudes (see task 1.3 above).The process consists in constructing surfaces of latitudes and longitudes (models) by interpolating trough the coordinates of the control
points, and then materializing those surfaces as matrices of values. Once an old chart is georeferenced it becomes possible to trace the mesh of meridians and parallels which is implicit to the representation. The visual inspection of such mesh is a simple yet extremely powerful way of qualitatively assessing the main geometric features of the charts. With this technique it is possible to quickly detect the distortions affecting the
different regions, to identify the areas where astronomically-observed latitudes may have been incorporated
and to make a preliminary assessment of planimetric accuracy, to be re-evaluated using finer techniques. Fig.
4 (left) shows the mesh of meridians and parallels in an excerpt of the Cantino planisphere. Notice how the
parallels in the Mediterranean are tilted counter-clockwise as a result of magnetic declination. This result
demonstrates that the region was charted using uncorrected magnetic directions and estimated distances only.
On the contrary, parallels are approximately straight, equally-spaced and east-west oriented along the African coast, where astronomical observations were made during the last quarter of the fifteenth century. Most
georeferencing operations will be made using the application MapAnalyst, which is freely available on the
Internet. All charts will be subjected to this preliminary analytical process.
Latitude error (degrees)
Task 1.5 – Assessment of planimetric accuracy
Once a chart has been georeferenced it becomes possible to assess the planimetric accuracy of its apparent
latitudes and longitudes. The use of the word ‘apparent’ comes from the fact that no pre-Mercator nautical
charts depict meridians or parallels and that only some of them incorporate observed latitudes. Portolan
charts were constructed on the basis of magnetic courses and estimated distances only, with no concern for
2
the latitudes and longitudes of the places. Latitude charts are typically hybrid representations, constructed on the basis of latitudes,
courses and distances, with no concern for
1
longitudes. However the assessment of northsouth and east-west relative accuracy of the
charts can reveal important geometric features
0
related to the construction methods and
sources of cartographic information. Good
examples are those illustrated in Fig. 3 of Part
-1
B1, for the Cantino planisphere, and in Fig. 2
-10
-5
0
5
10
15
20
25
30
35
40
45
Longitude (degrees)
of part B2, where the distribution of latitude
errors along the Mediterranean for a chart of
W Mediterranean Dulcetti
Greece & Aegean Sea, Dulcetti
Blac Sea, Dulcetti
Eastern Mediterranean, Dulcetti
Angelino Dulcetti (1339) is shown. In the first
Mediterranean African coast, Dulcetti
case, the method permitted to identify some of
the sources of information; in the second, to
Figure 2 – Distribution of latitude errors with longitude in the
discriminate the various regional representa- chart of Angelino Dulcetti (1339). Notice the regional clusters of
tions from which the chart of Angelio Dulcetti data and the scale differences among them, materialized by the
may have been assembled.
inclination of the regression lines.
Task 1.6 – Numerical modelling
Considering that pre-Mercator nautical charts were constructed, according to the textual sources, by transferring directly to the plane the information collected by the pilots at sea and on land, consisting in values of
latitudes, courses and distances, simulations of the process can be made numerically. The method is based on
the concept of empirical map projection, as explained briefly in Part B1. Two types of data are required for
the model to run: a sample of tracks connecting places on the spherical surface of the Earth, supposedly representative of the routes used to construct the chart; and the spatial distribution of magnetic declination at
time the navigational information was collected. In the case of the portolan charts, a dense and complex
mesh of maritime routes was likely used in the construction. Because these superabundant data become geometrically inconsistent when transferred directly to a plane surface, an optimization process is applied to the
positions of the points, which are incrementally adjusted until a minimal overall error is obtained. This pro-
16
cess intends to mimic the graphical adjustments likely made by the cartographers in the
earliest phases of portolan chart development. That is not the case when representing
larger oceanic areas, when the charts were
usually constructed on the basis of known
maritime routes, such as the one connecting
the North Atlantic to the Brazilian coast or to
the Indian Ocean. Concerning the distribution
of magnetic declination, contemporary observations will be used whenever possible
(e.g. those made by João de Castro, 1538),
complemented with the outputs of a geomagnetic model, such as Korte & Constable
(2005) or Jackson et al. (2000). Numerical
simulations of the geometry of charts or
groups of charts will be made in order to
assess the various aspects related to their
genesis, construction details and evolution.
Figure 3 – User interface of the application EMP – Empirical Map
An example of a output of a numerical
Projections (reproduced from Gaspar, 2011).
simulation of the Cantino planisphere is
illustrated in Fig. 4. A computer application will be developed, intended to replace my EMP model (Fig. 3)
with a more efficient and user-friendly one.
From the comparison between the outputs of the model with the geometric features determined by cartometric analysis, partial conclusions are expected to be drawn pertaining to the construction and updating of
the medieval and early-modern charts (specific objectives 1.1 through 1.5 and 2.3 through 2.5).
Figure 4 – Grid of meridians and parallels implicit in the Cantino planisphere (left), together with the output of a numerical simulation based on a set of maritime tracks supposedly used in its construction(right). Notice how the strong
distortion of meridians near the Red Sea, a feature shared by most charts of the sixteenth century, is correctly replicated by the model. Adapted from Gaspar (2010, 150; 176).
Task 1.7 – Physical examination of charts
Two main techniques – common in the study of old textual manuscripts and paintings, but never applied to
maps –will be used to examine a selected number of charts, in order to improve their legibility and detect
marks of use and construction: tangent lighting and ultra-violet photography. The easiest one, the visual exam of the charts under tangential light, is to be applied routinely to all originals chosen for physical analysis.
The aim is to detect marks of use and construction (pricking marks, traces and wax stains), to be interpreted
in the light of the navigational and cartographic methods of the time. Ultra-violet (UV) photography will be
used whenever the chart legibility is poor or to highlight content not perceived with a naked eye. If, after the
examination with UV light, it is concluded that important content can only be revealed through multispectral imaging, the use of such technique will be considered. However the process is complex and expensive, and will only be applied if I consider that the expected results are of critical importance to the project’s
goals. With the outcome of this task I expect to shed considerable light onto two unrelated questions: the use
17
of charts at sea (specific objectives 3.1 and 3.3), revealed by the compass marks and wax stains detected on
the parchments; and the graphical processes used in chart drawing, revealed by the marks left by underlying
construction patterns (objectives 1.1 and 2.1).
Task 1.8 – Assessment of cartographic genesis and evolution
This task can only be started after the studies of the relevant individual charts and textual sources are completed. The assessment will focus on the charts and periods considered to be the most significant, in which
technical evolution is concerned: up to c.1350 and around 1450, for portolan charts; and up to c.1600, for
latitude charts. In the course of this task, I expect to be able to provide answers to the issues of specific objectives 1.1 through 1.5 and 2.1 through 2.5, pertaining to the genesis and their technical evolution of the
charts over time, as well as to the one concerning the incorporation of the shape and size of the Earth into
nautical cartography (specific objective 2.6).
b.2. Study of textual sources
A number of medieval and early modern sources, many of them unedited, will be scrutinized. The set includes medieval texts and Iberian treatises, rutters and logbooks of the sixteenth century containing information about chart construction and use. The parts considered to be the most important for the project will be
translated and included in an Anthology. The analysis of the selected texts is to be articulated with the tasks
dedicated to the study of charts, most especially when assessing the construction methods, technical evolution and use, aiming to pursue specific objectives 1.1, 1.4, 1.5, 2.1 through 2.6 and 3.1 through 3.3. A list of
textual sources, including treatises, rutters and logbooks of various origins is in Extra Annex 2.
Task 2.1 – Analysis of medieval sources
All known medieval sources referring to the use or construction of nautical charts will be re-visited and reinterpreted. Among them are the manuscript known as the Liber de existential riveriarum (c.1200), containing the earliest known reference to chart making, Benedetto Cotrugli’s De navigatione (1465) and some Italian portolani, such as the one contained in the Liber and Lo Compasso da Navigare (c. 1250).
Task 2.2 – Analysis of early modern navigational treatises and manuals
A considerable number of navigational treatises and manuals have been produced during the sixteenth century, most of them in Portugal and Spain. Some important examples are O Livro de Marinharia de João de
Lisboa (c. 1515), Martín Fernández de Enciso, Summa de geographia (1519), O Livro de Marinharia de
André Pires (c.1520), Martín Cortés de Albacar, Breve compendio de la sphera y del arte de navegar (1551)
and Francisco da Costa, Tratado da Hidrografia & Arte de Navegar (c.1599). To these Iberian titles a number of European texts will be added, such as William Bourne, A Regiment of the Sea (1574), and Edward
Wrigth, Certaine Errors in Navigation (1599). A carefully selection will be made of those texts containing
information about the construction of charts and their use at sea.
Task 2.3 – Analysis of early modern rutters and logbooks
Most rutters and logbooks to be analysed are Portuguese and Spanish. This is because I am mainly interested,
in order to assess cartographic evolution, in the navigational data collected along certain coastal regions in
the Atlantic and Indian Oceans, which were surveyed by Iberian pilots. Contrarily to the nautical treatises,
the majority of these sources remain unedited and unstudied.
b.3. Supporting tasks
Task 3.1 – Development of software
To facilitate the research tasks described above, the following computer-based application will be developed
by an external software house, under detailed technical specifications : i) a database containing digital reproductions of charts and related information: date; authorship; coverage; places names; bibliography; historical
context; etc. The system will be fed and accessed by all members of the project; ii) a computer application
intended to make more efficient and user-friendly the cartometric analysis of the charts, using the data stored
in the database; iii) a computer application intended to replace my Empirical Map Projection (EMP) program
with a more powerful and user-friendly version, which can be used by any researcher inside and outside the
project.
Task 3.2 – Management
This task comprises all supporting activity related to the administrative, logistic and outreach component of
the project. It will mainly involve the PI, the Research Manager and the host institution.
18
b.4. Dissemination of results / outreach
The results of the project are to be disseminated through international publications and meetings. A dedicated internet page will be developed and made available during the first weeks.
Task 4.1 – Articles & papers
Articles and paper proposals will be submitted to top level journals and forums dedicated to the History of
Cartography (e.g. Imago Mundi and International Conferences on the History of Cartography), History of
Science (e.g. Annals of Science and the conferences of the European Society for the History of Sciences) and
History of Navigation (e.g. Journal of Navigation and the conferences of Royal Institute of Navigation). An
average number of one article and two paper presentations per year per member is expected, from the second
year on. An unpredictable number of lectures as invited speakers are to be added to the above talks.
Task 4.2 – Preparation of a book
One collective book will be prepared and proposed to an international editor. Its purpose is to complement
the on-going History of Cartography series (The University of Chicago Press) with new material pertaining
to the genesis and technical evolution of nautical charts.
Task 4.3 – Organization of meetings
Two workshops of the series “History of Iberian Cartography” will be organized in the second and fourth
year. The on-going “Spring Lectures in the History of Cartography”, to which well-known international researchers are invited, will also continue on a yearly basis. Two international workshops, specifically dedicated to the dissemination and discussion of the results of this project, will be held in the third and fifth year.
During the last meeting an innovative exhibition will take place illustrating the geometric features, construction methods and use of nautical charts in the Middle Ages and Renaissance, using digital techniques.
Timeline (selected tasks)
Tasks
Designation
Participants
1.1 - 1.8
Study of charts
2.1 - 2.8
Study of textual sources
PI, SR, PD, ST
3.1
Development of software
PI, SR, PD, ST
4.1
Articles & papers
PI, SR, PD, ST
4.2
Preparation of a book
PI, SR, PD
4.2
Organization of meetings
PI, SR, PD
Year 1
Year 2
Year 3
Year 4
PI, PD, ST
Spring Lectures in the History of Cartography
Workshops History of Iberian Cartography
MEDEA- CHART Workshops
PI - Principal Investigator (Joaquim Gaspar) SR - Senior Researcher (Henrique Leitão) PD - Postdoctoral researchers ST - PhD Students
19
Year 5