This is the accepted manuscript of the article published online in the Journal of Early Modern History in 7 March 2019: https://brill.com/view/journals/jemh/23/1/article- p1_1.xml Early Modern Nautical Charts and Maps: Working Through Different Cartographic Paradigms Joaquim Alves Gaspar & Henrique Leitão Centro Interuniversitário de História das Ciências e da Tecnologia Faculty of Sciences, University of Lisbon Abstract Of all the technical and scientific developments that made possible the European maritime expansion, the nautical chart is perhaps the least studied and understood. This fact is very surprising as it was with the information contained in those charts, and later imported to geographical maps and atlases, that the newly discovered lands were first shown to the European nations. There was, however, a deep incompatibility between these two cartographic paradigms – the nautical charts and the geographical maps – which remained unsolved throughout the sixteenth century and beyond, despite the attempts to harmonize the technical principles of Ptolemy’s Geography with the advances of nautical cartography. An eloquent symptom of such incompatibility was the difference between what was understood as an accurate depiction of the Earth, in the eyes of cosmographers and geographers, and what was considered by the pilots as an accurate nautical chart. The misunderstandings around these issues during the early modern period and the unsuccessful attempts at reconciliation were, in great part, the cause for some polemics among cosmographers, cartographers and pilots, such as the conflict in the Casa de Contratación around the charts of Diego Gutiérrez, a fact not entirely understood by historians. At the core of the difficulty lies the circumstance that only in the present day has the true nature of the nautical chart, as a navigational tool, started to be clarified. How the differences between geographical maps and nautical charts contributed to shape the History of Cartography in various periods, and how they are related to conflicting scholarly objectives and practices, is the subject of this essay. We will show, using the results of cartometric analysis, that not only were those artifacts constructed using different principles and with different purposes, but that they belonged to incompatible cartographic paradigms, and we will argue for the relevance of this fact for the history of science. 1 Introduction Late-medieval and early-modern maps and nautical charts are complex artifacts that have often challenged – and eluded – the understanding of scholars. Still, the historical relevance of these documents is such that the results of their study often transcend the boundaries of specialized fields, such as the History of Cartography or the History of Navigation. It is the purpose of this essay to shed a new light onto the differences between geographical maps and nautical charts historically, and to explain how these differences contributed to shape not only the history of cartography but also to affect the history of science at large. The importance of focusing attention onto the subject is obvious: from the 15th, and well into the 18th centuries, nautical cartography was the single most important source of information for the construction of a global and geographically coherent depiction of the world, most particularly the depiction collected during the sea voyages about the coasts, islands and the shapes of landmasses abutting the oceans. The image of the Earth that was conveyed during this period, as well as the discourse about the orbis terrarum, was inevitably influenced by the origin of the information. The consequences of importing geographical information from nautical charts into geographical maps, more specifically those resulting from the fact that they belonged to distinct cartographic paradigms, are deeper and historically more relevant than previously thought. 1 In fact, although these two models have co-existed since the Middle Ages, it was only in the early modern period that their inherent differences became significant. While in broad terms the differences between maps and charts are not new for historians, their full historical significance and implications to knowledge-making scientific practices of the time have not yet been fully explored. In fact, it was only in recent years that techniques of cartometric analysis and numerical modelling, especially designed to assess and simulate the geometry of old charts, were successfully developed and tested. One of the most important conclusions that can be drawn from the application of such techniques – this is indeed a crucial point to note – is that these two cartographic models, that of the geographical map and that of the nautical chart, were often incompatible with each other: not only their geneses, evolution and use were associated to different social and professional circles and practices but, more importantly, they were designed according to different principles and exhibited different geometries. 1 The term ‘paradigm’ in the present context has no relation to the homonym Kuhnian concept that was proposed in The structure of Scientific Revolution (Chicago: The University of Chicago Press, 1962). By choosing this word, we aim to emphasize the relevance of aspects that are beyond the strict technical matters of map construction, such as the professional context of chart making and the use of charts in navigation. Throughout this text we will use the word ‘cartographic model’, instead of ‘cartographic paradigm’, whenever we focus on the technical component only. 2 Secondly, we will show that to fully understand nautical charts historically, it is imperative to know how navigation took place on board ships. Not only were those charts constructed using the navigational data collected by the pilots by means of the contemporary navigational methods, but also, they were used to support navigation exactly the same way. 2 This can be stated in a somewhat metaphorical, albeit expressive manner: one needs to look at these charts not only with the pilots’ eyes but, more importantly, with the pilots’ hands. In other words, the full comprehension of old nautical charts requires a full cognizance of the contemporaneous practice of navigation, a fact that has not been sufficiently explored by historians. The intimate connection between nautical cartography and navigation is an aspect that will become evident in this essay. Hence, a brief, but necessary, detour into the technicalities of early modern navigation might be useful. Finally, we will note how pilots, cartographers, cosmographers and mathematicians of the early modern period were, most of the time, incapable of grasping the intrinsic differences between nautical charts and geographical maps, and how this difficulty often led to misunderstanding and conflict. This article will make evident that, at the core of the misunderstanding, lies the circumstance that information of geographical interest was being acquired, structured, and displayed according to different processes and with different purposes. Whereas for pilots, the nautical chart was an “instrument to navigate” -- whose construction and use were to be understood in the scope of that specific activity-- for geographers and other scholars, nautical charts were treated as maps, that is, as attempts to represent the geography of the world. Thus, the very same object was perceived in a distinct way by two professional groups.3 An important aspect of those conflicting views concerns the two different concepts of accuracy that were at play. Whereas for geographers and learned cartographers, the accuracy of a map was measured by the agreement between the latitudes and longitudes represented on it and their correct values on the surface of the Earth (what we call today absolute positional accuracy), for pilots and nautical cartographers, the 2 It is now consensual among most historians of cartography that pre-Mercator nautical charts were constructed with compass courses, estimated distances and – from the beginning of the sixteenth century on – observed latitudes. This kind of functional relation is not applicable to the present-day practice, where the surveying methods used to acquire the information for the charts are usually distinct from the navigational methods. 3 Our topic is therefore relevant to the classical debate around the role of artisans vs. scholars in early modern Europe. For a recent and innovative statement of the issues at stake, see: Matteo Valleriani (ed.), The Structures of Practical Knowledge (Dordrecht: Springer, 2017). See especially the introductory article of Valleriani, “The Epistemology of Practical Knowledge”, pp. 1-19. See also Pamela H. Smith, The Body of the Artisan: Art and Experience in the Scientific Revolution (Chicago: The University of Chicago Press, 2004); Pamela O. Long, Artisan/Practitioners and the Rise of the New Sciences, 1400- 1600 (Corvallis: Oregon State University Press, 2011); Lisa Roberts, Simon Schaffer and Peter Dear (eds.), The Mindful Hand: Inquiry and Invention from the Late Renaissance to Early Industrialisation (Amsterdam: Koninklijke Nederlandse Akademie van Wetenschappen, 2007); Pamela H. Smith, Amy R. W. Meyers and Harold J. Cook (eds.), Ways of Making and Knowing: The Material Culture of Empirical Knowledge (Ann Arbor: The University of Michigan Press, 2014). 3 accuracy of a chart was measured by its efficacy in the accomplishment of a specific task –navigating a ship in a sea voyage. Thus, a chart was navigationally accurate when the latitudes of the places, as well as the compass courses and distances between them were accurate, and not necessarily when the shape of the landmasses matched their “true” form. 4 What made this situation highly prone to ambiguity and conflict was the fact that a specific object – the nautical chart – when transported from the realm of pilots to the realm of geographers and learned cartographers, did not carry a clear (and, much less, an explicit) explanation of its nature. Hence, the two professional groups interested in geography - pilots and scholars - were not solely separated by social barriers but also by a deeply contrasting interpretation of the meaning and content of a specific object, and of the processes by which geographical information was acquired. 5 Several technical studies that support our work have been published in recent years, but the full implications of these studies for the history of science – namely, the fact that those two cartographic paradigms collided – have not been sufficiently explained before. That is the purpose of this article. 6 Maps and charts When a present-day topographical map and a nautical chart depicting the same region are put side by side, the differences are evident: while in the topographical map the information is concentrated inland, leaving the wet areas almost totally empty, in the nautical chart the information is mostly located at sea and in the coastal areas, leaving the dry land areas virtually blank. The reasons for this difference are related to the distinct purposes of the two kinds of representations: while topographical maps are intended to show all types of general information located on land, nautical charts are constructed with the explicit goal of supporting marine navigation and display only the features that can be useful for such purpose. 7 A subtler difference, noted by the late American cartographer Arthur Robinson, concerns the way maps and charts are used: 4 We will see later in this article that this desideratum could only be met for some specific routes on the chart. 5 This social barrier mostly applied to Iberian pilots and cosmographers and was usually not a feature in Northern European countries. We are indebted to Sarah Tyacke for this observation. 6 For a general explanation on the use of cartometric methods of geometric analysis and numerical modeling of old charts, see Joaquim Alves Gaspar, From the Portolan Chart of the Mediterranean to the Latitude Chart of the Atlantic: Cartometric Analysis and Modeling. Unpublished doctoral dissertation (Lisboa: Universidade Nova de Lisboa, 2010). For a discussion around the nature of early modern nautical charts and its connection with the contemporaneous navigational methods, see Joaquim Alves Gaspar and Henrique Leitão, ‘What is a nautical chart, really? Uncovering the geometry of early modern nautical charts’, Journal of Cultural Heritage, 29 (2018), 1: 130-136. 7 Other less obvious differences, not perceived by the casual map reader, are the distinct vertical references used to reckon depths and elevations, and the map projections adopted in the two types of representations. 4 “maps are to be looked at while charts are to be worked on”. 8 Although Robinson’s aphorism is wonderfully synthetic and expressive, it only focuses on the use of present day maps and charts, and does not capture the whole spectrum of differences and its historical significance. Looking into the history of maps and charts makes one realize that, not only their geneses and evolution are distinct, but also their dissimilarities are conspicuous, significant and historically relevant. While in general qualitative terms this is not new for historians, the subject has never been addressed technically. When browsing through the published volumes of the History of Cartography, most especially those dedicated to the Middle Ages and Renaissance, it is notable how terrestrial maps and nautical charts have been traditionally studied separately. 9 Moreover, the geometry and construction of old nautical charts are almost exclusively addressed in a qualitative way, and the fundamental question of what happened when information from nautical charts was imported to geographical maps is not discussed. The result has been that some incorrect interpretations of those charts and maps have been perpetuated in the historiography, as we will indicate later in this article. 10 The Renaissance chart of the Atlantic Navigation in the European waters during the Middle Ages and early modernity was usually made near the coast, using the available portolan-type charts and the information registered in the pilots’ rutters. Sometimes it was necessary to move away from the coast, to reach a distant island or cross a larger body of water, but these open water tracks seldom took more than a few days. In those situations, the position of the ship was determined using the information about the course steered and the distanced sailed, estimated by the pilot. This method, designated in the present day as dead reckoning, was known by the early modern Portuguese and Spanish pilots as the ponto de fantasia (point of fantasy), a colorful name that expressed the uncertainty of the estimation process (Fig. 1, left). Although considerable errors could be made with the point of fantasy, they seldom represented a serious problem for navigation 8 Arthur Robinson, Joel Morrison, Philip Muehrcke, A. Kimerling and Stephen Guptill, Elements of Cartography, Sixth Edition (New York: John Wiley & Sons, 1995), 15. 9 In the first and third volumes of the History of Cartography there are eleven chapters totally or partially dedicated to the history of nautical charts. See J. B. Harley and David Woodward (ed.), The History of Cartography, Volume One, Cartography in Prehistoric, Ancient and Medieval Europe and the Mediterranean (Chicago: The University of Chicago Press, 1987); David Woodward (ed.), The History of Cartography, Volume Three, Cartography in the European Renaissance, Parts 1 and 2 (Chicago: The University of Chicago Press, 2007). 10 There is a single chapter in the History of Cartography where the mathematical aspects of chart construction are addressed: John Snyder, ‘Map Projections in the Renaissance’, Volume Three, Part 1, 365-381. The author postulates that both the portolan chart and the Renaissance nautical chart are based on the cylindrical equidistant (or equirectangular) projection. This is a repetition of an erroneous interpretation that originated in the sixteenth century, propagated to the present day and was adopted by respected authors in other important international publications. 5 because the correct position of the ship could be easily found as soon as the coast was again in sight. 11 By mid-fifteenth century, Portuguese ships were already navigating away from the coast, in long-distance voyages that often took several weeks or months. In these new conditions, positioning the ship by dead reckoning was inadequate. As time elapsed from the last known position, the accuracy of the point of fantasy steadily degraded to the point of becoming almost useless. To face the new problem, a new navigational method had to be found. The solution was the introduction of astronomical navigation, made possible by the adaptation of the instruments of observation used by astronomers on land – the astrolabe and the quadrant – and the development of very simple procedures that could be effectively used by uninstructed pilots on board, to observe the heavenly bodies and determine latitude. 12 The point of fantasy then gave place to the so-called ponto de esquadria (set point), where the latitude became the prevailing element of navigational information. The position of the ship started to be determined on the chart as the intersection between the line representing the course steered and the parallel representing the latitude (Fig. 1, right). The superlative importance of this technical development cannot be over emphasized, as the reliability of the new method permitted to establish regular oceanic routes from Europe to India and beyond, for the centuries to come. We will see next how the introduction of these techniques had a major impact on the geometry of nautical charts. 11 The practice of navigation and the use of nautical charts in Europe, during the Middle Ages and Renaissance, is addressed in various studies, most of them focused on the Mediterranean. See Eva Taylor, The Haven-Finding Art: An History of Navigation from Odysseus to Captain Cook (London: Hollis & Carter, 1956); David W. Waters, The Art of Navigation in England in Elizabethan and Early Stuart Times (London: Hollis and Carter, 1958); Patrick Gautier-Dalché, ‘L’usage des cartes marines aux XIVe et XVe siècles’, in Spazi, tempi, misure e percorsi nell’Europa dal Bassomedievo: aati del XXXII Convegno storico internazionale, Todi (Spoleto: centro Italiano di studi basso Medievo – Academia Tudertina, 1996), 97-128; James Kelley, On Old Nautical Charts and Sailing Directions: Technical Essays (Melrose Park, PA: Sometime Publishers, 1999); Piero Falcheta, ‘The Use of Portolan Charts in European Navigation During the Middle Ages’, in Europa im Weltbild des Mittelalters: Kartographische Konzepte (Berlin: Academic Verlag, 2008), 269-276; Eric Ash, ‘Navigation Techniques and Practice in the Renaissance’, in Woodard & Harley (ed.), The History of Cartography, Volume Three: Cartography in the Renaissance, Part I (Chicago & London: The University of Chicago Press, 2007), 509-527. Several Portuguese and Spanish sources contain descriptions of navigational methods in the early modern period, the oldest being the anonymous Portuguese manuscript of ca.1508 known as the Livro de Marinharia de João de Lisboa. Among the best-known treatises explaining how to determine the position of the ship on the chart are those of Alonso de Chaves, Quatri Partitu en cosmographia practica y por otro nombre llamado espejo de navegantes (1520-38), chapter 2; Martín Cortés, Breue compendio de la sphera y de la arte de nauegar (1551), chapter XIII; Rodrigo Zamorano, Compendio de la arte de navegar (Sevilla,1581), chapters 20 and 21; and Francisco da Costa, Arte de Navegar (1596), chapter LIII. 12 A good introduction to the genesis of astronomical methods of navigation is Luís de Albuquerque, ‘Astronomical Navigation’, in Armando Cortesão, History of Portuguese Cartography (Coimbra: Agrupamento de Estudos de Cartografia Antiga, Junta de Investigações do Ultramar, 1971), 221-442. 6  N N PD PD d C C  PF SP Figure 1 – Point of fantasy (left) and set point (right). The point of fantasy PF is the intersection between the line representing the course, C, and the arc of circle representing the distance sailed from the point of departure, PD. The set point SP (right) is the intersection between the line representing the course and the parallel of latitude  (the horizontal line). During the fifteenth century, nautical charts used by Portuguese pilots in voyages of exploration and trade adopted the same cartographic conventions as the portolan chart of the Mediterranean. Whenever the value of magnetic declination was small, these charts could still be used with the new astronomical methods of navigation. That was indeed the situation in the Atlantic, north of Cape Vert, during the fifteenth century. But as ships progressed to the south, the value of magnetic declination markedly increased, and it eventually became obvious to pilots that the old cartographic model was no longer compatible with the new navigational methods. The incompatibility of the traditional charts with the set point method, under the effect of magnetic declination, resulted from the fact that magnetic declination tilted the east- west lines implicit in their geometry (parallels of latitude) relative to the east-west direction (Fig. 2, left). On the contrary, a chart based on the set point method always represented points with the same latitude on the same horizontal east-west line (Fig. 2, right). 13 A new and better chart had thus to be designed, this time based on the set point method. The Cantino planisphere, drawn by an anonymous Portuguese cartographer in 1502, is usually considered the oldest extant nautical chart incorporating the latitudes of the places (Fig. 3). With this chart, a new cartographic standard was established, adopted by many other world maps of the beginning of the sixteenth century, such as the Caverio planisphere (ca. 1504), the Hamy-King planisphere (ca.1504) and the two printed world maps of Martin Waldseemüller (1507; 1516). 13 For a detailed explanation see Gaspar, From the Portolan Chart of the Mediterranean to the Latitude Chart of the Atlantic (note 6), 13-21; ‘Blunders, Errors and Entanglements: Scrutinizing the Cantino Planisphere with a Cartometric Eye’. Imago Mundi, 2012, 64, 2: 184-186. 7  N N PD PD d C C   C m C m  PF1 SP SP1 PF Figure 2 – The influence of magnetic declination on the point of fantasy (left) and the set point (right). PF1 and SP1 are, respectively, the point of fantasy and the set point, as affected by magnetic declination . Notice how magnetic declination only affects the longitudinal position of the set point, thus conserving the observed latitude . Figure 3 – The Cantino planisphere (1502) depicts the world as it became known after the exploration voyages in the end of the fifteenth century and beginning of the sixteenth to Africa, Greenland, Newfoundland, Brazil and India. It is usually considered the first ‘latitude chart’, that is, the first nautical chart representing places according to latitudes. Biblioteca Estense Universitaria (C.G.A.2), Modena. How accurate are those world maps? Or, more precisely, how well do the shapes of the landmasses agree with the equivalent shapes on a modern map? Occasionally, historians have tried to give an answer to this question by comparing the coastlines depicted on them with the corresponding lines in a modern map. Such exercise is equivalent to what is called today an assessment of absolute positional accuracy, where the geographical positions shown on an old map are compared with the ones accepted as true. The process is, however, theoretically meaningless unless the two 8 representations share the same map projection. 14 A possible way out of this difficulty is to assume that the geometry of those early modern planispheres is identical – at least approximately – to the one of the cylindrical equidistant projection centered at the Equator (the so-called plate carré), where meridians and parallels form a square grid. That was indeed the (erroneous) interpretation of most map historians until very recently. While we can confidently assume that the parallels of latitude that are implicit in those early modern planispheres are approximately oriented in the east- west direction, which is a direct consequence of the use of the set point method, a similar assumption cannot be made for the meridians, whose orientation reflects the convergence of meridians and, as we will show, the effect of magnetic declination. 15 Figure 4 – Coastlines of Africa and Brazil in the Cantino planisphere (left), compared with the corresponding outlines in a modern map (right). Notice the eastward displacement of Africa and Brazil, and the longitudinal stretching of Africa. The projection of the modern map is the equirectangular projection centered at the Equator, commonly known as square chart or plate carrée. Indeed, if we compare the representation of the South Atlantic in the Cantino planisphere with the corresponding one in a modern plate carré map, the result is unexpected (Fig. 4): Brazil and Africa appear both displaced to the east and the African continent is markedly stretched in the east-west direction, making the Isthmus of Suez look enormous, a feature that was replicated in all European cartography of the 14 A map projection is any systematic arrangement, on the plane, of the meridians and parallels defined on a geometrical model of the Earth (a sphere or an ellipsoid). Different map projections exhibit different geometries, depending on the geometrical or analytical process used to transform geographical coordinates ( ) into plane coordinates (x, y). For example, in all cylindrical projections, meridians and parallels are straight and perpendicular to each other, forming a mesh of rectangles. Other projections may depict meridians and parallels as curved lives. Thus, comparing the geometries of an old chart with the one a modern map, in order to assess its accuracy, only makes sense when they share the same map projection. 15 This is easily perceived by the visual inspection of the geographical grid of meridians and parallels implicit to those representations, which can be estimated using a sample of control points of known latitudes and longitudes. For an explanation of the process see Gaspar, From the Portolan Chart of the Mediterranean (note 6), 45-54. 9 sixteenth century and beyond. 16 An attractive reason that could be invoked for this apparent distortion is the inadequacy of the surveying techniques of the time, when longitude could not be determined accurately. An alternative explanation, proposed by some authors, is that the eastward displacement of Brazil was purposeful, aiming to fool the Spanish as where the dividing line of Tordesillas should pass. 17 None of these reasons resists a detailed geometrical analysis. N N SP SP*  C C Cm Cm  PD  PD  SP* SP Figure 5 – The effect on the set point of an eastward magnetic declination for a course in the northeast and southeast quadrants (right). The graphs illustrate how the ship’s position SP was marked on an ordinary chart using the compass (magnetic) course Cm, sailed from the point of departure PD, and the observed latitude . N is the chart’s North, from where all courses are reckoned. If magnetic declination were corrected for, the resulting set point SP* would lie to the east or to the west of SP, depending on the quadrant of the course. Notice (right) how the eastward displacement of SP (here representing the Cape of Good Hope), relative to its true position SP*, is explained by the eastward magnetic declination in the South Atlantic, as correctly perceived by João de Castro. The main cause for the apparent displacement of Brazil and Africa on the charts of the sixteenth century was correctly identified in 1538 by the nobleman and navigator João de Castro (1500-1548), during a voyage from Lisbon to India. During this voyage, Castro made systematic observations of magnetic declination with the purpose of confirming (or disproving) the conjecture that its value varied linearly with longitude. 18 After describing its spatial variation along the route from Lisbon to Brazil and to the 16 A detailed cartometric analysis of the Cantino planisphere is in Gaspar, From the Portolan Chart of the Mediterranean (note 6), 141-174; idem, ‘Blunders, Errors and Entanglements (note 11)’, 186. 17 Indeed, some Portuguese charts of the sixteenth century deliberately manipulated the coastline of Brazil with diplomatic purposes. However, that is not the case of the Cantino planisphere. See Avelino Teixeira da Mota, Reflexos do Tratado de Tordesilhas na Cartografia Náutica do Século XVI (Coimbra: Centro de Estudos de Cartografia Antiga, Junta de Investigações do Ultramar, Separata LXXX, 1973), where the idea that the positions of Newfoundland and Brazil on the Cantino planisphere were manipulated is refuted. 18 During his voyage from Lisbon to India, João de Castro collected 37 daily values of magnetic declination, which were determined along the route using the instrument designed by the mathematician Pedro Nunes. These values were registered in Castro’s “Roteiro de Lisboa a Goa” (1538), together with other navigational data. See João de Castro, ‘Roteiro de Lisboa a Goa’, in Luís de Albuquerque (eds.), 10 Cape of Good Hope (where the compasses pointed to the east of the geographic north), and from the Cape of Good Hope to India (where the compasses pointed to the west), Castro concluded that the eastward displacement of both Brazil and Africa on the charts was caused by using uncorrected compass courses. Castro illustrated his interpretation with examples, where the effect of magnetic declination on the position of the ship, as determined using the set point method, is correctly explained (Fig. 5).19 A critical point to stress is that latitude charts were constructed, like portolan charts, by transferring directly to the plane the uncorrected compass courses measured at sea. That was done, not because pilots were unaware of the phenomenon of magnetic declination but because that was the simplest thing to do. Correcting all charts for magnetic declination would imply time-consuming and costly surveys, in order not only to measure the true geographical directions between places but also to determine the spatial distribution of magnetic declination. 20 Such an endeavor only became truly necessary when the longitude problem was solved in the second half of the eighteenth century, making it possible for the old latitude chart to be abandoned and the Mercator projection to be fully adopted by mariners. Coming back to the initial question about the accuracy of the early modern planispheres, it seems clear that the modern criterion of absolute positional accuracy, based on the latitudes and longitudes of the places, should not be applied here. Knowing that charts were made for supporting navigation, an appropriate way of assessing their navigational accuracy would be to evaluate to what extent they were fit to that purpose, at the time they were used. More specifically, how accurate were the latitudes, the magnetic courses and the distances between places? 21 Incidentally that was the mistake, not only of present-day historians, but also of the cosmographers of the sixteenth century, as described next. Obras Completas de D. João de Castro (Coimbra: Academia Internacional de Cultura Portuguesa, 1962), Vol. I, 113-296. 19 Idem, 198-207. Castro’s text contains a remarkably clear discussion about the effect of the uncorrected magnetic declination on the geometry of charts. A lengthy study of his ideas and contribution for the early modern science, emphasizing his experimental work, is Reijer Hooykaas, ‘Science in manueline style. The historical context of D. João de Castro’s works’, idem, Vol. IV, 231-426. 20 The earliest known cartographic representation of isogons (that is, lines of constant magnetic declination) is in the unsigned chart of c.1585, attributed to the Portuguese cartographer Luís Teixeira, representing the western margin of the Pacific Ocean, which was most likely part of a now lost planisphere composed of four sheets of parchment. Our interpretation is those lines might have been used during the late sixteenth century as an aid to navigation, rather than to correct magnetic courses. See Joaquim Alves Gaspar and Henrique Leitão, ‘Luís Teixeira, c.1585: The Earliest Known Chart with Isogonic Lines’, Imago Mundi, 70, 1, 2018, 221-238. 21 Two unrelated factors make such an assessment not easy to accomplish: the fact that only the specific (often unknown) routes used to construct a certain chart were supposed to be represented accurately; and the difficulty in knowing with sufficient accuracy the spatial distribution of magnetic declination in ancient times. An assessment of the latitudes, courses and distances of five Portuguese charts of the 15th and 16th century is in Gaspar, From the Portolan Chart of the Mediterranean (note 6), 108-119, 151-166. 11 The Lopo Homem incident Around 1560, the Portuguese cartographer Lopo Homem, in a note addressed to the king, harshly complained about the new official master chart that had been established by the Cosmographer Major, the mathematician Pedro Nunes (1502-1578). 22 According to Lopo Homem, such chart had been prepared using the eclipses of the sun and moon −to determine longitude− with the purpose of showing that from Lisbon to India, and to the Moluccas, the real distances were shorter than the ones shown on the charts. According to the cartographer, however, the charts made according to the new pattern were “so wildly distant from all truth and navigational science that many ships were lost and pilots were forced to buy their charts in Castile”. 23 Figure 6 – Excerpt of a nautical planisphere by Lopo Homem (1554). The black outline represents the coastline of a modern plate carré map. Although the longitudinal stretching of Africa was reduced, when compared with the Cantino planisphere, the eastward displacement was not completely eliminated. This may be explained by the error, of about 10 degrees, made in the determination of the longitude of Diu. Istituto e Museo di Storia della Scienza, Florence, IMSS, n. 0946. By the time this note was written, pilots were perfectly aware, as João de Castro was in 1538, that the longitudinal distance between Lisbon and India, as shown on contemporary charts, was exaggerated. In his Tratado em defensam da carta de marear (Treatise in Defense of the Nautical Chart), of 1537, the same Pedro Nunes had already complained about the fact, which he attributed to the incompetence of the pilots, “who represented as straight which had been sailed with so many detours”. 24 22 The royal master chart, or Padrão del Rey, was the official cartographic standard from which all charts used by the pilots at sea should be based on. This chart was supposed to be kept up to date, concerning the new geographical discoveries and surveys. 23 The undated letter of Lopo Homem is kept in the Bibliothèque nationale de France (ms. Coll. des Cinq Cents de Colbert 298, fols 6 r-8 r) and was transcribed by Luís de Matos, Les Portugais en France au XVIe siècle (Coimbra: Universidade de Coimbra, 1952), 318-322. 24 Pedro Nunes, Obras, Vol. I: Tratado da Sphera; Astronomici Introductorii de Spaere Epitome (Lisboa: Fundação Calouste Gulbenkian, 2002 [First edition: Lisboa, 1537], 120-184 at 129. 12 Although Pedro Nunes was right about the exaggerated longitudinal distance between Lisbon and India, as represented on the charts, he missed the point when he blamed pilots for the apparent mistake. In 1547, he was appointed cosmographer major, becoming responsible for the royal master chart. Then he finally had the opportunity to correct what he considered to be a major imperfection, by ordering the longitude of Diu (in India) to be determined by astronomical methods. The result was a new cartographic standard in which the longitudinal distance between Lisbon and India was partially reduced, as in the planisphere of the same Lopo Homem (1554) shown in Fig. 6.25 The examination of the Portuguese charts from 1554 onward makes clear that Nunes’ new master chart was not adopted by all. We have compared the longitudinal width of Africa at the latitude of Cape Guardafui (in the easternmost tip of Africa) in a series of nautical planispheres of the sixteenth and seventeenth centuries, and the results are revealing. While the apparent longitude of Cape Guardafui remained more or less constant from 1502 to 1628, the width of the Mediterranean increased during the same period, making the longitudinal distance between its eastern margin and Cape Guardafui decrease markedly. This means that the Mediterranean was artificially stretched in the east-west direction, with the likely purpose of cosmetically reducing the size of the Red Sea and, with it, of the Isthmus of Suez. 26 This fact suggests that the new pattern introduced by Nunes failed after all, most probably owing to the opposition of pilots and cartographers. Coming back to the comments of Lopo Homem, why did the cartographer consider Nunes’ standard “wildly distant from all truth and navigational science?” What kind of errors would make the new master chart so unfit for navigation? Clearly, he was not referring to the distances measured on them – particularly the one from Lisbon to India – which were usually not to be trusted.27 The problem was, most likely, the orientation of the African coastline, which no longer agreed with the uncorrected compass courses. Although the allegations of Homem – that many ships using the new charts were lost in their way to India and that pilots were forced to buy their charts in Castile – are obvious exaggerations, the apparent failure of Nunes in imposing the new standard demonstrates that the arguments of the cartographer had considerable weight. 25 The astronomical observations ordered by Pedro Nunes in Diu were reported by the humanist and historian Jerónimo Osório, in De Rebvs Emmanuelis Regis Lusitaniae (Olysippone: 1571), 424. 26 See Gaspar, ‘From the Portolan Chart to the Latitude Chart : The Silent Cartographic Revolution’, Cartes et Géomatique, Bulletin du Comité Français de Cartographie, 216 (2013), 67-77. 27 It was shown before that it was not possible to conciliate, on a latitude chart, both the course and the distance between places owing to the effect of magnetic declination A critical point to note is that the courses between places was a much a more important piece of information for the pilots than distances because it affected on a much larger degree the safety and effectiveness of navigation. For a more detailed explanation see Gaspar, From the Portolan Chart of the Mediterranean (note 6), 13-21. 13 Who was then right and who was wrong in this dispute? Was it Pedro Nunes, who used astronomical methods to determine the longitude of India and enforced the result onto the official cartography? Or was it Lopo Homem, who insisted in keeping the traditional representation and supported his position with practical arguments concerning the safety of navigation? Because the requirement of Pedro Nunes (correct longitudinal distance) could not be reconciled with the one of Lopo Homem (correct compass courses), the answer to the question depends on the use of the charts. As cosmographer major, Pedro Nunes was responsible for keeping the official master chart and approving the nautical charts used in navigation. But charts were also made for purposes not related to navigation, for example, for diplomatic purposes. This line of reasoning suggests that Nunes may have been driven by the intention of applying the Ptolemaic prescription of map construction, that is, the use of latitudes and longitudes, to nautical charts. In other words, he may have intended to enforce the use of a positionally accurate plate carré into nautical cartography. A better explanation is suggested by an attentive reading of Nunes’ Treatise in Defense of the Nautical Chart of 1537, where a detailed analysis of the geometry of the contemporaneous charts is presented and the necessity of representing correctly the distances in the east-west direction is emphasized. 28 However, nowhere in the text is the influence of magnetic declination discussed, although Nunes was perfectly aware of the phenomenon. This suggests that either he was convinced that courses were corrected before being transferred to the charts, or alternatively, he did not consider the effect to be geometrically relevant. Knowing that the eastward displacement and stretching of Africa in the charts was mostly caused by the effect of magnetic declination, this may explain why Nunes insisted that the distance to India should be corrected: not to depict the peninsula according to its true longitude but to conserve longitudinal distances. Once again, two distinct concepts of accuracy were at stake here: the modern concept of absolute accuracy, applicable to the geographical depictions of the Earth; and the concept of navigational accuracy, applicable to nautical charts. This episode also demonstrates that even when close contact between scholars (cosmographers) and artisans (pilots and cartographers) was enforced, the full understanding by the former of artisanal practices could be difficult to achieve. Charts with multiple latitude scales Another example of the clash between the geographical and the nautical paradigms in early modern cartography is the dispute in the middle of the sixteenth century around the making and use of charts with multiple latitude scales. This ingenious expedient, 28 Pedro Nunes, Obras, Vol. I: Tratado da Sphera, 129-133. An article containing a detailed discussion of of the Tratado em Defesam da Carta de Marear (1537), and its reflex in Europe, is being prepared by the present authors and will be shortly submitted for publication. 14 first used by Pedro Reinel ca. 1504 to represent Newfoundland, was intended to conciliate the three elements of navigational information in the location of a given region – the latitude, the compass course and the distance – in the presence of magnetic declination. The idea was to locate that region on the chart according to a compass course and a distance, measured from some origin (that is, using the method of the point of fantasy), and add a secondary latitude scale only applicable to that region. In the case of Pedro Reinel’s chart, the origin of the track used to put Newfoundland on the chart was the Azores archipelago. The secondary latitude scale is tilted as to indicate the direction of geographical north in the northwestern Atlantic (Fig. 7).29 Figure 7 – Excerpt of a chart by Pedro Reinel (ca. 1504) with two different latitude scales. The tilted scale near Newfoundland, indicating the direction of geographical north, was only applicable to the region. Bayerish Staatsbibliothek, Munique (Cod Ican 132). In 1539 Pedro de Medina (1493-1567), an influential humanist and author of the celebrated Arte de Navegar (1545), arrived in Seville with a royal authorization for making charts in the Casa de la Contratación. This authorization triggered a strong conflict between the Piloto Mayor Sebastian Caboto (ca. 1474-1557), who was responsible for the official master chart (the Padrón Real), and the cosmographer Diego Gutiérrez (ca. 1485-1554), who held the effective monopoly of chart production and sale. The problem was that the charts with multiple latitude scales produced by Gutiérrez, with the approval of Caboto and the support of pilots, did not comply with the Padrón Real. The dispute lasted for several years and involved other members of 29 The reason for this expedient was the fact that it was geometrically impossible, as referred before, to conciliate the three elements of information, in the presence of magnetic declination. For a more detailed explanation, see Gaspar, From the Portolan Chart of the Mediterranean (note 6), 118-119. 15 the Casa, including the cosmographer Alonso de Chaves (ca. 1492-1587). It was finally resolved by a decree of Prince Philip (the son of Charles I of Spain), in 1545, who ordered that all charts should comply with the royal master chart. In other words, the use of multiple latitudes scales was banned from official cartography. This well-known incident has been discussed by several authors, who focused on the aspects related to the “scientific truth” and how it was assessed in the sixteenth century. 30 But once again, what was really at stake here were two distinct and irreconcilable concepts of cartographic accuracy: the geographical (absolute) and the navigational. While most Spanish cosmographers considered that nautical charts had to comply with the Ptolemaic geographical standards (that is, accurate latitudes and longitudes), pilots were unable to reconcile them with the practice of navigation. The problem was technically complex, and to consider that the reason rested on only one of the sides of the argument would be an oversimplification. Charts with a secondary latitude scale for Newfoundland continued to be produced, and it is curious to realize that the only surviving Spanish chart of the kind was made by the same Diego Gutiérrez in 1550, well after the royal order forbidding their production was issued (Fig. 9). Also remarkable is the fact that the magnetic course and the distance measured from the Azores to Cape Race (the southeastern tip of present-day Avalon Peninsula in Newfoundland) in those charts are very accurate, a fact that eloquently explains the support of the pilots.31 An alternative solution that would satisfy the cosmographers’ requirement would have been to correct magnetic declination in all courses measured with the marine compass. This would make positions determined with the point of fantasy and the set point to be compatible, implying that the main latitude scale of the chart could be used in the whole charted area. But Spanish pilots were not willing, or prepared, to determine magnetic declination on board, and the process would imply costly changes in navigation and nautical cartography. Not only pilots would have to start using corrected compass courses, but also new surveys and new charts would have to be 30 Ursula Lamb, ‘Science by Litigation: A Cosmographic Feud’, Terrae Incognitae, 1, 1 (1969): 40-57. See also Alison Sandman, ‘Mirroring the world: sea charts, navigation, and territorial claims in sixteenth century Spain’, in Merchants and Marvels: Commerce, Science, and Art in Early Modern Europe, ed. Pamela Smith and Paula Findlen (New York, Routledge, 2001), 83–108’; Antonio Sánchez, La espada, la cruz y el Padrón, Madrid: CSIC (2013), 229-261; María Portuondo, Secret Science. Spanish Cosmography and the New World (Chicago and London: The University of Chicago Press, 2009). 31 Five charts of the sixteenth century were analyzed: Pedro Reinel (ca. 1504); Lopo Homem (ca. 1550); Diego Gutiérrez (1550), an anonymous Portuguese chart of the Atlantic kept in the Bibliothèque nationale de France (ca.1560); and Sebastião Lopes (1583). In all of them the error in the magnetic course between the Island of Terceira (Azores) and Cape Race (Newfoundland) is less than 2 degrees, the corresponding error in the distance is less than 10.5 Castilian leagues (0.6 latitudinal degrees), and the latitude error of Cape Race, as measured in the secondary scale of latitudes, is less than one degree. To estimate the value of magnetic declination in the region, during the sixteenth and seventeenth century, the geomagnetical model CALCS7K of Korte and Constable was used. See M. Korte & Catherine Constable, ‘Continuous geomagnetic field models for the past 7 millenia: 2. CALS7K’. Geochemistry, Geophysics, Geosystems, 6, 1, 2005. 16 made to comply with the new standards. Browsing through the Iberian charts of the sixteenth century quickly makes clear that such corrections were never made. Figure 8 – Excerpt of the only surviving chart by Diego Gutiérrez (1550), with three different latitude scales. Two of them are shown in the figure: one on the right, for Newfoundland, which appears rotated; and the other on the left, for North America. The third latitude scale, not shown in the figure was intended to be used in the western Atlantic and Caribbean Sea. Some decades later, near the end of the sixteenth century, the instrument maker and mathematician Andrés García de Céspedes (1560-1611) was appointed cosmographer of the Casa de la Contratación with the mission of reforming navigation and cartography. Contradicting the position of his predecessors, Céspedes clearly aligned with the pilot’s points of view by asserting that because the primary purpose of charts was to support navigation, they should be made in accordance to the navigational practices and the pilots’ wishes.32 As a university-educated scholar, he knew perfectly well that the ad-hoc projection used in nautical cartography, where uncorrected compass courses, distances and latitudes were transferred directly to the plane of the chart, was not geometrically consistent. However, no better alternative in which rhumb lines were represented by straight segments could be found at the time. Unfortunately, neither his improved Padrón Real nor any charts based on it survived to the present day. 33 32 For a more detailed discussion on the solutions proposed by Céspedes see Alison Sandman, ‘An Apologia for the Pilot’s Charts: Politics, Projections and Pilots’ Reports en Early Modern Spain’, Imago Mundi, 56, 1 (2004): 7-22. 33 García de Céspedes’s Hydrographia, bounded to his Regimiento de Navegación (Madrid, 1606), is a lengthy and rich treatise where the author explains his ideas about how to improve navigation and nautical cartography in Spain. Although considerable emphasis is put on the aspects related to the location of the Moluccas, whose political relevance was still felt at the time, the author makes a serious 17 The Mercator projection The clash between Ptolemy’s cartographic prescriptions and the nautical chart reached a dramatic climax with the work of Gerard Mercator (1512-1594). In 1569, the Flemish mathematician and cartographer presented his map of the world with the revealing title ‘’New and improved description of the Earth properly adjusted for use in navigation’’. 34 In this map, meridians and parallels form a regular mesh of rectangles where the spacing between parallels increases with latitude in such a way that rhumb- lines are represented by straight segments making true angles with the meridians. In theory, this solution would enormously facilitate the planning and execution of navigation because all courses between places – and not only a few, as in traditional latitude charts – were supposed to be accurately represented. Moreover, Mercator’s map was based on the latitudes and longitudes of the places, exactly as geographical maps. Apparently, this fact alone would resolve the long-standing incompatibility between Ptolemy’s paradigm and nautical charts.35 But was Mercator’s world map truly adjusted to navigation at the time it was proposed? As a matter of fact, it was not, owing to its intrinsic incompatibility with the navigational methods of the time, still based on compass courses. Only after the longitude problem was solved and the distribution of magnetic declination was known could the old model be abandoned for good and new charts constructed using the Mercator projection. A relevant question that could be posed is whether the new world map of Mercator – or rather, some larger scale nautical charts based on it – could still be used for navigation, assuming that pilots were prepared to use corrected courses. We do know that the graticule of meridians and parallels calculated by Mercator, to serve as a reference for latitudes and longitudes, was sufficiently effort to understand the pilot’s needs and to contribute to improving the information used to construct the charts. The contributions of two authors to the subject are profusely cited and praised in the treatise: Pedro Nunes’s Tratado en Defensam da Carta de Marear (1537), where the geometry of the contemporary charts is discussed; and João de Castro’s Roteiro de Lisboa a Diu (1538), where the reason for the apparent distortion of Africa on the charts is explained. For a technical discussion of Cespedes’s work, see Víctor Navarro Brotóns, ‘Astronomia y cosmografia entre 1561 y 1625. Aspectos de la actividad de los matemáticos y cosmógrafos Españoles y Portugueses’. Cromos, 3: 2 (2000), 362-368. For a discussion of Céspedes’s contribution to the improvement of navigation and cartography, see Alison Sandman, ‘An Apologia for the Pilot’s Charts: Politics, Projections and Pilot’s Reports in Early Modern Spain. Imago Mundi, 56: 1 (2004), 7-22. 34 The original title reads: Nova et aucta orbis terrae description ad usum navigatium emendate accomodata. 35 The purpose of Mercator appears to have been more ambitious than the use of his novel projection in navigation. In the legend Inspectori Salutem, three objectives are enumerated: ‘Firstly, to spread on a plane the surface of the sphere in such a way that the positions of places shall correspond on all sides with each other both in so far as true direction and distance are concerned and as concerns correct longitudes and latitudes […]; [secondly] to represent the positions and the dimensions of the lands, as well as the distances of places, as much in conformity with very truth as it is possible so to do […]; [thirdly] to show which are the parts of the universe which were known to the ancients and to what extent they knew them’. For a transcription of the Latin text of Mercator’s legends and its translation into English see ‘Text and translation of the legends of the original charts of the world’, Hydrographic Review, 9:2 (1932): 7-45 at 11.13. The Walther Ghim biography is in Gerardus Mercator, Atlas sive Cosmographicae Meditationes de Frabica Mundi et Fabricati (Duisburg: 1595). 18 accurate for navigational purposes.36 The question is: how accurate are the geographical coordinates of the places, as well as the directions between them? Two types of errors affect the latitudes and longitudes of Mercator’s world map: those that originated in an insufficient knowledge of the geography of the regions, such as in the depiction of South America; and those associated with the distortions of the cartographic sources used in the compilation. Only the second type of error is relevant for the present discussion. 37 Figure 9 – Grid of meridians and parallels implicit in the Mercator world map of 1569 (left) compared with its own geographical graticule (right). Notice the convergence of meridians and parallels in the northern hemisphere and the irregularities of the supposedly straight and parallel lines. Reproduced from Gaspar, ‘Revisiting the Mercator World Map of 1569’ (2016), p. 8. Figure 9 illustrates the grid of meridians and parallels implicit in the Mercator world map of 1569 (left), which was estimated using a sample of places with known latitudes and longitudes, side by side with its own geographical graticule (right). If the coordinates of the places were all exact the two grids would be coincident, which is not the case. Several errors can be spotted by visual inspection of the map’s implicit geographical grid (left): the convergence of meridians in the northern hemisphere, which is not supposed to occur in a Mercator projection; errors in the latitudes, especially in the northern Atlantic and the eastern Mediterranean, which are reflected in the irregularity of the parallels; and the unequal spacing of meridians in various regions, most especially in northern Africa and the Mediterranean. This irregularity is better perceived by comparing the coastlines of Mercator’s world map with those in a 36 The method used by Mercator to calculate his projection was the object of two recent articles where it was demonstrated, historically and numerically, that he has most likely used a table of rhumbs to calculate the spacing of the parallels. See Joaquim Alves Gaspar & Henrique Leitão, ‘Squaring the Circle: How Mercator Constructed His Projection in 1569’. Imago Mundi, 66, 1 (2013), 1-24; Henrique Leitão & Joaquim Alves Gaspar, ‘Globes, Rhumb tables, and the Pre-History of the Mercator Projection’. Imago Mundi, 66, 2 (2014), 180-195. 37 A detailed assessment of the accuracy of Mercator’s world map is in Joaquim Alves Gaspar, ‘Revisiting the Mercator World Map of 1569: an Assessment of Navigational Accuracy’, The Journal of Navigation, 69 (2016), 1183-1196. 19 modern Mercator representation, as done before in this article with the Cantino planisphere. This exercise makes clear that the eastward displacement and stretching of Africa and Brazil, affecting the Cantino planisphere, are still present in Mercator’s world map (Fig. 10). Figure 10 – Excerpt of Mercator’s world map of 1569, depicting part of the Atlantic and Indian Oceans, compared with the coastlines of a modern Mercator’s map (white lines). Notice the longitudinal displacement and stretching of Africa and Brazil, identical to the one in the traditional cartography of the sixteenth century, and the artificial enlargement of the Mediterranean, shared by other planispheres of the sixteenth century. How should the persistence of these distortions be interpreted, knowing that the novel projection was intended to represent latitudes, longitudes and rhumb line directions accurately? The discrepancy is easily explained by Mercator’s erroneous assumption that both the latitudes and the longitudes were approximately correct in contemporary cartography. This should not come as a surprise given the fact that Mercator shared with his contemporaries the false idea that the meridians and parallels implicit in the contemporary charts formed a square grid, a conviction that is clearly stated in one of the Latin legends of the map: 38 “On the navigator’s charts the degrees of longitude, as the various parallels are crossed successively towards the pole, become gradually larger with reference to their length on the sphere, for they are throughout equal to the degrees on the equator, whereas the degrees of latitude do not increase at all”. 38 See Gerard Mercator (1932), ‘Text and translations of the legends of the original chart of the world by Gerhard Mercator issued in 1569’, Hydrographic Review, 9, 2, 7-45. For a discussion about the misinterpretation of the geometry of the early modern nautical charts, see Gaspar, Blunders, Errors and Entanglements (note 11), 186; idem, From the Portolan Chart of the Mediterranean (note 6), 33-34; idem, ‘The Myth of the Square Chart’, e-Perimetron, 2, 2: 66-79. 20 Under this assumption, all that was needed to transfer the geographical information from the available charts to the projection was to take note of the latitudes of the places (as measured using the latitude scale of the charts), as well as of the apparent longitudes, whose scale was supposed to vary linearly in the east-west direction. Coming back to the question posed above: could Mercator’s world map still be used for navigation at the time it was proposed, assuming that the pilots were prepared to use corrected compass courses? The answer to this hypothetical question is no, because the latitudes and longitudes of the places on the map, and hence the courses connecting them, were not accurate enough for navigational purposes. To construct an accurate Mercator’s representation, new global surveys were – once again – required, this time based on the latitudes and longitudes of the places. 39 Again – as in the case of Nunes – not being fully aware of the artisanal context in which nautical charts were constructed and used led Mercator to misinterpret the information imported from them. Final remarks Pre-Mercator nautical charts are complex objects whose inner geometrical properties are only now starting to be understood. Rather than attempts at a faithful geographical depiction of the world, they were tools aimed to accomplish a specific purpose – to support marine navigation – a fact that is often mentioned in contemporary sources. Thus, they carry inside their internal geometry the imprint of the activities of pilots on board, that is, a “signature” of their artisanal origin and of the techniques used to navigate. Any hope for a better understanding of these cartographic artefacts requires a full acquaintance with the contemporaneous navigational practice. More specifically, any attempt at evaluating the accuracy of early modern charts using modern concepts, or comparing them to modern representations, should be made bearing in mind their purpose and the way they were used on board. Ever since the Middle Ages, geographical maps and nautical charts have co-existed as two distinct representations of the known world, but it was only in the first decades of the sixteenth century that the confrontation between the two models became a critical issue. Following its first printed edition in 1478 and the numerous other editions in the next decades, Ptolemy’s Geography was widely disseminated and studied in Europe. At the same period, because of the oceanic voyages of exploration, 39 Attempts at using the Mercator chart for navigation are known to have been made from the late sixteenth century onward, by the Dutch and English. On these experiments, see Sarah Tyacke, ‘Chartmaking in England and Its Context, 1500-1600’, in David Woodward (ed.), The History of Navigation, Volume III – Cartography in the European Renaissance (Chicago & London: The University of Chicago Press, 2007), 1722-1753, at 1743-1745. For a general discussion on the resistance of pilots against the adoption of the Mercator projection in navigation, see Mark Monmonier, Rhumb Lines and Map Wars: A Social History of the Mercator Projection. Chicago & London (2004): The University of Chicago Press. 21 nautical cartography was undergoing enormous changes, and became the single most important source of geographical information on a planetary scale. Inevitably, scholars compared and tried to harmonize Ptolemy’s data and cartographic prescriptions with the image of the orbis terrarum conveyed by nautical cartography. This effort proved to be extremely complex and a source of constant tension between those working with maps and those working with charts. The clash brought into play many different factors, such as authority, imperial ambitions, social and professional environments, etc. But, at its core, lay a crucial and yet extremely subtle technical issue that was not grasped by those protagonists: the fact that the two representations were impossible to reconcile. The attempt of Gerard Mercator to construct a chart ad usum navigantium, based on the latitudes and longitudes of the places, was a serious effort to reconcile the two cartographic paradigms. We have seen, however, how his achievement was much ahead of the time it was proposed. In its broadest sense, the argument of this paper is about knowledge production in early modern Europe. By focusing on cartography, it adds to the already abundant literature on artisanal practices and their role in the shaping of science during that period. More specifically, it shows how certain objects - nautical charts - were deeply influenced by the practices of the artisans who used them, but, simultaneously, how such influence was almost invisible to contemporary scholars and present-day historians. Our argument puts in evidence the deep level of complexity in the interactions between scholars and craft practitioners. Even when contact and cooperation between these two groups was intense and stable – in fact, even when this contact was enforced by legislation – and when the artifacts (instruments, charts, etc.) were the same, deep incomprehension might arise. One final consequence of our study is bringing to light the existence of a major debate that took place in Europe around the clash between Ptolemy’s geographical paradigm and nautical charts. It is no surprise that this debate has been almost unnoticed by historians and that its nature was not correctly understood. In a general sense, this may have to do with the difficulty in knowing, and fully integrating in the historical discourse, the practices of artisans. But in a deeper sense, the reason is mostly technical. Only today, with the use of digital cartometric methods of geometrical analysis, can we fully understand the true nature of nautical charts, and therefore, the technical reasons behind those misunderstandings, and conclude: first, that historically, nautical charts can only be understood in the context of the specific navigational methods they were intended to support; and second, that a nautical chart should not be considered as a true geographical map but as an instrument for navigation. Acknowledgements: this project received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement 714033-Medea-Chart / ERC-2016-STG). 22