Certain arguments against the hypothesis that portolan charts were genuine late-medieval cartographic products

ABSTRACT

A sample of 12 representations of the Adriatic Sea on portolan charts, created between the late thirteenth century and the late 16th, were subjected to a series of tests in order to examine their navigational applicability. However, the by-products of their results strongly suggest that it is impossible that portolan charts were, in terms of their geometry, genuine late-medieval cartographic products. It was demonstrated that, during the 300 years of their continuous production, their accuracy was not improved, and their anticlockwise tilt was not chronologically aligned with the dynamics of magnetic declination across the research area. Moreover, it was explicitly demonstrated that bearing measurements which could have been performed aboard ship (from sailor's perspective) cannot be used to create charts which display such levels of geometric accuracy, not only in the late medieval period, but in general.

ABSTRAITE

Un échantillon de 12 représentations de la mer Adriatique sur les cartes portulans, créées entre la fin du treizième siècle et la fin du 16^ème siècle, ont été l'objet d'une série de tests pour examiner leur adéquation à la navigation. Toutefois, les produits dérivés de ces résultats suggèrent fortement qu'il est impossible que les cartes portulans soient, au niveau de leur géométrie, de véritables produits cartographiques de la fin du moyen âge. Il a été démontré, pendant les 300 ans de leur production continue, que leur précision ne s'est pas améliorée et que leur inclinaison dans le sens inverse des aiguilles d'une montre n'était pas chronologiquement aligné avec la dynamique de la déclinaison magnétique dans la zone de recherche. De plus, il a été explicitement démontré que les mesures de relèvement qui auraient pu être réalisées à bord des navires (du point de vue du marin) ne peuvent pas être utilisées pour créer les portulans avec un tel niveau de précision géométrique, ni à l'époque médiévale, ni en général.

KEYWORDS:

• Portolan chart origins
• portolan chart accuracy
• terrestrial navigation
• Adriatic Sea

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The Pisan chart is the property of Bibliothèque nationale de France, France, Call No.: GE B-1118 (RES). It was labelled ‘Pisan’ because it was once the property of a Pisan family, but it was probably produced in Genoa (Crone, 1953, p. 30).

2 The Cortona chart is the property of Biblioteca dell'Accademia Etrusca di Cortona, Italy, Call No.: Port.105; the Avignon chart is the property of Archives départementales de Vaucluse, France, Call No.: Port 01, 3E 54, 888, while the Lucca chart is the property of Archivio di Stato, Lucca, Italy, Call No.: Cornice 194/I.

3 The Catalan Atlas or Carta Catalana (property of Bibliothèque nationale de France, France, Call No.: Espagnol 30), was made by Jewish cartographer Abraham Cresques and his son Jehua Cresques in 1375, and represented a gift from the king of Aragon to the king of France (Freiesleben, 1979, p. 85; Taylor, 1957, p. 116)

4 The original of Benedetto Cotrugli’s De Navigatione Liber is the property of Yale University, Beinecke rare book and manuscript library, Call No.: MS 557.

5 Roman emperor Antoninus Pius ruled between 138 and 161 AD, while Caludius Ptolemy died in approximately 168 AD (Kotruljević, 1464 (2005), p. 217).

6 Original text:  … Lo qual Ptholomeo fo nel tempo de divio Antonino et scripse mirabilmente. Fo grande astrologo et geometra, et ordinò et divise le mesure et le proportioni dello mare, cielo et terra, et misurò tucto per lo compasso cieleste et descripse lo mappamundo donne nui havimo la cartha dello navigare, la qual c’insegna lo andar per mare et non ci lassa errare … nulla de mino semo multi obligati ad Ptholomeo perché allo giorno de ogi non erramo. Ançi, quello facievano in prima a casu et zero arbitrio, hora havimo reducto per venti et mesure.

7 This record might be a glimpse into an idea about how their contemporary users managed to use the charts without exposing them to rapid wear.

8 The projected extent of the Adriatic Sea basin along the Y axis on a Mercator map φ₀= 42.7° made to a map scale of 1:1 (which approximates its latitudinal extent in reality with an accuracy of about 0.5%) is 646 km.

9 The average map-scale factor in Table 2 (SCF_AVG) is, for the sake of simplicity, put in its shorter form (as its ‘true’ scale factor divided by 1,000,000), in accordance with which millimetres on a chart can be converted directly into projected kilometres in reality, and vice versa.

10 For the sake of graphical-content standardization, 2×SD ellipses have also been used in Figure 1, on which they are significantly larger; the reason for this is that earlier charts were substantially less similar to one another, which affects the cumulative result.

11 Geomagnetic field values from the CALS3k.4 model for the Adriatic Sea (centred at φ = 42°N and λ = 16°E) were obtained from the Geomagia web-calculator: http://geomagia.gfz-potsdam.de, GEOMAGIA50.v3.2 database, on 15th November 2019.

12 There is a possibility that knowledge of the dipole and orientation properties of magnets already existed in antiquity in the Mediterranean. Cotrugli quotes Aristotle's work On Stones (which has not been preserved to modern times): ‘The angle of a magnet is the ability to attract iron to the zoron, that is, to the north, and thus sailors use it. The angle of this magnet, opposite to that, attracts to the aphon, that is, to the south pole. And if you bring the iron closer to the angle of zoron, it turns the iron towards zoron, and if it approaches the opposite corner, it turns towards the aphon’. Cotrugli then adds to it: ‘Thus, according to this authority [Aristotle], we have a double magnet that serves for different poles.’ (Kotruljević, 1464 (2005), p. 153).

13 Atmospheric conditions such as air humidity, fog or clouds, which additionally constrain the viewshed, were neglected in the viewshed analysis.

14 Points along the equal-distance lines were plotted 3 nautical miles (M) apart from each other, because the pilot test conducted beforehand showed that, for an observer whose eye-height is 3 m, due to Earth’s curvature, the horizontal component of the viewshed ends (or the horizon appears) at an approximate distance of 3 M (5.6 km).

15 Quarta de sciloccho verso levante.

16 Per schilocho.

Additional information

Notes on contributors

Tome Marelić

Tome Marelić, PhD, is a research assistant at Department of Geography, University of Zadar, Croatia. His prime research interest is quantitative (cartometric) analysis of old maps and charts with the application of GIS and statistical software. His current research is focused on geometric properties of Adriatic Sea basin coastline renderings on late medieval and modern navigational charts.