MEDIEVAL DISTORTIONS : THE PROJECTIONS OF ANCIENT MAPS

Estimates of the map projection employed for an ancient map is a prerequisite for a variety of other studies. The preliminary evaluation presented here has yielded empirical equations for the Hereford map and illustrated the agreement of a Portolan chart with an oblique Mercator projection. HE study of ancient maps provides one of T the fascinating aspects of historical geography. Such maps can be analyzed for many purposes and from several points of view. The following comments refer only to the estimation of the map projection implied by the ancient mappaemundi and portolan charts. Evaluation of the map projection of these old maps is of assistance in the determination of the accuracy of the maps, and may provide insight into their method of construction. Modern theories regarding the ancients’ perception of the world also may require consideration of the map projection employed for maps. The maps in the two classes under investigation do not contain any indication of the terrestrial graticule of latitude and longitude. This has led some students to conclude that the maps are not based on any map projection.1 This point requires clarification. Certainly the lack of the graticule does not imply the absence of a projection. Even modern maps are occasionally published without this grid.2 More telling is the high probability the sphericity of the earth was unknown to, or was not considered relevant by, the individuals who constructed the maps. If this is the case the maps would be constructed as though the earth were flat. Inconsistencies between the plotting and the observational information then might arise; these inconsistencies could be attributed to the (unavoidable) errors in one or the other, or both. For a small Accepted for publication May 16, 1965. A. E. Nordenskiold refers to these maps as paratropical; Periplus, An Essay on the Early History of Charts and Sailing-Directions (Stockholm: Bather translation, 1897). R. E. Dahlberg, “Maps without Projections,” The Journal of Geography, vol. 60 (1961), pp. 213-18. area the errors and inconsistencies might be quite small and could go unnoticed. Inconsistencies are not necessary or inevitable, however. No set of observational information specifying the location of any terrestrial position by not more than two independent measures will lead to inconsistencies when plotted. This is true whether the earth i s considered round or flat. In either of the above events it is correct to say that the map is not based on a map projection only in the sense that the cartographer involved was not consciously employing a map projection.3 But, as one learns from any elementary work on map making, every map requires a map projection. The ancient maps therefore are implicitly referred to some map projection. The next difficulty, it seems, occurs if it is assumed that this implicit projection is one of the now-known projections. For example, the portolan charts have been compared with charts drawn on Mercator’s projection and on the square pr~jection.~ Suppose that the match is sufficiently poor to conclude that the chart is not drawn on either of these two pro3 Similar comments apply to an engineering survey of a small area. * H. Wagner, “Das Ratsel der Kompasskarten im Lichte der Gesamtentwickelung der Seekarten,” Verhandlungen, XI Deutsches Geographentages, Bremen, 1895, pp. 65-87; E. Steger, “Untersuchung uber italienische Seekarten des Mittelalters auf Grund der kartometrischen Methode” Dissertation, Gottingen, 1896; M. Fiorini, Le projezioni delle carte geogrufiche, Bologna, 1881; A. Brewing, “Zur Geschichte der Kartographie,” Zeitschrift fur Wissenschaftliche Geographie, I1 (1881), p. 168 ff; M. A. Clos-Arceduc, “L’Enigme des Portulans: Etude sur la Projection et le mode de construction des cartes a rumbs du XIV“ et du Xv” Siecle,” Bulletin, Comite des Travaux Historiques et Scientifiques, Section de Geographie, LXIX (1956), pp. 215-31. The square projection is also known by the names plate carrbe, simple cylindrical, and cylindrical equal-spaced projection.