i,6o 



NATURE 



[December 15, 1904 



mosaic of the •sixth century or to Matthew Paris's thirteenth 

 century " England." The chief errors which Ptolemy had 

 imparted to the shape of the Mediterranean are corrected. 

 The main features of the great inland sea are jjresented 

 with a correctness and a minute detail which, at the most 

 casual glance, immediately distinguish portolan work 

 from any preceding variety of cartography. No attempt is 

 made to fill up the interior of the lands — continental or 

 insular — of which the coasts are portrayed ; such attempts 

 are made later, it is true, but they are obvious and confessed 

 additions to the primitive, normal, or typical portolan. 

 But, along the shores in question, all points important for 

 navigation are drawn with great care ; small islands, bays, 

 cliffs, and headlands — of no great general importance, but 

 vital to the coaster — are often depicted in disproportionate 

 size; all the ports especially suitable for calling, watering, 

 and revictualling are indicated with the especial honour of 

 red colouring ; even shallows are frequently marked, denoted 

 by a sign still used at the present day ; the very large number 

 of shore-names testifies to the minute knowledge under- 

 lying the work. Thus along the north coast of the 

 Slediterranean we have (by a.d. 1320) about 620 names; 

 on the coasts of the Black Sea and Sea of Marmora about 

 260; on the coasts of Asia Minor and Syria about 160; on 

 the north coast of Africa about 240 ; in all some 1280, 

 without counting island names — which are very numerous— 

 or the names which fringe the western coast of Europe to 

 the mouth of the Elbe, and the western coast of Africa to 

 Cape Xun, or Non, at the extreme south-west of Morocco. 

 In respect to these shores — let us say from Hamburg almost 

 to the Wady Draa, and from Gibraltar to Azov, Poti, 

 Batum, .^lexandretta, Jaffa, and the Nile — the portolani 

 soon become fixed in the pattern they permanentlv retained, 

 a pattern which gradually triumphs' over every other — even 

 the revived Ptolemaic, to which scholars clung so 

 desperately and so unhappily. We may therefore regard 

 the great mass of these works as mere copies of a 

 few normal or typical designs which were completed (at 

 least in all their essential parts) before the outbreak of 

 the Hundred Years' War, and a good twenty years before 

 the battle of Crecy. How- closely the orig'inal type was 

 followed may be guessed from the fact that the portolan 

 colours — used according to certain definite rules — are un- 

 altered for long periods of years, and through scores of 

 examples. Thus red or reddish-brown is alwavs kept for 

 the Red Sea, and long after the Turkish conquest of Rhodes 

 that island regularly appears in white with a black cross. 



Instead of lines of latitude and longitude (or substitutes 

 for such lines, as we find in the " Palestine " of Marino 

 Sanuto, c. a.d. 1310), a net of loxodromes is employed on 

 (or has, at any rate, been added to) the portolani even of 

 the earliest time. These loxodromes are straight lines in 

 the direction of the various winds, proceeding from a number 

 of crossing-points regularly distributed over the map. But 

 m this loxodrome net-work, in sharp contrast to all other 

 features of the portolan map-type, there is almost infinite 

 variation ; one seldom comes across two designs of exactly 

 similar character in this respect. 



A distance-scale, with the same unit of length, occurs on 

 all the portolani ; this unit (which has been called the portolan 

 mile) IS estimated with much care by Nordenskjold at 5830 

 metres ; while of all known mediaeval measures, that which 

 corresponded most nearly with the " portolan mile " seems 

 to have been the Catalan legua. A Catalan league 

 therefore, it is suggested, may have furnished the basis of 

 the portolan measure, and the portolan tvpe of map may 

 have originated (in part at least) among Catalan mariners'. 



Baron Nordenskjold, indeed, does not hesitate to ascribe 

 to the portolani an entirely Catalan parentage. But, 

 admitting that one germ of the first true maps mav have 

 existed at Barcelona or some other centre of Catalonian 

 trade and seamanship, I cannot but think that another 

 germ still more active and important w-as to be found in 

 Italy, and -above all in the north-west— in Genoa and Pisa. 

 For, remembering the indications in Dati's " Sfera," we 

 may agree with Theobald Fischer that map sketches of 

 portolan type, and with the practical object of helping 

 navigation, were almost certainly drawn in Italy, and by 

 Italians, before 1300. Remembering, also, that of the 



NO. 1833, VOL. 71] 



existing portolani all the earliest examples are unquestion- 

 ably Italian — and that, of some 500 known, 413 were 

 executed by the countrymen of Carignano and Vesconte — 

 we shall not be ready to deprive Italy of the first place in 

 the creation of the oldest scientific maps. Even if that 

 creation w-as, as seems probable, an " Homeric " feat — the 

 piecing together (with additions and improvements) of a 

 great number of small sectional coast-surveys — yet this 

 earlier stage, only recorded in Italian manuscripts, seems 

 no less due to the seamen of the peninsula. 



Can we throw any other light upon the origin of the 

 portolani? 



In 1881 Fiorini suggested that West-European mariners, 

 such as those of Italy, learnt from the Byzantines the art 

 of making and using maps founded on careful draughts- 

 manship and close study of distance (i.e. portolani of a 

 kind) as early as the eleventh century. This idea has been 

 accepted by Theobald Fischer, and has been treated with 

 great respect by other scholars. Vet it is surrounded by 

 difficulties. For no Greek portolan has yet been found, nor 

 is Greek influence anywhere to be detected in the language, 

 legend-allusions, contours, or other details of the early 

 portolani. Fragments of Latin, fragments of Italian and 

 Catalan dialects, fragments of a lingua franca composed 

 of various Romance tongues — these are the media through 

 which the early portolan draughtsmen convey information. 

 But of Greek they make no use, and of Byzantine geo- 

 graphy, history, harbours, or coast routes they show no 

 special knowledge. We may give weight to the fact that 

 the Byzantine navy was one of the chief Christian weapons 

 in the ninth, tenth, and early eleventh centuries; that 

 Constantinople was then the greatest trade centre in 

 Christendom; and that the seamen of the Greek islands 

 were very prominent in .Mediterranean navigation in the 

 age of the Byzantine revival (c. 86o-!o6o a.d.). But all 

 this is far from proving a Byzantine right to the " inven- 

 tion " of the portolan coast-chart,' even in the primitive 

 form of sectional pilot-maps of limited areas. 



It only remains to say that all genuine progress in geo- 

 graphical delineation followed the lines of the portolani; 

 that the accurate methods employed by them for coast-w-ork 

 were gradually applied to the interior of countries ; that in 

 spite of the contempt show-n for them by most of the learned 

 in the so-called Renaissance period, they were at last known 

 by their fruits and vindicated by the success of their type. 



.Ancient classical or pre-Christian maps were not w-ithout 

 certain merits, though we can only judge of them by the 

 tw-o remaining examples, the Peutinger table, originally a 

 road-map of Augustus's Empire, and the designs illustrating 

 the " Geography " of Claudius Ptolemy of Alexandria — both 

 surviving only in manuscripts of the central medijeval 

 period. After the modern age of oceanic discovery had 

 passed through its earliest and most difficult stages, the 

 Renaissance editions of Ptolemy (from 1474) played a very 

 important part in delaying geographical progress and re- 

 tarding the history of civilisation. But in the time of the 

 early portolani (say from 1300 to 1400) neither the work of 

 the Alexandrian astronomer nor the road-maps of the Roman 

 Empire were adequately known in western Europe. The 

 sixteenth and seventeenth centuries were not so innocent. 



Designs of the portolan type do not seem to have existed 

 even in the best ages of classical geography and e.xploring 

 activity; the old peripli were sailing directions, not drawn, 

 but written ; and the only .Arabic scheme of the sort which 

 has yet been found is certainly copied from a Christian 

 — and Italian — original. 



It is in the portolani, and especially in such a work as the 

 Laurentian design of 135 1, with its revelations of the Azores 

 and the Madeira group, and its still more startling sugges- 

 tion of the true shape of .-\frica, that we may find, perhaps, 

 the chief geographical teachers of Henry the Navigator and 

 his Portuguese. Never better than in these long-neglected 

 charts does the history of civilisation illustrate man's 

 change from empirical to scientific, from traditional book- 

 learning to the investigation of nature. The portolani long 



^ To NordenskjOld's wild theory, " Facsimile Adas," p. 48, that Marinus 

 ofTyreisthe real original portolan draughtsman, and that the ^Ia^inus 

 maps which Masudi saw before A.D. 956 were really portolani, we need nol 

 pay attention. 



