I.
THE MYSTERY OF THE PORTOLAN CHARTS
Portolan charts were mostly made of the coasts of countries bordering the Mediterranean and Black Seas, and the Atlantic Coasts of France, Spain and North Africa. They first appeared during the last part of the 13th century and were made through the beginning of the 17th century. With respect to these coastal lands, the portolan charts are very accurate as to relative latitude and longitude. Indeed, they have been found to be as accurate as a Mercator projection for this area. Such accuracy requires the use of either a Mercator projection or some similar projection to be incorporated into the original map - “the map projection is not accidental but must have been intentionally designed into the charts as part of their construction.”
4 However, Mercator did not develop his projection until the mid 16th century, almost three hundred years after portolans first appeared. Prior to Mercator, Europeans did not have the ability to make such accurate maps.
Moreover, in order to have a map of this accuracy, it is necessary to have accurate data for longitude and latitude for the many places on a map or chart. By the 13th century, people in Europe could get latitude, but no one, so far as we know, was able to determine longitude on a regular basis. Because the Mediterranean goes primarily from West to East, finding longitude is essential to making an accurate map of this area.
Finding longitude requires the use of an accurate, practical method of timing to determine how far East or West of a prime meridian a particular location is. Using the eclipse of the moon is one such method, but it is not practical because it does not occur frequently enough. Europeans did not have a practical method until after Galileo, peering through his telescope, discovered the four major moons of Jupiter in 1610, and after these moons had been sufficiently plotted so that their positions could be used as a universal time mechanism for determining the time at the prime meridian. This universal time could then be compared to local time to determine longitude for any particular place on the earth. Plotting these moons took about forty years of astronomical study. But even after this development in Western Europe, the portolan charts continued to be the most accurate maps for this area until the 18th century.
5 The earliest known and existing portolan chart is the Carte Pisane, made about 1290. While this chart is not as developed as ones which were made a few years later, within the Mediterranean it is as accurate as the more developed portolans. A good example of an early, more developed portolan, is the Dulcert Portolan of 1325. (
Plate 1). This chart accurately shows the coasts of North Africa, Spain, France, the Mediterranean and Black Sea. It also includes England and Ireland, although these are not as accurate. And it goes East as far as the Persian Gulf and the Tigris River.
The portolans are called charts rather than maps because they primarily show the coasts accurately, as opposed to the interiors, and are believed to have been used by sailors during the 13th to 18th centuries. Although the portolan charts are accurate for relative longitude and latitude, none of them show either. Instead they have 32 lines emanating from centers placed throughout the chart. Later portolans have these lines emanating from a compass rose with the lines explicitly representing directions. These lines are called rhumb lines, and are believed to have been inserted to help sailors plot their course on these charts.
There is no known mapping history which accounts for the sudden appearance of these highly accurate charts at the end of the 13th century. As Brown, in “The Story of Maps” states:
“[Around 1300] suddenly come several copies of a full-blown chart of the Mediterranean and the Black Sea, with not a shred of evidence as to who made it or how it evolved. But it was much too accurate and detailed to be the work of any one man or any one group of navigators, nor could it represent the surveys of any one generation; the area is much too large and the details too complex. For three hundred years, from 1300 to 1600, this early chart of unknown origin was copied and recopied with few alterations or improvements.”
6 Indeed, the charts made in the first half of the 14th century tend to be more accurate than later portolan charts. We of course recognize that improvements to science tend to be developed over time. Yet, although portolan charts were made until the beginning of the 17th century, there was no improvement in these charts in terms of accuracy after 1340. Or, as Tony Campbell, a leading authority on portolan charts, states: “By 1340, because of the work of the Vescontes and Angelino Dalorto, the portolan charts had reached what must have been generally accepted as to their final form, as far as their constructional underpinning, drafting techniques and hydrography were concerned. The toponymy [place names], however, would continue to be steadily amended….”
7 The fact that improvements to the portolan charts were not developed by map-makers and surveyors shows that people making portolan charts in the period of the 13th through the first half of the 17th century had no real idea how to develop a map, but were simply copying earlier maps handed down from prior civilizations. And there is no report of anyone ever making a survey to make a more accurate portolan chart for this area of the world. Emphasizing this copying is the fact that neither latitude nor longitude lines are reflected on these charts, even though the charts accurately portray relative latitude and longitude between the places depicted.
Some scholars have attempted to explain these charts as having been developed from sailors consistently sailing to the same ports and taking sightings of the land they passed. But, as probably most sailors would tell, it is impossible to measure distance that accurately just by sailing by. In addition, this theory does not explain how relative accurate longitude between places on the Atlantic Coast and Mediterranean points was obtained because this would entail sailing through the land, including the Pyrenees. Nor would aggregating the measurements of many sailors explain this as the concept of improving accuracy by calculating the arithmetic mean was not introduced into scientific practice until the end of the seventeenth century.
8 One of the features of the portolan charts is that virtually all of them have a rotation of about 9 degrees West of true North. That is, North on the portolans is 9 degrees West of true North. We can see this on the portolans by taking two known locations on the same latitude and drawing a line between them. In the example (
Plate 2), a line is drawn between Gibraltar and the center of Rhodes as both are on the 36th parallel, and if true North on the map were correct, this line would be horizontal. Instead, we find that there is an angle of about 9 degrees between the line from Gibraltar to Rhodes, and a line drawn at right angle to true North as shown on this portolan.
Finding true North is easily accomplished and was certainly known to the men of science in the 13th through the first half of the 17th century. The most basic way, known to ancient peoples, is using a gnomon (stick planted on a level surface) and measuring the suns’s shadow during the day with the midpoint being true North, or through the use of the North Star (and plotting its slight divergence from true North). Thus, any map maker, who had the ability to develop a portolan chart from surveyed locations, would have certainly known how to make a chart with true North at the top.
Some have said that the 13th to 17th century portolan chart makers deliberately made North 9 degrees West of true in order to make it easier for sailors to plot compass courses because in large parts of the Mediterranean the declination (difference between true North and magnetic North) was about 9 degrees East, and rotating the charts West would compensate for this. In this way, it would not be necessary for a sailor plotting a course on a portolan chart to compensate for declination in determining his compass course.
This justification for the 9 degree rotation seems unlikely. In the first place, in the 14th century declination was not known.
9 If declination was not known in the 14th century, it seems highly unlikely that the makers of the portolans at that time adjusted their charts to compensate for declination. Moreover, when sailors gradually discovered that the compass did not point to true North, they blamed the compass makers for making an inaccurate compass. The result was that compass makers started to adjust the compass card so as to make the compass point to true North at that specific location.
10 According to an English instrument maker, Robert Norman, “The Newe Attractive” (1581), compass makers in Sicily, Genoa and Venice made compasses with no adjustment because there was no “variation” (as declination was then called) from true North at those locations. In Denmark and Flanders, the compass card pointed “3 quarters of a point to the Eastwards of the North” and sometimes as much as a whole point East of North. Those made in Spain, Portugal, France and England pointed a half point to the East.
11 These facts indicate that declination was not understood in Europe during the period when the portolan charts were developed in the 14th and 15th...