View Diagrams  Download Paper  Download Images 


In order to ensure the research carried out previously is fully recorded this text sets down the evaluation of the internal construction of the charts and atlases, their accuracies and foibles and thus ensures all of the research regarding the 1311 chart and the atlases is comparable to the many Portolan Chart texts that are on the website. It is perhaps wise for researchers who have not studied any of the previous texts to do so, as even in a cursory glance will unfold the story line regarding these charts, its atlases and its near neighbours by time scale and their problem areas. There remains however the enigma of how Petrus Vesconte drew atlases of the whole Mediterranean Sea basin and the Atlantic coast area when his only extant chart concerned the Eastern Mediterranean Sea basin from Corsica/Sardinia to the Black sea. The Riccardiana MS 3827 is still my research contender for the obviously required whole chart, because 2 years later he produces a complete atlas. But as most Portolan Charts are severely distorted the factors causing that must be evaluated and explained using a chart and its measurements. The appendix section is that evaluation. I am not a follower of the historian’s magnetic agenda and plainly state the facts of the distortions.


ChWr/1&ChWr2; Windrose construction on a portolan chart. Revisisons to the origins of the charts format as drawn. This text includes ChWr/2 as an appendix.
ChCs/1; Is the Riccardiana MS 3827 chart the earliest of the P Vesconte charts? Thus a comparative study of five portolan charts and a geographical Mercator chart. Who begat whom?
ChCs/2; Charts from 1300AD to 1350AD compared. Is the Riccardiana MS 3827 the master template?
ChCPS/1; The Carta Pisane. BNF Res Ge B 1118 Amb. A full and detailed examination of the chart.
ChMAT/1; Melange a trios avec l’une, MS 3827; Cortona, Vesconte 1311 and Dulceto 1330.
ChDIS/1; 10 distorted charts- LCN problems and critiques of HOC/1/19 p380-386 and LCP p 510-511.
ChCOR/1; the Cortona chart LCP C2; investigated and the Black Sea enigma quantified. This text has ChBLS/1 as the appendix for the Black Sea investigation.

There are many other texts which describe in detail the Portolan Charts and their attributes and thus I suggest a check on the complete web site chart section texts is carried out.


  • One degree of Latitude equals 75 Roman Miles of 8 stadia = 600 stadia
  • One degree of Latitude equals 90 miliaria of 6.667 stadia = 600 stadia
  • At 36N, one degree of latitude = 5/4ths x one degree longitude
  • Therefore 75 Roman Miles x 4/5ths = 60 RM per longitude degree
  • Therefore 90 miliaria x 4/5ths = 72 miliaria per longitudinal degree

The basic wind rose is 92 geometric units and Petrus Vesconte has drawn the wind rose using the scale bar and it is thus a basic 920 miliaria per quadrant section. The overall wind rose is thus 1840 x 1840 miliaria; latitudinally this gives a distance measure of 20.444 degrees and longitudinally a distance measure of 25.555 degrees. ChWr/1&ChWr2 texts.


The first extant and provenance for a Petrus Vesconte chart is dated 1311 and as diagram ChPV/1/D01 illustrates is limited to the eastern Mediterranean Sea basin and Black Sea areas. The texts such as ChMAT/1 and ChCPS/1 discuss this chart but do not include the distance measure evaluation to follow.
The basic setting out on vellum is a rectangle some c42 x 52 cms with the north/south measurement equalling 1840 miliaria. That is the basic Windrose setting out distance measure which Petrus Vesconte detailed in his 1318 atlas as text ChWR/1&2 indicates and explains. Thus the actual rectangle within the bordure is actually1840 x 2260 miliaria with the Black Sea forming an extension in the north east corner.


The latitudinal northern border could easily have been extended to encompass the Sea of Azov, but obviously the exact wind rose measurement was required at 1840 miliaria. The east/west setting out likewise could have been such that the Black Sea was fully encompassed but Petrus Vesconte chose to add only 420 miliaria to the western edge and thus set out the bordure. That measurement 420 miliaria is six times the smallest wind division of 70 miliaria and thus is 6/5ths of the major division of 350 miliaria. That echoes the change from Roman Miles to Miliaria, 75 to 90 for the degree of latitude measurement; 90 x 1.233Km and 75 x 1.4791Km, or the basic 111km distance measure of a degree of latitude which is suitable for investigating Portolan Charts.
The chart is set at c10.75 degrees skewed to Windrose north. That skewing has already been explained in previous texts and stems from the errors in placing Genoa north of its latitudinal position, caused by the map distortion in the west, and then the mis-measurement of the individual sections of the islands situate along the 9E longitude, Corsica and Sardinia to the North African coastline. Thus it is quite certain that Petrus Vesconte had drawn not only the corresponding western section of the Mediterranean Sea basin as a separate chart but also must have commenced with a chart which encompassed the whole Mediterranean Sea and Black Sea areas. This logic is confirmed later in the text.
Thus the mathematics of the Windrose, and the trignometrical ratio’s which arose there-from, 35/30/20/7, to facilitate the drawing of the 22 ½ degree subdivisions cannot be ignored as the overall bordure indicates those facts precisely and predate the 1318 atlas.
The two extant earlier (?) charts, “Carta Pisane” and “Cortona” have been fully evaluated and they use a scribed circle(s) to set out the wind-rose(s) and thus Petrus Vesconte has introduced or had knowledge of a far simpler and more useful methodology to establish the basic Windrose parameters. The fact that our knowledge of the methodology only comes from his 1318 atlas page which illustrates the setting out shows that in fact we are missing a large amount of data as this 1311 chart exemplifies by its measurements. Why wait until 1318 to explain the methodology when it was known before 1311. We are missing something here!
However, it should be noted that I do not consider the “Carta Pisane” to be a Portolan Chart as such, but in all probability a “map” which has had two very crude wind rose diagrams added, as well as a very curious set of squares and rectangles which appear to obscure original features on the “map”. The Cortona Chart could be described as a peregrination map, having text appended on the reverse for “Holy Land” travel all as described in text ChCOR/1. Thus it is probably also a converted map.


How a chart was probably constructed, that is if it was not a direct copy of an existing chart or map of the Mediterranean Sea basin, can be evaluated from the basic graticule which we can be assess from its draughtsmanship. There are two diagrams included here-in which illustrate the longitudinal divisions and latitudinal divisions separately before combining them on larger scale plots to illustrate the complexity of the whole layout. But, before discussing them in detail I would again draw your attention to the “Cortona” chart text, ChCOR/1 which has an appendix essay, ChBLS/1 describing the Black Sea and its illustration on various charts, including reference to the geographical plot. It includes this 1311 chart, but unfortunately due to the degradation of the “Carta Pisane” chart it is not included. However it serves to illustrate the complexities of the research required.



Commencing in the west, it is evident that the 9E longitude is capable of being accurately plotted from Genoa, through Corsica and Sardinia then to the east of the Iles de Galite and the North African coast to the east of Tabarka. That alignment can then be utilised at the eastern limit of the Mediterranean Sea to draw the 36E longitude and can be accurately positioned by the Gulf of Issus. When drawn it is positioned correctly and the charts apparent accuracy assured by the fact that it passes northwards correctly through the Peninsula of Crimea. Thus we have two parallel lines which can be used to determine the longitudinal measurement, provided we know where it was taken and used.
The most important latitude in history within the confines of the Mediterranean Sea is the 36N latitude which actually passes from the Strait of Gibraltar, to Malta, Rhodes and thence to the Gulf of Issus. Thus it can be plotted on the chart and immediately indicates that it is set at 90 degrees to the putative 9E and 36E longitudes. Therefore we have the beginnings of a graticule which can be used to assess the whole chart and its veracity.
Firstly, using the scale bars drawn on the chart we can measure the distance along the 36N latitude from the 9E to the 36E lines, that is 27 degrees, and determine the longitudinal measure, which in turn as we know the ratio at the 36N latitude we can determine the latitudinal measure as well. From the scale bar the 27 degrees longitude are each 72 miliaria at 36N and thus the latitudinal degree is 90 miliaria. Thus we understand that the scale could also have been 75 Roman Miles for the latitude and 60 Roman Miles for longitude at 36N.
It is therefore possible to draw the longitudinal lines across the chart and ascertain its internal accuracy. Study the lines and it becomes immediately apparent that the chart in terms of longitude is accurate, albeit skewed by c10 degrees. Only one area having been drawn awry, the Black Sea. The main areas of the Mediterranean Sea and the Adriatic Sea are correctly drawn, with the Black Sea being slightly extended in longitude as was generally found in the ChBLS/1 appendix text.
What it also indicates is that the 72 miliaria longitudinal measurement is constant.
Given that these areas were the subject of accurate distance measures taken in the Roman Era, this should not surprise anybody. The Roman road system followed the coast lines and from documents such as the Itinerarium Antonini and the Tabula Peutingeriana we can readily read of their achievements in survey measuring.
However, I caution everybody here not to jump to the conclusion that the 1311 chart just appeared as an accurate unit. The 1313 atlas belies that fact and gives cause for concern.



The latitudinal layout is all together different, having an accuracy and a distortion but visually acceptable. From the 36N latitude southwards to Alexandria at 31N the spacing is as predicted at 90 miliaria or 75 RM. The 37N latitude is also correct, but then the system goes awry. The measurements along the 9E longitude already indicated as generally in error on most Portolan Charts show that part of the 38N line is correct. The 39N line is misplaced as are the following lines to 45N. The main area of problematic latitudes, the Aegean Sea, identified also on many other charts by the fact that from the 36N latitude to the mouth of the Dardanelles, which is set at 40N it is in fact drawn one degree or 90 miliaria north; thus ensuring the Black Sea is mis-positioned. The following four diagrams have detailed distance measures and explanation for the discrepancies there-on.


This diagram illustrates Corsica/Sardinia/Sicily/Italy and the Adriatic Sea with the north eastern corner of Tunisia, Cape Bon and Malta. This section of most Portolan Charts is perhaps the first part to illustrate a distortion which affects the whole setting out to the east. The diagram has a full graticule appended and dimensions for the sections. The first task was to evaluate whether the 9E line and the form of Italy were correctly positioned. Thus when the angular alignment from Genoa south to Reggio di Calabria is checked against a geographical chart, the 1311 chart is an acceptable setting out. That is, Italy has been set correctly with its western coastline to both the longitudinal and latitudinal graticule even though it is skewed. But when the distance measures for the 9E line are assessed a picture emerges of consistent error. From Genoa to the N African coast, chart scale is 635 miliaria but geographically it is c800Km or c650 miliaria. The calculated distance is 659 miliaria and as such could be considered accurate given the scale of the chart, but that is not the whole storyline. Genoa to Cape Corse (43N) should be 1.4 x 111Km = 155.4Km = 126 miliaria and is drawn as 145. Corsica is drawn as 130 miliaria, but is in fact 146 miliaria and the 10 miliaria “Bouches de Bonifacio” is drawn as 25 miliaria. Sardinia is drawn as 225 miliaria when it is 211 miliaria thus ensuring that Sardinia to the N African coast is drawn 125 miliaria instead of the geographical 162 miliaria. Therefore, the latitudinal lines plotted against the chart form indicate the errors. But those errors are not consistent and as discussed vary by latitude.
Sicily has been drawn with its northern coast and eastern coast accurate for length, but with a minor distance difference of 20 miliaria for the southern coast. However given that the Italian mainland and Sicily are drawn correctly from the 36N latitude a check against the distance from Reggio to Venice indicates an extension of 90 miliaria; the east coast of Italy is extended by 80 miliaria, although the southern coastline is correct at 250 miliaria. Hence we can observe why the extended graticule is necessary for the Adriatic Sea, particularly as Genoa to Venice is lengthened by 55 miliaria. Thus Venice has been positioned probably one degree latitude northwards, and Genoa is already out of position compounding the error.


However given the scale of the chart, its size being only c42 x 52cms, many errors indicated upon the diagrams could be considered irrelevant as we are discussing only millimetres difference on a small freehand chart dated 1311AD. But as these self same errors can be shown to exist in the later Portolan Charts it is correct to identify them now.
As an aside, but relevant to the foregoing text I offer the following example. Although the Itinerarium Antonini is suspect because of various copy errors over the years there is one rather telling distance measure given twice; “Caralis Galatam usque insulam stadia DCCCCXC” (990) and then later, “Insula Galata a Caralis de Sardinia stadia DCCXXX” (730). But, Galata to Tabraca is given in both sections as stadia CCC (300). The actual distances are 1026 stadia and 353 stadia. Should any cartographer use one or the other these distances errors will occur? Thus perhaps the distortion at 9E has a Roman base?



The basic problems of Portolan Charts, the skew and latitudinal excesses is shown here, and, on what in fact is the first extant Portolan Chart we consider as the forerunner of the genre. The North African coastline is drawn reasonably correctly with the longitudes and latitudes correctly positioned. However the actual positioning of Berenice and Tolemais is incorrect and should show Berenice where Tolemais is noted and Tolemais should be on the northern coast. But that is a minor discrepancy compared to the Aegean Sea errors.
From the 36N line the 37N line is correctly positioned, but, then there is the extra degree of latitude inserted between the 36N and 40N lines such that it is 450 miliaria and not 360 miliaria the correct distance for 4 degrees latitude. The Peloponnese and Euboea Island have expanded to fill the void and Saloniki has been positioned far north of its position and this in turn forces Constantinople northwards and thus the whole of the Black Sea is wrongly set out. South of this area both Crete and Rhodes are correctly positioned, and as the 37N latitude is correct, then both Akri Taernus and Akri Malea are basically correct at the southern side of the Peloponnese peninsula.


There is little discrepancy with the Eastern Mediterranean Sea as drawn. The basic graticule is correct with 90/72 miliaria setting out. Alexandria is correctly positioned at 30E/31N and the islands of Rhodes and Cyprus are quite correct on the graticule. The diagram has appended the scale bar used for these large scale diagrams to enable a check to be carried out by researchers interested in the finer points of the Portolan Chart.



The final large scale plot is of the Black Sea or Pontus Euxine. It could be represented by the longitudinal setting out based on the whole Portolan Chart but in fact as discussed in text ChBLS/1, the Black Sea has been stretched longitudinally such that a distance measure of c75.5 miliaria should be applied to the longitudes. The western limit of the Black Sea is correctly positioned on the 28E line and Constantinople is also acceptable for 29E/41N. The southern limit of the Black Sea has both sections following the 41N and 42N latitudes correctly and thence the overall north/south graticule is 90 miliaria and positions the Sea of Azov correctly. Obviously, it is not practical to evaluate the western seabord as the Danube River mouths explode into the Black Sea, rather colourfully. However it is an excellent visual rendering of the sea as the previous texts will illustrate.



The scale of the individual pages of the 1313 atlas precludes highly accurate measurement on any putative graticule that we may wish to append. However it can indicate intent and possible errors, such as the western coastline of Iberia from the 37N line at Cape St Vincent, then to the 43N line of Cape Finisterre where the distance measure by scale bar units appears to be c72 units. However, if we construct the longitudinal lines from Cape St Vincent, 9W, to Majorca, 3E, which aligns correctly to the north east Spanish coastline, then if this was a miliaria plot the latitudes should be 90 and the longitudes 72 miliaria.
They are not, and in fact it is quite simple to opine that Petrus Vesconte has used a Roman Mile numeration for latitude and longitude. That is 75/60 for the Iberian plot and then suddenly it is changed it to 90/72, approximately at the 2E longitude and thence to the Italian Mainland which features partially on this atlas page. This is confirmed by the fact that the southern coast of Italy has been drawn as 250 miliaria its geographical length. The distances from Cape Santa Maria di Leuca to both Ancona and Trieste are acceptable when miliaria measurements are used.


This change of measurement will of course cause a large distortion when the latitudinal measure suddenly grows from 75 to 90 units and thus dependent upon the base line chosen, the only normal line being the 36N latitude, it can only mean that the northern coastline, that is France to Genoa moves northwards and any further part of the chart will distort accordingly. This will drag the N African coastline northwards and induce the twisted format into the chart, particularly as the southern French coast distance to Genoa is suspect.


This atlas page should exhibit the typical distortions found in the measurements from Genoa to N Africa; but it does not!!
The whole setting out of this area of the Mediterranean Sea can be shown to be highly accurate and exhibits no distortion within its construct. If true then this section of a chart should not be able to influence the skewing by raising the N African coastline further!


But this is only correct if it is accepted that the scale bar as drawn has a 10% error which can be simply proven by assessing the charts distances. If a 10% addition is applied, as the diagram illustrates, the distance measures are excellent, and the chart is accurate. In fact this is not the first chart to have suspect scale bars, some with two or three bars appended show a variation to each which can have a disproportionate effect on the distances.
However, it does confirm that the correct distance measures were known for this section of the chart and could thus have been drawn avoiding the distortions emanating from the mis-measurement of this are. (See the next chart for confirmation).


On plotting the putative graticule it was quite apparent there were two problems on this page of the atlas. Firstly, the standard latitudinal expansion of the Aegean Sea was a fact, increasing its length latitudinally by one degree. Secondly when the distance measures were checked against the scale bar it indicated a graticule of 100/80 miliaria when it should be 90/72 miliaria. Thus the scale bar is a 10:9 increase in size for this chart.

QUESTION; have the scale bars for diagrams D09 and D10 been drawn on the wrong pages as a swap of 10% would counteract the problems noted?



Yet again it would appear that the scale bar has been mis-drawn and thus the graticule is overlarge with latitudes 8:7 increased. Thus it is approaching a square chart as drawn. But with such a small scale bar for a very large sea area it is possibly a simple error by Petrus Vesconte in drawing the scale bar on the chart.



A full evaluation of the Black Sea coastline as drawn on Portolan Charts is included in text ChCOR/1. This atlas chart follows the typical enlargement of the Black sea from 910 miliaria to 1040 miliaria However it is to all intents and purposes a visually acceptable representation. The increase in length for this atlas page, probably a library edition, would not have been problematic with its 10% expansion, but if for mariners could be problematic..



The scale bars on the charts are meant to be a regular gradation such that a 200 miliaria length on one chart becomes 150 miliaria and then 100 miliaria on subsequent charts. Thus they are a 1: 1 ½: 2 increase in scale and the distance measures cannot be transferred chart to chart. The fact that Petrus Vesconte does not appear to inform the users of the atlas of this fact, particularly as the wind rose dimensions are obviously the same on each page of the atlas, led me to believe they are library editions, easier to study than a full size Portolan Chart, but not as helpful to a mariner.
Then, as Petrus Vesconte drew a small scale atlas chart of surprising accuracy, negating the distortions in the main central area of Genoa to N Africa which if correctly applied the distortions to the Levant would be abated, led me to believe that Portolan Charts were copied with their distortions from the master copy or template. But of course, given the varying scales of these atlas pages, which could not be set alongside each other, it begs the question if the one chart non distorted area was actually noted and could be have been acted upon and thus it could have provided an unparalleled accuracy. I can only conclude it was not recognised and these are just copies of a master chart, but of what date?


The 1318 atlas comprises 9 charts of varying scales, but obviously as the sheets are the same size, yet again the wind rose diagrams are also equal but will vary in size if measured with the atlas page scale bar. The scale bars appear to be in a logical reduction such that if the largest is denominated 1, then we have 1.5 and 1.75 reduction for the change from 200 to 400 miliaria. The 9 charts are individually examined and their idiosyncrasies are exposed. They are presented as an overlay on the actual charts and maintain the libraries millimetre scales for further research.





Because of the rather schematic drawing of Great Britain, it is not possible to be as precise as I would wish, but, suffice it to say an estimate is quite possible which indicates varying latitudinal distance measures and consistent longitudinal measures. However, as was shown on the similar chart from the 1313 atlas, we have latitudes for France of 75 units but wildly varying longitudes. The 75 units have already been noted as probably 75 Roman Miles, the equivalent of 90 miliaria which in fact may well be the actual measurement for two degrees of Great Britain.



When a putative graticule is drawn for this chart the intent becomes very clear in that the latitudes are 75 Roman Miles and the longitudes 60 Roman Miles. Thus the scale bar for this atlas chart is perhaps still Miliaria as would be expected but could have changed to Roman miles. It should also be noted that the graticule comprises near rectangles for the Iberian section.
Thus the beginnings of a major distortion are to be seen as when the 75 Roman Miles are changed to 90 miliaria using the same scale bar it represents a 20% enlargement and thus the north/south distance of the Mediterranean Sea expands to misplace Genoa etc.



Immediately with the change of scale the graticule changes to c90 x c72 miliaria and thus this scale bar should be considered as representing miliaria not Roman Miles. Between Cape St Vincent and Majorca, 9W to 3E, there is a consistent longitudinal measurement and from 35N to 42N a consistent and correct latitudinal measurement.



It is here that we meet the normal distortion on a Portolan Chart, that of the latitudinal measures from 37N to 39N, south of Sardinia, where they total 130 miliaria instead of the correct 180 miliaria. The overall graticule by scale bar is 90 x 64 miliaria and not the correct 90 x 72 miliaria. If the portion of the chart covered on both D18 and D19 around the 3E longitude is studied the position of this distortion becomes very clear. It should also be noted that on D18 between 42/43N it is 50 miliaria, whereas on D19 it is a correct 90 miliaria.



I have always doubted the wisdom of correcting charts as the originals can have an accuracy, albeit distorted, and all that you finally achieve is a much more accurate but distorted Portolan Chart. However in this instance and as it is the general distortion point of Portolan Charts that is along the 9E longitude from Genoa to N Africa, it is sensible, just once to illustrate diagrammatically the problems of distance measure errors.
But I must plainly state yet again, these distortions are not to do with magnetism.
The diagram is an overlay to Petrus Vesconte’s chart and I have accentuated the original coastline in black ink and then set out the corresponding geographical positions in red ink.


On diagram D19 the graticule was set out via the scale bar and is shown to be 90 x 64 miliaria. But when the latitudes are corrected from Majorca, 3E, to Sardinia, 9E, the 64 miliaria longitude measurement is maintained. However when Italy and Sicily are corrected the longitudes become 72 miliaria, the correct distances apart. The eastern section from 3E to 9E should read as 6 x 72 miliaria as Montpellier, France is 43° 37N and 3° 52E and thus accords with the 3E and 4E longitudes through Majorca and Menorca. Thus this section of coastline from southern France to Montpellier and Genoa has been foreshortened.
These two facts, the distance along the 9E line and the distance from 3E to 9E, both being short measures distorts the chart in such a manner as to convince previous researchers who have not actually examined the charts internal measurements that the reason is a magnetic deviation applied to the chart. It is not so!


The wildly varying graticule, both latitudinally and longitudinally indicates the induced distortion evident from the previous chart. But, as has been shown on Diagram ChPV/1/D04, the distance from Venice to Cape St M Leuca is quite acceptable and hence the skewing of Italy has caused the severe graticular format.



Being mostly open Mediterranean Sea, I have set out the 36N and 16E lines to establish a graticule which appears to be a standard 90/72 miliaria grid. However as the chart scale is small enough to induce an inaccuracy of measurement I will not proceed further.



This chart introduces the distortion evident by following charts of most cartographers with an over measurement from the 36N line to the mouth of the Dardanelles at 40N latitude. This is typically one degree latitude expansion. The correct measures are however used for the 35N to 36N latitudes with a graticule of 90/72 miliaria. This is a major distortion.



There are two scale bars on the chart with slightly differing lengths as I have indicated. However, the graticule can be set out as 90/72 miliaria for the chart. It would be churlish not to accept the intent as correct.



The length of the Black Sea is drawn as 1040 miliaria, which is grossly over-length. Using the scale bar, which is again a very small unit length; the graticule could therefore be taken as 100 x 80 miliaria. This is of course proportionally correct and thus as has been shown previously it illustrates the scale bar is probably 10% in error. Actual length 910m.



Firstly, it is important to recognize the small size of this atlas and thus the simplest of minor errors becomes a major factor. Thus I have not tried to follow the graticular alignments other than in a general manner to illustrate the overall intent and resultant charts.
As I have already opined, these are probably pocket book library editions of little use to mariners, but for the educated a boon to understand the form of the Atlantic coast and Mediterranean Sea basin. Bearing in mind that nobody had seen the actuality of the form shown on these charts as is possible in the 21st century and thus it would not matter to those educated persons that the presentation to our eyes is askew through distortions. They would not be aware of this fact, as it is hard to recognize if you have never seen the correct image set down. I cannot believe that the Francesco Beccari chart where-on he accepted the complaint of mariners concerning the distortion between Sardinia and N Africa was the first such comment. I am also perturbed that previous researchers appear to have ignored that comment and not back traced the Portolan Charts to study the distortion affects that are so very apparent.
Unfortunately Petrus Vesconte, probably through no fault of his own, has failed to bequeath us an overall Portolan Chart using the correct distance measures which he surely must have known. Although given the fact that it appears he was a copyist, then I must question his original capacity for cartography and suggest he was perhaps an individual who later in his draughtsmanship career became a mathematical cartographer.The only other possibility is that it is a copy of another cartographers chart and he did not recognize the veracity of that section. I temper that comment with the fact that the eastern section of the Riccardiana MS 3827 chart appears to be either a copy of or the forerunner of the Petrus Vesconte 1311 chart.
Finally, the 1318 Atlas is a minor masterpiece, suffering from the vagaries of measurement, the uncertainty of Roman or Miliaria measures there-on, but nevertheless a master piece of cartography and visual presentation from the medieval age.


This chart is signed Perrino Vesconte, and is from the 1321 Atlas
Covering the British Isles, France and Iberia it is drawn at a small scale and only by reviewing the previous charts can an educated assessment be made of the graticule.
Basically it appears to be a 75 x 60 graticule which indicates as per the previous charts the usage of the Roman Mile and not the Miliaria. Again, Iberia is shortened North/South, Cape St Vincent to Cape Finisterre being some 50 RM short, whilst France is 35RM short latitudinally. But it must be stressed that the scale bar is quite small for complete accuracy and of course there is the fact that the scale bar could be drawn erroneously.
The island of Hibernia has been afforded an increase in size, probably due to the measurements which were taken as named equivalents but actually wildly differing in length. It must also be acknowledged that as Rome did not survey this island it is an estimate from mariners of this age when the charts were being prepared. Even Al Idrisi has no information.
Thus on the 1313 atlas the British isles are still no more than overall ideas apropos shape and size.



This chart is similar to D26, but of far greater accuracy. The basic graticule is probably 75 x 60 Roman Miles and expands to 90 x 72 miliaria in the Mediterranean Sea area. The distortion caused by the change of measurement is very apparent as the longitudes traverse Iberia and thus what was a simple 75 x 60 graticule expands to a 90 x 72 graticule and thus the 43N latitude is swept northwards. This is in fact a 6:5 expansion through the wrong interpretation of the measurements as has been shown on the previous charts.



I have previously opined that the origin of Portolan Charts is from basic Roman maps of the Mediterranean Sea basin. The originals were then probably developed by Monks in the north west of Italy where they had access to the Roman texts of the Agrimensores and Surveyors and many other such technical treatises. Thus the basic map was founded on Roman measurement, the Roman Mile of 1.4791KM, of 600 stadia.
That morphed into the Miliaria when confusion apropos the 500 and 600 stadia latitudinal degree was in the minds of the draughtspersons, persons who were not aware of the nomenclatures erroneous nature; a mile was not always the same mile. Hence a latitudinal degree of 500 stadia when equalled to 75 Roman Miles changed to 600 stadia and it became 90 units, as both were latitudinal degrees and must be equal. Then realisation in the distant past that the two were not equal, the 500 stadia and 600 stadia represented the same distance but were thus not the same length, the 75 Roman Miles was still 90 miliaria but they were reduced to 6.667 stadia each. Thus in our terms, a Roman Mile is 1.4791KM and a Miliaria is 1.233KM. Would that in the past the mile was named differently country by country.
But if a later cartographer was unaware of this equalisation and was presented with a map having a scale bar appended, and that map was a copy of a Roman original for Iberia, then the scale bar would be Roman Miles and from Cape St Vincent to Cape Finisterre would be 6 x 75 = 450 units. Then if you were aware that from France due south through the Balearics to N Africa was “XXX” miliaria and plotted that on a copy of the original to the east of the Iberian Peninsula, using the map scale bar as miliaria, immediately the distortion problem would arise as in fact you could be increasing the distance by 75:90 or 20%.
Study the early charts as per this text and the scenario they present is one of a chart being drawn to a miliaria scale but using Roman mile distances as miliaria. Thus there is no doubt there had to be a second origination for the Portolan Charts, one that understood the Roman Mile/Miliaria conflict and redrew the chart to a single scale. That could be simply achieved by drawing the original scale bar and subdividing it into 90 parts not 75 parts. But as most scale bars are made up of 100 units (2 x 50) there was thus the opportunity for the 10% error to percolate into the system as is shown by this text.
But there remains the visual effect of the distortions.


If a Portolan chart or Atlas was drawn using the Iberian scale of 75RM which was in fact 90 Miliaria then this 90 was used it would produce the distortion of the chart which previous researchers have concluded must be a magnetic deviation. See the appendix for facts.
But not one single person in the Medieval World, or Roman World for that matter actually knew the geographical form of the Mediterranean Sea basin. Yes, they had a good idea of its general form from these maps or charts, but could not have determined the actuality. Thus when it became necessary to redraw the map or chart to a single miliaria scale the distorted shape originally produced just continued as nobody actually knew any better.
Petrus Vesconte has proven that the distortion along the 9E longitude was totally unnecessary as it was possible to draw a distance measure accurate chart and negate the Francesco Beccari problem.
Whence came the corrected miliaria map to be used as a template for future Portolan Charts is still to my mind answered by the opinion that various monastic establishments produced their own maps for sale to, firstly educated persons and then when the City States became seafaring nations in the 9th century a new clientele existed, the mariner and his factor to facilitate trade in the Mediterranean Sea basin. Therefore we have the possibility of several original map formats which gradually coalesced into a single format which was found to be the most “accurate” and usable by mariners, but probably not at sea as I have previously shown in text ChUg/1. I thus consider chart development not proven by timescale.


Return to the 1311 chart of the eastern Mediterranean Sea and study the diagrams which show that the scale bar produces a miliaria graticule of 90/72 units. It also shows the two distortion points, one on the 9E longitude and the second in the Aegean Sea, 36 to 40N.
What would the western section have shown?
Given that the next Petrus Vesconte chart available to us is the 1313 atlas and a single sheet there from covers this missing area (ChPV/1/D08) and it is obviously a near perfect fit to provide the missing chart. The graticule changes as it crosses the Iberian Peninsula from Roman Mile to Miliaria and the 4E longitude through Menorca actually changes direction to accord with the angular alignment of the 9E longitude of the 1311 chart.
Study the Riccardiana MS 3827 chart and the similarities to the 1313 atlas are very apparent, even to the change of direction of the 4E longitude through Menorca. Follow through on the 1318 Atlas chart of Iberia, it has the Roman Mile units of 75 x 60 and thus I opine that Petrus Vesconte is copying an original Roman based chart for his basic templates.


The basic map was a Roman original using the Roman Mile as its scale which was carried through from the original survey possibly by Marcus Vipsanius Agrippa as told by Pliny the Elder. Confusion reigned over the number of stadia in a degree and it resulted in the miliaria being born from the Roman Mile. It took centuries for the erroneous pairing to be altered and as the geography of the Mediterranean Sea was unknown the ensuing distortions caused by the measurements continued to be seen in Portolan charts today.

M J Ferrar August 2017.


On Diagram ChPV/1/D08, I indicated the putative graticule for the Iberian peninsula and noted that it was drawn as a 75/60 miliaria grid and that it is obviously a complete mistake in that they are the Roman Mile measurements for latitude and longitude at 36N but taken as Miliaria. Hence by drawing the peninsula as 75/60 units it was reduced in scale by 90/72 or 5/6ths.


Thus on diagram ChPV/1/D28, to visually explain the distortions of Portolan Charts and the reason behind the all important 9E longitude being slewed anti-clockwise, I have included an overlay of the original 1313 Atlas chart (part of) and then drawn the revised positions of the peninsula’s geographical points using the 90/72 graticule. Thus there are two diagrams upon the page set vertically above each other.
The use of the 90/72 graticule was obviously the intention of Petrus Vesconte as the distance from the 5W line to Skikda at 7E, i.e. 12 degrees measures 864 miliaria. There are then the latitudes through the Balearic Islands drawn at 90 miliaria compared to the west coast of Iberia drawn at 75 miliaria.
This is the crux of the information concerning the major distortion of charts.
If Iberia is drawn 5/6ths scale then the proximity of N Africa, the distances being well known and correct on the chart, will be altered. Thus if the general alignment of 37N from Cape St Vincent to Carbonaras is foreshortened, that is from 9W to 2W, instead of being 7 x 72 miliaria it is in fact 7×60 miliaria, 504 to 420 miliaria, then as the diagram illustrates the coast of N Africa must move northwards to maintain the correct distance apart. Therefore, the greater the error within the longitudes of the Iberian Peninsula then greater is the distortion in the position of the N African coastline and the subsequent slewing of the latitudes and longitudes anti-clockwise to the east.
Thus it is quite a simple explanation of the 9E distortion angle south from Genoa. But even when the Iberian Peninsula is drawn correctly at 90/72 miliaria the visual effect of the original error as the 1311 Portolan Chart indicates, was continued.
That begs the question; was the 1313 Atlas in fact drawn before the 1311 chart? It would have been very simple to start an atlas, draw a chart and then complete an atlas only to date it later.
Commencing at 9W the west coast of Iberia is north/south and the slewing gradually increases as you progress east. As I have already opined, not a single cartographer before the age of technical surveying, latitude plotting, etc., could have known that what was being drawn was incorrect. Sail the Mediterranean Sea and without doubt a mariner who had seen a Portolan Chart would be convinced of its veracity. However, try to sail long distance using a Portolan Chart, and as my text ChUg/1 indicates, disaster will ensue.
It must be borne in mind that a general distortion will always occur when a set of Latitude Longitude distance measures are used which are basically correct at 36N, the centre line of the Mediterranean Sea. Consider this; at 36N the ratio is 90/72; at Genoa, 44.4N it is 90/64.3 and at Alexandria 31N it is 90/77.145 miliaria. Thus a chart of the Mediterranean Sea will automatically have a 10% error to the north of 36N and a 7% error south of 36N. Given the length of the Mediterranean Sea basin as shown on Portolan Charts, from 9W to 36E, which totals 45 x 72 = 3240 miliaria; then at 44.4N it is 2893 miliaria and at 31N it is 3471 miliaria. Those differences are the equivalent of 5.4 degrees at 44.4 N and 3 degrees at 31N. Hence if correct measurements on coastlines are utilized and those are plotted against a rectangular graticule of 90/72 miliaria the internal layout will require to be adjusted, just to make the chart visually acceptable as a coherent unit. Mis-use the data given, that is Roman miles as Miliaria and the recipe for distortion is written large on the Portolan Chart.
If diagram ChPV/1/D19 is again studied, there-on are longitudes which are 64 miliaria, corresponding to Genoa (64.3) but using 90 miliaria per latitude from 36N to 45N. The corrected diagram ChPV/1/D20 can still only be read as 64 west of the 9E line and 72 miliaria east of the line. These are inbuilt distortions which will grossly affect the visual outcome of a Portolan Chart. Compound the problem with the extra latitudinal degree inserted in the Aegean Sea and Constantinople moves further north. The Black sea is thus mis-positioned and given it is normally drawn over-length 910/1040 miliaria, there are multitudinous distortions on these visually acceptable charts.
They are a feast for the eyes, no doubt about that, but as they are all copies of one or two basic charts, which were constructed from erroneously used data, then the fact that it is possible to evaluate most of the early charts with similar distortions only serves to confirm that fact. The last point to be made is of course that these distortions have absolutely nothing to do with any magnetic deviation and the use of a basic magnetic compass. That I am afraid to say is the biggest failure in the previous research carried out into Portolan Charts.


This research has led me to believe that the 1313 Atlas may have been so dated but is in fact a copy of a much earlier atlas or portolan chart. The 1311 chart changes the scenario because it exemplifies the known traditional Portolan chart. Why so? If the charts of Petrus Vesconte are mere copies of previous maps or charts, then why does his enigmatic wind rose diagram of 1318 contain such geometrical data, and is thus the arbiter of the wind rose diagram and thus it could be argued Petrus Vesconte perhaps knew the details of the layout of a Portolan Chart. But did he? Was he purely a copyist, did the fact that he drew a 1313 chart with very strange Roman and miliaria units intertwined mean he was not cognisant of the actuality, merely that of a copyist. How does one draw a chart with Iberia at two totally different scales unless you are copying a chart you have no actual knowledge of or its construction? The two scales are actually incompatible and should have raised alarm bells.
I actually think that we do not have enough very early charts to draw any conclusions from for their origination and provenance and thus consider what a Portolan Chart actually is or is not. It appears to me that the Carta Pisane and Cortona charts are maps converted to look like Portolan Charts. It also appears to me that Rome plays a large part in the origination. If as it appears many researchers consider that our sample is precise enough to deliver a definitive statement regarding these charts, then I think it is the time for them to place on the record, via texts on web pages their arguments apropos the origination, the magnets and their distortions identified here-in, and thus prove this text to be unacceptable. Only then will students of Cartography have the confidence to know that the corpus of information so far presented by all parties is acceptable. If not then we should be able to judge the texts which can be assigned to the rubbish bin. If you are certain in the veracity of your research then ensure it can be widely read and understood.
The early works of Petrus Vesconte are incompatible with geographical fact due to the scalar changes, the distortions and the mis-measurement, but the overall presentation is somewhat better than the knowledge, a feast for the eyes.

M J Ferrar August 2017.