The starred sky in science and art: the early fifteenth-century representations of the Ptolemaic stars catalogue

One of the most interesting lectures in the recent Apotelesma congress was the one given by Felice Stoppa, the author of the famous site about celestial maps and atlases Atlas Coelestis, rewarded in November 2005 by the Griffith Observatory Star Award.   This is the English translation, which I believe could be of interest for many. I should thank Felice Stoppa (the author) and Lucia Bellizia (Apotelesma president) who shared it with us.

The purpose of this report is to identify the first documents and the first artworks anticipating or influencing the celestial cartography of the sixteenth and the seventeenth century and to find a common thread that somehow  connect them each other.

Therefore I will examine a small number of metal celestial globes, all of Arabian production, and the earliest printed maps representing in a so-called scientific way the starry sky as described by Ptolemy in his Almagest. We will see that around these documents have been unfolded real stories, sometimes in a fortuitous way, which lead to anticipate as a possible conclusion to this research, that the history of celestial cartography cannot be considered complete  because it is still possible, even in our days, as already happened, to find in libraries and antique markets, new documents or artifacts that may call into question our current certainties.


Was towards the end of 1876 that Ferdinando Meucci bought a small globe of brass he had just found in the market at Florence (Images 01 and 02).

figura 1
1- The Arabian celestial globe
image 2
2 - Detail of the flat map as reconstructed by Meucci

He decided to buy it even before a previous examination in order to establish its authenticity.

The globe was, in fact, completely oxidized and the lines of the traditional images of the constellations and some words in Arabic Kufic characters could be barely read. In the Regio Museo Fisico of Florence which Meucci had been running from 1844,  the collection of ancient astronomical instruments was including invaluable Arabic astrolabes too, but the lack of a globe of the same period impelled to hazard even a misguided purchase.

After perfunctorily cleaned from the patina it was covered,  Meucci proceeded to determine a first approximation of the year of construction of the globe. In order to do this, he verified for which age the stars were positioned  using the method of the difference in longitude caused by the precession of the equinoxes: “It was great my surprise, when, because of the longitude of the stars, I could recognize that it should have been built at the end of the eleventh century of Christ. Was indeed well distinguishable the  position of the bright star Regulus (the Lion heart)  at 16 ° 40′ Leo, which is to say, advanced of 14 ° 10′ from the  position of 2 degrees and 30′ reported by Ptolemy, as it seems , 140 years after Christ; so that, by reasonably calculating according Albategnius, a time of 66 years that fixed stars take to proceed of one degree, resulted that this globe was built in 1075, or a few years after …. ” The dating is confirmed by the Arabic translation  of the inscription in Kufic characters engraved around the Antarctic Circle, which also reveals the place of production, the authors and the purchaser who commissioned the globe: “Made this globe supplied with its stand,  for the charged with double Vizierate, Supreme Qayid, Abù Isà Ibn Labbun, his servant Ibrahim I bn Said as Shali, the weights inspector  in Valencia with his son Muhammad, and he placed the fixed stars in that, according their size and diameter. This was done in the beginning of Safar of year 473 to the Hijra (of the Prophet), God bless him and grant him the perfect peace.” (Image 03).

figura 3
3- Reconstruction of the inscription of the Arabian globe.

The beginning of Safar of the year 473 from the  corresponds with the end of July 1080 of our age. The inscription was translated, as well as the others on the globe, by Professor F. Lasinio, orientalist at the Institute of Advanced Studies, Department of Philosophy and Philology, of Florence.

The Arabian Celestial Globe in  Florence is therefore the oldest known representation of the Ptolemaic heaven.
The Globe, 209 mm in diameter, is composed of two hollow brass hemispheres welded together, and reproduces 1015 stars (both figured and formless) and divided into six sizes, each represented by circles of different diameter. On it are engraved the circles of ecliptic and equator divided by 1 degree marks, each five numbered in Kufic numeral letters. The zero of the scale of the equator coincides with the point of intersection of the equator with the ecliptic which is the beginning of Aries. Twelve great circles of declination divide it in portions of 30 degrees, namely in its 12 signs of the zodiac, the ecliptic.

There are 47 constellations represented, the Crater is missing, and, curiously, the area containing the six formless stars of Aquila, which is the one coinciding with the non-Ptolemaic constellation of Antinous, is highlighted by a engraved heart-shaped line.  The characters represented in the constellations are drawn from the front, as used in flat maps representing the sky as seen from Earth, but because the global surface used is convex, we find that the stars are symmetrically arranged on characters’ anatomical parts  in respect to the way they are traditionally represented in the constellations.  It is a minor annoyance because often the meaning of star names has to do with their anatomical location.

Meucci condensed the results of his study of the globe in a booklet edited by Tipografia dei Successori Le Monnier, consisting in a short but dense 13 pages report, followed by three double pages collecting in a catalogue the constellations and the stars engraved on the globe.
The catalogue also includes a column with Arabic names of the constellations, followed by their phonetic transcription and translation into Italian and by a  section relative to the stars showing the number and position by longitude and latitude, and comparing the relevant data reported on the globe with those of Ptolemy catalogue.

Meucci employs part of the booklet introduction in order to draw a story of the cartography from the beginning to his time and reminds the following definitions, useful to understand the constellations catalogue he will offer in the following pages: ” Some stars were called formless (external) by the ancients because they were around the figures representing the constellations, while those contained in them were said figured (internal) “.

The book is supplemented by the two tables we are describing: a flat paper, of size 91×45 cms folded several times faithfully representing the data of the globe, which is represented divided into twelve gores of 32.5 cm in diameter, tangent at the ecliptic and  a second card representing the Arctic Circle of the the globe, void of stars, but bearing the inscription in Arabic Kufic characters.

The production date of the globe is between 1075 and 1080, an interesting period because particularly close to 1054, the year of the huge explosion of a supernova, the one which created the famous Crab Nebula,  currently visible through a telescope in the Southern horn of Taurus.
It is estimated that the explosion generated a new star  that should remain visible even in daylight for several weeks. In addition to its brightness, this supernova is famous in the history of science for the fact that it is not present in any of the chronicles of the time, absence that philosophers of science highlight to prove their thesis that scientific  and philosophical paradigms of a given age often exclude from consideration those phenomena that cannot be explained by their theories: the system of Aristotelian-Christian world, in force throughout the Middle Ages relegated the appearance and the occurrence of physical phenomena in the sublunary sphere, excluding them  therefore  from the crystalline and unchanging spheres  occupied by  planets and fixed stars. For this reason  it was unconceivable that the temporary appearance of stars beyond the sphere occupied by the Moon could be mentioned and so the supernova of 1054, even if was seen, was soon forgotten.

Meucci’s globe could be a good example to verify the validity of this philosophical theory. I have therefore isolated the area of sky around the Southern horn of Taurus trying to identify all the stars of globe matching them to those numbered by Ptolemy in his catalogue. The purpose of this work was to isolate the possible presence of a spurious star near the position currently occupied by the Crab Nebula. The operation is very complex in practice because in this area of the globe are given fewer stars compared to those of Ptolemy catalogue and the combination is not mechanical. At the end of my examination only a star of the globe could be compatible with Ptolemy catalogue and quite close to the Crab Nebula position. So I asked advice to Prof. F. Richard Stephenson, historian of astronomy, an expert in historical studies on the Crab Nebula and professor at the University of Durham in the UK.  Stephenson starting from a modern map of this area of sky, calculated for each star with a magnitude lower than  5.5 the shift caused by the precession of the equinoxes and was  finally able to show that every star of the globe appearing in the Taurus head was already known to Ptolemy, thus excluding that the star I spotted could be the Crab supernova. Even if my research ended with a negative result, however, was useful to confirm that for Ibrahim  Ibn Said Sahli as Sahil, the weights inspector in Valencia, , the manufacturer of the globe too, was not worth remembering  an astronomical phenomenon which the vision prevailing in the world, the Aristotlean one, also valid for the Arabs, assumed impossible.

It will be not the same for another Stella Nova, the one discovered by Tycho Brahe in 1572 in Cassiopea constellation, for which the astronomer, not being able to observe proper motion and parallax,  conjectured a distance similar to fixed stars. The appearance of this star produced so much fuss that the star was reproduced for many decades in almost all the stellar maps, even when it was long gone, as evidenced by the drawing made by Bayer for its Uranometria of 1603. But in this case we are in a different historical era, between Copernicus and Galileo, and the framework of the Aristotelian system, also dear to the Catholic Church, spurred by new discoveries was demonstrating the deep cracks.

The Globe of Galileo Museum in Florence is the oldest but not the only instrument of Arabian origin, representing Ptolemy’s sky. In the not so far 1958 a similar situation to Meucci experience was revived in Paris, where, in an auction, made its appearance another small  copper globe, 175 mm in diameter, dating  the year of 539 Hijra, ie 1144 of the Christian era, work of Yunus ibn al-Husain al-Asturlabi. It was described it in his detailed report Marcel Destombes, collector and historian of astronomy, whose work also informs us of the existence of another manufact preserved in the Département des Cartes et Planes de la Bibliothèque Nationale de Paris, a  anonymous unsigned globe, but attributed by Destombes to the same author of the globe in Florence of which it would be  a few years older.

According Destombes the positions of the stars of the two  globes made by Ibrahim ibn Said as Sahli were taken from the Toledan Tables, calculated for 1067 by the Arabian astronomer Arzachel  (Abu Isaac ibn Ibrahim an-Yaya Naqqach).

At least another dozen of Arab metal globes, these ones all from the East, separate the artifacts we have seen  above from early renditions on map of Ptolemaic spherical sky. I call here “Ptolemaic” all those celestial representations showing the stars according to their position in latitude and longitude, with reference to the tables compiled by Ptolemy, or derived from them. Therefore I’m excluding from my analysis, though beautiful, everything coming from an artistic interpretations as to the various translations of medieval Aratea, where the stars are located in the constellations only with an ornamental function and in dependence on the poetic text which they illustrate.

We have to wait until 1515 for works on paper comparable to those of Arab globes. The author is the  German artist from Nuremberg, Albrecht Durer, painter, xylographer, but mathematician too who, with similar characteristics to the celestial globes, shows us two of more beautiful representations of the Northern and  Southern sky (Images 04 and 05)

figura 4
4- Durer, Celestial Map of the Northern Sky, 1515
Figura 5
5- Durer, Celestial Map of the Southern Sky, 1515

The two maps are woodcuts, that is they are printed from woodcuts, which allow to reproduce on paper countless copies of the document. It is therefore in their elevated circulation that these documents have an advantage compared to previous globes.

The two tables of Durer follow Ptolemy tradition  as  content and accuracy. In the corners of the Northern hemisphere are depicted Aratus, Manilius, Al-Sufi and Ptolemy who inspired the author.

The Northern sky map is, between the two, the richest and the only showing  the twelve zodiacal constellations, identified both by name and the astrological symbol. The author follows Vitruvius and Aratus tradition in naming some constellations, such as the Cygnus, which is still called Avis.

Durer respects tradition, typical of the production of three dimensional globes, to track the constellations of the sky as if they were seen from outside (convex projection), while stars are identified by a number following  Ptolemy’s Almagest .

The two plates, square, each side of 35 cm, reproduce in stereographic convex polar projection the two starry vaults represented by circles  sided 35 cm, having as centre the projection of their ecliptic poles and as circumference the projection of the ecliptic.  This one is divided into segments of thirty degrees, which show the zodiacal constellations. The longitude of about 1020 stars  located here can be read with a degree of accuracy using the graduated scale located in the outer circumference. In both hemispheres the borders of the Milky Way are shown by dotted lines.

Of stars coordinates and their positioning  were in charge two scientists of the time, Stabius and Heinfogel, whose names are placed with Durer’s one, in the cartouche in the lower left of the original version of the Southern plate, dated 1515: Ioann Stabius ordinavit, Conradus Heinfogel stellas posuit Albertus Durer imaginibus circumscripsit.

Johannes Stabius (approx.1460-1522), an Austrian mathematician, was in charge to transfer on the flat map,  therefore two dimensional, the stars drawn on the globes. He chose the stereographic projection, non-linear in latitude, but which allows  to maintain the relative position between stars and  do not improbably stretch the constellations placed on the edge of the ecliptic.

Conrad Heinfogel, a German astronomer of the time, dealt with the calculation of the positions of the stars. It is actually a calculation and not a direct observations of the starry sky: the reduction proceeded from a previous catalogue data on which were made the corrections induced by the precession of the equinoxes, namely a correction for the assumed time rather than the one determined today. Assuming that Heinfogel  worked with Alphonsine Tables, whose compilation was organized by King Alphonse  X of Castile and was made in Toledo around 1252 by about fifty astronomers, and assuming that he applied the correction for the equinox precession known at his time, we get that the stars of Durer plates were placed for a year around 1500, a period surely compatible with the year of actual production of the maps. However, if we would start from Ptolemy’s Almagest tables using the correction for the precession known today, we wrongly evaluate the effective date, obtaining in fact, an year very close to 1450.

We have said that Durer maps are the first to appear on paper and printed for mass production, but they are not the first samples  because they are based on two manuscripts, drawn on parchment and dated 1503, published in 1943 by a private collector  of Hamburg. About the two square  – each side is 67 cm- maps (Pictures 06 and 07)

image 7
6- Heinfogel, Northern Celestial planisphere, 1503

figuta 8
7 - Heinfogel, Southern Celestial planisphere, 1503

now  at the Germanisches Nationalmuseum in Nuremberg with the bookshelf Inv.-Nr. 5576 Hz and Inv.-Nr. 5577 Hz, talks Zofia Ameisenova in a book published in Poland in 1959, where the images of two planispheres are published for the first time. They are further described by Adèle Lorraine Wörz in her thesis of Doctor of Philosophy obtained in 2006 in the Oregon University. To the best of my knowledge the two plates have never been reproduced in other books. The double manuscript, work by the same hand and representing the two celestial hemispheres, is certainly a forerunner of Durer’s work, has the same setting of the reference grid, uses the same projection, compared to the tables of 1515, however, is enriched by the presence of projections of the two polar circles, while is absent the reproduction of the Milky Way. In addition,  the rectangular cartouche in the upper right corner of the Southern hemisphere has a reference to Heinfogel for the position of the stars. Also applies in this case the consideration of the equinoxes precession, which we already did for 1515 tables because the position of the stars is almost identical. The stylistic interpretation of the constellations, very schematic and rigid, in some cases quite naturalistic, for example  the Crater in the Southern hemisphere represented as a wooden tub, does not give the impression that the author could be the same Durer. However the authorship is attributed to him by some art historians, who see some similarities between the allegorical images surrounding the maps with the artist’s later works.

In my opinion it is more likely the work of technical reduction made by the mathematician and the astronomer produced first of all the 1503 manuscripts and the Durer used them to produce in fair copy his xylographies of 1515.

In addition to the two tables of Nuremberg we know two other manuscripts, much older, which may have inspired Durer work,  (Pictures 08 and 09).

figura 8
8- Vienna manuscript, Northern celestial planisphere about 1430-1440
figura 9
9- Vienna manuscript, Southern celestial planisphere, about 1430-1440

They are currently stored at the Austrian National Library in Vienna, collected with other astronomical material to form the so-called “Vienna manuscript”, with the bookshelf MS 5415. The first plate represents the celestial Northern planisphere, and it is about 37x28cms., bookshelf MS 5415, fol. 168 r, and already appeared several times  in the literature concerning the celestial cartography. The second document, the one dedicated to the Southern sky, has hardly ever been published since it appeared only at low resolution in Zofia Ameisenowa limited edition book. John Brian Harley and David Woodward mention it in their monumental work The history of cartography written in 1992, where, however, incorrectly indicate the Vienna Library bookshelf, which made not easy  for me finding the map and partly could also justify its absence in historical literature, including the recent one. I have to thank the intervention of the librarian of Vienna library, Dr. Eva Farberger who was willing to help me in order to locate the document without knowing the bookshelf, found it and sent to me. This second manuscript, almost square, its sides are 31×28.5 cms., bookshelf MS 5415, fol. 170 r.

The origin of these manuscript maps, dating from around 1440, is presently unknown. About their possible author have been made different assumptions, the most reliable, in my opinion still to test, is the name of Johannes von Gmunden (about 1384-1442), Austrian mathematician and astronomer. This thesis is particularly supported by Paul Kunitzsch, a scholar of history of medieval astronomy, and is based on the affinities of these plates  with a star catalogue, this definitely attributed to Johannes von Gmunden, collected in the same manuscript MS 5415 from folio 217 to 251.

The Northern planisphere reproduces by convex projection Ptolemaic Northern and zodiacal constellations showing their Almagest star classification number, the stars are plotted according their ecliptic projection. The ecliptic is divided into segments of thirty degrees whose alternation is highlighted by the notches’ color,  first black and then red; strangely enough the computation of degrees inside the segments is measured  by an uncommon progression of six by six .

The plate was elaborated with great precision and accuracy  and  strongly departs in terms of scientific content from  constellations’ drawings accompanying in the same period Aratea and Hyginus fables, so that it can be considered the forerunner of all subsequent scientific papers . It anticipates of at least seventy years Durer plate of 1515 and compared to this, generally regarded as the prototype of celestial charts, it provides more information because it also shows the projection of the North pole, the Boreal circle and also, and this will be particularly important, a portion of the equatorial one.

The presence in the plate of the ecliptic projection and of its points of intersection with the ecliptic allows us to appreciate and calculate the value of the equinox precession according which the stars were placed, and then to determine with some precision the date production of the map. By identifying in the constellation of Virgo the star  Eta Virginis, the number 6 in the manuscript table, which repeats Ptolemy’s Almagest number, we can easily calculate which is far from the  autumnal equinox point of about -3 degrees of ecliptic longitude, therefore with a difference of about 8 degrees than as currently shown for example in the Millennium Star Atlas for the year 2000. The precession of the equinoxes gives a difference in ecliptic longitude of 50.33 “per year, then 8 ° is equivalent to about 570 years.  This is therefore the difference between the Millenium Star Atlas for 2000 and the manuscript, which can be therefore dated  1430, with a  difference  of only 10 years from the period indicated by  Kunitzsch.

Remains, however, at this point the problem  that the value of the precession of 50.33 ” per year  I used in my calculation is the one currently known and not the one used in the fifteenth century. So why do not we see the same discrepancy  we would get using the same value for Durer plates?

We have seen, unlike Durer and 1503 plates, that Vienna Northern sky map shows  a section of the equatorial circle through which I was able to operate the precise calculation of the precession. The presence of this section, together with the accuracy of the position of stars for the 1430-40 can lead to think that Gmunden did not calculate the positions for reduction from another ancient catalogue but somehow he used real observations, produced by himself or by another astronomer of his time. If this hypothesis is true, the Vienna manuscript would be almost like a picture of the sky of his time, so the stars would be included almost spontaneously in respect to the intersection of the ecliptic and the equator of that moment. It would be- after many centuries- the first handwork based on recent observations. That’s what Adèle Lorraine Wörz wrote in his thesis  conjecturing that the author of the Vienna manuscript used a recent catalogue of the period, the Khaqani zij , elaborated by the Islamic astronomer al-Kashi. He had been invited by Ulug Beg in Samarkand in 1420, where he played a key role in the realization of the famous astronomical observatory,  opened around 1429, and in equipping its observation instruments.

The work performed by Ulug Beg, al-Kashi and about sixty other scientists led to the publication of  Sultan tables (Zij-i Sultani), which appeared in 1437 but were improved by Ulug Beg just before his death in 1449. Even in the implementation of these plates were used the data of al-Kashi.Khaqani zij .

The possible link between Vienna manuscript and Samarkand astronomers’work is further confirmed by the use of Arabic names of many stars, some of which are used for the first time in the two Vienna planispheres.

The Vienna Manuscript, whether the hypothesis I mentioned is true or not, is without doubt more informative and more original than the 1503 and 1515 planispheres. This is demonstrated by the presence of the Milky Way, the  equatorial circle and the polar circles, the correlation of stellar positions with the grid reference, the numbering each six degrees, the name of the stars and the originality of representation of the constellations style. Elements that combined together make of it an archetype from which not even Durer could refrain from taking inspiration to achieve his  two most well-known maps.

Felice Stoppa

Translated by Margherita Fiorello, CIDA certified member, for heaven astrolabe blog @ year 2010. If you want to be notified the next time I write something, subscribe to my RSS feed.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s