Episodic Survey of the History of the Constellations

The illustrations on this page have been compiled from a variety of sources. They are reproduced in accord with 'fair use' provisions unless copyright is otherwise noted. If advised that copyright has been infringed I will immediately remove the particular illustration(s).

Return To Section Index Page

G: Greek Constellations

15: Farnese celestial globe

The Farnese Atlas. Once the only known celestial globe surviving from classical antiquity. The Farnese Atlas is an incomplete globe of the known sky. The term Atlas to describe any representation of the celestial sphere or terrestrial globe derives from the Farnese Atlas.

An early surviving depiction of the Classical Greek constellations that is substantially complete is the Farnese celestial globe (commonly called the Farnese Atlas (The Atlante Farnese)). (Images of the individual stars are not shown.) It is a large decorative, free-standing, celestial globe, sculptured in raised relief. It is the only large celestial globe surviving from antiquity. It has been restored in the mid 16th-century from a damaged condition. It is basically a statue of the figure of Atlas in a kneeling position holding a celestial globe on his back (i.e., between his shoulders). (It is the oldest surviving statue of Atlas in existence.) Atlas is bearded, has a cloak draped over his left shoulder, and is holding the globe with both hands. The position of the hands serve to obscure part of the globe. (The sculpture is 2.1 metres tall and the diameter of the globe is 65 centimetres.) According to astronomer/historian Ed. Krupp the globe also symbolises the cosmic axis around which the heavens turn.

According to the Museo Archeologico Nazionale in Naples, the figure of the Atlas Farnese was originally placed in the library of the Trajan Forum in Rome.

The Atlas Farnese can be viewed as marking the beginning of a new approach to Atlas iconography. Atlas was originally depicted as a young Titan standing upright and shouldering the firmament (not the celestial globe) at the Western of the four corners of the (flat) earth. With the Atlas Farnese the figure of Atlas is depicted as an old man on his knees lifting a globe.

On the Farnese globe the 42(41) constellations depicted (carved in reliefs) are precisely positioned against a grid of reference circles. The grid of reference circles on the Farnese globe are: the equator, the ecliptic with the zodiacal band (the zodiacal belt is a triplet of rings - the ecliptic being the middle of the 3 rings), the tropics (tropic of Capricorn, and tropic of Cancer), the colures (the solstitial colure), the Arctic circle, and the Antarctic circle. The Farnese globe was a display item and not a usable sky globe (i.e., the reference circles are drawn inexactly). The constellations depicted are consistent with the description of the sky in Aratus' Phaenomena.

The name Farnese Atlas was gained from its acquirement by Cardinal Alessandro Farnese (1520-1589) in the early 16th-century (and its exhibition in the Palazzo Farnese (Farnese Palace) in Rome). (Cardinal Farnese was a notable collector and patron of the arts.) It is now located in the Museo Nazionale Archeologico (previously named the Royal Museum of Naples, now the National Archaeological Museum) in Naples. (I am unsure of the date when it was transferred from Rome to Naples.) The Farnese collection at this museum is rated one of the finest collections of objects of Roman antiquity in the world. The collection passed to Naples when Elizabeth Farnese (1692-1766), the Queen Consort of Spain (wife of Philip V), who owned it by inheritance, passed it on to her son, Charles, who, in 1735, became the first monarch of the new Bourbon dynasty of Naples. The extensive Farnese collection was incorporated into a budding museum at the northwest corner of the old city premises that had originally housed a cavalry barracks and then the University of Naples.

Exactly where or when it was discovered is not known It would appear that the statue was unearthed in an archaeological excavation in Rome (possibly on the Palatine Hill) in the early 16th-century. (There are an estimated 500 Roman buildings on the Palatine Hill.) What is known for certain is that Cardinal Allessandro Farnese purchased it in 1562 from the antiquarian dealer Paolo del Bufalo. 

The statue is believed by art historians to be a Roman copy of an earlier (presumably) Greek original. The style of execution of both the statue and the constellations show that it is a copy of a Hellenistic original. It is generally thought that the existing white marble sculpture was made in Rome circa 150 CE and is a late copy of a Greek original made circa 200 BCE. The classical philologist Georg Thiele proposed the Farnese Atlas was a Roman (Hadrianic, dating to between 117-138 CE) copy of a Greek original. The date of the Farnese Globe remains uncertain. (Artistically the figure of Atlas has equal important with the globe. Also, it appears the globe was not used as a functional astronomical instrument by the Romans.) In Greek mythology Atlas was sentenced by Zeus to hold up the sky.

The Farnese Atlas first came to modern notice in the early 16th-century when it became part of the antiquities collection in the Farnese Palace in Rome. There are some slightly divergent accounts concerning its recovery. One account states it had been found in a damaged condition in the late 15th-century. Another account states that it was uncovered in the ruins of ancient Rome during the course of excavations in the mid-16th-century. The source of the Farnese Globe remains controversial. Circa 1555 it underwent extensive restorations. There is no evidence that the globe itself has been modified or reworked by this restoration work. The first detailed study of the Farnese Globe was made by the Italian astronomer and historian Francesco Bianchini (1662-1729). This study, likely carried out circa 1694-1695, remained in manuscript form. He also had detailed engravings made.

The Farnese Globe is quite austere in its details. (Note: Erwin Panofsky and Fritz Saxl write ("Classical Mythology in Mediaeval Art." Metropolitan Museum Studies, Volume 4, Number 2, March, 1933, Pages 228-280): "With the exception of the figure of Atlas, which was added in the Renaissance, it is a Roman copy of a Greek original.")

The north polar region of the celestial globe has some damage (i.e., there is a hole in the top of the globe) and so the constellations Ursa Major and Ursa Minor are missing. The globe depicts 42(41) of the 48 classical Greek constellation figures but not the stars comprising each constellation. (some writers state only 41 constellations are depicted.) For example, Sagitta is absent from the Farnese Globe. However, the stars may have originally been painted on the globe. One of the 42 constellations is obscured by one of the hands holding the globe. The constellation figures are depicted facing outward from the perspective of an observer positioned outside the earth and looking back towards earth (rather than inward, toward the centre of the globe). This means that the constellation figures are depicted back to front (i.e., are shown inverted) in comparison to the way an observer sees them from the earth (which would be a fixed point at the centre of the sphere). Naturally the earth is not represented but would be a tiny sphere centrally located inside the celestial sphere.

The constellations, which are shown in relief, are considered inaccurately depicted. One commentator has stated the constellations are depicted with moderate accuracy. (The Farnese globe seems to be connected - not with technical astronomy - but connected with the astronomical poem of Aratus of Soli. However, in his study Antike Himmelsbilder (1898) Georg Thiele denied this.) In his recent study of the Farnese globe the astronomer Dennis Duke concluded there is no reason to regard the globe being as a particularly high quality map of the constellations in the night sky. Included on the globe are a grid of traditional reference circles which include the Ecliptic (Zodiac), (Celestial) Equator, Tropics (Tropic of Cancer and the Tropic of Capricorn), Colures (Equinoctial Colure and the Solstitial Colure), Polar Circles (Arctic Circle (of star that never set) and Antarctic Circle (of stars that never rise). However, the First Point of Aries and the First Point of Libra do not coincide with the equinoctial colures. Also, the rectangular figure north of the zodiacal constellation Cancer appears to depict a constellation that is not mentioned by Aratus, Hipparchus, or Ptolemy. This apparently unrecorded constellation has yet to be identified.

The place of origin of the Farnese globe is uncertain but indicated to have been at a latitude of 38 degrees, which is close to the latitude of Athens and Pergamon. One modern estimate dates the statue of Atlas to 73 BCE and (from the position of the constellation figures to the globe's equinox) dates the constellation figures to circa 370 BCE. The interpretation of this dating is the celestial globe is probably a decorative attempt to depict the constellation figures described (circa 275 BCE) by Aratus in his astronomical poem Phainomena. The American astronomer Bradley Schaefer has recently announced (January 2005) his analysis of the globe and concluded that the date of the constellations depicted is indicated as circa 125 BCE. He believes this analysis has been strengthened by comparing the constellation figures with Hipparchus' Commentary. Schaefer's conclusion is that the constellation figures on the Farnese Atlas are an accurate depiction of Hipparchus' lost star catalog. Whilst Schaefer's work and conclusions has drawn some sharp criticisms it has been the belief of a number of scholars that the Farnese globe probably embodies the astronomy of both Eudoxus and Hipparchus.

If constructed, no celestial map has survived from ancient Mesopotamia. Likewise, no celestial map has survived from ancient Greece. A number of celestial globes have survived from the Roman period. These are based on Greek uranography. According to the astronomer and researcher Susanne Hoffmann there are several hints pointing to uranographies in both cultures. The Mesopotamian scheme of uranography used what may be termed equatorial coordinates. The Greek scheme of uranography used ecliptical coordinates.

There are only three celestial globes which survive from Graeco-Roman times. All three globes reproduce all or most of the constellations known in Graeco-Roman culture. These are: (1) the Farnese Globe; (2) the Mainz Globe; and (3) the Kugel Globe. Of these only the Farnese Globe is a large globe. Both the Mainz Globe and the Kugel Globe are miniature globes. Also existing are the Stuttgart Celestial Globe and the Berlin Celestial globe (fragment Sk 1050 A). It is currently believed that Archimedes celestial globe from the 3rd-century BCE has not come down to us. However, the work of Dr. Ulrich Kühne (2011) on fragment Sk 1050 A might indicate otherwise (See the discussion: Page 11-16). Basically, the globes are the only visual records of the constellations, to survive from antiquity.

The Farnese Globe, dated circa 150 CE, but likely based on the constellations as depicted in the 2nd-century BCE, is discussed above. The Farnese Globe is the largest of the celestial globes surviving from Classical antiquity. A number of constellations are either damaged or missing. However, the Farnese Globe is the archetypal model for the constellation illustrations. Georg Thiele, Antike Himmelsbilder (1898), demonstrated a direct link the constellations depicted on the Farnese Globe and those found in the earliest surviving Aratean manuscripts. As example: The 9th-century Germanicus translation in the Leiden manuscript, and the 2 early 9th-century Aratus latinus manuscripts.

The Mainz Globe is dated to circa 150-220 CE. It is a small brass globe that was acquired circa 1997 by the Römisch-Germanischen Zentralmuseum in Mainz (Germany). It is believed to have originated in Roman Egypt. It is a complete celestial globe in that it depicts all 48 Classical constellations with relative precision. It originally formed part of a sundial and was positioned on top of the gnomon. The entire Milky Way is depicted on the Mainz Globe and this is the first complete depiction known.

The Kugel Globe is dated circa 300-100 BCE. It may be the earliest celestial globe to survive from Classical antiquity. It is a small globe made of silver and is held in the Gallery J. Kugel Antiquaries in Paris (France). It was acquired circa 1996. It is thought to have originated as a copy of an existing sphere that had been repaired. This is because it appears to have reproduced the repair rivets on the original globe in the mistaken belief that they were celestial symbols. This would indicate that the artisan who made the copy was not knowledgeable in astronomy.

Appendix 1: The Farnese Atlas image with the fig-leaf add-on

Regarding the image of the Farnese Atlas depicted above. The exact origin of the electronic image remains unknown. The image appears as Plate IIa in Le Ciel des Romains by André Le Boeuffle (1989). However, with the exception of the fig-leaf add-on the image matches Tafel II in Antike Himmelsbilder by Georg Thiele (1898). The Farnese Atlas photographs appearing in this book were taken by the Berlin-based photographer Paul Schahl. The image above was possibly derived, either directly or indirectly, from Paul Schahl's photograph in the book, or his photograph was perhaps accessed from another source. The date of modification of an original photograph, by way of adding-on the fig-leaf, remains unsolved but is most likely quite modern (and perhaps original to Boeuffle's 1989 book). The solution would depend on confirming the particular photographic source for the modification and scanned image. A likely source is the (unmodified) photograph of the Farnese Atlas, comprising Plate 19, in La Science Antique et Mediévale edited by René Taton (1957). However, this is credited to the Italian firm of art photographers Fratelli Alineri (established in Florence in 1852 by, as the name indicates, the Alineri Brothers). It is a match for the photograph comprising Tafel II in Thiele's book. (The Farnese Atlas image with the fig-leaf add-on has been published in the well-known American monthly astronomical magazine Sky and Telescope.)

Appendix 2: Early surviving depictions of Greek constellations?

Early Greek astronomy is known primarily through its literary sources. There are only a very few surviving depictions in Greek classical art of their constellations. There are also Roman mosaics showing the constellations. A constellation depiction survives from Harran, a major ancient city in upper Mesopotamia, located on the west bank of the Balikh River. Pottery depicting constellations has been recovered from the ancient city of Rayy (Iran). Rayy dates back to circa 300 BCE. The constellation depictions were likely due to the influence of immigration of Greek and Macedonian settlers following the military conquests of Alexander the Great.

To know how the pre-Classical Greeks viewed the night sky (i.e., before Aratus) researchers must rely on visual depictions of the sky, such as those found on ceramic pottery. However, these artifacts are relatively rare, and what has been recovered through excavation generally only show one or two constellations. Also, these earliest Greek artistic representations of constellations are fragmentary and can be problematic in terms of what they represent.

Scenes located in the sky are not uncommon in late Archaic (540-480 BCE) and Classical (480-323 BCE) Greek art, primarily in krater (large vase) painting, and are mostly indicated by the presence of stars as locational devices. However, these scenes are most usually mythological only in nature. They do not directly depict constellations or reflect the physical realities of the sky.

Amongst the Greeks, Romans, and Etruscans, the depiction of constellations, the sun, the moon, and the planets, when represented in human form are surrounded by either luminous rays or by luminous circles (nimbi and aureoles).

"The scenes of myth and daily life that decorate Athenian vases often have a pronounced sense of time, which is depicted in simple pictorial terms that are meant to be easily recognized. Night, for instance, can be signified with lamps, torches, and the presence of the appropriate nocturnal deities, Selene the moon goddess, and Nyx, the very personification of night. Similarly, Helios the sun god and Eos the goddess of dawn indicate daytime. The great frequency of temporal motifs on vases suggests that time was integral to the narrative construction of many vase paintings. Moreover, the deliberate references to time on Athenian vases can often be explained as an essential feature of the specific subject portrayed." (Source: Udell, Jennifer. "Time of Day on Painted Athenian Vases". In: Heilbrunn Timeline of Art History. New York: The Metropolitan Museum of Art, 2000 –. http://www.metmuseum.org/toah/hd/time/hd_time.htm (October 2004))  

An early identification of a (non allegorical) constellation on a Greek vase (sherd from a krater ("large vase")?) is given by Samuel Birch on Page 346 of his book History of Ancient Pottery (1858) : "The constellation Pegasus appears once with the moon." Seemingly, another reference (but possibly otherwise) to this states the example, possibly one of the oldest constellation images from Greece, comes from a pottery (krater ("large vase")) fragment from the Late Geometric period (760 to 700 BCE) found at Pithekoussai on the island of Ischia in Italy, but it only depicts what may be the constellation Boötes ("the Herdsman"). (George Huxley, philologist, Professor at Queen's University, Belfast, has tentatively identified it with the constellation Boötes. See: Coldstream, J[ohn]. N. and Huxley, G[eorge]. L. (1996). "An Astronomical Graffito from Pithekoussai." (Parola de Passato, Volume 51, Pages 221-224).)

Another fragment of pottery, dating to the 5th- or 4th-century BCE, excavated at Canosa (southern Italy, a pottery centre for pottery made for the Greek market) depicts a bull and the stern of a ship and has been interpreted as the constellations Taurus and Argo.

In 2014, John Barnes, a classical archaeology doctoral candidate at the University of Missouri analysed a two-handled wine cup on display at the Lamia Archaeological Museum in Greece and believes it depicts one of the earliest Greek depictions of the constellations. This item of ancient pottery (dating to 625 BCE), called a skyphos, has long been thought to depict a random assortment of animals. The artifact, which dates back to 625 B.C., was originally discovered in a debris-filled trench next to a temple in the 7th-century BCE acropolis of Halai, which is located about 40 kilometers (25 miles) north of Thebes, Greece. About a third of the wine cup (the base area including one handle) is missing. The skyphos depicts an array of animals: a bull (with only the back half preserved), a snake, a hare or small dog, a large dog, a scorpion, a dolphin and the front half of a panther or lion. The skyphos is displayed with a label stating a simple animal scene is shown. John Barnes believes it is more likely that the animals are constellations: The bull is Taurus; the snake is probably Hydra (rather than Serpens or Draco, two other serpent constellations recognized by the Greeks); the rabbit is Lepus; the dog is Canis Major or Canis Minor; the scorpion is Scorpius; the dolphin is Delphinus; and the lion is Leo. However, the animals are not arranged on the skyphos in the order they appear in the sky. See: "Asteras Eipein: An Archaic View of the Constellations from Halai." by John Barnes (Hesperia, Volume 83, Number 2, 2014, Pages 257-276).

Too few Greek artifacts have survived to enable the history of the development of constellation figures to be traced.


Copyright © 2001-2018 by Gary D. Thompson

Return to top of page.

This Web Page was last updated on: Saturday, December 9, 2017, 11:30 am.

This Web Page was created using Arachnophilia 4.0 and FrontPage 2003.

You can reach me here by email (but first delete the obvious attempt in the email address to foil the spammers): garyREMOVE.thompson10@bigpond.com

Return To Site Contents Page