CHAPTER 6

ERATOSTHENES AND THE INVENTION OF THE DISCIPLINE OF GEOGRAPHY

Shortly after his accession in 246 BC, Ptolemy III hired Eratosthenes of Kyrene to be librarian at Alexandria, and, eventually, tutor to his son, the future Ptolemy IV.¹ Eratosthenes had been studying and writing in Athens for nearly 20 years, specializing in philosophy and mathematics, and had a reputation as a philologos, a learned scholar.² Within a few years his academic career would move in a remarkably different direction. As librarian he had access to essentially everything that had been written in Greek, which allowed him to do the research that led to his two profound works on geography: the Measurement of the Earth and the Geography.

The Measurement of the Earth was the earlier work, surviving in only a handful of fragments, mostly preserved by mathematicians.³ In modern times it was long thought to be a part of the later Geography, and it was not until the early twentieth century that it was recognized as a separate treatise, although ancient commentators were well aware of the fact.⁴ The complete scope of the work is unknown, but it probably was brief and devoted to the methodology of calculating the size of the earth.

Suggestions for the circumference of the earth had existed since at least the fourth century BC,⁵ but these were probably more intuitive than the result of any actual calculations, with the exception of a figure of 300,000 stadia attributed to Dikaiarchos. Eratosthenes refined the rough methodology used to determine that figure, basing his conclusions on the knowledge that Alexandria and Syene (at the First Cataract of the Nile) were on the same meridian (the difference between their meridians is only about 100 miles), and that they were 5,000 stadia apart.⁶ This was not an estimate, but the result of a royal survey commissioned by the government (whether for Eratosthenes’ benefit is unknown).⁷ It was realized from the report of Philon⁸ that Syene lay on the summer tropic (actually a few miles to its north, but this made no difference), and thus there was no shadow cast at the summer solstice, something that could be determined by the use of a gnomon (essentially a measuring stick).⁹ In Alexandria at the same time a gnomon cast a shadow that formed an angle of one-fiftieth of a circle, which meant that the distance between the two points was one-fiftieth of the earth's circumference.¹⁰ This is a simplified summary of the technique, which relied on a knowledge of ratios, triangles, and a method to measure angles. The result is a circumference of 250,000 stadia, which Eratosthenes adjusted to 252,000, perhaps to provide a number divisible by 60.¹¹ His figure was not automatically accepted: in the following centuries both Hipparchos and Poseidonios offered their own calculations, resulting in smaller totals.¹² Conversion of Eratosthenes’ distance to modern measurements is impossible since the length of the ancient stadion varied, yet it seems remarkably close to an accurate number.¹³ Three hundred years later Pliny noted the spectacular nature of Eratosthenes’ feat, “expressed by such a subtle argument that one is ashamed not to believe.”¹⁴ It was now possible to understand the earth as a whole, and, more importantly, to locate the oikoumene properly on its surface, positioning known points, and thus creating an accurate view—whether by a map or merely descriptively—of the world.

Eratosthenes’ subsequent work was his Geography. This was probably the earliest use of the word (Greek geographia), a new term created by him. The treatise was in all likelihood completed before the Roman advance onto the Greek mainland in 218 BC, an event seemingly unknown to Eratosthenes. It survives today in approximately 150 fragments or paraphrases, more than 100 of which were recorded by Strabo, who essentially provided a summary of the work in the first two books of his own Geography. The actual treatise probably existed no later than the second century AD: it was familiar to Pliny in the previous century, and thereafter Arrian may have been the last to see one of the rare surviving copies.¹⁵

The Geography was not a lengthy work: only three books are known. There is little evidence that Eratosthenes did much fieldwork beyond visits to Arkadia, Achaia, and Rhodes, and it is probable that the entire treatise was written using the resources of the Alexandria library, supplemented by data received from merchants and seaman in the city.¹⁶ There are about 20 sources cited, from Homer to Eratosthenes’ colleague, the mathematician Archimedes, who dedicated his Method of Mechanical Theorems to him. Most of Eratosthenes’ sources are recent, but the pioneers of geographical thought, such as Homer, Anaximandros, Aeschylus, Hekataios, and Herodotos, were also credited. The work (as summarized by Strabo) is the major source for many of the geographical explorers of the fourth and third centuries BC, such as Pytheas, Androsthenes, Patrokles, Megasthenes, and others. Whether or not a map was included is still debated.¹⁷

The first book begins with a summary of the history of geography, from Homer to the explorers of the early third century BC. There appears to be a special emphasis on the issues surrounding Homer's role as a geographer, although this may be due to the summarizing of the treatise by Strabo, who was trained as a Homeric scholar and who disagreed with Eratosthenes about the poet's importance. Eratosthenes then continued with a discussion of the physical nature of the earth and its formative processes, citing sources such as Xanthos of Lydia and Straton of Lampsakos. The book closed with a consideration of an increasingly popular genre—fantasy geography—in which mythical locales were created, often for allegorical reasons, and generally placed beyond the limits of the known world. The concept may have originated with Plato's Atlantis, but became common after the explorations of Pytheas and Alexander, as they had penetrated into the strange extremities of the inhabited world.

The second book described the size and shape of the earth. Eratosthenes emphasized that his views depended on a spherical earth upon which all the inhabited portions might not be known, demonstrating that these were still new ideas that were not totally accepted.¹⁸ The arguments in this book are based on the conclusions reached in the Measurement of the Earth, and the Geography may have included a summary of it, perhaps without the most technical data. The book also dealt with the concept of the oikoumene, stressing that it was longer east–west than north–south, and assuming that one could sail west from the Pillars of Herakles and reach India, ideas that had been gestating for some time.¹⁹ There was also an exposition of the five zones, essentially as they had been set forth by Eudoxos of Knidos.²⁰ But unlike the sparse surviving comments of earlier theories about the oikoumene, the actual diction of Eratosthenes is preserved, demonstrating (as he had with the term “geography”) his continuing role as an innovator of the language.²¹ He marked off an area between two parallels that went around the entire earth, with the oikoumene lying between them, and called this shape a spondylos, or spindle whorl.²² In addition, the oikoumene, which was an island in the middle of the spindle whorl, was shaped like a chlamys. The word he used, chlamyoeides, was new: a chlamys was an outer garment worn by horsemen, most familiar on the riders of the Parthenon frieze. It gave a shape to the inhabited world, a rectangle narrowed at the corners, for it was believed that the inhabited world tapered at its extremities. By using common domestic terminology for his complex concepts, Eratosthenes (who, after all, was better known as a philologist than a geographer) made his cosmic speculations more palatable to non-specialists. Book 2 may also have included some topographical data, especially about the location of Syene, the crucial point in understanding the size of the earth, but most of the topography was reserved for Book 3.²³

The heart of the Geography was Book 3, in which Eratosthenes described the topography of the entire inhabited world, beginning with India and heading west, the reverse of previous and future descriptions: he may have done it this way because of the recent interest in India. He began by establishing his prime parallel and prime meridian (Map 8). The parallel had already been outlined by Dikaiarchos, but Eratosthenes refined and straightened it, extending it from the Pillars of Herakles through the Sicilian Strait (modern Straits of Messina), the southern Peloponnesos, Attika, Rhodes, and the Issic Gulf (the modern Gulf of Alexandretta at the northeastern corner of the Mediterranean).²⁴ Sardinia, the most anomalous point of Dikaiarchos’ parallel, was eliminated, and Eratosthenes’ line runs somewhat to the south of that of Dikaiarchos. It was necessary that the parallel go through known points, and thus it was not a straight line, but Eratosthenes was aware of this and the deviation was not great.²⁵ East of the Gulf of Issos, however, it was somewhat more problematic, defined as “along the entire Tauros” as far as India, passing through two points known to Alexander's expedition, Thapsakos on the Euphrates and the Caspian Gates.²⁶ The data reported to Eratosthenes told him that the easternmost part of the Tauros (the modern Himalayas) turned to the north, and he realized that this was an error, so he corrected it by making the mountain range extend due east-west. In fact the Himalayas run northwest to southeast, through 10^o of latitude, and Eratosthenes’ adjustment still resulted in the placement of India too far to the north. Relying on the 15,000 stadia provided by Patrokles for the north–south extent of India, Eratosthenes concluded that the southern end of India (modern Cape Comorin) was on the same parallel as Meroë. In fact that parallel is about 9^o north of the one through Cape Comorin, but nonetheless it was an amazing feat to connect two places so far apart, with thousands of miles of unknown territory between them. It was probably the first time such had ever been attempted.

The concept of a meridian—a north–south line across the oikoumene—had existed since the fourth century BC, but there is no evidence of any attempt to plot one before Eratosthenes. Dikaiarchos, however, had come close with his consideration of the relative position of Syene and Lysimacheia. Inspired by the north–south course of the Nile and his previous use of this line to determine the circumference of the earth, Eratosthenes created a prime meridian that was based on the river from Syene to Alexandria.²⁷ It could be extended to Meroë, also on the Nile and essentially due south of Syene. Farther south was the parallel of the Cinnamon-Bearer Territory, which crossed the Nile. This hypothetical point was said to be 3,000 stadia south of Meroë (somewhere in southern Sudan), but there is no evidence that it was ever plotted on the ground. It was believed to be the southern end of human habitation, because beyond this line it was too hot for people to live. These calculations demonstrated how Eratosthenes’ developing grid could determine unknown and inaccessible points on the surface of the earth.

Map 8. The prime parallel and prime meridian of Eratosthenes.

To the north the meridian ran through Rhodes, Lysimacheia, Olbia, and eventually to the parallel of Thule. The line from Alexandria to Lysimacheia, presumably a shipping route, had already been noted by Dikaiarchos, and thus it could be extended to the north. Because of its dependence on a sailing lane from Alexandria, the meridian swung west around the end of Anatolia and then headed northeast across the Black Sea to the Greek settlements on its north shore. It was not easy to plot beyond this point, for even though it had long been known that various Skythians lived to the north, there were no details and no known points or any obvious way to extend the meridian in the manner that it could follow the Nile to the south. Implicit in this, but not actually stated, was that Syene and the parallel of the Cinnamon-Bearer Territory were far closer to the equator than Olbia was to the north pole (and thus the oikoumene was in the southern part of the northern hemisphere). For the far north, all that Eratosthenes could do was to run the meridian north from Olbia to the parallel of Thule.

The creation of a prime meridian and a prime parallel—of utmost importance to understanding the location and nature of the inhabited world—allowed Eratosthenes to position the chlamys-shaped oikoumene on the surface of the earth. The parallel and meridian crossed on Rhodes, which thus became the nexus of geographical thought. Since the latitude of Syene was known (it was at the summer tropic, the boundary between the temperate and burned zones), and many distances along both lines were available from travelers’ reports, or, in the case of Syene-Alexandria, had been measured on the ground, it was now possible to suggest the actual size of the inhabited world: 30,000 stadia north–south by 70,000 east–west. This was a rough figure dependent on many variables: for example, even though the Pillars of Herakles were used as the west end of the prime parallel, it was obvious that land extended to the west both north and south of the Pillars. Eratosthenes, or perhaps Strabo, believed that the “bulge of Europe” added 3,000 stadia to the west.²⁸ There were also issues about the placement of India; moreover, Thule, whose parallel was now used as the northern limit of the inhabited world, remained enigmatic and its existence even came to be questioned. For various reasons the length of the oikoumene was far too long, something that affected Renaissance exploration, since it made the westward route from the Pillars to India appear shorter than it was, an error that actually became worse over time as the far east was added to the oikoumene. Nevertheless the perfection of a prime meridian and a prime parallel was another of Eratosthenes’ stupendous feats, for it allowed an understanding and sizing of the inhabited world, locating it on the surface of the earth. New parallels and meridians could now be created, even extending to places far from where any Greek had ever been, and the grid of the oikoumene could be filled in. As one might expect, much of the extent of the new lines was theoretical or even non-existent, especially the eastern meridians. Few points were plotted at the extremities, such as north of the Caspian Gates or in interior Africa. In theory, however, it was possible to determine the location of any point where a parallel and meridian crossed. Yet the major flaw of the system, as Hipparchos of Nikaia was soon to point out, was that it really depended on information received and known locations, even though some latitudes could be determined by Pytheas’ method of the height of the sun at the solstice.

Having placed the inhabited world on the surface of the earth, and determined its size, Eratosthenes then considered its actual nature. He divided it into units that he called sphragides (gem or seal stones), a word also used in Egypt to describe a plot of land.²⁹ Their exact shape is not certain, perhaps rectangles with rounded corners, but this does not easily apply to any of the four sealstones that Eratosthenes proposed. The idea may have occurred to him because India, the first sealstone, was an easily defined unit.³⁰ This sort of thinking showed the influence of Euclidean geometry, a relatively new discipline in Eratosthenes’ day.³¹ The idea is important because it looks at geography in terms of land forms, not, as Ephoros had, through ethnicity, yet the concept of sealstones was not truly viable and did not survive as part of geographical theory after Eratosthenes. In fact, Eratosthenes himself began to have reservations, since the sealstones only seemed to work in the east, perhaps because of the large physical units that existed there.

The four sealstones were India, Ariana, Mesopotamia, and Arabia/Egypt/Aithiopia; the amalgamated quality of the last indicates the breakdown of the concept. There was no problem with India and Ariana, but there were serious difficulties with Mesopotamia. The peculiar fourth one, mentioned only once, became virtually impossible.³² There is no evidence of any sealstones farther west, and the idea did not work at all in the Mediterranean world.

Within the concept of sealstones, Eratosthenes began his discussion of the inhabited world, moving from east to west, and mixing topographical and ethnographic data. About 400 toponyms are known, from the familiar to the still unlocated.³³ The Indian section relied on the explorers of the era immediately after Alexander, such as Patrokles and Megasthenes, and not only considered topography but the routes between India and the Mediterranean;³⁴ yet the description of the second sealstone, Ariana, is scant. Mesopotamia was examined more thoroughly but the increasing problems with the sealstone concept tend to burden Eratosthenes’ arguments and data.

As was normal for his era, Eratosthenes believed that the Caspian was an inlet of the Ocean.³⁵ Egypt was described in some detail, and there are comments about coastal Africa as far as the Carthaginian outposts on the Atlantic.³⁶ The remainder of the Geography, from Anatolia across southern Europe to the Iberian peninsula, is sparsely preserved, because these regions, especially west of the Greek peninsula, were better known from other sources by Strabo's time. In fact Strabo, perhaps unfairly, took Eratosthenes to task for his ignorance of the west.³⁷ Rome is nothing more than an improperly placed toponym between two parallels.³⁸ Moreover, Eratosthenes, expectedly, relied on Pytheas for the far northwest, whom Strabo believed should be discredited; yet without Eratosthenes (through Strabo), Pytheas and his achievements would be far worse known than they are today. In fact, the most difficult issue in understanding Eratosthenes is that in large part he must be approached through the lens of Strabo, two centuries later, when geographical knowledge had changed significantly, especially in the western Mediterranean, something that Strabo was not generally sensitive to.

The end of the Geography is an interesting discussion of ethnicity, and how at the time of Alexander a Hellenocentric orientation had begun to change to a more global point of view with the realization that the Greeks did not have a monopoly on culture and civilization.³⁹ It was now understood that the oikoumene was only a small part of the earth, and the Greek world itself was only a portion of the oikoumene. There were advanced cultures—particularly in India—who had not even known about the Greeks. Thus geography was beginning to change traditional views about the peoples of the earth.⁴⁰ Eratosthenes gained great fame for his Measurement of the Earth, yet his Geography was soon forgotten. In 218 BC, as he was finishing the work, the Romans gained their first foothold on the Greek peninsula. Seventy years later Carthage was destroyed. The topographic reality of the inexorable spread of Roman power resulted in major changes in the Mediterranean world, and within a century of its completion the Geography was seen as obsolete. Later scholars with extensive bibliographic knowledge, such as Plutarch or Athenaios, made no mention of it. Yet the fact remains that, however quickly the Geography became out of date, Eratosthenes had invented a new academic discipline.

Hipparchos of Nikaia

The most serious early critic of Eratosthenes was Hipparchos, active in the second century BC; the latest date associated with him is 126 BC.⁴¹ He was from Nikaia in northwestern Anatolia but spent most of his academic career on Rhodes. Hipparchos was not a geographer, but a mathematician and astronomer. He wrote a single work of relevance to the history of geography: Against the “Geography” of Eratosthenes, which nonetheless is more mathematical than geographical. It survives in 63 fragments, 55 of which are from Strabo's Geography, and thus there is the same difficulty in assessing Hipparchos’ contribution as there is with Eratosthenes: in fact the two are often tangled together in Strabo's recension, coupled with Strabo's own comments.

Hipparchos’ treatise was in three books, the same as the work that he sought to refute.⁴² His complaint was that Eratosthenes based his distances and positioning of toponyms on hearsay reports, and he believed, quite reasonably, that the only accurate way to locate places was through mathematics and astronomy, not travelers’ accounts.⁴³ He had mathematical skills that Eratosthenes did not, most notably an ability at what would today be called trigonometry, and developed an astronomical method of positioning places on the surface of the earth.⁴⁴ The flaw in his reasoning is not his methodology, but that implementation of such a technique would require a competent observer on the ground at each point, something obviously impossible, at least at that time. Thus Hipparchos was beholden to Eratosthenes’ system of information received—especially for remote places—while realizing, as had Eratosthenes, its flaws.

Nevertheless Hipparchos made a number of alterations to Eratosthenes’ scheme of the oikoumene, most of which were improvements, although some actually made Eratosthenes’ data worse. Hipparchos was able to adjust the main parallel somewhat, moving it to the south in the western Mediterranean,⁴⁵ and he made other corrections in this region (with the fall of Carthage in 146 BC, the west had become better known to Greeks). He realized that something was wrong with the positioning of India and the Imaos Mountains, but eliminated Eratosthenes’ partial correction and reverted to the older, more erroneous data.⁴⁶ This may have had a cascading effect elsewhere, since he validated the assumption that Byzantion and Massalia were on the same latitude, when their parallels are actually 2^o apart. But he did provide as many astronomically determined locations as possible (although these were few) and outlined the techniques for doing so.⁴⁷

Hipparchos’ contribution to geography is not so much his critique of Eratosthenes’ methodology as his understanding that an accurate conception of the oikoumene was only possible through precise positioning of toponyms and features, and that this depended on astronomy, not travelers’ information. It was impossible to apply this theory consistently—something that Hipparchos himself understood—but nonetheless he created the basic theory of mapping.⁴⁸ Yet like all such endeavors before the eighteenth century, Hipparchos was hampered by the difficulties in determining longitude, although he realized that observation of eclipses at separate points on the same parallel would provide the information. His statement on the matter, as quoted by Strabo, is a neat summary of the issue:

It is not possible for anyone—whether an amateur or scholar—to undertake geographical research without determination of heavenly phenomena or the eclipses that have been observed. How far Alexandria-next-to-Egypt is north or south of Babylon—or what the distance is—cannot be determined without an investigation of the latitude. Similarly, the displacement to east or west cannot accurately be learned without comparing the eclipses of the sun and the moon.⁴⁹

Other Developments in the Second Century BC

A number of other issues affected geographical scholarship in the second century BC but these were essentially tangential to the mainstream. Homeric studies became popular during this era, and it was assumed that Homer had made important contributions to geography, a view already endorsed by Eratosthenes at the beginning of his Geography.⁵⁰ In fact, the juxtaposition of Homeric studies and topography had long existed, and even served as a way of popularizing geography.⁵¹ Krates of Mallos (in Kilikia), Pergamene ambassador to Rome in the 160s BC, addressed a number of topographical problems in the Homeric poems, believing that the wanderings of Odysseus took place in the External Ocean, emending Homer to this end.⁵² He also considered the location of the Aithiopians, something ambiguous in the Homeric poems.⁵³ These arguments were more philological than topographical, yet of particular interest is that Krates used a globe to illustrate his theories, the first evidence of such a device, and of any attempt to represent visually the entire earth.⁵⁴ Whether he actually made one, or just suggested that it could be constructed, remains unclear. It was 10 feet in diameter and showed how limited the oikoumene was within the entirety of the world. The globe could also show the parallels and meridians, and how the latter converged at the poles, although whether this was Krates’ or Strabo's suggestion is not certain. Krates’ topographical views were driven by his Homeric exegesis, and he theorized that the earth consisted of four land masses,⁵⁵ separated by streams of the Ocean (thus reconciling the differing Homeric views of the Ocean as both expanse and river), a view that was popular in medieval times but hardly affected ancient geographical scholarship.⁵⁶

Another Homeric scholar of this era was Demetrios of Skepsis, a small town in the Troad (at modern Kuşunlu Tepe, about 20 miles southeast of the site of Troy⁵⁷). Demetrios wrote an exhaustive commentary on the Homeric Catalogue of Ships, 30 books long, which was a major source for Strabo. He also considered topographical issues in the tale of the Argonauts.⁵⁸ Although it is difficult to imagine how he could have filled 30 books (as long as Ephoros’ universal history) on 60 lines in the Iliad, he was particularly concerned about the location of Troy.

His younger contemporary, Apollodoros of Athens, who relied heavily on Demetrios’ work (Strabo found him derivative), was also interested in the location of places and peoples mentioned in the Homeric poems.⁵⁹ Hestiaia of Alexandria, the only woman topographer known from antiquity, was probably also from this era: she, too, wrote on the Iliad and seems to have performed fieldwork in an attempt to locate the site of Troy, realizing that the landforms may have changed since early times.⁶⁰

These Homeric scholars of the second century BC, whose topographical efforts are preserved today almost solely in the Geography of Strabo, were not truly geographers. Yet Strabo himself was trained as a Homeric scholar, and the relevant passages in his Geography are the best surviving example of this type of Hellenistic commentary on Homer.⁶¹ As knowledge of the inhabited world expanded, the Homeric scholars attempted to extend Homer's geographical reach to fit the current reality. They wanted Homer to have been aware of the extremities of the earth, particularly the western Mediterranean and what was beyond the oikoumene. Yet they were attempting the impossible: Homer's world was limited to southern Italy and points east. Nevertheless they made important contributions to the study of topography, essentially inventing it as it is still practiced today.

At the end of the second century BC, Artemidoros of Ephesos wrote a Geographoumena in 11 books, probably at that date the longest on its topic.⁶² Portions survive in an epitome by Marcian of Herakleia, and a papyrus fragment of his text has attached to it a controversial, but perhaps ancient, map of part of Iberia.⁶³ Artemidoros may have written the first geographical account of western Europe, using contemporary Roman data. Strabo relied on it regularly, and it was one of his most cited authorities, largely for distances, although he did not place Artemidoros in the first rank of geographical scholars. Much of his western data was from autopsy, but Strabo found him less valid in regard to India.⁶⁴ Artemidoros calculated the east–west length of the oikoumene at 68,645 stadia, a lower figure than usual, probably because he rejected any information from Pytheas.⁶⁵

There was also further interest in the tides. Pytheas had been the first to connect them with the moon, and extensive Greek travels on the Ocean, especially by those with Alexander and the Ptolemaic explorers of the Red Sea, had increased exposure to the phenomenon. Seleukos of Seleukeia (at the head of the Persian Gulf), active in the mid-second century BC, seems to have been the first to write a treatise on the topic.⁶⁶ He further developed Pytheas’ idea that they were associated with lunar activity, and attempted to establish a full theory of the tides. Tidal activity is still not fully understood, and Seleukos’ theory was inevitably incomplete and erroneous. Hipparchos and Poseidonios attempted to verify it but were unable to do so, and Seleukos’ efforts came to be virtually forgotten, surviving in only a few oblique references by Strabo. Seleukos was one of the profound but little known personalities of antiquity: he seems to have been the last to support the heliocentric theory of Aristarchos of Samos, even providing a proof.⁶⁷