2

ERATOSTHENES’ CHLAMYS-SHAPED WORLD: A MISUNDERSTOOD METAPHOR

Klaus Zimmermann

Just about every treatment of Greek geography mentions the two attempts of hellenistic science to give the reader an idea of the landmass surrounding the Mediterranean by comparing it with an article of daily use: Erato-sthenes’ οικουμ?νη χλαμυδοειδ?ς taken up by Strabo¹ and Posidonius’ οικουμ?νη σφενδονοειδ?ς. While for an understanding of the latter comparison we only have to look at the one sling known from antiquity,² there is no definite archaeological evidence to help us interpret the meaning of the adjective ‘chlamys-shaped. Given the way in which the garment was draped, figural representations allow only partial conclusions on its shape when spread out. Further, the cut of the chlamys may have varied across time as well as across different regions. Thus, modern scholars focusing on ancient geography usually content themselves with citing the known metaphor and its references.

In order to reconstruct the idea Eratosthenes had in mind, there are two questions we have to consider separately: 1. What shape did the chlamys have? 2. How (if at all) can Eratosthenes’ geographical knowledge be harmonized with this picture? On that basis, the third and last part of my paper will be devoted to the question: What did Eratosthenes actually want to express with this comparison?

I

As Tarbell at the beginning of the twentieth century noted, ‘we are in the habit of applying the name “chlamys” with a great deal of confidence to all small brooch-fastened outer garments represented in Greek art.’³ This practice is based on literary evidence: Ovid (Metamorphoses 14.393-4), Suetonius (Tiberius 6.3) and Isidore (Origines 19.24.2) mention the brooch as a characteristic feature of the chlamys. In addition, Ovid (Metamorphoses 2.733), Lucian (Timon 30) and Pausanias (5.27.8) describe the chlamys as a typical garment of the god Hermes. Thus, the link to archaeological material is established. Representations of Hermes, Oedipus, the Niobids and others from the classical period show that the garment in question was put around the left shoulder and closed on the right, that it was four- cornered and that it had a rectangular shape (figs. 1-2).⁴

Fig. 1. Heuzey 1922, 122 fig. 61.

Fig. 2. Heuzey 1922, 123 fig. 62.

However, in contrast to the classical form an obviously Macedonian variant with a circular lower edge seems to have prevailed in hellenistic times. There is evidence for this both in ancient literature⁵ and visual arts.⁶ But above all, we owe to Strabo, Pliny and Plutarch descriptions of Alexandria which compare the outline of that city to a chlamys.⁷ Beginning with Strabo, we learn more about the dimensions of Alexandria than about the exact shape of the garment:

The shape of the area of the city is like a chlamys; its long sides are washed by the two waters, having a diameter of about thirty stadia, and the short sides are the isthmuses, each being seven or eight stadia wide and pinched in on one side by the sea and on the other by the lake.

This passage provides only a very rough outline of the structure we have to imagine: a more or less straight long side in the north (the coast of the Mediterranean), two narrow sides, whose angles in respect to the coastline are left unmentioned, and finally a second long side in the south (approximately corresponding to the shore of the lake), which may have been slightly curved towards the south, as the use of the terms διαμετρον⁸ and ισθμοι seems to indicate.

A somewhat more precise description is furnished by Pliny:

metatus est eam (sc. Alexandriam) Dinochares architectus.. .ad effigiem Macedonicae chlamydis orbe gyrato laciniosam, dextra laevaque anguloso procursu. (Pliny, Naturalis Historia 5.62)

It (sc. Alexandria) was laid out by the architect Dinochares. in the cornered shape (ad effigiem laciniosam) of a Macedonian chlamys with a circular contour and a projecting corner on the right and on the left sides.

It is true that the exact meaning of the adjective laciniosus is not quite clear, lacinia designating the corner⁹ as well as the hem of a cloth.¹⁰ Anyhow, the pattern Pliny talks about must have corners that protruded beyond the coast sector: the somewhat clumsy addition dextra laevaque anguloso procursu hardly makes sense if in fact the circular contour (orbis gyratus) simply met a straight base line.¹¹ Hence, we have to imagine the outline of Alexandria approximately as given in Fig. 3.¹² Now the question is whether the angulosiprocursus were typical of the Macedonian chlamys, thus coinciding likewise with Fig. 3, or whether the corner on the right and left side in Pliny’s description has to be understood as a divergence from the basic form of the chlamys (orbe gyrato), corresponding grosso modo to Fig. 3 with the dashed lines from A to E and from B to F. For the moment, this question will best remain unanswered.

Fig. 3. Cf. Tarbell 1906, 284 fig. 1.

Fig. 4. Cf. Tarbell 1906, 284 fig. 2.

In connection with the references just mentioned, it is Plutarch’s narrative of the foundation of the city that clarifies things:

They drew a rounded area the inner arc of which was continued by straight lines as from the seams towards the shape of a chlamys, narrowing the size evenly.

Taken on its own, this passage is not easy to understand either. Following Tarbell, I propose to translate the expression ? ?ντος περιφ?ρεια as ‘the circular contour on the landward side’.¹³ The ευθειαι βασεις can hardly be anything other than the narrow sides of the city’s area.¹⁴ But does ?ξ ισου συναγειν το μ?γεθος mean a narrowing in respect to the maximum breadth C-D (as in Fig. 3)? Or do we have to imagine two parallel lines rising up from the circular arc towards the north and meeting the Mediterranean coast in a right angle (as do the dashed lines in Fig. 3)?¹⁵ In a sense, this interpretation would also be compatible with συναγειν το μ?γεθος, the imaginary full circle being narrowed ?ξ ισου.

Here we can rely on Pliny who shows us the shape the perimeter of Alexandria actually had: its corners projected to the right and to the left sides. The marking out of an area εις σχ?μα χλαμυδος, described by Plutarch, must have led to the contour referred to by Pliny. In other words: the chlamys familiar to those who created the comparison with the outline of Alexandria had approximately the shape given in Fig. 3. In contrast to the classical rectangular chlamys, the change is not only brought about by cutting off, then rounding, the rear edges to get rid of the bothersome tips hanging down.¹⁶ Apparently, the front edges were also reduced in order to avoid an unnecessary gathering of folds at the right, open side (see Fig. 2). One may wonder at what point and at which angle those cuts reached the long side of the garment. In any case, a shape like the second one proposed by Tarbell (Fig. 4)¹⁷ certainly did not apply to the chlamys with which the outline of Alexandria was associated, for it is contradicted not only by Strabo’s reference to isthmoi between the two waters, but also by our knowledge of the topography and the earliest building activities of Alexandria.¹⁸ The shape given in Fig. 3, however, might fit the archaeological data to some extent.

So much for the Macedonian chlamys as we know it from the comparison with Alexandria. The fact that the characterization of the oikoumene as χλαμυδοειδ?ς, attested five times in Strabo’s second and eleventh books, does go back to Eratosthenes emerges clearly from 2.5.6 C 113: after localizing the oikoumene roughly on the northern hemisphere, then within a quadrilateral of equator, polar circle and a perpendicular meridian circle, the author writes:

The inhabited world is a chlamys-shaped island in this (sc. quadrilateral), being less in size than half of the quadrilateral.

A few sentences later Strabo closes the quotation as follows:

In essential accord with all this are also the views of Hipparchus. For he says that, having taken as hypothesis the measurement of the earth as stated by Eratosthenes, one has to subtract the inhabited world from the earth.

Berger’s inclusion of the preceding passage among the fragments¹⁹ is, therefore, very probably correct.²⁰ And it seems rather unlikely that a metaphor applied only twice in ancient literature, once to the shape of hellenistic Alexandria, once to the oikoumene by an hellenistic author working at Alexandria, is based on two different variants of chlamydes. Thus, in the following, we shall deal with the question how, or rather whether, Eratosthenes’ state of geographical knowledge can be reconciled with the shape of the chlamys identified above.

II

The reconstruction of the ‘map’ of Eratosthenes as it has appeared in the manuals of early geography since the nineteenth century (Fig. 5)²¹ offers a somewhat deceptive certainty, with its detailed course of the coast and its latitudinal and longitudinal lines.²² There is no doubt that Eratosthenes adopted Dicaearchus’ division of the oikoumene by a parallel through the Straits of Gibraltar and the Taurus range,²³ that he added a meridian through the Nile and the Borysthenes, which intersected the main parallel at Rhodes,²⁴ and that he calculated the east-west²⁵ and north-south²⁶ dimensions of the oikoumene from the distance between prominent landmarks. We also know that he drew Libya as a right-angled triangle to the west of the Nile,²⁷ that he localized the southern end of India nearly on the same latitude as the extreme south of Libya,²⁸ and that he assumed an almost north-south direction for the Indian east coast, thus giving the subcontinent the shape of a rhombus.²⁹

Fig. 5. Bunbury 1879 I, pl. X facing p. 650.

The northern coast of Europe might have been completed by a straight line between the two last known (or taken for known) points, following the same principle as for the south-western coast of Libya. The idea of an open Caspian Sea, established by Patrocles’ expedition, may have provided a further argument for a more or less straight course of the coast between the extreme north of Europe and the eastern end of the Imaos range, as the reconstruction shows.

On the basis of this - not very precise - data we now have to ask: what does such a pattern have in common with a chlamys? A first look shows that only the west of Eratosthenes’ oikoumene might be tolerably compatible with the chlamys-shape we determined from the description of the outline of Alexandria: a curved line from the extreme south to Brittany, a break of the coast grosso modo to the north-east up to an imagined point, from where the boundary of the continent had to run inevitably in an eastern direction.³⁰

However, this is where the correspondences end. The widest extent of the Rhodes meridian does not even approximately coincide with the middle of the oikoumene as could be expected from the symmetrical object of comparison.³¹ As for the two thirds east of the meridian, neither does the reconstruction of the unknown north nor the regions south of the Rhodes parallel agree in the least with the chlamys-shaped western portions: the indentation of the Indian Ocean, and the south-eastern stretch of India almost to the latitude of the Cinnamon country and beyond the eastern end of the Imaos range, stand in obvious contrast to Libya in the west.

One does not gain much by rotating the image of the chlamys slightly clockwise against the Eratosthenic system of axes (i.e., rotating the figure slightly anticlockwise) and taking the Indian east coast as a second narrow side of the chlamys, with Brittany and the cape of India as extremities: the respective realities of Libya and the Indian Ocean are even less reconcilable. Moreover, given the importance of the bipartition of the oikoumene by the Rhodes parallel,³² it seems rather unlikely that Eratosthenes disre-garded completely his system of axes when talking about the form of the oikoumene.

Yet, nothing compels us a priori to believe that the Greeks oriented every geographical concept to the north, as we do.³³ As Berger suggested, one could fit the oikoumene with the south ‘upwards’ into a chlamys with circular bottom,³⁴ assuming a circular north coast and equating the hypotenuse of right-angled Libya with one of the front cuts of Tarbell’s second diagram (Fig. 4). In my doctoral thesis on the Greeks’ ideas of Libya I adopted this interpretation (Fig. 6), albeit without being completely convinced of it: first of all, we have seen above that Tarbell’s second sketch does not fit with the chlamys-shape applied to the outline of Alexandria. Moreover, there remain the same problems as with the northerly-oriented chlamys: the gross asymmetry of the regions west and east of the widst extent and an India which drops completely out of the picture. Finally, right before the reference to the chlamys-shape, Eratosthenes compares the northern hemisphere from the equator to the polar circle with a σπ?νδυλος, the head of a kind of artichoke (κιν?ρα),³⁵ and it is hard to accept that with two directly neighbouring metaphors the author oriented one to the north, the other to the south.

Fig. 6. Zimmermann 1999, 122 fig. 20.

So, back to the northerly-oriented chlamys in accordance with Fig. 3: is it probable that Eratosthenes adopted (if indeed it already existed) the comparison of Alexandria with a chlamys in order to illustrate his idea of the oikoumene, focusing only upon the west and disregarding generously the obvious divergences in the east? Thus, we have arrived at the third and final point: what could Eratosthenes have had in mind with such a comparison?

III

An image that is at best valid for half of the evidence one wishes to explain is confusing rather than illuminating. Descriptions by means of simple geometrical figures - rectangle, triangle, rhomboid etc. - are sufficiently general to function even despite major divergences. The reader understands by abstraction what an author wants to express comparing, e.g., Italy to a triangle.³⁶ However, the more specific and distinctive the shape of the object of comparison, the more confusing and unhelpful it is in the points at which it varies from the object described. Thus, Eratosthenes’ attempt to explain to his readers the completely unknown northern and the totally asymmetrical southern shape of the oikoumene with reference to a chlamys does not really make sense. In fact, a look at the context of the Eratosthenes-passage may suggest another interpretation of the term χλαμυδοειδ?ς which would be of considerable importance for our understanding of the Cyrenean’s geography. In the fragment of Eratosthenes, the section immediately before the comparison with the chlamys reads as follows:

So let us presuppose that the island lies in the aforesaid quadrilateral. We must then take as its size the figure that is obvious to our senses, which is obtained by subtracting our hemisphere from the entire size of the earth, then from this area its half, and in turn from this half the quadrilateral in which we say the inhabited world lies; and it is by an analogous process that we must form our conception of the shape of the island, accommodating the manifest shape to our hypotheses. But since the segment of the northern hemisphere that lies between the equator and the circle drawn parallel to it next to the pole is like an artichoke in shape, and since the circle that passes through the pole, by bisecting the northern hemisphere, also cuts the artichoke in two and thus forms the quadrilateral, it will be clear that the quadrilateral in which the Atlantic Sea lies is half of the artichoke’s surface. The inhabited world is a chlamys-shaped island in this, being smaller in size than half of the quadrilateral.

As the whole passage and in particular the italicized sentence indicate, Eratosthenes - whose special interest in the geography of the globe is sufficiently proven by his measuring of its circumference³⁷ - did not care about the exact outline of the oikoumene in the text referred to by Strabo, but about its general shape according to its position on the northern hemisphere of the earth. With this intention in mind, we have to look at the real chlamys-shape (Fig. 3) once again. We only have to change the straight line AB into the curved, dotted line and the pattern, projected on the three-dimensional surface of a cone or globe, would cover perfectly what Eratosthenes is calling ‘half an artichoke’.³⁸ Or, vice versa, the quadrilateral between equator, polar circle and meridian containing the oikoumene would roughly assume, in a two-dimensional projection, the shape of a spread-out chlamys as we reconstructed it. The resemblance is too evident to be merely casual. Actually, the comparison of the oikoumene to a chlamys seems to be based on the obvious chlamys-shape of the quadrilateral circumscribing it. One only wonders why Eratosthenes applied this image to the oikoumene itself, considered as an island (ν?σος) with, as we have seen, a rather irregular physical shape.³⁹ The tertium comparationis of both — chlamys and oikoumene — must, in fact, have been something other than a strikingly similar outline. There remains only one aspect linking the chlamys to the oikoumene as well as to its quadrilateral: the way in which the inhabited world, like the garment, had to be imagined on a three-dimensional body. If this is what Eratosthenes meant by χλαμυδοειδ?ς,⁴⁰ irregularities of the coast lines would be of no relevance. The aim of his comparison was to explain to the reader that in reality the oikoumene was not a flat surface but a curved one, located on the northern hemisphere of the globe like a chlamys put around the shoulder of its wearer (Fig. 7).⁴¹

Fig. 7. Eratosthenes’ chlamys-shaped world.

One of Eratosthenes’ main purposes in his Geography was to update older maps of the earth on the basis of new scientific and empirical evidence.⁴² As Heidel put it, ‘we know that Eratosthenes made a map, and we are sure that it was the first map in which definite cognizance was taken of the sphericity of the earth. Just how did his epoch-making work affect the picture of the earth as it had been depicted by his predecessors?’⁴³ Yet, the comparison to a chlamys seems to indicate that Eratosthenes realized the defectiveness of a flat projection using a straight line, like Dicaearchus’ diaphragma, intersected by straight meridians parallel to each other. It cannot be excluded that the geographer drew the conclusions of his doctrine designing his map ‘in the shape of a chlamys, i.e., based on a curved main parallel, with its vertical ‘seals’ (σφραγ?δες) gradually converging to the north. The total lack of references in later sources is the essential shortcoming of such an hypothesis. It might be more probable that the famous map still followed the same straight parallel(s) as earlier specimens and that it was by the comparison to a chlamys Eratosthenes tried to call his readers’ attention to the problematic nature of flat projection not yet overcome at his time.⁴⁴

Once more we regret not having additional and more precise information from Strabo, whose own commentary on the ο?κουμ?ν? χλαμυδοειδ?ς may at best be regarded as an example of his helplessness in the face of the Cyrenaean’s theories:⁴⁵

Its shape (sc. that of the oikoumene) is described as roughly similar to that of a chlamys; for we discover a considerable contraction in its width at its extremities, and particularly at its western extremities.

In other words: for the completely different east this reasoning does not work. This fact must have been as unavoidable to Strabo as it has been to us in the above examination. Still on another occasion Strabo explains the comparison with reference to the tapering of the extremities.⁴⁶ It is significant, however, that the same detail is used by Agathemerus as argument for the Posidonian οικουμεν? σφενδονοειδ?ς.⁴⁷ Apparently, already in antiquity there was some confusion in understanding the geographers’ metaphors. If actually the tapering of the extremities depends on an Eratosthenic statement, it may have served either to illustrate the shape of the spread-out chlamys (cf. dextra laevaque anguloso procursu) or to explain the fact that the island itself was ‘smaller in size than half of the quadrilateral’. Strabo himself, seeking for the meaning of χλαμυδοειδ?ς, seems to have introduced the narrowing of the eastern and western portions of the oikoumene as a reason for its chlamys-shape, though without meeting Eratosthenes’ concern.

Against the understanding of χλαμυδοειδ?ς proposed above there are, prima facie, two objections to be raised, which I will discuss briefly:

1. Eratosthenes’ well-known tendency for two-dimensional objects of comparison (triangular Libya, rhombic India)⁴⁸ does not speak against the use of three-dimensional metaphors by the same author. The artichoke illustrating the globe segment between equator and polar circle is pretty three-dimensional and there is no doubt that the image traces back to Eratosthenes. There is, however, some difficulty in not supposing identical meanings for two examples of the same, otherwise unattested chlamys- metaphor, both born at Alexandria in early hellenistic times, i.e., hardly independent of each other. If in fact, as Pliny seems to suggest,⁴⁹ the comparison of the city’s outline to a chlamys (thus, a purely two-dimensional image) goes back to the foundation, one might hesitate to admit that an Alexandrian scholar, picking up the image about a century later, would have changed fundamentally its meaning. Yet, as Préaux has pointed out,⁵⁰ in the sources based on material from Alexander’s time, notably in Arrian’s report on the foundation of Alexandria, there is no hint at the chlamys-shape. We may therefore suppose that either a later hellenistic writer on the history of Alexandria or Strabo himself created the image of the chlamys-shaped city, having heard about Eratosthenes’ chlamys-shaped world without catching its proper sense. A desire to parallel, by means of this metaphor, the microcosmos of the metropolis to the macrocosmos of the whole oikoumene, is not unlikely to have played a certain role therein.

2. My hypothesis substantially depends on the assumption that Erato-sthenes, after all we can say about his knowledge of the oikoumene, may hardly have tried to explain its physical outline on a two-dimensional map by comparing it to a spread-out chlamys. Of course, images do not abso-lutely correspond at 100 per cent to their author’s reality, simplification and abstraction being essential parts of illustration. The perfect inequality of the south-east and the south-west with elements as marked as rhombic India and triangular Libya seems to exclude, however, that by mere simplification Eratosthenes could get to the idea of a chlamys, hoping that his readers also did. But how about Posidonius’ sling-shaped oikoumene which seems clearly to be based on the rhombic shape of the weapon’s middle section, holding the projectile until the throw? Does not this metaphor also use a symmetrical object of comparison incompatible with the south-eastern extension of the Indian subcontinent (Fig. 8), thus weakening the above argument against the traditional interpretation of the chlamys-shape? It does, in fact, if we reduce the adjective σφενδονοειδ?ς to a merely twodimensional image. Again, do we necessarily have to understand Posidonius like this? We may ask ourselves why — instead of saying simply ‘rhombic’ - the Stoic, interested as was Eratosthenes in the geography of the globe,⁵¹ likewise chose an object significant above all for its threedimensional use. Did he think about an oikoumene ‘wrapping’ the globe like a sling its projectile (Fig. 9)?⁵² In this case too, a three-dimensional understanding would render somewhat more plausible the choice of the highly original object of comparison, subordinating at the same time the divergence in the south-east. Yet, the works of both authors being lost, the three-dimensionality neither of Posidonius’ nor of Eratosthenes’ metaphor can be definitively proven. It has to remain an hypothesis which derives its attractiveness essentially from the fact that other conceivable explanations do not satisfy at all.

Fig. 8. Zimmermann 1999, 124 fig. 22.

Fig. 9. Posidonius’ sling-shaped world.

The relation between oikoumene and globe is identical with the one between the chlamys and the body of its wearer: if the interpretation proposed above is correct, the image of the chlamys should be seen as an attempt by Eratosthenes to establish a link between both objects of his geographical efforts -geography of the globe and cartography of the oikoumene. Behind that metaphor, we perceive the author’s awareness of the difficulty of reproducing a spherical body in a two-dimensional projection, of rendering imaginable the special quality of a curved surface to the reader of a book as well as to the viewer of a map. Strabo himself focuses on this problem a little later on, presenting as an ideal of cartographical reproduction the globe according to Crates, and as a suitable expedient the flat map with parallel latitudes and longitudes⁵³ As everybody knows, only Ptolemy was to resolve the problem some 200 years later with his cone and spherical projection. It is all the more noteworthy how close Eratosthenes, with his metaphor, had already come to this form of representation which is still used today.

Acknowledgements

For comments and suggestions I am indebted to W. Ameling (Jena) and K. Geus (Bamberg) as well as to W. Huβ (Bamberg) who gave me the opportunity to present a first version of this paper at the colloquium ‘Zur Geschichte und Kultur des Hellenismus’ in memoriam H. Bengtson (Bamberg, 22—24 June 2000). M. Hilgert (Jena) kindly assisted me in preparing the English text.

Notes

¹ Fr. II B 27 Berger = Strabo 2.5.6 C 113; cf. the adoption of the comparison at Strabo 2.5.9 C 116, 2.5.14 C 118, 2.5.18 C 122, 11.11.7 C 519; further, Macrobius, Commentarii in somnium Scipionis 2.9.8: denique veteres omnem habitabilem nostram extentae chlamydi similem esse dixerunt.

² A specimen dating about 800 bc and coming from Lahun in Egypt (Flinders Petrie 1917, 36 with pl. LI no. V 14) shows a lozenge-shaped middle section for holding the projectile which corresponds exactly to the description of the literary sources: ?οσειδωνιος δε ο Στωικος σφενδονοειδ? και μεσοπλατον απο ν?του εις βορραν, στεν?ν προς εω και δυσιν (Posidonius fr. 200 a Edelstein, Kidd = Agathemerus 2, Geographi Graeci Minores [ed. C. Müller, Paris 1855—61, hereafter GGM] II, 471) — the addition τα προς ευρον δ’ ομως πλατ?τερα <τα> προς τ?ν ?νδικ?ν notes the obvious divergence from the object of comparison, the Indian subcontinent in the south-east reaching the latitude of East Africa; even more distinct is Dionysius’ comparison with two opposite cones mentioned by Eustathius immediately after the sling-shape in Posidonius (Eustathius, Commentarii in Dionysium periegeten 1, GGM II, 217 = Posidonius fr. 201 Edelstein, Kidd); see Zimmermann 1999, 123—4 with fig. 22 (= Fig. 8 below).

³ Tarbell 1906, 283.

⁴ For details see Heuzey 1922, 115—38; Bieber 1928, 69—72; 1967, 29—30, 32; Pekridou-Gorecki 1989, 88—9 with fig. 63; Losfeld 1991, 176—81.

⁵ Tarbell 1906, 285.

⁶ Heuzey 1922, 140; Bieber 1928, 69 with pl. XXXV fig. 1 (Ephebe from Tralles); 1967, pl. 32; Préaux 1968, 182.

⁷ Further Diodorus 17.52.3; Eustathius, Commentarii in Dionysium periegeten 157, GGM II, 245; Scholia in Aratum vetera 236 p. 192 Martin. Besides Tarbell 1906, 285—6, cf. also Berger 1903, 405; Bernand 1966, 51—2; Préaux 1968, 176—84; Fraser 1972 II, 26—7 n. 64; Pekridou-Gorecki 1989, 135; Losfeld 1991, 182.

⁸ See Tarbell 1906, 286.

⁹ Heuzey 1922, 140: ‘.le mot lacinia étant un terme spécial, réservé à la toge, pour désigner les deux pointes formées par la rencontre du bord rectiligne avec la courbe extérieure’.

¹⁰ See Glare 1982, 994 s.v.

¹¹ In this sense, however, Préaux 1968, 177, 181—2.

¹² See the first reconstruction of Tarbell 1906, 284 fig. 1; similarly Aujac, Harley, Woodward 1987, 156 fig. 9.5 (after the commentary of Jones 1917—32 I, 435 n. 3 [ad 2.5.6]).

¹³ Tarbell 1906, 285 n. 1.

¹⁴ Préaux 1968, 181, supposing here, as in Pliny, a circle-segment directly meeting the coast (see n. 11 above), considers the ευθειαι βασεις as sections of the coastline converging from the extremities (the angles with the εντος περιφερεια) to a point in the middle of the eastern harbour. One wonders, however, why an almost straight coastline should be artificially divided in two. If, in fact, the outline of Alexandria had been just some kind of semicircle, neither Pliny nor Plutarch would have had to describe it in such an intricate way.

¹⁵ In this sense, Bernand 1966, 51 (‘une pièce d’étoffe rectangulaire ayant trois côtés droits et le quatrième arrondi aux angles’) who rightly emphasizes that it is a spread-out chlamys Plutarch is talking about (52 and again 1995, 59).

¹⁶ See Heuzey 1922, 139; Bieber 1928, 69—70; 1967, 35.

¹⁷ See Tarbell 1906, 284 fig. 2.

¹⁸ Cf., e.g., Hoepfner, Schwandner 1994, fig. 225 (facing p. 238).

¹⁹ See Berger 1880, 219 (referring to von Humboldt 1836—52 I, 124, 145—6); Thomson 1948, 163; Losfeld 1991, 181; hypercritically, Thalamas 1921b, 176—8 deletes this fragment as well as the rest of Berger’s section II B.

²⁰ The fact that Strabo himself is not the originator of the comparison already follows from the phrase λ?γεται δε και χλαμυδοειδες πως το σχ?μα a little later on (2.5.9 C 116). Nevertheless, the comparison has been attributed either to Strabo himself or to Strabo’s age by Bunbury 1879 II, 229; Tarbell 1906, 286; Aujac 1966, 201 with n. 1; Préaux 1968, 182; Dilke 1985, 64; Aujac, Harley, Woodward 1987, 156; there is no mention of the chlamys-shape in Heidel’s chapter on Eratosthenes (1937, 122—8).

²¹ This is Bunbury’s (1879 I, pl. X facing p. 650) version, repeatedly copied up to the present day (e.g., Olshausen 1991, map 4).

²² See, e.g., Thalamas 1921a, 212—14; 1921b, 163—7; Aujac, Harley, Woodward Klaus Zimmermann 1987, 157.

²³ Dicaearchus fr. 110 Wehrli = Agathemerus 5, GGM II, 472; Eratosthenes fr. III A 2 Berger = Strabo 2.1.1 C 67—8.

²⁴ Eratosthenes fr. II C 2 Berger = Strabo 1.4.2 C 62—3.

²⁵ Eratosthenes fr. II C 18 Berger = Strabo 1.4.5 C 64.

²⁶ Along the meridian through Rhodes (see n. 24 above).

²⁷ Strabo 17.3.1 C 825; for Eratosthenes’ authorship see Zimmermann 1999, 120—1.

²⁸ Eratosthenes fr. III A 2 Berger = Strabo 2.1.2 C 68.

²⁹ Eratosthenes fr. III B 5 = Strabo 2.1.22 C 78; Eratosthenes fr. III B 7 = Strabo 2.1.31 C 84; Eratosthenes fr. III B 11 = Strabo 2.1.34 C 87.

³⁰ For an equation of the hypothetical northern coast with the collar of the chlamys, cf. Berger 1880, 219—20 (referring to Mannert 1829, 89, 116).

³¹ As a consequence of Alexander’s campaign Eratosthenes abandoned the old idea of the symmetry of the oikoumene already put forward by Anaximander; see Olshausen 1991, 94.

³² Eratosthenes fr. III A 2 Berger = Strabo 2.1.1 C 67: εν δε τω τριτω των Γεωγραφικων καθισταμενος τον τ?ς οικουμεν?ς πινακα γραμμ? τινι διαιρει διχα απο δυσεως επ’ ανατολ?ν παραλλ?λω τ? ισ?μεριν? γραμμ?.

³³ See Podosinov 1992, 66; 1993, 34.

³⁴ Berger 1880, 220; 1903, 406.

³⁵ See Liddell-Scott-Jones 1996, 951—2 s.v. κιναρα; Suppl. 177 s.v. κιναρας; cf. Edictum Diocletiani 6.2 Lauffer: σφονδυλοι κιναρων.

³⁶ Polybius 2.14.4—12.

³⁷ See, e.g., Aujac, Harley, Woodward 1987, 154—5; more recently Geus 2000, 77—82; 2002, 225—38.

³⁸ Yet Berger (1903, 406) noticed the resemblance of Plutarch’s description to Ptolemy’s first, conic projection (therefore see, e.g., Dilke 1985, 77—8).

³⁹ Clarke (1999, 212) emphasizes the difference between the quadrilateral and the inhabited world, lying within that quadrilateral.

⁴⁰ Cf. his creation of σφαιροειδ?ς to describe the shape of the whole earth (Thalamas 1921a, 105—6; 1921b, 161). Adjectives in -ειδ?ς may indeed have a rather figurative (‘in the way of ...’) sense, as the use of σωματοειδ?ς (Polybius 1.3.3—4) for the history being a coherent whole since 218 BC(see Clarke 1999, 119) shows.

⁴¹ E.U. readers are invited to examine the obverse of the one- to five-eurocent- pieces, which shows a rather similar pattern.

⁴² Strabo 2.1.2 C 68: διορθωσαι τον αρχαιον γεωγραφικον πινακα; see, e.g., Bunbury 1879 I, 619 n. 2; Heidel 1937, 122; Olshausen 1991, 93—4.

⁴³ Heidel 1937, 125.

⁴⁴ Cf., however, Thalamas 1921a, 4: ‘Il (sc. Eratosthène) s’en est tenu à une géométrie générale de la sphère, sans aborder même le problème des projections.’

⁴⁵ For immediate use of Eratosthenes by Strabo, see Thalamas 1921a, 189; 1921b, 126—7.

⁴⁶ 2.5.14 C 119.

⁴⁷ Posidonius fr. 200 a Edelstein, Kidd = Agathemerus 2, GGM II, 471; see Eratosthenes’ chlamys-shaped world: a misunderstood metaphor Berger 1880, 220.

⁴⁸ Cf. already Pseudo-Scymnus 112—14, GGM I, 198 (see Clarke 1999, 63, 103).

⁴⁹ Naturalis Historia 5.62: metatus est eam Dinochares.. .adeffigiem Macedonicae chlamydis etc.

⁵⁰ Préaux 1968, 177—8.

⁵¹ For the probable origin of the metaphor in ?ερι ωκεανου, see Clarke 1999, 172—3.

⁵² It is perhaps worth mentioning that in one of his two measurements of the earth, Posidonius got the smallest circumference of 180,000 stadia known to Strabo (fr. 49 Edelstein, Kidd = Strabo 2.2.2 C 95; see the editors’ commentary on fr. 202, p. 722—3).

⁵³ Strabo 2.5.10 C 116—17; cf. 2.5.1 C 109.

Bibliography

Aujac, G.

1966	Strabon et la science de son temps, Paris.

Aujac, G., Harley, J.B., Woodward, D.

1987	‘The growth of an empirical cartography in hellenistic Greece’, in J.B. Harley, D. Woodward (eds.) The History of Cartography I, Chicago, London, 148—60.

Berger, H.

1880	Diegeographischen Fragmente des Eratosthenes, Leipzig.
1903	Geschichte der wissenschaftlichen Erdkunde der Griechen2, Leipzig. Bernand, A.
1966	Alexandrie la Grande, Paris.
1995	Alexandrie des Ptolémées, Paris.

Bieber, M.

1928	Griechische Kleidung, Berlin, Leipzig.
1967	Entwicklungsgeschichte der griechischen Tracht2, Berlin.

Bunbury, E.H.

1879	A History of Ancient Geography among the Greeks and Romans from the Earliest Ages till the Fall of the Roman Empire, 2 vols., London.

Clarke, K.

1999	Between Geography and History. Hellenistic Constructions of the Roman World, Oxford.

Dilke, O.A.W.

1985	Greek and Roman Maps, London.

Flinders Petrie, W.M.

1917	Tools and Weapons, London.

Fraser, P.M.

1972	Ptolemaic Alexandria, 2 vols., Oxford.

Geus, K.

2000	‘Eratosthenes’, in W. Hübner (ed.) Geographie und verwandte Wissen- schaften, Stuttgart, 75—92.
2002	Eratosthenes von Kyrene. Studien zur hellenistischen Kultur- und Wissen- schaftsgeschichte, Munich.

Glare, P.G.W.

1982	Oxford Latin Dictionary, Oxford.

Heidel, W.A.

1937	The Frame of the Ancient Greek Maps, New York.

Heuzey, L.

1922	Histoire du costume antique, Paris.

Hoepfner, W., Schwandner, E.-L.

1994	Haus und Stadt im klassischen Griechenland2, Munich, Berlin. von Humboldt, A.
1836	—52 Kritische Untersuchungen über die historische Entwickelung der geogra¬phischen Kenntnisse von der Neuen Welt und die Fortschritte der nautischen Astronomie in dem 15ten und 16ten Jahrhundert, 3 vols., Berlin.

Jones, H.L.

1917	—32 The Geography of Strabo, 8 vols., The Loeb Classical Library, London.

Liddell, H.G., Scott, R., Jones, H.S.

1996	A Greek-English Lexicon, with a revised supplement, Oxford.

Losfeld, G.

1991	Essai sur le costume grec, Paris.

Mannert, K.

1829	Einleitung in die Geographie der Alten und Darstellung ihrer vorzüglichen Systeme, Leipzig.

Olshausen, E.

1991	Einführung in die historische Geographie der Alten Welt, Darmstadt. Pekridou-Gorecki, A.
1989	Mode im antiken Griechenland, Munich.

Podosinov, A.V

1992	‘OpHeHTacHq flpeBHHx KapT (c flpeBHefimHx BpeMeH do paHHero cpeflHeBeKOBbq)’, VDI 203, 64—74.
1993	‘Die Orientierung der alten Karten von den altesten Zeiten bis zum frühen Mittelalter’, Cartographica Helvetica 7, 33—43.

Préaux, C.

1968	‘Alexandrie et la chlamyde’, CE 43, 176—87.

Tarbell, F.B.

1906	‘The form of the chlamys’, CPh 1, 283—9.

Thalamas, A.

1921a	La géographie d’Eratosthène, Versailles.
1921b	Etude bibliographique de la géographie d’Eratosthène, Versailles. Thomson, J.O.
1948	History of Ancient Geography, Cambridge.

Zimmermann, K.

1999	Libyen. Das Land südlich des Mittelmeers im Weltbild der Griechen, Munich.