MAPS AND THE HISTORICAL TOPOGRAPHY OF BEIRUT

MICHAEL F. DAVIE


Scanned from: Berytus, vol. 35, 1987, pp.141-164. by Elie Wardini.


The illustrations belonging to the article are:



ABSTRACT

Five maps of Beirut, draughted in the mid- 19th century were used to help explain the initial Bronze Age choice for urban settlement, then to discover the Hellenistic and Roman grid street-pattern which survived in the disposition of the modern souks. The maps were also used to propose possible sites for the hyppodrome of the city, and to locate the mideaval defences of Beirut.

This article examines how maps of Beirut, draughted in the early - and mid-19th century, can be used in the fields of urban historical geography, historical topography and archaeology of the city. Maps are, by definition, conventional renderings of phenomena observed at a precise geographical point, at a given time in History. They are thus essentially historical documents, although unfortunately rather infrequently used by historians. Maps of Beirut and its environs illustrate, albeit schematically, the urban physiognomy at the date of the survey and, if correctly interpreted, will inform researchers on the local topography of the city, its economic and spatial organization as well as documenting or even locating now-disappeared sites of historical or archaeological interest. They can also be used to understand the choice of the site for the original town. Several maps, produced by the Military during the first half of the 19th century corroborate information obtained much later by archaeological discoveries. Descriptions of the town, written by visitors, pilgrims or consular agents also help in identifying elements visible on the maps. As they were draughted before the modern in-filling of the port area and the destruction of the defensive walls, they illustrate the town as it stood since the Middle Ages.

The article will firstly describe the documents that were used, then present a description of the local topography which conditioned the choice of the site for urban settlement. The second part will present the methodology applied to discover the Hellenistic-Roman grid-system which survived through time in the pattern of the souks. This will shed some light on the urban physiognomy of the Roman colony of Berytus. The maps will locate and identify the defences of the town: the walls, castle and forts which maximized the advantages of the local topography. It will also reflect on the possible location of the hippodrome, the temple of Venus and the main cemetery. Thirdly, the maps will be used to propose a plan of the Crusader castle of the city, and to suggest the location of other edifices, one of which may have become Fakhr ad-Dín's palace. Medieval sources will be used to help in the identification of these sites.



THE MAPS



Two distinct groups of documents were used for this study. Those drawn between 1831 and 1840, are sketches perhaps produced from memory, on which some details of military interest were added. Those drawn in 1841-1842 are real topographical maps, containing all the information required for military operations in the area and for policing an occupied town. These maps were surveyed by military engineers, with state of the art instruments: in the 10 years between these dates, the whole of the Levant had suddenly become the theatre of intense rivalry between the European powers. Western military, political or economic interests in the Near East were no longer content with vague sketches of strategically important areas.

1. The pre-1840 maps

Two maps, well-known by historians of 19th-century Lebanon, are used:

Beirut, the antient Berytus, by H .A. Ormsby, 1831. Published by the British Admiralty, 1839. British Library number SEC.5.(1243); reproduced in Du Buisson 1921 and Jidejian 1973.

Plan of the town & harbour of Beirout, ancient Berytus. Published by J. Wyld, October 5, 1840. British Library number 48940.2; reproduced in Chevallier 1972.

2. The post-1840 maps.

Three maps were discovered at the Public Record Office in Kew (Great Britain).

Plan of the town and defences of Beyrut and its vicinity, by C.F. Skyring, Royal Engineers. P.R.O. number WO. 78 1000 (9).

--Plan of the envirions of Beyrout. Scale 3 inches to 1 mile, by Major Rochfort Scott. P.R.O. number WO. 78 1000 (1).

--Plan of Beyrout, the ancient Berytus Scale i2 inches to 1 mile, by Major C. Rochfort Scott. P.R.O. number WO 78 1000 (2).

A detailed description of these hitherto unpublished maps has been presented in Davie, 1984, together with re-drawn maps of the areas.



Historical background

The mapping of Beirut during the decade between 1830 and 1840 is explained by the vital interest the area took on following Ibrahím Pasha's territorial expansion and the subsequent defeat of the Ottoman army. Ibrahím Pasha briefly occupied the Sinai, Palestine, Syria and Lebanon and then pushed into Anatolia (1831-32). This remodelling of the Ottoman Empire was too sudden and dangerous for the European powers who had other plans for the area, in which they would have had the dominant role. This situation, which was considered by them to be intolerable, led to direct military intervention in the Levant by the European powers. An Allied fleet, composed of British and Austrian ships was summoned to Beirut, and troops disembarked at Jounieh. The city was bombarded from the sea on orders from Commodore Sir Charles Napier on September 11th, 1840, and it was subsequently occupied after the defeat of the Egyptian troops posted along the coast.

The political and military aspects of the Allied intervention are well documented, as are the Eastern Question and Ibrahím PAsha's Syrian Campaign and its influence on Lebanon. Reference should be made here to the studies by M.S. Anderson 1966. The Eastern Question I774-I92~; P.M. Holt 1966. Egypt and the Fertile Crescent, 1576-1927; K.S. Salibi 1965. The Modern History of Lebanon; D. Chevallier 1971. La societe du MontLiban a l'epoque de la revolution indistruelle en Europe, M.E. Yapp 1987. The making of the Modern Near East, 1972-1923; M. Jouplain 1961. La question du Liban. However, the history of the British or Allied presence in the town is less well-known. During their brief stay in the area, the Royal Engineers were able to survey and draught maps covering the whole of the Beirut peninsula, as well as detailled plans of the town and its immediate environs. Previous maps were rendered quite obsolete and useless. The maps are thus a record of the town and the surrounding country-side in the early 19th-century. Our hypothesis is that the town had changed little with no major expeditions being launched against it, and no large-scale destruction (human or natural), since the 13th century, with the end of the Crusader presence. The topography of Beirut peninsula did not change either. The definitive appearance of the city took shape towards the end of the 17th century. Changes linked to phases of local or regional economic expansion or decline produced minor modifications in the urban structure, as elsewhere in the Middle East (Abdel Nour 1982; Raymond 1985).

The maps can thus help locate sites, or at least their remnants, dating from the 11th century. The disposition of the souks give information regarding the earlier Roman and possibly Hellenistic urban structure. Normally, the internal structure of cities changes little through time, if they are not destroyed, because of topographical constraints and ease in re-using pre-existing elements. The history of Beirut does not indicate massive remodelling of the city during any of the post-roman eras. Rather, continuity seems to have been the dominant theme. The defences occupied the same general locations, the castle and forts are at the same places, the main roads leading to the city are immutable, as are the main cross-roads. These maps are thus historical source-material of prime importance, as they illustrate the urban physiognomy of the city before its rapid expansion towards the end of the 19th century before the demolition of its defences and the beginning of its urban sprawl, overrunning gardens and mulberry plantations which surrounded it. They also detail the pattern of the souks before the wide "modern" streets were cut through the town during the last days of the Ottoman presence and the first decade of the French Mandate.



Reliability of the maps

While the first two maps, dated 1831 and 1839 must be used with caution because of errors, the other three are documents that accurately reflect the situation on the ground. Maps 4 and 5 are especially valuable as they were surveyed under "normal" conditions, (i.e. after the end of military operations against the Egyptians). They pre-date the civil war of 1841 -42. However, map 3 seems to have been surveyed with considerable haste as important details in the town plan were omitted 'due to the route'. Compared to cadastral documents drawn during the first years of the French Mandate, the post-1840 maps show complete similarity in physiognomy of the streets, although some of the smaller impasses are not included. Overall, the reliability of the maps is excellent, and they can be used without correction. The pre-1840 maps are quite useless in this respect, and the streets 1. seem to have been drawn only to fill space inside the walls, equally imaginary in shape and size. They are however useful as they note the presence of ancient ruins (columns, etc.), on the outskirts of the city, south of the walls. They also tantalizingly hint at the location of the defences, before their repair or re-building, some time between 1773 and the early 1830s.



THE SITE

The maps were draughted by military engineers whose interests were essentially to prepare documents that could be used in the event of military action in the area. All the topographical details are marked: cliffs, hills, hillocks, stream incisions, vegetation (both natural and planted), important obstacles as well as all visible landmarks. The information can thus determine the reasons for initially choosing a site for settlement.



The Beirut Peninsula

One must examine its geographical context in order to understand why Beirut existed. Beirut peninsula occupies a complex site wedged between the base of the Mount Lebanon range, and the sea, which surrounds it on two of its three sides. This triangularshaped area is far from being uniform. Two hills (named the Asharafiyyeh and Ra's-Bayrut hills for convenience) occupy the northern part of the peninsula. The Ra's-Bayrut hill falls steeply into the sea at the Rawsheh cliffs, while the eastern edge of Ashrafiyyeh was cut by the Beirut river forming an escarpment at Siufe and Karm az-Zaytun. These two hills overlook low-lying areas to the North and South. To the South a more-or-less flat plain extends towards Burg al-Baragneh and Shuayfat. The soils are rich and fertile with mainly colluvial deposits washed down from the overlooking hills and mountain. However, to the South-West, sand dunes resulting from sediments brought to the shore by sea currents and blown inwards by the prevailing south-westerly winds have encroached on this plain. With vegetation unable to colonize the area, the dunes have slowly advanced, rendering agriculture precarious on the fringes of the fertile plain.

To the North, the two hills overlook a gently-sloped area limited by a westwardlystretching coastline, which, although generally straight, is also complex. To their West, the coastal plain narrows at 'Ayn al-Mrayseh then widens slighly between today's American University of Beirut sports-ground and the Bain Militaire, to disappear completely at Rawsheh. Low marine-erosion cliffs form the beach. The only natural harbour is at the 'Ain al-Mrayseh creek. To the East of the hills, the coastal plain disappears under an escarpment, a marine erosion cliff. However, above it, another marine erosion surface is very well developed and extends to today's Gemmayzeh, Rmayleh and Karantína areas. To the West of the two hills, this same surface constitutes part of the plain mentioned above, and includes the Mazra'a, Baribír and Harj-Museum areas. It also includes the Hamra, Sanaya'ah and Clemenceau sectors.

To the East of the peninsula, the Beirut river cuts diagonally across the plain after emerging from the mountains. Its incision is quite deep at H. azmiyyeh-Gisr al-Basha, but diminishes as it approaches the sea. Between the two hills, this marine terrace dips imperceptibly towards the North and South. It joins the lower marine terrace (the lower coastal plain) in the area of the Beirut Municipality and rue Weygand.



Topography and Movement

General topography explains some of the reasons which conditioned the choice of site. Movement along the Levantine coast was blocked by natural obstacles. The Mountain was avoided because of the difficulty of movement, while the coast was preferred because of its generally flat terrain. In the case of the Beirut peninsula, movement was limited to a very narrow corridor, between the mountain on one side and the sand dunes on the other. The latter blocked all movement along the coast, right up to the western hill, at today's Ramlet al-Bayda. One could not follow the beach in that general area because it is blocked by the Rawsheh cliffs. Access had to have been northwards or towards the North-East. However, the Beirut river, to the North-East, blocked movement along the axis of Shuwayfat-Hazmiyyeh-Gdaydeh with some fords passable but only during the summer, at low water. Bridges could of course have been built, but they were limited by the technology in use. Long-spanned stone bridges were out of the question, since they required pillars at short and regular intervals, resting on the alluvium of the river-bed. This was precisely the problem with the Beirut river, near its estuary, where it separated into several channels before reaching the sea. It is possible to imagine a passage along the banks of Beirut river. But if one was coming from the North, there would be little sense crossing at the Hazmiyyeh area, where the banks are steep and the river deep and dangerous. Today's Burg-Hammud and Nab'a areas were marshy and seasonally flooded. Coming from the South following the left bank of the river, one would have a very narrow passage between the river, with over-flowing banks in the winter, and the eastern escarpment of the Ashrafiyyeh hill making an uncomfortable and sometimes dangerous route. Quite naturally then, movement would have been northwards between the two hills, where the topography was relatively flat, well drained and without any obstacles or potential dangers. The hills are some distance away and relatively low. This axis is today's rue de Damas and rue Basta. These general topographical considerations must be combined with a detailed examination of Beirut peninsula's coast in order to understand the rationale for the site of the city.



Topography and Settlement

One must firstly examine whether the two hills could have been the site of an urban nucleus. The top of the Ashrafiyyeh and Ra's -Beirut hills are ideally suited for defence and communications: they command an uninterrupted view of the coastal plain and the sand dunes to the south, towards Shuwayfat, as well as the entire coastline northwards to 'Amshit. They also command the pass between the hills as well as the river fords and bridge. Was there an important city in the bronze age ? Scarcity of water may have been one factor limiting settlement. .

There are no springs as the hills are of marl (Ashrafiyyeh) or limestone (Ra's-Beirut). Water could only have come from rain, collected from roofs and stored in cisterns or from ground water tapped from wells. The geological structure of these areas did not permit this last solution in antiquity.

Another reason is that the soils are not particularly fertile on the hills. Ra's-Beirut i covered by sandy soil and sand dunes fixed by vegetation and pedogenized, which are very porous, tending to dry up in summer. The summits of the hills, whether Ashrafiyyeh or Ra's-Beirut, are small, and rather steeply sloped on all sides, a problem for agriculture. Farmers would have had to descend further down to the flatter and richer areas, the Hamra-Qantari-Basta-Sodeco-G'aytawi marine terrace, to the North or to the BArbir-Museum terrace to the South, abandoning the protection of the site. The concept of a hill-top site was apparently abandoned (or perhaps not even considered) in favour of one, on the northern coastal plain, which was fertile, well-drained, flat and wide. Along with these advantages there were no major disadvantages. One could exploit the sea as well as the land for food. The sea permitted some form of trading, while the road between the two hills channelled movement from- the South. The easy crossing of the river as well as the flat marine terrace to the East of the Peninsula (the KarantinaGemmayzeh area) channelled movement from the North. But the reasons for chosing the ideal part of the coast must be identified.



The northern Coast-Line

The West coast of the peninsula (the Khaldeh and Ramlet al-Bayda sector) is unfavourable for urban settlement because it is exposed to the prevailing south-westerly winds and swell; the waves hit the shore with their full force, as they are parallel to the coast. The currents are northerly and rapid. They are the cause of both extensive shore erosion and sedimentation (Sanlaville 1977). Any port or jetty would be filled-in very quickly, especially as the sea is quite shallow here. The isthmus at Tyre is the result of the same erosional-depositional processes triggered by Alexander's building of the mole. Elsewhere, the shape of the promontory deflects the currents and creates dangerous eddies, especially near Ramlet al-Bayda. Another reason for avoiding this sector, is the obstacle of sand-dunes.

Only the northern part is favourable for urban settlement, on the lower coastal plain. However, several sites could have been chosen, such as the 'Ain al-Mrayseh creek, with abundant water (thus its name), and space for urban construction. However, the Ra'sBayrut hill immediately behind it (the French Embassy compound, Bliss Street and AUB's Main gate) render its defence problematic. Walls would have had to be built far from the creek, with observation towers scattered at regular intervals to monitor movement along the coast and between the two hills.

A second possibility can now be examined. Between 'Ain al Mrayseh and Zaytuneh is a cove (the St-George's Hotel pleasure-port). The site was well protected from the waves, and all of the area between the Phoenicia, Martinez and Cadmos hotels could have been used for agriculture or building. The 1840 maps show that it was in use then as a fishing port. However, it did not develop earlier in history for the same reasons as in the 'Ain al Mrayseh site vulnerability. The Holiday Inn Hotel and rue Georges Picot areas overlook the site at close proximity.

Much further East, near Karantina (Quarantine), is another favourable site. At the base of the marine erosion cliff is an abundant fresh-water spring. However, the coastal plain is minuscule, and the only possible urban site would have been on the upper plateau surface (i.e. the general area of the Electricity building and the railway station). If we accept the idea that the populations which settled the coast of the Levant used the sea as much as the land, that they were fishermen as well as farmers, the Karantina site offered many disadvantages. Water was available, but it was at the base of a steep cliff. Access to the sea was also a problem, as all goods or sea-food had to be hauled up the cliff to the plateau. There was also no space for repairing their vessels. However, defence was easy: the cliff would interdict any offensive from the North side, while normal fortifications could defend the other sides. The site was also quite far from the Ashrafiyyeh hill, while being one of the most fertile sectors of the peninsula.

Finally, the sector between Zaytuneh and the Rivoli cinema must be examined. A study of the 1840 maps show that it was a shallow bay open to the North. In the centre was a small island, while further East was a narrow elevated promontory which overlooks the island and the cove. Inland to the south-west is a spur of land which stretches down from the hill, ending at the western side of the bay. Here, the coastal plain is at sea-level, bordered only by a low marine erosion cliff. All the conditions for the site's defence are present: the cove between the promontory and the island could easily be protected; the former overlooks the sea on three sides, while the island could conceivably be used as a refuge for the population should the coast be occupied. A second line of defence could be exploited by using the spur, today's Grand Sérail hill. There is of course the problem of the water-supply. This problem was apparently solved very early in the history of the city, whose very name bears testimony to the presence of 'wells', be'eroth. But water was not impossible to find. In the area between the two hills is a source (at Ra's al-Nab'a) whose water ran down the slope to the cove. The cove(s) would have provided well-protected fishing harbours. But Beirut was also at the outlet of a passage between the two hills the only route along the coast. It was thus an obligatory staging-post, conveniently situated halfway between Sidon and Byblos, and less than a day's march from Shuwayfat to the South and SarbaKaslik to the North. The original site would thus have been at the foot of the promontory, perhaps opposite the small island. Perhaps some sort of fortification existed . on the promontory itself (the site of the mediaeval castle). However, the tell of Beirut is slightly inland and away from the port cove. This could indicate movement away from the shore at some date with perhaps two urban nuclei, an "upper city" near the promontory, and a "lower city" on the cove.



THE HELLENISTIC AND ROMAN CITY

Among the problems posed to urban historical geographers is the one of the survival through time of an ancient street pattern and its recognition in the modern city. Some of the roads, of course, are conditioned more by local topography than by history: fords, natural obstacles, marshy land etc. all determine the direction of movement. The influence of these conditions on cities like Paris, Lyon or London are well documented (see Rouleau 1975 and examples in Carter 1983). However, many street patterns are entirely man-made, with the chessboard pattern being the favoured solution for elementary urban planning. The Greeks, perhaps borrowing and improving on prototypes born in the Middle East, used it extensively in their colonies, the apoika:

The whole city was constructed in relation to a coherent and controlling plan. It was not the . result of gradual and unorganized extension, but was conceived as a complete whole. (...) The dominant feature of the rigid system according to which the city was laid out (was) the complete dominance of rectangularity. The public buildings and the main squares were set within this grid lisystem. The (whole) city was orientated generally along a North-South axis, and was laid out regardless of topographical conditions (Carter 1983 :21).

Numerous examples of colonies and cities laid out according to this system exist in the Middle East and adjacent areas (e.g. Miletus, Olynthos, Priene, Alexandria, Dura Europos). The invention of the chess-board disposition is attributed, erroneously it seems, to Hippodamus (1983:24). The Romans adapted the general idea, but stressed the importance of two main axes: the Decumanus and the Cardo, intersecting at the main cross-roads of the city. The main buildings of the city such as the basilica, forum and temples, were located along these two main roads, which intersected with the walls at four main gates. The eastern and western gates were aligned with a particular compass bearing, defined by the position of the rising sun on the day of the foundation of the colony (Mouterde 1966:24).

Berytus was founded as a Roman colony towards the year 10 AD (1966:23), in order to settle parts of the V Macedonia and VIII Augusta (or Gallica) Legions. This installation was built on the ruins of the Hellenistic city, which in turn was located on or near the site of a much earlier settlement, dating perhaps from the Neolithic period. Mouterde (1966:25) states: "as it happened when a city already existed on the site chosen to settle veterans, the Berytus colony must have occupied land adjacent to the already existing city." No doubt the pre-Roman city was located near the port, to the North. The alignement of columns there suggests (Lauffray 1948:59) that the town also extended westwards to the base of the hill, now occupied by the Grand Sérail, the suspected location of a temple (du Buisson 1923:123-124, Lauffray 1948: 5, note 2). The new city could thus have been built on the ruins of the town destroyed by Tryphon in 143 BC. The columns indicate that a grid-pattern already existed and that the Roman colony simply reproduced the previous urban pattern. This therefore raises the question of the mythical and ritual founding of the city, and the cosmological meaning of the position of the gates.



The surviving grid-pattern

Our hypothesis is that the Roman (and thus part of the Hellenistic) grid-pattern survived man-made or natural destructions, in a deteriorated form. One such destruction was the massive earthquake of 551 which flattened the city. The pattern is still visible on a map as portions of streets, more or less orientated parallel to a certain compass bearing, with other streets having a perpendicular bearing. This does not imply that all streets necessarily intersect: portions of streets can have an orthogonal bearing with other streets, without actually meeting.



Previous studies

Lauffray (1948:7) noted that some of the souks in Beirut were suspiciously ancient in plan. One of the main souks, suq al Haddadín, which starts at Bab ad-Darkeh, the main southern gate of the city, was quite straight, and ended close to the main mosque (Jami' al 'Omari), once the cathedral of St. John, built on Byzantine and perhaps Hellenistic foundations (Lauffray 1948). He concluded that the souk was superimposed on the Cardo of the city. The Decumanus was thus the souk starting at Bab as-Saraya heading westwards; in the 19th century it was widened to become rue Weygand. These two axes divided the city into four equally-sized quarters, with the main gates being Bab as-Saraya (the eastern gate), Bab ad-Darkeh (the southern gate), and Bab Idris (built around 1860, the western gate). The northern gate did not exist, as the Decumanus led to the port, protected by a fort built on an outlying island and perhaps another built on a spur overlooking the eastern part of the cove.

only the post-184() maps were used to discover the Hellenistic or Roman streets. It must be stressed that Lauffray (1948) inferred the same results from archaeological evidence and logical deduction. The methodology proposed here is only an illustration of how good maps can be used. Traces of an orthogonal road system dating to Roman times have been found in several near-eastern cities. Sauvaget (1934, 1941, 1949) was able to reconstruct the town plans of Laodicea (Lattakia), Damascus, Antioch and Aleppo. However, the problem in Beirut is not as simple as in Lattakia. Successive accumulations during the Arab and Ottoman periods partly affected the urban plan. Berytus was apparently not built according to one grid-system. The local topography as well as older urban vestiges probably forced the Roman town-planners to use grids of different orientations (Lauffray 1948:8).



Methods used and results

The methodology used consisted of measuring the heading of each street or portion of street contained inside the walls of Beirut, as defined by the map. Each street, identified by a number, thus has a value, in degrees, in relation to an arbitrary North heading. The information for the street elements was fed into a computer. After preliminary sorting, a list of streets intersecting at right angles was obtained, by one degree intervals. A bargraph of the frequency of the pairs of streets having the various orthogonal headings was produced, showing the predominance of 0-90deg., 2-920, 5-95deg. and 12-102deg. headings. A map of the streets falling into these categories was drawn (fig. ~). The first point is the surprising similarity with the results obtained by Lauffray (1948). A reconstruction of the physiognomy of the city inside the walls is thus possible. A detailed examination of the maps shows that the grid-system is not visible near the port area. This poses several problems:

1) Was the city square or rectangular ? In other words, was the city uniformly gridded or did it comprise more than one urban nucleus ?

2) Why is there no clear surviving grid-system to the North of the Decumanus ?

3) Can unknown monuments or constructions be identified inside or outside of the city based on the maps ?

Firstly, the shape of the town was conditioned by the disposition of its defences, which, in turn, depended on the topography of the site. Lauffray (1948), as well as the pre-1840 maps or the paintings by Montfort (Du Buisson 1921) indicate fallen columns outside the mediaeval walls, suggesting that the Roman city extended southwards. Perhaps the area of Zaytuneh, where traces of ruins were found (Mariti 1787: 52) suggest another extension, this time westwards. A temple dedicated to Venus and a Roman altar (Lauffray 1948:6 note 2) was located on the spur dominating the city, near where the Grand Sérail now stands. Columns and extensive ruins once covered the area (Buckingham 1825:448). Thus it seems that Roman (and perhaps Hellenistic) Beirut extended much farther than the mediaeval walled city. However, this does not imply that the walls of the Roman colony extended so far: the extensions were suburbs, i.e. 'under' the walls of the city.

One can pose the question whether there were any walls at all around Roman Beirut. However, the fact that the city was officially founded to settle veterans implies that the military authorities preferred a fortified city with a street-plan they understood and had imposed on many other sites. The grid system worked and there is no reason why Berytus should have been an exception, especially in an area already characterized by walled towns.

The walls, then, must have taken into account the local topography and contemporary military technology. This changed little between the Hellenistic period (or much earlier, for that matter) and the early 19th century. A city was an area to protect with walls, which an invader had to penetrate in order to conquer. Warfare was thus quite static, and effective defences had to maximize advantages of local topography. This implies that defences which were effective in Roman times remained so until the late Ottoman era. The walls depicted on the 1840 maps are thus more or less the ones in use during the Crusader period or earlier.

The defences of the town were dictated by several topographical features. To the West is the spur of the Grand Sérail, dominating the plain on which the town was built. The weakest point of the city was there. Whoever controlled the spur could monitor the entire city. Watch-towers or bastions would have been the rule all along the ridge of the spur. It might have been possible to build the defences much further to the West, integrating the hill into the walled city. However, this would have created a much larger town than the population warranted, and excessively long defensive walls. The city would also have been cut in two by high ground whose steep sides were quite useless for urban construction. The problem of defending the top of the spur would not have been solved either. The most economical military solution must have been either defending the base of the spur or defending the ridge. The differences in the shape of the walls on the 1830 and 1840 maps suggests that the defences that ran along the ridge of the spur were built some time between 1773 and the early 1830s by Gazzar Pasha, re-using pre-existing outlying forts. The walls at the base of the spur were thus built on mediaeval or RomanHellenistic foundations. In this case, the walls would have followed a northerly direction from Bab Ya'qub towards the harbour. The spur could have been fortified by a bastion (the Burg 'Umm Dabbus ?), which could have been in visual contact with other watchtowers, either inland (Burg Abi-Haydar, Burg Barajneh) or along the coast (Burg alKhodr, Tabarja etc.).

To the South of the city an uninterrupted extension was not possible. The mediaeval walls would have met the degraded marine erosion cliffs which limit the general area of Zoqaq al-Blat, Basta-Tahta, Ghalghul, Debbas and Tabaris. These areas dominate the lower plateau (a marine terrace) on which the city was built. The defences would have had to be above the cliff, i.e. along its edge, or well away from its base, out of range of missiles that could be launched against the walls. Lack of archaeological evidence or toponyms suggesting earlier ruins or walls argues against defenses above the cliff.

The eastern defences could have been anywhere, with size of the population dictating their extent, and the necessity of respecting a more-or-less square format for the city. The topography here is quite flat, with no major hills or other natural features to exploit for defence. A detailed examination of the walls to this side shows that it is the least preserved part. There are too many changes in direction, suggesting an uncoordinated effort to rebuild (or repair) the walls, which were probably ruinous or demolished when the maps were draughted. Houses were built on the walls themselves, and in some cases, the walls of the houses replaced the fortifications. The last repairs or reconstructions can not be dated with any degree of precision, although some portions of the walls were rebuilt by Fakhr ad-Din as his palace was located to the East of the city, inside the walls. Gazzar Pasha could also have been responsible for some of the re-building in this sector.

The main defence on this portion of the wall was the castle, built on the spur overlooking the cove. This flat-topped hill was limited on three sides by vertical cliffs falling down to the sea, while on the landward side, a partially filled-in moat separated it from the plain on which the town was built. It monitored the port area, as well as its approaches: St-George's Bay northwards to Dbayyeh, Nahr al-Kalb, and Ra's et.-Tayr, near Kaslik. The eastern wall, of which some portions are parallel to the Cardo, was straight and connected with the castle, but unlike the western defences it is no longer identifiable on our maps.

Thus the walls, shown on the 1840 maps indicate the location of the Roman (and Hellenistic) defences, without excluding the extramural suburbs extending inland or along the coast. Conceivably, the corner of the southern and eastern walls was fortified, perhaps by a tower built further inland. It is tempting to suggest that the tower was located at modern Burg al-Kashef, with foundations in the Roman era. This tower is exactly symmetrical to Burg 'Umm Dabbus in relation to the Bab ad-Darkeh gate, the main Cardo gate of the city. It would be surprising if the mediaeval builders in Beirut, Crusader, Arab or Turk did not reuse ancient foundations, as well as the stones from ruined edifices that littered the plain or hills. Saleh Ibn Yahya, writing in the 11th century, described the walls of Beirut noting reinforcements of ancient columns placed in the walls by Crusader builders.

Secondly, the fact that the grid-system does not extend to the North of the Decumanus is indicating several possible solutions. It never existed there, because of the older town around the port area. The Hellenistic and Roman towns were founded to the South of the site, perhaps occupying the plateau to the South of the slight escarpment identified by Sanlaville (1977) as being a marine cliff. There was a lower city, around the port, perhaps without any distinctive urban pattern, and an upper city, rigidly laid out. The limit between the two towns would have been the Decumanus, or the escarpment.This idea is similar to the one developed by Mouterde (1966: 25), who suggested that the two cities existed side-by-side. If the Hellenistic grid system did exist northwards, perhaps it was obliterated by the earthquake and subsequent tsunami in 551, or during the various battles which took place for control of the port and castle, from the Crusader era onwards. Perhaps the grid system extended to the sea only to the West of the small outlying island, towards Bab as-Santiyeh and Ras esh-Shamiyyeh. In this general area, the marine cliff is less pronounced, and roads leading down from the 'upper city' could reach the sea without any problem. The original town would have been small and concentrated around the useful part of the coast, the port cove between the island and the castle-spur. The newer Roman city would have occupied a much larger area and been more concerned with the defence of the urbs. The western walls extend in a straight line from the base of the Grand Sérail plateau down to another small cove in the vicinity of Bab as.-Santiyeh, surrounding the low-lying plain extending down to the sea and heading perhaps to a watch-tower or a secondary fort on Ras esh-Shamiyyeh. The view from there (occupied by the Bahri Restaurant in pre-1975 days), indicates strategic importance for the site in respect to the town's weakest point, its maritime approaches.

In any event, the city was not perfectly square. If the colony was in fact a square, limited to the North by the Decumanus, the 'real' Decumanus of this square would have been somewhere in the vicinity of the Greek-Orthodox cathedral of St. George, which Lauffray (1948) identifies as a possible site for the Berytus law-school. This solution fits with the idea advanced by Mouterde (1966: 25) that Berytus was built alongside the port city, of Hellenistic origin, which survived Tryphon's destructions. Yet, there are no traces of gates for this 'new' Decumanus To the West, a gate would have been opposite today's Grand Sérail, a difficult position: roads would have had to climb the slope of the spur or run alongside the walls once outside the gates, of which no traces have survived. These considerations indicate one of three solutions. The colony could have been rectangular, with the grid system extending to the port, though no trace of the system has survived in this area. It could have been rectangular with three of the quadrants gridded, and t-he last quadrant being unplanned, occupied by the indigenous population, and not the veterans. The colony could also have been a square, located to the South of a preexisting site, roughly to the North of the Decumanus. This colony would have occupied the plateau to the South of the slight escarpment, a marine erosion cliff, making an 'upper city', urbanized in a grid system, and a 'lower city', near the port, with another system, perhaps anarchical. This does not exclude the possibility that other gates existed in this rectangular ( ?) city. Perhaps a sea-road, described much later by Mariti (1787: 56) could have extended to the base of the castle entering the city through Bab as-Saraya. It is also possible that another gate (near Bab as.-Santiyeh) could have existed on the other side of the city, leading to Zaytuneh and 'Ain al-Mrayseh. Coincidentally, the Bab as-Saraya-Bab as.-Santiyeh axis is parallel to the Decumanus.



OTHER SITES

Just as the maps inform us on the urban topography of the city during the late Roman period, they also illustrate aspects of its ancient suburbs. Thus a space between the eastern wall and cultivated countryside, suspiciously clear of any vegetation or constructions, has survived to this day as the Place des Canons. A careful examination of the maps shows that its northern part is occupied by a Muslim burying-ground. The southern part stops quite suddenly at the Burg al-Kashef, with the middle section showing signs of slight elevations. These are quite distinct on the map and were apparently drawn because they once existed on the ground.

Two hypotheses can be advanced:

1) The slight elevations are perhaps the traces of a tell, distinct from the one further seawards, on which the cemetery was established. It could tentatively be dated from the Iron Age; a movement inland has been documented at other sites of this period previously lying directly on the coast.

2) Another solution could be that the elevations are the traces of the hippodrome. Mouterde (1966:40) suggested that the Beirut hippodrome was at Wadi Abu-Gamil, based on the shape of streets visible on aerial photographs of the area. It is our opinion that the site proposed by Mouterde is very unlikely, and that the oval shape visible on the aerial photographs is merely coincidental. Our maps show that the rue de l'Armee (the 'southern track' of the hypothetical hippodrome) did not exist in 1840, but only the northern track (i.e. the Wadi Abu Gamil), is visible as a path. Furthermore, this site is dominated, to the South, by the abrupt cliff of the Grand Sérail, and its base would have suffered from erosional-depositional processes, as well as periodic flooding during the rainy season. The slopes would also have been too steep to be used without extensive engineering work, of which no archaeological trace has survived. Our opinion is that the hippodrome was located to the East of the city, perhaps at the Place des Canons, and that the slight elevations could conceivably be the traces of the spina or of the tiers of seats. However, Lauffray (1948:6) notes that near the Petit Sérail, the underlying rock is only 1 m from the surface. The site is much simpler than the one of Wadi Abu Gamil and geomorphological problems do not exist to hamper construction or maintenance.

Furthermore, it is close to the main gate of the city and the main coast-road. The Wadi Abu Gamil site, albeit close to a possible gate (near the future Bab Idris), was less appealing, as the western roads headed to the general area of Qantari, characterized by advancing sand-dunes and sparse population. The 'Ain al-Mrayseh road, more to the North, is quite distant from Wadi Abu Gamil, and one must make a detour to reach it. The eastern road, on the contrary, would have conveyed far more people as part of the North-South road, via Burg al-Barajneh and Shuwayfat to Sidon. The hippodrome would thus have been visible to all those entering or leaving the colony, while the Wadi Abu Gamil site is hidden by the spur of the Grand Sérail. Since the slight elevations could also be the traces of the town's refuse-heaps, as is observable elsewhere (Raymond 1985 :149), only excavation can confirm these hypotheses.

Other possible sites do exist for the hippodrome. Our translation from the original Italian text by Mariti (1787:46f.) is worth presenting here.

(...) "We went to the Gemès. This is a place not far from the town, to the East, and is a place appropriate for the rustic amusements of the Beirutis. One can find cafes, and one can also be refreshed by iced drinks. In parts of this area, rare animals are exhibited (...), while elsewhere one can find dice-players. In another corner, one can dance and everywhere there are persons to entertain the visitors. All the area is turned over to the idle.

The disposition, as well as the location of the site are quite pleasant and I would be inclined to believe that the custom of coming here is quite ancient. Perhaps the amphitheatre was here or perhaps even one of the other public edifices with which Baruti was graced. (...) The place is called Gemès because of the tall trees which grow there. (...) They are worthy of admiration, because of their beauty and age. (...) We know these trees as sycamores, Ficus Sycomorus Linn."

This area became the red-light district of Beirut during the middle of the last century, and survived until its total destruction in 1975-83. Because of the apparent wellestablished tradition of the place for pleasure and entertainment, perhaps one should consider it as a possible site for the hippodrome, in agreement with Mariti. The general area would correspond to a rectangle limited by today's rue Gemmayzeh, rue Pasteur, the Ecole des Freres du Sacre-Coeur and the Place des Canons, being located further to the East than the previous site. However, the cafes, cabarets and bars, direct descendants of the Gemès pleasure-grounds, suggests the possibility of a Temple of Venus in the area (H. Salame-Sarkls: pers. comm), perhaps adjacent to the hippodrome.

The area, located between the coastal road (leading to the Decumanus ) and the rue Gemmayzeh (which leads to Bab ad-Darkeh and the Cardo) would also have included the main suburbs, where one would expect to find the polluting industries: tanneries, pottery-kilns and metal works. The general wind-direction (from the South-West) would have rid the atmosphere of smoke and odours. Being on the 'main' coastal road of the Levant, no doubt there were also stables, rest-houses, taverns and bordellos. These suburbs would also have included the cemeteries. Mariti (1787:49) mentions very ornate sarcophagi having been unearthed in the general sector of the North-East corner of today's Place des Canons, not far from the Muslim burying-ground of the maps. Curiously, no ancient cemetery seems to have been documented to the West of the city.



THE MEDIAEVAL CITY

Just as old maps shed light on the urban topography of late Roman period Beirut, they also clarify the city's physiognomy during the mediaeval period. As the town was neither systematically razed nor rebuilt from the 7th-century onwards, certain elements survived and must have been re-used by its various masters. Thus if the street lay-out did not evolve too drastically, still following the lines of the Hellenistic and Roman grid-pattern, then other sites may have survived, such as the castle and various forts.

The city walls were probably built on, or in close proximity to the Roman defences. As was stated earlier, warfare techniques did not evolve much between the Roman and Mediaeval periods, and local topography remained the determining factor to take into consideration when defending a position. Portions of the Crusader fortifications survived until the draughting of the maps and can be identified, for example, by the circular shape of a tower on the spur of the Grand Sérail, to the West of the town. Its shape and origin can be confirmed by the close examination of lithographs by Montfort (Du Buisson 1921), Flandrin and Toudouze. Perhaps the Burg 'Umm Dabbus and the Burg al-Kashef were also simply re-used by the Crusaders as outlying forts, on the foundations already in place. The function of Beirut did not change much either: it was still a main staging-post on the coastal Levant axis, as well as a port of some importance. Its defences would be more or less identical to those of the Roman period.

One can also locate the mediaeval port. Rey (1871) describes it as being defended by two towers "of Genovese origin", between which a chain could be drawn in order to close its entrance to marauding pirates and enemies. This port was located to the West of the Island Fort, protected by a jetty, the traces of "an ancient mole" mentioned on the maps. A detailed examination of the maps and of contemporary drawings show that the two towers of the island are built on rock foundations with no passage between them possible. The hypothesis of Rey has to be corrected in the light of Saleh. Ibn Yahya's description of the port: A chain was drawn between the castle and the main wharf, where perhaps a small tower was built, the Bab as-Sinsileh. The passage, "for small vessels only", no longer existed in the early 19th century, and a bridge of 4 arches joined the island to the town. But by that date, the ancient mole had been severely dismantled by the sea, and the port was open to the North. Fakhr ad-Din also contributed to its destruction by filling in the port, for fear of maritime raids and reprisals by the Ottomans (Chebli 1946 :141; Jouplain 1961 :1 15).

Thus the port could have consisted of two parts: the first port between the main landcastle and the island-fort, and a larger artificial port between this island and Ras eshShamiyyeh, from where the jetty was built. This second port would have been of Crusader origin, while the former would have been the original (Neolithic? Bronze Age ?) one. The Sea Fort to the West of the main land-castle was also built (or re-built) by the Crusaders. Paintings and etchings show distinctive Crusader architecture for its foundations and lower levels.

The Main Land Castle can also be examined on these maps. It was severely modified over the years by the successive occupiers of the city, and no doubt only few Crusader elements survived until the 19th century, and then only in foundations. Wilbrand of Oldenburg described the state of the castle in 1212 (text as translated into English by Boase 1967 :65 ): "On the lowest part it is fortified by the sea and precipitous rocks, on the other, by a ditch and walls (...) Two strong walls overlook the ditch, in which towers have been powerfully erected (...) the stones of which are linked by great iron bars. In one of the towers recently built (...) (was) a most ornate palace. It is well planned and sited, looking out on one side to the sea and the ships leaving the harbour, on the other on meadows, orchards and very pleasant places. In the midst of the palace is a well (...) and from it a fountain rises into the air..."

In 1422 Ghillebert de Lannoy described the same castle (our translation from the Old French, in Du Buisson 1923: 246):

"(...) On the other side of the port (...) is a small castle, built on a rock rising from the water (...) Towards the meadows, it is also built on a rock. There one can see two ditches, without water. However, towards the sea, this ditch ends under the cliffs of the castle. On this rock, there are two square forts, one nearer the sea than the other, and surrounded by walls".

The maps show in stunning detail the ditch and the walls. The latter are of course not original, and were built, rebuilt or repaired several times after the 15th century. An understanding of the logic behind the choice of the site for defensive purposes is possible. The castle occupies a small promontory protected on three sides by steep cliffs. The fourth side (the moat) is also steep. It was no doubt natural and the fact that there was no water in it suggests that it was carved out by wave-erosion before the last variation of the sea level.

The maps also permit confirmation of a hypothesis concerning the location of Chaufor or Mont Chafort. According to Crusader chronicles, the Castle of Beirut was captured after its moat had been occupied and its base mined. It had also been bombarded from a hill-top which dominated it. Rey (1883: 523) suggested that Mont Chafort was the Grand Sérail spur. Mesnil du Buisson (1921:242 note 1) thought that the hill was the Muslim cemetery, the tell of the city. An examination of the toponym indicates that it is of Arabic origin (shufra, shafra, meaning cliff edge, brow). This corresponds exactly with the physical aspect of the Muslim cemetery, which overlooks the sea, castle and moat, and whose northern side is quite abrupt. Another toponym is perhaps of Crusader origin: the Bab as.-Santiyyeh (or Sambatiyyeh) could derive from St. Etienne or from Santa; the latter would in turn come from Terra Santa (Daher 1985:63). Could it be an indication of a cemetery West of the town, which became the Muslim cemetery of Santiyyeh?



The Palace

This site, known as the Palace of Fakhr ad-Din, stunned many a Western visitor by its architecture and decor, and even more by its gardens, laid out and maintained by Florentine gardeners, especially brought to Beirut for the purpose. The size of the palace, as well as its general position in the town suggests that it was not built entirely by Fakhr ed-Din, but that he rebuilt or restored portions of a pre-existing structure.

Our translation of the text by Mariti (1787: 22) again provides a good description of the site:

"One can observe, to the North-West (sic) a palace that is said to have belonged to the Grand Emir Faccardino, because it is thought that it was built by him. This palace is large and covers an extensive area, but the building of such an edifice goes back rather earlier than the days of the Grand Emir. I would say that it is of Sarrasine origin. It was perhaps the palace of the Lords of Baruti when the Latin Christians were the masters of this country, or perhaps the residence of a cavalry corps, for reasons we will examine below.

One can be sure that Faccardino restored it at least in part, and remodelled it according to European tastes with the help of masters brought from Tuscany. They embellished the courtyard, which is in the middle of the construction. They paved the courtyard with grotesque motifs using multi-coloured marble.

In the middle of the courtyard there is a large pool and all around it are small walls which were decorated with vases. Certain pedestals which one can still see were the bases of statues found in the various ruins of the country. That is at least what is said here (...).

In the front of the courtyard one can observe an empty space, a sort of arch. It was a terrace, a sort of gallery used for relaxation or to seek the freshness of the Summer, as is the oriental habit. Above were the quarters of the women.

In this palace, one can observe underground bays, which resemble our wine cellars. One can find kitchens and cisterns, as well as stables for horses. In another part now live the Grand Emir's guards.

Next to the palace is a garden, laid out as those of Europe. Today, however, it is abandoned, but there are still a few lemon- and orange-trees, which form a small grove disposed with taste and symmetry (...).

This palace has two exits, one leading to the outside of the town, the other leading into it. This, at least, is its current state. Towards the exterior, and facing the palace, there is a large stable formed by five naves, and which once must have lodged a corps of cavalry in Latin times. (...) I believe that it was the seat of the Jerusalem Templars of perhaps the Teutonic Order. Even the stables are not absolutely of Faccardino's time, who lived in the 17th century. The walls of the construction are of an extraordinary thickness, and one can observe that they are not all at the same level, i.e. at the height of the horses, but that they wanted to build another edifice above it. Either the building was never built, or it was destroyed by some unknown event. One can say the same about what is the rest of a tower that is annexed to the palace. The walls of this tower are of a great thickness.

Towards the town, facing the gate of the palace is a beautiful mosque built by the Grand Emir Faccardino (...)."

The map (fig. 4) gives a clear illustration of the plan of the palace in the early 1 840s. It was, however, modified during the 18th and 19th centuries, according to the whims of the local pashas. The statues and vases were perhaps destroyed by less religiously tolerant masters of the city than Fakhr ad-Din, or taken away by looters.

If the idea is correct that the Fakhr ad-Din palace was a "Crusader" edifice, repaired and partially rebuilt, the city would have been protected by two important buildings. They would have defended the eastern side to the city as well as its approaches from the coast. A fort or tower defending the Grand Sérail spur would have been in visual contact with the Castle, the Burg al-Kashef, Burg al-Hod. r and Burg Abi Haydar. Possibly a tower existed somewhere on the Ashrafiyyeh plateau to observe the fords of the Beirut river and to receive signals from the South. A similar tower also existed on Ras Beirut hill to watch over the sand dunes and the sea to the West.

But the Grand Sérail spur was the weakest point of the defence of the city: anyone who controlled its top could dominate the city, and monitor all movement. Bombing any point could be done with accuracy, and the walls forced with ease, since we have seen that the city was defended mainly on the eastern side. This situation lacks all logic. The spur is the key to the city, and it would be surprising not to find any signs of military construction there. The 1830 map by Wyld mentions walls and ruins in the sector, while Buckingham noted extensive ruins and columns there. They could either be the traces of the Palaces of the (Ma'n) Emirs, or of a Crusader fortification. Perhaps the Burg 'Umm Dabbus was the lone survivor of this fortification.



CONCLUSION

The five maps draughted between 1830 and 1842 are documents of prime importance for the reconstruction of the urban physiognomy of Beirut. They have permitted an identification of the site of the first urban nucleus, as well as aspects of Roman and Mediaeval defences. The documents can be used as base-maps for further investigations on the city, especially for the Mamluk and Ottoman periods. Many Muslim authors have left descriptions of Beirut, and their information could be placed on maps to further understand the internal organization of the town. Precise dating of the public edifices, portions of walls etc. could be advanced. Maps of the functions of the souks could be established and their evolution would help in understanding the socio-economic history of the area. It is hoped to produce a continuation of this study of Beirut using maps and sources covering the town's later history.

On another level, research could be conducted on the military organization of the peninsula, with a study of possible locations for watch-towers and roads. This study could be combined with an examination of the distribution of the Hans and of other relays mentioned by the maps or various visitors. The regional organization of the coastal plain could be elucidated. The maps could also help research on the location of ancient religious edifices. Mar Elias Btina, for example, is located in an unusual site, away from the fertile plain, in the middle of the dunes. Could it be an ancient site converted into a Christian shrine? Mar Mitr (mentioned on the maps as Saint Demetrios Convent, and with no indication of today's cemetery) is also situated on a curious site for a church, built exactly at the base of a cliff, just under the Ashrafiyyeh hill. Could it (or its predecessor) have been built at the entrance of an ancient sacred cave, re-used as a Christian shrine ? Or was it built near or on a spring (of which no trace exists anymore) venerated for some as yet unknown reason? Do other sites exist on or near Beirut peninsula which could inform us on pre-Christian religious activity?

Finally, the maps here studied must be compared to other maps of the Levantine coast, so as to define a model for the spatial evolution of the urban centres. The cities of the interior are now well documented (Aleppo, Damascus, Cairo), but the coastal cities are still relatively unexplored. Beirut is the least well-known of them all.



BIBLIOGRAPHY

Abdel Nour, A. 1982. Introduction a l'histoire urbaine de la Syrie ottomane (XVI-XVIIIe siecle). Beyrouth: Publications de l'Universite Libanaise.

Boase, T.S.R. 1967. Castlesand Churches of the Crusading Kingdom. London: Oxford University Press.

Carter, H. 1983. An Introduction to Historical Geography. London: Edward Arnold.

Chebli, M. 1946. Fakhreddine II Maan. Prince du Liban (I~72-I6~). Beyrouth: Imprimerie Catholique.

Chevallier, D. 1971. La societe du Mont-Liban a l'epoque de la revolution industrielle en Europe. Beyrouth: IFAPO.

--, 1972. Signes de Beyrouth en 1814. BEO 25:211-228.

Daher, M. 1985. Bayrut wa Gabal Lubnan. Beirut: Dar al-Nada (in Arabic) [Translation of text in Russian by A. Kremski].

Davie, M.F. 1984. Trois cartes inedites de Beyrouth. Annales de Géographie (Beyrouth, Universite Saint-Joseph) 5:37-82.

Du Buisson, R. du Mesnil. 1921. Les anciennes defenses de Beyrouth. Syria 2:2~S-2~7 and ~ I7- ~27-

--, 1923. Rapport sommaire sur la mission archeologique de M. le Comte du Mesnil du Buisson en Syrie. Bull. Soc. Fran,c des Fouilles Archéologiques V: 121-128.

Dussaud, R. 1921. Le peintre Montfort en Syrie, 1837-1838. Syria 21:1-33.

Jidejian, N. 1973. Beirut through the Ages. Beirut: Dar el-Machreq Publishers.

Jouplain, M. 1961. La question du Liban. Etude d'histoire diplomatique et de droit international. Jounieh.

Lauffray, J. 1948. Forums et monuments de Beryte. BMB 7-8:1-78.

Mariti, G. 1787. Viaggio da Gerusalemme per le coste della Soria. 2 vols. Livorno: T. Masi e Co.

Mouterde, R. 1942-43. Emplacement du forum de Beyrouth. MUSJ, 25: 33-40.

--, 1966. Regards sur Beyrouth, Phénicienne, Hellénistique et Romaine. Beyrouth: Imprimerie Catholique.

Raymond, A. 1985. Grandes villes arabes a l'epoque ottomane. Paris: Sindbad.

Rey, E.G., 1871. Etudes sur les monuments de l'architecture militaire des Croisés en Syrie et dans l'ile de Chypre. Paris: Picard.

Rouleau, B. 1975. Le trace' des rues de Paris. Formation, typologie, fonctions. Paris: Editions du CNRS.

Sanlaville, P. 1977. Etude geo-norphologique de la region littorale du Liban. 3 vols. Beyrouth, Publications de l'Universite Libanaise.

Sauvaget, J. 1941. Alep. Essai sur le developpement d'une grande ville syrienne, des origines au mileiux du XIXe siecle. 2 vols. Paris.

--, 1934. Le plan de Laodicee-sur-Mer. BEO 4.

--, 1949. Le plan antique de Damas. Syria 26:314-358.

Yapp, M.E. 1987. The Making of the Modern Near East, I792-I92~. London & New York: Longman.

----------------

al@mashriq                       960325/960614