Version classiqueVersion mobile
OpenEdition Books

Les bâtisseurs sur les deux rives de la mer Rouge

 | 
Jean-François Breton

Chapter 3. Building materials (8th cent. BC to 1st cent. AD)

Texte intégral

3.1. Stone (Jean-Claude Bessac)

1The majority of the South Arabian monumental productions of this period stand at the foot of the mountains in the eastern part of Yemen, around the Ramlat as-Saba’atyn desert. The South Arabian builders limited themselves to the use of stones from the nearby surroundings of their cities because of the impossibility of long-distance transport. Except for certain decorative stones such as the gypseous alabaster and chalcedony, the radius of the area from which they quarried stone for dressing did not exceed fifteen kilometres. For this reason, each site adapted its techniques and its choice of construction to the local geological characteristics. For example, in the part of the Jawf valley which includes the cities of Barâqish (ancient Yathill), Ma‘în (ancient Qarnawu), Khirbat Hamdân (ancient Haram), Kamna (ancient Kaminahu), al-Baydâ’ (ancient Nashqum) and as-Sawdâ’ (ancient Nashshân), only sedimentary rocks were used. The size of the stone blocks and their hardness vary greatly, as they include firm limestone with average-sized grains as well as the very hard and breakable varieties with fine grains (stones known as “cold”), and (intermediate between the two) oolithic limestone, slightly shelly (but never soft), and also sandstones. Hard stone clearly prevails in all of the South Arabian monuments.

  • 1 Bessac 2009: 42-43.

2From one city to another in the Jawf valley, the choice of stone is related to the possibilities of access to the various quarries1. In Mâ’rib, in addition to hard and cold limestone, the interior faces of the walls often consist of a very soft volcanic conglomerate which takes the place occupied elsewhere by mud-bricks. Further to the south, the builders of the city of Tamna‘, situated in a granitic zone, primarily used this very resistant stone, provided by nearby quarries. This is also the case for Hajar Yahirr where granite and schist are commonly used. In these two cities, only some paving and a very few constructions are made of white limestone. In Shabwa, hard sandstones are employed in foundations. Hard and cold limestone, mined only a few kilometres from the site, predominates in the walls. Firm limestone is represented in quantity in the form of plating and in the sculpture of the gargoyles of the royal palace. No soft stone, such as tuff, is available in this area except for the surroundings of Mâ’rib.

  • 2 Bessac 1998a: 174-176.

3The stone types in the Jawf valley catalogued below will be presented mainly from a technical and not from a geological point of view. Since the conditions of surveys in the Jawf have prevented, up to now, archaeologists from exploring the ancient quarries, the lithostratigraphic character of stones could be approached only through information gathered from the constructions. Three main varieties of stone were employed: lumachellic limestone, sandy limestone and oolithic limestone. The last of these corresponds to the dominant raw material for dressed stone in the Jawf. Within this group, secondary categories of limestone can be differentiated based on two essential factors: the homogeneity of the ooliths and the cohesion of the cement which binds them. Although from a geological point of view, these would seem to be merely variants of the same formation, it is preferable to present them separately in order to better elucidate the study of the technical problems. Five kinds of limestone will be analysed2.

3.2. Lumachellic limestone3

  • 3 Bessac 1998a: 174.

4This rock is very dense; the shells inside it which give it its character of lumachelle are very fragmented and drowned in a fine calcareous cement, compact and dense, therefore practically devoid of vacuum. Beige rosy calcareous cement gives the dominant colour to the material. Its hardness varies little from one site to another. Its hardness is estimated between indices 8 and 12 according to the French AFNOR standard (French norm of stonemasons). This rock thus belongs to the hard and cold category of stone. The compactness and the smoothness of its grain give the stone a hard shiny surface when polished. Once thus treated, its natural colour becomes darker and it takes on the aspect of granite. Lumachellic limestone is generally used for monoliths whose broad sections and long lengths testify to the regularity of the benches in the quarries: it is common to find stones more than 3 m long and 40 cm to 60 cm in section.

5This is the most frequently used type of stone in the construction of the archaic (8th-7th cent. BC) temples. It seems that this material was valued more for its fine polishing qualities, which bring out the engravings, than for its size (fig. 4.1).

3.2.1. Sandy limestone

6Sandy limestone consists chiefly of round and angular grains of quartz with occasionally other minerals, such as feldspar and mica. The Ma‘în buildings are constructed of a slightly sandy limestone with very fine grains (approximately 0.1 to 0.2 mm), with mixed fossils inside. Its beige grey colour is rather uniform. It sometimes includes silica lines inside natural beddings. Its sandy texture does not preclude polishing.

3.2.2. Oolithic limestone

7The general characteristic of oolithic limestone is the presence of a large proportion of carbonate of lime, formed chiefly by the accumulation of shells or calcareous skeletons of marine organisms. In the Jawf buildings, oolithic limestone is a beige, pink or white stone, often ribboned with beige lines. The cement which binds together the oolithic grains, mainly carbonate of lime, imparts only a medium degree of hardness to the whole of the stone. This stone corresponds to n° 6/7 in the French AFNOR standard, i.e. a firm stone.

3.2.3. Hard oolithic limestone

8This variety of oolithic limestone has almost no fossils visible to the naked eye. Moreover, its ooliths are almost molten in the cementing material which binds them. The general colour of this limestone varies between cream-coloured white and light beige. Its hardness is estimated between indices 7 and 10 in the French AFNOR standard; it thus belongs to the varieties of hard limestone. It is used in particular in the upper parts of the city-wall of Ma‘în. To judge by the height of courses in the buildings, the thickness of the natural benches in the quarries should vary between 20 cm and 40 cm. These stones were used in the part-worked rubble wall in the temple of ‘Athtar dhû-Risaf in as-Sawdâ’, and at Mâ’rib in the city-wall and in the necropolis (7th-6th cent. BC).

9It is difficult to know what South Arabian term was used for these various limestones. The term blq may apply to this “hard oolithic limestone”. The word is common in Ma‘în and in Barâqish building inscriptions on city walls. But it may apply also to limestone of every type, as blq is the only epigraphical term although many different limestones are used for building.

3.2.4. Cold oolithic limestone

10It is almost the same material as the preceding one, except that its ooliths cannot be distinguished from the more calcareous cement. It is a thin and brittle homogeneous cementing material which, like flint, gives a conchoidal profile to the splinters. Its natural hardness gives it a classification between indices 10 and 13 in the French AFNOR standard. It is a cold stone, the hardest that can be found in the Jawf for building.

3.3. Sandstone

11A variety of firm sandstone, sometimes rather susceptible to deterioration depending on the resistance of each of its layers, is represented in the Mâ’rib necropolis. The absence of any petrographical analysis does not allow us to distinguish “calcareous sandstone” (calcareous cement is usually in the form of crystalline calcite which has been deposited from solution) and “sandy limestone” with a high proportion of silica. Thus one will generally stick to the denomination of “sandstone” or “sandy limestone”. “Firm” indicates an average degree of resistance, neither too hard nor too smooth under the action of the dressing tool.

12Different qualities of sandstones are used in the foundation of the city-walls in Hadhramawt. At Shabwa, blocks of red or green sandstone are used in various places (western sector, Dar al-Kafîr, al-Hajar). This uncoursed square rubble is of two sizes, either large-sized blocks measuring 0.80 m by 0.70 m (west sector) or small rubble (0.20 m x 0.30 m). Coursed square rubble sandstone is used to build the walls in al-Barîra (wâdî Jirdân).

13In the zone A necropolis of Awâm/Mâ’rib, sandstone is employed in the walls of tomb n” 6, in zone Β in tomb n° 9 and in the foundation of the walls of building n° 10. This sandstone is cut according to the type of masonry, which is generally smaller than with the preceding limestone. As this sandstone is rather smooth and not very breakable, the technique of its extraction was certainly different from that of the hard limestone. It is possible that this technique somehow influenced the choice of small-sized blocks over larger blocks. On the other hand, it does not seem that there was a systematic attempt to reduce the volume and the weight of these stones.

3.4. Ornamental stones4

3.4.1. Chalcedony

14Chalcedony with its white milky aspect, sometimes moderate with red, gold or green veins, looks like the gypsum alabaster with which it is often confused. Chalcedony is not found around Shabwa; it had to be imported from quite a distance away. Curiously, as far as we know, this very beautiful stone is attested in large blocks only in the royal palace of Shabwa. The chalcedony is here represented by three large blocks in the staircase. This is a rather surprising place to find such a decorative rare stone, and such a difficult stone to cut, especially in a massive form. These three blocks may come from an older prestigious monument of the site. Chalcedony is also known at Shabwa and other sites in the form of plaques decorated with recessed panels imitating wooden decorations and sometimes with inscriptions. These archaeological witnesses are usually considered to be late in Shabwa.

3.4.2. Alabaster (gypsum)

15Many gargoyles in Shabwa and Mâ’rib are made of this gypsum alabaster, white-coloured with greyish veins. Most of them terminate in a bull’s head with horns. By contrast, calcareous alabaster (or Egyptian alabaster) is unknown in the Shabwa-Mâ’rib area.

3.5. Various other stones

3.5.1. Granite

  • 5 Breton-Arramond-Coque-Delhuille&Gentelle 1998: 12-23.

16Granite is the name given to a group of plutonic acidic igneous rocks which have consolidated at a great depth below the earth’s surface, causing them to cool very slowly. Because of this slow rate of cooling, the rock is completely crystalline in structure. The exposed rock is usually termed an outcrop. Granites are widely attested in the Wâdî Bayhân-Markha area: the outcrops are mainly oriented SSW to NNE according to the main fault lines (where wâdîs later arose, like Wâdîs Bayhân and Surbân). Hornblende (a silicate of lime, iron and magnesium) is also widespread in Wâdî Bayhân5.

17When freshly quarried, granite contains some amount of moisture known as quarry sap, and while the rock is in this condition it is readily worked. Because of this, it is preferable to have the stones worked directly at the quarries. Granite as a facing material is particularly suitable in positions where abrasion is likely to occur. The durability of granite explains the good conservation of some inscriptions, such as RES 4329 in Hinû az-Zurayr walls or RES 3881 and Jamme 2437 in the South Gate of Tamna‘.

18Some granite quarries have been identified north of Hajar Yahirr and in Wâdî Dura’: tool marks are clearly visible in the quarry faces. Stones from these quarries were commonly used at Hajar Yahirr, capital of ‘Awsân, for the extra-muros temple and for intra-muros houses.

19Granite blocks could be set in uniformly coursed ashlar masonry as in al-Mi’sâl: courses are 60 cm to 74 cm high (building C) in Hajar al-Dhaybiyya (Wâdî Dura’). In Tamna‘, all the buildings (city-wall, domestic houses and temples) are built in granite of different colours, black, grey, pink or orange.

3.5.2. Volcanic tuff

  • 6 Brunner 1989: 28.

20Tuff, a rough, light, spongy stone, is used in the foundations and the internal faces of the tombs in Mâ’rib. It is dark coloured, hardly more durable than mud-brick. It probably comes from lava flows close to the ancient site of Mâ’rib6. Its extraction can be carried out in the open air, using only an ordinary tool for earthworks, by trenching a perimeter around the blocks. Underground mining is also possible with this very soft rock because it is sufficiently homogeneous to allow deep galleries. However, due to the rapid deterioration of this stone, which is not very resistant once dressed, it hears no traces of tools which could give information about quarrying techniques.

3.5.3. Slate

  • 7 Dostal 1983: 80 and plate V (Houses at al-Khâlef).

21The main characteristic of slate is its distinct cleavage caused by intense lateral pressure; under this condition it has lost its original bedding planes. The colour, caused by the presence of iron and carbonaceous substances, varies across green, grey and brownish tints. Because it splits easily into thin slabs, slate is typically used in roofing; for stairs; and for façades, mainly in parallel lines of horizontal rows, slightly oblique, so as to protect the building from rain. Many such slate slabs have been discovered in the Royal Palace of Shabwa; some are 50 cm long and 4 cm thick. Similar slab rows are used in ‘Asîr tower houses or granaries, mainly in mud buildings7.

22Slates are also used as treads in stairs and as paving in domestic buildings in the Markha ‘Abadân areas. In Hajar al-Dhaybiyya (wâdî Dura’) houses, the main entrance, the central corridor and some rooms are paved with thick slabs of slate.

3.5.4. Pebbles

23The wâdîs near the cities carried a huge amount of pebbles which were sometimes used in buildings. The most striking evidence is an early staircase (n° 1) in the royal Palace of Shabwa. Under this staircase, there is a massif of pebbles, 6 m high, 10.50 m long and 4 m wide. It is made of thick layers of pebbles intercalated with layers of stone splinters. The height between each layer of splinters corresponds to the height of each course of masonry of the platform of building A (0.40 m/0.50 m).

3.6. Plaster and gypsum

24To produce lime from limestone by burning in a kiln requires the maintenance of a high temperature (ca 900° C) for a period of three to six days. It is generally considered that plaster (which is made from lime) was introduced in the Ancient Middle East only in Graeco-Roman times. However, some archaeological sources make it clear that the burning of limestone was already taking place in archaic architecture of South Arabia.

25Plaster is used as a basic binder in the tombs of Mâ’rib, but not in a systematic way. It appears in two forms. The first gives a very refined and white plaster grout used to bind and stabilise the stretchers, between the bedding surfaces and the resting ones, thus acting as a kind of shock absorber to counteract the tendency of a stone to break under a heavy load. This grout is to be found in particular in the walls of tomb n° 9 of sector Β and in tomb n° 1 of sector A in Mâ’rib. The second category of plaster is coarser and granular. Its use can be identified in the front wall of tomb n° 1 in sector A in Mâ’rib to repair small splinters in the joints of the stones.

  • 8 Wright 1985: 373.

26Gypsum is “a term used both for a building material and for the common material, calcium sulphate, from which it is produced. Various crystalline developments of the latter material give the rock belenite or true alabaster (not the Egyptian alabaster which was calcite or ‘onyx marble’)”8.

27Gypsum for producing plaster is found at various places in the Yemeni Highlands, close to Sanaa, near Shibâm al-Ghirâs. At Shabwa, the mortar always contains lime. In the current state of research, the use of plaster in the Mâ’rib necropolis is the oldest testimony to its use in South Arabia (7th-5th centuries BG). In Mâ’rib, probably during the 3rd and the 2nd centuries BG, lime mortar was used as cement in the hydraulic works, especially in the northern and southern sluices.

3.7. Stone quarries at Shabwa and Axum

28Few stone quarries have been carefully studied, some around Mâ’rib or Shabwa, but none in the Hadhramawt and Jawf valleys. In this last area, some engraved quarry marks have been observed on stones, at Ma‘în, at al-Baydâ’ and at Barâqish. The quality of their engraving looks very crude and should not be compared with carefully incised mason’s marks. The former crude quarry marks generally appear on one of the inner faces of the rough stones of the ramparts exhibiting no special finishing, at least on the inner face; but they do not exist on finished dressed stones. It seems that the stone cutters did not identify their production in this way. Thus, it is clear that these quarry marks were incised by stone workers (not stone masons) far away from the building site, and probably even in the quarries themselves.

29If this assumption is right, these individual or collective marks were probably intended to identify various quarrymen for their own payment. It is also possible that the contractors sought by means of these marks to identify the product. These may also have figured in the quality control of the work and of the material (fig. 3.1).

  • 9 Bessac 1998b: 231-234.
  • 10 Bessac 1998b: 231.

30Many quarries have been studied around the city of Shabwa, providing information about the process of distribution of suitable stones and their use in the various buildings of the city through history9. On the eastern side of the city, the spur of the rocks of al-‘Aqab raised by the extrusion of rock salts comprises hard sandstone beds useful for building. These beds of sandstone, 30 cm to 60 cm thick, were often split up into blocks of an eighth to a quarter of a cubic metre by tectonic pressure. This natural building stone is the nearest to the city and at the same time the easiest to extract, because nature has already carried out the main roughing out. Unfortunately, no tool marks are visible10.

  • 11 Pirenne 1990: 49-50.
  • 12 Bessac 1998b: 232.

31Located 3 km west of the city, Sh’ib al-Layl was certainly the main quarry near Shabwa. The quarry is at the bottom of a narrow valley ending in a cirque at the foot of a high cliff dug out to form a vast shelter. At the bottom of this shelter, an inscription11 mentions that a king took stones to build his house Sha’ban. As the limestone at this level is unsuitable for extraction, only the very hard limestone coming from 150/200 m higher, on the high edges of the Khushm Ghalîb plateau, was mined at Shi’b al-Layl. The main fallen isolated blocks, already detached from the cliffs, were the first to be exploited. Given the proximity of the city, the best blocks available must have been used first. Nevertheless, an enormous amount of rock remained unused while roughing out. Very close by, 2 to 3 cubic meter blocks were also abandoned after a preliminary roughing12.

32The traces on these blocks form three parallel lines of strokes on their upper face. The impacts, quite unequal and unclear, were probably made from the blunted points of some heavy tools, probably hard stone percussion tools. The principle of fracture is founded on the progressive weakening of the stone, by making small holes in a line along the striking surface. Then the actual splitting is done with wooden or metal wedges driven into these small holes. Above the Shi’b al-Layl cirque, the calcareous plateau of Khushm Ghalîb, some 9 km long, rises high above the plain, but only its most northern projection was exploited for quarrying and especially for the transformation of limestone into lime.

33Above the cliffs, the edge of the plateau is skirted by a path going down the slope to reach the bottom of a narrow valley facing Shabwa. Along the path of the descent, there remain retaining walls, built with large blocks. On the plateau, the layout of the path is very clear because the stones were carefully removed all along, forming a clear path approximately 2 m wide. At the places where the rock formed strong protuberances, the limestone was strongly heated in situ in order to reduce it out as lime, which is easier to break, and then levelled. These traces of calcinations (but no trace of lime kilns) appear especially where the pathway starts its descent.

34Long lines of limestone domes, a few metres in height and some twenty metres in diameter, with their slopes covered with fallen debris and white traces all around, are definitively lime kilns. All around, the bedrock is fractured and allows the quarrying of blocks approximately 40 cm in length, 15 to 20 cm in width and 20 cm in height. At this site, the stone yield could be estimated roughly at one cubic meter for fifteen to twenty cubic meters of extracted rock. Thus, the bulk of the production consisted of very small blocks from 2,000-5,000 cm3 in volume, intended only for the lime kilns.

35The extraction was practised all around the kiln domes. Above the cliffs, on the plateau, there remain traces of many lime kilns. Unfortunately, their upper part no longer exists, and there remain only circular shapes from 2 m to 3 m in diameter, sometimes delimited by large stones. Inside the ovens are found whitish brittle lime and incompletely calcinated stone fragments. The single building identified on the plateau evidently represents the remains of a stone hut, probably as shelter for quarrymen. In addition, several well-cleared spaces of a few square meters probably mark the camping sites of the quarrymen.

3.8. Earth and clay

36In South Arabia, there is no particular type of earth which is specially favoured or indicated for the manufacture of mud bricks. Whatever earth is at hand near the building is generally suitable. A supply of earth of the requisite plastic quality is mixed with water to form a plastic mud and very often with vegetal material, i.e. straw or dung. This raw material is prepared on the site, and is struck into units using a wooden box mould or frame, although no such frame has ever been discovered.

37In Yemen, most of the excavation reports provide very little information about the material composing the mud bricks, their colour, their geological origin, and their form. Most of the bricks were moulded in orthogonal forms: i.e. as plain flat parallelepipeds, the height being the smallest dimension. It seems axiomatic that the dimensions of the moulded bricks should be standardized and that any observed variation should therefore have some chronological or regional significance. However, we cannot yet produce any such chronology of brick manufacture in South Arabia due to the lack of excavations of the walls’ deepest layers (see infra).

3.9. Wood

  • 13 König 1987.
  • 14 König 1987: 42-46.

38The built-up areas of South Arabia, from ‘Asîr13 to Hadhramawt, were not in well-watered regions. Some of them just around the Ramlat as-Sab’atayn desert, and others like the Jabal al-Nisiyîn granite areas, were not particularly fertile. Rainfall, largely under 100 mm a year, would not allow trees to grow abundantly. Furthermore, afforestation and deforestation of the region has changed noticeably over the last century, and evidently changed very considerably over three thousand years. Nowadays, few trees are to be seen in the areas around Tamna‘, Hajar Yahirr and Shabwa. In the highlands of Yemen, we have clear evidence of cultivation in earlier times; this has drastically reduced the forests to solitary representatives of any species. In the ‘Asîr, north of Yemen, there is nowadays practically no trace of Juniperus14 and very little Olea europaea forest. Today it is rare indeed to find traces of the Juniperus belt that once existed in Northern Yemen.

39Yet there is now considerable evidence that wood was used extensively in ancient buildings. Because of the dryness of the climate, wood has been fairly well preserved in many buildings. Archaeological excavations in Shabwa, in Hadhramawt and in Tamna‘ have provided large quantities of wood, either in its original condition or as charcoal. Hundreds of beams and posts have been discovered in the royal palace of Shabwa. As the wood was burned to charcoal under thick mud brick layers, the shape and the mortise-and-tenon assemblage have been rather well preserved. It is now clear that the floors of well-off houses were constructed with a wooden framework.

  • 15 König 1987: 37-48.

40It is worth mentioning the trees that were probably used in antiquity. First would be the Zyziphus or Paliurus Spina Christi or Jujuba (in Arabic: ‘ilb) whose extensive use is attested in Shabwa building and in Mâ’rib. Varieties of acacias were widespread in the lowlands of Yemen: negrii, gerrardii, tortilis, etbaica and tamarisks. In the higher escarpment over the Red Sea, the most characteristic tall trees were Ficus vasta and Ficus sycomorus, but the dominant species was Cordia abyssinica. Above 1,500 m, in the middle rainfall zones (1,500-1,800 m altitude), Juniperus Phoenica and Juniperus Fxcelsa (above 2,000 m) are found in the Khawlân as-Shams, Hujariya and ‘Asîr mountains (Jabal as-Sawdâ’, near Taif (alt: 2,600 m), northwest of Abha and Zahran al-Janûb)15. Juniperus Excelsa also appears in Eritrea above 2,000 m (on the mountain line, east of Asmara), in Ethiopia, and in the Red Sea hills of Sudan.

  • 16 Serjeant & Lewcock 1983: 44.

41In describing the Ghumdân palace of Sanaa, Al-Qamah lists the timbers used in its construction: sawn juniper, teak (or sâj “Indian teak”), hard lote-tree wood (sulb al-sidr) and Labakh Mimusops Schimperi16.

42If there were enough material to assemble a corpus of ancient wood, this would constitute some indicator of ancient vegetation.

3.10. Ore and Metals

43Silver and lead mines have been found in Jabalî mountain, some 50 kilometres north of Mâ’rib, but a pre-Islamic use is not yet attested.

44Copper mines were largely used in Yemen near Mâ’rib, at al-Baydâ’. In this latter area, narrow trenches scattered in the outcrops, some five meters deep, show antique mining techniques. This copper was used extensively in the South Arabian bronze craft industry. Samples show that the bronze ornaments (architectural decorative elements, plaques and statues) contain some 95% to 97% copper. Tin would have been imported, but its origin and trade remain uncertain.

  • 17 Bessac 1998b: 235.

45Iron mines. Approximately ten kilometres south of Shabwa, the Jabal Harash is made of ferruginous soft sandstone. This rock, naturally very fissured, cannot be used as building material. But in spite of its low content of iron, it was exploited in order to extract iron ore. The iron ore exists in three forms: in surface deposits on rocks, in concretions on limestone at the base of the sandstone, and isolated in irregular nodules from 2 to 5 cm in diameter. Only this last variety was of interest to the ancient metallurgists17.

46In the middle of Jabal Harash, in an approximately one-kilometre diameter zone, ten sites of iron ore extraction have been found and there are probably more. The majority are open pit mining, sandstone being quarried out from the many natural cracks; the sandstone was then broken up in order to extract the ore from it. The mining sites form working faces, often irregular, a few metres in length, with sandstone scraps piled up at the foot of the faces. Many of these faces are engraved with short inscriptions in South Arabian letters.

47One mining site, unusually, consists of three tunnels 5 m to 8 m deep, of more or less rectangular section from 0.8 m to 1 m high and 0.60 m to 0.80 m wide. At the entry of the deepest tunnel, in addition to some South Arabian letters, there is a series of short vertical lines separated by a point. These could be marks for counting quantities of ore—baskets, for example—extracted at the site. Most probably these tunnels should be considered as test galleries intended to assess the potential mining yield, rather than as true mining galleries. At the mining sites, fragments of iron stems have been found some 10 to 12 cm long and 1 to 1.2 cm thick. They could be percussion tools, whose point would have broken off. Such tools could have been used for breaking sandstone to extract the iron ore from it.

48Following the identification of these sites of extraction of iron ore, surveys in the surroundings of Shabwa unearthed several heavy slag fragments of a twisted aspect characteristic of the processing of rough iron in low hearths with natural ventilation. Thus, in spite of the scarcity of iron tools at Shabwa, we know that iron manufacture was locally assured, probably just enough to produce iron tools for erecting local buildings, but not enough for export.

  • 18 Lepage & Mercier 2005: 235.

49In Tigray, one can note the presence of gold in the alluvia of the Weri River and primary gold is found in the locality of Adi Hoza to the northeast of Hawzien. Some copper, of abnormally high content, is present in the sediments and comes from the intrusion of granites18.

Notes

1 Bessac 2009: 42-43.

2 Bessac 1998a: 174-176.

3 Bessac 1998a: 174.

4 Bessac 1998b: 247-248.

5 Breton-Arramond-Coque-Delhuille&Gentelle 1998: 12-23.

6 Brunner 1989: 28.

7 Dostal 1983: 80 and plate V (Houses at al-Khâlef).

8 Wright 1985: 373.

9 Bessac 1998b: 231-234.

10 Bessac 1998b: 231.

11 Pirenne 1990: 49-50.

12 Bessac 1998b: 232.

13 König 1987.

14 König 1987: 42-46.

15 König 1987: 37-48.

16 Serjeant & Lewcock 1983: 44.

17 Bessac 1998b: 235.

18 Lepage & Mercier 2005: 235.

© Centre français des études éthiopiennes, 2015

Conditions d’utilisation : http://www.openedition.org/6540