How Petra was Built:
An analysis of the construction techniques of the Nabataean freestanding buildings and rock-cut monuments in Petra, Jordan
BAR International Series 1460

(How Petra was Built, Shaher M. Rababeh, 2005, p223-227)


The previous chapters have dealt with the surviving evidence concerning Nabataean construction techniques. Data collected were used to provide the bases for a systematic study of the building materials and to determine the specifications of the various types of Nabataean construction techniques applied in both the freestanding buildings and the rock-cut monuments. The Nabataeans had their own distinctive architectural style which reflected both Hellenistic and Oriental influences, the origins of which it is possible to suggest. However, determining the origins of the techniques used to build them is not easy. I found that some of these construction techniques were used in the same way as throughout the Greco-Roman world, and were probably borrowed. However, there are some technical features used by the Nabataeans which indicate that they further developed and adapted these techniques rather than merely borrowing them. At the same time, there are features not found elsewhere. Others are the earliest known examples. Therefore, there is in fact a whole range of possibilities, and the question is how it developed and varied, and where within this range Nabataean construction practice lay. It is not at present possible to establish chronological stages for the different technical aspects because the monuments are packed into a short period (from the last quarter of the 1St cent. BC to the early 2nd cent. AD), and their exact dates of construction within this period are still not absolutely certain.'

Nabataean construction techniques before those used at Petra have not previously been defined (in I.a.2.1), and before the present study their possible relationship to those used by their predecessors, the Edomites, have not been considered. The experience of the Nabataeans digging wells to store water underground is the only architectural skill reported in Diodorus' account of 312 BC.2 What I found is that very few construction features can be attributed to the Edomites, the earlier inhabitants of Petra, in which some local technical features were developed. Three sources of evidence show influence from Edomite architecture which can be recognized in Nabataean architecture. The Nabataeans' knowledge of carving cisterns reported in Diodorus' account could have been inherited from the Edomites. The walls of the earliest Nabataean freestanding houses were built of wadi stones and clay, not ashlar masonry.3 This type of construction is somewhat similar to the walls recovered at Umm el-Biyara, Tawilan, and Buseireh.4 Moreover, it is probable that the method of climbing up rock faces with the aid of ropes and holes was inherited from earlier knowledge in the area, as it is found in the Neo-Babylonian rock relief of Sela'5 (in III.b.2). Although the Edomite evidence does not include complex public buildings, some Edomite influence can be detected in Nabataean architecture. So continuity like that already detected in religion' and ceramics' can now be suggested.

' McKenzie 1990:121-2 and more recently 2004: 559-568. Diodorus Siculus 19.94-100.

3 Parr 1990: 16-17.

4 cti.="4n.',.4; 1Q05. 116

There are also other technical features that the Nabataeans inherited from the Iron Age practised elsewhere in the region. It is shown here that, contrary to Hammond's suggestion,' the use of wooden beams as stabilizing aids embedded between the courses of ashlar masonry walls in Petra can be attributed to regional tradition. This technique has been in use in the Levant since the Iron Age,9 but its use was restricted to small buildings and sometimes in more primitive constructions using field stones. The Nabataean builders were able to develop its application into a number of sophisticated versions used with monumental ashlar masonry (in V.c.2).

The main point is that the construction techniques used and developed by Nabataean builders were affected by two important factors: their outside contacts, and the availability of the building materials.

The first factor is the influence of contemporary Hellenistic and Oriental architecture in the surrounding areas. Petra was a trading emporium and had strong ties with the West and East as is clearly shown by its architecture and other remains (in I.a.2.3). It was previously assumed that there was relationship between the Nabataeans and the Seleucids.'0 Consequently, it was thought that some of the Hellenistic influence in Nabataean architecture might have come from the Seleucids, not just Alexandria. But as there was no surviving Hellenistic architecture in Antioch it was not possible to confirm or refute this possibility. However, my detailed analysis of Nabataean contacts (in I.a.2.2) shows that the strongest relationship and direct exchanges were with Ptolemaic Alexandria, and not with Seleucid Antioch. The relationship with the Seleucids was one of confrontation from as early as 312 BC" and continuing until the end of their empire. By contrast, contact with Alexandria was significantly increased by the trading activities in which the Nabataeans were involved. These observations not only accord with the architectural styles of Petra, which show that Alexandria was the main Hellenistic centre of architectural types from which the

5 Dailey and Goguel 1997: 171.

6 Parr 2003: 33; Bartlett 1990: 34.

7 Bienkowski 1990: 103.

Hammond 1995:215-21.

Thomson 1960: 60-2; Reich 1992: 8; Shiloh 1979: 61. 16 Robertson 1943: 221.

Diodorus Siculus 19.94-100.


Nabataeans borrowed their style:2 but also provides some evidence, for the first time, that perhaps there was not much influence on it from the Seleucids.

I found that there was also contact between the Nabataeans and the Parthians, and with the occupants of Arabia. The historical evidence and the involvement of the Nabataeans in the silk route and incense trade (in La.1.2) suggest a strong relationship existed between them. This contact is reflected by the continuing use of crowsteps," and some of the characteristics of the

Nabataean temples (in I.a.2.3).14 But there are no technical studies of Parthian architecture published and nor is the architecture in Arabia well published, so it is not possible to confirm or refute the possibility of technical exchange. However, all these contacts provide the possibility of exchange of building techniques. The presence of foreigners among the people of Petra, as reported by Athenodorus in Strabo's account,15 makes this even more likely.

The second factor which affected the Nabataean construction techniques was the availability of the building materials. The survey of the materials made in chapter 2 contributes to our understanding of the basic building materials used by the Nabataeans. It appears that the supply of the major building materials in bulk followed a logical economic model. Generally, the Nabataean preference was for local resources, as they chose the most readily available and cheapest supplies based on their location. The exceptions occurred when buildings required special materials not found in the region, such as wood and marble. The most readily available building material was sandstone. The rock-cut monuments were carved in the Smooth, Honeycomb, Tear, and Disi sandstone layers, and these layers were also quarried to provide blocks for freestanding buildings. This use differed from the type of stones normally available and used in classical monumental architecture elsewhere, where limestone was the most common stone. However, the sandstone of Petra is not a high quality construction material compared to limestone or granite. Undoubtedly, this had a notable effect on the construction techniques used. Advantages of it were its availability and its relative ease of cutting which would have been factors in its use. But its softness also made it quick to erode, and it was weaker than limestone. Despite its friable nature and lack of resistance to weathering, the Nabataeans still used it because it was the only stone available. However, because they were aware of its weaknesses, they used better stone when essential, hard limestone for the slabs paving the Colonnaded Street. Softer fine limestone was used for bases and capitals of columns as it could be more finely carved than the

12 McKenzie 1990.

15 Anderson 2003:7.

'4 Schmid 2001: 379; Colledge 1986: 9-12; Netzer 2003: 112-3; Clarke 1999: 205-25; 2000: 123-7; 2003: 171-5.

'5 Strabo 16.4.2.

sandstones. Marble slabs were used as veneer for walls and in paving the cellas of the temples. It was also used for statues and ring bases, but rarely for column capitals.'6 To modern eyes the colour of sandstone is an attraction, but it was usually stuccoed. In addition to its decorative use to add classical details:7 stucco was also used extensively as a finish on the sandstone walls and column-drums, regardless of the type of sandstone.

The examinations of the quarrying of sandstone for the freestanding buildings and the calculation of the amount of it have thrown light on the adequacy of supply (in III.a). Contrary to what might have been assumed,'" it was discovered here that the primary quarries (those whose sole function was as quarries) would have yielded approximately half of the stone blocks for the freestanding buildings in the city. The other half was obtained from quarrying for the rock-cut monuments, and from levelling building sites. This pattern of use on this scale was not normal elsewhere. Other sites rarely had so many rock-cut monuments, although rock cutting was sometimes necessary at them, such as to level a site or to form a theatre. This stone was probably used in buildings.19 Therefore, the most distinguishing feature of quarrying in Petra, as detailed in chapter 3, is the contribution of the stone blocks for buildings that the rock-cut monuments made. It is unique to Petra and it arose because of the topography of the site.

The very high faces of the primary and tomb quarries of Petra required specific techniques to cut and transport the product (in III.a). The only other case known of quarries with very high faces is at Syracuse.29 But the marble quarries for the Parthenon would have raised some of the same problems.'-' The Nabataean quarrymen, as often was the case elsewhere, used the trench and wedge technique to extract the blocks, and gave the quarries a stepped shape (in III.b). Contrary to what has been suggested,22 carved terraces were used at Petra to create a horizontal working platform, in place of scaffolding. The use of this technique came from the lack of wood for scaffolding. This technique was also used to carve both the exterior and interior of the rock-cut monuments. In small monuments the cliff was cut to a vertical face and the decoration carved at the same time, whereas larger monuments would have required the work to be done in two stages. The reasons for this were practical ones, such as the work space available on the ledge or the requirement of stone blocks for use in freestanding buildings under construction. In some cases, where the extraction took place in the centre of cliffs, the

'6 There is only one capital of marble found at Petra McKenzie 1990: Pl. 39c.

Zayadine 1987:131-43.

Pfltiger 1995: 289; Browning 1980: 168.

19 Martin 1965: 146-55; Orlandos 1968: 15-20.

20 Peschlow-Bindokat 1990: Plates 26-27; Ginouvas and Martin 1985: Pl. 10; Martin 1965: 147-9; Orlandos 1968: 20.

21 Martin 1965: 147, Plate XI12; Korres 1995a, b.

22 Browning 1973: 50; Pfliiger 1995: 292.



quarrymen used ladders, ropes and slots to reach the working level. The sandstone quarries at Petra could not yield single colossal blocks like other harder stones elsewhere. The normal maximum size measures 1.5 x 2 x 0.6 m, and weighs approximately 4 tons (the largest 7 tons, in III.c.1). Notably, although the quarries were located as close as possible to the intended building sites, the stonecutters roughly shaped the stone blocks to reduce their weight in the quarry, as was common at other Greco-Roman sites.23 The weight of the blocks quarried suggests that the Nabataeans used one or more transportation devices to move the extracted stones to the final destination, such as sledges, sledges on rollers, or wagons with wheels.

The Nabataean stonemasons used two different techniques to prepare the stone surfaces: block preparation for freestanding buildings and finishing the surfaces of the rock-cut monuments (in IV.a, b). In both techniques the processes of dressing were carried out in a sequence, and in each stage different tools were used. The Nabataean tools were similar to Greek and Roman ones,24 such as picks, hammers, mallets, bush hammers, and pointed flat and claw chisels. The most notable tool used was the claw chisel which left the diagonal coarse lines as technical feature which is undoubtedly one of the most characteristic features of Nabataean stone masonry. As it is unusual for claw chisel marks to be as visible as they are at Petra, this could be explained by their use as preparation for stucco. A rasping metal tool like a hairbrush or comb, similar to the so-called chemin-de-fer,25 might have been used to carve the fine lines. Rasps or grinding stones were used to smooth the final surfaces. In addition, a drill was used in the preparation of decorative elements.

The properties of the sandstone influenced the size of blocks used for building by the Nabataeans, both in walls and columns. In terms of size, wall blocks (average) measure c. 70 x 30 x 30 cm and weigh about c. 150 kg. The walls of the buildings were built using two methods normal elsewhere: header and stretcher or two-skinned construction26 (in V.a.2). No opus reticulatum has been found in Petra, unlike in the Herodian buildings (end of 1st cent. BC) at Jericho.27 This indicates that the Nabataeans did not use the Roman system of wall construction, nor were there Roman builders involved in Nabataean building projects at Petra. The Nabataeans certainly knew the technique of the use of dowels and clamps (in V.a.3), and the size of the cuttings for clamps recovered at Khirbet et-Tannur suggests Egyptian influence.28 However, the Nabataean builders at Petra concentrated on the use of mortars and avoided the use of

23 Abu Dayyah 2001: 524, 9.

24 Adam 1994: 45-6; Korres 1995a: 76-7, Fig.10.

25 Rockwell 1993: 62, Ginouves and Martin 1985: 74.

26 Tomlinson 1961: 133.

27 Netter 1977: Fig. 12; Fischer 1998: 37. Ward-Perkins 1981: 310 states that opus reticulatum was rare in the East, but was used at Jericho.

28 Arnold 1991- 125-7.

clamps. This was probably because clamps and dowels may exert a concentrated force on a small part of a block, which would be too much for sandstone, causing it to flake or chip. Their unique use was in limestone walls of Khirbet et-Tannur, but not in sandstone walls of Petra. This shows that their selective use of techniques is related to the properties of the stone and their knowledge of them.

The proportions the Nabataeans used for their column drums were dependent on the properties of sandstone. This caused the use of drums with a small height (in V.b.1). Consequently, for large columns the large diameter and a small height of the drums make them like discs (c. 0.7 m high, c. 1.5 m diam.). However, although their height is small, the wooden pin technique was used for centering these Discs. It is also for the Normal column drums of smaller diameter. Alphabetic and numeric symbols were also cut in the bottoms and sides of the columns, as in Greek architecture. The weights of the column drums (Normal drums 350-450 kg, Disk drums 2 tons, and those of the Qasr el-Bint 7 tons), and the weight of the wall blocks indicate that systems of simple and compound pulleys must have been used in lifting them, as elsewhere in Greco-Roman world29 (in IV.d). The use of lifting equipment is also evidenced by the presence of handling bosses, slots on the side of the column drums, and the horizontal grooves in the bottom of some column drums. Notably, the use of slots and horizontal grooves are unusual and could be related to the friability of the sandstone.

The types of material available and their properties were also the essential factors affecting the roofing techniques discussed in chapter 6. The lack of wood and the availability of sandstone presented a serious challenge to the Nabataean architects in roofing their buildings. They concentrated on the use of the sandstone by using technical features, paralleled elsewhere in the Greco-Roman world. Vaults, domes, series of arches supporting slabs, and relieving arches were used in roofing cisterns, rooms, doorways and basements (in VI.b, c). However, the Nabataeans avoided using sandstone for long horizontal beams (lintels and architraves), since it is friable and cannot withstand high tensile forces. This is the best explanation for the absence of the architrave blocks from the pronaoi of the buildings excavated so far. In certain circumstances, as the builders could not use slabs to cover longer intervals than 1.1 m (table in Fig. 6.37), they used wood, probably local juniper and olive timber (in VI.c.2). But the solution to the problem of spanning larger distances involved importing timber, such as cedar, to which they had access to through their trading activities and wealth (in VI.d).

Although Nabataean stonemasons did not deal with colossal single blocks in constructing their buildings, the rock-cut monuments they carved were monolithic units.

29 Coulton 1974: 1-19.

1-1A1-1EK KABAtitti

The larger rock-cut monuments in Petra have substantial dimensions which are unique in antiquity. Of all the construction techniques which appeared and were developed in the Greco-Roman world, Nabataean rock-cut monuments are still extraordinarily distinguished in terms of stone masonry. This is also a product of the availability of sandstone and the local landscape.

No one in the ancient world, outside Petra, ever carved a two-story rock-cut monument on this scale.3 Although the architectural composition of el-Khazneh is found in wall-paintings, it is now generally accepted that the form of it and other buildings in Petra was influenced by the architecture of Alexandria. Fragments of this architecture survive in Alexandria and they are from freestanding buildings. This means the form of the Khazneh was influenced by freestanding buildings, but the technique of carving it from the rock was not necessarily borrowed at the same time. It is worth noting that the techniques used in constructing the freestanding buildings at Petra are completely different from those used in carving the rock-cut monuments. This suggests that the architectural form of the rock-cut monuments, such as el-Khazneh, was borrowed but not the carving technique.

To carve a rock-cut monument is a difficult task. The method of carving a huge facade as one unit full of architectural details is actually much more rigid than carving them separately and then building them, because the stonemasons dealt with elevated surfaces in the former and with blocks on the ground in the latter. As carving began from the top down (whereas buildings were normally designed from the bottom up) the whole design had to be worked out accurately in advance and there was no scope for changes during construction. The rock-cut technique would require a tradition which comes from gradual development. This would have certainly required craftsmen with knowledge of the local geomorphology. It seems more likely in principle to start this development from the simplest monuments and move in the direction of complexity. For this reason, contrary to what Schmid suggested,3I it is possible that carving some of the simpler (and usually smaller) rock-cut monuments in Petra namely the Pylon, the Step, Proto-Hegr, and Hegr type tombs began before the larger ones. This is now clear after the remarkable discovery of the tombs in front of el-Khazneh [treasury], which are classified stylistically as Proto-Hegr type. These were definitely carved before el-Khazneh, probably before the last quarter of the first century BC. Moreover, the quantity of Nabataean rock-cut monuments in Petra and Hegra suggest considerable local activity. The masons32 named at Hegra suggests the presence of family craft development in the society and

3 The Persian tombs at Naqsh-i Rustam only have a massive relief above a lower storey, see Boardman 2000: Pl. 2.27a, b; Fedak 1989.

31 Schmid 2001: 388.

32 Schmidt-Colinet 1987: 143-50; McKenzie 1990: 14-5. They used the family tree from mason's names at Hegra as very important evidence for determining the organisation of the sculptors and their styles (schools).

also show that the rock-cut monuments there were the work of Nabataean craftsman. This is also suggested by the masons' marks in Nabataean letters at Petra. Although the tools discussed in chapter 3 and 4 are similar to Greek and Roman tools, the other technical features such as using horizontal platforms for working and the way of carving would have been developed the Nabataeans' own experience in carving simple rock-cut monuments.

The creativity of the Nabataean masons was significant in overcoming the limits of the physical environment by using a combination of rock-cut and freestanding techniques, the "mixed technique", (in iii.c.2). The use of sandstone inset elements in the rock-cut monuments can be explained because either a fault appeared in the rock or a mistake was made by the masons, while limestone inset elements were used for finer details. In some other cases the architect found difficulty in positioning the whole of his design on the site chosen, for example the Palace Tomb was partly built and partly carved because the natural downward slope of the hill from south to north did not enable the architect to fit the dimensions of the facade on the cliff. The use of the "mixed technique" is also found elsewhere occasionally, as in the Tomb of Absalom in the Kirdon Valley in Jerusalem33 and in tomb 13 at Etenna in southern Anatolia,34 but it is not as frequent or not on the same scale as at Petra.

In addition, a mixed technique was also used in roofing the temples, in which the architect combined the use of the local technique of the terrace roof with the classical pitched and tiled roof (in VI.d.1). This sort of technical combination reflects the innovative ability of the Nabataean architects. Similarly, their architectural style developed from a mixture of eastern, classical, and local features, as in ed-Deir.35

One of the most significant discoveries this analysis of Nabataean construction techniques has revealed is that the earliest examples for some construction techniques survived at Petra. The first is the use of wooden tie-beams across arches (in V.c.3), as later used in Muslim architecture from the Umayyad period36 up to the present. This suggests the possibility of a continuation of the Nabataean architectural tradition into Islamic architecture. Although there is no evidence recovered so far which proves this continuity from the Nabataean to the early Islamic Period, evidence for such continuity has been found from Nabataean to Umayyad sculptures, suggesting that such local continuity was possible.37 In addition, the use of ring bases around columns (in V.b.1), which are unique to Petra in antiquity is later found in the Dome of the Rock in Jerusalem,38 in which, notably, the

33 Fedak 1990: 143, Figs 205-6.

Cevik 2003: Figs. 13, 14, 15.

3' McKenzie 2001: 97; Lyttelton 1974: 79.

3' Creswell 1989: 29.

37 J. McKenzie, personal communication, 2004.

38 Creswell 1989: 29.



use of wooden tie-beams connecting the arches also occurs39 (in V.c.3).

The use of pendentives in constructing the domes of the Baths (in VI.b.4) at Petra is also the earliest known example (lst cent BC), and one might want to speculate from where the technique came. Since there are no earlier Nabataean domes, the possibility of influence from elsewhere arises.40 The location of the sites with, second and third centuries AD, pendentive domes, as well as the Petra one, suggests either Egypt or Mesopotamia as the possible source of influence; not the Aegean where true domes of masonry are rare or Roman Italy where the method of construction is different (concrete) and there are no pendentives in spite of the survival of many domes.

The Nabataeans show more creativity in adapting and finding new technical solutions in the rock-cut monuments than in the freestanding buildings. Sandstone carving as a separate technique is the product of the Nabataeans, because sandstone was the only stone available in Petra. It seems that this high availability created the interaction between their technical learning and the landscape, and led the Nabataeans to be involved mainly in working sandstone. The use of this stone had the same dominant effect on the techniques they used and developed, as marble or limestone had in Sicily, Greece and Italy, or concrete in Italy. This unique aspect brought the masons to use several technical methods in carving.

However, more technical features were adopted for use in the freestanding buildings than in the rock-cut monuments. The simple explanation for this is that the techniques used in the Nabataean freestanding buildings were borrowed from other freestanding ones, whereas in the larger rock-cut monuments they borrowed the form from other freestanding buildings and not from similar rock-cut ones. In the simple rock-cut facades they used Oriental architectural elements, and thus, the technique was developed locally, as is attested by the inscriptions at Hegra.

My final and most important point is that the Nabataean construction techniques show amazing exploitation of the possibilities of the available materials. It seems that the factor of outside contact mainly determined the architectural style, while local materials determined the technique used. Although the Nabataeans borrowed some technical elements through their outside contacts, the building materials available controlled the way they used all of them. More simply, trade routes were good to import ideas, but the available materials changed these ideas. The builders were forced to develop the technical elements borrowed to make them suitable for their own use as the local materials required. Although the technique is borrowed, the result is Nabataean because of the properties of the materials and their limitations. Therefore, a lot of their techniques relate basically to the materials, from which everything came out. Perhaps the most straightforward example of this phenomenon is the use of the orthostates in the Qasr el-Bint (late 1st cent. BC, the earliest freestanding monumental building in Petra). This use could be an initial borrowing from Greek practice (coming with the Greek forms) which was later rejected because the quarries could not provide larger blocks because the stone was not suitable. A further example is the use of timber. It is expensive and so Nabataean builders used stone slabs in roofing. Another example is stone dressing which was not new but how the Nabataeans did it (with diagonal tooling) is unique to Nabataean architecture. All this leads to sum up that the Nabataeans had their own construction techniques, just as they had their own architectural style. It is right that they got some of their techniques from local sources or the Greco-Roman world, but they took them and they developed them in ways specific to Petra.

This thesis has provided a number of contributions to understanding how the Nabataeans established their own construction techniques for both freestanding buildings and rock-cut monuments in Petra. It is hoped that this thesis will be relevant not only to academic work but also to restoration work at Petra and other classical sites as well. The hope from this study is also to be useful to modem architects and designers who are using the classical tradition in their designs, but desire to know more about the techniques used in carving or constructing them.

It is, however, more than that. There are a number of areas on which I would like to do further research to improve the contributions made, using this thesis as the starting point. In focusing on connecting the study of comparisons on the wider Mediterranean world, not all Nabataean buildings at other sites were analysed. One issue for further research is to enlarge this kind of study for more Nabataean, classical sites, and possibly those in Iran (Parthia). Perhaps the most prominent area for further research is one focusing on the calculation of the total quantities of man power required for carving the monuments. This will extend to examining the organisation of supplies and the logistics of the operation. Surveying all the rock-cut monuments to calculate the volume of stone extracted more accurately is vital. That is for the future, and if this study inspires others to attempt to combine quantitative analysis with their more qualitative studies of Nabataean construction techniques, it will have achieved one of its future aims.

39 Creswell 1989: 29.

40 McKenzie (in Orenaration).


By Steve Rudd: Contact the author for comments, input or corrections.


Click to View