The International Umm al-Biyara Project

• Welcome to the International Umm al-Biyara Project
• Outline of the International Umm al-Biyara Project
• Selected bibliography of the International Umm al-Biyara Project
• Preliminary Report on the 2014 Season
• Preliminary Report on the 2013 Season
• Preliminary Report on the 2012 Season
• I. Acknowledgments
• II. Introduction
• III. 2012 excavations – a. Structure 26
• III. 2012 excavations – b. Structure 20
• IV. Conclusions and perspectives
• Preliminary Report on the 2011 Season
• Preliminary Report on the 2010 Season

Preliminary Report on the 2012 Season

by Stephan G. Schmid and Piotr Bienkowski

III. 2012 excavations – b. Structure 20

One of the structures that provided the most intelligible results during the 2010 survey was Structure 20 (ST 20 on fig. 2; no. 2 on fig. 3), standing on the NE edge of the plateau, prominently overlooking the city centre. The results from both the survey season as well as from the first excavations carried out in 2011 strongly suggest this structure was a bathing installation. The structure consists of several clearly visible rooms. From the S end of the building, a partially rock-cut and partially built water channel brings in water to the structure (a on fig. 3). The water is collected in a substantial basin, identified by its greyish hydraulic mortar containing charcoal fragments which improved its waterproofing qualities (b on fig. 3).

These observations were confirmed by the excavation of the water basin during the 2012 season. The interior surface, i. e. the one for storing water, originally measured 3.5 m x 1.5 m, but its height so far remains unknown (figs. 11. 12). Water is brought into the tank at its SE corner. For the time being, only one outflow could be definitively identified. On the SW corner of the tank, a lead pipe measuring 51 cm in length and 3.5 cm in diameter leads the water out of the tank (fig. 11). At this spot, but slightly further to the N, a small stone basin was discovered, measuring 75 cm x 50 cm. The basin has an inflow on its E side and – on a slightly lower level – an outflow on its W side (fig. 13).

It seems, therefore, perfectly reasonable to suggest that the stone basin originally took water coming from the tank. Its actual position a few centimetres further N is most probably due to one of the several earthquakes that hit the region, including the one of 363 CE that was responsible for the destruction of the excavated structures. Therefore, one would expect to find a water pipe taking the water from the stone basin, possible similar to the clay pipe recorded in 2011 which led water to the two individual bathtubs. However, this hypothesis has to be verified in a future season. On a slightly lower level, but on the same alignment as the water tank, is a huge area which yielded a large number of hypocaust and tubuli fragments during the 2010 survey.

For that reason in 2011 we excavated part of a hypocaust room about seven metres N of the water tank. In order to find out more about the space between these two installations, we decided to expose another space, immediately adjacent to last year’s excavation and, therefore, closer to the water tank (c on fig. 3). It turned out that we are dealing with a smaller hypocaust area, with both a floor- and a wall-heating system, measuring roughly 2.5 m x 2.5 m (fig. 14). At this stage, we do not know if this was a separate, and smaller, hypocaust room, or a narrower continuation of the one excavated in 2011.

According to the average organisation of Roman type bathing installations, the first variant would seem the more probable (cf. below). Some of the suspensurae, forming small pillars usually with a rectangular base and then disk-shaped elements, were still standing in situ, and in the collapsed debris of the walls many fragments of tubuli mixed with plaster were found. The hypocaust pillars are built on a layer of large, uncut stones or bedrock, and the space between the pillars was covered with a dense layer of black ash. Between the large stones some pottery belonging to the last quarter of the 1st c. CE was found. This, together with the abundant greyish mortar, would indicate a construction date for the hypocaust room towards the late 1st c. CE. Since the heating system of that small room has a clear entrance in its S wall (top centre on fig. 14), the heat must have been transmitted from this direction and this would be the place to look for the praefurnium of the bathing installation. According to the same logic, the narrower heated area excavated this year should correspond to the caldarium of the bathing installation, i. e. the hot room, while the heated area excavated in 2011 would be the tepidarium, the warm room – if we assume that they are separate areas, which is not confirmed. Also, the supposed praefurnium would be situated very close to the water tank. This would well correspond to the usual layout of Roman bathing installations: Usually, the praefurnium is also used in order to heat water in a metal container [21] Schiebold 2006: 25; Yegül 2010: 91–92; Baatz 1978: 91. and, therefore, should not be too far away from the main water source.

The presumed central room of the bathing installation is a huge hall measuring – in its currently known dimensions – 6.4 m x 6.9 m (d on fig. 3; fig. 15). This major hall features two pillars on its S and N sides. Already last year we supposed that the hall underwent major changes during its use. Indeed, this year too there are some indications that the pillars may not belong to the original layout of the hall. It is obvious that none of the pillar stones bond into the wall behind them. Each pillar simply butts against the wall and, therefore, must belong to a second phase.

This puts into question how the room was originally roofed, if at all. It would seem very difficult to cover a space of 6.4 m x 6.9 m using the construction techniques known in the area and without arches or pillars. Therefore, it seems possible that the huge hall was, in fact, an open courtyard in its original phase. In a second phase the pillars were added and these most probably supported arches. This is indicated, among other evidence, by the order of the collapsed debris as shown in fig. 16. The fallen stones clearly allow for tracing the arches that once covered the hall and helped to support the roof and the first floor.

The arches were covered by impressively large stone slabs, some of them remaining almost complete, although broken, when the room collapsed (fig. 17). They can reach up to 2 m in length. In this room too, the massive collapsed debris indicates a violent destruction that most probably can be assigned to an earthquake (fig. 18). In any case, pillared rooms such as our central hall in its final stage do occur elsewhere in bathing installations in the wider region, as may be underlined by comparable halls in the western bathhouse in Beth-She’an [22] Skythopolis; cf. Hoss 2005: 131 cat.no. 21 and corresponding figures. .

At the N end of the building, two rectangular structures measuring 90 x 140 cm and 160 x 220 cm respectively were clearly visible on the surface already in 2010. During the 2010 season cleaned the smaller one which turned out to be a bathtub. During the 2012 season, we excavated the second, larger bathtub (e on fig. 3). Already in 2011 it was observed that the water from that bathtub was used to flush the adjacent latrine. To that effect, a small channel leads from the lowest point of the bathtub in its NW corner into the defecation channel of the latrine. In order to augment the pressure, the water was led through a lead pipe that was bedded into a clay pipe, which itself was bedded into hydraulic mortar. The bathtub is constructed from well cut sandstone slabs that form the floor as well as the walls of the structure, that are bedded into thick layers of greyish hydraulic mortar (fig. 19).

The facts that the two bathtubs are not aligned, neither constructed on a common level, and that apparently stones formerly belonging to other contexts were reused for their construction, all indicate that they belong to a later phase of the building. They were definitely in use when the whole building collapsed in the earthquake of AD 363. This can be confirmed by some interesting discoveries. The sequence of collapsed material within the second, larger bathtub was very clear. On top came the architectural debris related to the collapse of the building, followed by fragments of wall plaster and, more or less directly on the floor slabs of the bathtub, a series of marble fragments belonging to the sculptural decoration of the building (fig. 20). Interestingly, several of the fragments discovered in the larger bathtub excavated in 2012 belong to sculptures already found in 2011. This is the case with the lower right leg that fits the torso of a small boy holding a water jar (fig. 21). As pointed out elsewhere [23] Schmid, Fiema, Bienkowsky and Kolb 2012: 81; Schmid and Bienkowski 2012: 258. , this type of a basin or fountain figure is well known from villas related to the Roman upper class [24] On the type see Kapossy 1969: 41–42. . Such is the case with a statue of the same type found at Rome within the so-called horti Lamiani and, therefore, certainly belonging to a luxury installation of a senatorial or even the imperial family [25] Häuber 1986; Häuber 1991: 106–107 no. 51; 287 no. 213. .

A quite interesting connection to one of the pleasure villas of the Roman emperors can equally be drawn: From Domitian’s villa at Castelgandolfo near Rome comes an architrave decorated with relief figures of Venus and small Erotes [26] Liverani 1989: 48–51 no. 19. . One of the Erotes is depicted in exactly the same type as our statue from Umm al-Biyara [27] Liverani 1989: 50 with n. 55. . Another reflex of this statuary type of a boy with a water jar can be found on a Roman sarcophagus, found near Rome and now conserved in the Louvre in Paris [28] Baratte – Metzger 1985: 49–55 no. 15. . The front side of that sarcophagus is depicting the myth of Aktaion who surprised the goddess Artemis while taking a bath and who was punished by the goddess by being killed by his own dogs. The relief scene of Artemis taking a bath makes use of several well known statuary types: Artemis is depicted using the famous „Crouching Venus“ type; behind her stands a Eros gushing water on her back from a jar and he is depicted in the same type as our statue from Umm al-Biyara. The scene on the sarcophagus probably gives a quite good idea of the manifold possibilities such statues could be used for in fanciful Roman villa gardens.

The type of our statue shows a strong concentration around the city of Rome and more generally in Italy. Only a few examples have been found outside Italy, for instance on the island of Ibiza, within the thermae of Thuborbo Maius and Timgad [29] Häuber 1986: 87. and recently two more statues were discovered in the baths of Sagalassos [30] Only one is illustrated so far, cf. Waelkens, Poblome and De Rynck 2011: inner dust jacket. . Our example would be only the sixth statue found outside Italy.

Additional fragments that belong to another small torso discovered in 2011, possibly depicting a statuette of a Herakliskos, were also found within the bathtub excavated in 2012 (fig. 22). Most probably, we are dealing here with a variant of Herakliskos (young Herakles) modelled on the Herakles Farnese type [31] On the type see LIMC IV 1 [Zurich/Munich 1988] 786–787 s. v. Herakles [O. Palagia]; Wrede 1981: 242 no. 128. .

Interestingly, in the above cited horti Lamiani in Rome, one of the good comparisons for our statue type was found, although in a slightly bigger scale [32] Häuber 1991: 100 cat. no. 35. . It seems, therefore, that the combination of fountain and other decorative garden sculpture as witnessed in the bathing installation on top of Umm al-Biyara corresponds very well to the standard decoration of wealthy Roman villas and palaces.

In addition to the fragments found in the huge bathtub, also within the main hall of the bathing installation some fragments were discovered, i. e. in the same room as the sculptures found in 2011. A substantial part of the back head and of the water jar (figs. 23. 24), obviously belonging to the statue of a boy holding a water jar mentioned above, were discovered near the collapsed arch on the W side of the hall. On the other side of the hall, near the E arch, the lime stone relief panel with the depiction of a bearded male deity, wearing a cloak upon his left shoulder and holding a spear or a sceptre on his left side, was found (fig. 25). The relief measuring 21.2 cm x 19.0 cm x 10.1 cm corresponds very well to larger examples found elsewhere in Petra and is most likely depicting Zeus/Jupiter or Serapis. Normally, such reliefs go by pair and are flanking doors.

If we try to apply concrete functions within a typical bathing installation to these rooms and structures, the main hall with the adjacent basins (d and e on fig. 3) should be interpreted as the frigidarium, the cold water bath, while the not yet excavated space south of the main hall would be the apodyterium, the room where the users of the bath get undressed.

As in 2011, we were able to carry out a small sounding beneath one of the floor slabs of the huge room. Despite the small number of sherds found, some painted Nabataean sherds may give a chronological pointer to the construction of the installation. The painted sherds cover the 1st c. CE [33] Phases 2c, 3a and 3b after Schmid 2000. , the latest one being one single sherd of phase 3b, ranging from ca. 70/80 to 100 CE. Since this picture corresponds very well to the results from comparable soundings carried out in 2011, we would like to propose a construction date for the bathing complex towards the end of the 1st c. CE. It remains to answer the questions as to when the different changes occurred that we observed, such as the construction of the bathtubs or the additional construction of the pillars and arches.

For the time being it is not yet possible to give precise answers, but only some hints. In what first appeared to be a small room immediately E of the larger bathtub (f on fig. 3), it was again shown that the bathtub is a later addition. Clearly visible on fig. 26 is the fact that the two lowest courses of the E wall of the larger bathtub (top on fig. 26) do not bond with the major N wall of the huge main room of the bathing installation (left on fig. 26). Only the third course of that wall is bonded, and that only at the spot where the earlier N wall of the main room was cut back. The lower N wall of the presumed room does not bond with any of the other walls and must have been a later addition. The presumed small room E of the huge bathtub turned out to have served as a kind of rubbish dump. Two main layers filled it from bottom to topsoil, the lower one consisting of yellowish earth, densely packed and containing smaller stones, the upper one being of greyish and ashy consistency (the upper limits are clearly visible on fig. 26). In particular, the upper layer contained a huge amount of pottery, even more impressive when taking into account the small size of the structure, measuring only 1.4 m x 1.6 m. Besides the fact that the lower layer contained less pottery, there was no difference in the composition and chronology of the pottery between the two layers. Both showed a majority of painted pottery belonging to phase 3c, with some sherds belonging to phase 3b, but no other painted pottery was recorded (fig. 27). This would indicate a date somewhere around 100 CE. For example, the massive destruction layers from the Nabataean houses on az-Zantur, tentatively identified as being the result of the Roman annexation, showed exactly the same composition [34] Schmid 2000: 34–35. 141–146. . No matter whether the dumped material from Umm al-Biyara can be related to a historical event or not, the fact that so much pottery was thrown away might indicate that some changes occurred to the building. Also, the frequently observed use of greyish hydraulic mortar for the construction of the two bathtubs would support a date within the early 2nd c. CE. The pottery from the above mentioned dump contained almost exclusively painted and unpainted Nabataean fine ware. A remarkable element are several fragments of at least five different unguentaria (unguent flasks) (fig. 28). Their presence in a bathing installation is of course perfectly logical, since after the different cold and hot baths people would use perfumes before getting dressed.

To a presumed last phase before the earthquake of 363 CE belongs another change in the structure. In front of the small bathtub a tiny rectangular „room“ was constructed, leaving a small entry space open towards the NW pillar of the room (fig. 29). Whether this was an attempt to create more privacy, a kind of a separate apodyterium for a specific individual, or whether it was an attempt to create a smaller space easier to heat, cannot be decided at the moment.

Although the overall plan and layout of our bathing installation are not yet completely known, a few preliminary observations can be added. For the time being, one would tentatively ascribe this installation rather to the block type of Roman bathhouses [35] For a short definition see Hoss 2005: 30–32. . Quite good parallels can be found for instance in Hazeva, Mezad [36] Hoss 2005: 146–147 cat.no. 59 and corresponding figures. dating to the 3rd and 4th c. CE. There, the overall layout is comparable as well as the positioning of specific units, such as the pools/bathtubs and the latrine. In general terms, the block type of bathing installations is more prominent in military and in private baths compared to public bathing installations [37] See Brödner 1983: 179–197; for a detailed study of private baths, including many good parallels, see Haan 2010. . This definition may, however, also include private baths of the highest level, as can be underlined with the thermae of emperor Domitian’s villa at Sabaudia [38] Mielsch 1997: 71 fig. 43; Lafon 2001: 368. 371 fig. 96; Jacopi 1936. . There, the thermal complex clearly belongs to the block type and is arranged in a similar way as on Umm al-Biyara. In general terms, more privately used baths have a certain tendency to belong to the block type, while public baths would manifest other plans, more linear, in order to force visitors to follow an obligatory sequence and to optimise visitors’ use of the installations. Therefore, also the baths near the „Teatro Marittimo“ in the famous villa Hadriana, clearly a rather private sector of the complex, belong to this type [39] See Verduchi 1975 passim and especially 92–95 for comparisons with other structures, including again Domitian’s villa at Sabaudia. .

The chronological elements we have so far for our baths, i. e. a first building phase in the last quarter of the 1st c. CE, followed by at least one manifest remaking, as well as smaller changes, and a massif destruction during the 363 CE earthquake, are well paralleled by the chronology of the bathing complex adjacent to the „Great Temple“ in the city centre of Petra [40] Joukowsky 2007: 96–98. . The major difference consists in the fact that the baths in the city centre were somehow reused afterwards while for the baths on Umm al-Biyara we don’t have such evidence fort he time being.