Click on small images to see larger version.

The first season of the Wadi Ramm Recovery Project ran from August 2nd through 23rd, 1996. Funding for the project was provided by the University of Victoria Alumni Association, the Archaeological Society of British Columbia, and the generous donations of private individuals. In addition, Barbara Reeves received a Von Rudloff Travel Scholarship and assistance from the Graduate Students' Society of the University of Victoria. The Department of Antiquities of the Kingdom of Jordan provided salaries for the workmen and the use of a lorry for transportation of finds at the end of the season. ¹

Project directors were Dennine Dudley and M. Barbara Reeves, both of the University of Victoria. Luay Mohamadieh served as Department of Antiquities representative and draughtsman. The ceramics were analysed by Dr. Khairieh 'Amr, also of the Department of Antiquities. Naif Zeben Ahmad carried out the cleaning of the painted plaster and the bronze figurine. The Wadi Ramm Recovery Project ran as a collaborative effort with that of Laurent Tholbecq (University of Louvain and IFAPO) who conducted a simultaneous examination under a separate permit of the Nabataean temple and the complex behind it.

The purpose of the Wadi Ramm Recovery Project was to document two structures at the foot of Jebel Ramm, described as a Nabataean house and bathhouse (hereafter referred to as the Eastern Complex). The Eastern Complex is located west of the Government Rest House of the modern village of Ramm; it is situated on the slope of the foothill of Jebel Ramm just east of the Nabataean Temple (pl. 1). The structures were cleared out by the Department of Antiquities in the 1960s, but any records have since been lost. According to Mohammed Ataiyk, a local man who participated in this effort, the excavations took place in 1964 and involved a number of areas in Wadi Ramm besides the Eastern Complex. The interior walls of the two buildings had been exposed by this work, but the lines of the outside walls were still lost in the sand. As stated above, there is no documentation concerning the site; the only exception is two photographs from the Department of Antiquities Archives, one showing the excavation team and the other showing general work in progress. Additional probes excavated by Kirkbride in 1959 have only been published in vague terms; her report may refer to the Eastern Complex or the structures west of the temple. ²

Wadi Ramm has been associated with an important commercial site mentioned in Pliny (NH 6.32.157). The geographer Ptolemy (Geog. 6.7.27) includes a settlement called Aramaua in his list of inland cities of the northwestern corner of Arabia, which is considered to refer to Wadi Ramm. The site has also been linked with Iram, the "many-columned city" (Quran Sura 89:7). There are numerous trans-literations of the name of this site: Ramm, Rum, Rhum, Iram; we have followed the present government designation in our choice of spelling.

Pl. 1. General view of site looking west, Eastern Complex in Centre, Nabataean temple behind (Photo: Dudley)

The site is not specifically identified in any historical documents and only information from the archaeological and epigraphical evidence provides any clues to the date of the structures there. The only structures at Wadi Ramm which have been properly documented are the aqueduct and shrine at Ain Shellaleh and the Nabataean temple. The temple was first excavated by Horsfield and Savignac in 1934 and was later worked on by Kirkbride. ³ On the basis of inscriptions in the area the temple has been assigned to the goddess Allat, a principal deity of the Nabataean people associated with water supply. The date of the temple is now assigned to the reign of the Nabataean king Aretas IV, in the first century B.C. or early first century A.D. ⁴ The presence of the Eastern Complex was noted in the first archaeological surveys of Wadi Ramm; Savignac notes the presence of "ruins" on his 1932 map, and by 1934 some of the wall outlines had been measured and drawn. ⁵ Although Savignac and other scholars had noticed the presence of the Eastern Complex in the course of their surveys and excavations, no valuable information regarding it has previously been published.

The 1996 field season of the Wadi Ramm Recovery Project began on the 3rd of August with a general examination of the site. It was observed that the complex had sustained damage since the previous summer, presumably in the November 1995 earthquake; some wall faces had collapsed and a number of blocks had fallen. Consolidation of the structure is urgent. The floor levels (both paved and plastered) and the thresholds, are the first priority for conservation. The plaster which remains in situ also requires attention. All of the walls are in need of protection, since the sandstone is weathering at an alarming rate. The damage sustained by the Eastern Complex as a result of earthquake shock and weathering proves how fragile the complex is.

Prior to our excavation season a map of the ancient structures in the area had been prepared by a survey team from IFAPO. The mapping included everything from Ain Shellaleh to the Eastern Complex. Numerous points on the structures were surveyed in, and the data were used to generate dotted outlines of the structures. Since the architecture was further defined during the excavation, however, the computer map was only useful as a general reference. Any top plan generated strictly from this information would be inaccurate. Through the use of a sketch plan and computer print-out as a base, each room in the Eastern Complex was designated with a capital letter for purposes of mapping and recording (fig. 1), ⁶ and the division of responsibility within the structures was discussed.

Since part of our research design was to assist the development of the site for tourism, decisions were made about cleaning up the site. It was decided to consolidate all of our dump with the northern dump from the 1964 excavation to facilitate later removal and to keep it out of sight of the approach to the site. A low wall around the site was begun using tumbled stone; this will delineate the site and restrict the access of vehicles. Piles of dumped rock already around the complex were removed or pushed back.

The Villa ⁷

The part of the Eastern Complex renamed "Villa" consists of two rectangular structures (A, B, C [16.5 x 7.5 m], and H, I, J [13.8 x 6.8 m]), as well as the two paved rooms/courtyards D and P. One 5.0 x 2.0 trench (Square 01) was excavated in the junction of walls W08, W09, W11 and W12. There are traces of wall lines around the periphery of the structures, but these features await future excavation. Walls are preserved to a maximum of about 1.5 m; the upper stones had been removed and/or dumped in the previous excavation.

The western rectilinear structure (A, B, C) was excavated first. Room A was originally separated from Room B by two piers, which created three doorways between the rooms. Rooms B and C are linked by a doorway at the western end of W05; a second doorway in the eastern end of the same wall is possible, but the weathered and tumbled condition of the upper courses here makes interpretation difficult. No doorway was located in Wall 06 which would have linked the structure directly with the paved courtyard in Room D.

Wall construction in the western structure is of large regular blocks laid mostly in stretchers mortared together with a crumbly grey mortar and with some cobble courses. The eastern end of W06 and the western face of W09 in Room C are constructed mainly of cobbles and are probably later mending. The blocks in the piers (W03) between Rooms A and B are carefully cut (pl. 2). On the northern side of the piers the blocks are laid in ashlar courses and show typical Nabataean dressing; on the southern side of the pillars, however, the construction appears the same as that in the rest of the structure with cobbles and mortar. In the later phase also, the northern façade of W03 seemed to have been of primary importance since the blockage (W04) between the piers is set flush with the northern face, but not with the southern. The blockage is not retained to the same height as the piers, but it is unknown if this was the original intent or the result of the 1964 excavations. Blocks which had fallen from the W03 piers over the winter were replaced, since their position was known and their location is stable; they should be anchored, however, or another tremor will dislodge them again.

Pl. 2. W03/W04 looking south, stone pillars and later phase blockage (Photo: Dudley)

A probe (Square 01) which was carried out in order to explain the association between the two rectangular structures revealed a grand threshold (A01) in the entrance to Room B, which appears to be the only entrance into the western structure (pl. 3). The threshold is 1.38 m wide and consists of a low step running between the door jambs and a wide level (0.05 m below the step) which retains sockets in the centre and the inside corners for the door supports. The elevations of the thresholds to the exterior and to Room A are the same. The level adjacent to the threshold included a mix of ceramics from the first to the fourth centuries A.D., as well as numerous fragments of bone and wall plaster. A cobble layer adjacent to the upper part of the foundation revealed only 10 sherds dating from the first to the early second century A.D. The soil beneath yielded only a single bone fragment at 0.50 m of depth. No flooring was found, however, so it is possible that the stratigraphy is disturbed. Room B served as a central hall for the structure; all interior doorways open out of this room. The offset entrance and the grand construction make it possible that this building was subsidiary to the Villa's more linear eastern structure.

Pl. 3. A01 doorway into Room B, showing threshold and exterior paving (Photo: Dudley)

The eastern rectangular unit consists of Rooms H, I and J. The structure lies alongside the western building and is separated from it by corridor A04. Each building shares doorjamb blocks with the corridor, and thus they must date from the same construction period (pl. 4). The eastern structure was entered by a doorway (now blocked 1.40 m wide) centrally located along the southern wall (W12) of Room H. This door is matched by a doorway (also blocked, 1.39 m wide) in the centre of W13. These centrally placed doorways are more characteristic of public structures. The upper courses of the outside/southern side of W12 are set back 0.24 m from the face of the lower two courses and form recessed ledges on both sides of the doorway (pl. 5). These ledges may have carried decorative panels, or provided visual interest in themselves, thus creating a grand façade. The probe (Square 01) showed that the area south of Room H was paved with large unshaped cobbles; excavation in Room G is required to determine the extent of the pavement. W14 abuts W11 and W16 and therefore dates to a later phase of the structure. This wall seems to have been built to decrease the size of Room I since there is no trace of a doorway in Room J.

Pl. 4. A01 to left, A04 at top; note that the corner block serves as a doorjamb for both entrances (Photo: Dudley)

Traces of a flagstone pavement were visible in Room D prior to the clearing. Excavation showed that the floor was in good condition; while many of the pavers are cracked or slipped, not very many are missing (pl. 6). It was discovered that W01 does not extend north along Room D but forms a corner with W06. W09 also forms a corner with W06 and with W11 creates a corridor (A04 1.23 m wide) which provides access to Room D from the southern side of the Villa. A second corridor (A03, 0.82 m wide) leads west off Room D between W06 and a newly discovered wall, W10. The whole northwestern corner of the Villa appears not to have been previously disturbed and awaits future excavation. The corridor (A04) leading to the south would also be interesting to excavate, since it appears to be the only area within the Villa that was not cleared out in the 1964 project; this exploration cannot take place, however, until the exterior faces of the walls are stabilised.

Pl. 5. Room H, exterior from south; note ledges on wall at either side (Photo: Dudley)

Room P is a paved area to the north of the bath-house. This courtyard is reached by a series of doorways from the south which offer access first into the corridor O and then into Room P itself. The door jamb blocks are part of the original construction of the Bath and the Villa and thus indicate simultaneous construction of the two buildings. Although the walls around the pavement are retained only to a height of 0.65 m, it is unlikely that P was an open courtyard, since the southern access seems to have been so carefully controlled. The pavement is bounded by walls W16, W19 and W20 and extends north past the line of W21 towards W15, which forms the other edge of the room. The central square of the courtyard is set approximately 0.05 m below the surrounding pavement (pl. 7). More definition is also required to explain Room S, which seems to be a corridor but does not appear to connect with corridor O.

Pl. 6. Room D, pavement looking south (Photo: Dudley)

Since the structures had already been cleared out, we were not expecting many finds; however, a number of interesting pieces were recovered. Fragments of decorative plaster were found in all rooms of the Villa. Along with numerous pieces in the fill, plaster was discovered in situ on corners and along the lower parts of some walls. The plaster can be dated to the earlier phase of the structure since, for example, it is clearly visible between the piers and the later blockage of W03/W04. Many of the plaster fragments are moulded, and a number of pieces are painted in black and gold (Buckets 960017 and 960019). Carved stucco in various geometric patterns was observed at the temple, and the wall plaster there was painted, as is recorded in Savignac's reports. ⁸

Also of interest is a significant number of seashells which were found in the area between W21 and W15. Similar shells were found in quantity near the principia of the Roman fort at Humeima; there it was suggested that the shells represented the remains of luxury foods supplied to the officers of the fort. ⁹ Shellfish, being imported by pack animal from Aqaba some 40 km away, are a luxury food item, a fact which is in keeping with the nature of the Villa complex.

Pl. 7. Room P, pavement looking north (Photo: Dudley)

The most significant find which was recovered from the Villa was a small bronze statuette of Venus (Registered Find 960010.01, pl. 8). The statue was recovered during the cleaning of W13 in Room H, 2.83 m from W11. The figure stands 7.5 cm high and weighs 66 grams. She stands with her weight on her left foot and the right knee bent. Her hair, parted in the centre, is knotted on top of her head and again at the nape of her neck; she twists her hair with her right hand and one lock drapes over her left shoulder. The figure is missing one foot and the left hand, which may have held a mirror, ¹⁰ but is otherwise in good condition. The figurine is of a common type which has its origins in the Aphrodite of Knidos. No exact parallels have been identified, but a similar figurine has recently been recovered in a first-century or early second-century A.D. house in Sagalassos, Turkey. ¹¹ Such Venus figures are not unknown in Nabataean contexts; Negev reports that a bronze Venus statue was found in the Oboda excavations. ¹² The Nabataean deity Allat took on the attributes of Venus because of their respective associations with water; thus our figure may represent the deity to whom the nearby temple was dedicated.

Pl. 8. Bronze statuette of Venus (Photo: Dudley)

It is clear that the Villa was not a habitation of the type implied by the previous designation of "house." Buildings of a more domestic nature may possibly be found in the "village" below, which also awaits excavation. The axial plan of the eastern rectangular structure and the arrangement of the corridors are unlike any of the published Nabataean residences. The grand nature of the architecture and the finds recovered therein indicate that this part of the Villa originally served a more public or official function. ¹³ Many questions remain to be answered, especially concerning the Villa's relationship with peripheral areas such as E, F, and S and the specific function of the complex. Only future excavation can resolve these issues. In a recent publication, the Near Eastern villa is described strictly in terms of imitation of Roman building practices and as "rare" before the fourth century A.D. ¹⁴ The Wadi Ramm Eastern Complex is proof of elaborate secular architecture in the Near East pre-dating the "boom" ¹⁵ in Roman-type villas and may provide clues to the indigenous architectural tradition.

The Bath ¹⁶

The bath-house is located in the southeastern quadrant of the Eastern Complex. The full extent of the bath-house cannot be known until full excavation is completed, but it is clear that the bath's central core consists of Rooms N, R, Q, W and L/M, all of which are linked by doors (fig. 1). Rooms K, T, U, V, X and Area D were probably service or ancillary areas.

All of the bath's rooms were cleaned by removing the large and small tumbled stones and hoeing the soil flat (the exceptions are Area D, where no stones were removed, and Room X, which contains the collapsed outside wall of Room R and from which only small stones were removed). In Rooms M, R, and W, which underwent further excavation, the blocks which might be used in future conservation work were labelled and stored in a lapidarium. In the other rooms, the blocks in good condition were lined up against the wall they fell from. In all rooms, irregular undressed blocks were sent to the dump pile.

Room W (the calidarium)

Fig. 1. Sketch Plan of Eastern Complex

Room W is located on the eastern side of the bath's central core. This placement is contrary to the recommendations of Vitruvius, who advised that heated rooms be placed on the west or south where they could benefit from the afternoon light and heat (5.10.1). At Ramm's bath, however, the close proximity of Jebel Ramm meant that the whole complex was in full shadow by mid-afternoon and hence unable to benefit from a west-facing calidarium (pl. 1). The placement of a calidarium on the east can also be observed in some of the (probably contemporaneous) Herodian bath-houses, though shadow would not have been a causal factor in their design. ¹⁷

Room W is 4.40 m N-S x 4.58 m E-W (pl. 9). The walls are badly worn, but Nabataean-style dressed sandstone blocks seem to have been used throughout. Most of the blocks are large and are generally fitted tightly together in level courses, though small stones were occasionally used for levelling. Between the two faces of a wall was mortared rubble fill. Each wall is ca. 0.83 m thick. The walls now stand 0.72-1.48 m above the room's floor.

There were three rectangular grooves for flue pipes in the southern wall, two in the eastern wall, and two in the northern wall. The western wall contained no grooves. Some of the grooves were carved into the centre of wall stones and some near the sides. This may imply that the grooves were carved after the walls were raised. The wall grooves vary greatly in dimensions because of weathering. One of the best preserved is 0.15 m wide x 0.08 m in diameter. ¹⁸ A probe (described below) revealed that the vents extend 0.54 m below the top of the floor surface into the hypocaust and end in a shelf (i.e., at an uncarved stone block). A vertical slab of rock (0.24 m thick) covered the outside edge of the vent (pl. 10).

Square projections (ca. 0.41 m in length and width) were built into each corner of the room (continuing below floor level). These were possibly associated with the system of roofing, of which no other traces remain. The western wall was covered in at least two layers of a hard, sandy, off-white plaster and traces of the same wall plaster adhere to the corner projections. The northwestern corner projection retains traces of a thin mud plaster directly over the wall stones; this is consistent with Negev's description of Nabataean plastering technique. ¹⁹ On the southeastern projection the wall plaster extends halfway down the lowest layer of the plaster flooring, indicating that the walls were plastered before the floor was laid.

Pl. 9. Room W, floor and basin (Photo: Dudley)

The room's floor has been greatly damaged, but intact sections of it curve through the room (pl. 9). The floor extended right up to the plaster on the western wall but stopped 0.10-0.12 m short of the eastern wall. This, coupled with the grooves in the other walls, hundreds of fragments of box flues in the room, and a broken box flue (tubulus) found in situ on the eastern wall, suggest that the western wall (and possibly the northern and southern walls) would have been lined with box flues. The lowermost box flues presumably sat half on the stone slabs beneath the floor with their other half open to the hypocaust. One long face would have been mortared to the stone wall and the other face coated in plaster to form the inner surface of the room. No complete box flues were found, but from the fragments it is possible to surmise that a typical box flue would have been 0.29 m high x 0.14 m long x 0.10 m wide x 0.007 m thick. Irregular, oblong vents were cut into the short sides. ²⁰ The external faces were not scored.

The calidarium had a plaster floor, 0.17 m thick, laid in 6 layers. The top layer is orange hydraulic plaster. Beneath this are two layers of light grey plaster, another layer of orange plaster, another layer of grey plaster and, at the bottom, a layer of irregular cobbles set in light grey mortar. The two surfaces of orange hydraulic plaster (which should be the upper surface of the floor) suggest that the floor was relaid once. The floor of the room extends up to the lip of a rectangular built-in basin.

The basin was positioned along the northern wall of the room (pl. 9). It has an internal size of 2.01 m x 0.76 m and is framed by a raised lip 0.20-0.29 m wide. The lip was constructed from flat cobbles and tiles or bricks set in a light grey mortar. The original height of the lip is not known, but one section along the northern wall is preserved to a height 0.28 m above the bottom of the basin. In this section, flat tiles (ca. 0.03 m thick) were laid in courses with 0.025-0.035 m of the light grey mortar between; four of these courses are preserved. On the interior vertical face of these courses is a layer of grey mortar (ca. 0.03 m thick) faced with orange plaster. The interior of the basin was constructed of light grey plaster covered (inside and out) by orange hydraulic plaster. A circular hole 0.20 m wide apparently chopped into the eastern end of the basin reveals that the grey plaster undercoat of the basin was 0.17 m thick. The bottom of the basin is currently recessed beneath the floor level and uneven, with its deepest part in the southwestern corner, but since most of the upper coat of orange plaster has eroded, the original depth and evenness of the basin is not known.

The whole room was cleared down to the (highest) floor level. Then, in order to determine the construction of the floor (described above) and the hypocaust, a small probe (1.00 m E-W x 1.50 m N-S) was excavated through and beneath the floor. This probe was located in the southeastern corner of the room (abutting the northern side of the southeastern corner protrusion). The probe incorporated an almost intact section of floor, a vertical hypocaust vent, and the location of the in situ box flue. The probe revealed that the plaster floor sat on top of a layer of large flat stone slabs. Two complete stone slabs (plus the edge of another) were exposed in the probe. The slab abutting the corner projection is 0.33 m wide x 0.55 m long and the slab north of it is 0.51 m wide x 0.58 m long; both are ca. 0.13 m thick. There is a gap between the slabs and the wall to allow hot air to enter the box flues which partially rested upon them. The slabs were burnt on the bottom from the hypocaust gasses.

Pl. 10. Room W, top view of hypocaust probe, facing east (Photo: Dudley) Pl. 11. Room W, side view of northwestern pila in hypocaust probe; note two-stage capital (Photo: Reeves)

When the two stone slabs were lifted, four hypocaust pilae were exposed (pl. 10). The narrow slab next to the projection rested on just two pilae but the other slab's corners rested on four pilae in the common manner. The pilae were made of sandstone. The southeastern, southwestern and probably the decayed northeastern pilae had capitals consisting of a single rectangular block. Of these, only the south-eastern capital was fully exposed. It was 0.35 m N-S x 0.32 m E-W x 0.18 m thick and sat over a pila base 0.33 m N-S x 0.25 m E-W. In contrast, the northwestern pila had a two-stage stepped capital (pl. 11); the first stage was 0.07 m thick and the second was 0.20 m thick. One-stage capitals were possibly used around the walls and two-stage ones in the centre of the room. The pilae are at least 0.78 m high, but their bottoms could not be reached since the probe became too cramped to excavate.

The construction of the hypocaust suggests that this is a very early bath-house. Stone pilae are very rare and their presence is often indicative of an early bath. The use of architectural ceramics (such as tile) did not become widespread in the Near East until Roman soldiers moved into a territory. ²¹ Thus Near Eastern baths built in the first century B.C. until the first century A.D. were usually constructed with stone pilae (sometimes in combination with tile ones). ²² For the purposes of dating Wadi Ramm's bath, it is also significant that Nabataean dressed stones (in original placement, not re-use) were used throughout the bath, including in the hypocaust construction. This strongly suggests that this bath was constructed by the Nabataeans before their territory was annexed by the Romans in A.D. 106. The high concentration of first-century A.D. and first-century B.C. ceramics throughout the Eastern Complex supports such a date. ²³ Moreover, the type of stone used for the pilae also supports this dating. The pilae were constructed of sandstone, which is not a fire resistant material. Pilae which are not fire resistant tend to be early. ²⁴ It thus seems highly likely that Wadi Ramm's bath was constructed in the first century B.C. or first century A.D., earlier than any known hypocausted bath in Jordan. ²⁵

Part of the bath's hypocaust may also have been built in tile during the original construction or during a later renovation. Broken kiln-fired tiles found in the disturbed soil above the floor of this room may have come from broken installations such as the basin, from the (not yet located) furnace, or from tile pilae. At baths in Masada and lower Heriodium, tile and stone pilae were used concurrently. ²⁶ Two fragments of circular tile (ca. 0.19 m in diameter, 0.057 m thick) and many fragments of rectangular tile were found in Room W. One complete tile (0.210 m long x 0.099 m wide x 0.017 m thick) closely resembles those used in the hypocaust of a Nabataean villa-bath of the first century A.D. in Wadi Musa. ²⁷

The only door to Room W was 0.98 wide and was located in the centre of the western wall. It was blocked in a later period (before the wall plaster fell from the doorway's walls). The reason for this blockage (which cut off Room W from the rest of the structure) is not known, but it may be contemporary with the blockage of the door between Rooms Q and R, the blockage of the western doorway in Room N, and the establishment of a cooking area in Area M. At this stage, with Room W blocked, the building probably no longer functioned as a bath-house.

Room R

Pl. 12. Room R and corridor between Rooms R and N, after cleaning (Photo: Reeves)

Room R has a circular interior (pl. 12, 3.40 m in diameter) and a square exterior. The internal walls are constructed from Nabataean-dressed blocks with concave interior faces. The care taken in carving these blocks, in fitting them tightly together in level courses, their size (up to 0.91 m long x 0.37 m high), and the fact that no irregular blocks were used and that the diagonal dressing always runs in the same direction (bottom left to upper right), supports the belief that this bath-house dates from the site's first (Nabataean) phase. Up to five courses of wall blocks remain. Wall plaster up to 0.03 m thick still adheres on the room's interior. The interior faces of the walls sit on a circular (Nabataean-dressed) rim which extends 0.05 m out from the wall towards the centre of the room. On the western and southern sides of the room are two short corridors leading into Rooms N and Q respectively. The rim stones continue into the corridors and become the frontmost flagstones. Thus the top of the rim is level with the flagstone corridors. It is also probable that the robbed-out floor of this room was at the same level. The floor was probably made of flagstones like the corridors. One possible flagstone (0.38 m x 0.35 m x 0.05 m) with a convex face was found in the room in a disturbed location. Traces of light grey plaster (0.002 thick) on top of the rim may suggest a plaster surface covered the flagstones.

On the basis of arrangements in other baths, it was hypothesised before excavation that this circular room might be a heated tepidarium or sweat room. ²⁸ However, a small probe (1.20 m x 0.75 m) along the eastern wall and overlapping the southern corridor revealed no hypocaust. Instead, the room's foundation begins directly below the assumed floor level. The foundation is composed of large (undressed) irregular stones and cobbles in firm light brown soil. The foundation is deepest (1.00 m) under the wall, less deep (0.50 m) under the corridor, and thinnest towards the centre of the room. Beneath the foundation is sterile orange-brown sand. None of the 23 pottery sherds found within the probe were diagnostic.

No hydraulic plaster was found in Room R, so presumably the room did not contain a washing basin. The function of this room will become clearer after the remaining rooms in the bathing circuit (N, Q and L/M) are excavated. In a later stage in the building's existence, the corridor between Rooms R and Q was blocked. Two glass beads reminiscent of those found in Ramm's temple ²⁹ were found within the room.

Room N

Room N was not excavated; instead it was cleaned by removing small stones, hoeing the soil flat, and lining up the large blocks along its walls. The room is 1.96 m N-S x 3.50 m E-W. There are doors in its western, eastern, and southern walls. The construction of the north-western doorjamb shows that the bath and Villa were built together. ³⁰ In a later stage the western door was blocked. This blockage is now only 0.38 m high, presumably because it was cleaned out in, or since, 1964 to provide an entrance to the bath. ³¹

The northern wall of the room suffered damage in the November 1995 earthquake. Much of the interior face of this wall fell down, but the exterior face remained intact. When the tumbled stones were removed, a flat sandstone slab (MPL = 0.112 m, MPW = 0.132 m, 0.044-0.094 m thick) bearing a Thamudic inscription was found beneath. The text of the inscription is still being deciphered. It could not be determined whether this stone had been below the wall before the earthquake, or was part of the tumbled wall core

Room Q

Room Q was cleaned in the same fashion as Room N. It is 3.20 m E-W x 4.25 m N-S. The rectangular walls (i.e., not the walls shared with R) vary between ca. 0.80 and 0.88 m thick. All the wall faces are constructed of large blocks laid in level courses without smaller levelling stones. There are doorways in the northern, western, and eastern walls. The doorways in the northern and western walls were intentionally blocked, but the eastern doorway seems to have been inadvertently blocked by tumbled stones and soil. Patches of white plaster (0.02 m thick) adhere to the eastern wall. A tumbled block reveals the wall plaster was laid in three layers.

Given that the doorsill visible under the blocked doorway in Room W probably continues into Room Q, the floor level in this room is probably equal to the floor level in W. The soil currently on top of Q's floor may be hiding fixtures. Given its placement between the calidarium and (presumed) frigidarium. ³² Q is likely to be the bath's sweat room or a tepidarium. Next season's excavation will look for benches under the soil and a hypocaust beneath the floor.

Room M/L (the frigidarium?)

Prior to the start of the season, Areas L and M were thought to be separate rooms. Preliminary excavation has suggested, however, that this area comprises one room which has a large rectangular basin (ca. 1.90 m N-S x at least 2.55 m E-W) in its centre. Only the northern edge and part of the eastern and western edges of the basin were uncovered. The basin is edged with large flat stones ca. 0.27 m wide. The tops and sides of the basin were apparently covered with orange hydraulic plaster with grey beneath (very little of which now remains). On the northern outside edge of the basin are the tops of 3 large flat stones (ca. 0.36 wide) parallel to the basin wall, perhaps the tops of stairs or benches. The soil in Room M was only cleared to the tops of these stones, so the function of these stones and the appearance and height of the northern side of this basin awaits further excavation. Moreover, except for part of the outside rim of the basin, most of the basin (and all of area L) is still covered in tumbled stones and soil. Clearing this area (hypothesised to be the bath's frigidarium) and delineating the basin will be a priority in the next season of work. A fuller description of the area will be published after this excavation is carried out.

In a later period (the fourth and fifth centuries A.D., to judge from the pottery), Room M may have been used as a cooking room. A high concentration of charcoal, ash and bone was retrieved, along with storage jar sherds at the level of the top of the bench/stairs. A broken basalt grinding stone was found in the southeastern corner of the room, next to the edge of the basin. ³³

Typology and Decoration

The order in which the core bathing rooms were used cannot be determined with certainty until all these rooms have been excavated. At the moment, it is tentatively hypothesised that bathers passed from N (a small apodyterium?) to R to Q (a sweat room?) to W (the calidarium) back to Q to M/L (a frigidarium?) and back to N. The bath thus conforms to Krencker's Ring Type. ³⁴

The lack of decoration within the bath-house is extremely puzzling. The baths in the Herodian palaces were lavishly decorated with wall frescoes and floor mosaics. ³⁵ The Nabataean bath-houses at Petra and Wadi Musa were also lavishly decorated. ³⁶ There is no indication, however, that the bath-house at Ramm was decorated with wall frescoes. None of the in situ or disturbed wall plaster in Rooms W, R, Q or M/L showed signs of paint. The floor in Room W was simply plaster, not a mosaic. There are two main possibilities for this lack of decoration. The first is that there were decorations but they were removed in Antiquity or in the 1964 excavation. All of the rooms excavated in 1996 had been previously excavated in 1964. It is possible, for example, that paint was applied to an upper layer of plaster that had fallen off and was removed in 1964, leaving only the lower (undecorated) coats of plaster on the walls today. We found fragments of painted plaster in the discard piles from the previous excavation but have no way of knowing what area they came from. Another possibility for the lack of decorations is that the rooms were simply not decorated in Antiquity. It is possible that the inhabitants of Wadi Ramm did not want or could not afford such embellishments. Next season's work, which will include the excavation of a room apparently not cleared out in 1964 (Area L), will help to decide between these possibilities. At the moment, the lack of decoration in this structure remains a mystery, especially given the discovery of painted plaster in the Villa (see above).

Ancillary Areas

Room K does not have a door opening into the main bathing rooms. In addition, a channel, presumably for water, curved around its north-western corner. For these reasons, Room K was originally thought to be a service room. In the course of defining the walls for drawing, however, it was found that the room had a carefully laid flagstone floor (pl. 13). The quality of this floor belies the service-room hypothesis.

Pl. 13. Room K, floor and channel in northwestern corner (Photo: Reeves)

Room K was a large room (ca. 6.04 m E-W x 5.74 m N-S). The absolute size of the room and the thickness of the walls cannot yet be determined, since one or both faces of all the walls are still buried in the sand. ³⁷ The walls are best preserved in the northwestern corner where a channel curves around the corner of the room 1.00 m above the flagstone floor. The channel is greatly worn but the best preserved section is 0.10 m wide. The interior track of this channel has been dressed in the Nabataean style, presumably to hold plaster. 0.60 m of the channel are in situ on the western wall and 0.58 m on the northern wall. Two other disturbed channel blocks had fallen to the base of the northern wall. Water may have travelled through the channel in this room en route to the basins in Rooms L/M and W.

It is not clear whether Area D (not excavated) is an internal room or an external space. Except for the outside wall of Room W, no wall lines are clearly visible. Many architectural blocks can be found in the area, including nine sandstone column drums (one of which is a base). Most of the drums are 0.67 m in diameter but one is 0.62 m in diameter. The surfaces of the drums have been dressed, presumably so they could be decorated with a stucco imitation of carved fluting, as was done in Ramm's Nabataean temple. ³⁸

Water System

The water for the structures on the temple hillock probably came not, as earlier scholars had assumed, from the spring of Ain Shellaleh, ³⁹ which is associated with a well known Nabataean aqueduct and shrine. Instead, water probably came from the closer spring of Ain Abu Rmeileh, located 340 m to the west of the eastern complex and ca. 70 m above it in elevation. ⁴⁰ In a cursory search we found 12 sandstone canalisation blocks lying on the steep slope between Ain Abu Rmeileh and the temple. All of the blocks were in a disturbed location, presumably as a result of earthquakes and the erosion of the precipitous upper face of Jebel Ramm. Each block was ca. 0.65 m long, 0.25 m wide, and 0.22 m high with a channel 0.08 m wide and 0.04 m deep. Two channels were also found carved into massive (presumably undisturbed) boulders. These are probably the remains of an aqueduct which carried water from Ain Abu Rmeileh to the temple hillock. ⁴¹ Trying to reconstruct the course of this aqueduct will be the focus of study in a future season.

Concluding Remarks

The first season of the Wadi Ramm Recovery Project was very successful. Thanks to the 1996 explorations, information on this prominent yet neglected Nabataean complex has finally been made available to scholars. The project achieved not only its primary goal of recording the extant ruins but also made some important discoveries about this complex. A new season is being planned for 1997 which will add to our knowledge of this structure by exposing the remaining core bathing rooms (the suspected frigidarium, sweat room, and apodyterium). In the Villa, one priority of the coming season will be to define the architecture of the peripheral areas. Extending the architectural plans will assist our understanding of the nature and purpose of the complex.

DEPARTMENT OF GREEK AND ROMAN STUDIES
UNIVERSITY OF VICTORIA
VICTORIA, BC V8W 3P4

¹ The Wadi Ramm Recovery Project is accredited by the Committee on Archaeological Policy of the American Schools of Oriental Research and licensed by the Department of Antiquities of the Kingdom of Jordan. The co-directors are grateful to Dr. Ghazi Bisheh, Director General of the Department of Antiquities, for his generous support, and to Dr. Pierre Bikai, Director, and all the staff at ACOR for their assistance to our project. Dr. John Oleson of the University of Victoria and the Humeima Excavation Project deserves special thanks for providing unending support and assistance. The following abbreviations are used in the text:

NEAEHL: E. Stern, A. Lewinson-Gilboa and J. Aviram, eds., The New Encyclopedia of Archaeological Excavations in the Holy Land, 4 vols. (New York 1993).

SHAJ: Studies in the History and Archaeology of Jordan.

² D. Kirkbride, "Region du Wadi Ramm: Communication," RB 67 (1960) 230-231.

³ M.R. Savignac, G. Horsfield, "Le temple de Ramm," RB 44 (1935) 245-278; D.V.W. Kirkbride, "Le temple nabatéen de Ramm," RB (1960) 65-92.

⁴ L.T. Geraty, L.A. Willis, "Archaeological research In Jordan," in L.T. Geraty, L.G. Herr, eds., The Archaeology of Jordan and other Studies, Presented to Siegried H. Horn (Berrien Springs 1986) 54.

⁵ M.R. Savignac, "Notes de voyage: Le sanctuaire d'Allat à Iram," RB 41 (1932) fig. 1; "Le sanctuaire d'Allat à Iram (suite)," RB 43 (1934) pl. XXXV.

⁶ The stone-by-stone plans are currently being inked.

⁷ This section was written by Dennine Dudley.

⁸ Savignac and Horsfield (above, n. 3) give some idea of the effect produced by the shape of the moulding (251 fig. 3), although the backs of all our fragments show that the plaster had been applied to a planar surface; some of the painted plaster at the temple is described at 257-258.

⁹ J.P. Oleson et al., "Preliminary report of the Humeima Excavation Project, 1995" (unpublished) 11.

¹⁰ Two of the ritual gestures associated with these Venus figures are arranging the hair and looking in a mirror; see J. Charbonneaux, Greek Bronzes, trans. K. Watson (London 1962) 123.

¹¹ M.-H. Gates, "Archaeology in Turkey: Sagalassos," AJA 99 (1995) 231-232.

¹² A. Negev, Nabatean Archaeology Today (New York 1986) 112.

¹³ See, e.g., B. Kolb, R.A. Stucky, "Preliminary report of the Swiss-Liechtenstein excavations at Ez-Zantur in Petra 1992, The Fourth Campaign," ADAJ 27 (1993) fig. 14.

¹⁴ J.J. Rossiter, "Villa," in E. Meyers, ed., The Oxford Encyclopedia of Archaeology in the Near East, vol. 5 (Oxford 1997) 300-301.

¹⁵ Rossiter (above, n. 14) 301.

¹⁶ This section was written by M. Barbara Reeves.

¹⁷ E.g. Machaeus: Stanislao Loffreda, "Preliminary report on the second season of excavations at Qal'at El-Mishnaqa Machaeus," ADAJ 25 (1981) 85-94; the "Great Public Bath-house" at Masada: Y. Yadin, "The excavation of Masada - 1963/64 preliminary report," IEJ 15 (1965) 29-36. In addition to the calidarium, the apodyterium, frigidarium and tepidarium seem to be in the same cardinal positions at Masada as at Ramm.

¹⁸ This is close to the dimensions of the vents (0.15 m2) in the Roman bath at Humeima, 40 km to thenorth; see J.P. Oleson, "The Humeima Hydraulic Survey, 1989: Preliminary field report," EMC/CV 34 (1990) 158.

¹⁹ Negev (above, n. 12) 51: "The interior of the room was then covered with several layers of plaster, beginning with mud plaster to fill the joints and cover the coarse face of the stone, and ending with layers of finer plaster. The walls were then whitewashed with fine lime plaster to form a base for frescoes." Traces of mud plaster were also found on the walls in Room N.

²⁰ The cuts resemble those from the Herodian bath-house at Masada (see Yadin [above, n. 17] pl. 8b), but not those from the Roman villa at Jalame (see G.D. Weinberg, ed., Excavations at Jalame: Site of a Glass Factory in Late Roman Palestine [Columbia 1988] 248-249, with pls. 8-10).

²¹ D.B. Small, "Late Hellenistic baths in Palestine," BASOR 266 (1987) 59-74; D. Barag, "Brick stamp-impressions of the Legio X Fretensis," BJb 167 (1967) 250; cf. Negev (above, n. 12) 49.

²² E.g. Masada: Yadin (above, n. 17) 29-36; Jericho (Baths in Herod's Second Palace): Ehud Netzer, "Jericho: Hellenistic to Early Arab periods," in NEAEHL II.681-91; Ramat Hanadiv: Yizhar Hirschfeld, "The early Roman bath and fortress at Ramat Hanadiv near Caesarea," The Roman and Byzantine Near East. JRA Supplementary Series no. 14 (Ann Arbor 1995) 29-55.

²³ Within Room W, the pottery in the (highly disturbed) fill above the floor ranged in date from the first to the fourth century A.D. The only pottery sherd found within the hypocaust probe dated from the period of the late first century B.C. to the early first century A.D.

²⁴ Inge Nielsen, Thermae et Balnea: The Architecture and Cultural History of Roman Public Baths (Aarhus 1990) I.14.

²⁵ The city bath at Petra may be earlier, but the excavation in the 1960s (still unpublished) did not reveal evidence of a hypocaust (F. Zayadine, "Decorative stucco at Petra and other Hellenistic sites," SHAJ III [Amman 1987] 137-139). For discussions of this bath, see J. McKenzie, The Architecture of Petra (Oxford 1990), and R. Wenning, Die Nabatäer-Denkmäler und Geschichte (Göttingen 1987) 226-227. Dr. Zayadine has suggested that this is not actually a bath but part of a palatial residence.

²⁶ Masada: Yadin (above, n. 15); Herodium: Ehud Netzer, Greater Herodium. Qedem 13 (Jerusalem 1981) 47-49. A combination of stone and tile pilae was also used in a heated hall in the Promontory Palace at Caesarea, but it is not yet known if this combination dates from Herod's construction phase or is the result of later renovations (see K. Gleason, ed., "Excavations of the Promontory Palace at Caesarea Maritima 1992-96: A preliminary report," JRA [forthcoming]).

²⁷ The villa and bath were excavated in 1996 by Khairieh `Amr of the Jordanian Department of Antiquities. Unfortunately, this was a salvage excavation. Once excavation was completed, the landowner demolished these important Nabataean remains to build shops for Petra's tourists. Barbara Reeves was fortunate to have been given a tour of the site by Dr. `Amr just prior to its destruction.

²⁸ Circular rooms in baths are often heated, e.g. in two Herodian-period baths at Herodium; see G. Foerster, "The Mountain Palace-Fortress," in NEAEHL II.619-621; E. Netzer, "Lower Herodium," in NEAEHL II.623.

²⁹ Kirkbride (above, n. 3) pl. IXb

³⁰ See above, discussion of access to Room P.

³¹ This is the main entrance used by tourists today.

³² See discussion of Room M/L below.

³³ Grinding stones, cooking pots and ash were also found in a later (First Revolt) stratum in the frigidarium of Herod's bath-house at Machaerus; see Loffreda (above n. 17) 87.

³⁴ D. Krencker, E. Krüger, H. Lehmann, and H. Wachtler, Die Trierer Kaiserthermen (Augsburg 1929).

³⁵ E.g. Machaerus: Loffreda (above, n. 17); Masada: Yadin (above, n. 17) 29-36; Jericho: Netzer (above, n. 22); Herodium: Foerster (above, n. 28) II.619-621; Netzer (above n. 28).

³⁶ Petra: A. Barbet, "Les caractéristiques de la peinture murale à Pétra," SHAJ V (Amman 1995) 383-389, Zayadine (above, n. 25) 131-142; Wadi Musa: Khairieh `Amr, personal communication, 1996.

³⁷ Room K was not excavated. The room was merely described and the soil was pulled back in the northwestern corner to photograph the floor and channel.

³⁸ See Savignac and Horsfield (above, n. 3) fig. 3.

³⁹ E.g. A. Stein; see S. Gregory and D. Kennedy, Sir Aurel Stein's Limes Report. BAR International Series 272 (Oxford 1985) I.303.

⁴⁰ On the location of both springs in relation to the temple hillock, see Savignac (above, n. 5) 43 pl. XXXV.

⁴¹ This aqueduct may be the one mentioned by A.S. Kirkbride and L. Harding ("Hasma," PEQ 79 [1947] 13) who described Ain Abu Rmeileh in this way: "Consists of three springs. On the rock terraces above the main spring are the remains of walls and buildings and quantities of sherds; a rock-cut channel leads from this spring to a ruined masonry cistern." Since we only came across the description by Kirkbride and Harding after our return to Canada, we were unable to confirm their details. Near the spring, however, we had noticed a (disturbed) door jam block.