This document is part of the online version of the book Amheida II: A Late Romano-Egyptian House in the Dakhla Oasis / Amheida House B2 by Anna Lucille Boozer, which is available at http://dlib.nyu.edu/awdl/isaw/amheida-ii-house-b2/. It is published as part of the NYU Library's Ancient World Digital Library and in partnership with the Institute for the Study of the Ancient World (ISAW). Further information about ISAW's publication program is available on the ISAW website. Please note that while the base URI of this publication is stable, the exact content available at that address is likely to change over time.
An Architectural Analysis and Reconstruction of House B2
Anna Lucille Boozer
This chapter engages the archaeological evidence with regional and temporal domestic architectural comparanda. Although it is not possible to produce a definitive reconstruction for all of the B2 attributes, I consider likely scenarios and offer a potential reconstruction. I will review: (1) the size, (2) orientation, (3) overall layout, (4) specific attributes and features of this structure, (5) placement of this structure within its local, regional and Mediterranean perspectives, and (6) the reconstruction of House B2.↩#p546
Size
House B2 is a building of moderate size, measuring 11 x 11 m (121 m²). It is at the lower end of large houses in our current spectrum of Romano-Egyptian housing types (Figure 1.5). A much grander end of the local domestic spectrum can be found in Amheida house B1, which has a 295 m² footprint. The local Dakhlan range in house sizes extend well beyond these two structures, although the largest Kellis houses are at least a century earlier than B2. The later Kellis houses are more comparable to B2.↩#p547
Excavated domestic units from Kellis Area A are between 99.2 and 129.65 m², without their exterior courtyards. In Area B, the elite domestic complex B/3/1 is 626 m² and the others are 306.25 m² and 415.6 m²; in Area C the domestic units are 101.4, 110.25, 129.9 and 175.25 m².1 These comparative dimensions indicate that Amheida House B2 falls well within the norms of excavated structures in the oasis.↩#p548
Looking to the broader Egyptian context, we find that B2 fits within the lower end of large houses in the housing spectrum. Four houses recently exposed at Douch, in Kharga, are approximately 300 to 350 m².2 Houses at Marina el-Alamein, on the coast of Egypt, range from 192 to 456 m².3 Archaeological research from the Fayum suggests that the typical house from this region contained about 50 square meters of living space for the inhabitants, although there are certainly houses of comparable sizes to structure B2 there as well.4↩#p549
We lack an informative range of house comparisons from all periods of Pharaonic Egypt. Most of our data derives from Middle and particularly New Kingdom structures. Shaw’s extensive study of Amarna workmen’s houses is perhaps the best comparison, although most scholars consider Amarna to be an atypical site. Shaw suggests that 65 per cent of the population lived in houses smaller than 100 m², whereas about 21 per cent lived in houses between 100 and 200 m², and as little as 8 per cent lived in houses larger than 200 m².5 Houses from Memphis and Thebes show that most individuals lived in very small houses, certainly under 100 m².6 The average house size in Deir el-Medina was 72 m² and the houses exhibited a range between 40 and 120 m².7↩#p550
It should be noted that it is likely that archaeologists have left the extreme lower end of the housing spectrum unexcavated in the above examples. Many people probably lived in very small one or two-room structures that are currently difficult for us to identify from the surface, or even upon excavation, as houses. Given this range of data it is fair to say that B2 is a modest but not a poor dwelling within the Dakhlan spectrum and a relatively large house within the Egyptian, Romano-Egyptian, and Roman Mediterranean spectrum.↩#p551
Orientation
The ancients were aware of the advantages and disadvantages of various architectural orientations and often built their houses and planned their cities accordingly. In Roman Egypt papyrological research indicates that there was a strong preference for entrances to be oriented towards the north. The west is the second most preferred direction. The tendency to orient houses towards the north agrees with a long-standing principle evident in the domestic architecture of the Pharaonic period. By designing north-facing houses, it was possible to provide houses with thermal comfort in Egypt’s long and hot summer season. A northern orientation reduced exposure to the light and heat of the sun and admitted cooling air movement from the prevailing north–northwesterly winds that are found across Egypt and in Dakhla in particular.8↩#p552
Egyptologists, particularly of New Kingdom Egypt, regard the northern orientation of houses to have been a general rule.9 Houses from Dura Europos on the Euphrates, a warm climate not too dissimilar to Egypt, are also oriented towards the north.10 Moreover, most of these houses from Dura Europos have courtyards located immediately to the north of a principal room (known as the andronitis) and its flanking rooms.11 This room organization allows the principal living quarters to be exposed to the prevailing northwesterly winds. A similar arrangement can be found in Egypt, and most houses that opened to the east or to the west will have had a courtyard located to the north of the principal living areas rather than to their south. This arrangement would open up the living quarters to more cooling air.12 The northern orientation preferred for houses also can be found among workshops, as can be seen from papyrological evidence at Oxyrhynchus.13 The northern orientation of the Egyptian houses and those from Dura Europos contrasts with contemporary houses found from around the northern Mediterranean, where winter posed a greater hindrance to comfort than summer and houses were oriented to catch the warming sun rather than the cooling wind.14↩#p553
House B2 is oriented with its entrance to the north, although it (with Area 1 generally) is slightly off-axis, potentially to reduce the amount of sand that might have come with the north winds. The central room of B2 (room 7) is oriented to catch the winds. It is also worthwhile noting that the ceramics production area north of House B2 is oriented such that it captures the strong north winds, which would have aided production.↩#p554
Figure 6.1: Cluster plan and linear plan diagrams (A. Cervi).
Layout
Footprint
Davoli distinguishes three types of Fayum house footprints: (1) rectangular buildings, (2) square or quasi-square buildings, and (3) complicated buildings with irregular exterior walls.15 The first and second types of houses are the most common in the Fayum, and the rooms within these general footprints could be arranged in a variety of ways.16 Some sites within the Fayum do not conform to this typology, such as the houses from Hawara, but these three types provide a good general understanding of the domestic footprint possibilities currently known to us during the Roman Period.17↩#p555
B2 conforms to the second type of house footprint, since it is square in plan view, measuring 11 x 11 m. Amheida House B1 and several Kellis houses also conform to this footprint, indicating that B2 fits both local and more general housing footprint norms found in Roman Egypt.↩#p556
Access Analysis
Interpreting room arrangement has been an important part of archaeological analysis for a long time.18 More recently, scholars have explored this method for analyzing Roman domestic architecture.19 These analyses show that a clustered patterning of rooms is not common for houses in Pharaonic Egypt, where linear plans dominate.20 A linear plan is one in which each room was entered through the previous room in a direct line (Figure 6.1); it was thus necessary to travel through all intervening rooms to get to a room at the end of the house. A clustered plan had one or more rooms that provided access to several rooms, so each room was more directly accessible than rooms in a linear plan (Figure 6.1). During the Roman Period, Fayum houses often have elongated footprints and do not have a single room through which most other rooms were accessed. Karanis and Soknopaiou Nesos both contain houses that demonstrate the linear access common to the Fayum. Despite the predominance of linear forms, there are also several houses from Karanis that had clustered plans, or combinations between clustered and linear plans.21↩#p557
B2 is square in plan view, and the rooms are organized around a large central room (room 7). This central room provided access for all of the major rooms in this structure except for one (room 3). In this respect, B2 demonstrates a simple clustered plan of access rather than a linear plan of access (Figure 6.1). Karanis House C127 shows a similar type of simple cluster plan (Figure 6.1). Kellis Houses 1 and House 3 also have clustered access plans.22 At Amheida, House B1 shows a bifurcated clustered access plan, with two rooms providing access to most of the rooms in the structure (Figure 6.1). These comparisons suggest that B2 conforms to local housing types and that this form could be found in other areas of Egypt, although it may have been less common in the Fayum.↩#p558
Vertical Reconstruction
Ptolemaic and Roman houses in Egypt typically had between one and four stories, although one seven-story house has been attested in the papyri.23 It seems that the average house found in cities and villages had two stories.24 Roman Period houses from Karanis and Soknopaiou Nesos had at least two stories, and vaulted ceilings were present only in the underground rooms.25 Forty-seven per cent of village houses and forty per cent of urban housing for which the number of stories is attested (n=109) had two stories.26 Although the sample size was quite small, these data suggest that most Egyptian houses of the Roman Period emphasized a vertical component rather than spreading over the horizontal axis.↩#p559
The recovered evidence from B2 suggests that this structure had only one story, although a staircase provided access to the roof. This reconstruction is consistent with evidence recovered from other domestic structures in the Dakhla Oasis, such as Kellis Houses 1–3 and Amheida House B1. These results suggest that Dakhlan houses are somewhat different from the Romano-Egyptian norm, as far as our present information goes.↩#p560
The staircase placement is somewhat variable in multistory houses found across Roman Egypt. In the houses of Davoli’s first type, which had a rectangular footprint, the staircase was located at the center of the house. In the houses of the second category, which had a square or quasi-square footprint, it was located in one of the corners.27 B2 is at variance with Davoli’s typology in that the staircase (room 8) is located in the center of the structure rather than one of the corners. Even so, the staircase location off the central room (the aithrion, see below) conforms to the general association between stairs and aithria across Roman Egypt.28↩#p561
In summation, the single-story reconstruction of B2 suggests that this house is consistent with known Dakhlan domestic architecture. The staircase placement in B2 does not conform to the typology Davoli established for the Fayum, but it does conform to the frequent juxtaposing of aithria and staircases. These results suggest that B2 shares more in common with local housing forms with respect to its vertical components than it does with houses located elsewhere in Roman Egypt. Even so, B2 falls within the generally understood Romano-Egyptian housing spectrum.↩#p562
Architectural Components
The specific architectural components of B2 can be compared with a range of excavated sites across Roman Egypt as well as more broadly within the empire. Moreover, papyrology provides us with descriptions of urban and rural houses from the first through seventh century, which can enhance our interpretations. Byzantine house descriptions tend to be more elaborate than those from the early and high Roman Period. Since most of these descriptions served to identify the house or the part of the house under discussion in legal cases, only major architectural features are mentioned.29 The most common domestic features mentioned in texts include the number of stories, the aule (yard), any purgoi (towers), wells, the aithrion (interior court), the pylon (gatehouse), the symposion (dining room) and sometimes outbuildings, such as dovecotes or stables.30 The first extensive attempt to integrate papyrological and archaeological data about housing in Roman Egypt was made by Luckhard.31 Unfortunately, van Minnen still had reason to complain about integration nearly a century later.32↩#p563
Caution is always recommended when mapping papyrological terms onto archaeological evidence, since the possibility of misattribution is high and can lead to much confusion.33 Even so, some named features are understood clearly from the texts and can be employed with some caution. The terms most relevant and least risky for exploring B2 are the aule (yard) and aithrion (interior court) so they will be employed below. The other terms attested either are not relevant to B2 or could cause confusion, such that it seems prudent to avoid them in the present examination. At a later date, if these terms and domestic architecture become better understood, it may be become possible to accurately map them on to B2.↩#p564
One of the few features that Romano-Egyptian houses appear to share is access to an open courtyard.34 Courtyards could be private or could be used by a group of houses and could vary in size. It was more common for traditional Egyptian houses to relegate courtyards off to the rear or side of houses (aule), but they may have been included as a central courtyard (aithrion).35 Despite the variation between size, location, and privacy, the presence of this space and the ways in which the courtyard was used is consistent. Courtyard activities typically included cooking and food preparation, agricultural work, and animal husbandry. Kitchen chores comprised milling grain and cooking food, both on both open fires and in closed ovens. Feeding troughs and shelters for animals were often in the same courtyard, but they also could be located in a separate one.36↩#p565
These activities left behind clear indications in the archaeological record. Archaeological examples of courtyards from Karanis houses preserve consistent traces of domestic animals as well as clay ovens for food preparation.37 Some of the features commonly found in Karanis courtyards included jars and bins for storing grain, ovens or stoves, bread ovens, or animal troughs.38 Architecturally, these spaces could have either informal beaten earth floors or more formal pavement preparations.↩#p566
Aule: Exterior Courtyard
The aule was a court or yard exterior to the covered house, and aulai are frequently attested in papyrological material in both urban and rural contexts.39 Many houses had this kind of open space, and the frequency of their attestation makes it likely that the term aule can be linked with the exterior work yards found at sites such as Karanis.40 From documentary sources it is clear that Romano-Egyptians considered the aule to be a separate entity from the house itself, as can be found in descriptions of houses by individuals from Karanis. For example, a woman, Tasoucharion, lists her property for the census and describes a “house with a courtyard”.41 Likewise, the courtyard of Hawara House E is described in the following terms; “the courtyard forms the western side of the house.”42 Documentary records also make it clear that aulai could be sold separately from the house and that individuals could own more than one aule.43 The peripheral location of aulai found at sites such as Karanis suggest that these spaces could be exchanged between neighboring parties quite easily.44↩#p567
C2, the exterior courtyard adjacent to B2, seems to follow the general descriptions of an aule, while room 7, which may have been open, would be considered part of the house proper. If room 7 was partially or fully open it is likely it would have been considered an aithrion (see below). The exterior courtyard space south of B2 conforms to a standard that appears to have been common in Roman Egypt. It was an exterior, multi-use space for food preparation, craft activities, and animal husbandry. As seen from the previous chapter, this space was much less formally constructed than its counterparts for food preparation within the structure (rooms 5 and 7).↩#p568
Courtyard C2 may have been considered part of the property of the household that lived in B9, east of B2, but it may have changed hands during the occupational history of the house.45 One possibility might be that C2 was originally communal, but was sold to B9, which led to the construction of the bread oven in room 7 of B2 to replace the bread oven found in the exterior courtyard. It is impossible to verify such conjectures on the basis of the archaeology at this time.↩#p569
Aithrion: Central Courtyard
The term aithrion is confined to the papyri from Egypt, and consideration may be given to a specifically Egyptian explanation of what this architectural unit might have been. Just as the Pharaonic and Romano-Egyptian house tended to be oriented towards the north in order to take advantage of the cooling effects of the prevailing winds, also the aithrion might have had a connection with the north winds.46↩#p570
Documentary evidence of aithria is unevenly distributed between cities and villages. Most aithria houses come from urban contexts, while most of the village attestations of aithria houses come from Tebtynis in the Fayum.47 No documentary sources mention the sale of an aithrion separate from that of a house, which suggests that the aithrion was an integral feature of the house.48 Documentary sources make it clear that the aithrion could contain domestic equipment and that aithria were used in a similar manner as the aule.49↩#p571
The aithrion (court) formed an integral, central part of the Romano-Egyptian house.50 Houses could have both an aule (exterior courtyard) and an aithrion, but sometimes aithria house descriptions do not mention an aule. This absence may indicate that the presence of an aithrion meant that an aule was unnecessary. The aithrion usually served rooms that opened on to it. The aithrion was thus usually centrally located, and more than a single room opened onto it. Houses usually had one aithrion, but sometimes they had two (or more?). The space occupied by the aithrion extended from the ground floor up to the roof. The aithrion was often located close to a stairway or stairwell.51↩#p572
It is debatable if aithria were open or closed spaces. Daniel argues that the aithrion was not an inner courtyard or Lichthof that was open to the sky above, but was a covered hall that received only a small amount of daylight and was ventilated in such a way that it also cooled adjoining rooms.52 This hall might have been covered by a flat roof or by an air-trapping shed that took advantage of prevailing north winds.53 The archaeological and papyrological evidence for these suggestions is ambivalent.↩#p573
Local examples of aithria houses can be found at Kellis. Kellis House 3 has an open room (room 6) that functioned in a similar way to the aithrion in B2 and was labeled as a “living room” by the excavators because it provides obvious gathering place and access point for inhabitants.54↩#p574
The Fayum also contains archaeological examples of aithria houses. Karanis House C168 has a well preserved central courtyard and dates to sometime in the middle of the first century CE, through perhaps as late as the end of the second century CE.55 This area was used in a similar way as the exterior yards that were more commonly found at Karanis:↩#p575
Every house in Karanis had its courtyard or shared one with its neighbor. It was usually situated at one end or at the side of the house, where there was direct access to a street or passageway. Occasionally, as in C168, it was in the center of the house with the remaining rooms grouped around it, so that it had no communication with the outside. The courtyard was enclosed by walls but it was unroofed, although sometimes roofs were built over animal pens or bins.56
Grenfell and Hunt (1898–99) as well as Rubensohn (1902) excavated the houses of Theadelphia (Kharabet Ihrit) in the search for papyri. As a result, the archaeological data is poorly presented and little remains visible on the site surface today.57 Rubensohn only described and published adequately two houses from Theadelphia (Batn-Ihrit), and both of these houses contained aithria.58 The first house was preserved for four meters in height, and had an L-shaped plan divided into four rooms. The aithrion was located in the middle of the house and provided access to a stairway leading upstairs.59↩#p576
Rubensohn’s second house was much larger and more complicated than the first house and appears to have been occupied from the second to the fourth century CE. An aithrion was also found in the middle of the building. It is notable that this house appears to have been modified so that this aithrion became more of a centralized access point than in its early occupation.60 The published evidence of houses from Theadelphia denotes a degree of wealth and indicates an architectural and decorative repertoire that was not entirely Egyptian in origin.61 As such, the houses from Theadelphia provide informative comparisons to B2.↩#p577
Additional houses can be found in the Fayum, although limited comparisons can be made since they have not been well published. For example, although the majority of houses in the Fayum only have an aule, Soknopaiou Nesos (Dîme) also has evidence of aithria houses.62 Philadelphia (Kom el-Kharaba el-Kebir) houses also provide evidence of internal courtyards.63↩#p578
Central courtyards can be found elsewhere in Egypt, although they are not always used in the same way as the one from B2. Two out of the four Romano-Egyptian buildings recently uncovered at Kysis (Douch) provide evidence of potentially unroofed or partially-roofed courts.64 House III has a large interior court with columns as well a raised inner portion that does not find ready parallels in Egypt. It may be more closely linked to Late Antique luxury houses in Syria.65 House I also has an inner peristyle that was probably lightly roofed between the columns and the walls, but may have been unroofed over the center or had a light framework placed there.66 The excavator suggests that it is possible that these structures were public buildings or served administrative purposes.↩#p579
The coastal, Graeco-Roman town at ancient Leukaspsis or Antiphrae (Marina el-Alamein) had houses with central courtyards. This feature, among others, is similar to the more lavish dwellings found in Dakhla. The more sumptuous dwellings contained a central court surrounded by two or three portico wings, while the smaller dwellings had just one wing of a single column. Rooms were built around these central courts and, typologically, the excavators identify these houses as portico and peristyle houses.67 The excavators attributed this style to a Greek and Roman influence upon domestic architecture, which appeared to be the dominant spatial scheme in the town as it was replicated consistently. These peristyle houses had highly complex plans. Orthogonal and axial plans predominated, indicating a Roman tradition rooted in Hellenistic architecture of the eastern Mediterranean.68↩#p580
Unfortunately, the excavators have not published comprehensive reports of their excavations yet, so we cannot at this time venture thorough comparisons between Amheidan domestic forms and those at Marina el-Alamein. Likewise, they have not yet published data on smaller structures similar in scale to B2. Even so, the houses at el-Alamein offer a more opulent and classically pronounced variety of architecture than we find at Amheida and particularly at B2. Moreover, Marina el-Alamein occupies a different ecological zone than Dakhla, which would lend itself to different architectural types.↩#p581
The central courtyard is a common feature in contemporaneous houses found throughout the Roman Mediterranean. In North Africa, these houses are not Roman style atria houses but rather peristyle houses that devote considerable space to open areas around which other rooms are organized.69 Like the Marina el-Alamein houses, these structures show much more influence from Classical housing norms than does B2.70 Courtyard houses are quite common in the Roman East as well. For example, the well-known Christian House from block M8 at Dura Europos looks quite similar to B2 in plan view and contains a central open space.71↩#p582
A central open room was a feature signifying a Roman Mediterranean (or Classical) house plan rather than the traditional Egyptian plan, which typically had only external open areas off to one side or behind the house proper.72 Even so, it is clear that there is a range in how this influence was expressed in the architecture. B2 shows less classical influence than houses at Marina el-Alamein and some areas of North Africa. It fits more in line with houses found at Kellis, Theadelphia, and some from Karanis. There may be some significance attached to a general clustering of courtyard houses to the second century CE and later, but we do not have enough of these structures excavated at this time to make general conclusions about phased attributes.↩#p583
The presence of an aithrion would have a significant impact on the appearance of a domestic structure. This broad open space would provide light and fresh air to the rooms grouped around it, while houses lacking this space would be substantially darker and more closed. A negative attribute of such openness in Dakhla is that the oasis is subject to strong blowing sand and sandstorms as well as extreme highs and lows of temperature. Such environmental considerations would have impacted the appeal of this type of architecture in Egypt’s various regions.73↩#p584
The ways in which individuals used the aithrion would have changed the ebb and flow of spatial usage in the house, since practical activities would occur in the center of the home rather than at the peripheries. Although the presence or absence of an aithrion would impact the appearance and spatial usage of houses profoundly, the distinction between these types of houses should not be exaggerated, since a small aithrion house may have little if any more space than a house without an aithrion and with an aule.74 The presence of an aithrion should not necessarily be understood as an economic advantage, but rather a social difference.↩#p585
Based upon the lack of archaeological data indicating the presence of a roof, room 7 may have been open. Likewise, bread ovens typically were placed in unroofed rooms with their draft hole opposite to the prevailing wind.75 This situation occurs in room 7 with the presence of bread oven (F19), further justifying the suggestion that this space was open. Moreover, as it has been proved that Courtyard C2 belonged to B9 and not to B2, the aithrion in B2 was essential for daily domestic needs. Of course, absence of evidence is not evidence of absence, so it is not possible to prove beyond a doubt that room 7 was unroofed. An additional possibility is that room 7 once had light roofing that was not load bearing but would have shielded this space from wind and sun. Such conjectures are impossible to verify without additional data from unexcavated houses. At this time, House B2 fits within the limited range of aithria houses known from Roman Egypt, and it is unclear if, and how frequently, this range of houses could have light screening frames.↩#p586
Entrance System
House B2 has an L-shaped entrance, opening to the north (room 9). The shape helped to keep out sand and trash deriving from the busy street outside the house. An L-shape entrance system is not uncommon in Dakhla, and it is a housing feature with a long history in Egypt. Examples can be found from several houses at Kellis.76 The lack of a view from the entrance into the rest of the house makes it clear that such vistas were not considered to be an important aspect of the architecture for this structure; they may indeed have been negatives due to the strong wind. Because of the immediate sharp turn to the east, individuals who entered the structure would not have had a view inside the house until they had already turned from the entrance into room 7 to the east. Even once they entered that room, no orchestrated architectural vistas, as could often be found in Roman houses, were available to visitors.77 The turns in this entrance would have concealed the interior of the house from the outside, enabling a greater sense of privacy, visually and audibly removing the house from the chaos, dirt, and noise that were likely present on this main street into Amheida.↩#p587
The north doorway to the street in room 9 was quite worn and shows signs of repair and modification. For example, it contains limestone steps leading up to the street, mirroring the rising street level found in excavations from Street S1.78 These stairs have deep gouges in them that may be signs that individuals were sharpening knives or shaping other tools upon these stones. Stone was readily available for construction, but it was more expensive than mud brick: it is therefore common for houses from Roman Egypt to restrict stone to the use of thresholds and lintels of entrance doorways, as we find here.79↩#p588
When entering the house from the street, the wall facing the street retained no visible signs of decoration, although it is likely that it was once covered in mud plaster. There were small adobe barriers constructed on either side of the doorway in the street, which may have protected the house from sand, refuse, and traffic.80↩#p589
The importance of the dodged entrance likely is due to the environmental conditions surrounding B2. Because B2 is located on the edge of the desert and is facing into the sand-laden wind from the north, it required protection from these elements. The turns in the entrance to B2 would have concealed the interior of the house from the outside, enabling a greater sense of privacy as well.↩#p590
Fireplaces and Hearths
Room 5 contained a simple fireplace (DSU 40) superimposed above a hearth (F46). The fireplace consisted of a pair of charred, small mud bricks and a mud plaster coating. The base was a prepared mud plaster surface. Husselman provides us with a succinct description of a simple fireplace similar to that found in room 5:↩#p591
Cooking was done on a simple fireplace constructed usually of two adobe bricks laid flat on the floor about 15–20 cm apart, set up against the courtyard wall. A small fire would be built between the bricks, on which a round bottomed cooking pot would rest.81
These food preparation features were standard in Roman Egypt. The use of multiple spaces for food preparation is not uncommon in Roman Egypt. Houses from Soknopaiou Nesos, for example, typically had one or two exterior courtyards with ovens, mortars, and millstones.82↩#p592
The hearth (F46) (diameter c. 72 cm) was a rounded adobe construction also charred from use. Basic cooking chores could take place on this hearth. These cooking features were typical within Roman Egypt, and parallels can be found both locally and in the Fayum. Local parallels to the hearth (F46) found in room 5 can be found at neighboring sites, such as Ain el-Gedida and Kellis.83↩#p593
Bread Oven
The oven in room 7 (F19) represents the most common type of bread oven found in Roman Egypt and is attested locally at Kellis (Figure 4.2).84 Husselman provides a standard description of this type oven as found at Karanis:↩#p594
In addition to the fireplace there was usually an oven in the courtyard for the baking of bread. It was made of baked clay, circular in shape, and was .5 m to .75 m in height. The diameter of the base was approximately 1m and the walls curved at the top to an opening about 40–50 cm in diameter. It was generally placed in the corner of the courtyard and mud brick walls were built to enclose it, with the space between the walls and the circular oven filled with broken bricks and plaster. Each oven had a draft hole in its base. In use a fire would be made in the oven to provide a bed of coals on which flat loaves of unleavened bread would be baked.85
This type of oven is what Yeivin calls a “later type” oven.86 Depraetere’s more recent comparative study contains a revised typology, which would make this oven a Type II oven.87↩#p595
Type II ovens were typically situated in a corner or built up against a wall in the courtyard of domestic structures with the draft hole opposite to the direction of the prevailing wind.88 As mentioned above, the aule and aithrion were the primary areas used for cooking and other domestic tasks, so the bread oven would be in close proximity to other cooking facilities. In order to secure the oven in the courtyard, the space between the walls and the circular oven was filled with debris.↩#p596
The B2 oven (F19) consisted of a ceramic body enclosed within a square of mud-brick walls that are one brick thick. The oven itself has a 65 cm interior diameter and is preserved to a height of 60 cm. There is no clear bonding pattern to exterior walls. There is a 14 cm diameter hole at the bottom of the feature to the south. A draft-hole was located in the lower part of the oven, extending from the interior of the feature to the exterior, thereby cutting through both the ceramic and the mud brick wall.89 It is 8–17 cm above the interior base of the feature; 46–80 cm west of the east wall of the feature. The diameter of the draft hole is 10 cm. At 0–13 cm above the base of the feature there are lines running laterally along the ceramic. This oven description fits well within the norms of Type II ovens.↩#p597
The Courtyard C2 bread oven (F4) was poorly preserved. Only one course was preserved for most of it and only two courses on its east wall. Together with F3 to the west, F4 was built into the southwest corner of Courtyard C2. Including all of the walls, the feature itself consisted of a rectangle of mud-brick 138 x 102 cm. The diameter for the circular interior of the oven was approximately 85 cm, but it was very poorly preserved and present only in patches. The area between the circular interior and the rectangular mud brick exterior was filled with ash and mud brick. The draft hole of the oven was visible in the east wall of the feature. It was 7.5 cm from the bottom of F4, as measured from the exterior. The top part of the draft hole was missing, so it was not possible to say how high up it extended. The north–south orientation of the draft hole was 38–63 cm north of F5, the south boundary of the feature. The platform (F3) for F4 was preserved to the west. Together the two features measure 106 cm north–south and 225 cm east–west. The platform (F3) was poorly preserved, with only 1 course visible above clean sand below it. It consisted of a small dividing wall 99 cm east–west built to the north and another, 96 cm long north–south, built to the west. Each wall is only 1 course high and 1 brick thick. The base of F3 consisted of blackened mud bricks in no clear bonding pattern. The northwest 20 x 32 cm was not preserved. This oven appeared to be a more elaborate version of the Type II oven than the B2 bread oven. Unfortunately, it is not well preserved, so it has not been possible to say more about its construction details or how the inhabitants used it.↩#p598
The oven tops were covered to hold the heat longer. Yeivin describes two such oven covers that were recovered from Karanis. They consisted of an oblong dried-mud plate with a mud ridge along the middle that was notched to make it easier to grasp.90 A possible bread oven cover was found discarded in Street S1 just outside of B2, and a deteriorated one was found in room 5.↩#p599
Ovens were fueled with stalks and branches in order to build up a flame. Dung was added at a later stage in order to sustain the flames and conserve wood. Wood was sparse in Egypt, but dung was readily available and usually at close proximity, since animals were often kept within the same space as the cooking implements.91 Ovens fueled by dung never reached high temperatures, but they were sufficient for cooking bread and other basic requirements.92↩#p600
When the oven was used, bread baking probably took place on the interior surface of the ceramic, where it was possible to plaster flat, disc-shaped loaves to the ceramic walls with water or milk.93 Depraetere’s interpretation of oven usage contrasts with Husselman’s opinion that flat loaves of unleavened bread would bake on top of a bed of coals fired in the oven.94 The oven from B2 had two rows of circular burn spots, possibly caused by baking bread loaves, which adds credence to Depraetere’s theory (Figure 6.2). There is a row of burn spots 10–20 cm above the bottom of the feature with diameters of 10 cm. The second row of burn spots is visible 21–33 cm above the bottom of the feature with diameters of 12 cm. Although I have found no mention of such marks on other ovens from Roman Egypt, the lack of such comparanda is not surprising, since ovens have been studied in detail only recently, and Depraetere himself cites the lack of specific information provided about excavated ovens.↩#p601
Figure 6.2: Photo of burn spot detail on bread oven (F19), room 7.
The primary advantage of the Type II oven over the Type I oven is that it was less fragile and provided a more even baking of bread.95 According to Depraetere, the Type II oven was probably bought or ordered from a potter’s workshop; he did not believe that they could be constructed without specialized labor.96 However, ethnographic research suggests that it is possible that such ovens were domestic products. Ikram’s ethno-archaeological research in contemporary Middle Egypt demonstrates that the women in each family produced domestic pots in a large number of forms ranging from stoves (kanoon) and ovens (fourn), to large storage vessels (soma).97 Unfortunately for archaeological purposes, domestic potting leaves very little residue in the pot fabrication area. Once the mud dries, it turns to dust similar in appearance and compaction to the earth commonly found around mud brick houses. Likewise, the tempers used by domestic potters—straw, dung, ash, and reused brick—do not suggest any special activity.98 As Ikram suggests, it is difficult to verify archaeologically that domestic pots were produced within the home, but it is important not to lose sight of such possibilities.↩#p602
For example, this ethnographic research indicates that women may have had seasonal work demands for domestic potting that would have created constraints upon their time that were not present at other times of the year.99 In many ancient societies, such tasks became an enactment of specific identities that were defined, created, and emphasized through meaningful activities.100 Furthermore, the potential of constructing such ovens within the home demonstrates that such types of production can have a considerable effect upon the household itself and should be considered as part of our understanding of larger-scale social issues such as economic and social structures. In other words, ethnographic evidence for the household production of items such as this oven suggests that local households might have been self-sufficient for a modicum of daily practices and did not require as many large-scale dependency relationships as they might have if they had to rely upon workshops for required household items. In turn, the expertise of specific women potters may have been noticed among the community, and other families may have called upon their assistance, as Ikram describes in her ethnographic study. Such potential interconnections may have helped strengthen community relationships and the individual identities of women.101↩#p603
Room 11 Feature
Room 11 is a small sub-area located in the north of room 7 just west of the oven feature. Room 11 contains two different components; a storage area and an upper portion that is poorly preserved but probably also functional (Figure 6.3). The storage portion of room 11 is a barrel-vaulted, rectangular space, 1.25 m2. No significant artifacts were associated with this feature that could clarify its function, although the interior of the feature is quite low at only 28 cm high. This storage area was accessed through an aperture of 56 cm in the east of the feature.↩#p604
The function of the upper portion of the feature is unclear due to the poor preservation of the feature at this elevation. The preserved remains show that rubble was placed above the barrel-vault to create a flat surface and that a plastered channel rests on top of this rubble. This channel follows the south and east sides of this feature and probably would have continued along the north side. The rest of this feature was not preserved. It is possible that this feature would have been used for the preparation of food or liquid. For example, the plastered area could have provided a surface upon which the bread loaves from the oven could cool. It is also possible that the top of the feature may have been used as a variety of water pot stand, a feature not uncommon in domestic contexts, but these are usually made out of stone and have hollowed out spaces on top.102 A similar feature can be found from a temple precinct at Ain Manawir in the Kharga Oasis. The immediate context of this feature was external to the sacred areas and therefore was located in an area of the site that was probably devoted to storage and production. The excavator did not know what its function was.103↩#p605
Figure 6.3: Photo of room 11 and bread oven (F19), room 7.
Storage areas are commonly found in aithria.104 Storage spaces are generally near oven features and can be found at comparable sites, such as Kellis. Kellis House 1 has two ovens, which had replaced earlier hearths, in close proximity to a storage bin.105 Likewise, courtyards in Karanis usually were equipped with areas used for grain storage.106 Room 11 likely served this function, although it is a bit different from standard varieties found in Karanis. Moreover, the oven next to room 11 was added after the initial construction of B2, suggesting that room 11 initially had an independent usage separate from the oven. This feature will be re-examined as a broader range of comparanda becomes available, which appears likely since the unexcavated house west of B2 appears to contain a better preserved version of this type of feature.↩#p606
Roofing
This house appears to have had barrel vaults over six rooms.107 Barrel vaults are a type of roof that generates an outward thrust against the walls that support it. The supporting walls are often thick and strong in order to absorb this thrust. A more elegant method of compensating for the outward thrust is to build two or more vaults parallel to each other so that the outward thrusts from each vault negate each other. In this situation, only the outer wall must be quite thick or reinforced by buttressing.108↩#p607
Although barrel vaults are mostly associated with the Romans, they are known from Dynastic Egypt. Examples of barrel vaults can be found at Dendera (ca. 3500 BCE), the Ramesseum of Rameses II (ruled 1279–1213 BCE), and numerous other funerary contexts. In Dynastic domestic contexts, vaulted rooms were rare and flat roofs were more common, except for underground cellars and perhaps a few very small rooms.109↩#p608
In Roman Egypt, many sites conform to the Pharaonic roofing tradition. Karanis houses, unlike the ones from Amheida, seem to have been covered mostly with flat roofs constructed from palm logs and palm reed mats and mud, except in the case of underground rooms, which were generally covered by vaulted ceilings.110 The houses in Kellis that are contemporaneous with this structure appear to be largely flat-roofed structures, although it is only slightly later that we find that barrel vaults are the dominant roofing type there. The “kitchens” of houses from Kellis Area A often have flat roofs consisting of palm, despite the presence of barrel vaults elsewhere in the same structure.111 There may be some preference for flat roofs over areas with hearths in Dakhla during the third century because smoke could exit through the roof more easily than through mud-brick barrel vaults, although there is not enough of a sample at this time to confirm such conjectures. Certainly the food preparation areas found in Amheida House B1 were located in barrel vaulted rooms 4 and 8, although neither of these rooms had a bread oven (Figure 2.5).112↩#p609
Room 5 may have had a flat roof covering at least part of the room, as some organic debris was found to the north in that room and also as there must have been some point of connection between the stairs and the rooftops over the other rooms in this structure. At this point it is impossible to prove the presence or absence of a light jarid roof over any part of room 7. Certainly there was no evidence of any roofing materials from this room, but we will have to wait until we unearth additional structures in Area 1 before we can make statements with greater certainty.↩#p610
Rooms 4 and 11 had slightly different types of vaulting than the other vaulted rooms. These roofs have been discussed in the previous chapter.↩#p611
Stairways and Steps
B2 contains a stairway as well as a series of steps leading from the house to the street. Both the stairway and the steps conform to standard building types in Roman Egypt, and comparanda can be found locally in Dakhla. The placements of stairway F29 and steps F50 have been discussed already but the construction methods employed will be discussed below.↩#p612
The steps in room 9, F50, were added at a later date in the building’s history to compensate for the rising ground in the street. This situation is common in Dakhla because of the strong sand-laden wind that raises the ground level. Moreover, the deposits of debris and other goods in the street also raised the surface level. Steps connecting an entrance door to the street were often made of stone in Karanis, much as they were in B2 at Amheida.113 The stone was used preferentially in the entrance due to heavy wear during frequent usage. Kellis House 3 also had stairs added at a later date to its entrance (room 1) in order to compensate for the rising surface level of the street.114↩#p613
The stairway (F29, in room 8) conforms to a standard type of stairway found in Romano-Egyptian houses, and its placement corresponds to that in other aithria houses in Roman Egypt.115 B-level Karanis houses, which were probably contemporaneous to our B2 structure, usually did not have underground rooms but had stairways to upper floors and roofs that were commonly used for functional tasks.116 Husselman’s description of these stairways conforms to our own from B2:↩#p614
The method of constructing stairways remained the same throughout the existence of the town. The stairs consisted of several short flights built around a central core, each flight leading to a landing…The steps were built of adobe bricks as a general rule.117
The staircase and steps from B2 conform to standard types found throughout Roman Egypt and represent a neutral attribute in the architecture with respect to building traditions.↩#p615
Storage Spaces
There were a number of informal and formal spaces within this structure created for the provisioning of storage. The formal spaces included spaces beneath the stairs (rooms 4 and 10) and a vaulted storage area within room 7 (room 11), which is discussed above. The informal storage space was a storage pit with two storage vessels located under a trap door (room 6).↩#p616
Due to the poor vertical preservation of this structure, we cannot reconstruct the access to room 4 with certainty. There was an opening between room 4 and room 5 that could easily allow for access between those rooms if that opening represented a doorway. This possibility can be corroborated by comparanda within the oasis. At Kellis, Houses 1 and 2 from Area A both have walk-in cupboards in close association with the stairs.118 Alternatively, a large amount of comparanda from Karanis suggest that room 4 and room 10 were both accessed by means of trap doors within the stairway above them.↩#p617
Regardless of the means of access, the spaces beneath flights of stairs were often used as small storage rooms. Under-stairs storage spaces usually had mud-brick floors, a height of approximately 1 m, and an entrance through a trap door.119 In the German excavations at Elephantine there are many houses with small storerooms located underneath the staircase. These rooms had pavements fitted for preserving goods from insects and rot.120 The storage spaces from Karanis share similar characteristics. The under-stairs area can be interpreted as an Egyptian feature, since Greek documents from Syene use an Egyptian loan-word for this feature.121↩#p618
In addition to the formal storage spaces found in rooms 4, 10, and 11, we have simple storage spaces. In room 6 a wooden lid covered an unlined pit with two storage jars. Buried storage jars can be found throughout Roman Egypt, including at Berenike, Karanis, and Kellis.122 Simple methods of storage at Karanis included underground vaults reached by trap doors in floors, or bins constructed by walling off corners of rooms or courtyards. Large jars and even open and unlined pits also served for storing small amounts of grain for domestic use.123 Bins made by building a wall across the corner of a room are found in almost every house in Karanis, as well as in the courtyards.124 Trap doors at Karanis were usually set in corners of rooms so that people could use footholds cut in the walls of the lower rooms.125 The room 7 storage pit from B2 seems to be the simplest type of construction available, since it lacked evidence of pivots, ropes, or multiple doors and was completely unlined. The doors of such storage pits in Karanis consisted of planks of wood that were set into acacia frames.126 Although the contents of the storage pit were invisible to visitors, the lid itself became a marker. The number and placement of lids indicated the presence of stores to anyone who entered the house.↩#p619
Windows
The wall heights in structure B2 are not sufficiently preserved to determine if there were windows present and, if so, where they might be located. On the basis of comparanda, it seems unlikely that there were windows of any substantial size in this structure. Comparanda from Karanis suggest that windows only served to admit light and air, being set high in the walls immediately below the ceiling, rather than offering views.127 It is likely that B2 once contained similar windows, particularly in room 3, since it was not accessed directly through room 7 and there would have been no natural light source for that room.↩#p620
Wall Niches
Niches were common attributes of walls in Roman Egypt, serving as storage spaces for small objects or to place lighted lamps. Alternatively, niches could serve as household shrines, similar to a Roman lararium, which can be found throughout the empire.↩#p621
The wall heights of B2 are not sufficiently preserved to determine the presence of niches, but given the breadth of parallels, it is certain that they would have been present. Niches were usually located below first floor windows at Karanis, but there are few noted commonalities in their location beyond correspondence with windows.128↩#p622
It is possible to conjecture that room 7 contained one niche on wall F12 since it is the only wall of any substantial length in that room. Two objects (Inv. 11048 and Inv. 10347) were found on top of one another amid the wall collapse and may have fallen out of a niche together. The statue fragment (Inv. 11920) also probably once rested in a niche and may be most closely associated with wall F12. The other rooms may have had niches on their outer walls, although all of these suggestions are highly tentative.↩#p623
Discussion
An archaeologist must always bear in mind that only a small quantity of material remains for us to study out of the vast array that once existed. Developments that initially appear singular may be indicative of a type that has yet to be discovered or now may be lost to us. In an area of nascent study, such as Romano-Egyptian houses, we ought to remain mindful of the many comparative examples have been lost to us through neglect, destruction, and inaccessibility. The isolated ruins of houses in Roman Dakhla, therefore, may offer the possibility of shedding light on housing types that, rather than being a local specificity, may have been common in many now-inaccessible areas of Egypt or even further afield in the Roman Empire. Even in the case of Pompeii, which often is considered the site paradigm for Roman housing due to its high state of preservation, archaeologists are uncertain whether an innovation should be described as new to the Roman world or new to Vesuvian housing.129↩#p624
New data from Amheida and Kellis suggest that we begin to revise our expectations for housing styles in Roman Egypt. Likewise, the new discoveries at Marina el-Alamein, Douch, and Tebtynis reveal that classical style housing existed in Roman Egypt and that we should not expect all houses in Egypt to reflect only the vernacular Egyptian (Pharaonic) style found in many of the published houses of Karanis and Soknopaiou Nesos.130 The papyrological record from Egypt corroborates this conjecture, as it preserves numerous references to houses that appear to have possessed a classical layout and they also document a rise in housing prices during the second and third centuries CE, perhaps due to a change in domestic structure choices.131↩#p625
Egyptian Housing Spectrum: Diachronic Comparisons
There are some Pharaonic houses that bear a slight resemblance to the B2 house, although they are significantly earlier in date and distant in location. The closest parallel is the Khentkawes houses that were built for the mortuary priests of Queen Khentkawes, a late 4th Dynasty king’s mother who has a major monument (sometimes called the fourth pyramid) at Giza. Like Area 1, these structures were planned identical houses along a road. The houses all had the same plan originally, and extend between a road that runs next to the causeway leading from the cultivation to her tomb (on the south) and one that runs parallel to it and which marks the outer wall of the complex (on the north). They are about 11 x 24 m, so they do not demonstrate the square plan we find in B2.132↩#p626
There are also much larger Middle Kingdom houses at Kahun and Abydos that bear a slight resemblance to the B2 house. These are also planned houses of funerary functionaries. In the Kahun houses, attached to the pyramid of Senwosret II, there is a long and slightly convoluted path to the central main rooms. The path leads to a central open court (with a colonnade on the south to catch the northern breeze), south of which is a transverse hall and then three rooms, the “bedroom” to the west and a room of unknown function to the east, which leads to a passage into the “kitchen areas” (which have their own open court). There is also a smaller version of this sequence of rooms west of the main internal courtyard. The Abydos house is similar, and only a few generations later (Senwosret III).133 These comparisons are considerably more distant in time than the New Kingdom houses at Amarna and Deir el Medina that most archaeologists of Roman Egypt use as comparisons to Roman housing traditions. Despite such a time disparity, such comparisons between Romano-Egyptian houses at Amheida and older Pharaonic models should be kept in mind for future research on diachronic change within housing traditions.↩#p627
Romano-Egyptian Housing Spectrum
The layout of House B2 is typical of what appears to be a local Dakhlan domestic architectural tradition. This can be seen by comparison with recent excavations at Kellis by the DOP. Results from these excavations indicate that Roman Dakhla domestic architecture of the second, third, and fourth centuries CE typically consisted of a single-story structure with barrel vaulted roofs and a central room that was often partially or completely unroofed.134 In other words, these houses have a clustered plan of access, a central open room, and a preference for horizontality. We can discern some variations in this plan over time in the oasis, as can be seen by comparisons to houses from other periods at Kellis and another house from Amheida.↩#p628
Contemporaneous areas of Kellis, namely Areas A and C, show that the dominant domestic architectural form consists of a rectangular or square structure, with a central open room. Smaller rooms cluster around this central open area, and a staircase provides access to the roof.135 The earlier houses in Area C at Kellis tend to have flat roofs, while the later houses in Area A tend to have vaulted side rooms and two central rooms that may be vaulted or covered with flat roofs.136 Our own B2 may represent a transitional phase in which the open central courtyard is retained but vaulted rooms, which became more common during the fourth century in the Oasis, are also present. Indeed, Kellis Area A houses may bear more similarity to structure B1 at Amheida, which clusters around two central rooms and appears to be some decades later in date than structure B2.137↩#p629
A closer examination of the Kellis houses yields additional points of commonality.138 Area A structures drew upon Classical models in their arrangement, as indicated by the placement of living and work spaces around a central open courtyard. The Kellis Area A houses are single-story elongated structures dating to the late third to fourth century.139 Houses from this area contain well-preserved barrel vaulted roofs over the rooms that surround a central open courtyard. The main core of House 3 from Area A bears a striking resemblance to Amheida’s B2. House 3 has a central, open court (room 6) around which the unroofed stairs (room 7) and main barrel-vaulted living quarters (rooms 2–5, 8–10) clustered.140 Likewise, House 5 in Area A/9 dates to the fourth century, is somewhat elongated, has a central room through which other rooms were accessed, and has barrel vaults over all of the rooms.141↩#p630
Kellis House C/2/8 resembles Amheida’s B2 in its square plan and arrangement of rooms around a central open room. From the surface, Kellis House C/2/10 likewise bears a close overall plan to Amheida’s B2, but it has not been fully excavated, and thus comparisons between these two structures must be tentative at this time.142↩#p631
Area B at Kellis indicates a different phase of construction from Areas A and C and consists largely of extensive complexes. House B/3/1, with finds largely dating to the late first to early third centuries CE, appears to take on a more Classical style layout than is commonly associated with Egypt, as it also has a clustered plan of access and appears to have a central open area (room 1b).143 As in Amheida House B1, the organizing principle seemed to have been around two rooms rather than around one room, which may have been an attribute of wealthier Dakhla houses. If this is the case, it would be a tradition with great longevity, as the occupation of Amheida House B1 falls in the fourth century CE.↩#p632
Food preparation areas in Dakhla show high variability; sometimes they are integrated into the house proper while at other times they have been added on to the house.144 B2 bifurcated domestic cooking needs between rooms 5 and 7, while the aule (exterior courtyard) (C2) contained signatures of food preparation, craft production and animal husbandry. Amheida House B1 shows a similar spread of food preparation areas within and outside of the house. Food preparation areas were almost never located within Roman Period houses in the Fayum, a common practice since Pharaonic times.145 This evidence suggests that Dakhla houses were used somewhat differently from Fayum houses.↩#p633
Moving beyond the Dakhla Oasis to other regions of Egypt, we find some large differences in household forms. Comparisons between B2 and houses from the Kharga Oasis indicate that the two oases may have drawn on different sources for housing types. The unexcavated Roman houses of Umm el-Dabadib, located 38 kilometers north of el-Kharga in the Kharga Oasis, are laid out in regular blocks and consist of at least two stories of barrel-vaulted rooms.146 These houses resemble the houses of the Fayum more closely than the houses of Dakhla, as is particularly evident in the emphasis on vertical form. Likewise, the North Kharga Oasis Survey revealed that most domestic units were multistory (at least a ground and an additional story) with barrel-vaulted roofs. Wall niches were common. A central room appears to have been a particular characteristic of this architecture.147 These houses therefore show a preference for a vertical component that is more typical in the Fayum, but they tend towards a more clustered patterning of rooms, which may have been typical of Dakhla.↩#p634
The four Romano-Egyptian buildings recently uncovered at Douch provide uncertain data regarding the houses, as the boundaries between them were not always clear.148 The average size of these structures is significantly larger than that of B2, as they are approximately 300–350 m², depending on where one determines the boundaries might be located.149 Moreover, it is uncertain if all of them are domestic, although House IV and House II are most likely houses. House III has a large interior court with columns as well as a raised inner portion that does not find ready parallels in Egypt. It may be more closely linked to Late Antique luxury houses in Syria.150 In this respect, these houses from Douch resemble B1 and other luxury houses recently uncovered in Dakhla, although the excavator suggests that it is possible that they were public buildings or for administrative purposes. The evidence from Kharga suggests that there were regional variations on house plans and that sites within close proximity to one another probably influenced conceptions of house design.↩#p635
The Fayum remains the region with our most extensive evidence of Romano-Egyptian housing. Despite the apparent wealth of data, these houses have not been published properly (see Chapter 1). Even so, Davoli’s extensive review of Fayum settlements and houses is particularly instructive in drawing comparisons between Dakhla and the Fayum. She argues that the general plans of houses from the excavated areas reveal that there were no radical changes in private architecture between the Hellenistic and Roman Periods. She goes on to describe the houses as typically with a square or rectangular plan and a wide-ranging number of rooms. The size and location of the rooms varied, and they did not have a uniform shape. The ground floor could be composed of a single room and a central-pillared stairway leading to upper stories and, in some cases, a basement. Most houses were multi-story structures.151 As mentioned earlier, food preparation typically took place in an exterior yard next to the house.↩#p636
From this description it is clear that the published Fayum houses differed considerably from Dakhlan houses. Fayum houses generally lacked a clustered plan of access, central open areas, a horizontal component, and food preparation areas incorporated within the house. Houses at Karanis and Soknopaiou Nesos have a strong vertical component, while B2—and, indeed, Dakhla houses in general—tend to favor a more horizontal component. Likewise, “kitchens” were almost never located within Roman Period houses in the Fayum, while B2 incorporated these areas into the house proper.152 Amheida House B1 also had food preparation areas integrated within the main household core (rooms 4 and 8), since these rooms had a hearth, instruments for grinding, and storage bins.153↩#p637
The houses in Syene, in the south of Egypt, are likely contemporaneous to B2. These Syene houses look quite different from B2, however, as they do not appear to have a clustered plan of access to rooms.154 Furthermore, textual sources indicate that the Syene houses were likely three- or four-storied structures, although there is some archaeological evidence for one-story structures dating to an uncertain period.155 For example, Syene House 5 was a rectangular structure (9 x 10 m) with three rooms and a staircase in its northeastern corner and a courtyard to the north.156 In the case of these features, many known Syene houses share similar features with houses from Karanis.157 Syene House 7, which may date to the ninth century, may have had an interior courtyard that one had to pass through to enter other rooms, although the plan for this house is certainly far from clear and it was much later in date than B2.158↩#p638
Roman Housing Spectrum
Roman houses are notoriously difficult to define because the empire spanned such diverse geographical, environmental, ethnic, and economic regions. House designs are contingent upon the environment as well as individual and local tastes, so the archaeologist must command substantial evidence in order to suggest a trend in Roman architecture as a whole. Scholars generally recognize a few key signatures as typical features for Roman housing, namely a central corridor or court, axial symmetry, and a more elaborate reception room.159 B2 conforms to this general description in some aspects of its layout, namely the central court (if it is that) and the proportional design. There are signs that some rooms were once painted (rooms 1, 3, and 6) and that room 7 was better cared for than other rooms. Outside of this information we cannot say much about the possibility of an elaborate reception room within B2, but it seems unlikely that there was a formal room dedicated to the reception of guests.↩#p639
Ellis has suggested that there are strong regional groupings in the Roman Empire, citing commonalities between Syrian styles and Karanis.160 Vernacular housing types in northern Syria tend to emphasize the vertical plane, while houses and apartments within the multi-story vernacular houses in Italy (e.g. Ostia and Rome) emphasize the horizontal plane.161 These significant differences suggest singular regional developments in Romano-Egyptian housing, although some overlap between Dakhla and the Fayum is evident. For example, Theadelphia had houses with courtyards in the middle of them, more centralized access to rooms and stairways leading from the courtyard area to the roof. There was also a room located underneath the stairs in one of these houses that was accessed directly from the ground floor.162 Unfortunately, these excavations were never completed, and it is not possible to verify this data today. Dakhlan houses do not seem to conform to the apparent Karanis norm of houses that place a strong emphasis on the vertical component and and a linear plan of access. Recent research by the DOP at Kellis and the current work at Amheida suggest that Dakhlan houses fit within a regional norm, but are closer to the horizontal house type found more commonly in other areas of the Roman Empire.↩#p640
Figure 6.4: Reconstruction of House B2 (N. Warner).
Recent results from Dakhla suggest that Alston’s assertion that the lower classes were not able to adopt Romanized housing should be re-examined, since this expectation is not true for all regions in Roman Egypt.163 The occupants of B2 appear to have belonged to a lower level management and services class. Although they were not “lower class,” they certainly were not among the most wealthy within the local spectrum. This new data suggests that less wealthy individuals may have had a modified version of a Roman house type, which was also in line with local housing norms. Despite the strong resemblance of this house to Classical house models, there are some elements in common between this structure and traditional Egyptian domestic architecture. For example, the staggered entrance system, the mud brick building materials, storage areas, and the preference for rectilinear forms show continuity with past housing forms.164↩#p641
The discovery of a particular housing type within the oasis can be taken as an indicator that a wide variety of houses may have existed in Roman Egypt and that there were probably regional housing types. Some of these housing types are currently lost to us due to issues of accessibility, destruction, or lack of scholarly attention. The Nile Valley, in particular, is largely archaeologically lost to us during this period and would have provided significant data on the dominant house types of that region. Even so, results from Dakhla have clearly indicated the need to move beyond the Fayum standard when discussing Romano-Egyptian housing norms.↩#p642
Conclusions and Reconstruction
I will review the relevant data for the reconstruction in the following paragraphs, giving particular emphasis to several features of B2’s overall plan (Figures 6.4, 6.5).↩#p643
Barrel-vaulted roofs are indicated over all of the rooms except for rooms 7 and 5. Room 5 likely had some form of light, flat roof over at least the northern portion of the room. No evidence of roofing could be found in room 7 and comparanda suggest that this room could be open. The possibility of a very light, non load-bearing roof should be kept in mind in future excavations, although it is not represented in this reconstruction.↩#p644
The house is shown as a single story structure. Local comparanda suggest that the majority of houses in Dakhla had a single story. Moreover, the lack of significant debris among the recovered collapse suggests that House B2 had a single story.↩#p645
The access to room 4 is uncertain, and this uncertainty is deliberately left ambiguous in the reconstruction.↩#p646
Windows were added to rooms in areas that seemed logical, but there was no surviving evidence to indicate where these may have been located.↩#p647
The staircase had two turnings, or landings to enable the ceiling height commonly found in the oasis. Of course, more than one landing could not be proved due to the poor preservation of upper levels of this structure, and the possibility of a single landing must be kept in mind, given the comparanda at Karanis.↩#p648
The reconstruction shows parapets and screening with palm. This feature is also a conjecture, as it would provide some light protection to activity areas on the roof.↩#p649
The overall reconstructed layout of B2 closely resembles local housing forms. This local style may be a particular regional style in Roman Egypt. It is possible to find comparanda to this form across the Roman Empire, although local specificities dictate that these comparisons are often partial.↩#p650
Figure 6.5: Reconstruction of House B2 (N. Warner).
4 On the Fayum house sizes, see Hobson 1985. On Tebtynis houses, see Gallazzi and Hadji-Minaglou 2000, Gallazzi 1999. For an overview of Fayum houses by site, consult Davoli 1998.
7 Valbelle 1985:117. Within the Roman Mediterranean domestic spectrum, the average is roughly 75.2 m² and the average property area for a Pompeian house is 271 m². B2 is larger than the average for the greater ancient Mediterranean, although in some locales it would be on the lower end of the spectrum. On the Roman Mediterranean average, see Alston 1997:53. On Pompeiian houses, see Wallace-Hadrill 1994:81, table 4.2. New works on this average are constantly emerging, so we may expect these averages to shift dramatically over time.
17 On the non-conformity of Hawara houses with Davoli’s typology, see Uytterhoeven 2010:322.
18 Hillier and Hanson 1984. Space syntax analysis evaluates room arrangements in order to map out access and the permeability of architectural spaces (Bowes 2010:20).
19 Laurence 1994, Laurence and Wallace-Hadrill 1997, Grahame 2000.
35 On the traditional location of courtyards in Egyptian houses, see Davoli 1998:47. On the central courtyard (aithrion), see Alston 1997:53, Husselman 1979:49-54.
42 P.Ashm. I 7 (187/186 BCE), discussed in Uytterhoeven 2010:323.
43 On aule sold separately from the house, see P.Oxy. III 505; XIV 1696; 1697. On individuals owning more than one aule, see P.Mich. VI 428. It is also important to bear in mind that, due to inheritance practices, it was possible to buy and sell fractions of houses. These fractions could be dealt with practically through the use of “virtual” fractions rather than real fractions of houses. The archaeological record does not provide evidence for the physical division of houses (see Muhs 2008). On these inheritance practices and the impact of marriage on houses, see Pestman 1969, Pestman 1961.
53 Ibid.:133. Luckhard and Husson suggested that the aithrion was the central part of a Ptolemaic house excavated at Ghoran in the Fayum. They believe it was open to the sky above (Luckhard 1914:58, n. 1, Husson 1983:30-31, fig. 1). The excavator suspected that unit E (the central “aithrion”) might have been covered with a ceiling higher than the surrounding roofs (Jouguet 1901:392). Houses from Tel el-Amarna (ca. 1350 BCE) had houses with various rooms opening onto a central hall that had windows above the surrounding roofs (Spence 2004:135, fig. 5). These clerestory windows admitted light and promoted air circulation (Endruweit 1994:66-70).
55 Husselman 1979:9. House C168 does not have the same sort of squared-off plan and clustered plan of access that we find in B2. It was one-story in height and contained underground areas as well as stairs to the roof.
62 On these houses in general, see Davoli 1998:45-50. Davoli believes that it was impossible for Fayum and Dakhla houses to have internal courtyards (pers. comm. 2014).
72 Alston 1997:5, Davoli 1998:47, although a central hall is evident in the domestic architecture of Pharaonic Egypt (Roik 1988:209-211). On the central hall in houses at Tell el-Amarna, see Badawy 1968: 93-94, 97-127, Endruweit 1994:57-78, Spence 2004:131-132, Tietze 2010:95-103.
73 It is worthwhile noting that the Bedouin village to the north of Amheida has houses with central courtyards today. These interior courts are often used for functional domestic purposes as well as sleeping during hot months.
76 Rooms 5–6 in House 2 also formed an L-shape (Hope 1987:161), as did rooms 1 to 6 in House 3 (Hope, Kaper, Bowen and Patten 1989:fig.1).
77 For example, the houses from Pompeii are often arranged such that a dramatic vista was available to individuals when they first enter the house or courtyard (Hope, Kaper, Bowen and Patten 1989:107-122).
78 House 3 at Kellis had a flight of four mudbrick steps leading to the street level as well as a low, semicircular wall added in the street to help keep out windblown sand (ibid., Hope, Kaper and Bowen 1992:41-42). Houses from Karanis often show signs that mudbrick stairs were later replaced by stone steps (Husselman 1979:40).
79 On the restricted use of stone in Romano-Egyptian houses, see Lewis 1983:51.
80 Even more elaborate structures, such as the elaborately painted Kellis House B/3/1, also share this trait (Hope and Whitehouse 2006:315).
83 E.g., from Ain el-Gedida, Mound 1, room B1 had a hearth (diameter c. 90 cm), Mound 1, room B6 has a hearth (diameter not given), room B10 had a hearth (diameter c. 45 cm), and room B19 (diameter c. 58 cm) had a hearth (Aravecchia forthcoming).
99 For example, in the prehistoric societies of the American Southwest, women experienced increased workloads and demands on their time when they developed pottery vessels that were used as the primary food processing and storing containers (Crown and Wills 1995).
122 On Berenike storage jars, see Sidebotham and Wendrich 2001/2. On Karanis trap doors and storage jars, see Husselman 1979:43, plates 7, 35a, 51a, 83a.
138 Area C at Kellis was a vernacular and industrial zone comparable to Area 1 at Amheida. Structures from Area C date to the second and third centuries CE and tend to be uniform in plan (Hope et al. 2006:29).
143 Hope and Whitehouse 2006:318, Hope et al.:23-31, Hope 2007a:33.
144 There is some reason to believe that the Manichees may have had a prohibition on cooking inside houses. It seems that Manicheans may have lived in the Kellis 1–3 houses, and this may explain, in part, the location of the food preparation areas outside of the house. On the Manichean community at Kellis, see Gardner 1993, Gardner and Lieu 1996, Gardner 1997a, Gardner 1997b.
155 On three- and four-story structures at Syene, see Husson 1990:127. On single-story structures at Syene, see Jaritz and Rodziewicz 1994:119. Syene House 5, founded in the first century CE, seems to have had two rectangular halls in the south and at least one large court north of it. A staircase connected the open area with the first floor and gave acess to the western compartments through a vaulted corridor (Von Pilgrim et al. 2004:129). Unfortunately, the entire house has not been excavated and published. On Syene House 5, see also Jaritz and Rodziewicz 1994:especially 121, Jaritz and Rodziewicz 1996:especially 233-234, 237-238.
