At the ancient city of Dura-Europos, a Roman garrison town located along the southwestern bank of the Euphrates (near the modern-day village of Salihiyeh in Syria), private homes were architecturally adapted across the late 2nd and 3rd c. CE by different religious groups to serve the needs of their communities.Footnote 1 The products of such adaptation include the city's Mithraeum, Synagogue, and Christian Building. All three buildings remained in use until the mid-250s CE, at which time the Sasanians lay siege to the city, the population was displaced, and the site was buried.
The Mithraeum, Synagogue, and Christian Building all began as private residences and share a similar architectural development. Yet, the Christian Building's domestic origins have received unique attention and form the subject of ongoing emphasis. This attention has been cultivated and maintained across decades of scholarship in two ways: first, in efforts to situate the structure within a developmental model of Christian architecture that endorses a material progression from the private homes described in the New Testament to basilica;Footnote 2 and second, through the use of terminology that presupposes a quasi-domestic character for the building.Footnote 3
Through a critical reexamination of the archeological and material evidence of the architectural adaptations made to the building at Dura by a Christian community in the 3rd c., this article contends that this emphasis does not do full justice to the archaeology of the structure. In fact, acts of architectural differentiation, enacted during the building's renovation to accommodate Christian community use, crucially distinguished the Christian Building from the domestic structure that had preceded it and from Dura's houses more broadly. Drawing upon quantitative analysis and comparison of three-dimensional reconstructions and daylight simulations of the structure before and after renovation, we show how these acts of architectural differentiation reconfigured the space such that visitors could use and experience it in ways that were categorically distinct from its domestic antecedent. The changes effectively divested the building of the key architectural features that constituted Durene house space and disrupted habituated patterns of occupying the city's residences. Disentangling the material reality of the structure from modern imaginings, the Christian Building emerges as a product of its unique built environment. Thus, we show the long-held view of the structure as occupying a pivotal place in a seamless trajectory of Christian architectural development is untenable. The contextual approach that takes seriously practices of architectural and embodied differentiation emerges as fruitful for understanding religion and the built environment at Dura more broadly. It also calls on us to reconsider the material origins of early Christianity and the domus ecclesiae as a category of ritual space.
The Christian Building in the context of Dura
In the city of Dura, the Mithraeum, Synagogue, and Christian Building were all located on the same street, along the western fortification wall (Fig. 1). The three structures share a similar story, having origins as private houses that were later architecturally adapted to accommodate community use. The Mithraeum underwent two major building phases between the late 2nd and early 3rd c.Footnote 4 The first, undertaken sometime around 169 CE, consisted of razing portions of the existing house and installing an engaged altar, raised podium, and benches. Then, sometime around 209–211 CE, the podium was extended, and an entranceway was cut to provide access between the sanctuary and a suite of chambers to the north.
The Synagogue followed a similar pattern of adaptation.Footnote 5 Little about the first phase is archaeologically discernible. In the second phase of adaptation, which took place sometime in the late 2nd c., walls were removed to create a hall of assembly and a small Torah niche. Most of the structure's eastern portion, which included several chambers and the triclinium, remained untouched. During the third phase of renovation, in 244/45 CE, these chambers were removed to accommodate the expansion of the forecourt and sanctuary. The house to the east, House H, was also annexed into the synagogue complex, and four of House H's chambers were left untouched. This change prompted the expedition's architect, Henry Pearson, to hypothesize that House H's chambers were needed to rehouse people who lived in the original building, after the structure's chambers were removed.Footnote 6
The house that became the Christian Building is widely thought to have been constructed sometime around 232/233 CE (Fig. 2a).Footnote 7 This date is suspect, however, as it is primarily (and problematically) based on an inscription in an undercoat of plaster in Room 4B, rather than on the construction of the building itself.Footnote 8 About a decade later, the structure underwent numerous architectural adaptations, both internally and externally (Fig. 2b), and these are discussed in detail later in this article. Like the city's Synagogue and Mithraeum, the Christian Building remained in use until sometime between 254 CE and 256 CE, when the structure was sealed beneath the rampart and the city besieged by the Sasanians.Footnote 9 The nature of the site's abandonment thus provides it with a firmly pre-Constantinian terminus ante quem and makes the Christian Building at Dura the archaeological record's only securely dated example of a pre-Constantinian Christian community space.Footnote 10
Of the three structures, only the Synagogue is thought to have continued to house people after it was converted into a community cult space.Footnote 11 There is no evidence indicating that the Christian Building ever simultaneously served as a domicile and a Christian community space after the renovations were completed. Nevertheless, since its excavation in the early 1930s by a joint team from Yale University and the French Academy of Inscriptions and Letters, scholars have attributed special importance to the domestic origins of Dura's Christian Building.
Evolutionary models and the limits of language: critiquing the category of domus ecclesiae
The discovery of Dura's Christian Building and its initial publication occurred in a period of scholarship invested in tracing an evolutionary progression from the private homes described in the New Testament to later Christian community architecture.Footnote 12 The category of domus ecclesiae emerged from these discussions as a technical designation for homes architecturally adapted for Christian use in the ante-pacem period, conceived of as a sort of architectural intermediary that still maintained domestic associations.Footnote 13 Understood as emblematic of the presumed house-based nature of early Christian gatherings,Footnote 14 Dura's Christian Building served as the prime example of this architectural phenomenon.Footnote 15 Even as later scholars have refined the definition of the domus ecclesiae,Footnote 16 two misleading assumptions from that period have persisted: first, that houses renovated for Christian community use were ubiquitous across the Roman Empire in the 3rd c.; and second, that such houses at that time would have been designated domus ecclesiae or οἶκος τῆς ἐκκλησίας.Footnote 17 However, neither the archaeological nor the literary evidence supports these assertions.
Dura's Christian Building is the only Christian community building previously identified as a domus ecclesiae that remains securely dated to the ante-pacem period; all the others have been securely re-dated to after the Edict of Milan.Footnote 18 Archaeologists and art historians working more recently in the fields of Roman and Late Antiquity have moved away from evolutionary models of architectural development;Footnote 19 Kristina Sessa's survey of the literary evidence has demonstrated that the term domus ecclesiae and its Greek equivalent οἶκος τῆς ἐκκλησίας are not attested before the 4th c. and that ancient authors did not use the term to denote a specific architectural category.Footnote 20 Nevertheless, the terms domus ecclesiae and “house church” are still widely invoked in relation to Dura's Christian Building as a way of emphasizing its domestic origins.Footnote 21 New Testament and early Christian scholars in particular routinely uphold the building as critical evidence for continuity in the evolutionary development of early Christian church architecture from private home spaces.Footnote 22 However, as we will show, a closer examination of the earliest uses of the phrase domus ecclesiae's Greek equivalent, οἶκος τῆς ἐκκλησίας, reveals that it did not denote a building architecturally adapted for Christian use. Here, we build upon Sessa's findings regarding the architectural indeterminacy of Eusebius's use of the phrase.Footnote 23 In doing so, we seek to highlight how dissimilar the structures Eusebius refers to as οἶκος τῆς ἐκκλησίας were from Dura's Christian Building.Footnote 24
Books 7, 8, and 9 of Eusebius's 4th-c. Ecclesiastical History (Hist. eccl.), as well as a passage in his Martyrs of Palestine, contain the earliest undisputed attestations of the phrase οἶκος τῆς ἐκκλησίας.Footnote 25 In both works, Eusebius uses the phrase τῆς ἐκκλησίας οἶκου to denote a physical building occupied by Christians without offering any qualification for the materiality of the structure to which he is referring.Footnote 26 For instance, in Hist. eccl. 7.30.19, he uses it to describe the 3rd-c. Christian space in Antioch from which Paul of Samosata is evicted after being deposed as bishop of Antioch.Footnote 27 In 8.13.13, the phrase τῶν ἐκκλησιῶν τοὺς οἴκους denotes the buildings belonging to Christians that Constantius I did not destroy.Footnote 28 In Hist. eccl. 9.9a.11 and the Martyrs of Palestine 13.1, Eusebius's use of the phrase is similarly enigmatic: in the former, he uses οἴκους ἐκκλησιῶν to refer to the structures that Maximus did not prohibit Christians from holding meetings in or “building” (οἰκοδομɛῖν);Footnote 29 in the latter, he employs οἴκους ɛἰς ἐκκλησίας δɛίμασθαι to denote the structures that Christians built themselves (δɛίμασθαι).Footnote 30 In both these passages, οἴκους τῆς ἐκκλησίας is the direct object of a verb (either οἰκοδομɛῖν or δέμɛιν) that can mean either to build a new structure or to revamp an existing one. As such, it is unclear whether the structures Eusebius is referring to are domiciles that served as spaces where Christians assembled (like the meeting houses described in the New Testament), purpose-built constructions, or preexisting spaces that were architecturally adapted to accommodate Christian community use. Consequently, Hist. eccl. and the Martyrs of Palestine do not provide conclusive evidence that Eusebius is employing the phrase οἶκος τῆς ἐκκλησίας to describe constructions that are akin to the Christian Building at Dura.Footnote 31
Indeed, in his later works, Eusebius uses the phrase to refer to entirely new constructions, not architecturally adapted ones. In the Life of Constantine (Vit. Const. c. 337 CE), Eusebius designates two post-Constantinian Christian buildings erected by imperial powers οἶκοι τῆς ἐκκλησίας.Footnote 32 In Vit. Const. 3.43.3, he identifies the structure Constantine's mother, Empress Helena, “raised” (ἀνɛγɛίρασα) on the Mount of Olives as a “sacred house of the church” (ἱɛρὸν οἶκον ἐκκλησίας).Footnote 33 In Vit. Const. 3.58.3, he attests to Constantine's construction of “a very large church building” (οἶκον ɛὐκτήριον ἐκκλησίας μέγιστον) in the place of the Temple of Aphrodite at Heliopolis (Baalbek).Footnote 34 In both instances, Eusebius pairs οἶκος τῆς ἐκκλησίας with a verb that is typically used to connote the construction of a new structure, not the adaptation of an older one. In Vit. Const. 3.43.3, the verb is ἀνɛγɛίρɛιν, which means “to raise a new structure.”Footnote 35 In Vit. Const. 3.58.3, the verb governing the sentence is ἐπάγɛιν, which often implied a colonial action or bringing something unfamiliar to a community.Footnote 36 Moreover, Vit. Const. 3.58.3 describes Constantine as “setting in their midst (καταβαλλόμɛνος) also a very large church building for worship,” which implies that it was a new structure.Footnote 37 Further confirming that Eusebius is not referring to a renovated structure in Vit. Const. 3.58.3 is the later church historian Sozomen's account of Constantine's actions at the Temple of Aphrodite. Sozomen describes Constantine as first destroying the Temple of Aphrodite, then erecting a new ἐκκλησία on its ruins.Footnote 38
Admittedly, we can only speculate as to what architectural form Eusebius imagined each time he described a Christian space as an οἶκος τῆς ἐκκλησίας. Yet Helena's and Constantine's imperial constructions occupied landscapes long treated as sacred, such as the Mount of Olives, where they purportedly attracted people away from competing cults. It seems improbable that Eusebius would have considered them architecturally analogous to the structures that the martyrs built near the copper mines of Palestine “to serve the assemblies” that he mentions in Martyrs of Palestine 13.1.Footnote 39 Consequently, it appears that Eusebius indiscriminately applied οἶκος τῆς ἐκκλησίας to Christian space regardless of architectural form.Footnote 40 This would explain why Eusebius retrospectively applied it to Paul of Samosata's οἶκος τῆς ἐκκλησίας (Hist. eccl. 7.30.19), even though it predated Constantine's imperial constructions.Footnote 41 As Sessa has argued, this was likely because Eusebius was writing at a time when there was not an established terminology for Christian architecture.Footnote 42
In our view, the continued reference to Dura's Christian Building as a domus ecclesiae or house church, terminology that emerged out of evolutionary understandings of early Christian material origins, has created blind spots in scholars’ interpretations of it. Even as scholars using domus ecclesiae as a category have considered renovation a requisite characteristic, they have presumed an underlying continuity between the domus ecclesiae and the structure that preceded it, regardless of the degree of renovation it underwent.Footnote 43 The idea that the Christian Building maintained its domestic function and associations has also been bound up with other problematic assumptions about the threat of persecution in pre-Constantinian Christianity in ways that have implicitly marked such Christian structures as different from other renovated private spaces, such as the Mithraeum and Synagogue at Dura.Footnote 44 As a result of such macro-historical and translocal assumptions, scholars have not trained their attention on how practices of architectural differentiation operated on a local scale or considered the implications of these changes for the Christian Building at Dura.
In fact, a critical reexamination of the architectural adaptations made to the structure indicates that the renovations to the Christian Building designed to accommodate new patterns of community use set it apart from Dura's houses in both form and function. Such practices of architectural differentiation disrupted habituated patterns of residential occupation characteristic of domestic space at Dura, engendering new ways of accessing, using, and moving within the building. Taken together, the accumulation of these differences – both to the structure itself and to how bodies inhabited it – reshaped the building's function to such an extent that labeling it a “house church” is misleading on several levels.
In making this argument, we do not mean to imply that ancient buildings were either “domestic” or “cultic,” or that they can or should ever be neatly categorized as one to the exclusion of the other. Indeed, in antiquity, there was fluidity and overlap between private and public, cultic, commercial, and household spaces.Footnote 45 This was particularly true at Dura, where structures were not only regularly readapted, but also often served multiple functions simultaneously.Footnote 46 Rather, we want to highlight the ways in which the adaptations to the Christian Building reflect an intentional move away from household space and to raise questions about how Durenes might have experienced the space from a locally situated perspective.Footnote 47 In doing so, we seek to integrate the Christian Building more fully within Dura's wider adaptive built environment, showing how a deeply contextual approach sheds new light on this much discussed structure.
Methodology
In what follows, we detail aspects of the renovations to Dura's Christian Building, focusing on the ways and extent to which those adaptations differentiated it from the space that had preceded it and from Dura's house spaces more broadly. We consider how acts of architectural differentiation created new ways of using, perceiving, and experiencing the building. Our analysis relies upon multiple modalities. We employ quantitative methods to consider what would have prompted Dura's population to think of it as domestic. To do this, we analyze architectural elements and their combinations in Dura's houses. We then use this information to evaluate the degree to which the architectural adaptations to the Christian Building differentiated it from other domestic spaces at Dura. We also employ three-dimensional modeling and daylight simulations to consider the effect that the renovations had on the embodied and perceptual experience of visitors to the Christian Building.Footnote 48 Underlying our approach is the understanding that “spatiality is not simply defined by the architectonic setting, but emerges in the co-presence of moving human bodies and material culture.”Footnote 49 That is to say that architecture, sensory experience, and embodied practices all work together to engender spatial form.Footnote 50 Architectural differentiation reshaped experiential and usage patterns. In turn, such patterns worked in tandem with architecture to determine the impression the Christian Building gave to its visitors. Focusing on this combined impact enables us to make more comprehensive and historically situated judgments about the ways that the renovations to the Christian Building changed how Durenes experienced it.
Ultimately, the evidence discussed in the remainder of this article indicates the Christian Building came to bear neither a material relationship with the specific domestic structure that had preceded it nor the key architectural features that commonly constituted Durene household space more broadly. To make this judgment, we seek to understand the Christian Building through a locally situated perspective, rather than imposing contemporary perceptions of space onto it. This is challenging as patterns of perception and embodiment are deeply habituated and culturally produced, and the temporal remove that separates modern contexts from Dura's is vast.Footnote 51 However, it is possible to gain insight into Durene sensibilities about their built environment through a methodology that is sensitive to the local dynamics of space at the time.
In antiquity, the boundaries between household and cultic space were permeable.Footnote 52 Nonetheless, we seek to establish what features likely constituted household space at Dura with two assumptions based on such distinctions: first, that the markers of domestic space are discernable in household architecture, and second, that the architectural characteristics commonly associated with household space are in some way delineable from those of other types of space at Dura. The distinctions we apply are not meant to construct a false binary between cultic and household spaces, or even to set these two categories in opposition to each other. Rather, we undertake this approach in order to think critically about what features often constituted a house at Dura and to see how the Christian Building compared. Relatedly, “domestic” and “domesticity” are themselves fraught categories, particularly within household archaeology, which has rightly noted how this terminology wrongly projects modern assumptions about domesticity onto ancient material where it has no place.Footnote 53 Yet, in this study, we occasionally employ the term “domestic” to refer to household space at Dura in order to interrogate the category of space through which scholarship has historically treated the Christian Building.
In order to establish what elements served as markers of domestic space, we select structures at Dura originally categorized as houses and completely excavated by the Yale University and the French Academy of Inscriptions and Letters between 1928 and 1936. Among these 89 structures (not including the Christian Building), we exclude 27 because documentation is prohibitively scant, because the building did not exist simultaneously with the Christian Building, because it was significantly adapted by the Roman military for their own use, or because it was differentiated by an attached shop.Footnote 54 As to the last category, it is often not possible to delineate the extent of commercial enterprise in a given building due to the excavation methods employed at the time.Footnote 55 At the same time, we know that housing associated with the Roman military was often materially, and in some cases geographically, differentiated from civilian homes.Footnote 56
Portions of the Christian Building were first unearthed during the fourth season of excavation at Dura by Yale University and the French Academy of Inscriptions and Letters (October 1930–March 1931). Field director Maurice Pillet, who led the expedition's first four seasons of excavation, identified the structure as “the Edifice of Tower 17” because its connection to the Christian community at Dura and its significance were not yet apparent.Footnote 57 The remainder of the building was uncovered during the fifth season of excavation (October 1931–March 1932), with some final archaeological work continuing until the summer of 1933.Footnote 58 Following the fifth season, in 1934, the then field director, Clark Hopkins, published a preliminary and incomplete account of the findings from the excavation with Paul V. C. Baur.Footnote 59 Though work on the Christian Building continued, the publications slowed after that. During the seventh season (October 1933–March 1934), Pearson, a graduate student of architecture at Yale University at the time, studied the standing remains of the Christian Building and investigated the subsoil under the structure. He produced plans, elevations, and sections of the Christian Building and its various phases. Pearson's drawings and findings from this season did not appear in print, however, until Carl H. Kraeling published the Final Report on the Christian Building in 1967.Footnote 60 Though Kraeling had visited the site in the 1930s, the majority of his magisterial account was not based upon firsthand knowledge of the structure, but rather on the careful analysis of materials contained in the Dura archives at Yale, especially Pearson's and Hopkins's accounts.Footnote 61
Due to the excavation methods employed by Yale University and the French Academy of Inscriptions and Letters in the early 20th c., our ability to provide a nuanced picture of other aspects of the Christian Building is limited. For example, excavators primarily recorded architectural remains, with little attention to small finds.Footnote 62 Yet, small finds are imperative to understanding the function of spaces within the Christian Building and tracing changes in the use of the space over time.Footnote 63 As such, it is challenging to elucidate patterns of use before and after the structure was adapted that are not discernable from the architectural remains. It is also likely that the renovations to the Christian Building happened somewhat gradually, but there is no documentation of the phasing of the renovations beyond the account provided in the excavation reports and the stratigraphic relationships immediately visible in photographs from the excavation.Footnote 64 Presumably, then, the people who interacted with the space as the renovations were at varying levels of completeness would have experienced the space differently. It seems likely that the domestic features of the original structure were more visible during the initial stages of architectural adaptation than after the successive stages of renovation were complete. Similarly, we have little information about the property's history prior to its renovation and what collective memories Dura's community may have held about it. Stories about the Christian Building's formation must certainly have informed how visitors perceived and interpreted it, but these unfortunately escape the archaeological record.Footnote 65
The early excavations at Dura were also not primarily concerned with documenting the houses they uncovered. There is no final report on Dura's domestic structures, and the preliminary excavation reports document the houses with various levels of detail and accuracy. The reports generally give the sizes of structures, the number of rooms excavated, and any notable architectural features found within them. Thus, the points of comparison between Dura's domestic spaces are limited to the features that excavators consistently recorded, like cisterns or staircases. Jennifer A. Baird's monumental The Inner Lives of Ancient Houses provides the most thorough analysis of Dura's houses, and our attempt to establish what key architectural features constituted Durene household space relies on a quantitative analysis of the data that she meticulously compiled in her monograph.Footnote 66
Woven throughout the discussion of the archaeological material in this article are the results of daylight simulations that compare how natural light flowed before and after the renovations. These simulations were executed using revised three-dimensional models of the Christian Building before and after it was architecturally adapted. We have detailed our methodology for creating the daylight simulation models, provided explanations of what the different types of simulations show, and included additional renders in the Supplementary Materials (Suppl. Figs. 2a–9b). There were certainly other sensory phenomena beyond light that shaped experiential and usage patterns at the Christian Building. Unfortunately, because the information that excavators recorded almost exclusively pertains to the architectural findings, it is not possible to reconstruct olfactory and auditory details.
Acts of architectural adaptation as practices of differentiation
When the residence was architecturally adapted to accommodate use by the Christian community at Dura and to facilitate that community's ritual needs, its original configuration changed. Room partitions, ceiling heights, seating installations, floors, and walls all underwent alterations that differentiated the Christian Building from local examples of domestic space. The majority of these renovations affected the courtyard and three rooms at the western end of the building, namely 4A, 4B, 5, and 6 (Figs. 2a and 2b).
During the renovations, partitions between rooms were modified. Changing the size and number of rooms delineated the interior of the Christian Building in ways that departed from the spatial and visual patterns of local domestic space. After the renovations, the Christian Building had six ground-floor rooms (excluding the courtyard), while the average Durene house had seven (excluding the courtyard). While this might not seem like a great difference, it resulted in two spaces – Room 4 and Room 6 – that had particularly uncommon proportions.Footnote 67 The partition wall between Room 4A and Room 4B had been taken down, combining the triclinium with its adjacent antechamber. The resulting Room 4 spanned nearly the entire length of the house and contained 35% of the structure's total floor area.
Room 6 also had uncommon proportions – in this case, its small size and shape. Even before the renovation, the room is thought to have had two entrances, which was rather unusual for a room of such size at Dura.Footnote 68 Then the extensive renovations to make it into a baptistery further considerably reduced the room's usable space.Footnote 69 The floor was excavated to bedrock to construct a foundational bedding for a baptismal font, a process that included adding an extra layer of hydraulic plaster.Footnote 70 A sizeable canopy and basin were laid down. After that, the floor of the room was raised. A new ceiling was then created out of large wooden beams and plaster (Fig. 6a; Suppl. Fig. 1). After the walls were plastered over, they were painted from floor to ceiling with two registers of figural scenes.
The conversion of Room 6 into the baptistery exaggerated the room's already odd dimensions. The installation of the baptismal font against the west wall reduced the total usable floor space of the room by approximately 30%. A new ceiling, bisecting the original room horizontally at a height of approximately 3.60 m from the floor, created a floor for a new “upper story room” (Room 6B), and reduced the wall height in the baptistery by 2.0 m. As such, walking from Room 5 to Room 6 involved a 36% decrease in ceiling height. To imagine the sensory implications of the re-ordering and resizing of this space for visitors, one must consider how these adaptations functioned within the structure. The Christian Building occupied approximately 229 m2, making it 14.5 m2 (5%) larger than the average Durene home.Footnote 71 The creation of one large room (Room 4) and one exceptionally small room (Room 6) produced a spatial polarity between “functional” rooms within the structure that is atypical of Dura's domestic architecture.
The form and arrangement of the architectural decorations that were added to the Christian Building during the renovations were also uncharacteristic of the design of Dura's households. This is especially the case for the decorative modifications made to Room 6, which indicate an appropriation of Durene cultic spatial aesthetics and reflect a shift away from formations of domestic space. Individually, many of the elements added to Room 6 bear a strong similarity to features of the decorative program at both the Mithraeum and the Synagogue at Dura. For example, the drop ceiling in Room 6 and the vault of the font's canopy were both painted dark blue and decorated with motifs of white dots that looked like stars, closely resembling those painted on the reveal of the Mithraeum's niche and the vault of its later shrine.Footnote 72 As Kraeling has noted, both the design and the decoration of the canopy also bore an unmistakable similarity to the Synagogue's Torah shrine.Footnote 73 While wall paintings are not unique to Durene cultic space, the most comparable instances of floor-to-ceiling painted panels like those within the room in the Christian Building come from Dura's Synagogue and Mithraeum. Room 6's painted program framed the baptismal font, above which was a painting of the Good Shepherd.Footnote 74 At the Synagogue and Mithraeum, the placement of paintings likewise worked in tandem with architectural features to create a focal point at one end of the room. In the case of the late Mithraeum at Dura, wall paintings, depicting the life and cosmogony of Mithras, surrounded a niche containing the cult image.Footnote 75 In the case of the Synagogue, paintings of biblical scenes covered the walls and framed a niched Torah shrine, set against the western wall. Given these similarities, the phenomenological associations of being in a room surrounded by images would have prompted viewers to conceptualize the space in a way analogous to neighboring cultic spaces.
These points of visual convergence also suggest that those responsible for the Christian Building's renovation took strong cues from the Mithraeum and Synagogue and may have even employed some of the same workshops.Footnote 76 The use of figural decoration in cult spaces across the wider Dura community evidences the city's view of the utility of figural art for didactic purposes and its appropriateness for community cult spaces. Given that there is no evidence that the community at Dura had other dedicated Christian worship spaces at its disposal, the presence of floor-to-ceiling wall paintings (in a room containing a cult like image, as we have in the Mithraeum) points to the Christian community's desire to articulate their own space through the local idiom of community cult space rather than household space.
The distribution of architectural decoration within the Christian Building also deviated starkly from what was found in the majority of Dura's household spaces. Typically, windows, lintels, and doorframes received the most elaborate decoration at Dura. Within the interiors of the city's houses, the most intricate decorative treatments, such as plaster molding and wall paintings, most often took the form of discrete panels found within the triclinium. The distribution of decorative elements at the Christian Building departed from this pattern. After the renovations, Room 6 was the most sumptuously decorated, while Room 4, part of which had previously served as the triclinium, was left relatively unadorned. It thus seems likely that the local Durene visitor would have found Room 6's wall paintings and the absence of such decor in Room 4 to be highly unusual for a domestic context. The choice to richly decorate only Room 6 of the Christian Building, while leaving the rooms where users would most expect to find decorations undecorated, strongly points to a lack of interest in maintaining the trappings of household space. The insertion of wall paintings thus constituted one practice of architectural differentiation.
The removal of the building's cistern and its food preparation area further delineated the Christian Building from Dura's domestic spaces. Cisterns and food preparation areas were architectural features especially characteristic of local house architecture. As Baird has noted regarding Dura's cisterns, “[a]ccess to water, and management of water, would have of course been an important everyday domestic concern at Dura as it is everywhere, whether it was ensuring availability of potable water during the hot summers or managing the torrential downpours of spring.”Footnote 77 Processing and preparing food is likewise an inescapable reality of daily life. The absence of these elements meant that visitors would have interacted differently with the space than with the houses where they dwelled.
Prior to the renovations, the cistern was located in the northwest corner of the courtyard, but it was paved over when the courtyard floor – originally beaten red earth sprinkled with a dry plaster surface – was tiled with limestone.Footnote 78 Kraeling, mistaking the cistern for a cesspool, thought that this adaptation indicated the house's “transformation into a maison sacrée,” with the paving of the courtyard serving to “dignify” it for “formal use.”Footnote 79 When the installation is correctly identified as a cistern, however, the decision to tile over the feature suggests a more dramatic change in the structure's use.Footnote 80 While cisterns are found in both houses and temples at Dura, many of the cisterns that do appear in temples were originally constructed as part of the building's former use as a residence, as is the case for the Synagogue.Footnote 81 Making a cistern entirely inoperative points to the likelihood that the Christian Building was no longer meant to function as a domicile.
Indeed, no other adaptations to the structure during the renovation compensated for cistern water access. One likely reason for this is that without residential use of the building, smaller quantities of water were needed. Indeed, while the baptismal font that was added during the renovations was clearly outfitted to hold water, as evidenced by the fact that the font was coated with same material as was used in the Roman baths at Dura, water was likely needed less frequently for this ceremony, and was thus obtained and removed through other means.Footnote 82
The adaptations to the Christian Building also seemingly divested it of a food preparation area, and in particular, of any type of oven installation. And, while food production was not a uniquely domestic activity at Dura, it certainly was embedded in the architectural fabric of homes.Footnote 83 Within the city's domiciles, three different types of kitchen technologies are archaeologically attested: tanurs, fire pits, and brick hearths.Footnote 84 Domestic spaces contained at least one of these types of installation, and sometimes all three existed in the same house, or even in the same room.Footnote 85 After renovation, however, none of these cooking installations seems to have been present in Dura's Christian Building.
None of the excavation reports documents the discovery of a tanur, fire pit, or brick hearth in the Christian Building. Nonetheless, it is inconceivable that prior to the renovation, the private house did not have a food preparation area. Where would it have been located? The presence of reed matting, wood rafters, and wood beams in the construction of the roof would have made it too flammable for cooking with a heat source.Footnote 86 Two more likely locations were the courtyard or the alcove under the stairs. If a clay oven once existed in the courtyard, it would have been undetectable after the space was renovated. The alcove under the stairs (Room 7B), floored with packed earth and left open to the courtyard, also could originally have been a food preparation area. If that was the case, it would not have remained so after the Building's renovations, when the floor in the courtyard was raised roughly 65.5 cm higher than the floor in the area under the staircase (Room 7B).Footnote 87 This change in flooring level, along with the addition of a plaster-rubble bench in the northwest corner of the court, covered the lowest steps of the staircase (Room 7A) and the vestibule (Room 8), as well as the single step that led from the courtyard to Room 6 (Fig. 3a–b).Footnote 88 The plaster-rubble bench added to the northwestern corner of the court also obstructed access to Room 7B. All this suggests that the Christian community did not consider it necessary to keep this area accessible for everyday use.
It is important to recognize that cooking facilities within Dura's houses were not put out of sight in the same way that they were at Pompeii's and Ostia's wealthy homes.Footnote 89 Rather, their positions in courtyards and under the stairs of Dura's homes meant that food preparation areas were clearly visible. They were also audible and smellable: the noises and aromas they created were an integral part of Durene home life. Removing them from the Christian Building reconfigured the functionality of its structure, thereby reshaping how visitors experienced it. In recent years, household archaeology has moved towards a framework of understanding domestic architecture through the quotidian activities it facilitates.Footnote 90 If we imagine that certain architectural markers are associated with households because they support the daily activities that most commonly occur within them, the renovations to the Christian Building likely had a profound effect on restructuring the space. By divesting the structure of features necessary to carry out routine functions of domestic daily life, namely the cistern and a food preparation area, the renovations further differentiated the Christian Building from Dura's other household spaces. The practical consequence was that the Christian Building ceased to function as a domestic residence.
Combinations of features
Of course, visitors to the Christian Building did not experience each of the structure's architectural elements discretely. Instead, these features worked together to inform how visitors inhabited the space. Establishing the frequency with which particular combinations of architectural features appeared together within Dura's domestic contexts makes it possible to determine the extent to which the adaptations to the Christian Building differentiated it from Dura's domestic spaces.
As Baird's careful survey of the houses at Dura noted, wall paintings, courtyard staircases, and cisterns are the three architectural features that are recorded most consistently across the majority of the households excavated at Dura.Footnote 91 As such, the points of comparison between the 62 buildings in the dataset that can be quantitatively expressed are related to these three features. Here, we will use the maximum likelihood estimation method to compute how often cisterns, courtyard staircases, and wall paintings appeared together or separately in certain combinations in Dura's households to discern whether a visitor to the Christian Building would have expected to encounter certain features together. By applying this method, we can get a sense of how the addition of certain features during the renovation differentiated it from the town's other domestic spaces.Footnote 92
The Christian Building had a courtyard staircase that pre-dated the renovations and wall paintings that were added during the renovations. Only 9.5%, or seven, of Dura's houses (including the Christian Building) had both a courtyard staircase and wall paintings or pictorial graffiti. Thus, this combination was dissonant with the architecture of over 90% of Dura's household spaces (Fig. 4). When we take into account the removal of the cistern in the courtyard, the difference between Dura's Christian Building and local household spaces becomes even starker. No other house at Dura in this dataset had the combination of figural wall paintings plus a courtyard staircase but no cistern.
The question of residential use: the “Upper Room” and the roof
Despite the extent to which the architectural adaptations differentiated the Christian Building from Dura's domestic spaces, scholars occasionally posit two possible areas of the structure that may have facilitated its simultaneous use as a residence and community space: the so-called second-story room, known as the “Upper Room” (Room 6B), and the roof. However, neither of these arguments is convincing.
The upper story (Room 6B) was not original to the house. Located above Room 6, it was created by installing a new ceiling in Room 6 during the renovations.Footnote 93 When the new ceiling of Room 6 was essentially retrofitted around the canopy, Room 6B was formed in the space above.Footnote 94 That is, Room 6's new ceiling formed the floor of the Upper Room (6B). Based on our reconstruction, the ceiling height in this Upper Room (6B) was approximately 1.75 m.
Judging by Pearson's sectional elevation (Fig. 5), we could conclude that the Upper Room's floor (which, as mentioned, also functioned as Room 6's new ceiling) was sturdy and that the room had a window.Footnote 95 However, in Room 6B, there was no window to the west, and any theories about a window to the south have been entirely speculative due to the state of preservation of the building.Footnote 96 Pearson's reconstruction of the ceiling in Room 6 also made it far thicker than the excavation documentation indicates. In fact, his section drawing (Fig. 5) shows the ceiling of Room 6 as significantly thicker than the roof of the structure. However, the ceiling of Room 6 was actually only approximately 0.27 m thick, roughly 31% thinner than the roof, which had a thickness of approximately 0.37 m (Fig. 6a–b).Footnote 97 Further, the construction of Room 6's ceiling indicates that it was not a structurally sound floor for human occupancy, nor was it ever intended to be (Fig. 6a).Footnote 98 The thick wooden beams running north and south, which formed the “structural” joists of Room 6's ceiling, were in fact only inserted firmly at their north end; at the south end, they were merely tension fit. The rafters, which were placed perpendicularly upon the thick beams, do not appear to have ever been affixed to them. In addition, only the terminal east and west ends of the rafters were embedded in the walls of the room. A thin matting was placed on top of these rafters, followed by a thin mud coating upon the matting. This coating would have prevented the matting and rafters, as well as the reeds which were packed in between the thick beams, from moving when the ceiling was plastered and sealed off. Our revised section drawings of Room 6's ceiling and the Christian Building's roof (Fig. 6a–b) reflect these details. The packed mud on the matting and rafters was considerably thinner than the layer of tamped earth used for the roof construction. Moreover, because the rafters used in the ceiling construction were not affixed to the beams, any movement on top of the packed mud layer of the ceiling of Room 6/the floor of Room 6B would have damaged the decorated plaster ceiling of Room 6. Enough movement on top of the “floor” of Room 6B – Room 6's ceiling – and it would have cracked and eventually collapsed. The low ceiling height, lack of light, and structurally unsound floor of Room 6B all provide conclusive evidence against human occupation of the Upper Room (6B). The idea that the second story would have served a domestic purpose is unlikely anyway, since functional second stories were rare among Dura's households. In fact, the only two houses excavated at Dura that certainly had upper stories had both been adapted for use by the Roman army, thus suggesting that the utilization of second stories within households was likely limited to this group.Footnote 99
The question of whether the roof of the Christian Building may have provided an area for continued domestic use of the structure is less easily answered. Like most structures at Dura, the Christian Building had a flat roof. It was reached via an enclosed staircase (7A) accessed through a doorway off the courtyard. This doorway swung inward towards the stairs (see Fig. 3a–b) and could be barred from inside the staircase (7A). Though the stairs were only preserved up to the third landing, Kraeling rightly suspected that they continued up to the roof over the vestibule, thus providing access to the roof over the other chambers of the building.Footnote 100 The staircase, the doorway, and the door's locking mechanism are all thought to have been original to the residence and remained accessible after the renovations to the Christian Building.
The functionality of the roof, both before and after the building's adaptation, remains elusive, however. The picture is again further complicated by Pearson's sectional elevation drawings (Fig. 5), which provide a reconstruction that does not align with Kraeling's descriptions or the excavation photographs. If the details Kraeling provides regarding the roof construction are correct, as seems to be the case, it was certainly occupiable. The general consensus based on other Near Eastern case studies is that Dura's roofs were flexible spaces, which could have been used for a range of activities, including processing foodstuffs, storage, and sleeping, depending on the season.Footnote 101
Whether the functionality of the Christian Building's roof changed after it was adapted is not known. Given the option to bar the door of the staircase (7A) shut, the roof space could certainly have been separated from the activities happening on the ground floor of the Christian Building. Considering how little we know about the nature of the roof's functionality, certainly individuals may have continued to occupy it in a way consistent with how they had prior to the structure's adaptation, yet this seems unlikely given that access to water was limited after the cistern was removed.
Experiencing the adaptations: movement and illumination
On one hand, renovations, such as the addition of wall paintings and the removal of the cistern, differentiated the Christian Building in function from Dura's domestic spaces and foreclosed possibilities of occupying the structure in ways that were consistent with how it had been used prior to its conversion. At the same time, these adaptations also engendered new ways for visitors to inhabit and engage with the structure, which bore little resemblance to patterns of residential occupation representative of domestic space at Dura.
The results of our light simulations indicate that the way natural light flowed within the Christian Building also changed radically as a result of the renovations. When the courtyard was repaved with stone tiles, plaster-rubble benches were added in the southwest and northwest corners. Paving the courtyard raised the height of the floor by 8 cm. This new tile surface had a high light reflectance (Fig. 7a–b; Suppl. fig. 3a–b). Unlike the original beaten red earth sprinkled with dry plaster, which reflected light diffusely, the tile reflected light specularly (Fig. 7a–b; Suppl. fig. 3a–b). The specular reflection of light off the tiles meant that light entered rooms with entrances off the courtyard at the western end of the building more intensely, namely Rooms 4, 5, and 6 (Fig. 7a–b). Moreover, when the low windows were added to Rooms 4 and 5, the light that reflected off the courtyard also entered these rooms rather than reflecting off the walls of the southwest corner of the court (Figs. 7 and 8). The implication of these changes was that, if their doors were open and their windows were unshuttered, a greater area of each of these rooms was illuminated and the rooms could be used at more times throughout the day without the need for a lamp or candle (Fig. 7a–b; Suppl. fig. 4a–b).Footnote 102 Moreover, because a greater area of the western rooms off the courtyard was illuminated, more people could comfortably congregate in the rooms without needing a secondary source of light.
These changes in lighting worked in combination with other architectural strategies to disrupt habituated patterns of residential occupation typical of Dura's domestic spaces, and to afford new ways of using and moving within the building. For example, several of the Christian Building's features after renovation were designed to monitor and mediate natural light within rooms. The two gypsum-paneled windows in Room 4, for instance, had draw cords.Footnote 103 The low window installed in the north wall of Room 4 had wooden shutters, and the low window in the east wall of Room 5 likely did as well (Fig. 8a–b).Footnote 104 Users of the building could employ the cords and shutters as well as the doors onto the court to darken these rooms if they wished.
The modifications made to the Christian Building's windows indicate an impulse to restructure the movement of bodies within the space. During the renovations, two low-level wood-shuttered windows were inserted: one in the southeast corner of Room 5 and another on the north wall of Room 4, west of the door to the courtyard (Fig. 8a–b).Footnote 105 This type of aperture is evident in only one other instance at the site, and it is not in a private residence, but rather in the principia, or Roman military headquarters building, in Block E7.Footnote 106 Pivot holes in these windowsills at the Christian Building indicate that the shutters swung inward towards the interiors of the rooms.Footnote 107 In doing so, the windows provided visual access from the rooms onto the court and also facilitated communication between people standing in the court and those within Room 4 and Room 5. These windows would have controlled the flow of people and sound between the courtyard and Room 4, as well as between the courtyard, Room 5, and the baptistery (Room 6), which could be accessed through Room 5. Moreover, the restriction of the features to two rooms suggests a distinct concern with regulating people's departure from Rooms 4 and 5 and the court. Such a conclusion hints there was a ritualized pattern of movement between these three spaces following the structure's renovation.
Adaptations to the Christian Building's doorways suggest a similar concern with regulating movement between rooms. The baptistery (Room 6) had two entranceways, both located in its south wall. The first doorway provided access between Room 5 and Room 6, opening into Room 6.Footnote 108 Though this doorway appears to have been original to the structure, the proportions of the door and evidence of later embellishment suggest that one face of the doorframe was adapted during the conversion of the building. Namely, a lavish molded plaster trim was added to the southern face of the doorway in Room 5, but not to its northern face in Room 6. This strongly suggests that the renovation was made with the expectation that visitors would now enter Room 6 through Room 5 but would not enter into Room 5 from Room 6. Moreover, the plaster trim made the doorway far more elaborate than the entryway between Room 5 and the courtyard. As Kraeling noted, this degree of decoration was atypical of the internal doorways of Dura's houses; doorways between the courtyard and a house's larger rooms, or between the triclinium and its adjoining rooms, usually received the most elaborate decoration.Footnote 109 The choice to decorate the doorway between Room 5 and 6 rather than the doorway leading from the courtyard to Room 5 further hints that after the renovation, people likely entered Room 5 through the doorway at the west end of the north wall of Room 4 and then carried on through Room 5 into Room 6. The fact that the doorway into Room 4 from Room 5 was left undecorated lends further support to this conclusion. The second entrance to the baptistery was a narrow door, set 0.5 m from the east end of the wall, which connected Room 6 with the courtyard. This door also opened into Room 6 and, like the entrance to the staircase, could be barred from the inside using a vertical lock bar.Footnote 110 The baptistery's two doorways are puzzling; it was highly unusual for a Durene home to have a chamber of such small size with two entrances.Footnote 111 Moreover, the locking mechanism for the door between Room 6 and the courtyard was cut into the existing doorframe,Footnote 112 hinting that the mechanism was not original to the structure but added later as part of the renovation of the baptistery. The two doors would have enabled a traffic pattern unlike those of other Durene domestic chambers of such size. The juxtaposition of the size and decoration of Room 6's two doorways suggests that, after entering Room 6 from Room 5, individuals exited Room 6 into the courtyard. This conclusion also stands in opposition to the received opinion about the liturgical use of the space, which has supposed that initiates processed in the same direction as the women depicted in the lower register of the wall painting, located at the west end of the north wall.
The removal of the built-in benches from the triclinium (Room 4A) and the addition of plaster-rubble benches in the southwest and northwest corners of the courtyard and Room 6 also suggests an interest in regulating visitors’ movements.Footnote 113 Benches in the triclinium were very prevalent in Durene houses the size of the Christian Building. Their removal points to an effort to structure visitors’ movements through the space by means of visual cues, subliminally signaling it was not appropriate to linger in the triclinium. It also suggests a recapitulation of the courtyard space for use by visitors. Paving the courtyard and adding benches to it would have made it more welcoming than the assembly room (Room 4). As Kraeling noted, benches indicate the presence of groups larger than the typical family, suggesting an effort to accommodate a larger number of people than were typically present in household gatherings.Footnote 114
The results of the daylight simulations also present new clues about how the reconfiguration of natural light worked together with other adaptations to provide visitors to the Christian Building with an embodied experience of individual rooms that was categorically different from the one provided by its domestic antecedent. This is especially true for Room 4. Prior to the renovation, the partition that separated Rooms 4A and 4B contained the light that entered the windows along the southwest wall of Room 4B within that small chamber. At the same time, light almost exclusively entered Room 4A – the house's triclinium – unidirectionally through the doorway off the courtyard when open and only softly lit the room (Fig. 7a). In the late afternoon and evening, this light would have gently illuminated the low benches that lined the walls of the room, where guests would recline during dinners (Fig. 9a; Suppl. figs. 5a, 6a, 7a). The lighting of the room, in addition to the U-shaped benches, encouraged guests to recline along the benches, facing the center of the room. Typically, in the triclinia of Dura's domestic spaces, there was no such singular line of sight. While a specific axial view may have been more privileged in cases where the triclinium was decorated with murals, for the most part, guests were able to orient themselves freely towards the center of the room.
The architectural adaptations to Room 4A and 4B disrupted these habituated practices of domestic inhabitation. Removing the partition wall between Rooms 4A and 4B created an unobstructed line of sight within the space. Upon the removal of the partition, a new floor was packed down. This floor covered the original plaster floor and the plaster benches that had formerly lined the south, east, and west walls of Room 4A, originally the structure's triclinium, creating a downwards slope of 0.07 m towards the east wall of Room 4. This slope would have led the visitors’ eyes downwards towards the slightly rhomboidal dais made of low rubble and plaster that was subsequently installed against the east wall of Room 4 (0.97 × 1.47 × 0.20 m) (Fig. 2b).Footnote 115 These adaptations oriented bodies to stand facing the east wall, not to recline towards the interior of the room as the previous space had prompted them to do and as they normally would have done in Dura's triclinia.
The change in how natural light flowed into Room 4 after the renovations also encouraged people in Room 4 to assume this new posture when they inhabited the space. After the adaptations, light entered the room from three or four points: the door off the courtyard if left open, a new low window beside the door, and the two windows on the west wall (Fig. 7b). In the afternoons, the light from the windows along the west wall (roughly 2.35 m above floor level) streamed in and reflected off the room's east wall.Footnote 116 This illuminated the northeastern corner of Room 4 and in the late afternoon created a sort of soft floodlight above where the dais was located (Fig. 10b; Suppl. fig. 9b). When the door off the courtyard was left open, the light over the dais was even brighter and more comparable to a spotlight than a floodlight (Fig. 9b; Suppl. figs. 5b, 6b, 7b). This lighting effect was most intense around 3:30pm in the afternoon and lasted until around 5:30pm, though it was more subdued in the winter (Suppl. figs. 5b, 7b) than in the summer (Figs. 9b, 10b; Suppl. fig. 6b). The area of the wall directly above the dais was also brightly illuminated immediately after sunrise, around 7:48am (Fig. 11b; Suppl. fig. 9b), though primarily by light that poured in from the low window and the entrance onto the courtyard. Tower 17 and the wall directly west of the building blocked the morning light from entering through the two windows at the western end of Room 4. The results of the daylight model, therefore, lend support to Kraeling's hypothesis that the dais functioned as a platform where a person sat or stood when addressing an assembly. (It is also possible, however, that an object may have been placed upon the dais and not a person.) These findings help to explain why the rubble dais was set against the east wall of Room 4 and immediately beside the doorway to Room 3, instead of being centered on the wall. Because of where the light hit the east wall, had the dais been installed against the center, it would have been more difficult for the occupants of the room to see the person or thing standing upon it.
The axial organization of bodies towards the east wall might explain why the community responsible for the renovations did not feel it was necessary to remove Room 4's original decoration entirely. A small portion of a plaster Bacchic frieze, which is thought to have originally decorated all four of the room's walls before the renovations, remained at the east end of Room 4's north wall afterwards.Footnote 117 If our reconstruction is correct – that when the Christian community congregated in Room 4 they faced the east wall – then the frieze left in situ would have been largely out of view, except for when exiting the room.
However, Room 4 received consistent light throughout the day, and Durenes had other forms of lighting technology available to them. Thus, it is probable that the Christian community at Dura also gathered in Room 4 at times of day other than the early mornings and late afternoons and that these configurations did not have the same axial nature as the early morning and late afternoon gatherings. Even so, with the removal of the triclinium's (Room 4A) benches, these configurations would in no way have resembled how individuals had congregated in the space prior to the renovation.
The daylight simulations also suggest that the renovations to the Christian Building engendered new patterns of movement between rooms with ritualized uses. In the late afternoon, when the dais in Room 4 would have been most illuminated, the baptismal font in Room 6 was softly illuminated, but the lower half of the northwest wall, where there was a series of wall paintings, would have been well lit (Fig. 9b; Suppl. figs. 5b, 6b, 7b) if either or both of the doors was open or even slightly ajar.Footnote 118 It is unclear whether either entrance to Room 6 would have been open during the baptismal ritual, but given the illumination of these spaces in the late afternoon, it is possible that certain types of ritualized activities were performed in Room 4 then and that the baptism was performed around the same time. If this reconstruction is correct, then it would represent another way in which habituated patterns of movement, facilitated by the orchestration of light, were altered as a result of the renovations to the Christian Building.
Conclusion
This investigation has sought to challenge the well-established interpretation of Dura's Christian Building as a house church or domus ecclesiae by reconsidering the archaeological remains of the building in relationship to Dura's domestic architecture. Our reexamination reveals that the renovations to the Christian Building constituted acts of architectural differentiation that engendered new ways of experiencing and perceiving the structure that bore little resemblance to the established practices of residential occupation characteristic of Dura's houses. With our shift of attention away from the domestic origins of the structure and toward micro-level archaeological analysis, Dura's Christian Building emerges more fully as a product of its unique built environment, one in which structures were commonly readapted, reused, and repurposed.
If this study of Dura's Christian Building provides any direction forward, it is that Christians took their architectural cues from the spatial vernacular of their immediate contexts. As such, it would be helpful to reconsider the post-adaptation form of the Christian Building at Dura in relation to the city's other cultic architecture, and not just the Mithraeum and the Synagogue. What other insights might spaces such as the Temple of Aphlad or the Temple of Artemis provide about what early Christianity at Dura looked like in practice? It is also time to revisit early Christian architecture more broadly, by sponsoring other studies that work at the intersection of the material record, embodied practice, and sensory experience, with an eye toward developing a more textured sense of how ancient religious communities perceived and interacted with their built environments.
Acknowledgments
We are grateful to our many colleagues who provided us with their encouragement and invaluable feedback on this article. Foremost among them are D. An, J. A. Baird, D. Bohac, S. Bond, C. E. Bonar, A. H. Chen, S. J. Davis, M. E. Doerfler, D. Frankfurter, F. Harley-McGowan, A. McGowan, R. MacMullen, V. Marinis, C. Moss, L. S. Nasrallah, C. R. Rice, K. B. Stern, N. Smith, our anonymous peer-reviewers, and the editors at JRA, most especially G. Woolf. We are also appreciative to YUAG and L. Brody, who kindly provided us with access to the Dura collection and archive. We are especially indebted to our steadfast mentors S. J. Davis and F. Harley-McGowan who helped us turn an idea, conjured in their seminar nearly a decade ago, into this article.
Supplementary Materials
The Supplementary Materials contain a description of our methodology for creating the daylight simulations analyzed in this article, an explanation of what the different types of daylight simulations included visualize, and additional renders. To view the Supplementary Materials for this article, please visit https://doi.org/10.1017/S1047759424000126.