Navigation – Plan du site

AccueilNuméros95Dossier : Sur les routes de Syrie...Religion in Early Fourth Century ...

Dossier : Sur les routes de Syrie et d’Asie Mineure – Hommage à Maurice Sartre

Religion in Early Fourth Century Roman Palestine: The Kfar ʿOthnay Mosaics

Glen W. Bowersock
p. 65-72


Deux mosaïques de Palestine occidentale, découvertes récemment en Israël et datées vers la fin du iiie siècle de n. è. (ou juste au début du ive), méritent une analyse comparative : l’une est incontestablement chrétienne mais l’autre reste indéterminée, peut-être juive ou chrétienne ou païenne. L’analyse propose des explications liées à la diffusion des cultes dans la société palestinienne à un moment de transition.

Haut de page

Texte intégral

1My friendship with Maurice Sartre and my admiration for his wide-ranging scholarship on the Roman Near East both began many decades ago. I have been inspired by his mastery of epigraphy and his sovereign command of historical analysis. From his post in Tours he has launched several of the best scholars of the next generation under his expert tutelage. His is a legacy that will endure. I went to Tours to pay tribute to him upon his retirement, and I now offer him these modest reflections on the religious environment in Palestine in the late 3rd and early 4th cent. This was long after the bruising rebellion of Bar Kokhba but still well before Constantine transformed the new Byzantine empire into a Christian realm. It is astonishing that two important mosaics of almost the same date, one with no indication of faith and the other with explicit reference to Christianity, should have been discovered in the region within a few years of each other. This paper will be devoted to the Christian mosaics at Kfar ʿOthnay, but a brief allusion to the enigmatic, though undoubtedly more beautiful, ones at Lod, which have now been published by the Israel Archaeological Authority, is necessary to introduce what exists at Kfar ʿOthnay.

2The magnificent mosaic carpet at Lod was uncovered in 1996 and has been enjoying a triumphal tour while the museum destined to house it at its original site has been under construction. It began its tour at the Metropolitan Museum in New York and has passed through, among other places, San Francisco, Paris, Berlin, and Waddesdon Manor in England. It has now ended its journey at the Hermitage in St. Petersburg before returning to Israel. Although the traveling mosaic did not include all the finds on the site, it did include the large and impressive central panels, which, with one exception, are in excellent condition (fig. 1).

Figure 1.

Figure 1.

The Lod mosaic carpet, including the fragment on the northeast side

© Israel Antiquities Authority.

  • 1 The full publication of the Lod mosaic under the auspices of the Israel Antiquities Authority, edi (...)
  • 2 The centurion must have come from the nearby VI Ferrata at Caparcotna, which had moved to Udhruḥ b (...)

3Curiously, only nine years after the discovery at Lod, another astonishing mosaic of approximately the same date appeared during work at the Megiddo prison, close to the site of the Roman legionary camp for the Sixth Ferrata at Caparcotna. The site, known under various names to classical and rabbinic sources, is called Kfar ʿOthnay, and a summary publication by Yotam Tepper and Leah Di Segni promptly appeared just one year after the discovery of the mosaic. 1 Although the two mosaics are very different in quality —one is a masterpiece and the other is not— they appear from the archaeological testimony to be close in date. Excavated finds, such as coins and ceramic fragments, point to the late 3rd cent. ce for both sites. Although the Lod mosaic shows no clear evidence for the building that might have contained it or what the religious beliefs of those who congregated there might have been, the Kfar ʿOthnay mosaic is unambiguously Christian (fig. 2‑3). Both towns presumably had mixed populations of Jews, Christians, Samaritans, and pagans, although the relative strength of any of these is beyond telling. The mosaic images at Lod point to influences from Africa, but the austere simplicity of the Kfar ʿOthnay mosaics shows only the role of the occupying Roman force. A donor of one of the panels identifies himself as a centurion. 2

Figure 2.

Figure 2.

The Kfar ʿOthnay floor, with three inscribed mosaics

© Israel Antiquities Authority.

Figure 3.

Figure 3.

The Kfar ʿOthnay mosaic of Akeptous with reference to God Jesus Christ

© Israel Antiquities Authority.

4The two sites are just under 90 km apart, at the southern and northern extremities of the Plain of Sharon, near the coast in western Palestine. Lod lies not far to the southeast of the modern coastal city of Tel-Aviv and Kfar ʿOthnay lies well to the north of it. Lod became famous in late antiquity for its shrine of St. George, and it appears on the Madaba mosaic map with both its indigenous name of Ludea (Lydda) and its Graeco-Roman name Diospolis. This was the name it had already received when it became a city under Septimius Severus. The site also appears in the extant copy of an ancient map that is known as the Peutinger Table. It reappears yet again as Diospolis in the great mosaic at the Church of St. Stephen at Umm er-Rasās in Jordan, where the major cities on both sides of the Jordan valley are represented with icons that evoke their most famous buildings. For Diospolis the building depicted at Umm er-Rasās was most probably the Church of St. George, for which the late antique city was renowned. On the other hand, the name of Diospolis for Lod is unique in Palestine, although half a dozen cities of that name are known from Roman and Byzantine Egypt. That would tend to reinforce hints in the mosaic itself of influences from Africa, particularly in the animals it depicts.

5Kfar ʿOthnay to the north was initially called Caparcotna and later Maximianopolis, commemorating Diocletian’s colleague at the end of the third century. The presence of a Roman legion eventually imposed the name Legio or Legeon, from which modern Lejjūn was formed. The Lod mosaic is artistically superb, although puzzling because it lacks any indication of its purpose or the religion of its owner. By contrast, the Kfar ʿOthnay mosaic explicitly belongs to a Christian hall dedicated to worship, although it is not part of a church —which is hardly surprising in view of its early date. Such a hall is often called in the scholarly literature an aula ecclesiae, and that was what Tepper and Di Segni intended by their designation “prayer hall.”

6Although the mosaic at Kfar ʿOthnay is technically inferior to the masterpiece at Lod, it is extraordinary in its own way. It consists of four panels placed within a floor in the middle of which is some kind of raised platform. In the largest of the panels two fish appear in a roundel and at the side a Greek text reveals that a Roman centurion called Gaianus qui et Porphyrius generously paid for the mosaic from his private resources. It also identifies the man who laid the mosaic as Broutios, i.e. Bruttius:

Γαϊανὸς ὁ καὶ Πορφύρι(ο)ς (ἑκατονάρχης) ἀδελφὸς ἡμῶν φιλο|τειμησάμενος
ἐκ τῶν ἰδίων vvv|vacat ἐψηφολόγησε · Βρούτι(ο)ς ἠργάσετ[o]

“Gaianos, also called Porphyrius, centurion, our brother who made this benefaction from his own funds, laid down the mosaic. Bruttius did the work.”

7On the opposite side of the central platform another panel has two Greek texts, one asking that four Christian women be remembered. They are named as Primilla, Kyriakê, Dôrothea, and Chrêstê. The other Greek text, facing in the reverse direction from the commemoration of the women records that a god-loving woman by the name of Akeptous presented the table (τὴν τράπεζαν), by which an altar-table is unquestionably meant, to God Jesus Christ:

ημονεύσατε | Πριμίλλης καὶ Κυρι|ακῆς καὶ Δωροθέας, |
ἔτι δὲ καὶ Χρήστη<ς>

“Be mindful of Primilla and Kyriake and Dorothea, and also of Christe.”


  • 3 These three mosaic inscriptions are reviewed by Denis Feissel in the Bulletin épigraphique 2008. 6 (...)

Προσήνικεν | ̓Ακεπτοῦς | ἡ φιλόθεος | τὴν τράπε‖ζαν Θ(ε)ῷ ̓Ι(ησο)ῦ Χ(ριστ)ῷ | Μνημόσυνον 3

“The god-loving Aceptous presented the table to God Jesus Christ as a remembrance.”

  • 4 Aupert & Masson 1979, p. 367 n. 27; Bulletin épigraphique 1990, 809.
  • 5 Ignatius, Epist. Eph. 18.2; Epist. Rom. 3.3. For recent discussion of the date of Ignatius, see Ba (...)

8This third text is remarkable both for the donor’s name, which is a variant of the more common Accepta in a feminine form ending in –ous, and for the designation of Jesus Christ as God. Feminine names such as Akeptous are both rare and relatively late, as Olivier Masson has demonstrated. 4 They are not found before the second century or so, and are most common in Egypt. The unqualified Christian identification of Christ with God is similarly rare and only occurs outside doctrinal debates before Constantine. The letters of Ignatius to the Ephesians and the Romans, from the first half of the second century, provide the most unambiguous and earliest examples of Christ as God. 5

  • 6 Kennedy & Falahat 2008, p. 151-169. Ca. 303-4 at Udhruḥ: p. 160.

9It has seemed to most scholars that the raised platform in the floor with the mosaics must be precisely the altar-table that is mentioned in the mosaic panel that lies right beside it. In her recent book on Palestinian mosaics in antiquity, Rina Talgam raises the possibility that this platform might be the remains of a square pillar supporting arches implied by engaged pilasters in the walls, but she does not press this point. More significantly she also expresses some doubt about the archaeologists’ dating of the mosaic floor to the late third century. She admits that on the grounds of artistic style she would be inclined to a somewhat later date, but she firmly acknowledges, “Art historians should not ignore the strong arguments made by the excavators.” In fact, the archaeologists’ dating is now incontestably confirmed by the presence of the VI Ferrata legion at its new camp in Udhruḥ, near Petra, by about 304, as revealed in a Latin inscription found at the site. 6 It could no longer have been at Caparcotna after that. Furthermore, philological considerations are also fully consistent with the archaeologists’ dating, not least the name Akeptous and the aberrant spelling for the verb describing her gift, προσήνικεν for προσήνεγκεν. The identification of Jesus Christ with God similarly militates strongly against a date later than the third century.

10The centurion donor, who was undoubtedly attached to the adjacent legionary camp of the Sixth Ferrata, offers, through his personal generosity (φιλοτειμησάμενος ἐκ τῶν ἰδίων), a glimpse into the spread of Christianity in the Roman army before Constantine. His Latin name Gaianus reveals little apart from his Roman identity, but his agnomen Porphyrius presumably individuated him in some way, just as the same name given to the Neoplatonist philosopher who was originally called by the common Semitic Malchus.

11The reference to the altar-table presented by Akeptous leaves no doubt about the purpose of the space in which these mosaics appear. This was the table from which the Eucharist would be celebrated in this fledgling Christian community. Whether the hall was part of a house or villa or a free-standing meeting place is impossible to determine, but it was manifestly not in anything like a church. The implements for the Eucharist are not mentioned or depicted, but they must have included some kind of chalice as well as vessels from which the wine would have been poured.

  • 7 Talgam 2015.

12It is this table (τράπεζα) which directs us back to the contemporary Lod mosaic. On the east side of the principal mosaic at Lod there is another fragmentary image which is not perfectly aligned with it but appears to confront a visitor moving out from the main space. It seems improbable that this fragmentary piece served to greet visitors upon entering the area with its animal mosaics, although Rina Talgam thinks of this fragment as somehow promising an opulent banquet: 7 “The small panel projecting eastward from the northern carpet, and apparently relating to a doorway, was intended to express wealth, status, and pleasure by mean of the presentation of valuable luxury items connected with the drinking of wine and banqueting.”

13She has rightly observed that what can be seen in the panel is an elegant three-legged table upon which sits an even more elegant krater, which she supposes may be made of glass. The object on the table is very different from the purely decorative krater sprouting tendrils, just below the maritime scenes of the principal mosaic, in a fanciful rectangular panel that may have replaced an original wall. Talgam notes that on either side of the table in the east fragment are two amphoras on low stands, and she plausibly describes these containers as “not storage amphorae but elegant banquet vessels that contained wine.” She conjectures that their yellow color implies they were made of metal.

14Although this small and neglected panel at the northeast edge of the main mosaic at Lod understandably did not go on tour and is not therefore so well known, it may actually help with the riddle of the mosaic and its building. As Talgam noticed, it appears to be related to a doorway at the eastern side of the great hall, but its orientation implies that it would be seen by someone leaving the hall to go into the adjacent room to the east. It could hardly greet someone who was entering from the east, because it is oriented towards a viewer on its west side. Besides, since the hall itself is so grand, it would be hard to believe that guests would leave it for the kind of sumptuous banqueting that Talgam imagines as promised by the krater and amphoras.

  • 8 Greg. Naz., Orat. 5. 2.
  • 9 Lampe 1969, p. 1484.

15But Kfar ʿOthnay raises another possibility. A table with a krater on it may be exactly the kind of τράπεζα that Akeptous presented —a Eucharist table. The krater itself, intricately made of glass or metal, would be the chalice for the ritual. In fact, this very conjunction of table and krater turns up explicitly in the second invective of Gregory of Nazianzus against Julian. He recounts with evident relish instances of sacrilege committed by the emperor’s acolytes and the consequent illnesses, ruptures, and deaths that their deeds have brought upon them. Gregory invokes violations of holy places (περὶ τοὺς θείους οἴκους), and he specifically speaks of desecration of altar tables and eucharistic chalices (περὶ τὰς ἱερὰς τραπέζας … περὶ τοὺς μυστικοὺς κρατῆρας). 8 The documentation in Lampe’s patristic lexicon leaves no doubt that μυστικός was a word used in reference to the Eucharist. 9

  • 10 In the New York Review of Books blog for 14 February 2011 ( (...)

16Accordingly, in the small east panel at Lod, with its chalice and table, flanked by special amphoras of exceptional manufacture, presumably for communion wine, guests who were leaving the main hall through a doorway to the east would become aware of the sacred space they were entering. That may therefore have been an aula ecclesiae, certainly more elegant than the one at Kfar ʿOthnay but equally in a private building, not in a church. The excavator at Lod, Miriam Avissar, has suggested that the context might have been Christian on the basis of the images of fish and the great sea-monster, which she sees as a reference to Jonah. If her interpretation of the main mosaic and mine of the east fragment are on the right lines, the Lod building must have been, as conjectured by others in the past, an opulent private villa. The testimony of the east panel suggests that the villa incorporated a pre-Constantinian church-hall, which would provide another of the rare examples of early Christian places of worship outside a church building. I have suggested elsewhere that the central medallion in the great hall of the Lod mosaic may represent Isaiah’s peaceable kingdom. 10 In view of the undeniably Jewish origin of Isaiah’s vision it would be agreeable to think that in the mixed society of Lod those guests who came to the villa could as readily have been Jewish as Christian, and that they would have felt comfortable in this environment.

  • 11 I am extremely grateful to Gideon Avni for sending me a photograph of this new mosaic. It appeared (...)

17In 2009 new excavations at the Lod site uncovered another mosaic that offers a dramatic confirmation of the foregoing interpretation of the krater on a table in the east fragment. The new image shows a krater comparable to the one on that fragment as well as to those on mosaics at Petra and Maʿin, but with the significant addition of two doves, which can only be understood as Christian symbols of peace (fig. 4).11

Figure 4.

Figure 4.

The new Lod mosaic with doves

© Israel Antiquities Authority.

18The mosaics at Kfar ʿOthnay and Lod, of approximately the same date, afford together a hitherto unprecedented glimpse into the complexity of pre-Constantinian Christianity in Palestine. Their social links with the Roman army at the northern site are balanced by links with the surrounding Jewish community at the southern one. The visual parallels at Daphne near Antioch and at Maʿin in Jordan had their roots in the common heritage of the Hebrew Bible, and the Petra Church exemplifies the longevity of the third-century depictions of animals and objects. If the space beyond the surviving fragment to the east of the main mosaic at Lod really was an aula ecclesiae, then the Lod building was certainly, as the excavators and others have supposed, a private villa rather than a civic building.

19The owner would have received guests into a great hall from the diverse population of Lod. These guests, who were as easily Jewish as Christian, shared traditions that were not only religious but inflected with African and Dionysiac accents from the paganism of the Roman empire. It would not be overbold to imagine that its aula ecclesiae was just as tasteful and impressive as the reception hall itself.

The editor thanks Dr Chadi Hatoum (UMR 7041 ArScAn) for the abstract’s and keywords’ Arabic translation.

Haut de page


Aupert (P.) & Masson (O.) 1979 « Inscriptions d’Amathonte, I », BCH 103/1, p. 361‑389.

Avni (G.), Bowersock (G.), Gorzalczany (A.), Schwartz (J.) & Talgam (R.) 2015 The Lod Mosaic: a Spectacular Roman Mosaic Floor, New York.

Barnes (T. D.) 2008 « The Date of Ignatius », The Expository Times 120. 3, p. 119‑130.

Kennedy (D. L.) & Falahat (H.) 2008 « Castra Legionis VI Ferratae: a building inscription for the legionary fortress at Udruh near Petra », JRA 21, p. 151‑169.

Lampe (G. W. H.) 1969 A Greek Patristic Lexicon, Oxford, p. 1484.

Sartre-Fauriat (A.) 2001 Des tombeaux et des morts : monuments funéraires, société et culture en Syrie du Sud du ier s. av. J.‑C. au viie s. apr. J.‑C., Beyrouth.

Talgam (R.) 2014 Mosaics of Faith, University Park.

Talgam (R.) 2015 « The Lod Mosaic », Avni et al. 2015, p. 50-109.

Tepper (Y.) & Di Segni (L.) 2006 A Christian Prayer-Hall of the Third Century ce at Kefar ʿOthnay, Jérusalem.

Haut de page


1 The full publication of the Lod mosaic under the auspices of the Israel Antiquities Authority, edited by Gideon Avni, appeared in 2015, with contributions by Amir Gorzalczany, Joshua Schwartz, Rina Talgam, and myself (Avni et al. 2015). An excellent treatment of the Lod finds appears in Talgam 2014, p. 65-70, with color images. For the Kfar ʿOthnay mosaic, see Tepper & Di Segni 2006, together with p. 73-75 of Talgam 2014, Mosaics of Faith, cited above. Leah Di Segni decisively counters Chr. Markschies’ attempt to question the date of the Kfar ʿOthnay mosaic in SEG 61 (2011), no. 1441.

2 The centurion must have come from the nearby VI Ferrata at Caparcotna, which had moved to Udhruḥ by 303-4: Kennedy & Falahat 2008, p. 151-169. This date therefore constitutes a terminus ante for the mosaic and confirms the late 3rd‑cent. date proposed by Y. Tepper and L. Di Segni.

3 These three mosaic inscriptions are reviewed by Denis Feissel in the Bulletin épigraphique 2008. 634, where the last word of the first text, as initially published, is corrected from ἠργάσετ[αι]. The three inscriptions were reprinted in SEG 56, 1900 and 1901, and SEG 58, 1740.

4 Aupert & Masson 1979, p. 367 n. 27; Bulletin épigraphique 1990, 809.

5 Ignatius, Epist. Eph. 18.2; Epist. Rom. 3.3. For recent discussion of the date of Ignatius, see Barnes 2008, p. 119-130.

6 Kennedy & Falahat 2008, p. 151-169. Ca. 303-4 at Udhruḥ: p. 160.

7 Talgam 2015.

8 Greg. Naz., Orat. 5. 2.

9 Lampe 1969, p. 1484.

10 In the New York Review of Books blog for 14 February 2011 ( and more definitively in my introduction to Avni et al. 2015.

11 I am extremely grateful to Gideon Avni for sending me a photograph of this new mosaic. It appeared in the final Lod publication (Avni et al. 2015). The two birds are doves, not pigeons, which were consistently associated with tombs and the dead: Sartre-Fauriat 2001, p. 69‑72.

Haut de page

Table des illustrations

Titre Figure 1.
Légende The Lod mosaic carpet, including the fragment on the northeast side
Crédits © Israel Antiquities Authority.
Fichier image/jpeg, 219k
Titre Figure 2.
Légende The Kfar ʿOthnay floor, with three inscribed mosaics
Crédits © Israel Antiquities Authority.
Fichier image/jpeg, 91k
Titre Figure 3.
Légende The Kfar ʿOthnay mosaic of Akeptous with reference to God Jesus Christ
Crédits © Israel Antiquities Authority.
Fichier image/jpeg, 273k
Titre Figure 4.
Légende The new Lod mosaic with doves
Crédits © Israel Antiquities Authority.
Fichier image/jpeg, 1,0M
Haut de page

Pour citer cet article

Référence papier

Glen W. Bowersock, « Religion in Early Fourth Century Roman Palestine: The Kfar ʿOthnay Mosaics »Syria, 95 | 2018, 65-72.

Référence électronique

Glen W. Bowersock, « Religion in Early Fourth Century Roman Palestine: The Kfar ʿOthnay Mosaics »Syria [En ligne], 95 | 2018, mis en ligne le 01 mai 2021, consulté le 12 mai 2021. URL : ; DOI :

Haut de page

Droits d’auteur

© Presses IFPO

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search