The ankh in early Christian iconography: some reflections in light of recent discoveries at Kellis

Gillian E. Bowen

Monash University, Victoria, Australia



This paper has been prompted by the recent excavation of two early fourth-century churches at Ismant el-Kharab, ancient Kellis in the Dakhleh Oasis, Egypt, both of which featured cruces ansatae as part of their decorative scheme.1 The crux ansata, a Christian adaptation of the Egyptian ankh, has long been attested from funerary contexts that have been tentatively dated to the late fourth and fifth centuries but its use within an ecclesiastical context during the early fourth century is hitherto unknown. The discovery leads me to consider whether the Kellis material, which comes from a securely dated context, might be used to refine the dates of other monuments which employ the crux ansata motif. What follows is description of the Kellis cruces ansatae together with an overview of the evidence for dating the churches. Comparative motifs from the mausolea of Bagawat in the neighbouring Kharga Oasis will be considered in order to determine their possible contemporanity. From there the study will be extended to the Nile valley and will focus upon those painted shrouds from Antinoopolis that bear the crux ansata in an attempt to determine whether the interred were Christian or pagan. In so doing, I shall address the issue of whether the use of the crux ansata/ ankh was peculiar to Christians by the fourth century or whether it continued in pagan funerary iconography.


Excavations carried out thus far at Ismant el-Kharab indicate that Kellis was abandoned by its occupants in the closing years of the fourth century CE. Evidence for this is suggested by the papyrological, numismatic and ceramic data. The last dated text, P. Kell Gr I 26, was written around 389 (Worp 1993, 77-79)2 although a horoscope cast for the year 392 has also been found (Hope 1997, 14). The volume of coinage struck after 365 that has been retrieved from the site is considerably reduced by comparison with that struck prior to the death of Constantius II (361). A mere five specimens post-date the reign of Theodosius I (379-95) and only three specimens of the enormous Salvs Reipvblicae issue, struck between 388-94, have been identified (Bowen 1999, 43-47). A study of Egyptian coin hoards has shown that the Salvs Repvblicae type accounted for in excess of 90% of the coins in circulation in Egypt in the early years of the fifth century (Bowen, in press). There is no evidence to suggest that their absence from Kellis can be attributed to a lag in the circulation of new issues in the oasis. Moreover, the latest ceramics found are entirely at home within a fourth-century context (Hope 2000, personal communication). In light of this cumulative evidence, the churches at the site can be placed firmly within the fourth century.

A Christian community is attested in the village from at least 280; this is determined from an onomastic study based upon dated textual evidence (Bowen, unpublished PhD dissertation). The foundation of the two churches in question has been assigned to the early part of the fourth century based on numismatic evidence. The smaller of the two structures, a small chapel with an adjoining north room, is a domus ekklesiae, that is, a pre-existing building that was converted for Christian worship prior to the formalisation of the liturgy, and the construction of purpose-built churches using a basilica form (Bowen 2000, 33). The chapel has yielded a mere 15 coins and of those six are third-century tetradrachms, a currency that became obsolete following the monetary reforms of Diocletian in 2963 but which appears to have circulated for at least a further 20 years at Kellis (Bowen, unpublished PhD thesis). A coin dating to the reign of Maxentius (306-12) was retrieved from the fill just below the foundation wall of the apse; hence the structure was converted some time after 306. The other church is a large, purpose-built basilica. Once again, an approximate date for its foundation was established primarily upon numismatic evidence.4 The structure yielded in excess if 150 coins but only a third are identifiable. Of these, seven are third-century tetradrachms and several others date early in the reign of Constantine I. Scrutiny of the composition of Egyptian coin hoards deposited during the fourth century indicates that the issues concerned were obsolete in the currency pool by the mid-fourth century (Bowen, in press) and therefore a foundation date for the basilica in the reign of Constantine I is proposed.

The cruces ansatae in the decorative scheme of the churches

In both churches the painted decoration was confined to the apse and the wall of the sanctuary which faced the nave. The paintings in the apse of the chapel are preserved below the cupola; those on the sanctuary’s outer wall are badly damaged although sufficient remains to determine that this comprised imitation columns and geometric designs. Two representations of the crux ansata were incorporated into the decorative scheme of the apse. These formed the centrepiece of two geometric designs set on either side of the entrance, within the body of the apse. The cruces, which are painted red on a yellow background, are small and poorly executed, comprising a circle set atop a cross, the bar of which is longer than the vertical shaft. In both examples, the shaft projects above the cross-bar. The decoration within the basilica is badly preserved and the representations of the cruces ansatae were retrieved from amongst the myriad fragments of decorated plaster which formed part of the debris that fell on and around the bema immediately in front of the apse. A fragment of the capital of an engaged column preserved two small cruces ansatae of the same type as those in the chapel. These once formed the frieze immediately below the top of the capital. The most impressive representation, however, was retrieved in several plaster fragments, not all of which have been recovered. It appears to have been located on the sanctuary wall. The crux is outlined in deep maroon on a white plaster background (Fig 1). It comprises two circles between which is a further circle of maroon dots. The outer circle of the crux is surrounded by another circle of dots but these are small and are set closer together than those of the inner circle. The cross-bar and the shaft were probably of equal length. Maroon dots are painted between the outline of the cross-bar and shaft. The cross-bar has a slight flare at the end but this is not reflected in the shaft. The height is around 28 cm and the diameter of the circle approximately 15 cm. There is no indication that the decoration in the church was replaced and it can be assumed that it was original. 5

Figure 1: The crux ansata from Kellis


The only other representations of cruces ansatae on wall paintings that have been well documented are those in the Christian tomb chapels at the necropolis of Bagawat that served Hibis, the ancient capital of Kharga Oasis. Although the site was visited by Cailliaud, Edmondstone, Hoskins, Wilkinson and Schweinfurth in the nineteenth century, and described by W. de Bock in his monograph published in 1901 (Fakhry 1951, 4),6 excavations did not begin until 1907 under the auspices of the Metropolitan Museum of Art. The expedition concentrated on photographing and classifying the 263 mud-brick chapels; however, many burial chambers were excavated as were numerous Christian graves that had been dug between the chapels (Lithgoe 1908, 203-208; 1909, 122). There is considerable variation in the state of preservation of the chapels, all of which were classified into ten groups and described in detail by Fakhry (1951).7 Fakhry (1951, 9) could not determine the chronology of the chapel types, examples of which are found scattered throughout the necropolis.8 Seven of the chapels have substantial wall paintings on the interior and a further 20 have representations of the crux ansata painted either as the original decoration or dipinti left by visitors. Most of the surviving Christian symbols are on the interior of the chapels but some are represented on the exterior, either in painted form or moulded in mud and painted red. Each one of the painted chapels has at least one representation of the crux ansata motif and is therefore presumed to be Christian. It is worthy of note that no traditional, pharaonic funerary iconography is represented within any of the chapels, although the exterior design of several, including those with cruces ansatae, incorporates cavetto cornices and torus mouldings; one chapel has a winged-sun’s disc sculptured on the lintel.9

The cruces ansatae at Bagawat

The cruces ansatae from Bagawat show a variety of types. A brief description of the shape, elaboration and position within the chapels follows.10

No 30, the Chapel of the Exodus.
This small chapel is in the north of the necropolis. The interior is painted in its entirety; it has geometric designs on the walls but the dome is decorated with biblical narratives. A large crux ansata is painted within each of the pendentives and a further one was placed at the top of each of the arches (Fakhry 1951, 39, 53, 55, pls IV, XVI and XVIII). On the façade depicting of the Promised Land, a single crux ansata surmounts each of the three entrances although the elongated loop of these examples is more reminiscent of the ankh. Greek and Latin crosses were added at a later date (Fakhry 1951, 39) (Fig. 2).

Figure 2: Crosses from No. 30

No 80, The Chapel of Peace
Tomb 80 is one of the most southerly in the necropolis; it is only slightly larger than No. 30. Fakhry (1951, 67) draws attention to the superior quality of its painting compared with that in the other decorated examples, and notes that both the artistic style and the clothing identify it as Byzantine. The chapel incorporates a single crux ansata in its decorative scheme; this, however, is a prominent feature for it is held in the left hand of E I R H N H , symbol of Peace (Fakhry 1951, 73). EIPHNH grasps the crux by the top of the circle in much the same manner the gods of pharaonic Egypt held the ankh. The crux is rendered in outline with the circle contiguous with the cross-bar. The extremities of the cross-bar flare at the top and the vertical widens slightly at the base.

No 25
Tomb 25, together with 23 and 24,
11 are the largest and most important of the mausolea. They occupy the prime site on a high ridge in the north of the necropolis with commanding views of the surrounding region. Several paintings within this chapel are unfinished; others are so poorly preserved that they are unrecognisable (Fakhry 1951, 79). Amongst those preserved is a depiction of the sacrifice of Isaac by Abraham. A prominent feature of this tomb chapel is the large number of elaborate cruces ansatae that are painted above each of the niches and above a scene which may represent Paul and Thekla (Fakhry 1951, 79, 87). A crux ansata, executed in relief, was placed over the niche above the central arch (Fakhry 1951, 79); this may have been part of the original design although it is not impossible that it was a later modification. It does indicate, however, that this, the largest complex in the necropolis, was Christian. The cruces are all rendered in outline with the area between the outlines filled variously with dots and geometric shapes (Figs. 3a , 3b and 3d). In one example the circle comprises six concentric circles of varying thickness and there are solid triangular blocks of colour within the shaft and cross-bar (Fig. 3b). The terminals of the cross-bars flare and the vertical shaft widens slightly at the base. On two examples the circle is separated from the shaft (Figs 3a and 3d). A representation of another form of the cross, a variation of the chi-rho (Fakhry 1951, 37) (Fig. 3c) is also depicted in this chapel but Fakhry gives no indication of its location or whether it was part of the original decoration or a later addition. The chi-rho post-dates Constantine’s victory at Milvian Bridge (312).

Figure 3a: Crux ansata from No. 25


Figures 3b, 3c and 3d: Crosses from No. 25

Nos 172, 173, 175, 210
The small chapels 172, 173 and 175, are just to the north of the necropolis church;
12 No. 210 is ca. 60 m to the east. All of these chapels are decorated with vine motifs that incorporate depictions of the crux ansata (Fakhry 1951, 36-37, 88-99). Fakhry illustrates only two examples, a triple-circle type with a solid centre from No. 172 (Fig. 4) and one from No 210 which is incorporated into the vine. This example is painted red with blue hatching in both the circle and the shaft (Fig. 5). Cruces ansatae in No. 210 are also to be found above a seated figure on the south wall, and above each phoenix that decorates the four pendentives (Fakhry 1951, 97, 99).

Figure 4: Crux ansata from No 172


Figure 5: Vine and crux ansata, No. 210


No 24
The only paintings to be preserved in this imposing tomb are three representations of the crux ansata. The top comprises three circles with a small crux ansata in the centre; small dots are set within the other two circles. The vertical shaft is trapezoidal and the individual arms of the cross-bar are attached directly to the shaft and flair at the terminals (Fakhry 1951, 39, 119) (Fig. 6)

Figure 6: Crux ansata from No. 24

No 49
At the side of each of the light apertures in the chapel are two red cruces ansatae. They vary in height from 36 -38 – 48 cm (Fakhry 1951, 34, 124) (Figs. 7a and b). The shaft of one is squat with the vertical flaring at the base; on the other example that Fakhry has illustrated, the cross-bar flares at the terminals and the vertical widens at the base.

Figures 7a - 7b: Cruces ansata from No. 49

No 53
The decoration comprises a crux ansata on each of the walls (Fakhry 1951, 125). These are neither described nor illustrated by Fakhry.

No 88
The façade of the tomb chapel is coated with white plaster and in the middle of each of the side arches is a crux ansata (Fakhry 1951, 134). This example is neither illustrated nor described by Fakhry.

No 89
Red cruces ansatae are found over all three niches and within each of the pendentives (Fakhry 1951, 137). The design is not described or illustrated by Fakhry.

No 90
This is one of the largest and best preserved of the chapels in the necropolis. On the exterior in five of the six arches is a large crux ansata above which is a small Greek cross (Fakhry 1951, 137). There are no paintings on the interior walls.

No 105
Inside the niche which faces the entrance is a crux ansata; over the niche are two further examples that are painted yellow and decorated in red. The word E
r w n , Love, is written on the wall to the west of the niche (Fakhry 1951, 140). Two examples are illustrated by Fakhry (1951, 38). In one, the outer circle has a series of curves within giving the appearance of segmentation (Fig 8a). The vertical shaft has a criss-cross pattern, and widens at the base; the cross-bar, which widens at the terminals, commences each side of the shaft and is undecorated. The other example (Fig. 8b) has a continuous scallop design on the outer edges of both circles and the cross bar. The vertical shaft has a zig-zag design. The shaft and cross-bar widen at the base and terminals. In both examples the cross-bar is attached to the vertical shaft rather than being executed in a single stroke. These are the only examples of such.

Figures 8a - 8b: Cruces ansatae from No. 105


Traces of a crux ansata can be seen above the niche in this chapel; it has been defaced and graffiti written over it (Fakhrey 1951, 140, Pl. XLII B). There is no illustration or description provided by Fakhry.

No 109
Cruces ansatae
are painted in each of the pendentives; others are to be found on the north and south walls with seven around the niche facing the entrance (Fakhry 1951, 140-41 and 37 Fig. 24). Two illustrated by Fakhry have a small crux ansata in the centre of the inner circle (Figs 9a and 9b). The lower section of one is a simple cross and the other is a cross in outline; neither the cross-bar nor the vertical widen out. The others are crudely executed, simple circle atop a Latin cross.

Figures 9a – 9b: Cruces ansatae from No. 109


No 127
There are scant remains of plaster on the façade of this chapel; the upper part of a large crux ansata is preserved in the middle of the south arch (Fakhry 1951, 144). It is neither described by Fakhry nor illustrated.

No 130
Two large moulded cruces, painted red, are found at the side arches of the façade; neither is illustrated by Fakhry. The remains of paintings were noted on the interior at the north-east corner but the chapel was half filled of wind-blown sand rendering it was impossible to view the structure in its entirety (Fakhry 1951, 144).

No 134
The upper wall facing the entrance is the only section of this chapel to be whitewashed. Three cruces ansatae are painted on this section; one is on the curve of the arch and the other two are at either side of the niche (Fakhry 1951, 145) Again, Fakhry provides no description or illustration.

No 159
The three arches of the façade are decorated with large red cruces ansatae (Fakhry 1951, 151). They are probably moulded but Fakhry neither described nor illustrated them.

No 161
The decorated façade of this chapel is unique at Bagawat for it combines the crux ansata motif with that of the Latin cross (Fakhry 1951, 152-53). The cruces are moulded in mud and painted red; they are set within each of the three arches, the centre one above the entrance whilst the two Latin crosses are placed above each of the side arches. The chapel itself is an illustration of the blending of traditional Egyptian, classical and Christian styles. The façade has a cavetto cornice at the top and above the entrance door are Ionian columns. The circle and the lower section of the cruces are outlined and the upper section of the cross-bar curves slightly in line with the circle. There is no widening of the extremities. This chapel alone indicates that the cavetto cornice does not in itself indicate a pagan burial.

No 192
On the interior, facing the entrance, is a decoration representing a niche with two small ones to either side. In the three niches and over the arches are cruces ansatae (Fakhry 1951, 162). No illustrations or description was given.

No 194
On each corner of the first chamber is a red and yellow crux ansata; in the second there are two on each of the side walls and three on the wall facing the entrance (Fakhry 1951, 164, 35 Fig 37). One type (Fig. 10a) is elaborately decorated. It comprises four concentric circles, the centre is outlined in red, the second circle, again outlined in red, has small dots on its inner edge; the following circle has a series of dots set within at regular intervals and the outer band, which is red, has small circular protrusions on its outer face. The lower section is yellow with small lines painted on the exterior of each face, except the base. There is a slight widening of the cross-bar; the vertical is the same width throughout. The second example illustrated (Fig. 10b) shows a large circle above a small lower section. The circle rests on the vertical shank above the cross-bar and the latter curves to follow the lines of the top section. The width of the cross-bar is the same throughout; the vertical widens slightly at the base.

Figures 10a – 10b: Cruces ansatae from No. 194

No 200
A crux ansata decorates the wall that faces the entrance and one is found at each pendentive (Fakhry 1951, 166, 37). One type (Fig. 11a) appears to be crudely executed, elongated ovals with a line for the vertical shaft and a cross-bar formed by a vine-motif. The other example illustrated (Fig. 11b) is a chi-rho with two dots placed in the horizontal sections. The letters A and W were placed in the extrados above the entrance (Fakhry 1951, 166).

Figures 11a – 11b: Crosses from No. 200


No 201
A large crux ansata painted in red and yellow is found at every pendentive; several other examples are also recorded (Fakhry 1951, 37, 167). Those illustrated include (Fig. 12a) a type with three concentric circles, the inner being a block of red paint, enclosed within a ring of yellow and an outer circle of yellow, outlined in red with red hatching and red dots placed at intervals. The base, which sits directly onto the top, is of yellow, outlined in red and has the same thin red diagonal lines which decorate the outer circle. The cross-bar flares at the terminals and the vertical widens considerably. The second illustration (Fig. 12b) is a variation upon the former. It has three concentric circles, the inner and middle ones in yellow outlined in red. The outer circle is also of yellow but has red hatching. The vertical shaft is as the former but within the cross-bar are two triangles outlined in red.

Figures 12a – 12b: Cruces ansatae from No. 201

No 213
A large crux ansata is modelled over a triangular niche in the west wall; it is not illustrated by Fakhry (1951, 172).

No 230
Three aches form the façade and within each is a crux ansata modelled in mud (Fakhry 1951, 175).

General observations and the proposed date of the Bagawat chapels

Before considering the role of the Kellis crux ansata as an aid to dating the Bagawat tomb chapels, it is pertinent to consider the other chronological indicators for the necropolis. The mud-brick chapels are poorly dated and their period of use has been suggested to range from the fourth to sixth (Fakhry 1951, vii) or seventh centuries (Badawy 1978, 96). Hauser (1932, 50) alone suggested a commencement date in the mid-third century, based upon the contents of a single tomb (see below). On architectural considerations, P. Grossmann (1989, 327) regards a rock-cut tomb with a chapel façade, located to the north of the necropolis where the mausolea are situated, to be amongst the earliest; this he dates to the late third or early fourth century.13 De Bock (1901, 15),14 having studied the necropolis, estimated the interval between the earliest and latest tombs to be about a century. Winlock (1941, 49), who worked with the Metropolitan Museum team, concurred and by considering the numismatic evidence, suggested that the necropolis was appropriated by Christians during the reign of Constantine and was abandoned sometime before the mid-fifth century when Hibis was sacked by the Blemmeys.15 Coins found buried with the dead in the mausolea, and in the ordinary graves, range in date from Constantine I (312-37)16 to Arcadius (395-408); in general, the issues were in mint condition (Hauser 1932, 40) suggesting that the interment took place within a short time of striking. Although the coins serve well as a dating tool for the graves, they provide only a terminus post quem for the chapel tombs as they contained multiple burials.

The date of the decoration of the chapels is speculative. I have assumed that they were painted at the time of the initial interment, although this cannot be verified.17 Attention has focussed upon the two of the best-preserved examples, the so-called Chapel of the Exodus and the Chapel of Peace. These have been dated largely on art historical considerations. Badawy (1978, 243-45) suggested a fourth-century date for both and noted a similarity of themes in Roman catacombs and in the domus ekklesiae at Dura Europos. John Beckwith (1970, 71) placed them well within the fifth century without further comment, whilst Margaret Riddle (1983, 163) admitted that a fifth-century date was probable, but emphasised that further excavation was needed within the necropolis before a more positive date could be assigned. She also noted the need to pay greater attention to early Christian sepulchral art in order to draw a comparison. As no further excavation has been undertaken in the necropolis, and no fourth-century sepulchral art from Egypt has come to light with which to make comparisons, one is left with the elaborate crux ansata from Kellis. As no exact parallel can be drawn, one can do no more than argue that the elaboration of the symbol had certainly begun under Constantine.

The incorporation of the Latin cross on the façade of No.161 may advance the chronological investigation, for this form of the cross is exceptionally rare at Bagawat. The symbol is representative of the crucifixion of Jesus, and therefore of salvation, and one would expect it to be appropriate within a funerary context. It was presumably used by Christians from at least the time of Clement of Alexandria (ca. 190-215) who refers to it as the sign par excellence of the Lord (Stromateis 6). Reproductions naturally increased from the time of Constantine (Eusebius Ecclesiastical History, 9.8) particularly after the announced discovery of the wood from the `true cross’. Not only is the Latin cross rare at Bagawat, it is non-existent at Kellis. Moreover, there is no representation of this form of the cross at the only other well-dated fourth-century church: the small Constantinian church at Shams el-Din, Kharga Oasis. The walls of its kathesterion are covered with graffiti amongst which the crux ansata features prominently (Wagner 1987, Pls VII-XIV). The absence of the Latin cross prompts the question of whether this symbol was deemed inappropriate within churches or cemeteries in fourth- and fifth-century Egypt or whether its use in the chora was not in vogue.

To further the investigation, one can perhaps attempt to determine the date by which the crux ansata was superseded by other forms of the cross. Rufinus, Sozomen and Socrates Scholasticus, writing from around the first to the sixth decade of the fifth century respectively, all refer to the destruction of the Serapeum in Alexandria in 391 by zealous Christians. They are reported to have destroyed the busts of Serapis which adorned the walls, entrances, doorposts and windows of the houses and replaced them with a painted sign of the `Lord’s cross’. That this was a crux ansata rather than another representation of the cross, is suggested by its juxtaposition within the texts to a discussion of the discovery of the hieroglyphic ankh signs interpreted as meaning the `life to come’ by those who remembered how to read the script. The exercise is said to have been instrumental in converting numerous pagans, including priests, to Christianity (Rufinus Ecclesiastical History, XI.29; Sozomen Ecclesiastical History ,VII.c.15; Socrates Ecclesiastical History, V.17). One might assume from these texts that the symbol was still in use in the early fifth century. Can this be verified by archaeology?

Very few fifth and sixth century churches preserve any form of decoration. Where they do I have found that the crux ansata is not included, with a single exception, that at Dendera where a small crux ansata is placed in a palm tree which forms the centrepiece of a sandstone shell (Atalla undated, 26-27). P. Grossmann (1991, 691) dates the church to the mid-sixth century. By contrast, the crux is found in all three, well-preserved, fourth-century churches: those at Kellis and Shams el-Din (mentioned above) and at Dush, ancient Kysis, in Kharga Oasis18 where a painted crux ansata was found by the French mission (Grimal 1994, 399). It can be assumed that the crux ansata also featured in fourth-century monastic decoration; the symbol is attested on the bindings of some gnostic codices from Chenoboskion (Doresse 1960, 24). Moreover, it forms the centrepiece of the lintel of the entrance to the saint’s tomb at the Monastery of Saint Psote that is purported to date to the fourth-century (Atalla n.d., 46), although the evidence for such a date is not stated. It is perhaps telling that the earliest surviving monastic murals, those at Bawit and Saqqara, which have been dated tentatively to the sixth century; have no cruces ansatae in their decoration; other forms of the cross, however, are well represented (Walters 1974, 251-328).19 It is also worthy of note that whilst representations of the ankh were not erased from temples converted for Christian worship, which is thought to have taken place from the fifth century, other forms of the cross, usually the Greek version, were cut deeply into the sandstone blocks alongside the traditional ankh. Such can be observed at Karnak and Philae.

Other representations of the crux ansata are found on grave stelae and in Coptic textiles. The latter may have been specifically produced for a funerary context. Both categories of artefact are too poorly dated to be of assistance.20 Grave stelae from the Coptic Museum, Cairo and the Graeco-Roman Museum, Alexandria, which feature the crux ansata and are assigned a blanket third-fourth century date in the catalogue (Atalla, n.d.36-47),21 would certainly not be out of place stylistically within a fourth-century context. Where these symbols appear with another form of the cross, it is the Greek or the chi-rho, not the Latin form. Regrettably, there are insufficient dated examples to draw any firm conclusions for the general discontinuation of the use of the crux ansata.

Returning to the tombs of Bagawat, at present it is impossible to state whether those in question were built by Christians, other than those which incorporate Christian symbols into the mouldings of the façade, or whether they were pagan tombs appropriated by Christians and decorated accordingly. Hauser (1932, 50) notes that all of the bodies were placed with their heads to the west; most were simply laid upon the ground although some were placed on boards or biers with legs. Only one mausoleum was found to contain coffins; there were three, all decorated in traditional pharaonic style (Hauser 1932, 44-50). Two were in a battered condition and had been modified for re-use; Hauser (1932, 50) suggests that the third had probably also been reused. Again, the bodies were placed with their head to the west. Amongst the few grave goods was a bronze coin of Nero that had been gilded and mounted on a disc with a perforation in the top to form a pendent (Hauser 1932, 46). Hauser (1932, 50) identifies the deceased as pagan, on the basis that Christians would not reuse coffins with pharaonic iconography, nor would they keep a coin of Nero, whom he describes as `the archenemy of the Christians’. On the basis of the coffins discovered in this tomb, Hauser (1932, 50) redated the commencement of the necropolis to the mid-third century. It is of interest to note that in January, 2000, intrusive Christian burials were found in a mausoleum at Kellis, the walls of which were resplendent with pharaonic funerary iconography. The bodies were simply wrapped and laid in the ground in an east-west direction, head to the west I the Christian fashion. Although the deceased were not in coffins, the discovery does illustrate that Christians were not averse to burial within pagan tombs. It is to be hoped that with further excavation of the Roman mud-brick mausolea at Kellis, and perhaps those at Amheida in Dakhleh Oasis,22 it will be possible to refine the chronology of the Bagawat necropolis, and add to our knowledge of the reuse of pagan tombs by Christians.

The Painted Shrouds from Antinoopolis

A second category of funerary artefact that depicts the crux ansata is the painted shrouds from Antinoopolis. This class of artefact, which is classified under the category of mummy portrait, has received considerable attention of late. There are four published examples of shrouds that feature the crux as a major iconographical element; a fifth employs the symbol as a minor role. The designation Christian has been disputed with several authorities reluctant to commit one way or the other because of what Aubert (2000, 148) terms `ambiguous iconography’, that is, the incorporation of pagan elements on some examples. The date of the portraits is also hotly debated. If it can be shown that the shrouds that incorporate the crux ansata were commissioned by Christians and that they date to the third century, this will be the earliest representation of the symbol within a Christian context.

Description of the relevant shrouds

Before a further discussion is entered into, a brief description of the relevant iconography represented on each of the shrouds is given. The shrouds are identified according to Parlasca's catalogue numbers.

Parlasca II 73, No. 41823
A women holding a crux ansata in her left hand; the right hand raised, palm facing. The circle of the crux ansata is wide with a very short vertical shaft above the cross-bar. The cross-bar flares out at the extremities. The vertical shaft is obscured by the woman's hand. The extant shroud terminates just below the waist.

Parlasca II 73, No 420.24
A woman holding a crux ansata in left hand; the right hand is raised in the same manner as that above. The shroud is fragmentary and insufficient of the crux ansata survives to describe it in full; the flair of the cross-bar does not appear to be as pronounced as in the former.

Parlasca II 73, No. 419.25
A woman holding an crux ansata in left hand; the right hand is raised as above. The crux is narrower and more delicate than that of number 418. A longer shaft separates the circular top from the cross-bar than that of number 418. Again, the cross-bar flares out at each end. The entire length of the shroud is preserved. The front of the shroud has geometric motifs; the side panels are decorated with pairs of birds, garlands, and two cruces ansatae which flank the feet. The shape of the latter is identical to that held by the woman. No obvious pagan imagery is represented, however, the incorporation of vignettes is reminiscent of that on traditional pharaonic coffins of the Roman period and, as such, the shroud is of a known funerary type.

Parlasca II 73, No. 424.27
A boy holding a crux ansata and a pigeon or dove in his right hand and a pomegranate in his left. The crux has a thick top with a straight cross-bar. A large winged scarab, the Egyptian symbol of regeneration, adorns the shroud in the region of the boy's waist.

Parlasca II 73, No. 42228
A man holding a goblet against his chest in his right hand and a floral wreath resting against his left shoulder in the left hand. Above the man's right shoulder is a crux ansata with a short vertical shaft above a straight cross-bar. In the corresponding position above the left shoulder is a seated figure which has been identified by Parlasca (1977, 74) as the god Ptah.

Dating the shrouds

Since their discovery in the nineteenth century, mummy portraits have been variously dated from Ptolemaic times through to the closing years of the fourth century CE (Borg 1995, 229-31; 1996, 19-27). The method used to assign a date is art historical, that is, comparison with jewellery, costume and the hairstyles depicted on the paintings with those of Roman sculptures (Borg 1995, 229-30). The first attempt to date the paintings individually on such grounds was undertaken by H. Drerup in 1933 and his results gained wide acceptance (Borg 1995, 231; 1996, 22-25). Drerup's chronology formed the basis of that adopted by K. Parlasca whose catalogues of the portraits form the standard reference works. According to Borg (1995, 231), certain of Drerup's conclusions are `questionable', for the chronology of the Roman portrait sculpture from which the comparisons were drawn was not in itself secure, especially for the third and fourth centuries. She (1995, 231) notes that where the coiffure proved to be uncertain for dating, Drerup discarded his systematic methodology and resorted to what she terms a subjective, stylistic approach. Borg (1995, 232) considers hairstyles to be the main determinant for dating the portraits and that it is imperative that these be analysed separately from stylistic grounds, which she maintains could merely reflect a poor quality work. Borg has herself undertaken a reappraisal of the dates ascribed to the paintings. Following Jucker,29 Borg (1995, 232) acknowledged that a revival of second-century coiffure during the Constantinian period renders it difficult to distinguish fourth-century from second-century portraits and sculptures. Drawing upon recent research which has shown a greater diversity and intricacy of hairstyles for women in the second century than in the fourth, Borg (1995, 232) believes that sufficient variation has been noted to distinguish them chronologically.30 The results of Borg's (1995, 233) research have shown that the earliest mummy portraits can be dated to the reign of Tiberius, 14-37, and the latest to Severan period, 193-236.31 She (1995, 231; 1996, 23) rejects the argument that the portraits continued to be used until the edict of Theodosius, which placed a ban on pagan cults in the closing years of the fourth century. Parlasca (1997, 127) rejects Borg's claim that the portraits ceased before the mid-third century and maintains that the genre continued throughout the fourth century. The chronology is certainly not settled; more emphasis is now being given to the style of dress and jewellery with the result that two schools of thought have emerged (Walker 2000, 34-35). Although the Antinoopolis portraits which bear cruces ansatae have been variously dated, they are regarded as being contemporary (Parlasca 1977, 72-75; Doxiadis32 1995, 46, 117-21). The various dates assigned to the paintings are as follows: Severan period,193-235, (Doxiadis 1995, 117-21), second quarter of the third century (Parlasca 1977, 72-75), second half of the third century (Thompson 1973, 437), 250-300 (Walker and Bierbrier 1997, 160), and fourth century (Aubert 2000, 147-48).

The religious persuasion of the subjects of the shrouds

In considering the possible religious persuasion of the deceased only those belonging to females that have no overt pagan iconography are discussed at this point. The shrouds of the males, both of which incorporate pharaonic symbols, are discussed later.

There appear to be two areas of contention for accepting the three female subjects of the shrouds as Christian. The first is whether the crux ansata held by each of the women here represents the pharaonic ankh, albeit in a modified form, or whether it had, by this period, been adopted exclusively by Christians. The second problem is encountered by those who accept a third-century date for the shrouds. The argument that has been raised in relation to an early date is, either that Christianity had not reached the chora by that time, or that Christians would not overtly display their affiliation to an outlawed religion.

The ankh in traditional funerary iconography

The crux ansata, as depicted on the three shrouds, is the dominant motif. In pursuing the discussion, consideration must be given to whether the ankh enjoyed equal prominence in traditional funerary iconography. An investigation has shown this not to be the case and consequently, the evidence presented here is mostly negative. In the vignettes from the Book of the Dead the ankh is shown only in a subordinate position and in no spell is the deceased shown holding the sign. In Spells 17 and 125 an ankh rests on the knees of squatting deities in an upper register; in Spell 175 the ankh is placed on the knees of the squatting Thoth; in one register of Spell 16 two animated ankh-signs are shown holding ostrich-feather fans; Spell 146 depicts two examples alternating with two ba birds above the portals of the House of Osiris; in Spell 183 Thoth offers an ankh and two was-sceptres to Osiris, whilst in Spell 186 the goddess Opet carries an ankh, and in a vignette of Osiris, Isis holds an ankh in her left hand.33

A further illustration of the insignificance of the ankh within a late pharaonic funerary context may be found in the nature of amulets that were placed within the wrappings of mummies. Indeed, Carol Andrews (1984, 38) expressed surprise at the infrequency of the symbol. Although it is impossible to go through every publication that lists amulets placed within mummy wrappings, a single example will suffice: the mummy of Djedhor, found at Abydos by Petrie (1902, 37-38, Pl. lxxviii), and dated ca. 380 BCE. None of the 87 amulets wrapped between two layers of bandages of the intact mummy was an ankh (el-Mahady 1989, 151).

Evidence for the use of the ankh during Roman Period is similarly sparse. It is absent from the catacombs of Kom el-Shoqafa in Alexandria which combines Hellenistic and Egyptian funerary iconography. It is not found in the decorated, first-century CE tomb of Padiosiris in the Dakhleh Oasis. The ankh is depicted infrequently on the hundreds of funerary stelae from the cemetery of Kom Abu Billou, which served the town of Terenuthis in the Delta, and dates to around the third century CE. Many stelae from this cemetery incorporate elements from the traditional Egyptian funerary repertoire. Anubis is by far the most popular element of Egyptian iconography on the stelae although falcons, embalming scenes, the judgment before Osiris, Ma’at, uraei and offerings are also represented (Abd el-Hafeez, Abd el-Al, Grenier and Wagner 1985, passim). The ankh only appears in the hands of Ma’at as she sits on the balance in scenes representing the judgment before Osiris (Abd el-Hafeez et al 1985, 83 fig. 1).

Where the ankh is incorporated into the decoration of coffins and cartonnage in Roman-period burials, its appears in an ancillary position and is confined to the area of the feet. Examples of such are to be found on the painted wooden coffin of a young woman from Akhmim, dated to 50 BCE-50 CE, where the ankh is shown in front of the feet34 and on the richly decorated coffin of Soter from Thebes, dating to the early second century. Inside the lid of the coffin, which is resplendent with an extensive range of Egyptian funerary iconography, the ankh is placed around the neck of the cow at the foot end.35 A painted shroud, provenance unknown, dated variously to the first or late second century, is decorated entirely on the body field with Egyptian funerary iconography (D'Auria et al 1988, 204, fig. 154).36 The decoration incorporates a winged scarab with two ankh-signs with small circular rather than looped tops, placed on the upper wings. No further ankh–signs appear on the other fifteen entries for the Roman period in the catalogue of the Boston Museum of Fine Arts. The ankh is not represented on the published cartonnage of the mummies from Kom Ombo in the Musée Guimet d'Histoire Naturelle, Lyon,37 nor on the cartonnage found on the burials at Dush in Kharga Oasis or those at Ismant el-Kharab in the Dakhleh Oasis. It is also absent from the cartonnage mummy coverings on display in the Graeco-Roman Museum at Alexandria. Whilst the search is by no means exhaustive, it can be suggested that by the Roman period, the ankh was not used a funerary symbol. On this criterion, I would argue that the women in question were Christian. Should this not be so, the question which needs to be addressed is why the ankh, which was virtually absent within a funerary context, should suddenly be chosen as the dominant iconographic element on the shrouds of these women, albeit in the form of the crux ansata. For them it must have been imbued with profound significance. If it were a pagan symbol why is it incorporated on so few shrouds?

The raised-hand gesture

Before addressing the chronology of the shrouds, it is pertinent to consider the hand gesture adopted by each of the women, all of whom have their right hands raised, palms facing out, with the tips of the fingers placed at breast height. Aubert (2000, 148) recognises a parallel in the gesture of Isis to Osiris, a gesture which she identifies as representing protection and veneration. Doxiadis (1995, 46), in interpreting the gesture as Christian, draws upon a twelfth-century wall painting. A far earlier Christian parallel, however, is to be found in the figure of Noah at Bagawat (Fakhry 1951, 77, Fig. 69). In fact, many of the representations of biblical characters and personifications in the Chapel of Peace have both hands raised in a similar gesture.38 The Bagawat parallels show clearly that the raised-hand gesture was adopted by Christians of the fourth and/or fifth centuries.

The male portraits

Both shrouds of the males incorporate pagan elements as well as the crux ansata. This introduces the possibility that the symbol was used by non-Christians. Neither the pigeon nor the pomegranate are out of place within a Christian context, however. The pigeon or dove was popular in Christian iconography and could represent the soul, the Holy Spirit or the Church itself (Murray 1996, 144) whilst the pomegranate may symbolise eternal life (Doxiadis 1995, 46). The winged sun’s disk is problematic as is the Ptah figure. One can argue either that the deceased represented were pagan who happened to incorporate a Christian symbol in their shroud or the opposite, that they were Christians who were reluctant to abandon some of the traditional customs. The latter should not be surprising, especially if one accepts an early date for the shrouds. The simple fact that mummification continued to be practised is testimony to that reluctance.39 The combination of architectural features in the necropolis of Bagawat is further testimony to the blending of the two traditions.

The complications of assigning an early date

With authorities at odds in designating a date to the Antinoopolis shrouds, one can do no more than reflect on the possibilities, or complications, that each suggested date presents. Whilst to us it may seem foolhardy to overtly express one's Christian beliefs during the early Severan period because of the persecution which took place in 201/2 (Eusebius H.E. VI.1.I), the de facto tolerance of Christianity thereafter, until 250, would present few difficulties. Indeed, if Eusebius (H.E. VI.21 3-4) is to be relied upon, Alexander Severus adopted a pro-Christian stance and his mother, Julia Mamaea, was deeply influenced by Origen’s teaching. With the exception of the period 253-62, Christians were free to practice their religion without undue harassment until the persecutions under Diocletian. Moreover, papyrological documentation attests a Christian presence in Antinoopolis by the third, if not the second century.40 If the shrouds can be shown to date to the first half of the third century, as Borg and Parlasca argue, the use of the crux ansata by Christians can be pushed back by at least a century.


Whilst the exercise was inconclusive in refining the date for the period of use of the necropolis of Bagawat, it has certainly shown that the crux ansata unlike its pharaonic model, the ankh, played a major role in early Christian funerary iconography. It would not be too presumptuous to say that where the crux ansata is shown dominant, the chances of it being from a Christian context are high.



  1. The site of Ismant el-Kharab falls within the concession granted to the Dakhleh Oasis Project, the director of which is Anthony J. Mills. Excavations at the site are conducted under the direction of Colin A. Hope, Monash University, Melbourne, Australia. I am most grateful to Dr Hope for entrusting me with the excavation and publication of all three of the Kellis churches.
  2. The text is dated by reference to the praeses of the Thebaid, Septimius Eutropius who is known to have been in office in December 389 (Worp 1995, 78).
  3. Dr William Metcalf of the American Numismatic Society argues convincingly that the tetradrachms were recalled soon after the reform. From his sample of 15,000 specimens, he notes that the wear becomes progressively greater on issues struck prior to Year 12 of Diocletian; the Year 12 issues are all in mint condition (personal communication, 1998).
  4. Certain architectural features within the basilica also indicate a date prior to the formalisation of the liturgy.
  5. The only other representation of the crux ansata to be found within the churches is a graffito in the outer room of the basilica.
  6. Materiaux pour servir à l’archéologie de l’Égypte Chrétienne, St Petersburg.
  7. Fakhry (1951, 9) numbered the tombs according to location working generally from north to south then south to north.
  8. P. Grossmann (1989, 414-38) has attempted to place some of the tomb types in a relative sequence.
  9. Tomb 7. Fakhry (1951, 103) notes that this is not amongst the oldest type of chapel in the necropolis but that the owners were clearly followers of the ancient religion.
  10. The first seven chapels listed are those that have elaborate designs; the remainder is arranged according to Fakhry’s (1951) catalogue numbers.
  11. P. Grossmann (1989, 434) regards Chapel 24 as a small church.
  12. P. Grossmann (1989, 433) identified the structure (No. 180) as a place where memorial meals for the dead were consumed, rather than a church..
  13. P. Grossmann (1989) refrained from assigning a date for the latest tombs. He simply reiterates Winlock’s suggested date of abandonment, ca. 450.
  14. Cited by Winlock (1941, 49).
  15. The incident is recorded by Evagrius, Ecclesiastical History I, 7.
  16. Many of Constantine’s coins stuck before his conquest of the east circulated in Egypt.
  17. In some of the chapels the decoration is incomplete. One can only speculate upon the reason and the time that the painter ceased his work.
  18. The excavators date to the end of the fourth century on numismatic evidence (Grimal 1994, 399).
  19. The cross is usually held by representations of saints. There are no representations of saints, or of Christ at Bagawat, although the Annunciation to Mary is depicted in the Chapel of Peace.
  20. An elaborate fragment of textile which incorporates the chi-rho monograph inside the circles of two cruces ansatae that are set within arches (inventory number 2023), is assigned fourth to fifth century (Atalla n.d., 36-42). The design within the crux bears similarities with some examples from Bagawat.
  21. The inventory numbers of those stelae in the Coptic Museum are:4302, 8016, 8578, 8682, 8552, 8015, 8566, 8585, 8612, 7730, 8556; those from the Graeco-Roman Museum are: 11852 and 23565.
  22. The excavation of Amheida, ancient Trimithis, will be undertaken by Professor Roger Bagnall and Dr Lynn Meskell of Columbia University under the auspices of the Dakhleh Oasis Project.
  23. Athens, Museo Benaki, Inv. 6877.
  24. Paris, Musée du Louvre, Département des Antiquités Egyptiennes, Inv. AF 6487.
  25. Paris, Musée du Louvre, Département des Antiquités Egyptiennes, Inv. AF 6440.
  26. The latter is noted by M-F Aubert (2000, 147-48) who comments on this shroud as illustrative of the interface between paganism and Christianity.
  27. Paris, Musée du Louvre, Département des Antiquités Egyptiennes, Inv. AF 6488. I am indebted to C. Ziegler, Conservateur général, Chargé de Département des Antiquitiés Egyptiennes, the Louvre, for providing me with a colour scan of this shroud.
  28. Paris, Louvre, Dep. Gr. Et. R. Inv. P.215.
  29. In a review of K. Parlasca, Ritratti de Mummie, III, in Gnomon 56 1984, 554ff (not viewed), Jucker cast doubt upon Parlasca's identification of certain of the portraits assigned to the reign of Constantine I based upon insecure dates of his parallels (Borg 1995, 232).
  30. Borg (1995, 232-33) notes that the women's hairstyles are complex and change rapidly making it possible to give a fairly accurate date. The similarities they bear to those of Roman sculpture in relation to the smallest detail is indicative of their derivation from Roman fashions. She notes that the closest parallels to the portraits are those sculptures which come from other Roman provinces in the east which similarly reflect Roman high fashion. Borg (1995, 233) claims to know of no examples in sculpture where the hairstyle continues longer than fashion dictated. Whilst acknowledging that provincial Egypt could be influenced by indigenous styles, she states that very few of the mummy portraits do not have parallels in Roman sculpture. The model for Borg's revised dating is as follows:
    1. There was a broader range in women's hairstyles in the second century than in the fourth century;
    2. fourth-century styles never copied second-century coiffure exactly;
    3. none of the hairstyles shown in the portraits belong exclusively to latter third or fourth century;
    4. stylistic qualities attributed to a later date could be poor paintings.
  31. Borg (1995, 233) dates the latest men's hairstyle to the first decade of the third century and the latest women's to ca. 235.
  32. Doxiadis uses Borg's dates throughout.
  33. The vignettes and their accompanying captions are taken from R. O. Faulkner, Book of the Dead, Guild Publishing, London, 1972. All of the papyri are from the British Museum and their registration numbers followed by the page number on which the illustrations appear in Faulkner are as follows: Spell 16, 10479/11 (p.42); Spell 17, 9901/8 (p.48); Spell 125, 9901/3 (p.34); Spell 146, 10470/12 (p.138); Spell 175, 10470/29 (p.174); Spell 183, 9901/2 (p.178); Spell 186, 10470/37 (p.187), Spell 189, 10470/37 (p.187) and the Osiris vignette, 10470/30 (p.26).
  34. British Museum EA 29587; for a description and illustration see Walker and Bierbrier (1997, 35 Cat. No. 9).
  35. British Museum EA 6705; described and illustrated by Walker and Bierbrier (1997, 149-50).
  36. Boston Museum of Fine Arts No. 54,993.
  37. See A. Schweitzer, "Les Parures de Cartonnage des Momies de Kom Ombo au musée Guimet d'Histoire naturelle de Lyon", La Revue du Louvre et des Musées de France 3, 1992, 18-27.
  38. This is similar to the orant pose.
  39. The bodies of three Coptic monks buried below the walls of St Mark's Monastery, Deir el-Medineh were found to have undergone mummification. No attempt of evisceration had been made but the bodies were covered with natron and other preservatives. See E. Prominska, "Ancient Egyptian Traditions of Artificial mummification in the Christian Period in Egypt", in A. Rosalie David (ed.) Science in Egyptology, Manchester, 1986, 113-121.
  40. P. Ant. I 7, I 8, I 9, II 54 and III 112.



Abd el-Hafeez, Abd el-Al, J-C. Grenier and G. Wagner, Stèles funéaires de Kom Abu Bellou, Paris.

Andrews, C., 1984 Egyptian Mummies, British Museum Press, London.

Atalla, N. S., n.d. Coptic Art Lehnert & Landrock, Cairo.

Aubert, M.-F. A.,2000 Portraits of the Later Third Century AD From Deir el-Bahri and Antinoopolis, in S. Walker (ed.), Ancient Faces: Mummy Portraits from Roman Egypt, 145-48.

Badawy, A., 1978 Coptic Art and Archaeology, Massachusetts Institute of Technology, Massachusetts.

Beckwith, J., 1979 Early Christian and Byzantine Art, Penguin Books, Hammondsworth.

Borg, B., 1995 Problems in the Dating of Mummy Portraits, in E. Doxiadis, The Mysterious Fayum Portraits, Thames and Hudson, London.

Borg, B., 1996 Mumienportrats,Mainz on Rhine.

Bourguet, S. J., 1965 Early Christian Painting, trans. S. W. Taylor, Weidenfeld & Nicolson, London.

Bowen, G. E., in press The Churches, in C. A. Hope and G. E. Bowen (eds) Dakhleh Oasis Project: Preliminary Reports on the 1995-1999 Field Seasons, Oxbow Books, Oxford

Bowen, G. E., 2000 The Small Church at Ismant el-Kharab, Bulletin of the Australian Centre for Egyptology, 11, 29-34.

Bowen, G. E., 1999, The Coinage, in C. A. Hope and A. J. Mills (eds), Dakhleh Oasis Project:Preliminary Reports on the 1992-1993 and 1993-1994 Field Seasons, Oxbow Books, Oxford, 43-47.

Bowen, G. E., 1998 The Spread of Christianity in Egypt in Light of Recent Discoveries from Ancient Kellis, unpublished PhD dissertation, Monash University, Melbourne.

Doresse, J., 1960 Des Hiéroglyphics à La Croix, Nederlands Historisch-Archaeologisch Institut Te Istanbul, Istanbul.

Doxiadis, E., 1995 The Mysterious Fayum Portraits, Thames and Hudson, London.

D’Auria, S. P. Lacovara, C.H. Roehrig, 1988 Mummies & Magic, Museum of Fine Arts, Boston.

Fakhry, A., 1951 The Necropolis of el-Bagawat in Kharga Oasis, Government Press, Cairo.

Grimal, N., 1994 Travaux de l’Institut Français d’Archéologie Orientale en 1993-1994, Bulletin de l’Institut Français d’Archéologie Oriental 94, 383-400.

Grossmann, P., 1990 Early Christian Architecture in the Nile Valley, in H. Hondelink (ed.), Coptic Art and Culture, Shouhdy Publishing House, Cairo, 3-18.

Grossmann, P., 1989 Some Observations on the Late Roman Necropolis of al-Bagawat, in Al-Sahraa al-masriya, Gabaana al Bagawaat fi al-waaha al-khaariga (transliteration of the Arabic), Egyptian Antiquities Organization, Cairo, 414-38.

Hope, C. A., 1997 The Find Context of the Kellis Agricultural Account Book in R. S. Bagnall (ed.), The Kellis Agricultural Account Book, Oxbow Books, Oxford, 5-14.

Gardiner, A., 1950, The Baptism of Pharaoh, Journal of Egyptian Archaeology 36, 3-12.

Lithgoe A.M., 1908 The Oasis of Kharga, Bulletin of the Metropolitan Museum of Art ,Vol III, Number 11, 203-208.

Lithgoe, A.M., 1909 The Egyptian Expedition, Bulletin of the Metropolitan Museum of Art ,Vol IV, Number 7, 119-123.

Murray, P. and L. Murray, 1998 The Oxford Companion to Christian Art and Architecture,Oxford University Press, Oxford.

Parlasca, K., 1977 Reperatorio D’Arte Dell’Egitto Greco-Romano, Vol II, Rome.

Parlasca, K., 1997 Mummy Portraits: Old and New Problems, in M.L. Bierbrier (ed.), Portraits and Masks, British Museum Press, London.

Petrie, W. M. Flinders, 1902 Abydos, Part I, The Egypt Exploration Fund, London.

Riddle, M., 1983 Frescoes in Dakhleh and Kharga Oases, in G. H. R. Horsley (ed.), New Documents Illustrating Early Christianity, 3, Macquarie University, North Ryde, 162-64.

van Moorsel, P., 1990 Forerunners of the Lord, in H. Hondelink (ed.), Coptic Art and Culture, Shouhdy Publishing House, Cairo, 19-29.

Wagner, G., 1987 Les Oasis D’Egypte,L’Institut Français d’Archéologie Orientale du Caire, Cairo.

Walker S. and M. Bierbrier, 1997 Ancient Faces: Mummy Portraits from Roman Egypt, British Museum Press, London.

Walker, S., 2000. A Note on the Dating of the Portraits, in Ancient Faces: Mummy Portraits from Roman Egypt, S. Walker (ed), British Museum Press, London, 34-36.

Walters, C.C., 1974 Monastic Archaeology in Egypt, Aris and Phillips, Warminster.

Winlock, H. E., 1941 The Temple of Hibis in el-Khargeh Oasis, The Metropolitan Museum of Art, New York.

Worp, K. A., 1995 Greek Papyri from Kellis: I, Oxbow Books, Oxford.

Return to 18 December programme

Return to Encounters with Ancient Egypt homepage