The analysis of the religious and social significance of baptismal initiation in Christian Africa clearly cannot be based on surviving literary works alone. The archaeological remains of baptismal spaces (fonts and the spaces connected to them) need to be studied both apart from and in relation to the available documents. This paper begins by classifying and describing a representative sample of these structures (as well as their decoration), and drawing some preliminary conclusions about their particularity, mostly by comparing them to structures in other parts of the Christian world (Italy and Gaul, in particular). The second part of this paper identifies the possibilities of and problems with trying to reconstruct the baptismal rite in these spaces, and offers some initial speculation about the theological significance, social construction, sacramental efficacy, and overall symbolism of the baptismal rite based upon the analysis of the extant archaeological remains.

However, prior to beginning the work of description and classification, we need to acknowledge some serious methodological problems. Most of the extant examples of North African baptismal architecture date to the fifth, sixth, and seventh centuries (a period when we have few contemporary writings on the subject of baptism, either theological or liturgical, for purposes of comparison). The limited chronology of these examples raises two particularly vexing issues. The first is whether we can presume that earlier baptismal structures were in any way similar to these later models, especially considering the history of invasion and occupation by indigenous African, then Vandal, and finally Byzantine Christians. If North African baptistery architecture developed in a manner similar to that in other parts of the world, we should assume no great degree of continuity in either style of building, the interior arrangements of space or its decoration between the earliest construction (now lost) and the later examples which we may study.

The second problem is in some ways parallel to the first; whether we can assume any conformity of liturgical rites, not only across the chronological span from the time of Tertullian (early third century) through the centuries well after Augustine. In addition to the simple passage of time, we are also dealing with several different Christian communities who lived in the same geographical area over these centuries, and perhaps even successively occupied particular church buildings (Donatist, Catholic, Vandal Arian, and Byzantine Christians). Thus, on one hand we are lacking relevant material evidence, especially from the earlier period (third and fourth centuries), while on another, we are lacking descriptive liturgical and theological texts from both the Catholic and sectarian sources, specifically from the later period (fifth and sixth centuries).

Related to all of this, of course, is a third matter -- the lack of any compelling reason either to assert or deny important liturgical or even social distinctions between baptismal rites (or spaces) between these different groups of Christians. The theological and polemical issues are well-documented, but give us little to go on if, for instance, we try to distinguish an actual Donatist font from a Catholic, Arian, or Byzantine one. Similarly, based on existing literary evidence we have little or no reason to assume significant liturgical distinctions among these different groups. The power of tradition, mutual influence, and even attractive innovation, may have softened or blurred the particularities of architectural spaces even more than the liturgies celebrated within them.
 

Part I. Classification and description of extant baptismal spaces in Christian Africa, based on representative examples and compared with similar spaces in other geographical regions. The distinctions among the baptisteries can be made according to several criteria:
 

A. Free standing (detached) vs. adjoining baptistery spaces.
 

The attachment or integration of the baptistery space into the plan of the adjunct basilica is the most common pattern in North Africa. Few completely free-standing baptisteries are found in Africa, compared to Italy, Gaul, or even Syria (e.g. the Lateran baptistery in Rome, both the orthodox and Arian baptisteries in Ravenna, Sta. Thecla in Milan, or the fifth-century baptistery at the pilgrimage site of Qalat Siman). The baptistery of Jucundus, attached to the fourth-century "Church of Bellator" in ancient Sufetula (Sbeitla --f igs. 1-2), is one of the rare exceptions. Generally, North African baptisteries are separate spaces which are nevertheless attached to or adjoin the main church hall, either next to the apse or just to the side of the rear or west entrance (cf. Naples' S. Giovanni in Fonte, or the baptistery attached to the basilica of the Holy Sepulchre in Jerusalem). These rooms are accessible to the nave by means of a communicating door or doors. An outstanding exception is the baptistery in Bulla Regis, which was placed in the rear center of the nave itself (figs. 2-3).

The particular architectural relationship between baptismal spaces and main basilica naves seem to have three basic patterns:
 

1. On the left (or north) side of the apse (e.g. Kélibia, Mustis, Timgad, Belalis

Minor);

2. Directly behind the apse (Mahktar -- Hildegun's church, Sbeitla - Baptistery of Vitalis, Henchir Soukrine);

3. At the rear and to the left (north) of the nave (Thurburbo Maius, Carthage - Dermech I); or the right (Timgad);
 

In addition, two exceptions should be noted:
 

4. In the midst of the nave, to the rear (Bulla Regia);

5. Off to the right in the cella of a former temple: Sbeitla III.
 
 
 
 

B. The number and type of adjoining rooms.
 

The baptistery proper was the small room (or detached building) which held the font, which was ordinarily placed in its center. However many, if not most, attached baptisteries were also connected to several adjoining rooms as well as the main hall. These adjoining rooms may have had functions related to the initiation rite proper (renunciation of Satan, anointing, disrobing, vesting, or consignation) or to the preparation of catechumens (training, scrutinies, or exorcism). In some cases these contiguous or annexed rooms had apses as well as connecting doors leading either to the nave or to the room containing the font (e.g. Mahktar, Sbeitla -- Jucundus and Sbeitla III, Carthage -- Dermech I, and Hippo).

The existence of these adjoining rooms distinguishes North African baptisteries from most of the baptisteries found elsewhere in the West (Italy and Gaul), which typically have only separate entrance areas or alternating shallow niches and apses radiating from an ambulatory around the perimeter of an open and centralized room (e.g. Ravenna's Orthodox Baptistery, Milan's Baptistery of Sta. Thecla, Albenga's baptistery, or S. Giovanni in Fonte in Naples). Those non-African baptisteries which have adjoining rooms (e.g. Aquileia, Poitiers, Gerasa, or Qalat Siman), seem either to be parts of larger cathedral complexes, or designed to manage the movements of pilgrims, and do not appear to be directly connected to aspects of the baptismal liturgy proper. However, some scholars have interpret these adjoining rooms in North African as evidence of a more elaborate or distinctively patterned baptismal liturgy or of a particular emphasis on the sacrality and efficacy of the rite itself, or of having something to do with the African baptismal controversy between Catholic and schismatic bishops.[1] Such an interpretation has little to support it from the ancient documents.
 

C. The shapes of the fonts.
 

The fonts in these baptisteries were usually centrally placed in a small room are especially notable for the wide variety of font shapes that has been discovered (some of them unique). Some scholars have suggested that certain shapes are characteristic of particular chronological periods, and that font shapes generally progress from square or rectangular to round, hexagonal, cruciform, and finally to poly-lobed.[2] Since no extant examples of fonts in North Africa pre-date the Vandal era (fifth century), scholars have strained to argue that the Vandal-era fonts are typically quadrilateral while the later Byzantine fonts are poly-lobed.[[3] Notable exceptions which refute this scheme are the hexagonal and polylobed fonts which probably pre-date the Vandal eras (e.g. Sbeitla III and Jucundus, respectively). A reliable chronological development of font style is hard to determine, at least for the North African baptisteries.

Examples of the different font shapes include:

1. Square (Mahktar -- late Vandal era --f ig. 4);

2. Round (Djilma, Tebéssa and Mustis -- fig. 5);

3. Hexagonal (Damous El-Karita, Bir Ftouha?, and the "Supermarché baptistery, Tabarka, Timgad -- fig. 6);

4. Octagonal (Bir Bou Rekba) ;

5. Mixed round and hexagonal (Dermech I -- figs. 7-8);

6. Cruciform (Thurburbo Maius --f ig. 9, Bulla Regia, Henchir Soukrine -- fig. 10, the font from El Kantara on Jerba -- now in the Bardo -- fig. 11, Oued Ramel);

7. Poly-lobed -- trefoil, quatrefoil, or rosette (Kélibia -- figs 12-13, Bekalta -- fig. 14, Belalis Minor --f ig. 15, Henchir Soukrine);

8. Mixed Round and Rosette (Tébessa, Sbeitla III -- fig. 16, Sidi Mansour, Upenna);

9. Labial shaped (Sbeitla -- Vitalis -- fig. 17-18 and Jucundus).
 

All these fonts have access in and out by means of steps. Most of them have three steps to down into the deepest part of the basin and out again (e.g. Kélibia, Timgad, El Kantara, Mustis, Thurburbo Maius, Sbeitla III, La Skhira, Dermech I, Tébessa, and Henchir Soukrine?) but some have four (e.g. Sbeitla -- Vitalis and Jucundus, and Bekalta) or even five (Bulla Regia). Often the steps are in two flights (e.g. Sbeitla -- Vitalis, Henchir Soukrine), but some have steps on three or four sides (e.g. Carthage -- Dermech I, Thurburbo Maius), or even steps that seem to form seats or benches around the interior of the font (Kélibia, Timgad, Mustis, and Bekalta). In one case (Bulla Regia), two flights of steps appear to be used by the candidates for descent into one end and ascent out of the other end of the long, enclosed, arm of the cross-shaped font, while the two other flights of stairs are the only access to the base of a deep, short basin (possibly left dry as a place for clergy or assistants to stand).
 

D. Decoration of the fonts.
 

The baptismal fonts usually were constructed of marble or of other fine stone and set into pavements covered with the mosaic patterns typical of the region (e.g. Dermech I, Sbeitla -- Jucundus, Bir Ftouha, Henchir Soukrine, and Bulla Regia). In some cases the fonts themselves were covered with decorative mosaic, including the exceptionally beautiful and well-preserved fonts from Timgad, Sbeitla, Bekalta, Oued Ramel, and Kélibia (as well as some other, unpublished fonts, e.g. one just outside of Sbeitla, on the road to Kairouan). Although richly decorated baptisteries are well-known elsewhere in the West (e.g. Dura Europos, Naples, Ravenna, Albenga), the decoration of the fonts themselves appears to be limited to the baptisteries of North Africa.

Both pavements and fonts were decorated with geometric patterns, but also often with figurative images, including some iconographic themes common to baptisteries in Naples and Ravenna. However, while we don't know how walls and ceilings of African baptisteries were decorated, the available remains of Italian baptisteries also show certain, consistent narrative images on their walls and in their domes. Naples' S. Giovanni in Fonte had a programmatic decoration which paralleled that of the baptistery at Dura Europos and included images of Christ stilling the storm and walking on the water, the empty tomb, the woman at the well, and the miracle at Cana . The two Ravenna baptisteries both contained dome medallions showing the baptism of Jesus surrounded by a procession of the apostles bearing jeweled crowns. Extant African baptisteries have neither human figures nor narrative images. The baptisteries from Kélibia and Oued Ramel did have a descending doves in their the fonts.[4]

Common to both the African and Italian baptisteries, however, were the more symbolic or simply decorative representations of birds (chickens, quail, pelicans), vases, flowers, fruit, animals, fish, dolphins, and trees (palms, fig, and apples, and olives), all augmented by a variety of border patterns, repeated crowns or scallop shells. The pavements of the fonts at Bir Ftouha, Oued Ramel, and Henchir Messaouda shows harts drinking from streams of living water (Ps. 42), a theme also found in both the Ravenna and Naples baptisteries (and in other floor pavements from North Africa, e.g. Chigarnia). In addition, the baptismal fonts in Sbeitla and Kélibia have the crosses or christogram supplemented by the alpha and omega on their inner sides, with a chi-rho with the alpha and omega in the center of the basin (cf. the baptistery in Naples, also the unpublished baptistery just outside of Sbeitla). Slightly different, the font from Bekalta has a cross from which hang an alpha and omega in its center. The font at Henchir Soukrine includes Latin crosses, fish, and birds within its mostly geometric pattern.[5]

Although the African and Italian fonts share some decorative themes, there are significant differences in their general presentation, richness of composition, and the materials and colors that were used. The fifth-century mosaics in the Naples and Albenga baptisteries, like their sixth-century successors in Ravenna, have an opulence and formality quite different from the geometric patterns of the font at Timgad, and the garden-like beauty (and even relative simplicity!) of the fonts from Kélibia and Bekalta. The Italian examples repeat certain luxurious motifs, including draperies, jeweled wreaths, thrones, peacocks, rich architectural details, star-studded skies, and sumptuously-dressed apostles or martyrs. The colors are dark and rich (reds, greens, and deep blues), and gold tesserae are abundantly used. Wall, floor, and ceiling areas not covered by mosaic (either glass or stone) tend to be covered by colored marble or sculptured stucco. By contrast the designs and colors of the African fonts are lighter, simpler, and less deliberately ostentatious. The feeling is entirely different.
 

E. Inscriptions.
 

Several of the African baptisteries contain inscriptions which offer some insight into both the social location as well as the theological significance of baptismal architecture in the region. Some of these inscriptions are merely dedicatory, for example the baptistery connected to the fourth-century "Basilica of Bellator" in Sbeitla with its dedicatory inscription to the martyr-bishop Jucundus (whose relics may have been deposited in the font after its use as a font was discontinued). This inscription, "HIC INVENTA EST D[E]P[OSITIO] S[AN]C[T]I IUCUNDI EP[I]SC[OPI] PER INQUISITIONE[M] AMACI EP[I]SP[OPI," was found in the aisle of the baptistery structure and probably dates from the sixth century, commemorating the discovery of the martyr's remains by his successor bishop, Amacius.

The fifth or early sixth-century Vitalis baptistery in Sbeitla which was modeled on the earlier (Jucundus) baptistery, contains an inscription which only refers to the votive dedication of the individual(s) who likely paid for the structure: "VITALIS / ET CARDE / LA VOTU / MSN."

However, some of the inscriptions in the baptisteries are more theologically interesting. The bapistery of Henchir Soukrine shows the two-word remnants of a longer inscription: FONTES and BAPTISMA. The font at Djemila, was inscribed with an allusion to the significance of the rite: "[ACCEDI]TE AD D[E]U[M] ET ILLUMINAMINI."[6] The seventh-century baptistery from the basilica of Bekalta is surrounded by a text which has been reconstructed to read: "[GLORIA IN EXCELSIS DEO] / ET IN TERRA PAX (H)OMINIBUS BONE / BOLUM [TATIS L]AUDAMUS T[E]." This reference to the gospel of Luke's infancy narrative suggests an association of baptism with incarnation and birth, perhaps even the possibility that Christmas or Epiphany were considered appropriate occasions for baptism.

Like the Bekalta font, the font found at Kélibia was also surrounded by text which was incorporated into the mosaic pavement. This inscription, broken into four distinct lines corresponding to the four different "lobes" of the font itself was essentially dedicatory but had a few interesting features: "S[AN]C[T]O BEATISSIMO CYPRIANO EPISCOPO ANTISTE / CUM S[AN]C[T]O ADELFIO PRESBITERO HUIUSCE UNITATIS / AQUINIUS ET IULIANA EIUS CUM VILLA ET DEOGRATIAS PROLIBUS / TESELLU[M] AEQUORI PERENNI POSUERUNT." At the edge of the bordering mosaic (probably the entrance to the font) we see a more simple inscription: "PAX, FIDES, and CARITAS."

The longer inscription, around the perimeter of the font refers to two clerics, one (Cyprian) a "high priest" (if antiste = a contraction of antistite?) in addition to an "episcopus." Certainly this must refer to the blessed St. Cyprian, martyred in the mid-third century. The other, a mere presbyter named Adelfius who seemed to be united to him in some particular way. Perhaps this person was Adelfius de Thasvalte, who is recorded as being present with Cyprian at the council of Carthage in 256, or he may be merely the namesake of that person and a priest attached to the Catholic church in Kélibia in the sixth century.[7] One cannot simply assume that this text makes particular reference to the Donatist schism and re-emphasizes the role of the bishop as well as the "unity" of this Kélibia congregation with the Catholic cause, but it nonetheless seems a very real possibility. Added to this is the exchange of "pax" for "spes" in the familiar formula of "faith, hope, and love" (I Cor. 13.13), which might be an additional layer of emphasis on the Cyprianic "peace of the church," referring to the Catholic unity under one bishop and in one communion.
 

Part II. Reconstruction of the rite in the space and analysis of the symbolism, theological significance, social construction of the rite, and matters pertaining to sacramental efficacy.
 

A. Reconstruction of the rite in space and architectural symbolism.
 

The North African textual sources have no parallel to the kind of articulate and detailed catechetical documents produced by Ambrose, John Chrysostom, Cyril of Jerusalem, or Theodore of Mopsuestia. Therefore the basic outlines as well as the details of the baptismal rite in North Africa must be pieced together and hypothetically constructed from a number of different sources, including Tertullian's treatise On Baptism, various letters of Cyprian which address particular problems or issues (especially in regard to the baptismal controversy), Optatus of Milevis, and a number of Augustine's homilies (particularly those addressed to catechumens and neophytes).[8] In addition to the style of these writings, which seem to be more intent on the interpretation of the rite as symbol or its prefigurement in the Hebrew scriptures, the lack of clear, descriptive information also might be due to an effort to protect the sanctity and secrecy of the rite (disciplina arcani). We have some evidence that even Augustine (in the fifth century) was reticent about saying too much in public or in the presence of the unbaptized.[9]

Even so, from the brief glimpses and indirect references that we can harvest out of these diverse types of literary documents, we can discern the broad outlines of a rite that was very similar to that of other western churches. The enrollment of catechumens usually happened at the beginning of Lent in preparation for baptism during the Easter vigil. Catechumens who were being prepared for immanent baptism received the new title "illuminati" through a rite that included the imposition of hands, making of the sign of the cross, and the ingestion of salt. Sponsors vouched for their moral worthiness, reliability, and general good conduct. Throughout Lent, those enrolled were regularly instructed, asked to refrain from the baths and from sexual intercourse, fast, keep long sleepless vigils, and remain kneeling in church as if they were penitents. They were also subjected to regular exorcisms and scrutinies in the presence of the congregation. At the beginning of Holy Week, those who were still in the process underwent a last exorcism, physical examination, scrutiny, and standing barefoot on an animal skin, each individually renounced Satan and the powers of evil.

When the day for baptism came, the water in the font was sanctified, candidates were stripped, led to the font, and immersed three times in the pool in accordance with the Trinitarian interrogation. Bishops would ordinarily have officiated at any baptism, but Tertullian gives some evidence that priests and even deacons might be authorized to baptize, as well as lay persons in certain cases.[10] After their immersions, the candidates were anointed, had hands laid on, were signed on the forehead in the sign of the cross, and were blessed (possibly all in one single gesture). All of these subsequent rites of consignation symbolized the gift of the Holy Spirit. The newly baptized were then re-robed and allowed to join the congregation for the Eucharist at the climax of the paschal vigil. The neophytes (Tertullian calls them "novices") were given a cup of mixed milk and honey in addition to the bread and wine, in order to symbolize both their status as newborn children and their entrance into the Promised Land.

All these various ritual elements could be superimposed on the plan of a basilica with its adjoining baptistery. Noel Duval attempted this very thing for the baptistery of Vitalis in Sbeitla (fig. 19).[11] Ancillary rooms might have served the purposes of such pre- or post-baptismal rites as the initial disrobing, anointing, imposition of hands, and re-robing after the three-fold immersion in the font (it is impossible to be certain about the use of any of these spaces).

The number of steps into and out of the font may have been as symbolic as convenient; three steps to represent the three immersions or days that Christ stayed in the tomb, four for the four rivers of paradise or the four gospels. In one case (Bulla Regia) the font appears to be divided into three parts, each with its flight of steps, possibly permitting the descent of assisting priests, deacons, or deaconesses. The placement of the steps themselves (on eastern and western sides of the fonts, for instance) would have been more than a matter of managing the flow of movement in and out, but also symbolized passage from an old life to a new one. Immediate access from the outside into the baptistery would have allowed candidates to enter symbolize their leaving a "secular" world, while egress from baptistery directly into the basilica nave would be necessary, once candidates were re-robed, were ready to process triumphantly in to the midst of the congregation at the most dramatic moment in the liturgical year.

Like the number of steps into the font, other distinguishing architectural features of each baptismal space would have had symbolic as well as functional value. The relative position of the baptistery to the apse in some cases would have meant that the newly baptized would enter from the front and down the aisles, in full view of the congregation instead of from the back. The shapes of the fonts themselves would have reinforced the theological symbolism of womb and birth canal in some instances, or tomb and cross in others. Hexagonal or octagonal fonts were most likely designed to capitalize on the number symbolism of the sixth and eighth days of the week (either crucifixion or new creation).

Obviously, local traditions varied and each congregation individually developed or elaborated aspects of their rites according to the available spaces and the symbols that had come to be particularly valued over time. Possibly because of communities' varying wealth , or simply following prevailing traditions, some baptisteries were simple, some more complex; some more beautifully decorated, some more spare. Whether form followed function, or function form is probably a question without an answer. Moreover, although wide variance in the shape of liturgical space does not necessarily imply significant liturgical differences or local innovations, we can be sure that space imparts both symbolism and basic direction to liturgy.
 

B. The decoration of the spaces and baptismal symbolism/typology.
 

As described above, the mosaic decoration of the fonts and pavements which surrounding them primarily employed natural images (birds, fruit, flowers, trees) Some of these images very likely were only decorative, others more symbolic in their significance. Vines with grapes, vases, fish and marine creatures, and birds of all kinds were standard decorative programs for wealthy Roman dining room or courtyard pavements. Nevertheless, in these liturgical spaces, we might expect that more than decoration was intended by the images chosen.

The font from Kélibia serves as a good case study in the combination of decorative and symbolic images. At the pavement level the decoration consists of a geometric border inside of which are scrolling grape vines growing out of four urns placed in each of the corners. Four birds mark the centers of each side of the square (fig. 20). The font itself is divided into four segments, perhaps referring symbolically to the four rivers of paradise, or to the four gospels.

The top of the font shows only simple rosette-like or circle-in-square designs along with the four-part inscription. The first band of the inside of the font begins to become more complex, however. Here we see four primary images at the center of each lobe: a dove with breath, water, or oil coming from its beak (cf. the doves in the Ravenna dome medallions) and flanked by lilies and christograms with the alpha and omega; a cross under a ciborium between two quail (?) and four roses; a box-like ark, also between two quail; and a chalice filled with water and flanked by birds and roses (fig. 21-22). Separating each of these four images is a single lit taper, perhaps representing paschal candles. Each of these images, ark, chalice, cross under ciborium, and dove are elements of or allusions to the baptismal rite, or the eucharist which immediately follows it.

Separated from the first level by colorful bands and a geometric border, the second level (or register) functions as a step or seat. This register too has four central images, but three are identical and show a small fish beneath a dolphin which itself is beneath a christogram with an alpha and omega. The fourth of these compositions replaces the fish with a bee (fig. 23). Four different trees separate each of these images, each possibly representing one of the four seasons: date palm (spring); olive (winter); fig (summer); and apple (autumn), or (arguably) four trees in paradise. The fish are clearly a baptismal symbol and call to mind Tertullian's famous claim that we are all fishes who are (after the example of the Big Fish), born in water and only safe in water.[12] Dolphins are a familiar resurrection symbol (possibly drawn from the iconography of Dionysus). Because it is unique, the bee draws special attention. The bee might allude either to the neophytes' first eucharist of mixed milk and honey, or even refer to a section of the Exultet, the hymn in praise of the paschal candle which was sung in the liturgy of Holy Saturday. The use of the paschal candle in North Africa is attested to in the City of God (15.22). The Exultet, which tradition has dubiously ascribed to Augustine, might have originated in Africa and could have included the laudes cerei ("praise of the bee"), attested to in the Gelasian Sacramentary, but logically derived from an earlier source.[13]

Directly below this level is a band of white crowns (?) on a dark blue background, and in the basin of the font is the christogram with the alpha and omega surrounded by a wreath-like band. Taken together, the imagery recalls much of the typological and symbolic imagery found in the writings of the African fathers. Consider, for instance, this quote from Cyprian's epistle 73 (and recall that the baptistery was dedicated to this same saint): "The Church is like Paradise: within her walls she encloses on the inside fruit-bearing trees. Of those trees any which do not produce good fruit are cut out and cast into the fire. And those threes she waters by means of four rivers -- that is, by the four Gospels; by them she generously spreads in a saving and heavenly flood the graces of baptism."[14]
 

C. Multiple baptisteries in one region as evidence of competing sects.
 

The existence of more than one baptistery in a small city -- or even in a larger urban area (e.g. Belalis Minor, Sufetula, and Carthage), suggests two possibilities: either the deputizing of presbyters for the purposes of baptism by the fifth century onwards, or the existence of more than one bishop in an area and thus the presence of a competing Christian sect.

Carthage, with its five or more baptisteries in a fairly small geographical area, was also a populous and complex community (even linguistically divided?). Such a cosmopolitan center quite plainly needed more than one "parish" church, and clergy who could attend to the sacramental needs of these individual communities. The presence of multiple baptisteries, therefore, serves as evidence that the Bishop of Carthage shared his baptismal authority with other ministers who could have baptized simultaneously in different parts of the city.[15]

On the other hand, we know that competing sects (Catholics, Donatists, and Arians, for instance) had competing bishops, churches, and without question built their own baptisteries. But although we know that, we have very little evidence that these groups built discernibly different basilicas, baptisteries, or fonts. The only really useful, distinguishing evidence comes from inscriptions found in the churches themselves.

Several examples of such inscriptions have led scholars to identify certain churches as Donatist or Vandal (Arian) in particular. The basilica at Timgad is a well-documented Donatist bishopric, based on a dedicatory inscription to its Donatist bishop, Optatus (consecrated in 388), whom Augustine identified as the "undisputed master of Southern Numidia," and with the so-called Circumcellions in central and southern Numidia.

In Makhtar, one of the two extant basilicas contained a mosaic tomb cover near its entrance which bore the Vandal name "Hildegun," leading scholars to surmise that the church itself was Arian. By process of elimination, the other basilica was deemed to be Catholic. Unfortunately Hildegun's church had a baptistery, while the other basilica's baptistery (if there was one) has not yet been discovered.

In Sbeitla, three churches have baptisteries, and two are part of the same cathedral complex (Jucundus and Vitalis). Jucundus, the martyr whose relics may have been placed in the font after the community built its new baptistery, was an identified Catholic bishop, listed as attending the Council of Carthage in 416, and martyred during the Vandal invasion. By process of elimination, again, that would leave the third basilica (Sbeitla III), as the likely Donatist church. The baptistery in this presumed Donatist church had an apsed ancillary room, which led Duval and others to conclude that Donatists gave particular emphasis to the efficacy and exclusivity of the rite, particularly in the context of the baptismal controversy with the Catholic church.[16]

Mustis was the episcopal seat of Maximian, the Donatist schismatic and foe of Primian. Surely, since this purely intra-Donatist schism was mostly settled by the Council of Carthage in 411, is is difficult to tell what congregation was served by the existing (Byzantine) building on the site. So we are left to wonder. Similarly, the records of that same council (411) report that Bulla Regia had both Catholic and Donatist bishops. However the late date of the excavated basilica (sixth century) indicates that it was probably Catholic, given the repression of Donatists (as well as Catholics) during the Vandal era, and of both Vandals and Donatists after the Byzantine reconquest.

An imperial edict in 535 restored the privileges of the Catholic church and of the see of Carthage to what they had been before the Vandal conquest. Rome was given the rights of a patriarchal see over Carthage, while Carthage was given control over the internal affairs of the African church, both in Proconsular Africa as well as Numidia. Donatists, Arians, and Jews were proscribed, although they probably didn't disappear overnight (or ever completely), and a Donatist presence has been argued for well into the Muslim era.[17] The Byzantine occupiers undoubtedly wielded their influence over new construction and even the saints to whom the churches were dedicated (e.g. the Basilica of Gervasius and Protasius in Sbeitla). No doubt the occupying armies and their families wanted to honor their familiar saints, but such dedications might also have been an effort to suppress the local cults of the dead.

Analyzing the architecture and baptisteries of known or presumed Vandal and Donatist churches, however, yields no obvious distinguishing features that would assist in identifying other churches as Donatist or Arian. In fact, no reason exists to conclude that Catholic, Arian, and Donatist rites or liturgical patterns were substantially different from one another. Optatus of Milevis implied that Catholics and Donatists shared basic ecclesiastical practices. Presumably, no obvious distinctions existed between spaces either, since churches apparently changed sides through various stages in the controversy. Ecclesiology might diverge, but architecture not necessarily follow.
 

D. The connections between martyria and baptisteries: theological symbolism and social construction.
 

This section will have to remain unwritten for the moment. I simply wanted to mark it, for future discussion. To be considered: burials in baptisteries and the parallel symbolism in the funerary mosaics and fonts (candles, birds, flowers etc.).

1. See N. Duval, Les ruines de sufetula, 78.
 

2. For instance see J.G.Davies, The Architectural Setting , 2ff.
 

3. See Duval,
 

4. Oued Ramel, see Lapeyre, "La Basilique," 205.
 

5. F. Bejaoui, "Découvertes d'archéologie chrétienne," 1943-45
 

6. J.G. Davies, 39.
 

7. E. Palazzo, "Iconographie et Liturgie," 118-119.
 

8. Consult additional papers by Wm. Harmless and Graeme Clarke in this session.
 

9. Harmless, Augustine and the Catechumenate, 170-171.
 

10. de Bapt. 17.
 

11. Duval, diagram of the flow of movement through the Vitalis baptistery, Les ruines, .
 

12. de Bapt. 1.
 

13. E. Palazzo, "Iconographie et Liturgie;" Alw 34, (1992); O. Perler, "Die Taufsymbolik der vier Jahreszeiten," Mullus, Fst. T. Klauser, (1964); P.A. Février and C. Poinssot, "Les cierges et l'abeille," CA 10 (1959); and C. Courtois, "Sur un baptistére dans la region de Kélibia," Karthago VI (1955).
 

14. Ep. 73.10.3. Trans. G. Clarke.
 

15. de Bapt. 17.
 

16. Duval, Les ruines, 78.
 

17. Frend, The Donatist Church.