The First Vaulted Churches in

charles anthony stewart University of St. Thomas, Houston

he transition from wood roofs to stone vaulting Five barrel-vaulted churches survive in Cyprus: the Pan- marks a watershed moment in medieval architec- agia Chrysiotissa (Afentrika), Asomatos church (Afentrika), tural history. In the West this change took place Agia Varvara (Koroveia), Panagia Afentrika (Sykhada), and T 5 over course of the Romanesque period in Spain or northern Panagia Kanakariá (Lythrankomi). They are concentrated Italy.1 In the East scholars generally agree that vaulting co- on the northeastern panhandle known as the Karpas Penin- incided with the shift from longitudinal to centrally planned sula (Figure 1). This narrow stretch of land lies about 45 buildings during the reign of the Emperor Justinian I (482– miles south of the Anatolian coast and 60 miles west of Syria. 565).2 This simple explanation would be sufficient if not for Unfortunately these basilicas are in a poor state of repair; all the conspicuous absence of Justinianic-type churches outside are in ruins, and their frescos are disappearing. Preservation the imperial cities of Constantinople and Ravenna.3 It seems and systematic excavations could place these monuments that the building innovations of Justinian’s architects had in their rightful place in architectural history.6 Until then, limited impact on structural design, including vaulting, in the this article offers observations and new evidence toward a eastern Byzantine provinces. For example, the Nea church plausible construction sequence and chronology; as dem- in Jerusalem and St. Catherine’s at Sinai, although both onstrated here, eighth-century builders on Cyprus experi- sponsored by Justinian, were constructed as traditional mented with different ways of vaulting longitudinal space wood-roofed basilicas.4 Another case in point, the island of and developed designs that became hallmarks of Cypriot me- Cyprus did not have centrally planned churches until the late dieval . eleventh century. However, several Cypriot basilicas were refitted to support barrel vaults in the eighth century. Since these structures have longitudinal ground plans, the shift to Early Christian Churches on Cyprus vaulting on Cyprus cannot be associated with the develop- Over the past forty years many ancient churches have been ments that took place in Constantinople. They exemplify the excavated on Cyprus. All sixty-five known Early Christian localized history of a different construction technology. churches, dating from the late fourth to mid-seventh cen- tury, display the same traits: they have at least three aisles 7 Journal of the Society of Architectural Historians 69, no. 2 (June 2010), 162–89. ISSN divided by colonnades supporting wood roofs. In this pe- 0037-9808, electronic ISSN 2150-5926. © 2010 by the Society of Architectural Historians. riod there are no centrally planned, domed, or vaulted All rights reserved. Please direct all requests for permission to photocopy or reproduce article content through the University of California Press’s Rights and Permissions web- buildings. The conservative nature of Cypriot church build- site, http://www.ucpressjournals.com/reprintInfo.asp. DOI: 10.1525/jsah.2010.69.2.162. ing can be explained by the insular character of the Church.

JSAH6902_02.indd 162 4/16/10 4:12 PM Figure 1 Map of Cyprus indicating principal sites of the eighth century. See JSAH online to view Cyprus in Google Maps

During the First Council of Ephesus (431) the Cypriot - Agios Spyridon (Tremithous) and Agios Mamas (Morphou), bishop was granted autocephaly, meaning that he was not preserved the layout of their original wood-roofed struc- subject to any outside patriarch. The Cypriot Church could tures, but their imported marble colonnades were replaced maintain its own internal appointments and customs. Be- by square-piered arcades made of local stone.9 In these two cause of its independent and powerful hierarchical system, cases, the walls, apses, floor mosaics and opus sectile, and heresies such as Arianism and Monophysitism did not affect liturgical furnishing were reused after the renovations. By Cyprus as they did other Byzantine provinces. Likewise, its recycling the shell of the damaged church and using local church architecture was not affected by outside influence. materials, Cypriot builders were able to renovate quickly. The island possessed both limestone quarries and vast cedar Around this same time experiments in barrel vaulting were forests that could supply the main materials for traditional conducted on the . basilica construction. Other materials used for decorating Today the peninsula is remote, barren, and sparsely the interior of churches, such as marble , opus sec- populated, however in the late fourth century it had thriving tile, and tesserae, could be imported from other provinces cities, deep harbors, and thick forest. One of the island’s by ship, across the Mediterranean Sea. twelve bishops resided on the peninsula in the city of Car- The Cypriot Church’s three-hundred-year building tra- pasia. This see was founded by Epiphanius (320–403), the dition was disrupted in 649, when the Arab armies invaded. archbishop of Salamis-Constantia, who appointed his disci- From that time until the Byzantine reconquest of 965, the ple Philon as its first bishop.10 The architecture of the early island was a neutral state within the political sphere of both fifth-century cathedral now known as Agios Philon repeated the Arab Caliphate of Damascus and the Byzantine Empire. the layout of other Early Christian basilicas on the island, This era is referred to here as the “period of Cypriot neutral- having three-aisles, three apses, and wood roofs. These el- ity.”8 Sometime after the initial Arab raids, seven churches ements are also evident at other early churches nearby, in- were rebuilt in a slightly different manner. Some, including cluding Agia Trias (Yialousa) and the Panagia Kanakariá

T h e F i r s t Va u lt e d C h u r c h e s in Cyprus 163

JSAH6902_02.indd 163 4/21/10 1:50 PM (Lythrankomi).11 These structures exhibit how Early Chris- testifying to a large population during the Hellenistic and tian architecture fared in the Middle Byzantine period on Roman periods. The principal ruins today consist of four this part of the island. At Agios Philon, the cathedral was buildings: the Panagia Chrysiotissa, the Asomatos church, abandoned in the seventh century and subsequently rebuilt Agios Georgios chapel, and Prophitis Elias.15 Seven feet as a cross-in-square church in the eleventh century. The northeast of the Panagia church is a large archway leading sixth-century church of Agia Trias was destroyed in the mid- to a subterranean rock-cut tomb. Within are several Hel- seventh century and never rebuilt. In contrast, the sixth- lenistic arcosolia niches cut into the walls, later inscribed century Panagia Kanakariá was damaged in the seventh with crosses, and on the floor are scattered remains of Afri- century and subsequently reconstructed as a barrel-vaulted can red-slip pottery (Figure 3). Since the sixth century basilica. church was partially constructed over this tomb, it would Near the Panagia Kanakariá church, four other wood- seem that the ancient necropolis was later converted to a roofed Early Christian basilicas were transformed into Christian shrine.16 Such a concentration of churches indi- barrel-vaulted structures. In order to carry heavy vaulting, cates that Afentrika was a major cult center in the early cross-shaped piers replaced aisle colonnades. The new fea- Middles Ages. tures of this scheme—square imposts, round transverse The Panagia Chrysiotissa currently is a small single- , and windowless clerestories—became key features in aisle chapel, probably rebuilt in the sixteenth century, to Cypriot church architecture. The best-preserved examples, judge from its pointed arches, vaulting, and rounded spring- the Panagia Chrysiotissa and the Asomatos church, are lo- ing corbels.17 Its name means “the Virgin of Gold,” perhaps cated at the site of Afentrika, five miles east of Carpasia. referring to a lost decoration or icon. Surrounding this cha- pel are the ruins of an earlier Early Christian basilica that was refitted to support a barrel vault (Figure 4). The nave apse The Panagia Chrysiotissa (Afentrika) and southern aisle wall still stand to a height of 8 feet. In The ruins of Afentrika lie near the north coast of the Karpas Georgios Soteriou’s 1935 photograph, portions of the south- Peninsula. Rising over the site is an acropolis where either ern aisle’s barrel vaulting are shown intact; however by 1946 a temple or citadel once stood (Figure 2).12 In 1889 David this had fallen, as documented by A. H. S. Megaw.18 Hogarth proposed, based on his reading of Diodorus of Sic- The Early Christian church (first phase) was a three- ily (first century BCE), that Afentrika was a later town built aisled wood-roofed basilica with three apses (Figure 5). It had from the ancient city Urania (Ουρa´νια).13 Since there are the same general layout as the Early Christian churches of no other ancient harbors in the area, Hogarth’s hypothesis Agia Trias (Yialousa) and Agios Philon (Carpasia).19 From its is still plausible.14 Today vast numbers of pottery shards are external wall surface, the Panagia Chrysiotissa’s ground plan found, spread from the coast to the ridge of the acropolis, measured 55 by 75 feet (17.6 × 24 m, not including the apses).

Figure 3 Panagia Chrysiotissa, Cyprus, sixth-century entrance of the Hellenistic subterranean rock-cut tomb, later converted to a Christian Figure 2 Schematic map of Afentrika (Urania), Cyprus shrine, looking northwest (author’s photo)

164 jsah / 69:2, june 2010

JSAH6902_02.indd 164 4/16/10 4:12 PM Figure 4 Panagia Chrysiotissa, exterior view, looking east (author’s photo)

Its three apses were connected by passages, a shared feature of the island’s earliest churches.20 The nave arcades had seven columns carved from local limestone with responds at each end. The original south wall and central apse were con- structed with courses of alternating stretchers and headers.21 Three large windows (5 feet by 2 feet 6 inches) pierced the Figure 6 Panagia Chrysiotissa, surviving apse window, showing a reduction in the second phase, looking east (author’s photo) apse above the synthronon (Figure 6). Today earthen mounds spread out beyond the apse and southern aisle, presumably the remains of adjacent rooms lie underneath. One of these The builders replaced interior colonnades with pier arcades, was uncovered in 1964, exposing foundations of a chapel each consisting of four cross-shaped piers with square re- flanking the south aisle apse, perhaps functioning as a parek- sponds at each end. The piers’ imposts carried transverse klesion.22 arches forming five bays in each aisle. The southwesternmost When this church was rebuilt as a barrel-vaulted basilica, was left in situ to bolster the new archway; it had to the original walls and apse were reused (Figure 7). The altar, be removed in order to center the arch and was then re- synthronon, and chancel screens were also reincorporated.23 erected (Figure 8).24 An additional pier was built along

Figure 5 Panagia Chrysiotissa, cutaway reconstruction of original Figure 7 Panagia Chrysiotissa, cutaway reconstruction of vaulted structure, first phase structure, second phase

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 165

JSAH6902_02.indd 165 4/16/10 4:13 PM the interior of the original south wall to support the aisle vaulting; this resulted in a blind arcade whose niches mirror the nave arcade (Figure 9). The north aisle foundations, which are currently buried, did not have a similar blind ar- cade.25 In order to mitigate the stress caused by the barrel vault, the builders added a thin wall around the original apse and reduced its window apertures (see Figure 6). The height of the nave vaulting was not uniform. The vault of first three bays (from the western wall) curved at a height of 32 feet. Megaw observed that the remains of the fourth bay did not continue this curvature and appeared more vertical. This caused him to hypothesize a transverse barrel vault running perpendicular to the nave and formed a transept.26 This suggestion is highly unlikely. In Early Chris- tian basilicas, such as St. Peter’s (Rome) or Santa Croce (Ravenna), the transept usually crosses the middle or east- ernmost bay. But here at the Panagia Chrysiotissa, Megaw proposed that the fourth bay was the crossing square, which does not conform to any known design on Cyprus or any- where else. Athanasios Papageorghiou cast doubt on this hypothetical transept when he supervised the clearing of the site in the 1960s.27 He partially exposed the eastern bay’s north wall, confirming that there were no transept Figure 8 Panagia Chrysiotissa, remains of original column re-erected arms. So if the church did not have a transept, how should under the southern arcade, looking north (author’s photo)

Figure 9 Panagia Chrysiotissa, blind arcade in the vaulted southern aisle, looking southeast (author’s photo)

166 jsah / 69:2, june 2010

JSAH6902_02.indd 166 4/16/10 4:13 PM fashion the vaults. Builders used stones of varying sizes for the arcades; most are rectangular limestone ashlars measur- ing about 14 by 7 inches (35 × 18 cm), facing a rubble core. Slightly smaller blocks were used for the vaulting. The uni- formity of the stones suggests that they were mined from a limestone quarry rather than from spolia.28 The impost blocks and transverse arches were made of finely carved ash- lars of a standard size. To fill the gaps between ashlars, small pieces of ceramic and tile were embedded in the lime mortar (Figure 10). Little is known concerning the Panagia Chrysiotissa’s internal decoration. The original floor surface has not been uncovered. Inside the sixteenth-century chapel, a marble step demarcates the bema; it was part of a chamfered cor-

Figure0 1 Panagia Chrysiotissa, southern aisle impost showing lime nice in the original church. A similar beveled cornice can mortar with tile infill (author’s photo) be seen in the Early Christian apse supporting the semi- dome. Gypsum plaster and frescos survive on the apse and south wall. Blue and red pigments can still be seen but are the apparent vertical transition be interpreted? Upon closer too faded to discern form or style.29 Furthermore, there is inspection, the fourth bay’s vaulting has a more subtle curve no trace of windows above the arcades in the nave. In other than the adjoining bays. It seems that the vaulting over the words, where one would expect to find a clerestory, there is two easternmost bays was taller than the rest of the nave. a solid vault. This greater height emphasized the sacramental function of the bays covering the bema (the sanctuary area reserved for the clergy). Asomatos Church (Afentrika) The builders of the Panagia Chrysiotissa left behind The Asomatos church lies less than 100 feet (30 m) south of clues regarding their methods. Surviving putlog holes in the the Panagia Chrysiotissa (Figure 11).30 The two churches nave and aisles indicate that wooden centering was used to have the same layout and style of construction. The exterior

Figure1 1 Asomatos church, Cyprus, exterior view, looking east (author’s photo)

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 167

JSAH6902_02.indd 167 4/16/10 4:13 PM Figure 12 Asomatos church, reconstructed elevation

dimensions of the Asomatos are 41 by 53 feet (12.5 × 16 m, between the Panagia and the Asomatos church is that the not including the apses)—somewhat smaller than the Pan- latter has much larger foundation stones. From the ground agia. The original basilica had three aisles, two arcades of to the impost level, ashlars measure approximately 28 by five columns with responds, three semicircular apses, and a 18 inches (70 × 45 cm); above the imposts the ashlar dimen- wood roof. The apse belongs to the original church display- sions are similar to the Panagia’s. ing finely-carved ashlar blocks laid with alternating the The west wall of the Asomatos church had three portals headers and stretchers. leading to each aisle. Because these did not have fasteners or Like the Panagia, the Asomatos was converted into a hinges, Camille Enlart proposed that a narthex once encom- vaulted basilica utilizing the original ground plan (Figure 12). passed the basilica’s western end.32 There is, however, no The three apses, opus sectile flooring, and the synthronon evidence of an abutting wall on surviving surfaces, casting were reused from the previous church. However unlike the doubt on his suggestion. Today only the entrance in the Panagia, the north, south, and west walls were completely south aisle is preserved; its round arch is slightly wider than rebuilt upon the earlier foundations. Within these walls, two space between the jambs, appearing somewhat like a horse- pier arcades were constructed, consisting of three cross- shoe arch (see Figure 11). The bottom of another doorway shaped piers with pilaster responds. These arcades were con- can be traced in westernmost bay of the south aisle, which nected by transverse arches that spanned across the nave was subsequently abandoned and filled in by the time the forming four bays. Both the north and south walls were re- imposts were laid. This might indicate that a parekklesion inforced with an adjacent interior blind arcade. Two levels of was planned flanking the south wall. putlog holes, the top measuring 10 feet (3 m) above the other, Since its superstructure and vaulting is intact, the Aso- initially held wood scaffolding used to center both the nave matos church provides important clues regarding how its vaults. Evidence for similar centering is also found in the builders converted the original wood-roofed basilica to a aisles. Since about 2 feet (60 cm) survives above the springing barrel-vaulted structure. A vertical stratigraphy was formed point of the nave’s barrel vault, it is certain that there were by the imposition of the second phase components (e.g., the no clerestory windows. internal pier arcades) adjacent to the first phase construc- The masonry of the vaulted phase is comparable to the tions (e.g., the walls). Because masonry between the two Panagia’s. The spaces between ashlars were filled with small phases differ in quality and appearance, it is possible to trace pieces of ceramic and tile embedded within the lime mortar. the construction sequence (Figure 13).33 First, at some point Only a few monuments on Cyprus were built in this man- the original building was damaged. Elements such as the ner, including the late-seventh-century walls at Salamis- internal colonnades, that could not be reused, were re- Constantia and the so-called Arab Tower at the Panagia moved, leaving the apses and foundations in place. Builders Limeniotissa at Paphos. The only other church built in this then constructed new walls, and square piers were placed fashion was the Cathedral of Agios Epiphanios (Salamis- along the north and south walls; these were connected by Constantia), when it was rebuilt as a square-pier wood- arches, forming a blind arcade. At that time they built the roofed basilica in the late seventh century.31 One difference internal nave arcades supported by square piers. Fourth,

168 jsah / 69:2, june 2010

JSAH6902_02.indd 168 4/16/10 4:13 PM Figure 13 Asomatos church, schematic sequence illustrating the process of vaulting. Top, left to right: original wood-roofed basilica; with damaged wood roof and columns removed; with original north, west, and south walls removed. Bottom, left to right: new walls were built and pier arcades were constructed in the side aisles to support the aisle vaulting; once the aisle vaulting was finished, the nave vault was completed; exterior view of completed vaulted basilica

transverse arches were built spanning aisle piers and the and transverse arches channeled the north-south lateral nave arcade. After the aisle vaults were finished, builders forces from nave vault to the aisle piers. In addition, the east- proceeded to construct the nave vaulting. west lateral forces were transferred across the arches between It is certain that the aisle vaults were crucial in support- the piers. This network of perpendicular arches foreshadows ing the larger nave vault. As seen in the surviving south aisle, the development of modular planning usually attributed to the vault thickens as it merges with the upper part of the nave the Carolingian period.35 To further stabilize the structure, arcade (Figure 14). This buttressed the superstructure below the outer walls abutted the inner blind arcades. There were the springing point of the nave vault (see Figures 11, 12). As other advantages to this support system. Aisle niches occu- a result, the thrust exerted by the main vault was partially pied less space than a thicker wall, preserving the original channeled outward from the inner arcade to the exterior interior dimensions. Moreover, wall niches mirrored the ar- walls. With these stable side vaults in place, the architects caded nave, contributing to a symmetrical and consistent could build the nave vaulting with confidence during the internal arrangement. final stage of construction.34 In the original wood-roofed basilica, the side aisles supported the tall walls that served as clerestories; in the second phase, side aisles buttressed the Agia Varvara (Koroveia) nave vault. Apparently stability was more valuable than light- Standing about 15 miles (24 km) southwest of the Asomatos ing at this later period. church and about 500 yards (450 m) from the southern coast Unlike the Panagia Chrysiotissa, where the original is another barrel-vaulted basilica called Agia Varvara (St. walls were preserved, at the Asomatos church builders con- Barbara) (Figure 15). Its earlier designation was Ayios Yeor- structed the new outer wall adjacent to the blind arcade in gios as marked on Horatio Kitchener’s 1885 map. Since the aisles. It would have been more economical if they simply then, the church has been conflated with an abandoned constructed a thicker outer wall, but the blind arcades in the settlement called Ayia Varvara about half a mile to the north, aisles served an important purpose. As explained, the vaulting and it is now referred to by this name. Neither site was

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 169

JSAH6902_02.indd 169 4/16/10 4:13 PM Figure 14 Asomatos church, exterior view of the south aisle apse (now fallen), looking west (author’s photo)

documented prior to Kitchener’s map. The nearest modern including the apses); these dimensions are comparable to town, Koroveia (about 1.8 miles north), appears on the 1570 those of the Asomatos church. But unlike the Afentrika map by Venetian Iacomo Franco as “Corouio.”36 At one churches, there are no blind arcades adjacent to the aisle point the church belonged to a coastal town, judging from walls, and putlog holes are absent. Perhaps the builders used the scattered building materials and ceramic shards extend- centering that rested on the floor. Moreover, the masonry is ing for half a mile westward from the church.37 Today the less refined than in the Afentrika churches. A rubble core land is occupied by a large goat farm. Agia Varvara lies in was faced with large ashlar blocks mixed with irregular ruins with much of its masonry strewn about within its stones, including shards of ceramics and tile. The imposts aisles. Since most of the foundations are buried, it is difficult and voussoirs appear to have been salvaged from other to assess the Early Christian layout. Fortunately the north buildings, judging from their awkward and various sizes. wall and pier arcade survive to a height of 8 feet (2.5 m), The surface would have appeared uneven and scabrous even which provides some sense of its later rebuilt phase. In its after the walls were plastered. final form the church was a three-aisle, three-bay, barrel- While a modest structure, the church is significant for vaulted basilica (Figure 16). From its external wall surface, the art history of Cyprus. In the northern aisle, on the east- its ground plan measured 41 by 44 feet (12.5 × 13 m, not ernmost arch’s intrados survives a faded and partially

170 jsah / 69:2, june 2010

JSAH6902_02.indd 170 4/16/10 4:13 PM Figure 15 Agia Varvara, Cyprus, interior view of the north aisle, looking east (author’s photo)

damaged fresco. Reconstruction is possible since the form Panagia Afentrika (Sykhada) and style consists of a repeating geometrical pattern (Figure Five miles east of Agia Varvara stands another barrel-vaulted 17). It is an example of a wheel-interlace design that is rare basilica known as the Panagia Afentrika (Figure 18). Its in Early Christian or Middle , but ground plan is approximately 41 by 45 feet (12 × 13.5 m, not more common in Umayyad mosaics of the eighth century. including the apses or narthex). It once stood near a settle- Its appearance in this barrel-vaulted church provides an im- ment called Sykhada, which is now deserted. Historic portant clue regarding its date.

Figure 16 Agia Varvara, hypothetical plan of the barrel-vaulted basilica

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 171

JSAH6902_02.indd 171 4/16/10 4:13 PM two arcades consisting of four columns with responds, made of local limestone. A “cleaning” of the church in 1931 ex- posed the synthronon, the original floor, and remains of a marble ambo and chancel screen.40 Unfortunately the floor today is covered by dirt and weeds, concealing the opus sec- tile tiles. When the church was converted into a barrel-vaulted basilica, all three apses and the church furnishings were reused. In order to support the vaulting, additional walls were constructed along the original northern and southern walls. Arcades divided the aisles, as at Agia Varvara; each Figure7 1 Agia Varvara, reconstruction of arcade soffit fresco, faded consisted of two cross-shaped piers, forming three bays. areas are damaged sections (author) Likewise, its large ashlar blocks were joined together with lime mortar containing rubble, ceramic shards, and tile. documentation for the town and church are lacking. The Clearly the builders here used spolia from the earlier toponym sykada (συ´κα) means “fig” and could possibly be church, including the columns and marble furnishings. the town where the ninth-century Demetrius, bishop of Unlike the other churches, neither piers nor blind arcades Chytroi, was born.38 The site was passed over by Hogarth, were added to support the aisle vaults. Instead, the trans- but visited by Enlart in 1897, who mentioned that the locals verse ribs were carried by square corbels jutting out from simply called it “the Panagia.”39 Today potsherds and ashlars the walls with no pilaster supports (Figure 20).41 Blind ar- are scattered in adjacent fields, attesting to an ancient cades were used on the west and east walls of the narthex, settlement. which belonged to the same phase of construction as shown Like the churches at Afentrika, the Panagia at Sykhada by the integration of the masonry between the narthex and has traces of an earlier sixth-century church (Figure 19). the aisles. Three entrances led from the narthex into the This was a three-aisled, three-apse, columnar basilica, with nave and aisles—these do not show signs of hinges or door

Figure8 1 Panagia Afentrika, exterior view, looking southwest (author’s photo)

172 jsah / 69:2, june 2010

JSAH6902_02.indd 172 4/16/10 4:14 PM Figure9 1 Panagia Afentrika, Cyprus, plan of the vaulted basilica (redrawn after Camille Enlart, L’Art gothique et la Renaissance en Chypre [Paris: Thorin, 1899], 399; and A. H. S. Megaw, “Three Vaulted Basilicas in Cyprus,” Journal of Hellenic Studies 66 [1946], 55)

clasps. The surviving door in the south aisle is a simple Little decoration survives at the Panagia Afentrika ex- post-and-lintel design. Another doorway pierced the south cept for fragments of fresco painting. On the south east- wall; its arch was wider than the jambs, similar to the door- ernmost soffit is a line drawing of a nimbed saint with an way at the Asomatos church. This door led to a subsidiary elongated head.43 Nearby the south wall interior bears chapel with an apse, like the parekklesion of the Panagia faint traces of other figures with haloes. Unfortunately Chrysiotissa.42 these are too faded to ascertain details or attribute a style.

Figure 20 Panagia Afentrika, south aisle vault with hanging corbels supporting transverse arch (author’s photo)

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 173

JSAH6902_02.indd 173 4/16/10 4:14 PM In the northwest corner are more vibrant colors, ultrama- of the Kanakariá church was a barrel-vaulted basilica.46 How- rine and scarlet, belonging to a border that covers earlier ever, Megaw challenged this hypothesis in his 1977 mono- fresco layers. graph, claiming that this rebuilt structure had a wood roof. He never provided clear reasons. Perhaps his interpretation was based on the awkward arrangement of the side aisles. For Panagia Kanakariá (Lythrankomi) instance, the south aisle is taller than the north aisle by over Eleven miles west of the Sykhada basilica stands the Panagia three feet and the south apse has pointed arches, indicating Kanakariá in a village called Lythrankomi (Figure 21). Its a post-Byzantine renovation. Furthermore each side aisle unique apse mosaic of the enthroned Virgin and child has incorporates a second arcade abutting the nave arcade. Al- been the subject of several art historical studies.44 The build- though these arcades are sandwiched together and are simi- ing itself is the result of several renovations. Portions of the lar in scale, they are misaligned by a few inches. This clumsy original sixth-century basilica are preserved in the central arrangement may have led Megaw to identify the current apse area and within the foundations. Six marble capitals lie side aisles and their arcades as later additions. In their ab- within the churchyard and probably belong to the original sence, the nave arcade could not have sustained the weight building. At some point, the church was rebuilt with square- exerted by a barrel vault. Perhaps this is why Megaw conjec- pier arcades replacing the original marble colonnade (Figure tured that only a lightweight wood roof could span the sec- 22). Unlike the churches at Afentrika, the Kanakariá’s piers ond phase building. He further postulated that a tall, are close together, following the intercolumniation of the rectangular tower rose over the bema, in an attempt to ex- previous church.45 The arcades stop at the bema, where plain the massive bema walls.47 walls separate the nave from the aisle, forming “bema walls.” Megaw’s reconstruction raises issues that bear further These walls are massive; their thickness equals the total analysis.48 First, he assumed that the inner nave arcades were width of the arcade piers. built prior to the outer aisle arcades. However, there is no Scholars such as Enlart, Andreas Dikigoropoulos, and structural evidence for this. In fact, both arcades are similar Athanasios Papageorghiou proposed that the second phase in style and construction. For instance, the average size of

Figure 21 Panagia Kanakariá, Cyprus, exterior view looking northeast, remnants of Early Christian church in foreground (author’s photo)

174 jsah / 69:2, june 2010

JSAH6902_02.indd 174 4/16/10 4:14 PM Figure 22 Panagia Kanakariá, plan of the vaulted phase (redrawn after Camille Enlart, L’Art gothique et la Renaissance en Chypre [Paris: Thorin, 1899], 402; and A. H. S. Megaw and E. J. W. Hawkins. The Church of the Panagia Kanakariá at Lythrankomi in Cyprus [Washington, DC, Dumbarton Oaks, 1977], fig. C)

their ashlar blocks (1.5 by 2 feet) and the molding of the weighty structure in the exterior masonry; but no such imposts are identical. Moreover in Cyprus traces of earlier traces exist. Tall rectangular lantern towers are uncommon plaster were usually left between two separate building in the eastern Mediterranean and not found in Cyprus.49 phases. But at the Kanakariá there is no trace of plaster be- While it is possible that the Panagia Kanakariá had a unique tween the adjoining arcades. While the north aisle arcade crossing tower (and evidence for this was later destroyed), had its plaster removed, the nave arcade’s surface was left it seems unlikely given the uniformity of architectural ty- alone, contributing to the appearance of two separate phases pologies on Cyprus. (Figure 23). If the two arcades were built during the same It is more reasonable to associate the second phase of the phase of construction, their combined strength was more Panagia Kanakariá with the other vaulted churches. Located than sufficient to support a vaulted nave. in the Karpas Peninsula, just five miles northwest of the Megaw proposed that the cross-shaped piers of the vaulted Agia Varvara, the Kanakariá church shared a similar second-phase supported wood trusses. However, arcades environment and history with the other basilicas. It exhibits with engaged piers are absent from all other wood-roofed three important traits with the Panagia at Sykhada. Both basilicas in Cyprus, while they are present in the barrel- have transverse arches in their side aisles that spring from vaulted churches. In these buildings, engaged piers sup- square corbels from the wall. Both have narthexes rebuilt ported transverse arches, which in turn carried the vaulting. with blind arcades that carry the barrel vaulting. And both It would have been superfluous for the Kanakariá church to churches have “horseshoe” arched doors in their south aisles, have transverse arches to carry a wood roof. Furthermore, like the Asomatos church (see Figure 11). the surviving clerestory of the second phase is rather low, Kanakariá’s arcade arrangement is also similar to the and such a massive support system would have been un- Afentrika churches. The Panagia Chrysiotissa and Asomatos necessary for such a small amount of masonry. Megaw’s basilicas have blind arcades adjacent to the north and south hypothetical reconstruction also included a problematic walls. The builders of the Kanakariá took a slightly different crossing tower. One would expect to find vestiges of such a approach by placing the additional arcades adjacent to the

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 175

JSAH6902_02.indd 175 4/16/10 4:14 PM Figure 23 Panagia Kanakariá, north arcade, easternmost pier, showing two abutting arcades, view from the side aisle looking south- west (author’s photo)

nave arcades rather than the outer walls. In other words, the problem of distributing the lateral forces of the nave vault was solved using a variation of the same solution employed at the Afentrika churches. There are two plausible reasons for the massive walls enclosing the Kanakariá’s bema. The current structure could have supported an elliptical dome (see Figure 22).50 But since no other barrel-vaulted church incorporated domes, this is unlikely. Perhaps the walls simply carried two barrel vaults (i.e., a double vault), with one course of voussoirs sitting im- mediately above the other (Figure 24).51 The use of double vaults can be found in the multiple-domed churches of Cy- prus, such as Agios Barnabas (Salamis-Constantia) and Agios Lazaros (Larnaka), and in , such as the so-called Great Church at Maden S¸ehri–Deg˘le (also known as Bin- birkilisi).52 Outside of Cyprus, the cross-and-square church Figure 24 Panagia Kanakariá, cutaway reconstruction of vaulted struc- of St. John of Pelekete at Trilye () also has a double ture (second phase)

176 jsah / 69:2, june 2010

JSAH6902_02.indd 176 4/16/10 4:15 PM Figure 25 Panagia Kanakariá, arrows indicating imposts for transverse arches over the bema, looking northeast (author’s photo)

vault over its bema, tentatively dated to the late eighth cen- it would seem that they were built in response to similar tury.53 At the Kanakariá evidence for a large triumphal arch stimuli around the same time. What forced this radical survives that would have held up this double vault. Several transformation? Previous scholars have suggested three imposts of various heights indicate that this arch was rebuilt explanations. It has been proposed that the Early Christian several times (Figure 25). As discussed, the Panagia Chrysi- churches on Cyprus were destroyed in one cataclysmic otissa’s vaulting over the bema was also different than the rest event, either invasion or earthquake, which necessitated of its nave. The change in vaulting marked the sacred space immediate renovation. Others have suggested that the where the altar stood. In Early Christian churches, ciboria churches gradually fell into disrepair due to economic de- (baldachins) hovered over altars to convey the idea of holi- cline and depopulation, and were later updated when con- ness (i.e., that which is “set apart”).54 Perhaps the bema’s ditions changed. And a third model sees a new population vaulting served the same purpose, replacing the ciboria, immigrating to the peninsula and introducing this innova- when these churches were rebuilt.55 tive rebuilding campaign. Megaw and Papageorghiou espoused the first view and argued that the precipitating event was the Arab incursions Impetus for Change of 649 and 650. This theory was first formulated by Megaw All the barrel-vaulted churches exhibit similar design layout in 1946. He wrote, “No doubt the churches remained in use and masonry techniques. Since they are close geographically, . . . until the Arab raids obliged the inhabitants of the coastal

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 177

JSAH6902_02.indd 177 4/16/10 4:15 PM settlements . . . to withdraw to the relative security of the catastrophic earthquakes on Cyprus, Ruggieri’s theory could hinterland . . . the churches must then have fallen into disuse be applied to the island’s first vaulted churches.63 Slobodan and ruin, if they were not actually burnt by the invaders.”56 C´urcˇic´ concluded that “the principal cause of the destruc- According to this hypothesis, the wood-roofed basilicas were tion of early basilicas in Cyprus was not the Arabs, but earth- destroyed in war and subsequently rebuilt. This thesis is quakes,” reasoning that “Cypriot builders believed in based on numerous historical accounts, from both Byzantine increasing the buildings mass as the optimum way of coun- and Arabic sources, that describe the violent Arab invasions tering the effects of earthquakes . . . vaulting permanently of Cyprus.57 displaced wooden trussed roofs, thick walls replaced rela- Andreas Dikigoropoulos countered, however, that the tively thin ones, while the general height of buildings was destructive force may not have been human. He reasoned reduced.”64 Recently Megaw also proposed that the Cathe- that “it would not be difficult for the Arabs to set fire to a dral of Kourion was destroyed due to earthquake damage, wooden-roofed basilica, but, on the other hand, it is difficult based on the displacement of its walls.65 The coins found in to see either why they should—the purpose of the Arab the ruins point to the 680s as the time of destruction. In raiders could be achieved without setting fire to religious general, historical documents seem to substantiate this the- buildings—or, if they did, why the new church [that was] ory; violent earthquakes were recorded throughout the erected did not reuse the columns of the burned building . . . Eastern Mediterranean region in the Early and Middle Byz- it seems to me more probable that the church and columns antine period, while few earthquakes were reported in the were destroyed in an unrecorded violent earthquake.”58 He Western Europe, where the wood-roofed basilica tradition believed that both earthquakes and plagues between 649 and continued.66 965 could account for depopulation and economic decline. The earthquake theory might also explain why damaged If this were the case, then Early Christian churches, perhaps columns were not reused when these churches were rebuilt. already damaged by earth tremors, gradually deteriorated However it is still difficult to identify earthquake damage due to lack of maintenance. within the majority of Early Christian basilicas.67 History Both historical sources and archaeology testify to some records no specific earthquakes on Cyprus between 370 and Early Christian buildings surviving the first Arab invasions. 1114, although several sources state that Hagia Sophia (Con- For instance, documentary and archaeological evidence sug- stantinople) was damaged by quakes in 558, 879, 989, and gest that the cathedral of Agios Spyridon (Tremithous) sur- 1348.68 Why would Cypriot builders have associated stone vived the initial Arab attacks.59 In 655 an ecclesiastical synod vaulting with earthquake resistance if the well-known vaults was held there in commemoration of the saint who built the of Hagia Sophia were heavily damaged by seismic forces?69 original structure.60 Evidence like this led Tassos Papacostas Likewise there is no scientifically based reason to suppose to propose that Early Christian sites were “peacefully aban- that barrel-vaulted churches were more resistant to earth- doned” as early as the seventh century. 61 If so, rural churches quakes than wood-roofed structures.70 Gravity pulls barrel may have collapsed because of disuse, earthquakes, and other vaults downward in a unilateral north-south direction. Since economic factors decades prior to the Arab invasions. nave vaults were buttressed by side aisles, additional lateral Unfortunately it is difficult to confirm peaceful aban- forces on the walls, such as seismic waves, would create a donment by means of archaeology. If sites were evacuated domino effect with walls toppling over the arcades. due to coastal raids, the buildings might appear to have been Perhaps the transition to stone vaulting had nothing peacefully abandoned, since there is little reason why ma- to do with structural concerns. Papageorghiou once sug- rauders would destroy empty churches. Archaeology is ill gested that these barrel-vaulted churches were introduced equipped to measure how long a building might have been by an external cultural influence. He wrote that “after re- vacated, without historical markers such as textual evidence, turning from captivity, a large number of Cypriots became inscriptions, or coin finds. These indicators have rarely been acquainted with the Early Christian architectural in Syria found associated with the destruction of Early Christian . . . and presented a new architectural type to Cyprus . . . churches on Cyprus. the vaulted basilica.”71 However, few scholars have ac- The earthquake theory has steadily gained momentum cepted this theory since churches in Syria were not refitted in recent years. In 1991 Vincenzo Ruggieri explained how with vaults like those in Cyprus. Syrian basilicas, such as Early Christian churches were susceptible to destruction by the fifth-century Julianos church at Umm el-Jimal or the seismic shock. In order to resist these effects, Byzantine Bizzos church in Ruweha were constructed from the start builders designed an alternative form of sanctuary—the cen- as vaulted buildings.72 While it is true that these vaulted trally planned domed church.62 Since history records many structures predate the Cypriot churches, they do not have

178 jsah / 69:2, june 2010

JSAH6902_02.indd 178 4/16/10 4:15 PM the same characteristics—such as cross-shaped piers, blind testify to Arabic occupation throughout the eighth century. arcades and horseshoe-arched doorways.73 Moreover, the Panagia Limeniotissa church was modified to The majority of Early Christian churches that have incorporate a trapezoidal tower that could have served as a been excavated remain unpublished. The lack of systematic watchtower, lighthouse, or minaret.77 Twenty-seven miles and rigorous excavation records have made it difficult to east on the Kourion beach, another Early Christian basilica evaluate these divergent and equally plausible theories. was desecrated, as evidenced by an Arabic inscription on one However, there is one exception. Excavations at Kalavasos- of its marble columns; soon afterward the basilica was aban- Kopetra, in southern Cyprus, have provided the best evi- doned.78 dence for the demise of Cypriot churches in the mid-seventh Similar evidence has been discovered on the east and century. Those investigations revealed the foundations of north coasts. Two inscriptions dated 649/650 at Soloi cathe- three Early Christian basilicas located 2.5 miles north of the dral clearly attest to successive invasion, ruination, and the southern coast. Coin and pottery evidence indicates that all rebuilding of the basilica.79 In Salamis-Constantia the ca- three churches functioned during first four decades of the thedral of Agios Epiphanios was destroyed by a gigantic fire. seventh century and were subsequently abandoned. One of The excavators there found large quantities of “black earth,” the churches was Kalavasos-Sirmata monastery. Within its burnt material, and a human skull in one of the drains.80 courtyard, less than seven feet from the basilica, remains of Later research in Salamis-Constantia’s gymnasium provided several individuals (men, women, and children) were discov- clear ceramic and coin evidence of a mid-seventh-century ered in a water cistern.74 The archaeologists found evidence destruction.81 A few yards away at the Campanopetra church, that their placement in the cistern coincided with the mon- other excavators found indecipherable letters smeared in red astery’s destruction. Although forensic analysis was not able paint within the sanctuary; they suggested that this was an to ascertain how these individuals died, it is clear that an Arabic inscription for the purpose of desecration.82 earthquake was not the principal cause for the monastery’s The Karpas Peninsula lies only five miles northeast of ruination.75 Bodies in the cistern would have contaminated Salamis-Constantia. Therefore there is no reason to disas- the water supply and precluded further use of the church by sociate the destruction of the original Karpas churches with earthquake survivors. The only viable explanations are the widespread demise of Early Christian basilicas through- plague or war. out the island. Regardless of whether the catalyst was earth- The early-ninth-century Byzantine chronicler Theo- quakes or Arab invasion, or both, the original basilicas were phanes recorded a plague in Constantinople in 745 when damaged sometime between 650 and 700. But while this empty cisterns were used for burials.76 However, the histo- might explain the need to rebuild the Karpas churches, it rian clearly explained that this behavior occurred only in does not answer why they were redesigned with barrel vaults. Constantinople, where land was limited, cemeteries were full, and cisterns were empty. In contrast, Cyprus was arid and sparsely populated; there were vast acres of non-arable The Shift to Vaulting land that could have been used for burials, and cisterns were According to conventional wisdom, secular Roman basilicas not usually condemned or contaminated. Moreover, there were utilitarian buildings. Their central naves accommo- was evidence that the Kalavasos-Sirmata monastery’s relics dated large assemblies of people who could move through- were visited by the faithful long after the church was de- out the broad, unencumbered spaces. Light and fresh air stroyed. This indicates that the destruction of the basilica were channeled to the interior through the nave’s clerestory. was not due to peaceful abandonment. The best possible ex- There are no historical references to any symbolic signifi- planation is that the Arab incursions of the middle of the cance attached to the tall naves in these early structures. The seventh century caused social upheaval and disuse of the site. first churches commissioned by Emperor Constantine in Further evidence of the Arab destruction of churches the fourth century were only slightly different in form from comes from other locations on the southwestern coast. Ex- the secular Roman basilica. The Early Christian churches cavations at Paphos uncovered systematic destruction of two of Cyprus belonged to this line of development. When faced churches, the Panagia Limeniotissa and the cathedral known with rebuilding their churches with vaults, however, the Cy- as the Chrysopolitissa. Sections of each church were reused priots omitted the clerestory while retaining the taller nave as workshops after the mid-seventh century. It is highly un- design of earlier churches. Why? Before this question is an- likely that desacralization of this magnitude would have been swered, two factors must be kept in mind. First, other Cy- carried out by the local Cypriot population. Arabic inscrip- priot churches were rebuilt in the established manner—with tions and tombstones discovered within these churches wood roofs, square-piers, and clerestory windows—during

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 179

JSAH6902_02.indd 179 4/16/10 4:15 PM the same period.83 It was possible for the Karpas builders to side aisles—one aisle was for women and the other for men.88 have done similarly and simply reconstructed their churches In the Karpas churches the apse passages of Early Christian with pier arcades supporting wood roofs. This would have structures were retained. These tripartite sanctuaries allowed been more expedient and less costly. Second, while domed the clergy to deliver the Eucharist to the congregation in the centrally planned buildings had been popularized in Con- aisles.89 Furthermore, just as the dome symbolized the stantinople over a century earlier, the Cypriots chose not to dome of heaven in Justinianic architecture, it is possible that follow this trend.84 Therefore the Cypriot builders con- the vault of heaven was suggested by the new vaulting in the sciously experimented with barrel-vaulted basilicas when Karpas churches. Christians interpreted natural rainbows as other types of churches were known and could have been representations of the cosmic arch delineating a cross-section constructed. of the sky. The barrel vault was a similar continuous arch, Vaulting would have been more practical than wood mimicking the sky above.90 roofs for several reasons. As Megaw and Papageorghiou ar- gued, vaulting was durable and fire resistant. Wood-roofed basilicas could not be maintained at a time of continual war- The Dating of the Barrel-vaulted Group fare, as in the period of Cypriot neutrality, when there were Until the 1960s, scholars disagreed on the date of the rebuild- at least eight major Arab raids on Cyprus.85 The basilicas in ing of these basilicas. Camille Enlart examined the Afentrika the Karpas peninsula were easily approached by sea-borne and Sykhada churches during his research trip to Cyprus raiders, and this necessitated a fire-proof construction that in 1895–97, and concluded that they were Romanesque was solid enough to stand against occasional attack. This and that the Panagia Chrysiotissa had “nothing Byzantine theory has also been applied to explain the vaulting of the about it.”91 He hypothesized that Crusader architecture in earliest Romanesque basilicas in Spain and France, where Syria and Palestine inspired these churches. This was un- warfare with Moors and Vikings was a continuous threat.86 derstandable, since massive and dark three-aisled churches The transition from wood to stone could also be a sty- were quite common in France during the twelfth century. listic preference. Vaulted churches and domed architecture The system of cross-pier arcades, transverse arches, blind were common in other parts of the Byzantine Empire. The arcades, and square imposts can be found in many French Karpas barrel-vaulted churches meshed with these wider churches, such as Saint-Lazare (Autun, begun in 1120). trends in the Eastern Mediterranean, while retaining the tra- However, while photographing and drafting ground plans ditional basilica form. And there is no need to posit an exter- of these churches, Enlart did not realize that the Asomatos nal prototype, since there were earlier barrel-vaulted and Sykhada churches had two building phases. His classi- buildings in Salamis-Constantia, such as the gymnasium fication system assumed that barrel-vaulted basilicas were (fourth/fifth century) and the so-called St. Catherine’s Prison Western phenomena; vaulted Byzantine churches could (seventh century BCE). While those examples lack trans- only be centrally planned. verse arches and engaged pilasters, they could have suggested In 1931 George Soteriou argued against Enlart, sug- the potential of vaulting to late seventh-century Cypriot gesting that the churches dated to the sixth or seventh cen- builders. tury, like the Early Christian churches of Syria.92 The use of Symbolic considerations may also have prompted the the synthronon and opus sectile indicated to him that they adoption of vaults in these Christian churches. There is evi- could not be Romanesque or Comnenian. Their basilical dence that changes in the ceiling were used to signify the plans and square piers convinced him that they pre-dated the changes in the sacred character of space below. In both the centrally-planned Middle Byzantine period (843–1204). Panagia Chrysiotissa and the Panagia Kanakariá the area of Despite this reasoning, Enlart’s opinion would continue to the bema is demarcated by a change in the vaulting overhead. prevail, with English-speaking scholars such as George Jeffery Likewise, the distinction between the taller nave vault and and Rupert Gunnis repeating his ideas.93 shorter side aisles marks the segregated areas reserved for the In 1946 Megaw published a short article providing mod- clergy and those occupied by the laity. Apparently the liturgy ified plans and photographs of the Afentrika and Sykhada in the seventh and eighth centuries on Cyprus was still a basilicas. He identified the two building phases of each processional ritual.87 The entire nave was used for the cele- church. He tried to compromise the extreme dates held by bration, from the Great Entrance at the west wall to the read- Enlart and Soteriou, hypothesizing that they were built in ing of scriptures in the center, and finally to the Eucharistic the tenth-century, when Cyprus was reconquered by the celebration in the bema under the eastern bay. While the Byzantine Empire.94 He provided no evidence to support his clergy performed the service, the laity was restricted to the claim, nor did he cite examples elsewhere in the empire of

180 jsah / 69:2, june 2010

JSAH6902_02.indd 180 4/16/10 4:15 PM the same period. Few scholars were persuaded. Megaw chose dating of the Afentrika churches were incorrect. Following this dating because he could not imagine churches being Dikigoropoulos and Papageorghiou, he suggested that they built during the tumultuous period of Cypriot neutrality were built even earlier, immediately after the first Arab inva- (649–965), when the island was in the sphere of Arab gover- sion of 649; but he, too, provided no explanation for his nance and isolated from the Byzantine Empire. These his- change of opinion.103 By accepting this redating, Megaw torical assumptions were first communicated by George Hill completely transformed his view, stating that the Karpas in 1940: “It seems improbable that any important buildings churches “attest to a substantial recovery in the Cypriot com- can have been put up during the periods of the Arab raids, munities well before the re-establishment of Byzantine rule that is, from the middle of the [seventh] century to 965.”95 . . . [and that] during the period of neutrality some notable Megaw assumed that Islam prohibited Christian worship and monuments were erected.”104 Perhaps Megaw’s work within that Christian Cypriots were left without churches for over the Panagia Kanakariá led him to reconsider his dating three hundred years.96 This theory contradicts many his- model. torical accounts, which attest to a thriving Cypriot church Several pieces of evidence suggest that the church of the during this period, and Megaw’s protégé, Dikigoropoulos, Panagia Kanakariá is the earliest of these barrel-vaulted quickly modified this thesis in light of his own archaeological churches. First, the incorporation of a clerestory level below discoveries.97 the springing of the nave vault attests to the Kanakariá build- During the 1950s the most notable Early Christian sites, ers’ desire to maintain an Early Christian feature. The addi- such as Agios Epiphanios (Salamis-Constantia) and Agia Tria tion of the clerestory prevented the aisle vaults from (Yialousa), were excavated by Dikigoropoulos. Altogether he adequately supporting the nave vaulting at its springing uncovered five square-pier wood-roofed basilicas built on point. All later builders of vaulted churches would omit the the foundations of earlier Early Christian churches.98 The clerestory. Second, the Kanakariá’s aisle arcades, flanking the evidence indicated that these were immediately rebuilt after nave arcades, are evidently experimental. The architects un- the previous churches were destroyed, during the period of derstood that aisle vaults had to be completed prior to build- Cypriot neutrality. In order to reconcile the evidence with ing the nave; however, they were unsure of how to integrate Megaw’s architectural model, Dikigoropoulos suggested that these two construction phases. They solved their problem by the Arab raids, coupled with earthquakes and plague, caused constructing two arcades, one to hold up the aisle vaults and an economic downturn which stunted architectural develop- a second to hold up the nave vaults. These two arcades would ment. He concluded that the modest square-pier basilicas be built alongside each other. Their solution worked, but is were patchwork constructions, mere shadows of the previous decidedly less refined than the arrangement in the Afentrika structures. In contrast, his analysis of the Karpas barrel- churches, where the aisle arcades are against the north and vaulted churches indicated a reversal of Cypriot fortunes. He south walls. At the Kanakariá the rebuilt nave arcades fol- wrote, “the size and type of the churches of Aphendrika fam- lowed the original intercolumniation, suggesting that the ily suggest that they were built at a time when either the fi- inexperienced builders used the previous foundations as a nances of those commissioning their erection could afford template for their reconstruction because they were unsure the cost, or when financial assistance from outside the Island of the stability of square piers. These successful experiments could be forthcoming . . . since we know that the Cypriots taught Cypriots that massive piers, buttressed by additional invoked the help of Byzantium between 780 and 806 we must arcades in the side aisles, were more than sufficient to carry conclude that the above churches were erected during this the nave vaulting. period.”99 Dikigoropoulos believed that only Constantinople All Early Christian churches have atria and narthexes. could save Cyprus from the devastation caused by Islam.100 At some point, atria were completely omitted from Cypriot Even with lack of historical evidence for this hypothesis, churches, while few narthexes were built between the tenth scholars were convinced by Dikigoropoulos’ arguments. and twelfth centuries.105 The Kanakariá and Panagia at This led to a new consensus that the Karpas churches were Sykhada rebuilt their Early Christian narthexes with barrel- rebuilt in the late eighth century.101 vaults; the omission of narthexes in the other barrel-vaulted Papageorghiou was the last archaeologist to work on the churches is a fundamental change that suggests a later devel- Karpas basilicas. While he did not carry out full-scale excava- opment. Because all five barrel-vaulted churches share other tions, he studied their structures and supervised a regional similarities and important differences, it is possible to trace survey in 1965. He was convinced that the vaulted churches a line of development. were built earlier in the eighth century but never provided The careful preservation of the Kanakariá’s older apse his reasons.102 Eventually Megaw conceded that his previous mosaics by the rebuilders, and the reuse of marble flooring

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 181

JSAH6902_02.indd 181 4/16/10 4:15 PM and cathedra, indicates that this church was renovated im- would seem, therefore, that they, too, were rebuilt not very mediately after it was destroyed. Megaw argued that an long after the initial Arab raids, most likely in the early eighth inscription in the Kanakariá’s southwesternmost pier com- century. Their masonry is similar to the late-seventh-century memorated a renovation that took place before the ninth square-pier basilica Agios Epiphanios at Salamis-Constantia, century.106 Its barrel-vaulted phase dates to the decades after incorporating ashlars of various sizes, with pieces of tile and the first Arab raids in 649 and 650, as argued by Dikigoro- ceramics in the mortar.111 The blind arches were fully inte- poulos and Papageorghiou.107 grated with their aisle vaults. These five eighth-century The Panagia Afentrika (Sykhada) most resembles the barrel-vaulted churches can now be placed within our gen- Kanakariá and was either built at the same time or slightly eral knowledge of Byzantine history. later. The similarities include a vaulted narthex supported by blind arcades and transverse arches in the aisles that spring from corbels jutting from the wall rather than imposts rest- Historical Context ing on engaged piers. The marble ambo, synthronon, and In 649 the Arab general Mu‘a¯wiya received permission from chancel screens were reutilized from the original basilica, the Caliph Uthman to conquer Cyprus. Sailing from Egypt indicating continuity of liturgical practice from the previous and Syria, two Muslim fleets converged at Acre forming an phase. However, unlike the Kanakariá, the builders designed armada of 1,700 ships.112 They then proceeded to the Cy- the nave arcades with wide arches that did not follow the priot capital, Salamis-Constantia.113 After the conflict, a intercolumniation of the previous building. Its large piers truce was declared between the Byzantine governor (archon) would carry both the aisle and nave vaults, without additional and Mu‘a¯wiya. Under the terms of the treaty, Cyprus would supports, such as an adjacent aisle arcade. Constructed of become a neutral state, paying an equal amount of taxes to rubble masonry and having only three bays, the Panagia Damascus and Constantinople. 114 Later, in 686, Emperor Afentrika resembled the nearby church of Agia Varvara. Both Justinian II and Caliph ‘Abd al-Malik reaffirmed the were probably built within years of each other, though there treaty.115 This unique arrangement lasted until 965, when are subtle differences. the Byzantine Empire reconquered the island.116 Over this At Agia Varvara engaged piers carry the aisle ribs, and three-hundred-year period, both the Arab and Byzantine its narthex was not rebuilt. Fortunately, Agia Varvara con- government maintained a presence on the island, but tains a fresco whose unique style can be dated. Its wheel- could not garrison soldiers there. Without the ability to interlace pattern is rare among Early Christian and Byzantine enforce laws, the outside governments were ineffectual. monuments and more common among eighth-century struc- Local government was empowered. As a result, the island tures in Umayyad Palestine. For example, a floor mosaic in developed along a different trajectory than other Byzan- the Panagia at Madaba (in present-day Jordan) has the same tine provinces.117 interlace pattern, with an inscription dated to 767.108 The The initial raids in 649 and 650 devastated many Early same pattern is also exhibited in fresco at Ag˘aç Kilisi (Ihlara) Christian basilicas, and those that survived were probably in Cappadocia. Nichole Thierry has persuasively dated Ag˘aç destroyed by earthquakes a few years later.118 In subsequent Kilisi to the late eighth or early ninth century based on the decades the policies of the Byzantine Empire exacerbated adjacent Sassanid simorgh pattern.109 Because the wheel- the decay of infrastructure. For example, in 691 Justinian II interlace pattern is unique, associated with Eastern designs, reneged on his treaty with the Caliphate. He devised a plan and aniconic in nature, its appearance at Agia Varvara places to remove the archbishop of Salamis-Constantia and a large the church in the eighth century.110 segment of the Cypriot population from Cyprus and resettle The Afentrika churches were probably the final Karpas them on the Cyzicus peninsula, across the Sea of Marmara churches to be refitted with barrel vaults. Unlike Agia from Constantinople.119 A cathedral for the Cypriots was Varvara, they were much larger and built with carefully cut established at the city of Artake, which the emperor re- ashlar blocks. The high quality of their masonry and their named Nova Justinianoupolis.120 For the next seventeen well-integrated blind arcades show refinement that the other years the Cypriot archbishop and the Byzantine Empire churches lack. Nevertheless, their later restructuring closely abandoned Cyprus to the Arabs, who apparently plundered followed their original ground plans and fully incorporated its monuments. The tenth-century Arab chronicler al- the earlier apses, walls, and flooring; this indicates that those Muqaddasi recorded that around the year 707, “eighteen elements were intact when rebuilding began. The conscious mule-loads of gold and silver” from Cyprus were used to pay reuse of the synthronon and original side passages in the for the construction of the Umayyad Great Mosque of Da- bema implies that liturgical practice remained the same. It mascus.121

182 jsah / 69:2, june 2010

JSAH6902_02.indd 182 4/16/10 4:15 PM After the deposition of Justinian II in the early eighth eighth centuries on the island. And yet these basilicas mark century, the Cypriot archbishop returned to his original epis- a clear shift away from Early Christian architecture: vaulting copal seat.122 Since the island’s capital was 25 miles (40 km) replaced wood roofs, square-pier arcades supplanted previ- from Lythrankomi and 30 miles (48 km) from Agia Varvara, ous colonnades, local stone replaced imported materials, there is reason associate the renovation of the Karpas churches mosaics gave way to frescos, and narthexes were eventually with the rebuilding of Salamis-Constantia that occurred at omitted. As a result, the light and airy interiors of Early this time. The Afentrika churches and the vaulted cathedral Christian basilicas were transformed into dark and fortified of Agios Epiphanios exhibit close similarities, such as ma- sanctuaries. These elements would characterize Cypriot sonry style, the use of pilasters supporting transverse arches, church design throughout the subsequent centuries. and tripartite sanctuary connected by apse passageways.123 As Robert Ousterhout has observed, “the adaption or These developments reflect an economic recovery in the modification of existing built forms provided a significant northeastern part of the island at a time when there was a impetus for the development of new building types and new flourishing Muslim presence on the southwestern coast. planning arrangements” in Middle Byzantine architecture.129 Islamic coins, inscriptions, and tombstones dating This generalization can be applied to the Karpas barrel- throughout the eighth and into the ninth century have been vaulted churches. Arab raids in the mid-seventh century led found at Paphos. One such tombstone has a Kufic inscription Cypriots to experiment with alternatives to colonnades and mentioning the hak¯ım (governor) of the island, dating to the wood roofing. Similar developments occurred many decades year 780.124 Scholars once believed that “the island was then later in western Europe. divided into two zones, the Arabs controlling the western It is tempting to propose that the Karpas churches part (around Paphos)” with the Byzantine Empire ruling the served as prototypes for Romanesque builders of France eastern part (Salamis-Constantia).125 They proposed a “line and Spain. When the German bishop Willibald visited Cy- separating the two communities and the possible existence prus in 723 he recounted, “Those Cypriots dwell between even of two different types of coins, Byzantine for the Byz- the Greeks and the Saracens, and were disarmed, because a antines and Islamic for the Moslems.”126 Since the majority great peace and friendship was then in force between the of Islamic artifacts had been found in Paphos, this theory Saracens and the Greeks there.”130 This passage implies gained momentum in the 1990s. However, in 2004 French that eighth-century Cyprus served as a bottleneck where excavators published their catalogue of coins discovered dur- Latins, Arabs, and Greeks could interact and exchange ideas. ing excavations at Salamis-Constantia between 1964 and However the words of one lone pilgrim are not enough to 1974. They record twenty-three Umayyad coins, but only establish a solid connection between the Cyprus and West- three Byzantine coins of the eighth century.127 While these ern Europe. Perhaps it is sufficient just to acknowledge how numbers of coins are small and provide only circumstantial different groups of builders arrived at similar solutions based evidence, they clearly indicate that Islamic influence per- on shared principles. The Karpas and Romanesque basilicas sisted throughout the island into the ninth century.128 Even arise from a common type—the Early Christian basilica, and with the Arabic presence the Church of Cyprus apparently were guided by a common cultural force—orthodox Chris- continued to thrive. tianity. In recognizing these common factors perhaps histo- rians can be less rigid as they differentiate between Eastern and Western architecture development. Conclusion The barrel-vaulted churches testify to the resiliency of the Cypriot Church after the initial Arab raids. While these structures exhibit a tension between tradition and innova- Notes This material was analyzed in the Republic of Cyprus from 2005 to 2007. tion, their unified typology sets the island apart from other Generous support came from the Samuel H. Kress Foundation Dissertation Byzantine provinces. The conscious recycling of older litur- Fellowship of the Society of Architectural Historians, Fulbright Fellowship, gical furnishings, floors, and apses, coupled with mainte- Cyprus American Archaeological Research Institute (CAARI), and Medieval nance of the basilica form, attest to the importance of Academy of America. I sincerely thank W. Eugene Kleinbauer and Diane maintaining older materials in the renovation process. In- Reilly of Indiana University, Alexis M. Clark of Duke University, and Pro- stead of constructing on virgin ground, builders made an fessor Marcus Rautman of the University of Missouri for reading and editing initial portions of this study. I also acknowledge assistance provided by the effort to reutilize established sacred spaces and preserve their Director of the Cyprus Department of Antiquities, Dr. Pavlos Flourentzos, cultic associations. Moreover, the basic layout indicates that and the Curator of Monuments, Dr. Marina Solomidou-Ieronymidou, for little changed in liturgical practice between the seventh and access to photography and fieldwork archives. Additional assistance was

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 183

JSAH6902_02.indd 183 4/16/10 4:15 PM provided by Dr. Thomas Davis and Vathoulla Moustoukki of CAARI. Pre- To´mou tiV epethri´daV tiV etairei´aV buzantinw´n spoudw´n 49, 185–236. liminary findings of this work were presented at the Byzantine Studies Con- 9. Seven Early Christian churches were rebuilt as square-pier basilicas with ference (Toronto, 2007) and the 4th International Cyprological Congress wood roofs: Agios Mamas (Morphou), Panagia Angeloktisti (Kiti), Agios (Nicosia, 2008), where I received significant suggestions from the audience. Heracleidos (Tamassos-Polikitico), Agios Spyridon (Tremithous), the 1. Hans Kubach, (New York: Abrams, 1975), 105, Southwest basilica at Amathous, and Agios Epiphanios (Salamis-Constantia). 128; Kenneth J. Conant, Caolingian and Romanesque Architecture 800– Regarding Agios Spyridon see Athanasios Papageorghiou, “E´ρευνα εις τον 1200, 4th ed. (New Haven: Yale University Press, 1978), 88–93; Roger Stal- ναw´ν του αγi´ου Σπυρi´δωνος εν Τρεμετουσιa´,” Kupriakai´ Âpoudai´ 30 ley, Early Medieval Architecture (Oxford: Oxford University Press, 1999), (1966), 17–33; for Agios Mamas see A. H. S. Megaw “Archaeology in 131–34; Edson Armi, Design and Construction in Romanesque Architecture Cyprus, 1958,” Archaeological Reports 5 (1958), 34; see also Andreas Dikigo- (New York: Cambridge University Press, 2004). ropoulos, “Cyprus ‘betwixt Greeks and Saracens,’ A.D. 647–965,” (DPhil 2. Richard Krautheimer, Early Christian and Byzantine Architecture, 4th ed. diss., Oxford University, 1961), 85–87. (New Haven: Yale University Press 1986), 202–3; Andre Grabar, “Christian 10. Polybius of Rhinocorura Vita Epiphanii (Patrologiae Cursus Completus, Architecture, East and West,” Archaeology 2 (1949), 95–108; Robert Oust- Series Graeca, 41.85). Carpasia was originally a sea port. This city was first erhout, “The Architecture of Iconoclasm” in L. Brubaker and J. Haldon, recorded by Strabo and seems to have flourished through the end of Late Byzantium in the Iconoclast Era (Aldershot: Ashgate, 2001), 6; Thomas Antiquity: Geographica 14.6.3; A. H. S. Megaw and Joan du Plat Taylor, Mathews, The Early Churches of Constantinople: Architecture and Liturgy (Uni- “Excavations at Ayios Philon, The Ancient Carpasia” Report of the Depart- versity Park: Pennsylvania State University Press, 1971), 155–79. ment of Antiquities Cyprus (1981), 209–50. 3. It is rather ironic that Hagia Sophia remained unreplicated until the advent 11. Papageorghiou, “Architecture paléochrétienne,” 318–19; Megaw and J. of Ottoman control over Constantinople; of course, Justinianic architecture du Plat Taylor, “Excavations at Ayios Philon,” 209–50; A. H. S. Megaw and was replicated in the west, such as the Palatine Chapel (Aachen) and San Marco E. J. W. Hawkins, The Church of the Panagia Kanakariá at Lythrankomi in (Venice); Robert Ousterhout, “The Holy Space: Architecture and the Liturgy,” Cyprus (Washington, D. C.: Dumbarton Oaks, 1977). in Linda Safran’s Heaven on Earth (University Park: Pennsylvania State Uni- 12. David Hogarth, Devia Cypria (Oxford: Horace Hart, 1889), 86. versity Press, 1998), 82; Master Builders of Byzantium (Philadelphia: University 13. Hogarth was not the first explorer to visit the site. The cartographer of Pennsylvania Museum of Archaeology and Anthropology, 2008), 16. Richard Pococke passed the ruins in 1837 and reported that locals called the 4. Yoram Tsafrir, “Procopius and the Nea Church in Jerusalem,” Antiquité place “Asphronisy”; see C. D. Cobham, Excerpta Cypria (Cambridge: Cam- Tardive 8 (2001), 149–64; George Forsyth, “The Monastery of St. Catherine bridge University Press, 1908), 258. The British diplomat Alexander Drum- at Mount Sinai,” Dumbarton Oaks Papers 22 (1968), 3–19. mond wrote twenty-five years later: “About two miles eastward are fine ruins 5. Two other churches located on the southern coast may also belong to the of a village, which they call Athendrae, though I cannot find it in any map I same group: the Panagia Limeniotissa (Paphos) and the Sarayia chapel have seen: however, in many circumstances it answers the description of the (Episkopi). This article focuses primarily on the Karpas churches, since ancient Carpasia, built by Pygmalion . . .”; Letter dated 13 Nov. 1750, in these latter two churches are pending publication by their excavator, Atha- David Martin, English Texts: Frankish and Turkish Periods (Albany, New York: nasios Papageorghiou. Institute of Cypriot Studies, University of New York, 1998), 98. Hogarth, 6. Archaeological excavations in are currently illegal, and Devia Cypria, 85; Papageorghiou, “Cities and Countryside,” 40, fig. 9. Dio- so the material presented here was obtained through past excavation reports, dorus, Bibliotheca historica 20.47. visual analysis, and archival photographs in the Cyprus Department of 14. Athanasios Papageorghiou, “Cities and Countryside at the End of Antiquities, Nicosia. All measurement provided in this article are given in Antiquity and the Beginning of the Middle Ages in Cyprus,” in The Sweet approximations. Accompanying illustrations are based on the author’s sketch Land of Cyprus (Nicosia, 1993), 40, fig. 9; J. du Plat Taylor and E. Dray, plans and measurements in the field. These plans were compared with the “Tsambres and Aphendrika,” Report of the Department of Antiquities Cyprus available sketch plans published by Camille Enlart and A. H. S. Megaw and (1937–39), 24–123. modified accordingly; Enlart, L’Art gothique et la Renaissance en Chypre I 15. Ibid, 86; Papageorghiou, “Cities and Countryside,” 40, fig. 9. The cha- (Paris: Leroux, 1899), 397–98; Megaw, “Three Vaulted Basilicas in Cyprus,” pel of Agios Georgios, while related to the Afentrika barrel-vaulted churches Journal of Hellenic Studies 66 (1946), 48–56. in both date and construction, will be treated in a separate publication being 7. In 1985 Athanasios Papageorghiou counted fifty-six Early Christian churches, prepared by the author. since then nine more have been found: Garrison’s Camp (Paphos), three at 16. Hogarth was unaware of this shrine but described a similar one at Rizo- Kalavasos-Kopetra, Agioi Pente (Geroskipou), two basilicas at Polis, the Pan- karpaso five miles away: “The tombs are all empty, and many, to judge from agia Limeniotissa (Kourion), and Agios Georgios (Nicosia): “L’Architecture the crosses cut on walls and roof, have been re-used in Christian times: one paléochrétienne de Chypre,” Corsi di cultura sull’arte ravennate e bizantina 22 is lined with plaster”; Devia Cypria, 92. (1985), 300; Filippo Guidice et al. “Paphos, Garrison’s Camp,” Report of the 17. Megaw, “Three Vaulted Basilicas in Cyprus,” 48–56. Department of Antiquities Cyprus (2004), 271–315; Marcus Rautman, A Cypriot 18. Georgios Soteriou, Ta Buzantina´ Mnh´mata thV Ku´prou (Athens: Ath- Village of Late Antiquity: Kalavasos-Kopetra in the Vasilikos Valley (Portsmouth, ens Academy, 1935), plate 15a; Megaw, “Three Vaulted Basilicas,” 50–52. Rhode Island, 2003); Demetrios Michaelides, “‘Ayioi Pente’ at Yeroskipou, a 19. Papageorghiou, “Architecture paléochrétienne,” 299–324; Megaw and New Early Christian Site in Cyprus,” Musiva & Sectilia 1 (2004), 185–98. du Plat Taylor, “Ayios Philon,” 221; Megaw and Hawkins, Panagia Kana- 8. Previous scholars have often used the appellation “Condominium Period” kariá, 24–30. for the time between 650 and 965 on Cyprus. This is a misnomer, since 20. The late-fourth-century cathedral of Agios Epiphanios at Salamis- neither the Byzantine Empire or the Arab Caliphate governed the island in Constantia is the earliest church on the island to have these passages. It is any direct or joint fashion; R. J. Jenkins, “Cyprus between Byzantium and a feature found in other Early Christian churches such as Agios Philon Islam, A.D. 688–965,” Studies Presented to David Moore Robinson on His Sev- (Carpasia). entieth Birthday (St. Louis, 1953), 1006–1014; 1994–1998. Costas Kyrris, 21. This is noted by Megaw, “Three Vaulted Basilicas,” 50. The large fifth- “Cyprus, Byzantium and the Arabs from the 7th to the early 8th centuries,” century church of the Campanopetra (Salamis-Constantia) was built in a

184 jsah / 69:2, june 2010

JSAH6902_02.indd 184 4/16/10 4:15 PM similar fashion: G. Roux, La basilique de la Campanopétra (Paris: De Boc- Chiat and K. Reyerson (St. Cloud, Minn.: North Star Press, 1988), 67–74. card1998). Portions of the walls at the 5th century church Qal’at Si’man 36. Andreas and Judith Stylianou, The History of the Cartography of Cyprus (Syria) and the 6th c. fortress at Haïda in North Africa also have this type of (Nicosia: Cyprus Research Centre, 1980), 57–58. The site of Agia Varvara construction; Krautheimer, Early Christian and Byzantine Architecture, 260. is closer to the prehistoric ruins of Nitoviklia on the coast than to modern 22. Papageorghiou called this chamber a chapelle des martyres; Koroveia. A Cypriot city called “Koroneia” (Κορw´νεια) was mentioned by “Παλαιοχριστιανικh´ και βυζαντινh´ . . . 1964,” 94–96; Vassos Kara- the sixth-century chronicler Stephanus Byzantinus in his Ethnica. georghis, “Chroniques des fouilles en 1964,” Bulletin de correspondance hellé- 37. The road leading to the church is a goat path used by the local Turkish nique 89 (1965), 300. Regarding the function of these chapels see Georgiana farmer; occasionally ruts made by tractor tires uncover examples of coins, Babi´c’s Les Chapelles annexes des églises byzantines (Paris: CNRS, Klincksieck, amphorae, and ancient glass. Hogarth provided the earliest description of 1969) and Slobodan Curcic’s “Architectural Significance of Subsidiary Cha- the church: Devia Cypria, 70. pels in Middle Byzantine Churches,” JSAH 36, no. 2 (1977), 94–110. 38. H. Gregoire, “Saint Demetrianos, évêque de Chytri,” Byzantinische 23. Athanasios Papageorghiou, “Η παλαιοχριστιανικh´ και βυζαντινh´ Zeitschrift 16 (1907), 209–12. αρχαιολογi´α και τe´χνη εν Κu´πρω κατa´ το 1964,” Apo´stoloV Barna´baV 39. The church was first recorded by Kitchener on his 1885 map, marked 26 (1965), 94–96. as “Aphendrika (ruins)” northwest of Sykhada. He also marked another 24. A similar development occurred at Andaval (Cappadocia) after the church a few yards away called “Agia Yeorgios (ruins),” which I have been church there was rebuilt as a barrel-vaulted basilica. Like the Panagia Afen- unable to locate; A Trigonometrical Survey of the Island of Cyprus (London, trika, this Cappadocian church still has its early Christian columns in situ 1885), Enlart, L’Art gothique, 399. between the later square pilasters: Nicole Thierry, La Cappadoce: de 40. This work consisted of removing loose stones and brush from the inte- l’Antiquité au Moyen-âge (Turnhout: Brepols, 2002), 78, fig. 46. rior space, and was not a scientific excavation: Megaw, “Three Vaulted 25. Papageorghiou, “Παλαιοχριστιανικh´ και βυζαντινh´ . . . 1964,” 94–96. Basilicas,” 55, fig. 13. 26. Megaw, “Three Vaulted Basilicas,” 51–52, fig. 7. Other scholars omit 41. The “hanging” corbel used to support a transverse arch is common in any discussion of this detail; Camille Enlart, L’Art gothique (Paris, 1899), the Binbirkilise region of Anatolia, perhaps as early as the sixth century; R. 397–98; Papageorghiou, “Η παλαιοχριστιανικh´ και βυζαντινh´ . . . 1964,” Lienhardt, “The Great Basilica, Church no. 1, at Bin Bir Kilise in Anatolia,” and “Η παλαιοχριστιανικh´ και βυζαντινh´ αρχαιολογi´α και τe´χνη εν JSAH 24 no. 4 (Dec. 1965), 300–303; W. Ramsay and G. Bell, The Thousand Κu´πρω κατa´ το . . . 1965,” Apo´stoloV Barna´baV 27 (1966), 221; Georger and One Churches (London: Hodder and Stoughton, 1909), 41–51. Krau- Jeffery, A Description of the Historic Monuments of Cyprus (London, 1918, rpt. theimer has argued the architectural technology and design of the Binbirkilise London: Zenon, 1983), 57–58; G. Soteriou, “Τα παλαιοχριστιανικa´ και region was derived from Syria; however, I have not found any examples of βυζαντινa´ μνημεi´α της Κu´πρου,” Praktika´ thV Akadhmi´aV Aqhnw´n “hanging” corbels in Syria; Early Christian and Byzantine Architecture, 162. (1931), 482; R. Gunnis, Historic Cyprus (London: Methuen, 1936), 167; Many barrel vaults at Binbirkilise have neither piers nor transverse arch Dikigoropoulos, “Cyprus ‘betwixt Greeks and Saracens,’” 230. support. The sixth-century basilica at Tolmeita in North Africa had a barrel- 27. Papageorghiou, “Η παλαιοχριστιανικh´ και βυζαντινh´ . . . 1964,” vaulted nave with transverse arches supported by wall corbels; however, the 94–96. corbels rest directly above the aisle piers, so they are not really hanging, but 28. A large quarry called Phylakes is located east of Afentrika and is still used inset within the wall supported by piers (ibid., 275). as such today; Hogarth, Devia Cypria, 85. 42. On the others side of the church, Enlart mentioned an external staircase 29. Papageorghiou, “Η παλαιοχριστιανικh´ και βυζαντινh´ . . . 1965,” 221. attached to the north wall, which does not survive. He explained that this 30. Hogarth called this church Agios Demetrios; its original name is uncer- led to the flat roof of the church. This suggestion is rather odd, since the tain; Devia Cypria, 86. church was clearly barrel vaulted, as Enlart noted. Perhaps this stairway led 31. Dikigoropoulos, “Cyprus ‘betwixt Greeks and Saracens,’” 179–81. to a later bell tower, or like so many converted churches on Cyprus, a mina- 32. Enlart, L’Art gothique, 398. ret; L’Art gothique, 399. Hogarth wrote, “That the Carpas was entirely 33. The first phase masonry has large, finely carved ashlars (measuring 24 × unmolested by the Turks is disproved by the existence of so many ruined 12 inches) joined together by smooth mortar; the second phase has smaller, churches in its area, and of Greek-speaking Mahometan villages like Galino- coarser stones (12 x 6 inches) joined by mortar mixed with tile and ceramics porni and Koroveia, sure traces of a forced conversion of the conquered”; pieces. Devia Cypria, 59. The Panagia Afentrika at Sykhada lies two miles between 34. Few scholars of Byzantine architectural history have mentioned the neces- and Koroveia. sity of side aisles or ambulatories in supporting vaults. In contrast, Stalley 43. The figure’s head is elongated so that its height is twice the length of its describes the significance of this technological understanding in the earliest width. Human representations in this style brings to mind the seventh- or Romanesque vaulted basilicas; Early Medieval Architecture, 133–34. Perhaps eighth-century figures depicted in the Red Monastery in Egypt; Elizabeth the structural necessity of aisles is demonstrated at the early-fourth-century Bolman, “Late Antique Aesthetics, Chromophobia, and the Red Monastery, Basilica Nova in Rome, where aisle vaults served as “proto-flying but- Sohag, Egypt,” Eastern Christian Art 3 (2006), 1–24. tresses”; Robert Mark, Architectural Technology up to the Scientific Revolution 44. Megaw and Hawkins, Panagia Kanakariá; Jakov Smirnov, “Хрйстанскiя (Cambridge: MIT Press, 1993), 87. Мозайкй Кйпра,” Vizantiiskii vremennik (1897), 1–93; Marina Sacopoulo, 35. The reliance on round arches in the arcades, vaulting, and transverse La Theotokos à la mandorle de Lythrankomi (Paris: Maisonneuve & Larose, ribs necessarily depends on the square, or more precisely, the half-square. 1975). Each half-square unit was repeated for each bay, simply because each bay 45. This spacing was followed during the second phase reconstruction of supported a network of round arches. The measurements of each bay Agios Spyridon (Tremithou); Papageorghiou, “Αγi´ου Σπυρi´δωνος,” 17–33. derived from the width the half-circle (i.e., barrel vault) overhead; François 46. Enlart, L’Art gothique, I.402; Dikigoropoulos, “Cyprus ‘betwixt Greeks Bucher, “Medieval Architectural Design Methods,” Gesta 11, no. 2 (1972), and Saracens,’” 189; Athanasios Papageorghiou, “Η παλαιοχριστιανικh´ και 37–51; W. Eugene Kleinbauer, “Pre-Carolingian Concepts of Architectural βυζαντινh´ αρχαιολογi´α και τe´χνη εν Κu´πρω,” Apo´stoloV Bapna´baV 27 Planning,” in The Medieval Mediterranean: Cross-Cultural Contacts, ed. M. (1966), 21, and vol. 29 (1968), 24.

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 185

JSAH6902_02.indd 185 4/16/10 4:15 PM 47. Megaw’s reconstruction of the Kanakariá’s second phase is based on its 61. Tassos Papacostas, “Byzantine Cyprus: The Testimony of Its Churches, similarities with the wood-roof phase of the Agios Epiphanios (Salamis- 650–1200,” (DPhil diss., Oxford University, 1999), 209–10. This echoes Constantia), such as the square-pier arcades and especially, the “walled Dikigoropoulos’ observations that “one would expect to find some evidence bema.” But as I argue throughout this article, all the barrel-vaulted churches that the abandonment of some 58%, if not more, of the Island’s rural settle- have similarities with both the second and third phases of Agios Epiphanios. ments took place as a result of sacking or destruction by raiders . . . such The use of a “walled bema” does not exclude the Kanakariá from the barrel- evidence is not, however, forthcoming”; “Cyprus ‘betwixt Greeks and Sara- vaulted group as Megaw seems to infer; Megaw and Hawkins, Panagia Kana- cens,’” 266. kariá, 30–33, plate F. 62. Vincenzo Ruggieri, Byzantine religious architecture (582–867) (Rome: 48. Few scholars have accepted Megaw’s reconstruction; S. C´urcˇic´’s review Pontificale Institutum Studiorum Orientalium, 1991), 140–50. in Speculum, 55, no. 4 (1980), 812–16; Athanasios Papageorghiou, “Constan- 63. “Byzantine Architecture on Cyprus,” Medieval Cyprus (Princeton: Princ- tinopolitan Influence on the Middle Byzantine Architecture of Cyprus,” eton University Press, 1999), 73. Jahrbuch der Österreichischen Byzantinistik 32, no. 4 (1982), 469–70. 64. Middle Byzantine Architecture on Cyprus: Provincial or Regional? (Nicosia: 49. There are rare examples of Early Christian churches with crossing tow- Bank of Cyprus Cultural Foundation, 2000), 10. ers known in Asia Minor, such as Meriamlik; Stephen Hill, The Early Byz- 65. Megaw, Kourion: Excavations in the Episcopal Precinct (Washington D. C: antine Churches of Cilicia and Isauria (Aldershot: Ashgate, 1996). However Dumbarton Oaks, 2007), 174–76. little is known regarding their roofing and they appear much earlier than the 66. E. Guidoboni, A. Comastri and G. Traina, Catalogue of Ancient Earth- rebuilt church of the Panagia Kanakariá. quakes in the Mediterranean Area up to the 10th Century (Bologna: Istituto 50. Megaw dated the current bema dome to the twelfth century; Panagia Nazionale di Geofisica, 1994). Kanakariá, 34–35. However, the last archaeologist to work on the church, 67. No archaeological method has been devised to distinguish between a Athanasios Papageorghiou, believes it must date earlier, perhaps to the tenth wall that has fallen due to an earthquake and one that collapses due to human century (personal communication, 24 Nov. 2006). demolition or simply age. The latest “earthquake-linked coins” at Kourion 51. Multiple courses of voussoirs are known in the earliest examples of Roman are dated to 687. This is a rather conspicuous timeframe for the coin record barrel vaults; Jean-Pierre Adam, Roman Building: Materials and Techniques, to end. According to the ninth-century Byzantine historian Theophanes, in trans. A. Mathews (Bloomington: Indiana University Press, 1994), 158–59. the year 691 the Emperor Justinian II resettled “the population of the island 52. C. A. Stewart, “The Barrel-Vaulted Basilicas of Cyprus,” Proceedings of of Cyprus” on the Cyzicus peninsula near Constantinople; C. Mango and the 4th International Cyprological Congress, Lefkosia, 29 April–3 May 2008 R. Scott, The Chronicle of Theophanes Confessor (Oxford: Oxford University (Nicosia, 2009). Press, 1997), 509, notated as Chronographia below. It is possible that the 53. Cyril Mango, Byzantine Architecture (New York: Harry N. Abrams, bishop of Kourion and his flock abandoned their church at this time. This 1978), 98, plate 140. is further discussed in the next section. 54. Richard Kieckhefer, Theology in Stone: Church Architecture from Byzan- 68. While an earthquake was reported in Palestine in 640, Cyprus belongs tium to Berkeley (New York: Oxford University Press, 2004), 75–76. to a separate fault line and was probably not affected. Only one historical 55. There is no evidence for the use of ciboria in Cyprus between the sev- earthquake was massive enough that it could have impacted Cyprus: in 796 enth and tenth centuries. a devastating tremor in Constantinople, Alexandria, and Crete was recorded; 56. Megaw “Three Vaulted Basilicas,” 56. N. Ambraseys, “The Seismic History of Cyprus,” Revue de l’Union Interna- 57. Costas Kyrris, “Cyprus, Byzantium and the Arabs,”185–236; Costas tionale de Secours 3 (1965), 25–48; Guidoboni et al., Catalogue of Ancient Kyrris, “The Nature of Arab-Byzantine Relations in Cyprus from the Mid- Earthquakes. But as I discuss below, the barrel-vaulted churches seem to date dle of the 7th to the Middle of the 10th Century A.D.,” Graeco-Arabica 1 to the earlier part of that century. For the historical sources for earthquakes (1982), 144–75; R. J. Jenkins, “Cyprus between Byzantium and Islam,” damage regarding Hagia Sophia, see W. Eugene Kleinbauer’s Saint Sophia 1006–1014; M. T. Mansouri, Chypre dans les sources arabes médiévales (Nicosia, at Constantinople: Singulariter in mundo (Dublin, N.H.: William L. Bauhan, 2001). Robert Schick, The Christian Communities of Palestine from Byzantine 1999), 31, 34. to Islamic Rule (Princeton, N.J.: Darwin Press, 1995); Bat Ye’or, The Decline 69. A general critique of Ruggieri’s earthquake theory was provided by of Eastern Christianity under Islam: from Jihad to Dhimmitude (Madison, N.J.: Robert Ousterhout; Speculum 68.2 (1993), 559–61. Fairleigh Dickinson University Press, 1996). 70. Experiments conducted by the Getty Institute and Macedonian scholars 58. Dikigoropoulos also wrote: “Various factors may have contributed to have tested the seismic strength of Middle Byzantine churches in a series of the abandonment of the great basilicas. Arab action in the seventies of the experiments. While central-planned structures are remarkably resilient to seventh century, when an Arab garrison may have been established on the seismic forces, the dome was the single most unstable feature. The swaying Island is not known. But it is not impossible that it was adverse to the main- of the dome led to walls cracking. Therefore we should question whether tenance of a large basilica . . . by a shrinking population. Earthquakes, which Byzantine architects viewed domed structures are “earthquake proof ”; P. are recorded in the region of Palestine during this period . . . may have struck Gavrilovi´c, “Seismic Strengthening and Repair of Byzantine Churches,” Constantia causing irreparable damage to the great basilica. Of this however, Journal of Earthquake Engineering 3, no. 2 (1999), 199–235; Predrag we have no record”; Dikigoropoulos, “Cyprus ‘betwixt Greeks and Sara- Gavrilovi´c et al., Conservation and Seismic Strengthening of Byzantine Churches cens,’” 182 note 2, 186 note 1. in Macedonia (Los Angeles: Getty Conservation Institute, 2004). 59. Papageorghiou recognized two different stages of rebuilding after the 71. “Παλαιοχριστιανικh´ και βυζαντινh´ . . . 1965,” 221; Athanasios Papa- Early Christian church of Agios Sypridonos was destroyed. He proposed georghiou, “Constantinopolitan Influence on the Middle Byzantine Archi- that the destruction was caused by the Arabs, but not the initial raids of 649 tecture of Cyprus,” Jahrbuch der Österreichischen Byzantinistik 32, no. 4 and 650; “Αγ i´ου Σπυρi´δωνος,” 17–33. (1982), 468–70; see also Papageorghiou’s entries for the respective churches 60. Theodore of Paphos, La Légende de St. Spyridon évêque de Trimithonte, in the Mega´lh Kupriakh´ egkuklopai´deia (Nicosia, 1985). trans. P. Van den Ven, (Louvain, 1953), 89; Acta Sanctorum, Maii VI, 72. A. Michel and M. Piccirillo, Les églises d’époques byzantine et Umayyade de 684–85. Jordanie (Turnhout: Brepots, 2001), 169. The rare examples of barrel-

186 jsah / 69:2, june 2010

JSAH6902_02.indd 186 4/16/10 4:15 PM vaulted churches elsewhere, such as Agios Georgios in Astypalaia (Greece) II resettled Cypriots on the Cyzicus peninsula, causing the Arabs to raid the and Belovo (Bulgaria), are even more dissimilar in their design and form; A. island in retaliation; A. Dikigoropoulos,“The Political Status of Cyprus A.D. H. S. Megaw, “Byzantine Architecture and Decoration in Cyprus: Metro- 648–965,” Report of the Department of Antiquities Cyprus (1940–1948), 101 politan or Provinical?” Dumbarton Oaks Papers 28 (1974),” 78 note 81; see note 42. (4) In 743, Walid-Ibn-Yezid raided Cyprus and deported scores of also Krautheimer, Early Christian and Byzantine Architecture, 162, 275. Cypriots to Syria; Agapius of Manbij, Universal History (Kita¯b al-‘Unwa¯n) 73. The closest comparanda to the Cyprus barrel-vaulted basilicas can be ed. and trans. A. Vasiliev, Patrologie Orientalis 8 (1912), 511–12. (5) In 773, found at Maden S¸ehri–Deg˘le, also known as Binbir Kilisi. This region con- Arabs capture Laherfavos, a prosopon (Byzantine representative) on Cyprus; tains several churches, like the so-called “Great Basilica” which were reno- Theophanes, Chronographia, 446. (6) In 805/6, Arab raids; Theophanes, vated with double barrel vaults (see note 41 above). Also the scale and form Chronographia, 481. (7) In 807 Humaid b. Ma’yüf, governor of the coast of of the seventh-century barrel-vaulted church of Küçük Tavs¸an Adasi (off the Syria, carried out the raids on Rhodes and Cyprus; Theophanes, Chrono- coast of Gölköy) in Turkey is the closest to the Karpas churches; V. Ruggieri, graphia, 483. (8) In 912, four-month Arab raid, led by Damianus, Emirate “La chiesa di Küçük Tavs¸an Adasi nella Caria Bizantina,” Jahrbuch der Öster- of Tarsus; A. Mas’¯udi, Mur¯uj al-Dhahab wa Ma’a¯din al-Jawhar VIII, ed. C. reichischen Byzantinistik 40 (1990), 383–403. These similarities necessitate a Barbier de Meynard and A. Pavet de Courteille (Paris: Impr. Impériale, closer examination. 1861–77), 282–83. 74. Rautman, Cypriot Village, 69–71, 90. 86. Stalley, Early Medieval Architecture, 131–34. 75. Ibid., 275–76. 87. Our best description of the seventh- and eighth-century liturgy comes 76. Chronographia 423. from Maximus the Confessor’s Mystagogia (Migne, Patrologia Graeca 91) and 77. These sites were never fully published as excavation reports; Megaw, the Patriarch Germanus I of Constantinople Historia Ecclesiastica (Paul Mey- “Reflections on Byzantine Paphos,” inKaqhgh ´tria: Essays Presented to Joan endorff, St. Germanus of Constantinople on the Divine Liturgy [Crestwood, Hussey for Her 80th Birthday (Camberley, England: Porphyrogenitus, 1988), New York: St. Vladimir’s Seminary Press, 1984]). 135–36; Dikigoropoulos, “Cyprus ‘betwixt Greeks and Saracens,’” 230; Vas- 88. Thomas F. Mathews, “An Early Roman Chancel Arrangement and sosKarageorghis, “Chroniques des fouilles en 1967,” Bulletin de correspon- its Liturgical Functions,” Rivista di archeologia cristiana 38 (1962), dance hellénique 92, 351; Vassos Karageorghis, “Chroniques des fouilles en 93–94. 1968,” Bulletin de correspondance hellénique 93, 564–66; Athanasios Papa- 89. Robert Taft, “The Frequency of the Eucharist throughout History,” georghiou, “Παλαιοχριστιανικh´ και βυζαντινh´ . . . 1968,” Apo´stoloV Concilium 172 (1982), 13–24. Barna´baV (1969), 82–88; Athanasios Papageorghiou, Iera´ Mhtro´poliV 90. Karl Lehmann, “The Dome of Heaven,” Art Bulletin 27 (1945), 1–27; Pa´jou, Istori´a kai te´cnh (Nicosia: Imprinta, 1996), 6, 55. T. F. Mathews, “Cracks in Lehmann’s ‘Dome of Heaven,’” Art and Architec- 78. The excavation this church, known as the Panagia Limeniotissa at Kou- ture in Byzantium and Armenia (Variorum Collected Studies Series CS 510, rion, has not been published; initial findings were reported inCAARI News 2001), 11–16. The rainbow became an important symbol in Christian art 15 (Dec. 1997), 4–5. denoting the sky above and the spiritual heaven. For example, at Hosios 79. The second inscription read “destroyed by the fire . . . the entire epis- David (fifth century) and Hagia Sophia (eighth century) in Thessaloniki copal palace . . . in the other places . . . of the island . . . other fires . . . the mosaics depicts Christ seated on the rainbow. As such the rainbow is a sym- fervor of John . . . and in great haste . . . rebuilt buildings that were destroyed, bol for heaven—Christ’s throne (Isaiah 66:1, Acts 7:49, Revelations 4:2–3). renovated roofs, they have decorations and have finished the work for the Often ancient writers do not make a distinction between a barrel vault and Glory of the Father of the Son and Holy Spirit the year 13 of the Indiction, a dome; see the discussion in H. Howe, “The Dome of Clement,” Transac- 371st year of the Era of Diocletian”; D. Feissel, “Bulletin épigraphique- tions and Proceedings of the American Philological Association 97 (1966), 265–67. Chypre,” Revue des études grecques 100 (1987), 380–81; J. Des Gagniers and 91. Enlart, L’Art gothique, I: 395–401. T. Tinh, Soloi. Dix campagnes de fouilles (1964–1974) I: La Basilique (Saint- 92. Soteriou, “Τα παλαιοχριστιανικa´ και βυζαντινa´ μνημεi´α της Foy, Canada: Presses Université Laval, 1985); Athanasios Papageorghiou, Κu´ πρου,” Praktika´ thV Akadhmi´aV Aqhnw´n (1931):482; Soteriou, Ta “Μια σu´γχρονη πηγh´ για τις δu´ο πρw´τες αραβικe´ς επιδρομe´ς κατa´ της Buzantina´ Mnh´mata, plate 10a, 11a, & 15a (mislabeled as Agios Philon). Κu´πρου,” Stasinos 9 (1986–88), 167–75. 93. Jeffery, Historic Monuments of Cyprus, 257–58; Gunnis, Historic Cyprus, 80. Munro, A. and H. Tubbs, “Excavations in Cyprus, 1890, Salamis,” Jour- 167. nal of Hellenic Studies 12 (1891), 102–3. 94. “Three Vaulted Basilicas,” 48–56. 81. George Hill, A History of Cyprus (Cambridge: Cambridge University 95. Hill, History of Cyprus, I. 322. Press, 1940), I.254; Dikigoropoulos, “Cyprus ‘betwixt Greeks and Sara- 96. “Three Vaulted Basilicas,” 56. cens,’” 186; Megaw, “Archaeology in Cyprus, 1956,” Archaeological Reports 97. John Hackett, A History of the Orthodox Church of Cyprus (London: 3 (1956), 30. Methuen, 1901); Andreas Dikigoropoulos, “The Church of Cyprus during 82. Roux, La basilique de la Campanopétra, 250. the Period of the Arab Wars,” Greek Orthodox Theological Review 11 (1965– 83. See note 9 above. 66), 237–79. 84. For a list of centrally planned churches of the eighth and ninth centuries, 98. See note 9 above. see Ousterhout, “The Architecture of Iconoclasm,” 3–18. 99. “Cyprus ‘betwixt Greeks and Saracens,’” 193. 85. Recorded raids from both Arab and Byzantine sources are: (1) In 648/49, 100. Theophanes recorded that Emperor Michael sent a grant to refugees invasion of Mu‘a¯wiya; Feissel, “Bulletin épigraphique-Chypre,” 380–81; M. fleeing Palestine; these funds were not for church building; Theophanes, Tabari, Ta’r¯ıkh al-Rusul wa’l Mul¯uk IV, ed. M. Ibra¯h¯ım (Cairo: Dar al Chronographia, 498. Ma’arit, 1972), 258; Theophanes Chronographia, 344; Hippolyte Delahaye, 101. Dikigoropoulos wrote: “Professor Soteriou has also revised his views “Life of St. Therapon,” Acta Sanctorum Maii 6, 684–85). (2) In 653/54, Sec- since then and now agrees to a dating within the period of the Arab wars”; ond Arab invasion led by Abu’l A’war al Sulami; A. Bala¯dhuri, The Origin of and “Mr. Megaw tentatively dated the Aphendrika churches to the period the Islamic State (Kita¯b Fut¯uh al-Bulda¯n) trans. P. Hitti and F. Murgotten after the Byzantine reconquest of 965; he has, however, revised his view since (New York: Columbia University Press, 1924), 209. (3) In 691–98, Justinian then and is now of the opinion that they belong to the period of the Arab

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 187

JSAH6902_02.indd 187 4/16/10 4:15 PM wars. I understand Mr. Megaw is going to comment on the Aphendrika Scientifique de Chypre, 2001), 29. Christides argued that these sources con- churches in a paper he is reading at the Byzantine Congress, at Ochrid, in fuse the initial treaty, whatever it contained, with the later treaty of 686; September, 1961”; “Cyprus ‘betwixt Greeks and Saracens,’” 191 note 1. Image of Cyprus, 31–33. I see no reason to why the later treaty would neces- However, no trace of Megaw’s paper exists, nor have I been able to verify if sarily depart from the earlier arrangement. he attended. Megaw mentioned that his views toward the barrel-vaulted 115. Theophanes, Chronographia, 363. basilicas changed, but provided no reason why; “Byzantine Architecture and 116. John Norwich, Byzantium: The Apogee (New York: Knopf, 1992), 195; Decoration in Cyprus: Metropolitan or Provinical?” Dumbarton Oaks Papers Titos Papamastorakis,“The Bamberg Hanging Reconsidered,” Delti´wn thV 28 (1974), 76 note 80. Cristianike´V Arcaiologike´V Etairei´aV 24, no. 4 (2003), 375–92. 102. Papageorghiou, “Παλαιοχριστιανικh´ και βυζαντινh´ . . . 1965,” 221; 117. Dikigoropoulos, “Political Status,” 94–114; “The Church of Cyprus,” Papageorghiou, “Constantinopolitan Influence,” 468–70; Athanasios Papa- 237–79; D. M. Metcalf, Byzantine Lead Seals from Cyprus (Nicosia: Cyprus georghiou, Mega´lh Kupriakh´ egkuklopai´deia (Nicosia, 1985). Research Centre, 2004), 69–79. 103. Megaw and Hawkins, Panagia Kanakariá, 31 note 130. 118. Gagniers and Tinh, Soloi, 118–23; Feissel, “Bulletin épigraphique,” 104. “‘Betwixt Greeks and Saracens,’” in Acts of the International Archaeo- 380–81; Rautman, Cypriot Village, 90; Megaw, “Three Vaulted Basilicas,” logical Symposium “Cyprus between the Orient and the Occident” September 1985, 56; Papageorghiou, “Μια σu´γχρονη,” 167–75; Papageorghiou, Iera´ ed. V. Karageorghis (Nicosia: Dept. of Antiquities, Cyprus, 1985), 518. Mhtro´poliV Pa´jou, 6, 55. 105. A. Papageorghiou, “The Narthex of the Churches of the Middle Byz- 119. Theophanes, Chronographia, 365. The sea route from Salamis-Constantia antine Period in Cyprus,” in Rayonnement Grec, ed. L. Hadermann-Misguich to Cyzicus is about 800 nautical miles. and G. Raepsaet (Brussels: Editions de l’Université de Bruxelles, 1982), 120. Dikigorpoulos argued that “Nea Justinianoupolis” was actually a 437–48. renaming of the city of Salamis-Constantia, after a renovation campaign of 106. Megaw and Hawkins, Panagia Kanakariá, 148–49. Cyril Mango, who Justinian II; “Cyprus ‘betwixt Greeks and Saracens,’” 31–35. This was transcribed the inscription for Megaw, suggested that the content served a proven incorrect by B. Englezakis; Studies on the History of the Church of funerary function unassociated with repairs; Ruggieri, Byzantine religious Cyprus (London: Ashgate-Variorum, 1995), 63–82; see also Hill, History of architecture, 268 n. 361. Cyprus, 228–29; Dikigoropoulos, “Political Status,” 94–114; Dikigoropou- 107. Dikigoropoulos, “Cyprus ‘betwixt Greeks and Saracens,’” 189–92; Papa- los, “The Church of Cyprus,” 237–79. georghiou, “Η παλαιοχριστιανικh´ και βυζαντινh´ αρχαιολογi´α,” 221. 121. K. A. C. Creswell, Early Muslim Architecture, Umayyads, Early ‘Abba¯sids 108. M. Piccirillo, The Mosaics of Jordan (Amman: American School of Ori- and Tu¯lu¯nids (Oxford: Clarendon Press, 1932), 101. ental Research, 1992), 25–27; L. Di Segni, “The Date of the Church of the 122. This resettlement took place either in the year 707 or 717. Regarding Virgin in Madaba,” Liber annuus (1992), 251–57. This pattern is not com- the return of the Cypriot archbishop from Cyzicus, see the following: mon in late antiquity, but widespread in Umayyad Palestine. It was described Englezakis, Church of Cyprus, 63–82; Kyrris, “Cyprus, Byzantium and the by Balmelle as a “pattern of rows of tangent and intersecting circles, in inter- Arabs,” 185–236. laced and tangentially interloped bands; C. Balmelle et al., Le Décor géométrique 123. Dikigoropoulos, “Cyprus ‘betwixt Greeks and Saracens,’” 179–81. de la mosaïque romaine (Paris, 1985–2002), 1: 368, plate 235. The same pattern 124. Megaw, “Reflections on Byzantine Paphos,” 135–50. A complete sur- is found on the island of Cos in mosaic, but is not dated; S. Pelekanides et al., vey of early Islamic artifacts on Cyprus has not been produced. Several Âu´nthgma ton Palaiocristianikw´n yhjidwtw´n dape´dwn teV Ella´ doV archaeologists have mentioned Arab graves discovered in their excavations, (Thessaloniki: Kentron Vyzantinon Ereunon, 1974), I. no. 250. but subsequently never published them (e.g., see Professor Daszewski’s com- 109. Thierry, Cappadoce, 140–42, plate 54. ments in Megaw, “‘Betwixt Greeks and Saracens,’” 518). Seventeen Arabic 110. For the dating of aniconic frescos: M. Chatzidakis et al., Naxos (Athens: inscriptions have been found in Cyprus dating from the eighth to ninth Melissa, 1989) and Thierry, Cappadoce, 114–42; J. Lafontaine-Dosogne, century; Christides, Image of Cyprus, 113–32; Athanasios Papageorghiou, “Pour une problématique de la peinture d’Église byzantine a l’époque icon- “Les premières incursions arabes à Chypre et leurs conséquences,” Ajie´rwma oclaste,” Dumbarton Oaks Papers (1987) 41: 321–37. Since aniconic art is eiV ton Konstanti´non Spurida´kin (Nicosia: Kypros, 1964), 152–58; Atha- based on Late Antique mosaic and opus sectile, I am not persuaded by the nasios Papageorghiou, Η Παλαιοχριστιανικh´ και βυζαντινh´ αρχαιολογi´α later dating proposed by L. Brubaker and J. Haldon who base their argu- και τe´χνη εν Κu´πρω κατa´ το 1968,”Apo´stoloV Barna´ baV 30 (1969), ments on book illumination; Byzantinum in the Iconoclast Era, 24–28. 82–88; Papageorghiou, “Μια σu´γχρονη,” 167–75. 111. Dikigoropoulos, “Cyprus ‘betwixt Greeks and Saracens,’” 179–81. 125. J. Nesbitt and N. Oikonomides, Catalogue of Byzantine Seals at Dumbar- 112. Theophanes, Chronographia, 344. A few Arabic sources provide more ton Oaks and in the Fogg Museum of Art (Washington D.C.: Dumbarton Oaks, conservative numbers such as 120 or 250 ships; Vassilios Christides, The 1991), 101. Image of Cyprus in the Arabic Sources (Nicosia: Archbishop Makarios III Foun- 126. A. Pitsillides and D. Metcalf, “Islamic and Byzantine Coins in Cyprus dation Press, 2006), 19–18. during the Condominium Centuries,” Epithrei´V tou Ke´ntrou 113. Dionysius of Tel-Mahré, Chronicle, A. Palmer, The Seventh Century in Episthmonikw´n Ereunw´n 21 (1995), 1–13. An argument against this thesis, the West—Syrian Chronicles (Liverpool: Liverpool University Press, 1993), based on historical sources, was provided in Christides, Image of Cyprus, 39. 174–75; J.-B. Chabot, Chronique de Michel le Syrien II (Paris, 1899–1910; rpt. 127. These numbers are based on the legible coin finds, coming from several Brussels 1963), 442. sites, such as the Campanopetra church, the Temple of Zeus (which was 114. This initial treaty is recorded by Al-Baladhuri (d. ca. 892) and converted into a church by the seventh century), and a seventh-century Abu’Ubayd al-Qasim b. Sallam (770–838). The historian Abi J’afar Muham- house called “L’huilerie”; O. Callot, Les Monnaies: fouilles de la ville 1964– mad ibn Jarir al-Tabari (838–923) recorded an additional condition: “And 1974 (Paris: De Boccard, 2004), 99–104. Thirteen of these Umayyad coins in the treaty between them [the Cypriots] and him [Mu’a¯wiya], was that they dated before 680. Of the Byzantine coins, 194 were from Constans II should marry women of our enemies only after our authorization”; M. Man- (641–668); seven from Justinian II (685–695); five from Leo III (695–698); souri, Chypre dans les sources arabes médiévales (Nicosia: Centre de Recherche eleven from Tiberius (698–705); two from Constantine V (741–775). No

188 jsah / 69:2, june 2010

JSAH6902_02.indd 188 4/16/10 4:15 PM coins from subsequent Byzantine emperors were discovered, except for one from mid-seventh to the tenth century, found throughout the island; dating from late tenth century, Basil II (976–1025). Therefore, it is rather “Islamic and Byzantine Coins,” 1–13; Pitsillides and Metcalf, “Some more curious that Pitsillides and Metcalf make the statement that “there were no finds of Islamic and Byzantine coins from the condominium centuries,” copper coins from after 691 from the site [Salamis-Constantia]: none was Epithrei´V tou Ke´ntrou Episthmonikw´n Ereunw´n 23 (1997), 1–7. identified”; “Islamic and Byzantine Coins,” 10. 129. Ousterhout, Master Builders of Byzantium, 86. 128. In addition to the coin finds at Salamis-Constantia, Pitsilides and 130. Huneberc of Heidenheim, Vita seu Hodoeporicon S. Willibaldi, Itinera Metcalf record thirty-three Islamic and fourteen Byzantine coins dating Hierosolymitana in Acta Sanctorum, July, II (1721): 505a.

T h e F i r s t Va u lt e d C h u r c h es in Cyprus 189

JSAH6902_02.indd 189 4/16/10 4:15 PM